APPLICATION OF STRESS-WAVE THEORY TO PILES

This Page Intentionally Left Blank

PROCEEDINGS OF THE SIXTH INTERNATIONAL CONFERENCE ON THE APPLICATION
OF STRESS-WAVE THEORY TO PILES/SAO PAULO/BRAZIL/ 11 - 13 SEPTEMBER 2000

ss
Edited by
Sussumu Niyama
Institute for Technological Research - IP2: SLio Paulo, Brazil
Jorge Beim
PDI Engenharia, Rio de Janeiro, Brazil

A.A. BALKEMA/ROTTERDAM/BROOKFl[ELD/2000
This conference was organized by the Brazilian Society for Soil Mechanics and Geotechnical
Engineering

The texts of the various papers in this volume were set individually by typists under the supervision of each of the
authors concerned.

Authorization to photocopy items for internal or personal use, or the internal or personal use of specific clients, is
granted by A.A. Balkema, Rotterdam, provided that the base fee of per copy, plus per page is
paid directly to Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923, USA. For those
organizations that have been granted a photocopy license by CCC, a separate system of payment has been arranged.
The fee code for users of the Transactional Reporting Service is: 90 5809 150 3/00

Published by
A.A. Balkema, PO. Box 1675,3000 BR Rotterdam, Netherlands
Fax: +3 1.10.413.5947; E-mail: balkema@balkema.nl; Internet site: www.balkema.nl
A.A. Balkema Publishers, Old Post Road, Brookfield, VT 05036-9704, USA
Fax: 802.276.3837; E-mail: info@ashgate.com
ISBN 90 5809 150 3
02000 A.A. Balkema, Rotterdam
Printed in the Netherlands
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Table of contents

Preface XIJI
Organization xv
Acknowledgements XVII

1 Wave mechanics and its application to pile analysis

Keynote lecture: Some wave mechanics applications 3
G.G.Goble
Analysis of bearing capacity of rock-socketed piles based on wave equation theory 11
L.B. Cai
Pile acceptance based on combined CAPWAP analyses 17
R. F: Stevens
Theoretical study on effect of pile shaft resistance on rebound during pile driving 29
R.P Chen & Y M. Chen
Time effect in determining pile capacity by dynamic methods 35
M. R. Svinkin
Set-up considerations in wave equation analysis of pile driving 41
CWCho, M.WLee & M.F:Randolph
Drivability and performance of model piles driven into cemented calcareous sand 47
D. Bruno, M. F: Randolph, C WCho & H A Joer
Automatic signal matching with CAPWAP 53
E Rausche, L. Liang & B. Robinson
Combining static pile design and dynamic installation analysis in GRLWEAP 59
E Rausche, B. Robinson & J. Seidel
Rules of thumb for field and construction engineer in relation to impact pile driving 65
W J Lucieer

V
2 Driving equipment and recent developments -New technologies for quality
assurance of piles
Keynote lecture: Pile driving equipment: Capabilities and properties 75
F: Rausche
Hammer system design using wave equation analysis for testing cast-in-situ concrete piles 91
M A Mukaddarn, WM.Iskandarani & M. Hussein
Analysis of pile load transfer using optical fiber sensor 99
J.-H.Oh, W-J.Lee, S-B.Lee & W-J.Lee
Set-up effect of cohesive soils in pile capacity 107
M.R.Svinkin & R.Skov
Modulus of elasticity and stiffness of composite hammer cushions 113
M. R. Svinkin
A new technique to drive piles: Down-the-holepiling 117
A. Benarnar
Simplified neural network models for estimating soil resistance using dynamic pile test 121
R. Liang & N 0.Nawari
Pile capacity prediction using neural networks technique 127
A. S Dyrninsky, E. Parente-Ribeiro, C.Rornanel & J. W Beim
Hammer and pile cushion optimalisation 135
G.Jonker & R.J.van Foeken

3 Pile integrity and low strain dynamic testing

Keynote lecture: 0resundlink, grouting work under West Pylon: Integrity test of a giant pile 145
J. Rome11
A study for the shape of pile with reflect-wave method 153
Yang Wu
Evaluation of the performance of an existing foundation via PIT tests 157
S B. Foa, J. H. l? Pereira, R. P Cunha & J. Carnapum de Carvalho
Sonic integrity test of piles-integrity effected by basement excavation in Bangkok soft clay 163
NThasnanipan, A.WMaung, PTanseng & Z.ZAye
Non-destructive integrity testing on piles founded in Bangkok subsoil 171
N. Thasnanipan, A. WMaung, T.Navaneethan & Z.Z.Aye
Estimation of pile head stiffness using sonic integrity testing 179
K. Irnada, T. Matsurnoto & Y:Nakata
Integrity testing of cast-in-situ concrete piles associated with the construction I87
of New Haccho Bridge
X Michi, T.Matsurnoto & K Matsuda
Presentation of low strain integrity testing in the time-frequency domain 193
J. R Seidel

VI
First experiences in the application of the stress wave theory to foundations in Uruguay 20 1
A.Gutikrrez, L.Abreu, Ch.H o f i a n n & D. Hasard
Detection and prevention of anomalies for augercast piling 205
G. Likins, G. Piscsalko, E;: Rausche & C.M. Morgano
Recent advances and proper use of PDI low strain pile integrity testing 21 1
G. Likins & E;: Rausche
Stress wave propagation velocity at early ages 219
C Restrepo
Examination of a new cross-hole sonic logging system for integrity testing of drilled shafts 223
S.G. Paikowsky, L. R. Chernauskas,L.J.Hart, C D. Ealy & A. I;:DiMillio

4 Pile-soil interaction
Keynote lecture: Identification of soil-pile model interaction parameters from recorded 233
time-displacement signals
A. I. Husein Malkawi & I. M.Ayasrah
Load transfer analysis from increasing energy dynamic load tests in concrete piles driven 241
in very soft clay formation
J.Balech & NAoki
Dynamic pile testing and finite element calculations for the bearing capacity of a quay wall 249
foundation - Container terminal Altenwerder, Port of Hamburg
F: Kirsch, B. PlaPrnann, ir:Huch & W Rodatz
The shaft dynamic response of a pile in clay: Induced pore pressure 255
A. Benamar
An investigationof pile diameter influence in the bearing capacity on Dynamic Load Test 26 1
(DLT)
G I? Bernardes, C S Andreo & C Gongalves
Pile set-up in sands 267
J. I? Seidel & M. Kalinowski
Modeling pore pressure generation during dynamic testing of deep foundations 275
l? L. Pinto

5 Codes
Keynote lecture: The performance of the dynamic methods, their controllingparameters 28 1
and deep foundation specifications
S G. Paikowsky & K. L. Stenersen
LRFD design codes for pile foundations - A review 305
G.G.Goble
Pile testing competitions - A critical review 313
J. M.Arnir & B. H. Fellenius

VI I
The need for quality assurance in the dynamic pile testing industry 319
J. 2? Seidel
High strain dynamic pile testing, equipment and practice 327
G.Likins, E Rausche & G.G.Goble
Dynamic load testing and Statnamic load testing for acceptance and design of driven piles 335
in Japan
Matsumoto, K Fujita, 0.Kusakabe, M. Okahara,N Kawubata & S. Nishimura
Ethics and money. Are they compatible? 345
H. Goldemberg & J.J. Goldemberg

6 High strain dynamic testing of driven and cast in situ piles - Dynamic testing
of large piles
Keynote lecture: Pile acceptance criteria for large diameter and cast in situ piles 35 1
R. E Stevens
Analyzing the bearing capacity mechanism of large diameter diving casing cast-in-situ 369
concrete piles by using high strain dynamic testing
Xi Liang
Analysis of dynamic load tests on steel rails piles 375
EM.A. Lima, J.C.A.Cintra & NAoki
A discussion of penetration matching on high strain dynamic pile testing 383
D.Xu, S. Wu & L.Xiao
Experience gained and difficultiesin performing dynamic load test in composite piles made 389
with steel rails
G.P Bernardes, C.S.Andreo & C.GonCalves
The application of high strain dynamic pile testing to screwed steel piles 393
J.G.Cannon
Case study on the application of high strain dynamic pile testing to non-uniform bored piles 399
J.G.Cannon
High capacity dynamic load tests for bored piles in Sydney shale 403
D.J. Klingberg & fl Mackenzie
Predicting uplift deflection from dynamic pile testing 407
WG.Chambers & D.J.Klingberg
Applicability of dynamic load test on a toe improved pile 41 1
J. Sakimoto,N Kitu, S. Nishimura & TTakeda
Dynamic pile testing practice in Finland 415
H,Jokiniemi, J. Hartikainen & I? Korkeakoski
Driving behavior of large diameter steel pipe piles 42 1
1-A?Lee & J. -SLee

Vlll
Dynamic load test of cast in place pile using a free fall hammer 429
S. Niyama, G. Cde Campos, S. Navajas, S. C.Paraiso, C M. C.Costa & G.E. Barbosa
Dynamic testing of large auger pile using free fall loading system in a harbour work 435
S. Niyama, S.Navajas & G.C.de Campos
Dynamic load test on high capacity pile socketed in basaltic rock 441
S. C Paraiso, C M. C.Costa & E. Pinto Soares

7 SPT measurements and special field monitoring test

Keynote lecture: Frequency characteristics of stress wave and penetration during SF'T 45 1
K. Fujita
The application of energy conservationHamilton's principle to the determination of energy 457
efficiency in SPT tests
N.Aoki & J. CA.Cintra
Correlative study of Smith damping coefficient and SPT blow count 461
R.1 Liang
Stress wave theory application to standard penetration test in Japan 469
K. Fujita & M. Ohno

8 Vibratorypile driving - Vibration in pile driving

Keynote lecture: Vibratory driving analysis 479
A. E. Holeyman
Computation of ground waves due to piling 495
C.L. Ramshaw,A. R. Selby & l? Bettess
The effect of pile impedance on energy transfer to pile and ground vibrations 503
M. R. Svinkin,B. C.Roth & IT?R. Hannen
Determination of modulus of subgrade reaction in a pile with a vibrating apparatus 51 1
M. Hilmi Acar
Analysis of crosswise vibration of pile driving 517
R.l?Chen, B.Zhu & YM.Chen
Monitoring and control of dynamic effects of pile installation prior to pile driving 525
M. R. Svinkin
Full-scale field-test study of dynamic soil resistance of vibratory driven sheet piles 533
K. Viking
Accelerations of a driven pile and the surrounding soil 541
E.L. Hajduk, S.G. Paikowsb, I! Holscher & l? B.J. Barends

9 Statnamic and other similar techniques

Keynote lecture: Statnamic, the engineering of art 55 1
R Middendorp
IX
Keynote lecture: Three-dimensional finite element analysis of statnamic load test 563
I:Boonyatee~M. Kimura & F: Zhung
Lateral statnamic load testing of model piles 569
M.~imura& ~ ~ o o n y a t e e
Application of the Stress Wave method to a~tomaticsignal matching and to statnamic 575
predictions
G. Esposito, WMG.Courage & R.J.van Foeken
A comparative study of static, dynamic and statnamic load tests of steel pipe piles 583
driven in sand
A, ~hibata?N Kawa~ata,~ ~ a k i yY:aYoshizawa,
, M.~ u ~ a s&h 1:i Ma~s~moto
Case studies of statnamic load testing in Japan 591
S. Nishimura, 1:Matsumoto, 0.Kusakabe, K Nishiumi &: YYoshizawu
Statnamic and dynamic load tests for large diameter steel pipe piles supported by a thin 599
bearing layer at Nagoya port in Japan
Y:~ikuchi,S Nishimura & M. ~atsuta
Statnamic load testing using water as reaction mass 609
M. D.~ustason?~ * ~ ~ a I!n~ei ds d, e n d ~ &
r pAG. ~ u l l i n s
Introducing statnamic load testing in Europe: Case studies in the Netherlands 617
G.JJ,van Ginneken & I! Middendorp
The advantages and disadvantages of dynamic load testing and statnamic load testing 625
P Middendorp, G.J.J. van Ginneken & R.1 van Foeken

10 Case histo~ies, pile set-^^ and c ~ ~ ~ e l a t i~o n se ~teste ethe hods

~
- P ~ e ~ i c ~~i oen~ i a ~ i ~ i ~

Keynote lecture: Improving the reliability of pile bearing capacity prediction by the dynamic 635
increasing energy test (DIET)
~ ~ o k i
Correlation analyses of dynamic and static loading tests for nine piles 65 1
Y:M. Zheng, J , M. Zheng & B. Chen
~ a c k - ~ ~ yofs steel
e s pile driving records for quality assurance 657
B. R. Danziger & J. S. Ferreiru
Evaluation of pile set-up from penetration per blow 665
G.Axelsson & S Hintze
Comparative analysis of dynamic and static test of foundation pile 673
GZhou &LWu
Dynamic load test and elastic rebound analysis for estimation of the bearing capacity of piles 677
in residual soil
RL R.de Albuquerque & D.de Carvalho
Strain dynamic testing on pressure-grouted piles 683
Liu Xi-An & Zhang Yao-Nian
X
Assessment of the interface between dynamic and rapid loading tests 689
M. B. Karkee, X Sugimura & 7:Horiguchi
Dynamic load testing on 102 steel pipe piles for bridge foundations on mudstone 697
M. Hayashi, 7:Matsumoto& M. Suzuki
Behavior of short CFA piles in an overconsolidated clay based on static and dynamic 707
load tests
A. C.M. Kormann, R R. Chamecki,L. Russo Neto, L.Antoniutti Net0 & G.19Bernardes
Static and dynamic testing of the ‘Campile’ - A displacement, cast-in-situ pile 715
D.J.Klingberg & r! Mackenzie
Is DLT the final word? Correlation between DLT and SLT 719
H.Goldemberg & J.J.Goldemberg
Results of an international pile dynamic testing prediction event 725
A. Holeyman, J. Maertens, N. Huybrechts & C Legrand
Preparation of an international pile dynamic testing prediction event 733
A Holeyman, J. Maertens, h? Huybrechts & C Legrand
Case studies of dynamic load testing in Japan 741
X Wakiya,K. Nishiumi, M. Hayashi,A. Shibata, S.Nishimura & TMatsumoto
Case studies of high capacity CFA pile testing in Australia 75 1
S. Baycan

11 Supplement
Keynote lecture: Retrospective of Sonic Integrety Tests - Its application to the control quality 757
on piles
J. J. Goldemberg

Author index 77 1

XI
This Page Intentionally Left Blank
 Applicationof Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Preface

The application of dynamic testing based on the wave equation theory was first introduced in Brazil
in 1981. Differently from other countries, its first application was on offshore piling. At that time, a
great number of jacket type platforms began to be installed in the country by the Brazilian Oil
Company, in waters with depths varying from 50 to 250 meters. The use of this technique also on
land job sites experimented a continuous growth since then. In 1989, low strain integrity testing was
introduced. In 1994, a code on the dynamic load testing of piles was issued by the Brazilian
Technical Standards Association - ABNT A couple of years later, in 1996, the same ABNT edited a
revision of the Standard for Design and Execution of Foundations, which now accepts the use of
dynamic load testing as one of the alternatives for bearing capacity evaluation. In the case of driven
piles, it also allows for a reduction of the safety factor, from the usual value of 2 down to 1.6,
provided that dynamic testing is performed on at least 3% of the piles on the job site, and that it is
applied since the beginning of the piling.
The present event in the Southern Hemisphere, and particularly in South America, certainly will
contribute to increase the utilization of this technique, thus helping improve the quality of pile
installation in this region.
The support of the Brazilian Society for Soil Mechanics and Geotechnical Engineering - ABMS
in hosting this conference had the purpose of trying to make the geotechnical and foundation
engineering communities and the wave equation theory users work more closely together. The
conference structure was modified, eliminating the traditional presentation of the papers, thus
allowing more time for discussions. After five conferences, the first one organized by Dr
H. Bredenberg of the Swedish Geotechnical Institute in 1984, we hope that discussions based on the
papers and lectures presented in this proceedings might have contributed to improve the understand-
ing and bettern interpretation of dynamic tests, not only for the executioners but specially for the
users of the the tests.

Sussumu Niyama
Jorge Beim
Editors

Xlll
SPONSORSHIP
Brazilian Society for Soil Mechanics and Geotechnical Engineering

SUPPORT
FINEP

XIV
 Applicationof Stress-WaveTheory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Organization
ORGANIZING COMMITTEE
Niyama, Sussumu Chairman
Beim, Jorge Secretary
Andreo, Cristiana
Aoki, Nelson
Bernardes, George
Campos, Gisleine Coelho
Cintra, Jos6 Carlos
Danziger, Bernadete
Falconi, Frederico E
CONFERENCE ADVISORY COMMITTEE
Lopes, Francisco Rezende
Velloso, Dirceu de A.
D6court, Luciano
Golombek, Sigmundo
Mello, Victor E B. de
INTERNATIONAL,ADVISORY COMMITTEE
Altaee, Ameir - Canada
Amir, Joram M. - Israel
Burbano, Germiin J. - Spain
Fellenius, Bengt H. - Canada
Fujita, Keiichi -Japan
Goble, George G. - USA
Goldemberg, Juan J. - Argentina
Gravare, Carl-John - Sweden
Gutierrez, Alvaro - Uruguay
Holeyman, Alain E. - USA
Holloway, D. Michael - USA
Hussein, Mohamad H. - USA
Kusakabe, Osamu - Japan
Liang, Xu Ding - China
Gonplves, Cliiudio
Mello, Jayme Ricardo de
Kormann, Alessander Morales
Merighi, Valmir
Navajas, S6rgio
Paraiso, S6rgio C.
Valverde, Skrgio
Likins, Garland - USA
Matsumoto, Tatsunori -Japan
Mendiguren, Eugenio - Argentina
Middendorp, Peter - Netherlands
Paikowsky, Samuel G. - USA
Pinto, Paul0 - Portugal
Randolph, Mark - Australia
Rausche, Frank - USA
Sakai, Tomoaki - Japan
Seidel, Julian F! - Australia
Towsend, Frank C. - USA
Van Impe, William - Belgium
Van Weele, A. E - Netherlands
Veiga, Jose - Chile
xv
This Page Intentionally Left Blank
 Application of Sfress-Wave Theory to Piles, Niyama L? Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Acknowledgements

The editors are grateful to the following persons who helped to review the manuscripts and thus
greatly assisted in improving the overall technical standard and presentation of the papers in these
proceedings:

Kormann, Alessander Morales

Danziger, Bernadete
Gravare, Carl-John
Velloso, Dirceu de A.
Lopes, Francisco Rezende
Goldemberg, Hermh
Amir, Joram M.
Seidel, Julian I?
Svmkin, Mark R.
Randolph, Mark
Kusakabe, Osamu
Middendorp, Peter
Goble, George G.
Holeyman, Alain E.
Paikowsky, Samuel G.
Paraiso, Skrgio C.
Matsumoto, Tatsunori
L W s , Garland

XVll
This Page Intentionally Left Blank
1 Wave mechanics and its application to pile analysis
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Keynote lecture: Some wave mechanics applications

George G.Goble
George G.Goble Consulting Engineer, University of Colorado, Boulder, Colo., USA

ABSTRACT: The solution to the one dimensional wave equation is used to determine the maximum impact
force at the top of a concrete pile for a given ram weight, cushion stiffness, and pile impedance. The result is
used to generate a set of curves that can be used to assist in the selection of a pile cushion as limited by maxi-
mum compression stress. A second problem, the wave equation modeling of the Statnamic load testing sys-
tem, was also studied. The testing process was modeled using a wave equation computer program. A com-
parison of the commonly used damping model, Smith damping, with the Rausche model can be used to
evaluate the appropriateness of the two models. This study indicates the desirability of the Rausche model and
the necessity for fbrther study to recommend appropriate values for the Rausche damping constant.

1 INTRODUCTION plays now generally used for presenting dynamic

The interest in one dimensional wave propagation
theory for understanding pile driving is more than a
century old. Perhaps, the pile driving problem was a
major motivation in deriving and solving the one di-
mensional wave equation since those people that
were active in that effort were mostly Civil Engi-
neers. The important work done in this area that re-
lates directly to current applications to pile driving
started shortly aRer the Second World War and con-
tinued to the present.
measurements, for many of the usefbl concepts for
visually evaluating those measurements, and for cur-
rent approaches to integrity testing of deep founda-
tions. Since the basic work is complete and many di-
rect applications of the theory well established it is
unlikely that fbrther hndamental developments will
occur. It is appropriate to note that, while the use of
the closed form solution has been very important in
developing conceptual understanding of the mechan-
ics of pile driving, it has been less usefil for obtaining
quantitative results.

Today it can be said, with confidence, that the use of
one dimensional wave propagation theory is gener-
ally accepted and widely used in the pile driving in-
dustry. The hallmark of these applications is that
they have centered on usefil applications of the the-
ory. In this paper, two topics will be discussed, one
dealing with the closed form solution of the wave
equation and the other, the discrete solution. Useh1
results will be presented in each case.
The one dimensional wave equation was derived and
solved in the nineteenth century. Probably the most
important contributions to the use of the closed form
solution were made by Professor Fischer of Upsula
University beginning in the early 1950’s. He devel-
oped many useh1 applications of the solution in-
cluding several graphical approaches. A summary of
this work was presented at the Second Conference
on the Application of Stress Wave Theory to Piles
(Fischer 1984). This work laid the basis for the dis-
Probably the most remarkable work in developing
quantitative solutions was done by E. A. L. Smith,
then Chief Engineer of the Raymond Company, at
that time, the largest pile driving company in the
world. Before 1950, he published notes on the de-
velopment of a numerical solution to the wave
propagation problem applied to pile driving (Smith,
1950). Shortly aRer, a proprietary program was op-
erating on an electronic digital computer. Smith
called this program the “Wave Equation” and it was
widely publicized (for example; Smith, 1957; Smith,
1960). This program may have been the first appli-
cation of electronic digital computers to a civilian
engineering problem.
The implementation of wave equation analysis has
been painstakingly slow. A public domain program
was developed in the early 1960’s at Texas A&M
University and it was widely used in the offshore in-
dustry (Samson, et a1 1963). Today, the

3
GRLWEAPTMprogram (Goble & Rausche 1976,
GRLWEAPTM,1998) is widely used in the United
States and that usage is increasing rapidly. It is be-
coming standard practice to perform driveability
analyses both during the foundation design stage and
at the beginning of construction. However, outside
the United States usage is much less common.
2 PILE CUSHION SELECTION
A direct application 'of a closed form solution to the
problem of selection of a pile cushion for driving a
concrete pile will be presented. This solution can be
used with the aid of graphs and it may be found to be
usefiil in practice. Clough and Penzien (1975) solved
the problem shown in Figure 1. The pile driving
system is modeled by three elements, a rigid ram, a
linear spring (cushion) and an elastic pile. The solu-
tion provides the maximum delivered impact velocity
induced at the pile top during the first wave passage
(among other things). This solution is used to gener-
ate curves for selection of the cushion stiffness re-
quired to limit the induced compression stress to
some selected value.
i3x c dt
where A is the cross sectional area of the pile, E is its
elastic modulus, c is the velocity of wave propaga-
tion, o is the stress at the top of the pile, e is the
strain at the top of the pile, x defines the distance
from the top of the pile to some point on the pile, and
t is time. This expression is substituted into the ex-
pression for the driving force, F, in Equation (1). It
is then written in terms of the displacement of the
pile top and the resulting expression is rearranged
and expressed in terms of the pile top velocity ,v.
k k
v+-++-v =0 (4)
Z m
In Equation (4), the weight of the ram has been ne-
glected and Z is the pile impedance, ENC. This fa-
miliar equation form has the solution
v= (A sin o t + B coso t) (5)
The variables k/m and k/Z in Equation (4) have been
replaced by a2and 205, respectively, and

If the initial conditions of the pile top velocity,v, and

the ram impact velocity, vh, are imposed, the expres-
sion for the velocity at the top of the pile becomes

Figure 1. Pile Driving System Model

The equation of motion for the ram in contact with

the cushion can be written

F=mg-muh (1)

where the variables are defined in Fig. 1. Continuity

of displacements requires that the ram motion equal
the pile top motion plus the spring deformation.

This expression is differentiated twice and substituted

into Equation (1). The resulting expression is solved
for F and then expressed in terms of the stress in the Figure 2. Values v/vh for pile stiffness, ram mass, with
pile as given by one dimensional wave mechanics. given impedance

4
Maximum values of v/vh were determined for a range
of values of ram mass, cushion stiffness, and pile im-
pedance.
A parameter study was made for a 300 mm square
pile with a full range of practical values of ram mass,
cushion thickness, and concrete modulus. Values of
the ratio of the pile top velocity to the ram impact
velocity, dvh, were determined for a range of values
of ram mass and cushion stiffness, and four values of
pile impedance. An.example of the results is shown
in Figure 2 where the surface of v/vh is given for a
range of values of k and m with a specific value of
pile impedance, 2. Such a representation cannot be
5
used quantitatively so contours for four values of Z
are given in Figures 3 through 6.
In order to use the curves, trial cushion stiffness and
ram mass is selected, and they are divided by the pile
area. Likewise, the pile impedance divided by the
pile area is also determined. Using the appropriate
figure, the pile velocity-ram velocity ratio is deter-
mined from the curve. This quantity is multiplied by
the anticipated ram impact velocity to obtain the
maximum particle velocity transmitted to the pile top.
The particle velocity times the pile impedance gives
the maximum impact force transmitted to the pile.
 Figure 7. Wave equation model

It should be noted that the model does not include
the hammer cushion or the helmet mass. Further-
more, the result has not been tested against field ex-
perience and such testing must be done to assure the
usehlness of the results.
There can be little hope of ever accurately knowing
the actual hammer efficiency prior to going to the
field, Even beginning driving the hammer efi-
ciency will only be known when measurements are
available.

The actual quake that exists during pile installation is

3 WAVE EQUATION DAMPING MODEL
dependent on the size of the pile cross section and on
~01.1properties. Quake values can be assigned based
Wave equation analysis is now used routinely in land
on the pile cross section but when a large quake is
pile driving practice in the United States and also in
due to soil properties it cannot be predicted in the
controlling the installation of piles for offshore pe-
current state-of-the-art. Fortunately, this character-
troleum recovery platforms. The model used for
istic occurs infrequently. Experience with dynamic
wave equation analysis is shown in Figure 7. The
measurements and signal matching analysis indicates
three principal problems that limit the accuracy of
that this problem usually occurs during extended
wave equation analysis are a lack of knowledge of
driving and is probably pore pressure dependent. It
the driving efficiency for a particular pile driving
disappears in re-strike testing. A solution to this
hammer, an occasional, surprisingly large quake
problem will depend on a better understanding of soil
value, q, and a lack of accuracy in the soil damping
behavior.
constant, j.

6
Damping constants have been selected based on tra-
dition using the original recommendations of Smith
(1960). They are selected based on soil type and
they show a wide range of accuracy based on signal
matching analyses (Rausche et al 1994). This vari-
ability can strongly affect the results of wave equa-
tion analysis. Rausche et a1 (1994) suggested a dif-
ferent model that is based on the work of Coyle and
Gibson (1970). Coyle’s research, fbrther proven by
Herema (1979), indicates that the dynamic resistance
has a strongly nonlinear relationship to the velocity of
pile motion. The study by Rausche et a1 indicated
that the direct application of the Coyle research
caused numerical problems. A suggestion by
Rausche avoided those problems.
The traditionally used Smith model states
R, = R,(1+ j,v)
where Rt is the total soil resistance, R, is the current
value of the static resistance, and j, is the Smith
damping constant. This model becomes viscous
when the static resistance is equal the ultimate resis-
tance. Based on laboratory testing Coyle showed
that a more appropriate law would be
R , = R,(1+ j g v N ) (9)
where j, is the Gibson damping constant and N is an
exponent, typically less than 1.0. The difficulty with
this representation is that numerical problems arise
when the velocity changes sign. Rausche suggested
that damping be represented by
v clay, and (d) Rausche damping for clay. The results
R, = R,[1+ j R v r --
v, R,
where j, is the Rausche damping constant, vx is the
maximum velocity achieved up to a particular time
during the blow, and R, is the maximum static resis-
tance actuated prior to the time under consideration.
Both the Smith and the Rausche model are available
in GRLWEAPm.
The Statnamic method (Janes et a1 1994) uses a
charge of slow burning explosive to generate a force
between a large mass and the pile top. As the mass is
accelerated upward a downward acting force is in-
duced against the pile top. This force has a slower
rate of increase than the typical ram impact used in
dynamic testing. Force pulses of the length of 80-
100 ms are common. The resulting test must still be
7
evaluated as dynamic but it is slower than the tradi-
tional dynamic test.
The Statnamic test was modeled using a mass of the
typically used size as a “ram” in GRLWEAPTM. It
was dropped on the top of the pile with a cushion
having a very soft spring. The spring constant was
selected by trial-and-error so that a force pulse of
about 100 ms was generated. The magnitude of the
peak force was controlled by the ram drop height.
The test pile used in the study was a closed-end steel
pipe with a length of 24 meters and an ultimate ca-
pacity of 4000 kN. An example of the force-time
and velocity-time record for one of the tests is shown
in Figure 8.
Figure 8. Statnamic Test - Force and Displacement vs.
Time
Four cases were studied, each with a range of applied
forces. They are (a) Smith damping for sand, (b)
Rausche damping for sand, (c) Smith damping for
are given in Figures 9 through 12.
In Figure 9 the force-displacement results for the
case of the pipe pile in sand is shown. The wave
equation results include the top force in the pile as a
hnction of time together with the pile top displace-
ment. The maximum applied force was 5200 kN. In
this case a maximum displacement of 55 mm was
achieved. The method that has been recommended
for use in determining the static capacity was to use
the force at the time of zero velocity and adjust this
force by the mass of the pile times the acceleration at
that same time. This method was first suggested by
Nara (1970) and was used by the Case Research
project (Goble & Rausche, 1970). It can be seen in
Figure 9 that this method gives quite good results
when compared with the specified static capacity of
4000 Kn. The applied forces were 5200 kN, 4840
kn, and 4180 kN. In all three cases the capacity at
maximum displacement was about 4000 kN. The
adjustment for the acceleration at zero velocity (iner-
tia force) would be small due to the low mass of the
pile.
Rausche damping was applied to the same sand ex-
ample and the results are shown in Figure 10. Forces
of 6400 kN, 5960 kN, and 5060 kN were applied.
The results for the two larger forces, given by the
capacity at zero velocity, are reasonably close to the
static capacity. However, the case of the smallest
applied force gives a predicted static capacity of
about 4800 kN, about 20 percent larger than the
given capacity.
Examples of the modeling of Statnamic tests in clay
for the same pile that was analyzed above are given
in Figures 11 and 12. Figure 11 shows the results for
three Statnamic tests of different applied force levels
8
in a clay soil. The usual Smith damping constant
result is given in Figure 11. The predicted static ca-
pacity for all three Statnamic load levels are all quite
close to the known static capacity.
Figure 12 shows the result for Rausche damping in
clay for applied forces of 8000 kN, 7150 kN, 6070
kN, and 5450 kN. The associated predicted static
capacities are 6400 kN, 6400 kN, 5830 kN, and 5240
kN, respectively. All of these values are quite high
with the largest 60 percent too large.
This study showed that it was possible to induce
Statnamic-like forces using a wave equation analysis
by GRLWEAPTMwithout modification for this par-
ticular application. The force time record is of the
appropriate length and a similar shape. The force-
displacement curve has the appropriate shape and
appearance.
It is well-known that the Smith damping representa-
tion does not produce results that agree with experi-
ence. (Rausche et a1 1994) has shown that the Smith
damping constant does not agree with measurements
for the assumptions used. Some results reported for
Statnamic (James et a1 1994) show a character of re-
sult that agree with the wave equation analysis. In
particular, the case presented with Rausche damping
for sand has the observed result. As the applied
Statnamic force is increased the agreement between
the evaluation method used and the actual capacity is
improved. The conclusion has been reached that a
substantial permanent set should be achieved.
The results presented here support the idea that the
Rausche damping constant should be studied hrther
and that recommendations should be developed for
the required damping constants.
CONCLUSIONS
Curves are presented to assist in obtaining cushion
stiffness requirements for concrete piles to limit the
induced compression stress at the top of the pile.
Impact stresses can be determined for a selected set
of ram mass, cushion stiffness, and pile impedance.
These values have been found for reasonable ranges
of the variables using a closed form solution of the
one dimensional wave equation. They have not been
checked against field measurements and such checks
should be performed. It should also be noted that the
analysis does not include the pile cushion and the
helmet.
A wave equation study is reported on the modeling
of the Statnamic test. The model matches the varia-
tion of the induced force at the pile top quite well.
Smith and Rausche damping representations have
been studied. The Rausche damping representation
seems to match the field observed results. In view of
the observed poor comparison between the Smith
representation and field observations it would be de-
sirable to determine appropriate Rausche damping
constants by additional studies.
REFERENCES
Clough & Penzien 1975. Dyiianiics of Structures. McGraw-
Hill, Inc., New York.
Coyle, H. M. & G. C. Gibson 1970. Empirical damping con-
stants for sands and clays. Journal of Soil Mechanics a i d
Foundations Division, ASCE.
9
Fischer, H. C. 1984. Stress wave theory for pile driving appli-
cation. lecture at the Second International Conference 011
the Application of Stress Wave Theory on Piles, Stock-
holm, Sweden.
Goble, G. & F. Rausche 1970. Pile load test by impact driving.
Highway Research Record, No. 333.
Goble, G. G. & F. Rausche, 1976 Wave equation analysis of
pile driving - WEAP program. Volumes 1 through 4,
FHWA #IP-76-14.1 Through #IP-76-14.4.
GRLWEAPTM- Wave equation analysis of pile driving 1998.
Goble Rausche Likins and Associates, Inc., Cleveland,
Ohio.
James, M., A. Sy, & R. Campanella 1994. A comparison of
statnamic and static load tests on steel pipe piles in the
Fraser Delta. Proceeding of the Annual Svniposium on
Deep Foundations, Vancouver Geotechnical Society, Van-
couver, B.C., Canada.
Heerema, E. P. 1979. Relationships between wall friction dis-
placement velocity and horizontal stress in clay and in sand
for piled driveability analysis. Ground Engineering.
Rausche, F., G. Likins, and G. Goble 1994. A rational and us-
able wave equation soil model based on field test correla-
tion. Proceeding International Conference on Design and
Construction of Deep Foundations, Orlando, Florida, De-
cember.
Rausche, F., G. Thendean, H. Abou-matar, and G. Goble
1994. Investigation of dynamic and static pile behavior
from modified standard penetration tests, Federal Highway
Administration Report.
Samson, C. H., T. J. Hirschjr., & L. L. Lowery, 1963. Com-
puter study for dynamic behavior of piling. Journal ofthe
Structural Division, ASCE, Volume 89, No. ST4.
Smith, E. A. L. 1950. Fundamentals of electronic calculation -
pile driving impact. International Business Machines.
Smith E. A. L. 1957. What happens when liammer hits pile.
Engineering News-Record 159, September, 5.
Smith, E. A. L. 1960. Pile driving analysis with the wave
equation. Journal of the Soil Mechanics and Foundation
Engineering Divisioii, ASCE, No. 86, August.
This Page Intentionally Left Blank
 Applicationof Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Analysis of bearing capacity of rock-socketed piles based on wave equation

theory
Lai-bing Cai
Fujian Academy of Building Research, Fuzhou, People’s Republic of China

ABSTRACT: Based on wave equation theory, a dynamic loading model is used to analyze the behaviors of
rock-socketed piles in this paper. The results of dynamic loading tests are analyzed by CAPWAP. From the
analyzed results of twelve dynamic loading tests on rock-socketed piles, the behaviors of shaft resistance of
the socket and toe resistance are discussed, a relationship of shaft resistance of socket to the length of socket
and the rock strength and the construction of piles is presented. And the parameters of dynamic loading tests
are estimated. Some suggestions for dynamic loading tests of rock-socketed are made to analyze the bearing
capacity.

1 INTRODUCTION vanced in this paper. Through analyzed the results

of CAPWAP, the bearing behaviors of rock-
Smith (1 960) initially made numerical analysis on socketed pile are discussed and a method is given
one-dimensional wave equation of pile foundation to determine the bearing capacity.
with the finite difference scheme. From then on,
many scholars put up with their numerical models
and gave the calculating formulations all over the 2 DYNAMIC LOADING MODEL
world, so wave equation analyzing method was
widely applied to engineering practice. Basing on During wave equation analyzing, the total driving
the previous research results, Goble et al. (1980) resistance R can be broken up into two distinct
modified the model with better formulation, and portions: the static resistance R, and dynamic re-
put out the program of CAPWAP, which can better sistance R,
simulate the fact of pile-soil interaction. Thereby, R = R , +R, (2.1)
it was virtual stage for wave equation taken into The static resistance and the dynamic resistance
engineering practice. At present, wave equation are represented by
analyses are used to simulate the process of driv-
u<q
=;I
+U,
ing pile, judge the feasibility of driving pile and (2.2 )
the damage of pile, and determine the bearing ca- R, R,,
pacity of pile. R, =], * u = J * R* V, (2.3)
Rock-socketed pile foundations can provide an
effective and economical means of transmitting In above equations (2.2) and (2.3), k, is soil stiff-
large concentrated structural loads through over- ness, R,, is ultimate static resistance, q is loading
burden soils to underlying rock, especially under quake, J , is viscous damping factor, J is Smith
the condition of overburden soil being soft and damping factor.
rock embedding shallowly. The bearing behaviors When the pile exerts a force on the soil, it
of rock-socketed pile change with rock geological causes the soil surrounding the pile to move. By
condition, construction technology, sediments of analyzing energy formula, as the pile motions are
pile bottom, rock-socketed depth and so on. It small such that a shear failure of soil around pile
would be taken much manpower and many finan- does not occur, a wave is generated in the soil
cial or material resources to determine bearing ca- = R . U + E,,,
Eprle
pacity of rock-socketed pile by static loading tests. (2.4)
Based on the theory of wave equation, a dy- Where Ep,le is energy in pile, E,,, is energy of
namic loading model of rock-socketed pile is ad- wave generating in soil.

11
 A radiation damping dynamic model is built to
resolve the energy dissipating in pile-soil inter-
faces, where the energy is radiated rather than
consumed for soil shearing, shown as Figure 1.
Figure 1. Radiation damping dynamic model for rock-
socketed pile
For the above model, the soil support dashpots
work only during dynamic event, not during static
event. The governing equation (2.2) and (2.3) are
changed in that the pile motion variables U , v are
replaced by the relative variables U,, v,.
The motion variables us, v, of the soil support
mass are calculated simply by
ur,, =u,,,-1 +ut,l-l * A t
(2.5 1
U,,I =?'f,,-l +(R,*uc,,-1 .J,)/(J, +w/ A 4
Where A t: time increment, J,: shaft soil radiation
dashpot or toe soil radiation dashpot, Mr: shaft
soil support mass or toe soil support mass.
The relative variables ur, v, can be written as
u,,z = a , - a f ? I
(2.6 )
U,,, = U, - U ' ,I
3 BEARING BEHAVIORS OF ROCK-
SOCKETED PILE
The bearing capacity of rock-socketed pile can be
divided into three components: shaft resistance of
overburden soils Q s k , shaft resistance and toe re-
sistance of the socket, shown as
QILk = Q , k t-~<,faI,+<if,Ai (3.1 )
Where c: circumference of pile, <,, <,: are re-
spectively shaft resistance coefficient and toe re-
sistance coefficient.
The behaviors of shaft resistance and toe resis-
tance of socket are expounded as following.
3.1 Shaft resistance
The behaviors of shaft resistance of socket can be
influenced by the roughness of the socket wall.
12
However, the construction of pile influences the
roughness. Two cases are taken into account: case
(I), for hand-dug belled pile, explosion is used to
mine the rock, which results in the very roughness
of socket wall and relatively fractured rock under
pile bottom. The same outcomes arise for percus-
sion-drilled pile. Case (11), for bored pile, the
socket wall is relatively smooth and the rock un-
der pile bottom is relatively less shattered. Com-
pared to case (11), the shaft resistance raises and
the toe resistance falls somewhat in case (I).
According to the results of the static load
test^[^]"^^, the relations between the shaft resis-
tance and the relative displacement for the above
two cases are shown as Figure 2.
Figure 2. Relation between shaft resistance and relative dis-
placement
In case (I), the shaft resistance R, increases
with the relative displacement, and it shows a
harden trend. For dynamic analyses of rock-
socketed pile, the trend can be regarded as an
elastic-plastic relation, and the elastic quake may
value between 4mm and 6mm, and the ultimate
static resistance ranges (0.15-0.20) v;Y is un-
confined compressive strength of r ~ c k ) [ ~ ] _ [ ~ ] .
In case (11), the trend differs by the shaft resis-
tance decreasing after the mobilized displacement.
And the quake qs takes value between 2mm and
4mm, the value of ultimate static resistance is
about (0.15-0.20) [61,[71.
3.2 Toe resistance
Toe resistance raises slowly with the displacement
unless the concrete of pile comes to failure. It needs
a large displacement for toe rock to reach plastic
failure. The fractured rock makes the difference of
toe resistance: a more fractured rock produced more
reduction of toe resistance. Therefor, a high value of
toe resistance happens in case (I) against case (11).
The shaft resistance and the toe resistance should be
adjusted according to the construction of rock-
socketed pile.

Figure 3. Behaviors of hand-dug belled pile
4 DYNAMIC ANALYSES OF ROCK-
SOCKETED PILE
Based on the above model, the results of twelve
dynamic load tests on rock-socketed piles are dis-
cussed. The parameters of piles and the results of
CAPWAP are given in Tab. 1 and the CAPWAP
results of TPl are shown in Figure 4.
4.1 Analyses of bearing behaviors
Piles TP1 to TP4 were hand-dug belled piles, and
the toe rock was medium weathered mar1 breccia,
which belonged to soft rock. Note that the ratio
t,(=R,/Rt) of shaft resistance over toe resistance
has a value of 1/16-1/13. The results of CAP WAP
for pile TP1 shows that the shaft resistance of 1.8
time (D-d) (d, D are respectively shaft diameter
and belled bottom diameter) length over belled
bottom is lower than other portions of socket,
shown as Figure 3. It comes to conclusion that the
belled bottom reduces the shaft resistance above
it. This result reasonably conforms to the results
of static loading tests[61.Medium weathered gran-
ites were observed as the bearing strata of pile
TP5 to pile TP 11. The ratio t, achieves values of
1/24 to 1/20 for TP5-TP7 and 1/36-1/30 for
TP8-TP11 (see Table 2).
Pile TP12, 72.0m length, was socketed into
medium weathered granite about 2.0m. Length of
strongly weathered granite was up to 45.0m and
the corresponding mobilized shaft resistance was
94kPa, which shared about 80% total load. The
socket only bore 20% total load.
As stated, the bearing behaviors of rock-
socketed pile depend on the length of pile, the
construction, the rock type, and the roughness of
socket wall. A long pile (l/d>40) lowers the mo-
bilization of the socket to a degree. The socket
wall of hand-dug belled pile or percussion-drilled
pile is rougher than that of bored pile, which
causes high shaft resistance. And the shaft resis-
tance differs with the rock type. The behaviors of
rock-socketed piles with long length are coinci-
dent to the results of static loading
4.2 Analyses of dynamic parumeters
Tab. 2 presents the results of dynamic parameters
of radiation damping dynamic model for the twel-
ve tests. Final values of dynamic parameters are
observed as following: (0.4-1.5)EA/c for shaft
soil dashpot Jss; (1-20)EA/c for toe soil dashpot
Jst; the weight of the soil in a cylinder with same
length and diameter equal to 3 time pile diameter
for shaft soil support mass Ms; the weight of the
soil in a cylinder with same toe plate and depth
extending to 5 time socket diameter for toe soil
support mass Mt.

Table 1. Parameters of piles and dynamic analyzing results

Pile Diameter Length Lengthof Toebeanng Loadofpile Ratio Rocktype
No. d(m) Z(m) socket(m) Q t O Q(kN) tq(Qt/Q) ofsocket
TPI'') 1400 11.33 4.50 24700 34200 0.722 Marl breccia
900 10.12 0.80 9900 12600 0.786 Marl breccia
1400 10.10 1.30 27000 34600 0.780 Marl breccia
1400 10.83 1.OO 25000 31000 0.806 Marl breccia
1200 17.30 1S O 15200 25000 0.608 Granite
1 200 6.42 0.80 1 8800 26400 0.712 Granite
800 7.48 0.60 5100 7550 0.675 Granite
1200 8.08 1 .oo 19900 25500 0.780 Granite
1200 7.14 1 .OO 16000 22000 0.727 Granite
loo0 7.93 1 .oo 15600 21900 0.712 Granite
1000 8.35 1 .oo 13200 19500 0.677 Granite
1300 72.17 2.00 3000 24500 0.112 Granite

13
 (I) The piles were hand-dug belled piles, and the belled bottom diameters of TP1, TP2, TP3, and TP4
were respectively 2600 mm, 1900 mm, 3000 mm, and 2600 mm;
The piles were percussion-drilled piles;
( 3 ) The piles were bored piles;

Soil Depth Depth Ru Force Sum Unit Resist. Smith Quake

Sgmnt Below Below in Pile of Ru w. Respect t o Damping
No. Gages Grade at Ru Depth Area Factor
m m kN kN kN kN/m kN/m2 e/m mm
34200 .O
1 1.0 1.0 22.3 34177.7 22.3 21.68 4.93 ,520 2.000
2 2.1 2.1 33.4 34144.3 55.7 32.52 7.39 ,520 2.000
3 3.1 3.1 44.5 34099.8 100.2 43.36 9.86 ,520 2.000
4 4.1 4.1 55.7 34044.1 155.9 54.20 12.32 .520 2.000
5 5.1 5.1 152.1 33892.0 308.0 148.04 33.66 ,520 2.000
6 6.2 6.2 1082.9 32809.1 1390.9 1054.17 239.69 .520 2.000
7 7.2 7.2 1815.4 30993.8 3206.3 1767.16 401.81 .520 2.000
8 8.2 8.2 1819.5 29174.3 5025.7 1771.15 402.72 .520 2.000
9 9.2 9.2 1677.9 27496.4 6703.6 1633.31 371.38 .520 2.000
10 10.3 10.3 1677.9 25818.6 8381.4 1633.31 326.23 .520 2.000
71 11.3 11.3 1118.6 24700.0 9500.0 1088.88 153.53 .520 2.000

Average Skin Values 863.6 840.71 178.94 .520 2.000

Toe 24700.0 4652.48 ,441 2.150

Soil Model Parameters/Extensions Skin Toe

.374 , 8 2 6 - S m i t h Type
Case Damping Factor (% of loading quake) 27 100
Unloading Quake
Reloading Level (% of Ru) -100 50
soil Plug Weight ( kN) 20.00
Soil Support Dashpot .SO0 4.029
soil Support Mass ( kN) 150.00 300.00

Figure 4. The results of CAPWAP for TPl

Table 2. Shaft resistance of socket, toe resistance and dynamic parameters

Pile Shaft Toe Unconfined Shaft damping Toe damping Shaft soil Toe soil
No. resistance resistance compression dashpot dashpot support mass support mass
(kPa) (kPa) .f;,J(MPa) Jss(EMc) Jst(ENc) MdkN Mt (kN)
TPl 340 4600 0.50 4.00 150 300
TP2 290 3500 0.60 19.80 60 200
TP3 250 3800 2.0-8.0 0.45 19.00 150 300
TP4 300 4700 0.45 9.30 50 150
TP5 630 13400 0.60 6.50 50 120
TP6 710 16500 10.0-40.0 0.80 12.00 80 150
TP7 550 1 I500 1.oo 17.00 50 80
TP8 490 17500 0.60 14.00 100 150
TP9 460 14100 1.20 11.60 150 200
TPlO 560 19x00 10.0-40.0 1S O 5.50 100 200
TPl1 480 16800 I .oo 16.00 100 200
TP12 120 2260

14
5 CONCLUSIONS
Utilizing wave equation, a radiation damping
dynamic model is suggested herein for rock-
socketed pile. By analyzing the results of twelve
dynamic loading tests, the following conclusions
are made.
1. Radiation damping dynamic model resolves
the energy dissipating in pile-rock interfaces for
rock-socketed pile.
2. Bearing behaviors of rock-socketed pile
depend on the length of pile. A long pile (l/a340)
has less mobilizing action of the socket, but for a
short pile (1<20m and d 2 8 0 0 m m ) the socket will
undertake 70%-95% total load.
3. The ratio ts of shaft resistance to toe resis-
tance depend on the rock type and the construction
of pile. For soft rock, the ratio tS achieve values
of 1/20 to 1/10, but for granite, it ranges between
1/40 and 1/20.
4. For hand-dug belled pile, belled bottom re-
duces the shaft resistance of socket and the influ-
ent length over bottom ranges about (1-2)(D-d).
5. The bearing capacity of rock-socketed pile
can be determined by dynamic loading test with
rational dynamic parameters.

REFERENCES

[ l ] Shi Pei-dong. Vertical Bearing Capacity of

Rock-Socketed Pile. Chinese Journal of
Geotechnical Engineering, Vol. 16, No.4,
1994.
[2] Robert G.Horvath & T. Cameron Kenmey.
Shaft Resistance of Rock-Socked Drilled
Piers. Deep foundation, 182-2 1 1.
[3] Code for Design and Investigation of Building
Soil and Foundation, DBJ13-07-91, Fujian
province in China.
[4] Technical Code for Building Pile Foundation,
JGJ94-94, China
[ 5 ] CAPWAP Manual. Goble Rausche Likins and
Associates, Inc. 1993
[ 6 ] Feng Shi, Chun Liu & Lai-bing Cai. Experi-
mental Research on Bearing Capacity of
Belled Piles, Fujian Academy of Building Re-
search, 1997
[7] Zhang Yiao-nian & Gong Yi-ming. Research
on Bearing Capacity of Large-Diameter Pile
in Fuzhou, Fujian Academy of Building Re-
search, 1995

15
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyarna & Beirn (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Pile acceptance based on combined CAPWAP analyses

Robert E Stevens
Fugro-McClellund Marine Geosciences Incorporuted, Houston, Tex., USA

ABSTRACT: The pile driving hammer selected for a particular installation may be large enough to drive the
piles to design penetration, but not large enough to overcome the long-term static capacity. In clay, the skin
friction during driving is generally much smaller than that mobilized under static loading because large excess
pore pressures are generated during continuous driving. CAPWAP analyses are used to estimate the
distribution of the soil resistance to driving along the length and at the toe of a pile during continuous driving
and after a set-up period. It is shown that by combining these results, it is possible to proof test a pile without
the expense of mobilizing a larger hammer to the site. Examples are provided to illustrate the use of
combined CAPWAP analyses to interpret the results of redrive tests in clay and to estimate the static pile
capacity when refisal occurs during continuous driving.

1. INTRODUCTION varied until a best match between measured and

File driving is monitored using strain transducers
and accelerometers attached near the pile top. The
energy transmitted to the pile is obtained by
integrating the product of the measured pile top
force and velocity. The ratio of the transmitted
energy to the rated hammer energy is defined as
the system efficiency. The ram momentum is
obtained by integrating the measured pile top force
until the measured velocity equals zero. The ram
impact velocity is obtained by dividing the ram
momentum by the ram mass. The soil resistance to
driving is determined from the measured force and
velocity and a damping coefficient that is a
function of soil type. The cushion stiffness and
cushion coefficient of restitution are obtained by
using force-time characteristics (rise time and
peak-to-peak time) obtained in a pre-installation
parametric study.
A signal matching program, such as the
CAse Pile Wave Analysis Program (CAPWAP)
developed by Rausche (1970), is used to estimate
soil quake and damping parameters, and the
distribution of the soil resistance to driving along
the length and at the toe of the pile. The pile is
divided into continuous segments and calculations
are made using a traveling wave algorithm. Either
the measured pile top force or velocity is used as a
boundary condition, and the complementary
quantity is computed and compared with the
measured quantity. The set of soil parameters is
17
computed pile top force or velocity is obtained.
During continuous driving, the clay
surrounding a pile is remolded and large excess
pore water pressures are generated. Because the
excess pore pressures decrease rapidly with radial
distance from the pile, water will begin to flow
laterally out of the disturbed zone and the clay will
consolidate. As pore pressures dissipate, pile
capacity increases. Field measurements (Bogard
and Matlock, 1990) have shown that the time
required for driven piles to regain f i l l capacity can
be relatively long. Redrive tests are valuable
because a substantial increase in capacity occurs
within a relatively short period of time after
driving is terminated.
2. CASE HISTORY NO. 1
Our first case history is for a redrive test performed
with a hammer that was large enough to mobilize
the full soil resistance. Two 14-in.-square precast
prestressed concrete piles having a length of 75 ft
were driven with a Vulcan 010 hammer. Soil
conditions consisted of fill to 8 ft, very soft to soft
clay to 33 ft, medium-dense sand to 44 ft,
underlain by firm to stiff clay. To aid the
installation, a g-in.-diameter pilot hole was drilled
to a depth of 40 ft. The maximum resistance
encountered during initial driving occurred in the
sand stratum, even with pre-drilling. The piles
were driven to a depth of 66 R, with a final blow
count of 10 blows per foot (bpf).
18
19
Figure 3. Stresses and Soil Resistance Measured for Redriving of Pile 271

20
Figure 4. Stresses and Soil Resistance Measured for Redriving of Pile 480

21
Figure 5. Soil Resistance Determined from CAPWAP Analyses
The soil resistance (RMX) at the end of
continuous driving, as shown in Figs. 1 and 2, was
about 91 kips (0.40 MN) for Pile 271, and 102 kips
(0.45 MN) for Pile 480. Also shown are the
maximum compressive ( C S X ) and tensile (TSX)
driving stresses. The lower compressive stress
plotted is the average of two strain transducers, and
the higher compressive stress is the maximum
value obtained from either transducer. The piles
were redriven 5 fi (1.5 m) afier a set-up period of
4.7 days. The restart blow counts were 52 bpf for
Pile 271 and 41 bpf for Pile 480. The maximum
22
soil resistance, as shown in Figs. 3 and 4, was 295
kips (1.31 MN) for Pile 271 and 337 kips (1.50
MN) for Pile 480, resulting in a set-up factor of
about 3.3 in less than 5 days. Also of interest is
the shape of the soil resistance versus penetration
curve during redriving. The soil resistance
decreases about 40 percent as the pile is redriven
3 feet, but is about twice the soil resistance
measured at the end of continuous driving.
 B5-2 A5-4 A5-2 A5-1 A1-4

Blow Count, bpf 20 16 19 17 16

Penetration, ft 383 3 83 380 385 383
Hammer Efficiency, % 85 87 78 -- 80
System Efficiency, % 51 57 51 -- 50
Transmitted Energy, k-R 439 496 443 -- 430
Stress, ksi 18.8 22.0 20.7 -- 21.0

B5-2 A5-4 A5-2 A5-1 A1-4

Blow Count 2013" 2613" 7816" 5 013 8013 'I

Penetration, ft 383.25 383.25 380.5 385.25 383.25

Hammer Efficiency, YO 80 81 72 82 --
System Efficiency, % 44 40 51 56 56
Transmitted Energy, k-ft 380 3 50 44 1 485 485
Stress, ksi 20.0 22.0 20.0 22.0 23 .O
Delay, days 0.7 2.3 5.3 7.95 13.35

Figure 6. Summary of Driving System Performance Data

3. CASE HISTORY NO. 2
Our second case history is for a series of redrive
tests performed with a hammer that was not quite
large enough to mobilize the full soil resistance.
Sixteen 72-in.-diameter open-ended pipe piles
were driven to 383-R (117-m) penetration with a
Menck MRBS 8000 hammer. Soil conditions
consisted of alternating strata of silty clayey sand
or sandy clayey silt and stiff to very stiff sandy
silty clay.
The distribution of the soil resistance was
determined from CAPWAP analyses performed for
a series of redrive tests. Results are presented for
continuous driving at the end of driving for four
different piles driven to final blow counts of only
16 to 20 bpf. Results are also presented for the
beginning of redrive are for five different piles,
each driven about 3 inches. The soil resistance
along the length of the pile is presented in Fig. 5
for continuous driving and after set-up times
ranging fi-om 0.7 to 13.4 days. This plot shows
that the full soil resistance was not mobilized for
the redrive tests performed after the two longest
delays, as indicated by a change in slope in the soil
resistance along the bottom 15 ft (5 m) of the pile.
A summary of the driving system performance
data is presented in Fig. 6.
4. CASE HISTORY NO. 3
Our third case history is for a series of redrive tests
performed with a hammer that is clearly not large
enough to mobilize the full soil resistance. The
soil resistance mobilized during a series of redrive
tests performed on a 1.6-m-diameter steel pipe pile
driven to a penetration of 26 m in a very silty clay
is presented in Fig. 7. The four lower bound soil
resistance profiles are for continuous driving. The
pile was redriven by applying only two
consecutive hammer blows with a PMJ-400
hydraulic hammer after delays of 6 minutes, 15
minutes, 33 minutes, 2 hours, and 66 hours. The
resistance generally increases with time, but the
resistance after a 33-minute delay appears to be
slightly smaller than the resistance mobilized after
a 15-minute delay, and the resistance after a 66-
hour delay is less than the resistance mobilized
after a 2-hour delay. The maximum resistance
overcome by the pile driving hammer is about
3300 kips (14.7 MN). The hammer is too small to
mobilize the full soil resistance during the redrive
tests. This is shown very clearly for the redrive
test performed after the 66-hour delay. Almost no
soil resistance was mobilized over the bottom
quarter of the pile.
In our combined CAPWAP analyses, we
have assumed that the soil resistance mobilized
during continuous driving is a lower bound, i.e.,
the soil resistance mobilized on a particular pile
segment is assumed to be the larger of the actual
resistance mobilized or the resistance mobilized
during continuous driving. The soil resistance
mobilized after a 66-hour delay in the combined
CAPWAP analysis is about 4400 kips (19.6 MN),
as shown in Fig. 8. In the standard CAPWAP
analysis, the resistance mobilized was only 3305
kips (14.7 MN).

23
Figure 8. Soil Resistance Determined from Combined CAPWAP Analyses

25
 LENGTH - - - - - - - - - - - --=
PENETRATION - - - - - - ---- -- -----=.270ft
-
EFFECTIVE LENGTH - -- -
EQUIVALENT AREA - - - - - - - -- -
SPRING----------- - -- =928.1 kips/in.
Hammer Properties - - - - - - - - - = VUL. 040,060,560
EFFICIENCY - - - - - - - - - - -
ENERGY - - - - - - - - - - - ---- 120,180J3_0@ ft-klps
RAM WEIGHT ( W i ) - - - - - - - - --- 40;60,62.5 kips

-- - - - - -~ 1 5 %
PERCENT Ru AT PILE TIP
Qp - - .- - .- - - - - - - - = 0 , 1 0 i n .
Qs - - - - - - - - - - - --=O.lOin.
JP - - - - - - - - - - - - r0.10 sec/fl
Js - - - - - - - - - - - - =0.10 sec

0
Blow Counts, Blows p e r Foot

Figure 9. Driving Resistance-Blow Count Curves

To help put this in perspective, Fig. 9
shows that the resistance mobilized by a Vulcan
060 hammer is about 3300 kips (14.7 MN), and the
resistance mobilized by a Vulcan 560 hammer is
about 4100 kips (18.2MN), assuming the same
soil and pile parameters. Comparing the three
hammers shown in Fig. 9, we see that by
mobilizing a hammer having 50 percent more rated
energy results in only a 10 percent increase in the
maximum soil resistance overcome, and
mobilizing a hammer having 150 percent more
rated energy results in a 37 percent increase in the
maximum resistance overcome.

26
5. CONCLUSIONS

CAPWAP analyses were used to estimate the

distribution of the soil resistance to driving along
the length and at the toe of a pile during
continuous driving and after a set-up period. Our
final case history was for a series of redrive tests
performed with a hammer that could overcome a
maximum resistance of about 3300 kips (14.7
MN). The soil resistance mobilized after a 66-hour
delay is shown to be about 4400 kips (19.6 MN)
using combined CAPWAP analyses. In the
combined CAPWAP analyses, the soil resistance
mobilized on a particular pile segment is assumed
to be the larger of the actual resistance mobilized,
or the resistance mobilized during continuous
driving.

REFERENCES

Bogard, J.D. and Matlock, H. (1990), "Application

of Model Pile Tests to Axial Pile Design,"
Proceedings. Twenty-second Annual Offshore
Technology Conference, Houston, Vol. 3, pp. 271-
278.

Rausche, F. (1970), "Soil Response from Dynamic

Analysis and Measurements on Piles," Ph.D.
Dissertation, Division of Solid Mechanics,
Structures, and Mechanical Design, Case Western
Reserve University, Cleveland, Ohio, 320 p.

Roussel, H.J. (1979), Pile Driving Analysis of

Large Diameter Hiph Capacity Offshore Pipe
Piles, Ph.D. Thesis, Department of Civil Engi-
neering, Tulane University, New Orleans.

27
This Page Intentionally Left Blank
 Applicationof Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam,ISBN 90 5809 1503

Theoretical study on effect of pile shaft resistance on rebound during pile

driving

Chen Ren-peng & Chen Yun-ming

Geotechnical Engineering Institute, Zhejiung Universio, Hungzhou, People’s Republic of China

ABSTRACT: In this paper the effect of the shaft resistance on the rebound at pile-top during pile driving is
studied. The soil around pile shaft is assumed to be rigid-plastic and that under pile-tip is assumed to be ideal
elastoplastic. The driving force acted on pile-top is simplified to be a triangular impact. The kinematics
equation of pile-tip is established. With the one-dimension wave equation, the movement of pile-tip and pile-
top are obtained. The rebound at pile-top can be written in a very concisely form. It proves that the shaft
resistance decreases the rebound at pile-top. When pile is long enough or the soil around pile is very stifc the
rebound decreases obviously. The neglect of the shaft resistance will bring on a large amount of errors.

1 INTRODUCTION the integration. Fortunately such problem has been

During pile driving, that rebound at pile-top gives
more information about resistance than penetration
has been recognized (Chen et a1 1996, Uto et a l ,
1992) The theoretical model of point resistance,
obtained from PD- and PDA- measurements, has
been developed in the recent years (Chen et a1 1996,
van Weele et a1 , 1994)) This makes it possible to
estimate static point resistance during pile driving
In the method, shaft resistance is neglected for the
reason that transverse vibration of pile greatly
decreases the shaft resistance Driving resistance is
assumed to be mainly consisted of point resistance
Soil under pile-tip is supposed to be a ideal
elastoplastic material, with the ultimate static
resistance of X, In order to make the analysis more
simple, impact force caused by hammer, is
simplified as a triangular force The driving model is
shown in Figure l(a) The kinematics equation of
pile-tip can be easily established and solved Then
point resistance can be written as
Re hoinid
N, =c,- Z (1)
to
Where C,is a dimensionless constant, almost equals
t o 1 3. to is the duration of contact between hammer
and pile, Z is the impedance of pile By double
integrating the recorded acceleration of the pile as a
function of time, the pile-top settlement can be
obtained For it may take approx 200ms before the
pile-top is at rest after each blow and only then the
permanent deflection is reached, it is imperfection in
solved by IFCO BV (van Weele, 1994).
In the equation (1), impedance of pile is known
before pile driving, and to for each blow can be
obtained from the pile-top velocity. The rebound can
be obtained from the pile-top settlement
Furthermore, the total settlement and the permanent
settlement for each blow still can be presented.
Figure 2 shows the typical diagram of the point
resistance vs. depth, which clearly indicates the
thickness of each soil-layer and agrees well with the
results of CPT cone resistance. The new method
requires only a single, small and robust sensor and
the data collection and interpretation is done
automatically by the field computer (van Weele et al.
1994). PDA- analyses in combination with the
CASE- or CAPWAP-methods are different as they
require more and also more complicated
instrumentation. It also shows that the dropheight,
cushion stiffness and soil damping have little effect
on the point resistance compared with other method,
such as CASE- and CAPWAP method. The method
has been put into use (van Weele et al. 1994), and
seems very helpful. The defect is that the shaft
resistance is neglected. In practice, soil resistance
includes shaft and point resistance during pile
driving. If the pile penetrates through stiff soil or the
pile is very long, the shaft resistance is always very
great. The neglect of the shaft resistance certainly
brings on great errors. In this paper, the shaft
resistance is taken into account and the method is
improved.

29
2 THEORETIC ANALYSIS
2.1 Theoretic model ofpile driving

Figure 3. Driving Force

Simplified as a triangular force

(a) pile driving model
by Chen (1996). No
shaft resistance considered
Figure I . Pile Driving Model
(b) analysis model presented in
paper, shaft resistalce is
considered.
shaft. is pile length and is the stress wave
velocity, which is described as c = 4~I p , where E
is the elastic modulus and p is the density of pile
material. The reflection wave at pile-top is not
considered. Up-going velocity is supposed to reach
pile-top without attenuation.

2.2 Kiriemafics equafiori of pile- f i p

The kinematics equation of pile-tip can be described

as
C V ( f )+ K [ l / ( f )- U , (I)] = Fd ( t ) + I;, ( t ) (2)

where U ( f ) and V ( f ) are the movement and the

velocity of pile-tip respectively, U p ( f ) is the
permanent movement of pile-tip. F d ( t ) is down-
going force. Z is the impedance of pile. C is the
damping coefficient of soil at pile-tip and K is the
stiffness of soil at pile-tip. <,(t) is the up-going
force, which can be described as

fi;, (1) = F‘, ( t )- ZL’(f) (3)

Figure 2. Point Resistance vs. Depth Estimated by
the Method (Chen 1996)
As shown in Figure l(b), soil at pile shaft is
postulated to be rigid-plastic, and soil at pile-tip is
postulated to be elastoplastic and modeled by Smith
soil model (Smith, E.A.L., 1960). The soil at pile
shaft is homogenous, and the shaft resistance
distributes evenly around the pile. The driving
model is shown in Figure l(b). The driving force is
simplified to be a triangular impact force, as shown
in Figure 3. In Figure 3, F/Jis the amplitude of the
driving force. tl is the loading time. In Figure I(b),
R, is the static shaft resistance and R, is the static
point resistance. C, is the damping coefficient at pile
The11 EqL1atiorr (2) 1s rewitten as
[C+ z]V(t) + K [ U ( f ) [ I p ( t ) ]= 2Fd (I)
- (4)
or
[(’ + 2 1 9+ K [ l / ( t )-U,(r)]
= 2 F , (f) (5)
dt
The movement of pile-tip U(t) (or soil at pile-tip)
can be divided into three stages: elastic movement,
plastic movement and rebound. When the movement
of pile-tip U ( f )is less than the maximum elastical
movement of soil at pile-tip (denoted as Qp),there is
only elastic movement of soil. The static soil
resistance increases linearly with the elastic
movement of pile-tip until it reaches at the
maximum static resistance R,. Equation ( 5 ) is
simplified as

30
 the multiplication of velocity caused by driving
[ C + Z ] U ' l l O + K U ( t ) =2F,(t) force and damping coefficient at pile shaft.
dt
Therefore at pile-tip down-going wave caused by
When the movement of pile-tip U(t) exceeds Qp, soil resistance at pile shaft is
plastic movement takes place. Then the maximum
static soil resistance remains constantly. Equation
( 5 ) is written as

[C + 2 1 9+ I?, = 2Fd(f) (7)

dt
When the movement of pile-tip U(t) reaches to its
limit, then rebound takes place. The static soil t <to
resistance decrease when the movement of pile-tip
U(t) decreases. The movement of the pile-tip can be
described with Equation ( 5 ) .

Equation ( 5 ) can be solved step by step with the

condition of the deformation compatibility at the end
of each stage.

2.3 Lblocrty arid movemerit at pile-tip

In the following, seven dimensionless parameters are

introduced:
Figure 4. Stress wave propagating in pile
m ? = C ( K t o ) , n i ? ~ = C(Kto),
~, T=2L (cto).
With the assumption of the Smith damping law, Down-going force Fd(t ) can be described as
there exist
F b ( t ) = F ( t )+ ZL;:i (1) (11)
Furthermore, up-going wave reaching pile-top is
expressed as

where J p is the Smith damping coefficient L:;la, is

I-2
the maximum velocity of pile-top, where
z
J;&, = F,, / <fo
When pile hammer impacts on pile-top, the impact
force propagates downward through pile When the Substituting Equation (12) into Equation (1 l),
stress wave reaches point x, as shown in Figure 4, Equation ( 6 ) , Equation (7) and Equation ( 5 ) step by
the shaft resistance is excited simultaneously Then step, dimensionless form of the movement and the
the tensile and compressive stress wave propagating velocity at pile-tip are derived by
in the opposite directions are generated The tensile
wave propagates downwards and the compressive
propagates upwards The amplitudes of the two
kinds of stress wave equal to the half of the
amplitude of the shaft resistance When the driving
force reaches to the pile-tip, the amplitude of the
shaft resistance equals to the half of the total shaft
resistance, i e R,/2 The dynamic shaft resistance is

31
 (7 - 1 ) - n , 2m m21
- ~ - _ _ _ _ _ _ _ _
2m
( n , + -- - - - - q e
mm -'
(rl - 1) (rl - 1) "1 (77 - 1) 77 "177

2 ( r - 1) (n+ n,) 1112,(r -1)

m(17-1) m miii1(r7--1)

Figure 5 Typical solution ofEquation (2) Figure 6 Velocity and movement of pile-top
and pile-tip
where T = t / f ( , , r,, and z, are the dimensionless time
when plastic movement and rebound take place and
respectively, w i and 11'2 are integration constants, .ye/
is residual movement when rebound takes place = ('I + "1 7'() - '1
2 - 1112, +1 (16)
From Equation (1 3) and Equation ( 1 4) rn and T, can
be written as where. m mI-nT2 From down-going velocity and
7
velocity at pile-top, up-going velocity can be derived

32
With the previous assumption, up-going velocity and
will reach pile-top without attenuation. Figure 5
shows typical solution of Equation (2).
/('C = ~ ff2
(@I + fir1 1

as
Assuming there is no residual compression in pile,
CI lie hoitnd
the final residual movement of pile-tip is equals to RI =
(1 -A(,',) fg
z
that of pile-top The rebound of pile-top can be
expressed as or
lie /,oilrid = [ J , ! , ~ ~-
,~ ( +~
Q, i ~ (18)
~ lie~
hoimd =
RJtg 1
-_____
-

z CI
where (JAlc,r and [J:lc,x are the maximum
movement of pile-top and pile-tip respectively
lJ,!,lay and [J!'lay can be derived by integrating
velocity at pile-tip and pile-top as shown in Figure 6 If 3, = 0 , Equation (23) can be rewritten as
Because velocity caused by up-going soil resistance, lie boiriid
L'!'(f), is little compared with J " ( f ) and L',,(t), it
R, = ~ z
f0
can be assumed that the movement of pile-top which is the same as that obtained by Chen et a1
reaches its maximum when up-going velocity, (1 996) Commonly, during pile driving there exist
P;, (f)reaches zero. 1~~=(10-15)ms, Q,3 (5-1O)mm, ,J!, (0 5-1 5)s/m,
L&., = ( 2 5-4 0 ) d s Therefore the ranges of the
p i s denoted as the dimensionless time when up- dimensionless parameters can be known
going velocity reaches zero. As shown in Figure 5 i?=(O 1-0 5), tl,=(O 03-0 02), ~ ( 1-00 5 ) ,
there exists m,=(O 08-0 25)/ri, m2=(0 15-1 5), n?2~=(005-0 5 ) ,
7'=( 1 0-2 0) Substituting the dimensionless
parameters into Equation (25) the range of c'd can be
known C'(,=(O 5-0 9) s 0 7 ('>can be estimated
by Equation (24) In-situ measurement also shows
that CllSvaries little, and equals to 1 3 (Chen et a1
1996) In Equation (26), the rebound of pile-top,
duration time of driving force ill can be obtained
from the acceleration signal

where

and

f f 2 =--a1
p
+ ( T 2 p') +-+
- n?Zr(I - p)'
(21)
47' 2 2m1q (1 - 17)
Equation (1 9) can be written as
h W O 4fO
u;1,,, - ~J,,,,, = a1 7 -a:! 7 (22)
Substituting Equation (22) into Equation (]8), the Figure 7. Effect of shaft resistance on rebound and
static resistance at pile-tip can be obtained estimated point resistance
Re houizd
K , =,'C ~ z+ ( ' d l i l (23) The curve of the point resistance verus depth is very
10 similar to that of cone resistance of CPT. With the
where curve, it is convenient to decide the depth of the
,'C = (ff1 + m , > 1- (24) bearing stratum and pile capacity.

33
4 EFFECT OF SHAFT RESISTANCE ON Uto, K , Fuyuki, M & Omori, H 1992 New
REBOUND development of pile driving management
system Proc. 41h hit. Cut$ on Applrcatroii of
The effect of 2 on I?, is shown in Figure 7. When 2,
Str.e.ss-wnve Theory to Plies 1992 3 5 1-356, the
=O, it means that the shaft resistance is neglected
Hague
When 3, =1, it means the shaft resistance is the same Weele, A F & Schellingerhout, A F G 1994 Eficient
as point resistance. It can be seen from Figure 7, the driving of precast concrete piles Pruc. (’mf
shaft resistance has a very obvious effect on
DeveIcpnmrt 117 Geotech. Eig. 1994 Bangkok
rebound. If the error of the neglect of the shaft
resistance is supposed to be less than ten percent, iz
must be less than 0.16, which means shaft resistance
is less than thirteen percent of point resistance. Soil
resistance at pile-tip obtained with the consideration
of soil resistance at pile shaft is stronger than that
obtained without the consideration of soil resistance
at pile shaft

Equation (23) shows the influence of shaft resistance

on rebound. The shaft resistance distinctly decreases
the pile-top rebound. The greater the shaft resistance
is, the greater the decrease is. When A =0.7, the
rebound is only half of that neglecting the shaft
resistance. It was also proved in-situ measurement
that when the driving of the pile is stopped for some
reasons, the rebound will decrease greatly during
redriving. It is due to the increase of the shaft
resistance during the break.

5 CONCLUSION
In this paper the effect of the shaft resistance on the
pile-top rebound during pile driving is studied. The
soil around pile shaft is assumed to be viscous-rigid
and that under pile-tip is assumed to be viscous-
elastic. The driving force acted on pile-top is
simplified to be triangular impact. The kinematics
equation of pile-tip is established. With the one-
dimension wave equation, the movement of pile-tip
and pile-top are obtained. The rebound at pile-top
can be written in a very concisely form. It proves
that the shaft resistance decreases the rebound at
pile-top. When pile is long enough or the soil around
pile is very stiff, the rebound decreases obviously.
The neglect of the shaft resistance will bring on a
large amount of errors.

REFERENCES
Chen, Y.M. et al. 1996. Determining soil resistance
from pile driving by using an accelerometer.
Proc. .jth Irzt. Cot!$ on Application of Stress-
wave Theory to Piles : USA.
Smith, E.A.L. 1960. Pile driving analysis by the
wave equation. J. Soil Mech. & Fo~iid.Eiig.
Dzv. ASCE, 1960 86(4).

34
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 1503

Time effect in determining pile capacity by dynamic methods

M. R. Svinkin
VibraCoizsult, Cleveland, Ohio, USA

ABSTRACT: Knowledge of pile capacity as a function of time is important for proper performance of pile
foundations. This paper shows the advantages of the dynamic methods in determining pile capacity and
points out the necessity of considering the time effect for correct assessment of the accuracy of dynamic
methods. The prediction of pile capacity in pre-driving wave equation analysis can be made by the use of
variable damping as a function of time. Pile capacity obtained from a static loading test cannot be accepted
as a unique standard because the static loading test yields the pile capacity at the time of test only, due to the
consolidation phenomenon. Dynamic capacity testing has this same limitation. Any comparison of static and
dynamic tests has to be made for tests performed within a +art duration.

1 INTRODUCTION This method is a convenient tool in the pile driving

Pile capacity changes after pile installation.
Accurate and reliable determination of pile capacity
is very important for design, construction and
reliability of pile foundations.
Static analysis and the static loading test are
traditionally used for calculation and verification of
pile capacity. Contemporary dynamic methods have
advances in evaluation of the hammer-pile-soil
system and in data acquisition during pile driving
and restrikes. Therefore during the two last
decades, dynamic methods have become an integral
part of pile capacity prediction and measurement for
numerous projects.
Dynamic methods have certain advantages and
some uncertainties in their application. Wave
equation analysis of driven piles is a prevalent
method of pile driving stress calculations. Besides
driveability analysis, the wave equation method is
used for prediction and determination of pile
capacity during both the design stage and for
construction control during pile installation.
Unfortunately in most cases, computed pile capacity
differs substantially from results of both static and
dynamic load tests.
Dynamic measurements of force and velocity at
the pile head during pile driving, followed by a
signal matching procedure, is the most common
method for dynamic determination of pile capacity.
industry. However, though dynamic methods have
been used in practice for years, actual reliability of
dynamic methods is vague because their comparison
with static loading tests is made incorrectly in most
cases.
This paper considers the time effect on prediction
and determination of pile capacity by dynamic
methods after pile installation.
2 PILE CAPACITY VARIATIONS WITH TIME
Pile capacity determined at the end of initial driving
(EOID) in various soils changes with time. During
pile installation, the soil around the pile experiences
plastic deformations, remolding , and pore pressure
changes. Excess pore water pressure developed
during driving reduces the effective soil shear
strength and ultimate pile capacity. After the
completion of pile driving, soil reconsolidation,
manifested by the dissipation of excess pore pressure
at the soil-pile interface zone, is usually
accompanied by an increase in pile capacity (soil
setup). The amount of increase in pile capacity
depends on soil properties and pile characteristics.
In saturated sandy soils, ultimate pile capacity may
decrease (soil relaxation) after initial driving due to
dissipation of negative pore pressure. Changes of
strength in soil after driving and the time required

35
for return of equilibrium conditions are hghly
variable and depend on soil type, and pile size and
type.
Piles have to withstand design loads for a long
period of time. Therefore, the consequences of soil
modification around the pile are essential with
respect to changes of pile capacity. The
phenomenon of time-dependent strength gain and
loss in soils related to pile driving has been studied
and published, for example Axelsson (1998), Chow
et al. (1998), Fellenius et al. (1989), Long et al.
(1999), Randolph et al. (1979), Seed & Reese
(1955), Skov & Denver (1988), Svinkin (1996a),
Thompson & Thompson (1985), Thorburn & Rigden
(1980), Tomlinson (1971), Wardle et al. (1992),
York et al. (1994) and others.
Pile capacity as a function of time is displayed, for
example, in Figure 1. Static loading test (SLT) as
well as dynamic testing (DT) yields the pile capacity
at the time of testing, Svinkin (1997, 1998b). By
way of illustration, results of DT and SLT are
shown in Figure 1 for two identical cylindrical,
1372 mm x 127 mm, prestressed concrete piles, TP1
and TP2. The depth of penetration of each pile was
approximately 24.4 m. The soil consisted of about
25.6 m of mainly gray clays followed by a bearing
layer of silty sand. The water table was at the
ground surface. A Delmag D 46-13 hammer was
employed for initial driving and restrikes. Each of
the piles TP1 and TP2 was tested 2, 9 and 22 days
after the end of initial driving. The difference was
that three restrikes were made for TP1 and three
SLTs were made for TP2. Pile capacity from three
SLTs was a function of time as was the pile capacity
obtained from DT.
3 DYNAMIC FORMULAS
The well-known dynamic formulas have been
criticized in many publications. Tested data in
Figure 1 help to explain the causes of unsatisfactory
prediction in pile capacity by dynamic formulas.
Dynamic formulas using maximum energy, pile set
and maximum displacement from DT do not take
into account the time between SLT and DT. In the
case of a few SLTs made on one pile, like three
SLTs performed on pile TP2, what would be the
reliability of pile capacity prediction by the energy
approach methods? Which SLT should be taken for
comparison? Currently, there are no answers to
these questions. In principle, dynamic formulas
cannot predict time-dependent pile capacity (Svinkin
1998b).
36
Figure 1. Pile capacity versus time for prestressed
concrete piles in clayey soil, after Svinkin & Woods
(1998a)
4 WAVE EQUATION ANALYSIS
The main goal in using the wave equation method is
to provide a better prediction of the pile capacity, as
a function of pile penetration resistance, than can be
obtained from classical dynamic formulas.
The wave equation method was originally
suggested by Smith (1960) to compute the pile
capacity at the end of driving. This method is also
used for prediction of pile capacity at restrike
(RSTR) performed at any time after EOID. By
adjusting wave equation analysis input with results
of dynamic measurements, some researchers , for
example, Hunt and Baker (1988), York et al. (1994)
have obtained good correlation between computed
and observed pile capacities. However, in most
other cases, computed pile capacity differs
substantially from results of static or dynamic tests.
Existing dynamic models of the pile-soil system
mainly use a velocity-dependent approach for
calculation of the dynamic resistance as a damping
component of the total resistance during pile driving.
There are various linear and nonlinear relationships
between the damping component and the velocity.
For certain pile capacity, the dynamic resistance
depends only on pile velocity and the damping
coefficient.
Statistical analysis indicates no correlation between
the pile penetration resistance and velocity values.
The pile-soil system changes with time after the
completion of driving, but the pile velocity is only
a pile property and remains in the same range for
EOID and RSTRs. The largest values of pile
velocity measured at the upper end of the pile and
calculated along a pile shaft depend only on pile
parameters and energy transferred to the pile and
cannot reflect regain in soil strength and pile-soil
adhesion after EOID. This is the first cause of
unsatisfactory prediction of pile capacity with time
after EOID .
The basic disadvantage of many models is the
attempt to select the model parameters directly from
actual soil properties. This can yield acceptable
results for some cases, but in general this approach
is not successful in finding good correlation between
predicted and actual pile capacity after EOID. The
use of the constant damping coefficients for
calculation of the dynamic resistance is the second
cause of unsatisfactory prediction of pile capacity
with time after EOID.
Neither the pile velocity nor the damping constant
can reflect time-dependent variation of the pile-soil
system after EOID, Svinkin (1996~). The existing
approach of computing the dynamic resistance does
not take into account soil consolidation around the
pile after EOID and therefore cannot provide
determination of pile capacity as a function of time
after pile installation.
For the idealized Smith wave equation model, it is
important to find an appropriate combination of
parameter values, mainly paying attention to soil
variables, in order to achieve the accurate prediction
of pile capacity. Probably, there is only one
direction to enhance prediction accuracy of the
dynamic resistance with the velocity dependent
approach. Variation of the pile-soil system after the
completion of driving can be taken into account by
a variable damping coefficient which should be
considered as a function of time and other
parameters characterizing soil consolidation around
Smith damping as a function of time for various soil
the pile. For example, the soil shear modulus or the
types is shown in Figures 2 and 3. It can be seen
frequency of the fundamental mode of the pile-soil
that the shaft damping coefficient in clay is much
system could be considered, Svinkin (1996b). It is
higher than in unsaturated sand, but upper values of
assumed that the variable damping coefficient is
this coefficient in saturated sandy soil (sand with
independent of pile velocity. Inclusion of variable
high damping) are close to ones in clay, Svinkin
damping is thought to be the next step in the
(1995a, 1995b).
development of Smith’s model with the velocity
The idea of variable damping has been confirmed
dependent approach for representation of the
by results of statistical analysis performed by Liang
dynamic resistance.
and Zhou (1997) who have found that the damping
The damping coefficient as a function of time can
coefficient is affected by the time.
be found on the basis of back calculations using the
Soil damping is the key parameter for adjustment
wave equation model of the pile-soil system with
of wave equation solutions with time-dependable soil
known capacity. The five soil damping options,
properties in pre-driving analysis. The use of the
available in GRLWEAP program (GRL Manual
variable damping coefficient gives an opportunity to
1997), were investigated: Standard Smith Damping,
compute the time-dependent pile capacity by the
Viscous Smith Damping, Case Damping, Coyle-
wave equation method.
Gibson Damping, and Coyle-GibsodGRL Damping.
A trend of the damping coefficient increase with
time after EOID was found for all the considered
dynamic soil models and this trend is independent of
the damping resistances (Svinkin 1996b). Standard

37
5 DYNAMIC TESTING AND ANALYSIS

Dynamic testing followed by a signal matching

procedure has obvious advantage in determining pile
capacity at any time after pile installation. Since
dynamic testing is often used to replace the static
loading tests, it is important to ascertain the
adequacy of both SLT and DT. Design methods
predict pile capacity as the long term capacity after
soil consolidation around the pile is complete.
Independently of the time elapsed between the
driving of the test pile and the static loading test, the
ratio of the predicted ultimate load to the measured
ultimate load from static loading test is used for
approximate evaluation of the reliability of design
methods, Briaud and Tucker (1988). According to
the traditional approach, the main criterion for
assessment of the pile capacity prediction based on
Figure 4. Typical dynamic and static tests capacity
dynamic measurements is the ratio of capacities
correlation
obtained by dynamic and static tests or vice versa
(Figure 4).
It is necessary to point out that a ratio of DT/SLT
or vice versa, taken for arbitrary time between is taken into account for restrikes using the DT but
compared tests, is not a verification of dynamic is not in the SLT.
testing results. It is well-known that dynamic testing Static Loading Tests and Dynamic Testing present
methods yield the real static capacity of piles at the different ways of determining pile capacity at
time of testing, Rausche et al. (1985). This is not a various times after pile installation, but for valid
predicted value. Moreover, the static capacity from correlations two principal conditions have to be the
SLT is considered as a unique standard for same for both kinds of tests. I ) static and dynamic
assessment of dynamic testing results. Unfortunately, capacities must be compared at the same time after
that is a major error. As a matter of fact, pile pile installation in both SLT and DT methods, and 2)
capacity from Static Loading Tests is a function of the ultimate pile capacity is obtained in the SLT only
time and the so-called actual static capacity from if it provides the fully mobilized pile capacity (long
SLT is not a constant value. As it was shown in term capacity), similar to the DT, Svinkin (1997).
Figure 1, SLT, as well as DT, yields a different pile The adequacy of SLT and DT have to be
capacity depending on the time of testing, as confirmed by proper correlation of time. Due to the
measured after pile installation. consolidation phenomenon in soils, comparison of
For a few separate piles, it is possible to find SLT and DT can only be made for tests performed
published information regarding the time between immediately one after another. In practice, it is
static and dynamic tests. However, for the general sometimes difficult to make two immediately
case of assessment of reliability of the DT, the ratio successive tests, but nonetheless the time difference
of restrikes to SLT results has been considered for between both comparable tests should not exceed 1-2
various pile types, soil conditions and times of days during which soil setup changes only slightly.
testing lumped together as shown in Figure 4. What Closely time correlated comparisons of SLT and DT
is the real meaning of such mixture? Nobody have to be made in order to clarify the reliability of
knows. It is not a verification of dynamic testing at pile capacity by dynamic testing in soils with time-
restrikes and it is not assessment of real setup factor dependent properties.
because everything is lumped together without taking
into account the time between different tests. Such
a comparison of the pile capacities from SLT and 6 CONCLUSIONS
DT is invalid for piles driven in soils with time-
dependent properties because the soil properties at Contemporary dynamic methods are appropriate
the time of DT do not correspond to the soil tools to determine the time-dependent pile capacity:
properties at the time of SLT i.e. soil consolidation The prediction of pile capacity in pre-driving wave

38
equation analysis can be performed by the use of
variable damping as a function of time. Variable
damping is the key parameter to enhance accuracy of
wave equation solutions because this damping takes
into consideration soil consolidation after pile
installation.
The main criterion for accurate assessment of pile
capacity prediction based on dynamic measurements
of force and velocity at the pile head during driving
is the ratio of capacities obtained by dynamic and
static tests. Such a ratio, taken for arbitrary time
between compared tests, in not a verification of
dynamic testing results.
Dynamic testing and analysis yield the real, not
predicted, static capacity of piles at the time of
testing. The static capacity from a static loading test
is not a unique standard for assessment of dynamic
testing results. Both static loading test and dynamic
testing yields the pile capacity at the time of testing.
In soils with time-dependent properties,
comparison of static loading test and dynamic testing
must be made only for tests performed immediately,
in short succession.
Dynamic testing and analysis is a great tool to
determine the time-dependent pile capacity after pile
installation.
REFERENCES
Axelsson, G. 1998. Long-term increase in shaft
capacity of driven piles in sand. Proc. Fourth
Inter. Con. on Case Histories in Geotechnical
Engineering, St. Louis, 8-15 March, Paper 1.25.
Briaud, J.L. and L.M. Tucker 1988. Measured and
predicted axial response of 98 piles. Journal of
Geotechnical Engineering, ASCE, 114(9): 984-
1001.
Chow, F.C., R.J. Jardine, F. Brucy & J.F. Nauroy
1988. Effects of time on capacity of pipe piles in
dense marine sand. Journal of Geotechnical and
Geoenvironmental Engineering, ASCE, 124(3):
254-264.
Fellenius, B.H., R.E. Rker, A.J. O’Brien & G.R.
Tracy 1989. Dynamic and static testing in soil
exhibiting set-up. Journal of Geotechnical
Engineering, ASCE, 115(7): 984-1001.
CRL and Associates, Inc. 1997. GRLWEAP - Wave
Equation Analysis of Pile Driving, Manual,
Cleveland, Ohio.
Hunt, S.W. & C.N. Baker 1988. Use of stress-wave
measurements to evaluate piles in high set-up
conditions. In B. Fellenius (ed.), Proc. Third
Inter. Con. on the Application of Stress-Wave
39
Theory to Piles, Ottawa, 25-27 May: 689-705,
Vancouver: BiTech Publisher.
Long, J.H., J. Kerrigan & M.H. Wysockey 1999.
Measured time effects for axial capacity of driven
piling. Proc. I999 Transportation Research Board,
I 3 January, Washington, D.C.
Randolph, M.F., J.P. Carter & C.P. Wroth 1979.
Driven piles in clay - the effect of installation and
subsequent consolidation. Geotechnique, 29(4):
361-393.
Rausche, F., G.G. Goble & G. Likins 1985.
Dynamic determination of pile capacity. Journal
of Geotechnical Engineering, ASCE, 1985,
11l(3): 367-383.
Scov, R. & H. Denver 1988. Time-dependence of
bearing capacity of piles. In B. Fellenius (ed),
Proc. Third Inter. Conf. on the Application of
Stress- Wave Theory to Piles, Ottawa, 25-27 May:
879-888, Vancouver: BiTech Publisher.
Seed, H.B. & L.C. Reese 1955. The action of soft
clay along friction piles. Transactions, ASCE,
122: 731-754.
Smith E.A.L. 1960. Pile driving analyses by the
wave equation. Journal of the Soil Mechanics and
Foundation Division, ASCE, 1960, 86: 35-61.
Svinkin, M.R., C.M. Morgano & M. Morvant
1994. Pile capacity as a function of time in clayey
and sandy soils. Proc. Fijlh Inter. Con. and
Exhibition on Piling and Deep Foundations,
Bruges, 13-15 June: 1.11.1-1.11.8, Rotterdam:
Balkema.
Svinkin, M.R. 1995a. Pile-soil dynamic system with
variable damping. Proc. 13th International Modal
Analysis Conference, IMAC-XII, Beyond the
Modal Analysis, Nashville, 13-16 February, 1:
240-247, Bethel, Connecticut: SEM.
Svinkin, M.R. 1995b. Soil damping in saturated
sandy soils for determining capacity of piles by
wave equation analysis. Proc. DFI Annual
Member’s Conference, Charleston, South
Carolina, 16-18 October: 199-216, Englewood
Cliffs: DFI.
Svinkin, M.R. 1996a. Discussion of ’Setup and
relaxation in glacial sand’ by York et al., Journal
of Geotechnical Engineering, ASCE, 122(4): 3 19-
321.
Svinkin, M.R. 1996b. Soil damping in wave
equation analysis of pile capacity. In
F. Townsend, M. Hussein & M. McVay (eds.),
Proc. Fifh Inter. Con. on the Application of
Stress- Wave Theory to Piles, Orlando, 11- I 3
September: 128-143, Gainesville: University of
Florida.
Svinkin, M. R. 1996c. Velocity-impedance-energy
 relationships for driven piles. In F. Townsend,
h.1. Hussein & M. McVay (eds.), Proc. Fifth
Inter. Con$ on the Application of Stress-Wave
Theory to Piles, Orlando, 11-13 September: 870-
890, Gainesville: University of Florida.
Svinkin, M.R. 1997. Time-Dependent Capacity of
Piles in Clayey Soils by Dynamic Methods. Proc.
XIVth Inter. Conf. on Soil Mechanics and
Foundation Engineering, Hamburg, 6-12
September, 2: 1045-1048, Rotterdam: Balkema.
Svinkin, M.R. & R.D. Woods 1998a. Accuracy of
determining pile capacity by dynamic methods.
Proc. Seventh Inter. ConJ and Exhibition on
Piling and Deep Foundations, Vienna, 15-17June:
1.2.1-1.2.8, Rickmansworth: Westrade Group
Ltd.
Svinkin, M.R. 1998b. Discussion of 'Probability
method applied to dynamic pile-driving control' by
Liang & Zhou, Journal of Geotechnical and
Geoenvironmental Engineering, ASCE, 122(4):
319-321.
Thompson, C.D. & D.E. Thompson 1985. Real
and apparent relaxation of driven piles. Journal of
Geotechnical Engineering, ASCE, 1985, 11l(2):
225-237.
Thorburn, S. & W.J. Rigden 1980. A practical
study of pile behavior. Proc. 12th Annual Ofshore
Technology Conf , Houston.
Tomlinson, M.J. 1971. Some effects of pile driving
on skin friction behavior of piles. Proc. Institution
of Civil Engineers: 107-114, London.
Wardle, I.F., G. Price & T.J. Freeman 1992. Effect
of time and maintained load on the ultimate
capacity of piles in stiff clay. Piling: European
practice and worldwide trends, Proc. Institution of
Civil Engineers: 92-99, London: Telford.
York, D.L., W.G. Brusey, F.M. Clemente &
S.K. Law 1994. Setup and relaxation in glacial
sand, Journal of Geotechnical Engineering,
ASCE, 1994, 120(9): 1498-1513.

40
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 1503

Set-up considerations in wave equation analysis of pile driving

C.W.Cho
Visiting researcher: University of WesternAustralia, Perth, W A ,Australia
M.W. Lee
Piletech, Consulting Engineers, Seoul, Korea
M.ERandolph
University of WesternAustralia, Perth, WA., Australia

ABSTRACT: Set-up effects should be accounted for to obtain reliable estimates of driving resistance using
WEAP analysis because both the bearing capacity of driven piles, and other dynamic parameters, show a tert-
dency to change with time. Unfortunately, there are no reliable methods that take set-up effects into consid-
eration in wave equation analysis, even though much research has been conducted in this area. Therefore it is
desirable to propose new soil input parameters to account for time dependent characteristics. For this purpose,
statistical analyses of dynamic measurements for both ‘end of driving’ and ‘restrike’ conditions have been
undertaken. The paper recommends new parameters that take account of set-up effects for wave equation
analysis, and quantifies the reliability of the recommended values by means of statistical analyses on inde-
pendent pile tests. It is shown that the recommended quake and damping parameters are more reliable than
existing suggestions for wave equation analysis of driven piles.

I INTRODUCTION trols the static stiffness, while the damping coeffi-

WEAP (wave equation analysis of pile driving: GRL
(1996)) is based on the ideas developed originally by
Smith (1960) to predict drivability and bearing ca-
pacity at the time of driving. WEAP, however, has
also been used to predict the bearing capacity after
driving. In dynamic pile loading tests, the former is
referred to as an EOID (end of initial driving) test,
while the latter is referred to as a restrike test. Even
though the pile-soil system changes due to time de-
pendent effects such as consolidation, single values
of soil parameters are input in WEAP. This means
that the pile capacity calculated by WEAP can be
different from the actual value.
The prediction of driving resistance using WEAP
has been found to be reasonably reliable at EOID
using parameters suggested in the literature. How-
ever, due to the changes in the ground condition
around piles after driving, the bearing characteristics
also change with time and there is much greater un-
certainty regarding the bearing capacity, and other
dynamic parameters, at restrike conditions. In spite
of this, there are many cases reported where the ca-
pacity calculated by WEAP, without considering the
time effect, was compared directly with that meas-
ured from a static pile load test carried out some
time after the initial driving.
In WEAP, soil properties are represented as unit
values of shaft or base resistance, and quake and
damping coefficients; quake is a parameter that con-
cient is a parameter for representing dynamic en-
hancement of the soil resistance.
Hunt et al (1988), York et a1 (1994) have reported
good correlations between computed and measured
pile capacities by using dynamic measurements.
However, those methods have some limitations in
practical applications because it is possible to use
them only after carrying out dynamic tests. Mean-
while, Thendean et a1 (1996) did not clarify the time
lag between the restrike tests and static tests, and
used the same damping and quake properties for
WEAP analysis at EOID and restrike. In their study,
the change of set value was modelled only by
changing the unit values of soil resistance at EOID
and restrike in WEAP. At worst, it is often reported
that the bearing capacity deduced at EOID is consid-
ered as the long term static capacity in practice.
Svinkin and Woods (1998) have suggested a
variation of damping coefficient with time through
back calculation by WEAP to account for time de-
pendent characteristics of the soil parameters (see
Fig. 1). In their study, the damping coefficient was
considered, but the quake was disregarded, essen-
tially implying that only the dynamic soil properties,
and not the static parameters, are affected by time.
However, according to experience in the field, the
quake actually varies significantly with time.
To use Svinkin and Wood’s method in practice,
the function of damping with time has to be deter-
mined in advance, since different curves of the type

41
shown in Fig. 1 were obtained for different pile and
soil conditions. Thus dynamic measurements are
needed for the particular site, but, WEAP is mostly
used in preliminary design where dynamic data are
unavailable. Consequently it is desirable to suggest
new soil input parameters to account for time de-
pendent characteristics. To do so, it is first necessary
to analyze and appraise the soil parameters measured
for the same piles at both EOID and at any time after
driving.
Figure 1. Change in damping coefficients with time
(after Svinkin and Woods, 1998)
Statistical analyses of dynamic measurements
(quake, damping coefficients, set-up factor) are pre-
sented in this paper, which recommends new pa-
rameters to consider time effects in wave equation
analysis and confirms the reliability of the values
recommended. Although time effects may be di-
vided into two types, set-up and relaxation, only the
former will be dealt with in this paper.
2 DATA AND ANALYSIS METHOD
Dynamic pile loading tests for the same pile at dif-
ferent times after driving were carried out at 28 sites.
To avoid bias in the statistical analysis, a maximum
of only 2 data points were used from any one site.
For the determination of the set-up factor, the ratio
of bearing capacity at restrike to bearing capacity at
EOID, 54 piles from 28 sites were investigated and
analyzed, while to estimate time effects on quake
and damping parameters, 46 piles were analyzed due
to the lack of some data. The piles comprised 18
steel pipe piles, 23 concrete piles and 5 H-piles. The
diameters of piles ranged from 3 0 mm to 600 mm.
Time lags between EOID and restrike ranged from
1 day to 30 days (averaging about 8 days). Further
details of the data are given by Cho (1998).
Data for statistical analysis were drawn from
CAPWAP (Case Pile Wave Analysis Program,
42
GRL,1996) analyses of dynamic measurements at
EOID and restrike. Only the datum at the final re-
strike was used in cases where multiple restrike tests
were undertaken. Since many of the restrike tests
were not able to fully mobilize the pile capacity due
to limitations of either hammer capacity or material
strength of the piles, the set-up factors were deter-
mined from the larger of (i) the ratio of failure load
(interpreted from the static response computed using
CAPWAP, and using Davisson’s criterion for fail-
ure) at restrike to that at EOID; or (ii) the ratio of the
value (toe resistance at EOID + shaft resistance at
restrike) to the bearing capacity at EOID.
Because of the lack of data, it was not possible to
account for each different pile type, apart from the
toe condition (closed or open-ended). Soils for the
pile base response were divided into 4 types: cohe-
sionless soil (S), cohesive soil (C), silty sand (SM),
and sandy gravel (SG); while soil for the shaft pa-
rameters were divided into 3 types: cohesionless soil
(S), cohesive soil (C), and silty sand (SM). The silty
sands for the pile shaft responses were mostly
weathered residual soils, which are commonly used
as the bearing stratum. The soil type for the shaft
response was based on the soil layer that provide the
majority of shaft resistance for the pile.
CAPWAP is a trial and error method of signal
matching, which is known to lead to subjective, non-
unique, sets of parameters. As such, it is possible to
fail to discover any distinct trends in the parameters,
since each parameter determined from CAPWAP is
influenced by the other parameters. In order to
minimize operator-dependent trends in the statistical
analyses, only data with high matching quality (MQ
in CAPWAP) were analyzed.
In order to confirm the validity of the parameters
deduced from the CAPWAP analyses, they were
used in WEAP analyses for real sites (with new pile
data independent of the original set of data). Wave
equation analyses were performed for a total of 24
pile tests at 6 sites (4 tests per site).
In wave equation analysis, adjustments of the
maxima of force (FMX) and the transferred energy
(EMX) on the basis of the value analyzed by CAP-
WAP were made in order to increase the accuracy of
the compared data. The adjustment was considered
sufficient when the calculated and measured EMX
values agreed to within 10 %, following the guide-
lines in FHWA (1996). Other input parameters (ex-
cept the soil parameters) were based on the CAP-
WAP data and test conditions in the field. New soil
parameters were then used as the input data for
WEAP to match the measured drivability.
For the reliability analysis of the results of wave
equation analysis, using the new parameters drawn
from the statistical data, a log-normal probability
function was used. The log-normal function was
chosen to allow consideration of the non-
symmetrical Characteristics and distribution of the
ratios. The calculated capacities by WEAP were di-
vided by the measured capacity, and then the ratios
were evaluated by the log-normal function. From the
log-normal function, mean standard deviation, and
coefficient of variance were evaluated. The coeffi-
cient of variance is defined as the standard deviation
divided by the mean.

3 RESULTS AND DISCUSSION

The set-up factors for shaft resistance, determined

from the statistical analysis, are shown in Fig. 2 for
the 3 soil types. In Fig. 2, m, SD, and N stand for
mean, standard deviation and number of data, re-
spectively. The set-up factor increases as the soil
particle size decreases, and their mean values are 1.5
in cohesionless soils and silty soils, 2.5 in cohesive
soils. Although these values are larger than those in
WEAP (GRL, 1996) or FHWA (1996) as shown in
Table 1, they are still likely to be conservative, be-
cause most restrike tests failed to mobilize the full
(ultimate) bearing capacity of the piles.

The drilling logs for the data with high set-up

factors in Fig. 2 were analyzed. It was found that the
sites comprised mostly alluvial deposits. As a result,
it is suggested that set-up effects are influenced by
not only the particle size but also the formation his-
Figure 2. Deduced set-up factors for shaft resistance tory. Cho (1998) reported that mineral formation and
relative density highly influence set-up effects in
alluvial deposit soils even if they are sandy soils.
Table 1. Shaft set-up factors (WEAP, FHWA) Fig. 3 shows the change of shaft quake with time
and soil type, and also the values suggested in
Predominant soil type Set-up factors WEAP. There is no distinct trend in the change of
in the shaft (conservative) shaft quake, which lies in the range 2.0 i0.2 mm.
Clay 2.0 This is marginally lower than the value of 2.5 mm
Silt -Clay 1.o suggested in WEAP.
Silt, Sand - Clay 1.5 Fig. 4 shows the change of shaft damping coeffi-
Sand - Silt, Fine Sand 1.2 cients (Smith damping coefficient) with time and
soil type, and also the values suggested in WEAP. It
Sand, Sand - Gravel 1.0 may be seen that again there is no distinct trend in
the change of shaft damping coefficient. The range
Table 2. Ratio of velocity at EOID and at restrike of shaft damping coefficient represents about
0.65 i0.1 s/m regardless of time since driving, and
VMX EMX Toe Toe is higher than the values suggested in WEAP. This
ratio ratio Veloc- Energy may be because the soil type in the field is not as
simple as in the WEAP classification (and generally
- F D D includes some clay fraction). A possible reason for
the lack of time effects on the damping coefficient is
that the appropriate value is closely linked to the pile
Remarks VMX EMX Vtoe Etoe maximum velocity (VMX), which for the shaft is

43
dictated by the maximum energy (EMX). These
vary little between EOID and restrike (Table 2). As
such, it is inferred that the damping coefficient for
the shaft, unlike that for the pile toe, is not highly in-
fluenced by set-up effects.
Fig. 5 shows that the toe quake at restrike is
smaller than that at EOID, implying that the quake
decreases with time. This trend is particularly appar-
ent in cohesive soils, because set-up effects are more
significant in fine-grained soils (see Fig 2). From
those results, it seems that Svinkin and Wood’s
method has some limitations in accounting for set-up
effects. According to Fig. 5, there is no distinct trend
in the relationship between soil type and quake. As
shown by Davisson (1973), the quake is also related
to diameter of pile and shape of toe, etc, and cannot
be expressed in terms of soil type only.
Therefore, the change of the quake with pile di-
ameter, D, was explored through regression analysis,
as indicated in Table 3. Table 3 shows that although
the correlation factor (r2) is low, the general trend is
apparent, and that the quake (D/52 at EQID, D/94 at
restrike) is much higher than the value suggested in
WEAP (D/120). In particular, there is a large differ-
ence in the value at EOID. From the analysis, it is
concluded that the quake decreases with time and its
change is large in clay soils. The toe quake for
closed-ended piles is also smaller than that for open-
ended pile, which is presumably related to the com-
pressibility of the soil plug.
Fig. 6 shows that the toe damping coefficient in-
creases with time. This is assumed to be linked to
the reduction in particle velocity and transmitted
energy at the pile toe, as indicated in Table 2. Thus,
the non-linear nature of real damping means that a 4 RECOMMENDED NEW PARAMETERS AND
higher (linear) coefficient is needed for low veloci- THELR RELIABILITY
ties than for high velocities. Fig. 6 also shows that
the increase in the toe damping coefficient is small As shown in the previous section, there are signifi-
in fine-grained soils, perhaps due to limited consoli- cant differences between the analysis results and the
dation in the time between EOID and restrike. Fig.6 values suggested in WEAP. Therefore, new pa-
shows that the toe damping coefficients at EOID lie rameters for both EQID and restrike, have been rec-
in the range about 0.25 f 0.02 s/m, and the values ommended in the light of the statistical analyses, as
are smaller than that suggested in WEAP. detailed in Table 4.

Table 3. Relation of diameter of pile and the quake

All data

44
 The dynamic parameters used in WEAP are not
inherent properties of the soil but changeable ones
with time or test conditions. Consequently, it is in-
evitable that suggested values will not represent ac-
curately a given set of conditions in the field. Previ-
ous suggestions for the dynamic parameters in
WEAP have been given by Reese et al (1964) and
Coyle et a1 (1970). From this viewpoint, the recom-
mended values in Table 4,which are based on reli-
able data from various sites, should prove reasonable
for application in WEAP. However, it should be
emphasized that the recommended values are at best
average values for the sites in question and are
strictly applicable only over the relatively limited
range of pile diameters (76 % between 400 mm and
500 mm).
The set-up factors in Table 4 are the averages of
the data excluding data beyond 1 standard deviation
from the mean in the statistical analyses. Consider-
ing that bearing capacities at restrike were mostly
determined from dynamic tests where the full pile
capacity was not mobilised, the set-up factors are re-
garded as conservative values.
It is interesting that the suggested values in WEAP
are more similar to those recommended at restrike in
this study rather than ones at EOID, although WEAP
was originally developed for drivability analysis at
EOID. The recommended values in Table 4 can be
appropriately adjusted with site conditions, and are
particularly useful for the determination of bearing
resistance at restrike.
To analyze the reliability of the recommended pa-
rameters, a comparison was performed between key
predictions (such as pile capacity) calculated by
WEAP with the parameters recommended in this
study and those calculated by WEAP using parame-
ters suggested in the WEAP manual.
The ratio of the values calculated by WEAP to the
measured ones using CAPWAP were then obtained.
Finally the ratios were used for the reliability analy-
sis using a log-normal probability function.
The results of the reliability analysis using a log-
normal probability function are shown in Fig. 7 to
Fig. 10. In those figures, each initial stands for as
follows:
RI : mean,
SD : standard deviation,
CV : coefficient of variance (=SD/M),
ES : WEAP calculation with the parameters
suggested in WEAP at EOID,
RS : WEAP calculation with the parameters
suggested in WEAP at restrike,
ER : WEAP calculation with the parameters
recommended in this study at EOID,
RR : WEAP calculation with the parameters
recommended in this study at restrike.
Function value : probability density distribution
Fig. 7 and Fig. 8 show that the capacity ratios
based on the recommended parameters in this study
are closely grouped around unity, while those based
on parameters suggested in WEAP show a much
broader range. This suggests that the bearing ca-
pacities predicted using the recommended parame-
ters are more reliable than those using the suggested
ones in WEAP.
From Fig. 9 and Fig 10, it is found that there are
some differences in the predicted and measured
maximum force, FMX, using the suggested pa-
rameters in WEAP, while the new parameters pro-
posed here give improved accuracy. Either set of
parameters lead to good prediction of maximum en-
ergy, EMX.

Table 4. Recommended parameters for WEAP

I EOID I Restrike I

2.0

Soil with
2.0
~
0.65
0.65 ~
DnO
D/80 ~ 1:; ~
( .O)*
* : (1

2,0 o.65 Df70 0.25 2.0 0.65 D/180 0.5

Gravel

Soil
I 2.0 I 0.65 I D/60 I 0.25 2.0 0.65 D/90 0.5 1.5

1.2 times 1.2 times

Remarks smaller in smaller in
open pile open pile
* means the values suggested in WEAP

45
 5 CONCLUSIONS

The bearing capacity of piles after driving changes

with time. Set-up effects need therefore to be con-
sidered in order to deduce the bearing capacity from
restrike data using WEAP. A statistical analysis of
data from 46 piles tested at initial driving (EOID)
and restrike indicates that there are significant dif-
ferences between the parameters deduced from this
study and those suggested in the WEAP manual.
Accordingly, new soil parameters (damping, quake)
for wave equation analysis at EOID and restrike are
recommended in this paper.
WEAP analyses were carried out using both the
recommended parameters and those previously sug-
gested and their results were compared with the
measured values through reliability analysis. It was
found that the new parameters recommended in this
paper showed higher reliability compared with those
suggested in the WEAP manual.

REFERENCES

Cho, C. W.1998. A study on the time dependent capacity gain

of driven piles. Ph. D Thesis: Hanyang University. Seoul.
Korea.
Coyle, H.M. & Gibson, G.C. 1970. Empirical damping con-
stants for sand clay. Journal of the Soil Mechanics. and
Foundation Division, ASCE, Vol. 96. SM3 : 949-965.
FHWA. 1996. US FHWA Dpt. of Transportation. Design and
construction of driven pile foundation. Workshop Manual :
16.1 - 17.70.
GRL Associates, Inc. 1996. GRLWEAP/CAPWAP User Man-
ual
Hunnigan, P. 1984. Large quake development during driving of
low displacement piles. Proc. of the 2nd Int. Conf. on the
Application of Stress-Wave Theory to Piles : 126-133.
Hunt, S.W. & Baker, C.N. 1988. Use of stress-wave measure-
ments to evaluate piles in high set-up conditions. Proc. of
the 3rd Int. Conf. on the Application of Stress-Wave Theory
to Piles. BiTech Publisher. Ottawa : 689-705.
Parmar, H. & Brown, D. 1996. Comparison of dynamic and
static of evaluating static pile capacity. 6th Int. Conf. and
Exhibition on Piling and Deep Foundations. Bombay : 4.6.1
- 4.6.6.
Smith, E.A.L. 1960. Pile driving analyses by the wave equa-
tion. Jour. of the Soil Mechanics and Foundation Division.
ASCE. V01.86 : 35-61.
Thendean, G., Rausche, F., Svinkin, M. & Likins, G. 1996.
Wave equation correlation studies, Proc. of the 4th Int.
Conf. on the Application of Stress-Wave Theory to Piles :
144-162.
Reese, J.L. & Forehawd, P.W. 1964. Prediction of pile capacity
by the wave equation. Journal of the Soil Mechanics and
Foundation Division. ASCE. Vol. 90. SM2 : 1-25.
Svinkin, M.R & Woods, R. D. 1998. Accuracy of determining
pile capacity by dynamic methods. 7th Int. conf. and Exhi-
bition on Piling and Deep Foundations : 1.2.1-1.2.8.
York, D.L., Brusey, W. G., Clemente, F.M. & Law, S.K. 1994.
Setup and relaxation in glacial sand. Journal of Geotechni-
cal Engineering. ASCE. 120(9) : 1498-1513.

46
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) K3 2000 Balkema, Rotterdam, ISBN 90 5809 150 3

rivability and performance of model piles driven into cemented

calcareous sand

D. Bruno, M. E Randolph, C.W.Cho & H.A. Joer

Ceiztrefor Offshore Focindutioiz Systems, Uiziversityof Western Australia, Perth, W A ,Australia

ABSTRACT: This paper reports the results of a centrifuge model study of pile drivability in cemented cal-
careous sand. The tests were carried out at a scaling ratio of 1 :150, on piles of equivalent prototype diameter
1.42 m,driven to depths of 20 to 24 m into cemented sand with cone resistances ranging from 14 to 60 MPao
Two samples were prepared, one with three layers of gradually increasing cone resistance, and one with a high
strength intermediate layer approximately 2 pile diameters thick. Different pile toe configurations were ex-
amined ranging from open-ended (with no change of pile section near the toe) to closed-ended with a conical
tip and either flush with the outer pile shaft, or slightly oversize. The effect of the different toe configurations
on the drivability was examined. It was found that pile refusal occurred in the deepest (most highly cemented)
layer, and that drivability became marginal for the closed-ended piles in the intermediate strong layer (in both
cases, cone resistance was around 40 MPa). The hard driving was consistent with high measured static ca-
pacities of the piles, and the observed drivability was found to be consistent with results from a numerical
drivability study, leading to recommended soil parameters for cemented calcareous sediments.

1 INTRODUCTION of open and closed-ended piles in cemented calcare-

ous sands. Two different stratigraphies were inves-
The low shaft capacity of driven open-ended piles i n
tigated, one with three layers of gradually increasing
calcareous soils has led to alternative construction
strength, and one with an intermediate strong layer.
techniques such as drilled and grouted, or grouted
driven, piles. An alternative approach, which may Three different pile toe configurations were tested,
open and closed-ended, and with either a flush or
also be useful for grouted driven pile construction, is
oversize toe closure.
%cs drive closed-ended piles (de Mello et al, 1989).
GIssed-ended piles have rarely been driven offshore,
because of concerns of refusal. In calcareous soils,
2 EQUIPMENT AND PILE DETAILS
the main concern is the presence of intermediate ce-
mented layers (prior to the founding stratum) that
The model tests were carried out on the geotechnical
could cause premature refusal for a closed-ended
centrifuge at UWA, an Acutronic Model 661 with a
pile.
40 g-tonne capacity and platform radius of 1.8 m
Grouted driven piles, where post-installation
(Randolph et al, 1991>.
grouting is carried out on driven piles (Barthelemy et
The model pile tests were conducted at an accel-
al, 1987), are an economically attractive alternative
eration of 150 g using a combined pile-driving and
‘es drillled and grouted piles because of the much static loading actuator (de Nicola and Randolph,
shorter construction period, While the original tech-
1994). Figure 1 shows a photograph of the pile-
nology for grouted driven piles was based on open-
driving actuator mounted on the centrifuge strong-
ended piles, with grouting conduits pre-installed in-
box, prior to a pile test. The mechanical driving ap-
side each pile (Rickman and Barthelemy, 1988), a
paratus is in the form of a detachable carriage that
closed-ended configuration would allow greater reli-
houses the driving ram and moves down with the
ability and potential for remedial actions.
pile during driving. A small potentiometer measures
The potential advantages of driven closed-ended
the gap between the ram (in its retracted position)
piles, either grouted or ungrouted, must be weighed
and the pile cap, and a feedback loop is used to con-
against the higher risk of premature refusal, com-
trol the position of the carriage in order to maintain a
pared with open-ended piles of a similar diameter.
specified drop-height. Table 1 summarises the key
The work reported in this paper was aimed at as-
mechanical properties of the system.
sessing, through physical model tests, the drivability

47

Figure 1 Model pile-driving hammer on strong-box
Table I . Key properties of model pile-driving system
Mechanical property Range
Vertical displacement 240 mm
Horizontal displacement 150 mm
Hammer driving frequency 1 -20Hz
Ram drop height 0-20mm
Ram mass 37 or 54 g
Static pile test driving speed 0-3mmh
Maximum vertical driving load 6kN
The model pile was fabricated from a thin-walled
cylindrical stainless steel section with the outer sur-
face roughened by sandblasting. The main section
of the pile, 200 mm long has internal and external
diameters of 8.5 and 9.5 mm respectively. An addi-
tional instrumented section, 60 mm long, has an in-
creased external diameter of 10.5 mm. All tests
were conducted at a nominal acceleration level of
150 g, so that the 9.5 mm diameter of the model pile
represents a prototype pile diameter of 150 x 9.5 =
1425 mm (or 1.425 m). The pile was instrumented
with a single pair of half-bridge strain gauges located
230 mm from the tip of the pile, which ensured they
remained above the sample surface even at the deep-
est pile penetrations. These gauges provided the
measurement of total bearing load in the pile during
static testing as well as the transient stress-wave
force during dynamic testing.
The pipe pile could be fitted with two different
conical driving shoes, rendering the piles closed-
ended. The driving shoes, shown in Figure 2, are
made from stainless steel and were sandblasted to
the same degree as the model pile. The 9.5 mm di-
ameter shoe is flush with the model pile - these piles
are referred to as CE (closed-ended), whilst the piles
fitted with the 10.5 mm shoe are referred to as CEO
(closed-ended oversize). The open ended piles (no
shoe attached) are referred to as OE.
48
Figure 2 Conical tips for closed-ended piles
A Hopkinson Bar, in the form of a strain-gauged
PVC strip attached at the pile head, was used to
measure the transient stress-wave velocity during the
dynamic tests, (Bruno and Wandolph, 1998). This
arrangement obviates the need for miniaturizing an
accelerometer, as would be used in the field. The
Hopkinson Bar can be seen attached to the pile in
Figure 1 . Essentially, the strip acts as a transmission
line whereby the strain in the strip, induced by the
movement of the pile head, is converted to a veloc-
ity. This conversion is only valid for uni-directional
waves. Hence the strip is made from PVC, which
has a low wave-speed, and is made sufficiently long
so that it is able to transmit several return waves
from the pile toe before receiving reflected waves
from the end of the strip.
3 S O L CONDITIONS
The model tests were carried out in calcareous silt
dredged from the sea bed on the North-West Shelf of
Australia. The material was dried in an industrial
kiln and sieved to give an effective particle size
range of 0.1 to 0.4 mm, with a mean particle size of
just under 0.2 mm.
The dry soil was mixed with 4 to 10 % (by
weight) of Portland cement, and then with water to
give a water content of 30 %. The slurry was mixed
for a further 30 minutes before being transferred to
the strongbox and placed under a consolidation
press. A 100 kPa surcharge was applied to the soil
surface for around I hour. This process was re-
peated for each layer of cementation, with new lay-
ers being poured on top of the pre-consolidated bot-
tom layers. The sample was then sealed and left for
at least 7 days in order for the cement to cure.
Cement contents of 4, 6 and 10 % were used, to
give nominal unconfined compressive strengths of
250, 500 and 1000 kPa, and corresponding cone re-
sistances, qc, of 15, 27 and 50 MPa. In practice,
some variation from the above values of qc was
found, with the second of the two samples have
somewhat higher strengths than the first.
The nominal layers depths (at equivalent proto-
type scale) were 0 - 1 I m (4%), 11 - 22 m (6 %)
and 22 - 36 m (10 %) for Sample 1, and 0 - 11 m
(4%), 11 - 1 4 m ( l O % ) , 1 4 - 2 4 m ( 6 % ) a n d 2 4 -
34 m (10 %) for Sample 2. Four cone penetration
tests were carried out ‘in-flight’ in each sample, us-
ing a 7 mm diameter cone. These showed excellent
repeatability, and gave the average cone profiles
shown in Figure 3. As may be seen, the ranges of
cone resistances for the different cement contents are
13 to 18 MPa, 21 to 30 MPa and 35 to 60 MPa re-
spectively (although it appears that there may have
been some shortcomings in preparing the base layer
for Sample 1).
Figure 3 Average profiles of cone resistance
4 TESTING PROCEDURES
The standard ‘strongbox’ (or soil container) used on
the UWA centrifuge is an aluminium segmented
box, with internal dimensions 390 mm wide by 650
mm long and 325 mm high. The total soil depth in
each sample was between 240 and 250 mm, and six
driven pile tests were carried out in each sample,
(two of each type) down the centre-line of the box.
The minimum pile spacing between tests was 7 di-
ameters, and a minimum of 14 pile diameters was
maintained from the end-walls of the box.
For Sample 1, it was intended to perform dynamic
and static tests in each of the three different cemen-
tation layers, but this was only achieved in the first
two layers owing to premature refusal before enter-
ing the deepest layer. For Sample 2, the intention
was to test the piles at a penetration of 1 pile diame-
ter, d, past the top 10 % cement layer, as well as at
an ultimate penetration into the deepest layer.
Again, however, premature refusal of some of the
piles frustrated this plan, as described later.
All piles were driven into the soil samples using
the model pile driving actuator, with a blow fre-
quency of 10 Hz and drop heights ranging from
10 mm initially, to 17 or 18 mm at deeper penetra-
tions. Dynamic tests were performed prior to static
loading, usually at a drop height of 17 mm.
49
The model piles were driven continuousPy be-
tween one test penetration and another. The blow
count for each pile was recorded by monitoring the
number of times the hammer was fired between each
sampling point. Two ram masses were used in the
centrifuge tests. In Sample 1, the first two open-
ended piles were installed with a 37 g mass; this was
later replaced with a 54 g mass, which was used
throughout the remainder of the testing programme.
In the centrifuge model, the scaling law for mass
is given by:
where N is the g level in the centrifuge tests. Al-
though an acceleration of 150g was applied to the
soil samples, the g level is affected by the distance
from the axis of the centrifuge. The hammer was
initially positioned around 325 mm above the sam-
ple surface, moving down with the pile to a final
height of around 175 mm. Since the acceleration
field varies linearly with the radius of rotation, the g
level is considerably less at the position of the ac-
tuator. Hence, the actual g levels at the initial and
final actuator positions are approximately 1104 g and
122 g respectively. This corresponds to an average
prototype ram mass of 94 tonne (for the 34 g mass)
and 137 tonne (for the 54 g mass), and average rated
hammer energies in the range 980 kJ to 3000 k9.
Thus the model hammer is equivalent to a Menck
8000 or 11 2500 offshore hammer.
After installation of the piles, static load tests
were conducted either in tension (T) first or com-
pression (C), up to vertical displacements of ap-
proximately 40 % of the pile diameter (0.4d). The
tests were a11 performed at a (model) rate of 0.05
m d s e c , which ensured a drained condition around
the pile. The static pile capacity, Qs,was taken as the
load measured at a pile head displacement of O.ld,
although this was usually close to a distinct yield
point and dramatic reduction in loading stiffness.
The transient stress-wave force, IF; and velocity, v,
were measured throughout pile driving as well as
during dynamic tests carried out prior to static iload-
ing, (Bruno & Randolph, 1998’). The Case analysis
was used to estimate pile capacity (or mobilised soil
resistance), Rs, during installation, as given by the
following expression:
(2)
where Z1 and Z 2 are the impedances of the upper (in-
strumented) and lower (embedded) pile sections and
the other notations are standard terminology. This
modified form of the Case formula has been found
to give accurate estimates of soil resistance (Bruno
and Randolph, 1999), although rigorously an addi-
tional correction should be made to the returning
stress-wave, (F - Z I V ) ~ + ~toUallow
~ , for downward
travelling waves reflected from the interface between
the upper and lower pile sections.
The deduced mobilised soil resistance, R, is sen-
sitive to the value adopted for the damping parame-
ter, J,. However, here it was possible to calibrate
this parameter from the static tests, and a value of
j, = 8.B provided the best estimate to the measured
static capacities.

5 RESULTS

5.1 Driving Characteristics

Typical driving records are shown in Figures 4 to 7
for an open-ended and (flush) closed-ended pile in
Samples 1 and 2. In addition, the driving record for
one of the over-sized closed-ended piles in Sample 2
is shown in Figure 8, as it reveals an interesting phe-
nomenon associated with the strong layer.
In Sample 1, the open-ended pile was driven with
the smaller ram mass, and thus lower energies and
showed much easier than for the closed-ended pile

(comparing Figures 4 and 5). The easier driving of

the open-ended piles is more apparent for Sample 2,
where both open and closed-ended piles were driven
with the same ram mass. Figures 6 and 7 show that
the closed-ended pile essentially reached refusal at
the high strength layer between 1 1 and 14 m.
Eventually, the pile was forced through the layer, but
only after a very large number of blows, at a rate of
advance that would be impractical in the field.
(Note that the velocity monitoring instrumentation
failed in the latter stages of penetration of the closed-
ended pile in Sample 2, hence the lack of informa-
tion on delivered energy.)
Comparison of the soil plug monitoring for the
open-ended piles showed that the plug advance was
very similar for each pile, with an incremental filling
ratio of about 0.8 over most of the driving record.
This accounts for the easier driving of the open-
ended piles.
The breakthrough of the closed-ended pile in
Sample 2, which occurred only after the tip was fully
through the strong layer, was also seen for one of the
over-sized closed-ended piles, but this time while the
tip was still within the strong layer (see Figure 8).

50
Figure 8 Driving record for over-sized closed-ended pile,
Sample 2
Interestingly, however, the pile refused mid-way
through the subsequent (6 % cement) layer. The im-
plication is that the breakthrough halfway through
the strong Payer must have carried a large fragment
of that strong layer down into the weaker zone be-
IOW, eventuafly leading to refusal of the pile.
5.2 Mobilised Soil Resistance and Pile Capacity
Profiles of mobilised soil resistance for the different
pile types are compared with pile capacities deduced
from static load tests in Figures 9 and 10, for Sam-
ples 1 and 2 respectively. The mobilised soil resis-
tance has been obtained from Case analyses of the
stress-wave data, using a damping parameter of
3 , = 0.2.
Agreement between the dynamic and static esti-
mates of pile capacity is reasonably good at low ca-
pacities, but the dynamic analyses underestimate the
measured values at high capacities. This may be at-
tributed to the very small movements with each dy-
namic blow once the pile capacity exceeds 25 to
30 MN. It should be noted that, in general, the de-
livered hammer energy was higher for Sample 2 than
for Sample 1, as the pile-driving hammer was serv-
iced in the intervening period.
6 PILE DRIVABILITY
Drivability studies have been undertaken using
WEAP (GRLWEAP, 1997) in order to ascertain ap-
propriate soil parameters to use for cemented cal-
careous soils. The hammer has been modelled as a
Menck 12500 hammer delivering energies consistent
with average values shown earlier in Figures 4 to 8.
Since WEAP cannot model the soil resistance
within open-ended piles explicitly, the open-ended
pile has been modelled by increasing the specific
weight of the pile over the plugged length of the pile,
as suggested in the WEAP manual.
51
Figure 9 Profiles of dynamic and static resistance, Sample 1
Figure 10 Profiles of dynamic and static resistance, Sample 2
Shaft friction was taken as 0.43 % (open) and
0.73 96 (closed) of the cone resistance, while quake
and damping values of 2.0 mm and 0.65 s/m have
been assumed along the shaft. At the pile base, the
resistance has been taken 40 % (open - assumed to
act over the complete base) and 60 % (closed) of the
cone resistance, q,, while quake and damping Rave
been taken as d50 = 30 mm) and 0.25 s/m respec-
tively (consistent with Cho et al, 2000).
The resulting profiles of pile capacity are shown
in Figure 1 1 , and are reasonably consistent with the
measured capacities shown in Figure PO. Results
from the drivability analyses are shown for an open-
ended and closed-ended pile for the Sample 2 stra-
tigraphy in Figures 12 and 13.
For both piles, good agreement is observed with
results shown earlier in Figure 6, with only slight in-
crease i n blow-count as the open-ended pile passes
through the intermediate high strength layer, but the
closed-ended pile effectively refusing in the layer.
The predicted reduction in blow-count just below
this layer is less dramatic than measured, partly due
to the coarse depth spacing of the drivability data,
but possibly also reflecting overprediction of the
shaft friction in the strong layer.

7 CONCLUSIONS
This paper has described a study of the drivability of
open and closed-ended piles into cemented calcare-
ous soils. The study was undertaken with a view to
assessing the feasibilityof drivingclosed-ended piles
in such soils, with or without post-installation
52
grouting, in view of the low shaft resistance of
driven open-ended piles in calcareous sediments.
As expected, closed-ended piles were found to be
much more difficult to drive than open-ended piles,
although both pile types met refusal in thick layers
of cemented material with cone resistance in excess
of 35 to 40 MPa. It was found that open-ended piles
could be driven relatively easily through a layer of
strong material only 2 pile diameters thick. Closed-
ended piles essentially refused in such a layer,
breaking through only after more than 8000 blows.
Drivability studies showed that the measured
driving characteristics could be simulated using con-
ventional values of quake and damping parameters,
unit shaft friction of 0.4 (open) to 0.7% (closed) of
the cone resistance, qc, and toe bearing resistance of
0.4qc (open) to 0.6q, (closed). These values are rea-
sonably consistent with field experience reported by
de Mello et a1 (1987), and give confidence in the va-
lidity of the centrifuge model study.
ACKNOWLEDGEMENTS
This work was undertaken for Heerema Marine
Contractors as part of research into grouted driven
piles, and is presented with their kind permission. It
forms part of the activities of the Centre for Offshore
Foundation Systems, established and supported un-
der the Australian Research Council’s Research
Centres Program.
REFERENCES
Barthelemy, H., Martin, R., Le Tirant, P., Nauroy, J.F. & Cipri-
ano de Meideros, J. 1987. Grouted driven piles: An eco-
nomic and safe alternate for pile foundations. Proc. 19th
annual OffTech. Con$ (OTC), Houston, Texas, 421-436.
Bruno, D. & Randolph, M.F. 1998. Dynamic testing of driven
piles in dense sand. Proc. International Conference CEN-
TRIFUGE 98 (ISSMFE Centrifuge 98), Tokyo, 5 19-524.
Bruno, D. & Randolph, M.F. 1999. Dynamic and static load
testing of model piles driven into dense sand. J. Geot. Eng,
Div, ASCE, 125( 1 I), 988-998.
Cho C.W., Lee, M.W. and Randolph, M.F. 2000. Set-up con-
siderations in wave equation analysis of pile driving. P roc.
6’” Itit. Con$ On Application of Stress-Wave Theory to
Piles, Sao Paulo.
De Mello, J.R.C., Arnarai, C.D.S., Maia de Costa, A., Rosas,
M.M., Decnop Coelho, P.S. & Porto, E.C. 1987. Closed-
ended pipe piles: Testing and piling in calcareous sand.
P roc. Annual O f J f o r e Technology Conf., Houston, Paper
OTC 6000.
De Nicola, A. & Randolph, M.F. 1994. Development of a
miniature pile driving actuator. P roc. Itzt. Cot$ Cetitrifuge
‘94,Singapore, 473-478.
Randolph, M.F., Jewell, R.J., Stone, K.J.L. & Brown, T.A.
1991. Establishing a new centrifuge facility. Proc. Itit. CO$
Cetirrifuge 91, Boulder, Colorado, 3-9.
Rickman, J.P. & Barthelemy H.C. 1988. Offshore construction
of grouted driven pile foundations. Proc. Int. Cot$ on Cal-
careous Seditiietits, Perth, 1 , 3 13-319.
 Application of Stress-Wave Theory to Piles, Niyama Ei Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Automatic signal matching with CAPWAP

E Rausche & B. Robinson

Goble Rausche Likins and Associates Incorporated, Cleveland, Ohio, USA
L. Liang
Pile Dynamics Incorporated, Cleveland, Ohio, USA

ABSTRACT: CAPWAP@ Provides the most reliable means of analyzing dynamic pile top force and velocity
records from the Pile Analyzer@ (PDA). This is a signal matching approach which requires that certain soil
parameters are adjusted until measured and calculated pile top variables reach a reasonable match. The
number of unknown soil parameters depends on the depth of pile penetration and therefore the computational
effort can be appreciable if the pile is long. The process can be either done in an interactive manner or
automatically with great time savings. Current practice requires that the automatic results are checked by
interactive analysis.
In an attempt to make the automatic solution reliable, several additional matching parameters have been
included in the CAPWAP model. Among these variables, the most important is the final set (inverse of blow
count) of the hammer blow analyzed. Unfortunately, since restrike data is usually analyzed by CAPWAP for
long term static capacity predictions, final set is not always accurately known. For this reason a study was
undertaken to evaluate the accuracy of the automated CAPWAP results including blow count matching
compared to the traditional approach. More than 30 cases where static load tests and restrike tests had been
performed were analyzed using the automatic procedure provided by the Windows based CAPWAP Version
1999-1. This program also calculates the total dynamic resistance (the sum of damping and static resistance)
allowing for an assessment of the ratio of total dynamic resistance to static resistance and its relationship with
soil type.

1 INTRODUCTION (Goble Rausche Likins and Associates, Inc., 1999) is

Dynamic pile testing has two distinctly different
goals: (a) monitoring the installation of impact
driven piles to avoid pile damage and assure
sufficient pile penetration for bearing capacity at the
time of installation and (b) dynamic load testing for
an assessment of the long term bearing capacity of
either a driven pile or a drilled shaft. The following
paper deals with the analysis of dynamic load test
records, i.e., force and velocity as a function of time.
Dynamic load testing requires measurement of
pile top force and velocity and therefore the pile top
displacement is also known. Because of stress wave
effects caused by the rapid loading of the pile, a plot
of measured force vs. measured displacement does
not resemble the static load-set curve. For the
calculation of the static load-set curve it is therefore
necessary to reduce the dynamic force to a static one
by removing dynamic effects of both pile and soil.
This calculation is usually done by signal matching
(Rausche et al., 1972) a process that has been
continuously improved (Mure et al, 1983, Hannigan
et al., 1987, Hussein et al., 1991). Today, CAPWAP
the most widely accepted computer program for the
calculation of the static load set curve from dynamic
test records. The latest version, a Windows
program, includes a blow count matching option.
This paper briefly describes the fundamental features
of CAPWAP and presents a correlation study, which
investigates the potential benefits of the expanded,
automated search. The correlation utilizes
information of GRL’s data base which has been
described by Thendean et al. (1996) in a paper that
discussed the performance of an earlier CAPWAP
version. The present paper also briefly investigates
the relationship between total capacity and static
capacity.
2 THE: CAPWAS PROCEDURE
With two measurements at the pile top available,
both input to and response of the pile top are known,
however, one part of the system, the soil, which
produces the response is unknown. In order to
calculate the soil properties, a so-called inverse

53
analysis has to be performed which identifies the
unknown parameters of a soil model (Figure 1).
This inverse analysis is commonly called a Signal
Matching Analysis (Balthaus, 1986, Reiding et al.,
1988), or a System Identification (Klingmuller,
1984). The solution has to be achieved iteratively:
an assumption of the unknown soil parameters is
made and tested by performing an analysis with one
of the measured quantities as a top boundary
condition. If there is disagreement between the other
measurement and its calculated counterpart the
calculation is repeated with a corrected set of soil
model parameters. Obviously, the more realistic the
soil model, the better its capability t o match the
measured quantities. On the other hand, a very
sophisticated soil model may have too many
unknowns and may not be uniquely defined by the
matching process. For that reason, the relatively
simple Smith soil model (Smith, 1960) has been
most successfully employed for pile dynamic signal
matching,
The traditional iterative matching procedure can
be summarized as follows:
1. Data Input: select a record with appropriate
energy and data quality
2. Data Check and adjustment (normally
automatic)
3. Build pile model (normally automatic)
4. Check and change resistance distribution
5 . Recheck data adjustment
6. Check damping parameters
7. Check quakes and unloading parameters
8. Find absolutely best match quality
9. Produce output
An important part of the matching procedure is the
evaluation of the match quality, i.e. quantiQing the
difference between measured and computed
quantity. In CAPWAP the match quality is the
normalized, weighted sum of the absolute values of
the differences between computed and measured
values of all analyzed time steps. Normalization is
achieved with respect to both maximum pile top
force and the number of data points. The match over
a 3 ms time period, following the first return of the
stress wave from the pile toe, is given a double
weight because of its importance for total capacity
determination.
CAPWAP can either be used in an interactive
mode or automatically. The automatic procedure
searches for a best match using a step by step
procedure that is also recommended to the analyst
for interactive signal matching. In other words, the
automatic CAPWAP is not a standard minimization
software which would search in a more or less
random manner for a set of soil parameters that
produces a minimum difference between computed
and measured pile top variable. Experience has
shown that such a relatively mindless procedure may
lead to unsatisfactory results. On the other hand, the
54
automated CAPWAP procedure produces capacity
results that are very similar to those obtained by
experienced engineers working interactively on a
computer. On occasion, however, the automatic
method calculates an unsatisfactory resistance
distribution near the pile toe. An experienced
analyst must therefore always check the solution by
means of additional trial analyses. For the
simplification of the interactive matching task the
CAPWAP program does provide difference
minimization routines for individual soil resistance
parameters.
3 MATHEMATICAL MODEL
The pile is modeled as a series of uniform, elastic
segments, typically 1 m long, of equal stress wave
travel time. Calculations involve tracking the
upward and downward traveling waves and their
reflections where segment properties change or soil
resistance effects exist. The simplicity and strictly
elastic nature of this model is, unfortunately, a
disadvantage when modeling non-linear or non-
elastic situations such as cracks in concrete piles or
certain types of mechanical pile splices.
The effect of the soil, resisting the pile motion, is
modeled as a series of N lumped forces at intervals
not greater than 2 m which depend on pile velocities
and displacements. The parameters of this Smith
soil model are the unknown quantities that
CAPWAP must determine. In the standard analysis
situation, the displacement dependent (static)
resistance forces are represented by both a stiffness
and a capacity value. The velocity dependent
(dynamic) resistance forces are calculated using a
damping factor. For the resistance forces acting on
the shaft, soil stiffness and damping parameters are
chosen proportional to the static capacity values and
in this way, the number of shaft unknowns is kept to
a manageable N + 2 values. For the toe an
additional 3 unknowns have to be determined. To
produce a good signal match over a long record time
period several additional parameters had to be
defined. The most important ones allow for a
modification of the static soil stiffness and plastic
limit (upward directed capacity) for the rebound
phase of the pile and therefore have little or no effect
on the calculated total static pile capacity.
 Under certain conditions, particularly when the
pile set under a hammer blow is very small, the
assumption that soil resistance only depends on pile
motion becomes inaccurate because the soil motion
then has a magnitude comparable to that of the pile.
The CAPWAP radiation model helps improve the
calculated soil model for such cases by representing
the soil surrounding the pile by a mass and a dashpot
(Likns et al, 1996).
4 DATA CHECK (RECORD LENGTH AND SET
MATCH)
In earlier versions of CAPWAP, the analyzed record
length was generally set to 25 ms after the first
return of the impact wave from the pile toe. This
relatively short record length saved computer time
but did not always allow for an accurate calculation
of the final set. Today modern personal computers
and more sophisticated operating systems provide
the analyst with high computational speeds and huge
memory space at a low cost. It has therefore become
possible to economically analyze dynamic pile
records over longer time periods and to perform
many more trial analyses for more reliable results.
The longer analysis time period assures that the
calculation can be carried out until the pile stops
moving, i.e. until the pile velocity becomes zero and
the displacement has reached the final set. To be
sure, the recommendation for Pile Driving
AnaIyzerB users is a record length of 200 ms for
normal land piles. Figure 2 is the example of a pile
top force, velocity and displacement record whch
includes major vibrations after 100 ms. The record
also indicates a final displacement value that
matches the pile set or the inverse of blow count.
Ideally, the pile top set per blow is verified by
independent observation so that the double
integration of the acceleration can be checked and, if
necessary, corrected. This two point data check and
adjustment process is automatically performed in
CAPWAP by slightly shifting the zero line of the
acceleration. This zero line shift starts at impact. has
a somewhat higher magnitude for a few milliseconds
and a lower magnitude over the remainder of the
record. The magnitude of this zero line shift is only
a small fraction of the maximum measured pile top
acceleration.
5 NEW DEVELOPMENTS
With a record that correctly double integrates to final
set and with computational speeds allowing for an
analysis to final zero velocity at the end of the
record, the final pile set or its inverse, the blow
count, can be calculated for all segments of the pile.
As long as the final set is the same for several
55
Figure 2. Typical force-velocity curves and
displacement matching.
consecutive blows, this condition requires that all
points along the pile achieve the same final set equal
to that at the pile top. It is reasonable to require that
the CAPWAP signal matching process also produces
a match of calculated with observed pile top set; in
other words, the average of all the sets of all
segments equal the observed set. For lack of
computing power this requirement had not been
imposed on previous analyses. (It should be added
that the most accurate method of calculating the final
set of the pile is a residual stress analysis (RSA),
which repeats the analysis several times thereby
calculating the stresses locked into pile and soil.
This analysis method is available as an option in
CAPWAP. Regrettably, RSA is infrequently used
because of its complexity.)
6 PROGRAM PERFORMANCE
The enhanced analysis procedure is available in the
Windows based CAPWAP. To check whether or
not the blow count matching (BCM) procedure
would help improve the program performance, most
of the data presented by Likins et al. (1996) was
reanalyzed. The reanalysis involved adjustment of
the acceleration values for both a zero final velocity
and a final displacement matching the observed
blow count at a later record time. The soil model
parameters were then re-initialized to reset all soil
parameters to the CAPWAP start up conditions.
Any modifications to the pile profile were removed,
the analysis count was reset to 1 and the initialized
data file was saved. These steps assured the same
results would be obtained as if the analysis had been
started from the very beginning. With the
reinitialized data file the automatic procedure was
then run first with blow count matching and then
again without blow count matching. The analyst did
not interfere with the automatic analysis process in
any way. Capacity, computed final set and match
quality results were then subjected to a variety of
correlation studies. As in the 1996 study,
correlations were done using the ratio of CAPWAP
divided by static load test capacity. Also, the time
Table 1. Calculated CAPWAP capacity divided by static load capacity at
different time ratios
Figure 3. Log-normal probability density finction
for CAPWAP capacity
factor was introduced, i.e. time between restrike test
and pile installation divided by time between load
test and installation. Thus, a time factor less than
one indicates that the dynamic load test was
performed prior to the static test. For a meaningfd
correlation, the time factor should be close to unity.
Indeed, Figure 3 and Table 1 show that the data
marked ''close'' with time factors between 0.33 and
1.25 yielded the best correlations. For "No BCM"
the mean was .95 and the coefficient of variation .20.
Blow count matching (BCM) significantly improved
the correlation to a mean of .97 with a COV or .18.
Even the other time factor categories showed a clear
gain in accuracy and precision.
7 QUALITY OF BLOW COUNT MATCH
Computed blow counts are presented in the form
of histograms of calculated divided by observed
final set (inverse of blow count) in Figure 4, both for
"No BCM" and "BCM.". Clearly, the calculated sets
improved although, in quite a few cases they did not
change appreciably compared to those cases where
blow count match was not attempted. It is
concluded that either the observed blow count was
not accurate - and since these are all restrike tests it
would be expected that observed blow counts are
generally inaccurate - or the dynamic data, the pile
model or soil model did not accurately enough
represent the test conditions.
56
8 MATCH QUALITY
Obviously, a number of automatic CAPWAP
predictions are not satisfactory. In the data set under
consideration, one prediction was nearly twice the
static capacity and one was one half the static load
test result. Ideally, the match quality number would
reflect the reliability of prediction. In fact,
Hannigan, et al. (1 987) presented good correlations
with one exception whose match was not
satisfactory. It was therefore concluded that match
quality is an indicator of the reliability of prediction.
To firther study the relationship between match
quality and capacity prediction, Figure 5 was plotted
which is normalized capacity vs. match quality. The
cases presented were done with BCM; match
qualities were therefore slightly higher than those
achieved without BCM (BCM adds the final set
error to the quality of the signal match.) Obviously,
there is no correlation whatsoever between match
quality and capacity prediction. However, it would
be wrong to assume that match quality for a
particular data set does not matter because for each
case the program determined the best possible
match or lowest MQ value. The match quality
number for a particular case is therefore specific and
may be used to judge the reliability of only that one
data set. 7t is not possible to make a general
requirement on match quality: in one case even an
MQ = 24 yielded an acceptable result. However, it
was probably more a matter of luck that a good
correlation was achieved. In general, results with
MQ > 5 should be considered with suspicion. In all
cases the CAPWAP analyst must check the results
calculated by the automated routine and determine
whether or not additional MQ improvements are
warranted and possible.
9 PREDICTED SOIL MODEL PARAMETERS
Table 2 presents damping and quake values
calculated by either BCM or No BCM.. The mean
values of the calculated shaR damping differed little
(.74 vs .72 s/m), however, the blow count matching
procedure produced less scatter (COV .49 vs .63).
Figure 4. Computed to observed set comparison with and without blow count matching

values of approximately 2.0 (1.93 for BCM and 2.06

for No BCM) with significant scatter. This result
matches the D/60 recommendation currently in
GRLWEAP for certain types of soils.

Table 2. Dvnamic soil parameter comparison

IDamping (s/m) I Quake I
Shaft Toe1 I I
Shaft/O.lin Toe/(D/120)
No blow count matching

0.28
Mean

Figure 5. Match quality comparison with capacity 0.45 0.65

prediction, blow count

This is not significant since the data represented a

variety of soil types. Toe damping is generally Max 1.85 4.56 2.97 6.72
assumed to be independent of soil type. Its Min 0.14 0.02 0.41 0.36
magnitude is, however, highly dependent of the
magnitude of end bearing since viscous damping is Mean 0.74 0.84 1.17 1.93
divided by toe resistance to produce the Smith
damping value. The new CAPWAP routine St Dev 0.36 0.83 0.56 1.32
produced much more reasonable results than the COV 0.49 0.99 0.48 0.68
previous code with mean values of .84 vs 3.79 s/m
and COV’s of .99 vs 2.95.
Calculated q u a k e v a l u e s w e r e n o n -
dimensionalized by their GRLWEAP recommended 10 TOTAL, STATIC AND DYNAMIC
values. Thus, a calculated skin quake of 2.5 mm RESISTANCE
would be presented as 1.0 as would be a toe quake
value equal to D/120 (where D is the diameter or
width of a displacement pile). The non-dimensional CAPWAP calculates soil resistance as the sum of a
calculated shaft quakes were 1.OO and 1.17 for No static plus a damping resistance. The maximum
BCM and BCM with a slightly greater scatter for the static resistance component is equated to the static
new calculation method, probably because variation bearing capacity according to Smith (1950).
of quake values often help to improve the blow Another approach would be the calculation of peak
count match. Toe quakes were very similar for the total resistance, i.e. the sum of maximum static plus
two calculation methods with mean non-dimensional peak dynamic resistance, multiplied by a reduction

57
factor to account for dynamic resistance losses. A
justification for this approach is the difficulty of
separating static from dynamic components by
signal matching when the pile displacements are
small. The static, displacement dependent
components then differ littfe from the damping,
velocity dependent components which easily
introduces errors in the calculation. Worse yet, in a
hard or very dense soil or in a rock, the static toe
resistance components sometimes appear to be
velocity dependent and could therefore be
misinterpreted as damping resistance by the
traditional CAPWAP approach, leading to an
underprediction of static capacity. To check for
possible improvements in capacity prediction,
various methods of i~terpretationof the total shaft
and/or toe resistance values were explored. Table 3
shows the most promising method which adds the
calculated static shaft resistance to the total, factored
toe resistance. Sorting the results by dominant soil
type, a marked improvement of the traditional
approach was achieved for sands, where the mean of
the ratio of predicted capacity to Ioad test capacity
would be 1.02 with a COV of 0.24. Further
exploration of this method is warranted. However,
at this time too little experience is available (only 14
cases for the sands) and the time factors should also
be considered in future studies.
Table 3 : CAPWAP static shaft resistance and total
toe resistance divided by static load test capacity for
different soil types
Clay 6 1 1.30 0.50 I 0.38
SandvClav
4 2
4 1 1.23
t
0.22 I 0.18
SiltyClay t 8 1 1.08 1 0.20 0.18
Rock I
6 1 1.15 I 0.18
1 1
0.15
14 1 1.02 0.25 0.24
8 1 1.03 0.24 0.24 I
1
8 1 1.12 0.24 0.21
4 1 1.93 1.51 0.78
I 5 1 1.071 0.15 0.14
11 CONCLUSIONS
The correlation between CAPWAP predicted pile
bearing capacity and static load test capacity can be
improved if not only the difference between
computed and calculated pile top quantity but also
the difference between calculated and observed blow
count is included in the match quality evaluation.
The improvement over the traditional method, which
ignored the calculated blow count, is significant and
since, with modern computers, the additional
computational effort is minor, blow count matching
should always be done. Great care should be taken
58
in the measurement of restrike blow count. It is
believed that lack of accurate blow count
measurement limited the improvement over the
traditional CAPWAP approach. On the other hand,
the signal matching process itself already
incorporates blow count matching to a certain
degree, since the measured velocity and therefore the
top displacement are imposed as top boundary
conditions. The improvement achieved with the new
CAPWAP program should therefore be primariiy
attributed to a more accurate analysis over a longer
time period.
As in earlier correlation studies, the time factor
(time between load testing and installation divided
by time between restriking and installation) proved
to have the greatest effect on the accuracy of the
CAPWAP prediction. Obviously waiting times
comparable to those of the static test assures the best
possible prediction of long term bearing capacity by
CAPWAP.
REFEENCES
Balthaus, H., 1986. Zw bestirnmung der tragfaehigkeit von
pfaehlen mit dynamischen pfahlpruefmethoden.
Dissertation, TU Braunschweig, Germany.
GobIe Rausche Likins and Associates, Inc., 1999,
CAPWAP background report, Cleveland, OH.
Hannigan. P.> and Webster, S. W., 1987. Comparison of
static load test and dynamic pile testing results, Second
Int. Symposium, Deep Foundations Institute, May.
Hussein, M., and Rausche, F., 199 1. Bearing capacity of
deep foundations from dynamic measurements and static
tests - ten correlation cases, Piletalk International '9 1 ~
Malaysia.
Klingmiiller, O., 1984. Dynamische pfahlpruefung also
nichtIineare systemidentifikation: dynamische probleme -
modellierug und wirklichkeit. tagungsband. Curt-Rsch-
Institut, TU Hannover, Oktober.
L k n s , G., Rausche, F., Thendean, G., and Svinkin. M.,
1996. CAPWAP correlation studies. Proceedings of the
jthhnt. Cod. on the application of stress wave theor). to
piles, F. Townsend ed., Orlando, Fiorida.
Mure, J.N., Kightiey, M.L.. Gravare.C.J.t and Hermansson,
-
1983. CAPWAP an economic and comprehensive
alternative to traditional methods of Ioad testing of piles,
Paper 16, Piling and ground treatment for foundations,
Thomas Telford, pp 167- 174.
Rausche, F.: Goble, G.G., and Moses, F., 1972, Soil
resistance predictions from pile dynamics, JozirnaI ox the
Soil Mechanics and Foundation Division, ASCE.
Reiding, F. J., Middendorp, P., Schoenmaker. R.P.,
Middeldorp, F.M., and Bielefeld, M.W., 1988. FPDS-2:
A new generation of foundation pile diagnostic
equipment, Proceedings, Third Int. Cod. on the appl. of
stress wave theory to piles, Bengt Felienius, ed., Ottawa,
Canada.
Smith, E.A.L., 1960. Pile dnving analysis by the wave
equation, Journal of Soil Mechanics and Foundations,
ASCE, 86>August 1960,35-61.
 Application of Stress-Wave Theory to Piles, Niyama & Beim [eds) (9 2000 Balkema, Rotterdam, ISBN 90 5809 150 3

Combining static pile design and dynamic installation analysis in G

E Rausche & B. Robinson

Goble Rausche Likins and Associates Iiicorporcrted, Cleveland, Ohio,USA
J. Seidel
Monush University, Melbourne, Vic.,Australicr

AE3STRACT: Wave equation simulations of pile installations by impact driving have become an important
part of the pile design process in many countries. In its traditional form, the wave equation approach replaces
a dynamic formula, i.e. it leads to a bearing graph which is a relationship between pile bearing capacity and
blow count. In addition, the simulation calculates realistic dynamic pile stresses. GRLWE.Q, probabiy the
most widely used wave equation program, also enables the user to calculate blow count stresses vs depth in a
so-called driveability analysis. T h s procedure requires soil resistance vs depth input which thus far had to be
precalculated in a separate analysis either manually or using another computer program. The program has
now been expanded to include t h s pre-analysis. T h s paper presents the method itself, its correiation with
static tests and another similar method. It also briefly discusses limitations and special considerations that
make this method somewhat different from other static pile analysis formulas..

1 INTRODUCTION for consideration of different pile lenshs dunng

-4s originally proposed by Smith (1960) the wave
equation method generates a so-called bearing graph
w h c h is a relationship between pile bearing capacity
and blow count. In addition, maximum tension and
compression pile stresses are usually plotted. These
results are well suited for the selection of hammers,
hammer performance parameters and cushions for a
given pile type and soil condition.
The GRLWEAP program (Goble et al. 1999) not
only calculates bearing graphs but also offers a
convenient method for predicting the blow count as
a function of pile penetration. However, for
meaningful results, this so-called driveability
analysis requires a much more detailed soil
parameter input than the original approach. The
required static and dynamic soil resistance input
parameters should reflect the various layers that the
pile penetrates. For realistically calculated blow
counts vs depth, it is necessary that these soil
parameters are determined with as much accuracy as
possible based on quality geotechnical information.
From such improved analysis, it is possible to obtain
the best estimate of both the total number of blows
required to install the pile, and the total installation
time. In addition. the GRLWU..P program allows
installation and hammer and pile cushion
parameters, which may be adjusted as driving
resistance varies to most closely model the pile
installation process.
For user-friendliness the GRLLWAP program has
now been expanded to accept soil stren,uth and soil-
type data input. Based on this information, an
automated static formula approach has been devised
which estimates the static soil resistance parameters
and calculates a load-movement curve at the design
depth. The program automatically selects the
dynamic soil resistance parameters including
resistance loss factors, which relate static resistance
to driving to static long term capacity values. The
complete GRLWEXP driveability approach is
described in this paper and its benefits and
shortcomings are discussed.
It has to be emphasized, however, that there is no
universal approach which is hlly reliable, and that
the pile driveability analysis requires the local
knowledge of an experienced geotechnical engineer
to be most accurate. It is also evident that the quality
of the prediction is a fbnction of the quality of the
soil information whch is available. The method
presented here should not replace the design
calculation that the geotechcal normally performs

59
for File type selection.. Instead, GRLWE.AP's
analysis is an additional effort in the preparation of
the driveability analysis.
2 G E N E R U CONSIDERATIONS
Driven pile analysis is generally performed in two
steps (Hannigan et al. 1996): an initial static analvsis
which is followed by a dynamic or wave equation
analysis. Static analysis can be based on one of the
many proposed approaches, whch calculate shaft
resistance and end bearing for a particular pile
penetration depth. Depending on the quality of the
available soil strength data, the variability of the soil
properties over the site, and the realism of the
calculation procedure, the reliability of the results
obtained from such analyses varies significantly.
Statically calculated capacities are indirect estimates
based on site soil parameters estimated from the
foundation investigation, not based on the additional
information provided by the installation process.
Because of this inherent shortcoming, engineers
often prefer the use of a dynamic formula andior
wave equation analysis which provide a more direct
method of capacity evaluation for each individual
pile. In these approaches, blow counts from pile
installation observations, plus a calculated
relationship between blow count and beanng
capacity (the bearing graph) is used to make a pile
capacity prediction. The driveability approach can
also be used prior to pile installation to predict rather
than confirm the installation process. In this case,
field observations of the installation process are not
available to improve the reliability of the analysis
procedure.
In traditional methods of static pile analysis. the
requirement that the analysis provide a safe design is
paramount. Efficiency of design is something for
whch all engineers should strive, however. it is of
less importance than providing a safe structure.
Conservatism in the design is provided in several
ways:
1. In conservative interpretations by the engineer
of the site soil strength parameters;
2. In upper limits on strength values imposed by
the design method;
3 . In a conservative appraisal of the data on Lvhich
the design method is premised;
4. In the application of safety factors or strength
reduction factors.
The last of these factors affects only the maximum
structural load which can be allowed on the
foundation element during its service life. However,
the first three of these factors ensure that most
design methods applied in practice will
underestimate ultimate pile capacity.
Driveability analyses are conducted as part of the
design process for a number of reasons:
60
1. To evaluate the ability of a given pile driving
hammer to drive the pile to the nominated capacity
and/or penetration;
2. To estimate the final blow count;
3. To evaluate the stresses which will be induced
in the pile during the installation process.
In such an analysis, a conservative choice of soil
strength parameters has an ambiguous meaning. For
instance, when assessing tension stresses or the
required penetration, an underestimate of soil
resistances would usually produce a conservative
analysis, whereas in assessing, compression stresses,
an overestimate of soil resistances would generally
be conservative. Assessment of the bearing capacity
which can be achieved by a given hammer may not
be significantly affected by the assumed distribution
of resistance. Because of this ambiguity, and
because the consequences of an incorrect assessment
of soil resistance distribution does not have the same
implications as traditional static design with regard
to the safety of the structure, it is most appropriate to
undertake driveability analysis using the most
realistic assessment of soil strength parameters.
In the GRLWEAP static calculation method, it
was therefore attempted to avoid a static resistance
bias as much as possible. Again, the assessment of
the pile's bearing capacity should always be done
with prudent limits or reductions as dictated by
experience.
Having calculated the static soil resistance based
on in-situ soil strength parameters, an estimate has to
be made of the static soil resistance that is actually
present during pile driving. For example, pore water
pressure changes in the g o u n d during pile
installation tend to change the effective stress regime
and therefore the resistance acting on the pile.
Unfortunately, no matter how accurately the soil
exploration and static analysis was conducted,
estimating the Soil Resistance to Driving (SRD)
based on its static capacity may generate significant
errors and although experience values are available,
the selection of so-called gain-loss or capacity
reduction factors is one that should always be
carehlly reviewed. These uncertainties are greater
for sensitive soils such as marine clays and it is
common practice to estimate both a lower and an
upper bound SRD and calculate lower and upper
bound blow counts by the wave equation.
When pedorming a dynamic analysis using
GRLWEAP, it is not only necessary to calculate the
static resistance and its distribution, additional
dynamic soil resistance parameters, damping and
quake, both at shaft and toe have to be estimated.
Actually, GRLWEAP recommendations (Goble et
al. 1999) are rather simple for standard analyses
where little is known about the soil. Based on these
commonly accepted rules, unless dynamic test
results indicate otherwise, only the shaft damping
factor is a function of soil type. Toe damping is
probably independent of soil type because the In Equation 1, k is the lateral earth pressure
dynamic resistance component at the pile toe is more is the mid-layer vertical
coefficient, CT,,,,~~ effective
a fimction of inertia forces caused by the soil being stress and d is the friction angle at the soil-pile
displaced around the pile toe than with forces of interface. The effective stress requires that the
viscous flow. Similarly, shaft quakes are assumed
independent of soil type with no evidence that this
assumption affects the accuracy of wave equation
predictions. For the toe quake, experience indicates
that pile size has to be considered and, in the case of
rock, the hardness of that material. With soil type
and pile size known, it is therefore a simple task to
assign the necessary additional paramerers
automatically.

3 THE GRLWEAP STATIC ANALYSIS

To estimate the ultimate static capacity of a pile, two

quantities must be calculated: the shaft resistance
and the toe resistance. Many methods of static
capacity estimation exist for cohesionless and
cohesive soils. These range from empirical methods
based on SPT and CPT values (e.g. Meyerhof, 1976
and Nottingham and Schmertmann, 1975) to semi-
empirical effective stress and total stress methods
(e.g. Fellenius, 1991 and Tomlinson, 1980). All of
these methods require that soil type and some soil
strength parameter (SPT-N value, CPT cone
pressure, friction angle, unconfined compressive
strength, etc.) is known. Unfortunately, the methods
often Sive no clearly defined solutions for certain
ranges of soil strength parameters, because their
experience base is limited.
WhIe it is certainly desirable, and theoretically
more accurate, to know as much about the
geotechnical properties of a site as possible, hnds
allocated for soil exploration studies typically are
kept to a minimum, usually in a false sense of
economy. Particularly for small projects, only a few
soil borings with depth, soil type and SPT N-value
are available to the deep foundation designer. In
such instances, a sophisticated analysis method will
not add much to the quality of the prediction. In
fact, there is often a wide gap between what should
be known about the soil to satisf) a method’s input
requirements and the available data. For example,
specific weight, friction angle and pile adhesion
often must be known, however, the soil exploration
yielded only soil type and SPT N-value. The buoyant weight of the soil is calculated and that
requires knowledge of water table and specific
designer therefore has to choose an empirical weight. The depth of the water table is therefore an
approach to convert soil type and SPT N-value to the additional input into GRLWEAP. The specific
required parameter, before the calculation method weight (as well as the friction angle) is based on SPT
can be employed. This two-step approach has been -value and soil type. Thus, an extensive soil type
automated in GRLWEAP. table was developed whch also serves to estimate
GRLWEAP estimates shaft resistance based on an dynamic soil resistance parameters such as quake
effective stress approach. The unit shaft resistance and damping. A portion of that “click-on“ table is
a t a point along the pile is therefore calculated from shown in Figure 1. The table contains much more
detail than necessary for a simple SPT based soil and

61
pile analysis. However, it contains information
necessary to expand the analysis methods in the
future.
The unit toe resistance in kPa is estimated
according to Meyerhof, 1976 using the simple
empirical expression:
cot = min(200N; 12,000)
where N is the SPT N value in the strata at the
bottom of the pile. Of course, it is possible to
improve this very simple method in future updaIes
and as experience is gathered with this approach.
Soil data should be entered where layers change,
where the water table is encountered, or where major
changes in N-values occur; also, data must be
entered when the soil type changes, and at minimum
intervals of 3.5 m.
As discussed, the N-value and soil type is used to
estimate unit weight, friction angle (effective hc:ion
angle for clays) and horizontal stress based on
correlation studies by Kulhawy et al. (1939 and
1991), Schmertmann (1975 and 1973): and
Robertson and Campanella ( 1933). The lateral earth
pressure coefficient is estimated from the horizontal
stress for sands and from the effective friction angle
for clays and from the diameter of the pile. This
value must fall between the Rankine active and
passive earth pressure coefficients.
Both calculated unit shaft resistance and end
bearing values are then input into the G R L W L V
driveability analysis. Since soil type is known. the
damping factor for the shaft can be automatically
assigned. Toe resistance calculations from unit toe
resistance also require the knowledge of an effective
toe area (for open profiles such as pipes and H-piles
these may be the plugged areas), which can be
immediately converted to toe quake. As discussed,
shaft quake and toe damping are independent of pile
size and soil type.
For a particular pile depth, the static analysis
method can be coupled with the static load-set curve
calculation method proposed by Fleming ( 1992).
This additional result requires estimation of the
settlement characteristics of the soil. Again. soil
type and SPT -values are converted to provide the
necessary input. Obviously, calculation of
settlements is an even more challenging task than
bearing capacity prediction, particularly when only
SPT information is available. These results must
therefore be used with appropriate caution. X
correlation study on this result had not been
completed by the time t h s paper was finished.
4 CORRELATION WITH STATIC LOAD TESTS
Using data from the database presented by Liluns et
al. (1996), a correlation study was conducted on 53
62
cases for which static pile load test was available in
a variety of soil and pile types. The proposed
GRLWEAP static capacity method was used. It is
noted that the predictions are automated, and thus
operator independent. The ratio of predicted static
failure load to actual capacity as evaluated by the
Davisson offset method averaged 0.92, with a
coefficient of variation of 0.31 (see Table 1). By
comparison, the same data was also subjected to
analysis by the SPT97 approach which is popular in
the South-Eastern United States (Townsend et al.
1997), which yielded a mean of 0.31 and a
coefficient of variation of 0.39. Figure 2 shows the
capacity ratios of GRLWEM and SPT97,
demonstrating that the two methods show similar
tendencies of underprediction or overprediction.
This may be an indication of the failure of the basic
soil data, primarily the SPT -value, to provide
correct and sufficient soil strengqh information.
Clearly, for the particular set of correlation cases
investigated, GRL’ltrEAP’s static method is slightly
less conservative than SPT97 and also less scattered
(Coefficient of Variation 0.3 1 .vs 0.39).
GRLWE,i\P SPT-97
I Minimum I 35 I 33 I
I Standard Deviation I 29 1 32 I
Coefficient of
Variation
5 Sb%&MY . & CONCLUSIONS
I
A static analysis method has been presented that has
been incorporated in the dynamic wave equation
analysis program GRLWEAP. The method is based
on SPT -values plus an assessment of soil type. An
effort has been made to make this method:
1. hlly automated and therefore independent of
user judgement. Covering the complete range of
possible N-values and soil types.
2. Unbiased, i.e. neither conservative nor non-
conservative for capacity, stress and blow count
predictions.
3. Generate automatically the dynamic soil
resistance parameters necessary for both bearing
graph and G R L W E A P whole-installation
driveability analysis.
Figure 3. Comparison of G R L W ~ A Pwith SPT97 correlation.
4. Capable of calculating a prediction of pile-top
load-movement for a specific pile embedment.
Results from this analysis method compare
favorably with similar methods. However. the user
of this approach is advised of the shortcomings
inherent in any static formula approach.
REFERENCES
Fellemus. B H . 1991 Pile Foitndaiions Foitndation
Engineering Handbook. Second Edition H S Fang, ed .
Van Nosuand Reinold. publisher. New Yorh 13 5 11-
536
Flemng. W G K 1992 A ne\v method for srngle pile
settlement prediction and analtsis Geotechniqzir 42
411425
Goble Rausche Lihins and Associates 1999 GJU KE4P
Wave Equation .tfanual. Clex eland. Ohio
63
Hannigan. P.J.. G.G. Goble. G. Thendean. G . E . . L h s . & F.
Rausche. 1996. Design and construction of dnven pile
foundations. U.S. DOT Federal Highuay .4timrnisiranon
Reporr No. FHW.4-41-96-03;,
Kulhany. F.H.. C.W. Jackson. & P W. Ma\-ne. 1989. First
order estimation of KO in sands and clays. Forindanon
Engineering: Ciirreni Principles anti Practices. Vol. 1 ~
Ed. F. H. Kulhawy. ASCE. New York. 171-12-4.
Kulhay.. F.H. & P.W. ivlaye. 1991. Relative densic.
SPT and CPT interrelationships. Calibration C h m b e r
Tesring, Ed. .A. B.Hyang, Elsevier. New York. 197-21 1
Likins. G.. F. Rausche, G. Thendean. 22 M.Svinkin. 1996.
CAPWAP correlation studies. Proceedings of rhe F$h
Iniernntional Conferenct! on the ,4ppiication of Stress-
Wnve 2'77eop to Piles. Orlando. 447-464.
kleyerhof. G.G. 1976. Bearing capacity and settlement of
pile foundations. ASCE J. of the Geoiechnicni
Engineering Division. 193-228.
Nomngham and Schmertmann. 1975. missing.
Robertson. P.K. & R.G. Campanella. 1983. Interpretation
 of cone penetTaQon tests Sand. Can Geot J . 20 (1).
718-733
Schmertmann. J H 1975 Measurement of m-situ shear
strength. Proceedings of Confirence on In Srtzt
itlensitrement ofSoil Properties ASCE. New York
Schmertmann. J H 1975 Guidelines for cone penetranon
test performance and design. FHWA-TS-75-209
(report). US Department of Transportation. 145
Smith. E .A L . 1960 Pile dnving analJsis bc the n a \ e
equation. Joiirnnl of Soi I Mechanrcs and Foiindarions
XSCE. 86. August 1960.35-61
Tomlinson. bl J . 1980 Foundntion Design and
Construction. Fourth Editlon. Pitman Ad! anced
Publishmg Program
Tonnsend. F et a1 1997

64
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Rules of thumb for field and construction engineer in relation to impact

pile driving
W. J. Lucieer
Hollandsche Beton Groep n.x, Rijswijk, Netherlands

ABSTRACT: Simplification of the process of the complex shock waves propagating through a pile which
is driven by an impact piling hammer has to be done very carefully. Concerning the pile driving process the
object is to obtain insight by means of a simple calculator using simple equations.
A long experience based on the method of characteristics of the stress wave theory and field measurements
shows the possibility of very simple Rules of Thumb. The method of characteristics offers simple equations.
Derivation, background and subjects of approximation are described, together with information what is very
close to reality and what is not. Care should be taken in the interpretation of the outcome of the Rules, be-
cause the process is related to time depending soil properties.

1 INTRODUCTION per &second in steel and 4 in concrete. So for a

25 meter long steel pile it takes the front 5 millisec-
If the impact pile driving is considered, then insight onds to reach the tip of the pile.
in the process is the first important item. Simplifying During pile driving there is a fixation on the dis-
the piling rig, hammer with housing or cage and fal- placement of the top of the pile. That can be moni-
ling weight (ram), hammer cushion, pile cap, op- tored and the intention of the piling crew is that the
tional pile cushion, pile and soil to only a ram and a pile mainly disappears under ground level. But there
pile is a first action. (Figure 1) is no displacement after each blow if there is no set
During driving the of the tip of the pile. For a zero set of the tip there is
ram hits the- pile only an elastic displacement of the top returning to
RGM with an impact ve- zero before next blow occurs.
"t
locity v, . As long as Looking from another point of view it can be for-
the ram is in contact mulated that pile driving is the art to obtain a set per
Top- PILE
with the top of the blow of the pile tip. Hence, through the relation with
m-r =Ground level pile, a shock wave the top of the pile a final displacement per blow re-
is generated from mains for the pile top.
the surface of con- In other words: What are the possibilities for the
tact between ram displacement of the pile tip? Let us focus on that.
"IL and pile and which
is fed by the energy
accumulated in the 2 PILE DRIVING PROCESS
Tip +I-~

ram. The front of

the shock wave
Figure 1. Simplification of pile travels ,
driving system with a velocity of
5125 m/s through a
steel pile or about 4250 m/s through a high quality
concrete pile. In the case of solid rams the duration
of the hit takes a maximum time of say 5 to 6 milli-
seconds. The best thing for the process is thinking in
milliseconds. Then it can be said that the front of the
shock wave travels with a velocity of about 5 meter
The process can be modeled in the simple way as
mentioned in the first part of the Introduction, in
which force waves are propagating through the pile
and ram. Then pile and ram are simplified to one
dimensional elastic bars. The force waves can be
modeled in the way of the method of characteristics
introduced by Voitus van Hamme in 1974.
Through the pile, and even so through the ram
and other parts, there is a flow of two force waves
originating after each hammer blow and mainly dis-

65
appeared before next blow. One travels downward 2. Set s, displacement of the tip of the pile, which
FJ, the other upward F?. determines the blowcount for the concerned pile
Let us concentrate on the pile and forget for an in- penetration depth.
stance the waves traveling through the ram. 3.Allowable stresses in the pile.
First there only originates from the moment the 4. The velocity of the plane could be calculated
ram hits the pile a downward force wave FJ , which for every cross section. So, for every plane, the dis-
propagates in the direction of the pile tip. (Figure 3). placement could be calculated. This is most times of
As the front of the wave passes ground level the soil minor practical value.
reacts with friction forces in an obstructing way to
the downward traveling force wave. If the size of the
3.2 Equations
force in the downward traveling wave is larger than
the friction force over a certain height along the Following equations are used in the method of char-
shaft of the pile then the force of the downward acteristics, where:
propagating wave is decreased with half the size of F = force “1
the friction force. (The other half of the friction FJ= downward force wave “I
force is added to the upward traveling force wave.) F?= upward force wave “1
If the size of the downward traveling force wave FL v = velocity Wsl
is large enough to overcome all the individual fric- Z = mechanic impedance of element [Ns/m]
tion forces along that part of the shaft of the pile,
Zr = impedance of ram
which is in the soil, then that remaining force
(FJ minus % Wdyn) meets the point or tip resistance
force. If the size of mentioned remaining force is Zp = impedance of pile
larger as half the maximum point resistance then the E = Young’s modulus of elasticity [N/m2]
plane of the point of the pile can accelerate, obtain A = cross section area element [m’l
velocity and hence displacement. As long as the size c = wave velocity in element [m/sI
of the force of the downward coming force wave is m = mass [kgl
larger as half the size of the resistance force gener- m, = mass of element i [kgl
ated at the tip of the pile, the tip will displace and a
final set will be obtained. EA E
Z=- or Z=;A

General equation 1 shows the mechanic impedance

3 MODELING WITH METHOD OF
CHARACTERISTICS
3.1 General
In this method ram, pile cap cushion and pile are
modeled as one dimensional (vertical positioned)
elastic bars. Each bar is subdivided in elements. Be-
tween the two planes of each element the downward
and upward wave can travel undisturbed. Only be-
tween the lower plane of the upper element and the
upper plane of the following lower element could
occur a change in cross section area or mechanic
impedance andor an external force, such as a con-
centrated friction force, could act on.
With this method, intentionally used for computer
programs, it can be simplified with larger elements
with the length of parts, which does not change of
cross section area (or mechanic impedance). Then
we have to keep in mind that the accuracy will be
less, but will be enough for insight in the process.
The method calculates with the three quantities:
force, velocity and impedance.
From those the following important practical items
could be calculated:
1. F,, maximum force of the downward running
force wave originating from the plane between ram
and pile as a result of the impact velocity of the ram
and mechanic impedance of ram and pile.
of a one dimensional elastic bar. This relation shows
that if the material is known then the constants
E and c are a fact. So the impedance is linear related
to the cross section area A of the element, pile, ram
etc. In practice area A will change and is in the pile
driving process a very important “quantity”. (re-
mark: The impedance represents as well the spring
characteristics as the mass of the concerned element
or bar).
Equation 2 gives a possibility to calculate for in-
stance a pile cap or an impact block with membrane
like shapes. This equation is the equation to calcu-
late the impedance of an element with a shape of a
potato. For the simplified pile driving process we do
not need this equation, but it could give more in-
sight.
In this equation - m, - represents the total mass of
the element and - delta ti - the total time in reality
that the three dimensional waves (compression and
shear waves) travel from one side to the other side of
the element.
(3)

66
Equation 3 gives the maximum force F, of the
downward traveling impact wave originated in the
plane between ram and pile. From pile driving meas-
urements of steel to steel contact between ram, pile
cap (no cushion !!) and pile it is known that this
equation is very close to reality.
What could this equation say ? First, the force F,
is linear related to the impact velocity of the ram.
Second, with a chosen hammer, a particular ram
with a particular maximum impact velocity v,,
maximum force F, is depending on the impedance of
the pile Z,. ( =Practical statement). Because in prac-
tice the rate of impedance of hammer and the im-
pedance of the chosen pile varies from say 10:1 to
4:l. Therefore the impedance of the pile and hence
the (minimum) cross section area of the pile is very
dominant to determine the size of the maximum
force introduced in the pile.
ZV=FJ.-F~ (5)
Equations 4 and 5 are the relations for every cross
section and related to the forces of the downward
and upward propagating waves passing the plane of
the cross section.
So the total force in and the velocity of the con-
cerning plane could be calculated.
Figure 2 . Calculation of resulting upward and downward
force waves from known waves arriving at virtual plane be-
tween two successive elements.
In the case the ram, pile or any other part of the pile
driving system is modeled as elastic one dimen-
sional bars, then the bars could be divided in ele-
ments. Over the element there is supposed that no
disturbances, such as a friction force, or change in
cross section area or in material properties occur,
Those disturbances and changes occur between the
lower plane of element 1 and the upper plane of
element 2, see Figure 2.
Suppose a virtual plane with zero thickness,
which is situated between element 1 and 2. At a cer-
tain time the force with size F.J~of the downward
67
force wave from element 1 arrives at the virtual
plane. Even so the force with size F?, of the upward
traveling wave from element 2 arrives.
Those two forces are the known forces. From those
two forces the two departing forces could be calcu-
lated with equations 6 and 7. In these equations there
are two terms after the sign of equality, which are
contributing to the downward traveling wave (equa-
tion 6) or to the upward traveling wave (equation 7).
That is the way the wave comes through the virtual
plane (refraction) and the way the wave reflects on
(reflection) the virtual plane. Equation 6 shows first
a refraction term and second a reflection term. In
equation 7 the order of the terms is reversed.
For the simplified process only the downward
traveling wave is of interest. So if there is no upward
wave then in equation 6 only remains the refraction
term (first term).
4 SHAPE OF FORCE WAVE
Next equations 8, 9 and 10 are functions of the
downward force wave, which originates from the top
of the pile after the ram hits the pile.
t
in which n(t)= integer ( 2 delta-TJ
Hr
delta-Tr = 7 ; Hr = Height of ram ,
Figure 3. Force - time graphs at top of pile for down-
ward force wave in case of simplification (I), approxi-
mation ( 2 ) and according method of characteristics ( 3 )
 c= wave velocity in ram; t= time

Ft
F=Ft - r t ; if O < t l T (10) T is a rather fixed time for solid rams, when there is
From measurements in the field it is known that contact with the pile. This time is about 5 to 6 milli-
the force-time curves do not change so instantly as is seconds.
shown in Figure 3. Accelerations of 10,000 - Then the Rest Impulse is:
20,000 m / s 2 are possible in steel to steel impacts. So
the curves should be more “rounded”.
Curve 3 in Figure 3 shows the best curve in the
case upward force waves are absent or could be ne- The Rest Impulse is a quantity for the set of the pile
glected. This curve is calculated according equa- tip, which is worked out in section 13.3, Set of Pile.
tion 9, following the method of characteristics. In
this example the height of the steel ram is 2.50 meter
andZ,= 10Zp.
Curve 2 is an easy to calculate approximation of
curve 3, “touching” the lower values of curve 3. So
the impulse, represented by the area below the
curve, is lower as for curve 3.

Curve 1 is used in the simplified method and easy to

determine. This straight line represented by equa-
tion 10 has an intersection on the force F axis at the
value F, and an intersection at the horizontal axis at
time t= T milliseconds. In the field T is the time
from the moment the ram hits the pile till the time
the ram looses contact with the pile. For hammers
with steel to steel contact or a piling system with a
hard cushion this time T=: 5 to 5.5 millisecond. In
the case of a soft cushion T z 6 millisecond.
If the impact force graph is represented by Figure 4. Simplified force - time graph with Rest Impulse,
curve 1 then the total impulse and the Rest Impulse which determines the set of the pile.
are to be calculated by means of very simple equa-
tions, which suits in the option to simplify the pile
driving process. 6 MODELING RAM/HAMMER

Hammers have diverse outlook. Very simple ham-

5 RESTIMPULSE
The simplified Total Impulse transferred from the
ram during its contact with the pile is represented
by the area enclosed by the force and time axes and
the inclined graph of downward force-time relation
(Figure4). The impulse is needed to overcome the
total sum of friction forces Wdyntotal and the point re-
sistance F, at the pile tip. In other words: the down-
ward traveling force F will be decreased by half the
total sum of friction forces Wdyntotal and half the
point resistance Fp,See section 12 and 13. Then the
remaining rest force F,, will cause the set of the pile
tip. This gives equation 11:
1 1
Fr-m=F- Twdyntotal - zFp O < t < A T (11)
This results in the equation of the Rest Impulse:
mers have only a ram of cast steel, which are guided
by two leaders of the rig (Dutch rig). Usually ham-
mers have a housing or a cage in
which a ram can move. For a Die-
sel hammer the ram is the piston.
Most of the hammer types have
solid (not hollow) rams. The piston
of Diesel hammers have over their Real Model
height about the same cross section Figure 5 . Model of
area. The ram of steam hammers is steam driven ram
hollow. See Figure 5. Even so the
onshore hydraulic hammer of manufacturer Menck.
Many of the hammers of manufacturer IHC have a
solid ram existing out of three parts, in which the
middle one has the largest cross section area.
To calculate the representing impedance of the ram
in the case of no uniform cross section area, the
cross section area with the largest area represents the
impedance of the ram. In the case of steam hammers
the height of the largest cross section area (lower
part) is minor to the height of the upper part and

68
hence the cross section area of the upper part repre- Table 1. Material properties ram
sents the ram. Material Young’s Modulus Wave velocity Type
With previous information and the material prop- N/mL x 10’ m/S
erties from table 1 the representing impedance of the
Steel 206 5 125 1
ram could be calculated with equation 1. Perlitic
Table 1 shows possible material properties suit- cast iron 175 4960 2
able for a manufactured ram. 1 Hydraulic, steam and KOBE Diesel hammers
2 Piston Diesel hammers
7 EXAMPLES OF DYNAMIC HAMMER DATA
Table 2. Dynamic hammer data
Next table 2 shows some examples of hammers, Type Mass A, Z, Maximum Re-
marks*
used in the field, with the data needed to use in the
Ram m, vt
simplified calculations according the rules stated in ___-- ____
ton m 2 m s / m m/s
section 16, Rules of Thumb.
IHC-S70 4.5 0.17 7.0 6.3 Hydrau-
lic
D- 100 10 0.30 10.6 4.5 Diesel
8 IMPACT VELOCITY AND HAMMER MRBS3000 30 1.02 40.9 4.7 Steam
ENERGY * Remarks give the process of lifting the ram. The ram of the
IHC-S70 hammer is also pushed down, resulting in a smaller
Maximum impact velocity vt of the ram is a very stroke of the ram. The ram MRBS3000 is free falling. The fall
important item. It could be obtained from the kinetic of the piston of the Diesel hammer Delmag D-100 is slowed
hammer energy mentioned by hammer manufactur- down by the compressed gas.
ers. This is only done with modern hydraulic ham-
mers. Mostly the potential energy (U= mgh) is men- “accumulator” for the energy coming down from the
tioned, based on the maximum uplift of the ram or ram.
piston. For free falling rams (such as steam and most Many times the pile cap of onshore hammers
hydraulic hammers) it could be stated that at impact contains a hammer cushion and in the case of driv-
!hmv2= mgh in vertical condition. For Diesel ham- ing concrete piles a pile cushion. Large hydraulic
mers the kinetic energy has to be related to a compa- offshore hammers have no cushion. The influence of
rable uplift h= 0.9 - 1.0 meter and a free fall there a cushion can not be neglected.
after, which means the maximum vt = 4.5 m/s. (Effi-
ciency < 35% !!).
10.2 Cushion
There are two kinds of cushion used in driving, a
9 DETERMINATION OF SRD
hammer cushion and a pile cushion.
The hammer cushion is used to protect the ram
What is Soil Resistance during Driving or SRD?
against the enemy metal fatigue due to the high re-
There are two kinds of resistance acting on a pile in
petitive shock loads. It could contain blocks of hard
the soil. That is during the driving process the fric-
wood, coiled laminated metal, mattresses of steel
tion (W,,, acting along the shaft of the pile and cables, plastic dolly, etc. and may be covered by a
the point resistance (F,) acting on the tip. Those
thick steel plate for a better contact and protecting
quantities are dynamic forces. Hence the definition
the filIing material.
of the SRD is:
The pile cushion is meant for a good fit to the pile
SRD= Wdyn total + Fp (15) cap and protection of the top of the pile from dete-
In section 12 and 13 will be explained in what way rioration due to cracks. This pile cushion is the soft-
est of the two cushions and replaced for every new
the Wdyn and F, are obtained. pile. Mostly made of soft wood in pieces of more
than one layer and compressed during driving to
10 INFLUENCE OF PILE CAP/CUSHION 10% - 20% of its original height.
Comparing to steel to steel impacts, the effect of
10.1 Pile cap the cushion on the force - time graphs results in the
The pile cap is an extra part between ram and pile following:
and can easily interrupt the propagating force wave a) The maximum downward force F, decreases
to the pile. The cause could be a bad or no contact with 80% - 90%
between ram or pile. Sufficient mass of the pile cap b) The “head” obtains an increase in width in
could prevent those bad or no contacts. If that is the the order of 10% - 20%.
case then there is no problem in the wave propaga- In the simplification item a) is the important one and
tion and the pile cap can be used as intermediate can be expressed in the cushion factor q,. Then
equation 3 has to be multiplied by this factor.
69
So, to express the influence of the cushion, equa- the friction force Wstat, acting on the pile, could-be
tion 3 becomes: calculated.
In the simplified process the dynamic force wd,
is at least half the static force.
(16)
1

Table 3. Values for cushion factor qc Wdyn = Wstat (18)

Cushion qc Examples of hammer type If the pile is an open ended pipe pile then, due to
no 1.O Hydraulic offshore hammer the friction acting on the inside wall of the pile,
hard wood 0.9 Steam offshore hammer
Plastic dolly 0.9 Hydraulic onshore hammer
%WStathas to be added to Wdyn.
Plastic dolly 0.9 Diesel hammer on steel (sheet) piles
Cable mattress* 0.9 Diesel hammer on steel (sheet) piles
Soft wood 0.8 Diesel hammer on concrete tiles
12.2 Equations
* 6 layers of flat coiled cable

11 MODELING PILE CROSS SECTION

CHANGE

In the simplification only the downward traveling

force wave is of interest. The force of the upward
traveling wave is supposed to be zero. (In case of
many changes in cross section care has to be taken
into account). So from equations 6 and 7, only in Figure 6. Friction force acting on virtual plane
equation 6 the first part is of importance. (See also
Figure 2). In the method of characteristics the pile is divided in
Hence, if the wave travels from element 1 to ele- elements. Below ground level the friction forces wd
ment 2, the downward wave force in element 2 is: are acting on the virtual planes between the ele-
ments. (See Figure 6). Force wd represents the fric-
tion acting on the lower half shaft area of element 1
together with the upper half shaft area of element 2.
It is supposed that force wd is acting in upward di-
12 MODELING FRICTION rection. Then next equation could be derived.
12.1 Soil Equilibrium of forces related to virtual plane:
Soil is the most difficult subject for the dynamic
process to obtain reliable values. Many times the in- Fd1 + Frl - w d = Fd2 + FE (19)
formation over the soil is poor. For the driving proc-
ess soil below the ground water table or below sea
level is of interest. If the soil is far above ground wa-
ter table then driving could be a problem. Bore piles
could in that case be preferred. Interpretation of the
soil information gathered by tests is usually exe-
cuted by soil engineers, specialists, with the option
that the soil is acting in a static behavior, not dy-
namic. That is why only a few specialists could
translate the static values to dynamic ones. The
problem could be that much experience is needed.
Concerning friction, static values could be ob-
tained from tests according the Dutch Cone Penetra-
tion Test (CPT-test), the Standard Penetration Test
(SPT-test), laboratory tests on borings etc. and velocity of virtual plane:
The CPT-test gives graphs for friction and penetra- 1
tion pressure. The SPT-test needs a translation to ob- F.11 - FT2 -5Wd
tain friction and pressure values from the number of V= Z (24)
blows per foot of the sampler in the bore hole.
If the values of the friction [kN/m2] of the different What may be said in the simplified process, that the
soil layers are obtained for the whole length the pile downward traveling force wave F is decreased by
is in the soil, then over the height of every soil layer half the dynamic friction force wd.

70
So when force wave F arrives at the tip of the pile Or
this force is decreased with half the total sum of fric-
tion forces Wdyn , acting along the shaft.

13 MODELING POINT RESISTANCE

13.1 Soil From Figure 4 and equations 1 1,12 and 13 follow
From the soil values obtained by CPT, SPT - tests or that the integral equals the Rest Impulse I,.
borings / laboratory tests (section 12) the Cone resis- So the set is:
tance (pressure) CT, could be determined for different
depths. Then the point resistance (force) F, acting on 2
S= -1r
the tip of the pile, with area A,, can be determined in zP
case of:
- Sheet piles and open ended tube piles with small Or with known values:
wallthickness.
Then F,= CT, x A, . On rock 0, = undrained
compression strength (33)
- Closed ended tube piles, concrete and wooden
piles. F, is influenced by the soil around the tip.
The point pressure CT, is determined by means of 14 DETERMINATION OF BLOWCOUNT
the 0.7D - 4D method, an averaged pressure re-
lated to 4D (D= pile diameter) below the tip and If the set s of the pile tip is calculated from the Rest
8D above it. Impulse, see section 13 equation 33, then it is possi-
Then F,= CT, x A,, ble to calculate the blowcount in blows per 0.25 me-
ter or per foot penetration of the pile.
13.2 Velocity Pile Tip The blowcount n is:
If F, is a calculated value, then velocity v, at the pile 0.25 0.3048
n = y- blowd0.25 m (n = blows/foot)(34)
tip could be derived.
~

From general equations 4 and 5 equations 25 and 26

can be derived for the pile tip.
15 MAXIMUM SRD FOR A PILE
Fp =FJ. + F? (25)
The maximum SRD is based on the maximum
vpZp = F.J - F? (26) downward force traveling towards the pile tip over-
coming friction forces and the force acting on the
Then from equations 25 and 26 we obtain pile tip. The maximum SRD is referred to a practical
vpZp = 2F&- Fp ; if ~ F .-JFp >O (27) refusal of 120 blows/0.25 ni (=: 150 blows/foot).
(Above this limit damage to most hammer types
could be expected, because with a little increase of
resistance the blowcount increases progressively).
E
3 1 1 SRDmax = 0.8 q c vt ; Ap Only Point (35)

E
SRDmax = 0.7 q c vt ; Ap Friction + Point (36)
With A, = cross section area close to pile tip.

13.3 Set of Pile (tip) 16 RULES OF THUMB

If the velocity at the tip of the pile is a known value
at any time during the process of wave propagation, The equations have the reference numbers as men-
then the total displacement or set per blow could be tioned in previous sections.
derived according the following equations: RT1. Maximum downward force in pile:
T
S= Jvpdt
t=O
71
RT2. Simple force - time curve: Blowcount n= 495 blows/0.25 m. Critical, very
close to refusal.
Check SRD,,,= 8376 kN (Equation 36)

RT3. Downward force after change of cross sec- Example 2

tion: Data as example 1.
Now larger hammer Menck MRBS3000; hardwood
cushion in pile cap.
F,= 13004 kN (Equation 16); 0=157 N/mm2.
F,= 13004 - Xx13788 -Xx1240 5490 kN
RT4. Dynamic friction (section 12) Set s= 4.18~10”m if T= 0.006 s (Equation 33)
1 Blowcount n= 60 blowd0.25 m. Fair driving.
Wdyn = Wstat Check SRD,,= 983 1 kN (Equation 36)
RT5. Set per blow:
Example 3
Data as example 1. Wall thickness= 32 mm;
(33) A,= 0.1 194 m2 ; Z,= 4 . 7 9 7 ~ 1 0Ns/m.
~
F,= 11890 kN (Equation 16); 0=100 N/mm2.
RT6. Maximum SRD: F,= 11890 - Xx13788 -Xx1791 =4100 kN
E Set s= 1 . 7 7 ~ 1 m
0 ~if~T= 0.006 s (Equation 33)
SRDmax = 0.8 q c vt ; Ap Only Point (35) Blowcount n= 141 blowd0.25 m. Critical.

E 18 CONCLUSIONS
SRDmax =z 0.7 q c vt; Ap Friction + Point (36)
- The maximum force F, of the downward propa-
gating wave can be calculated very close to the
Maximum SRD for Diesel Hammers related to a re- real value with a simple equation.
fusal blowcount of 120 blowd0.25 m: - If there is more insight in the time related pile

SRDmaX~2000Xmr [kN], mr inton (37) driving process, then there will be more under-
standing for the specialist.
For most Diesel hammers: m, = half total mass of - After understanding the process the simplification
hammer. (Example: Delmag type D62 mr= 6.2 ton; could be made more complex for obtaining better
SRD,,, z 12000 kN) results.
- Interpretation of soil values remains a special
profession and will therefore influence the re-
17 EXAMPLES USING RULES OF THUMB sults.
Example 1
Given: Hammer Delmag D 100, cushion: Plastic 19 REFERENCES
dolly; Pile open steel tube, diameter D= 1.22 m
Voitus van Hamme, G.E.J.S.L. & et al. 1974. Hydroblok and
(48”), wall thickness = 22 mm, length= 28 m; Final improved pile driving analysis, De ingenieur, nr. 18, vol.
penetration 26m depth. 86, May : 345-353.
Soil: sand , CPT-values Depth 0-12 m De Sitter, W.R. & Lucieer, W.J. 1985. Measurements of local
oc=2.5 MN/m2 =Friction 2%= 50 kN/m2, Depth friction during driving. In Holm et al. (ed.) Proc. Second
12-36 m o,= 15 MN/m2 =Friction 2%= 300 kN/m2. intern. con$ on the application of stress-wave theory on
Then: piles. Stockholm 27-30 May 1984. Rotterdam: Balkema:
25-32.
nD= 3.83 m; A,= 0.0827 m2 ; Z,= 3 . 3 2 4 ~106Ns/m. Lucieer, W.J. 1985. Performance of single acting Diesel ham-
F,= 1240 kN; mer. In Holm et al. (ed.) Proc. Second intern. con$ on the
WStat=3.83(12~50+ 14x300)= 18384 kN outside application of stress-wave theory on piles. Stockholm 27-30
and inside; May 1984. Rotterdam: Balkema: 244-25 1.
Wdyn Xx18384 + Sx18384 = 13788 kN;
SRD= 15028 kN.
Wdyn ?4~18384 f Sx18384 = 13788 kN;
SRD= 15028 kN.
F,= 91 10 kN (Equation 16); Maximum stress
G= 110 N/mm2.
F,=9110 - Xx13788 -1/2~1240=1596kN
Set s= 0 . 5 0 ~ 1 0m
- ~ if T= 0.006 s (Equation 33)

72
2 Driving equipment and recent developments
- New technologies for quality assurance of piles
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

ynote lecture: ile driving equipment: Capabilities and properties

E Rausche
Goble Rausche Likens and Associates Incorporated, Cleveland, Ohio, USA

ABSTRACT: Deep foundations are expensive: they are labor and equipment intensive and pile material is
costly. If they are necessary, they become a significant part of the total cost of a project. For this reason, new
installation methods and associated machinery are continuously being developed. For prefabricated piles, the
installation method is still the age old hammering with or without driving aids such as jetting, spudding or
predrilling. Vibratory hammer installation is a newer development that promises great improvements in
productivity. For cast-in-place piles the development has gone to larger equipment with more powerful drills
and automatic casing machines or improved wet drilling techniques. However, this paper will only examine
the current state of the art of impact and vibratory pile driving equipment, summarizing what is available
today, what additional improvements can be expected in the near h t u r e and how this equipment can be
selected to optimize installation and return on investment.

1 HISTORY OF PILE DRIVING EQUPMENT
In the beginning there was the drop hammer. Pulled
up by humans (Fig. I), horses or other creatures it
was suddenly released and then pulled up again.
Guides or leads for the piles were always an
important part of the driving equipment.
The nineteenth century saw the general adaptation
of steam engines to a variety of tasks, including pile
driving. A first development the replacement of
human or animal power by steam power to pull the
rams up with a winch. Today, there are still many
winch driven drop hammers in use, typically with
ram weights between 1 and 5 tonnes. However,
steam power has generally been replaced by electric
motors. diesel or gasoline engines.
Figure 1 . Artist's concept of an early pile driver.
At the end of the 191h century, the mechanical
hammer was born. Steam pressure pushed against a
a piston attached to the ram. Then, after tripping a
valve, the steam pressure was vented and this
allowed the ram to coast up to its highest point and
then fall with relatively little energy loss. Even
today steam is still being used as a power source for
large hammers with ram weights up to almost 200
tonnes. However, more convenient air compressors
gradually replaced the steam boilers on hammers
with typically less than 15 or 20 tonnes ram weight.
Following the developments in general machine
industries, it was obvious that double acting steam
hammers should be built. They increased the
operating speed of the hammer, but also wasted a lot
of power. The differential hammer reduced that
problem. Ram weights of double or differential
acting hammers were limited to approximately 10
tonnes.
Adapting steam driven hammers to the more
convenient compressed air operation required little
change of the basic mechanical hammer. In contrast,
the next development, using hydraulic power,
required some new thinking. Merely repIacing steam
or compressed air by pressurized hydraulic fluid did
not succeed because the "venting" of hydraulic fluid
was not as simple as that of air or steam. Such
hammers suffered pre-admission problems, meaning
the motive fluid was admitted into the chamber

75
 before the hammer impacted, thereby self cushioning
the ram.
The first successfbl hydraulic hammers were
developed in Scandinavia in the 1960s for the
installation of regularly reinforced piles. The best
machine that met this objective had a heavy ram and
a short stroke. This was accomplished with a simple
hydraulic jack for a short upward ram motion
(typically at most 1 meter) and then a free fall of the
ram after quickly pulling the jack downward. These
machines have ram weights of nearly 20 tonnes.
They are being built by Banut, Uddcomp and others.
A parallel hammer development was engineered in
Holland and Germany for the offshore hammer
market. Their innovative design approach created a
computer controlled double acting hydraulic
hammer. The computer is provided with sensor
information of ram position and ram speed and
accordingly adjusts the hydraulic pressure for an
operator selected ram impact energy.
Self-monitored hydraulic hammers have been built
with ram weights reaching 150 tonnes by IHC and
Menck. Being hlly enclosed, they can be used
underwater which requires solving a variety of
additional problems as described below.
A variety of hydraulic hammers which either
incorporate features of both the drop hammer and the
computer controlled type or which cannot be strictly
put in either of those two categories have also been
developed. Some of these hammers are only
optionally fitted with sensors for computer control or
they are drop hammers with some downward
pressure assistance. Manufacturers of these
hydraulic hammers are BSP, HPSI, ICE, Junttan and
others.
Diesel hammers replaced the external power
source winch, steam, air or hydraulic fluid with an
internal combustion underneath the ram. Starting in
the 1920s in Germany, the first diesel powered rams
typically had a mass of '/2 or 1 Mg. Today, ram
masses of single acting hammers range from 0.2 to
20 Mg.
A double acting diesel hammer that limited the
high diesel hammer strokes and improved hammer
speed by creating a passive air pressure chamber at
the top of the hammer was also developed in the
United States after World War 11. Today closed end
diesel hammers are built by ICE with ram weights
up to 5 tonnes.
The vibratory hammer was developed in the
Soviet Union prior to and during World War 11.
These hammers were first built with relatively low
frequencies (less than 20 vibrations per
second[WS]) and were electrically powered. (After
the war US, French and other companies hrther
developed the low frequency hammers which today
generate centrifbgal forces of up to 4000 kN and
now use hydraulic power almost exclusively.) The
76
latest development allows for a gradual increase of
the eccentric moment which is advantageous during
the start-up of the hammer.
A French and later US development was the
Bodine hammer, a diesel engine driven Sonic Pile
Driver with 60 and higher VPS. Such frequencies
created resonance in piles with length of 20m or
more. Resonance in turn would greatly improve the
speed of pile penetration. Because of mechanical
problems this concept is still somewhat like a
utopian dream.
For the sake of completeness the Silent Pile Driver
which was developed in the last decades of the 20th
century should also be mentioned. It is really a
static machine which uses hydraulic pressure to push
piles into the ground. The required reaction forces
must be provided by previously installed
neighboring piles. The system is therefore
advantageous to sheet pile installation.
The so-called driving system includes a hammer
cushion, helmet with inserts (cap) and pile cushion.
While softwood and straw have been practically
abandoned for hammer cushions and pile cushions
respectively, little other progress has been made to
improve driving systems. Exception is made to the
systems that have eliminated hammer cushions
entirely.
The guiding systems also have undergone some
important development. Dedicated systems prior to
World War I1 were often barge or skid mounted and
had little freedom for adjustment. European
developments combined carrier and leads into a
dedicated system that requires little set-up time.
This was particularly advantageous where pile sizes
and lengths did not vary much, such as was the case
with the European segmental reinforced piles. In the
US, a greater variety of pile and hammer types made
the combination of simple pile guide with a general
purpose crane (often rented) more feasible. In the
nearshore and offshore environment, free riding
leads are used almost exclusively to align hammer
and pile while a template or jacket provide support
for pile and hammer.
Throughout the history of pile driving, a great deal
of creativity was expended on the development of
driving aids. The objective was the saving of pile
material while improving driveability. Mandrels,
jets, spuds, pre-augers, friction reducers and other
means were developed, improved, optimized and
occasionally abandoned.
2 MECHANICS OF IMPACT PILE DRIVJNG
Pile driving hammers are generally classified by
their potential energy prior to the ram beginning its
downward movement. The kinetic energy available
to do work on driving system, pile and soil is a
certain fraction of that potential energy. The ratio of
kinetic to potential energy is the hammer efficiency.
This concept of rating the hammer by a potential
energy and calculating the available kinetic energy
using a reduction factor has been replaced by a
measured kinetic energy for those hammers that
monitor ram position and speed immediately prior to
impact.
Energy is not the sole quantity that defines the
ability of a hammer to move the pile into the ground.
An alternate measure would be its momentum or the
product of ram mass and impact velocity. While this
physical quantity would better indicate the
maximum force that the ram can generate in the pile
and therefore the soil resistance force that it can
overcome, it is a measure that is only useh1 for
comparing hammer capability in very hard driving
cases.
Impact velocity is, however, one of the most
important parameters of the impact event. In the
absence of a cushion and helmet mass between ram
and pile top, the particles of the pile assume the
same velocity, vi, as the ram during the first instant
of the impact event. The corresponding force can be
calculated from
F = viz (1)
where Z is the pile impedance Z given by
Z=EA/c
with E being the pile's elastic modulus, A its cross
sectional area and c its wave speed given by
c = (E / p)" ( 31
with p being the mass density of the pile material
The presence of a cushion reduces the peak force
over time and spreads it over a greater time. A
helmet mass has a similar effect (Rausche et al.
1972). Even so, the maximum force in the pile at the
beginning of the impact event is primarily dependent
o n the impact velocity.
The greater the mass of the hammer, the greater its
ability to maintain the pile top force near the
maximum impact force. Therefore, a small hammer
will only generate a short force pulse which may be
ineffective in maintaining a sustained downward
movement and therefore pile penetration. On the
other hand, a slowly falling ram does not generate
the force necessary to overcome soil resistance.
Piles are relatively long, slender elastic bodies and
for this reason, their dynamic behavior is governed
by wave effects. In other words, in the first instance
after the ram has impacted, only the top of the pile
moves downward. It takes a certain time, the wave
travel time L/c (L is the pile length) before other pile
particles firther down the pile start to move. At the
77
pile toe a reflection occurs that depends on the
resistance that the material at the pile toe offers. If
there is no resistance at all the pile toe will move
twice the distance that the top moved during impact
and it will then pull the upper pile particles
downward. This creates a tension force which is
potentially damaging to concrete piles. If there is
an extremely high soil resistance such that the pile
toe cannot move then the force at the pile toe will be
twice the force that the pile top experienced during
the impact event. Theoretically, therefore, the
hammer can overcome a force which is twice the
impact force of Equation 1. In practice, because of
the need for all materials to compress before they
can exhlbit resistance, the limit of the soil resistance
is about 1.4 for piles with predominant end bearing
piles and 1.O for shaft resistance piles.
In the case of the high toe resistance a strong
compressive wave reflection occurs which pushes
the upper pile particles upwards. For large ram
weights, the ram will still have downward
momentum when the reflected wave reaches the pile
top. In the case of the hard driving pile it is possible
that another compressive reflection occurs at the pile
top which may cause the pile top force to grow
above its initial peak. (In addition, a new wave will
cause downward pile motions.) The benefit of a
large ram is therefore a longer lasting downward
motion of the pile, a fbrther increase of the pile toe
force and the chance for additional permanent pile
penetration into the ground.
Since it is uneconomical to use very large ram
weights, we normally do not see cases where the soil
resistance is much greater than twice the impact
force. However, damage at the pile toe when it is on
rock can easily happen when either the impact force
is very high or the ram mass is large.
Calculation of pile penetrations into the ground
and stresses along the pile are difficult and
inaccurate when considering simplified systems and
employing wave propagation theory. It is better to
use a wave equation computer program, such as
GRLWEAP (GRL 1999) to realistically model
hammer, driving system, pile and soil, and
mathematically simulate the pile driving process.
Obviously, such models require assumptions
regarding the performance of the individual
components of the system and it is therefore best to
measure forces and velocities at the pile top and
calculate the desired quantities based on these
measurements using CAPWAP (GRL 1996).
3 HAMMERPERFORMANCE
MEASUREMENTS
Pile Driving AnaZyzedC
The basis for the results calculated by the PDA are
pile top strain and acceleration measurements which
are converted to force and velocity records,
respectively. The PDA conditions, calibrates and
displays these signals and immediately computes
average pile force, F(t), and velocity, v(t), thereby
eliminating bending effects. The PDA calculates a
variety of results for each hammer blow, including
pile bearing capacity, pile stresses, pile integrity and
other important pile quality parameters using closed
form Case Method solutions based on the one-
dimensional linear wave equation.
The PDA also calculates the energy transferred to
the pile top from:
E(t) = ,,jtF(z)v(z) dz (44
The maximum of the E(t) curve is the most
important information for an overall evaluation of
the performance of a hammer and driving system.
This EMX value, also called ENTHRU, allows for a
classification of the hammer's performance when
presented as the rated transfer efficiency, eT, also
called energy transfer ratio (ETR), system efficiency
or global efficiency:
= EIW(/ER (4b)
where ER is the manufacturer's rated energy value.
Hammer Pwfix-niance Ana@er TM
The ram velocity may be directly obtained using
radar technology in the Hammer Performance
Analyzer ( W A ) For the HPA to work the ram must
be visible. The impact velocity results can be
automatically processed with a PC or recorded on a
strip chart. The result of the HPA is the velocity
immediately preceding impact. The kinetic energy
can be calculated directly from this measuremen!
(see appendix).
Smimeter nw
For open end diesel hammers, the time between two
impacts indicates the magnitude of the ram fall
height or stroke. This information is not only
measured and calculated by the PDA but also by the
convenient, hand-held Saximeter.
Both Saximeter and PDA calculate the stroke
(STK) of an open end diesel hammer using
STK = (g/8) TB2- h,
where
g is the earth gravitational acceleration,
TB is the time between two hammer blows,
hL is a stroke loss value due to gas compression
and time losses during impact (usually 0.3 ft
or 0.1 m).
70
A modification of the Saximeter is the Saximeter-
E which accepts timing signals from two proximity
switches installed in the hammer such that the time
between these two signals is a measure of the ram
velocity immediately preceding impact. Again, this
measurement can be simply converted to hammer
kinetic energy. The signals from the proximity
switches are sent to the Saximeter-E using telemetry.
Pile Installation Recorder nwfor Driveii Piles
(PIR-0)
The P E - D accepts signals from a variety of sensors,
among them time of occurrence of hammer blow,
time when ram passes two positions just before
impact (as for Saximeter-E), pile depth of
penetration, bounce chamber pressure. These
signals are evaluated and plotted vs depth as follows:
stroke for open end diesels,
impact velocity of all types of hammers, and
bounce chamber pressure for closed end
diesel hammers as an indirect measure of
stroke.
4 TECHNICAL DETAILS OF IMPACT
HAMMERS
Cable siisperided drop hammers
The system generally uses a crane mounted winch
for ram raising. Releasing the brake on the winch
causes the ram to drop. Of course, the falling ram
has to pull the winch cable and therefore accelerate
the winch which costs energy. Immediately prior to
impact, the crane driver is often tempted to reapply
the brake to prevent a spilling of the winch cable.
This action cushions the impact. It therefore must be
expected that approximately 50% of the potential
hammer energy is lost. Up to 3 m drops are not
uncommon. In general, these hammers apply
impacts of inconsistent magnitudes. Considering
nominal fall height times ram weight the full energy,
then energy losses are due to
inaccurate drop height,
fiction,
a winch drum inertia,
early catching of the ram, and
misalignment.
Free release drop hammers
Such a hammer is probably only found for special
applications, e.g. dynamic load testing. Free release
devices, although not a real technical problem, are
not a practical solution for normal pile driving tasks.
In fact it is advisable to transfer the support of the
ram from the crane to a stiff device such as the leads
prior to releasing the ram to avoid damage to the
crane. In this way a precise hammer fall height can
be achieved. Of course, the hammer efficiency
would be expected to be higher than that of the
winch brake released hammer. The energy losses are
then only due to
0 guide friction and
0 misalignment.
The hammer efficiency therefore may be assumed
to be quite high, e.g. 95%.
For those involved in dynamic load testing, the
development of free release devices has frequently
been of interest. The following solutions have been
encountered:
0 camhook,
Spread hooks, and
0 cut-off (torch, hydraulic shear) support cable.
Depending on the height, angle and friction factor
of the cam, the force to release the weight will be a
small percentage of the drop weight. Up to certain
loads such devices are commercially available.
The tests described by Seidel et al. (1982) utilized
a double hook device from which a 20 tonnes load
could be released by hand.
As mentioned earlier, whenever a heavy ram is
suddenly released, care must be taken that the energy
stored in the supporting carrier is carefblly released.
In a winch release this is automatically the case
because the weight is only gradually released.
However, a sudden free release should either be
done after the support of the weight is transferred
from crane to ram guide or by utilizing a ballast
between release device and crane hook. The inertia
force of the ballast slows the release of the energy
stored in the crane system.
Single acting aidsteam hammer
It was the mechanical steam hammer that simplified
the hammering process significantly. The ram was
simply attached to a single acting steam engine that
pulled the ram up. Near the top of the upward
motion the valves were switched, releasing the
upward pressure by venting the chamber. The ram
would coast upwards an additional distance before it
started to fall under gravity. Shortly before the
downfall ended the valves were switched again.
This allowed motive fluid to enter the chamber,
generate an upward directed force on the ram and
produce a new upcycle. Of course, allowing the
79
motive fluid to enter the chamber too early would
effectively cushion the impact of the ram prior to
impact (preadmission). This effect is equivalent to
the operator of a drop hammer applying the break
too early. If all settings are carefblly adjusted the
losses would be small, however, losses due to
preadmission are also possible if the cushion block
were too thin allowing for a longer than design drop
aRer the valves had been switched. On the other
hand, a very thick cushion block would reduce the
stroke and, in the worst case, prevent switching of
the valves. Then the hammer would not run at all. It
should also be mentioned that a sufficient aidsteam
pressure is important such that the ram coasts up to
the nominal fall height after the valves have vented
the chamber. In fact, in easy driving the upward ram
motion is sometimes so low that the hammer short
strokes (but in that case it is not a problem.).
There are two different arrangements of the steam
engine. One type attaches the piston directly to the
ram (Vulcan, Conmaco, MKT) and makes the
cylinder stationary. The other one uses a hollow ram
making it a movable cylinder while the piston is
stationary (Menck MRBS). There is another
difference between these two hammer types: The
former maintains constant pressure under the piston
until the chamber is vented; the latter works with a
constant volume of steam charge which is then
allowed to expand. Obviously, the former wastes
more energy but is faster, while the latter is more
economical.
Single acting air hammers are not much different
from steam hammers except that steam is hot and
compressed air is colder. The different temperatures
of these two motive fluids require that
manufacturing of piston and cylinder is done with
different tolerances because of the differences in
thermal expansion.
It is important that aidsteam hammers are
lubricated both internally with a lubricant injected
into the motive fluid and externally along the ram
guides. Energy losses primarily occur because of
0 guide and pistodcylinder friction,
0 low stroke in easy driving,
0 less than rated stroke when pressure is low,
0 preadmission, and
0 misalignment.
Typically, hammer eficiencies are assumed to be
67% for these hammers for wave equation analyses.
Transfer efficiencies recorded at the end of driving
average 55% on steel and 40% on concrete, and
timber piles (Figs. 2-3).
Drop heights of single acting air hammers
typically vary between 900 mm and 1500 mm. The
stroke of the hammer is generally controlled by a

Figure 2 . Transfer eficiencies of single acting
aidsteam hammers on steel piles at the end of
driving.
Figure 3. Transfer efficiencies of single acting
aidsteam hammers on concrete and timber piles at
the end of driving.
slide bar which is attached to the moving ram.
Several of these hammers have been equipped with a
variable stroke attachment which allows for
alternating between two different strokes. This is
particularly useh1 for concrete pile driving where
tender care is required in easy driving and high
It is not
energies are needed in hard driving.
uncommon that the hammer overstrokes on the short
stroke and therefore operates at apparently higher
efficiency. It is therefore reasonable to increase the
hammer efficiency (e.g. to 0.75 from 0.67) of short
stroked hammers in a wave equation analysis.
Double and differential acting aidsteam hammers
In order to generate hammer blows in quicker
succession it is desirable to shorten the stroke and
accelerate the ram on its downward motion by
pushing downward with active pressure. This has
been accomplished in double or differential acting
hammers. The difference between these two engine
80
types is subtle for the civil engineer: the former
being somewhat more power intensive than the
latter. While single acting hammers run typically at
50 to 60 blows per minute, double or differential
hammers run at 100 blows per minute or more.
There is an obvious limitation on the amount of
pressure that can be exerted onto the ram. If this
pressure would create a force in excess of the weight
of the hammer assembly itself (total hammer weight
minus ram weight) then the hammer would lift off
from the pile and an unstable driving situation would
result. Thus, hammer weight and operating pressure
must be matched.
During hard driving when the pile rebounds
strongly, the ram obtains a high upward velocity
which necessitates reduction of the operating
pressure. This means that the pressure is also
reduced during the downward cycle, causing a
reduced energy of the impacting ram. For this
reason, this type of hammer has on the average a
lower efficiency at the end of driving (when driving
is relatively hard) than other hammer types. The
following energy losses have to be expected:
0 guide friction and friction on piston in
cylinder,
e reduced stroke in easy driving,
e reduced stroke due to reduced pressure in
hard driving,
0 reduced ram velocity due to reduced pressure,
+D preadmission, and
misalignment.
Normally, for wave equation analyses a hammer
efficiency of 0.5 is therefore reasonable. Transferred
energies average 35% on steel piles and 30% on
concrete piles (Figs. 4-5).
These hammers run vary rapidly and have been
made very small for many purposes, such as jack
hammering. For pile driving their ram weights vary
typically between 2 kN and 90 kN.
DieseI hammers
In the 1920s, DELMAG built the first open end
diesel hammer which utilized diesel combustion
underneath the ram to provide the energy for the
upward ram motion. Thus, the ram serves as the
piston of a 2-cycle diesel engine. The advantage of
this hammer is a relatively light weight, sturdy
construction and no need for an additional external
power supply.
The hammer consists of four parts: a cylinder, a
ram or piston, an impact block that closes the bottom
of the cylinder and a fbel pump that injects he1 into
the cylinder prior to impact. There are four phases
that this hammer type undergoes prior to, during and

Figure 4. Transfer eficiencies of double acting
aidsteam hammers on steel piles from end of drive
measurements.
Figure 5. Transfer eficiencies of double acting
air/steam hammers on concrete piles from end of
drive measurements.
after the hammer impact. During the first phase the
ram first rises and then falls only subject to gravity.
When on the downstroke the ram closes the exhaust
ports. The air underneath the ram is being
compressed, and thus raises its temperature. This is
the second or compression phase. Impact occurs and
injected fuel combusts in the third, the impact and
combustion phase. Finally, the ram is driven
upwards during the expansion phase that ends when
the excess pressure is vented through the exhaust
ports and new air is sucked into the cylinder during
the ram’s upward motion which - being only subject
t o gravity - is part of the first phase of the next
hammer blow.
Open end diesel hammers are made or distributed
by APE, Berminghammer, DELMAG, and ICE.
Other makes still exist but they are no longer
manufactured.
Closing the top of the cylinder traps air above the
piston which causes the ram to move up and down
more rapidly (typically 80 blows per minute vs 40
blows per minute for the open end diesel hammer).
Like double acting air/steam hammers, the
maximum pressure on the piston is limited by the
hammer weight. Therefore the maximum stroke is
limited to that value that corresponds to a pressure
that causes an uplift condition. The manufacturer
rates closed end hammers somewhat below an open
end diesel of the same ram weight. Double acting
diesel hammers are made by ICE and MKT.
The behavior of diesel hammers is not as simply
predicted as that of other mechanical hammers. To a
much greater extent than airlsteam hammers the ram
drop height of diesel hammers is dependent on the
resistance that the soil offers. Typically in a refusal
situation the ram reaches its highest potential energy.
This is actually a desirable quality, since the hammer
strikes rather gently when driving is easy. In fact, it
maintains a certain minimum compressive pressure
on the pile from either gas compression or expansion
that protects concrete piles against damaging tension
stresses. However, the stroke variability also
complicates the prediction of energy output.
Another complication is that the compression of the
air charge (phase two) requires energy that is
temporarily stored in the compressed air and later
released, together with the combustion energy,
during the upward motion of the ram. The energy
stored in the compressed air is a hnction of the
compression ratio of the hammer, typically ranging
between 12 and 15. Typically, the compression
energy is between 20 and 30% of the potential
hammer energy. It may be much higher if the
hammer pre-ignites, which may happen if the
hammer is poorly maintained or overheats to such a
degree that either the fuel oil or the lubrication oil
combusts prior to impact. The ram then falls against
a very strongly increasing pressure which reduces
ram velocity plus kinetic energy. Note however, that
some energy already gets transferred to pile and soil
during the pre-compression phase.
Pre-compression is only possible if the pile can
offer a resistance that matches the compression
forces. For this reason, the hammer will not run if
the soil resistance is very small.
In diesel hammers, energy losses are typically
caused by:
pre-compression,
friction between ram and cylinder,
a impact and inertia losses of the impact block,
low stroke,
pre-ignition, and
misalignment.
 Considering pre-compression a design loss that
can be calculated, the remaining losses are typically
less than 20% of the potential energy. The appendix
contains a numerical example for the calculation of
pre-compression energy.
Ram weights range from 2 kN to 200 kN, rated
energies from 0.4 to 600 kJ. The hammer efficiency
typically averages 0.8. Transferred energies have
medians of 3 1 56 on steel piles and 25 % on concrete
piles (Figs. 6-7).
Hydrazilic drop hammers
Hydraulically powering hammers has become
attractive with the development of powerful
hydraulic pumps and other accessories. A direc!
conversion of aidsteam hammers to hydraulic
hammers proved unsuccessfd. because of the
problem of preadmission for incorrect valve settings.
82
The first successfhl hydraulic hammers were
simply drop hammers: a jack would pick up the ram
and then retract quicker than the falling ram. Since
these hammers were at first built with low drop
heights, energy losses were extremely small.
Examples are the Banut, ICE, Junttan, Menck MHF
series, Uddcomb and other hammers. Ram weights
are as high as 150 kN and rated energies reach 200
kJ.
Energy losses are caused by:
* guide friction, and
misalignment.
whereby guide friction is often insignificant due to
the short strokes.
Self-monitored hydraulic hammers
The MC and Menck MHU hammers are built
differently, however they share features which make
them very similar. A piston rod and piston are
connected to the ram. The piston is pushed up by
hydraulic power and downwards from its upper side
by active pressure. Electronics sense the impact
velocity and count hammer blows. A computer
outputs energy and blow numbers.
The energy that is used to control pile driving with
this hammer is the kinetic energy contained in the
ram just before impact. This energy is adjustable by
continuous variation of stroke andlor downward
pressure. The hammer’s rated energy is therefore a
net maximum value, after most losses. Hammer
efficiency for this type of hammer is therefore
defined differently from others: it only covers losses
occurring during the impact. The energy necessary
to accomplish the ram impact velocity is
inconsequential and therefore not monitored. It is
not surprising that the hammer efficiency is usually
set to 95%. This covers such relatively insignificant
losses as
* measurement errors of impact velocity, and
* misalignment.
This hammer type, being remote controlled,
completely enclosed and having an external power
source supply that can be supplied through hoses,
can also be employed under water. This explains
why the manufacturers removed the need for
cushioning by perfect machining of the impact
surfaces. Frequently used in offshore applications,
ram weights of these hammers reach 1500 kN.
Other hj*drmiichammers
There are a variety of other hydraulic hammers
available today. For example, BSP makes hammers
that are operated either with or without active
downward pressure. They may or may not include a
monitoring equipment. The variety of hammers in
this category preclude general statements of energy
losses and recommended efficiencies. Indeed,
Figure 8 shows that the combination of all hydraulic
hammers tested by the author’s firm generate an
efficiency diagram with two distinctly different
groups of performance: a lower area around 55% and
a higher one with an 85% mean. Additional data is
needed to produce a clearer classification of these
hammers.
Figure 8. Transfer efficiencies for a variety of
hydraulic hammers on steel piles.
5 DRIVING SYSTEMS
Driving systems are defined as the accessories
necessary to transfer the impact forces safely to the
pile top. They include a cushion for the protection of
the hammer, a helmet with adaptor inserts and a
cushion to protect concrete piles.
Hammer cushions
With a few exceptions, hammer impact is usually
cushioned by a relatively resilient material. The
reason is the need of protecting the ram from
excessive fatigue stresses. Often man made
materials are used and they require little exchange as
long as the cushion is protected by a well fitting steel
striker plate and is sandwiched between aluminum
sheets for heat extraction. If they are made of a stiff
material (softwood is not acceptable as a hammer
cushion), hammer cushions store and dissipate
relatively little energy.
Driving systems without hammer cushions
Certain hammer manufacturers have decided that it
is unnecessary to protect the hammer against high
83
pile stresses. Indeed, diesel hammers whose rams
strike the impact block directly and which have been
in existence for decades without damage to ram or
impact block have proven that well fitting impact
surfaces generate tolerable stresses. These hammer
manufacturers therefore devised a cushionless
driving system which features a well fitting and
machined impact surface for the ram such that high
contact stresses are avoided. This idea was first
incorporated in the LHC hydraulic hammers.
Pile cushions
Pile cushions are only necessary for concrete piles
where they serve two purposes. First they must
reduce the likelihood of stress concentrations at
points of steel-concrete contact. Second, they are
often designed to reduce the peak stress in the pile
by spreading the impact forces over time.
Historically, when hammers had limited drop heights
of less than 1 m, only the reduction of contact
stresses was important and straw was often
successfblly and economically employed. Today,
hammers with high strokes and/or high efficiencies
also require that the second condition be satisfied
and that requires a cushion block with low stiffness.
Plywood hlfills both purposes quite well, however,
it only lasts for about 1500 hammer blows. ,This
limit can be a problem when driving for an extended
time period with a follower under water where a
cushion exchange is not possible. Using thick
cushion blocks to lower the impact force by
spreading it over time leads to a rather wastehl use
of plywood. A combination of thin plywood
sandwiching a somewhat more resilient material can
then be more economical. Hamortex, a man-made
material consisting of rolled-up, aluminum coated
paper has been successfully employed for this
purpose.
Pile cushions are also made of a variety of other
materials. For example, in Latin America sisal and
hemp rope have been successfully used though with
greater energy losses than plywood.
He Ime ts
Helmets must align hammer and pile and securely
spread the impact forces over the pile top.
Depending on the type of pile, they must be fitted
with inserts that match the pile profile. Helmets also
accommodate the hammer cushion. It is important
that the hammer only allows for a small lateral
tolerance (e.g. 25 mm) on the pile top for good
hammer pile alignment.
 6 PILEGUIDES
Guiding of the pile is an important part of the pile
driving equipment. Historically, dedicated systems
with a stiff leg and a skid have fulfilled this role and
these systems are still used for some very large
projects. However, these systems are not very easy
to mo\re and therefore reduce productivity.
Swinging leads
A very simple system is the swinging lead which can
be cable supported from a multi-purpose crane and
which is stabbed in the ground for a second point of
support. Obviously, such a system is relatively
flexible, a feature that sometimes protects piles from
breakage when the pile encounters an obstruction
and wants to move laterally.
Fixed leads
Fixed leads are supported at the boom tip of the
crane with a hinge and near their bottom by a brace.
Normally, this brace can be extended or contracted
hydraulically to set the batter. Fixed leads are
probably the most common support system and
many dedicated lead-hammer systems that use a
dedicated carrier also achieve mobility and control
with such a lead type. Often these systems can adapt
to a compound batter.
Freeridiilg leads
Over water, piles are often driven through a template
which assures the correct location and direction of
the pile. In such a case leads are merely aligning the
hammer with the pile. These "caison-type" or
freeriding leads are supported at their top by the
crane. A disadvantage is for battered piles that the
transverse component of hammer and lead weight
has to be supported by the pile, thereby introducing
potentially high bending stresses in the pile which
are superimposed to the dynamic stresses from pile
driving.
7 UNDERWATER PILE DRIVING
Under water pile driving in water depths of 1000 m
or more can be done, however, only aRer a number
of formidable problems are solved. For example,
hammers and piles have to be run through the same
guiding system. The hammer outside diameter
therefore should not be greater than the pile
diameter. In fact, it is sometimes necessary that a
pipe extension is added to the hammer top for
additional guiding once the pile has been driven
below the lowest guide
84
Supplying high hydraulic pressures at high rates
of volume through long hoses pose another
challenge. One solution would be a remote hydraulic
power inside the guiding system of the hammer.
That means that the power pack also has to fit
through the guides and that electric power is
supplied over a long distance.
Ambient pressures are, of course, very high in a
1000+ rn depth environment. This means that the air
surrounding the moving ram becomes quite viscous
and care has to be taken that air passages are large
enough so as not to impede the ram movement. Of
course, the hammers' sensors will be able to indicate
the actual kinetic energy.
Finally, driving against a water column inside the
pile would dissipate valuable energy. Thus, excess
water must be allowed to escape through holes in the
pile near its top and an air cushion should be
maintained beneath the helmet.
8 VIBRATORYHAMMERS
An impact hammer generates a short duration
forcehelocity pulse in the driven pile. Between
hammer blows the pile and soil are practically at rest
for a time (typically 1 second) that is large compared
to the impact event itself (typically 20 to 50 ms).
The impact hammer therefore generates a complete
cycle of loading and unloading during the short
impact event
First built in the Soviet Union in the mid 20th
century, vibratory hammers became the installation
device of choice for non-bearing piles, because of
their superior speed of pile installation in many soil
types.
Vibratory hammers consist of pairs of rotors with
eccentric weights attached. Always two rotors are
synchronized such that their horizontal centrifbgal
force components cancel while their vertical ones
add. The vibratory machine therefore obtains an up
and down sine motion with a frequency equal to the
speed of rotation Obviously, the maximum vertical
force is the same in upwards and downwards
direction with a downward bias given by the weight
of the machine itself and the pile. If the eccentric
moment of all rotors in the machine is ME and the
frequency. f. then the downward maximum dynamic
force is
Originall\,. iibratory hammers had a relatively low
frequency of less than 20 Hz (1200 FWM), were
electrically powered, and of moderate size. Today
these standard machines are powered with hydraulic
aggregates reaching 600 kW (800 HP) with
centrihgal forces of up to 4000 kN (800 kips) and
frequencies typically up to 30 Hz. Larger outputs
may be achie\,ed with two machines running in
tandem.
Resonant pile drivers
The vibratory pile driver’s dream has always been to
achieve even better penetration rates with a machine
that would generate resonance in the hammer-pile
system. Resonance would generate forces much
larger than the centrihgal force of the hammer.
Such a high frequency machine has been built, used,
improved and retried, generally with the result that
pile driving indeed became very quick, however, the
machine would suffer and frequently break down.
The Boudine pile driver is equipped with two diesel
engines driving the eccenters directly at the pile top.
It is capable of frequencies up to 80 Hz which would
generate resonance in a pile of 15 to 30 m length,
depending on the relative weights of pile and
machine and on the resistance magnitude and
distribution.
Variable moment hammers
An interesting aspect of vibratory pile driving is the
response of pile and soil to various frequencies.
Typically, for somewhere above 3 and below 10 Hz
a resonance between pile and soil occurs which
generates strong soil motions. Nearby buildings can
sustain damage due to these vibrations. When
shutting the machine off, the standard vibratory
hammer again generates these lower damaging
frequencies. In addition, while higher frequencies
tend to densifL the soil, low resonant frequencies
may actually cause a loosening of denser materials.
It is therefore highly desirable to make a hammer
that does not cause a low frequency resonance and
since no motor can suddenly run at a high speed the
only solution is to increase the eccentric moment
aRer the desirable hammer speed has been reached.
Such variable moment machines are now being
produced in Europe. They are also effectively used
for soil densification (Massarsch 1992).
9 INSTALLATION AIDS
It is often not wise to force a pile through harder
layers into softer materials or to give the pile a much
higher strength than necessary merely for
driveability. Instead it may be possible to use
installation aids such as jetting or drilling. However,
such installation aids must be used wisely. They
may actually cause harm to pile and soil if not
properly executed .
85
Predrillir ig
In the United States one can often see a pile driving
hammer with a continuous auger attached on its lead
next to the hammer. Usually a hole that is only 80
percent of the pile size is drilled through the
potentially difficult overburden. It is important that
the hole is not drilled too deep. Strongly reducing
shaft resistance on a long pile and then driving
against a high end bearing may actually cause
damaging pile stresses and make the pile harder to
drive.
Jetting
Either on the side of a pile, or sometimes through a
center tube of a concrete pile, water may be used to
loosen and remove soil from around the pile tip. In
softer soils large cavities may accidentally be opened
up and that would, of course, detrimentaIly affect the
bearing behavior of the soil.
Spudding
Prior to driving large concrete displacement piles
(e.g. a 900 mm square) a hole can be driven with a
spud which is typically a steel section of the same
size as the concrete pile. The spud may be so heavy
that it can be repeatedly lifted and dropped to form a
hole in which the concrete pile is then inserted. The
concrete pile only needs enough hammer blows to
seat it into soil undisturbed by the spud.
Externalfric 1ion reducers
For pipe piles that have to penetrate competent soils
for deep penetrations it is advantageous to attach
shallow rings to the outer pile circumference. These
rings reduce the friction temporarily during pile
driving by creating a heavily disturbed zone in the
area above. Design details such as size of and
distance between the rings are either proprietary or
non-existent .
Internalfric fioii rediicers
For pipe piles. a so-called internal driving shoe, i.e. a
section of pipe with increased wall thickness but
identical outside diameter is attached to the pile
bottom. Above this section the soil penetrating into
the pipe during pile driving experiences reduced
lateral confinement and therefore reduced friction.
Electric potential
For offshore steel pipes that have to be driven into
clay, an electric current can produce a temporary
-- HHK 5 on Concrete Pile

-
Banui 7 tonnes on Concrete Pile GRLWEAP (TM) Version 1998.:

JUNllAN HHK 5 BANW 7Tonnes

Stoke 1 22 0 80 meter
3c,ency 0 950 0 Q50
ie1met 4 90 1 95 WJ
iammer C U S ~ M 1988 707 Wmm
JrleC~shlon 336 336 Wmm

skin Quake 2500 mm 2 500 mm

roe Quake 10 000 mm 10 000 mm
Skin Damping 0 650 Sedm 0 650 sedm
roe Damplog 0 500 sedm 0 500 sedm
'le Length 20 00 20.00 m
1
' 18 TOPArea 1600 00 1600 00 cm2

Sun Friction Sldn Fncnon

Pile Model Distribution Pile Model Dimibutlon

'i

00 (60 aa 45.0 E
m rra (aa Res. Shan = 50 % ReS.Shan-50%
Biow Count (biowsl.25m) (fIOpORlOn~lJ (Propomonal)
y r e 9.Bearing graphs for two different hydraulic s on the same pile.

-- DV 04632 on Steel Pipe

-
20onSteelfi~e GRLWEAP (TM) Vemion 199F3-i

- DELMAG D 4 6 3 2 VULCAN VUL 020

l &' 1 . : Emclency 0 800 0 670

Helmet 7 60 2400 W
HammerCushion 10687 931 Wmm

Skn Owke 2500 mm 2 500 mm

Toe Quake 2500 mm 2 500 mm
Skn Damping 0 650 seclm 0 650 sedm
Toe Damping 0 500 seclm 0 500 seclm
Pile Lsnpth 20 00 2000 m
PileTwArea 80000 800 00 cm2

Skn Fnnlan Sldn Fticnon

Pile Model DiSuibuIion Pile Modei Dimbution

Figure 10. Comparison of a diesel (D46-32) with an air steam hammer (V 020)

86
reduction of the strength of the clay. This can
dramatically reduce the installation time and effort.
However, this system apparently has not been
proven economical or safe enough for genera!
applications.
Steamfor permafrost
Pile driving in permafrost poses special problems.
The frozen ground behaves under hammer impact
like a soft rock. Resonant vibratory pile driving is a
actually a possible solution only limited by the lack
of reliable machines. For regular pile driving steam
injection thaws the permafrost temporarily and
therefore allows the pile to penetrate under normal
hammer impacts.
10 EQUIPMENT SELECTION
Selecting the proper equipment for pile installation
can make and break a project's technical and
economic success. An underpowered equipment
may cause excessive numbers of hammer blows and
pile fatigue and installation may not even be
possible. Selecting very powerfid equipment also
may cause damage and/or require unnecessary
capital expense.
While many "rules of thumb" have been tried and
dynamic formulas been used to estimate the hammer
size for a particular job, the wave equation, (e.g.
using GRLWEAF' [GRL 1999]), is still the most
reliable tool to size hammer and driving system for
reasonable blow counts and safe stresses. Very
briefly: for friction piles, blow counts should be less
than 80 blows for .25 m, for end bearing piles where
hard driving is expected to be of a relatively short
duration, this criterion can be relaxed to 200 blows
for .25 m. According to Hannigan et al, 1996,
compressive concrete stresses should be less than
85% of the concrete strength (after any subtracting
prestress). Tensile concrete stresses should amount
t o less than 70% of the equivalent yield strength of
regular reinforcement or prestress plus 50% of the
concrete tensile strength. For steel piles, stresses
should be less than 90% of yield.
A comparative analysis was made as a
demonstration of the value of the basic wave
equation output, the bearing graph. As shown in
Figure 8, a 20 m long concrete pile, 400x400 mm
square was anaiyzed under two somewhat different
hydraulic drop hammers. One of these hammers, the
Junttan HHK 5, has a ram mass of 5 Mg and a
maximum fall height of 1.22 m (maximum rated
energy 60 kJ). The other hammer, the Banut 7
tonnes, operates at a stroke of at most 0.8 m and
therefore has a rated energy of 55 kJ.
a7
Figure 9 shows that the heavier hammer, even
though it has a lower energy, requires approximately
20% less blow count to accomplish the same bearing
capacity. The compressive and tensile stresses were
considerably lower. Obviously, the lower stresses
are a result of the lower impact velocity of the
heavier hammer. The somewhat surprising result of
lower blow counts for the lower hammer energy is
caused by a relatively large quake (10 mm) at the
pile toe where 50% of the total resistance were
assumed to act. This large quake means that the toe
resistance has a high flexibility which can best be
overcome with a force pulse in the pile that extended
over a long time. The heavier ram does this b,etter
than the lighter one with the stroke.
Whde heavy rams do rather well in relatively easy
driving, high impact stresses help t o overcome a
higher soil resistance in hard driving. An example of
a diesel hammer with 4.6 Mg ram and an aidsteam
hammer with 9 Mg ram is shown in Figure 10.
Obviously the lighter hammer is performing better in
hard driving. This is not too surprising if rated
energies are compared which are 153 and 81 kJ,
respectively. In other words, the heavier hammer is
rated with a much smaller stroke (.9 m) than the
diesel (3.4 m). Obviously, an air/steam hammer
with a heavier ram and a higher stroke for
comparable energy ratings would make the aidsteam
hammer perform better than the diesel.
CONCLUSIONS
Equipment for driven pile installations has
undergone slow changes. Depending on where in
the world piles are being installed, equipment may
range from a simple winch driven drop hammer to a
sophisticated computer controlled hydraulic
hammer. The hammers and piles may be guided by
an old skid rig or a dedicated fixed lead system.
Depending on the type of hammer and driving
system, great variations of the system efficiency
must be expected. These eficiencies can be
accurately assessed in the field using a Pile Driving
Analyzer which calculates the so called transfer
energy from pile top force and velocity
measurements.
Predictions of pile blow counts, stresses or
capacities should be based on a wave equation
analysis which incorporates realistic models for
hammer, driving system, piles and soils. These
predictions should be verified by measurements.
REFERENCES
GRL, Goble Rausche Likins and Associates, Inc. 1996.
CAPWAP Background report and users manual. 4535
Renaissance Parkway, Cleveland, OH, 44128.
GRL, Goble Rausche Likins and Associates, Inc. 1999.
GKLWMP Background report and helpful data. 4535
Renaissance Parkway, Cleveland, OH, 44128.
Hannigan, P., G.G. Goble, G. Likins, G. Thendean, & F.
Rausche 1996. Design and construction control of driven
pile foundations. FHWA Publication No. FHWA-HI-96-
033, Contract No. DTFiLiTl-93-C-001 IS.
Massarsch, K.D. 1992. Static and dynamic soil
displacements caused by pile driving. Proceedings, 4"
International Conference on the Application of Stress-
Wave irheory to Piles: 15-24. F.B.I. B a d s , (ed.),
Balkema Rotterdam.
Rausche, F., & G.G. Goble 1972. Performance of pile
driving hammers, ASCE, Journal of the Construction
Division, Vol. 98, No. C02, Proc. Paper 9188, 201-218,
September 1972.
Seidel, J. & F. Rausche 1984. Design and performance of
dynamic tests of large diameter drilled shafts. Pd Strass
Wave Conference, Proceedings of the PdInternational
Conference on the Appiication of Stress Wave Theory to
Piles., 9-16, Balkema, Rotterdam.
APPENDIX
Energy considerations
This appendix summarizes the various forms of
energy in hammer, driving system and pile. It is
hoped that this description will help to explain these
concepts to the practitioner.
Energy is a measure of the amount of work that
can be done: push a mass over a rough surface, lift a
weight, compress a volume of gas, accelerate a mass
to a certain speed, compress a spring, etc.
Mathematically energy is defined as the product of
force and distance over which the force acts. For
example, the weight of a ram, WR, suspended a
distance h above a reference datum, has the ability to
do work equal to its potential energy
Ep=WR h (AI)
Since energy is "indestructible" (in the worst case
it is converted to heat, sound or other forms of
energy) after falling through a vertical distance h, the
ram now contains a kinetic energy E, while its
potential energy has become zero.
EK= !h.V;wR/ g = Ep WR h (A21
where g is the gravitational acceleration.
Therefore
vi = T ( 2 g h ) (A31
88
This is the ram impact velocity in the absence of
any fiction or other losses. The work that a friction
force R would be doing while the ram is falling is
EF=Rh (A4)
practically reducing the available energy fiom WR
h to (WR - R)h. Let us assume that the friction force
is a certain percentage, p, of WRYthen we can write
the reduced potential energy as
Where q = (1 - p) is the so-called efficiency of the
hammer while p indicates the fraction of the energy
that has been converted to heat rather than speed. Of
course, the impact velocity of the ram is now
vi = r ( 2 g h q ) (A6)
In addition to fiiction, other losses also occur
during the descent of a hammer. For example, a
diesel hammer compresses air which requires the
following energy:
ED= I (P dV)
=A 1 ( pIN[V, /VIexp)ds
= [A pIN/(exp - I)] [VN/A+ hJeXP A (A71
with
In this formula, V, is the initial volume of air that
is present in the diesel hammer chamber when
compression starts, A is the inside area of the
cylinder, pINis the initial pressure (atmospheric), s is
the distance that the ram has traveled after the
compression started, V is the associated volume, &
is the compressive stroke, and exp is the exponent of
adiabatic compression, typically 1.4 for clean air.
The chamber volume, Vc , i.e. the volume left over
during impact is the product of compressive stroke
and area A equal the initial volume. The
compression ratio is therefore Vm / V,. Note that the
atmospheric pressure also does positive work on the
top surface of the ram. This energy is merely the
product of atmospheric pressure times ram top area
times compressive stroke.
As an example, let us use a few realistic numbers
representing a typical diesel hammer with 2 Mg ram
mass. Assume = 394 mm, Vc = 2,584,240 mm3,
A = 80000 mm', exp = 1.35, atmospheric pressure
of 100 H a . Then V, = 2,584,240 + (394)(80,000)
= 34.1 106 and the compression ratio is 13.2.
Entering these numbers in the above equation and
considering the work of the atmospheric pressure
yields a pre-compression energy of 1 1.4 kJ.
If this hammer has indeed a ram weight of 20 kN
and if its maximum stroke is 3 m, then it would be
rated with 60 kJ. The pre-compression loss is
therefore 19% of the rated energy.
Energy in //it. driviiig sysfem
During impact the ram also compresses one or two
cushions. If the spring constant of the cushion is k
and the maximum compression x and the maximum
force F = k x, then the energy stored in the cushion is
E,=%Fx (As)
Let us compare a 20 mm thick plywood cushion
and a 200 mm thick plywood cushion with cross
sectional area 0.1 m and normal plywood modulus
(200 MPa). The corresponding cushion stiffness
values are k = (0.1)(200) / 20 = 1000 kN/mm and
100 kN/mm, respectively. A mass falling onto these
springs and producing a force of 1000 kN in the pile
underneath will produce the following respective
cushion compression values
x = 1000/k = 1 and 10 mm
and therefore store energies of
Ec = % 1000 (1) or % 1000 (10) = 0.5 or 5 kJ (A10)
(Actually, the force underneath the softer cushion
is probably lower than the force under the thinner
and stiffer cushion; this would make the effect
demonstrated here a little less extreme.) For a
concrete pile and cushion of this size it would be
reasonable to assume that the hammer rated energy
ranges between 20 and 60 kJ. The thick cushion
therefore stores a considerable amount of energy
during the compression phase. The thinner cushion,
o n the other hand generates rather insignificant
energy losses. Assuming a coefficient of restitution
of 0.5, which is reasonable for a plywood cushion,
half of the compression energy would be converted
into heat. The other half is returned to the ram
during its upward movement.
Hammer cushions are subjected to higher forces
than pile cushions because the helmet's high weight
causes high inertia forces. The helmet acceleration
is maybe 1,000 g and the helmet weight of our pile
with 0.1 m2 area may be 10 kN. 'The inertia force is
then 10,000 kN. The stiffness of this cushion
[assuming a thickness of 100 mm and a modulus
(wood with grain parallel to load) of 15,000 MPa] is
0.1(15,000,000)/100 = 15,000 kN/mm. The
compression x is therefore 0.67 mm and the energy
09
stored is 3.3 kJ of which one half may be converted
to heat (if it is a wooden block).
Let us now calculate how much energy it takes to
accelerate the helmet to its highest velocity which is
approximately equal to the maximum velocity that
we see at the pile top and which may reach 3 d s .
Again assuming a helmet weight of 10 kN or a mass
of roughly 1 Mg yields a helmet kinetic energy of
Em = % m v: = % 1(3)2 = 4.5 kJ (A1 1)
It may be assumed that this kinetic energy is still
doing useful work on the pile after the helmet has
started to slow down.
In summary, our hammer may have a rated energy
of 60 kJ. If it is a diesel hammer it stores in the
compressed gases roughly 10 kJ. The hammer
cushion stores approximately 3.3 kJ, the helmet 4.5
kJ, the pile cushion between 0.5 and 5 kJ. For a
single acting air/ steam hammer we would expect an
energy loss of 33% of 60 W or 20 kJ due to friction,
misalignment, pre-admission etc. For a diesel
hammer we would normally expect additional losses
(friction, misalignment) of 20% of (60 - 10) = 10 kJ.
For the diesel hammer there would also be kinetic
energy temporarily stored in the impact block. Its
magnitude may be comparable to the energy stored
in the helmet when it moves. However, most of this
kinetic energy is probably transferred to the pile at a
time when it still can do useful work on the pile and
soil. The energy transferred to a concrete pile may
therefore be approximately 30 kJ for a single acting
aidsteam hammer and 25 kJ for a diesel hammer.
For steel piles which have no pile cushion the
transferred energy would be about 5 kJ higher. The
transfer efficiencies would then be 0.5 and 0.45 for
the single acting hammers on steel and concrete
piles; for diesel hammers these numbers would be
0.45 and 0.4. These results are in reasonable
agreement with the measured transfer efficiency
shown in Figures 2 through 7 .
Energy losses are generally lower for hammers
with low impact velocity, obviously because inertia
and cushion forces would be lower and thus the
associated cushion strain energy and mass kinetic
energy losses. However, experience, measurements
and wave equation analyses all indicate that high
velocity hammers do rather well when driving gets
hard.
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3

arnmer system design using wave equation analysis for testing cast-in-situ
concrete piles
M.A. Mukaddam & W. M. Iskandarani
A1 Hai and A1 Mukaddamfor Geotechnical Works, Dubai, United Arab Emirates
M. Hussein
Goble Rausche Likins and Associates Incorporated, Orlando, Flu., USA

ABSTRACT: Dynamic load testing of drilled shafts has become a routine technique of construction quality
control of deep foundations. A study of the suitability of the hammer system to generate a net measurable pile
penetration or an estimated mobilised static resistance in the bearing strata without damaging the shaft has
been routinely required in many countries. In addition, hammer details must be appropriately chosen for a
successful high strain dynamic testing of cast-in-situ concrete piles. The Wave Equation Analysis using
GRLWEAP can realistically simulate such hammer impacts and pile penetration process. Based on the study
of several cases in the United Arab Emirates, the aim of this paper is to simulate the behaviour of some drilled
shafts under the impact of a free falling hammer system and validate such wave equation analysis with the
dynamic load testing results. Suggestions and recommendations on how to perform a wave equation analysis
on drilled shafts using a free falling hammer system is also presented in this paper.

1 HIGH STRAIN DYNAMIC PILE TESTING
EQUIPMENT AND ANALYTICAL METHODS
Static loading tests have traditionally been per-
formed to determine the characteristics of pile load-
movement relationship. Due to the expense involved
and time required, this type of testing is performed
on only selected piles on a very limited basis. Lower
cost and speed of testing make dynamic pile testing
according to Case method and CAPWAP analysis
particularly attractive (Hussein et al. 1988).
The dynamic testing method was developed at
Case Institute of Technology in Cleveland, Ohio in
the late 60’s (Rausche et al. 1985). This method pro-
vides an economical, easily portable system which
bases pile bearing capacity predictions on electronic
measurements of pile top force and acceleration. The
following are the main objectives of dynamic cast-
in-situ pile testing: (a) evaluation of pile static ca-
pacity (shaft friction and toe bearing), (b) determi-
nation of dynamic piles stresses under hammer im-
pacts, (c) assessment of pile structural integrity.
High-strain dynamic pile testing is based on the
iiieasurement of pile strain and acceleration under
impacts of a relatively large weight. Pile strains are
measured with strain transducers. Sensors are
moiuited around pile circumference, ideally at a lo-
cation approximately one to two pile diameters be-
low its head. Data analysis is based on field meas-
ured dynamic records, sophisticated modelling and
numerical techniques, and one-dimensional elastic
wave propagation principles.
Field testing is done according to Case method
procedures as applied with the Pile Driving Analyzer
(PDA). The Case method is a closed form solution
based on a few simplifying assumptions such as
ideal plastic soil behavior and an ideally elastic and
uniform pile (Linkins & Hussein 1988). Assump-
tions of a soil damping constant must be made at the
time of testing. The PDA is user friendly field data
acquisition system and computer that provides
power supply, signal conditioning, processing and
evaluation of the measured dynamic data. Generally,
testing results obtained in the field, include results
for evaluating hammer performance, dynamic pile
stresses, pile structural integrity, and pile resistance
and static bearing capacity. High strain dynamic pile
testing is part of many standards and specifications
such as ASTM - 1989, AASHTO - 1993 and the
British CIRIA guide.
Rigorous analysis on the field collected PDA data
is performed using the CAPWAP method (CAse Pile
Wave Analysis Program) (Rausche at al. 1985). This
method uses the velocity records as pile-top bound-

91
ary condition and performs a wave equation form of
analysis. A soil-resistance distribution is assumed,
and if it is correct then good agreement betweeii
measured and computed pile-top force is obtained.
By iteratively improving the pile-top force match,
the set of soil-strength parameters producing the
measured data is found. It is also possible to esti-
inate side-friction and end bearing distribution as
well as a simulated static load test curve. Alterna-
tively, the force record can be used as a boundary
condition and the analysis performed to match the
velocity measured and computed records.

2 PILING WORKS AND DYNAMIC LOAD

TESTS IN THE UNITED ARAB EMLRATES

Almost all piles in United Arab Emirates iiowadays

are made up of cast-in-situ bored reinforced concrete
piles. Two to three millions of piles are drilled every
year and tlie diameters of such piles vary generally
from 50 cm to 120 cm. Cast-in-place piles are pro-
duced by forming holes in tlie ground and filling
them with concrete. The design of such shafts is
usually based on the recommendation of the soil in-
vestigation and foundation report. Then normally 2
% of tlie working piles are tested statically andor
dynaiiiically to a maximum load equal to one and
half times to two times tlie design working load in
order to verify tlie adequacy of the piles. In this case,
the ultimate bearing capacity of the tested shafts is
not reached and the client is satisfied that a certain
ininimum static load was achieved, and this depends
on the safety factor desired for the test. In this case,
displaceinents under hammer impacts shall be ap-
propriate to mobilise maximum skin friction in the
shafts as it is required for a successful wave equation
analysis. In general, displacements in the range of
about 5 inm are acceptable to develop maximum
skin friction in the shafts (Bowles 1996), sometimes
even less.
3 HAMMER SYSTEM INPUT DATA FOR
GRLWEAP~
Testing of driven piles is conveniently done utilising
tlie pile driving hammer blows. For testing cast-in-
place piles, it is necessary to have a drop weight for
testing impacts (Likins & Hussein 1995). Hence, in
tlie absence of driven pile rigs, we have devised our
free-falling hammer system in which the drop
weight is connected to a mechanical crane having
release mechanisms. Such hammer system consists
of the following: a non uniform steel ram made in
segments which are bolted together having a total
weight of about 70.8 KN, a harmner guide con-
structed of steel hollow pipe allowing variable drop
heights attached to a hammer striker plate (Figure 1).
Figure 1. Free falling hammer system.
Such haininer weight has proved to be acceptable
for testing drilled shafts having diameters between
50 and 100 cm (Mukaddam & Iskandarani 1996).
For the length (L) and area (A) of each hammer
segment, i.e. segment 1 to 3, refer to Table 1.
Table 1. Soecifics of Hammer Serrments
A2 cm-
7
3846
L3 cm 15
. A3 an2 5024
A mechanical crane can lift this steel hammer to
the required height and then release it to drop freely
on the top of the pile to be tested. Few sheets of
plywood are placed as a cushion between the ham-
mer striker plate and pile top to control tlie pile dy-
namic stresses and reduce the local contact non-
uniformity.
In order to perform a successful test based on pile
specifics and soil conditions, wave equation analysis
is to be performed in order to design the test system,
i.e. size and weight of hammer, drop height and
cushioning. The GRLWEAP analysis model requires
some information related to our driving system, like
the ram weight, ram diameter and length, cushioning
conditions etc.. ..
Since our hammer is non-uniform, the ram di-
ameter Dr should yield the effective ram stiffiiess,
Kr, over the length (GRL & Associates 1993).

92
 Consider K, = A, E / L, ; i = Segment 1, 2 or 3.
By using tlie Kirchoff s Law, tlie effective ram stiff-
ness shall be
K’,= KIK2 / (Kl + K2)
S = A I A2 L’, / ( A I L2 + A2 L I )
T ~ UA’,
where L’, = LI 4- L2
Again K, = K’r I<i / (K’r + K3)
u s= A’, A3 L, / (A’r L3 + A3 L’r)
T ~ ~ A,
where L, = L’, + Lj
And finally D, = (4 A, / pi)”2
Based on the above, a GRLWEAP data file re-
lated to our hammer system was made available.
Such data are given in Table 2 below.
Material Plywood
Pile Cross section cm2 3 847
cushion TIlickness imn 18
WEAP Modulus MPa 207
Coeflicieiit of resti(iitioti 0.5
4 CASE STUDY (1)
For tlie proposed buildirig at Nabbaa area in Sharjah
Emirate, low capacity piles were drilled to a depth of
about 15.5 m below the site reference level. The in-
vestigated soil consisted in general silty sand occa-
sionally cemeiited and slielly within the top 9.8 m,
with the presence of sandstone formation at depths
beyond 14.4 m. Static capacity analysis conducted
by the soil investigation firm concluded a design
working load of 883 kN (324 kN for skin friction)
and 1226 kN (390 kN for skin friction) for pile hav-
ing diameters of 50 and 60 cm respectively. About
50 to 55 % skin friction of the desired load is coil-
s idered for wave equation analysis distributed more
or less uniforiiily over the pile shaft according to the
soil characteristics as per the investigated soil layers.
Wave equation analysis input information have
included a hammer efficiency of 30 %, skin aiid toe
soil quakes of 2.54 and 4.17 m m respectively for 50
cm pile diameter, skin and toe soil quakes of 2.5 and
(1) 5.0 mm respectively for 60 cm pile diameter, and
soil damping factors of 0.25 and 0.50 s/m for skin
(2 1 and toe resistance. The GRLWEAPR analysis re-
sults, called ‘inspector’s chart’, showed a relation-
(3) ship between tlie stroke and the blow count for a
constant capacity option, aiid the compression and
tension dynamic stresses for each hammer drop
height (Fig. 2)
For pile 6 having 50 cm diameter and hammer
stroke ranging between 10 and 100 cm, the constant
capacity analysis showed the necessary driving re-
sistance at each hammer stroke for the input capacity
(4) of 1325 KN. Based on this analysis, a 30 cm ham-
mer stroke was found to be reasonable to test 50 cin
diameter shafts since it shows a reasonable driving
resistance and keeps the driving compressive stress
in the range of 7.7 MPa. However, for 60 cm pile di-
ameter (pile 24), a hammer stroke of 50 cm was used
in order to verify a load of 1845 kN aiid to keep the
driving compressive resistance in the range of 9.5
MPa. Based 011 this, it is to be noted that such infor-
mation is very helpful for the PDA Engineer prior to
starting the dynamic load measuremeiits for the
control of stresses in the drilled shafts. It is also very
helpful for field inspection in such a way that when
the point of intersection of the stroke and driving re-
sistance plots below tlie inspector’s chart curve, the
desired pile capacity has not been obtained. How-
ever, any combination of stroke and driving resis-
tance plotting above the curve indicates that the re-
quired pile capacity has been reached (Haiinigan et
al. 1996).
During the field dynamic measurements coil-
ducted 011 pile 6, the CASE method has showed a
proof load of 1550 kN, a driving compressive stress
of about 7.0 MPa and a driving resistance of about
1500 blows/meter for a hammer stroke of about 30
cm (Fig. 3). The point of intersection of the stroke
and driving resistance is falling above the inspec-
tor’s chart curve, hence tlie desired pile capacity has
been obtained aiid thus field measurements have
confirmed our initial analysis.
The same applies to pile 24 where a stroke of 50
cm has showed a proof load of about 2750 kN, a
generated driving compressive stress of about 9.1
MPa and a driving resistance of about 2000
blows/meter, for which tlie intersection point falls
above the curve of the WEAP inspector’s chart (Fig.
4).
5 CASE STUDY (2)
The second case study consisted of the proposed
Building at Noor AI Aiiz East Area in Dubai Einir-
93
Figure 2. GIUWEAP analysis for case I

ate. Soil investigation works have consisted of

drilling four boreholes down to a depth of twenty
five meters each below tlie original existing ground
level. The subsurface condition consisted in general
of medium dense to very dense sand layers, occa-
sionally cemented, interbedded with layers of weak
calcarenite at depths between 1 1 and 13 m and at
depths between 18.5 and 20.5 iii. In order to verify a
static load of two times the working load of 1374
kN, high strain dynamic load measurement was per-
formed on one indicator pile TP having 60 cni
nominal diameter and a total length of about 16 ni.
Static capacity calculation have concluded a skin
friction to be about 75 % of tlie desired load to be
verified distributed almost in a triaiigiilar way over
the length of the shaft. A summary of the wave
equation analysis input values related to pile and to
soil data are summarised in Table 3.

94
 Figure 5. GIUWEAP analysis for case 2.

6 CASE STUDY (3)

In this last case study for the proposed Office De-

velopment at A1 IUiabeesi Area in Dubai Emirate,
shafts of 90 cni nominal diameter were having pile
toes at a depth of about 18 ni below the site refer-
ence level. Tlie scope of soil investigation works
consisted of drilling five boreholes down to depths
of 20 to 25 111 below tlie original existing ground
Figure 6. I’DA data for pile ‘IT
(Case 2).
level. Tlie soil strata consisted in general of medium
dense to dense sand with shells down to a depth of 8
meters, followed by dense to very dense cemented
sand layers and fine to medium grained sandstone
LCll@- in 16 down to the drilled depths of the borelioles. Tlie de-
I’ i I e lli amct er 111111 I 600 sign working load of the piles to be tested was 2649
kN and the proof load was 3973 kN. GRLWEAP in-
1’1-oof Load 2748 put details related to pile specifics and soil data are
Shalt I k s . Distribution
kN.s/m 2442
as summarised in Table 4 below.
______
Soil 2.5415 .00
0.2.510.50 ’I’ablc 4. Soil and oile data file [or GRLWEAP (Case 3 1
Total Length 111 12.2
Pile Diameter (nominal) 111111 900
Assuming a hammer efficiency of 30 % for our Working Load kN 2649 -
70.8 KN hammer weight with various drop heights Proof Load kN 3973
using 1 8 nini cushion thickness, wave equation Shaft Resis. Distribution %
’ Uniform & 70 YO
analysis have indicated a drop height of 65 ciii to Impedance kN.s/m 5494 __
mobilise the requested proof load with a maximum Soil Skiifroe Soil Quakes iiiiii 2.54/7.50--- -
compressive strength in the range of about 10.5 Mpa Skiid‘l’oe Soil Daimine mii 0.2510.50
(Fig. 5 ) .
During the dyiianiic load measurements, the Case
For a hammer efficiency of about 30 %, wave
method as presented in Figure 6 lias shown a maxi-
equation analysis has shown a ininiinuin drop height
iiiiini compressive stress of 9.5 MPa and a mobilized
of 100 cni for the 70.8 kN hammer weight to achieve
static load of more than 3000 kN. The observed set
a static proof load of 3973 kN within a driving coiii-
was about 1 . S iiiiii, i.e. about 667 blows/meter. By
pressive stresses limited to about 10 MPa (Fig. 7).
checking tlie intersection points of tlie observed
However, during the field dynamic load measure-
blow count and the hanimer stroke, such point is
ments conducted on the first pile 73, the Case
slightly falling near the curve of the inspector’s
cliai~.Hence, the capacity achieved has been verified method has shown a mobilised static load above
aiid our initial analysis lias been confirmed. 6000 kN aiid a inaxiinuin coinpressive stress of 1 1.7

95

Figure 9. GRLWEAP analysis for case 3 (35 % hammer effi-
ciency).
MPa for an observed blow count of about 1000
blowdmeter (Fig. 8). As far as the inspector’s chart
is concerned for such assumed efficiency, the inter-
section point falls below the curve which means that
the capacity was not verified which is in contradic-
tioii with the field data. Definitively, the hammer has
96
performed with slightly higher transfer efficiency,
however, as the capacities mobilised were in excess
of that required, attempt was made to adjust the
hammer performance. By conducting wave equation
analysis considering 35 % hammer efficiency, it is
concluded for pile 73 that the intersection point be-
tween the measured blow count and stoke falls al-
most on the inspector’s chart for this tested pile, for
a maximum compressive stress of 11.2 MPa which
confirms the PDA field data (Fig. 9).
Wave equation analysis with 35 % hammer effi-
ciency for the second pile to be tested (pile 56) hav-
ing the same diameter has showed that a stroke of
about 80 cm is appropriate to mobilize the desired
static capacity with a driving stress of about 9.8
MPa. Furthermore, the PDA field measurements
(Fig. 10) have showed an activated static capacity of
about 4220 kN, a maximum driving stress of about
9.5 MPa and a measured set of about 0.5 inm for
which the intersection point with the stroke falls al-
most on the curve of the inspector’s chart. Hence,
our initial analysis has been confirmed.
7 RECOMMENDATIONS ON WAVE properly, wave equation analysis shall lead to a SUC-
EQUATION ANALYSIS cessful dynamic test.

The following are our conclusions for a successful

wave equation analysis on drilled shafts using our REFERENCES
free falling hammer system:
Bowles, J. 1996. Fotindntioii Anmlysis cud Design. Fif'th Edi-
1. For a free falling hammer system using a me- tion. Washington: McGraw Hill.
GRL & Associates 1993. GRL WEAP Mciniid. Cleveland:
chanical crane, the hammer rated energy can be GRL.
taken in the range of about 30 to 35 %. Hannigan, P.J., G.Goble, G.'I%endean, G.1Likins & 1~.Kaiischc
2. It is difficult to predict accurately the hammer 1996. Design ~ n Construction
d (d'Urilvn Piles i;ounc,'tr-
efficiency for a reliable wave equation analysis. lions. National Highway Institute, Course no 1322 I &
Hence, it is advantageous to have more than one 13222. Washington: US DO?' Federal Highway Admini-
shaft to be tested dynamically on a given project in stration.
order to adjust the GRLWEAP data based on the dy- Hussein, M., G.Linkins & F.Rausche 1988. 7'esling Methock q'
iianiic data of the first tested shaft. Hence, the ability Driven Piles. Cleveland: GRL.
Linkins, G. & M.Hussein 1988. A suniinai-y of the Pile Driving
to produce more accurate GRLWEAP analysis for Analyzer Capacity Methods. The 1 I"' Pile Driving Analyzer
the other shafts to be tested. User's Diiy Setiiitiar, Clevelatzi 1988. Cleveland: GRL.
3. For siiiall shafts to be tested with our current Linkins, G. & M. I-Iussain 1995. High Strain D namic Testing
hammer system, a plywood cushion having a thick- of Drilled Shalis and Cast-in-Situ Piles. 20" Annud Men7-
ness of about 18 to. 20 mm may be acceptable to hers Conjereiice tind Meelitig, Chirlesioti, 16-18 Oclobes
IYY.5. South Cxolina: Deep Foundation Institute.
produce a high set and mobilise more capacity by Mukaddain, M. & W. Iskandarani 1996. High Strain Dynamic
increasing tlie impact force and energy transferred. Testing of Cast-in-Place Piles in the UAE. 5''' iti[erna~ioncl/
Increasing the cushion thickness between the striker Conference oti [he Application of Stress Wove Tlieory 10
plate and the shaft will definitively decrease the Piles, Orlcrticlo, 11-13 Septeiiiher 19Y6. 805-822. Orlando:
driving stresses, but will also increase the required Townsend.
hammer stroke. Rausche, I:., G.Goble & G.Linkins 1985. Dynamic Determina-
tion of Pile capacity. J011~tIfdof'Geo-technical Etigineeritig,
4. It is not recommended to place plywood cush- Vol. I I I , No. 3, March 198.5.
ion between the striker plate and tlie hammer. By
doing so, the transferred energy to the sensors level
will also decrease.
5 . Since we are testing working shafts to one and
a half times the design working and not to their ul-
timate capacities, it is iiiiportant to have a pile top
displacement under each hammer impact in the
range of about 3 mm or above in order to ensure that
almost 90 % of the ultimate skin friction resistance
has been mobilised. Hence, the percentage of skin
friction with respect to total load to be verified may
be properly assumed for GRLWEAP.
6. Drilled shafts are probably the most difficult
pile types to perform dynamic testing and analysis
on, since such shafts may be non-uniform. Hence,
wave equation analysis may be used as an additional
quality control tool to confirm the findings of the
field dynamic load measurements, especially in the
absence of static load tests.

8 CONCLUSION

Wave equation analysis is commonly used to study

the dynamics of cast-in-place shqfts under high-
strain impacts. Such analysis can be used not only to
decide on the hammer weight, drop height and
cushion at the pile top but also as an additional
quality control tool to confirm the field findings of
the high strain ineasurenlents. Hence, if conducted

97
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Analysis of pile load transfer using optical fiber sensor

Jeong-Ho Oh & Won-Jae Lee

Department of Civil Engineering, Korea University,Seoul, Korea
Sang-Bae Lee
Photonics Research Center, Korea Iizstitute of Science and Technology,Seoul, Korea
Woo-Jin Lee
Department of Civil Engineering, Korea University,Seoul, Korea

ABSTRACT: Distribution of axial strain in model piles was measured by Fiber Bragg Grating(FBG) sensor
to investigate a possibility of measuring and analyzing the load transfer mechanism of pile foundation by
Fiber Optic Sensor(F0S) system. Since Fiber Bragg Gratings(FBG) of different wave lengths can be
multiplexed in an optical fiber, the installation of sensor system aiid the measurement of strains are relatively
simple compared with the system consisted of strain gages. In this study, Fiber Bragg Grating(FBG) sensors
and electric strain gages were embedded in the piles and the distributions of load transfer measured by two
sensor systems were compared. It was observed from the test results that the magnitudes of axial load
measured by both systems showed insignificant difference and that the distributions of axial load by FBG
were smoother than those by strain gage. Under the environmeiits of laboratory testing, survival rate of
embedded FBG system was higher than that of strain gage. Therefore, it was concluded that the use of FBG
sensor has a great potential for the measurement of pile load transfer with accuracy.

1 INTRODUCTION and the relationship between the variation of

wavelength of FBG and strain.
Through the monitoring of geo-structures engineers
usually obtain the information on the state of those
structures and reevaluate the design parameters used. 2 FIBER OPTIC SENSORS
While tlie strain gages usually measure the strains
accurate enough for general purposes, simultaneous
measurement of strains at several locations causes 2.1 Fiber Optic Components
difficulties of enibeddment in a structure and of the Fiber optic cable transmitting the light is coniposed
line connection to measuring system at remote of core, cladding and buffer, as shown in Figure 1 .
location. While the use of conventional strain gages Most of the cables are glass while plastic cables are
is still popular, the applications of FOS system to used in some limited applications. The core carries
civil engineering structures have been gaining the optical signal and the size of core varies,
ground. depending on cable types. Single mode cable, used
Several types of FOS have been developed with for long range transmission of light and high
tlie advance of fiber optic sensor technology and bandwidth systems, has a core size of 8 to 10irllr.
they have been mainly applied to monitor the Multi-mode cable, used for shorter range, has a core
behavior of the composite structures. FOS can size of 62.5pn1.The cladding keeps the light within
measure an infinitesimal deformation with its high the core, and prevents external light from entering
sensitivity to compressive and tensile deformations,
aiid the multiplexing of several sensors in an optical
fiber significantly reduces the complex arrangement
of lead lines to remote data acquisition system.
In this research, static pile load tests were
performed on model piles in which two sensor
systems were embedded and the load transfer results
by both systems were compared. A tension test of
steel bar and a compression test of mortar specimen
embedded with them were also carried out to study
tlie method of attachment of sensor in the structure Figure 1. Structure oi kibei Optic Cable

99
the core. This layer, typically 125p,11in diameter, is
glass similar to the core, but with a slightly different
composition. Surrounding the cladding is a buffer
layer that protects the inner glass. The buffer
diameter is 900/,,,, made of ceramic or plastic.
2.2 Tjpes of Fiber. Optic Seiisors
Fiber optic sensors are divided into two broad
categories,intensionietric and interferometric sensor.
Intensionietric sensor measures the amount of
light transmitted through the fiber. In its simplest
foim, damage may be detected by a fiber breaking
and transmission going from on to off. Strain sensors
based on intensity variation due to microbending
have also been developed; however, they are
relatively insensitive, inaccurate and have a limited
range of detection.
Interferometric sensors have been developed for
high-sensitivity applications, such as acoustic
sensors and magnetic field sensors, and are usually
based on single-niode fiber. There are several types
of interferometric sensors: Mach-Zender, Michelson,
Extrinsic Fabry-Perot(EFP) and Intrinsic Fabry-
Perot(1FP) interferometric sensors. In addition to
these, a third category, and one of the most attractive
types of fiber optic sensors for embedded
application, is tlie Fiber Bragg Grating.
The comparison of different types of optical fiber
sensors is summarized in Table 1. The higher the
number, tlic better in its functional category.
Table 1. Comparison of Sensors
I I I I I 1
Though most of the sensors are good in terms of
sensitivity to strain, FBG sensor is the best one when
it conies to the multiplexing. Multiplexing refers the
ability to distribute sensors spatially through one line
of optical fibers and to demodulate all sensors using
the same demodulation instrument. It is the most
crucial advantage of FOS sensor over the common
electronic strain gage.
From above table, FBG sensor is considered as a
proper sensor system to measure the load transfer
mechanism of pile due to its multiplexing, high
sensitivity, and simple fabrication. So, FBG sensor
was selected as fiber optic sensor for present study.
2.3 Fiber- Br-agg Gmting sensor(FBG sensov)
FBG sensor is made by exposing tlie fiber of about
lcin to the interference pattern of two coherent UV
beams as shown in Figure 2. The absorption of UV
light in tlie fiber changes the chemical bonds in the
glass, thus causing the change in the refractive index
of the glass. The resulting spatial modulation of
refractive index produces the Bragg grating and acts
as a narrow band in-line wavelength notch filter.
When the light signal from broad band sourcc is
transmitted, tlie grating reflects strongly at tlie
wavelength, ~~,
for which the Bragg resonance is
satisfied, as shown in Figure 2. This Bragg
resonance is determined by the condition:
1
where, i\ : spatial pitch of the grating
n : refractive index of the fiber
I_ : the laser wavelength
8 : the angle between the beams
The refractive index is dependent of the type of
fiber. Spatial pitch of the Grating is controlled
during the manufacturing process and Bragg
wavelength can be generated within almost any
spectral band of interest. The linear distance
between the gratings can be adjusted for the desired
sensor placements on the structure. This
characteristic is called multiplexing, as is depicted in
Figure 3.
When the deformation is occurring in the structure,
strain at the grating shifts the wavelength at which
the resonance condition is satisfied. The input
spectrum from passive broadband optical source is
reflected by FBG and the output narrowband
spectrum includes tlie information on the change of
resonance wavelength. Strains at each FBG sensors
are calculated by filtering an output spectrum and
comparing it with that before deformation. This
inherent 'wavelength-encoded' nature of the output
100

Figure 3. Multiplexing of FBG sensor
of FBG has distinct advantages over other sensing
schemes.
There are two constraints on the Bragg
wavelengths of the sensors in the array. The
wavelengths of all gratings niultiplexed in an optical
fiber should be within the bandwidth of the source.
There should not be 'cross-talk': the wavelengths of
adjacent gratings must be such that they do not
intersect or cross over when the structure is loaded
to its maximum load.
3 TESTING PROCEDURES AND APPARATUS
A series of laboratory tests were performed to
investigate the possibility of measuring load transfer
mechanism of pile by fiber optic sensor. For all the
specimens tested, both optical fiber sensor and
conventional strain gage were installed at
corresponding locations. Preliminary tests, such as
uniaxial tensile test of steel bar and compressive test
of cubic concrete specimen, were carried out. Static
pile load tests on three model piles installed in
calibration chamber were perfomied and the
distributions of load transfer measured by both
measuring system were compared.
3.1 Fdmcirtion of FBG sensors
There are two typical methods in grating fabrication.
One is the holographic writing technique and the
other is the phase mask method. The former is not
suited to write gratings of a precise wavelength,
since it is difficult to set the interfering angles for a
desired wavelength. Therefore, phase mask method
simpler for stable grating fabrication is used in this
study. A phase mask grating is fabricated by
esposing the UV transmitting silica mask plate to e-
beams, followed by dry plasma etching. The grating
period is similar to that required for producing the
reflection grating in germanium-doped silica. The
phase mask diffracts the induced UV beam into Oth,
1st order beams making interference patterns on the
fiber core which causes an increase in the local
101
refractive index as in the holographic writing
technique. With phase mask method, one can
reproduce gratings of the same wavelength that is
determined only by the mask period, which is very
crucial for mass production.
Fiber Bragg Gratings used in the study were
fabricated at Photonics Research Center at Korea
Institute of Sience and Technology (KIST). The
germanosilicate fiber (KIST-1104, single mode at
1.55 Dm, Ge: 18.5 mol.%, B: 22moI.%) was drawn
with a modified chemical vapor deposition (MCVD)
method. A KrF excimer laser (Lambda Physik,
h =248nm, maximum pulse p o w e ~ 6 5 0mJ) was
used as a UV source. The fabrication set-up
consisted of a UV laser source, cylindrical lense,
mirror and phase mask is shown in Figure 4.
Figure 4. Fiber-grating fabrication set-up
3.2 Uniaxial tensile test for steel bnv
Both conventional strain gauges and FBG sensors
were attached to a steel bar with an elastic modulus
of 220 GPa and with a dimension of 3cm dia. and
50cm length. The distance between strain gages was
lOcm and FBG sensor was grated at two measuring
points. One side of steel bar was machined flat with
width of lcm and the other one is fabricated with a
groove to ensure the bonding between steel and
FBG. Optical fiber with FBG is laid straight in a
groove and section of optical fiber except the Bragg
gratings is coated by a resin epoxy. And gratings
were coated by paraffin to make it respond
elastically. The coating with resin epoxy and
paraffin provides the moisture protection for the
optical fibers.
Tension loading and unloading to the
instrumented steel bar were applied by universal
testing machine, and the strain measurements by
strain gage and FBG were analyzed in terms of
linearity and hysteresis. The FBG sensor signal
encoding the information on the wavelength was
measured using an OSA(Optica1 Spectrum
Analyzer) or a spectrometer. Strain at each FBG
sensor is detemiined by analyzing the variation of
resonance wavelength from previous one.
3.3 Compression test on cement mortar specinien
To observe the performance of FBG sensor
embedded in the concrete structure, a stress-strain
relationship was measured for a cement mortar
specimen. Two steel bars attaching a FBG or a strain
gauge in the mid-length were embedded in mortar
specimen in the direction to the applied load. The
mortar specimen was a rectangular hexahedron
(1 5cm" 15cm"20cm). The steel bars were first
iiiserted in a mold and then a fluid mortar was
poured.
Compression loading was applied on the mortar
specimen by universal testing machine and the
strains measured by FBG and strain gage were
compared.
3.4 Model Pile test
3.4.1 Sciiizple pvepamtion
bar with a length of 50cm for loading test was
inserted into the ceiiter of model pile. Model pile
was cured for 7 days and the loading test was
performed.
3.4.3 Static load test on moclel piles
Two tension tests and one compression test were
performed, following a procedure of "quick
maintained load test". The interval of maintained
load at each loading stage was 5 minutes and the
upward displacement was checked at 0, 2.5, 4 and 5
minutes. The schematic drawing of tensile loading
test system is shown in Figure 6. Load applied by an
oil jack of 5 ton maximum capacity was measured
by a proving ring which can be loaded up to 3 tons,
and displacements at pile butt was measured by two
dial gauges with accuracy of 1/200mm.

Model ground was prepared by compacting the

layers of weathered granite soil, aiid then model pile
was installed in compacted soil through excavation
aiid fabrication of pile material. Weathered granite
soil was obtained from Poi-dong, southern part of
metropolitan Seoul. Its optimum moisture content
and maximum dry density are 16% and 1.9t/m3,
respectively. Weathered granite soil dried in the
oven at 105 for 24 hours is mixed with water to
have optimum moisture content. Then it is
Figure 5. Procedure of model pile installation
compacted in five layers up to lOOmm below the
upper plate of the chambcr, using a metal rammer of
7.5kg weight with a drop height of 60cm. And sands
are deposited on the surface of weathered granite
soil and filled tlie chamber thoroughly by raining
method. Air-dried sand with water content of less
than 2% is poorly graded clean fine sand, and
classified as SP by the USCS. An upper chamber
plate was placed above the surface of sand deposit
aiid tied strongly to prevent a leakage of chamber
pressure applied to tlie model ground in horizontal
and vertical directions.

3.4.2 Imtallcition ofniodelpile

Figuie 6. Tensile Loading 'I est System
Tlic procedure to install a model pile in a calibration
chaiiiber is illustrated in Figure 5.
A hole was drilled with an electric screw whose Table 2. Description of three model piles
diameter, pitch, and length are 4cm, 6cin and 40cm,
respectively. Using this screw drill, compacted
weathered granite soil was excavated up to 40~111
below the surface with 4Cm diameter. Before
pouring a mortar, the slime occurred during an
excavation was removed by coiiipressed air of 1
kgicm2 pressure. A mortar mixture of cement, water,
expansive agent and additional materials is poured
and compacted by five layers in a drilled hole.
Immediately after the injection of mortar, tension

102
 Three model piles were tested and Table 2
presents the details of test condition.
Piles P1 and P2 were tested in tensile loading system
and a compressive load test was performed for P3.
For the investigation of load transfer mechanism,
electronic strain gages and FBG sensors attached on
a reinforcing bar were installed in the test piles. To
protect the FBG sensor and strain gauges, the
reinforcing bar was embedded in a thin-wall metal
pipe with a slightly larger diameter than that of
reinforcing bar. Metal pipe was removed
immediately after the installation of model pile. For
tension tests, four strain gages were attached to a
reinforcing bar at 5 , 15, 25, 35cm depth. For
compression test, sensors were attached to a
1-einforcingbar at depths o f 5 , 20, and 35cm.
Figure 8. Stress-strain measured by FBG sensor and stmin gage

4 TEST RESULTS AND ANALYSIS

4.1 Uiiici.~rd teiisrori test f o r steel bar
Figure 7 showing the relation between the shift of
wavelength of FBG and the strain measured by
strain gage was obtained from the uniaxial tension
test for steel bar. The dependence of the return
wavelength on strain is obtained by the regression of
the measured values, 1 .28pndU strain. Here 1
1-1 strain is an variation of 1 part in 106. In other
words, wavelength shift is converted into strain by
m 111t iplying 780.
It can be noted from Figure 8 that the strains
mcasurcd by FBG sensor are matched to those of the
strain gauge quite well. It is also shown that the
trend of stress-strain reIationship is similar for both
measurements and that there is an insignificant
difference in strain values under the loading and
unloading steps. An elastic modulus calculated from
stress-strain relationship in Figure 8 is
approximately 2 15 GPa and it is similar to 220 GPa
given by the manufacturer.
4.2 Embedded FBG seizsor 111 mortur specirneii
Loading test on mortar specimen was carried out for
preliminary examination of measured strain values
by two sensor systems in the case of embedding
them in concrete.
One of the most important factors to be
considered in any embedded sensor is the interfacial
bonding between fiber and the surrounding host,
since effective, accurate, and repeatable
measurements are strongly dependent 011 an
effective load transfer from host material to sensing
fiber. In this test, the compression load was applied
to the mortar specimen that had been cured for 14
days after two sensor systems were embedded The
unloading was performed in order to examine which
sensor works better under the cyclic loading
condition. FBG sensor and a strain gauge bonded at
the mid-height o f two steel bars were embedded 111 a
mortar specimen of rec t aiigul ar hex ahedi-on. Fi gu re
9 shows the test results, which compare the
measuring values in FBG sensor and a strain gauge.

Figure 9. The comparison of stxain values

Figure 7. Relation between wavelength shift and strain

103
It is observed from Figure 9 that there are some Compared with the results of previous test iii
differences between two strain values especially in section 4.2, the measured strain values of FBG
large loading rather than in small loading stage. sensor are observed higher than those of strain gauge
Such a relatively large difference in measured strain since less stiff coating material may help FBG
appears to be due to an eccentric application of load sensor respond elastically. Therefore, it can
011 both ends of mortar specimen. Since two steel beconcluded that the method to install the FBG
bars, on which two sensor systems were attached, sensors in the structure to be monitored has a
were vertically embedded in a mortar specimen with significant influence 011 the performance of sensor
a horizontal spacing of 5cm, the distribution of strain systern.
occurred within cross-section was not likely to be
uiii forin.
It is also noted from Figure that the strain values
of FBG sensor are smaller than those of strain gauge
at load larger than 5 tons. This trend appears to
occur because stiff epoxy, which was used as a
coating material of FBG sensor, may restrict the
elastic behavior of FBG sensor.
It may be concluded from Figure 9 that FBG
sensor appears to work better than strain gage in
unloading stage. FBG sensor represents the
difference of 16 iiiicrostrains due to loading-
unloading while strain gauge shows 50 microstain
difference at load of 15 ton, as depicted in a dotted
rectangular in Figure 9. Also, the curve obtained by
the FBG seiisor is observed to be similar to the
common stress-strain curve. This indicates that FBG
sensor is more adequate than conventional strain
gage for measuring stress or strain in various
structures that are exposed to exteiiial environments
and long-terni monitoring. Also, because the trend of
FBG sensor measurement shows comparatively less
hysteresis, there seems to be insignificant problem in
interfacial bonding between FBG sensor and host
materials.

4.3 Load transfer meclinnism of model pile

As mentioned previously, three model pile tests

were performed to compare the measured load
transfer by two different sensor systems. However,
the different type coating material, which appeared
to be less stiff than that used in previous
compression test on mortar specimen, was used for
the pile load tests.

4.3.1 Coinpai-isori of strain values

The comparison of strain values measured by two

sensor system are made by plotting them with
respect to loads, as shown in Figure 10. It is
observed that the difference between two values is in
the range from 5% to 10%, and both measurements
show similar trends. It is also showed that FBG
sensor presents smoother trend in measured strains
while slightly abrupt changes in measured values by
strain gage are detected as loading increases. Minus
Figure 10. Comparison of sham values(P1, P2, P3)
values of strain in the test results of P3 mean that
wavelength of FBG sensor is shortened in
compressive condition.

104
4.3.2 Distribution of axial pile loads

Tn order to evaluate the axial pile loads from

measured strains, Young’s modulus is determined
fi-oiii the strain gauge located at 5cm below the
ground surface as a reference for determining the
relation between the measured strain and the load in
the pile. Despite of careful installation of tension
bar, a strain gauge located at 25cm depth of PI pile
and a strain gauge positioned at XScm depth of P2
pile were found not to work after the curing. For P3
pile, strain gage positioned at 35cm depth was
brolteii during load test. From above observations, it
appears that the durability of FBG sensor is superior
to that of strain gauge and FBG sensor may have a
strong possibility to be a useful tool for monitoring
geo-structures in a field.
Axial loads evaluated from load transfer tests on
test pile PI, P2 aiid P3 at each loading stage are
obtained in Figure 11.
When it comes to the load transfer niechanism,
the distribution of axial load on three piles are in
excellent agreement except. at the depths where
strain gages were broken. In pile load transfer
curves, variation of load with depth is represented by
a coiinectioii of dashed line to indicate that strain
was not measured due to the strain gage probleiii.
Froiii model pile tests, it was concluded that the use
of FBG sensor in nioiiitoriiig the strain in piles
shows great potential for the aiialysis of pile load
transfer with the accuracy.

5 CONCLUSIONS
FBG seiisor used in this research as a representathe
of fiber optic sensors is considered to be a promising
sensor system that can replace traditional strain gage
for the measurement of pile load transfer. It showed
the advaiitages of durability aiid sensitrvity The
dependence of the return wavelength oii strain was
l . 2 8 p d P strain and there showed a rclatively good
agreement i n correlation betwccn two sensor
systems.
The quality of Bragg grating fabrication aiid the
method of installation of FBG sensor into a structurc
should be certified for the usc of optical fiber sensor
in practice. It is cleared that strain values measured
by FBG sensor are smaller than those of straiii gauge Figure 1 1 . Comparison o f Pile axial loads ( P I . I J 2 . P.3)
because o f a stiff adhesive used to bond the FBG
sensor on steel. But, this problem is resolled by presented smoother trend 111 measuring 1 alues,
using a less stiff adhesive to help Bragg grating slightly almpt trend of strain values of strain gage
sensor respond more elastically Therefore, i t is was observed as load iiicreased during pile load test.
concluded fiorn above result that the iiiethod to From the observations 011broken strain gages during
install FBG sensor has a significant influence on the installation, i t appears that the durability of FBG
accuracy of strain values iiieasured sensor IS supei-ior to that of strain gauge aiid that it
It was observed that the difference between tMo has a promising potential as a useful tool for
strain values ineasui-e by FBG and strain gage I S inoiiitoring geo-structures iii a field
ranged from 5% to 10%. Whereas FBG sciisor In this study, i t is shown that FBG sensor has a

105
sufficient potential to measure the load transfer
incchanisni of pile. However, since the wavelength
shift of FBG sensor is sensitive to temperature as
well as strain, it needs to take consideration on the
side effect due to temperature when evaluating pile
load precisely. Also, more advanced embedding
method should be developed to use FBG sensor in
the field.

REFERENCES

Aboutaha, M., Roeck, G. D. and Van Impe, W. F.(1993),

"Bored versiis displacement piles in sand-experimental
study". Deep Foundation on Bored and Auger Piles,
pp. 157 - 162.
Claus. R. 0. and Muiphy K. A(1993), "Optical fiber sensors
for the quantitative nieasureiiient of strain in reinforced
concrete structures", Experiments in Smart Materials and
Structures, ASME, Vol. 181, pp.97- 101.
Ilandi-idge A.. J.H Cole, and G.H Siege1 Jr.. (1984) Optical
Fiber Sensors. Technical Digest-Symposium on Optical
Fiber Measurements, pp.49 - 54.
Kltyner, I. M., and Krizek, R. J. (1995), "Mathematical Model
for Bore-Injected Cement Grout Installations", Journal of
Ceotechnical Engineering, Vol. 121, No. 11, pp.782 -788.
Kulhawy, F. E-I. and Turner J. P.( 1989), "Load transfer in deep
foundations under repeated axial loading", Foundation
-
Engineering, Vol. I . pp.5 1 1 525.
Worse, T. F. and Reinhart L. J.( 1993), "Esperimental results on
embedded optical fiber sensors in concrete", Experiments in
Smart Materials and Structures, ASME, Vol. 18 1,
-
pp.53 60.
Oh. .I€3. . (1999), "Measurement of Pile Load Transfer using
Fiber Bragg Grating Optical Sensors", M. Sc. Thesis, Korea
Ilniv.. Seoul, Korea.
O'Neill. M.W. and Raines. R. D.(1991), "Load transfer for
pipe piles in highly pressured dense sand", Journal of
(ieotechnical Engineering Division, ASCE, Vol. 117, No.
S. pp. 1208 - 1226.
Ostrandci-, K. A.( 1990), "Application of fiber optics in strain
n i e a ~ ~ i r e n i e noft ~structures", M. Sc. Thesis, University of
Colorado, Boulder, U.S.A.
Sirkis. J. S.(1999), "Optical fiber strain sensing in engineering
mechanics", short course notes, SPIE's International
Syniposiuiii on Smart Structures, CA.
Rcinhart. I-. J. and Morse, T . F.( 1993). "Experimental results
on embedded optical fiber sensors in concrete",
Espcriiiients in Smart Materials and Structures, ASME,
Vol. 1S1. pp.53-60.
Yoon. S. S.( 1998), "An experiiiiental study on the
characteristics of bearing capacity of auger-cast pile
installed using expansive mortar", M. Sc. Thesis, Korea
Univ.. Seoul, Korea.

106
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Set-up effect of cohesive soils in pile capacity

M. R. Svinkin
VibraConsult, Clevelund, Ohio,USA
R. Skov
CP Test a Js, Vejle,Denmark

ABSTRACT: Knowledge of pile capacity over a long period of time after the end of initial driving is
important for proper design, construction and estimation of the cost of pile foundations. In this paper,
assessment of pile capacity as a function of time has been performed for cohesive soils. On the basis of an
existing formula, a new relationship between pile capacity and time after pile installation has been derived.
This relationship takes into account pile capacity at the end of driving and an actual time after pile
installation. Derived results can be used as a guide for evaluation of long term capacity of piles in cohesive
soils.

1 INTRODUCTION Assessment of pile capacity as a function of time

Piles have to withstand design loads for a long
period of time. Therefore the consequences of soil
modification around the pile are essential with
respect to changes of pile capacity. During pile
installation, the soil around the pile experiences
plastic deformations, remolding, and pore pressure
changes. Excess pore water pressure developed
during driving reduces the effective soil shear
strength and ultimate pile capacity. After the
completion of pile driving, soil reconsolidation in
cohesive soils, manifested by the dissipation of
excess pore pressure at the soil-pile interface zone,
is usually accompanied by an increase in pile
capacity (soil set-up). The amount of increase in
pile capacity and the time required for return of
equilibrium conditions depend on soil properties and
pile characteristics. For example, the disturbed zone
around a pile is more or less proportional to the soil
volume displaced during driving and dissipation of
excess pore water pressure occurs faster in friction
soils.
The phenomenon of time-dependent strength gain
in cohesive soils related to pile driving has been
studied and published, for example, Fellenius et
al. (1989), Randolph et al. (1979), Skov & Denver
(1988), Seed & Reese (1955), Svinkinet al. (1994),
Thorburn & Rigden (1980), Tomlinson (1971),
Wardle et al. (1992), and others.
is, of course, important in the design and
construction of pile foundations. Having knowledge
of general tendencies of pile capacity with time after
driving would certainly be beneficial in economical
standpoint. Such information may be used during
construction to reduce the design penetration and
pile capacity at the end of initial driving and also to
choose relevant time for dynamic testing at restrikes
or static loading test. In this paper, an attempt is
made to present the relationships between pile
capacity and elapsed time after the end of initial
driving (EOID) for cohesive soils and to show some
benefits for estimation of pile capacity from this
approach.
2 EXISTING FORMULA
Skov and Denver (1988) found the following
formula for ultimate pile capacity, R,(t), as a
function of relative time between different tests
Some designations in this formula are different from
those given in the original expression. The pile
capacity at the first restrike, RRSTR.,, is the lower
limit for appreciable increasing in pile capacity when

107
some time elapsed after initial driving results in
developing soil set-up. Quantity, t, is a time elapsed
from the end of initial driving and capacity RRSTR-1 is
determined for t=t,. A factor, A, is dependent on
soil conditions.
Consolidation of cohesive soils around the pile
after pile installation requires much more time in
comparison with other soil types like sand and
gravel to regain in soil strength and pile-soil
adhesion after EOID. For this reason the existing
formula is pertinent for clay and cohesive soils.
Case histories presented by Svinkin et al. (1994)
confirmed that formula (1) is a good indicator of the
pile capacity versus relative time relationship after
pile installation. Besides, it was shown that the
factor, A, depends not only on the soil but also on
the pile type and size.
The application of formula (1) usually requires to
make the first restrike in 1-2 days after pile driving.
It might be especially convenient for dynamic testing
at construction sites where many piles should be
tested by restrike in a short period of time.
Nevertheless, the existing formula yields relative
set-up versus relative time for assessment of the pile
capacity after the first restrike. This is a
contradiction to the set-up approach commonly used
in geotechnical practice and inconvenient for certain
construction sites to restrike piles on 1-2 days after
pile installation.
3 PROPOSED FORMULA
The main goal to derive a new formula for
evaluation of the set-up effect of cohesive soils in
pile capacity is to take into consideration the pile
capacity at EOID and the actual time elapsed after
pile installation.
For a soil set-up straight line passing through two
points corresponding pile capacity at EOID, REOID,
and pile capacity at any time after pile installation,
R,(t), a formula with a logarithmic time scale can be
written
Time is calculated in days after pile installation.
The time for EOID is taken 0.1 (2.4 hours) that
negligibly affects increasing in the pile capacity at
EOID but gives an opportunity to use the
logarithmic time scale.
108
After simplification formula (2) becomes
R"(t>
- 1 = B rlog,,(t) + 11 (3)
RE,,,
Formula (3) is similar to formula (1). However,
proposed formula (3) has certain advantages. This
formula (a) uses the traditional set-up formulation,
(b) takes into account the actual time in days passed
after pile installation, (c) provides determination of
the soil set-up independently of the time of the first
restrike.
4 CASE HISTORIES
For verification of formula (3), in the following case
histories the existing and proposed expressions were
used to calculate the pile capacity as a function of
time after EOID.
4.1 Case I
Three piles were considered in this study from total
number of seven prestressed concrete piles tested for
a bridge approach (GRL Report 1987, Svinkin et al.
1994). A pile description is presented in Table 1.
The depth of penetration of each pile was
approximately 24.4 m. The soil consisted of about
25.6 m of mainly gray clays followed by a bearing
layer of silty sand. Water table was at the ground
surface. A Delmag D 46-13 hammer was employed
for initial driving and restrikes. For each pile, 3 to
4 dynamic testings were performed after pile
installation. For piles TP3 and TP4 static loading
tests (SLT) were made as well. The elapsed time
after EOID, the penetration resistance and the time
dependent ultimate capacity of piles tested are shown
in Table 1. Pile capacities from dynamic testings
were determined by CAPWAP analysis and the
Davisson criterion of failure load was used for static
loading tests (GRL Manual 1993).
Pile capacity calculation according to formula (1)
was made by Svinkin et al. (1994). Quantity, t,,
was equal to 2 days for piles TPl and 1 day for
piles TP3 and TP4. Measured data and results
calculated in accordance with the existing formula
have been plotted in ordinates Ru(t)/RR,,R.l and
loglo(t/tl)as broken and straight lines, respectively,
in Figure 1. Note that the set-up coefficients from
the field tests generally match well to those obtained
from the existing formula.
For piles under consideration, the factor, B, has
 Table 1. Static and Dynamic Data for Prestressed Concrete and Steel Piles in Clayey Soils
Pile Time
Test after Penetration Ru Factor Set-up Set-up
No. Description EOID Resistance Measd Calcd
(days) (blows/0.3 m) (kN) B
TP1 1372 x 127 mm EOID 38 752 1.60 1 1
Cylinder RSTR- 1 2 > 240 245 1 3.26 3.08
Prestressed RSTR-2 9 > 240 2927 3.89 4.13
Concrete RSTR-3 22 > 240 3545 4.71 4.75
~~

TP3 610 x 6 1 0 mm EOID 10 267 2.37 1 1

(305 mm D. void) RSTR-1 1 21 912 3.42 3.37
Prestressed RSTR-2 10 72 1530 5.73 5.74
Concrete RSTR-3 18 144 1672 6.26 6.35
SLT 31 1841 6.90 6.91
TP4 762 x 762 mm EOID 14 200 3.50 1 1
(475 mm D. void) RSTR-1 1 23 890 4.45 4.45
Prestressed RSTR-2 4 60 1299 6.50 6.61
Concrete RSTR-3 9 > 240 1517 7.60 7.84
RSTR-4 18 168 1601 8.00 8.90
SLT 32 2273 11.37 9.77
B-2 HP 310x94 EOID 12 489 1.14 1 1
RSTR- 1 2 36 1201 2.46 2.48
RSTR-2 6 60
RSTR-3 7 72 1512 3.09 3.10
SLT 15 1397 2.86 3.48
RSTR-4 16 48 2002 4.09 3.50
RSTR-5 132 > 120 229 1 4.69 4.55

EOID - end of initial driving

RSTR - restrike
SLT - static loading test

been found on the basis of back calculations using
formula (3). This factor ranges from 1.6 to 3.5
(Table 1). Scattering of the factor, B, is the same
for the existing and proposed formulae. The set-up
coefficients derived from both the field tests and
from formula (3) are given in Table 1 and depicted
in ordinates Ru(t)/REOIDand Time after EOID in days
(logarithmic scale) as broken and straight lines,
respectively, in Figure 1. Good agreement is found
between calculated and measured values of pile
capacity as a function of time after EOID.
4.2 Case 2
Initial data for this case were taken after Fellenius et
al. (1989). An H-pile 310x94 (mm, kg/m) with
length of 47.2 m was driven and restruck by a
Vulcan 010 h a m e r with a nominal energy of 44 kJ.
Restrike No. 4 of this pile was performed by drop
hammer with nominal energy of 65 kJ. The soil at
the site consisted of about 6.1 m of miscellaneous
earth fill followed by about 19.8 m of soft to
medium stiff compressible post-glacial silty clay and
clayey silt underlain by about 27.4 m of glacial
material deposited on dolomite bedrock. The water
table was about 2.5 m below grade. The H-pile was
founded in the glacial material.
Five restrikes were made for pile tested. Pile
capacities from dynamic testings were determined by
CAPWAP analysis (GRL Manual 1993). The static
loading test for this pile did not show a plunging
behavior. Failure load from the static loading test
was 1397 kN (Davisson 1972), 1957 kN (Butler &
Hoy 1977), and 2535 kN (Fuller & Hoy 1970). The
capacity from the static loading test was evaluated
from the Davisson criterion. The results of dynamic
and static tests are shown in Table 1.
Pile capacity calculation according to formula (1)
was made by Svinkin et al. (1994) and shown in
Figure 1. It can be seen the calculated set-up line
averages measured values of pile capacity.
Measured soil set-up and calculated set-up in
accordance with formula (3) are presented in Table 1

109
Figur,e 1. Pile capacity-time relationship for prestressed concrete and steel piles in clayey soils
and displayed in Figure 1 as well. For the H-pile,
the proposed formula provides better fit to results
tested than the existing formula.
5 SUMMARY
The application of the proposed formula shows that
the magnitude of gain in pile capacity in cohesive
110
soils depends on soil conditions, pile material and
dimensions. However, the same equation with a
different factor, B, can be applied for assessment of
soil set-up in cohesive soils during relatively long
elapsed time after pile installation.
The existing and proposed formulae demonstrate
similar tendency of the set-up effect of cohesive soil
in pile capacity. However, there are substantial
differences between two approaches.
 The existing formula yields relative set-up versus
relative time for assessment of the pile capacity after
the first restrike. This is a contradiction to the set-
up approach commonly used in geotechnical
practice. Also, it is inconvenient for certain
construction sites to restrike piles on 1-2 days after
pile installation. Moreover there is no standard time
for the first restrike. If this time is different for
various piles, the existing formula yields different
assessment of soil set-up at the same site and
obtained results of increasing in pile capacity cannot
be compared.
The proposed formula uses the traditional set-up
formulation, calculates the pile capacity at the actual
time after EOID, and provides determination of the
soil set-up independently of the time of the first
restrike.
The proposed formula provides determination of
pile capacity as a function of time after pile
installation using pile capacity values obtained
through dynamic testing at EOID and one restrike or
one static loading test. This approach is identical
for any construction site. Obtained information
would be beneficial to choose sensible pile
penetration depth at EOID and search the proper
times in days after EOID to verify set-up in cohesive
soils with an additional dynamic testing or the static
loading test.
6 CONCLUSIONS
Determination of long term pile capacity is
important for proper design and construction of pile
foundations in cohesive soils.
A new relationship like a linear equation in a
logarithmic time scale has been derived to predict an
increase in pile capacity with time after pile
installation. Soil set-up for several piles in cohesive
soils was verified on the basis of the existing and
proposed formulae. The latter has certain
advantages. Obtained results showed that changes of
pile capacity with time in cohesive soil may be
predicted well.
Results presented in this paper certainly do not
mean that pile capacity will change with time exactly
like shown above. However, the demonstrated pile
capacity versus time relationships can be used as
guide for assessment of pile capacity with respect to
time. Presented results give a chance to safe
significant amount of time, energy and materials
taking into account the gain of pile capacity from
soil set-up. Derived relationship can also be useful
in choosing representative times for both static
loading test and dynamic restrike testing.
111
REFERENCES
Butler, H.D. & H.E. Hoy 1977. Users manual for
the Texas quick-load method for foundation load
testing. FHWA, Office of Development,
Washington.
Davisson, M.T. 1972. High capacity piles. Proc.,
Lecture Series, Innovations in Foundation
Construction, ASCE, Illinois Section.
Fellenius, B.H., R.E. Riker, A.J. O’Brien, & G.R.
Tracy 1989. Dynamic and static testing in soil
exhibiting set-up. Journal of Geotechnical
Engineering, 115(7): 984-1001.
Fuller, R.M. & H.E. Hoy 1970. Pile load tests
including quick -load test method, conventional
methods and interpretations. HRB 333: 76-86.
GRL and Associates, Inc. 1993. CAPWAP - Case
Pile Wave Analysis Program, Manual, Cleveland,
Ohio.
GRL and Associates, Inc. 1987. Dynamic pile tests
pe@ormed during June and July, 1987, Advance
Pile Test Program, Louisiana DOT, Project No.
450-36-06, Cleveland, Ohio.
Randolph, M.F., J.P. Carter & C.P. Wroth 1979.
Driven piles in clay - the effect of installation and
subsequent consolidation. Geotechnique, 29(4) :
36 1-393.
Skov, R. & H. Denver 1988. Time-dependence of
bearing capacity of piles. In B. Fellenius (ed),
Proc. Third Inter. Con. on the Application of
Stress- Wave Theory to Piles, Ottawa, 25-27 May:
879-888, Vancouver: BiTech Publisher.
Seed, H.B. & L.C. Reese 1955. The action of soft
clay along friction piles. Transactions, ASCE,
122: 731-754.
Svinkin, M.R., C.M. Morgano & M. Morvant
1994. Pile capacity as a function of time in clayey
and sandy soils. Proc. Fifth Inter. Con$ and
Exhibition on Piling and Deep Foundations,
Bruges, 13-15 June: 1.11.1-1.11.8, Rotterdam:
Balkema.
Thorburn, S. & W.J. Rigden 1980. A practical
study of pile behavior. Proc. 12th Annual OfSshore
Technology Con., Houston.
Tomlinson, M.J. 1971. Some effects of pile driving
on skin friction behavior of piles. Proc. Institution
of Civil Engineers: 107-114, London.
Wardle, I.F., G. Price & T.J. Freeman 1992. Effect
of time and maintained load on the ultimate
capacity of piles in stiff clay. Piling: European
practice and worldwide trends, Proc. Institution of
Civil Engineers: 92-99, London: Telford.
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Modulus of elasticity and stiffness of composite hammer cushions

M. R. Svinkin
VibraConsult, Cleveland, Ohio, USA

ABSTRACT: Elastic properties of cushion materials are used to determine elastic properties of composite
hammer cushions. Equivalent stiffness of the composite hammer cushion depends mostly on the
characteristics of soft cushion material and only the thickness of a soft material, not the total cushion
thickness, should be taken for calculations. Equivalent modulus of elasticity of the composite hammer
cushion depends on the modulus of elasticity of a soft material and a ratio of stiff to soft layers thicknesses.
A change of soft and stiff layers thicknesses and their ratio may be used as a tool to increase force
transmitted to the pile.

1 INTRODUCTION of elasticity and stiffness of different laminated

Hammer cushion is installed in a well on a top of
the drive cap (helmet) under the anvil which is
struck by the hammer ram. The drive cap is
employed to hold the pile head in position under
hammer and to transfer impact energy to the pile.
The hammer cushion is used for two opposite
purposes. On the one hand the hammer cushion
must transfer the hammer energy to the pile without
excessive energy losses, but on the other hand the
hammer cushion has to prevent pile and hammer
damage from driving.
Various materials such as wood, rope, polymers,
fibers, aluminum and others are placed in the
hammer cushion well. Combination of materials,
like aluminum with conbest or micarta, are
frequently used for hammer cushions (Practical
guidelines 1984; GRL Manual 1997; SPS 1999;
Penn State Fabricators 1999).
Laminated materials such as aluminum and
conbest or aluminum and micarta have a relatively
constant elasticity during relatively long life,
consistent and predictable energy transfer and more
uniform driving results.
The transfer of hammer energy to the pile and
protection of pile and hammer from possible damage
during pile driving depend on the moduli of
elasticity and the stiffness of the materials used to
composite hammer cushions.
The purpose of this paper is to show how moduli
cushion materials effect equivalent modulus of
elasticity and equivalent stiffness of the composite
hammer cushions.
2 TWO CUSHION MATERIALS
A composite hammer cushion has alternate layers of
soft materials like conbest or micarta and layers of
stiff materials like aluminum or steel.
Stiffness of a soft layer, &oft, and a stiff layer, btf,
may be written
where Esoft= modulus of elasticity of soft material;
Estf = modulus of elasticity of stiff material; A =
cross-section of cushion materials; tsoft= thickness
of soft material layer; tstf = thickness of stiff
material layer.
It is common that Estfis considerably greater than
Esoftand Gtf is equal or less than tsoft.
2.1 Equal number of layers
Soft and stiff material layers are in series.
Therefore equivalent stiffness of the composite
hammer cushion, keq, is

113

where n = number of layers of each material;
remaining parameters are the same as defined
previously. Equation (2) may be rewritten as
Relationship between equivalent stiffness and
equivalent modulus of elasticity is
Equating the right pars of equations (3) and (4), we
obtain
Equation (3) can be simplified to analyze a
contribution of each cushion material to elastic
properties of the composite hammer cushion. Since
in the denominator of equation (3) is 30-
term tsoftEstf
100 times greater than term tstfEsoft,
the latter term
can be neglected. After simplification, equation (3)
becomes
According to equation ( 6 ) , equivalent stiffness of the
composite hammer cushion depends on the modulus
of elasticity, the layer thickness, and the cross-
section of a soft material.
After analogous simplification, equation ( 5 ) takes
the form
(7)
where a! = tstf/tsoft.
It can be seen that equivalent modulus of elasticity
of the composite hammer cushion depends on the
modulus of elasticity of a soft material and a ratio of
stiff to soft layers thicknesses.
Calculation of the equivalent stiffness and the
equivalent modulus of elasticity using simplified
equations (6) and (7) has an error margin about 3 %
in comparison with results of equations (3) and (5).
114
2.2 Unequal numbers of layers
A stiff material has usually one additional layer in a
combination of conbest or micarta with aluminum.
Equivalent stiffness of the composite hammer
cushion, keq, can be expressed
where all parameter are as defined previously.
After transformation and simplification,
equation (8) becomes equal to equation (6).
3 THREE CUSHION MATERIALS
Composite cushions of three materials like
aluminum, micarta and steel rope are sometimes
used. Such cushions consist of two soft and one
stiff materials connected in series. Assume that each
cushion material has one layer and two soft layers
have the same thickness, tsoft. Also, assume a ratio
of the materials moduli of elasticity as
Estf> > Esoft> where Esoftest is modulus of
elasticity of the softest material.
Equivalent stiffness of the composite hammer
cushion, k,,may be written
After transformation and simplification, equation (9)
takes the form
According to equation (lO), equivalent stiffness of
the composite hammer cushion filled with three
materials depends on the modulus of elasticity of
two soft materials, the thickness of a soft material
layer and the cross-section of cushion materials.
Relationship between equivalent stiffness and
equivalent modulus of elasticity is
Equating the right parts of equations (10) and ( l l ) ,
and, assume Esoft=2Esoftestfor simplicity, we obtain
 (12) 4.2 Three cushion materials

Equivalent stiffness of the composite hammer

For assumptions taken, equivalent modulus of
elasticity depends on the modulus of elasticity of the
softest material and the ratio of stiff to soft layers
thicknesses.
4. DISCUSSION OF RESULTS
cushion in equation (10) depends on the moduli of
elasticity, the thicknesses, and the cross-section of
the softest and soft material layers. According to
equation (12), equivalent modulus of elasticity
depends to a lesser degree on the ratio of stiff to soft
layers in comparison with two cushion materials.

4.1 Two cushion materials

5 CONCLUSIONS

Equivalent modulus of elasticity and equivalent Proper determination of elastic properties of

stiffness of the composite hammer cushion are used composite hammer cushions is important for the
application of the wave equation method to piles.
in wave equation analysis of pile driveability and
pile capacity. It is common that the total thickness Equivalent stiffness of the composite hammer
cushion depends mostly on elastic properties of a
of hammer cushion is taken for wave equation
analysis. soft cushion material: the modulus of elasticity, the
layer thickness and the cross-section.
Equation ( 6 ) shows that only the soft material
The total thickness of composite hammer cushion
effects the equivalent stiffness and only the thickness
is usually taken into account for wave equation
of a soft material should be used in calculations.
analysis of pile driveability and pile capacity. Since
Decreasing the total thickness of the composite
a soft material mostly effects the equivalent
hammer cushion to the actual thickness of a soft
stiffness, only the thickness of a soft material should
material increases force transmitted to the pile and
be used in calculations.
provides more realistic consideration of hammer
Equivalent modulus of elasticity of the composite
cushion properties in wave equation analysis of pile
hammer cushion depends on the modulus of
driveability and capacity.
elasticity of soft material and the ratio of stiff to soft
According to equation ( 7 ) , the ratio of stiff to soft
layers thicknesses.
layers thicknesses effects the equivalent modulus of
A change of soft and stiff layers thicknesses and
elasticity. Values of E,, are increased with
their ratio may be used as a tool to increase force
increasing the aluminum or steel thickness and
transmitted to the pile. This may improve pile
keeping the same conbest or micarta thickness. For
driveability without switching to more powerful
example, an enlargement of the aluminum layer
hammer for certain driving conditions.
thickness from 1 to 3 inches with the same conbest
thickness of 1 inch increases two times the value
of E,.
REFERENCES
Thus, equivalent stiffness depends on the thickness
of soft cushion layers and equivalent modulus of
GRL and Associates, Inc. 1997. G R L W A P - Wave
elasticity depends on the ratio of stiff to soft layers
thicknesses. Therefore a change of layers
Equation Analysis of Pile Driving, Manual,
Cleveland, Ohio.
thicknesses provides certain flexibility to regulate
Penn State Fabricators. 1999. Conbest cushion
elastic properties of the composite hammer cushions
and, under certain driveability conditions, gives an
blocks - Information, New York.
opportunity to increase force transmitted to the pile
Practical guidelines for the selection, design and
for account of the layers thicknesses change instead
installation of piles. 1984. Committee on Deep
of switching to more powerful hammer. Such a Foundations, ASCE.
Specialty Piling Systems, Inc. 1999. Hammer
hammer will increase force transmitted down the
pile in limits allowable by the pile impedance. It is cushion materials - Information, Slidell,
reasonable to change the thickness of soft and stiff Louisiana.
layers of the composite hammer cushion as the first
step in increasing dynamic force applied to the pile.

115
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 1503

A new technique to drive piles: ‘Drive-the-holepiling’

A. Benamar
Laboratoire de Me‘canique, Universite‘du Havre, France

ABSTRACT: A new pile driving technique by which the hammer is inside the pile has been developed by con-
tractors. An experimental set-up is used in laboratory to drive a tubular pile with the new driving technique by
impact hammer inside the pile, near the toe. This new driving technique enables to use long hammers inside the
pile. The ‘down-the-hole’ piling process is tested in laboratory and compared with top pile driving using a quite
similar energy. The hole pile driving is performed by a ram falling into the pile bottom. The advantages of this
device are essentially the reduction of wall thickness of piles and the noise compared with the hammer at the
pile top, It is also concluded for an improved driving efficiency. In comparison between top piling and ‘down-
the-hole’ technique, it is observed that in top piling the displacement at the pile head decreases when drop
height increases, while in the new driving technique the displacement is not af‘fected by such parameter.

1 INTRODUCTION hole” driving pulls the pile down (tension stress

wave). The results of the two driving techniques are
The recent improvement in the estimation of load ca- compared and some advantages of the new device are
pacity of piles by dynamic methods has resulted from enlightened.
the use of the wave equation. This approach enables a
rational analysis to be made of the stresses in the pile
during driving. Dynamic methods have certain ad- 2 PRINCIPLE OF (( DOWN-THE-HOLE ))

vantages and some uncertainties in their application. DRIVING SYSTEM

It is used in dynamic methods that driving resistance
is carried out by the analysis of stress waves meas-
ured near the pile head. In the traditional top-pile-
driving the incident wave is quickly influenced by the
shaft friction wave making the evaluation of driving
resistance difficult. A new technique to drive tubular
piles by impact hammer inside the pile has been de-
veloped and tested in France by some contractors.
This new “Down-the-hole’’ Piling technique devel-
oped for offshore application is tested in onshore site
and has confirmed in field tests a consistent reduction
of noise (20 to 30 dBA) and the opportunity to save
steel (Arentsen et al. 1996). This new driving tech-
nique enables to use long hammers inside the pile.
Also a reduction in driving time and a higher bearing
capacity has been observed during a test program of
this technique. The technique must enable more pre-
cise bearing capacity calculations.
An experimental set-up is used in laboratory to
drive a tubular pile by the two techniques : pile top
driving and (( Down-the-hole )) technique (Benamar
1999a). The main difference between the two driving
techniques is that top piling pushes the pile into the
soil (compression stress wave) while “Down-the-
The system developed by contractors in field tests in-
volves independent penetration of the pile tip and the
pile itself, using a special (( driving head D. This driv-
ing head is connected to a slender hydraulic hammer.
The recoverable pile tip is in fact a part of the driving
head (Roger et al. 1997). The ram of the hammer
hits the central part of the driving head without any
cushion. This central part, with conical tip, penetrates
the soil independently from the pile wall. The princi-
ple of such device is to involve independent penetra-
tion of the pile tip and the pile itself. This technique is
efficient when the pile is driven through successive
soft and hard soil layers. The steel to steel ram and tip
causes a high peak force for penetration of hard soil.
Where a closed bottom pile is required, a driving
head with a sacrificial plate can be used. When driv-
ing inside the pile, a slender hammer is required to
drive small diameter piles. This new driving technique
enables to use long hammers inside the pile. The
choose of conical pile tip is not evident and geometri-
cal shape obviously governs the pile rate penetration.

117
3 . LABORATORY DRIVING TESTS
The driving tests are performed in a tank of 1.50 m
diameter and 2.00 m high, filled with a loose sand.
The density of sand is about 1.2 and of a range 0/5
min. In this first step of the experimental study the
L’Down-the-hole” piling tests are performed without
using the device allowing uncoupled penetration of
pile tip and pile shaft. The hammer strike simply in
the bottom hole of pile (figure 1). The pile model
made of steel, is 1.50 in long and 0.42 m of diameter.
The conical pile tip with an angle of 40 degrees is
welded to pile shaft. The driving system is composed
of a 2 in long tube surrounding the top of the model-
pile, and guiding a free falling mass which impacts on
top or inside the hole pile, without any cushion. The
test pile is instrumented with strain gauges in order to
collect the stresses near the head and the toe of the
pile. A displacement sensor is placed at the pile head.
Figure 1 : Laboratory Hole-Pile-Driving device
The “Down-the-hole” Piling technique is tested
and compared with top pile driving by the hammer
involving the same energy In driving piles it is essen-
tial to control the hammer energy and the way it is
transmitted to the pile and the soil. The energy in-
volved in the two types of driving tests is ranging
from 13.5 to 34.6 J. In top piling the hammer pushes
the pile into the soil, allowing a compression stress
wave, while in “Down-the-hole” driving the hammer
pulls the pile down, creating a tension stress wave
The hammer used is enough long to develop a S U E -
cient blow duration to mobilize the total shaft resis-
tance Because the center of gravity of the pile and
118
hammer system is much lower, directional stability
during driving is improved. Subsequently tip deflec-
tion, torsion and bending pile stress can be avoided
4 STRESS WAVES PROPAGATION IN THE
PILE
The typical time-force signals recorded during down-
the-hole driving at top of pile is shown on figures 2 .
Unlike the top piling where the initial stress wave is
of compression and propagate from the pile head to
the tip, in “Down-the-hole’’ driving the initial wave
at pile tip is of tension and propagate upwards to the
top. The tension wave TI (figure 2) is then reflected
at pile top as a compressive wave C I . After the first
up and down travel of the stress wave in the pile, its
amplitude is quite reduced by shaft and tip interaction
with surrounding soil. The initial tension wave inter-
acts firstly with the soil near the pile tip. The reflected
waves from the pile tip are compressive stress and
superpose with initial tension wave induced by rain
stroke. The up and down time travel of the wave is
measured on figure 2 and agree with theoretical value
2L/C where L is pile length (1 S O m) and C the wave
celerity (5 180 ids). Because of the nature of driving
which impacts the pile tip, the shaft friction is less so-
licited and the radial dissipation may be lesser com-
pared with pile top driving. The force measured
shows the weak shaft friction of pile during Hole Pile
Driving. The main difference between the two driving
techniques is that top piling pushes the pile into the
soil (compression stress wave and positive Poisson
effect) while down-the-hole driving pulls the pile
down (tension stress wave and negative Poisson ef-
fect). The negative Poisson effect must be more
marked for thin piles.
In classical driving the shaft friction is firstly mobi-
lized in the pile-soil interaction, leading to a subse-
quent disturbance of the surrounding soil. The initial
compression wave travel downwards to the pile tip
(figure 3). The soil layer is than compressed creating
a high stress field around the pile and an increasing in
shaft resistance is observed. But in “Down-the-hole”
driving the tip resistance is first mobilized and the ini-
tial tension wave strongly attenuated travels upwards
to the pile top (figure 4). The energy transfer from
the pile to the soil is made through the tension wave
transmitted and reflected at the pile tip.
The stress waves measured on the pile allow to
identify the incident and resistant waves using the
stress wave theory in an unidimensional elastic media.
Stress wave measured in top driving show that the
wave duration increase after the successive reflections
at the pile extremities (Benamar 1999b). This effect is
caused by the high amplitude of shaft resistance. But
in “Down-the-hole” driving the initial tension wave
is weakly influenced by friction waves during
propagation along the pile.
 stresses amplitudes and the unit friction are more im-
portant in the top piling because of the densification
of the sand around the pile by the initial compression
wave. This effect may be more marked in dense sand
involving hard driving This characteristic allows an
interesting save of steel by the reduction of wall
thickness of piles with consequent cost reduction.

Figure 2 : Force wave measured at pile top

Figure 4: Wave propagation in Down-the-hole Piling

Table 1 . Stress levels involved during driving

L Down-the-Hole-Piling
Maximal Energie Maximal
Force (kN) 0 Force (kN)
13.46 22 12 13.85
17.31 27.61 15.58
19.23 33.65 19.04
21.15 3 1.90 22.50
37.04

Figure 3 . Wave propagation in top driving

Table 2. Unit skin friction during driving
Down-the-Hol
5. PILE - SOIL INTERACTION

In comparison between top piling and down-the-hole 20.58 15.58

driving, it is observed that in top piling the displace-
ment at the pile head decreases when the embedment
increases, while in down-the-hole driving the dis-
placement is not affected by such parameter, in spite
of the densification of soil around the pile The pile
penetration is less influenced by the embedment in
“Down-the-hole” driving This effect can be ex-
plained by the weak influence of the shaft friction on
the penetration and by the particular transmission way
of the hammer energy to the pile-soil system Consid-
ering the transmitted hammer energy to the pile, the
stress levels in the two driving techniques are com-
pared Table 1 show the measured force of initial
stress wave with different hammer energies in the two
driving techniques. With the same driving energy, the
19.23 16.80 19.04 09.05
21.15 10.40 22.50
The unit shaft friction w is evaluated using the am-
plitude of the initial stress wave before and after
crossing the pile shaft. The results seem to be compa-
rable with those obtained in field measurements. In
table 2 are reported the skin friction values measured
in the two driving processes with different hammer
energies. It is clear that for a similar energy the unit
shaft friction is more intense in top driving. The dif-
ference amplitude seems to decrease linearly with the
hammer energy for the three results presented. When
the hammer hits pile tip the shaft friction is less solic-

119
ited and the radial dissipation may be lesser com-
pared with pile top driving. The friction developed
around the pile shaft has a more important influence
on the pile penetration than the driving energy.

6. CONCLUSIONS

The experimental study performed aims to compare

the two driving techniques and set the essential char-
acteristics of “Down-the-hole” driving. At this
stage, preliminary conclusions from the laboratory
test results can be stated. The advantages of the new
piling technique are mainly the possibility of reduction
wall thickness of piles and the improved driving efi-
ciency Because of the particular character of hammer
blow the interaction at pile-soil interface may be dif-
ferent than the classical modeling. Another advan-
tages of this new device is the possibility to use
long hammers. But it is not concluded the easy driv-
ing through hard soil layers as it was in field tests.
Apart from savings in steel, this new technique elimi-
nates many problems encountered in top driving. The
possibility of the “Down-the-hole” uncoupling pile
driving allows to consider an economical implemen-
tation of tubular piles. This technique is being im-
proved and extended to uncoupling pile driving and
tested in hard driving conditions.

REFERENCES

Arentsen, D , J L Pautre & S El A b a 1996 New

technique to drive piles by impact hammer inside
the pile I-’roceedings of the 6th International
Conference and Exhibition, Deep Foundation
Institute, Bombay, India, I 5 - I 7 January I996
Benamar A 1999a Influence of driving impact on the
axial response of piles Proc 4th C’onference of
the European Association for Stuuctural 1Jynam-
ICS, Eurodyn’99, 7 - 10 .June 1999, Prague pp
1019-1023, Balkema
Benamar A 1999b Dynamic pile response using two
pile driving techniques Proc 9‘” Int Conf on Soil
Dynamics and Earthquake Engineering, 9- I 2
A t g IY99, Bergen, Norway
Roger, P , J L pautre & D Arentsen 1997 The pile
bottom technique (or driving without its disad-
vantages) Revue Pavaux, Nr 727 , pp 76 -80
Paris

120
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 1503

Simplified neural network models for estimating soil resistance using

dynamic pile test
R. Liang & N.O. Nawari
CivilEngineering Department, University of Akron, Ohio, USA

ABSTRACT: Neural network formulations are attempted to predict the soil resistance under dynamic
conditions. These models are based on simple input data from the dynamic pile test. In this investigation, only
HST data and the geometrical properties of the pile are invoked in the models. Thus making the models
effective and appropriate for the daily practice. The models involved are simple 3-layer and 4-layer
backpropagation networks. Data from more than 90 driven piles, representing different subsurface conditions
were utilized in the investigation. The results and comparisons with other methods indicate that the neural
network models fullill the reliability and adequacy desired in the prediction of soil resistance under dynamic
loads.

1 INTRODUCTION A more rational approach for determining the soil

Presently, there are a large variety of formulas
available for estimating the soil resistance fiom the
dynamic pile test. For instance the Engineering
News, Hiley, Janbu, Danish and Gates formulas
among others. The vast majority of these formulas
are derived largely based on Newton's law of
impact. In these formulas, the driving energy
transmitted from the ram to the pile is determined
using impulse-momentum principles. The soil
resistance is then obtained by equating the energy
transmitted to the pile to the work done by the
penetration resistance. The criticism of the use of
such formulations is that they over-simplify the pile
driving behavior and that the characteristics of the
soil are not taken into consideration.
Despite the oversimplification of the physical laws
used to drive the pile driving formulas, they continue
t o enjoy popularity among practicing engineers
because of their simplicity. It is undeniable that there
is some form of relationship between those
parameters used in these pile driving formulas and
the soil resistance. Due to the complexity of the
problem, setting a suitable regression model to
represent this relationship will also be arduous if not
imp0 ssible.
resistance under dynamic conditions is the wave
equation model (Smith, 1960) and the application of
the methods described by Forehand and Reese
(1964) and Chow et.al. (1988), among others. The
use of these techniques was developed into a
commercial computer code called CAPWAP
(Rausche et al., 1972).
Employing the wave equation model in
determining the soil resistance and taking account of
the soil characteristics, CAPWAP provides a much
more reliable estimation than pile driving formulas.
However, this approach is time consuming due to
the iteration process and more profoundly, it needs
extensive experience to achieve reliable results.
An alternative approach to the problem is the
application of neural network paradigms to estimate
the soil resistance under dynamic excitations. Neural
networks are computational models that are
massively parallel adaptive system. The essential
advantages of the neural network stem from the fact
that they are capable of learning complex nonlinear
relationships from a large store of given examples
and similar to human intellect, are fault receptive.
Moreover, they do not assume any a priori physical
low or regression model. Availability of field
measurements which involve the use of the dynamic

121
pile test data made it possible to investigate the use
of neural network for estimating the soil resistance
using the same parameters as the simplified dynamic
test data (High Stress Testing, (HST)).
2. NEURAL NETWORK APPROACH
2.1 General
Recently some investigators suggested artificial
neural networks as a model for the prediction of the
behavior of axially loaded piles, using Hiley formula
(Chan et al., 1995; Teh et.al. 1997). No attempts
have been made so far to investigate the application
of artificial neural networks in the prediction of the
soil resistance under dynamic loading using a
combination of HST data and soil properties from
simple soil tests.
The main objective is to develop optimal models
using only simple input data. These data include the
geometry of the pile, soil properties in terms of type
and SPT-N values and HST data. The models
involved are backpropagation neural networks.
2.2 Neural network structure
Artificial neural networks are nonmodel-based
pattern recognition and approximation methods.
They are also known to be noise tolerant; they are
adaptive and they can learn and generalize.
The network architecture affects the optimal
training and its ability to generalize. The general
criterion in designing neural networks is to start from
the simplest structure that can provide the essential
consistency and adequacy. In this study, varieties of
neural network models are introduced for the design
of friction piles. Namely, Feedforward
Backpropagation (see figure 1) are implemented. It
also addresses the data preparation and data
inception along with the problems involved with the
artificial neural networks. A description of these
architectures can be found in many publications in
neural network, for example Dayhoff (1990) and
Hecht-Nielsen(1990) among others.
In this research, a three-layer and four-layer
feedforward backpropagation networks were
investigated. Choosing an appropriate number of
hidden neurons is extremely important aspect in the
backpropagation networks. However, there is no
exact method for determining the number of hidden
layer neurons. Using too many will increase the
training time and may cause the overfitting problem
122
(memorizing the training pattern rather than
generalizing the prediction). On the other hand,
using fewer hidden neurons often increased the
likelihood of learning algorithm becoming trapped in
a local minimum. Thus, it is imperative that we use
absolute minimum number of hidden neurons, which
will perform adequately.
One rough guideline for choosing the number of
hidden neurons in many problems is the geometric
pyramid rule. It states that, for many practical
networks, the number of neurons follows a pyramid
shape, with the number decreasing from the input
toward output. The number of neurons in each layer
follows a geometric progression. Other investigators
(Berke and Hajela, 1991) suggested that the nodes
on the hidden layer should be between the average
and the sum of the nodes on the input and output
layers. These are only rough approximations to the
ideal hidden layer size. The best approach, found by
the authors, to find the optimal number of hidden
neurons is to start with a few numbers of neurons,
then slightly increase the number of hidden neurons,
until no significant improvement is noted.
A variety of different transfer hnctions were
investigated to achieve best performance in training
as well as in testing. The optimal performance was
obtained from t ansigmoid-logsigmoid-logsigmoid-
linear activation h c t i o n s in the first, second, third
and output layers respectively.
Figure 1: Typical Feedforward Backpropagation
Neural Network
3 . DATA PREPREOCESSING
This phase involves data preparation and
organization before feeding the network model. The
data used in the current neural network development
is primarily from the GRL report. Basically, a total
number of 200 pile cases were provided in the GRL
report. However, after more detailed examination of
the completeness of each pile case, only 94 piles
cases have been accepted. These piles cases are max. Force, FMX, max. velocity, VMX, max.
classified into three groups: steel pipe pile, H-pile, displacement, DMX, and max energy transferred,
and pre-stressed concrete pile. Table 1 shows the EMS (iv) Soil properties in terms of soil types and
number of pile cases in each pile group, as well as SPT-N values. In addition, differentiation between
EOD and BOR is made by assigning different case
the number of cases in neural network training and
numbers. For EOD, the case number is 0, whereas,
validation.
for BOR, the case number is 1. The case damping
value is also included in the input vector.
Table 1 Number of pile cases for each pile group Table 2 lists the data structure required by the
Pile type Notation Total No. for No. for neural network model. The input parameters are
No. Training Testing basically the SPT-values with depth, Pile length,
cross-sectional area, circumference and the amount
Steel Pipe pile 27 16 11 of steel reinforcement.
Pipe
piles Table 2: Input Data Structure
H shape H pile 28 17 11 Pile NO. 1 2 3 ... 27
steel
Pile 21.33 36.72 24.53 ... 19.81
piles
Pre- PSC pile 39 23 16
stressed
concrete
piles
(m2)
Database preparation for the training of the neural Toe Area 0.082 0.164
network represents a crucial step in the neural (m2)
network modeling. The performance of the network Circumfer 1 .O 18 1.436
model rests solely upon the input training pattern. ence (m)
The best source of the input training pattern in Shaft Soil 2 8
complex phenomenal like the soil-structure Type
interaction in axially loaded piles, is an experimental Toe Soil 1 7
investigation in which a large number of cases are Type
properly tested. The measured data depicts the Length 20.12 35.75
inputs and outputs variables for the training of the Below Gage
neural network. These data were then divided up (m)
into training and testing subsets. The training subset Penetra- 20.12 30.2

""i~
constitutes 60% of the database whereas the testing tion (m)
40%. Blow 30 17
The development of the training network starts Count
with the selection of a number of different
combinations of input variables to evaluate the most
reliable neural network model. Then, about 20% of
the patterns in the training set is extracted before
training and used for the cross validation to satisfl 53.55
the adequacy of the generalization of the proposed
neural network paradigms.
For the pre-processing phase, all input and output
data are normalized to values between 0-1. This
scaling was performed in accordance with the A numerical codes has been developed for soil types
probability approach (Master, 1993). to facilitate their representation in the input vector.
The suggested neural network model is supposed Table 3 shows these numerical code assignments.
t o utilize a combinations of HST data and soil
properties , namely this include: (i) Pile geometrical
data, such as pile length, cross-sectional area, and 4. E S U L T S AND DISCUSSION
circumference (ii) Pile driving record, including pile
length below gage, pile penetration depth, blow During the training phase, the measured axial pile
count per foot (ii) Selected HST data, including capacities are compared with the capacities obtained

123
Table 3 Numerical Codes for Soil Types

Figure 3 : Predicted and measured soil resistance for

H piles.

Figures 2 to 4 present the measured pile capacities

versus the results predicted by the backpropagation
networks for steel pipe piles, H piles, and prestressed
concrete piles. These results show that the regression
surfaces generated by the feedforward
backpropagation networks in the case of steel pipe
piles and H piles were sufficient to predict, with high
accuracy, the total soil resistance under dynamic
loading. The simulation results for concrete piles
exhibit a lower coefficient of correlation than the H-
piles.

Figure 2: Predicted and measured soil resistance for

steel pipe piles

by neural network models after the training phase,

the neural network models are capable of
reproducing the target output values with minimal
error. Next, the reliability of the trained model in
producing correct responses for a new set of data is
examined. Once the training, testing and validation
phases are successhlly accomplished, the neural
network obtained can be used as a practical design
tool for driven piles.
A three-layer and four-layer feedforward
backpropagation networks were investigated with
different number of neurons in the hidden layers. The Figure 4: Predicted and soil resistance for
results indicated that the four-layer network
concrete pipe piles
produced better results with optimal number of
neurons in the middle layers of 6 &d 3 neurons.

124
CONCLUSIONS
Simplified backpropagaton neural network model is
introduced to estimate the soil resistance under
dynamic loading conditions. The approximation of
the physical behavior of the pile in these models is
based only on basic data, like pile length, pile cross
sectional area, circumference and HST data along
with soil characteristics. This combination of data
seems to enhance the prediction accuracy of the
neural network models. The presented neural
network structures provide promising results that
will be beneficial in the practice.
The chief preference of the neural network
methods over pile driving formulas is that it does not
require any priori assumptions, either in terms of the
physical behavior or the mathematical models.
It has been shown that the four-layer
backpropagation network model satisfies a
reasonable accuracy for the practical applications
and reveals a promising potential of the neural
network approach for estimating the soil resistance
under dynamic excitation.
Master, T. (1 993), “Practical Neural Network
Recipes in C++”, Academic Press Inc., New
York.
Rausche, F., Moses, F. and Goble, G.G. (1972),
“Soil resistance predictions from pile dynamics”,
Journal of soil mechanics and foundation
division, ASCE, 98, pp.917-937.
Smith, E.A. L. (1960), “Pile driving analysis by the
wave equation”, Journal of soil mechanics and
foundation division, ASCE, 86, pp. 35-61
Teh, C.I., Wong, K.S., A.T.C. Goh, and S.
Jaritngam, ( I 997), “Prediction of Pile Capacity
using neural networks”, J. Comp. In Civ. Engrg.,
London, pp. 1-8 .

REFERENCES
Berke, l., and Hajela, P. (1991), “Application of
neural nets in structural optimisation”,
NATO/AGARD Advanced Study Institute, Vol.
23(1-11), Berschtesgaden, Germany, pp.73 1-745.
Chan, W.T., Chow, Y.K. and Liu, L.F. (1995)
“Neural Network: An alternative to pile driving
formulas”. Computers and Geotechnics, Elsvier,
Amsterdam, pp. 135-156.
Chow, Y.K., Karunaratne, G.P., Wong, K. Y. and
Lee, S. L., (1988), “ Prediction of load-carrying
capacity of driven piles”, Canadian Geotechnical
Journal, 25, pp. 13-23
Dayhoff, J.E, (1 990), “ Neural Network
Architecture”, Van Nostrand Reinhold, New
York.
Forehand, P. W. and Reese, J.L., (1964), “Prediction
of pile capacity by the wave equation”, Journal
of soil mechanics and foundation division,
ASCE, 90, pp. 1-25.
Hecht-Nielsen, R, ( 1990), “Neurocomputing”,
Addison-Wesley Publishing Company, Reading
Liang, R., (1 998) “Development and Implementation
of New Driven Piles Technology”, The Ohio
Department of Transportation and the US
Department of Transportation, Federal Highway
Administration

125
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3

ile capacity prediction using neural networks technique

A. S. Dyminsky & E. Parente-Ribeiro
Federal University of Parand, Curitiba, Brazil
C.Romanel
Catholic University of Rio de Janeiro, Brazil
J.W. Beim
PDI Engenharia, Rio de Janeiro, Brazil

ABSTRACT: Dynamic pile tests are widely accepted as a rapid and accessible way of estimating the pile
bearing capacity, through the application of stress wave theory. Mathematical methods were proposed to pro-
ceed the interpretation of the signals obtained from pile instrumentation. The simplified CASE method pro-
duces results that have to be confirmed by more elaborate programs, like the CAPWAPB. This program, on
the other hand, is time consuming and its precision depends to some extent on the professional experience of
the user. To overcome these limitations, the authors are proposing to couple the CAPWAP program with the
neural networks technique. To accomplish this, a data file was assembled by collecting information from
more than two hundred Brazilian dynamic test reports. This data was used to train and test neural networks in
order to obtain the pile bearing capacity that would be found by the CAPWAP analysis.

1 INTRODUCTION tance, maximum elastic deformation (quake) and

Dynamic load tests have become widely used in the
evaluation of the bearing capacity of foundation
piles. Their execution is relatively simple, compared
to the traditional static load test, and much more
economic.
However, the analysis of the data from a dynamic
load test is complex. The signals collected by the
sensors are usually the deformation and the accel-
eration of the pile, which are later converted to force
and velocity. These signals are analyzed using the
wave propagation theory, in order to indirectly esti-
mate the bearing capacity of the tested pile. Mathe-
matical methods based on the wave equation were
developed for making this analysis, as described by
Gongalves et al. (1 996).
The CASE method utilizes a closed form mathe-
matical solution, to calculate the total soil resistance
mobilized by the dynamic loading of the pile. It con-
siders the soil as an ideally plastic medium, and the
pile as perfectly elastic and homogeneous. It relies
on an estimated factor to eliminate the velocity de-
pendent component of the soil resistance.
The CAPWAP method is based on the model first
devised by E.A.L.Smith (1960). The pile is divided
into segments, and the soil-structure behavior of
each one is analyzed. The soil reaction, considered
as a passive force, is represented by elasto-plastic
and visco-linear components. The soil model has
three unknowns for each point: ultimate static resis-
damping constants.
In the CAPWAF' analysis, the soil parameters are
initially estimated. Those parameters are then used
to calculate the response of the system to the applied
dynamic load. The calculated curve is compared to
the measured curve, and the soil parameters are in-
teractively changed until a reasonable match of the
two curves is obtained. The adjustment of the soil
parameters during the analysis depends to a large
extent on the user of the computer program. He
needs to have a good notion of the effects on the cal-
culated curve of the changes in soil parameters. He
also has to judge if the final model is realistic, and if
the parameters are within reasonable limits.
2 PROPOSED MODELLING
The soil-pile model adopted by the CAPWAP
method is far more refined than that of the CASE
method, and it allows for more reliable estimates of
the bearing capacity. On the other hand, it demands
a higher analysis time, and previous experience fiom
the engineer in the utilization of the method.
A less complex and time consuming method than
the CAPWAF' analysis was therefore sought, but
with the same reliability. The neural network tech-
nique was the opted alternative. It does not use com-
plex pre-determined mathematical models to inter-
pret the dynamic pile-soil behavior. It is instead

127
based on the simulation of the problem using exist-
ing experimental data, and would allow for bearing
capacity predictions to be made in the field, simulta-
neous with the collecting of data.
Other authors have already used neural networks
to determine the bearing capacity of piles using data
from load tests. For example Teh et. al. (1997) used
digitized samples of the force and velocity curves as
inputs to the network, and Kiefa (1998), utilized
general regression neural networks to obtain the
shaft and tip resistance of driven piles in cohesion-
less soils. A different and innovative approach is
used in this work, as described in the following ses-
sions.
3 DATASET
The choice and use of a consistent data set is of the
utmost importance in modeling physical phenomena
through neural networks. Only the use of enough
consistent information, with a good space distribu-
tion, will allow the neural network to adequately
simulate the problem at hand.
In the present case, a set of reports of dynamic
load tests executed in Brazil between 1996 and 1999
was available. From these reports a good amount of
data could be extracted, such as: characteristics of
the tested pile, of the driving equipment, of the soil
where the pile was executed or driven, as well as in-
formation about the force and velocity curves versus
time, pile set, bearing capacity estimated by the
CAPWAP method and other information relative to
the analysis.
Next, a data bank was assembled using an elec-
tronic spread sheet, where the information was gath-
ered, compared and was checked for coherence. The
data included tests on steel piles, reinforced and
prestressed concrete piles, and cast in situ piles, with
diameters varying from 0.18 ni to 1.O m.
As the majority (about 70%) of the load tests in-
volved concrete piles, it was decided to use data
only from this kind of pile in this study, to avoid bad
data distribution. Besides that, in this preliminary
stage only load tests executed on piles with diame-
ters ranging from 0.18 m to 0.70 m were considered.
The few piles with larger diameters were excluded,
as well as those piles for which soil information (like
SPT tests close to where the pile was installed) was
not available.
Thus, a set of 60 load tests results was obtained,
of concrete piles with diameters varying from 0.18
m to 0.70 m, with a data space distribution adequate
for training (51 load tests), and for generalization
and result prediction tests (two subsets with 6 and 3
load tests, herein identified respectively as TEST0 1
and TEST02).
128
4 MODEL PARAMETERS
It is known that several factors can influence the
bearing capacity of a pile and, consequently, the re-
sult of a dynamic load test obtained by the CAP-
WAP method. However, it is impossible to precisely
assess the relative importance of the several inter-
vening factors, due to the complexity of the dynamic
phenomenon.
Thus, several variables were selected to be used
as inputs to the neural networks, in an attempt to
study the importance of each one in the deterniina-
tion of the total mobilized resistance by the CAP-
WAP analysis. Those variables are listed on Table 1,
together with their code names.
Table 1. Parameters used as inputs to the neural networks
Code Parameter descriution
AR Pile cioss section area
LE Pile length
EM Pile elastic modulus
EMX Maximum energy transferred to the pile
DMX Maximum pile displacement at the sensors
location (plastic + elastic displacements)
Set Pile set due to the blow analyzed by CAP-
WAP (plastic displacement)
DMX-Set Elastic portion of the total pile displace-
ment due to the blow used in the analysis
Aveiage number of blows of the SPT sam-
SPTs11a.A
pler along the pile shaft
SPT,,, Number of blows of the SPT sampler at the
depth corresponding to the pile toe
5 NEURALNETWORKS
Artificial neural network is an alternative computa-
tional paradigm. It relies on learning a set of exam-
ples presented to the network. It has a parallel dis-
tributed structure that makes it very fast to operate.
It also has the ability to generalize, i.e., to produce
reasonable outputs for inputs not presented during
the learning process. Like the brain, a neural net-
work is composed of a set of neurons connected with
each other by synapses. The artificial neuron is a
non linear element interconnected to other neurons
by a synaptic weight. The way in which neurons are
interconnected is called network architecture. More
detailed description of artificial neural networks can
be found in the literature (Haykin 1994).
The network architecture used in this work is the
multi-layer perceptron which is a feed-forward net-
work. It consists of neurons distributed in layers
completely interconnected.
A very common architecture, with one hidden
layer, was used in this research. It is composed by an
input layer of nl source nodes (nl is the number of
input parameters), one hidden layer composed of 1
to 10 neurons, and an output layer with 1 neuron,
with linear activation function. This type of neural Each neural network was initialized 5 times, i.e.
network was chosen based on the authors previous was trained and tested S different times. This proce-
experience in this kind of problem (Dyminski et al. dure was done to mitigate the effect on non optimum
1995, 1999a and 1999b). The purpose of the hidden local minima in the network error surface. The stop
neurons is to intervene between the external input criterion used in the networks training was SO ep-
and the network output and, in this case, the se- ochs or a minimum error gradient.
lected activation function used the logistic sigmoid.
The algorithm used to train the neural network
was the LM ( L ~ v e n b e r g - ~ a r q u ~ method,
dt) that 6 RESULTS
has shown to be faster than other training algorithms
such as the error backpropagation with delta rule. The average training and test errors can be ob-
Because of the modeling aspects commented on item served in Tables 3a-e. The best results for each type.
4, the authors have chosen to train and test several of neural network are written in boldface. Those
neural networks using different sets of input pa- values were calculated considering the iowest errors
rameters. attained by the network, especially during the test
The root mean squared error (rmse) was used as a phases where its generalization capacity was under
figure of merit to compare the networks. This error evaluation. The standard deviation (s) has to be also
was calculated for the training patterns and the test considered, because i f s assumes a high value then
patterns. Because the input and output data pre- the neural network training and test may not be sta-
sented to the network was normalized to lie between ble.
zero and one, the rmse represents also a percentage
error. ons side ring N patterns presented to the net-
works with desired output value or target, tj, the
Number of NNl NN2
networks produce outputs, oi, that should be as close
neurons Mean Error Standard Mean Error Standard
as possible to tj. Their difference is used to calculate
the error as shown below. ("h) deviation ("3) deviation
01 Tr 8.0 0.0 8.5 0.1

The neural networks were organized according to

the parameters used as their input, as shown in Table
2. Each configuration had networks of I , 2, 3, 5 , 8 05 Tr 3.9 0.7 3.1 0.8
and 10 neurons in the hidden layer trained and TS 32.0 36.9 21.3 14.0
tested. Almost all network types have AR,LE, and 08 Tr 2.5 0.3 2.1 1 0.5
EMX as input, because the first two parameters are TS 43.6 54.0 28.1 I 28.2
necessary to define the geometry of the pile and the 10 Tr 1.9 0.5 1.5 1 0.3
last one to define the energy applied in the system. TS 20.0 13.1 21.5 1 11.9
But the neuron network NN9 does not have the pile
geometry parameters as input because, in this case,
the authors tried to represent the simplified energy
approach formulation (Paikowsky & Chemauskas Number of NN3 NN4
1992) which basically uses those three parameters to neurons Mean Error Standard Mean Error Standard
define an equation that estimates the pile resistance. ("/.I deviation (%I deviation
01 Tr 8.0 1 0.0 9.0 0.0
Table 2. Type of neural networks according to its input pa-
rameters.
Tvpe of neural network Input Parameters
"1 AR, LE, EMX, Set
"2 AR, LE, EMX, DMX
"3 AR, LE, EMX, DMX, Set
"4 AR, LE, EMX, (DMX-Set)
"5 AR, LE, EM, EMX, DMX, Set
"6 AR, LE, EMX, Set, SPTsha8,SPT,,,
"7 AR, LE, EMX, DMX, Set, SPTsha8,
SPT,,,
"8 AR,'LE, EMX, (DMX-Set), SPTShaA,
SPTtO,
"9 EMX. DMX. Set

129
Table 3c. Training (Tr) and test (Ts) mean errors and their re- Three different network types - " 3 , NN6 and
spective standard deviations - "5 and " 6
1 I I NN9 - had their training and test results plotted in
Number of I NNS I NN6 I Figures 1.a - f. The horizontal axis corresponds to
the pile resistance values (Rs), normalized between
0 and 1, and the vertical axis refers to the neural
network output.
The number of neurons in the hidden layer for
each network was chosen based on the results shown
in Table 3, thus resulting a two neuron hidden layer
for " 3 and a single neuron hidden layer for NN6
and "9. Some dispersion can also be observed in
the cases shown in Figure 1, but the scattering is al-
most the same in these three neural networks, phe-
nomenon that could indicate that the energy applied
to the system, and its resulting deformation, are the
most important parameters for estimating the total
pile capacity.
A different approach can also be adopted to solve
the same problem, selecting neural networks with a
larger number of neurons (8 to 10) in the hidden
layer. Although they may present a weaker stability
than the smaller networks and be more susceptible
during training to overfitting risks and local mini-
mum effects, larger networks can be quite success-
ful, as shown in the results from Table 4, obtained
using neural networks similar to "3, "6 and
NN9 but possessing a 10 neuron hidden layer.
Many initializations were carried out (more than
10 times) in order to choose the best network of each
type, with different networks reaching the optimum
point, regarding both the training and test errors, at
different number of iterations. Table 4 shows the
training and test errors for the revised "3, "6
and NN9 networks, considering the set named
TEST01 (06 dynamic load tests) for the generaliza-
10 Tr I 0.0I 0.1 I 0.7 I 0.2 tion (or prediction) test.
TS 20.0 1 8.4 I 27.5 I 15.2 Given the good results obtained with this ap-
proach, a new and independent data set (TEST02
Table 3e. Training (Tr) and test (Ts) mean errors and their rce- with 3 pile load tests) was used to confirm the gen-
eralization capacity of such revised networks. As
shown in figures 2a-f and 3a-c, the training and test
errors thus computed give strong evidences that the
10 neuron hidden layer networks are quite able to
successfully predict the load capacity of the piles
analyzed.

Table 4. Ten neurons in the hidden layer networks training and

test errors and the number of iterations to reach them.
Neural Network Number of Training Test
Type iterations error (%) error (%)
"3 50 0.3 5.6
NN6 26 0.001 6.8
"9 150 1.4 9.2

130
 Figure la. Training results for network " 3 Figure Id. Test results for network "6

Figure lb. Test results for network "3

Figure 1e. Training results for network "9

Figure lc. Training results for network "6

Figure 1f. Test results for network "9

For these specific cases, the generalization capac-

ity of the larger networks was superior to the smaller
networks (Figure 1) considering both data sets
TEST01 and TEST02. Finally, it could be observed
that the revised "3 and "6 networks yield better
results than the correspondent "9. This can be ex-
plained by the fact that the increase of neurons in the
hidden layer gives networks extra degrees of free-
dom, thus improving their ability to deal with a
higher number of input parameters ( 5 parameters for
"3, 7 for "6 but just 3 for "9).

131
Figure 2a. Training results for the revised (10 neurons) "
3 Figure 2d, Test results for the revised (10 neurons) "6 net-
network using data set TESTOl in the generalization process work using data set TESTO 1 in the generalization process

Figure 2b. Test results for the revised (10 neurons) "3 net-
work using data set TESTO 1 in the generalization process Figure 2e. Training results for the revised (10 neurons) "
9
network using data set TESTOl in the generalization process

Figure 2c. Training results for the revised (10 neurons) NN6
using data set TEST0 1 in the generalization process Figure 2f. Test for the revised ( neurons) NN9 net-
work using data set TESTO 1 in the generalization process

7 CONCLUSIONS AND DISCUSSIONS from SPT tests) did not affect the neural network re-
sults in a very significant way, as indicated by the
Observing the results shown on Table 3, it is possi- small differences detected when comparing the error
ble to conclude that the soil characteristics (drawn estimates from networks " 3 , " 6 and "9 and

132

Figure 3a. Test results for the revised (10 neurons) NN3
network using data set TEST02 in the generalization process
Figure 3b. Test results for the revised (10 neurons) NN6 net.-
work using data set TEST02 in the generalization process
Figure 3b. Test results for the revised (10 neurons) NN9 net-
work using data set TEST02 in the generalization process
the other configurations considered in this research.
Evidently, this does not mean that the soil type
can not influence the distribution of shear resistance
along the shaft or it can not affect the relationship
133
between the shaft and pile tip resistance, since these
more detailed aspects of the dynamic soil-structure
interaction were not studied in this paper.
On the other hand, results point out that the en-
ergy applied to the system (EMX) and the elastic
and plastic displacements (DMX and Set) are the
most important input parameters that must be care-
fully dealt with for a correct interpretation of dy-
namic pile load tests.
It is also very important to study different neural
networks architectures and to initialize them several
times to prevent local minimum effects. Neural net-
works with just a few neurons (1 to 5 neurons) in the
hidden layer could satisfactorily predict the total pile
resistance for every architecture considered in Table
2. Neural networks with a higher number of neurons
in the hidden layer (8 to 10) have shown yet smaller
training and test errors, for some specific configura-
tions.
The study of the influence of the pile type on its
load capacity behavior will be described in a future
paper, where results from concrete and steel pile
load tests will be jointly analyzed.
8 ACKNOWLEDGMENTS
The authors wish to thank Prof. Alessander C.
Morales Kormann, from the Federal University of
Parana, for helpful discussions about PDA tests and
the CAPWAP method.
REFERENCES
Dyminski A.S.; E. Parente Ribeiro, C. Romanel, C.E. Pedreira,
“Modeling Residual Soils Direct Shear Tests Using Neural
Networks”. Proc. Of X v I CILAMCE, Curitiba, Brazil., pp
684-692 (1 995) (in Portuguese).
Dyminski, A . S., Romanel, C. A “Simple Neural Network for
soil Modeling”, Applied Mechanics in the Americas, vol.
06, pp. 459-462 Jad1999.
Dyminski, A.S., Ribeiro, E.P., Weingaertner, C., Quoirin, N.R.,
Romanel, C. “Modeling The Mechanical Behavior Of Cu-
ritiba Experimental Site’s Clay” P i - 0 ~ofXx Cilarnce, SBo
Paulo, Brazil., Nov/1999.
GonCalves, C., Andreo, C.S., Bemardes, G.P. “Dynamic Pile
Test” Edited by Estacas Benaton, 1996 (in portuguese).
Haykin, S., “Neural Networks - A Comprehensive Founda-
tion.” New York: MacMillan College Publishing Company
(1994).
Kiefa, M. A . A . (1998) “General Regression Neural Networks
for Driven Piles in Cohesionless Soils” ASCE - Journal of
Geotechnical and Geoenvironmental Eng., Vol. 124, No.
12, December11998.
Paikowsky, S.G. and Chemauskas, L.R. (1992) “Energy Ap-
proach for Capacity Evaluation of Driven Piles” Proc. of
the Fourth Int. Conf. on Application of Stress Wave Theory
to Piles, The Hague, Netherlands, Sepd1992.

Smith, E.A.L. (1960), “Pile Driving Analysis by the Wave

Equation” - Journal of the Soil Mechanics and Foundation
Division, ASCE, SM4, vol. 86, 34-61.

Teh, C.I., Wong, K.S., Goh, A.T.C. and Jaritngam, S.(1997)

“Prediction of Pile Capacity using Neural Networks” ASCE
- Journal of Computing in Civil Eng., Vol. 11, No. 2,
ApriV1997.

134
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

armner and pile cushion optimalisation

G.Jonker & R. J.vm Foeken

IHC Hydrohammer and TNO Building and Construction Research, Rijswijk, Netherlands

ABSTRACT: The pile cushion has an important fbnction in the protection of the concrete pile fiom tensile
forces. In The Netherlands the most common applied pile cushion is composed of two or three layers of soft-
wood. Recent research by IHC and TNO has shown that the cushion behavior determines to a large extend the
driving performance and energy transfer to (concrete) piles. It has also shown that different type of hammers
need different type of cushions. This paper describes the experimental research, field testing and stress wave
simulations which have been performed to understand the behavior of the drive setup and to determine the op-
timum drive cap configuration for IHC hammers.

1 INTRODUCTION the behaviour. The results of these investigations will

Due to the sensitivity of concrete piles for tensile
stresses the use of a hammer cushion and/or a pile
cushion is a must for every pile driving hammer. The
pile cushion is used mainly to reduce the peak impact
force and to reduce the risk of pile damage by tensile
stresses. Part of the impact energy is absorbed due to
the hysteresis effect in the pile cushion and is trans-
formed in to heat. Hammer cushions are mainly
added to reduce the noise f+om the steel to steel im-
pact on the striker plate or anvil and to protect the
hammer fiom rebound forces. The impact resistance
of the hammer cushion material should be high with a
long service life.
For diescl hammers a soft pile cushion is required
to significantly reduce the peak force, since the im-
pact energy of the first blow can not be reduced. But
for hydraulic impact hammers the impact energy can
be manually controlled and therefore the stresses in
the pile can be controlled.
The mechanism of energy transfer fiom the impact
hammer to hammer cushion, helmet, pile cushion
and the concrete pile is discussed. Full scale tests
have been done to monitor the behavior of the driv-
ing set up and have been compared with stress wave
predictions. Hammer and pile cushion materials have
been evaluated in laboratory conditions to optimise
be shown and discussed.
2 PRINCIPLE OF ENERGY TRANSFER FOR
HYDRAULIC IMPACT HAMMERS
The impact energy of a hydraulic hammer can be
controlled, however the direct impact of the ram on a
concrete pile is still not desirable. The high impact
stresses will damage the pile head and the peak
stresses may lead to cracking of the pile. In general a
pile cushion of softwood is placed between the pile
head and the helmet. Also cushions of plywood are
used for spun piles. The drive setup of only an anvil
with a pile cushion gives a single spring mass system,
see figure 1.
This basic drive set-up can be extended with a
hammer cushion between the ram and the anvil. The
anvil is now called a helmet. The main function of the
hammer cushion is to reduce the noise fiom the steel
to steel impact. The use of a hammer cushion, a hel-
met and a pile cushion gives a double spring mass
system. In general the stifhess of the hammer cush-
ion is five to ten times stiffer than the pile cushion. In
case of a too soft pile cushion the velocity of the
helmet will exceed the velocity of the ram and will
move away from the ram. In this system the helmet is

135
accelerated more than the ram because of the soft
pile cushion spring. The helmet can move 1 to 20
mm before a reaction force is activated in the pile
cushion. Because of this high velocity, there will be a
gap between the hammer cushion and the helmet just
after impact. After some milliseconds the gap is
closed and a second impact will occur.
Figure 1: Basic drive set-up, a single spring sys-
tem or double spring system
When measuring during the driving in the field
with a PDA system, in the pile a force with the shape
of a ‘camel’ is clearly shown, see figure 2.
Figure 2: ‘Camel” shaped pile response.
136
For soft pile cushions the peak helmet velocity is
higher than the impact velocity and the pile head ve-
locity. During driving the pile cushion is compressed
permanently and the cushion thickness is reduced
continuously till it is completely burned. During the
Figure 3: Impact velocity and peak velocity of
helmet and pile head as function of penetration.
driving the height of the pile cushion will decrease
due to its compression under the repetitive hammer
blows and as a consequence the s t f i e s s will increase
and the peak velocity of the helmet will reduce. An
example of the impact velocity of the ram, the peak
velocity of the helmet and pile head is shown in fig-
ure 3 as h c t i o n of penetration for an IHC-SC-110
hammer with a 7900 kg ram and max net rated en-
ergy of 110 kNm. This was a pile cushion composed
of a combination of multiplex and formwork board.
Different pile cushions have been investigated in
the field:
o Three fir softwood layers, total thickness of 180-
190 mm.
e Two fir softwood layers, total thickness of 120-
130 mm.
a One softwood layer with a plywood cover at
both sides.
0 Two plywood covers, 40 mm of softwood and
two plywood covers, total thickness of 110 mm.

For hydraulic hammers, especially double acting

hammers the preloading of the pile cushion spring is
very low to virtually zero. This is contrary to diesel
hammers where the compression of the he1 before
impact prestresses the pile cushion and at impact
time the actual stifftiness of the pile cushion is much
higher than without prestress of the hammer. Due to
the different preloading conditions perform stiffer
pile cushions better in combination with IHC hydrau-
lic hammers.

3 PILE CUSHION BEHAVIOUR AND ENERGY

TRANSFER IN FIELD TESTING

In the field under normal driving conditions, the

pile and helmet are instrumented with accelerome-
ters. At the pile also strain sensors are fixed to the
pile at about 2 times the pile diameter below to top,
see figure 4.
With the accelerometers, which are placed at op-
posite sides, the displacement of the helmet and the
pile head is calculated, by integration of the signal to
Figure 4: Position of the sensors on IHC ham-
mers.

velocity and to displacement. The results of the dis-

placements are shifted in time due to the stress wave
propagation. The measured force in the pile is plotted
against the displacement difference of helmet and pile
top.
Also the proximity switches are measured to de-
termine the impact velocity of the hammer. The dif-
ferent measured results are shown below, respec-
tively the proximity sensors A and B, the acceleration
of the pile head and helmet, the force and velocity
times impedance curve at the pile head.
Furthermore the displacement of the helmet and
the pile head, the force in the pile head and both the
displacement signals are shown in one figure. Finally
the load versus compression of the pile cushion can
be determined.

Photo 1: Sensors mounted on the pile and helmet.

137
 Figure 5 : Measurements made in the field

For each blow a measurement as shown above has
been made. The tangent stiffness of the pile cushion
at 5 , 10, 15 and 20 MPa stress level is shown as
hnction of blow number in figure 6. After about 950
blows the pile cushion was replaced with a new one.
During these test no damage of the pile head has
been noticed when using plywood cushions. The
maximum tensile stresses in the pile were about the
same as with softwood.

138
 The results shown in figure 6 are for a pile with a
cross section of 0.25 m2, a IHC SC-110 hammer and
preloaded (4.0 MN) plywood cushion with an initial
height of 105 mm and one normal plywood on top of
the pile to protect the pile head.
Figure 7: Eeciency reduction when applying
softwood.
Figure 6: Tangent stifhess as hnction of blow
number
In another test when driving a pile (350x350 mm')
with an SC-75 at 60 % impact energy, the IHC ply-
wood cushion was changed after about 825 blows
and over a penetration of 0.5 m the driving was done
with a fresh 130 mm softwood cushion. Afler about
70 blows and a penetration of 0.5 m a new ply-
wood/sofiwood 1 10 mm cushion was used for the fi-
nal driving.
The results of measured efficiency (transferred en-
ergy divided by impact energy times 100 Yo)in the
pile are shown in figure 7.
139
During the driving of 25 concrete piles (290x290
mm') with a SC-40 at 65 % impact energy, four dif-
ferent cushions have been randomly used at a small
site in the Netherlands by the Dutch contractor Ken-
nis:
3 layers of softwood (height of 180 mm)
2 layers of softwood (height of 120 mm)
4 layers of plywood and 1 softwood layer of
30 mm (height of 105 mm)
4 layers of plywood (height of 72 mm)
Assuming that the number of blows to drive the
piles with 2 layers of softwood, with is common
practice, is set to 100 % (483 blows). The number of
blows using the 3 layers of softwood is 113 %. With
4 layers of plywood and 1 softwood layer it gives 89
% and for only plywood it gives 78 %. When 30 piles
are driven on one day, the contractor could drive 3
more piles per day.
During these test no damage of the pile head has
been noticed when using any of the plywood cush-
ions.
4 LABORATORY RESEARCH FOR HAMMER
CUSHION BEHAVIOR
During driving the temperature of the hammer
cushion rises as a consequence of the hyteresis in the
behavior of the cushion material under loading and
unloading conditions. Small test specimen with a di-
ameter of 80 mm and a height of 25 to 75 mm have
been tested under pseudo static loading conditions at
different temperatures (20°, 50" ,80" and 95"'. In 0.3
seconds a stress level of 30 [MPa] is applied. This is
 Behavior at 95 " (e=0.70; E25 = 900 MPa)
Behavior at 50 " (e=O.76; E25 = 1700 MPa)

Figure 8: Behavior of MC-901 material at differ-

ent temperatures (E25 is youngs modulus at stress
level of 25 MPa).

kept constant for a very short period and the load is
removed over a period of 0.3 seconds. A lot of alter-
native hammer cushion materials have been investi-
gated. The main selection parameter was the coeffi-
cient of restitution measured during the tests and the
Young's modulus at different stress levels.
Presently IHC is using a nylon based material,
variation on the MC-901. For this materials the ob-
served behaviour is shown at different temperatures
(20", 50" '80" and 95" Celcius)
The measured stress-strain curve for loading and
unloading of the cushion material are fitted with the
following function:

140

The third degree fbnction gives the best results
and is required when describing cushion behavior.
This material model is incorporated in the
TNOWAVE-R stress wave simulation program. For
unloading a linear behavior is assumed.
The tangent stifiess at a stress level of 25 MPa
and the coefficient of restitution as hnction of tem-
perature is shown in figure 9
shown in figure 10. Under impact conditions the
stiflkess of the material will increase approximately
by 30 %.
5 PILE DRIVING PREDICTION
During the driving of a concrete pile with a SC-75
hammer, the behavior of the drive setup was meas-
ured. With the TNOWAVE-R program a prediction
has been made for this pile using the measured be-
havior as input.

First a signal match has been made for the soil pa-
rameters using the PDA results from the measure-
ments on the pile. The behavior of the hammer
cushion is based on the measurements for MC-901 at
80" C. The pile cushion was an plywoodlsofiwood
110 mrn cushion. The results of the pile, pile cushion
and hammer cushion are shown below.

Figure 9: Tangent stiffness and coefficient of res-

titution (upper line) as hnction of temperature.

Figure 11: PDA measurement during driving

Figure 10: Behavior of MC-901 under impact

loading at 20" (e=0.66; E25 = 5400 MPa)

A second test was performed to measure the ef-

fect of the duration of impact. This was done in a
small scale impact hammer test set up. A mass of 80
kg is dropped from a height of 100 mm. The dis-
placement iS measured by double integration of ac- Figure 12: Calculated pile behavior
celerations. The results from the impact test are

141
The PDA measurement was done about 3000
blows on the hammer cushion. This means that the
temperature of the hammer cushion will be about 80
"C This has been adopted in the stress wave simula-
tion.
Shen, H.R., Yu, Z.C, Tang, N.C., Studies on Pile
Driving Cushion MaterailS. Proceedings ofthe Int.
Conf. On Deep Foundations, Beijing, 1-5 Septem-
ber 1986.

In figure 13 and 14 the behavior of the pile and

hammer cushion are shown as applied in the stress
wave simulation.

Figure 13: Modeled hammer cushion behavior

Figure 14: Modeled pile cushion behavior

6 CONCLUSIONS

The presented research is still ongoing. The meas-

ured data during the field tests will have to be com-
pared with the test results on the hammer cushion
material for all tests performed.

The efficiency of the double acting hydraulic SC

hammers has significantly improved by using ply-
wood/softwood cushions. This has been demon-
strated on a lot of sites in The Netherlands and Bel-
gium.

REFERENCES

Pile driving analysis-simulationof hammers,

cushions, piles, and soil. Texas transportation in-
stitute. Texas highway department. Research re-
port 33-9, study 2-5-62-33, Piling behaviour.

142
3 Pile integrity and low strain dynamic testing
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Keynote lecture: Oresundlink, grouting work under West Pylon: Integrity test
of a giant pile
Jan Rome11
Department of Civil Engineering, Skanska Teknik AB, Gothenburg, Sweden

ABSTRACT: Integrity tests were performed at the West Pylon caisson of the high bridge in the 0resundlink
between Sweden and Denmark. The purpose of the tests was to investigate the grout filling works under the
0.9 in thick concrete bottom slab. During the grouting works a stomi occurred which swept away approxi-
mately 1% of the underbase grout causing voids and gaps in the interface between the bottomslab and the
grout. This circumstance jeopardised the bearing capacity of the caisson. The bond between the slab aiid the
grout should be at least 80% according to contract. The investigation showed that this demand was not ful-
filled. First after intense grouting works the 80% connection was achieved. The Non-Destructive tests using
low strain integrity method were covering the 35 x 37 m bottom slab. Tests were also perfomied in the verifi-
cation phase after the repair works were completed.

1 INTRODUCTION ness of this void was = 0.8 m. The minimum thick-

Sundlink Contractors erected the bridges of the Ore-
sund fixed link between Copenhagen in Denmark
and Malino in Sweden. The total bridge project con-
sisted of a Western and Eastern Approach Bridge of’
= 3.3 kin length each and a High Bridge of 1080 in
length. The approach bridges have spans of 120 in
aiid 140 m whilst the cable stayed high bridge has a
main span of 490 in and its Pylons are 202 in high.
The two Pylon Caissons were prefabricated in an
existing dry-dock in Malino. The caissons were rein-
forced concrete structures of outer dimensions 35 x
37 i n and height 21 m (East) and 22.5 m (West) re-
spectively. Each caisson consisted of 42 compart-
ments approximately 5.5 x 5.5 in.
Beginning April 1997 the West Pylon Caisson
was transported to its final location in the bridge
aligninent by a purpose-built catamaran and using a
heavy lifting system. At the location, the caisson
was submerged almost completely into a foundation
pit which had been previously dredged. The caisson
was placed temporarily on three foundation pads,
concrete slabs of size 3 x 3 in with a thin plywood
plate i n between. The foundation level of Pylon
Caisson WP was . 18.5 m, whilst the surrounding
water depth in the Oresund was = 8.00 m.
After placing. the void between the underside of
the caisson bottom slab and the limestone level of
the foundation pit was grouted. The average thick-
ness in some local spots was 0.35 m whilst the
maximum depth was up to 1.0 in. The underside of
the bottom slab had an inverse pyramidal shape to
assist during grouting as the grout should fill the
void completely under the bottom slab. In order to
transfer horizontal loads due to ship impact into the
surrounding soil the grout was extended to 1.2 in
above the foundation level around the circumference
of the caisson.
After grouting. the 0.9 in thick bottomslab and
the 0.8 in ;hick grout layer could be considered as a
35 x 37 in-square pile with a length of 1.7 in.
During the underbase grouting of the West Pylon
Caisson scheduled to last about 5 days a storm oc-
curred. It was discovered by diving inspections, at a
late stage when the grouting operation was almost
completed, that the foundation was damaged. A sand
ridge in the foundation pit outside the North-West
part of the Pylon Caisson was discovered upon coin-
pletioii of the underbase grouting giving the first
indication of possible damage to the foundation.
The repair works of the WP-caisson can be di-
vided into three phases:
1. Investigation phase.
2. Repair works.
3. Verification phase.
The contract stated that an 80% contact between the
bottom slab and the underbase grout had to be ful-
filled. This turned the problem into the investigation
phase.

145
2 INVESTIGATION PHASE
2.1 General
The underbase grout situation under the West Pylon
Caisson was investigated intensively. Several differ-
ent methods were used to clarify the situation under
the base slab prior to repair works and to determine
the best method for performing the repair of the de-
fective underbase grout layer.
The investigations clarified the following two main
subjects:
1. Detection of voids in the interfaces: bottom
slab - underbase grout, underbase grout - limestone
and inside the underbase grout layer.
2. Verification of the structural capacity of the
entrapped sand and the existing underbase grout
layer.
To investigate the situation, the following tests and
investigations methods were used:
0 Core drillings in 7 drilling stages.
0 Endoscope investigations.
0 Impact-Echo measurements.
2.2 Cove drillings
Core drillings were performed as part of the repair
works. Records were kept from all drilling works
being executed and in general terms all boreholes
were used for the execution of the repair as well.
The core logs record the different observations such
as grout, sand, water and loss o f core etc. Including
the investigation drillings a total of 465 drillings
through the caisson bottom slab had been recorded
and were to be treated during the repair works. As
the bottom slab of the WP-Pylon Caisson was a
highly post tensioned structure drilling had to be
minimised and had to be executed under extreme
caution and by taking special measures to limit
structural impairment.
2.3 Endoscope investig~rtions
A total of 220 endoscope investigations were per-
formed. The videoscope investigations were carried
out to clarify the "loss of core" reported in the drill-
ing reports. The main subjects of the endoscope in-
vestigations were:
a Verification of the bond between bottom slab
and underbase grout layer.
0 Evaluation of size and where possible. evalua-
tion of extent of voids under the base slab.
* Investigation of grout or sand material being
present.
0 Investigation of the bond between underbase
grout and the limestone.
146
2.4 Impuct-Echo somdings
2.4.1 General
Non Destructive Tests were performed in 665 loca-
tions covering all 42 compartments. The Impact
Echo Method was used for the NDT investigations.
The purposes of the NDT tests were:
0 To increase the knowledge about the situation
existing underneath the base slab prior to the repair
works. To clarify the presence and spatial distribu-
tion of voids and sand in the underbase grout layer.
a To use the status prior to repair works for later
verification of the repair works by comparative
studies.
The Impact-Echo method is based on stresswave
analysis of an impulse response. The chosen system,
PIT-Collector (Pile Integrity Tester) from GRL is
basically used for integrity tests on piles. The
stresswave from an impulse of a small sledge-
hammer of steel is registered with an accelerometer.
The acceleration is integrated to velocity. The ve-
locity record is studied in the time as well as in the
frequency domain. A simple sketch of the execution
of the measurements is shown in Figure I and Equa-
tion 2. l below.
Figure 1. NDT-measurement system. Low strain
testing system GRL.
A t = 2 S l c and At= 1 I f
=3
s = c l ( 2 x f, (2.1)
S = thickness of the concrete bottom slab.
c = wave speed in concrete (assumed to 4500 d s ) .
f = frequency according to the peaks in the velocity-
time curves.
4t = time elapsed between two peaks in the velocity-
time curve.
Z = acoustic impedance.
Figure 2. Calculation of R-factor between concrete and water

Figure 3. Calculation of R-factor between concrete and sand or grout 50%-water 50%.

Impact-echo tests are based mainly on the propa- x l e 1. Characteristics for different materials.
gation and reflection of compression waves (P- Material P-Wave Density Acoustic R-factor
waves). The P-waves propagate into the solid mate- speed (kg/m3) Impedance
rials along spherical wavefronts. An area and not (mls 1 (kdni2s )
only the measuring location influence the measured Concrete 4500 2400 10.8*106 0.00
response. The measuring positions were located in Grout 100% 3225 2000 6.45*106 0.25
principle with a minimum distance of 500 mm to
surrounding walls and ribs to avoid interferences. As Grout 80%- 5.46*106 0.33
Water 20%
described above, the velocity (integrated from the
acceleration) of the stresswave was studied in time Grout 60%- 4.46" 106 0.42
Water 40%
as well as frequency domain.
Grout 40%- 3.46" 1 O6 0.52
Water 60%
2.4.2 T 7eloci(j)-Timecurve Sand 2000 1700 3.40*106 0.52
The velocity-time curve represents the velocity re- Grout 20%- 2.47*106 0.63
sponse for the impulse in the time domain. The Water 80%
curve was used to calculate a reflection factor. The Sand loose 1500 1500 2.2j*106 0.66
reflection factor, R, is defined in Equation 2.2. Water 1480 1000 1.48*106 0.76

R = (22 - Z l ) /' (22+ Z1) (2.2)

where Z1 is the acoustic impedance of the concrete
and 22 is the acoustic impedance for the material
under the slab. Z is defined as the density times the
P-wave speed. Some Z's for different materials were
assumed. The R-factors with an interface to concrete
were then calculated. The R-factor for concrete-
water interface, for instance. tells that 76% of the
incoming wave is reflected. The theoretical R-
factors for different materials are presented in Table
1.
The reflection factors of the measured curves
were estimated by scaling the amplitude in the ve-
locity curves dividing the amplitude of one reflec-
tion with the one coming before. Then the result was
classified with the Figures listed in the Table above
as a guideline. Examples of this interpretation are
presented in the Figures 2 and 3 below. The first
peak was neglected since initial effects tend to over-
shoot the measured value.
The measurements showed that it could also be of
interest to study the number of oscillations (reflec-
tions) in the velocity-time record. The more oscilla-
tions there are the greater is the difference in acous-
tic impedance between the interfacing materials.
This can also be studied in the Figures above.
2.4.3 ~'elocih~-speccrzrn~
czwve
The velocity is transformed to the frequency domain

147
using Fast-Fourier-Transformation (FFT). This re-
sulting curve represents the velocity response for the
impulse in the frequency domain. In this curve the
peaks are studied. The important details are to find
the fiequencies where peaks appear and the shape of
the peaks. The shape of the curves tells that a sharp
peak appears when the reflection against the under-
lying material is high i.e. reflection against air, water
or very loose sand. A damping material such as sand
or grout causes a more rounded and broader peak. If
the grout has very high acoustic impedance the
stresswave tends to travel through the grout with
only small amount of reflection.
In hnpuct-Echo. Non-destwcti\*e Evuluution of
Concrete m d Mmonry, Sansalone and Streett
( 1 9971, Chapter 15, Figure 106, are described the
different shapes of peaks for different degrees of
bond between two concrete layers. This can in our
case be compared with different degrees of connec-
tion between the concrete slab and the underlymg
grout. The investigations showed that voids and
pockets between the concrete and the grout most
ol'ten caused the problem. Findings of washed in
sand were very limited.
To compare theory with practice, five curves with
different shapes of the peaks were chosen. Such
curves from the West Pylon are presented below in
Figure 4. A short description of significant features
of each type of peak is given below.
Peak A. The response had a nice rounded shape.
The peak could often not be distinguished.
Peak B. The response was broad and had a simi-
lar look before and after the peak. The peak could be
hard to distinguish. Often it was possible to notice a
response from the interface grout-limestone (the first
vertical line in the Figure)
Peak C. The appearance of the peak was more
difficult to distinguish if just looking before the
peak. The fall of the peak was though quite obvious.
The response before the peak was more "filled up"
than after.
Peak D. The peak was more rounded than in peak
E. The fall of the peak was steeper than the rise. The
response around the peak became broader.
Peak E. Both the rise and the fall of the peak was
very steep. The response before and after the peak
had offen low amplitude. The peak was very sharp.
Figure 4. Classification of velocity curves in frequency domain.
148
The interpretations that can be done are presented
in Table 2 below. The degrees of connection were
evaluated by studying the Sansalone and Streett ref-
erence.
Table 2. Interpretation of bond between concrete and
grout.
Degrees of connection Classification
between slab and grout Peak
100 % A
80 - 100 % B
25-75 % C
0-25 % D
0% E
The results of the measurements were classified in
five main groups A-E where A represents full con-
tact and E represents no contact. Eight subgroups
were also established. All thirteen groups were in-
terpreted further in order to establish three areas.
@ Full or close to full contact between slab and
grout [A, ( B N , ( A B B, (GIB].
Partial contact between slab and grout [(B)C,
c , (DKI.
@ No or low contact between slab and grout
[(CID, D, ( W , (D)E, El.
The NDT-measurements were correlated to core
drillings and endoscope investigations, as the NDT-
measurements solely could not tell the actual physi-
cal reason for a reduced degree of connection.
2.5 ~otrc1usion.sofflie iizvestigutionphase
Using the information available from all investiga-
tion results the situation under the Pylon base was
described by establishing three separate categories:
1. Area with good contact.
2. Areas with small gaps 0 - 2 mm.
3. Areas with gaps > 2 mm.
The average size of the voids under the baseslab
observed by the endoscope was < 10 nim but in cer-
tain areas the voids were o f 3 0 - 120 mni thickness.
Regarding the interpretation of the low strain meas-
urements most importance were attached to the ve-
locity-spectrum curve. The structural capacity of the
existing underbase groutfill was verified by per-
forming unconfined compression strength tests on
grout cores. Grout specimens were taken from vari-
ous depths and locations from the underbase grout
fill of the West Pylon Caisson. The average com-
pressive strength was sufficient compared to the
requirements. The load carrying capacity of the en-
trapped sand was tested with respect to the static and
cyclic loading conditions. Based on the test results
the designers demonstrated that entrapped sand in its
in-situ state provided adequate bearing capacity and
thus no further remedial measures on entrapped sand
was required i.e. soil fracture grouting was not
needed.
3 FOUNDATION REPAIR
The foundation repair works, injection works, were
performed between 20th November 1997 and 2 1st
February 1998. The injection works for filling the
voids under Pylon baseslab were executed according
to the following five principal steps:
1. Sealing of limestone fissures.
2. Hydraulic connectivity test (HCT). Prior to
injection works tests of the hydraulic connectivity
between adjacent boreholes checked the connections
of void areas.
3. Installation of sleeve pipes.
4. 1st Injection Step, i.e. filling injection. The
grout was injected at low pressure, up to 2 bar ex-
cess pressure to fill voids and fissures. Later on the
pressure in this step was increased to 6 bar.
5 . 2nd Injection Step, i.e. control of filling injec-
tion. The 1st filling injection was controlled by an
additional injection step at an excess pressure of up
to 10 bar. The 2nd injection step was repeated in
most of the areas pending intermediate NDT-test
results.
3 3
In total, almost 15 m of injection grout, 10.1 m
in the filling injection step 4 and 4.8 m’ in the repe-
tition injection steps 5, were introduced under the
baseslab of the West Pylon Caisson.
4 VERIFICATION PHASE
4.1 Core drillings
Verification drillings were performed after comple-
tion of the injection works. The main purpose of
these drillings was verification of the injection
works by endoscopy. However core drilling using a
geotechnical drilling technique was executed to
achieve a good core recovery on the verification
drillings. For verification purposes in total 11 drill-
ings were performed. The location of the drillings
were selected considering the following criteria:
0 Size of the voids being filled by the injection
works.
8 NDT-test results prior and post injection
works.
149
0 Coverage of the area to be repaired.
0 Structural requirements and foundation de-
mands regarding support.
The core drilling results were evaluated espe-
cially with regards to:
0 Loss of core (comparison of core recovery
with drilling depth).
0 Evidence of injection grout.
0 Filling of the voids under the Pylon Caisson
baseslab with injection grout.
0 Bond of injection grout or existing underbase
grout to the baseslab.
4.2 Endoscopy Iizvestigatiom
The boreholes of the verification drillings were en-
doscoped and video recorded subsequent to the core
drilling. The following subjects were especially con-
sidered when endoscoping the verification drillings
in addition to the main criteria established in the
investigation phase:
0 Bond between the bottom slab and injection
grout.
0 Bond between injection grout and existing un-
derbase grout.
0 Any evidence of filling of voids, cracks and
fissure by injection grout.
4.3 Non Destructive Testing
Non-destructive tests were carried out continuously
during the injection works. The measurements were
used, first to control the grouting works as an active
design, second, to verify the filling injection works
after completion. The tests were performed in sev-
eral steps as the grouting works were executed in
nine separate zones. The number of measuring posi-
tions taken in the zones subsequently to the comnple-
tion of the repair works were 595 positions covering
all areas where injection works were performed.
The results of the measurements were classified
according to the investigation phase in five main
groups A-E and eight subgroups where A represents
full contact and E represents no contact.
The voids under the bottom slab were filled by
injection grout during the repair works. The impact
echo method was used to verify the foundation con-
ditions of the Pylon caisson during and after com-
pletion of the repair works.
Examples of improvements after the repair works
are presented in Figures 5a-c. The result was inter
preted as (E)D curve evaluated prior to repair works
improving to a (C)D after the first treatment result-
ing into a B after the repeated treatment.
The measurements were performed when the ini-
tial part of the curing process in a treated area was
completed. The interpretations of the curves were
compared to the previous measurements taken prior
to the repair works at the same location.
Figure 5a. Velocity spectrum 9708 14. Evaluated as
(E)D. i.e. area with no/little contact.
Fig 5b Velocity spectrum 980209. Evaluated as
(CID, i.e. area with no/little contact.
Figure 5c. Velocity spectrum 980218. Evaluated as
B, i.e. area with good contact.
Figure 6. Definition of pattern.
The NDT-measurements were presented on
drawings both prior to injection work and after. The
drawing prior to grouting works is presented in Fig-
150
Figure 7. Conditions according to NDT prior to
grouting works.
Figure 8. Conditions according to NDT after grout-
ing works.
ure 7 and the drawing after completion of the grout-
ing works is presented in Figure 8. Definitions of the
patterns in the drawings are to be found in Figure 6.
In order to verify the NDT-measurements the
results were compared with the observations ob-
tained by the verification drillings, core descriptions
and endoscope inspections of the verification bore-
holes. Verification dr1bgS Were performed in
eleven Positions.
4.4 Correlation Nondestructive Evaluation of Concrete and
Masonry, Chapter 15, Figure I06. Ithaca, New
The comparison of NDT, verification drilling and York: Bullbrier press.
endoscope investigations demonstrated a clear COT- Steinfeld, B et al. 1997-98: J.’ariolls~~o,~-ofi,-ially
relatioii in six of the eleven positions. In the other published texts. Sundlink Contractors, Malrno.
five positions, a general improvement was noticed.
However, due to lack of connections from injection
holes into very enclosed voids a total filling was not
achieved. Because of the amount of injection holes
especially in these areas these local gaps/ cracks
must be very limited. Otherwise the NDT-results
would not indicate such a clear general improve-
ment. As described in Section 2.4.1 above an area
and not only the measuring location influences the
impulse response. The improvement was also veri-
fied by the placed grout quantities. Therefore the
Impact-Echo method was decided by the project to
be a reliable instrument for the quality assurance of
the repair works of the West Pylon Caisson.
By comparison of NDT test results prior to and
after the injection works a major improvement in
contact to the baseslab was demonstrated all over the
foundation area.

5 CONCLUSIONS

The repair works were successfully executed. The

voids and gaps which remained under the Pylon
Caisson baseslab from the underbase grouting op-
eration were filled providing an adequate support of
the Pylon base.
For independent verification of the repair works
core drilling, endoscopy and NDT-tests were used.
Good correspondence between the NDT-test results
and the other verification methods was demon-
strated.
The NDT-test results demonstrated that no large
areas of little or varying contact remain. The existing
voids and gaps under the baseslab were filled inde-
pendently of their width and the resulting total
baseslab contact was considered to be 80% or big-
ger.
In general the verification drillings and endoscopy
performed demonstrated a successful filling of the
voids below the bottom slab of the Pylon Caisson
and verified a good contact and support of the Pylon
Caisson base. Injection grout was placed in the areas
of bigger voids and gaps and in small fissures also.
The repaired foundation fulfills the design re-
quirements with regard to the transfer of the gov-
erning design loads into the limestone.

REFERENCES

Pile Dynamic, Inc May 1997. P. I. T C’ollector U.sers

Munzrul2.70 Cleveland, Ohio.
Sansalone, M.J. & Streett, W. B. 1997. Impact-echo.

151
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

A study for the shape of pile with reflect-wave method

Yang wu
Fujian Academy of Building Research, Fuzhou, People’s Republic of China

ABSTRACT: In this paper, based on stress-wave theory and self-adaptive strain-wave theory, the basal rela-
tionship between unit velocity-pulse responds of the pile top and the shape of pile (wave impedance) is devel-
oped. The definientia of coefficient for pile integrity P and coefficient for overall reflection of pile top R and
its expression are proposed, the principle for a method of identifying shape of pile is systematically described.
The change of pile’s cross-section (wave impedance) and the shape of pile can be evaluated by this method.

1 INSTRUCTIONS 2 COEFFICIENT FOR PILE INTEGRITY P i

The technique of low strain pile integrity testing It is assumed that the length of pile is L, if the
is applied extensively during recent years, especially pile is partitioned to N units, then the length of every
in quality testing, the reflect-wave method is one of unit is A L=L/N. Assume that wave impedance of the
the method that is adopted widely at present. The re- ith unit is Zi, based on the boundary-conditions of
flect-wave method is a principle that bases on some displacement, velocity and force on interface, and
character to assess the pile integrity, such as the based on the relationship among reflect-wave,
shape of incidence-wave and reflect-wave of veloc- transmission-wave and incidence-wave, then the co-
ity echo wave, phase and the arriving time of wave. efficient for reflectance of interface between the ith
However, nowadays this technique can only used to unit and the (i+l)th unit can be obtained as follows:
determine the nature of pile quality, such as necking, r.I = ______
Z;- ‘;+I
increase of cross section or full crack of pile shaft, (i=o, 1,. .....,NI (2.1)
but can not used to analyzed quantitatively the defect
z;+ zi+*
Here, assume Zo=O, then ro= -1. It can be obtained
of pile quality. The tester can estimate but by experi-
from Eq. 2-1 as follows:
ence whether the pile can be used as engineering
pile. So, at present, the problem that need eagerly re-
solve of the pile integrity testing is how to quantita-
tively identify the pile integrity, quantitatively de-
scribe the shape of pile, quantitatively identify the
degree of defect of pile shaft.
In this paper, based on stress-wave theory, the
basal relationship between the unit velocity-pulse re-
spond of pile top and the shape of pile (wave imped-
ance) is developed. The definientia of coefficient for
pile integrity P and coefficient for overall reflection
of pile top R and its expression are proposed. A
method of identifying shape of pile is systematically P i is named the coefficient for pile integrity on
described. The change of pile’s cross section (wave interface, it can be obtained from definiens:
impedance) and the shape of pile can be evaluated.
The problem of pile integrity testing with quantita-
tively calculating to identify is resolved. In fact, it can be found that P is the synthetic reflec-

153
tion of A, E and C at the defected interface of pile. If In Eq.3-5 and Eq.3-6, defined as:
E and C are not change, then R’ (1,O) = 1 R(i,j) = R’ (ij) = 0
6 . = & (i=1,2,......,N) (2-7) When i<O orj<O
ni i=j=O
Namely, P i shows the change of pile across sec- The transmission equation can be obtained by
tion for defect. -
Eq.3-5 Eq.3-6:
R(1) = r,
3 ASCERTAIN THE COEFFICIENT FOR OVER- R(k) = rk (I -ri2)+ r,R(k-m) n
m-1
(1-r?>
ALL REFLECTANCE OF PILE TOP R(k) i=l ?/?=I i= 1
For instantaneous vibration, it is assumed that the (k =2,3 ,* * * * * * ,N) (3.7)
time of first reflection which run to toe of pile and
back to top of pile is N T for vibration. It is apparent Where f j (I -ri2)= 1.
i=l
that the time interval y is the time that wave propa-
gate come-and-go one time in each unit, the expres-
sion of unit velocity -pulse respond of pile can be
4 THE METHOD OF I ~ E N T ~ F YSHAPE ~ G OF
given as follows:
PILE
Vd(t) = 6 (t) + R(i) 6 (t-i T ) (3.1) The Eq. 3.1 can be discretelized, then:
/=I
A’
where 6 (t) is the unit pulse vibration, defined as Vd(i z ) = 6 (i ) + C R(k) 6 (i z -k T )
1-1
1 t=O
(i=O, I ,* * ,N)
0..
(4.1)
6 (t) = (3.2)
The equation can be shortened as:
0 tfO N
R(i) (i=1,2,.**.*-,N)is the overall coefficient for Vd(i) = 6 (i) + R(k) 6 (i-k) (4.2)
k=I
reflection that arrives pile top for the time t= i T . The Eq.3-2 joins into Eq.4-2, then
The key of estimating shape of pile is to ascertain
R(i) (i=1,2,*..**.,N).It is assumed that R’ (ij) is re-
gard as overall coefficient for incidence-wave of (4.3)
..............................
downward (ij) location at the same time and R(ij) is
regard as overall coefficient for incidence-wave of so
upward (ij) location at the same time, shown as
R(1) =Vd(I) / 6 (0)
Fig.1, Fig.2 and Fig.3.
For i=O, namely on pile top, the overall coeffi- R(2) = v & ? ) / 6 (0) (4.4)
..............................
cient for reflectance R(i) is as follows:
R(1) = R(0,I) = R(1,O) ~ R(N) = VdW) / 6 (0)

i
If the velocity echo of pile top V(i) is already
R(2)=R(0,2)=R( 1,l) (3.3) known, commonly, on the I& of beginning seg-
..............................
R(N) = R(0,N) = R( 1,N- 1) ment, namely the segment of dashed frame in Fig.4,
R ’ (0,l) = R(1) the unit pulse echo Vd(i) can be obtained by adaptive
strain-wave arithmetic(here omit).
R ’ (0,2)=R(2) (3.4) The interfacial coefficient for reflectance, namely:
..............................
R’ (0,N) = R(N)
~

For i=1,2,..*-.*,N, it can be obtained the follows (4.5)

by using direction of arrow adding:
R(i,O) = r,R’ (i,O)
1 ..............................
R(i,l) = r,R’ (i,l)+(l-rJ R(i+1,0)
R(i,2) = r,R’ (i,2)+(1-r,) R(i+l,l) (3.5)
..............................
~

R(i,N-i) = rlR’ (i,N-i)+( 1-rJ R(i+l ,N-i+l) The shape of pile can be evaluated by interfacial

‘
R’ (i,O) = (I+r,-l)R’ (i-1,O) coefficient for reflectance. The impedance ratio B i
R’ (i,l) = (l+rl-I)R’ (i-l,O)-rl..lR(i,O) can be calculated by Eq.2-5, then the wave imped-
~
R’ (i,2) = (I+r,-l)R’ (i-l,2)-ri-lR(i,I) (3.6) ance of pile can be obtained:
.............................. Zi+l = B iZi (i=1,2,***.-.,N) (4.6)
R’ (i,N-i) = (l+r,-l)R’ (i-l,N-i)+(l-r,)R(i+l,N-i+l) The change of wave impedance reflects the in-

154
Fig. 1 the diagram of entering wave and leaving wave
for a certain section Fig.4 the velocity-respond V(t) and incidence-wave

5 CONCLUSION
If the velocity respond of pile top V(i)
(i=O, 1, ..-*.* ,N), is already known, then pulse inci-
dence-wave Ie(i) can be also distilled; the unit pulse
respond Vd(i) (i=O,l ,..*.**,N)can be calculated by
Self-adaptive strain-wave arithmetic; the interfacial
coefficient for reflectance rk(k=1,2,.. .,N) can be cal-
culated by V,(i) taking into Eq.4.5; then the wave
impedance at each interface also be calculated by
Eq.2.5, the shape of pile is ascertained by Eq.4.6 or
Eq.4.7.

Fig2 the reflection and transmission diagram of

wave propagating in pile shaft

Fig.3 the decomposing diagram of reflection and

transmission of wave propagating

tegrity of pile. If both E and C are not change, then

Aj+l P iAi (i=1,2,.--.-.,N) (4.7)

155
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 1503

Evaluation of the performance of an existing foundation via PIT tests

S. B. Foi, J. H. E Pereira, R. RCunha & J.Camapum de Carvalho

Depurtment of Civil Engineering, University of Brasilia, Bruzil

ABSTRACT: This paper presents a simple case history in which the Pile Integrity Test (PIT) was used to
evaluate the integrity of preexisting bored piles founded in the porous clay of the Brazilian Central Plateau.
This study, accompanied by a geotechnical investigation program (auger holes and SPT tests) was further
adopted to evaluate the bearing capacity of deep foundations which were under suspicion of structural damage.
The study aims to clarify the present conditions of the INEP building foundations at the University of Brasilia,
in the Federal District, Brazil. It was observed that the PIT test program furnished an excellent prediction of
the piles length, concrete quality and integrity. The PIT apparatus has demonstrated to be a tool of fast execu-
tion and low cost, being important to the effective cost for the historical case studied herein.

1 INTRODUCTION concrete which are measured with the sonic test are:
8 Dynamic modulus of elasticity (using measures

One of the most important concerns of the
Engineering is the quality control, and the attendance
to the specification of structural elements in service.
At the present time, it is observed that both the safety
and reliability of any geothecnical work are strongly
influenced by the executive process. In this context,
of resonance frequency - normalized by ASTM
C215/60 - and with the determination of the speed of
propagation of sonic and ultra-sonic pulses in the
concrete, not yet normalized);
Integrity of the structural element;
0 Maintenance of the compressive resistance.

it is also advocated Mello (1981) an isolated factor
of safety for the execution of the engineering work.
The field tests in structural elements can, in a
general way, be divided into destructive and non
destructive tests. The destructive test demands the
molding of special test samples, which don’t
necessarily supply direct information of the concrete
from the real structure. In this case, it becomes more
interesting to use extracted samples of the structure
in service. However, this procedure can damage the
structure, even if in a localized manner. To avoid this
problem, non destructive tests have been developed
to be used directly onto the structure. They are the
most widely used methods in the engineering and one
of their great advantages is the accomplishment of
multiple tests in the same place. These methods can
also be performed with time, in order to evaluate the
variation of the characteristics of the structure during
its useh1 life.
One of the most used types of non-destructive test
is the sonic (or ultra-sonic) test. They have been
continuously and increasingly used in the last years.
They supply the knowledge of the concrete fissures
plus the material’s deformation.
Some of the internal characteristics of the
The Pulse Echo Method or Pile Integrity Test
(PIT) is one of the tests most widely used for the
post installation inspection of the integrity of
concrete piles, drilled shafts, augered piles (CFA’s)
and, eventually, timber piles. This test is also known
by its easy execution, with great reliability allied to a
low cost. It appears, then, to be a potential tool to
the effective cost of quality control of deep founda-
tions.
The PIT test has been commonly used in deep
foundations to evaluate the integrity of piles with
suspicion of constructive defects, as fissures, cracks
or variation of the shaft cross sectional area. The PIT
might also be used to evaluate the pile length in cases
were the wave propagation speed can be assume
This speed is a function of the density and the quality
of the concrete, hence, it can be correlated with the
characteristic resistance of the pile. It can also be
monitored with time, thus allowing the determination
of the quality of the foundation along its normal life.
The PIT test assumes that the stress (or acoustic)
wave propagates in one dimension. This will require
a foundation with a minimum ratio in term of
1engtMdiameter about 5 (Pile Dynamics, 1998) to
have the interpretation model physically “fit”.

157
Moreover, in piles with a large diameter in relation Table 1 Geotechinical Darameters of the porous clay (Salles.
to their length, it may be difficult to pin point any
eventual defect close to their heads. To locate even- Parameter Unity Range of Values
tual defects in this position, several PIT tests at dif- Moisture content YO 30-34
ferent locations around the pile top would be re- Degree of saturation YO 50-86
quired. Void ratio -- 0.9-2.0
Liquid limit % 25-78
This paper, therefore, describes the use of the PIT Plasticity limit YO 21-35
test as a post-construction testing alternative for Plasticity index YO 4-43
quality control. The PIT was used to determine the Effective cohesion KPa 11-23
pile length of existing deep foundations of the INEP Friction angle Degrees 28-32
building - in the campus of the University of Brasilia Coefficient of Collapse % 0-12
(UnB), and their integrity. These foundations were Coefficient of earth -- 0.3-0.6
Pressure at rest
installed in 1994, together with the floor plain of this Coefficient of
same building. The construction of this building was cm/s
Permeability
stopped at that time, and has recently started (1999)
demanding the study of the existing foundations,
since their length have not been recorded in the Table 2. Summary of soil conditions.
original blueprints, and especially because the new Thickness SPT -N
architecture of the building was changed-thus im- Lqer Classification
(ni) (bl/0,3 m)
posing new live and dead loads.
01 very soft red silt 4 to 5 2 to 10
2 PROBLEM DESCRIPTION
red sandy clay
2.1 Geotechnical site characterization 02 with gravel 1to2 7to8
The Brazilian capital city is located in the Central
Plateau of Brazil, within the “Federal District”. This sandy silt 3 to5 22 to59
district has more than 86% of its area covered by , O3 compacts to stiff
latosoils and lateritic soils, commonly known by
“tropical” soils. The upper layer has a variable
thickness (in the present case of 5 m) and displays a
low mechanical resistance with a reduced SPT blow
count (from 2 increasing to around 15)
Due to geological factors and physical-chemical
processes which have taken place during the
formation of this particular soil, it is also collapsible
under saturation. Then, it becomes problematic the
spread footing use in the Federal District, and, due to
this, deep foundations are the predominant type of
support for the buildings in this region. More details
of the upper Brasilia “porous” clay can be also be
found in Cunha & Perez (1998) and in Sales et a1 ,
1999.
Table 1 below displays some of the main
geotechnical characteristics of this clay, and Table 2 Figure 1. T!-pical profile
presents the geotechnical characteristics found in two
SPT loggings at the site. Figure 1 presents the cross The building columns are made of wood (log
sectional profile found in the two aforementioned trees) transfer the loads from the roof and the
SPT loggings at this same site, displaying the N blow basement to the foundations. This type of structure
counts per 30 cm. solely allows the transfer of axial load from the su-
per structure to the foundation (hence the soil).
Thus the main concern of the designers was the
2 2 INEP Buildiig Strzrctiu-e (new) axial loads on top of the preexisting founda-
tions.
The INEP building has an area of approximately The foundation of each “wing” was constituted of
2686 m2 and it is constituted of a central module bored piles with 0.35 m in diameter, linked su-
linked to four outlying “wings”, thus forming an perficially by capped blocks. These blocks were de-
octagon. This building is composed of one basement fined with two and four piles, as depicted in Figure 2.
and a single floor.

158
2.3 PIT tests
The testing program consisted of PIT tests in order
to determine the length and the integrity of some of
the piles. Due to time restrictions only two piles
foundations were tested. This was done because no
(post-construction) information could be obtained
from the original blueprints.
Two blocks in particular were selected, due to its
location and load capacity within the complex
Blocks PI6 and P41 (related to pillars 16 and 41) are
those studied herein, and can be easily found in
Figure 2.
3 TEST METHODOLOGY
The PIT test is based on the acoustic wave
propagation at a low deformation level. These
waves are measured by using a PEM (Pulse-Echo
Method) equipment. The waves are longitudinal
and they induce displacements in the same
direction of their propagation. They generate
dilatation and compression effects in the material,
when passing to all sort of material types (solids,
liquids or gaseous). Their propagation speed is
proportional to the square root of the Young’s
159
Modulus and inversely proportional to the square
root of the pile’s material density. This is due to the
fact that the elastic parameters supply the restoring
forces of the displaced regions, and the density is
related to the inertia of the system.
The test consists of measuring the response of a
low strain impact caused by a small hand-held
hammer. Each blow generates a compression wave
that spreads from the pile top in direction to its
bottom. This wave can be reflected during its
journey at each cross sectional variation of the
shaft, thus generating second order waves
propagating with the opposite direction.
The traces of the acoustic wave signals are
captured by an accelerometer, which is attached to
the pile head. This signal is hrther converted into
an electric signal, which is amplified and sent to a
conditioning center where is transformed from the
analog to the digital format. All this information is
stored into a data acquisition system which is built
in the PIT unit system. Once this stage is ac-
complished, the data can be transferred to a PC for
hrther processing of the signal (interpretation
stage).
The verification of the pile integrity is made
through the interpretation of the shape of the
obtained (reflected) acoustic wave recorded at the
pile head From the tests, it is possible to obtain
graphs of velocity and force versus time (which can
be later translated into pile length) or frequency
Due to the aforementioned considerations, any
cross section variation of the pile shaft, or variation
of its structural density, will cause a change in the
shape of the velocity/force measured curve Those
changes, and the curve itself, allow for the
establishment of conclusions regarding the condi-
tions of the pile, and the location of any eventual
structural damage, bulging, etc along its shaft
3 1 Pi.eparatron of the pile top
One of the most important steps for any PIT test is
the preparation of the pile top surface It is well
known that the accelerometer (and the hammer
impact blow) should be placed onto a horizontal
and free of dirt surface Hence, the accelerometer
can be attached only after a horizontal and
reasonably free of dirt top surface is prepared
Generally this attachment is made with the use of a
thin layer of a soft paste (vaseline, etc ), allowing a
good recording of the impact and post impact
wave An imperfect installation will undoubtedly
hamper the quality of the signal (which will have
low amplitude, noise, etc )
The generated acoustic wave will be of clear and
sharp magnitude solely when an intense impact, of
short duration, is produced at the pile head
Therefore, it shall be again emphasized that the
correct preparation of the “impact surface” is a
fundamental step for the success of this test
In order to carry out the tests at the INEP
building, it was necessary to excavate the capped
block down to its base (at the block/pile interface)
This was done in order to open a “window”, or
lateral niche, where the hammer blow could be
applied This small recess was flattened out by
using a thin layer of mortar, thus allowing the
The aforementioned grading criteria was also
Figure 3. Disposition of the lateral niche.
used with the piles tested herein, as presented next.
160
accelerometer to be fixed in the vertical position.
Figure 3 depicts, schematically, the position of the
lateral windows in each studied pile of the pile
group.
4 CRITERIA FOR SIGNAL EVALUATION
The PIT test should be used in precast concrete
piles, which are driven into place with the use of
standard crawling cranes. This is so because this
type of pile has, in general, an almost perfect cross
sectional geometry, thus reducing the effect of the
reflected “secondary” acoustic waves. Never-
theless, more recently, some authors started using
the PIT in the evaluation of cast in place piles
(Arias, 1977, Rausche & Seitz, 1983, Berger &
Cotton, 1990, Cunha & Costa, 1998 and others).
With this objective, Cunha & Costa, 1998 have
proposed a simple alteration to the evaluation crite-
ria (grading letters) presented by Rausche et al.
(1 988) with the PIT acoustic wave results. Cunha
& Costa (1998) modified this grading criteria to in-
clude an “extra” letter C’, as follows, to be specifi-
cally used in the case of cast in place piles:
A- No significant defect is apparent in the
records. It shows a perfect reflection in the
pile toe, and no other significant reflection
prior to this point. The pile is considered as
physically sound;
B- No significant defect is apparent in the
records, however the pile toe reflection was
not observed, thus indicating that the full pile
length wasn’t tested. The pile is considered
as partially tested;
C- The record showed significant impedance
changes, but the pile toe reflection was
clearly observed. The pile is considered of
limited value;
C’ - The record showed significant impedance
changes, but the pile toe reflection was
clearly observed. In the case of cast in place
piles, this may refer to cross sectional
changes, hence the pile can be considered as
physically sound, if it refer to “bulges” on the
pile’s shaft;
D- The record showed significant impedance
changes and the pile toe reflection was not
observed. The pile is considered as
questionable and additional tests or its
substitution is suggested.
5 ANALYSIS OF THE RESULTS has been extremely useh1 to evaluate the group ca-
pacity, by means of static method, of this and other
An analysis of the pile integrity can be made with blocks of the INET?building, given the new ways
the velocity signal, by using some of the criteria imposed by recent architectural modifications.
exposed in the previous item. It is known that It is finally observed that the piles have part of
velocity tracing signal is positive (above the axis) their load sustained by lateral friction in the con-
during the time of the hammer impact. They should crete/soil interface. This lateral friction reduces and
also reduce to nil at the end of the blow. absorbs somehow the energy of the acoustic wave,
In relation to the reflection of the wave at the thus turning negative (totally below the axis) the fi-
pile toe, it is shown that it will change in sign if the nal stage of the acoustic wave. This part of the
material of the pile is less stiff than the surrounding signal is the one related to the reflection of the
soil. Hence, in this case, velocity signal becomes wave in the pile toe, as can be seen through Fig-
negative. Moreover, it is also known that secondary ures 4 and 5.
wave reflections are intrinsically related to changes
in the impedance of the shaft (cross sectional or
density changes). The impedance is basically de-
fined as the result of the multiplication of the cross
sectional area by the elasticity modulus divided by
the wave velocity. Hence, any reduction in the
elasticity modulus or the cross sectional area will
result in a “tension” reflective wave (positive sig-
nal), whereas any increase of above variables will
result in a “compressive” reflective wave (negative
signal).
With this theoretical “background” in mind, it
was possible to conduct the interpretation of the
results of the PIT tests carried out at the INEP
foundations. Hence, Figures 4 and 5 present the
results of the two casts in place augered piles
studied herein. It is plotted the acoustic wave Figwre 4 Pile PIG - speed wave sign.
velocity signal versus the pile length (actually time.
but by assuming the propagation velocity it is
possible to convert time into pile Iength). These
tests adopted the value of 4000 m/s as the
propagation speed of the acoustic wave through the
concrete piles. This was done due to previous
experience (Cunha & Costa, 1998) with this
equipment under similar (pile) conditions.
The curves of Figures4 and 5 do present the
positiveinegative signal variations explained before,
thus indicating that either cross sectional or
material density changes exist along the shaft of
these piles. These variations occur between the
initial hammer impact and the thin wave reflection
i n glength of the pile). In this particular
~ i n d ~ ~ a tthe
case, however, it is believed that such impedance F i p r e 5. Pile P4 1 - speed wave sign.
changes are related to cross sectional changes of
the piles, caused by the pile installation method.
The excavation of the soil was done mechanically 6 CONCLUSIONS
with a standard auger from a local company, which
leads to slightly imperfect (with a rough surface) Low strain PIT tests are simple and fast to perform.
boreholes. It is further observed in Figure 5 that They can be used in new or preexisting
the cross section of the shaft has increased some- foundations, and do solely require the use of the a
how after the initial few decimeters of the piIe. simple and portable equipment in the field.
In both cases, however, the piles have been In the case reported herein, this technique has
graded with the letter C‘ i.e., they are physically proven to be extremely important to the effective
sound besides the cross sectional changes. They cost to evaluate the structural and geometric (un-
also have a length of around 8 m according to these known) characteristics of preexisting cast in place
few PIT records. This (previously unknown) length piles. This evaluation was very important to access

161
the bearing capacity, by means of static method, of
the existing pile groups, which will be subjected to
new superstructure loads
In the foundations tested herein, the PIT tests
detected sound piles with a reasonably good quality
concrete. The piles had a length of around 8 m,
and a shaR of primarily bulges ( in a random
manner). These results, although few for the site,
were extrapolated to all other existing piles, given
the fact that fbrther PIT tests could not be per-
formed (due to time constraints).
The very simple example presented in this paper,
therefore, is another indication that the PIT has a
large potential for usage into all sorts of deep
foundation problems (underpinning, structural
changes, and others), which are commonly found in
the Federal District. This is so due to its fast,
reliable and effective cost characteristics.
Rausche, F. e Seitz . J. 1983. bitegrip testing of sliaff.s arid
caissons. Specialv Session on Shafts and Caissons.
ASCE Annual Convention,
Rausche, F.. Likins, G.E. e Hussein, M. 1988. Pile Ziitegriry
b,v loiv a i d high straiii impacts. Seminar on the Applica-
tion of Stress Wave Theory on Piles.
Rausche, F.. Beim, J.. Beim, G. e Morgano M. C.. 1999.
Custom tailoring qualily assurance ,for augercast piles.
XI Panamerican Conference on Soil Mechanics and
Geotechnical Engineering . Volume 3 Brazil. p. 1466-
1370.
Sales. M.M.. Cunha, R.P. & Jardim. N.A. 1999. Analysis of
Piled Footing Tests 011 a Tropical Porous Clgv. XI
Panamerican Conference on Soil Mechanics and Geo-
technical Engineering. Volume 3 Brazil. p. 1395-1402.

ACKNOWLEDGMENTS

The authors would like to express their gratitude

to the local engineering company EMRRE, for all
the field support in the INEP building. They also
acknowledge the financial support given by the
local institution FAD-DF for the acquisition of the
PIT, and the laboratorial support given by the
Geotechnical laboratory of the University of Brasi-
lia.
The first author, in particular, thanks the CAPES
for the scholarship she was granted to pursue her
MSc studies in the Geotechnical Post Graduation
Program of the University of Brasilia.

REFERENCES

Araki, M.S. 1997. Aspects related to the proprieties of the

collapsible soils of the Federal District. M.Sc. Thesis.
Department of Civil Engineering. University of Brasilia.
(In Portuguese).
Arias, R.P. 1977. Evaluation of drilled sliaji integrity b,v non
destructive methods. M.Sc. Thesis. University of Austin.
Texas.
Berger, J.A. e Cotton, D.M. 1990. Low strain integrity test-
ing of deep foundation. Procedure. Deep Foundation In-
stitute Annual Meeting, Seattle
Cunha, R. P. e Costa, F. L. 1998. Estimation ofphysics in-
tegrity of piles assented on a porous cla-v in Brasilia us-
ing PIT tests. XI Brazilian Congress on Soil Mechanics
and Geotechnical Engineering . Volume 3 - Brasilia -
Brazil. p. 1647-1654 (In Portuguese).
Mello, V.F.B. 1981. Proposed Bases for Collating Experi-
ences for Urban Tuiineling Design. Symposium on Tun-
neling and Deep Excavations in Soils, S2o Paulo, p. 197-
235.
Globe Rausche Liluns and Associates,1987. Catalogue of
Computed Pile Top Velocitiespom Travelling Wave Re-
jections Noiiuniform Piles Cleveland, Ohio.
Pile Dynamics, Inc. 1998. PIT User's Manual. Cleveland,
Ohio.

162
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Sonic integrity test of piles-integrity effected by basement excavation

in Bangkok soft clay
N.Thasnanipan, A.W. Maung, PTanseng & Z.Z.Aye
Seufco Company Limited, Bangkok, ~ ~ u i ~ a ~ ~ d

ABSTRACT: Modern buildings in Bangkok nowadays are designed with basement facility and different ele-
vation of pile foundation which require excavation works in the immediate vicinity of the constructed bored
piles. Adverse effects on piles from adjacent excavation works in terms of excessive movement of piles and
pile damages have been observed in sonie projects. Assessment of bored pile damages using sonic integrity
test is focused in this paper with demonstration of test results from three case histories. Method of remedial
works for damaged piles and additional testing for pile capacity justification where available are also briefly
discussed. Stress induced in piles due to the soil movement caused by excavation works is also examined
with the aid of computerized finite element program Location of cracks detected by sonic integrity tests is
correlated with the analyzed bending moment in pile, which exceeds the cracking moment capacity of pile.

1 INTRODUCTION 2 GENERAL DESCRIPTION OF BANGKOK

Due to the prevailing condit~oiiof subsoil (existence
of considerably weak stratum at top part) piles are
conimonly used for foundation of buildings and
structures in Bangkok. Deep-seated large diameter
bored piles of over 35 m depth with 0.8 -1.8n1 in di-
ameter have been maiiily used for heavy structures
such as high rise buildings, elevated expressways,
flyovers or overpass bridges and more recently for
uiiderground subway stations of the first Bangkok
MRTA system. As a part of pre-planned quality as-
surance regime and retrospective iiivestigatioi~,in-
tegrity of the piles are tested to obtain the informa-
tion with regard to the potential deficiencies of the
constructed piles which may have formed during
actual pile construction process or may have been
attributed by other activities after construction of the
piles. Designs of the modern buildings in Bangkok
frequently call for the basement facility and for the
different elevation of pile foundation, which require
excavation work mainly in top soft clay fayer. In
some projects, adverse effect from excavation works
adjacent to constructed piles has been observed.
This paper presents the assessment of bored pile
damages which were induced by adjacent excavation
SUBSOIL
Bangkok subsoil consisted of alternating clay and
sand layers of Quaternary deposits extending down
to about 550m depth where bedrock generally exists
(Balasubramanium, 199 1). Made-Ground consists
predominantly of Fill-Materials, Clayey Sand or
Silty Clay with some cement block rubble and rock
fragments, is coinmoiily found up to 4m depth. Soft
to very soft, highly compressible dark gray marine
clay lies beneath Made-Ground and in some areas it
lies under weathered crust layers of 2m thick. De-
pending on the location, this layer is extended up to
12-18m. About 2m thick Medium Clay layer can be
observed between Soft Clay and underlying Stiff
Clay. Generally Stiff Clay layer occurs directly un-
derneath Medium Clay and its depth goes up to 22m.
Below Stiff Clay layer, First Sand layer 5-8m in
thickness can be found. This First Sand layer,
however, is absent in some areas. Stiff to Hard Clay
layer underlies First Sand and it is found to be about
51-13thick. Second Sand layer generally occurs at
depths between 45 to 65111. The Generalized Fence
Diagram of Subsoil covering the centraj. area of
Bangkok is illustrated in Figure 1 ~

using sonic integrity testing method.

163

Figure I . Generalized fence diagram of Bangkok subsoil
3 BORED PILE CONSTRUCTION IN
BANGKOK
In Bangkok, bored piles are constructed mainly by
two methods, dry and wet process. Dry process is
applied conimonly for small piles of diameter 0.35
to 0.6Om with relatively shorter lengths. In dry pro-
cess, a pile is bored to the required depth within iin-
permeable layers (usually up to 25111) by either ro-
tary or percussion method whilst temporary casing
(about 15m in length) is installed in soft clay layer
for protection of soil caving. After installing rein-
forcement cage, concrete is poured into the borehole
under dry condition. Tripod method, sliell with drop
weight is also cornmonly used in dry process.
Wet process method as its name implies makes
the pile under wet condition by using drilling slurry
(bentonite or polymer). Temporary casing of 14-
15m in length is also used as a support in soft clay
layer and soil inside the casing is normally exca-
vated by auger applying rotary drilling method.
Drilling is continued with the bucket under the
slurry from the top of sand layer to the final depth.
Treiiiie concreting is necessary for pile installed un-
der wet process. Bored piles constructed by wet
process are generally of large diameter (0.8m 10
1.8111 diameter) deep-seated (30m to over 60m) and
are norinally founded in either first or second sand
layer. Bored piles having conipressive cylindrical
strength of concrete in a range of 240-280 ksc (24-
28 MPa) with reinforcement ranging 0.5% tol.2% of
pile sectional area, are cornnionly used in Bangkok.
4 SONIC INTEGRITY TESTING IN ~ A N G K O K
Sonic integrity test is widely used for integrity test-
ing of both driven and bored piles in Bangkok. It is
employed as a part of quality control and or as a ret-
rospective investigat~onwhen some problem be-
comes apparent. For a quality control regime, the
number required for integrity testing depends upon
164
the technical requirement of the particular projects
ranging from minimum 10% to maximum 100% of
total constructed bored piles. A sonic integrity tester
with built-in computer having high quality signal
acquisition system and computer programs devel-
oped by TNO, PDI, IFCO etc. are commonly used
for pile integrity testing.
5 OVERVIEW OF EXCAVATION INDUCED
PILE DAMAGE IN 3ANGKOK
In general, pile defects are caused at two stages,
during the pile construction process and after con-
struction of the piles (post-pile-construction).
Thasnanipan et al. (1 998b) reported that integrity
of 285 bored piles (3.3 ?40 of 8,689) were found to be
of doubtful quality according to an assessment on
the results of sonic integrity test with the additional
information obtained from the pile construction rec-
ords of bored piles in Bangkok subsoil. The results
of the findings are summarized below.
Table 1. Suniinary of pile assessment on 8,689 bored piles
(Thasnanipan et al. 1998b)
Type of Defect Causes %
Poor concrete at Cutoff level near the ground 0.1
pile top (0-3in) or inadequate overcast
Size reduction Insufficient casing length or I .o
soft clay layer variation in
thickness
Cracks/discon~inuit Excavation works for pile 2.2
ies trimming and construction
activities
As can be seen in the above table, higher percentage
of defects is caused by post-pile-coiistruction activi-
ties and more commonly by movement of sur-
rounding soils induced by excavation works in the
vicinity of the piles.
Modern buildings in Bangkok frequently require
the excavation works for basement facility and for
the use of different elevations of pile foundation.
Excavation works are mainly carried out by using
cut slope, sheet pile wall, diaphragm wall and secant
pile wall. Since most of the bored piles are designed
mainly to carry an axial load, lateral and tensile
force which may impose on pile during basement
excavation are sometimes not considered or over-
looked. The lateral displacement of ground caused
by excavation naturally induced an addit~oiiallateral
force and bending moment in the pile. Once the in-
duced bending moment exceeds the cracking
moment capacity, piles are subject to be cracked.
Unexpected tensile force acting upon the pile due to
an excessive heave or uplift force induced by exca-
vation can also cause cracking of pile. In Bangkok,
pile cracks caused by external forces are n o r m a ~ ~ y
found to be at the level where Soft and Medium to
Stiff Clay boundary is present. Figure 2 illustrates
the typical features of external forces, which cause a
crack in the piles. Pile damage associated with sheet
pife wall excavation is presented in this paper.
Figure 2. Typical features of external forces causing pile
cracks in Bangkok subsoil
6 APPLICATION OF SONIC INTEGRITY TEST
IN PILE-INTEGRITY CHECKING
Sonic integrity test is simple, rapid, efficient and
cost effective in examination of the pile integrity. It
is reasonably reliable for the integrity testing of the
pile especially for the upper part where potential of
damage is higher as far as defect caused by the ex-
ternal effects (lateral displacement of soft clay due
to adjacent excavation works, unexpected force from
improper pile head trimming) is concerned.
6.1 Busic yrinciyles of sonic integrify test
Sonic integrity testing also called as low-strain in-
tegrity testing examines tlie response of a pile to a
light external impulse force. Either time-domain or
fiequeiicy-domain can be used in analysis of meas-
tired data by sonic integrity test. In Bangkok, time-
domain method is mainly applied in interpretation of
test data.
Soiiic integrity test is undertaken by striking the
head of the pile with a light hand-held hammer and
recording the response of the pile to this impulse
blow by means of a sensor or accelerometer placed
in good contact with tlie pile head. Stress wave the-
ory is a basic principle of this low-strain integrity
testing. The hammer blow induces a compressive
stress-wave into the pile, which propagates axially
along the pile shaft aiid reflects back toward the pile
head at a change of impedance within the pife. In
the case where pile impedance variation occurs due
to chaiiges in pife cross sectional area or properties
of pilc materials or presence of discontinuities, part
or ~ l i o l eof down ward wave reflects at the imped-
ance variations and returns to the pile head before
the first reflection from the pile toe.
165
Basically, following formulae are used in analysis of
sonic integrity test.
L =c TI2 (1)
where, L is a length from pile head to reflected sur-
face, c is a velocity of stress wave and T is a travel
time of the wave for length L.
Froin this formula, length L, can be computed
from the travel time of reflected wave and the wave
velocity of concrete which may be estimated from
concrete strength of piles or from the reference pile
of known length.
z =EA/c
E =pc2
where, z is an impedance of pile, E is Young
Modulus ofpile material, A is cross-sectional area of
pile, c is the propagation velocity of the stress wave
and p is density of pile.
From Equations (3)and (3)
z =pcA (4)
As can be seen from Equation (4), the change in pile
impedance could be due to variation in pile cross-
sectional area or in concrete quality.
6.2 I n t e r p ~ e ~qf~sonic
t~~~ integrity
~i test
In general the typical pile features detectable by
sonic integrity test includes;
Pile size variation
Q
Pile toe reflection (for pile length verification)
0
Pile material variation
I)
0 Soil influence
B Discontinuity andior cracks
The typicaf soiiic integrity signals of above fea-
tures have been reviewed aiid presented by Thas-
iianipan et al. (1998a). As some features have a
similar pattern to each other, it is always important
to check and compare the test data of pile in ques-
tion with those of other piles within the same job site
to estabfish “site signature”. In some cases correla-
tions should be made with similar projects where
damaged piles were investigated with firmed evi-
dence (coring or excavation to detect damage level).
Interpretation should also be made with sound
knowledge of pile construction technique, subsoil
condition and other factors (probieni encountered
during and post pile coii~tructioii)which may influ-
ence the soiiic test signals. Pile construction records
are also useful in interpretation of piles with de-
tected anomaly. Conclusion should be made after
careful and thorough review of test results in con-
junction with other available information. Further
investigation (if applicable excavate down to level of
detected crack for visual inspection) may require be-
fore final conclusion is made.
As a main focus of this paper, the typical features
associated with the pile damage (discontinuity/
cracks) detected by sonic integrity test is discussed
in detail with the data obtained from three cases as
presented in the following sections.
6.3 Pile damage assessment
From the results of sonic integrity test, degree of
damage or level of crack may be assessed. The as-
sessment of crack generally involved - setting of
amplification in the same range for all piles and
comparison of input impulse, amplitude of anomaly
and intensity of reflection from the anomaly (i.e.
multiple reflection). Additional information obtained
fiom further investigation as described in above sec-
tion is also used in pile damage assessment.
7 CASE STUDIES
7.1 Case I - Sheet pile excavation supported by soil
berm
7.1.1 Subsoil condition
A 12111 thick soft clay layer occurs beneath fill mate-
rial and weathered crust of 1.Oto 2.5m thick. Uncon-
fined shear strength (Su) of soft clay layer ranges
lt/m’ at top and 2.5t/ni2 at bottom whereas the natu-
ral water content is about 6040%. Medium clay
layer of 1 to 2m in thickness, having shear strength
of2.5t/ni2 to 3.5t/1112 underlies soft clay layer. Be-
low medium clay, at the depth between 18 to 27m,
stiff clay with traces of fine sand having SPT N
value of 15 to 37 is observed. Thick layer of fine
sand interbedded with clayey sand occurs under-
neath stiff clay layer.
7.1.2 Pile foundation
Pile foundation of eight high rise buildings in this
project comprises 904 cast-in-place bored piles of
diameter 1.0m and 1.5m, founded at 59111 below
ground level. Pile tips are embedded in dense sand
layer and the entire length of piles are reinforced
with steel reinforcement of 0.72% of pile sectional
area for the top part and reduced gradually to 0.23%
for the bottom section. To support the underground
water tanks, 88 pre-cast concrete piles (I-30cm)
having 21m length were also driven down to stiff
clay layer.
7.1.3 Excavation method and observed failure
Excavation was initially carried out by using sheet
pile (Type FSP 111) with soil berm support or slope.
Before reaching to -8.lm maximum depth of exca-
vation. sheet pile failure and associate severe lateral
166
movement as well as settlement of ground including
tension cracks of more than 300mm in width were
observed at some locations. Minor damages were
also investigated at the adjacent properties. Tempo-
rary raking struts were immediately installed in re-
sponse to these excessive ground movements caused
by sheet pile failure. A typical scheme showing ex-
cavation works of Case I is presented in Figure 3.
Figure 3. A picture showing slope excavation with sheet pile
support - Case I
7.1.4 Sonic integrity test andpile damage assess-
ment
More than 50% of piles were tested with sonic integ-
rity test. Test results were thoroughly reviewed and
“site signature”, most common features of the ma-
jority of piles, is established. It is to be noted that
test signals from piles of less potential for damage
(least effect from excavation) are selected in estab-
lishment of “site signature” or reference signal of
good piles. Based on the “site signature” of good
piles, defect piles were identified and assessed. Ta-
ble 2 shows the summary of pile damage assessment
by sonic integrity test.
Table 2. Summary of pile damage assessment by sonic integ-
rity test - Case I
f’ilc No. of piles with defect Defcct depth
Dia. Imst Less p r o ,ni below ground
(~n) promincnt prominent crack level (in)
crack crack
I .o 4 8 7 6.0 - 13.8
1.2 10 2 6. I - 7.6
1.5 1 9.9
The typical sonic integrity test results of undamaged
pile and damaged pile classified as least prominent
crack, less prominent crack and prominent crack are
shown in Figure 4. Majority of piles with crack de-
tected were located in the periphery of the excava-
tion zone. Piles, particularly with prominent cracks,
were observed to be deviated from their original lo-
cation up to 6OOn1m towards excavation.
 7.2 Case 11- Pile damage caused by sheet pile
excavation with one level temporary support

7.2.1 Subsoil condition

Soft Clay layer extends to 15m below ground sur-
face with undrained shear strength increases from
lt/m2 at the top and 2t/m2 at the bottom. Medium
Clay layer of 2 to 3ni in thickness is found between
Soft Clay and the underlying Stiff Clay. Stiff Clay
layer occurs up to depth 42ni with the SPT ‘N’ val-
ues ranges between 14 and 40.

7.2.2 Pile ,foundation

402 bored piles of diameter l m and 1.5m, founded at
55m below ground, being embedded in dense sand
layer were used as foundation in this project. Rein-
forcement steel bars of 0.5% pile cross sectional area
were provided for the entire length of bored piles.
Figure 4. Seismic test results of good pile, pile with least
prominent, less prominent and prominent crack (top to bottom 7.2.3 Excavation method and observed failure
in order) - Case I Sheet pile wall of 14111 in length with 1 level bracing
was utilized for maximum 8ni excavation. Though
the design required for installation of struts at l m
7.1.5 Remedial measures and pile capacity just@ below ground level, the actual excavation was car-
cation ried out up to 3m depth without installing the tempo-
Bored piles with cracks were cored to the depth be- rary struts causing sheet piles deflected in large
low the crack and grouted with non-shrink cement. magnitude.
Sonic integrity test was conducted few days after
rectifying the pile. Sonic integrity signals of the de- 7.2.4 Sonic integrity tests and pile damage assess-
fected pile tested before and after the remedial work ment
is presented in Figure 5. After the rectification Sonic integrity test was performed for all piles after
works. high strain dynamic load test was performed trimming to the designed cutoff level. It is evident
on two piles with large horizontal deviation and sus- that piles with crack detected by sonic integrity test
pected prominent cracks. The dynamic test results were mostly located in the vicinity of the excavation
indicate that tested piles could be capable of carry- boundary.
ing the design load with factor of safety higher than
1.5.

Figure 6. Sonic integrity test results of good pile and pile with
prominent crack - Case I1

A summary Of the Of damaged piles and

Figure 5. Sonic integrity test results on defected pile before
and after remedial work - Case I their degree of damage assessed are presented in Ta-
ble 3.

167
Table 3. Suinmary of pile damage assessment by sonic integ-
rity test - Case I1
PllC No of piles with defect
Dla Least Less P r o in t iI e n t
(111) proininent prominent crack
crach crack
10 5 15-94
1 5 14
7.2.5 Remedial measures
Damaged piles were cored to the level below the de-
fected crack and grouting was performed.
7.3 Case 111- Pile damage caused by sheet pile
excavation with two level temporary support
7.3.1 Subsoil condition
Underneath the top 2m of weathered crust, Soft Clay
layer can be observed up to 15m below the ground
surface. Medium Clay layer underlies Soft Clay
with the thickness of 2m. Undrained shear strength
of Soft Clay ranges between lt/m2 and 3t/m2
whereas it is about 50t/m2 in Medium Clay. Beneath
Medium Clay, Stiff Clay layer is found up to 25m
where the boundary of Hard Clay is investigated.
Within Stiff Clay stratum softer zone of Medium
Clay is identified at the depth between 19 and 2 l m
characterized by reduction in SPT values which ap-
pear consistent with increases in iiioisture content.
7.3.2 Pile joundation
A total of 143 bored piles of 0.6m diameter were in-
stalled with pile tip at about 26m below the ground
to support a residential apartment. Dry process was
employed to construct the piles. Cut-off level of
bored piles ranges from 0.85m to 5.95m depth. En-
tire lengths of all piles are fully reinforced with steel
bars grade SD40 of 0.35% of pile sectional area.
7.32 Excavation method and observed failure
Temporary retaining wall using FSP I11 sheet pile
type of 14111 in length with two level bracing placed
at 1111 aiid 3.5m below ground level was employed
for 6n1 depth excavation. A part of pile layout
within excavation area showing sheet pile wall and
bracing system is illustrated in Figure 7. No major
failure in terms of large tension cracks and ground
settlement was observed at the surface adjacent to
the excavation though some excessive sheet pile de-
flection was visually inspected at few locations.
Hence, it was not expected by the visual inspection
at site that excavation work has effected the integrity
ofpiles.
Defect depth
below ground
level (m)
84-21 0
Figure 7. Layout plan showing part of bored piles, sheet pile
wall and bracing system - Case 111
7.3.4 Sonic integrity tests and pile damage assess-
ment
Though sonic integrity test was employed as a part
of quality control regime in the original plan, it was
eventually applied for a retrospective investigation
after encountering many piles with defect. Hence all
piles were undergone sonic integrity test. According
to the test results, 84% of piles located within the
excavation zone were identified to be with crack.
Sonic integrity test results of good pile, pile with
less prominent crack, and pile with prominent crack
are illustrated in Figure 8. As can be seen in Figure
8, for pile with prominent crack, a distinct aiid sharp
reflection is evident at depth about 5.5m below pile
top. Repetition of reflection at multiples of this
depth can also be observed for this pile. These fea-
tures are comparatively not distinct for the pile with
less prominent crack. An excessive basal heave, in-
sufficient reinforcement to resist the tensile stress
and relatively small pile size are considered to be the
main reasons causing the cracks in this case. Table
4 shows a summary of pile damage assessment by
sonic integrity test.
Figure S. Case I I I - Seismic test results of good pile, pile with
less prominent and prominent crack (top to bottom in order)
168
Table 4. Sunitnary of pile damage assessment by sonic integ- Stiffness of foundation piles per linear meter ( E . S
rity test - Case 111 per pile spacing in row) was used as a parameter for
plate elements, where E and I is Young Modulus and
Pi Ic No. of piles with defect Defect depth
moment of inertia of pile respectively. The soil
Less p i below ground
Dia. Least between piles in a row is ignored. Mohr-Coulomb
(111) prominent prominent crack level (in)
constitutive model is adopted using undrained soil
crack crack parameters derived from soil investigation at site and
0.60 38 33 5.5 - 13.8 associated lab test results.

7.3.5 Remediul measures and pile capacity just$-
cution
Coring was carried out up to the depth below the
detected crack to verify the pile damage and then it
was grouted with non-shrink cement. To justify the
capacity of damage piles, high strain dynamic load
test was carried out on 25 piles in which 24 piles
with detected cracks and 1 pile without crack detec-
tion. Summary of dynamic load test is presented in
Table 5 .
9 CORRELATION BETWEEN RESULTS FROM
FINITE ELEMENT AND SONIC INTEGRITY
TEST
The analysis results indicate that bending moment in
piles induced by excessive ground movement due to
excavation works are higher than that of cracking
capacity of pile in all cases. Table 6 presents the
summary of the analyses.

Table 5. SLiininary of dynamic load test results of damaged Table 6. Moment capacity of piles and induced moment i n
piles -Case I l l piles due to excavation simulated by finite element method
Case Pile liebar Cracking Ultiniate Moment by
Mobilized capacity / Design ultimate No. of six (%)) moment inonlent FEM model
caoacitv of 240 ton uile
(111) (t.m) (Whitney) (t.m)
Less than 50% 1
Less than 10% 2 I 1 .O 0.75 30 79 149.7
More thal1 100% 22 11 1.5 0.50 102 188 117.2
255.7*
I11 0.60 0.35 6.5 9.2 13.3
8 mum ELEMENT ANALYSIS WITH Note: (*) moment i n pile located outside of sheet pile wall
within the active side of soil mass.
COMPUTER MODEL SIMULATION

Finitc element analysis was carried out using two- As can be seen in Table 7, the crack location in pile
dimciisional modeling. With the aid of PLAXIS indicated by sonic integrity test is generally consis-
computer program staged excavation was simulated tent with the location of computed bending moment,
and pildsoil movements as well as bending stress which exceeds the cracking moment capacity of
induced in the piles were examined for above pre- piles for all three cases.
sented projects. Model profile of Case I11 used in Table 7. Locations of crack and computed bending nioments
FEM analyses is presented in Figure 9. in piles in the vicinity of excavation boundary
Case Integrity tesl llcptli of maxi- Depth of bending
depth of crack in mum bending morncnt cxcecd
piles near csca\~a- nioinent b!, ing pile's cracking
lion boundar), (in) IFEM (in) Ill 0111ell t fro Ill
I'EM ( I l l \
I 12.7B 13.7 18.0 II.0KL 13.0
11 12.78Ll3.7 15.0 12.8 8L 14.5
Ill 11.5 & 12.0 15.5 1 3 . 0 KL 14 0

It is evident that detected cracks are located near or

at the depths where boundary of Soft and Medium or
Stiff Clay is present.
In case 111, FEM analysis results suggest that
piles are subject to some degrees of basal heave as
- ~ Finite element model profile of Case I l l
F i g ~ i l 9.
illustrated in Figure 10. This is found to be caused
by inadequate embedded length of sheet pile. As can
be seen in Figure 1 1 . the results also show that

169
bending moment induced by excavation exceeds the
ultimate moment capacity of pile . Sonic integrity
test results of some piles indicate the presence of
more than one crack at some depths, suggesting piles
would have experienced both bending and tensile
stress due to the excavation.
Figure 10. Finite element analysis results showing large degree
of basal heave caused by excavation with inadequate sheet pile
embedded length - Case 111
Figure 1 1 . Computed bending moment and deflection of pile
located closest to sheet pile - Case Ill
I 0 CONCLUSIONS
With the demonstration of three case studies. effec-
ti\eiiess of sonic integrity test in pile damage as-
sessment in Bangkok is presented in this paper. This
research study also proves that external forces due to
both lateral and uplift movement of the ground in-
duced by adjacent excavation can cause considerable
damage to constructed bored piles.
170
Observation of actual construction works on sites
suggests that over excavation prior to installation of
support and inadequate retaining system are the
mail1 causes leading to excessive ground movement,
a primary reason of pile damage. Excavation con-
tractors should be well aware of the consequences of
excessive soil movement induced by improper con-
struction practice.
The designer of the foundation should also be
aware of the potential problem on site and should
take the actual construction practice into considera-
tion in the design. For instance, if piles are expected
to experience the excessive tensile or bending stress
from the most practical excavation method, suffi-
cient reinforcement should be provided in the design
at the first place. The foundation designer should
closely be involved in the design of the retaining
system and construction control on site during the
excavation period so that any possible negative ef-
fect can be minimized if not entirely avoided.
REFERENCES
Balasubramaniutn, A S . I99 1, Inaugural Lecture on Contribii-
tions in Geotechnical Engineering, Soil Mechanics and
Foundation Engineering, AIT, Bangkok, March 199 1
Turner, M. J. 1997. Integrity Testing in Piling Practice, CIRlA
Report No. 144, Construction Indbstry Research and Asso-
ciation, London.
Thasnanipan, T., Maung, A. W. and Tanseng, P. 1998a. Dam-
ages to Piles associated with Excavation Works in Bangkok
Conjkrence on Problems qf Pile
Soft Clay, 6‘” In~ernutioti~~I
Fuz~ndutionsBzii/ding, Ufa, Russia. September 14-18, 1998.
pp. 91-98
Thasnanipan, T.,Maung, A. W. and Baskaran, G. 1998b.
Sonic Integrity Test on Piles Founded in Bangkok Subsoil -
Signal Characteristics and Their Interpretations, 4‘” Infer-
national Confirence on Case Histories in Geoiechriicul En-
gineering, St. Louis, Missouri. March 9-12, 1998. pp. 1086-
1092.
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Non-destructive integrity testing on piles founded in Bangkok subsoil

N.Thasnanipan, A.W. Maung, T. Navaneethan & Z.Z.Aye

Seufco Company Limited, Bungkok, Thuilund

ABSTRACT: Pile foundation is compulsory in Bangkok, Thailand, to support almost all types of structure
due to prevailing subsoil condition. Large bored piles of 600mm to 1800mm in diameter and barrettes
(0.8mx2.7m-1 Smx3.Om) with toe level up to 60m depth are commonly used to support heavy structures cov-
ering high-rise buildings, subway stations, elevated highways, etc. In order to test or check the quality of such
high load carrying foundation piles constructed in multi-layered soil conditions, three main types of non-
destructive integrity tests such as sonic integrity, sonic logging and high strain dynamic load tests are usually
employed. Sonic integrity test is the most common method in integrity testing on both driven and bored piles.
Sonic logging test is occasionally employed as a part of pre-planned site quality control system, High strain
dynamic load test is commonly used for pile capacity evaluation. This paper presents an overview of these
three commonly used non-destructive tests on piles in Bangkok subsoil. Some findings from the author's in-
volvement in research works are also presented.

1 INTRODUCTION 38m. Second Dense to Very Dense Sand layer occurs

at about 46m depth. Between these sand layers,
Three techniques namely sonic integrity, sonic log- Hard Clay layer is present.
ging and high strain dynamic load tests are the
commonly used non-destructive integrity tests ap-
plied in Bangkok, Thailand. The effectiveness and
suitability in terms of both cost and method of appli-
cation itself are the primary reasons of the greatest
growth in use of these tests. An overview of sonic
integrity, sonic logging and high strain dynamic load
tests conducted on piles particularly bored piles
founded in Bangkok subsoil is presented in this pa-
per.

2 TYPICAL SUBSOIL CONDITION IN

BANGKOK

Subsoil profile is relatively consistent at different

localities in Bangkok. A typical subsoil profile is
characterized by the alternating layers of clay and
sand deposits as soil succession shown in Figure 1.
Weathered Crust of Im to 3m thick occurs as upper
most layer underlain by Very Soft to Soft Clay
commonly known as Bangkok Soft Clay (BSC). Figure 1. Typical soil profile of Bangkok plain with pie-
Thickness of soft Clay layer varies from 12pL to zometric draw down conditions (after Thasnanipan et al.,
18m depending on the location. Underneath Soft 1998a).
Clay, lies Medium to Stiff Clay. Dense to Very
Dense Sand layer also known as First Sand layer is The depth of bedrock is not well determined in
found below Stiff Clay layer at depths about 25m to the Bangkok area but has been reported to be at least

171
550m deep (Balasubramaniam, 1991). Existing
pore water pressure conditions in upper part of BSC
are hydrostatic from nearly l m below the ground
level. Then the hydrostatic conditions change to pie-
zometric draw down near bottom level of BSC.
3 COMMON CONSTRUCTION METHOD
The common practice of bored pile construction
method in Bangkok can be divided into wet and dry
processes.
For the wet process, bored piles are constructed
using bentonite or polymer slurry as drilling fluid
over the full length of the piles, to ensure the stabil-
ity of the excavation at all stages. Top 15m-thick
soft clay was temporarily cased. Drilling for the
piles was done with a conventional drilling auger or
auger bucket. To clean the sediments at the borehole
base the airlift technique or cleaning bucket is ap-
plied. After lowering the rebar cage, concrete is
poured with tremie method.
Generally the dry process is employed for small
piles (0.35m to 0.6m in diameter and 25m in length).
This method uses a steel temporary casing of about
151~1 in length to protect the soft clay froin caving in.
Boring is carried out by rotary drilling or percussion
tripod rig.
Bored piles are commonly constructed with rein-
forcement of 0.5-1.2% of sectional area of pile.
4 COMMON DEFECTS CAUSED IN BORED
PILES
Common pile defects such as size reductiodnecking,
discontinuity, soil/slurry inclusions, etc. may be
caused during pile construction. Most of the cracks
in bored piles caused at post construction stage are
usually induced by construction activities associated
with basement excavation and improper trimming of
pile head to design cut-off level. Causes of defects
in bored piles during construction are;
Inadequate length to protect the soft clay layer
Delayed feeding of slurry and improper slurry
level maintenance for wet process
Long time duration of maintaining open borehole
Non continuous concrete pouring or disruption in
concreting
Lifting up of insufficient workable concrete or
hardened concrete upon extracting the temporary
casing
Contamination of concrete with drilling slurry or
soil
Soft pile toe due to improper base cleaning
172
5 THE SELECTION OF SUITABLE METHOD
In general, non-destructive tests are conducted to
verify the quality of the pile. Turner (1997) high-
lighted the factors in selection of pile testing in
CIRIA Report 144 as follows;
The perceived nature of possible features or de-
fects within the pile
The ability of the test method to detect the fea-
ture or defect under investigation
The cost of testing and examination
The ease of use and interpretation
Like other indirect measurement of pile integrity,
sonic integrity and sonic logging tests are only capa-
ble of identifying structurally significant features.
Rejection or acceptance of individual piles should
not rely only on the results of these tests. Further in-
vestigation, engineering evaluation and judgement
are highly recommended to confirm the defect de-
tected by these indirect techniques. Both methods
however are very useful for cost-effective screening
test to identify piles with potential defect.
Sonic integrity test is usually selected for both
quality check (control test) and retrospective investi-
gation. It is the cheapest in terms of cost and the
simplest in terms of testing process. The main ad-
vantage of this test is that since no pai-ticular meas-
uredpreparation is necessary during the pile con-
struction phase it is more flexible to select which
pile is to be tested. However, interpretation of sonic
integrity testing needs considerable experience and
knowledge in testing, subsoil condition and con-
struction method.
Sonic logging test is relatively expensive. It is
mainly employed as a pre-planed site quality control
testing. The major advantage of this method is that
test can be carried out shortly after the pile
construction. Hence, rectification measures can be
implemented while the foundation contractor is on
site. However, this method is generally not applica-
ble if pile integrity is in question due to post con-
struction activities as access tubes are usually
grouted after completion of the test.
High strain dynamic load test is usually selected to
verify the load carrying capacity of piles. Pile integ-
rity can also be determined by high strain dynamic
load test. The major advantages of this test in com-
parison with static pile load test are those of cost,
time and space requirement. In some projects, dy-
namic load test was applied to justify the capacity of
pile after rectifying its defect detected by other integ-
rity tests. Table 1 shows the suitability of test
method for different type of common defects.
 Table 1. The suitability of test method for different type of de-
fects.
Type of defects Test method and suitability
for defect detection
Sonic Sonic
integrity logging
~
Crack Suitable Possible
Pile size Suitable Possible
change
Poor quality con- Possible Suitable
Crete or Contami-
nated concrete
Soft pile toe - Not usually Suitable
6 SONIC INTEGRITY TEST
6.1 Overview of the test in Bangkok
Sonic integrity test, also known seismic test is the
most common method of integrity testing for both
driven and bored piles in Bangkok. In many projects
it is a part of the contractual requirement to conduct
sonic integrity test. Minimum 10 % to maximum
100 % of production piles are commonly tested. It is
also a reasonably acceptable method for applying as
a retrospective investigation in determining integrity
of the pile.
6.2 Interpretation of the test results
As has been reported in various literatures, it is to be
emphasized that sonic integrity test is highly de-
pendent on the experience in both field testing and
interpretation. The signal characteristics and their
interpretations of sonic integrity test on piles
founded in Bangkok subsoil were reported by Thas-
nasipan et al. (1998b). The authors presented the
interpretation of various features of sonic integrity
test signals in comparison with soil investigation
data and construction records.
6.3 Some findings from sonic integrity test in
Bangkok
Toe reflections are generally clear for the piles with
length to diameter ratio (L/D) within 40, but toe re-
flection was hardly detected for piles founded at
depth deeper than 30-35111 as reported by Thas-
nanipan (1998b). However, it was found that toe
reflection is detectable for some deep-seated bored
piles of about 50m in length if pile was tested at
early age. Figure 2 shows the significant toe reflec-
tion of 52.0m-long pile tested few days after con-
creting. This is considered mainly due to the fact
that at early stage the bonding between soil and pile
has not yet fully developed so that damping effect or
attenuation of the sonic signal by shaft friction is
relatively low.
High strain
dynamic
load test
Suitable
Suitable (if
severe defect) Figure 2. Sonic signal showing a toe reflection of bored pile at
Suitable (if 52.0m tested few days after concreting.
severe defect)
Use of temporary casing usually causes larger size at
the top portion (from pile top to casing bottom) for
completed piles in Bangkok subsoil. This is caused
by the size difference between casing and the drilling
tool (bucket/auger) as the top section of pile formed
by the outside diameter of the casing is usually 5-
lOcm larger than the lower section of pile with
nominal diameter formed in the borehole made by
drilling tool (Figure 3). Test signal shown in Figure
3 indicates a clear reflection at approximate 1 l m
from pile top (about 15m below ground level) with a
repeat or multiple reflection at about 2 times of that
distance (approximate 22m from pile top). As the
polarity of the reflection is negative, the first inter-
pretation is that this pile has decrease in cross sec-
tion or crack at 15m below ground level. However,
with the consideration of pile construction method, it
is concluded that negative reflection is caused by
variation in pile cross section at the bottom end of
15m long temporary casing. The sonic signals with
anomaly acquired on those piles are often misinter-
preted as indicative of defect piles, leading to an ar-
gument.
Figure 3. A typical pile shape constructed in Bangkok soil and
resulted sonic integrity test signal profile.
For the analysis of sonic integrity test, velocity of
sonic wave in pile material (concrete) is typically as-
sumed 4000m/s in most cases. Basically, lower
strength of concrete gives lower wave propagation
velocity. The trend of sonic wave velocity increment
with pile age derived from a number of 600mm di-
ameter bored piles with known lengths is presented
in Figure 4.
173
 overcome by drilling through the obstruction in the
worst case (but not possible if tube is bent), it should
be eliminated in the first place by careful fabrication,
installation and protection of the tubes.

7.2 Interpretation of test results

The results of a series of laboratory tests to deter-
mine the effects of various defects or inhomogenei-
ties within a concrete section were reported by Stain
and Williams (1991). The test results were carried
Figure 4. A curve of sonic wave velocity vs. pile age in Bang- out on small panels constructed to model various
kok sub soil (fc’ = 35 MPa). pile construction defects and anomalies. The results
from these panels were related to tests on control
panels formed from homogeneous concrete.
7 SONIC LOGGING TEST Faiella and Superbo (1998) reported the analysis
of the sonic logging test results of over 6800 piles
collected from 37 sites in Italy. Based on the results
7.1 Overview of the test in Bangkok of analyses, the authors presented the defect classifi-
Use of Sonic logging test for checking pile integrity cation criteria for piles monitored by different num-
has increased in Thailand in recent years. It is used ber of tubes as illustrated in Tables 3 (a) and 3 (b).
for cast-in-place foundation elements such as bored
piles, barrettes, diaphragm walls and caissons. Both Table 3 (a). Defect classification criteria monitored by means
of 2 tubes (Faiella and Superbo, 1998)
PVC and steel tubes can be used as access tubes.
Though cost of steel tube is higher than that of PVC, T,/T, Tk/D Type of defect
steel tubes are considered more suitable than PVC <1.15 - Non-homogeneous concrete
tubes for two reasons; (1) less potential for bending <3 Light
1.15-1.45 >3
and damage especially for deep-seated bored piles Probablv serious
and barrettes with heavy steel reinforcement and (2) <1 Light
1.45-2.0
better bonding with concrete. These access tubes are >I Probablv serious
<0.5 Probably serious
sometimes also used for pile base grouting. The ac- >2’o >OS Light
cess tubes are attached to the reinforcement cage, ~

which is installed prior to concreting as a normal Table 3 (b). Defect classification criteria monitored by means
practice for cast-in-place piling. of 3 and 4 tubes (Faiella and Superbo, 1998).
Two access tubes are required as a minimum for T,/T, P,/P, TJD Type of defect
sonic logging test. For a good coverage of the test
pile, recommended number of tubes for different <1.15 ~ p,<p, - Non-homogeneous concrete
p,=pt
sizes of bored piles is shown in Table 2. <3 Non-homogeneous concrete
pa<pt >3 Light
Table 2. Recommended number of tubes for different pile size 1.15-1.45
Pile Diameter Minimum No. of Tube spacing <3 Light
P,=P,
(.lm> tubes (degree) >3 Serious
D 5 750 2 180 <1 Non-homogeneous concrete
750<D I 1000 3 120 P,<P, 1-3 Light
1000<DI 1500 4 90 1.45-2.0 >3 Serious
1500< D 5 2500 6 60 <1 Light
P,=P,
2500 < D 8 45 >1 Serious
<0.5 Lieht
A typical problem encountered in sonic logging test P,<P, 0.5-3 Serious
is the access tube blockage which often comes to >2.0 >3 Serious
know only at the time sonic testing is conducted one P,=P,
<0.5 Light
week or more after pile construction. The blockage >0.5 Serious
of access tube is mainly caused by intrusion of con-
crete from leakage at either bottom cap or tube con- where:
nection and intrusion of soil or other material from T, = Travel time of sonic waves in the anomalous zone
the top cap of the tube. Sonic test instruments T, = Travel time of sonic waves in the sound concrete
(transducer and receiver) sometimes may not be able Tk= Thickness of the anomalous zone
to insert down to the bottom of the access tubes due P, = Number of measurement paths affected by sonic anomalies
to the bend of the tube itself. Though the problem P, = Total number of travel paths for each pile
associated with the access tube blockage can be D =Pile diameter

174
Srivanavit et al. (1999) reported the interpretation of
test signals in comparison with the actual integrity of
9 model tests conducted in Bangkok. Model piles are
of 800mm diameter with l m length. The summary
of the model test results including the description of
model piles as well as discussions by authors are il-
lustrated in Table 4.
Table 4. The summary of sonic logging model test results

7.3 Factors to be considered in interpretation of

sonic logging test
Turner (1997) pointed out the need of both theoreti-
cal knowledge and practical experience in interpre-
tation of the test results in CIRIA Report 144. Sug-
gestion is also made in this report that the anomaly
shown in sonic test signals can be caused not only by
changing in physical properties of the materials, but
also by factors within the measuring system itself.
These factors are;

e Free movement of the probes within access tubes

e Mismatched probe positions especially at pile toe
e Measurement resolution
e Incorrect position of access tube
e Air gaps or different material around access
tubes
e Aggregate variation (in the case of base grouting)

175
It is to be noted from the author’s suggestion based
on the above factors that only variations of transmit
time more than 15 to 20% of the norm should be re-
garded as warranting further investigation.
8 HIGH STRAIN DYNAMIC LOAD TEST
8.1 Overview of the test in Bangkok
High strain dynamic integrity test has become a
well-accepted method especially for evaluating the
pile capacity in the foundation industry today. A
large number of related technical papers and case
histories of the test have been published and it is a
part of standards and specifications such as ASTM
D4945-89 (Standard Method for High-strain Dy-
namic Testing of Piles). In Bangkok, dynamic load
test is applied for both driven and bored piles. A
summary of the number of dynamic load test carried
out during 1991 to 1997 is shown in Table 5. It is to
be noted that quantities shown are collected from the
available sources and the actual tested numbers are
likely to be slightly more than those indicated in the
table. As can be seen in the table, application of dy-
namic load test increases year by year in Thailand.
The pile driving analyzer (PDA) with computer
software by Pile Dynamic INC, USA is mainly used
in Bangkok.
Table 5. Yearly minimum quantity of dynamic load test con-
ducted in Thailand during 1991 to 1997
Year No. of dynamic load test done
1991 84
1992 134
1993 242
1994 369
1995 464
1996 473
1997 695
For driven pile it is usually performed at two stages,
during driving (initial driving monitoring) and some
period after installation of pile (restrike test). For
bored piles, a specially designed pile cap is normally
required as an integral part of the pile head to avoid
pile head damage. Ram weights of 4 ton to 20 ton
are commonly used in Bangkok. The common ram
weights used for different pile sizes are shown in
Table 6.
Table 6. Common ram weights used for different size of piles
Pile size (m) Ram weight (ton)
5 0.50 4.00
0.50 - 0.80 8.00
1 .oo - 1 s o 20.00
The dynamic load test is also occasionally used as an
additional investigation when pile is found with de-
176
fect by other integrity tests. Figure 5 shows the
photo of dynamic load test carried out on barrette of
size (1 .Ox2.7m) with toe depth 48.94m located at the
bank of water supply canal. A 20 ton hammer was
used to activate the 2050 ton test load in this project.
Figure 5. Dynamic load test carried out on a barrette
(1 .Om x 2.7mx 48.94m).
8.2 Comparison and correlation of dynamic and
static pile load test
A comparison between dynamic and static load test
results has been reported by various researchers.
Seidel and Rausche (1984) reported the results of
dynamic and static load tests performed on drilled
shafts of the West Gate Freeway in Melbourne,
Australia. A 20 ton hammer with drop heights be-
tween 1.6 and 2.5m was used for the dynamic tests
of 12 shafts ranging from 1 lOOmm and 1500mm in
diameter and 35m to 64m in length. The authors re-
ported that dynamic activation of static pile resis-
tance forces exceeded 3000 ton for some 1500mm
diameter shafts. Skin friction predictions from dy-
namic load tests and values obtained from instru-
mented shafts under static load tests were remarka-
bly similar and pile head load-movement
relationships obtained from both test methods were
comparable as reported by the authors.
Prebaharan et al. (1990) reported the results of
dynamic and static tests conducted on bored piles at
the Marina Bay Station of the Singapore Rapid Tran-
sit System. According to the report, dynamic load
tests were carried out for bored piles of lOOOmm
diameter and 25m to 50m length founded in Old
Alluvium of Singapore Island. Agreement between
dynamic and static load tests results permitted re-
placement of 44 out of 51 tests with dynamic load
test from originally planned static load tests as stated
by the authors.
Vasinvarthana and Kampananon (I 997) presented
the efficiency and reliability of dynamic load test
carried out in Bangkok. Figures 6 (a) and 6 (b) show
the load settlement curves of dynamic and static load
tests conducted in Bangkok on large and small di-
ameter piles respectively. As can be seen in the fig-
ures, load-settlement characteristics obtained from
dynamic load test agree well with those of static load
test for both large and small diameter bored piles.
Figure 6 (a). Load-settlement curves of dynamic and static load
tests on large diameter piles ($I 0.80m x 50.0m).
Figure 6 (b). Load-settlement curves of dynamic and static
load tests on small diameter piles ($I 0.35m x 18.91m).
9 CONCLUSIONS
Use of sonic integrity, sonic logging and high strain
dynamic tests in Bangkok has been presented.
Though sonic integrity test is a simple and cost
effective method, the reliability of the test results is
highly dependent on the experience of the person in
both field-testing and interpretation.
The results from sonic logging test conducted on
model piles in Bangkok helps to extend the knowl-
edge of the signal characteristics and interpretation.
High strain dynamic test has become a well-
accepted method especially for load testing of piles.
Agreement between static and dynamic load test re-
sults enhances the confidence in using the dynamic
load test as shown by rapid increment in number of
tests conducted in seven years in Thailand, mainly in
Bangkok.
177
ACKNOWLEDGEMENT
The authors wish to express their appreciation to
EDE Co., Ltd. for providing the test data. Thanks
are also given to Mr. Veera Vasinvarthana and Mr.
Natomon Kampananon STS Engineering Consult-
ants Co., Ltd. for giving permission for the use of
their previous works in this paper.
REFERENCES
American Society for Testing and Materials 1989. ASTM
04945-89 Standard inethod for high-strain dynamic testing
of piles, Philadelphia: American Society for Testing and
Materials.
Balasubramaniam, A. S. 199 1. Inaugural lecture on contribu-
tion in geotechnical engineering soil mechanics and fozrn-
dation engineering, AIT, Bangkok, 14 March 199 1.
Faiella, D. & Superbo, S. 1998. Integrity non destructive tests
of deep foundations by means of sonic methods - Analysis
of the results collected on 37 sites in Italy, Van Impe &
Haegeman (eds), Proc. 3rd Int. Geotechnical Seminar on
Deep Foundations on Bored and Auger Piles, BAP 111,
Ghent, Belgium, 19-2 1 October 1998: 209-2 13. Rotterdam:
Balkema.
Prebaharan, N., Broms, B., Yu, R. & Li, S. 1990. Dynamic
testing of bored piles, Proc. of the Tenth Southeast Asian
Geotechnical Conference, Volunw I , Taipei, ROC, 16-20
April 1990: 373-378.
Seidel, J. & Rausche, F. 1984. Correlation of static and dy-
namic pile tests on large diameter drilled shafts, Proc. 2""
Int. Col$ on the Application of Stress Wave Theory on
Piles, Stockholm, Sweden, 27-30 May 1984: 313-3 18.
Stockholm: Swedish Pile Commission.
Srivanavit P., Hiranoon J. & Pimsarn S., 1999. Integrity tests
on deep bored piles constructed in Bangkok subsoil by
sonic logging method & interpretation, The engineering
rechnology exhibition and syinposiurn, Organized by Engi-
neering Institute of Thailand, Bangkok, 1-2 November
1999: 1-8. (in Thai).
Stain R. T. & Williams, H. T. 1991. Interpretation of sonic
coring results: A research Project, Proc. 4"' Int. DFI ConJ
on Piling and Deep Foundations, Stresa, Italy, 7-12 April
1991: 633-640. Rotterdam: Balkema.
Thasnanipan, N., Baskaran, G., & Anwar, M.A. 1998a. Effect
of construction time and bentonite viscosity on shaft capac-
ity of bored piles, Deep Foundations on Bored and Auger
Piles, Van Impe & Haegeman (eds), Proc. 3"' Int. Geotech-
nical Seminar on Deep Foundations on Bored and Auger
Piles, BAP 111, Ghent, Belgium, 19-21 October 1998: 171-
177. Rotterdam : Balkema.
Thasnanipan, N., Maung, A. W. & Baskaran, G. 1998b. Sonic
integrity test on pile founded in Bangkok subsoil - signal
characteristics and their interpretations, Proc. 4"' Int. Cot$
on Case Histories in Geotechnical Engineering, St. Louis,
Missouri, 9-12 March 1998: 1086-1092.
Turner, M. J. 1997. Integrity testing in piling practice, CIRIA
Report 144, London: Construction Industry Research and
Information Association.
Vasinvarthana, V. & Kampananon, N. 1997. Efficiency and
reliability of dynamic load test), Seminar on Foundation '
97, Organized by Engineering Institute of Thailand, Bang-
kok: 79-94. (in Thai).
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Estimation of pile head stiffness using sonic integrtty testing

K. Irnada & Y. Nakata
Fuji Engineering Consultant Company Limited, Osaku, Japan
T. Matsumoto
Department of Civil Engineering, Kanazawa University,Japan

ABSTRACT: This paper presents a method for estimating the initial stifhess of the load-displacement of a
pile (pile head stifkess) fiom the Sonic Integrity Test (SIT). Basically, the proposed method is identical to
wave form matching commonly used for the interpretation of dynamic load testing. Validity of the proposed
method was examined through a series of SITs and static load tests (SLTs) of a model pile, 800 mm in length
and 25 mm in diameter, embedded in model grounds having different profiles of shear modulus of the soils.
The proposed method was further applied to the field tests of two cast-in-situ concrete piles where SITs and
SLTs were conducted. It is shown fi-om the model tests and the field tests that the proposed method gives us
a good estimate of the initial stifhess of the load-displacement curve of a pile.

1 INTRODUCTION having different profiles of shear modulus of the soils.

The proposed method is further applied to the field
Pile foundations should satis@ required qualities and tests of two cast-in-situ concrete piles where SITs
performances. It will be of vital importance to ensure and SLTs were conducted. It is shown from the
these requirements are met for constructed piles in model tests and the field tests that the proposed
the limit state design and the performance based method gives us a good estimate of the initial
design. stifhess of the load-displacement curve of a pile.
For cast-in-situ concrete piles, the sonic integrity
testing (SIT) has been widely used to assess qualities 2 INTEGRITY TEST AND STATIC LOAD TEST
and mechanical properties such as configurations, OF MODEL PILE
Young's modulus and the strength of piles after
hardening of the concrete. The static load test (SLT) 2.1 Model ground, model pile and test procedure
and rapid load tests such as the Statnamic test has The test apparatus for model tests is shown in Figure
been employed to obtain the performances of a pile 1. The model grounds were made in a cylindrical
such as the load-displacement relation as well as the
ultimate pile capacity. However, these load tests are
conducted on only a few piles at a construction site,
because these tests are time and cost consuming
compared to the SIT that may be conducted on to all
piles constructed at a site in a time and cost efficient
manner.
This paper presents a method to estimate the initial
pile head stiilbess (initial tangential stiffness of the
load-displacement curve at the earliest stages of
loading) of a pile by conducting a sonic integrity test
with the interpretation based on one-dimensional
stress-wave theory. Basically, the proposed method
is identical to the wave form matching commonly
used for the interpretation of dynamic load testing,
except that FEM is adopted to estimate the load-
displacement relation for the tested pile. Validity of
the proposed method is examined through a series of
SITs and SLTs of a model pile, 800 mm in length and
25 rnm in diameter, embedded in model grounds Figure 1. Model test apparatus.

179
acrylic box having an inner diameter of 250 mm and a length, L, of 8 0 0 Young's ~ modulus, E, of
height of 950mm. The soil of the model grounds was 1.65.1 O6 kN/m2, a density, p, of 2.1 8ton/m3and a bar
a dry Toyoura sand mixed with lime by a weight ratio wave velocity, c, of 8OOm/s, was placed in the soil
of 15% (weight of lime/weight of Toyoura sand = box after the completion of the bottom soil section.
0.15). Thereafter the middle and upper soil sections were
made. The number of tamping for each soil layer of
Table 1. Conditions of model grounds. 100 mm thickness in the three soil sections is listed in
Case No Number of tamping for each soil Number of Table 1.
layer of 1OOmm thickness soil sections A total of six different model grounds were made.
Upper Middle Lower The longitudinal wave velocity, VP,of each soil layer
section section section was estimated by generating a sonic pulse on the
Case- 1 0 0 0 1 ground surface, and cone penetration test was
Case-2 8 8 8 1 conducted to obtain a variation with depth of the tip
Case-3 0 0 40 2 resistance, 9.. In Cases-5 and 6, the sonic integrity
Case4 8 8 40 2 and the static load test were conducted on the model
Case-5 0 8 40 3 pile. This paper will present and discuss the test
Case-6 8 0 40 3 results of Cases-5 and 6 only.
Figure 2 shows the variations with the the depth of
The soil was compacted in the cylindrical box by the cone tip resistance, q., and the soil density, p.
letting a ram having a weight of 58.8N fall fiom a The shear wave velocity, VS,was estimated fiom
height of 20mm to make a soil layer of 100 mm the measured VPby means of Equation (1).
thickness. This procedure was iterated 8 times to
obtain the model ground of 800 mm height. 1- 2v,
Accelerometers (acoustic emission sensors) were v -- v4 2______
( 1 -Vs)
placed on the top of each soil layer of 100 mm
thickness to measure the longitudinal wave velocity, where v. is Poisson's ratio of the soil which was
VP,of each layer. The model ground consisted of assumed to be 0.3.
three soil sections as indicated in Figure 1. A model
pile made of Teflon, having a diameter, d, of 2 5 m , a

Figure 3. Measured variations in shear wave velocity with

depth of model grounds for Case-5 and Case-6.

2.2 Wave analysis of the integrity tests

A computer program KWAVE developed by
Matsumoto & Takei (1991) was used for the wave
propagation analysis for the sonic integrity tests of
the model pile. The pilehoil system adopted in
KWAVE is shown in Figure 4.
Rational soil resistance models proposed by
Randolph & Simons (1986) and by Deeks (1992)
have been incorporated in KWAVE for the shaft and
the base resistance, respectively.
The pile displacement and the displacement of the
(b) model ground for Case-6 soil adjacent to the pile is calculated separately so
that the viscous damping and the radiation damping
Figure 2. Variations in cone tip resistance and density with
depth.
are allowed for individually (Figure 5). Relative
displacement between the pile and the soil does not

180
occur until the mobilized shear stress reaches the For the base model (Figure 6), the spring constant,
maximum shear stress, zm,, that is expressed by the kb, the damping constant, C b , and the lumped mass,
slider in Figure 5 . According to Novak et al. (1978), Mb, are given as follows (Deeks 1992):
the spring constant, k,, and the radiation damping, er,
of the outer shaft resistance are estimated as follows: 8G
kb =- (4)
~ ( 1 v)d
-
k, = 2.756 l(d), d= pile diameter (2)
3.2p,VS - 3.2G
-
C b = p
c, = G I V,, G = shear modulus of the soil (3) n(1- v ) n(1- V)V,

0.1 - v 4
M~ = 2d3 ~

1-v

In the wave matching analyses of the sonic

integrity test signals, the values of the maximum shaft
resistance, zmas, and the maximum base resistance, q b ,
were taken as very large, because it was judged that
rmxand q b were not reached in the sonic integrity
tests fi-om the fact that permanent settlement of the
pile did not occur after the sonic integrity tests.
The wave propagation analyses were conducted
assuming various profiles of VSof the model grounds
until a good matching of the calculated and measured
pile head velocities were obtained. Note that all the
soil resistance parameters are determined from the
assumed VSand the measured ps.
Figure 4. Modeling of pile and soil used in KWAVE.

Figure 5. Shaft model (after Randolph & Simons 1986).

Figure 7. Calculated and measured pile head velocity.

Figure 6. Base model (after Deeks 1992).

181
 The calculated and measured pile head velocities
for Cases-5 and 6 are shown in Figure 7. The
assumed profiles of VS of the model grounds are
shown in Figure 8, together with the measured VS.
The profiles of VS estimated f?om the sonic integrity
tests seems to represent the measured results well,
although the estimated VS underestimate the
measured VS of shallower depths down to 350 mrn.
The underestimation of V . for the shallower depths
Figure 8. Variations in shear wave velocity with depth
estimated from wave matching analysis, together with
measured results.
Figure 9. Axi-symmetric FEM mesh for model ground and
pile.
182
might be due to loose contacts between the pile and
the surrounding soils.
2.3 FEM analyses of the static load tests
FEM analyses were conducted to estimate the load-
displacement curves for the model pile in Cases-5 and
6.
The axi-symmetric FEM mesh for the model
ground and the model pile is shown in Figure 9. The
lateral displacements along the center axis, the
vertical and lateral displacements at the bottom and
side of the ground were fixed to reflect the model test
conditions. Both the pile and the soil were model as
elastic materials. The shear moduli, G, of the soils
estimated from the sonic integrity tests (G = pYVs2)
and vs = 0.3 were used in the FEM analyses.
The load-displacement curves estimated from
FEM analyses for Cases-5 and 6 are compared with
the curves obtained from the static load tests in
Figure 10. The estimated load-displacement curves
are good measures for the measured results. It
should be noted that the measured load-displacement
curves shown in Figure 10 are initial portions where
an almost linear relation between the displacement
and the load was obtained.
Figure 10. Comparison of estimated and measured load-
displacement curves.
 3 INTEGRITY TESTS AND STATIC LOAD
TESTS OF CAST-IN-SITU CONCRETE PILES
Above, a method for estimating the initial stifThess of
load-displacement of a pile fiom the sonic integrity
test was demonstrated against the model pile. Below,
the proposed method is applied to two actual cast-in-
situ concrete piles.

3.1 Test description

The sonic integrity test and the static load test were
performed on two cast-in-situ concrete piles,
designated as pile D and pile E, constructed in two
different sites. The soil profile and the SPT N-values
at the test sites of piles D and E are shown in Figures
11 and 12, respectively. The specifications of the
concrete piles are listed in Table 2.

Table 2. Specifications of test piles and principal results of

static load tests.
Item Pile D Pile I
Pile Diameter (m) 1.2 1.5
Length at SLT(m) 23.5 44.9
Embedded length at SLT (m) 22.5 43.9
Figure 1 1. Soil profile and SPT N-values at the test site of pile Length at SIT (m) 17.6 36.4
D, together with the pile seating. Young's modulus (kN/m*) 2.27~1 O7 2.60~10'
Bar wave velocitv (m/s) 4000 4200
Principal Maximum load (MN) 16.7 19.5
results Pile head displacement at
maximum load (mm) 200 52
Yield load (MN) 12.7 16.7

A static the load test was performed on pile D,

having an original Iength of 23.5m, 4 weeks after the
pile construction. The top 5.9m of soil was
excavated aRer the static load test, and the
corresponding section of pile D was cut, so that pile
D had a length of 17.6m. Thereafter, the sonic
integrity test was conducted on pile D of 17.6m
length. Pile E had a similar situation such that the
original length of 43.9m at the static load test was
shortened to 36.4m before conducting the sonic
integrity test.
The measured load-displacement curves of piles D
and E are shown in Figure 13. In order to compare
the load-displacement curve to be estimated from the
sonic integrity test with the static load-displacement
curve, the measured load-displacement curve of pile
D with the original length of 23.5m was corrected to
estimate a load-displacement curve of pile D with the
shortened length of 17.6m, as follows. As the axial
forces of pile D were measured, the shaft friction
along the top 5.9m of pile D was subtracted fiom the
measured pile head load, and the elastic deformation
Figure 12. Soil profile and SPT N-values at the test site of pile
E, together with the pile seating.
of a pile length of 5.9m was subtracted fiom the
measured pile head displacement. The load-
displacement curve of pie E with the shortened
length of 36.4m was also estimated similarly.
Figure 13. The measured load-displacement curves of piles D
and E of full length, and the estimated load-displacement
curves for shortened lengths at integrity tests.

The corrected load-displacement curves of piles D

and E are also shown in Figure 13. Each corrected
curve will be compared with the load-displacement
curve estimated from the sonic integrity test.

3.2 Wave analyses of integrity test signals

Wave analyses of the sonic integrity test Figure 15. Axi-symmetric FEM meshes for piles D and E.
signals of piles D and E were performed using the
KWAVE program. In the analyses, Poisson’s ratio

Figure 16a. Load displacement curves and estimated initial pile

head stiffness of piles D and E.

was assumed to be 0.4 for all soils, and density was

assumed to be 1.9todm’ for sandy soils and
1.7todm3 for clayey soils. Wave propagation
analyses were iterated assuming various profiles for
Figure 14. Calculated and measured pile head velocities. the shear modulus, G. The results of the final

184
matching are shown in Figure 14. In Figure 14, the Osaka area of Japan. Pile diameters range from 1.2m
pile head velcity, v, is normalized by the maximum to 2.0m and the pile lengths range fiom 11.8m to
velocity, VO, induced by the impact of hammer. The 43.9m. In Figure 16, the pile head load, P, is
profile of G assumed in the final matching have been normalized by the ultimate pile capacity, P,, and the
shown in Figures 11 and 12 for pile D and pile E, pile head displacement, w, is normalized by the yield
respectively. pile head displacement, 4, at the yield load, Py.
Therefore, PIP, corresponding to w l 4 = 1 is defined
3.3 FEM analyses of the static load tests as PJP,,. Here, let the secant stifhess, m,, be defined
as P,/6;,.
FEM analyses of static loading of piles D and E were
conducted to estimate the initial stiffness of piles D
and E.
hi-symmetric FEM meshes for piles D and E are
shown in Figure 15. The piles and the soils were
modeled as elastic materials. The profiles of the
shear modulus, G, identified from the sonic integrity
tests and Poisson's ratio of 0.4 were used for the
elastic parameters of the soils.
The calculated load-displacement curves of piles D
and E are compared with the load-displacement
curves which have been corrected from the static load
test results. The calculated and corrected measured
initial stiffness of piles D and E are listed in Table 3.
The initial stiffness, m,,, estimated from the SITS
Figure 16b. Load displacement curves of cast-in-situ concrete
Table 3 . Comparison of estimated and corrected measured
piles constructed at different sites in Osaka area.
initial pile head stiffness of piles D and E.
Pile Initial stiffness Initial stiffness m,,lm,,n
from SIT from SLT
m,, (MN/mm) m,, (MN/mm)
D 8.4 1 7.22 1.16
F 14 41 12.56 1.15

overestimate the corrected measured initial stiffness,

rn,,,,, by 15% only. The overestimation of the initial
s t f i e s s may be attributed to the non-linearity of the
soils. This is because the shear modulus, G,
estimated from the SIT corresponds to a very small
strain level, which would be higher than G of the soils
surrounding the piles under the static load tests. The
authors are aware that some assumptions concerning
Poisson's ratio and soil density made in the wave
matching analyses, also be reasons for the
overestimation of the initial stifhess.
3.4 Discussion on the uses of estimated initial
st qfness
As mentioned earlier, it is common practice in Japan
t o perform only one static load test at a pile
construction site. Estimation of initial pile stifhess of
the other piles at the site by means of the SIT could
serve for assessment of the performance of the other
piles, by comparing the estimated initial stiffhess with
the initial stifhess obtained from the static load test.
Another use of the estimated initial stiffness is
discussed here. Figure 16 shows the load-
displacement curves of a total of 11 cast-in-situ
concrete piles constructed at different sites in the
Figure 17. Frequency distribution of m,lm, of cast-in-situ
concrete piles shown in Figure 16.
The frequency distribution of the ratio of the initial
stiffness, m,, to the secant stifiess, m,, of the piles in
Figure 16 is shown in Figure 17. The ratio, m,lm,,
ranges from 0.07 to 0.36 with an average of 0.21.
Although the number of the piles is limited, the
results of Figures 16 and 17 indicate that the initial
stifhess can be used for a rough but useh1 estimation
of the secant stiffness.
4 CONCLUSIONS
A method for estimating the initial stiffness of the
load-displacement curve of a pile (initial pile head
stiflkess) fiom the sonic integrity test has been
proposed. Validity of the proposed method was

185
examined through the sonic integrity testing and
static load testing of a model pile in different ground
conditions. The proposed method was further
applied to two cast-in-situ concrete piles constructed
at two different sites at which the sonic integrity tests
and the static load tests were conducted. The
proposed method estimated the initial stifftiness of the
piles with a good accuracy, 15% overestimation of
the measured values.
The uses of the initial stifftiness were discussed and
it is demonstrated that the initial stifhess can be used
for a rough estimate of the secant stifhess.

REFERENCES
Deeks, A.J. 1992. Numerical analysis of pile driving
dynamics. PhD Thesis, The University of Western
Australia.
Matsumoto, T. & M. Takei 1991. Effects of soil plug on
behaviour of driven pipe piles. Soils & Foundations,
JSSMFE, 3 l(2): 14-34.
Novak, M., T. Nogami & F. Aboul-Ella 1978. Dynamic soil
reactions for plane strain case. J. Mech. Eng. Div., ASCE,
104(EM4): 953-959.
Randolph, M.F. & H.A. Simons 1986. An improved soil
model for one-dimensional pile driving analysis. Proc. 3rd
Int. Con$ on Num. Methods in Offshore Piling, Nuntes: 1 -
17.
Randolph, M.F. & A.J. Deeks 1992. Dynamic and static soil
models for axial pile response. Proc. 4th Int. Con$ on the
Appl. of Stress- Wave Theory to Piles, The Hague: 3- 14.

186
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Integrity testing of cast-in-situ concrete piles associated with the construction

of New Haccho Bridge
Yuji Michi
Yoshimitsu Company Incorporated, Komatsu, Japan
Tatsunori Matsumoto
Department of Civil Engineering, Kanazawa University,Japan
Yoichiro Matsuda
Ishikawa Prefectural Government, Kunazawa, Japan

ABSTRACT: Cast-in-situ concrete piles were constructed for the foundations of New Haccho Bridge at
Komatsu City, Ishikawa Prefecture, Japan in 1997. The requirements for quality control during construction
of the piles were the quantity of concrete used for each pile, the outer diameter of the casing pipe and the
excavation length. For quality assessment of the concrete piles after hardening, Sonic Integrity Testing was
conducted on a total of 8 piles of one abutment after curing periods of 17 to 28 days. Wave analysis of the
Sonic Integrity Test signals showed that all the piles were constructed as designed. These signals will play the
role of the reference signals, if a disaster such as a great earthquake should occur in the future, The reference
signals would be compared with the Sonic Integrity Test signals that would be obtained after the disaster, in
order to detect the occurrence of pile damage easily and quickly.

1 INTRODUCTION m over Haccho River in Komatsu City, Ishikawa

Non-destructive inspection of piles such as the Sonic
Integrity Test (SIT) was used to detect the existence
of damage to piles after the Hyogoken-Nambu
Earthquake in 1995. The SIT was widely used,
because the SIT is easy to conduct on piles when
their heads are exposed and because the measured
signal (pile head velocity) is interpreted based on
one-dimensional stress-wave theory.
However, there were cases in which it was
difficult to determine whether a pile had damage or
not by means of the SIT (Hayashi 1996). This may
be attributed to the fact that the influence of soil
resistance was not taken into account in the
interpretation of the measured signals and that SIT
signals were not measured before the earthquake for
the purpose of comparison.
The New Haccho Bridge was constructed in 1997
after the Hyogoken-Nambu Earthquake. Cast-in-
situ concrete piles were adopted for the foundations
o f the bridge. Another objective of conducting SITS
was to obtain reference SIT signals for the
Prefecture. The bridge is sumorted bv two
abutments with cast-in-situ concrete piles.
2.1 Site condition
Figure 1 shows the profiles of the soil layers and the
SPT N-values at the construction site. The clay and
sand of the alluvium deposits are distributing
alternately to an elevation of -18.2 m underlain by
stiff layers of gravel and sand having SPT N-values
greater than 50. End-bearing cast-in-situ concrete
piles seated in the gravel layer were adopted for the
foundation piles. The piles had a length of 23 m and
a diameter of 1 m. The design ultimate pile
capacity, R,, of the piles is 2.3 MN which was
derived fiom the following empirical rule
(Specifications 1990):
Ru = q d A p

q d = 2.94 m / m 2 for N > 30 I

constructed piles, which would be compared with
Sonic Integrity Test signals that would be obtained
after a disaster in the future.
2 OUTLINE OF THE CONSTRUCTION WORKS
The New Haccho Bridge is a highway bridge on an
access route to Komatsu Airport, with a span of 26
in which q d is the ultimate toe resistance and A, is
the cross-sectional area of the pile.
A total of 8 piles were constructed at each
abutment. Figure 2 shows the layout of the piles at
one of the abutments, designated as Al. The date
indicated beneath each pile represents the date of
concreting of the pile construction process.
The requirements for quality control during the
construction of the piles were the quantity of
concrete used for each pile, the outer diameter of the

187
Figure 1. Soil profile and SPT N-values at the construction site, together with estimated soil parameters.

2.2 Integrity testing

The SIT was conducted on all the piles at abutment
AI 17 to 28 days after the concreting of the piles,
using the IFCO SIT system. Treatment of the pile
head of each pile had been finished when the SIT
was conducted. The concrete of the top 1.2 m of the
pile was removed and 20 steel reinforcement bars,
32 mm in diameter, were exposed as shown in
Photograph 1.
Measurements of the bar wave velocities, c, of a
concrete test piece, 200 mm length and 100 rnm
diameter, were done after different curing periods of
Figure 2. The layout of the piles in abutment A,.
the concrete by means of the ultrasonic wave test.
The test results are summarized in Table 1. It can be
seen that the bar wave velocity increased with
increasing curing period until 17 days, and almost
leveled off after that. Young’s modulus, E,, of the
concrete was estimated from the following relation:

2
E, = P C C

in which pc is the density of the concrete.

E, at the SIT was estimated as E, = 36.5 GN/m2
fiom the measured p, = 2.28 ton/m3 and c =
4000m/s.

Table 1. Results of the ultrasonic wave tests conducted on the

concrete test piece.
Photograph 1. Situation of the integrity testing.
Test date Curing period Wave velocity
(days) (ds)
casing pipe and the excavation length as designed. June 26, 1997 2 3590
The requirements were achieved under inspection of June 30’ l997 6 3690
a piling engineer of experience, and it was judged 3, 1997 9 3802
that all the piles had been constructed as designed. July 4, 1997 10 3891
June 11, 1997 17 4000

188
 3 RESULTS OF THE SONIC INTEGRITY TESTS
3.1 Measured SIT signals
The measured SIT signals (time vs. pile head
velocity) of all the piles at abutment A1 are shown in
Figure 3. All the SIT signals are similar, indicating
uniformitv of the d e s .

Figure 4. Modeling of pile and soil used in KWAVE.

Figure 3. Measured SIT signals.

It can be seen from Figure 3 that the amplitude of

the pile head velocity decays with time, showing a
large influence of radiation damping of the soil.
Reflection waves of the incident waves from the pile
toe are detected at a time around 11.5 ms. The bar
wave velocities, c, of the piles are estimated as 4000
m / s from this return traveling time of the incident
waves. The bar wave velocity, c, estimated from the
SITS is identical to that of the concrete test piece
measured by the ultrasonic wave test.
The pile head velocities oscillate periodically.
Although the soil resistance is a major influence for
this, another reasons for this may be considered to Figure 5. Shaft model (after Randolph & Simons 1986).
be the influence of wave propagation in the exposed
steel reinforcement bars above the pile heads and the
changes in the properties of the cross sections of the
piles due to the different volumes of the steel
reinforcement bars along the pile shafts.

3.2 Method of wave propagation andysis

Wave propagation analyses of the SIT of pile PI
were conducted. A computer program KWAVE
developed by Matsumoto & Takei (1991) was used
for the wave propagation analyses of the sonic
integrity test signals. The pile/soil system adopted
in KWAVE is shown in Figure 4.
Rational soil resistance models proposed by
Randolph & Simons (1986) and by Deeks (1992) Figure 6 . Base model (after Deeks 1992).
have been incorporated in KWAVE for the shaft and
the base resistance, respectively.
The pile displacement and the displacement of the occur until the mobilized shear stress reaches the
soil adjacent to the pile are calculated separately so maximum shear stress, zmax,that is expressed by the
that the viscous damping and the radiation damping slider in Figure 5. According to Novak et a1 (1978),
are allowed for individually (Figure 5). Relative the spring constant, k,, and the radiation damping, c,,
displacement between the pile and the soil does not of the shaft resistance are estimated as follows:

189
Table 2. Specifications of each section in the pile.
Section No. Section 1 Section 2 Section 3 Section 4 Section 5
Distances from pile head 0.0 to 6.1 m 6.1 to 7.4m 7.4 to 11.3m 11.3 to 12.6m 12.6 to 23.0m
Length (m) 6.10 1.30 3.90 1.30 10.40
Diameter (mm) 1000 1000 1000 1000 1000
Sectional area (m2) 0.785 0.785 0.785 0.785 0.785
Area of steel bars (m2) 0.016 0.024 0.008 0.012 0.004
Young’s modulus (GN/m2) 39.93 41.66 38.21 39.07 37.34
Wave velocity ( d s ) 4000 4000 4000 4000 4000
Density (ton/m ) 2.496 2.604 2.388 2.442 2.334
Impedance (kNs/m) 7836 8176 7499 7667 7328

k, = 2.756 /(d),
d = pile diameter (3) Poisson’s ratio of the soil, v, was taken to be 0.49,
because the undrained condition was assumed to be
c, =G/ V,, G = shear modulus of the soil (4) present during the SITS.
The parameters of the soil models were
For the base model (Figure 6), the spring determined by means of Equations (3) to (7), using
constant, kb, the damping constant, L‘b, and the the soil properties thus estimated. The soil
lumped mass, Mb, are given as follows (Deeks properties and the soil model parameters have been
1992): listed in Figure 1.
The pile was divided into a total of 5 sections to
(5) take into account the different volumes of the steel
reinforcement bars, as listed in Table 2. Young’s
modulus, Ep, of the composite pile of concrete and
3.2p,VS - 3.2G steel bars was estimated fiom the following
cb =-- (6)
n(1- v) n(1- V)V, equation:

(7) Ep =(E,A, +E,A,)/(A, +A,) (10)

In the wave propagation analysis of the sonic in which

integrity tests, the values of the maximum shaft
resistance, zmax,and the maximum base resistance,
q b , were taken as very large, because it was judged
that z ,, and q b were not attained in the sonic
integrity tests from the fact that permanent
settlements of the piles did not occur after the sonic
integrity tests.
3.3 Estimation of mechanical properties of the soils
and the pile
In order to utilize the above-mentioned soil models
in the wave analysis of the SIT signals, the soil
properties were estimated as follows. The soil
density, p,, was assumed to be 1.8 ton/m3 for the
clay, 1.9 ton/m3 for the sand and silt, and 2.0 ton/m3
for the gravel. The shear wave velocity, V,, of the
soil was estimated from the following empirical
equation using the SPT N-value (Imai & Yoshimura
1968):
Thereafter, the shear modulus, G, of the soil was
estimated based on the elastic theory of shear wave
propagation:
G = psV:
Ec: Young’s modulus of concrete (36.5 GN/m2)
A,: the total cross sectional area of concrete
E,: Young’s modulus of steel (206 GN/m2)
A,: the total cross sectional area of steel bars
Assuming c = 4000 m/s for each pile section,
density, p p , of each pile section was estimated as
follows:
Impedance, 2,of each pile section is defined as
follows:
Z = ( E p A ) / c = ( E p ( A ,+ A p ) } / c (12)
A criticism of this approximate procedure for
estimating the pile properties may be made
concerning the assumption that the bar wave
velocity is uniform along the total length of the pile.
The bar wave velocity varies according to the
changes in the properties of the pile sections.
However, in the common practice of the SIT, only
an average bar wave velocity is estimated fiom the
return traveling time of the incident wave along the
total length of the pile. Therefore, the bar wave
(9) velocity was assumed to be constant along the total
pile length in this particular case.

190
 The measured incident velocity at the pile head
was input as the boundary condition in the wave
propagation analysis of the SIT signals. As the data
sampling time interval was 0.025 ms, the pile was
divided into elements having a length of 100 mm.

3.4 Results of wave propagation analyses

Two wave propagation analyses of the SIT were
conducted with and without taking the influence of
the soil resistance into account.
The calculated pile head velocity in the later
analysis is compared with the measured one in
Figure 7. Some bumps detected in the calculated Figure 7. Comparison of the calculated and the measured pile
head velocities in the wave analysis without the influence of
velocity are caused by the changes in the cross the soil resistance.
sections of the pile, of which the amplitudes are
relatively small compared to the amplitudes of the
incident velocity and the velocity at t = 11.5 ms
caused by the reflection wave fiom the pile toe. It
can also be seen fi-om Figure 7 that the calculated
velocity caused by the reflection wave from the pile
toe is very large compared with the measured
velocity, strongly suggesting the importance of
consideration of the soil resistance even in the wave
analysis of SIT signals where amplitudes of
accelerations, velocities and displacements in the
pile are very small compared to dynamic load
testing.
The pile head velocity calculated fiom the wave
analysis with consideration of the soil resistance is
compared with the measured velocity in Figure 8. Figure 8. Comparison of the calculated and the measured pile
The calculated velocity is comparable with the trend head velocities in the wave analysis with the influence of the
of the measured velocity. The periodic oscillation soil resistance.
found in the measured velocity is not reproduced in
the calculated velocity. It should be noted that the resistance is indispensable even in the wave
wave propagation in the steel bars exposed above matching analysis of the sonic integrity test
the pile head was not taken into account in this wave signals. In order to take the influence of the soil
analysis of the SIT signals. It may be judged again resistance into account, it is preferable to conduct
fiom the wave analysis of the SIT signals that all the soil investigations such as seismic cone
piles were constructed as designed. penetration testing in parallel with the sonic
integrity test to estimate the shear wave velocity
of the soil.
4 CONCLUSIONS
If the possibility of the existence of piles not
The sonic integrity tests performed on cast-in-situ satisfying design requirements is found with SIT,
concrete piles for the foundation of New Haccho other types of tests such as borehole camera
Bridge have been described for the purposes of inspection and/or the Statnamic test and/or the
quality assessment of the constructed piles and lateral load test should be performed, depending on
obtaining integrity test signals from the piles as the position of the pile defects and the pile
reference signals which would be compared with performance required in the design, or change of
sonic integrity test signals that would be obtained design should be considered.
after a disaster in the fbture, in order to detect
occurrence of pile damages easily and quickly.
The principal conclusions from this study are as REFERENCES
follows :
Deeks, A.J. 1992. Numerical analysis of pile driving
1. All foundation piles for the New Haccho Bridge dynamics. PhD Thesis, The University of Western
Australia.
were constructed as designed. Hayashi, T. 1996. Fundamental study on qulity control
2. Consideration of the influence of the soil methods for concrete foundation piles. Dr. Dissertation,
Aichi Institute of Technology.

191
Randolph, M.F. & A.J. Deeks 1992. Dynamic and static soil
models for axial pile response. Proc. 4th Int. Con$ Appl. of
Stress-Wme Theoly to Piles: 3-14.
Matsumoto, T. & M. Takei 1991. Effects of soil plug on
behaviour of driven pipe piles, Soils & Foundations,
JSSMFE, 31(2): 14-34.
Matsui, T. & K. Oda 1996. Foundation damages of structures.
Special Issue of Soil & Faundations: 189-200.
Novak, M., T. Nogami & F. Aboul-Ella 1978. Dynamic soil
reactions for plane strain case. J. Mech. Erg. Div., ASCE,
104(EM4): 953-959.
Ohtsu, H., Y . Hatsuyama, A. Takeishi & K. Horikoshi 1997.
A study on pile foundations damaged by the 1995
Hyogoken Nambu Earthquake. Proc. Int. Con$ On
Deformation and Progressive Failure in Geomechanics
(IS-NAGOYA '97) ,Pergamon: 583-588.
Randolph, M.F. & H.A. Simons 1986. An improved soil model
for one-dimensional pile driving analysis. Proc. 3rd Int.
Con$ on Num. Methods in Ofshore Piling, Nantes: 1 - 17.
Specijkations 1990. Specification for highway bridges Part
IV: Substructures, Japan Road Association.

192
 Application of Stress-Wave Theory to Piles, Niyama 8,Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

resentation of low strain integrity testing in the time-frequency domain

J. €? Seidel
Monash Universi&, Melbourne, Vic.,Australia

ABSTRACT: The results of low strain integrity tests are usually presented either in the time doniaiii or in the
frequency domain. The information presented in either form is essentially the same, however, the two meth-
ods of display can highlight different aspects of the response. An alternative method of display in the time-
frequency domain is presented. This form of presentation derives from the mathematical concept of wavelets.
The time-frequency display does not present new information, but is an interesting alternative to the time do-
main and frequency domain, which can enhance the user’s understanding and interpretation of the pile-top re-
sponse.

1 INTRODUCTION flected from each change in impedance. The meas-

Low strain integrity testing techniques are coni-
iiioiily used to assess the construction quality of
cast-in-situ piles. In such techniques, the motion
generated at the pile-head due to a transient impact
is measured and interpreted. The motion may be
determined directly from geophone measurements.
Alternatively an accelerometer can be used, and the
signal integrated to derive the pile-head velocity.
The transient impact is typically imparted by a sinall
hand-held hammer, although alternative methods
have been used.
As the impact is generated by a relatively small
device. the resulting acoustic wave which transmits
along the pile length has low energy, and a relatively
high frequency content. It generates only low-strain
values in the pile, hence the term low-strain integrity
testing. The wave transmission and reflection is
typically interpreted using the theory of one-
dimensional wave mechanics. It is assumed that the
pile reasonably represents a rod with length signifi-
cantly greater than its diameter. This interpretation
may not be appropriate near the pile head where sig-
nificant dispersion of the wave from the impact oc-
curs.
In one-dimensional wave mechanics, the pile
impedance, Z, is defined as the product of mass den-
sity, p, pile area. A, and the speed of wave transmis-
sion, c. The impedance of a pile will be affected by
the cross-sectional area of the pile and by the quality
and composition of the pile material at any section.
If a pile with non-uniform impedance is impacted
at the pile head, the induced stress wave will be re-
ured pile response is affected at the time each re-
flection returns to the pile-head. The nature and
location of impedance changes may therefore be in-
terpreted from the pile-head velocity response. Be-
cause of the relationship between the physical prop-
erties of the concrete or grout and its impedance, the
existence of anomalies in the pile section may be in-
ferred. Two general caveats apply
1. That individual pile responses are judged
against a wide sample, or tlie complete popu-
lation of piles;
2. That the inferred anomalies be confirmed by
physical inspection.
It should be mentioned that energy is also radi-
ated from the pile shaft into the surrounding soil.
The effect of the soil is to absorb energy, and there-
fore to dampen the measured response. The greater
the soil stiffness, the greater will be the damping ef-
fect. This limits the penetration of the input wave
into the pile, and the effective length of shaft which
can be verified (Turner, 1997). In order to compen-
sate for the degradation of the input wave. amplifi-
cation routines are sometimes applied to the re-
sponse to assist with identification of features which
would otherwise be hidden.
2 CLASSICAL METHODS OF DISPLAY
The pile-head responses are typically displayed h
one of two forms - in the time domain or the fre-

193
(1 iic nc 1 do ilia i n . Gen era11y , proprietary systems d i s-
plal. data in only one or tlie other of these forms.
Testers tend to form their judgements based only on
that particular view of the data. A third form of dis-
pla) of the data will be presented in the latter sec-
tions of this paper.
2.1 Time doiiiciin
’fhe time domain display plots pile-top velocity
as the vertical axis against elapsed time as the hori-
zontal axis (see Figure 1). Records are interpreted
particularly by focussing on the response during the
first passage of tlie input wave down and then up the
pile (i.e. to a time of 2L/c, where L is the pile length
and c is the speed of the compression wave in the
pilc). This is not to say that the response beyond
2L/c is not used to assist in interpretation.
Figure 1. Typical response in time domain for pile with dis-
Crete impedance reduction
In Figure 1, the initial impact is indicated. Sub-
sequent features are all reflections from the p,ile
shal’t and the surrounding soil. From one diinen-
sional mave mechanics theory, it can be established
that a discrete reduction in pile impedance will gen-
erate a tensile reflection from the con~n~encement of
the reduction. A compression reflection will follow
where the pile returns to its normal dimension. This
can be seen as a characteristic positivehegative cy-
cle i n the time domain. The amplitude of this cycle
will be a fhction of the relative reduction in pile
impedance; of the length of the impedance reduc-
tion, and the wavelength of the input wave.
By tlie same logic, a discrete increase in pile im-
pedance will cause a negative/positive cycle in the
response viewed in the time domain.
The toe response is also a key feature which is
always sought in the time domain. If the toe re-
sponse can be identified, it is possible to establish
whether or not the pile is free of acoustic reflectors.
Assuming that energy reaches the pile toe dif-
ferent reflections can be expected. If the pile toe ap-
proximates a “free-end”, the reflected wave would
be tensile. and velocity increase will be detected at
2L/c. A fixed condition will generate a compressive
reflection, and a velocity decrease. Depending on
the stiffness of the toe condition, therefore, a spec-
trum of responses from large tensile to large com-
pressive could theoretically be generated. In prac-
tice. at such low strain levels, a tensile response is
194
more common. The magnitude of tlie response will
depend not only on the end-condition, but also on
tlie attenuation of the incident and reflected waves.
The toe response will be a half-cycle rather than the
full cycle resulting from discrete impedance reduc-
tions along the shaft.
If an amplification routine is applied to the time
domain signal, it is not possible to make an absolute
determination on the toe condition from the ampli-
fied response. However, the toe condition may be
judged relative to like piles on a project fi-om a com-
parison of responses as long as the piles are tested in
the same manner and transformed using the same
amplification level.
Although the indications of the impedance reduc-
tion and the toe response shown in Figure 1 are
clear, it is not always possible to locate these fea-
tures with sufficient accuracy. It may, for instance
be unclear whether a particular response emanates
from the pile toe or a defect just above the toe. This
may have a critical effect on the pile evaluation.
Part of this ambiguity arises from uncertainty about
the true speed of wave transmission. Values of 3600
m/s to 4000 m/s (and above) are typical for grout
and concrete. Stated accuracies in length determi-
nation of 10% are therefore commonly quoted.
There is another reason which can cause difficul-
ties in interpretation. When the hammer strikes the
pile-top, the energy that enters the pile is composed
of a spectrum of individual frequencies. The com-
posite of these frequencies results in the summed in-
put waveform which is indicated by the measured
velocity response.
Each of these individual frequency components
travels down the pile. Elasticity theory actually
shows that the transmission speeds for the different
frequencies are very marginally different. More im-
portantly. because the high frequency components
contain relatively little energy, these components are
attenuated faster than the low frequency components
which contain the bulk of the input energy.
The consequence is that the response received
from the pile has a different frequency spectrum
than the incident wave, being more dominated by the
low frequency components. The composite of the
individual frequency reflections will be a different
shape to the incident wave, complicating interpreta-
tion.
A shortcoming of the time domain method there-
fore is that it does not allow discrimination of the
frequency response.
2.2 Freqirency domain
As noted above, the hammer impact generates a
spectrum of waves with different frequencies and
amplitudes or energy contents. In order to view the
response of the pile in the frequency domain, it is
customary to measure not only the pile-head veloc-
 ity. but also the pile-top force generated by the
hammer. Both the force and velocity responses are
deconstructed to establish the frequency spectrum of
the hammer input and the pile/soil response respec-
tively. Depending primarily on the pile geometry,
the pile Lvill tend to resonate at particular frequen-
cies when the reflection from the pile toe or an in-
termediate feature is "in phase" with the input wave.
The result will be a heightened velocity response at
such frequencies. The ratio of peak velocity re-
sponse to peak force (called the mobility) at these
part i c 111ar i 11 put frequencies w i 11 be a inay i mum.
1 he frequency domain display plots pile-top mo-
~~
bility as the vertical axis against applied frequency
as the horizontal axis (see Figure 3).
Figure 2. Typical response in frequency domain for pile with
discrere impedance increase (after Ellway, 1987).
These records are primarily interpreted by estab-
lishing the frequencies at which local maxima of the
mobility occur. The frequency difference, Af, be-
tween successive maxima (harmonics) correspond to
intervals of c/2L. As noted previously, given the
uncertainties in establishing the exact value of the
wavespeed, the pile length, or the length to a feature
within the pile can be established to within 10% -
i.e. the same limitation as for the time domain analy-
sis.
For the same reasons discussed earlier about the
greater persistence of the low frequency compo-
nents, the overall pile length is most likely to be evi-
dent i n the low frequency area of the mobility plot,
whereas features higher in the pile will be more evi-
dent in the high frequency area (see Figure 2).
Another feature of the mobility plot is the pile-
head dynamic stiffness, determined as the inverse of
the slope of the initial (low frequency) portion. At
very low frequencies the inertial effects are said to
be negligible. Hence, the stiffness is suggested
(Davis and Dunn, 1974) to correlate to the initial
elastic loading stiffiiess of a static load test. This
suggestion is not universally accepted. However,
j List as conclusions may be drawn with respect to the
comparative toe responses of a large population of
piles tested using time domain analysis, a compara-
tively low pile-head dynamic stiffness may justify
further investigation.
195
As discussed earlier. this method of display is
dependent on the determination of the frequency
spectrum of both the velocity and force records.
This is normally done by a technique called Fast
Fourier Transformation (FFT). It is implicit in this
technique that the recorded transient signal is actu-
ally part of an infinitely recurring response. I n this
way, the signal can be transformed into an equiva-
lent set of infinite sinusoidal functions, each with a
defined amplitude and phase shift.
As the analysis is then based on infinite series,
the frequency response analysis does not contain any
temporal (time) information. To highlight this, Fig-
ure 3 shows two signals - the first consists of the su-
perposition of two frequencies (Sin1Ot and SidOt)
and the second contains these two frequencies in
separate succession. Although the resulting wave
traces are quite different, the FFT of both signals is
identical. This is a limitation of this method of
analysis.
Figure 3 . Comparison of FFTs for two distinctly different
wave forms (after I<umar and Fo~ifoula-Georgiou)
An alternative method of display, which simulta-
neously provides information in the time domain for
the individual component frequencies is presented.
3 WAVELET ANALYSIS
Fourier series are evaluated by the use of mathe-
matical functions called integral transforms. These
transforms are the result of multiplying the original
function with a kernel function integrated from -CO to
+CO. The kernel for a Fourier transform is
(3x1"'.exp(-ixy>.
An alternative mathematical technique, known as
Wavelet analysis, is based on a similar technique of
integral transformation. Wavelet analysis is being
applied to a vast array of applications in science and
engineering. Kumar and Foufoula-Georgiou ( 1994)
provide a good introduction to the use of Wavelets
in the related field of Geophysics.
The multiplying kernel function in this case is
called the mother wavelet, and the limits of integra-
tion are finite values at or slightly beyond the meas-
ured response. That is there is no implied assump-
tion of’the signal being an infinite series. Transient A. is a scale parameter which has the effect of di-
s i p a l s can be analyzed. Lee and Yaniamoto (1994) lating ( 0 1 ) or contracting (3,<1) the wavelet the
claim that every application using FFT can be for- function ( ~ ( f )
mulated using wavelets to provide more localized t is a localization parameter which translates the
temporal (or spatial) and frequency information. It wavelet so that the function j [ r ) can be analyzed
is this ability of wavelets to obtain frequency content around the point (or time), t.
of a process locally in time (time-frequency local- In order to preserve unit energy at all frequencies,
i za t i o n ) that is particular 1y import ant . and therefore not bias the integrand, the amplitude of
There are many different wavelet transforms the wavelet must be adjusted in accordance with
available - that is, it is not a unique function. A Equation 1 as the mother wavelet is dilated or con-
simple one-dimensional wavelet called the Mexican tracted. This concept is shown for the Mexican hat
hat wavelet, is shown in Figure 4. However. the wavelet for the cases of h=l, h<l and ?,>I in Figure
selection of wavelets is not arbitrary. A wavelet 5.
111us t sat i sfy t liree conditions :
1 . I t inust have unit energy
2 . I t must provide compact support. or suffi-
ciently fast decay so that it satisfies the re-
quirement of space (or time) localization.
3. It must have a zero mean - that is the integral
of the wavelet function from -cc to +cc is zero.
This is called the admissibility condition, and
ensures that there are both positive and nega-
tive components to the mother wavelet.
The Mexican hat wavelet in Figure 4 can be seen
to satisfy these conditions.

Figure 5. Dilation and contraction of the mother wavelet

3.2 Time locnlizntion

The temporal variation of frequency content is de-
termined by progressive translation of the wavelet
across the transient response using the parameter t.
For each value of h, the value of is varied so that
the wavelet passes across the entire signal, and the
integrand, Wf(h,t)determined for each combination
Figure 4. The Mexican hat wavelet of h and t. This process is sliown scheniatically in
Figure 6 for a single frequency of wavelet at 3 arbi-
trary positions.
The interested reader is referred to Kurnar and
Foufoula (1 994) and the extensive list of references
contained in that paper for a formal description of
the mathematical requirements that apply to wave-
lets.
However, the general procedures which allow
wavelets to establish frequency content and time lo-
calization are described hereafter:

3.1 Estahlishing frequency content

The general form of a wavelet transform of a func- Figure 6. Wavelet translation for time-frequency localization.
tionf(t) with finite energy is as follows:
h>0
Lastly. it is noted that when the parameters h and
I,Jq(h.t) = I,flZl) yJ&) n‘tf
t in the wavelet transform take on continuous values,
it is called a continuous wavelet transform. For
= S.j(u) \+,f([zr-t]1.2)ClZl (’) practical applications, it is necessary for the wavelFtt
\\.liere: and the signal to be discretized so that the integra-

196
 tion can be performed. This is then referred to as a
discrete wavelet transform. For the discrete trans-
form, the time steps are expanded in relation to the
dilation of the wavelet.
4 WAVELET ANALYSIS OF LOW STRAIN
INTEGRITY TEST SIGNALS
4.1 Previozu investigations
The application of wavelet theory to low-strain in-
tegrity testing has been pioneered by researchers at
Napier University, Edinburgh (Addison et al., 1997;
Adclison and Watson, 1997, Watson et al., 1998).
In Addison et al. (1997), the authors present fi-
nite-element generated responses for imaginary piles
10m in length, with and without a 20% impedance
increase at mid-length. They indicate that the 20%
iinpedance cliange is difficult to spot in the time
domain, whereas the defect can be observed clearly
in tile wavelet transformation. The author believes
that in practice a 20% impedance change would be
clearly discernible in the time domain if a toe re-
sponse as large as that indicated were obtained. It is
assumed that the poor reproduction of the paper does
not allow their contention that the change is clearly
evident in the wavelet plot to be confirmed.
At the time of these papers, Addison and his co-
workers appeared to favour use of the Daubechies
wavelet, possibly because its shape provides high
resolution. The resulting wavelet transformation has
a somewhat chaotic appearance, which is somewhat
disturbing to the uninitiated. Their data is typically
presented as one-dimensional transforms plotted
against time, although time-frequency fields are also
presented. It appears that at that time, a preference
for coiitinuous or discrete wavelet transformations
was not clear. The discrete wavelet approach has
significant computational advantages and does not
senerate redundant data. By comparison, the con-
tinuous form does not impose restrictions on dilation
incizments, but generates significant data redun-
dancy, and hence takes longer to perform.
The team at Napier University gives every indi-
cation of continuing with their interesting research in
this field.
4.2 Current Investigations
The author has also experimented with applica-
tion of wavelet analysis, exclusively to actual data
collected from field measurements.
The similarity between the wavelet technique and
cross-correlation technique, in terms of passing a de-
fined wave across the signal to be analyzed are evi-
deni. In wavelet analysis, however, the waveform is
dilated or contracted and an appropriate scaling fac-
tor applied to ensure constant energy.
197
Depending on the purpose of the time-frequency
analysis, it may not be necessary to impose the strict
mathematical requirements of wavelet analysis.
If the purpose is only to evaluate the relative
time-frequency behaviour, it may be possible, or
even preferable to apply a hybrid wavelet/cross-
correlation technique.
In cross-correlation techniques, the correlating
function should have a similar characteristic to the
feature to be extracted from the signal. Such a
choice will lead to the highest correlation values.
For low-strain pile integrity testing, the input wave-
form is reasonably well approximated by a half-
cycle of a sine wave function.
It is evidently that a half-cycle sine wave does not
meet all the criteria required for wavelets.
1. The energy may not be unity. However, if an
appropriate amplitude of the “mother” sine
wave is chosen, a unit energy could result.
Furthermore, if the resulting integrands are all
normalized prior to plotting, the energy in the
mother sine wave is irrelevant. It is only nec-
essary to ensure that the sine wave is appro-
priately scaled during contraction or dilation
so that energy remains constant under all
transformations.
2. The half-cycle sine wave does provide com-
pact support and fast decay, thus satisfying
the requirement of time localization.
3. As it has only positive components, the half-
cycle sine wave does not meet the criterion of
a zero integral in the interval -us to +us. The
author does not believe this to be critical for
the purpose of analysis of the low-strain sig-
nals.
As suggested before, the advantage of applying
the dilational and translation techniques of wavelet
analysis to a sine wave form rather than a true
wavelet, is in providing an optimum opportunity for
correlation with the measured signal. It is believed
that the fact that all the criteria for wavelets are not
satisfied is not relevant for a relative assessment of
time-frequency localization.
If, however, the purpose of the analysis is to al-
low deconstruction, modification and reconstruction
of the signal, as wavelet analysis allows, the use of
the sine-wave form could not be assumed to satisfy
these requirements.
5 RESULTS
In the following discussions, the term wavelet analy-
sis is used advisedly, taking into account the differ-
ences with classical wavelet analysis already re-
ferred to. The outcomes of this study, W P W
presented in one of two ways.
5.1 Time-doinciin presentation
As the result of time-stepping a wavelet of a single
frequency through the original signal. A modified
response curve, being the normalized integrand val-
ues of the signal and the wavelet results. The results
are presented in the time domain, and have a similar
form to the original data. This is shown in Figure 7,
where a set of data, and the wavelet-modified data
are compared. It can be seen that the original data
has been smoothened by the wavelet transforrna-
tion.. because the frequencj of the chosen wavelet
was lower than the high frequency perturbations in
the data. It is, if used in this manner, a useful filter-
ing technique. As always, filtering of data must be
undertaken with care to eiisure that significant fea-
tures are not removed. This algorithm has been in-
corporated in coinniercially available software
(Likins and Rausclie, 2000).
Figtire 7. Comparison of original signal arid modified signal
after wavelet anaiysis using a single fi-equency
Figure S. Comparison of time-domain data and \vavelet analysis in tlie time-frequency domain for pile with inid-
length defect
198
5.2 Time-fi-eqiiencydomciin preseimlion
As the result of time-stepping using a spectrum of
wavelet frequencies varying from approxiinately 300
Hz to 15 kHz. The wavelet integrands are normal-
ized to a peak maximum value of 1 for PVJ(h,r). The
array of results are then presented as a contour dia-
gram of the wavelet integrands with frequency as the
vertical axis and time as the horizontal axis. Exam-
ples of this presentation, also compared with the
original data in the time domain are shown in Fig-
ures 8, 9 and 10.
5.2.1 Pile with dcimnge
Figure 8 shows the time-frequency response plot
for a pile with a significant defect close to tlie pile
mid-length. The horizontal axis is time, at the same
scale as the time-domain plot also shown. The ver-
tical axis is frequency in Hz, ranging from approxi-
mately 300 Hz‘ to 11000 Hz. This is a descending
logarithmic scale - low frequency is at the top of the
plot and high frequency at the base.
The plot shows contours of equal PVflh,t). Areas
between the contour lines are shaded so that largest
values of PV‘h,t) are darkest. The contoured values
vary from +l to -1; the corresponding shades of
grey vary from black to white.
The plot shows that at approximately 0.92ms,
there is a “ridge” of maximum values generated
across all frequencies. This corresponds to the im-
pact. which can be seen directly underneath in the
time domain plot. As discussed previously, the im-
pact generates a spectrum of frequencies which enter
the pile. The peak value for Pvflh,t) occiirs at a peak
frequency of about 5500 Hz, corresponding to a
wavelength of approximately 0.7m. Another ridge
can be seen at approximately 3.00ms. This corre-
 Figure 9. Comparison of time-domain data and wavelet analysis in the time-frequency domain for
0.9kg (2 pound) haininer

Figure 10. Comparison of time-domain data and wavelet analysis in the time-frequency domain for
53kg (12 pound) hammer

sponds to the reflection from the pile toe. In this
case, the peak frequency occurs at approximately
3600 Hz. This downward frequency shift is charac-
teristic of this type of plot. As noted previously, the
higli frequency components of the incident wave are
iiiore quicltly absorbed, and the reflected wave is
expected to be more dominated by low frequency
components.
The 2L/c time is determined by measuring from
ridge to ridge, giving a wave return time of 2.1 1ms.
Another feature can be seen emanating from the
base corresponding to the high frequency compo-
nents at about 1.9ms. This is the damage indicator.
As stated by Ellway (1987), structural defects in the
pile preferentially reflect high frequency compo-
nents of the incident wave. Low frequency compo-
nents pass through with little reflection. The dam-
age indication cannot be seen in the low-frequency
range of the plot.
It can be seen that pile damage indications will be

199
rcstricted to the high frequency range, whereas soil
resi btance, which generates damping responses will
dominate the low frequency range.
5.2.2 Pile slruck with difji;r.ent haiiuners
Figures 9 and 10 show both time domain and
time-frequency domain graphs for the same pile
struck lvitli a 0.9kg (2 pound) and 5.5 kg (12 pound)
han: mer, respectively. A comparison of the two
tests is interesting. The time domain results are first
coin pare d.
The pile was initially tested with the lighter
liammer. Figure 9 indicates the possibility of some
impedance reduction above the pile toe. The reflec-
tions are not strong, and in the time-domain there is
overlap between this “damage” response and the
later response presumed to be the pile toe. This
makes it more difficult to make a definitive determi-
nation of the pile wavespeed.
In order to confirm (or otherwise) tlie damage, the
pile was struck with the larger hammer so that more
energy would reach the toe. Figure 10 shows, how-
evei.. that there is no clear reflection from above the
toe response, and fk-thermore, the toe response has a
very long rise time. Again, determination of the pile
wavespeed is unclear, and possibly in conflict with
the data from the smaller hammer, if tlie rise-to-rise
method is adopted.
The time-frequency responses are now examined.
First. it is clear that the input frequency of the signal
generated by the larger hammer (about 300 Hz) is
significantly lower than that generated by the
smaller hammer (about 800 Hz). The lower fre-
quency content of the returning wave is also evident,
particularly for tlie smaller hammer which is subject
to greater degradation.
As noted before, pile defects are more likely to be
detected with high frequency waves. The ridge at
about 7.3ms in the high frequency range of Figure 9
for the smaller hammer is the damage indication.
No such ridge can be seen in Figure 10, because of
the limited high frequency content of this signal.
The pile toe is seen clearly and definitively in
both plots, and the 2L/c time can be firmly estab-
lished as 8.1 ms for both tests, thus overcoming the
confusion from the time domain plots.
6 CONCLUSION
Lo\\ -strain pile integrity testing data is normally
plotted in either the time-domain or the frequency
domain. The information provided by both sets of
data is essentially the same, however, each method
highlights different features of the response. How-
ever, both methods of data presentation have limita-
tions. As discussed, frequency information is absent
from the time domain plots, and the frequency re-
sponse contains no temporal information.
200
Wavelet analysis is a mathematical technique
which allows the data to be viewed in the time-
frequency domain. Little research or investigation
of the application of wavelets to pile integrity testing
has been undertaken to date. However, it appears
that this approach may provide a more complete w y!
of viewing the data. Significant further work 1s
needed to refine our understanding of the application
of wavelets to low strain test data, and to its inter-
pretation.
The author has undertaken “hybrid” wavelet
analysis using some of the principles of wavelet ap-
plication using a half-cycle sine wave function. The
time-frequency responses presented provide deeper
insight into the pile-head response, and demonstrate
some physical principles previously established by
others. At this stage, it is a qualitatively useful
method of presenting low-strain data. It has poten-
tial to provide useful quantitative analysis with fur-
ther research.
REFERENCES
Addison, P.S., Sibbald, A. and Watson, J.N. (1997). Wavelet
analysis : a matliematical microscope with civil engineering
applications. Insight Vol 39 No. 7, July, 1997 : 493 - 497.
Addison, P.S. and Watson, J.N. (1997). Wavelet analysis for
low strain integrity testing of foundation piles. 5“’ Intl.
Conf. On Inspection, Appraisal, Repairs, Maintenance of
Buildings and Structures. 15-16 Mat, 1997, Singapore.
Davis, A.G. and D i m , C.S. (1974). From theory to field expe-
rience with the non-destructive vibration testing of piles.
Proc. Institution of Civil Engineers, Vol. 57, Part 2, De-
cember : 57 1-593
Ellway, I<. (1987). Practical guidance on the use of integrity
tests for the quality control of cast-in-situ piles. Ground
Engineering, Vol. 20, No. 7, October : 5-13,
I<umar, P. and Foufoula-Georgiou, E. (I 994). Wavelet analy-
sis in Geophysics : an introduction. Wavelets in Geophys-
ics, Foufoula-Georgiou and Kunar (eds.) Academic Press
Inc. ISBN 0-12-262850-0 : 1-43
Lee, D.T.L. and Yamamoto, A. (1994). Wavelet analysis :
theory and applications. Hewlett-Packard Journal, Decem-
ber, 1994 : 44 - 54.
Likins, G.E.L. and Rausche, F. (2000). Recent advances and
proper use of PDI low strain pile integrity testing. 6“’ Intl.
Conf. on the application of stress-wave theory to piles., Sao
Paulo, Brazil, Sept., 2000 (not yet published).
Turner, M.J. (1997). Integrity testing in piling practice.
CIRIA Report 144. 1997 0 CIRIA, London ISBN OS6017
473 5
Watson, J.N., Addison, P.S. and Sibbald, A. (199s). The
wavelet transform for pile integrity testing. Proc. 12‘” En-
gineering Mechanics Conf., La Jolla, California, May 17-
20, 199s. Mttrakami, H. and Luco, J.E. (eds). 0 ASCE. :
546 - 549.
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 1503

First experiences in the application of the stress wave theory to foundations

in Uruguay
A.Gutikrrez, L.Abreu, Ch. Hoffmann & D. Hasard
Laborutorio de Control de Culidud de Fundaciones, Instituto de Estructurus y Transporte, Facultad de Ingenieriu,
Universidud de La RepL2dicu,Montevideo, Uruguay

ABSTRACT: Since 1995 the LCCF (Foundations Quality Control Laboratory) of the Engineering School has
incorporated new advances in addition to the traditional foundation quality control technology combining soil
studies, pile driving diagrams, the extraction of core drills for concrete sample testing, static load test with
sonic integrity test, and lately the dynamic load test.
Here, some real cases are presented, in which sonic integrity testing (SIT) has been used successfully for dif-
ferent purposes .

1 BRIDGE OVER YERBAL STREAM 2 720 ANCAP TANK - LA TEJA REFINERY

This bridge suffered the effects of a serious flood-
ing; its foundations were exposed, and a part of its
superstructure was damaged. During the reconstruc-
tion of the bridge, one of the support pillars fell over
onto an adjacent pillar, damaging it, in such a way
that it also needed to be repaired. In order to deter-
mine the effects of this on the piles, SIT was used on
the possibly damaged piles and on the piles of unaf-
fected adjacent pillars that were not affected. These
tests thus differ from the usual pile test in that the
caps were already built. It was decided to place the
sensor and run the test on the upper surface of each
pile cap directly above the pile to be tested. Figure 1
shows the signal obtained by placing the sensor and
exiting cavities made especially below the pile cap
in order to help obtain clearer signals, that were used
in the data interpretation.
Figure 1. Signal registration with the sensor placed
o n the pile cap
Sonic integrity testing was used on a floating tank
roof in La Teja refinery of the National Administra-
tion of Fuels, Alcohol, and Portland (ANCAP). The
case deals with 44 pile driven foundations, all of ap-
proximately 10m in length, located on a grid, cov-
ering a circle of approximately 16m in diameter. The
decision to use a foundation of piles was based on
the problems generated by differential settlements in
these types of tanks. The problem arises when the
lateral wall strains impede the descent of the floating
roof which gets stuck, causing firther deformations
that can lead to the cracking of the tank walls and
extensive fuel loss with the consequent economic
and ecological problems.
Although specific ground studies haven’t been done
for this site, information about the soil properties
was available fkom previous studies carried out for
surrounding tanks and sites as a reference. Accord-
ing to this information, there is a heterogeneous
filling of essentially granulated soil, with low
strength zones and a water table at a depth of about
2m. The piling company supplied the pile driving
diagram for the first piles that shows there is a
higher strength layer between the depths of 5.0 and
6.0 m. The strength then dropped when the pile
reached 6 . 0 ~until the depth of approximately
lO.Om, where the strength increased again. See Fig-
ure 2.

20 1
 Due to this experience, in another one of the 6 de-
fective piles, pile 42, a stepped static load was ap-
plied, to see if the defect could be sealed with
preloading. A maximum load of 150 kN was reached
reacting against the piledriver, an integrity test was
carried out after each load step. In this case no im-
portant modification of the signal was seen. In the
attempt to reach higher loads the piledriver lost sta-
bility, which is the reason why an important part of
the strength was not achieved. Figure 4 shows the
signal registration before and after the application of
the maximum load reached, as well as the signal af-
Figure 2. Pile 1 driving diagram ter the construction of the adjacent pile, pile 42B. In
this case, the same type of signal obtained earlier
was maintained, with insignificant changes, showing
In this case it is seen, in agreement with the integrity a narrowed section in the pile.
signals, that 6 piles broke off at a depth of approxi-
mately 5.0 m. The piling company said there were
no signs of problems during the construction. In
addition, there were detected 5 more piles with a
negative reflection at a depth between 5.0 m and
7.0 m, but the signals did not show an actual break.
The remaining 33 piles gave very similar and normal
signals, fkom which the average signal was taken.
The average is shown for the pile 1 in Figure 3 , as a
reference signal.
In one of the 5 piles with a negative reflection, pile
1, the integrity testing was repeated after a reinforc-
ing pile was made, pile lB, alongside the first one.
In this case it was found that the new signal did not
have the negative reflection, and it was also com-
patible with the average signal of the piles without
integrity problems. It was inferred to be a minor
problem of local impedance reduction at the depth of
5 . 8 a most likely caused by a minor fissure, that
was sealed when the adjacent pile was placed.

Figure 4. Sequence of signals obtained on pile 42

before and after the application of the static load and
after the adjacent pile construction, 42B.

In fact, the high number of piles having a defect at

Figure 3 . Signal registration made on pile 1 before
after the construct~onofthe adjacent pile 1 ~in,
comparison with the average site signal.
the same depth (14%) in a relatively reduced area of
ground led to the conclusion that a possible system-
atic error in the piles construction may have been
committed*After contacts with the piling company
the conclusion was reached that the piles were rein-
forced only half way down their length (5m). This

202
way the company could use 12m bars easily avail-
able on the market by just cutting them in half with-
out the necessity of connections. Where the bars
ended, they were closed off by bending the bars,
forming a “nest” that, because of the quantity of
bars, left very little space for a proper concrete pas-
sage. It is important to keep in mind that the con-
crete used for these types of piles is a dry concrete
that contains thick aggregates. As mentioned above,
the zone where the narrow section appeared,
corresponded to the passage fiom a higher lateral
ground kiction to another with less resistance. For
this reason, if during the concreting the case was not
carefully raised, and was done abruptly, the pile can
be cut off with the invasion of saturated ground.
The high number of defective piles found at this
depth can then be explained by the difficulties re-
sulting iiom the passage of the thick aggregate con-
crete through the folded bars, and also due to the
transition from resistant soil to a lower lateral iiic-
tion during the case raising.
The foundation was reinforced by building several
new piles, following the proposal from the planner,
all made with continuous reinforcing bars up to the
pile toe. See Figure 5.
0 To evaluate the possible damages on pile number 11,
0 0 0
0 0 0 0 O0
0 0 0
0 0
0 0 0 0
0 0 0 0 OO
O0
0 0 0 0
42 0
42B 0
References:
Damaged pile
Reinforcing pile
0 Pile without problems
Figure 5. Arrangement of the foundation piles of
tank 720.
203
3 GRAIN UNLOADING WHARF EXTENSION -
COLONIA
During the piling phase of a grain unloading wharf
in the Nueva Palmira port, a loaded, 20,000 ton
barge accidentally hit a group of piles constructed
only 10 days before, reacting elastically. SIT studies
were done to see if the piles involved had suffered
structural damage.
In agreement with the information given by the con-
tractor, the piles constructed were drilled piles made
of reinforced concrete, all about 27m in length, 0.8m
in diameter, and fixed in about 4m of a high resis-
tance ground layer, belonging to the Fray Bentos
formation. The piles have steel casings in their up-
per section of 20m.
The signals obtained from the integrity testing pre-
sented a clear hammer impact, and a good reflection
of the pile toes. The lengths obtained corresponded
reasonably with the ones given by the construction
company. All signals indicated a lateral fiiction in-
crease below the depth of 20m, that corresponded
generally with the end of the steel casing and the di-
rect contact between the ground and the concrete.
None of the signals showed a negative polarity re-
flection, which would suggest the presence of any
breaks or narrowings in the shaft of the studied piles.
its signal was compared with the average signal
taken from the adjacent and unaffected piles (9,10,
and 12).
As shown in figure 6, the signal of pile 11 and the
average signal present very similar behaviors, giving
reason to affim that the pile did not suffer important
damages.
Figure 6. The signal of affected pile number 11, and
the comparison with the average signal
The tests allowed the integrity control of the affected
piles and the comparison of their signals with the
signals of the nearest piles. In this way the theory of
structural damage to the piles could be rejected.
4 MACHINE FOUNDATION IN THE TECHNO-
LOGICAL LABORATORY OF URUGUAY
(LATU)
shortly. Although there are still no official standards
for the tests, many state organizations like the
Transportation and Public Works Ministry (MTOP)
and the Petroleum National Company (ANCAP), re-
quire these controls in the sites under their admini-
stration.

As there were no plans or any information about this

site, SIT tests were done to determine the normative
conditions and thickness of the existing foundation
blocks of 2 dynamic test machines. On both foun-
dation blocks, the approximate dimensions were
1.5m by 2.0m. Five signal registrations were taken, a
central one, and four peripherals. Because of the
contractor’s interest in estimating the block’s default
thickness, and the inability to test the concrete, it
was decided to use a wave propagation velocity of
3,800 m/s to test the concrete.
All the signals presented similar behavior. An aver-
age signal was defined for each block and the thick-
ness was determined in each case. In one of the
blocks the thickness was estimated at 2.0 m k 0.2 m
and in the other at 1.6 m It 0.2 m. Figure 7 shows
the average signal of block number 2.

Figure 7 Average signal of foundation block

number 2.

5 DYNAMIC LOAD TEST (DLT)

After many difficulties in fine tuning the DLT, the

first investigation tests were done recently in piles
built in the Libertad - Dolores formation clays. The
short term objective is to incorporate this test into
customary foundation controls.

6 FUTURE PROSPECTS

The foundation quality control and the use of sonic

methods to determine the integrity and the bearing
capacity of piles, promises to become a reality

204
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Detection and prevention of anomalies for augercast piling

6.Likins & G.Piscsalko

Pile Dynamics Incorporated, Cleveland, Ohio, USA
E Rausche & C. M. Morgano
Goble Rausche Likins and Associates Incorporated, Cleveland, Ohio, USA

ABSTRACT: On many sites, augercast piles have significant advantages including cost and low vibration.
They also have some uncertainty due to the construction methods. Visual inspection during grouting is often
difficult. Static test piles have been used to confirm installation procedures and soil conditions, but are
restricted to only a small sample of piles. Furthermore, load test piles receive special attention during the
installation compared with typical production piling.

T o verify installation and increase confidence, Non-Destructive Evaluation (NDE) methods like low strain
Pile Integrity Testing are often specified for some percentage of production piles. Because such NDE
methods require testing the pile after the grout or concrete has hardened, the repair or replacement can be
expensive if problems are found. However, preventing defects through thorough monitoring during
installation is preferred over Pile Integrity Testing. Augercast piles can be automatically monitored for grout
volume pumped versus depth. With this information available during installation, if a low grout volume is
measured for any depth increment, the pile can be repaired immediately while the grout is still fluid.
Installation monitoring reduces the need for subsequent NDE tests.

While installation monitoring and NDE tests assess pile structural integrity, the bearing capacity cannot be
assessed by these methods. Bearing capacity of augercast piles can be evaluated economically by high strain
Dynamic Pile Testing methods used to test driven piles if a suitable drop weight is available.

Keywords: augercast pile, installation monitoring, integrity testing, dynamic pile testing

1 INTRODUCTION
The quality of augercast piles often depends on the
contractor’s skill. The most critical is the control of
auger withdrawal during grout placement (Roberts
1998). However, for a visual inspector, it is difficult
to accurately assess grout volume and auger depth
simultaneously. Automated measurement systems
can accurately monitor volume pumped as a fimction
of depth (Likins 1998). This information can guide
the operator during installation to produce a shaft
exceeding any minimum requirements.
Low Strain Pile Integrity Testing of selected piles
verifies the installation monitoring effectiveness.
This requires striking the pile top with a small
hammer and measuring both input and reflections
from pile non-uniformities or the pile end (Rausche
1992). Pile Integrity Testing is applied to any pile or
even all piles after the grout is hardened. There are
some limitations for long or jointed piles.
Bearing capacity can be evaluated on any pile on
site using High Strain Dynamic Pile Testing (Likins
2000a). Test Piles can be selected after installation
based on their site location (to assess soil variations),
on specific pile installation records, or at random.
The capability of installation monitoring and
subsequent low and high strain dynamic testing
allows the engineer to completely investigate the
foundation.
2 CURRENT INSTALLATION PROBLEMS
MONITORING AUGERED PILES
Visual inspection counts pump strokes as a fhction
of estimated auger depth to estimate volume versus

205
depth. Due to a large volume of information, there
are many possibilities for deficiencies or errors in
recording all the necessary information. Further, the
volume per pump stroke can be easily vary by as
much as 20% (Likins et al. 1998). Further, pump
strokes are usually manually counted per five foot
interval; this interval has marginal precision and
even the visual determination of the depth may have
one foot errors. Alternative methods of more
accurate inspection are desirable.
2. I Automated monitoring of augercast piles during
Installation
Automated monitoring (Likins et al. 1998) of grout
volume pumped versus auger position provides
information to the piling crew which can guide the
operator during auger withdrawal. The schematic in
Figure 1 shows the overall PIR-A configuration. A
small Control Unit conveniently located for the
crane operator acquires and processes all measured
data. The PIR-A depth monitor has a self-retracting
cable attached to the auger gear box. As the auger
advances or withdraws, the depth is measured by a
rotary encoder tracking the cable. A Magnetic Flow
Meter accurately measures pumped grout volume to
an accuracy within 2%. A Grout Pressure
Transducer installed in the grout line (usually near
the Flow Meter) continuously measures pressure in
Figure 1. PIR-A schematic.
206
the grout line. During augering, the auger torque is
measured.
In practice, after the operator inputs the pile name,
the PIR-A Control Unit handles virtually everything
else. During drilling, the operator observes the
auger's current depth and torque. Higher torque
(just below the crane stall torque) makes drilling
more efficient and reduces the spoils brought to the
surface. Geotechnically, the engineer can assess the
torque at auger rehsal to distinguish strong soils or
low torque.
Figure 2. P R display during grouting phase.
At the pile design depth, the operator initiates
auger withdrawal and grout injection. During
grouting, the measured grout volume pumped per
unit depth is displayed graphically as in Figure 2 for
pile 162. The operator can adjust the auger
withdrawal rate to keep the grout above the target
line. This can also be guided by keeping the
grouthnc display slightly above the depthhnc
display. If the grout per increment is low, the pile
can be re-augered immediately through this
increment and the pile re-grouted to eliminate
defects in the pile and subsequent remedial
measures. The operator presses "return" when grout
is observed at the surface and "done" when the pile
is completed.
Upon completion of the pile, monitoring results
for each depth increment are printed in the field.
Auger time, torque during drilling, and grout volume
and pressure during grouting are listed for every
depth increment. Summary information shows
pumped volume for the auger stem, grout "head" (for
the given diameter the head is the equivalent length
of extra grout pumped prior to withdrawal), grout
volume per pile shaft increment (the most important
information), and spill (grout pumped above ground
level). Printed results are available prior to moving
to the next pile location.
When the operator uses the PIR-A results to guide
the pile installation, the installed shaR should meet
any minimum guidelines established by the
engineer. Thus, the quality assurance of the pile is

Figure 3. Pile integrity tester system.
improved. By thorough and accurate inspection, the
need for further NDE testing such as low strain Pile
Integrity Testing is reduced.
Recording of other construction operations such as
grout arrival time on site and collection of grout
specimens for strength testing cannot be done
automatically. The construction process also
includes other activities such as screening,
installation of reinforcement, maintaining adequate
waiting times between casting of neighboring shafts,
observing for subsidence for previously grouted
piles, site excavation, etc. which require some
human supervision. If problems are observed in any
post grouting phase, the piles can be subjected to
low strain Pile Integrity Testing to assure integrity.
3 PILE INTEGRITY TESTING
Pile Integrity Testing (Rausche et al. 1988, Likins
et al. 2000b) uses a hand held hammer to impact the
pile top and generate a compressive stress wave in
the pile. Stress wave inputs and reflections (from
non-uniformities or the pile toe) are measured by an
accelerometer. Figure 3 shows a Pile Integrity Tester
(PIT), accelerometer and hammer. This method can
be applied to almost any solid concrete shaft. The
pile top is prepared by removing the upper concrete
if it has been contaminated during construction, and
making a smooth location and attaching an
accelerometer with a thin layer of bonding material.
Accelerations from several hammer blows are
normalized, integrated, averaged, digitally filtered
and displayed as velocities. Further processing
applies an exponential magnification function which
restores reflection details diminished by soil
resistance, pile material damping or pile non-
uniformities. The resulting signal is interpreted by
the skilled test engineer.
Figure 4 shows an example output for a pile with
LengthDiameter ratio of 40. An exponential
magnification is applied, increasing from unity value
at the left or pile top to a maximum multiplier (40x)
207
at the right for the pile bottom at 25 meters (example
assumes a stress wave velocity of 4150 d s ) . The
bottom plot shows a clear pile bottom reflection with
a steady velocity signal between the impact and pile
bottom, indicating a good pile shaft. The upper plot
for another pile on the same site shows a pronounced
velocity increase at about 16 m which indicates a
reduction in pile cross section or concrete quality. In
general, sharply defined changes in the velocity are
attributed to impedance changes, while slow changes
are usually caused by soil resistance. If soil
resistance effects are known from reference piles,
then unusual shafts can be identified.
Figure 4. PIT velocity records of a deficient pile
(top) and a normal pile (bottom).
4 COMPARATIVE RESULTS FOR AUGERCAST
PILES
Both Pile Installation Recording for Augered Piles
and Pile Integrity Testing have been employed on
recent construction projects. Records from these
methods are compared and interpreted.
4.1 Case histoly I
The site had 500 mm (20 in) augercast piles ranging
in length from approximately 18 to 20 meters (58 to
66 ft). Soil borings generally indicated medium
dense sands transitioning into dense sands at a depth
of approximately 13.7 to 15 m (45 to 50 ft). The
shafts were socketed at least one meter into the weak
bedrock formation. Following an initial static test
failure on this site (static test pile not tested by PIT
or PIR-A), Pile Integrity Testing was applied to
several other completed piles and the Pile
Installation Recorder (PIR-A) was specified for use
on all remaining production piles to improve quality
assurance.
 The Pile Integrity Tester's records of all the shafts
in Pier B18 indicated a characteristic increase of
impedance in the upper weak soils (evident by a
negative velocity to a depth of about 45 R) followed
by a decrease of impedance (evident by an increase
in velocity) beginning at this depth. Although the
impedance decrease cannot be precisely quantified
by PIT, it seems likely that the decrease is due to a
return to nominal diameter in the lower denser soils
from a larger shaft diameter in the upper less dense
soils. The PIT velocity record for Pile BlSA, Figure
5, indicates a velocity increase or reduction in either
cross sectional area or modulus at approximately
14.6 m (48 R). The PIT toe reflection in the rock
socket resulted in a negative velocity (compression)
at 17.7 m (58 R). This is typical for shaft sections
with either a high soil resistance or an increased pile
impedance.
Figure 5. PIT data (case 1).
The PIR-A result for the same Pile 18A showed
that between 15.2 and 14.6 m (50 and 48 R) below
the top that the grout volume per 0.6 m (2 A)
increment decreased from an oversized average of
about 0.18 m3 to 0.13 m3 (6.5 R3 to 4.66 ft3). The
minimum required volume was 0.154 m3 (5.45 R3 )
per increment. Although this one increment was
slightly low, the volume for the low increment still
had a grout ratio of 1.07 over the nominal hole
volume and adjacent increments were higher than
required; some redistribution of the fluid grout
between neighboring increments is likely.
record along the shaft is relatively flat and thus
indicates a relatively uniform shaft and no defects.
Figure 6. PIT data for pile 349.
The PIR-A result is given in Figure 7. The
volumes printed are for 610 mm (2 R) depth
increments. This depth increment was selected to
provide sufficient precision taking into consideration
that typically about 8 pump strokes would be
required for such a depth increment. As shown, the
grout pumped was relatively uniform and generally
above the volume required of 4.421 R3 per depth
increment representing the 1.25 grout ratio
requested. The grout ratio for two segments in the
middle to lower part of the pile were slightly below
the 1.25 grout ratio desired for the project, but still
well above the factor 1.0. Further, the adjacent
segments were above the volume required. The low
volumes pumped near the pile top (bottom two lines
in Figure 7) are acceptable since the grout return was
already observed. This early grout return is due to
depth
44 7-88 22 0
42 4.a 7 0
s10 63 8 0
L

38 "59 0 0
Examination of numerous PIR-A records 36 I45 8 0
demonstrated fairly uniform grouting rates, and by 34 4.84 8 0
implication no significant problems. In addition, 32 4-63 8 0
selected piles were tested by Pile Integrity Testing 30 4.49 7 0
for integrity after installation. Because no hrther 2s 1.273 8 0
difficulty was experienced due to adequate PIR-A 26 4.73 9 0
quality control and confirmed by Pile Integrity 24 4.34 e 0
22 3.01 9 0
Testing, there was no hrther static testing. 20 4-70 7 a
18 4.70 9 0
4.2 Case history 2 16 1-45 e 0
14 .U 8 0
It 4.56 7 0
10 4.49 8 0
Shaft 349 is a 450 mm (18 in) diameter shaft 13.9
8 4.66 9 0
m (45.5 ft) in length and is one of several hundred 6 4.91 8 0
piles on this project. All piles were installed using a 4 1-342 2 0
PIR-A. This pile was requested to be tested to 2 0.991: L 0
assure the PIR-A results were achieving the desired < 1999-07-26-M~S5
- 1
result and as a random check on the pile. The PIT
record for pile 349 is shown in Figure 6. The pile
toe is very clearly visible at the expected time after a Figure 7. PIR-A data for pile 349.
magnification factor of 35 is applies and the velocity

208
the extra grout head pumped prior to auger
withdraw al.
5 HIGH STRAIN DYNAMIC TESTING
Monitoring during installation does not assure
sufficient pile bearing capacity. Thus, either
additional static testing or high strain dynamic pile
testing (Hussein et al. 1996, Likins et al. 2000a) are
required. High strain dynamic pile testing with the
Pile Driving Analyzer@(PDA) has also been applied
since 1974 to drilled shafts and augercast piles with
increasing frequency using drop weights to generate
the loading. In some countries extensive dynamic
pile testing is now routinely performed on drilled
shafts and augercast piles.
The pile top may be temporarily extended above
the ground surface t o encase protruding
reinforcement or to facilitate attaching the strain and
acceleration sensors two diameters below the pile
top. A steel striker plate and minimal plywood
cushion are placed on the pile top to distribute the
impact over the entire top surface. The drilled shaft
or augercast pile is then impacted by a simple drop
weight guided by a short set of leads. A low drop
height is first applied to assess signal quality and
alignment of the weight with the shaft. After each
impact, the net permanent displacement or "set per
blow" is carehlly measured. The drop height is then
increased until either the set per blow exceeds 2 or 3
mm (to activate the h l l capacity), or until the
indicated capacity exceeds the required ultimate
capacity, The PDA monitors stresses directly and
additional cushion material can be inserted to avoid
overstressing. The measured strain and velocity data
can be analyzed on site by the signal matching
software CAPWAP@ to independently check if the
activated capacity exceeds the desired test load.
6 CONCLUSIONS
L o w strain Pile Integrity Testing (PIT) can detect
major defects in the pile shaft at low cost and with
little effort. However, particularly for very long
piles, PIT results may be difficult to interpret and
should not be the only means to verify the quality of
the foundation. As a minimum, geotechnical borings
and field installation observations should be
included in the evaluation process of the foundation.
For very long piles the method may not provide
conclusive evidence of integrity of the whole shaft.
Where the Pile Installation Recorder (PIR-A) is
installed on an augercast rig, it automatically records
the installation of all piles on a job. Grout volume
and grout pressure information from the PIR-A
during installation guides the contractor into
installation of quality piles. The PIR-A records are
209
used to judge pile consistency and acceptance. With
more accurate information available, augercast piles
are more readily specified and accepted by designing
engineers. If automated PIR-A installation records
of grout versus depth indicate a good shaft, then this
may reduce the need for hrther PIT tests. Thus, PIT
testing can then be restricted to shafts with
questionable PIR-A records or shafts with problems
observed after installation, during installation of
subsequent piles or during excavation. In addition, a
small percentage of randomly selected production
piles may also be tested by PIT for quality control.
Capacity of augercast test piles can be determined
by a static load test. However, dynamic pile tests
followed by CAPWAP analysis is a well proven
alternative for augercast piles when minimum
installation time is critical or if multiple tests are
desirable to evaluate site variability. Dynamic tests
on augercast piles usually require some pile
preparation and a drop weight to apply the impact.
REFERENCES
Hussein, M., Likins, G. & Rausche, F. 1996. Selection of a
hanmer for high-strain dynamic testing of cast-in-place
shafts. Proceedings,j$h international conference on the
application of stress-wave theory to piles. Orlando, FL,
USA.
Likins, G., Rausche, F. & Goble, G. 2000a. High strain
dynamic testing, equipment and practice. Proceedings,
sixth international confirence on the application of
stress-wave theory to piles. Sao Paulo, Brazil.
Likins, G., Rausche, F., Morgano, C.M., & Piscsalko, G.,
2000b. Advances in PDI low strain integrity test methods
and equipment. Proceedings, sixth international
conference on the ayplication of stress-wave theoiy to
piles. Sao Paulo, Brazil.
Likins, G., Piscsalko, G. & Cole, C. 1998. Pile installation
recorder tests for ACIPJCFA piles. Proceedings, 7ke
deep foundations institute 's aiigered cast-in place piles
specialty seminar: aiigered cast-in-place piles. Houston,
TX, USA.
Rausche, F., Likins, G. & Slicn, R.K. 1992. Pile integrity
testing and analysis. Proceedings. &iirth international
conference on the application ?f stress-wave theory to
piles. The Hague, the Netherlands.
Roberts, T. 199X. Quality control for augered cast-in-place
piles in Texas and Louisiana. Proceecling.y the deep
fixinclations institute '.c nrtgeretl cusf-in place piles
specialty serninar: aiigered cast-in-place piles. Houston,
TX, USA.
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama (a Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 1503

cent advances and proper use of P 1low strain pile intepty testing

G. Likins
Pile Dynamics Incorporated, Cleveland, Ohio, USA
E Rausche
Goble Rausche Likins and Associates Incorporated, Cleveland, Ohio,USA

ABSTRACT: Low strain Pile Integrity Testing (PIT) is a valuable tool to locate major defects in drilled shafts
o r concrete piles. It has demonstrated its worth in detecting defective piles when other methods are not
practical, It requires no advance planning or access tubes so can be applied to any concrete pile after
installation. Since the testing is quick and simple, the cost per test is quite reasonable. However, proper
application and interpretation are essential to obtain reliable results. The advent of digital processing has
opened up a wide variety of possibilities for enhancing the information contained in the data. Many of these
procedures will be discussed and examples shown depicting their application. Certain rules of processing are
presented and should be followed to assure good interpretation.

INTRODUCTION TESTING P RACT 1C E

Structural engineers are often faced with the question
of pile integrity. Difficulties during installation,
excavation procedures, slope failures or lateral
movements due to accidental impact can create doubt
about the integrity of any pile. In some cases, doubt
also arises due to lack of information about existing
piles when a new use or increased design loads are
required. In some cases lack of inspection on current
installations leads the structural or geotechnical
engineer to question the foundation adequacy.
If access tubes have not been installed in the
questioned piles (required preplanning and extra
cost) to allow for Cross Hole Sonic (CSL) testing,
then the costs for evaluation can quickly become
large. Static loading for every pile is obviously
prohibitively expensive. Coring every pile is likewise
not feasible, and even if it were it would only provide
a small sample that could easily miss a local defect at
some depth. Therefore, the only practical alternative
is often PIT. Since PIT is simple and low cost. it i s
conceivable to inspect every pile on a site if required
Often a statistical sample is made and if no problems
are found, then testing may end. If defective piles ar-e
located, then additional tests of neighboring piles ar-e
then justified. The selection of which piles to test can
be based on various observations. or made
considering redundancy of the pile, or at random
The pulse echo method (Rausche et al 1992) uses a
hand held hammer to impact the pile top and generate
a compr-essive stress wave in the pile Figure 1
shows a Pile Integrity Tester (PIT) for pulse echo
testing. The small battery powered main unit is
highly portable from pile to pile and extrcmel!
rugged. It uses a touch screen for input rather than ii
computer keyboard. Experience has proven this to
be a highly reliable method and extremely user
friendly. The screen displays intuitive menus to guide
the user and shows the graphical signals for field
interpretation. Stress wave inputs and reflection?;
(from non-uniformities or the pile toe) are measured
as a hnction of time by an accelerometer attached to
the pile top. The acceleration is integrated to
velocity by PIT, and then interpreted by the test
engineer.
The pile top surface is prepared by removing the
upper concrete if it has been contaminated with soil,
bentonite slurry or other foreign materials during
construction, and finding or making a smooth
location when a driven pile top may have been rough
cut. An accelerometer is then attached to the smooth
pile top surface as shown in Figure 1 with a thin layer
of a soft paste like vaseline, petro wax. etc.
Accelerations from several hammer blows are
normalized, integrated, averaged and displayed as
velocities. Further data processing includes

21 1

Figure I . Pile Integrity Tester (PIT) with hammer
aid accelerometer for the application of the pulse echo
method.
application of an exponentially increasing
magnification function. The magnification restores
reflection details which are diminished by soil
resistance, pile material damping or pile non-
uniformities.
To demonstrate the usefulness of PIT in
distinguishing a good pile from a defective one,
Figure 2 shows an example output with an
exponential magnification increasing from zero at the
left or pile top to a maximum multiplier (40 times in
this example) at the expected time of reflection from
the pile bottom on the right; the pile length is 25
meters (82 ft), and the stress wave velocity is 4 150
m / s (13,610 ft/s) in this example. The bottom plot
shows a clear signal from the pile bottom together
with a steady velocity signal benbeen the time ot'
impact and pile bottom reflection, indicating a good
pile shaft. The upper plot for another pile or1 the
same site shows a pronounced velocity increase
about 16 m (52 5 ft) which indicates a reduction in
pile cross section or concrete quality I n general.
Figure 2 PIT velocity \ S depth records o f a deticienr
pile (top) and norind pile (bottom)
212
relatively sharply defined reflections are attributed to
impedance changes, while slowly changing retlections
are usually caused by soil resistance If the etfect of
soil resistance is known from reference piles. then
unusual shafts can be identified This method can be
applied to almost any shaft
DATA ENHANCEMEhT AND Ah.4LYSIS
How do we obtain a clear graph such as the one in
Figure 3? The actual raw data for the defecti\.e pile
is shown in Fiyure 3
Figure 3 "Raw data" from a PIT test
Interpretation would be difficult at best since the soil
resistance has greatly dampened the reflecting
signals. The standard method to compensate for the
soil damping is to apply a magnification function that
begins shortly after the input pulse with a value of
unity and increases exponentially to some factor (40
in this example) at the pile bottom. The amplification
function should be shown graphically. The result of
the application of this exponential magnification
function is shown in Figure 4. In this case the curve
Figure 4: With only magnification
also gradually wanders fi-om the zero axis and i n part
resembles the magnification function
Slow changes are difficult to interpret Generally
they are the result of soil resistance effects
However, the real goal for integrity testing to
evaluate structural pile integrity so soil resistance is
not important. Therefore, a high pass filter (I-IPF) is
applied which removes low frequency events usually
caused by soil resistance. In the case of the extreme
example in Figure 4 (most data is generally near zero
even after magnification), the data is subjected to
high pass filtering which removes components with
less than a defined filter frequency. In general, care
must be taken to not apply E-IPF with too high a filter
frequency which could eliminate the frequencies of
interest. The most important frequency components
vibration of reinforcement bars protruding fi-or11the
pile top cracked pile top concrete or other- surtice
effects. Since the “noise” frequency is much hishei-
than the input frequency, if we select a wa\,elet
mother function with similar frequency content to the
input pulse, unwanted frequency content will be
suppressed as shown by the wavelet filtered Figur-e 7

Figure 7: PIT record ofFigure 6 with applied \ v a \ w l e ~

filter

Figure 5: With 40 magnification and 25 high pass filter

are those that make up the input pulse and the

reflections from changes of pile pi opei LIP,
Reflections from defects or the pile bottom h o u l d
have frequencies which are similar to that of the
fo rcing input func t I o n

The results shown in Figure 5 were further impr-o\,cd

t o the top curve in Figure 2 by use of the wabelet
analysis, although in this case the inipr-o\,ement I S
slight. The wavelet analysis (Rioul 199 1. Seidel
2000) is a specialized filter that et‘fecti\,el\
st i-engthens t he signa1 freq u en cy CO in p o ii e 11t s t h a t
match the input pulse and removes undesirable
frequency coin p o n en t s re su 1t i n 2 fir0 m 11o i s e
To more clearly demonstrate the effectiveness ot‘

Figure 8: Data with different locations for hammer

impact and accelerometer.

which is much more easily interpreted.

Figure 6: PIT record with significant “noise”
wavelet analysis, consider the record in Figure 6
which includes significant high frequency “noise”
components in the data caused by PerhaPS by the
Another consideration in PIT is the location of the
accelerometer and the applied hammer impact. The
three records shown in Figure 8 have different
characteristics.
The top curve has the hammer input applied at the
pile center while the accelerometer is attached at the
edge. The middle cuive has the same accelerometer
location but the force is applied at the eytreiile
opposite pile edge; obviously this configuration i s
less desirable. In the bottom graph. the hammei-

213

Figure 9: Same data as Figure S but with Wavelet
processing
input and accelei onieter loc(ttions <iie n t the I -1 'iriti
3/4 points of a inajoi diagonal The onl) siyi<il
processing in Figure 8 IS a five times magniticntion
The data can be further improved by wavelet filter5
as shown in Figure 9 Even the originall) poor i n p ~ i t
data i s now easily evaluated However, cciution I \
advised for cases where the defects near the pile top
produce a i ecord as shown i n the middle of' F i p r e S
Data quality is not the only consideration for- location
of accelerometer and hammer impact. For l a y e
shafts, the ratio of hammer to shaft diameter becomes
very small. Example 10 shows four- records for- an
SO0 inin (32 inch) diameter- drilled shafi I n all cases
the accelerometer was attached to the pile cenlcr
Figure 10 Accelerometer at center of SO0 iiiin pile,
hammer impact applied at N, S, W and E locations
214
location. The hammer impact location h a s La[-itxi
around the pile at the four compass directions
(specifically Nor-th, South, West. and East.
respectively). It is relatively clear that there is a local
defect near the pile top in the NW quadrant
For piles installed i n a soil bearing layer- lie i w
reflection is almost always tension because the soil
stiffness is less than the pile stif-fness Tension
reflections are observed in PIT data as a velocit!,
reflection at the toe \uith the same sign as thc impact
signal. In all example cases presented in this pqw
the pile toe produced such a tension reflections 'l'lic.
question often arises as to what magnification factor
to apply. In general one should try to make the
strongest reflection as high as the impact signal.
For similar piles in similar soils on the same site, it is
logical that the signals obtained by PIT should be
processed with similar parameters and result in
Figure 1 I : PIT records for two piles of different lengths
similar curves. It is relatively simple to identi5 piles
with unusual records and thus properties. However,
what if the piles are different? For example, the
results of tests on two piles installed in similar soil
conditions on the same site but with different lengths
are shown in Figure 1 1
The toe reflection for the shorter pile is the same
amplitude as the input using a magnification factor of
5 However, the longer pile has a smaller reflection
using the same magnification This is logical due to
the effect of extra soil damping on the longer pile
The binary magnification factors (2, 4, 8, 16, 32 )
are shown as the vertical markers inside the graphical
representation of the magnification ciirve For the
shorter pile in Figure 1 1 with magnification at the
pile toe, a magnification of 4 is at about 3 1 ft depth
However, for the longer pile, at 31 ft the
magnification factor is only about 2.5 while the factor
4 is only achieved at a depth of 47 fi. Thus the use
of similar magnification factors for different length
piles does not result in “consistent” niagnification
functions.
In Figure 12 a magnification of 15 has been applied
t o the longer pile resulting in a factor of 4 at about
31 fi so this then is a consistent magnification as a
function of depth. The toe response is now of similar
Figure 12 Pit records &ith consistent iiicigiiificdioii i;,i
different length piles
amplitude for both piles even though the lenyth\ ni c
different It appears that the longer pile I S only about
13 7 m (45 f t ) long rather than the planned I 5 S iii
(52 ft) While there IS some uncertaintv ( u p to 10” U )
in the wave speed of concrete for the wiie concrete
suppliei, the same wave speed i s usunllq u\ed t i ) i ‘111
piles at one site to plot the records If the roe
reflection occurs at a time that indicates more than
10% variation of wave speed, the pile length I \
questioned I n this case, the length of the \Iic)i t e i ~ I I C
matches well the toe signdl with the iioimnllj
assunied wave speed of 3,960 in/\ ( 1 i 000 ti/\) I 0 1
the longer pile the wave speed ~ ~ o u lhdn \ e t o tw
4,580 m/s ( 1 5,000 ft/s) to indtch the toe s ~ g n ~l ibli t h (1
pile length of 15 8 in (52 ft) Since this wnve speed
exceeds the average one bv inore t h n n 10” i t \iotiI(I
be concluded that the pile was indeed t o o h i t
There are further processing options for data Data
can be analyzed i n the frequency domain for iiiobilit!
and dynamic stiffness. However, determination of‘
length or shape is better made i n the time domain a s
presented in the exainples in this paper- since the
magnification function can only be used in the tinw
domain. Lack of a magnification severely limits the
effective length that can be analyzed i n the frecluenc:),
domain. Further, if something is clear in the
215
frequency domain, then it is even more apparent
when presented in the standard time record The
major advantage for the frequency aiialj,sis i n our
opinion is to obtain a quantification of shaft qualit),
by calculating the relative pile stiffriess (piles v, it11
lower stiffness are more likely to be defective.)
One other tool associated with PIT is to obtain an
impedance profile. This makes a shape visualization
as in Figure 13. However, experience has shown that
this is not often required and piles can be classified
without this extra effort for most piles.
Figure 13 : Processed signal and calculated profile
RECORD CLASSlFICATION
Before embarking on any type of pile integrity test,
there should be a clear understanding among all
parties involved as to the consequences of the test
and the action to be taken if the test generates a
question in the integrity of the foundation. The test
results can generally be placed into the following 4
categories and each one of them may require a
different course of action.
A - Clear toe response. No apparent defect.
B - Clear indication of serious defect. Toe
reflection is usually not apparent.
C - lndication of possible defective shaft, although
toe response is apparent.
D - lnconclusive data
Category A piles are good piles within the accuracy
limits of the method. For example, a it is generally
agreed that a defect that affects less than 20’10 01‘ the
pile cross section cannot be detected with certaint!,
(actually this a strength of the method since it does
not create questions where only minor problems
exist.) In this category pile integrity is satisfactoiy
and unless there is some othei- reason to suspect the
shafts, they are generally acceptable Of course, this
assumes that also the pile length indication i h
satisfactory.
Category B piles are somehow defective and some
contingency plan must be used Extra tests could be
made. If the defect is near the ground sur-t’ace.
excavation to expose and repair the dettctile pile
portion is generally possible. The shaft could a l s , ~be
cored and the defective are pressure grouted, a
subsequent PIT could confirm the effectiveness of‘
any repaii- measure The pile could be abandoned and
simply replaced The course of action inav depend
on the cost of a new pile vei-sus the cost to repaii-
Category C piles may also be assigned a reduccd
capacity. Also, other pile tests may be considered. or
excavation if the defect is located near the y-ound
surface. If the pile is a friction pile and the det’ect is
located far down the pile. the upper s o i l resistailcc
above the defect may render- the h i - c e in thc p i l e 21;
the defect as acceptable structur-ally and the ciclilii
inay in some cases be not serious
Category D piles may have poor data dcic to pooi.
concrete quality at the pile top. Tr-imming the pile to
a lowei- elevation where good quality concrete IS
assured and retesting the pile is an alternatiie Ollici
reasons for inconclusive results such as a long oi
irregular pile are discussed in the next section I t is
reasonable to accept a certain percentage of piles
with inconclusive records at a site where soil and pile
conditions produce complex recoi-ds.
LlMITATIONS
PIT is a very useful test for many sitiiatioiis I n ~ i i i e
cases it is the only practical alternative to doins
nothing which may not be acceptable i z hci-e
questions arise during constr-uction I t is certaInl\ ;i
simple and relatively inexpensive test and thus i s
widely used. It is intended to locate major defects
and to that put-pose it generally is adeqiratc l r i h ii
strength of the method that it does not generate
216
concern where there are localized defects of small
extent.
The method works particularly well on solid pile
cross sections, such as concrete or timber piles. It
has only limited applicability to steel piles. Of
course, concrete filled pipes can be tested.
In cases where there are mechanical joints or full
section cracks, the wave cannot usually cross the
resulting gap. Therefore only the portion above the
gap is really tested.
Defects which are very short compared with the
input pulse width and do not cover the entire section
can be misleading due to superposition of the waves
generated by both initial reduction and following
increase. A similar problem exists for gradual
changes which create relatively slow changes in the
record. Normal processing of PIT data then removes
the reflections from gradual changes since they are
similar in frequency content as the soil resistance
effects. Therefore if a pile gradually increases and
then quickly returns to the nominal cross section as it
enters a new strong soil layer, the sudden reduction
inay produce a record that indicates a defect. I t is,
therefore, important to compare any PIT result with
the general soil profile to assess the likelihood of
these situations.
There is a general rule of thumb that the test
effectiveness is limited to piles with lengths less than
30 times their diameter (L/D ratio) This is really
applicable to the length embedded in soil since there
is little signal degradation for shafts acting as
columns in water until below the mudline. In some
cases it might not be possible to “see” even this far
down the pile. However, with recent improvements
in electronics including lower noise and higher
precision 16 bit sampling (which gives 8 times more
resolution to older 12 bit samplers), PIT now
regularly tests piles with L/D ratios far in excess of
30.
The pile length effectively investigated by PIT
depends not only on soil damping, but also on the
non-uniformity of the pile, because any change in pile
quality or shape produces reflections. The first non-
uniformity detected will be more reliably analyzed
than additional sources of reflection farther down the
shaft. Highly non-uniform piles produce complex
records which are dif’ficult to analvze I n such c a w h
the records collected under comparable
circumstances can be compared Records thitt dill;,i
significantly from the “normal signature“ should Ix are obtained by avoiding the test on contaminated
hrther investigated. pile top concrete and by assuring meaninghl records
where the pile top is large relative to the hammer
PIT gives no information about load carrying size. In many instances it is helpful to vary hammer
capacity To confirm capacity a static loaad test or a size and impact and measurement location at the pile
dynamic pile test are required (Hussein 1996, Likins top. The former to produce records that have
2 000a). sufficient resolution and the proper frequency content
for a particular pile-soil system, the latter to find any
defects near the pile top but only at one side.
OTHER APPLLCATLONS, SOLUTIONS
The analysis and data interpretation process also has
PIT is essentially an after-the-fact test. The pile must
to be done carefully, particularly when choosing
be already installed and the concrete hardened I f a
magnification and filter parameters. Fully automated
major defect is present in the pile, it can be detected.
data processing is usually not advisable: each step of
but the consequences and costs of repair- are
the data processing procedure should be reviewed as
relatively high and it is obviously better to avoid an)’
to its effects on the records.
defects. An installation monitor for Continuous-
Flight- Auger Piles also called Auger-Cast Piles is an
The processed data should be plotted and the records
instrument that inspects the pile during installation
classified. Depending on the data interpretation
when correction of the defect is both simple and
certain additional actions may be required on the
relatively inexpensive (Likins 2000b). Even U i t h
construction site. It is important that such possible
such installation monitors, the need for PIT rernains.
actions are clearly established prior to conducting
for example, when problems occur after the
PIT.
installation such as lateral impacts, slope failures. or
often merely the need to evaluate existing piles
REFERENCES
Piles have also been often successfully tested iifici-
they had already been put to service under a structure
Hussein, M., Likins, G., and Rausche, F., (1 996)
(Hussein et al., 1992). However, the more complex
Selection of a Hammer for High-Strain Dynamic
the structure-foundation system the more difficult
Testing of Cast-In-Place Shafts.. Proceedings,
will be the data interpretation. For example. it is
Fifth International Conference on the Application
possible to tests piles located below a pile cap. but
of Stress- Wave Theory to Piles, Orlando, Florida,
the bottom of the pile cap will produce a mqui-
USA.
reflection. This is acceptable if the pile below the c a p
is uniform and of reasonable L/D ratio. Testing U I I ii
Hussein, M., Likins, G. and Goble, G., (1992)
pile below a structure will produce waves that travel
Determination of pile length under existing
up in the structure as well as down the pile To
structures, Proceedings, Deep Foundation
clearly separate reflections of pile from those of the
Institute annual Meeting, New Orleans.
structure it is often necessary to install two
accelerometers on the side of the exposed pile z t t\\v
Likins, G , Rausche, F., and Goble, G., (2000a)
different levels.
High Strain Dynamic Pile Testing, Equipment and
Practice, Proceedings, Sixth International
Conference on the Application of Stress- Wave
CONCLUSIONS
Theory to Piles, Sao Paulo, Brazil.
Modern digital electronics and signal processing
Likins, G , Rausche, F , Morgano. C M , and
techniques have made it possible to build equipment
Piscsalko, G (2000b) Detection and Prevention
and develop analysis methods that extend the natural
limits of the simple Pulse Echo Method Although of Anomalies for Augercast Piling, l’rocrcw’iiicy\.
Sixth International Cwference 011 he A p p f i u u i o i t
the PIT equipment meets all the demands that can be
of Stress- Wave Theory to Piles, Sao Paulo, Brazil
put on a modern equipment, successhl testing still
requires experience both during the test process itself
Rausche, F , Likins, G and Shen, R K ( 1 992) Pile
and when analyzing the data.
Integrity Testing and Analysis Protecrdinp
Fourth Infernational (’onftciznte o t i t h e
The field engineer has to make sure that clear records

217
 Applicafioti of Stress- W m e 7heoq: t o 1)ile.s. The
Hague, The Netherlands.

Rioul, 0 and Vetterli, M (1991) Wabelets and

Signal Processing IliLL S i p i d Proco\ jiiig
Magmine. Vol 8, no 4 October, p 14-38

Seidel, J.P., (2000). Presentation of low strain

integrity testing in the time-frequency dornain.
Proceedings, Sixfh Iuiferticrtiouial ( ' o i ? f i . l - ~ ~ i i c011
~e
fhe Applicatioii of Stress- Wave 7hc.o~~: to 1'ilcJ.s.
Sao Paulo, Brazil.

218
 Application of Stress-Wave Theory to Piles, Niyama 13Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Stress wave propagation velocity at early ages

C. Restrepo
Esninesa and Restrepo, Bogota, Colonzbia

ABSTRACT:

1 INTRODUCTION Table. 1
In order to ensure the continuity of a cast-in-place I Type I Nominal compressive strength IW/Cratio
concrete pile Pile Integrity Tests (PIT) can be per- Concrete
formed with certain advantages such as low cost and Concrete
Concrete 4000
non difficult pile head preparation. However, given Mortar 3000
that the stress wave propagation velocity is assumed Mortar 3500 0.60

during the PIT analysis, the ASTM standard requires

that concrete must have at least seven days of age,
moment what is considered that concrete’s wave
propagation velocity vs time curve starts to flatten.
Given the construction processes, sometimes an
early PIT analysis on a certain pile could be desir-
able to help the decision making in the construction
process. In order to investigate the stress wave veloc-
ity at early ages (younger that the seven days given
by the ASTM standard) the geotechnical instrumen-
tation Department at Espinosa & Restrepo ltda with
the collaboration of senior undergraduate students of
the Universidad Javeriana in Bogota Colombia
planned and performed a laboratory testing program.
The present paper describes the methodology, the re-
sults and conclusions of the tests performed.
2 TESTING PROGRAM
2.1 Probes preparation
Using concrete or mortar mixed by a commercial
concrete plant owned by Metroconcreto S.A based in
Bogota, Colombia; 4 concrete probes 5 meters long
and square section 0.20 wide were casted for each of
the following compressive strength mixtures: 3000
psi, (20.68 Mpa), 3500 psi (24.13 Mpa) 4000 psi
(25.58 Mpa) and mortars 3000 psi (20.68 Mpa) and
3500 psi (24.13 Mpa).
The waterlcement ratio for the concrete used were
a s follows:
During the probe’s cast, concrete or mortar the
materials were sampled for compressive strength
tests at ages 3 days, 7 days, 14 days and 28 days. The
probes were sampled and tested according ASTM
standards. To ensure proper curing all samples were
covered with wet fabrics to guarantee moisture con-
tent during the curing process.
2.2 PIT testing
Over all 4 probes for each material type PIT tests
were performed every day from day 1 to day 28 to
monitor the stress wave propagation velocity change.
During testing enough PIT registries were taken for
proper statistic analysis.
In order to obtain appropriate PIT curves a 450
grams hammer was used given that with such mass a
the pile tip was registered with good definition and
the length of initial peak was minimized; all PIT
measurements were performed by the same technni-
tian in order to achieve the highest homogeneity in
the PIT registries.
2.3 Error analysis
During the analysis 3 error types where statisti-
cally analyzed using Pile Dynamics software and
basing the analysis on velocity or length variations,
and the results obtained are revised as follows:

219
2.3.1 Heterogenity between probes within a con-
crete strength
Includes variations introduced at the casting mo-
ment by the fluid concrete mix vibrating of the
probes. To measure this error the coefficient of
variation of the velocities obtained using PITPC
software.
REF. RANGE AVERAGE STAT DISTRIBU-
TION OCURRENCE PROBABILITY (note Nol)
Whole sample 1.0% - 3.0% 2 CHI SQUARE
65%
REF RANGE AVERAGE STAT DISTRIBUTION OCURRENCE PROBABILITY
(note Nol)
2.3-2 Heterogeneity during PIT curves acquisition
Heterogeneity introduced by variations of the PIT
technician criteria, accelerometer disposition etc.
Such error was statically analyzed in terms of meas-
ured pile length based on the results obtained with
PITPLOT software. The variation of length error’s
magnitude obtained is as follows:
Table.3
According to the results the investigators con-
clude that the error introduced was reduced within
acceptable ranges.
2.3.3 Errors within the PIT analysis.
It is given by the variations generated by tests per-
formance variations, variabilty within the analysis
given variations on the technitian’s criteria who em-
ploys the software tools for a particular test.
For this particular testing program most of the
software tools such as high pass, magnification and
pivoting were not used given that the data acquisition
was closely controlled to obtain the best possible
curves.
220
3 RESULTS
Given the curves obtained for every ages within a
concrete or mortar strength vs wave propagation ve-
locity two intervals can be clearly differenced. The
first includes mixtures of ages between 1 to 7 days
were an strong dependancy between age and veloc-
ity.
All four probes data velocity vs mix age at early
ages (1 to 7 days) were averaged and regression
analysis were performed to modeled the age-velocity
relationship during early.
The following chart resumes the regression
analysis performed and the results obtained
at ages from 1 to 7 days.
Concrete type REGRESSION ANALYSIS
Type EQUATION R’
CONCRETE 3000 prl Log. V=234.99Ln(t)+2864.2 0.980
CONCRETE 3500 prl Polinomlal V=O.40581‘.6.4744t’+ 15.931t’+ 138.16t+3018.9 0.977
CONCRETE 4000 pd Pollnomlal V = l.2937t‘-18.688~’+56.102t’+209.44t+2374.9 0.998
MORTAR 3000 psi Log. V=256.05Ln(t)t2966.4 0.981
MORTAR 3500 prl Polinomial V = 15.452r’-228.6t*+ 1083.8t+ 1863.9 0.992
4 CONCLUSIONS
The possible errors within the testing program
correspond to normal conditions during the construc-
tion or PIT performance and they were reduced to
acceptable ranges.
The wave velocity is directly proportional to the
concrete‘s mix age during the first 7 days of curing
process.
According to R2 obtained the regression analysis
performed characterizes accurately the stress wave
velocity at early ages; for ages older than 7 days the
stress wave velocity reduces its dependancy of age
and becomes a property related to the compressive
strength of the concrete or mortar.
The correlation equations obtained between age
and stress wave velocity for every concrete or mortar
tested fits well reflection R2 values close to unity.
In general the result obtained with the mortar re-
flects a more consistent behavior; in the opposite the
concrete reflects results slightly more scatter. The re-
searcher attributes this behavior to the fact that in the
case of mortar probes the ratio between aggregates
particule size/cross sectional side is much lower than
for the concrete. i.e in concrete the ratio varies from
0.10 to 0.19 and for mortars it rages around 0.05 or
lower obtaining a much more uniform media for the
stress wave to travel
The results obtained with the 24.13 Mpa(3500
psi) reflects a reduction of the stress wave velocity at
28 days of age. The researcher concludes that exter-
nal effects could have influence such behavior.
There is a clear tendency for mortar to get higher
stress wave velocity than concretes.
22 1
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 1503

Examination of a new cross-hole sonic logging system for integnty testing

of drilled shafts
Samuel G. Paikowsky
Geotechnical Engineering Research Laboratov, University of Massuchusetts, Lowell, Mass., USA
Les R.Chernauskas & Leo J. Hart
Geosciences Testing and Research Incorporated, North Chelmsford, Mass., USA
Car1 D. Ealy & Albert E DiMillio
Federal Highway Administration, McLean, Va., USA

ABSTRACT Drilled shafts and other mixed or cast-in-place concrete deep foundation elements can be
costly solutions These foundations usually carry very high design loads, and often serve as a non-redundant,
single load-carrying unit These conditions have created a need for a high-level of quality assurance and
control applied to each in-place constructed deep foundation element
The non-destructive testing method, Cross-Hole Sonic Logging (CSL), currently offers the most reliable
technique for assessing the integrity of in-place constructed deep foundation elements Recent years have
seen progress in CSL instrumentation, taking advantage of the available computer technology The software
applications, however, have greatly fallen behind, thereby limiting the effectiveness and potential of the CSL
method and deep foundations integrity testing in general
A new, original CSL testing system by the name of PISA (Pile Integrity Sonic Analyzer) makes use of an
innovative software and data acquisition system, hence representing the state-of-the-art in deep foundation
integrity testing The PISA has the capability to show real-time graphical information during logging, in-
cluding planar tomography, which can identify the boundark! of a compromised zone within the foundation
element The equipment operates completely in a Windows graphical environment allowing alphanumeric
and graphical reports to be generated directly into word processing software The real-time graphical repre-
sentation during logging and the ease of reporting enables immediate, extensive on-site evaluation and deci-
sion-making
The PISA system was evaluated on different construction sites The case history presented in this paper
relates to a class ‘A’ prediction as tests were carried out on shafts in which defects were intentionally planted
The test results were submitted before the defects locations were known, both presented I the paper The ob-
tained results demonstrate the ease of use, accuracy of measurements and enhanced capabilities of the PISA
The systems’ abilities are shown to be superior to any other currently available commercial system

1 INTRODUCTION Cross-Hole Sonic Logging (CSL) is a common

Deep foundations integrity testing mostly applies
to foundations constructed on-site from concrete or
grout, such as drilled shafts, drilled mini piles,
pressure-injected footings, and pre-cast concrete
piles. Drilled shaft foundations usually carry very
high design loads, and often serve as non-
redundant, single load-carrying units. The integrity
testing is required for quality control during con-
struction to detect flaws in the pile (e.g. necking,
cracking, void, poor quality material, etc.) com-
mon in these cast-in-place concrete piles. As a re-
sult of the increasing design requirements on these
foundations, a need for a high-level of quality as-
surance and control has been created.
testing methods for determining the integrity of in-
place constructed deep foundation elements, such
as drilled shafts and caissons. A minor variation
of this method, called Single-Hole Sonic Logging
(SSL) can also be used on smaller diameter drilled
mini-piles and augercast piles. These methods are
both non-destructive testing (NDT) methods and
involve generating a sonic pulse with one trans-
ducer (transmitter) and picking the signal up with
another transducer (receiver). The transducers
typically consist of a geophone or accelerometer
The methods differ only in the number of tests per
pile and the locatiodorientation of the transducers
within the pile.
Significant improvements and advances in in-
strumentation, data acquisition hardware, and

223
computer technology have been made in recent
years. The software applications, however, have
greatly fallen behind and have not taken full ad-
vantage of the existing technological advances,
thereby limiting the effectiveness and potential of
the CSL method, as well as other deep foundations
integrity testing methods (Chernauskas and
Paikowsky, 1999).
A new state-of-the-art CSL testing system has
recently been developed that utilizes unique soft.
ware to take advantage of the new hardware (Amir
and Amir, 1998a) This system is called the PlSA
(Pile Integrity Sonic Analyzer). The PlSA is based
on a lightweight, portable, pen touch, computer
that operates in a Windows graphical environment.
This system is easy to use and efficient with regard
to its ability to make the collected data available in
a real-time manner. The following paper provides
the basic background theory on the CSL integrity
testing method, a description of the PISA system,
and a summary of a recent case history including
large size rock socketed drilled shafts, defects de-
tection, and verification.

2 OVERVIEW OF ULTRASONIC INTEGRITY

TESTING METHODS
2. I < iwss-Hole Sorm h g g ~ ~ g
Cross-Hole Sonic Logging (CSL) is the most
common integrity testing method for drilled or
cast-in-place foundations A piezoelectric trans-
ducer is used to generate a signal that propagates
as a sound (compression) wave within the con-
crete, while another transducer is used to detect the
signal Each transducer is placed into a vertical
PVC or steel tube that has been attached to the re-
inforcement cage and filled with water prior to the
concrete placement The water acts as a coupling
medium between the transducer and the tube A
typical tube arrangement and testing principles are
presented in Figure 1
The source and receiver transducers are lowered
to the bottom of their respective tubes and placed
such that they are in the same horizontal plane.
The emitter transducer generates a sonic pulse (on
the order of 10 pulses per second), which is de-
tected by the receiver in the adjacent tube The
two transducers are simultaneously raised at a rate
of around 300 mm/sec (1 ft/sec) until they reach
the top of the drilled shaft Typically this process
is repeated for each possible tube pair combination
(perimeter and diagonals) Figure lb shows the six
tube combinations that can be tested (logged) us-
ing a configuration of 4 tubes within a drilled
shaft Increased shaft diameter calls for a larger
number of tubes, which increases the number of
Figure 1 Typical CSL testing setup showing (a) transmitter
and rccen er at dffcreiit depths. aiid (b) plaii 1 ien of thc CSL
tubes 111111 possible test coinbiiiatioiis
combinations and thereby the resolution of the
testing zone.
In homogeneous, good quality concrete, the
stress/sound wave speed, C, is typically around
3,800 m / s (12,000 to 13,000 fi/s) and is related to
the modulus, E, and bulk density (unit weight, y,
and gravitational acceleration, g) as follows:
If for any reason the condition of the concrete is
compromised, the wave speed will be reduced
relative to that of the "good or sound" concrete
value. Figure 2 presents a typical sonic signal for
which the propagation time between the transduc-
ers is measured. The vertical axis is the signal
amplitude (microvolts) and the horizontal axis is
the time (microseconds). The point where the am-
plitude begins to rapidly fluctuate indicates the ar-
rival time of the signal to the receiver (a.k.a.
threshold time) Since the distance between the
two tubes is known, the wave speed of the con-
Crete between the tubes can be evaluated by the
fo 110wing relationship:
C = -L
t The testing can be performed only on drilled shafts
The wave speed in equation 2 is only an esti-
mate, as the identification of the arrival time, t, is
subjective and the distance between the tubes, L, is
known only at the top of the shaft. The signal ar-
rival times can then be plotted with depth to gener-
ate a log for the particular tube combination as
presented in Figure 3. In addition to the threshold
times, the energy of each signal may also be plot-
ted with depth. This information can be used to
compare signals of one zone to another where
lower energy and/or later arrival times correspond
to a compromised concrete quality andlor defect.
Advantages to this method include the direct as-
sessment of pile integrity and the ability to posi-
Figure 2. CSL typical testing signal.
Figure 3 Presentations of CSL test results in the form of
threshold time and energy with depth.
225
tion the transducers in different elevations to create
more signals, allowing the development of a to-
mographic presentation of the investigated zone.
The limitations of the method include detection of
defects only when they exist between the tubes.
for which access tubes were installed. Debonding
between the tubes and concrete is common if test-
ing occurs long after the concrete placement.
Testing in fresh concrete is also difficult as certain
zones may cure at a lower rate, creating difficulties
in the interpretation of the threshold time and en-
ergy. These zones may therefore be interpreted as
poor quality concrete.
2.2 Single-Hole Sonic Logging
Single-hole sonic logging (SSL) is a variation of
the direct transmission CSL method in which the
source and receiver are placed in the same tube
and the signal travels in a vertical direction (refer
to Figure 4). For drilled shafts and caissons, the
method is limited to defects adjacent to the tube
and is usually used only when a drilled shaft re-
quires integrity assessment after construction. Due
Figure 4. Typical SSL testing set-up showing transmitter and
receiver at different depths.
to high coring costs, a single hole is advanced (of-
ten down the middle) to the bottom of the shaft or
slightly below the depth where a defect is antici-
pated. It may also be desirable to perform SSL
during CSL testing to isolate the location of a de-
fect at a certain depth (i.e. distinguishing whether
the defect identified using CSL is adjacent to the
tube or in between the tubes). Recently SSL has
been performed within smaller diameter drilled
mini-piles and augercast piles (Amir and Amir,
1998b) The use of SSL in these foundation types
may become more commonplace in the near f&
ture, as research and experience provide insight for
the most efficient vertical placement of the tubes
to assess the lateral integrity. Brettman and Frank
(1996) describe a comparison between CSL and
SSL tests
3 THE PISA CSL/SSL TESTING SYSTEM
3. I Gene~crl
The PISA (Pile Integrity Sonic Analyzer) is a
modular system allowing for adoption, upgrade
and incorporation of additional integrity testing
technologies The current integrity testing options
available in the PISA include cross-hole sonic log-
ging (CSL) and single-hole sonic logging (SSL)
using CHUM (Cross-Hole Ultra Sonic Module)
and sonic echo (a k a small strain propagation)
using PET (Pile Echo Tester) module Additional
modules are currently under development
In addition to its modularity, two advantages of
the PISA integrity testing system over other sys-
tems include its software and portability The PISA
is the only Windows 95/98 based system and is
also compatible with Word 2000 The software is
updated periodically to incorporate new develop-
ments and algorithms that make data collection,
interpretation, and report preparation easier and ef-
ficient The PISA is lightweight (only 42 3 N (9-
1/2 Ib)) and self powered, hence can be easily car-
ried around from shaft to shaft or site to site This
feature is also beneficial for air travel The system
can be also used as a standard laptop, saving the
cost and space required for an additional personal
computer (PC) when using a dedicated CSL testing
system
Figure 5 presents a photograph of the PISA
system, including computer and sensors As a
scale, the width of the computer screen is 23cm (9
inches) Figure 6 presents the layout of the pile
screen, where one can enter the pile information
and select the tube orientatiodlocations Selection
of the desired tube combinations is accomplished
by drawing a line between any two tubes Real-
time graphical presentation of the concrete integ-
rity is provided during data collection If a suspect
226
zone is detected in this stage and the tomographv
option is enabled, the probes are lowered and
raised relative to each other around the suspect
zone, to fkrther investigate and delineate the area.
The signals can be examined and adjusted by
manually picking the points or using preset algo-
rithms to automatically determine the first arrival
time (FAT) as shown in Figure 7.
Figure 5 . PISA system components.
4 CASE HISTORY
4 I Rnckgrormd
Four drilled shafts were constructed at the Auburn
University in Auburn, Alabama as part of a re-
search study The shafts are 914mm (36in) d]-
ameter and approximately 11 Om (36fi) long Each
shaft was equipped with 4 access tubes and various
defects were installed during construction The de-
fects were constructed as soil inclusions formed by
sand bags made of a tough material and tied to the
rebar cage The cross sectional area of the defect
was based on the measured perimeter of the bags
following their installation The concrete was
poured into a dry hole which was cased the full
length of the shaft
Over a year after construction, Geosciences Test-
ing and Research, Inc (GTR) personnel tested the
shafts using the PISA system and submitted the re-
sults to Prof Dan Brown of Auburn University
The actual "manufactured" defects were then re-
vealed and comparisons were held between the
predicted and actual defects
4.2 Detected w. Planned Defects
Figures 8 and 9 summarize the defects as detected
by the PISA testing (on the right hand side) versus
those planned/manufactured during construction
(on the left hand side).
Figure 6. Layout of the pile screen.

Figure 7. Data collection screen.

227
 Figure 8. Presentation of manufactured defects with predicted results of PISA CSL testing system for Shaft
4 at the Auburn University test site.
The description of the manufactured defects
were for example: "Shaft 4 - soil inclusion at 11 to
13ft below the ground surface on the north side
(towards shaft 9), covers 20% of the cross-
section." As such, the defects outlined in Figures 8
and 9 are a reasonable approximation of the de-
scriptions.
Shaft 4: Two zones of defects were identified in
shaft 4 (Figure 8). The center of the upper defect
was suggested to be about 0.6m (2ft) above the
center of the actual defect with overlapping mar-
gins. The area of the defects was correctly identi-
fied though the test suggested that it is concen-
228
trated in two zones which may be the case
following the casting The lower defect was iden-
tified in the right location, however was marked as
approximately 25 6% of the cross-sectional area
versus the manufactured defect planned as 10% of
the cross-sectional area.
Shaft 9: Two zones of defects were identified in
shaft 9 (Figure 9). The upper defect was identified
as the right size but at a location approximately
0.6m (2ft) above the center of the actual location.
The lower defect was identified at the right loca-
tion and with the correct size
Shaft 2: A soft bottom in part of the cross-
 Figure 9. Presentation of manufactured defects with predicted results of PISA CSL testing system for Shaft
9 at the Auburn University test site.
section was identified in Shaft 2 This was not an
intentional defect and may have resulted from the
regular construction process
Shaft 7 A weaker zone was identified between
1 5 to 3 7m (5 to 12fi) along one segment of the
shaft (tube 2) No intended defect was installed in
this shaft The shaft, however, was loaded to fail-
ure in bendins and extensive tension cracks were
expected to be developed on the south side around
3 7m (12ft) This information, like all other infor-
mation, was provided after the tests results were
submitted
229
5 SUMMARY AND CONCLUSIONS
Four relatively small size sand inclusions were in-
stalled in two shafts out of four constructed shafts.
All the defects were identified in the tests con-
ducted over a year after construction. Three out of
four defects were identified in their approximately
correct size, the fourth defect was assumed to be
about 3 times the size of the actual defect. Two of
the defects were also identified within approxi-
mately O.Gm (2ft) of their actual locations.
Overall, the test results of Class 'A' prediction
provided accurate and reliable evaluation. The to-
mography feature of the testing equipment cer-
tainly allows an operator to estimate the extent of
the defected zone with a higher accuracy than ever
before.
The challenge of finding the defect seems to be
smaller than its accurate description. The latter,
however, is of great importance in order to be able
to conduct structural evaluation of the defected
shafi and hence to assess the need of remedial ac-
tion.
In summary, the PISA system represents a new
generation of CSL equipment capable of conduct-
ins non-destructive testing with ease and accuracy
not available before.

A C K N O K E DEMENT S

The evaluation of the PISA system was possible

with the support of the Federal Highway Admini-
stration (FHWA) The cooperation of Dr. Dan
Brown of Auburn University enabled the testing of
the drilled shafts

WFERENCES
Amr. E I and Aiiiir. J M . (1998a). "Testing Piles rvith Vir-
tual Iiistrunieiits". Proceediigs of the DFI conference.
Vieiuia. Austna. pp 4 1 1 - 1 1 4

Amir. E I aiid Ainir. J M . (1998b) "Rcceiit AdLances i i i Ui-

trasonic Pile Testing.' 3"' Iiiternatioiial Geotechiiical
Seminar 011 Bored and Augured piles. Ghent, Belgium

Baker. C N . Dnimnglit. E E . Bnaud, J . Mensah-Dnuiiiali.

F aiid Parikh. G . (1992). "Dnlled Shafts for Bndge
Fouiidatioiis". Final Report Under Contract No
DTFH6 1-88-2-00040, Federal High\%a) Administration.
Fcbruaq

Brcttmaii. T aiid Frank. M . (1996). "Coiiipansoii of Cross-

hole and Single Hole Sonic Iiitcgrit? Logging Methods".
Procecdlngs of the 5'" International Coiifcreiicc 011 t~ic
Application of Stress W ~ eLTheoq to Piles. Orlaiido.
Florida. September. pp 698-707

Clicmauskas. L R aiid Paikov slq . S G . (1 999). "Deep

Foundations Iiitcgnt! Testing Techniques aiid Case
Histories". Cir 11 Eiigmeeniig Practice Spriiig/Suiiimer
1999. ~ o l14. no 1. Boston Soclet) of Ciril Engneers.
SectiodASCE. pp 39-56

230
4 Pile-soil interaction
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Keynote lecture: Identification of soil-pile model interaction parameters

from recorded time-displacement signals
Abdallah I. Husein Malkawi & Izzaldin M.Ayasrah
Civil Engineering Department,Jordan University of Science and Technology,Irbid, Jordan

ABSTRACT: A series of pile load tests were conducted and time-displacement signals were obtained at
combinations of pile installation depths and sand’s initial densities. The signals were successfdly analyzed
employing the Logarithmic Decrement and a semi-iterative procedure for under-damped and critically damped
signals, respectively, in signal matching processes. The proposed signal matching procedure was shown to be a
powerful tool in evaluating soil-pile model parameters, more specifically damping constant and stiffness
Damping constants and damping coefficients were then plotted against pile depth, sand density, and pile static
bearing capacity that was known for every test soil-pile condition.
Based on this study, damping constants and damping coefficients were found to decrease with pile installation
depth, sand density and static resistance The static pile bearing capacity was also shown to stand for the
combined roles of depth and density.

1 INTRODUCTION and the roots of Eq.(2) are:

Damping of a material is defined as “The mechanism

by which the energy of vibration is gradually 3
converted to other form(s) of energy such as sound,
heat, noise, etc” The damping type prevailing in
pile driving problem is believed to be viscous where, (c ) is the damping constant, (k) is stiffness,
damping, in which the damping force is proportional (c) defines the damping ratio as (c/ccr), and (a,,)is
t o the velocity of the moving body, and the viscous the natural frequency.
damper is modeled by a piston-cylinder dashpot The The roots (bl, b,) are either real equal, real
dominating equation of free motion in a viscous distinct, or complex, defining the three possible types
media is of free damped motion, namely, Critically damped,
Over-damped, and Under-damped motions,
1 respectively.

where: in is the mass of the moving object, c is the 2 ANALYTICAL PROCESURE

viscous darnping constant of the surrounding media,
k is the stiffness, y is the displacement, and its first
and second time derivatives are velocity and
acceleration (v, a) respectively.
The solution of the above equation is in the form:
when substituted in Eq.(l) yields the following
characteristic equation:
mb2+cb+k=0
Described below are the analytical procedures for
under-damped and critically damped pile motions,
over-damped motion is far beyond our concern
Over-damped and even critically damped motions
are so impractical in pile engineering problems, and
the appearance of motions near the critical can only
be attributed to the relatively very small pile bearing
capacities and the reduced test pile scale It is
known, however, that critically damped motion is
quite similar to the over-damped with the only
2

233
difference in damping period, that the over-damped
motion endures more than the critically damped as
shown in Figure (1).
2.1 Under-damped Motion:
For such cases two methods can be used to calculate
the damping ratio, the first utilizes Hilbert transform
techniques either derived by Thrane (1 984),
Mohammad and Tomlinson (1988), or by Husein
Malkawi and Mohammad (1996), while the second is
the logarithmic decremental approach. The analytical
procedure for the signals in this research will be
based on the logarithmic decrement due to its
simplicity that it requires little calculations. The
logarithmic decrement (6) is defined as:
Y )
6= In( 1
Y2
hence, the damping ratio is calculated as.
6
where;yl and y2 are the amplitudes of the first and
second peaks of the oscillation, respectively.
Measuring the damping period- from the recorded
signal, between the first two successive peaks (zd),
then:
now the damping constant (C) is computed as:
c=cc = 217143
Cl'
Figure 1 Comparison between critically damped and
over-damped motions.
The values computed above along with the initial
conditions (yo, v,) are now used to obtain the
parameters of the under-damped equation of motion,
and then plotted over the experimental signal to
check for the matching between the derived and
experimental time-displacement signal.
As the value of damping constant, c is being verified,
and with the static pile capacity, R, in hands, the
viscous damping- Smith damping, coefficient (J) can
be computed as:
8
2.2 Csitically Damped Motion
For the critical damping condition the solution
algorithm is iterative; given the initial conditions (yo,
v,), the time displacement hnction becomes:
9
4
The natural frequency (a,) is adjusted in eq.(9)
until it fits best with the experimental signal, then for
the selected fitting equation:
5
3 TEST MODEL AND MATERIAL
3.1 Model Pile
The test pile is a steel, close ended, smooth pipe pile
with a geometry chosen so that its slenderness is
comparable to real pipe piles. The test pile is (90-cm)
long, having lengtwdiameter (L/D) ratio of (90/6.1 =
6
14.75), the (5 mm) thick pile-wall is capable of
sustaining piles section undeformed even after pile
driving. The test pile was also provided with a steel
pile cap, the cap has a semi-spherical notch on its
7 upper face where a steel ball (ball bearing) is placed
to insure pure axial loads, that it prevents bending
moments from occurring.
3 2 SU??dPrcpr~tres
The sand used in this study is fine to medium, poorly
graded sand (SP) named (Sweileh Sand), having the
gradation curve shown in Figure (2) This type of
sand has been chosen because it is easy to control
especially when it is a matter of density due to the
shape of particles (uniformity)
Using the sand bed preparation technique
described later, the dry density of the sand in the
loosest test state is (1 55 g/cc) and the angle of
internal friction varied in the range (26 - 30")
depending on the level of the effective stress
Whereas, in the densest, the dry density of the sand
is (1 72 g/cc) and the corresponding friction angle
laid in the range (35-40") Those friction angles
234
were obtained from direct shear test conducted on
loose and dense sand specimens at different
representative normal stress levels
to a maximum of (100-cm) because the resulting
dust may seriously affect the instrumentation of the
laboratory. Consequently, for the case of (70 %)
relative density, gentle vibrations are necessary to
achieve such high density.

Table 1 Selected Relative Densities and the

Corresponding Heights of Fall
Initial Sand D, P tl Falling
Cond. (%) (g/Cm3) Height (cm)
Loose 35 1.55 30

Medium 50 1.63 80

Dense 70 1.75 100”

Figure 2 Grain size distribution curve
3.3 Sand Container
The sand container is a square (100 x 100 cm2)
sectioned, (125 cm) high steel box, with (5mm) thick
wall, and an empty weight of 2.73 kN (278.37 kg).
Those dimensions were chosen in such away to
prevent the interaction of pile driving stresses with
the container’s either wall. This container was
manufactured in the engineering workshops of
Jordan University of Science and Technology. To
obtain precise measurements of the in tank sand
depths the tank was marked at corners and walls
mid-lengths with a metric graduation to the nearest
(0.5 cm)
* Vibration employed.
4 TESTPROGRAM
A general view for the test elements, instrumentation
and setup is shown in Figure (3). The first step in all
pile’s tests is the preparation of sand bed at the
desired test density (or relative density). The sand is
filled in the silo which is then elevated to an
elevation corresponding to the desired test relative
density obtained from Table ( l ) , and as a double
check, tank full weight is measured by the dartec
machine, and knowing the depth of sand in the
container the density is insured.

3.4 S m d Bed Preparation Techiiiqire

The model pile is to be installed in sand number of
times at combinations of installation depths and sand
initial densities Prior to every test a mechanical
technique named “Sprinkling Method” will be
employed to get the required sand density.
In this method sand is allowed to fall freely from a
silo elevated over the sand container, the desired
sand density will be function of the silo’s elevation
For every elevation, the corresponding density is
measured then the relationship between the falling Figure 3 Test elements and experimental setup
height and sand density or relative density is
determined
Table (1) summarizes the selected test relative First, test pile is driven by a square load wave of (7.5
densities, and corresponding densities for different kN) amplitude and a frequency of (20 Hz), see
silo elevations. The density is directly calculated Figure (4) The dynamic impact is applied to the pile
when final sand height in the sand container is via DARTEC machine ram which is now set to
measured, and container’s empty and full weights operate on the dynamic- strike mode, ram stroke
obtained from the DARTEC machine- when (displacement) and ram frequency are the only inputs
operating in the tension mode for those kinds of tests, chosen to be (8 mm, 20 Hz)
Silo elevation (height of fall) has been restricted respectively.

235

Figure 4 Pile driving input, square load-time wave.
To measure pile’s displacement with time a Linear
Variable Differential Transducer (LVDT) was placed
under the pile cap thus allowed to simulate pile’s
actual movement, the (LVDT) is connected to an
analogue storage oscilloscope from which the final
Time-Displacement signal was obtained. The
(LVDT) is one primary, two secondary coils and a
magnet, the motion of the magnet induces an
electrical current, this current is transformed into
displacement by the (LVDT) control panel, and the
signal is shown on the screen of the oscilloscope
The pile is forced to move for a definite forced
displacement depending on the strike of the machine
ram, then continues to move freely for a little
displacement upon which analytic works are later
conducted For the case in hands, forced
displacement can be easily measured as that
corresponding to constant Time-Displacement (t-d)
slope, i.e , for the initial linear part of the (t-d) signal
where the slope is a constant value Even when
running the machine in the dynamic mode, the ram
will move down as fast as specified by the frequency
input until it reaches a maximum forced displacement
at a high constant jacking rate, then rebounds up to
allow free damped motion of the pile
The same sand preparation technique, and test
setup were utilized in the static pile load tests An X-
Y adoptei- (Pen Recorder) was employed instead of
the LVDT to indicate pile loads and displacements
plotted on a millimeteric paper. Pile capacity is taken
as that corresponding to a displacement of 10% of
pile’s diameter
As for the expected experimental errors, they may
be comparatively insignificant The machine is
capable of measuring within 0 5 N (0 05 kg), sand
level in the container is precisely achieved by
applying gentle vibrations to the container’s walls,
and adjusted via a levelmeter allowing the
achievement of the desired sand depth with a
tolerance of few millimeters, and the oscilloscope
expected errors are probably the least.
5 DISCUSSION
Measured time-displacement signals are to be
analyzed utilizing the “Logarithmic Decrement”,
236
along with a semi-iterative parametric optimization
process The optimum parameters are fhrther judged
and checked through their match with the
experimental signals “Signal Matching” For the
purpose of evaluating damping coefficients it was
necessary to conduct static pile load tests at the same
test conditions Constant Rate of Penetration (CRP)
tests were conducted at a rate of (0 03 mm/s)
according to the ASTM (D1143-81) Ultimate pile
static resistances at combinations of pile installation
depths and sand initial density are shown in Table
(2)
Damping constants and damping coefficients are
then studied with different test parameters, pile
installation depth, sand’s initial density, weight of
soil column at the level of pile’s toe, and ultimate
static soil resistance
The under-damped motion is clear since it is
characterized by the oscillation (wave like motion),
while critically and over damped motions can only
differentiated by their damping period, as mentioned
earlier Below are detailed examples for signal
matching processes along with parametric evaluation
for two damping cases, under-damped, and critically
damped or near Each of which is analyzed
separately
Table 2 Static pile bearing capacity in (kN) as
obtained from CRP tests.
Depth 30 45 60
D,%
35 1.5 2.4 3.25
so 2.2 4.5 56
70 4.3 6.1 8.7
5 1 I I I i ~ e r . - ~ ~ ~ ~ i y e u (Refer
~ ( j t i otoi iFigure 5 )
The first step is to plot the initial tangent for the
curve starting from (0 0,O 0), this tangent match with
the initial linear portion that corresponds to constant
pile velocity, thus indicating the region of forced
motion The last point on the linear initial part of the
curve, is the first in the free movement region
The axes is now transformed to the point (t‘=O 0,
x‘=O 0) in the free motion part This point
corresponds to (t=75, x=6 4) The initial linear part
of curve is extracted form Figure ( S ) , the remaining
part represents pile’s free motion and is re-plotted on
the new axes ordinates (t’, x‘) as shown in Figure (6)
below The solid line represents the recorded
transformed signal, while the dashed is the deduced
Based on the Figures ( 5 and 6),
yo=O 3 min , v,,=89 inm/sR,= 8 7 kN
yl=O S inin, y2=0 22 mm, at z,1=36 ms
The logarithmic decrement, 6 = In yl/y2= 0 82 1
Damping ratio, = 0 1296 <
Damped frequency, atl = 0 174 rad /ms
Natural fi-equency, a,, = 0 175 rad h i s
 To check for the obtained parameters, the Similarly, Figures (7, through 10) show under-
deduced pile displacement fhnction is plotted against damped signals that were analyzed using the
the actual (experimental) signal to check for their logarithmic decrements and the equations described
match. The under-damped equation of motion is: above.

y ( t )= 0.63 e- 0.02268r cos(0.174t- 61.4)

Now that the equation of free pile damped motion
is verified and checked, the damping constant for the
system is computed as:

c= 2m<o,, = 362.9 N.s/m = 0.3629 kN.s/m, and

k = mo, = 245 kN.s/m.
and knowing the static pile capacity, the damping
coefficient is: 0.042 s/m.

Figure 6 Free under-damped pile motion at 60-crn

Figure 10 Time-Displacement signal for pile installed
pile installation depth in sand prepared at 70% D,.
at 60-cm, in sand prepared at 35 % relative density.

237
5.2 Critically DanipedMofron (Refer to Figure 11) Figures (13, through 15) are signals recorded and
The first step is the transformation of axes to the first suspected to be at Or near the
condition, and were all analyzed according by similar
non-linear point, which corresponds to the beginning
of the free motion The origin of the transformed approach.
axes has the ordinates ( t= 55, x= 4 3) as shown in
Figure ( I 2)
Initial conditions for the signal are
y,, = 0 2 mm, v, = 76 m d s ,
R,=4 5 kN
The equation of motion for the critical damping
condition is

y(f) = ( y , + (v, + w,,y,)f)e-' w,lf

Figure 13 Time-Displacement signal for pile installed

= 0.2 + (0.076 + 0 . 2 ~ , , ) t ) e - " ~ ~ '
at 45-cm, in sand prepared at 35 % relative density.
Now, the natural frequency which is the only
unknown that has to be varied until the above
equation fits best the experimental signal. The
optimal frequency for this case is ( a = 0.08
rad./ms. = 80 rad/s), then:

=1280 N s/m = 1 28 kN s/m

c = 21n a,,
and k = 51 2 kN.s/m
and, J = 0 284 s/m
where (R,) is obtained froin constant rate of Figure 14 Time-Displacement signal for pile installed
penetration static pile Tests at 30-cm, in sand prepared at 50 % relative density

Figure 15 Time-Displacement signal for pile installed

Figure I 1 Time-Displacement signal for pile installed
at 45-cm, in sand prepared at 50 % relative density
Figure 12 Free criticallydamped pile motion at 45-
cm Pile in sand Prepared at 50%
Dr
at 30-cm, in sand prepared at 35 % relative density
It is noted, however, that increasing the depth
from 30 cm to 45 cm in 35% and 50 % relative
densities (% Dr), have almost produced no effect on
the mode of motion and even on the model
parameters, that for those four combinations the
same damping constant, and stiffness were obtained
Table (3) summarizes the parameters obtained by
the analytical approach introduced herein
The variations of stiffness with depth and density
are not noticeable, specially for critically damped
motion (Figures 1 1 , and 13) One possible
justification is that the stiffness is very sensitive to
the precision by which the natural frequency is
computed, and treating those two signals as critically
damped endures some deviation from the real
frequency values, the error encountered in the

238
 frequency produce larger deviation in the value of seems to conclude the roles of pile installation depth
stiffness from the real because frequency is raised to and sand density.
the power of two. The influence of the precision of
calculated frequency is less significant on the errors
of calculated damping constant.

Table 3 Results of tile load tests

depth k C J I

45 180.0 2.40 0.967

60 115.2 0.512 0.1575
(1.63 g/cc) 5O%D,
30 53.8 1.296 0.64
45 51.2 1.280 0.284 Figure 17 Pile’s total damping constant Vs. pile
60 121.0 0.480 0.0857 installation depth for different test sand’s initial
(1.76 g/cc) 70%D, densities.
30 123.0 0.464 0.108
45 133.1 0.432 0.0708
60 295.0 0.363 0.042

Figures (16, through 19) may further help in the

study of those parameters with available test
variables. According to Figure (1 6), increasing pile
installation depth non-linearly reduces damping
constant for all test densities, and Figure (1 7) shows
that the role of pile installation depth and soil density
on damping coefficient, and damping constant can
not be considered through soil stresses (soil column
weight) at the level of pile’s toe, each one of which Figure 18 Influence of soil overburden pressure on
has to be considered separately, or look for an darnping coefficient, for all test densities.
alternative measure. This may be attributed to the
fact that, both-density and depth, affect many
variables other than the level of stress, such as soil
shear/compressive strengths, stiffness, and shear
modulus, so treating variables separately may make
the problem easier to handle.
Static pile bearing capacity is another important
factor in correlating to damping constants and
coefficients. Static pile resistance is obtained from
constant rate of penetration tests (as indicated
earlier), as shown in Figure (1 S), damping constant is
reduced by increasing the static capacity of the pile.
Yet, and as clear from this figure, the amount of
reduction is larger for looser sands than for the Figure 19 Pile’s total damping coefficient Vs. static
denser. pile capacity for different test sand’s initial densities.
As for the relationship between pile bearing
capacity and damping coefficient, it can be seen that
at any distinct sand density the relation is perfectly 6 CONCLUSIONS
linear, while for the whole set of data the damping
coefficient is a power function (ax”) of bearing On the basis of the experimental work, the signal
capacity, J = 4/R? process introduced in this study as well as the
Similar to damping coefficient, the damping correlative study, the following conclusions are
constant is a power hnction of Static pile bearing made:
capacity as shown in Figure (19), staic pile capacity The use of logarithmic decrement in under-
damped motion and the semi-iterative signal

239
matching in critically damped motions are verified to
be easy and powerfd in evaluating some of the soil-
pile model parameters such as; damping constant,
darnping ratio, and stiffness.
Both, damping coefficient and damping constant
decrease with pile depth of installation, sand’s
density, and static pile bearing capacity.
Static pile bearing capacity concludes pile
installation depth and sand’s initial density, while soil
pressure at the pile’s toe level does not.
The predictability of stiffness is dominated by the
accuracy of predicting the natural frequency.

ACKNOWLEDEMENT

This research is part of the second authors MS thesis

supervised by the first author and it was supported
by the Dean of Research at Jordan University of
Science and Technology.

REFERENCES

ASTM D 1 143-8 1, (1998) “Standard Method of

Testing Piles Under Static Axial Compressive
Loads”, Vol. 04.08, Philadelphia, pp. 179-1 89.
Clough, R. W., and Penzien, J., Dynamics of
Structures, McGraw-Hill, Inc., New York, N. Y.,
1975.
Husein Malkawi A. I. & Mohammad, K. S., (1996)
“Estimating Damping Constant of the Pile-Soil
System Directly From Measured Displacement
Using Hilbert Transform Technique”, Fifth
International Conference on the Application of
the Stress-Wave Theory to Piles, Orlando Florida,
September, pp. 37-54.
Litkouhi, S. & Poskitts, T. J., (1980) “Darnping
Constants for Pile Driveability Calculations”,
Geotechnique, 30(1), pp.77-86.
Mohammad, K. S. & Tomlinson, G. R., (1988) “A
Simple Method of Accurately Determining the
Apparent Damping of Non-Linear Structures”,
Proceedings of the 6th International Modal
Analysis Conference, Florida.
Rao, S. S., (1995) “Mechanical Vibrations”,
Addison-Wesley Publishing Company, New York,
N. Y., Third Edition, 1995., pp. 97-140.
Thrane, N., (1984) “The Hilbert Transform”,
Technical Review No.3, Bruel and Kjaer.

240
 Applicatjon of Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Load transfer analysis from increasing energy dynamic load tests in concrete
piles driven in very soft clay formation
J. Balech & N.Aoki
University of Siio Paulo, Siio Carlos, Brazil

ABSTRACT: This paper presents the results of dynamic loading tests performed on concrete piles driven in
very soft clay formation. The load transfer mechanism under blows of growing energy is described in terms of
mobilized resistances (friction and point) and displacements at the top and at the base of the piles. A compari-
son between measured static and dynamic load - displacement curves shows that they are in good agreement.

1 INTRODUCTION correlation with the resistance-displacement curve

obtained in the dynamic test. The measured load
1.1 Constant energy dynamic loading test transfer for all the piles are also presented (Balech,
The dynamic loading test is usually performed with 2000).
constant energy blows and each blow is analyzed by
CASE and CAPWAP methods when the monitoring
is done with PDA system (Gobble et al., 1996). The 2 SOIL DESCRIPTION
following results are obtained: a) total, friction and The soil profile for concrete hollow piles at sites A
point resistances; b) load distribution along depth; c) and B are presented in tables 1 and 2.
friction and point quakes; d) maximum displace-
ments, velocities, acceleration at various sections; e)
........1......Soil
Table
.....A........
....,. ...profile
............. at ................A
~....site ..,,........~..........."".', .................. .,... ........................
maximum force and energy. Soil characteristics i Thickness j
I N
......................................................................... ........... I?>........... .:. . B!OW!!C.~.
i.. ..
Soft, dark brown fine to medium 1.80 i 2/30
1.2 Gradual increasing energy dynamic test silty sand with roots
...................................................................................................................................
Very soft, gray, organic : 14.20 17/746
The dynamic loading test used in Brazil is related to micaceous silty clay
...................................................................................................................................
the gradual increasing energy blows procedure (Aoki Dense, gray, clayey fine and me- ! 2.20 10/30
1989, Bernardes 1989). Each blow is analyzed dium sand with gravels
...................................................................................................................................
according to with the constant energy dynamic load- Very stiff, mottled, micaceous: 0.20 j 2
ing test. From this analysis it is possible to obtain the ..c l a ~ e ~ .~s l~~l n, .a ~ s..es.dual.soll s~c .... ~................................i....... ..
variation of the many variables described in section
1.1, under growing energy levels. The rupture load Table 2. Soil profile at site B
....................................................................................................................................
can be characterized when the mobilized resistance Soil characteristics :i Thickness : N
remains constant after two blows of increasing en- i ............Im>.............;..Blows!cm..
..........................................................................
ergy (Aoki, 1997, Aoki & Cintra, 1997) Sofi, dark brown fine to medium I 1.00 1 1/64
silty clay with many roots
..................................................................................................................................
Very soft, gray, organic i 16.36 12/834
micaceous silty clay
......................................................................... ............................ :........................
1.3 Concretepiles in very soft clayformation Medium, gray, sandy medium ! 0.64 1 5/15
clay with gravels
This paper presents the analysis of typical results of ........................................................................ :................................. > .........................

gradual increasing energy dynamic loading test per-

formed on six hollow precast reinforced concrete Water level depth = 0.0 m (sites A and B) *impenetrable soil
piles, driven into gnaissic residual soil underlain by a
thick soft clay formation. All these piles are driven
with free fall hammers and the dynamic loading tests 3 PILE LOAD TRANSFER DIAGRAMS
were monitored and analyzed by using the PDA sys-
tem methodology. Static SML loading test and the 3.1 Piles andpile driving system
load-displacement curve for site B pile show a good The pile driving system is presented in table 3:

241
 Pile driving system Pile type Pile type
_A_ R
I

Hammer type Free fall Free fall

Hammer weight (kN) 50 34.2
Pile cap weight (kN) 3 2
Cushion thick. -hard wood (m) 0.30 0.30
Cushion thick. - plywood (m) 0.06 0.06

The main characteristics of the hollow reinforced

concrete piles are presented in tables 4 and 5 .
Table 4. Main pile characteristics

Concrete fck
Steel fvk 500 500

Pile Diameter Driven length Time after driving

(cm> (m) (days)
A1 50 19.3 20

3.2 Mobilized normalforces diagrams

When the hammer hits the pile the soil resistance is
mobilized and the maximum skin friction and point
resistance vary in accordance with the nature and
rheological characteristics of different soil layers
found along the depth. The normal force diagram re-
sulting from this impact, obtained from CAPWAP
analyses performed for each blow of increasing en-
ergy applied to the piles, is presented in figures 1 to
6. Except for pile A5, the normal force diagram for
all the piles is approximately constant along the soft
clay layer thickness, due to the low

Figure 4. Mobilized normal forces diagram: pile A4

Figure 1. Mobilized normal forces diagram: pile A1

242
 of the measured displacement at the top of the pile,
for the same energies of figures 1 to 6. In the case of
piles A l , A3, A4 and A5 (Figures 7, 9, 10 and 1 l), it
was observed that: a) in order to mobilize the maxi-
mum unitary local friction in the soft clay layer a
maximum displacement of about 10 mm, is neces-
sary; b) in the sandy layer, the maximum local fric-
tion is mobilized for displacement of about 20 mm.
For the piles A2 and A3 (Figures 8 and 9), the same
trend is observed but, in the last blows, the mobilized
skin friction diminishes and becomes equal to zero in
the case of pile A2. This fact can explained by the
presence of a very soft clay (muddy soil) in the upper
layer.

Figure 6. Mobilized normal forces diagram: pile B 1

mobilized soil resistance along the pile shaft. Al-

though being in the same site A, the different behav-
ior of pile A5 can be explained by a local variation of
the soil profile. In the case of piles Al, A2,A3 e A4
(Figures 1,2,3 and 4) the measured normal force dia-
gram can be explained by the transition between the
soft clay and the residual sandy soil layer. A sudden
loss of friction resistance can also be observed in the
case of last blow applied in pile A2,where all the re-
sistance was transferred to the point. For pile B1 em-
bedded in very soft clay (Figure 6) the friction resis-
tance was negligible and only point resistance was
mobilized for all the blows.

3.3 Mobilized unitary localfriction resistances

Figures 7 to 12 present the mobilized local friction Figure 8.Unitary local friction diagram: pile A2
resistances, measured at various depths below
ground (DBG) indicated in the figures, as a hnction

243
 Figure 12. Unitary local friction diagram: pile B 1

3.4 Local friction resistance aiong depth

The measured unitary local friction resistances at

each depth below ground (DBG), for the indicated
growing energy blows, are presented in figures 13 to
18. These diagrams show how the local skin friction
is mobilized along depth.
As a general trend, at each applied energy, the 10-
cal friction varies according to the nature and resis-
tance of the soil layer. It can be observed that, in the
case of piles A5 and B 1, the unitary local friction re-
sistance decreases with depth instead of increasing as
in the case of piles AI, A2,A3 and A4.
In the case of pile B1 the maximum skin friction
resistance is as low as 8 kPa corresponding to a very
soft clay with an SPT resistance equal to zero (actu-
ally 50 to 80 centimeters of sampler penetration, un-
der the static action of SPT hammer + rod weight).

Figure 13. Local friction diagram along depth: pile AI

Figure 11. Unitary local friction diagram: pile A5

244

Figure 16. Local friction diagram along depth: pile A4
It is emphasized that the CAPWAP analysis for
each blow of increasing energy was done without any
attempt to adjust the friction resistance distribution
t o the known soil conditions.
3.5 Point resistance x point displacement and set
Figures 19 to 24 present the relation between the
mobilized point resistance and the total and perma-
245
Figure 18. Local friction diagram along depth: pile B1
nent (set) displacements of the point of the piles, for
growing energy level. It can be observed that the
elastic displacement (quake of pile point) is depend-
ent on the applied kinetic energy level.
The shape of the total point displacement curve
suggests that the point of piles A1 and A4 has appar-
ently reached the layer which is impenetrable to per-
cussion.
3.6 Point percentage on the total resistance
The percentage of point resistance on the total mobi-
lized resistance, for each blow of growing energy, is
shown in figures 25 to 30.
Piles AI, A3 and A4 present the usual behavior
where the lateral friction is mobilized before the point
resistance for small system deformations and, then,
the pile resistance is gradually transferred to the base
(Kezdi, 1975).
Pile A2 shows the same usual behavior until the
energy level correspondent with the last blow. In this
event the residual lateral friction is reduced to zero
and all the resistance is represented by the point.
Pile A5 shows that the lateral friction is much
greater than the point resistance in all sequence of
growing applied energy.
Finally in pile B1 for all blows of growing energy,
the point resistance predominates.
4 CORRELATION BETWEEN STATIC AND
DYNAMIC PILE LOADING RESPONSE
All the equivalent static load transfer diagrams ob-
tained in this site, from dynamic loading test per-
formed with gradual increasing energy, show a gen-
eral trend that is in good resemblance to those
obtained with instrumented piles subjected to static
load test. The reliability on the application of a dy-
namic loading test procedure to obtain information
about the pile behavior under static loading has
grown with the increasing utilization of this type of
test all around the world. Otherwise the correlation
between static and dynamic load test results in the
same pile can be made and, in doing so, it would be
possible to improve our understanding of when dif-
ferent soil conditions occur. Such correlation was
done in this site and the results are presented in
figure 3 1 .
Figure 3 1. Resistance versus displacements curves: pile B 1
The dynamic test was performed in 1 day and the
static test was carried out three weeks after the dy-
namic one. The ultimate load of the pile is greater
than 1000 kN. The extrapolation done by the usual
method (Van Der Veen, 1953) indicated an ultimate
load of about 1050 kN in both cases.
5 CONCLUSIONS
The constant energy dynamic load test is inadequate
to measure the behavior of a driven pile as the only
piece of information obtained from this test is a point
in the resistance - displacement curve.
248
The so-called quake of the point (elastic displace-
ment resulting from the elastic deformations of the
soil under the base of the pile) is directly dependent
on the amount of energy reaching the soil under the
pile point. It can not be considered a soil parameter.
The ultimate load resulting from the static load is
in good agreement with the ultimate load obtained in
the gradually increasing energy dynamic load test.
The load transfer diagrams obtained in dynamic
load test show a good resemblance to those obtained
in instrumented static load tests. In the soft clay layer
the lateral skin friction grows up to a maximum and
thereafter diminishes to a residual value when the pile
is gradually subjected to increasing applied energy.
6 REFERENCES
Aoki, N. 1989. A new dynamic load test concept. In Japanese
Society of Soil Mechanics and Foundation Engineering
(eds): International conference on Soil nzechanics and
foundation engineering, 12. Drivability of piles, 2: 1-4. Rio
de Janeiro, Brazil
Aoki, N. 1997. The evaluation of the ultimate bearing capacity
of driven piles by using increasing energy dynamic load
tests (in Portuguese). Doctorate thesis. Sa"o Carlos School
of Engineering, University of Sa"o Paulo, Brazil.
Aoki N. & J.C.A. Cintra 1997. New interpretations of the dy-
namic loading curves for driven piles based on the energy
approach. In Almeida (eds), Recent developnzents in soil
and pavement nzechanics: 467-472. Rotterdam: Balkema.
Balech, J. 2000. Analysis of dynamic loading tests performed
in concrete piles driven in soft clay formation. (to be pub-
lished in Portuguese). Master of Science dissertation. Siio
Carlos School of Engineering, University of Sa"o Paulo,
Brazil.
Bernardes, G.P. 1989. Dynamic and static testing of large
models piles in sand. Doctorate thesis. Departnzent of
Civil Engineering. Norwegian h t i t u t e of Technologv,
Trondheim, Norway.
Gobble, G.G. & Likins Jr. 1996. On the application of PDA
dynamic pile testing. In F.C. Townsend (eds): Interna-
tional Conference on the Application of Stress- Wave The-
ory to Piles, 5: 263-272. Florida.
Kezdi, A. 1975. Pile foundations. In H.F. Winterkorn & H.
Fang (eds), Foundation Engineering Handbook: 556-600.
New York: Van Nostrand.
Van der Veen, C. 1953. The bearing capacity of a pile. Inter-
national Conference on Soil Mechanics and Foundation
Engineering, 3. v. 2: 84-90. Switzerland.
 Applicationof Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Dynamic pile testing and finite element calculations for the bearing capacity
of a quay wall foundation - Container terminal Altenwerder, Port of Hamburg
E Kirsch, B. Plapmann, T. Huch & W. Rodatz
Institute for Foundation Engineering and Soil Mechanics, Technical University of Braunschweig, Germany

ABSTRACT: In April 1999 the construction of a new quay wall in the port of Hamburg started. In a first stage
two new berths with an overall length of 1400 m will be built. In order to proof the bearing capacity of the
foundation dynamic pile testing is performed to a great extent. To date more than thirty different piles were
tested by the Institute for Foundation Engineering and Soil Mechanics of the Technical University of
Braunschweig. In order to check for possible set-up effects redriven tests were performed. For comparison
reasons static load tests were also carried out, one of which was done with an instrumented pile to allow skin
friction and end bearing to be evaluated separately. The CAPWAP results proofed to be very helpfid especially
in those cases where the testing situation differed from the later service condition. Special questions called for
detailed finite element analysis of the load-settlement behaviour of single piles in certain construction stages.

1 CONTAINER TERMlNAL ALTENWERDER 1.2 Sire conditions

The construction of the quay wall is done from the
1.1 The project dry. The 24,5 m draught necessary for modern con-
tainer ships will be achieved only after major dredg-
The port of Hamburg is one of the biggest and most
ing works (Fig. 1). Therefore it is necessary to install
important amongst the harbours in Europe. Its geo-
all load bearing elements of the quay wall by an ex-
graphical and infrastructural position makes it very
tensive pile driving operation with depths to ap-
interesting as a gate from the North Sea to Europe.
proximately 30 m.
This stands especially for the east of the continent
since the Baltic Sea cannot be shipped with the big
container ships of the latest generation. Since 1980
the total freight handling in Hamburg has grown from
63,l million tons to 75,8 million tons in 1998. In the
same time the share of contained freight has grown
from 1 1 % to 48 % (Freie und Hansestadt Hamburg
1999). This underlines the importance to provide of
sufficient space for containers and container ship
berths
As early as 1973 first planning started to realise
the extension of the port of Hamburg in the area of
Altenwerder, which is situated south of the well
known Kohlbrand-Bridge, one of the landmarks of
Hamburg. It did cost a lot of political endeavor until
in April 1999 the construction of the quay wall itself
could commence. In a first phase two new berths
with a length of approximatetely 800 m will be built
until the year 200 1. Finally a total length of 1400 m
will give four modern container ships the opportunity Figure 1. Cross section (Freie und Hansestadt Hamburg 1999).
t o unload their goods.

249
 the site investigation programme and as chosen for
design purposes are given in Table 1.
Approximately half way along the quay wall axis the
depth of the boulder clay reduces and the structural
elements of the foundation are embedded into it. This
was not addressed as a problem since the soil investi-
gation gave reason to expect sufficient bearing ca-
pacity. Figure 3 gives the geological situation along
the axis of the quay wall.
Table 1. Soil parameters
Twc ’{ y’ 9‘ C’ Es
kN/mi kN/in3 O kN/ni’ MN/iii’
Landfill 19 11 27.5 0 8.0
Sand 18 10 32.5 0 40.0
Clay 17 7 25.0 10 2.0
Moulmeat 14 4 17.5 10 1.5
Gravel 19 10 35.0 0 100.0
Boulder clay 22 12 30.0 20 25.0

Figure 3 . Geological situation (Miller 1999)

1.3 Construction details

Figure 2. View on the construction site (Freie und Hansestadt
Hamburg 1999).
In the first phase approximately 1300 cast-in-place
concrete piles with a total length of 32 km are driven
into the ground to support the quay slab and the
beam of the gantry crane tracks. In order to speed up
the construction of the wall and in particular the front
piles these structural elements are placed into 32 m
deep trenches and are then driven a firther five me-
ters to final position. Figure 2 gives a view on the
construction site in Summer 1999.
Subsoil conditions in the north of Germany are
characterized by the effects of the last glacial period.
Under deposits of marine sand and gravel with layers
of silt tertiary boulder clay with different stiffnesses is
encountered. Typical soil parameters as revealed by
The quay slab rests on tubular steel piles (0
1 2 1 9 16
~ mm) and a mixed sheet pile wall with dou-
ble planks (€32 975 A) and intermediate planks
(AZ18-10) as well as on a total of 1300 driven cast-
in-place concrete piles 0 51 cm. Horizontal stability
is ensured by 46 m long inclined steel piles (HTM
600/136) driven at 2,27 m intervals. The rear crane
rail, at a distance of 35 m from the front rail, is
founded on a row of inclined driven cast-in-place
concrete piles (Wittwer & Krefi 1999).
1.4 Pile instaiialioiz
The steel pipes and the double planks are placed into
a slurry trench filled with a cement-bentonite suspen-
sion. This mixture should hardens and achieves ap-
proximately the same qualities as the surrounding

250
soil. Afterwards the steel profiles are driven by a
15 tons Menck MHF 10-15 hammer for a hrther five
meters into the subsoil to ensure sufficient bearing
capacity. For the driving of the anchor piles and the
cast-in-place concrete piles hydraulic hammers of ap-
propriate capacity are used.
2 DYNAMIC PILE TESTING
2.1 lest r.eslrlls
Up to now 24 restrike tests were carried out on dif-
ferent concrete piles All test results were obtained
by the CAPWAP procedure. Figure 4 gives a graphi-
cal overview of the load calculated capacities split
into skin friction and end bearing.
Figure 4. CAPWAP results of the driven cast-in-place piles
A field test allowed the comparison between the
CAPWAP results and the results of a previously
performed instrumented static load test, where skin
friction and end bearing were measured separately.
The instrumentation consisted of a load cell at the
end of the pile and stress measurement sections in
different depths along the pile axis. Figure 5 shows
the load settlement curves calculated by CAPWAP
and those measured during the static load test.
Figure 5. Load-settlement response of a concrete pile.
25 1
The tests at the driven steel profiles and tubular
steel piles respectively were done at the end of the
driving operation and also as a restrike test three
weeks after installation. Up to now a total of 23 tests
with CAPWAP evaluation were carried out. Figure 6
shows an example.
Figure 6. CAPWAP result of a tubular steel pile
The reason for the need to gain information about the
skin friction distribution along the pile axis by CAP-
WAP analysis lies in the specific construction of this
quay wall foundation and the regulations of the com-
petent authorities. Each individual pile has to be able
to carry its load with an appropriate factor of safety
even in the case of a failure of the wall. Due to pos-
sible horizontal movement of the wall the soil behind
fails by developing shear bands. Thus it may loosen
and subsequently reduce its frictional capacity.
Therefore all resisting forces acting along the pile
above an assumed line of failure must not be taken
into account for the determination of the bearing ca-
pacity. This assumed line of failure starts at the toe of
the sheet pile wall with an inclination of approxi-
mately 1.2 and goes in upward direction into the
backing soil.
It is obviously important to distinguish between
skin fiction and end bearing in cases where piles are
being excavated and exposed to water later on.
2.2 Special soil condition
As shown in Figure 3 parts of the foundation are em-
bedded in the boulder clay which was addressed by
the soil investigation as a relatively stiff clay. In con-
trast the systematic pile tests revealed a significant
reduction in the bearing capacity of all structural
elements in that area. It became obvious that the clay
was far softer than expected. Carefbl adaptation of
the piles design became necessary and different
structural solutions were investigated by means of
dynamic pile testing. In case of the concrete piles
their lengths were extended as well as the redrive
method was adopted. For the steel piles a series of
different pile shoes with wings and filling sheets was
tried to cope with the situation (Fig. 7).
Figure 7. Different pile shoe designs
In this critical phase the dynamic pile testing proofed
its usefulness. Fast performance and the relatively
low costs of dynamic pile tests made the redesign of
the foundation during construction possible. In the
end this observational method using the different
CAPWAP results produced an optimised foundation.
3 SET-UP EFFECTS
For all types of piles a certain time dependent devel-
opment of the bearing capacity was expected. As the
design especially for piles embedded in the boulder
clay turned out to be difficult, more effort was spent
in the investigation of possible set-up effects.
Since the bentonite slurry hardens with time the
increase of the skin friction along the steel profiles is
evident. In fact the resisting forces became so high in
some cases that it was impossible to gain any vertical
displacement during the restrike tests.
For the concrete piles a gain in bearing capacity of
10% to 20% within two month after pile installation
was encountered (Fig. 8).
The reduction in the bearing capacity and its subse-
quent regain of pile D182 was object of intense
studies. In order to proof that the reduction could be
explained not by a general loss of strength of the sur-
rounding soil but by the disturbance of the ground
adjacent to that special pile by the recent cutting of
the slurry trench, numerical simulation of the con-
struction process was carried out.
252
Figure 8. Set-up of the driven cast-in-place piles
4 FINITE ELEMENT ANALYSIS
4.1 General remarks
Numerical simulation in geotechnics today is object
of profound investigation. The characteristics of the
soil as a material led to special solutions on the side
of constitutive laws for geomaterials as well as in the
practice of finite element calculations. Numerous
publications deal with the numerical simulation of
piled foundations and their problems (e.g. El-
Mossallamy 1999, Maybaum et al. 1999). For the fi-
nite element analysis of the subject case the ANSYS
program was used. In the last years specific imple-
mentations into the program were realised at the In-
stitute for Foundation Engineering and Soil Me-
chanics of the Technical University of Braunschweig
to adjust it to geotechnical problems (e.g. Vittinghoff
et al. 1997, Planmann et. a1 1999).
4.2 Niimerical modelling
The geometrical modelling of the problem was
guided by the use of specially designed macros,
which allow the easy variation of different parame-
ters. Due to the spatial nature of the situation a
three-dimensional discretisation was necessary. Using
the symmetrical nature of the problem the inclined
pile D182 was cut in half and the adjacent slurry
trench with the surrounding soil and appropriate
boundary conditions was modelled as shown in Fig-
ure 9. To allow for the calculation of the primary
stress field as well as for the different construction
stages double elementation became necessary. To
model the soil-pile interaction properly areas with
high stress gradients were meshed finer than other
zones. The finite element mesh consisted of a total of
1988 brick elements with second order shape func-
tions.
The soil behaviour is simulated by an elastoplastic
constitutive law. The Drucker-Prager yield criterion
governs the plastic deformations. The influence of
the choice of the constitutive law was subject of for-
mer studies (Maybaum et. a1 1999). Material proper-
ties of the slurry and the concrete were chosen ac-
cording to prior investigation (Heinrich 1998).
 The simulation of the construction process is done
by a step-by-step analysis beginning with the in-situ
stresses, then modelling trench and pile installation
and finally the stepwise loading of the pile The load
was applied in steps of 200 kN up to failure
Two different situations were investigated and the
findings were compared with the results of two dy-
namic pile tests. One took place before and one after
the cutting of the slurry trench adjacent to pile D 182
Figure 9 Finitc clcrncnt mcsh for pilc D182 with adjacent
trench
4 3 Results
With the finite element analysis it was possible to
proof that the reduction in the bearing capacity was
solely caused by the installation of the slurry trench,
which occurred in this particular case after the pile
driving and is therefore only a local effect in time In
general all pile installation took place after the cut-
ting of the trench and therefore no reduction of the
bearing capacity had to be expected
Figure 10 shows the influence of the open trench
o n the surrounding soil The shaded areas represent
Figpie 10 Horizontal displacement due to open trench bcforc
pile installdtion
253
the horizontal displacements of the soil towards the
slurry filled trench before pile installation It is appar-
ent that the pile is located in the area of influence
Figure 1 1 shows the vertical displacement of the
soil around the axially loaded pile in the state of fail-
ure
Figurc 11 Vcrtical displaccmcnt duc to loadcd pilc
The gained by finite ele-
ment analyses as well as obtained by the dynamic pile
tests before and after the slurry trench installation are
given in Figure 12
Figurc 12 Load-scttlcincnt ciin cs of pilc D 182
5 CONCLUSIONS
The subject project is an example of extensive use of
dynamic pile testing Adaptation of the pile geometry
due to the change of the soil properties became nec-
essary when pile tests revealed insufficient bearing
capacity The quick and economical nature of the dy-
namic pile testing allows a relatively easy redesign of
a piled foundation The results of dynamic pile tests
can be evaluated using a three-dimensional finite
This the Of local
effects due to different construction stages
REFERENCES
Freie und Hansestadt Hamburg 1999. Hamburger Hafen -
Zahlenspiegel.
Freie und Hansestadt Hamburg, Wirtschaftsbehorde Strom-
und Hafenbau 1999. Container Terminal Altenwerder -
Neubau Kaimauer, I. Bauabschnitt.
Miller, C. 1999. Aspekte zur Planung der Kaianlage. Hunsa
10/1999.
Wittwer, G. & Kref€t, R. 1999. Bau der Liegeplatze 1 und 2.
Hansu 10/1999.
El-Mossallamy, Y. 1999. Load settlement behaviour of large
diameter bored piles in over-consolidated clay. In G.N.
Pande et al. (eds), Numerical hdodels in Geomechanics -
NUWOG J87I. Rotterdam: Balkema.
Maybaum, G., Vittinghoff, T. & Rodatz, W. 1999. On the
bearing capacity and the servicability of piled rafts. In G.N.
Pande et al. (eds), Numerical Models in Geoniechanics -
NUMOG L7I. Rotterdam: Balkema.
Vittinghoff, T., Rodatz, W. & PlaRmann. B. 1997. Integration
of a Numerical Model in an Observation Programme. In
Jian-Xin Yuan (ed.), Coniputer Adethods and Advances in
Geoniecl~anics.Rotterdam: Balkema.
PlaBmann. B.. Kirsch, F.: Vittinghoff, T. & Lohr, M. 1999.
Implementierung infiniter Elemente im ANSYS Open
System und deren Anwendung bei Halbraumberechnungen
in der Geotechnik. In CAD-FEM GmbH (eds). 17. CAD-
FEN users' meeting.
Heinrich. 0. 1998. Einfld des Bodenkontinuums a& die
Entwicklung von Verschiebungsgrofien bei FE-Analysen.
Thesis, unpublished.

254
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3

The shaft dynamic response of a pile in clay: Induced pore pressure

A.Benamar
Laboratoire de Me'canique, Universite' du Havre, France

ABSTRACT: Driving of pile generates the transient shear in soil and can induce excess pore water pressure if
the soil is cohesionless. The shaft dynamic response of a single pile-model in clay is examined in a laboratory
study. Using the principles of stress waves propagating in a one-dimensional media, stress and velocity data
have been analyzed. A pore water pressure sensor is placed in the soil in order to assess the influence of such
parameter in the mobilization of shaft resistance during driving. The experimental study concerns both
drained and undrained clay. The parametric study show that the most significant effects associated with pore
water pressure are the reduction of the maximum skin fiiction allowed at the interface. This reduction being
less important than induced excess pore pressure.

1 INTRODUCTION test) are performed. Table 1 give the physical and

mechanical properties of the material obtained under
Pile foundations usually are necessary for most of a pressure of consolidation ( T ~= 500 kPa. The sam-
situations requiring construction of buildings on soft ple is 0.5 m in height and 0.2 m of diameter. The
ground. If driving is possible and allowed, driven clay is confined in a triaxial cell under a controlled
piles mostly are prefened. Due to pile driving in co- isotropic pressure ranging from 200 kPa to 500 kPa.
hesive, low permeable soils high excess pore water The hammer blow is obtained by various available
pressure can be induced around the pile. The state of falling masses. The impact of such masses on the
soil stress after driving a pile results from the addi- rod produces a rectangular pulse with various am-
tion of the following states: initial state of stress of plitude and duration (Benamar 1992).
the soil before pile driving, stress change during pile
driving, stress change after pile driving (reconsoli-
dation). The change of stress in the soil during and
after pile driving must be known for the estimation
of the shaft capacity of piles being driven in cohe-
sive bearing soils. In practice use estimation of ex-
cess pore water pressure in advance represent a
cheep help to avoid damages at buildings, at pile
driven before, or at slopes in the vicinity of the
driving area. This paper deals with the development
of excess pore water pressure induced by pile driv-
ing in cohesive soil.

2. LABORATORY DRIVING TESTS

In order to study the mobilization of skin friction

along driven pile in laboratory, a steel rod, 9 m long,
is driven through a clay sample reconstituted around
the pile model (figure 1). The clay samples are made
with kaolin and laboratory tests (triaxial and vane Figure 1: Experimental pile driving set-up

255
 Table1:Soil characteristics (oC= 500 Wa)
I Plasticity limit I 31 % 1 age allows seepage of water induced by excess pore
Plasticity index 23
Compression index 0.70
I Water content I 39 % I 3. EFFECT OF PILE DRIVING IN CLAY
I Shear strength 65 W a
Before each driving tests series ten hammer blows
are performed in order to avoid the "healing" effect
in the pile-soil interface and making the soil under
current driving conditions. The piling is operated at
a steady blow rate of one blow per minute approxi-
mately. Several clay samples are tested in various
conditions. At the end tests, water content and
undrained shear strength are measured using for the
second parameter the laboratory vane test. The evo-
lution of this characteristic with water content is il-
lustrated on figure 2. This evolution is quite linear
for the homogeneous reconstituted clay: the
undrained shear strength decreases linearly when the
water content increases.
Figure 2: Soil shear strength versus vater content.
A pore water pressure sensor is placed into the
soil sample against the steel rod in order to measure
excess pore water pressure at the interface during
driving. But the sensor is moved during the consoli-
dation about 2 centimeters from the interface. Nev-
ertheless, the measurements of excess pore water
pressure in the soil near the pile interface can give
interesting information concerning the evolution of
effective lateral stress during driving. Strain gages
are glued on the steel rod and stress waves are
measured in order to derive pile shaft resistance.
256
During driving tests the pile-soil interface can be or
not drained at the top and bottom sample. The drain-
pressure along the steel rod.
3.1. Displacement of soil due to pile driving
Driving pile through a soil layer is always combined
with deformation of the soil adjacent to the pile. In
non-cohesive soils voids decrease due to compaction
effects of driving, while the volume of cohesive soils
remains constant due to lower permeability and the
dynamic nature of solicitation. Near the pile partially
high excess pore water pressure arises. It decreases
after the end of driving process due to reconsolida-
tion of soil. According to Steenfelt et al. (1981),
during pile installation significant horizontal move-
ment of soil occurs even in a distance up to ten pile
radius f?om the pile axis. At this distance move-
ments of the magnitude of 0.05 times the pile radius
were measured. In the experiments reported the pile
displacement is measured but not soil displacement.
Results obtained for pile displacement are discussed
in further section.
3.2.Excesspore water pressure due to pile driving
Pile driving generates transient loading in the soil.
When the rate of loading exceeds the rate of pore-
water pressure dissipation (controlled by the perme-
ability), the dilative or contractive tendency of a soil
does not manifest, and shear takes place at constant
g lobal vo lume. The vo lume-confinement-shear do-
main of soil behavior prescribes that shear at con-
stant volume can only occur if the effective con-
finement changes, therefore, there will be pore-
pressure changes. In non cohesive, high permeable
soils pile driving induces void reduction since the
drainage is easy. On the other hand, in cohesive, low
permeable soils the deformation during pile driving
occurs at constant void and then develop excess pore
pressure, leading to a reduction of effective shear re-
sistance of the soil.
Due to the fact that pore pressure measurements
are difficult and expensive different methods to es-
timate excess pore water pressure due to pile driving
were given by several authors (Vesic 1972,
Randolph et al. 1979), using different stress-strain
soil theories, and assuming the pile driving being
modeled as the expansion of a cylindrical cavity.
Few authors suppose that the excess pore pressure is
constant within the plastic zone around the pile. But
local measurements showed a relatively wide differ-
ence between measured and calculated values. The
method according to Randolph et al. (1979) was
 modified by Cunze (1989) in consideration the soil tip resistance, the undrained Young’s modulus is re-
properties, stress state, changes of stress and of placed in the formula proposed by Cunze (1989) by
geometrical conditions. The excess pore water pres- the shear modulus G. The overburden load being de-
sure is expressed as a general function of the initial rived as the difference between the dynamic load
stress state, geometrical conditions and soil condi- stress (product of impedance pile Z = 12.7 kN.s/m
tions. He proposed the following equation for pre- and pile velocity) and the confining pressure.
diction of excess pore water pressure within the soil
plastic zone around the pile, due to the volume ex-
pansion caused by introducing the pile volume into Table 2. Estimated excess pore pressure for 4 tests
the soil: v I G’VO I P’nc I P’nc-P’cs I Au 1
(m/s) (Wa) (kPa) (@a) (kPa)
= 1.l(p’nc- p p ’ +1.
* c s ) -~ P 2 c u E (1) 0.99 105.6 77.4 42.6 135.7
rp1 -ro 1.21 195.7 142.8 78.6 161.8
With: 1.57 343.2 250.5 137.8 205.1
P ’ , ~= mean effective stress of normally consoli- 2.21 605.4 441.9 243.1 281.9
dated soil.
p’cs = mean effective stress of soil reaching the criti-
cal state. The excess pore pressure measured during these
C, = undrained shear strength four tests does not exceed 66 kPa. This large differ-
rpl= radius of the plastic zone ence between measured and estimated values is due
r = radius of the pile to the fact that Cunze equation is derived for both
p = distance to the pile axis. shaft and pile tip interaction and mainly based on
Vesic (1 972) expressed the radius of the plastic zone expansion of cylindrical cavity. These results show
as: in one hand that the equation proposed (1989) is
$I = G raj< more appropriate to estimate the excess pore pres-
sure at the immediate vicinity of pile tip, and in an
other hand that the excess pore pressure developed
Where G/C, is the rigidity index. around pile shaft during driving is very low in com-
The change of effective stress due to pile driving can parison with pile tip. The presented theory of pore
be estimated (Cunze 1989) using the following pressure prediction does not integrate all parameters
equation: affect the excess pore pressure, particularly the wa-
p ’ n c - p ’ c s zz p’nc(O.83-0.51L.L.) (3) ter content and the steady state water pressure in the
With: soil before piling. Note that the example treated by
Cunze (1989) concerns a clay with 90 ‘Yowater con-
1 1 s
p nc 2K0)
= -0-YO(1+ (4) tent, while the tests reported in this study are per-
3 formed on clay with a maximal water content of
c ~ = overburden
’ ~ ~ load 37%.
L.L.= liquid limit Figure 3 below shows the evolution of excess pore
IQ,= coefficient of earth pressure at rest. pressure for ten successive identical hammer blows.
The excess pore water pressure expected due to pile There are large values of excess pore pressure at the
driving is determined by equation (1) using four two first blows and following blows induce a rela-
driving tests as follows: tive low excess pressure. Because of no drainage in
soil: normally consolidated clay the soil the pore pressure reaches rapidly a maxi-
liquid limit LL = 0.54 mum value. The first blow corresponds in this situa-
plastic limit PL = 0.31 tion to restrike causes a weak displacement before
water content W=35% allowing large displacement at the two following
coefficient of earth pressure at rest KO= 0.60 blows. As excess pore pressure do not varies signifi-
undrained shear strength C, = 57 W a cantly after the three first blows, the displacement
rigidity index G/C, = 47 remains constant. As long as excess pore pressure
confining pressure P = 300kPa develop near the pile-soil interface, the effective
Pile: radius ro = 0.01 m stress must diminish. But this behavior is not ob-
inspection point (distance) p = 0.03 m served (figure 3). This situation can be explained by
the fact that no such large excess pressure develop at
From equation (2) we obtain rP1 = 0.07 and equa- the interface because of transverse drainage in the
tions (1) is used .to calculate excess pore pressure cylindrical clay sample. This effect is verified by
values represented on table 2 below, where v is the water content measures in the sample after tests.
pile velocity. Because in the experiments reported There is more water at midradius from the pile that
here the pile is sheared through the soil without any at interface.

257

"We 3: Of Pore Pressure and
pile displacement during driving.
It is evident that in piling the displacement in-
creases with hammer energy. Using driving tests re-
suits involving several hammer energies, the evolu-
tion of excess pore pressure and pile displacement
with driving energy is investigated (figure 4).
Figure 4: Influence of driving energy on pile dis-
placement and excess pore pressure
Contrary to the pile displacement, the evolution of
excess pore pressure is not affected by hammer en-
ergy and the same behavior as in figure 3 is pre-
sented here. The amplitude of pore water pressure
being more large for high energy. The excess pore
pressure reaches rapidly an equilibrium state and no
more important increase is possible.
258
3.3. Influence of excess pore pressure on pile shaft
resistance
The large capacity increases with time (set-up) are
considered unusual for sandy soil. In cohesive soil
the fine grain content seems to control the perme-
ability preventing normal rapid drainage conditions.
Low unit shaft friction was observed during driving
and was higher on restrike after reduction in pore
water pressure (Likins et al. 1992). The soil set-up is
the result of dissipation of the piling induced excess
pore pressure in the soil after the initial driving. Fig-
ure 5 shows the evolution of both excess pore pres-
sure and unit shaft friction during a driving se-
quence. The shaft resistance is derived using stress
wave measurements near the pile top (Benamar,
1996). Such quantity is calculated as the difference
between incident and transmitted stress wave in the
pile-soi] interface. There is a large decrease of ex-
cess pore pressure at the first blow before reaching a
steady state of constant excess pore pressure. But the
unit shaft friction does not follow the same evolu-
tion, showing a low variation. When the excess pore
pressure decreases the effective stress in the soil
must increase. The measured unit shaft friction does
not increase by the same amplitude than excess pore
pressure decreases. In clayey soil the excess pore
pressures are still dissipating. Therefore, the unit
shaft friction is smaller than the value expected after
some fixther wait. There is no direct relation be-
tween the variation of such two parameters. Thus,
the shaft resistance can not be directly related to the
effective stress in the soil as defined by Terzaghi's
concept. Note that the average pile displacement
corresponding to these measurements is 4.85 mm.
Figure 5 : effect of excess pore pressure on pile shaft
resistance
 4. EFFECT OF SURCONSOLIDATION ON IN-
DUCED PORE PRESSURE DURING DRIVING
The clay reconstituted sample was consolidated un-
der a confrning pressure of 500 kPa and driving tests
were performed under an isotropic stress of 300 kPa.
The variation of shaft resistance with the overcon-
solidation ratio (OCR) was investigated (Benamar et
al. 1993) and an increase of 30 % in unit shaft resis-
tance is noted for a value 1.67 of OCR. Driving piles
is than more hard in overconsolidated soils. The in-
fluence of overconsolidation on the induced pore
pressure in soil is carried out (figure 6). For an OCR
value of 1.67, the excess pore pressure is higher in
normally consolidated soil. This difference is much
more marked at the first blow, excess pore pressure
being four times more important in normally con-
solidated soil. In the rest of driving sequence the dif-
ference of induced pressure is less important, the ra-
tio being approximately equal to 2. The driving tests
are performed in each case after a wait, leading con-
solidation to be made. The surconsolidation makes
the soil sample in low stresses conditions and wave
stress is not sufficient to reach stress statelevel as in
normally consolidated soil. These results explain
perfectly the conclusions concerning the shaft fiic-
tion mentioned above. Nevertheless the increase of
unit shaft resistance does not reache the same pro-
portions as excess pore pressure. The such parameter
can not be directly related to the shaft resistance. An
effective stress method of analysis for estimating
dynamic pile skin fiiction taking into consideration
the induced pore pressure can not be adjusted, using
effective stress theory.
Figure 6: Induced pore pressure in normally consoli-
dated and overconsolidated clay
259
5. CONCLUSIONS
The experimental set up in laboratory allowed an in-
vestigation of the influence of pile driving in in-
duced pore pressure. Such experiments are difficult
to perform in field piling. The dimensions of driving
laboratory model can affect the interpretation of in-
duced pore pressure measures, but the general con-
clusions concerning the influence of various pa-
rameters seems to agree. The variation of pore water
pressure in the soil during driving do not induce the
same variation on the dynamic shaft resistance. The
experimental results show that very small excess
pore pressures are developed due to dynamic shear-
ing of the pile (pore pressures less than 2 % of the
consolidation stress). The pile driveability is not
strongly affected by the excess pore pressure which
partially dissipate in the pile-soil interface. The
measured pile displacements seem very low by
comparison with the applied dynamic energy, im-
plying high shaft fiiction. Development of excess
pore pressure depends on the initial state of stress in
the soil before driving. Parametric studies show that
the most significant effects associated with pore
water pressure are the reduction of the maximum
skin fiiction allowed at the interface and the altera-
tion of the pile response due to slippage. The shaft
resistance can be fully mobilized easily without pore
water pressure than with it.
REFERENCES
Benamar, A., P. Lepert & D. Levacher 1992. Physi-
cal and numerical simulation of lateral shaft fiic-
tion along a driven pile, Proc. 4th Int. Con$ on
the Application of Stress-wave Theory to piles,
The Hague, The Netherlands, Sept 1992., pp. 37-
42, Balkema.
Benamar, A. & D. Levacher 1993. CaractCristiques
de l'interaction latirale argile-pieu durant le bat-
tage. Annales de 1 YTBTP, juin 1993, No 5 14, pp
55- 65.
Benamar A. 1996. Dynamic soil resistance from pile
driving analysis. Proceedings of the Europeafi
conference on Structural Dynamics, Eurodyn '96
Florence, Italy, 5-8 June 1996, Vol. 2, pp. 1039.
1043, Ed. Balkema.
Cunze, G. 1989. A modified method for the
estimation of excess pore pressure generated bj
pile driving. Proc. Int. Con$ on Soil Mechanic:
and Foundation Engineering, Rio de Janeiro
August 1989, vol. 2, pp. 1097 - 1100, Balkema .
Likins, G., J. DiMaggio, F. rausche & W. teferra
1992. A solution for high damping constants ir
sands. Proc. 4th Int, Con$ on the Application o,
Stress-wave Theory to piles, The Hague, Thc
Netherlands, Sept 1992., pp. 117-120. Balkerna.
Randolph, M.F., J.P. Carter & C.P. Wroth. 1979.
Driven piles in clay - The effects of installation
and subsequent consolidation Geotechnique,vol.
29, Nr. 4, pp. 361 - 393.
Steenfelt, J.S., M.F. Randolph & C.P. Wroth 1981.
Instrumented model piles jacked into clay. Proc.
10. Int. ConJ:on Soil Mechanics and Foundation
Engineering, Stockholm, vol. 2, pp. 857 - 864.
Vesic, A. 1972. Expansion of cavities in infinite soil
mass. Journal of geotechnical engineering. ASCE,
vol. 98, SM2.

260
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 1503

An investigation of pile diameter influence in the bearing capacity

on Dynamic Load Test (DLT)
G.F? Bernardes
Department of Civil Engineering, U N E S t Guurutinguetci,Brazil
C.S.Andreo & C.GonGalves
Benuton Pile Foundation Company,SGo Puulo, Brazil

ABSTRACT: During the construction of five residential buildings in the city of Taubate, State of S3o Paulo,
it was possible to carry out one comprehensive investigation of the behavior of precast concrete piles in clay
shales. This paper describes the results of Dynamic Load Tests (DLT's) executed in three piles with different
diameters and with the same embedded length. The tests were monitored using the PDA@(Pile Driving
Analyzer) and the pile top displacement was measured by pencil and paper procedure. From the curves of
RMX versus DMX resulted fkom CASE@method, CAPWAPC@analyses were made for signals where the
maximum mobilized soil resistance was verified. The results were compared with the predicted bearing
capacity using the semi-empirical method of Dkcourt & Quaresma (1978) and DCcourt (1982) based on SPT
values and the description of the soil profile. Some comments related to the values of quake and damping
used for clay shales in the analyses are also presented

1 INTRODUCTION the high stress levels developed during driving

which might reach the pile element strength. Results
The semi-empirical methods of Decourt-Quaresma of Dynamic Load Test in soft clays (Gonqalves at al.
(1978) and D6court (1982) for prediction of the 1996) indicate that even for profiles with NI less
bearing capacity of single precast concrete piles than 3 a pile can have considerable bearing capacity
based on the SPT values is eequently used in Brazil if the effect of set up is taken into account.
because of its simplicity. For design purposes the The point unit stress resistance can be estimated
ultimate load (&It) is usually separated into two by the expression
components; the lateral or skin load (Rlat) and the
point load (hoe)as follow: qO=C.N, (kPa) (3)
where N, represents the average of SPT measured at
base level and the ones measured one meter above
where Alat and At, are respectively the lateral and and bellow the pile point. The characteristic factor
base bearing areas and (fs) and (qo) the Unit shaR and "C" can be estimated from Table 1 as a fkction of
point resistance expressed in stress units (Ha). the soil type.
According to Decourt (1982) the lateral unit stress
(fs) can be calculated by the equation
Table 1 : Characteristic factors "C"
Soil Type "C"
(Ha)
Clay 120
Clayey Silt 200
where NI is the average of SPT values along the pile Sandy Silt 250
shaft. In the above equation there is no distinction of Sand 400
soil type and it is recommended NI values of 3 and
50 as lower and upper limits of SPT values. Cintra The factors presented in Table 1 represent
and Aoki (1 999) suggest for precast concrete piles a average values obtained in static load tests
value between 15 to 35 for upper limit depending on performed in different geological formations. During
pile diameter. These values take into consideration the investigation of the behavior of precast concrete

261
pile in clay shales formed in Cenozoic age, Taubate
Basin, S.P. (Bernardes et a1.1999) it was possible to
v e r that
~ the bearing capacity prediction was too
conservative while the prediction for steel piles in
clay shales from "Ilhas" Formation in Bahia
(Velloso and Hammers 1982), showed exactly the
opposite situation. Therefore, it is recommend to
propose factors and constants to be used in this
semi-empirical method related to a specific
geological formation.
In this paper, the influence of the pile diameter in
the bearing capacity is investigated. The predictions
using DCcourt & Quaresma Method is compared
with the dynamic load test results executed in
precast concrete pile in clay shale.
2 GEOLOGICAL AND GEOTECHNICAL
CONDITION OF THE SITE
The area of study is located in Taubat6 City founded
over Terciary sediments formed in Cenozoic age
known as Taubat6 Basin. These sediments cover an
area of approximately 170 km in length and 20 km
wide forming the Paraiba River Valley. The
representative soil profile of the sediments near the
test area is presented in Figure la.
Figure 1. Soil characteristics and bearing capacity predictions
with depth.
The upper layer of sandy clay fill with dirt is 2.70
m thick overlying a 1.10 m thick layer of very soft
organic clay, black, with fragments of roots. For
262
depth between 3.80 to 11.00 meters it was
horizontally bedded clay shale formation, plastic,
dark gray to graphite. Underneath, the soil layer
consist of stiff clay shale with no sign of lamination,
high plasticity, green, until the end of SPT boring at
19.00 meters deep. The ground water table was
measured at one meter below the ground surface.
The bearing capacity predictions for the clay shale
layer based on DCcourt & Quaresma Method
(Fig.1b), gives 54.0 kPa and 4720.0 kPa for average
unit shaft and point resistance respectively.
3 DYNAMIC TEST PILE PROGRAM
Three pile with different diameters (23, 26 and
33 cm) were driven down to 9.00 meters deep. These
test piles were precast concrete piles having a yield
stress of 35.0 MPa, Elastic Young Modulus of
30.15 GPa and speed wave propagation of
3510.0 m/s. The driving equipment was a free fall
drop hammer with 20 kN ram weight for the 23 and
26 cm pile diameters and a 28 kN ram weight for the
33 crn pile diameter. A capblock of hard wood and a
cushion made with Masonite fiber plate were used
together with a steel helmet to reduce the high
impact stresses. This system guarantees a driving
efficiency between 45 to 55%.
The piles were monitored using a Pile Driving
Analyzer (PDA@) during pile installation and
restrike after six days. The pile top movement was
also measured by pencil and paper technique.
First, the dynamic load test method (Aoki 1989)
at restrike was used to plot the mobilized resistance
by CASE@ method (RMX) versus the maximum
displacement of the pile top (DMX), considering
damping factor of 0.4 s/m, in other to select the best
signals that guarantee full mobilization of the soil
resistance (Fig. 2). Further analyses of the mobilized
soil resistance are related to the three tests indicated
in Figure 2. One complete investigation of the
dynamic load test for 23 cm pile diameter is
presented elsewhere (Bernardes et al. 1999). The
signals of force (F) and velocity multiplied by pile
impedance (ZV) that will by use are presented in
Figure 3. As a simple procedure to check the
measured signals and the CAPWAPC? analyses, the
displacement from double integration of the
accelerometer was compared with the direct
measurement of the pile top displacement by paper
and pencil.
For all three piles, it can be seen from Figure 3
that after the time L/c there is a significant increase
of F and corresponding decrease of ZV related to a
high shaft resistance in the clay shale.
The results of the CApWApC@analysis made
with the signals from Figure 3 are presented in
Table 2.
Figure 2. ( b x ) versus (Dmx) from Case@Method with J=0,4

It can be seen from Table 2 that the analyses gave

a very little lateral quake (0.4 mm), probably related
to the high stiffbess of the clay shale, and point
quake in order of D/60. The lateral damping was
practically constant (0.240 s/m) and the point
damping was very small which as expected for
failure condition (Gonqalves et al. 1998). The
quality match ("Mqno") varies between 3.30 and
3.49.

4 ANALYSIS OF THE MOBILIZED SOIL

RESISTANCE

The simulated static load test from CAPWAPC for @

the test piles is presented in Figure 4.It can be seen

that increasing the pile diameter the s t s e s s of the
initial part of the load-displacement curve increase
as the maximum mobilized soil resistance.
This behavior might contribute for the evaluation
of pile displacement under working condition. For
one specific foundation, reducing the number of
piles and increasing theirs diameter can cause a
significant reduction in the foundation displacement.
The soil resistance distribution mobilized along
each pile element is presented in Table 3 . The last
five elements, embedded in the clay shale, support
most of the lateral resistance, between 80 to 90%.
This behavior was already expected based on the
dynamic signals from Figure 2.
According to Table 2, the lateral unit resistance is
equal to 127.0 kPa that is 135 YO higher than
predicted in Figure I-b. For this particular
investigation, it seems that the increase of pile
diameter and consequently the lateral soil
Figure 3. The measured force and velocity multiplied by
impedance.
displacement during pile installation, do not alter the
lateral unit resistance and the bearing capacity is
only proportional to the lateral area. According to
Eq. 2, a constant factor of 20 could be use instead of
10. For unit point resistance, the measured average
values of 4330.0 kPa is 10% lower than the
prediction using Eq. 3 with characteristic factor of
120 and the bearing capacity is also proportional to
the base area. Based on the effective stress analysis,
the skin bearing coefficient (Ns) and the bearing
capacity factor (Nq) are 2.2 and 55 respectively
which are congruent for s t Sclay.

263
Table 2: Summary of the CAPWAPCO analyses
Pile
Diam.
Wp Wr H Ef K i t DMX Qt Qs Js Jt MQno
(kN/m) (kN) (cm) (kN.m) (%) (kN) (mm) (mm) (mm) (s/m) (s/m) (%)
(cm)
23 0.95 20 80 8.7 54.4 687 14.7 3.9 0.39 0.249 0.114 3.46
26 1.25 20 100 9.9 49.5 813 13.8 4.0 0.40 0.239 0.098 3.30
33 2.05 28 140 18.4 46.9 1176 16.8 5.3 0.43 0.210 0.010 3.41
Obs.: W, = weight of pile; Wr = weight of ram; H = ram fall height; EMX = measured energy;
Ef = efficiency; Rult= ultimate resistance; DMX = top displacement; Qt = point quake; Qs = lateral quake;
J, = skin damping; J, = toe damping; MQno = signal match.

Table 3: Summary of the mobilized soil resistance

Skin Resistance Distribution / pile segment (kN)
Rlat Rlat f, Rtoe qo
1 2 3 4 5 6 7 8 9 (m) (shale) (shale)
~ i l t

(W
(W (W) (W (@a)
Diam.
4 23 cm 0.0 3.0 17.5 28.0 44.0 76.0 97.5 114.0 128.0 508.0 459.0 127.2 179.0 4308.0 687.0
4 2 6 cm 0.0 10.0 25.0 45.0 64.0 85.0 107.0 124.0 128.0 583.0 508.0 125.6 230.0 4332.0 813.0
4 33 cm 0.0 22.0 38.0 82.0 93.0 133.0 141.0 147.0 148.0 804.0 662.0 127.8 372.0 4349.0 1176.0
Obs.: Rlat = lateral resistance; Rlat(shale)= lateral resistance supported by the shale deposit; f, (shale) = unit skin
resistance; Rt, = point resistance; qo = unit point resistance; = ultimate resistance.

Pile Top and Point Load (kN) is proportional to the lateral and base areas. The
Decourt & Quaresma Method predicted very well
300 600 900 1200 the point resistance and under estimated the shaft
resistance. For fbture foundations construction in the
same region where site investigations identifjr this
h clay shale deposit, it can be use a factor of 20 can be
E used in Eq. 2.
W
E 5.0
.w
5
? 10.0
REFERENCES

3a Aoki, N.1989. A new dynamic load test concept.

Drivability of Piles, Proceedings of the Discussions
Session 14, XII ICSMFE, Technical Committes on
.,- 15.0
3 Pile Driving, 1:1-4.
p1

Bernardes, G.P.; Gonqalves, C.; Andreo, C.S. and

20.0
Figure 4. Simulated load deflection curves from Capwapc@.
5 CONCLUSIONS
Here, the dynamic load test with energy increasing
(DLT) was used to investigate the influence of the
pile diameter in the bearing capacity of precast
concrete piles in clay shales. The DLT tests were
also used to compare the resulted mobilized soil
resistance fiom CAPWAPCO analyses with the
predictions using the semi-empirical method of
DCcourt & Quaresma based on SPT values.
These analyses indicated that the unit shaft and
point resistance is constant and the bearing capacity
Fortunato, S.G.S. 1999. Evaluation of bearing
capacity of precast concrete pile in shale fiom
Dynamic Load Test (DLT). XI Panamerican Con$
on Soil Mechanics and Geotechnical Engineering:
3 :1491- 1496. Foz do Iguap: Brazil.
Cintra, J.C.A. and Aoki, N. 1999. Carga admissive1
em fimdag6es pro fundas. Internal Publication
EESC-USP: 61p. Siio Carlos: Siio Paulo.
Decourt, L. 1982. Prediction of the bearing capacity
of piles based exclusively on values of the SPT.
Second European Symposium on Penetrating Test:
1:29-34. Amsterdam: Holanda.
Decourt, L. and Quaresma, A.R. 1978. Capacidade
de Carga de Estacas a partir de valores de SPT. VI
Congress0 Brasileiro da Mecbnica dos Solos e

264
Engenharia de FundaqGes, 1 :45-54. Rio de Janeiro:
Rio de Janeiro.

Gonqalves, C.; Andreo, C.S.; Bernardes, G.P. and

Fortunato, S.G.S. 1998. Prova de carga dinhica
em estaca pre-fabricada apoiada em areia argilosa
densa. XI COBRAMSEG, 3:1535-1541, Brasilia: Siio
Paulo

Gonqalves, C.; Andreo, C.S.; Bernardes, G.P. and

Fortunato, S.G.S. 1996. Estudo da reduqiio do
comprimento de estacas pre-fabricadas atraves da
m a k e do "set-up" corn base na instrumenta$io
dinAmica com o PDA@. Terceiro Seminririo de
Engenharia de FundaqGes Especiais e Geotecnia:
2:35-46. Siio Paulo: Siio Paulo.

Velloso, P.P.C. and Hammers, M. 1982. Estudo da

cravaqiio de estacas metalicas em folhelhos da
formaqiio ilhas, na Bahia. VII Congress0 Brasileiro
de Mecbnica dos Solos e Engenharia de Fundaqbes:
3:325 - 338. Olinda: Recife.

265
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Pile set-up in sands

J. lc!Seidel
Monash University,Melbourne, Vic., AustraLia
M. Kalinowski
Baulderstone Hornibrook, Sydney,N S.K, AustraLiu

ABSTRACT: This paper describes the results of dynamic and static load testing at a site where open-ended
steel tube piles were driven into a dense to very dense sand layer. The pile capacities on driving were signifi-
cantly less than anticipated fiom the design calculations, with shaft resistances of 5 to 20kPa measured even
at 40m depth. Later testing confirmed development of shaft resistances in excess of 15OkPa. A mechanism
of pile whip or flexure is hypothesized. Resistance fiom the internal soil column was also investigated.

1 PROJECT DESCRIPTION which occurred in the Tertiary and Quaternary Ages.

The Bolte bridge is a new crossing of the Yarra river
in Melbourne Australia. The bridge is a six-span
structure which provides a new north-south fieeway
link to the west of the central business district. The
two river crossing spans were constructed by bal-
anced cantilever method fi-om the Piers 2, 3 and 4.
The construction for the outer spans used conven-
tional closure pours.
The geology at this site is a complex alluvial se-
quence of deposition, erosion and redeposition
An inferred geotechnical profile is shown in Figure
1.
The subsurface sequence encountered at Pier 4
consisted of the following sequence:
0 3m of recent fill
0 6m of medium dense sand (Port Melbourne
Sand, PMS)
0 6m of slightly overconsolidated marine
sediments (Coode Island Silt, CTS)
0 l l m of highly overconsolidated silty clay
(Fisherman’s Bend Silt, FBS)

Figure 1. Inferred stratigraphy at the Bolte Bridge site, showing approximate formation boundaries and in-situ
strength test results

267
 e 8m of dense clayey and gravelly sands (Mo-
ray Street Gravels, MSG - the lowest Qua-
ternary stratum)
e 3m of dense sands (Newport Formation, NF
-the highest Tertiary stratum)
e 13m of dense to very dense sands and sandy
gravels (Werribee Formation, WF)
e bedrock - Melbourne mudstone of Silurian
Age (Silurian, S)
The sequence at Pier 3 was generally similar, al-
though the upper layers are predominantly absent
because of the dredged river channel. A thin, ex-
tremely weathered Tertiary Age Basalt layer (TOV)
was also intersected at Pier 3. Estimates of
undrained shear strength and the results of SPT
testing at Piers 3 and 4 are also shown in Figure 1.
All piers are supported on large groups of driven
open-ended steel tube piles. The piles are 1200mm
diameter with a wall thickness of 20mm. No addi-
tional reinforcing plate was used at the pile toe. De-
sign pile founding levels (by others) were RL-43m
(43m below mean sea level) for the 56 piles at Pier
3, and RL-42m for the 18 piles at Pier 4 on the south
bank of the river. In both cases the target founding
stratum was the Werribee Formation comprising
dense to very dense sands. An ultimate capacity of
25MN was computed for piles at both piers. Allow-
ance had been made for downdrag in the design.
37 CLAYEY GRAVELLY SAND (SC-GC) very dense, grey-
green, fine to coarse sand, subanbdar gravel N>50 (33.45,
38 Increasing sand content
39
40 SAND (SM-SP) very dense, green-my, some silt, trace
clay; gravel medium to m a n e grained
41
Gravelly band noted at 41.4m Occasional geen-grey silty
clay layers
42 (SM-SC) slightly clayey, some gravel
Silty clay layer to 43.1rn
43 Gravel band comprising anbdarquam and mudstone noted
to 43.5171
(SP-SM) my, fine to medium grained, some silt with fri
44 able mudstone gravel
4s (SP) light grey, some silt, rare quartz gravel
46
Grey-brow1 clayey sand layer to 46.4m
(SP) Grey, coarse-grained, slightly gravelly
47
48
Light grey, trace of fines
Figure 2. Borelog excerpt from the founding Werribee Forma-
tion at Pier 4
Of particular importance, both to the pile design
and to the dynamic pile test data discussed hereafter
were the soil conditions in the founding Werribee
Formation. A borelog close to the initial piling lo-
cation at Pier 4 is shown in part in Figure 2 at the
depth of the Werribee formation. Note in particular
that the SPT N values were 50 or greater throughout
the layer. An automatic trip hammer device was
used. It is also emphasized that the formation at this
location is predominantly sand, with evidence noted
of some (<5%) silt and traces (4%) of clay. Occa-
sional layers of slightly clayey sand were found.
Piles were driven with a purpose-built 16 tonne
Juntann hydraulic drop hammer. Preliminary Wave
Equation analysis had indicated that this hammer,
with a maximum stroke of 1 . 2 would ~ be required
to reach the target founding depths. A hammer with
a larger energy rating would be required to demon-
strate the required ultimate capacity.
Testing at the site consisted of monitoring 15% of
piles using a Pile Driving Analyzer@and 3 static pile
load tests.
2 TRIAL PILE DRIVING
Driving of the first test pile commenced at Pier 4.
This pile, denoted TPl penetrated very easily to the
target depth. A hammer stroke of only 400 to
500mm was required - significantly less than the
maximum stroke of 1200mm. This was unexpected,
given previous wave equation predictions. The
hammer operated very well, with energy transfers
5135mm)
consistently in excess of 90%. Table 1 shows the
driving record for the last 3m of driving of TPl to
N>50 ( I 7,28,
target depth. Depth intervals are 250mm. The en-
221100mrn)
ergy figures are those determined from proximity
switches built into the hammer.
Table I Driving record above target depth for TPI
From(m) To(m) Blows Energy Set(mm)
30185mm)
(No) (t-m)
39.75 40.00 23 8.2 10.9
40.00 40.25 27 6.7 9.3
42.50 42.75 24 7.2 10.4
The pile was allowed to set-up at this depth for a
matter of some hours. As the indicated capacity was
still low, and the penetration per blow still large, the
268
pile was driven a further 4 metres beyond the target
founding depth. The pile was restruck at 48 hours (2
days) and 96 hours (4days), and demonstrated some
capacity increase, but insufficient to satisfy design
requirements. CAPWAP@ analysis of the data indi-
cated shaft resistances typically 5 to 10 kPa in the
dense to very dense sands on driving, increasing to
between 30 and 40kPa after 96 hours. These were
significantly less than the resistances predicted by
geotechnical analysis, as indicated in Figure 3.
Figure 3. Comparison of inferred shaft resistance distributions
for TPI at driving and restrike with design pile shaft resistance
Test-pile TPI was subsequently driven to a depth
of 48.5m, at which time the capacity indicated by
dynamic pile testing was still significantly lower
than expected. Despite this, a decision was made to
perform a static test on the pile at this depth. Reac-
tion was provided by anchors founded in the mud-
stone rock a significant depth below the pile toe. A
prediction of the pile capacity was made fiom
C A P W Aanalysis.
P~ The prediction and the results of
the load test are shown in Figure 4. Plunging failure
commenced at a load of 5.5 MN, and a load of more
than 6MN could not be sustained. The comparison
between prediction and static test is excellent.
Due to program requirements a decision was
made to drive all subsequent piles for the project to
the basement mudstone rock underlying the Wer-
ribee Formation sands. It was not feasible to leave
piles in the sands in the hope of further and SUE-
cient set-up effects.
269
Figure 4. Comparison between CAPWAP@
prediction and static
load test for TP 1
3 OBSERVATIONS DURING CONTRACT
DRIVING
The majority of further pile driving and testing
occurred at Pier 3 , the central river pier. Pile driving
was undertaken from a large sheet-piled cofferdam
which had been filled to above high water level. Re-
sults fiom testing of Piles 301 and 307 will be fur-
ther described in this section.
It is noted that the Pier 3 pile cap was approxi-
mately 12.5m x 52m in plan. Pile 301 was a corner
pile, and had two adjacent neighbor piles at 2.5m
centre-spacing. Pile 307 was an edge pile, again
with two adjacent piles at 2.5m centre-spacing.
3.1 Pile 301 test sequence
Pile 301 was tested on initial drive on 16 November,
1996. Subsequent restrikes were performed on 25
November (9 days), 30 November (14 days), and 8
December (22 days). In order to ascertain the rela-
tive amounts of internal and external shaft resistance
on these open-ended tube piles, Pile 301 had been
drilled out to within 6 metres of the pile toe between
the last two restrikes.
The force-time and velocity-time responses for
each case, including two blows for the final restrike
of this pile are compared in Figure 5. Note that the
pile was installed with the Juntann 16 tonne hydrau-
lic drop hammer. All restrikes were performed with
a 20 tonne trigger release drop h a m e r capable of
3.5m maximum drop height.
 The different hammer characteristics are evident
from comparing the input waveform of the fxst rec-
ord against all the other records.

3.1.1 Pile 301 - “Set-up”

It is evident to a person trained in the analysis of
dynamic pile testing records that there is a signifi-
cant increase in capacity, and particularly in shaft re-
sistance from the driving case to even the frst re-
strike after 9 days.
A capacity increase of this magnitude had not
been observed for the driving of the test pile at Pier
4 in very similar stratigraphy.
Table 2 summarizes the total capacity and the
shaft and toe capacity contributions for Pile 301 at
the time of each test.

Table 2. CAPWAP@
mobilized capacity summary Pile 301
End of Restrike Restrike Restrike Restrike
Drive #I #2 #3 #3
B10w2 Blow6
mobilized resistances (in MN)
Shaft 10.4 15.5 15.9 21.4 17.4
Toe 2.7 1.5 1.1 .5 4.0
Total 13.1 17.0 17.0 21.9 21.4

The shaft resistance distribution estimated by

CAPWAP@ analysis for all tests is shown in Table 3
for each 2m segment of the 5 l m of pile penetration.
It is noted that the shaft resistance distributions are
not purported to be exactly correct. However, the
total magnitude of shaft resistance is expected to be
reasonably accurate, and the distribution is expected
to be reasonably indicative of the variation of shaft
resistance with depth at the time of testing. The
availability of energy to hlly mobilize capacity can
also affect the apparent shaft resistance distribution,
particularly near the pile toe.
It is furthermore noted that these shaft resistance
values are computed on the basis of the external pe-
rimeter alone. No account has been taken of the in-
ternal shaft resistance or the sum of internal and ex-
ternal perimeters. That is, the implication is that all
resistance is external. The authors believe that this
is normal convention for dynamic pile testing. To
the authors’ knowledge, there is no substantive evi-
dence on the relative contributions of internal and
external shaft resistance.
In examining this table, it is evident that shaft re-
sistances at the end of driving were extraordinarily
low over most of the pile length, excepting the bot-
tom 6 metres (last 3 rows of Table 3), which corre-
spond to penetration into the weathered Silurian
rock. To a depth of 47 metres, pile shaft resistances
were uniformly less than 20 @a, and as low as 2
kPa in deposits that were found to be dense to very
dense.

270
 mobilized shaft resistances for Pile 301
Table 3. CAPWAPO As can be seen fiom Table 3, the estimated shaft
Depth End of Restrike Restrike Restrike Restrike resistances had generally, and substantially in-
below Drive #1 #2 #3 #3 creased between the second and third restrikes. In
ground B10w2 Blow6 the lower 24m of pile shaft, the inferred average
mobilized shaft resistance by depth (in kPa) shaft resistance increased from 155 kPa to 198 kPa
1 5 over this period. Even by the 6'h blow, at which
3 5 4 2 2 time there was a noticeable reduction of shaft resis-
5 5 5 13 13 tance, the average shaft resistance over the lower
7 6 6 7 22 22
9 9 7 II 25 26 24m had decreased to 16OkPa - still above the in-
11 10 9 11 27 16 ferred shaft friction at the time of the second re-
13 10 10 I1 26 16 strike.
15 9 12 16 18 19 It would appear that the loss of resistance due to
17 10 10 15 10 II removal of the internal soil column was more than
19 14 10 17 15 15 compensated for by other factors which served to in-
21 19 10 40 47 17
23 20 10 24 35 53 crease the external resistance in the intervening pe-
25 18 79 26 94 73 riod.
27 13 76 65 71 65 In addition, the average shaft resistance values
29 8 76 106 113 36 computed for Pile 301 are not noticeably less, and in
31 4 67 I06 81 39 some cases even exceed the shaft resistance values
33 2 89 160 165 150
35 4 105 178 216 167 for the other corner piles 315, 342 and 356 which
37 9 103 188 216 167 had the same configuration of surrounding piles.
39 16 17 93 216 167 The internal soil columns of these piles remained
41 13 17 161 216 167 intact.
43 13 17 186 216 196 Although no definitive recommendation can be
45 13 17 162 216 196 made from this data, it would appear that the internal
47 91 20 186 216 196
49 200 91 169 255 222 shaft resistance was not a major component of the
51 167 159 169 255 222 total shaft resistance for this project. If it had been a
significant component, some noticeable loss of shaft
For the first restrike, 9 days after driving, the resistance would have been expected, and the shaft
shaft resistance increased by approximately 50%, resistance for this pile would have been less than
however, this increase was very uneven. Almost no that measured for equivalent piles in the group.
increase was observed to a depth of 25m or beyond a
depth of 37m. Very large increases were, however, 3.2 Pile 307 - Load-settlement prediction
observed in the Moray Street Gravel and Newport
Formations. Although these formations contain A fkrther static pile load test was undertaken at Pier
some thick clayey bands, they are predominantly 3 on Pile 307 in order to confirm the validity of the
gravelly sands with little or no fines content. dynamic pile testing at this site, particularly in view
The second restrike at 14 days after driving sug- of the unexpectedly low shaft resistance values
gests continuing increases in shaft resistances below measured during driving.
29m, and a sudden and substantial rise in shaft re- Driving of Pile 307 was completed on 14 Decem-
sistance in the Werribee Formation sands. At this ber, 1996. Restrike testing of the pile was -per-
time, shaft resistances in the Werribee Formation formed on 8 January, 1997 - 25 days later. Pro-
were indicated as 160 to 180 kPa - well in excess of gramming constraints only allowed the pile to be
the design assumptions. statically load tested on 4 February - 52 days after
installat ion.
3.1.2 Pile 301 - Assessment of internal shaft resis- Two analyses were undertaken on the restrike
tance data for Pile 307 - on the first and fifth blows deliv-
After the second restrike was performed, a decision ered. The capacity estimates for the two blows were
was made to try to estimate the relative proportions similar, however, CAPWAP@analysis indicated a
of internal and external shaft resistance, for the pur- trend of decreasing shaft resistance and increasing
pose of estimating the reliability of the shaft resis- toe mobilization for the test sequence. The predicted
tance predictions. load-deflection responses for both blows were pro-
The internal soil column within Pile 301 was vided in advance of the static load testing as sug-
therefore removed to a depth of 6m above toe level. gested upper- and lower-bounds to the expected test
The internal water level within the pile was main- response. A comparison of predicted and measured
tained close to the pile head. The third restrike test responses is shown in Figure 6. The two predictions
was performed within 1 or 2 days after removal of bound the load test measurement. It is noted that the
the internal soil column. ultimate pile capacity was not actually measured in
either dynamic or static test, so no direct comparison

27 1
of ultimate capacity is possible. However, the load-
deflection prediction is, as stated, very good.
Figure 6. Predicted and measured load-movement responses
for Pile 307.
4 DISCUSSION
The project is notable for a number of reasons:
1 It provides additional documented 'Class A'
comparisons between dynamic and static load
testing, in particular for open-ended steel tube
piles.
2 The shaft resistances of the tube piles during in-
stallation was significantly lower than anticipated
in design, or than would be considered reasonable
by conventional geotechnical analysis.
3 The project provides some evidence, albeit indi-
rect, regarding the proportion of internal and ex-
ternal shaft resistance of steel tube piles
4 The shaft resistance of piles at this project in-
creased dramatically (but inconsistently) over
time, even in deposits that only contained a small
fiaction of fines
4.1 Load test comparisons
Predicting the behavior of open-ended steel tube
piles can be complicated because of physical differ-
ences in pile response during dynamic and static
loading. In particular the behavior of the soil col-
umn, as plugged or unplugged can be quite different
during static and dynamic loading. During a dy-
namic event, the pile will tend to move past the in-
ternal soil column because of the inertia of the soil
column. The pile therefore acts as an unplugged or
partially plugged pile. This has been referred to as
272
the "cookie-cutter" effect. As a result, the dynamic
response will include a distributed internal shaft re-
sistance response along the pile shaft-soil column
interface.
The inertial effect is not present during static
loading, and the soil column is therefore more likely
to move with the pile, in a plugged manner. Internal
soil-column resistance in this case will be equal and
opposite to the resistance developed at the pile base.
The internal resistance is concentrated above the pile
base.
It is interesting that despite these different be-
haviors, the predicted load-settlement responses
were in very good agreement.
4.2 Low driving resistances
The reason for the extremely low installation shaft
resistances measured in dense to very dense sands
(see Table 3) can only be speculated.
The local shaft resistance at the pile-soil inter-
face, z, is a hnction of the effective horizontal stress
at the interface, d h , and the interface fiiction angle,
6. Thus
. tan 6
z = ort1
Consider the driving of Pile 301, and in particular
the inferred shaft resistance of approximately 13 kPa
at a depth of 43m. Assuming an average bulk den-
sity of 19 kN/m3 to this depth, and a water table at
ground surface, the vertical effective stress at this
depth would be approximately 400 kPa. According
to Kulhawy (1 984), the horizontal effective stress
for this type of low-displacement pile should be ap-
proximately equal (i.e. K = 1.O, and (rrll= 400 H a ) .
The pile-soil fiiction angle, 6, is a hnction of the
soil fiiction angle, $, and the pile material, in this
case steel. Kulhawy and others recommend values
of 6/+ of approximately 0.7. For the dense to very
dense sands at this depth, a Eriction angle of at least
40" should be expected. From which 6 = 28" is es-
timated.
From Equation 1, T should be 210 kPa, although
design convention is to limit values to approximately
120 to 15OkPa. Even if the proposition of a critical
depth is accepted, the expected shaft resistance
would be at least 90 kPa at this depth. it is noted
that the value of 210 kPa actually agrees reasonably
with the shaft resistance values determined for the
third restrike of Pile 301.
This suggests that the long-term response of these
piles can be justified by geotechnical theory. The
key question is what caused the severe reduction (to
13 kPa) during pile installation.
The possibility of a reduction in interface fi-iction
angle is discounted. Such changes have not been re-
ported. The presence of mica is known to reduce
 interface friction angle, but this effect would be
permanent; would not cause a reduction of this mag-
nitude; and in any case, mica was not reported in
these sands.
Temporary increases in pore water pressure can
reduce the effective horizontal stress in fine-grained
soils, resulting in substantially lower installation re-
sistance, however, in this case, the sands are too
coarse to sustain elevated pore pressures. Note the
capacity of test pile 1 only recovered marginally
over a period of 4 days.
Another mechanism for reduction of lateral stress
on the pile must be found. Tomlinson and Poskitt
have both previously observed the phenomenon of
pile whipping, which can create a cavity around a
pile. This cavity can only be sustained if the soil
possesses some amount of binder which gives the
soil a temporary or permanent cohesion. A dense
soil in particular may be capable of self-support and
arching f?om the pile surface.
For the want of a more plausible explanation, the
phenomenon of pile-whip is believed to have caused
low contact stresses between the pile and the soil,
and hence the low observed driving resistances.
Given that the pile hammer and assembly had a
combined weight of 28 tonne; that the hammer was
not restrained by any leader system; and that free
pile lengths above ground of up to 20m were ob-
served during driving, sway and pile whip could be
feasible.
Support for this conjecture may also be found in
the driving shaft resistance distribution. The resis-
tance distributions for test-pile 1 shown in Figure 3
at the completion of driving and for various stages of
restrike show a local maximum at mid-pile length
(depth 20m) and local minima at the !A and % posi-
tions (8m and 32m respectively), especially at the
end of drive.
This resistance distribution was characteristic of
the initial driving behavior of most of the piles
tested, and this characteristic also persisted in some
of the restrike data. Figure 7 shows the resistance
distributions for Pile 301 at the end of driving, and
also at a penetration of 43.7m, 7m above final toe
level.
Figure 7. Shaft resistance distributions for Pile 301 during in-
stallation
273
It is interesting to observe that the characteristic
resistance distribution is linked to the pile penetra-
tion rather than the adjacent soil, suggesting a
structurally-induced geotechnical phenomenon. It is
proposed that the pile was flexing in a Mode 2 Euler
buckling shape. With this type of flexure, a local
minimum in movement (and maximum in shaft re-
sistance) could be expected at mid-length. Lateral
movement would be greatest at the !A and % lengths,
causing the greatest tendency for formation of a
cavity, and hence least shaft resistance.
Further research is required to substantiate this
hypothesis.
4.3 Internal shaft resistance
As noted earlier, the evidence at this site suggests
that internal shaft resistance was a small percentage
of the shaft resistance measured during dynamic
testing. The results of removing the internal soil
column were obscured by other effects which caused
the external shaft resistance to increase in the period
between the restrikes prior to and after the soil re-
moval.
Construction programming prevented restrike test-
ing immediately before and immediately after re-
moval of the soil column. Such a procedure would
have minimized other effects and made the compari-
sons of greater value. Further such testing is rec-
ommended in the fkture to provide some reliable
international data on internal shaft resistance for
open-ended steel tube piles.
4.4 Shaft resistance increases
The shaft resistance increases measured at this
site would appear to be linked to the very low initial
shaft resistances measured during installation and
discussed at length in Section 4.2.
In that discussion generation of locally elevated
pore pressures was discounted as a mechanism for
the low shaft resistances measured in the sands.
Equalization of pore pressures must logically there-
fore be discounted as a mechanism for the observed
subsequent shaft resistance increases.
It was hypothesized in Section 4.2 that the low
shaft resistances were due to a tendency for a cavity,
or zone of low contact pressure to develop around
the pile as a result of pile flexure during driving. It
was also shown that the final shaft resistance values
measured for Pile 301 were consistent with geotech-
nical theory. The mechanism of pile “set-up” in this
case then needs to be one which will restore the
contact pressure at the interface, removing the arch-
ing of the sands away from the pile.
It is believed that both the densification and vi-
bration effects of driving adjacent piles resulted in
increased horizontal stress at the pile-soil interface.
It is quite possible that these stress levels would de-
velop naturally over time with creep phenomena,
and that adjacent driving and vibration simply accel-
erated the process. Test-pile 1 which failed at a low
capacity, and did not demonstrate a substantial “set-
up” effect was the first pile driven at site, and would
therefore not have experienced the same vibration
and densification effects that occurred in the large
pile group of Pier 3.

5 CONCLUSIONS

This project provided some interesting and unusual

opportunities.
It is concluded that dynamic pile testing was able
to predict static load behavior for these driven steel
tube piles extremely well. This was demonstrated
by two Class A predictions.
Piles were installed through relatively clean sands
in a dense to very dense condition, and at significant
depth with unusually low shaft resistances. A
mechanism of pile flexure has been hypothesized as
responsible for this phenomenon.
Subsequent pile ‘set-up’ is believed to be due to
restoration of normal stress conditions around the
pile, promoted by vibration and driving of adjacent
piles.
Tests to determine the proportion of measured
shaft resistance which can be attributed to the inter-
nal soil column were inconclusive. However, there
is a reasonable suggestion that the internal shaft re-
sistance was not significant. Further research is rec-
ommended.

REFERENCES

Kulhawy, F.H. (1984) Limiting tip and side resis-

tance : fact or fallacy? ASCE Symposium on
Analysis and Design of Piled Foundations, San
Francisco, October 1984. Ed J. Meyer.

274
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Modeling pore pressure generation during dynamic testing of deep

foundations
I? L. Pinto
Department of Civil Engineering, Universidude de Coimbru, Portugal

ABSTRACT: Dynamic testing of deep foundations is becoming a common method to determine the static
capacity of piles and drilled shafts. Many of these foundations are placed underwater and it is expected that
the generation of pore pressure will affect the resulting capacity. The objective of the current work is to pres-
ent a numerical method to predict pore-pressure generation in this situation. An axisymmetric finite element
formulation is used with consideration of a continuous pore pressure field. A pile installed in saturated soil is
loaded by a sine wave pulse and the pore pressure is computed at the tip.

1 INTRODUCTION porous media was proposed by Zienkiewicz and

Shiomi (1 984). This formulation considers continu-
Large-strain dynamic testing of deep foundations is ous solid and fluid displacement fields as well as the
gaining popularity as a method to determine the pore pressure field. The governing equations are
static capacity of piles and drilled shafts. There has
been a strong evolution in understating the phenom-
ena involved since Smith (1960) proposed a rational
approach to the problem. The work of Rausche et al.
(1985) among others has provided useful tools for
the interpretation of test results. Nevertheless some
aspects require more research. Deep foundations are
often installed in saturated soils and it is expected
that the pore pressure generated by the compression
wave will have a direct impact on the interpretation
of the test results.
The model currently used by many researchers
where us is the displacement of the solid phase, uf is
and field engineers is based on the discretization of
the displacement of the fluid phase, p is the pore
the pile on an assemblage of masses connected by
pressure, CT"' is the effective stress tensor, n is the
elastic springs. The soil-pile interaction is consid-
porosity, pSand pf are the microscopic mass densi-
ered through linear-perfectly plastic springs and
ties of the solid and fluid phases respectively, K is
dashpots. This model has limitations on the consid-
the permeability tensor, Kf is the fluid bulk modulus
eration of pore pressure generation. Alternatively,
and b is the body force tensor. A dot over a vector
the finite element method presents great potential to
denotes time derivative.
perform such task (Pinto 1998).

3 FINITE ELEMENT FORMULATION

2 BACKGROUND
The solution of the governing equations through the
The theory developed by Biot (1955, 1956 and finite element method yields the following set of
1962) is used to provide the framework for the dy- equations:
namic behavior of saturated porous material. Satu-
rated soil is considered as a two-phase mixture,
where each phase has its own motion. The most M a + C v + K d=F(t) (4)
general formulation to solve the differential equa-
tions governing the dynamic behavior of saturated

275
where M is the mass matrix, C is the damping and K
the stiffness matrices. Each array is composed by the
following sub-matrices:
The time integration follows the Alpha scheme
introduced by Hilber et a1 (1977). If the integration
parameters are chosen correctly, it preserves uncon-
ditional stability and second order accuracy. The
numerical damping granted by this method affects
less severely the natural modes with period T
slightly larger than the time step At than Newmark’s
counterpart. It can also dissipate more effectively the
higher frequency (lower period) modes.
In the current case, the deep foundation and sur-
rounding soil were modeled using axisymmetric
9/9/4c elements. One of such elements is depicted on
Figure I . A continuous, quadratic approximation is
assumed for the displacement fields of the solids and
fluid. Pore pressure variation is assumed bilinear
within the element.
Figure 1 . Element 9/9/4c
This element is capable of providing accurate
solutions without incurring in problems such as spu-
rious pressure modes, m-sh locking or poor conver-
gence rate.
4 NUMERICAL MODEL
The numerical model shown in Figure 2 represents a
38.7-m long concrete pile, 0.52 m in diameter. It has
33.5 m embedded in the ground. The concrete
modulus is 41.6 GPa and Poisson’s ratio is assumed
to be equal to 0.2. The time that is required for the
compression wave to travel to the toe and back to
the pile head (2L/c) is equal to 18.6 msec.
For the analysis, the pile was modeled with 82
axisymmetric elements and 470 elements were used
to represent the soil mass.
276
The pile-soil interface was modeled as impervious
by slaving the horizontal displacements of the fluid
phase to the solids phase, along the shaft and tip.
This prevents the fluid from penetrating the pile. A
limitation of the current model is that no interface
elements were used.
Figure 2. FEM mesh
The soil is saturated, with the ground water table
at the surface. The water was considered incom-
pressible (Kf = 1.E 1 1 kN/m’)
Two soil types were studied. The first is repre-
sentative of a medium loose sand and the second of
a phosphatic clay. For comparative purpose, analy-
ses were performed with different reologic models.
The properties of the sandy soil are presented on Ta-
ble 1. Two extreme permeability coefficients were
considered to study its influence on the pore-
pressure generation.
Table I . Soil propertics for the elastic analysis
Es v n KD Ps Pf
MPa mls
20 0.3 0.37 1.0E-2 2.65 1.00
20 0.3 0.37 1.0E-10 2.65 1.00
A single pulse of harmonic load was applied at
the pile head using the following expression:
q(r)= A(l- COS( ~ t ) ) (6)
Two load types were considered: a quasi-static
with o = 50 rad/s and a typical dynamic pulse with
o = 1000 rad/s. The first load is comparable to the
duration of a “Statnamic” pulse. The time increment
used was equal to 5 msec, during 250 msec. On the
second load At = 0.25 msec and the response was
studied for 25 msec. The maximum load, corre-
sponding to 2A, was equal to 5.0 MN for the sandy
soil and to 2.0 MN for the clayey soil.
 5 RESULTS
Results for the quasi-static load are presented and
discussed first. The results show that for the elastic
solution, without interface elements, the skin friction
provides most of the resistance to the applied load
and only a small percentage of the load reaches the
tip (close to 3.5%). This load generates pore pres-
sure in phase with it. Figure 3 depicts the excess of
pressure, normalized by the hydrostatic pore pres-
sure, for soil with high and low permeability. As
should be expected, the less permeable soil has the
highest pore pressure increase.
Figure 3. Pore pressure ratio at the pile tip. Elastic soil and
slow load.
The elastic model is over-simplistic, but points
out the importance of the permeability in the pore
pressure generation. Later the soil was modeled as
elastic-perfectly plastic. The Drucker-Prager yield
surface was used, with c’=O and 4 ’=30”.
Figure 4. Pore pressure ratio at the pile tip. KD=lE-2 m/s.
For this particular case, Figure 4 shows that the
consideration of the plastic model has a very limited
influence on the pore pressure at the tip. The model
277
used has an associated flow law, with the consequent
dilative behavior of the plastified points in the soil.
This behavior causes higher confining stress around
the shaft, providing higher skin friction and limiting
the force reaching the pile tip. Also, as the soil di-
lates, the pore pressure is reduced. The combined ef-
fects explain the small departure of the plastic curve
from the elastic solution. It is expected that a more
complete model, with non-associated flow law and a
cap to control the volumetric plastic strains would
perform better.
On a second set of runs, the same problem was
studied on a clayey soil. The loading conditions
were A = 1000 kN and w = 50 rad/s. The soil prop-
erties, displayed on tables 2 and 3, are based on lab
and field tests performed on soft phosphatic waste
clay found in Polk County, Florida. The Cam-Clay
model was used in the analysis. The model is limited
in its ability to represent the behavior of soil under
cyclic load, but provides valuable insight on the dif-
ferences from a linear elastic representation.
Table 2. Elastic soil properties for clayey soil
Es v n KD P< Pf
MPa m/s
0.456 0.45 0.75 6.9E-5 2.65 1.00
Table 3. Soil properties for the CarnCIay model
CO M K h P’co
(kPa)
3.0 1.0 0.147 1.570 r(z)
The elastic solution shows a 10% increase of pore
pressure at the pile tip, resulting from the lower
permeability and high deformability of the soil.
Figure 5. Pore pressure ratio at the pile tip for the clayey soil
The impact of the plastic model is obvious in
Figure 5. As the soil surrounding the shaft plastifies,
more load is carried to the tip. The reduction of vol-
ume imposed by the elastic and plastic volumetric
strains causes the pore-pressure to raise more than
on the elastic representation.
 The results presented so far relate to a quasi-static
load. It is of great interest to study the importance of
the loading rate on the pile-soil response. For the
following set of analysis, the pile was loaded by a
single pulse, with the duration of 6.3 msec and the
response was followed for 25 msec. This scenario is
closer to a typical dynamic load test.
On the case of the sandy soil, the Drucker-Prager
model was used with the same properties as before.
Figure 6. Pore pressure response for dynamic loading condi-
tions on sandy soil.
The influence of the loading rate is captured on
figure 6. The excess pore pressure raises to close to
80% of the hydrostatic value, a significant increase
from the 5 % computed with the quasi-static load.
The finding strengthens the importance this pressure
may have on the interpretation of dynamic test re-
sults.
Similar findings are shown on Figure 7. Here the
clayey soil was used along with the CamClay model.
A significant increase of pressure is computed but
now it is quickly damped.
Figure 7. Pore pressure response for dynamic loading condi-
tions on clayey soil.
278
6 CONCLUSIONS
The purpose of the current research is to evaluate
the potential of a FEM coupled dynamic formulation
to predict pore pressure generation on piles installed
in saturated soil under dynamic axial loads. A small
study to determine the influence of permeability,
soil’s reologic models and loading rate is presented.
Results showed the importance of the permeabil-
ity on the pore pressure. For the elastic solution and
the particular soil and load conditions, a decrease of
KD from 1 .E-2 to I .E-I0 m/s caused the excess of
pore pressure to raise from 5% to 30% of the hydro-
static pressure.
The Drucker-Prager model with an associated
flow law did not show significant differences from
the elastic solution.
On the clayey soil, the CamClay model was able
to better represent the pore pressure generated when
the soil plastifies at the region of the tip.
Loading rate and inertial effects are visible for
both soils. The current approach seems to have po-
tential to model pore pressure generated during the
performance of a dynamic load test.
Some improvements are required. A dynamic in-
terface element may be implemented, to better
model the load transfer through skin friction.
Reologic models capable of modeling cyclic and dy-
namic soil behavior are also of interest.
Most importantly, calibration with values meas-
ured in field and laboratory tests is required.
REFERENCES
Biot, M.A. 1955. Theory of Elasticity and Consolidation for a
Porous Anisotropic Solid, J. of Applied Physics, 26, 182-
185.
Biot, M A . 1956. Thcory of Propagation of Elastic Waves i n a
Fluid Saturated Porous Solid - I Low Frequency Range, J.
Acoust. Soc. A m , 28, 168-178.
Biot, M.A. 1962. Mechanics of Deformation and Acoustic
Propagation in Porous Media, J. of Applied Physics, 33,
1482- 1498.
Smith, E.L. 1960. Pile Driving Analysis by the Wave Equation,
ASCE Journal of the Soil Mechunics and Foundation Divi-
sion, 86:4, 35-61.
Pinto, P.L. 1998. Coupled Finite Element Forinu1ution.s for
Dynariiic Soil-Structure Iiiternction, Ph.D. Dissertation,
University of Florida, Gainesville, USA.
Zienkiewicz, O.C. & Shiomi, T. 1984. Dynamic Behavior of
Saturated Porous Media; The Generalized Biot Formulation
and its Numerical Solution, Iiit. J . Num. Anul. Methods it7
Geomechanics, 8,7 1-96.
Hilber, H.M., Hughes, T.J. & Taylor, R.L. 1977. Improved
Numerical Dissipation for Time Integration Algorithms in
Structural Dynamics, Ecirthquake Engineering Nt7d Struc-
turul Dytiatiiics, 5 , 283-292.
Rausche, F., Goble, G.G. & Likins, G.E. 1985. Dynamic De-
termination of Pile Capacity, Journal of Geoteclznicrzl En-
gineering, ASCE, 11 1 (GT3), March, 367-383.
5 Codes
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 1503

Keynote lecture: The performance of the dynamic methods, their controlling

parameters and deep foundation specifications

Samuel G.Paikowsky & Kirk L.Stenersen

Geotechnical Engineering Research Laboratory, Department of Civil and Environmental Engineering, University
of Massachusetts, Lowell, Mass., USA

ABSTRACT: An ongoing project supported by the USA National Cooperative Highway Research Program
(NCHW) under the Transportation Research Board (TRB) of the National Academy of Science (NAS), is
aimed at rewriting AASHTO Deep Foundation Specifications for the year 200 1 . The AASHTO specifica-
tions are traditionally observed on all federally aided projects and generally viewed as a iiatioiial code of US
Highway practice, hence iiifluenciiig the construction of all the deep foundations of highway bridges through-
out the USA.
The new code is based 011 Load and Resistance Factor Design (LRFD) principles with resistance fac-
tors obtained from probabilistic analysis of data. A large database (PD/LT2000) is the backbone of the dy-
namic methods’ performance evaluation. This database originated with the work presented by Paikowsky et
al. (1 994), Pailtowsky and LaBelle (1994), and additional information acquired since.
A summary aiid careful evaluation of the large database is presented, detailing the performance of
various dynamic methods when compared to static load testing to failure. The parameters that control the ac-
curacy of the dynamic predictions are analyzed, suggesting the importance of certain mechanisms associated
with the pile penetration aiid the dynamic simulatioiis.
The controlling parameters and the statistical analyses are then utilized for the development of resis-
tance factors to be recommended for the new specifications.

1 BACKGROUND dure) and (ii) Overcome the generic difficulties of

National Cooperative Highway Research Program,
project, NCHRP 24-17, “LRFD Deep Foundations
Design” was initiated to: (i) Provide recommended
revisions to the driven pile and drilled shaft portions
of section 10 of AASHTO Specifications and (ii)
Provide a detailed procedure for calibrating deep
foundation resistance factors. The current AASHTO
specifications as well as other existing codes based
on Load and Resistance Factor Design (LRFD) prin-
ciples were developed using insufficient data, hence
they utilized mostly back-calculated factors. The
main challenges of the project are therefore: (a)
Compilation of large, high quality databases and
(b) Framework for a procedure and data manage-
ment to enable: (i) LRFD parameter evaluation and
(ii) Future updates. These challenges include two re-
quiremeiits: (i) Organization of the factors following
the design - construction - quality control sequence
(i.e. independence in resistance factors according to
the chronological stage and the evaluation proce-
applying the L W D methodology to geotechnical
app 1i cat i ons , i .e . incorporation of indirect var i ab i 1ity ,
(e.g. site or parameters interpretation), judgment
(e.g. previous experience), and other similar factors
The project team, headed by the first author, is
divided into three major groups dealing with static
analyses (University of Florida), probabilistic ap-
proaches and structural analyses (University of
Maryland), and dynamic analyses (University of
Massachusetts at Lowell). The present paper pro-
vides a background for design methodologies and
the LRFD. Database PD/LT2000 is presented and
analyzed. The state of practice and the selected dy-
namic methods are described, followed by an initial
evaluation of the signal matching technique and ex-
amination of the controlling parameters. The per-
formance of the dynamic methods is then provided,
categorized according to the controlling parameters.
The obtained results are used for the development of
resistance factors to be recommended for the new
specifications.

28 1
2 STATE OF STRESS DESIGN
METHODOLOGIES
2.1’ Working stress design
Working Stress Design (WSD) method, also called
the Allowable Stress Design (ASD) method, has
been used traditionally in Civil Engineering since it
was introduced in the early 1800s. The design loads,
Q, consist of the actual forces estimated to be ap-
plied directly to the structure.
where Q = Design load (F); Qr,// = Allowable design
load (F); R,,= Ql,lt= Ultimate geotechnical pile force
resistance; and FS = Factor of safety.
The factor of safety is commonly defined as the
ratio of the resistance of the structure (RJ to the load
effects (Q) acting on the structure. Table 1, last used
in AASHTO (1997), presents common practice in
which different factors of safety are used in conjunc-
tion with the level of control in the analysis of the
pile’s ultimate geotechnical resistance. Presumably,
when a more reliable and consistent level of con-
struction control is used, a smaller factor of safety
can be implemented, which leads to a more eco-
nomical design. Practically, however, the proposed
factors of safety in Table 1 (excluding the static load
test to failure) do not necessarily consider the con-
servatism (i.e. under-prediction) of the examined
method and hence their economical basis is ques-
tionable, (to be flirther discussed in section 9.3).
Table 1. Factor of safety on ultimate axial geotechnical capac-
ity based on level of construction control (AASHTO 1997).
Basis for Design and Type Increasing Design/Coiistructioii
of Construction Control Control
Subsurface Exploration X X X X X ing from uncertainties in resistance and to assure a
Static Calculation x x x x x Figure 1 shows the probability density functions
Dynamic Formula X for the load effect, Q, and the resistance, R. “Load
Wave Equation x x x x
CAPWAP Analysis X X
Static Load Test x x
Factor of Safety (FS) 3.50 2.75 2.25 2.00* 1.90
*For any combination of construction control that includes a
static load test, FS =2.0.
2.2 Limit states design
In the 1950’s the demand for a more econoinical de-
sign of piles brought about the use of Limit States or
282
Limit States Design (LSD). The two types of limit
states are the Ultimate Limit State (ULS) and the
Serviceability Limit State (SLS) both of which must
be satisfied when using the LSD method. Regardless
of the complexity of the analysis and calculation, all
limit states designs are carried out to satisfy the fol-
lowing criteria for ULS and SLS, respectively,
Factored resistance 2 Factored load effects (2)
Deformation 2 Tolerable deformation to remain ser-
viceable (3)
ULS pertains to structural safety and involves the
collapsing of the structure, or i n relation to piles,
when the ultimate bcariiig capacity of the soil is ex-
ceeded. SLS represents the conditions affecting the
function or service requirements (performance) of
the structure under expected service/working loads.
These conditions can include excessive deformations
and settlement or deterioration of the structure -
pile(s).
3 LOAD AND RESISTANCE FACTOR DESIGN
(LRW
3. I Principles
The design of a pile depends upon predicted loads
and the pile’s capacity to resist them. Both loads
and capacity have various sources and leveis of un-
certainty that engineering design methods and proc-
esses have historically compensated for by experi-
ence and subjective judgment. These uncertainties
can be quantified using probability-based design, or
safety check expressions, aimed at achieving engi-
neered designs with consistent levels of reliability.
The intent of the Load and Resistance Factor Design
(LRFD) method is to separate uncertainties in load-
prescribed margin of safety.
Effect” refers to the load calculated to act on a
particular element, (e.g. a specific pile). As the
loads are more deterministic than the resistances, the
load effect has smaller variability (smaller coef-
ficient of variation, translating to a narrower
probability density function) than the resistance. The
region where the two curves overlap is where the
load effect is greater than the resistance, indicating a
high probability of failure. The shaded area under
the probability density function for the resistance
represents the probability that the resistance will
have a value between a and b.

Figure 1 . An illustration of probability density functions for
load effect and resistance.
In LRFD, partial safety factors are separately ap-
plied to the resistance and load components. A fac-
tored (reduced) strength (capacity) of a pile is larger
than a linear combination of factored (magnified)
load effects. In this format, the strength is reduced
and the load effects are increased, by multiplying the
corresponding characteristic (noniinal) values with
factors, which are called strength (resistance) aiid
load factors, respectively. The nominal values (e.g.
the nominal strength, R,,) are those calculated by the
specific calibrated design method and are not neces-
sarily the mean strength or resistance, 3
and (see
Figure 1). For example, while and the distribution
represent the mean and performance of signal
matching analyses predictions from many case histo-
ries, R,, is the predicted value for the specific ana-
lyzed pile.
Based on considerations ranging from analyzed
case histories to existing design practices, a pre-
scribed value is chosen for the probability of failure.
The probability for failure of the designed pile fol-
lowing the application of the resistance and load fac-
tors should be smaller than the prescribed value (the
overlapping area in Figure 1). Practically, the resis-
tance factors are calculated to provide the prescribed
probability of failure.
3.2 Background information
The concept of using the probability of failure as a
criterion for structural design was presented in the
283
late 1920s by the Russians N.F. Khotsialov and N.
S. Streletskii (see Ayy~ibet al. 1998b) and was first
introduced in the US by Freudenthal (1 947). Recent
development of LRFD in civil engineering was initi-
ated in structural engineering (e.g. Ellingwood et al.
1980). Reliability-Based Design codes utilizing the
LRFD format are used by the American Institute of
Steel Construction (AISC 1994, Galambos & Ravin-
dra 1978) and by the American Coiicrete Institute
(ACI 1995). An effort was made by the National
Standards Institute (ANSI) to develop probability-
based load criteria for buildings (Ellingwood et al.
1982a,b) and ASCE 7-93 (ASCE 1993). The
American Petroleum Institute (API) extrapolated
LRFD technology for use in fixed offshore platforms
API (1989), and Moses (1985, 1986). Additional
comprehensive summaries for the implementation of
modern probabilistic design theory into design codes
include Siu et al. (1975) for the National Building
Code of Canada (1 977), Ellingwood et al. (1 980) for
the National Bureau of Standards, and the CIRIA 63
(1977). Ayyub et al. (1998a) provide details on
LRFD rules for ship structures that were developed
for the US Navy. The AASHTO LRFD Bridge De-
sign and Construction Specifications (1994) resulted
from work in NCHRP Project 12-33 (Nowak 1993,
1999) provide design guidance for girders.
3.3 LRFD performance and advantages
Experience has shown that adopting a probability-
based design code results in significant savings
and/or efficient use of materials. Reliability im-
provements are still under evaluation even though
the new codes are designed for reliability to be equal
to or better than the older codes. Experiences are not
yet well documented, but designers have commented
that, relative to the conventional working stress
code, the new AISC-LRFD requirements are saving
from 5% to 30% stecl weight, with a 10% typical
savings (Ayyub 1999). This may or may not be the
case for other industries. Specific benefits in pile de-
sign include the following:
1 . A more efficiently balanced design results in cost
savings and/or improved reliability.
2. Uncertainties in the design are treated more ra-
tionally and rigorously.
3. An improved perspective of the overall design
and construction processes (sub- and super-
structures), and the development of probability-
based design procedures can stimulate advances
in pile analysis aiid design.
4. The codes become a living document that can be
easily revised to include new iiiformatioii reflect-
ing statistical data on design factors.
5 . The partial safety factor format used herein also
provides a framework for extrapolating existing
design practice to new foundation concepts and
materials where experience is limited.
3.4 LRFD in geotechnical engineering
Early use of limit state design for geotechnical ap-
plications was examined by the Danish geotechnical
Institute (Hansen 1953, 1956) and was later formu-
lated into code (Hansen, 1966). Independent load
and resistance factors were used, with the rcsistance
factors applied directly to the soil properties rather
than to the nominal resistance.
Considerable effort has been directed over the past
decade to the application of LRFD in geoteclinical
engineering. LRFD approaches have been developed
in offshore engineering (e.g. Tang 1993, Hamilton &
Murff 1992), in general foundation design (e.g. Kul-
hawy et al. 1988, 1994, 1996), and in pile design for
transportation structures (Barker et al. 199 1, O’Neill
1995).
In geotechnical practice, uncertainties in resis-
tance principally manifest in site characterization,
soil behavior, and construction quality. The uncer-
tainties have to do with interpreting site conditions,
understanding soil behavior, and accounting for con-
struction effects. Uncertainties in extzrnal loads are
small compared with uncertainties in soil and water
loads and the strength-deformation bchaviors of
soils. The approach for selecting load and resistance
factors developed in structural practice, though a
useful starting point for geoteclinical applications, is
not sufficient. Work is needed to incorporate in the
LRFD formulation factors that are unique to geo-
technical design.
Philosophically, the selection of load and resis-
tance factors need not be made probabilistically, al-
though in current structural practice a reliability-
theory-based calibration is commonly used. This
approach focuses more on load uncertainties than re-
sistance uncertainties, and does not include many
subjective factors unique to geotechnical practice.
An expanded approach is needed if the full benefits
of LRFD are to be achieved for foundation design.
The National Research Council reported, “this.. .
subjective approach reflects the general lack of ro-
bust data sources from which a more objective set of
factors can be derived” (NRC 1995). The report
continues, “. . .realistically, because of the tremen-
dous range of property values and site conditions
that one niay encounter, it is unlikely that com-
pletely objective factors can be developed in the
foreseeable future.”
Today, the situation has changed somewhat, but
not entirely. The research team of NCHRP 24-17
gathered robust data sources on pile capacity from
which a more objective calibration of resistance fac-
tors can be made. Nonetheless, there remain uncer-
tainties associated with (1) site conditions, (2) soil
behavior and interpretation of soil parameters, and
(3) construction methods and their quality. These
factors are difficult or impossible to understand from
284
the pile databases alone. Knowledge factors need
therefore to be developed and combined with the re-
liability-theory-based calibration of the database re-
cords to achieve a meaningful LRFD approach.
3.5 LRFD forpiles
3.5.1 1999 AASHTO LRFD Bridge Design Spec$-
cations for Foundations
The ultimate resistance (I?,,) multiplied by a resis-
tance factor (4, which is equal to the factored resis-
tance (R,.), must be greater than or equal to the sum-
mation of the loads (Q,) multiplied by a modifier (71,)
and load factor (x)for the strength limits states:
(4)
where:
11, = factors to account for effects of ductility ( q ~ ) ,
redundancy ( q ~ and
) , operational importance (71).
The 1999 AASHTO LRFD Bridge Design Speci-
fications for Foundations provides the following
equations for the determination of the factored bear-
ing resistance of piles, QR:
for which:
Q,>= q,A, (7)
where $q = resistance factor for the bearing resis-
tance of a single pile specified for methods that do
not distinguish between total resistance and the indi-
vidual contributions of tip resistance and shaft resis-
tance; Q,,[,= bearing resistance of a single pile (F);
Q,= pile tip resistance (F); Qs = pile shaft resistance
(F); qr, = unit tip resistance of pile (F/L’); qs = unit
shaft resistance of pile (F/L’); A, = surface area of
pile shaft (L2); A, = area of pile tip (L2); and #q)qp,#qs
= resistance factor for tip and shaft resistance, re-
spectively, for those methods that separate the resis-
tance of a pile into contributions from tip resistance
and shaft resistance,
The resistance factors for use in the above equa-
tions are presented in Table 10.5.5-2 of the specifi-
cations. The table incorporates a factor h,, to ac-
count for the level of field capacity verification. As
an example, if a wave equation analysis is used to
determine the pile capacity with an assumed driving
resistance, a resistance factor of 0.65 is recotn-
mended. If pile driving formulas, e.g. ENR equa-
tion, are used to verify the pile capacity without
stress wave measurements during driving, a h, factor
 of 0.80 is recommended. This results in an actual
resistance factor of 0.52 to be used in the above
analysis/verification sequence (0.65 x 0.80).
3.5.2 Worldwide LRFD for dynamic methods
Review of the development of LRFD applications
for Geotechnical Engineering is presented by Goble
(1999). A short description of LRFD codes that in-
clude dynamic methods follow.
The Australian Standard for Piling-Design and
Installation ( 1995) provides detailed recommenda-
tions for resistance factors to be used with the dy-
namic methods. The recommended resistance fac-
tors range betwecn 0.45 to 0.65 for methods without
dynamic measurements (including WEAP) and be-
tween 0.50 to 0.85 when utilizing dynamic ineas-
urements with signal matching analysis. Selection of
the appropriate resistance factor depends on driving
conditions, geotechnical factors (e.g. extent of site
investigation), and extent of testing (e.g. low range
for ~ 3 of % the pile tested and high range for >15%).
In traditional structural design specifications, a
nominal value is given and the value used is based
primarily on engineering judgment and cannot ex-
ceed the nominal value. The Australian Standard is
therefore unique by providing a guide for choosing
the appropriate resistance factor. Interestingly, no
distinction is made regarding either soil type or time
of driving (i.e. EOD, BOR) when referring to the
signal matching based on dynamic ineasureinents.
The method by which the resistance factors were
generated is not provided in the code.
The AUSTROADS Bridge Design Code (1992)
provides resistance factors for four categories of dy-
namic methods. The range of resistance factors is
quite large and there is no explanation as to how the
resistance factors were obtained. Goble (1 999) pos-
tulates that the resistance factors were calibrated via
the working stress design method.
The Ontario Bridge Code (1 992) recommends
relatively low resistance factors with no differentia-
tion between the dynamic analyses. No information
is provided on how the resistance factors were ob-
tained.
The Canadian Bridge Code (1992) is brief in its
design requireinents for deep foundations. For dy-
namic load testing, resistance factors of 0.4 and 0.5
are recommended for routine testing and analyses
based 011 dynamic nieasurements, respectively.
3.5.3 Dificu1tie.s of the Existing LRFD Codes
All existing codes suffer from two major difficulties.
One is the application of LRFD to geotechnical
problems as described in section 3.4 (e.g. site vari-
ability, construction effects, etc.). The other problem
is lack of data. None of the reviewed codes and as-
sociated resistance factors were developed based on
databases enabling the calculation of resistance fac-
tors from case histories.
285
The current AASHTO specifications encounter
additional difficulty due to the irrational multiplica-
tion of the resistance factor by the modifier h,. This
procedure requires the interaction of two independ-
ent evaluations (e.g. static analysis and dynamic
methods) and results in unnecessary and confusing
conservatism. A clear separation of the resistance
factors on the basis of design and construction is re-
quired.
4 PD/LT2000 DATABASE
The database PD/LT2000 contains information re-
lated to 210 driven piles that have been statically
load tested to failure and dynamically monitored
during driving and/or restrike (389 analyzed meas-
urements). PD/LT2000 is comprised of the integra-
tion of databases PD/LT (Pailtowsky et al. 1994) and
PD/LT2 (Paikowsky & LaBelle 1994) with expan-
sion by an additional 57 pile cases (Stenersen 2000).
The data of PD/LT2000 were carefully examined
and analyzed following procedures described by
Paikowsky et al. (1994), resulting in detailed static
and dynamic pile capacity evaluations. Table 7 pre-
sents a summary of the data contained in
PDiLT2000 broken down according to site location
and soil type, pile type and capacity, driving behav-
ior and time of driving.
5 REFERENCE STATIC CAPACITY
5.1 Overview
The dynamic methods are assessed by comparing the
pile capacity of the evaluated method with a refer-
ence static capacity of the pile. The determination
of the pile’s static capacity based on load- displace-
ment relations is not unique. The test results depend
on the load testing procedures and the applied inter-
pretation method, both of which are subjective. The
following sections examine each of these factors and
its influence on the reference static capacity.
5.2 Failure Criterion
Past work (Paikowsky et al. 1994) have resorted to a
“representative” static pile capacity based on the as-
sessment of five interpretation methods; Davisson’s
Criterion (Davisson 1972), Shape of Curve (similar
to the procedure proposed by Butler and Hoy 1977),
Limiting Total Settlement to 25.4 mm and to 0.1B
(Terzaghi 1942), and the DeBeer log-log niethod
(DeBeer 1970). A single representative capacity
value was then calculated for the analyzed case as
the average of the methods considered relevant (i.e.
provided reasonable value). The development of a
framework for future modifications requires that the
evaluated resistance factors be based on an objec-
Table 2. Summary of the data in the PD/LT2000 database.

Pile Types
Location
Soil Types Soil Inertia Type of Data I Pile Capacities 1
Soil
Pile p p e No. Location No.
-Pile Northeast
-
Time

USA 44 EOD
&
Soiitheust
OEP BOR
USA 69 Clay
CEp 6, North /Till
USA 24 EOD
Voided 35 South &
- Cont ?te . USA I l0 BOR ‘s

254 9
EOD
305 5
Rock
356 8 Australia I 2
BOR
I sq.
Conc
406 1 Biwswick

L
(171 !?I) 457 8
EOR
508 8
Sand Kon
/Sill
610 16
I
762 5 Ontario 1 22
DD

Octagonal Sweden 1
3 DR
Cono.et e
NA 6
Tinibei- 2 ALT
Monotuhe 2
Total 2 10 210
Notes:
1. OEI’ - Open Ended Pipe Pile 11. No. - Number of Piles / Cases
2. CEP - Close Ended Pipe Pile 12. sq. Conc. - Square Concrete
3. EOD - End of Driving 13. NA - Non Applicablc / Unknown
4. BOR - Beginning of Restrike 14. USA - United States of America
5 . EOR - End of Restrike 15. Northeast USA - Federal Highway Regions 1, 2 & 3
6. DD - During Driving 16. Southeast USA - Federal Highway Region 4
7. DR - During Restrike 17. North USA - Federal Highway Regions 5 , 7, & 8
8. ALT - Alternate measurement 18. South USA - Federal I-Iighway Region 6
9. Blow Cts. - Blow Count 19. Noi-thwcst USA - Federal Highway Region 10
10. AI< - Area Ratio 20. Southwest USA - Federal IHiahwav Region 9

tive, repetitive procedure. 111 order to do so, the
static capacity of each pile in database PD/LT2000
was evaluated utilizing all five aforementioned crite-
ria and a representative capacity was assigned for
each pile. A statistical analysis was then carried out
by determining the mean and standard deviations of
the ratio of the representative pile capacity to the
method being evaluated. Figure 2 shows the histo-
gram and calculated distributions (normal and log-
normal) for Davisson’s failure criterion. Davisson’s
criterion was found to perform the best overall and
was therefore chosen as the single method to be used
when analyzing load-displacement curves. The
method provides an objective failure criterion and
was also found to perform well for piles exceeding a
diameter of‘ 610nini (examined through 30 pile
cases). The data presented in Figure 2 demonstrates,
however, that though small, a bias exists in the static
capacity used as a reference for the evaluation of the
dynamic methods’ performance.
5.3 Load Test Procedure
An additional factor to examine is the influence of
the static load testing procedure (loading rate) on the
designated pile capacity. This influence was exam-
ined in two ways.
Two detailed case histories from a research site
in Newburyport, Massachusetts were evaluated. A
pipe and pre-stressed concrete heavily instrumented

286

Figure 2. I-listograni and frequency distributions of KsD for
186 PDiLT2000 pile-cases in all types of soils.
friction piles were tested over a lengthy period of
time at a bridge reconstruction site. Both piIes were
tested using three types of static load testing proce-
dures: slow maintained (testing duration of about 45
hrs), short duration (testing duration of about 6-8
hrs), and static cyclic (testing duration of about 15
min.). Dctails about the piles and the testing are pre-
sented by Paikowsky and Hajduk (1999, 2000) and
Paikowsky et al. (1999). Thc interpretation of the
load-displacement relationships in both cases sug-
gested that the test type had an insignificant influ-
ence on the pile capacity, (referring to a failure crite-
rion irrespective of the displacement).
The effect of the test type was further investigated
utilizing a databasc containing information related to
75 piles tested under slow maintained arid static-
cyclic load testing procedures. In the static-cyclic
procedure, the piles were loaded to failure using a
high loading rate and then unloaded. The process
was repeated for four cycles. The testing procedure
and its interpretation method are presented by Pai-
kowsky et al. (1 999). A comparison between the pile
capacity based on Davisson’s failure criterion for the
slow maintained tests and the static-cyclic capacity
is presented in Figure 3. The obtained relations and
the associated statistical information suggest that
there is no significant influence on the static pile ca-
pacity based on the applied static load rate. The
static cyclic load test results were also compared to
the representative static pile capacity (based on the
aforementioned five methods) resulting in a mean
Ksc of 1.023 and a standard deviation of 0.057.
Figure 3. Comparison between pile capacity based on Davis-
son’s criterion for slow maintained load tests and static cyclic
load test capacity for 75 piles, (Paikowsky et al. 1999).
These evaluations led to the conclusions that
Davisson’s pile failure criterion can be used as a
method to determine the reference pile capacity irre-
spective of the static load-testing procedure.
6 DRIVEN PILES AND DYNAMIC METHODS -
STATE OF PRACTICE IN USA HIGHWAY
CONSTRUCTION
To address the state of practice in the USA highway
construction, a questionnaire was sent to the De-
partment of Transportation offices of all fifty states,
requesting information on the use, design and con-
struction of deep foundations. The main results
from 45 states and 2 Federal Highway officials re-
garding driven piles are:
0 Driven Piles constitute 75% of all bridge founda-
tions. From the people who responded, 64% prefer
the use of driven piles to any other foundation type.
In the driven pile categories: 21% PPC (Pre-cast
Pre-stressed Concrete), 52% H-Piles, 25% CEP
(Closed Ended Pipe) and 2% OEP (Open Ended
Pipe).
0 Dynamic Equations usage: 45% ENR (Engineer-
ing News Record, Wellington 1892) and 16% Gates
(Gates 1957).
0 WEAP (Wave Equation Analysis Program) is used
by 80%.
0 84% perform dynamic testing during construction
on approximately 1% to 10% of the piles.
0 82% use EOD (End Of Driving) conditions to set
the piles’ length, 18% use BOR (Beginning Of Re-
strike) only, and an additional 34% use both EOD
and BOR. Some specifically use EOD for end bear-
ing and BOR for friction piles (NH).
Q 36% do not consider pile freeze or relaxation when
assigning driving criteria.
The questionnaire's responses suggest an extensive
use of dynamic measurements, particularly at the
end of driving, extensive use of WEAP, and signifi-
cant use of the dynamic equations.
7 THE CHOSEN DYNAMIC METHODS
7.1 Oveiview
Prior to detailed analyses leading to the deterniina-
tion of resistance factors, two coniponents must be
established: (a) type of dynamic methods to be
evaluated and (b) the conditions under which these
methods need to be examined. Sections 7 and 8 ad-
dress these issues, respectively.
7.2 Methods of Analysis
7.2.1 Geiier-al
Table 3 presents a summary of the major available
dynamic methods for evaluating pile capacity. The
methods are subdivided according to the project
stage (i.e. design vs. construction) and the need for
data obtained through dynamic measurements. The
incorporation of dynamic equations and WEAP re-
flects the previously described state of practice.
The methods that require dynamic measurements
can be broadly categorized as those that utilize a
simplified analysis providing an instantaneous pile
capacity evaluation for each hammer blow, and
those that require elaborate calculations (i.e. signal
matching), traditionally carried out in the office.
7.2.2 WEAP
Based on Smith (1 960), the use of the WEAP (Goble
& Rausche 1976) during design is of great impor-
tance for achieving compatibility between the driv-
ing system, the pile, and the soil conditions. Driv-
ability study and pile stress analysis often determine
the pile type and geometry and tlie adequacy of the
proposed equipment. The evaluation of WEAP ef-
fectiveness for capacity predictions is difficult to as-
sess as a large range of input parameters is possible
and the results are greatly affected by tlie actual field
conditions. Examination of the method through
analyses making use of default values is probably
the best avenue. Other evaluations, including WEAP
analysis adjustments following dynamic measure-
ments (e.g. matching energy) seem to be impractical
in light of the other available methods, and lead to
questionable results regarding their quality and
meaning (Rausche et al. 1997, Rausche 2000). The
WEAP analyses are evaluated in this study in two
ways: (i) as a dynamic method for pile capacity pre-
288
diction, using WEAP default input values and the
pile's driving resistance at the end of driving (EOD)
compared to the static load test results and (ii)
WEAP as a pile design method examining the ana-
lyzed stresses at the design stage with the measured
stresses during construction. Such evaluation leads
to a strength factor (related to the allowed structural
stresses in the pile) that is beyond the scope of this
manuscript.
7.2.3 Dyiiamic Equations
The choice of dynamic equations came to address
the state of practice and reflect a range in equation
type and performance. While the Engineering News
Record (Wellington 1892) was proven to be unreli-
able through the years (e.g. Olsen & Flaate 1967) it
was founded on a solid theoretical basis and is used
in construction in about half of the states in the
USA. The equation's traditional foimulation (e.g.
Massachusetts State Building Code 1997) includes a
factor of safety of 6, which needs to be omitted
when developing a resistance factor. The evaluation
of the ENR equation is carried out, therefore, utiliz-
ing its basic formulation. The Gates equation (Gates
1957), while empirical, was found to provide rea-
sonable results in the past (e.g. Olsen & Flaate 1967,
Long et al. 1998). The equation was further en-
hanced by Richard Cheney of the Federal Highway
Administration, (FHWA 1988), based on a statistical
correlation provided by Olsen and Flaate (1967).
7.2.4 The Case Method
The Case method (Goble et al. 1970, Rausche et al.
1975) is often used in field evaluations, as it is built
into Pile Dynamics Inc.'s Pile Driving Aiialyzer
(PDA), the most coinnionly used in the USA. The
method is based on a simplified pile and soil behav-
ior assumptions (free end and plastic soil), resulting
in a closed forin solution related to the impact and
its reflection from the tip. With the years, the
method evolved to be implemented into at least five
different variations (GRL 1999). The Case method
utilizes a damping coefficient (Jc) that is assumed to
be associated with soil type. The influence of this
factor on the predicted static capacity depends on the
reflected wave from the pile's tip, and hence on the
driving resistance. The case-damping coefficient
was investigated through a back calculation (to
match the measured static capacity). Results, de-
scribed in the next section, suggest no correlation
between the soil type and the case-damping coeff-
cient. The coninion recommended practice suggests
the use of the method based on a specific site/area
calibration (GRL 1999). This approach, in conjunc-
tion with the application of the method for maxi-
mum resistance (RMX), has proven effective. Ac-
cumulated experience on extensive jobs in the
Boston area (e.g. GTR 1997, 1998) has denion-
strated the effectiveness of the Case method, when
Table 3. Summary of the available dynamic methods.

Categoiy
I Metlrod

WEAP
Advai7 rages Disadvnt 1rages Cotntiiell/

- Equipment Match - Non unique Analysis - Required for Construction

1960, Drivability Study Performance sensitive to - Required Evaluation foz-
Design Stage - -
Goble et al.
- Structural Stresses field conditions capacity predictions"
1976)
ENR
- Sound Principles - Needs to be examined
(Wellington - Unreliable
- Common use without a built in FS.
1892)
D):t?nmic
Eqziutions I Gates
(Gates 1957)
-
-
Empirical
Comnlon use
- Depends on original
database
- Found to be more reliable
than other equations

~ ~ - Correction
~ based On ~ Depends~on database ~ - Was found
l to be reliable
~ n
additional data
FHWA 1988
Signal Matching - Solid principle of
(e.g. CAPWAP) matching calculations Stationa'y 'Oil forces - Office Method
- Expensive
(Goble et al. to nieasurernents by - Found reliable a! BOR
- Requires time
1970) iimosine nisd. B.C.
Case Method - Requires local calibration - Was found reliable with
D):namic (Goble et al. - Simplified Analysis - Presumed dependency of local calibration
Measvrements 1970, Rdusche et - Field Method soil conditions found - How to obtain national or
al. 1975) baseless international calibration?
Energy
Approach - Simplified Analysis Shows capacity
Field Method
which may not be present - Ideal foi construction
(Pailtowsky -
at EOD
1982)
"Resuired by the research Dane1 andior DOT's state codes

calibrated. A statistical examination of local calibra-

tion was perfoinied in Florida by McVay et al. c

(2000). The results of this analysis suggested that for (9)

48 cases, the ratio between the static pile capacity to
the Case method prediction at EOD was 1.344 f
0.443 (mean f 1 S.D.).
As no generic conditions exist for the use of the
Case method, international or national calibrations
are unrealistic. As the projection of local calibration
(of good experience and practice) beyond the geo-
graphical location may be unwise and/or unsafe, the
Case method was excluded from the examined dy-
namic analyses.
7.2.5 Tlie Energy Approach
The Energy Approach uses basic energy relations in
conjunction with dynamic measurements to
determine pile capacity; the concept was presented
by Paikowsky (1982) and was examined on a limited
scale by Paikowsky and Chernauskas (1992).
Extensive studies of the Energy Approach niethod
were carried out by Paikowsky et al. (1994), and
Paikowsky and LaBelle (1 994). The underlying
concept o f this approach is the energy balance
between the total energy delivered to the pile and the
work done by the pile/soil system. The basic Energy
Approach equation is:
where R,, = maximum pile resistance, E,,,,, =
measured maximum energy delivered to the pile,
D,,,, = measured maximum pile top displacement,
and Set = permanent displacement of the pile at the
end of the aiialyzed blow, or I/nieasured blow count.
For further details regarding the Energy Approach
niethod see Pailtowsky et al. (1994) and Paikowsky
(1 995).
7.2.6 The Signal Matcliing Techniques
The response of the niodeled pile-soil system (e.g.
force at the pile top) under a given boundary
condition (e.g. measured velocity at the pile top) is
compared to the measured response (force
measured). The inodeled pile-soil system or, more
accurately, the niodeled soil that brings about the
best match (visual graphical match) between the
calculated and measured responses, is assumed to
represent the actual soil resistance. The static
component of that resistance is assumed to be the
pile's capacity at the time o f the driving. The

289
procedure of signal matching was first suggested by
Goble et al. (1970), utilizing the computer program
CAPWAP. Others developed similar analyses, (e.g.
Paikowsky 1982, Paikowsky & Whitman 1990)
utilizing the computer code TEPWAP. The TNO
program was developed by Middendrop and van
Wee1 (1986), which led to improvements and the
CAPWAPC program, which is in coininon use to
date. The signal matching technique is often refei-red
to as post driving analysis or office method. With
the availability of faster, portable computers, it
became reasonably simple to conduct the analysis in
the field as well. In difference from the ficld method
however, the signal matching analyses cannot be
carried out for each blow during driving.
8 THE CONTROLLING PARAMETERS
8.1 Overview
Preliminary examination of the parameters control-
ling the performance of the dynamic analyses is car-
ried out prior to a final detailed evaluation of these
methods, leading to resistance factors. Such exami-
nation iiifluences the sub categorization of the dy-
namic methods (according to the important control-
ling parameters), hence, directing the user to utilize
the appropiate resistance factor according to the
relevant conditions of the employed method. For ex-
ample, if soil type is a controlling factor and the ac-
curacy of the signal matching method is largely af-
fected by soil type, evaluation of the method for
different soil types will result in the development of
resistance factors, depending on the soil type. Con-
versely, if soil type does not control the accuracy of
the spccific dynamic method, categorization based
on soil type is neither desired nor perused.
The following sections outline the logic used
for the preliminary examination of the controlling
parameters, the analyses and the results. The ration-
ale presented in this section follow previous studies
by Paikowsky et al. (1994), Paikowsky (1995), Pai-
kowsky et al. (1995), and Paikowsky and Cher-
nauskas (1 996).
The evaluation of static capacity from pile driving
is based on the concept that the driving operation in-
duces failure in the pile-soil system, (i.e. a very fast
load test is carried out under each blow). Dynamic
analyses encounter three fundamental difficulties: (i)
correct formulation of the penetration process (e.g.
soil motion, soil plugging etc.), (ii) evaluation of the
static resistance out of the total resistance encoun-
tered during penetration, and (iii) time dependent
pile capacity, (Paikowsky 1995). The parameters
controlling the accuracy of the dynamic predictions
therefore reflect the ability of each method to ad-
dress the above difficulties.
Based on the concept of a pile loading to failure
under each blow, it has been traditionally assumed
that during high driving resistance (i.e. refusal) there
is not sufficient pile penetration to mobilize the full
pile capacity (e.g. Chellis 1961). As such, the dy-
namic methods are deficient under high driving re-
sistance, categorized as equal or above 12BPI
(Blows Per Inch) or approximately 5BPcm (Blows
Per cm), (MHD 1988).
The soil type is also believed to constitute a major
factor as soil damping parameters are commonly
employed to represent viscous resistance in the
modeling of the soil’s dynamic behavior. This vis-
cosity is assumed to be soil type dependent, and as-
sociated with intrinsic soil properties. As such, high
viscosity values are expected for cohesive soils and
low viscosity values are expected for cohesionlcss
soils. Naturally, under a given velocity, high viscous
values are associated with higher dynamic resis-
tance, and logically should prove more difficult to
accurately define the static resistance.
The effect of time is well recognized but poorly
quantified. Piles undergo a decrease or increase of
capacity with time, also known as relaxation and set-
up, respectively. While the resistance during driving
and its static component represent the conditions en-
countered during penetration, the major interest re-
mains the long-term ability of the pile to carry load
during its service life. The examination of the dy-
namic method predictions with static load tests (of-
ten carried out long after the driving) therefore re-
mains valid. The predictions can be assessed in
relation to the time in which the driving and/or the
dynamic data have been obtained (e.g. EOD or
BOR).
The sections below examine the importance of
each of the above three assumed controlling parame-
ters. The results are used to evaluate additional pos-
sible controlling factors, laying down the framework
for the detailed evaluation of the dynamic methods.
8.2The Eflect oJ’Soil Type
The effect of soil type is examined in two ways: (i)
the correlation between the parameters assumed to
be soil type dependent and soil type, i.e. damping
parameters and (ii) the accuracy of the predictive
methods relative to the soil type.
Figures 4 and 5 present the relationship between
soil type and Smith damping parameters used in ap-
proximately 370 CAPWAP analyses, from
PD/LT2000, for the tip and side pile resistances, re-
spectively. Figure 6 presents the back-calculated
case damping factors for 290 cases, from the PD/LT
database (Paikowsky et al. 1994), required to obtain
a match between the predicted capacity and the
measured static capacity. All three figures clearly
indicate that no unique relationship exist between
soil type and damping parameters, suggesting that
other mechanism control the value required for a
damping factor rather than the soil type.

A summary of the statistics obtained when exam-
ining the accuracy of the signal matching technique
(specifically CAPWAP) based on soil type is
presented in Table 4. The statistics shown are the
mean and standard deviation of a normal distribution
function for the ratio of the pile’s static capacity
(based on Davisson’s failure criterion) to the pile
capacity obtained in the CAPWAP analysis. The
major categorization based on three soil types at the
pile’s tip show no significant differeiices between
clay arid till vs. sand and silt. Although the cases for
piles found on rock provide different values, the
291
Figure 6. Tip soil conditions versus calculated Case damping
coefficient, Jc, based 011 static load test results for 290 PD/LT
pile-cases (Paikowsky et al. 1994).
numbers are based on a small subset of 15 pile-
cases, compared to 100 and 265 pile-cases for the
other categories.
Table 4 provides further examination of time of
driving and driving resistances as subsets of the soil
type categorization. Two sets are examined based on
the time of driving: EOD and last BOR, (i.e. in the
case of niultiple restrikes, only the last restrike is
considered for the analysis).
The presented results suggest that the time of driv-
ing significantly affect the performance of the
CAPWAP prediction, regardless of soil type. The
mean values for the BOR sets are closer to one,
while the mean values for the EOD are closer to two.
The coefficients of variation (the ratio of the stan-
dard deviation to the mean) show values of 0.33 and
0.39 for BOR, while the EOD ratios are 0.55 and
0.85, which indicates the existence of substantial
scatter.
Further evaluation of the records is carried out on
the basis of driving resistance. The division between
cases for which the driving resistance is smaller or
greater than SBPcm, examines the aforementioned
notion of refusal and the expected accuracy of the
dynamic methods. The results shown in Table 4 in-
dicate that the cases for which the driving resistance
is smaller than 5BPcni result in less accurate analy-
ses with larger scatter, compared to the cases for
which the driving resistance was above SBPcm.
Though driving resistance seems to be an important
factor, clear understanding of its influence on the ac-
curacy of the dynamic methods calls for additional
investigation, which is presented in a following sec-
tion.
Table 4. CAPWAP analysis based on soil type categorization.

Clay & Till Scind & Silt Rock

Meat1 1.352 1.5 17 0.930
Standard Deviatioii 0.723 1.085 0.172
Niiniher of Cases 100 265 15
Time of Driving EOD BOR(last) EOD BOR(1ast) EOD B OR(last)
Meotz 1.634 1.133 2.068 1.193 0.968 0.925
Sianda rd Deviation 0.899 0.444 1.765 0.391 0.132 0.203
Nziivher of Cuses 45 40 77 116 7 7
Blolv Count (BPcni) <5 25 <5 25 <5 25 <5 25 <5 25 < j 25

Meaii 1.127 1.725 0.750 1.315 2.191 1.458 1.126 1.283 1.070 0.952 0.671 0.879
Sia 11 durd Deviation 0.637 0.807 0.241 1.160 1.901 0.512 0.386 0.355 ----- 0.136 0.163 0.230
Nuinher ojCases 35 35 11 10 64 13 74 40 1 6 3 3
Notes: 1. EOD = End of Driving
2. BOR(last) = Beginning of the last rcstrike
3. BI’cni = Blows uer centiineter

In summary, while the performance of CAPWAP Table 5. Summary of static and dynamic based capacity gain
is not well correlated to soil type, other factors asso- data sets (Paikowsky et al. 1996).
ciated with soil type may be important (e.g. low
driving resistance in soft cohesive soils or gain of Static Daia
Dytiamic Data
capacity with time) but soil type itself does not ap- Sets L TT atid All Data
Set PD/LTT
pear to be important. The data of Table 4 suggests
that time of driving must be considered and driving C,, t,s* CEld tX** c,, t,5**

resistance need to be further examined. No. 15 5 7 6 22 11

8.3 The efect of Time Avg 0.389 385.0 0.348 21.3 0.376 186.6

Penetration of piles in fine-grained soils causes
compression and disturbance, resulting in soil resis-
tance during driving that differs from the long-term
pile capacity. Although factors such as thixotropy
and aging contribute to this phenomenon, the most
significant cause for gain of capacity with time is as-
sociated with the migration of pore water. Meas-
urements carried out on a model (Paikowsky & Hart
1998) and full-scale piles (Paikowsky & Hajduk
1999, 2000) show that pore pressure at magnitudes
similar to the total soil pressures create in clays
around the pile’s shaft zones of about zero effective
stresses, resulting in almost a complete loss of fric-
tional resistance. Pailtowsky et al. (1995, 1996) ex-
amined the static and dynamic gain of capacity with
time based on radial consolidation; a normalization
process was followed, allowing for a comparison be-
tween different pile sizes. Table 5 presents a sum-
mary of parameters describing the pile capacity gain
with time based on static and dynamic testing, The
presented data shows that while the rate of capacity
gain (normalized to the maximum capacity vs. time
Stdev. 0.1 19 226.3 0.068 7.9 0.106 237.9
* closed-ended piles only
*‘F
excluding the case from Denmark
on log scale) is similar based on static and dynamic
measurements (Cg, = 0.389, Cgtd= 0.348) the associ-
ated time for achieving 75% of the maximum capac-
ity (normalized for all piles to 254mm (lft) diame-
ter) is about 20 times greater. In other words,
dynamic testing (namely CAPWAP) while following
the physical behavior of capacity gain, exhibit this
gain much faster than the actual gain monitored by
the static load test results. The ramifications of
these conclusions are: (i) actual gain of capacity is
much slower than that exhibited by the dynamic
methods, (ii) scheduling of construction or testing
based on capacity gain should consider the reason
for time evaluation (i.e. actual loading in construc-
tion or dynamic testing as part of quality control),
and (iii) at present, the dynamic methods evaluation
should concentrate on the long term pile capacity.

292
 8.4 The Effect ojSoil Motion
8.4.1 Overview
Paikowsky and Chernauskas (1 996) had shown that
the stationary soil assumption, under which the
soillpile interaction models were developed, does
not reflect the physical plienomenon that occLirs dur-
ing pile driving. The use of pseudo viscous damping
serves as a inechanisin to absorb energy, but, as it
does not reflect the actual phenomenon, its correla-
tion to physical properties (e.g. soil type) or time of
driving cannot be achieved. If the motion of the dis-
placed soil is a major factor contributing to the en-
ergy loss during driving, a substantial portion of the
dynaniic resistance should be a function of two pa-
rameters: (i) mass/volume of the displaced soil that
is a function of the pile geometry, namely, sillall vs.
large displacement piles, and (ii) acceleration of the
displaced soil, (especially at the tip) that can be con-
veniently examined as a function of the driving re-
si st an ce .

8.4.2 Soil Acceleration

The energy loss through the work performed by the
displaced soil inass at the tip is directly related to the
acceleration of this inass. The detailed evaluation of
the soil's motion at the tip is beyond the scope of the
present paper and is described by Holscher (1995),
Holscher and Barends (1996), and Hajduk et al.
(2000). The indirect evaluation of these accelera-
tions can be performed through the driving resis-
tance, which is the measure of the pile's final dis-
placement under each hammer blow. With low
driving resistance, high acceleration and velocity
(i.e. free-end analogy) are developed at the tip. In
the case of high driving resistance (hard driving),
there is small acceleration at the tip, resulting in lit-
tle, if any, mobilization of the soil inass beyond a
radiating elastic wave. The corresponding energy
loss due to soil motion is, tlierefore, small.
To evaluate the blow count that identifies the tran-
sition between 'easy' driving with high soil accelera-
tion values and 'hard' driving with low values, the ra-
tio between the static capacity and the CAPWAP
prediction (Kw) is presented in Figure 7 against the
measured blow count values. The data were also
separated into intervals of 8 BPlOcm (2BPI) with
the ineaii and standard deviation of each group
graphed as a point and an error bar against the mid
point blow count of the interval. For example, for
driving resistance between 0 and 8BP 1 Ocm there
were 42 cases with a mean of 2.506 and a standard
deviation of 2.217 plotted at the center of the inter-
val, i.e. at 4BPlOcm. The data presented in Figure 7
shows that for the first two intervals (up to
l6BPI Ocm) the predicted capacity was substantially
lower than for all other intervals with a significantly
higher scatter. After approximately 16 blows per
10cm, the mean and standard deviation of the indi-
Figure 7. KSW versiis blow count for all pile-cases in
PD/LT2000.
vidual intervals fall well within the range of all
cases. The boundary of the dynamic method evalua-
tion based on driving resistance was defined, there-
fore, as 16BPlOcm (4BPI).
8.4.3 Soil Displacerneiit and Pile Areu Ratio
The volume of the displaced soil is identical to the
volume of the penetrating pile, except when pile
plugging takes place (Paikowsky & Whitman 1990).
The piles, therefore, can be classified as small (e.g.
H and unplugged open pipe) and large (e.g. closed
pipe and concrete) displaceinent piles. Additional
classification of open-pipe piles can take place ac-
cording to a tip-area ratio similar to that used for soil
samplers (Paikowsky et al. 1989).
As most soil displacement takes place at the tip
area, the classification of piles can be better served
by looking at the ratio between the pile's embedded
surface area and the area of the pile tip (Paikowsky
et al. 1994):
A = & = Szirface area in contact with soil
(1 1)
A,,,> Area of pile tip
According to this ratio, a pile that is traditionally re-
ferred to as a “large displacement” pile can behavc
like a small displacement pile if it is driven deep
enough. Because the frictional resistance of a pile in-
creases as the pile skin area in contact with soil in-
creases, the effect of the soil niobilizcd at the tip de-
creases. As the pile’s embedded surface area and the
skin friction increases, the energy losses resulting
from the mobilization of the soil mass at the pile tip
will decrease relative to the energy losses along the
side of the pile. For example, the area ratio for cylin-
drical (closed-end) piles is:

In which D = penetration depth and R = pile radius.

For the same pile diameter, this area ratio increases
linearly with depth. For example, a 356-mm (14-
inch) diameter pile will have an arca ratio of 69 at
the depth of 6.1 m (20 ft) and an area ratio of 360 at
the depth of 32 m (105 ft). It is clear that the cffect
of soil inertia at the tip in the second case will be
substantially snialler than in the first case, and the
pile may be classified as a “small displacement
pile”. A quantitative boundary of AR = 350 between
“small” and “largc” displacement piles was pro-
posed by Paikowsky et al. (1994).
Figure 8 presents the relationship between the arca
ratio and the ratio of the static capacity over CAP-
WAP prediction (Ksw) for 382 cases. The data were
separated into area ratio intervals of 175, with the
mean and standard deviation of each group graphed
as a point and an error bar against the mid point area
ratio of the interval. For example, for the 139 piles
with area ratio between 175 to 350, the mean was
1.656 and the standard deviation was 1.425 plotted
at thc center of the interval (i.c. at tlie arca ratio of
262.5). Figure 8 suggests that piles with area ratios
smaller than 350 present less accurate predictions
and larger scatters compared to the mean and the
scatter of all cases. Above an area ratio of 350, the
mean and standard deviation of the individual inter-
vals fall well within the range of all cases.
As the driving resistance may also affect the data
in Figure 8, the influence of the area ratio was fkr-
ther examined for piles with a driving resistance
greater than 16 BPlOcm at the EOD. The 71 cases
answering to these criteria are presented in Figure 9.
The data in Figure 9 suggests again that when ex-
cluding the easy driving resistance effects, the accu-
racy of tlie dynamic predictions are still lower with a
largc scatter for pilcs with area ratios smaller than
350. The boundary of AR = 350 between “small”
and “large” displacement piles was therefore con-
firmed based on database PD/LT2000.

294
 Table 6. Summary of the Performance of'the Dynamic Methods.

.z
c
Fh'WA
EOD
-5 Illoclified
GLlles EOD
135 1.073 0.573 0.534 0.45 0.35 0.27

62 1.306 0.643 0.492 0.60 0.47 0.37

BI.Ct. 16BPl O c i ~
1WAP EOD 99 1.656 1.199 0.724 0.48 0.34 0.25
Notes: EOD = End of Driving BOR = Beginning of Restrike
AR = Area Ratio B1 Ct = Blow Count
ENR = Engineering News Record Equation BP 1 Ociii = Blows per 1Ocni
I<SS = Ratio ofthe Static Load Test Results COV = Coefficient of Variation
to the predicted capacity

9 TIHE PERFORMANCE OF THE DYNAMIC
METIHODS CATEGORIZED ACCORDING TO
THE CONTROLLING PARAMETERS
9.1 The Amlysed Cases
The time of driving, driving resistance, aiid area ra-
tio proved to be the major controlling parameters of
the dynamic niethods. To facilitate the codes' sepa-
ration between design and construction, the database
was organized into these categories, followed by
subcategories of methods that use aiid do not use
dy ii aiii1c in eas uremeii t s, with subsets fo 11o w i 11g the
controlling parameters. Figure 10 presents for the
aiialyzed subsets, the number of cases in the set aiid
the normal distribution mean aiid standard deviation.
WEAP is utilized in the design stage. The analy-
sis is carried out for driving stress evaluation, lead-
ing to a load fs-ctor, not included i n this paper. The
use of the method for the evaluation of pile capacity
was examined through WEAP results for default in-
put values and the blow count at the EOD compared
to the static load test results. The presented data was
provided by GRL Inc. (Haiinigan et al., 1996).
For the construction category, tlie cases without
dynamic nieasureiiients evaluate the dynamic equa-
tions, specifically tlie ENR, the Gates, and the
FIHWA version of the Gates equation. The cases
with dynamic measurements evaluates both, CAP-
WAP and Energy Approach methods. The dynamic
methods are brolam down into subsets based on time
of driving, driving resistance, and area ratios. Judg-
ment arid statistics guidelines were used for the in-
clusion or exclusion of a few cases. For example, ex-
treme CAPWAP under-predictions (beyond 2 S.D.)
were observed at the EOD in one site. All cases 111-
chided easy driving and large area ratios and if in-
cluded in the general population of data, the EOD
statistics would have become 1.861 i 1.483
(niean+lS.D.). This site is included only in the sub-
category of blow count < 16 BP 1 Ocin and AR < 350.
9.2 The critical cases
The statistical analyses presented in Figure 10 al-
lows for the identification of the critical cases that
require calibration aiid developnient into resistance
factors. For exaiiiple, the CAPWAP cases include
(i) all data, (ii) EOD, (iii) BOR, and (iv) the worst
combination of soil motion effect (Blow count < 16
BPlOciii aiid AR < 350).
Table 6 presents a summary of the major catego-
ries of the dynamic methods that are identified from

295
 I
I
Dynamic, Analysis II

Construction Design

< 16 BPlOcm 2 16 BPlOciii < 16 BP 1Ocm 2 16 BP 1Ocm

*
1.306 f 0.643 0.876 0.4 I9 0.929 f 0.688 0.809 f 0.290
No. = 62 No. = 73 No. = 32 No. = 127

I I

iSignal Matching
(CAPWAP)
1.368 f 0.620
No. = 377
I
1

BOR (last)
Field Evaluation
Energy Approach
0.894 f 0367
N A = 171
I

I
BOR (last)
1.626 + 0.797 1.158 f 0.393 1.084 f 0.43 1 0.785 f 0.290
No. = 162 No. = 153

< 16 BPIOciii 2 16 BPIOcni

0.830 f 0.352 0.775 f 0.274
No. = 29 No. = 124
I I

AR < 350 AR 2 350 AR < 350 AR 2 350 AR < 350 AR L 350 AR < 350 AR 2 350
1.717 1.181 1.178 1 . 1 10 1.054 0.926 0.736 0.851
k 0.841 i 0.468 f 0.379 i 0.303 4 0.459 f 0.320 f 0.249 f 0.305
No. = 37 No. = 34 No. = 83 No. = 47 No. = 39 No. = 34 No. = 82 No. = 42

"All values represent the ratio of the static capacity based 011 Davisson's
failure criterion over the dynamic methods prediction, mean i: 1 S.D.

Figure 10. Flow chart presenting the sub-grouping of the dynamic analyses according to the controlling parameters
and the resulting, statistical parameters for a normal distribution function.

296

Figure 17. Histogram and Frequency Distributions for all
(384) FH WA modified Gates equation pilc-cases in
PDiLT2000.
Figure 10 as the cases that require calibration for a
resistance factor.
Histogram and frequency distributions were pre-
pared for tlie identified critical cases in order to ex-
amine the match between the actual data and the
probability distribution functions. Figures 1 1 thro-
298
Figure 18. Iiistogram and Frequency Distributions for EOD
default value GRLWEAP pile-cases, (99),data provided by
GRL (see Mannigaii et al. 1996).
ugh 18 present tlie data along with the calculated
iioi-inal and lognormal distributions.
9.3 Ititetviediate Conclusions
The data presented in Table 6 and Figures 1 1
through 1 8 lead to several preliminary conclusions:
(i) The sigiial matching procedure generally under-
predicts the pile’s capacity. Tlie method performs
very well for the BOR (last restrike) cases. (ii) The
simple Energy Approach provides excellent predic-
tion for evaluating the pile’s capacity during driving
(EOD). (iii) The above suggests that construction
delays due to restrike and costly signal matching
analyses need to be examined iii light of capacity
time dependency and economical factors. (iv) The
FHWA modified Gates equation provides very rea-
sonable predictions for evaluating the pile’s capacity
when dynamic methods are not carried out. (v) A
reasonably good match exists for most cases be-
tween the calculated lognormal distribution and the
observed data. For this reason lognormal functions
were used in calibration procedures of the resistance
factors. (vi) The traditional factors of safety pre-
sented in Table 1 can now be evaluated in light of
the available data. For example, the coefficient of
variation for the WEAP analysis is 0.724, which
practically means that the method is unsuitable for
the purpose of capacity prediction. The reduction in
the factor of safety from 3.50 to 2.75 in Table 1
when adding WEAP analysis to static calculations is
therefore unfounded. Tlie use of unspecified CAP-
WAP (general case) again does not justify the reduc-
tion of the factor of safety to 2.25 even though the
average prediction is conservative and hence the
ineaii case with a FS = 2.25 relates to an over predic-
tion ratio of 3.1 (1.368 x 2.25). hi comparison, the
use of FS = 2.25 with a specified CAPWAP at tlie
BOR is reasonable and is associated with an accept-
able probability of failure for single pile application
(approximately 3 cases out of 162, i.e. 1.85%, see
Figure 15).
 10 RESISTANCE FACTORS
10.1Metlzodology
Following Ayyub and Assakkaf (1999) and Ayyub
et al. (2000) the present project calibrates LRFD par-
tial safety factors using the First-Order Reliability
Method (FORM). FORM can be used to assess the
reliability of a pile with respect to specified limit
states, and provides a means for calculating partial
safety factors # and y, for resistance and loads, re-
spectively, against target reliability levels, PO.
FORM requires only the first and second moment
information on resistances and loads (i.e. means and
standard deviations), and an assumption of distribu-
tion type (e.g. normal, logiiormal, etc.). The franie-
work of the calibration process is presented in Fig-
ure 19.
In design practice, there are usually two types of
limit states: ultimate limit states and serviceability
limit state. Each can be represented by a perforin-
ance function of the foi-in:

in which X is a vector of basic random variables ( X I ,

X?,. . ., X,,) for strength and loads. The perfoniiance
function g(X) is sometimes called the limit state
ftinction. It relates the random variables for the
limit-state of interest. The limit state is defined
when g ( x ) = 0, and therefore, failure occurs when
g(x> < 0 (see Figure 19). The reliability index P is
defined as the distance from the origin of the space
of basic random variables (XI,X Z , . . , X,,)to the fail-
ure surface at the most probable point. The most
probable failure point is that point on the limit state
fLinction at which the probability density of the basic
random variables is greatest. This is also called the
design point. This relationship can also be used to
back calculate representative values of the reliability
index p from the current design practice.
In developing design code provisions for piles, it
is necessary to follow the current design practice to
ensure consistent levels of reliability over various
pile types. Calibrations of existing design codes are
needed to make the new design formats as simple as
possible and to put them in a form that is familiar to
users or designers. For a given reliability index p
and probability characteristics for the resistance and
load effects, the partial safety factors determined by
the FORM approach might be different for different
failure modes for the same or differing component.
For this reason, calibration of the calculated partial
safety factors (PSF’s) is important in order to main-
tain the same values for all loads at different failure
modes. In the case of geoteclinical codes, the cali-
bration of resistance factors is performed for a set of
load factors already specific in the structural code.
Thus, the load factors are fixed. In this case, the fol-

299
lowing algorithm is used to determine resistance fac-
tors: (1) For a given value of the reliability index p,
probability distributions and moments of the load
variables, and tlie coefficient of variation for the re-
sistance, compute mean resistance R using FORM.
(2) With the mean value for R computed in step 1,
the partial safety factor 4 is revised as:

where p~.,aiid are tlie mean values of tlie loads

and strength variables, respectively and y,,i = 1,
2,.. ., ii, are the given set of load factors.

10.2 Cnlctrlated factors

Figure 20 presents examples of the resistance factors
calculated based on the above procedure for the gen-
eral CAPWAP and Eiiergy Approach cases. As no
exact target reliabjlity has yet been established, the
factors were evaluated for target reliability values of
2, 2.5, and 3.0 associated with probability of failure
values of 2.3%, 0.62%, and 0.14%, respectively.
The factors were evaluated using load factors of 1.25 Figure 20. Calculated resistance factors for tlie CAPWAP and
and 1.75 for Dead Load (DL) aiid Live Load (LL), Energy Approach general cases showing the influence of the
respectively, and for DL to LL ratios ranging from 1 dead to live load ratio.
to 4. The obtained results presented in Figure 20
suggest very little sensitivity to tlie DL to LL ratio.
A parametric study was carried out for a generic co- fall into the general Eiiergy Approach case in Table
efficient of variation of 0.40 and dead to live load ra- 7. As a reasonable probability of failure for a siiiglc
tios ranging froin 1 to 10. No significant influence pile can be estimated to be approximately 1% to 2%
of the dead to live load ratio on the calculated resis- (considering tlie redundancy of pile groups), Table 7
tance factors was found. The large dead to live ra- presents representative resistance factors obtained
tios represent a wide possibility of bridge construc- from tlie average calculated from the target reliabil-
tion, typically associated with very long bridge ity level (p) of 2 and 2.5. These are prelimiiiary val-
spans. A summary of the calculated resistance fac- ues demonstrating the inethodology and by iio
tors for all tlirec targct reliability levels for the idcn- nicaiis represent the recoinmended or final approved
tified dynamic methods is provided in Table 6. AASHTO specifications.

10 3 Preliniirzniy Resr.s!aiice Factors 111.4Evaluution of the dyriarnic methods efliciericy

Table 6 was developed based on the cases presented
in Figure 10 and using the resistance factors pre-
sented in Table 6, the critical dyiiaiiiic cases were
reevaluated. The selected methods, their important
categories and reconimended preliiiiinary factors are
presented 111 Table 7. When reevaluating tlie calcu-
lated resistance factors and developing Table 7, the
ENR equation was omitted due to very poor per-
formalice The cases within a specific dyiiaiiiic
method that had similar resistance factors were
combined iiito one category. For example, the ROR
and general Eiiergy Approach cases had the same
calculated resistailcc factors so the BOR cases would
The resistaiice factors alone do not provide a meas-
ure for the evaluation of tlie efiicieiicy of tlie dy-
namic methods. Such efficiency can be evaluated
through the bias factor (mean of the ratio of the
measured over predicted), its coefficient of variation
(see Table 6) or the ratio of the resistance factor to
the bias factor, i.e. flnieaii Ksx, as proposed by
McVay et al. (2000). This ratio is provided for the
final selected cases iii Table 7. The efficiency val-
ues in Table 7 suggest that overall the higher effi-
ciency is obtained by the signal matching analyses
for the last restrike, followed by the Energy Ap-
proach at the eiid of driving (0.581 vs. 0.507).

300
Table 7. Preliminary resistance factors for the critical dynamic cascs.

Resistiince.fuctoi., 4 Repi.eseiztalive
Method Case 4/A4em KJ,V
f l = 2.0 P= 2.5 Resistaiice juctor, #
p i = 2.3% [ J i = 0.6%

Geiieral 0.68 0.54 0.61 0.446

Sigiiul EODc3.50
0.35 0.44 0.170
Matcliiiig . BC<l6BPlOcni 0'52
Dy ainic
Mea.siii.eiiieiits BOX 0.73 0.6 1 0.67 0.581

General 0.48 0.39 0.44 0.492

Elieigy
Approacli
EOD 0.60 0.49 0.55 0.507

Gates General 0.85 0.67 0.76 0.425

, Dynaiiiic
~ ~qiiatioiis F H WA Geiieral 0.42 0.33 0.38 0.404
iiiodified
WEAP EOD 0.48 0.34 0.41 0.248

11 CONCLUSIONS on restrikes seem to be marginal compared to the

Based on tlie presented data the following conclu-
sions are derived:
The compilation of a large database allows for
tlie evaluation of the dynamic methods, tlie ex-
amination of tlie Working Stress Design (WSD)
methodology (e.g. validity of tlie assigned fac-
tors of safety), and tlie development of new
methodologies such as the Load and Resistance
Factor Design (LRFD).
The dyiianiic methods performance is controlled
by the time of' driving and soil inertia, wliicli in
turn is controlled by tlie driving resistance and
the ratio of the soil displaced by the pile's tip to
the area of the soil along the shaft of tlie pile.
Tlie most commonly known dynamic equation,
the Engineering News Record (ENR) equation,
is shown to be completely unreliable and unrea-
sonable for use. In contrast, tlie Gates equation
and its variation, modified by the Federal High-
way Administration (FHWA), seen7 to provide a
reasonable assessment of tlie pile's capacity con-
sidering tlie absence of dynamic measurements.
The wave equation analysis performs poorly
when used for pile capacity evaluation. How-
ever, this conclusion should not be niistakcn
with the importance of the wave equation analy-
sis during the design stage. Tlie drivability study
and pile stress analysis often detei-niine tlie pile
type, geometry and tlie adequacy of the proposed
equipment.
Signal matching techniques (e.g. CAPWAP)
prove to be most reliable for long-term restrike
measurements. However, when evaluated on ef-
ficiency, the application of the signal matching
Energy Approach at the End of Driving (EOD).
These conclusions though representative of most
cases, cannot be based on statistical data alone.
For example, sites that exhibit a significant but
highly variable setup may economically justify
consistent and long-term restrikes along with
signal matching analysis.
Tlie field application of the Energy Approach
provides an exceptionally efficient evaluation of'
pile capacity during driving.
The framework for the development of resis-
tance factors as part of the LRFD methodology
seem to facilitate a design wliicli is better suit-
able for geoteclinical applications. The presented
work is only an initial stage in that direction.
A more complete code based on LRFD needs to
consider factors associated with subsurface vari-
ability, site-specific technology and previous ex-
perience, as well as amount and type of testing
during construction.
ACKNOWLEDGEMENTS
The presented research was sponsored by the
American Association of State Highway and Trans-
portation Officials (AASHTO), under project 24- 17,
in cooperation with the Federal Highway Admini-
stration (FHWA). Tlie panel of the research project
and Mr. David Beal of the NCHRP are acknowl-
edged for their stimulating demands. Messrs. Jerry
DiMaggio, A1 DiMillio and Car1 Ealy of the FHWA
are acknowledged for their interest, concern and
support. Drs. Gregory Baecher and Bilal Ayyub
from tlie University of Maryland contributed to sec-
tions 3.1 through 3.4, section 10.1 and performed tlie

301
calculations of the presented resistance factors. Drs.
Mile McVay and Bjorii Birgissoii of the University
of Florida and Dr. Ching Kuo of GeoStructures con-
tributed to a questionairre resulting in the informa-
tion presented in section 6. Dr. Frank Rausche of
Goble, Rausche, Likins (GRL) and Associates pro-
vided the data pertaining to the evaluation of
GRLWEAP.
DISCLAIMER
The research presented here is part of an ongoing
project, which is aimed at rewriting the AASHTO
Deep Foundation Specifications for the year 200 1.
The presented parameters are preliminary and do not
reflect in aiiyn'ay the recoiiiiiiended or approved pa-
rameters. The opinions and conclusions expressed
or implied in the paper are those of the eiiti-
ties/iiidividuals performing the research and are not
necessarily those of the Transportation Research
Board, the National Research Council, tlie Federal
IHighway Administration, tlie American Association
of State Highway and Transportation Officials, or
the individual states participating in the National
Cooperative Highway Research Program.
IGFERENCES
AASHTO, 1994. LRFD Bridge ntid Consrruclion Sjiec(/icir-
tio17.s. American Association of State Highway and Trans-
portat i o n 0ffic i a 1s, Washington, DC.
AASHTO, 1 997. Sfaiirlwci Specif;cutioti.s.fbl- Higlz~vuyBridges.
American Association of State Highway and Transporta-
tion Officials, Washington, DC., 16"' Edition (1996 with
1997 interims).
AASHTO, 1999. LRFD Bridge atid Cotisr~uclio~i Speci/icn-
tioizs. American Association of State Highway and Trans-
portation Officials, Washington, DC.
ACI, 1995, Builriiizg Code Requireiiients ,for Reiufol-ced Coti-
crefe. American Concrete Institute, Detroit.
AISC, 1994. Load and Resistance Factor Design. Muiiuul of'
Steel Cotistt*iicfioii, American Institute of Steel Construc-
tion, Chicago, IL.
API, 1989. Draft Recommended Practice for Planning, Design-
ing and Constructing Fixed Offshore Platforms - Load and
Resistance Factor Design. API RP2A-LRFD, American
Petroleum Institute, Dallas, TX.
ASCE, 1993. Minimum Design Loads for Buildings and Other
Structures. ASCE 7-93 (formerly ANSI A58.1).
AUSTROADS, 1992. A USTROADS Bridge Design Code, Na-
tional Office, AUSTROADS, Surry Hills, NSW, Australia.
Ayyub, B.M. 1999. Cost-Benefit Analysis of Risk-Infornied
Design of Structural Systems. American Society of Me-
chanical Engineers, the Research Committee on Risk
Technology, Washington, DC.
Ayyub, B.& Assakkaf, I. 1999. LRFD Rules for Naval Surface
Ship Structures: Reliability-Based Load and Resistance
Factor Design Rules. Naval Surface Warfare Center,
Carderock Division, U.S. Navy.
Ayyub, B., Assakkaf, I., & Atua, K. 199Sa. Development of
LRFD Rules for Naval Surface Ship Structures: Reliabil-
ity-Based Load and Resistance Factor Design Rules, Part
Ill - Stiffened and Gross Panels. Naval Surface Warfare
Center, Carderock Division, US.Navy.
302
Ayyub, B., Assakkaf, I. & Atua, K. 2000. Reliability-Based
Load and Resistance Factor Design (LRFD) of Hull Gird-
ers for Surface Ships. N u i d Engineers Joui.1~~1, ASNE,
May 2000.
Ayyub, B., Assakkaf, I., Atua, K., Engle, A., Iiess, P.,
Icaraszewski, Z., Kihl, D., Melton, W., Sielski, R., Sieve,
M., Waldman, J . , 61 White, G. 1998b. Reliability-based
Design of Ship Structures: Current Practice and Emerging
Technologies. T&R Report R-53, Society of Naval Archi-
tects and Marine Engineers.
Barker, R., Duncan, J.: Rojiani, K., Ooi, P., Tan, C., & Kini. S.
1991 . A4anuals j b i . (lie Design of Bridge Foutidations.
NCHRP Report 343. Transportation Research Board, Na-
tional Research Council, Washington, DC.
Butler, H.D., & Hoy, H.E. 1977. Users Munuu1,foi~the T a u s
Q u i ~ l i - L ~Metlzocl,ji,r
~d Fouiiclatioiz Load Testing. Federal
Highway Administration, Office of Development, Report
No. FHWA-IP-77-8, Washington, DC.
Canadian Geotechnical Society. 1992. Caiiadiuti Founckiliot?
Etigi~eer-ingA4~irziia1,3'd Edition. Bi-Tech publishers, Ltd.,
Richmond, British Columbia, Canada.
CIRIA 63. 1977. Rationalization of Safety and Serviceability
Factors in Structural Codes. Construction Industry Re-
search and Information Association, SWIP 3AU, Report
63, London.
Chellis, R.D. 1961. Pile Foundations, McGraw M i l l 2'ld ed. NY.
Davisson, M.T. 1972. Iiigh Capacity Piles. P~oceeditigs,Soil
A4ecliutiics Lecficre Series 011 Itiiiovatioris it7 Fourzdafiotz
Coristrucfioii. American Society of Civil Engineers, llli-
nois Section, Chicago, 8 1- 1 12.
DeBeer, E.E. 1970. I'roefondervindellijke bijdrage tot de studie
van het grandsdraagvermogeii van zand onder funderinger
op staal. English version, Geofech~ique,Vol. 20, No. 4,
387-41 1 .
Ellingwood, B., Galambos, T., MacGregor, J., & Cornell C.
1980. Developnient of a Probability-Based Load Criterion
for American National A58. National Bureau of Standards
Publication 577, Washington, DC.
Ellingwood, B., Galambos, T., MacGregor, J., & Coniell, C.
1982a. Probability Based Load Criteria - Assessment of
Cui-rent Design Practices. Jour-nal of [lie S r t u c t i d Divi-
sion, ASCE, Vol. 108, No. ST5, 959-977.
Ellingwood, B., Galambos, T., MacGregor, J., & Cornell, C.
1982b. Probability Based Load Criteria - Load Factors and
Load Combinations. Jourmd 03''the Stnictut.al Division,
ASCE, Vol. 108, ST5, 978-997.
FIHWA, 1988. FHWA Guide Specifications for Driven Piles.
Federal Highway Administration.
Freudenthal, A.M. 1947. Safety of Structures. T~ansactionsof'
the ASCE. Vol. 1 12, 125-1 80.
Galambos, T. & Ravindra, M. 1978. Properties of Steel for use
i n LRFD. Joul-tin1 U / S t ~ u c t w a Eiigiiieei.ing.
l ASCE, Vol.
104, NO. 9, 1459-3468.
Gates 1957. Empirical Formula for Predicting Pile Bearing Ca-
pacity. Civil Eiigiiiewirig, Vol. 27, No. 3, 65-66.
Goble, G . 1999. Geotechnical Related Development and 1111-
plementation of Load and Resistance Factor Design
(LRFD) Methods. NCHRP Synthesis 276, Transportation
Research Board, National Research Council, National
Academy Press, Washington, DC.
Goble, G . , Likens, G.,& Rausche, F. 1970. Dynamic Studies
on the Bearing Capacity of Piles - Phase 111, Report No.
48. Division of Solid A4echunic.s, Sfruclures, atid Mechani-
cal Desigiz. Case Western Reserve University.
Goble, G. & Rausche, F. 1976. Wave Equation Analysis of Pile
Driving-WEAP Program. Vol. 1-4, FHWA #IP-76- 14.1
through #IP-76- 14.4.
GRL. 1999. Pile-Driviizg Ana/yzet., PAK Users A4unual. Goble,
Rauschc, Likins and Associates, Inc.
GTR. 1997. Dynamic Pile Testing Report, Central Ar-
tery/Tunnel Project C07D2, I-90/Airport Ititerchange Ar-
rivals Tunnel - Phase I, East Boston, Massachusetts. Geo-
sciences Testing and Research, Inc. North Chelmsford,
MA.
GTR. 1998. Dynamic Pile Testing Report, Central Ar-
teiy/Tunnel Project C07D2, I-9O/Airport Intel-clznnge Toll
Plaza, East Boston, Massachusetts. Geosciences Testing
and Research, Inc., North Chelmsford, MA.
Hajduk, E., Paikowsky, S., Holscher, P., & Barends, F. 2000.
Accelerations of a Driven Pile and the Surrounding Soil.
Proceedings, 6'" International Conjerence on the Applica-
tions of Stress- Wave Theory to Piles, September 11-13,
2000. S2n Paulo City, Brazil.
Hamilton, J.M. & Murff, J. 1992. Selection of LRFD resistance
factors for pile foundation design. Proceedings Structures
Congress '92. American Society of Civil Engineers, San
Antonio, ASCE, NY.
Iiannigan, R., Goble, G., Thendean, G., Likins, G., & Rausche,
F. 1996. Design and Construction of Driven Piles Founda-
tions - Volume I1 (Working Draft). U.S. DOT Federal
Highway Administration. Washington, DC.
Hansen, B. 1953. Earth Pressure Calculation. Danish Technical
Press. Copenhagen.
Hansen, B. 1956. Limit Design and Safety Factors in Soil Me-
chanics. Bulletin No. 1, Danish Geotechnical Institute,
Copenhagen.
Iiansen, B. 1966. Code of Practice for Foundation Engineering.
Bulletin No. 22, Danish Geotechnical Institute, Copenha-
gen.
Holscher, P. 1995. Dynamical Response ojSaturated and Diy
Soils. Delft: Delft University Press.
Holscher, P. and Barends, F. 1996. In-sihi Measurement of
Soil-Motion near the Toe of a Dynainically Loaded Pile.
In F.C. Townsend, M. Iiussein & McVay, M. (ed), Pro-
ceedings, 5"' In tern n tion a / Coizfeimce of the AQp/icalion
of Stress- Wave Theory to Piles, September 11-13, 1996.
_nrla.1ld5, F1 76-36
Kulhawy, F.H. et al. 1988-1994. Reliablilty-Bascd Foundation
Design of Transmission Line Structures. Report EL-5507.
Electrical Power Research Institute, Palo Alto.
Kulhawy, F.H. et al. 1996. Proceedings, Uncertainty in the
Geological Environnzent. Madison, ASCE, NY.
Long, J., Bozkurt, D. & Kerrigan, J. 1998. Value of Methods
for Predicting Axial Pile Capacity. Ti~ansportation Re-
search Board Paper No. 99-1333, 76'" Annual Meeting,
Januury, 1999. Washington, DC.
McVay, M., Birgisson, Bjorn., Zhang, L., Pcrez., A & Putcha
S. 2000. Load and Resistance Factor Design (LRFD) for
Driven Piles Using Dynamic Methods - A Florida Per-
spective. Geotectinicul Testing Joio-nul, ASTM, Vol. 23,
NO. 1, 55-66.
MHD. 1988. Standard Specifications for Highways and
Bridges. Massachusetts Highway Department.
Middendoip, P. & van Weel, P. 1986. Application of Charac-
teristic Stress-Wave Method in Offshore Practice. Pro-
ceeclings oftlie 3'" Interriationu1 Coiljierence on Numerical
A4ethods in Oflshore Piling.
Moses, F. 1985. Implementation of a Reliability-Based API
RP2A Format, Final Report. API PRAC 83-22. American
Petroleum Institute.
Moses, F. 1986. Development of Prelimiilary Load and Resis-
tance Factor Design Document for Fixed Offshore Plat-
forms, Final Report. API-PRAC 95-22, American Petro-
leum Institute.
National Research Cot~ncil. 1995. Probabilistic Methods i n
Geotechnical Engineering. Washington.
National Research Council of Canada. 1977. National Building
Code of Canada. Ottawa.
303
Nowak, A. 1993. Calibration of LRFD Bridge Design Code.
Department of Civil and Environmental Engineering Re-
port UMCE 92-25. University of Michigan, NCHRP 12-
33.
Nowak, A. 1999. Calibration of LRFD Bridge Design Code.
Department of Civil and Environmental Engineering Re-
port UMCE 92-25. University of Michigan, NCHRP 12-
33.
Olsen, R. & Flaate, K. 1967. Pile Driving Formulas for Friction
Piles in Sand. ASCE JSMFC, Vol. 93, SM6, Nov. 1967,
279-297.
Ontario Ministry of Transportation and Communication. 1992.
Ontario Highway Bridge Design Code and Comnientaiy,
3'" ed.
Paikowsky, S. 1982. Use of Dynamic Measurements to Predict
Pile Capacity Under Local Conditions. M.Sc. Thesis,
Dept. of Civil Engineering Technion-Israel Institute of
Technology.
Paikowsky, S. 1984. Use of Dynamic Measurements for Pile
Analysis. Including PDAP-Pile-Driving Analysis Program,
GZA Inc., Newton, Massachusetts.
Paikowsky, S. 1995. Using Dynamic Measurements for the
Capacity Evaluation of Driven Piles. Civil Engineering
Practice, Journal of the Boston Society of Civil Engineers
SectiodASCE. Vol. 10, No. 2, 61-76.
Paikowsky, S. & Chernauskas, L. 1992. Energy Approach for
Capacity Evaluation o f Driven Piles. 4'" International
Coizference 011 the Application of Stress- Wave Theoty to
Piles. The Hague, Netherlands. 595-601.
Paikowsky, S. & Chernauskas, L. 1996. Soil Inertia and the
Use of Pseudo Viscous Damping Parameters. 5'" Interna-
tional Coiiference on the Application of Stress- Wave The-
ory to Piles. Orlando, FL. 203-2 16.
Paikowsky, S. & Hajduk, E. 1999. Design and Construction of
an lnstruiiiented Test Pile Cluster. Research Report sub-
mitted to the Massachusetts Highway Department, Sep-
tember, 1999. Boston, Massachusetts.
Paikowsky, S. & Hajduk, E. 2000. Theoretical Evaluation and
Full Scale Field Testing Examination of Pile Capacity
Gain with Time. Reseat-ch R e p o i ~to be submitted to the
Massachusefts Highway Department, December, 2000.
Boston, Massachusetts.
Paikowsky, S. & Hart, L. 2000. Development and Field Testing
of Multiple Deployment Model Pile (MDMP). Research
Report to be submitted to the Federal Highway Admini-
stration, April, 2000. FHWA-RD-99- 194 Washington,
DC.
Paikowsky, S. & LaBelle, V. 1994. Examination of the Energy
Approach for Capacity Evaluation of Driven Piles. US
FHWA International Conference on Design and Consfruc-
tioii of Deep Foundations, December 6-8, 1994. Orlando,
FL. Vol. 11, 1133-1 149.
Paikowsky, S., LaBelle, V. & Hourani, N. 1996. Dynamic
Analyses and Time Dependent Pile Capacity. 5'" Inlerna-
fional Confer.eiice on the Application of Stress- Wave The-
ory to Piles. Orlando, FL. 325-339.
Paikowsky, S., LaBelle, V. & Mynampaty, R. 1995. Static and
Dynamic Tinie Dependent Pile Behavior. Research Report
submitted to the Massachusetts Highwqy Department, No-
vember, 1995. Boston, Massachusetts.
Paikowsky, S., Operstein, V., and Bachand, M. 1999. Express
Method of Pile Testing by Static Cyclic Loading. Research
Report subriiilted to the Mussachusetts Highwaji Depart-
nient, October., 1999. Boston, Massachusetts.
Paikowsky, S., Regan, J., and McDonnell, J. 1994. A Simpli-
fied Field Method for Capacity Evaluation of Driven Piles.
FHWA Report No. FHWA-RD-94-042, September, 1994.
Paikowsky, S. & Whitman, R. 1990. The Effect of Plugging on
Pile Performance and Design. Canadian Geotechnical
Journal, Vol. 27, No. 4,429-440.
Paikowsky, S., Whitman, R. & Baligh, M. 1989. A New Look
at the Phenoinenon of Offshore Pile Plugging. Marine
Geotechnology,Vol. 8, No. 3, 2 13-230.
Rausche, F. 2000. Personal communication. February 2000.
Rausche, F., Goble, G. & Likens, G. 1975. Bearing Capacity of
Piles from Dynamic Measurements, Final Report. Ohio
Department of Transportation, Ohio DOT-05-75.
Rausche, F., Thendean, F., Abou-matar, H., Likins, G., and
Goble, G. 1997. Determination of Pile Driveability and
Capacity From Penctration Tests. Vol. 1-3, Final Report.
FH WA-RD-96- 179, Federal Highway Administration.
Siu, w., Parinii, S. & Lind, N. 1975. Practical Approach to
Code Calibration. Joumal of the Structural Division.
ASCE, Vol. 101, NO. ST7, 1469-1480.
Smith, E. 1960. Pile Driving Analysis by the Wave Equation.
Journal of Soil Mechanics atid Foundations,Aniericaii So-
ciety of Civil Engineers,August 1960. 35-61.
Standards Association of Australia. 1995. Australiam Stan-
dards, Piling-Design and Installation. Homebush, NS W.
Tang, W. 1993. Recent development i n geotechnical reliability.
Probubilistic tiielhods in Geotechnical Eiigineeriiig. Li, K.
& Lo, S-C. (Eds.), Balkema, Rotterdam. 3-28.
Terzaghi, K. 1942. Discussion of the Progress Report of the
Committee on the Bearing Value of Pile Foundations.
Proceediiigs, ASCE. Vol. 68: 3 1 1-323.
Stenersen, K. 2000. Load and Resistance Factor Design
(LRFD) for Dyiiainic Analyses of Driven Piles. To be
siibniittetl in Pnrlial Fiilfillimizt of the Requii-merits for. CI
Muster of Scieuce Degree. University of Massachusetts
Lowell, Lowell, Massachusetts.
Wellington 1892. Discussion of “the Iron Wharf at Fort Mon-
roe, VA.” By J.B. Cuncklee. Transactions, ASCE Vol. 27,
paper No. 543, Aug. 1892, 129-137.

304
 design codes for pile foundations - A review

G.G.Goble
GobLe Rausclze Likins and Associates Incorporated and Pile Dynamics Incorporated, University of Colorado,
Boulder, Colo., USA

ABSTRACT: Several Load and Resistance Factor Design (LRFD) codes for geotechnical applications have
been developed over the past few years. Some of these codes have covered all aspects of geotechnical design
while others have been primarily concerned with foundations. The basic structure of the driven pile parts of
these codes will be reviewed and critically compared, particularly with regard to the way capacity determina-
tion and dynamic testing is handled. In some cases dynamic testing is not even mentioned while in the other
extreme the resistance factor is governed by the amount of dynamic testing used. Also, the resistance factors
used in the various codes will be compared including a comparison of the load factors

1 INTRODUCTION The basic strength requirement of LRFD can be

The first Load and Resistance Factor Design codes
(LRFD) to be used were adopted in the early 60’s in
the United States and Denmark. They were devel-
oped independently and for quite different reasons
and neither of the developing organizations were
aware of the activity of the other. In the United
States the American Concrete Institute adopted a
code for the design of reinforced concrete buildings
(ACI 1963) that had a LRFD format. It covered all
aspects of the design of reinforced concrete elements
except foundations. Foundation design was ignored
except for the structural design of the reinforced
concrete footing. The footing size was selected to
satisfy the geotechnical requirements based on al-
lowable stresses, or in the case of deep foundations,
allowable loads.
The ACI’s motivation for the change was the
problem with the design of reinforced concrete ele-
ments based on an assumed linear elastic stress dis-
tribution on the cross section. In cases where rein-
forcement was in compression, this analysis was
very conservative. The economic advantages of the
new procedure were so great that the new code was
widely adopted within two years. The drive for this
change came from structural engineers.
The code change in Denmark (Hansen 1966) was
driven by geotechnical engineers. The motivation
was somewhat different than in the United States
and while it produced a similar the load and resis-
tance factors, were quite different.
stated as
where (Pk is the resistance factor for the k’” failure
mode, Rk is the nominal resistance in the kth failure
mode, yij is the load factor and Qij the load of the ith
load type in the jtil load combination. This notation
will be used throughout this paper.
In addition to the strength requirements covered
by Equation (l), there are other requirements that
must be considered. They include such behavior as-
pects as settlement, vibration, lateral displacement,
etc. In this paper the discussion will be limited to
axial pile strength since it is usually the primary de-
sign consideration. Many other factors must be con-
sidered in performing a complete design.
It is usehl to review the hrther development of
LRFD in the United States. The methods contained
in the original ACI LRFD Code were known as ul-
timate strength design at the time they were adopted.
After the new code appeared it was suggested that
the load and resistance factors could be generated
rationally using a probabilistic analysis (Cornell
1969). The resulting procedures were used in an
extensive study to develop load and resistance fac-
tors (Ellingwood et a1 1980). However, the original
load and resistance factors used in the ACE Code
were not determined in this manner but were se-
lected based on engineering judgement. There have

305
been some limited changes in the ACI load and re- the loads while in the second case the loads are
sistance factors since the 1963 version but those highly variable and must have a large factor of
changes have been small. They were made in the safety.
time shortly after the initial adoption. This paper will review and compare the load and
When probabilistic methods are used the statisti- resistance factors used in several recently adopted
cal distribution of both the various types of loads LRFD Codes. It should be noted that in order to
and the strength is used to select load and resistance meet the length limits it was necessary to simplify
factors such that a desired probability of failure is the presentation of the various factors. These codes
achieved. The two primary problems with this ap- are discussed in much greater detail in an NCHRP
proach are the lack of uniqueness when a single Synthesis (Goble 1999).
condition is used to generate three constants and the
limited data available on the statistical distribution
of the loads and the strength. It is not completely 2 LOADFACTORS
convincing that a probabilistic analysis is superior to
the use of judgement in selecting load and resistance During the examination of several different LRFD
factors. codes it was observed that some of the resistance
The load and resistance factors of LRFD replaced factors for a particular load combination had values
the single safety factor that foundation designers varying over a considerable range among the differ-
ent codes. Some of this variability may have its
have traditionally used in allowable stress design
source in differences in the load factors between
(ASD). In structural design by ASD, allowable codes. To make the comparison of resistance factors
stresses were normally used instead of a safety fac- easier, the load factors have been collected for the
tor and these stresses were selected and codified basic gravity load combination and they are given in
based on tradition and experience. In ASD struc- Table 1. There are several additional load combina-
tural design, the factor of safety was not visible to tions in the codes but this one is usually the most
the designer. But for most aspects of geotechnical important and a comparison between the values pro-
design, allowable stresses were not (and cannot re- vides a good general view.
alistically be) established. Instead strengths were
determined and a factor of safety was applied. Thus, Table 1. Load factors for selected codes.
in the geotechnical area there has been a tradition of Code Dead load Live load
evaluating the acceptability of a design based on a AASHTO Bridge Code 1.25' 1.75
calculated strength together with a factor of safety. ACI 3 18-95 1.40 1.70
The factor of safety should be based on the require- AISC and ANSI 577 1.20 1.60
ments of the strength variability and load variability. Ontario Bridge Code 1.201 1.40
When the conversion is made to LRFD, it is desir- Canadian Bridge Code 1.20 1.60
able that the design process produce designs that are Eurocode2 1.35 1S O
similar to those produced by ASD using a safety Danish Code 1 .oo 1.30
factor. This implies that the sum of the influence of Australian Code 1.25 1S O
the LRFD load and resistance factors should be API Code3 1.30 1S O
equivalent to the ASD factor of safety. There is a variety of load factors for the various
Safety factors used in driven pile design have tra- dead load types. The quantity 1.20 is the most
ditionally been selected by the geotechnical engineer common value.
based on the perceived need for safety in the pile de- This is a considerable simplification of the
sign. This is an inappropriate procedure since the Eurocode specification on loads and load factors.
same safety factor provides a margin of safety The definition of loads, load combinations, and load
against variability of the loads. Therefore, the factors is quite complex.
structural engineer should have a voice in the selec- The code of the American Petroleum Institute is
tions of the safety.factor. As an illustration of this concerned with the design of offshore petroleum re-
problem imagine two different design examples. covery platforms.
The first case is the pile foundation for a large multi-
story reinforced concrete building where the domi- The above brief summary shows the rather large
variation of Load factors now in use in the LRFD
nant load comes from the dead load of the structure.
Codes that were reviewed. The load factor on dead
Compare this case with the pile design for a dock load ranges from 1.OO in the Danish Code to 1.40 in
structure. The dock has very little dead load but the ACI Code. The most common value is 1.20 but
may be required to carry large live loads. In the first the Eurocode uses 1.35, almost as large as the ACI
case there is little requirement for factor of safety on

306
Code. The range of Values for load Factors on live 3 . I AASHTO LRFD Bridge Code (199 7)
load extends from 1.30 in the Danish code to 1.75 in
the AASHTO LRFD Bridge Code. It is quite under- Resistance factors for the soil strength limit state for
standable that load factors for live loads could be driven piles are tabulated in the LRFD Bridge Code
quite different among the various codes since the and that information is summarized here as Table 3
live loads could be more variable in one application and 4. In this table h, is a multplier defined by the
than another. The same conclusion cannot be installation control method. It is applied to a quan-
reached with regard to structural dead load. A dif- tity given by method of capacity determination. The
ference of 1.00 to 1.40, a factor of 40 percent, be- use of these resistance factors has proven to be un-
tween the Danish and the ACI Code seems quite clear to foundation designers. Consider some exam-
large. ples: Suppose that the capacity is determined by the
SPT-method at a site with piles having end bearing
and friction in sand. The specified resistance factor
is 0.45 L. If a bearing graph is used for verifying
3 RESISTANCE FACTORS
capacity during driving, h, is 0.85, giving a resis-
Some research has.been performed on the use of ra- tance factor of 0.38, a value associated with a factor
tional methods in determining resistance factors for of safety of about 3.7, a considerably larger value
single driven piles. . Goble, et al. (1980) reported a than would be used in current practice. In the same
set of resistance factors consistent with the require- case where PDA testing with static load test, the re-
ments of pile design. These factors were obtained sistance factor is 0.45, a value associated with a
by calibration with existing factors of safety. Goble factor of safety of about 3.0 again much greater than
and Berger (1994) and Berger (1989) reported on re- is normally used.
sistance factors obtained using available data bases
from Dennis and Olson (1983), Pennsylvania DOT Table 3. Resistance factors from the AASHTO
(Unpublished), and Rausche et a1 (1997) together LRFD Bridge Code
with a probabilistic analysis. These calibrations Method/Soil/Condition Resistance
were obtained using the software developed during factor
the National Bureau of Standards study (Elling- Skin friction:
wood, et al 1980) and, therefore, they are related to Clay 0.70 a,
the NBS load factors that were developed for build- Alpha method 0.50 a,
ing loads. These load factors are given in Table 1 as Beta method 0.55 A,
AISC and ANSI 577. Some of the results obtained Lambda method
by Berger are summarized in Table 2. A large vol- End bearing:
ume of pile capacity data has been assembled since Clay 0.70 A,
the work of Berger, and it would be desirable to re- Rock 0.50 a,
peat this analysis with the new data. In the work of Skin friction and end bearing:
Berger the accepted probability of failure was 3.0 Sand
standard deviations from the mean. This quantity is SPT-method 0.45 a,
known as the safety index. CPT-method 0.55 A V
Wave equation 0.65 a,
Table 2. Resistance factors for a safety index of 3.0 Load test 0.80 a,
Method Resistance factor
Static analysis 0.42 Table 4. AASHTO h, values
ENR 0.42 Installation control methods A V

Wave equation 0.50 Pile driving formulas 0.80

CAPWAP 0.73 Wave equation 0.85
PDA testing - 2% to 5% of piles 0.90
Barker et a1 (1991) generated the resistance factors PDA testing - 2% to 5% of piles with
for the static analysis methods contained in the Static test 1.oo
AASHTO LRFD Bridge Code (1994) using the ra- PDA testing - 2% to 5% of piles with
tional probabilistic approach on available estimates Signal matching 1 .oo
of basic soil property variability. This approach does PDA testing - 10% to 70% of piles 1.oo
not include the model variability, site variability, or
the fact that other means are normally used to de- If the SPT-method were used together with wave
termine capacity. They do not report any calibration equation analysis and static load test, then what
analyses of the other nominal strength determination should be used for the multiplier on L . There is no
methods nor do they mention the use of available guidance in the text of the LRFD Bridge Code.
pile capacity data bases. Which resistance factor should be selected?

307
 The concept of having the resistance factor re-
lated to the static analysis method is inconsistent
with current practice.
3.2 Ontario Bridge Code (1992)
The Ontario Highway Bridge Code defines the limit
states for strength and specifies resistance factors for
deep foundations subjected to axial compression
load as given in Table 5. The resistance factors
seem to be quite small when compared with the
LRFD Bridge Code. There is no discussion of the
methods used to obtain these values.
of elasticity of the pile material, C is the velocity of
stress wave propagation and the other two terms are
not defined. The source of this approach is not
given.
3.4 Florida DOT (1997)
The resistance factors recently developed by the
Florida DOT in a calibration study are given in Ta-
ble 7. This presentation is clear and presents a good
representation of the Florida practice, one of the
most modern in the United States.

Table 5. Ontario Bridge Code resistance factors for
deep foundations
Axial load Factor
Static analysis 0.4
Static test 0.6
Dynamic analysis 0.4
PDA test 0.5
3.3 Canadian Bridge Code (I 988)
The design requirements given in the Canadian
Bridge Code are quite brief. The soil limit state is
dealt with in three separate categories. For load tests
the resistance factor is specified as 0.5 for routine
testing and 0.6 for high-level testing and the two
levels of testing are defined. However, the defini-
tions are quite qualitative. If dynamic testing is used
the resistance factor is specified as 0.4 for routine
analysis and 0.5 for analysis based on parameters
obtained from dynamic field measurements.
Table 7. Florida DOT (1 997) resistance factors
Design methodology Resistance factor
SPT97 0.65
PDA 0.65
Wave equation analysis 0.35
Static load testing 0.75
The resistance factor for wave equation analysis
seems to be rather small when compared with the
value for SPT 97 a static analysis procedure. How-
ever, the Florida DOT follows a design process
where the static analysis is used only for determina-
tion of estimated pile length for bidding. It is inter-
esting that a resistance factor is used in this applica-
tion. This is the first presentation of such an
approach and it merits firther study. If data is col-
lected on the error in the length estimate on each job
it would be possible to adjust the resistance factor to
arrive at the best possible length prediction.

Table 6. Canadian Bridge Code (1988) resistance 3.5 Eurocode (I 994)

factors This document is lengthy and complex with a large
Type of unit Resistance factors number of descriptive, limiting conditions that are
Reinforced concrete 0.4 usually not stated quantitatively. The requirements
Cast-in-place concrete 0.4 for driven piles will be summarized briefly and
Expanded-base concrete 0.4 should be read with the understanding that only the
Prestressed concrete 0.4 more quantitative portions of the document are pre-
Steel H-section 0.5 sented. In the Eurocode, both driven piles and drilled
Unfilled steel pipe 0.5 shafts are handled in a single section and it will be
Concrete-filled steel pipe 0.4 discussed here in the same way.
, Wood 0.4 A nominal strength as used in North America
codes is not defined. Rather, the characteristic
The third category is geotechnical formula (static bearing capacity is defined to be
analysis) and the particular formula that is used is
not given but must be approved. The soil properties Rc, = SRc,
used in the formula are specified to be factored by
0.5 for cohesion and 0.8 for friction angle. The ca- where krn is the measured capacity, and the values
pacity resulting from the use of the factored soil of 5 are given in Table 8. The quantity 5 is deter-
properties is then factored by the resistance factors mined by the number of load tests used to determine
given in Table 6. The resulting capacity shall not be the capacity with values given for up to three load
greater than 2.5 EApK or the structural nominal re- tests. The values for 5 are specified without regard
sistance. The quantity E is defined as the modulus to the total number of piles on the job. Thus, a job

308
with 2,000 piles and one with 100 piles would have of soil properties from subsurface investigation in-
the same !-factor for three load tests even though for formation. It also recognizes the Danish Formula
the large job three tests would be a rather small for dynamic analysis of pile capacity based on driv-
sample. Only static load tests are mentioned in the ing resistance and the use of static load tests. It di-
code. No values are given for other methods of ca- vides piles into two safety classes with different re-
pacity determination such as dynamic test, wave sistance factors.
equation, or dynamic formula. There are no com- The following is quoted from the Danish Code of
ments regarding the various quality control proce- Practice: The coefficient cp given for piles and
dures. ground anchors and used for static design (per-
The characteristic bearing capacity is further re- formed with characteristic strength parameters) as
duced in a second step by component factors to ob- well as for assessment of the driving resistance, only
tain the design bearing resistance according to the apply to the load bearing capacity determined from
relationship shear strength tests on soil samples. In cases where
the load-bearing capacity is established by test
(3) loading, cp is defined for piles actually subjected to
test loading, while different values are given for the
where Rbk, and Rsk are the toe and shaft characteris- other (non-tested) piles. The resistance factors are
tic bearing capacities, respectively, and (pb and cps given in Table 10.
are the toe and shaft component factors, respec-
tively. The value Rtk is the total characteristic bear- Table 10. Danish Code (1 985) resistance factors
ing capacity where it cannot be divided between toe Safety Class Normal High
and shaft and cpt is the associated component factor Without test 0.50 0.45
on the total capacity. They have been inverted here With test 0.63 0.57
to more easily compare with the other codes. Actually tested piles 0.71 0.65
TABLE 8. Eurocode 6 factors'
Number of load 1 2 3 The resistance factors contained in the Danish
Code were selected to match current practice.
tests
Mean of R,,, used 0.67 0.74 0.77
Lowest of &,used 0.67 0.80 0.91 3.7 Australian Standard (I 995)
'The values of E contained in the Eurocode are The Australian Standard for the design and installa-
actually the inversi of those presented here. They
tion of piles is concise and specific. All aspects of
were inverted to more easily compare with other design are covered in addition to the axial strength
codes. requirements which are discussed here. The resis-
tance factors, given here in Tables 11 and 12 are
The values for ( p b , cps, and cp, are given in Table complete. No information is available as to the
9. The problem with multiplicative resistance fac-
methods used to generate the factors. It is interest-
tors is illustrated with this code. If the highest fac- ing that resistance factors are specified to be within
tors and the lowest factors are combined this gives a a range with suggestions contained in Table 12 for
range of resistance factors ranging from 0.70 to selecting the appropriate value within the range.
0.52. Structural design specifications have traditionally
given nominal strength values that were not-to-
Table 9. Eurocode cp values exceed quantities and lower values than those speci-
Component factors cp h (Pq cp t fied were used, as appropriate, in the judgement of
Driven piles 0.77 0.77 0.77 the engineer. This approach would imply that the
Bored piles 0.63 0.77 0.67 lower number in the given range is unnecessary.
CFA piles 0.70 0.77 0.71 However, the availability of the guidance given in
Table 12 makes the range of values useful to the de-
The resistance factors contained in Eurocode were signer.
developed by calibration to current practice. No Some of the factors of Table 12 will not be rele-
probabilistic calibration was performed. vant to all of the capacity verification methods con-
tained in Table 11. No recommendations are given
3.6 Danish Code of Practicefor Foundation regarding the use of measurements made at the end
Engineering (I 985) of driving (EOD) versus beginning of re-strike
(BOR). The specification does not deal specifically
The Danish Code discusses the use of static analysis with changes in the resistance factor with increased
and makes recommendations for the establishment numbers of tests.

309
Table 11. Australian Standard resistance factors Driven Bearing Piles (1999) including both ASD
Method of strength Range of values and LRFD versions. This code is for building de-
assessment sign in the United States and it can be used with ei-
Static test to failure 0.70 - 0.90 ther the ACI or AISC structural design specifica-
Static proof test 0.70 - 0.90 tions. Therefore, there are two different sets of
PDA test to failure with sig- resistance factors matching the different load factors
nal matching 0.65 - 0.85 of the two codes. The resistance factors for the ACI
PDA test to failure no signal loads were arbitrarily selected for inclusion in Table
matching 0 50 - 0.70
13.
PDA proof test with signal
matching 0.65 - 0.85
PDA proof testing, no signal Table 13. PDCA resistance factors
matching 0.50 - 0.70 Method of strength Range of values
Static analysis using CPT 0.45 - 0.65 determination
data Static load test'32 0.75 - 0.90
Static analysis using SPT PDA test3 0.70 - 0.80
data in sand 0.40 - 0.55 Wave equation 0.60
Static analysis using labo- Dynamic formulas 0.43
ratory data for cohesive 0.45 - 0.55 Static analysis - Clay only 0.43
soils 'A range of percentages tested from 0.5 - 10 percent
Wave equation 0.45 - 0.55 is specified
Driving formulae for piles 2Static tests can be replaced by dynamic tests at a
in rock 0.50 - 0.65 rate of 4 to 1.
Driving formulae for piles 3A range of percentages tested from 2-10 percent is
in sand 0.45 - 0.55
specified.
Proprietary displacement
piles, using well established
in-house formulae 0.50 - 0.65 This code is quite specific regarding resistance
factors for specified percentages of piles tested. Due
Table 12. Guide for selecting of resistance factors to limited space resistance factors are given here as a
Lower end of range Upper end of range range in some cases. In the code specific limiting
resistance factors are given for particular percent-
Limited site Comprehensive site ages of piles tested. This code also contains a
Investigation Investigation
Commentary that offers additional details and rec-
Simple method of Sophisticated design ommendations.
Calculation method
Average geotechnical Conservative
4 SUMMARY AND CONCLUSIONS
properties geotechnical properties
Published design Site-specific design The values given in the various codes can be com-
parameter correlations parameter correlations pared by assuming a specific live load/dead load ra-
Limited control Carefbl control tio and then determining the associated factor of
safety for the loads and resistance factors specified.
Less than 3% piles 15% piles dynamically A live load/dead load ratio of 0.3 was selected.
dynamically tested tested Then associated safety factors were determined and
Less than 1% piles 3% piles statically tested are summarized in Table 14.
statically tested For the case of the use of a static load test for ca-
pacity determination a large range of safety factors
The specification also includes some resistance are specified. The Florida, Ontario, Danish, Austra-
factors for the structural limit state for concrete and lia, and PDCA codes use values ranging from 1.70
timber piles. It is the most complete pile design to 2.08. In the case of the Australian and the PDCA
specification of all those reviewed. the range is controlled by the quality control used.
The Canadian Bridge code and the Eurocode use
considerably larger values.
3.8 PDCA LRFD Code (I 999)
When PDA testing is used the PDCA and Austra-
The Pile Driving Contractors Association (PDCA) lian codes are very similar in a range of about 2.0 to
has adopted recommended Design Specifications for 1.8. Florida, Ontario, and Canadian Bridge Codes

310
Table 14 Comparison of safety factors for the differ- port 343: manuals for the design of bridge foun-
ent codes, L/D = 0.3 dations. Transportation Research Board, National
Capacity Wave PDA Static Research Council. Washington, D. C., USA.
Determination Equation Load Canadian Standards Association. 1988. “Design of
method Test highway bridges.” CANICSA-5 6-8 8. Rexdale,
Florida 3.90 2.37 1.81 Ontario, Canada.
Ontario 3.12 2.50 2.08 Cornell, C. A. 1969. “A probability-based structural
Canadian code.” ACI Journal. December, pp. 974-985.
Bridge 3.22 2.50 2.50 Danish Geotechnical Institute. 1985. “Code of
Eurocode 2.68 practice for foundation engineering.” DGI Bulle
Danish 1.70 tin 36, 1 Maglebjergvej. DK-2800 Lyngby, Den
Australia 2.91-2.38 2.01-1.54 1.87-1.46 mark.
PDCA (ACI) 2.43 2.08-1.82 1.94-1.62 Dennis, N. D. & R. E. Olson. 1983. “Axial capacity
of steel piles in sand.” Proceedings of the Confer-
ence on Geotechnical Practice in OfSshore Engi-
use larger values in the range of about 2.5. Euro- neering. ed. SI. Wright. ASCE, pp. 389-402.
code and Danish does not mention the use of dy- Ellingwood, B., T. V. Galambos, J. G. MacGregor &
namic testing. C. A. Cornell. 1980. “Development of a probabil-
The use of wave equation is covered by all except ity based load criterion for American National
the Eurocode and Danish codes. Here the values are Standard A58 building code requirements for
somewhat more scattered ranging from about 2.4 to minimum design loads in buildings and other
3.9. structures.” National Bureau of Standards. Wash-
There are three clearly written design codes ington, D. C., USA.
available today, Florida, Australia, and PDCA. In European Committee for Standardization. 1994.
each of these cases the conditions for the use of “Eurocode 7: Geotechnical design - Part 1: Gen-
various capacity determination and quality control eral rules,” Central Secretariat, Brussels.
procedures are clearly linked to values of resistance Florida Department of Transportation. 1997. “1 997
factors. They are based on modern driven pile de- load and resistance factors for bridge design,
sign procedures. The other codes all contain some foundations.” September.
degree of vagueness in the definition of resistance Goble, G. G., F. Moses, & R. Snyder. 1980. “Pile
factor use. In some cases the code does not recog- design and installation specification based on the
nize modern driven pile design practice. load factor concept .” Transportation Research
Methods are available for determining load and Record 749, Transportation Research Board, Na-
resistance factors rationally by probabilistic analysis. tional Research Council. Washington, D. C, USA.
Studies should be performed using the many avail- Goble, G. G. & J. Berger. 1994. “Soil resistance
able data bases. It is absolutely necessary that the factors for LRFD of driven piles,” Proceedings,
selection of resistance factors recognize modern International Conference on Design and Con-
driven pile design practice. struction of Deep Foundations, Vol. 11. Federal
Highway Administration, Washington, D.C.,
USA.
REFERENCES Goble, G.G. 1999. “Geotechnical related develop
ment and implementation of load and resistance
American Association of State Highway and Trans- factor design (LRFD) methods,” Synthesis of
portation Officials, AASHTO LRFD bridge design Highway Practice, National Cooperative Highway
specrfications. Is‘ ed., 1997 Interim, AASHTO. Research Program, Transportation Research
Washington, D.C., USA. Board, National Research Council, Washington,
American Association of State Highway and Tran- D.C., USA.
portation Officials, 1994. AASHTO LRFD bridge Ontario Ministry of Transportation and Communi-
design specrfications. 1’‘ ed. Washington, D.C., cation. 1992. “Ontario highway bridge design
USA. code and commentary,” 3rded.
American Concrete Institute, 1963. “Building code Hansen, J. B. 1966. Code ofpractice for foundation
requirements for reinforced concrete.” ACI 3 18- engineering, Bulletin No. 22, Danish Geotechni-
63. Detroit, Michigan, USA. cal Institute, Copenhagen, Denmark.
Barker, R. M., J. M. Duncan, K. B. Rojiani, P. S.K. Pennsylvania Department of Transportation, Un-
Ooi, C. K. Tan, & S. G. Kim. 1991. NCHRP re- published Report. “Data base on pile driving rec-
ords and load test results.”

31 1
Pile Driving Contractors Association. 1999. “Rec-
ommended design specifications for driven bear-
ing piles,” St. Louis, Missouri, USA.
Rausche, F., G. Thendean, H. Aboumatar, G. E.
Likins, & G. G. Goble. 1997. “Determination of
pile driveability and capacity from penetration
testing,” Final Report, FHWA-RD-96- 179, Na-
tional Technical Information Service. Springfield,
Virginia, USA.
Standards Australia. 1995. “Piling - design and in
stallation,” 1 The Crescent, Homebush, NSW
2 140, Australia.

312
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Pile testing competitions - A critical review

J. M.Amir
Foundution Control Institute Limited, Herzlia, Israel
B. H. Fellenius
Urkkada TechnologyLimited, Ottawa, Ont., Cunadu

ABSTRACT: A number of pile testing competitions, using low-strain techniques, took place in different
countries between 1987 and 1996 and soon started heated debates by the professional community. Since the
outcome of these competitions was obviously contentious, they do deserve a critical review. Such a review
may serve as a basis for more productive competitions in the future. These, in turn, will highlight both
capabilities and limitations of existing methods and contribute to the advancement of new techniques.
The authors describe for each competition the testing scope and program, piles tested, the nature of the defects
installed, the participating parties, and the results obtained. They go on to analyze each competition,
specifically stressing those items that if done differently could have significantly improved the outcome.
Based on the lessons learned fiom these events, the authors present ground rules for fbture pile-testing
competitions. In addition, the importance of organizing competitions also in downhole testing applications is
strongly recom e n d e d .

1 INTRODUCTION 2 PILE TESTING COMPETITIONS

In spite of the rapid progress in piling techniques 2.1 Objectives

(and maybe because of it), defective piles and
drilled-shafts are still encountered at many
construction sites. Among all the methods designed
to test the integrity of bored piles, only two have
proved to be of real practical value: The sonic (echo)
and the ultrasonic (cross-hole) methods.
The sonic method was first applied a quarter of a
century ago (Steinbach and Vey 1975). Since then, it
has established itself as the leading method for
testing the integrity of all kinds of piles. With the
advent of handheld computing, sonic testing has
become more reliable and at the same time more
affordable.
The sonic method is based on pressing a sensor
against the surface of the pile head while hitting the
surface with a hammer. The hammer blow creates a
low-strain wave that travels down the pile and is
reflected fiom the pile toe, as well as fiom any
abrupt change in the pile impedance. The hammer
may be either plain or instrumented, and the results
may be analyzed and presented in either time or
frequency domain. An extensive treatment of the
sonic method is presented by Turner (1997). The
popularity of the method has brought a proliferation
of both instrumentation and testing laboratories.
Consequently, it naturally became a subject for
competition.
In principle, pile-testing competitions should be held
with the some or all of the following objectives in
mind:
1. Kindling the competitive spirit amongst
developers, manufacturers, and users of
equipment
2. Establishing the actual (as opposed to claimed)
capabilities and limitations of the method
3. Indicating where advances in the state of the x t
are required.
4. Serving as milestones to monitor progress in
both instrumentation and analysis tools
5 . Providing an opportunity for potential clients to
obtain reliable comparative data regarding the
performance of available instruments
3 COMPETITION OVERVIEW
Following is a brief’ summary of five such
competitions, held in Belgium, The Netherlands, and
the USA (three of them), respectively.
3.1 Ghent 1987
The frrst known integrity testing competition was

313
held in Ghent, Belgium in 1987 (De Jaeger et al.
1987). The Belgian Society for Soil Mechanics and
foundation engineering organized this event.
Altogether, the twenty test piles were constructed by
four different methods, five piles each:
1. Atlas helicoidal (sawtooth profile) piles,
diameter 430/530 mm
2. De Waal precast concrete driven piles, 320x320
mm
3. Socofonda CFA piles, diameter 460 mm
4. Fundex piles, bored with rotated casing and cast
in situ , diameter 390 rnm
The five testing firms that participated were given
the pile diameters, and told that the pile lengths
range between 11 m and 16 m. No defects were
knowingly produced, and the task in hand was to
determine the correct length.
The best overall length agreements were obtained
in the Socofonda and Fundex piles. The lengths
obtained for the Socofonda piles were 93% and
100% of the correct lengths. For the Fundex piles,
the spread was between 94% and 102%. Such results
are perfectly acceptable. On the other hand, poor
results were reported for the precast piles (82% to
120%) and the Atlas piles (101% to 125% for three
of the piles, with no results at all for the other two).
The testing of the Atlas piles also produced poor
results. Although the Atlas piles had the lowest L/D
ratio (26 to 30), all testers reported lengths that were
too high.
The conclusions from this exercise are as follows:
1. The precast driven piles were difficult to test
both due to higher shaft resistancc and to having
the highest L/D ratio (41 to 53).
2. The testers of the Atlas piles probably neglected
the fact that a helicoidal pile exhibits a wave
velocity that appears to be much lower than that
of a straight-shafted pile (Vyncke and van
Nieuwenburg 1987).
3. The CFA piles and the Fundex piles were the
easiest to test. This is probably due to their lower
L/D ratios (32 to 37) as opposed to the precast
piles.
3.2 California 1990
The California test program was carried out in the
fiamework of a research project for the Federal
Highway Administration (Baker et al. 1993). It took
place on two sites: Cupertino, with dry gravelly and
sandy soil and San Jose, with clayey soil under
groundwater. The piles had a nominal diameter of
915 mm and lengths ofbetween 7.6 m and 18.9 m.
‘l’here were five participants in the program,
applying four testing methods: Sonic echo (time
domain), transient dynamic response (fiequency
domain), cross-hole (ultrasonic) and radioactive
(gamma-gamma). All the participants were provided
in advance with the lengths and shapes of the piles
that they were to test.
314
3.3 Texas 1990
The Texas test program was a continuation of the
FHWA project (Baker et al. 1993). Altogether, nine
bored piles were constructed, seven of which had
known irregularities. All piles had a nominal
diameter of 915 mm, with lengths varying between
11 m and 24 m (L/D ratios between 12 and 26).
The irregularities were of different character and
magnitude. Four of the piles had a single planned
nccking at a dcpth of betwcen 3 and 18 m, the
reduction in cross section being between 12 and 50
percent. Three other piles had both increased and
decreased cross sections at various depths. In
addition to the planned defects, some unplanned
defects occurred and were recorded during
construction.
Five testing firms took part in this competition.
The testing firms received fbll information regarding
the subsurface conditions, as well as the lengths of
the two reference piles. No m h e r data about the
existence of defects was divulged.
The tests were conducted in two stages: In stage
one, only surface (sonic) methods were used. In
stage two, the contestants were allowed to lower
testing equipment into access tubes which were
prepared beforehand.
The results of the Texas program may be
summarized as follows:
1. At depths smaller than 7 m below the top of the
pile head, 80 percent or more of the testing firms
managed to identify all important defects in the
cross section.
2. The success rate dropped to 60 percent at a depth
of 9 m.
3. All participants failed when the defect (necking)
was located at a depth of 18 m.
4. Even at shallow depths, participants failed when
the reduction in cross section area was merely 12
percent.
5 . In general, an enlarged cross-section was more
difficult to fmd than a reduced one.
6. The participants had difficulty in determining the
length of the pile when there was a major
necking at mid-length, or when a defect existed
near the toe of a long pile.
7. As expected, no participant could identify a “soft
bottom” condition.
3.4 Delj? 1992
The Delft competition took place in conjunction
with the Fourth International Conference on the
Application of Stress-Wave Theory to Piles (Smits
1996). Twelve laboratories, employing six different
types of instruments, participated in the event. The
testing objects in this case were rather uncommon:
All of the ten test piles were made fiom precast
concrete and installed in the following way: First,
closed-end steel tubes were driven to a
predetermined depth. Then, a thin bed of sand was
placed at the bottom of the tubes and the precast
piles were lowered into the empty tubes and placed
on the sand-bed. The space around the piles was
then filled with a bentonite-cement mixture
supposed to represent the local soil stiffness.
The test piles had a nominal cross-section 250-
mm square. Two piles were straight shafied, with
respective lengths of 17 m and 18 m. Of the rest, six
piles were produced with the cross-section along a
given length either enlarged to 300 mm. square or
reduced to 200 mm square, or both. The two
remaining piles had a sawed notch, 10 mm. thick
and occupying one half of the cross section. Similar
notches were also produced in two of the piles with
enlarged section.
The testing circumstances were also noteworthy,
in two important respects: First, the participants
were not allowed to approach the piles, and the
notary public’s clerk was mobilized to hit the piles
with the hammer. Second, the participants were
given beforehand the exact shapes of all the piles,
and their task was to decided which of these shapes
best fits each of the reflectograms they obtained.
All the participants managed to achieve was a
correct fit for between three and seven piles, with a
mean success rate of 44%. The scores for the
individual piles varied between zero (straight shaft,
L = 18m) to 100% (straight shafi, L= 17m).
3.5 Texas 1996
This competition (Samman and O’Neill 1997) took
place on the campus of the University of Houston,
Texas. Altogether, twenty-two piles were tested.
Eleven of the piles had a diameter of 460 mm and
were bored to a depth of 4.6 m. The other eleven
piles were 760 mm in diameter and 7 m long in the
ground. Some of the piles were constructed with
polymer slurry while the rest were cast in the dry.
Six of the piles were regular piles, while sixteen
piles had planned built-in defects. These defects
were produced from 25 mm thick soft rubber mats,
laid horizontally. The mats were placed at different
depths, but not more than one per pile. Each
occupied between 10 and 50 percent of the total
cross section of the pile.
Eight laboratories took part: Two were from
government agencies, five were commercial and one
academic. The contestants were asked to report for
each pile whether it is sound or defective, and, in the
latter case, specify the depth and severity of the
defects. The participants were to submit two reports:
A preliminary report on the same day, and a final
report within five days.
The results, as can be expected, were far from
satisfactory: In the piles intended to be sound, only
315
7% of the tests confirmed the integrity. In the
anomalous piles, 82 % of the defects were found.
These rates improved somewhat in the fmal reports,
to 25% and 83%, respectively. The success rate for
the defective piles may seem impressive, but a
deeper look into the matter is far less encouraging:
Of all the reported defects, only 36% managed to fit
the depth of the defects within +20 percent. Of
these, only 37% provided the size of the defects
within -t20 percentage points. On the other hand, the
participants reported on the average 1.3 “phantom”,
i.e. nonexistent, defects per pile. None of the
contestants, or the testing instruments used,
demonstrated a markedly outstanding performance.
4 EVALUATION
4.1 Ghent 1987
The organizers declared that the piles were between
11 m and 16 m long. Thus, a tester who would have
reported a uniform length of 13.5 m (the mean of the
above limits) for all the piles would be accurate to
within f.10% in 90% of the piles.
While highlighting the influence of construction
method on testability, the Ghent competition totally
neglected the most important purpose of sonic
testing -- finding defects!
Publication of the test results was very complete,
and included a special seminar where the results
were presented and discussed.
4.2 California 1990
Since the organizers of the California testing
program gave the participants full details regarding
the planned defects, success rates have no meaning.
It is therefore questionable whether the California
test may qualify as a competition and therefore the
case is not pertinent to the present paper. (The
results are of course interesting in other contexts).
4.3 Texas 1990
While in California, the participants were given
information beyond that normally provided to
testers, in Texas they got too little. When a pile
testing firm is invited to a construction site, it is
customary and necessary to provide it with all
relevant information, such as soil data, pile
construction records and piling logs with the as-
made length and details of any irregular events that
may have happened. Testing under the “Texas
rules“, with no a-priori knowledge of the pile
length, is therefore the exception and detracts fi-om
the effectiveness of the results.
In other respects, the planning and execution of
the Texas tests was very effective. The tasks were
well graded from easy through difficult to
impossible. Thus, the performance of contemporary
systems was well defined. This competition proved
convincingly that the sonic method is a viable
technique for investigating pile integrity. It showed
that sonic equipment is able to identifj most
important defects where they matter most, that is in
the upper part of the pile. On the other hand, it
demonstrated that the sonic method is unable to
distinguish features that are relatively small or
located deep down the pile.
4.4 Deyt 1992
As expected, the Delft 1992 event triggered a
lengthy debate in Ground Engineering magazine
(Stain 1993, 1993a, Ellway 1993). The main
criticism was aimed at the following points:
1. Most routine sonic testing is done on cast in situ
piles, with an inherent variability of both
concrete quality and shaft resistance soil friction
and a rough top surface. Precast piles in an
artificial “soil” with smooth tops thus cannot do
not represent real-world life conditions.
2. The unusually high L/D ratio (72) is generally
considered to be beyond normal testing limits.
3. Most of the anomalies, and especially the saw
cuts, were outside the theoretical performance
envelope of the sonic method.
4. Actual testing was performed by inexperienced
people, not familiar with fine points of the test.
In all important respects, the Delft competition
did little to advance the state of the art, and in fact
was a large backward step from the Ghent affair.
With the whole setup being detached from the real
testing world, it had only reinforced the (erroneous)
belief that sonic testing is not to be taken seriously,
being based on little more than guesswork.
4.5 Texas 1996
In view of the poor results obtained in Houston
1996, the organizers declared that sonic testing “may
not be reliable enough to be regarded as a stand-
alone measure of the assurance of drilled shafts”.
Could it be that not the sonic method was to blame,
but the organization of the competition? In principle,
the Houston competition had the correct ingredients
to simulate a realistic testing assignment. The main
factor that detracted fiom the success of this
competition lay in the nature of the “defects”: To be
applicable, the sonic method utilizes a wavelength
that is large in comparison with the pile diameter. A
defect with a vertical dimension of 25 mm is
therefore well beyond the capability of the sonic
method unless it occupies all (or almost all) of the
cross section of the pile. Since the defects in
Houston occupied at most one half of the total pile
316
area, it took a lot of good luck to discover any of
them. The fact that some defects were placed as
close as 300 mm to the top only made things worse.
Moreover, a rubber sheet, such as that used to model
cracks in the piles, has a low stiffness when
unstressed. In contrast, a rubber sheet stressed by
the weight of the concrete has a considerable larger
stiffness that does not deviate enough from the
stiffness of the concrete. It would be have been very
hard to detect a reflection from the rubber sheet.
Since most competitors were keen to find defects,
and the nature of the defects made them practically
undetectable, the competitors found defects even in
perfectly good piles. This is the main explanation for
the participants discovering defects also in the good
piles (plain coin tossing would do markedly better!).
5 DOWNHOLE TESTING COMPETITIONS
Admittedly, the sonic method has a few basic flaws:
First, the wavelength used is about 3 m, which
provides rather poor resolution and second, both
input (hammer blow) and output (accelerometer
signal) are remote from potential defects. To
overcome these drawbacks, the industry has
developed instrumentation that is lowered into the
pile through-access tubes+
Historically, access tubes were first used for
testing piles with radioactive isotopes. This method
is now fast disappearing due to its limited range,
environmental limitations, and regulatory
requirements. It was largely replaced by ultrasonic
instrumentation, using wavelengths in the range of
50 mm to 100 mm. Modern ultrasonic equipment
(Amir and Amir 1998) combines long range (-3m)
with high resolution. With a suitable setup, it can
also perform tomographical imaging and produce
two-dimensional vertical sections. Another
technique, still experimental (Samman and O’Neill
1997), utilizes clear plastic tubes and a downhole
video camera.
In view of their obvious advantages, downhole
testing of piles has become the preferred method in
certain sectors such as bridges and high rise
buildings. The time is ripe to organize suitably
designed competitions which would greatly benefit
the piling industry.
6 RULES FOR FUTURE COMPETITIONS
To be effective, competitions must satisfy certain
minimum criteria. Based on the analysis of five such
competitions, The following rules are therefore
suggested:
1. The test program should be based on sound
theoretical foundations - participants must not
be asked to perform the impossible.
2. The tests should be carried out on real piles,
conventionally constructed in real soil.
3. The piles should have different lengths and
length-to-diameter ratios.
4. As a rule, organizers should create no more than
one anomaly per pile.
5. Anomalies should be carefully designed and
constructed to resemble, as far as possible,
anomalies that are actually encountered in
practice. This includes soil pockets and zones of
weak, honeycombed concrete.
6. The anomalies should be of different
magnitudes, with an importance ranging between
minor irregularities to complete discontinuities.
7. Tests should be carried out by experienced
personnel, familiar with the testing systems
8. Participants should be provided with normal
testing conditions.
9. Pile heads should consist of reasonably good-
quality concrete. Testers who desire to improve
the surface must be given an opportunity to do
--.
qc)
10. Participants should be provided with sufficient
soil data (borehole logs) and pile data in the
manner and to the extent usually provided to
testers on actual construction projects. This
includes the as-made lengths and any special
events observed during construction.
11.Participants should not get any data regarding
the special features installed in the piles.
12.In addition to the piles specially prepared for
testing, the competitors must be given an
opportunity, where possible, to test “ordinary”
control piles at the same site.
13.The integrity of the piles should be investigated
also by conventional methods, such as coring
and pile extraction, in order to provide reference
to both the integrity of the piles and the success
of the integrity testing competition.
14.The competitions setup and program should be
reviewed and sanctioned by a reputable
international body, such as APTLY, the Alliance
of Pile Testing Laboratories.
REFERENCES
Amir, E.l and Amir J.M (1998) Recent Advances In Ultrasonic
Pile Testing, Proc. 3rd Intl Geotechnical Seminar On Deep
Foundation On Bored And Auger Piles, Ghent ; 18 1-185
Baker, C.N., Parikh, G., Briaud, J.L., Drumwright, E.E. and
Mensah, F. (1993): Drilled Shafts for Bridge Foundations,
FHWA, McLean, Virginia
De Jaeger, J., Carpentier, R. and v.d. Broeck, M. (1987):
Integrity Tests, Ch. IV in Seminar on Pile Dynamic Testing
(Integrity And Bearing Capacity), Brussels
Ellway, K. (1993): Objectives Of Competition Are Unclear,
Ground Engineering, June; 8
Samman, M. M. and O’Neill, M. W. (1997): An Exercise In
Seismic Testing Of Drilled Shafts For Structural Defects,
ADSC Foundation Drilling, Dec. - Jan; 1 1- 17
Smits, M. Th. J. H. (1966): Ch. 3 - Pile Integrity Tests,
Application of Stress Wave Theory to Piles: Test Results,
Balkema, Rotterdam;25-54
Stain, R (1993): Test’s Integrity Is Questionable, Ground
Engineering JanuaqdFebruary, p. 7
Stain, R (1 993a): Competition Was Not Applicable, Ground
Engineering, April; 15
Steinbach, J. and Vey, E. (1975): Caisson Valuation By Stress-
Wave Propagation Method, .J. Geotech. Div. ASCE, Vol
101 No. GT4, April ; 361-378
Turner, M. J (1993): Integrity Test Usefulness Is Not The
Issue, Ability is, Ground Engineering, July/August pp. 27-
28
Turner, M. J. (1997): Integrity Testing In Piling Practice,
CIRIA Report 144, London
Vyncke, J. and van Nieuwenburg, D.(1987): Ch. I1 - Theorie
Van De Dynamische Proeven: A. Integriteit, Seminar On
Pile Dynamic Testing Integrity and Bearing Capacity,
Brussels.
Samman, M.M. and McNeill, M.W. (1997): Fiber-optic
inspection of drilled shafts, Foundation Drilling, Sept.-Oct;
16-19

7 SUMMARY

Pile testing competitions represent a major technical

and financial effort for organizers and competitors
alike. To profit from this investment, these
competitions should be planned very carefully. The
experience accumulated from pile testing
competitions in the past can serve as a good basis for
planning successful competitions in the future. Such
competitions should be open to all available testing
methods, both commercial and experimental. Such
events should be coordinated with APTLY and
published in full in a technical journal or conference
that is readily accessible to the general piling
community.
31 7
This Page Intentionally Left Blank
 Application of StreSs-Wave Theory to Piles, Niyama & Beim (eds) @ 2000 Balkema, Rotterdam, ISBN 90 5809 150 3

The need for quality assurance in the dynamic pile testing industry

J. F?Seidel
Monash Universily, Melbourne, Vic.,Australia

ABSTRACT: The theme of this Stresswave 2000 Conference “Quality Assurance on Land and Offshore
Piling” is one that is timely and worthy of exploration. This discussion paper has taken a broad view of many
aspects of quality assurance as they apply to engineering projects. The role of dynamic pile testing in the
quality assurance process is highlighted. It is concluded that as dynamic pile testing is a key aspect of quality
assurance in many foundation contracts, it is essential that dynamic testing also be subject to the same rigor of
quality assurance. It is asserted that the current ad hoc training process is unsatisfactory. The paper argues
for an industry standard for training and evaluation of competency. Several models are suggested. This
should be associated with a system of formal review for testers deemed to have only basic skill levels. A
scheme for prequalification for testing contracts, which applies conditions to testers with basic skill levels, is
also suggested.

1 INTRODUCTION ble). Submission of an NCR flags a problem which

Quality assurance has become a catch-cry in many
countries around the world. Like motherhood, the
pursuit of quality cannot be criticized. However,
unlike motherhood, the meaning of the term ”quality
assurance’’ is open to interpretation. Clients can also
try to effect quality assurance on their projects in
several ways.
In this section, a number of different methods and
applications of quality assurance will be discussed in
principle and with reference to both civil engineer-
ing projects and dynamic pile testing.
1.1 Qualily Assurance Systems
In some countries, the term Quality Assurance
(QA) denotes a system of checking, reporting and
documentation which that has become an integral
part of the contractual process. The purpose of the
QA system is to set in place a process which will en-
sure the final built quality of the product. Imple-
mentation of the QA system is usually the responsi-
bility of the entity (contractor, specialist sub-
contractor etc.) performing the work, There is no di-
rect supervision of the contract by the client or his
representative. Any aspect of the work which does
not conform with the specification is the subject of a
Non-Conformance Report, NCR (a particular termi-
nology used - equivalent terminologies are possi-
requires explicit consideration and rectification be-
fore consequent work can proceed. All work or rec-
tification must ultimately meet the contractual and
specification requirements.
Although ensuring the final construction quality
by this process is a laudable goal, the reality is that
quality assurance techniques do no more than ad-
dress consistency of processes and adherence to tar-
get outcomes. Quality Assurance as a system is well
suited to processes such as manufacturing, which are
repetitive, devoid of subjective decision and where
the product can be tested or measured to ensure
compliance with specification. Mass production of
consumers products is an ideal application.
By inference, QA techniques are not well suited
to processes which are one-off, involve subjective
decisions and experience, and where the end-product
is not always available for direct measurement.
Most civil engineering projects would fall into the
latter description.
Foundation engineering, in particular, is a field in
which the knowledge and experience of each mem-
ber of the team - designer, construction engineer,
foreman and crew are critical to successful con-
struction. Each project, and even each pile within a
project brings new challenges because of the
uniqueness of the ground and groundwater condi-
tions. For instance, the pile driver must be respon-
sive to changes in pile blow count, and must be

319
aware of conditions that will potentially damage the
pile during installation. The driller must prevent
collapse of the shaft, and must be sensitive to the
drilling equipment to ensure that a suitable founding
layer has been attained. These are all aspects of the
process which are based on knowledge, experience,
subjectivity, good practice, and even intuition, which
cannot be defined in a contract specification, or
measured for compliance. and are therefore outside
the ambit of a QA System.
This is not to say that QA Systems do not have a
role, or should not be an element of the engineering
process. Rather, it is important that the limitations
of these systems be understood and acknowledged in
the formulation of a more global approach.
What must be avoided is a belief that by virtue of
implementation of a QA System, that the end prod-
uct is necessarily of high quality. In the author’s ex-
perience, there is a real danger that the application of
such techniques can actually reduce the quality of
construction. The QA System requires such a pleth-
ora of documentation that time which should be
spent “at the coal face” ensuring real quality is spent
filling in forms to document compliance.
Dynamic pile testing can, and often is an integral
part of QA systems for driven piles, usually in the
aspect of confirmatory testing, which will be dis-
cussed subsequently. It is also possible and desir-
able, that the wider capabilities of dynamic pile
testing be utilized in setting parameters or guidelines
for the installation process, i.e. as an integral part of
the construction control process. However, this use
is generally underutilized, as it is not as easy to pre-
scribe this type of application in the System.
It is also noted that dynamic pile testing is also
applied to only a limited percentage of the piles in-
stalled on a contract (usually 15% or less). An in-
ference must be made that the quality of the test
sample is similar to the quality of the remaining
piles installed.
1.2 Qiiality Asszirance - sirpervision
A more traditional approach to ensuring a quality
outcome for construction projects has been by means
of independent supervision of the construction proc-
ess. Typically, the client employs a person knowl-
edgeable in construction techniques to oversee the
construction in either a part-time or full-time capac-
ity. Duties typically include taking spot measure-
ments of critical dimensions. and ensuring compli-
ance with the specification and with good practice.
This method is still in wide use in some countries,
and with particular clients.
The primary benefit of this approach is that the
quality of the construction process and the com-
pleted product is assessed by a person independent
of the process itself, and whose exclusive role is to
check for quality. QA systems. relying as they do
320
on self-regulation and the professional integrity of
each player in the process, are open to criticism on
this account.
The potential disadvantages of this approach are :
e it is an additional cost to the project;
the liability for any problem may be partially
transferred from the contractor to the super-
vising body;
the effectiveness of this process is dependent
on the knowledge and experience of the par-
ticular superintendent.
Particularly for foundation engineering, where
many proprietary systems are used, and where the
skills of foundation construction are very special-
ized, the ability, knowledge and experience of the
person providing oversight will generally be inferior
to that of the specialist foreman and crew undertak-
ing the work. In this case, the supervision is not an
effective approach for quality assurance.
In the particular case of dynamic pile testing, it is
unlikely that a construction supervisor would be able
to effectively oversee the testing process.
1.3 Quality Assurance - professional review
Quality assurance can also be applied to the engi-
neering design and construction processes. This can
be in the form of either a design check, review or in-
dependent design. The independent design would
normally be performed by an external organization,
but checks and reviews could be undertaken both in-
house or by an independent external engineer. The
use of professional reviews is common for larger or
more complex projects.
It is implied in this process that the design or con-
struction records be transparent and verifiable, and
that independent reviewers with the necessary skills
to interpret this information are available. Further
comments on this aspect with regard to dynamic pile
testing will be made later.
1.4 Quality Asszirance - Preqzicdifcation
On larger and more complex projects, it is also
common for contractors or consultants to demon-
strate that they have the requisite skills and experi-
ence to successfully complete a project or provide
advice. Without prequalification, the organization
cannot even bid to provide services. Prequalifica-
tion might typically require the organization to pro-
vide details of previous relevant projects completed;
available resources; experience and qualifications of
key field and office personnel and references.
Prequalification is adopted by some statutory
bodies with regard to dynamic pile testing organiza-
tions. In the absence of any effective alternative, the
process is typically based on whether the organiza-
tion has provided testing services for a reasonable
length of time, or on the advice of independent refe-
rees. To .the knowledge of the author, this approach
is not widespread for dynamic pile testing. It is also
applied to organizations rather than individual test-
ers.
1.5 Quality Asstiva~ce- Direct Testing
A key element of quality assurance in civil engi-
neering projects is the testing of constituents, indi-
vidual elements or in rare cases, large or complete
systems. With specific reference to foundations, the
quality of concrete is routinely tested by standard
compression testing, or chemical testing. Weld
quality is also evaluated by inspection and ultrasonic
testing. There are either absolute or statistical stan-
dards with which the tests must comply. Such rou-
tine testing of constituents fits well within Quality
Assurance Systems or more traditional contractual
arrangements, as compliance or non-compliance is
easily demonstrated.
The testing of individual piles is also commonly
stipulated as part of a project specification. In this
case, a variety of techniques can be used. For driven
piles, the final pile set and (optionally) temporary
compression are measured for every pile to obtain a
measure of axial pile capacity using a driving for-
mula such as Janbu, Hiley, ENR or Gates. Although
these methods are applied to the entire population of
piles, they are known to be simplistic estimates of
capacity, and hence subject to high levels of uncer-
tainty.
In many countries, for all but the smallest proj-
ects, more sophisticated, and hence more reliable
testing is performed in some combination of static
load testing, dynamic pile testing, StatnamicB pile
testing, Osterberg testing and various pile integrity
testing techniques. The particular testing strategy
developed for each site is a function of economics,
available technologies, pile type, site conditions and
historical precedent.
Testing of piles as individual components of the
final system also fits well within the framework of a
Quality Assurance system in which quantitative out-
comes can be measured against the required specifi-
cation.
Engineers generally recognize that despite con-
ducting a targeted pile testing program, some un-
certainty still exists for the following reasons:
e Only a fraction of the contract piles are tested
using high-level techniques. The capacity of
untested piles must be extrapolated or other-
wise inferred;
a Every testing method provides an estimate of
the (axial) pile capacity at the time of testing,
and each method has an associated uncer-
tainty.
These uncertainties are accommodated by speci-
fication of ultimate capacities which incorporate an
appropriate factor of safety or through prescribed
321
load factors and strength reduction factors under the
Limit State approach.
Nevertheless, there is an important implied as-
sumption that the capacity estimate provided is the
best, unbiased and ‘correct’ estimate available
within the constraints of that technique.
2 QUALITY ASSURANCE -THE TESTING
PROCESS
As noted, direct testing is a key element of quality
assurance in many civil engineering projects. It is a
particularly common technique in foundation engi-
neering due to the uncertainties introduced by the
natural stratigraphy which is not known perfectly in-
advance of any foundation contract.
As verification is a key element of the overall
construction process, it would be ilIogical if the
testing were not itself subject to the principles of
quality assurance. This section considers quality as-
surance of the testing process in the particular con-
text of static and dynamic pile load testing.
2.1 Static Load Test@
Standards and Specifications which set minimum
requirements for compliance exist as part of the
framework of quality assurance for testing. For
static load testing, ASTM D 1 143-81(1 994)el
(ASTM, 1994) is the U.S. National Code of Practice
which sets standards for the equipment, calibration,
procedures and records required for testing piles un-
der axial static compressive loads. Equivalent
Codes of Practice for static pile load testing are
mandated in many other countries.
If tests are performed in accordance with the ap-
propriate standard, it is generally assumed that the
load-settlement response is a true representation of
the pile load-settlement response. This may not,
however, be entirely correct.
Fellenius (1 984) describes the large errors which
can be introduced by using a manometer attached to
the hydraulic jack which is simultaneously a load
application device and a load measurement device.
It is more appropriate to use a separate load cell so
that the load estimate is true and unbiased. AS2 159
(Standards Australia, 1995) mandates the use of a
load cell for static pile load testing.
The complexities of interaction effects between
the test-pile and reaction pile, anchors or reaction
weights are usually ignored, and will not be dealt
with here, other than to note that both the inferred
capacity and settlement characteristics can be af-
fected. Engineering analysis is required to correct
these physical effects.
It should also be noted that there are many alter-
native procedures for static pile load testing. Stan-
dard methods are known variously as Maintained
Load Tests, Quick Maintained Load Test and Con-
stant Rate of Penetration Test (other names are
used). Within these broad categories, an infinite
number of specific test regimes are possible. Due to
the different loading paths. any pile subjected to the
various tests will exhibit a different load-settlement
response. The significance of the different re-
sponses, the separation of elastic, plastic and creep
components, and the extrapolation to service be-
havior is a matter for engineering analysis and inter-
pretation.
Furthermore, Fellenius (1 980) notes that the in-
terpretation of ultimate capacity from a static load-
movement curve is not unique. Application of the
many constructions proposed (e.g. Davisson Offset
Limit , Brinch-Hanson, Chin-Kondner) result in sig-
nificantly different estimates of ultimate capacity.
It can be appreciated, therefore, that correct inter-
pretation of the simple static load test is more com-
plex than it would first appear. Analysis of the test
data would generally be performed by a specialist
geotechnical engineer using accepted methods that
are in the public domain.
In general. however, the assumptions made in the
analysis of the load test data are transparent (it is,
after all only a correction which is applied to the
measured response), and verifiable. The fraternity
of local geotechnical engineers is usually large
enough to enable a professional review to be made,
if required.
Furthermore, as noted previously, in the majority
of cases, the static load test response is taken to be a
true representation of the pile load-settlement re-
sponse. For contractual purposes, it is often neces-
sary only to ensure compliance with the specifica-
tion of the peak applied load and the settlement at
one or more defined loads. Further analysis is not
undertaken, and the test outcome is accessible and
open to direct and immediate interpretation by both
structural and geotechnical engineers.
2.2 Dynamic Pile Load Testing
Just as for static load testing, Standards exist
which prescribe the requirements for dynamic pile
testing methods. ASTM D4945-96 (ASTM, 1996)
sets standards for the equipment, calibration, proce-
dures and records required for testing piles using
dynamic impacts. AS2 159 (Standards Australia,
1995) stipulates requirements for the use and inter-
pretation of dynamic pile testing. The Institution of
Civil Engineers (ICE, 1996) have published specifi-
cations m d good practice guidelines on inter alia
dynamic pile testing. Other similar documents with
national authority exist, although fewer than would
exist for static load testing.
The interpretation of dynamic load tests has both
similarities and differences to the interpretation of
static load tests. They are similar to the extent that
322
correct interpretation requires specialist geotechnical
knowledge. However, the following important dif-
ferences exist:
The direct output of a dynamic test is not a
load-settlement response, but usually pile-
head strain-time and acceleration-time re-
sponses;
* The measured test response is a dynamic re-
sponse, and the static behavior which is to be
determined must be extracted from the test
using either simplistic or more complex ana-
lytical or numerical techniques.
The fact that the direct test outcome bears no resem-
blance to the load-settlement response is significant.
This means that most structural or geotechnical en-
gineers are unable to interpret the test result. To this
extent the results are not transparent, and the tech-
nique is therefore considered “black box” technol-
ogy - sometimes with the attendant negative conno-
tations.
Interpretation of these pile-head time records is a
specialized technique, which is generally known or
understood by the small number of practitioners who
are providing dynamic pile testing services. This is
not to say that information on how to interpret dy-
namic pile testing records is not well published and
available in the public domain. However, the reality
is that the technique is so specialized that those not
directly involved simply “leave it to the experts”.
Local professional review is also not generally
available, as the only potential reviewers are likely
to be testing or construction competitors, and com-
mercial and professional sensitivities about releasing
data have even resulted in claims of intellectual
property over test records. This results in a problem
with verification of results.
Dynamic pile testing suffers from the problem
that if the process is not typically transparent or veri-
fiable, the client is not in a position to independently
assess the skill, understanding and knowledge of the
tester. He also cannot assess whether the estimate
which has been made is actually the best estimate,
given the constraints of the technology.
It appears to be a universal experience, judged by
the author’s personal communications with col-
leagues around the world, that not all practitioners
providing dynamic pile testing services have ade-
quate skills. The author is aware of cases where dy-
namic pile testers have given gross errors in advice
due either to poor data quality which is undetected
or ignored, or due to misinterpretation or incorrect
analysis of the test records. For obvious reasons,
these cases are not detailed here. The author’s expe-
riences are not unique.
These errors in advice (if detected by the client)
can not only affect the client’s confidence in the
practitioner, but also their view of the reliability of
the dynamic testing method in general. This has an
unfortunate and undeserved flow-on effect to all
practitioners.
This is to say nothing about our responsibility as
professionals to provide professional advice, and our
moral obligation to ensure the integrity and safety of
the structure and people who could be affected by a
collapse.
We must collectively address the challenge of
quality assurance within the dynamic pile testing in-
dustry world-wide.
3 QUALITY ASSURANCE IN THE DYNAMIC
PILE TESTING INDUSTRY
The previous section has addressed quality assur-
ance of the testing process in the context of the
foundation industry and with regard to dynamic pile
load testing in particular.
The following section discusses some possible
strategies to ensuring that dynamic pile testing serv-
ices are provided in accordance with the principles
of quality assurance. This discussion is placed in the
context of current practice.
3.1 Training
A fundamental requirement for ensuring compliance
with the principles of quality assurance is that the
individual providers of advice are suitably compe-
tent.
Four stages have been identified in the general
development of competence. These can be applied
to the specific case of dynamic pile testing:
0 Stage 1 : Consciotisly incompetent. The per-
son is not able to perform a task in a compe-
tent manner, and is aware of their inability.
This person is not a danger, because he/she
will generally be prudent enough not to pro-
vide advice.
e Stage 2 : Unconsciously incompetent. The
person is able to perform tasks at a basic
level, but is unaware of what they don’t
know, and the implications of their advice.
This person is a danger, because he/she will
provide advice without the necessary skills to
assess whether this is provided on a sound ba-
sis.
e Stage 3 : Consciotrsly competent. This person
has reached a stage where they have achieved
a basic to advanced understanding. They are
also aware of what they don’t know and the
possible implications of their lack of knowl-
edge. In general, this person will provide ad-
vice within the limits of their knowledge, and
seek assistance in areas outside their compe-
tency. They would only be dangerous when
they do not seek appropriate advice outside
the limit of their expertise.
323
* Stage 4 : Unconsciously competent. This per-
son has become an expert; has an intimate
knowledge of the subject area which they can
apply without effort and can apply their
knowledge to areas beyond their direct expe-
rience.
In developing strategies for quality assurance in
the dynamic pile testing industry, all four stages of
competency must be catered for and addressed.
Systems should be in place both to identify the stage
of competency, and to prescribe an appropriate level
of autonomy or independence. It will be seen from
the previous descriptions that the Unconsciously In-
competent person - unaware of their own limitations
is the person that requires the greatest attention.
The key to competency is knowledge. Knowl-
edge is typically acquired by one of three methods -
by formal or informal training, education, mentoring
or reading; through experience; and finally by mak-
ing mistakes. All are powerful methods of learning.
Obviously, in a contractual environment, the last
method is undesirable, and should be avoided by one
of the first two methods. Experience, unfortunately
can only be acquired over time. Training is the tra-
ditional way to quickly develop competency.
There are no current standards or guidelines for
training within the industry, and no standards or cer-
tification which can demonstrate competency. Be-
cause of the highly specialized nature of this field,
the subject matter is not generally covered in an un-
dergraduate engineering degree. Because of the
limited interest and again because of the scarcity of
people qualified to teach in this area, there are no
formal post-graduate courses available in pile dy-
namics (to the author’s knowledge). There is, there-
fore, no formal qualification or certification which
can be obtained which indicates competency in this
field.
There is no requirement to either hold a degree or
equivalent undergraduate qualification in Civil En-
gineering, or in Engineering in general. This is not
to say that such a qualification is necessary, although
such a degree provides obvious background to some
of the principles employed in dynamic pile testing.
Guided experience, self-help and training can pro-
vide an equivalent level of competency over time.
An individual or organization new to dynamic
pile testing could expect one of the following train-
ing experiences:
* An on-site training program of 1 to 3 days,
generally provided in association with pur-
chase of new equipment. Training is con-
ducted by the equipment manufacturer, or a
designated agent. This training would only be
an introduction to use of the equipment and
underlying theory. At the conclusion of this
training, the user would be at either Stage 1 or
Stage 2 competency in the above model
 0 An informal training program conducted on-
the-job and in-house by a tester from the same
organization. The trainer in this case may
have been given the original training by the
equipment manufacturer, or may have been
given similar informal in-house training. In
this way, 2'Id or 3rdgeneration training is quite
common. There are no controls on the quality
of such training. It would typically concen-
trate more on the particular types of projects
encountered, and practical aspects of testing,
and would give less broad overview, princi-
ples and theory than the manufacturer's
training program. Manuals will generally be
made available to supplement the training.
The quality of the training will only be as
good as the competency of the trainer. Any
misconceptions of the 2'Id or 3rd generation
trainer will be passed on to the new trainee.
This.current ad hoc method of training is not an
ideal, and is not consistent with a quality assurance
approach.
In order to comply with quality assurance princi-
ples in training, it would be necessary to provide a
consistent minimum standard of training to all new
practitioners. If the training is to be industry-wide, it
would be desirable to provide this through an inde-
pendent industry organization or educational institu-
tion. The practicalities of providing uniform train-
ing to a dispersed world-wide market would tend to
suggest the need for the material to be delivered in
distance education mode. Concentrated workshops
could be an alternative model.
3.2 Testing Conipetency
The purpose of training should be to establish basic
competency in the providers of testing advice so that
at the completion of the training period they are
moving into Stage3 - Consciously Competent. Of
course, this stage will not be fully reached without a
period of field experience to reinforce the principles
learned during training.
With further experience, discussion with peers,
attendance at industry seminars and conferences,
there would be an expectation that the tester would
progress to advanced and eventually expert status.
Under a quality assurance philosophy. there
would be a need not only to provide and undertake
training, but to assess competency both of those that
undertake training, and those that are already pro-
viding dynamic pile testing services
e It needs to be established that testers have
achieved and maintain a basic level of com-
petency, and
0 Those testers moving from Stage 2 to Stage 3
(hence from a basic level to an advanced
level) should receive an appropriate level of
review from a tester with either advanced or
expert status.
3.2.1 Evaltiation of competency
Assessment of competency could be formulated
in different ways. Some of these proposals may be
difficult to develop in practice:
By a mentoring scheme in which assessors
(accredited by an independent industry group)
would evaluate the competency of an appli-
cant over a period of time. Both field and
analysis skills could be assessed.;
By an independent review panel (comprised
of acknowledged industry experts) that could
assess a submission from an applicant based
on examples of the applicant's work; a list of
projects completed; details of training under-
taken and referee reports. An interview proc-
ess could also be part of such a scheme;
By standard examination. This could cover
both data acquisition and data interpretation
skills, but could not assess practical skills on
site. A multiple-choice format would give a
most objective assessment of capability
Any method used to assess competency should be
able to effectively distinguish three levels of com-
petency - basic, advanced and expert. All ap-
proaches should be capable of giving feedback to
applicants on areas of weakness so that targeted
training and improvement is encouraged.
As skills in dynamic pile testing need to be rein-
forced by regular practice, it may be necessary to in-
stigate a system of regular review and re-appraisal,
particularly for those with only basic skills, and for
those who may test on an infrequent basis.
3.3 Review
Having established a system of assessing compe-
tency, it is a logical consequence to ensure that those
practitioners assessed to have at most basic compe-
tency be required to obtain review from others with
either advanced or expert status.
The ideal situation would be for such partnering
to be undertaken in-house with someone with higher
competency. This person would be required to "sign
off' on all testing and analysis undertaken by their
junior.
Where a person of higher qualification was not
available within the organization, arrangements
would be required for review by any eligible re-
viewer. The arrangements for such review, includ-
ing commercial and legal aspects, would be a matter
for joint agreement. It is anticipated that standing
arrangements would be made with a particular re-
viewer to streamline the process. This would lead to
an effective mentoring arrangement and encourage
transfer of knowledge.
 3.4 Preqzral~fic~rtion
As noted earlier, prequalification is a quality as-
surance technique which is used for contracts re-
quiring high levels of expertise or resources. Some
statutory authorities already require prequalification
for providers of dynamic pile testing services.
It is suggested that this system could be more
widely adopted, and that acceptance be based on the
levels of competency assessed by the evaluation
process. The requirements for review, as suggested
in the previous section could be formalized in the
conditions for prequalification. It would be impor-
tant that providers not be excluded from providing
services, but rather that their provision of services
should be accepted conditional on demonstration of
effective and timely review arrangements.
4 BASIC SKILLS
The continuing development of dynamic pile testing
equipment has enabled the operator to have a vast
array of information available in real-time. Algo-
rithms have been developed which provide critical
feedback to the operator at critical stages of testing
alerting the operator to potential hazards (such as
development of damage, excessive stresses, exces-
sive bending etc.) Many calculations which were
previously undertaken manually are now automated,
freeing the operator to concentrate on more critical
observations.
Although these developments are positive, there
is a danger that the operator will lose the ability to
critically evaluate the results which are provided by
the equipment. No algorithm is flawless, and no
equipment, however smart the software, can repli-
cate the abilities of an expert to critically evaluate
the pile responses and make appropriate judgements
and decisions. Without a detailed understanding of
the basis for the computations, the limitations of
these computations, and the implications of changes
in parameters or assumptions cannot be known.
If the operator is to remain in effective control of
the testing process, and not become a slave to the
equipment, it is imperative that the operator develop
skills to make the necessary independent critical
judgements. Only by developing these skills can he
progress to advanced and then expert level.
Following is a tentative list of the types of skills
necessary for effective dynamic pile testing. Many
of these relate to recognition of conditions from the
forceivelocity and upward/downward wave re-
sponses:
o Recognition of valid and invalid data in a va-
riety of scenarios;
8 Field measures required to rectify poor data;
325
The principle of proportionality; pile imped-
ance;
The relationship between mass density,
modulus and wavespeed;
Material properties;
Allowable stresses;
Principles of one-dimensional wave mechan-
ics;
Computation of pile wavespeed in easy and
hard driving conditions;
Recognition of easy, moderate and hard driv-
ing conditions;
Identification of high compression stress lev-
els at the pile head and at the pile toe;
Estimation of pile toe stress;
Identification of high tension stresses before
and after the 3L/c time;
Identification of damage intensity and loca-
tion;
Recognition of a broken pile, pile joints and
changes in section;
Recognition of response from end-bearing
piles, and piles with small and large shaft re-
sistance;
Estimation of end bearing, shaft resistance,
and shaft resistance distribution;
Recognizing high and low cushion stiffness;
Recognizing bending and poor hammer-pile
alignment;
Understanding the effect of resistance, or im-
pedance changes close to the transducers;
Principles of valid data adjustment;
Pile compression and tension capacity;
Estimating Case damping factor;
Basis of the Case method of capacity deter-
mination, and factors influencing capacity es-
timates;
The implications and conditions for pile un-
loading;
Pile set-up and relaxation;
Mobilization and under-mobilization of ca-
pacity;
Hammer performance and transfer efficiency;
5 CONCLUSIONS
This paper argues the need for dynamic pile test-
ing to be subjected to a rigorous industry-wide qual-
ity assurance scheme, compatible with the founda-
tion industry which the testing serves. The scheme
should be one which delivers real, fundamental
quality, and not one which merely produces docu-
mentation to satisfy administrative requirements for
an audit trail.
Ideas for a possible comprehensive model have
been suggested. The challenge to adopt such an ap-
proach should be taken up by all players in the field
of dynamic testing including practitioners, industry
groups and clients.
Without adopting the principles of quality assur-
ance, the reputation of the industry as a whole will
continue to suffer at the expense of alternative test-
ing methodologies.

REFERENCES
ASTM ( 1994) ASTM D 1 143-8 I ( 1 994)e 1 Standard Test
Method for Piles Under Static Axial Compressive Load.
American Society of Testing and Materials, West Consho-
hocken, PA
ASTM (1996) ASTM D4945-96 Standard Test Method for
High-Strain Dynamic Testing of Piles. American Society
of Testing and Materials, West Conshohocken, PA
Fellenius, B.H. (1980) The analysis of results froin routine pile
load tests. Ground Engineering, September, 1980 : 19-3 1.
ICE (1996). Specification for Piling and Embedded Retaining
Walls. Institution of Civil Engineers. Thomas Telford
Publishing, London ISBN 0 7277 2566 1
Standards Australia (1995) AS2159-1995 Piling - Design and
installation. Standards Australia, Sydney. 0-7262-9884-0

326
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

High strain dynamic pile testing, equipment and practice

G. Likins
Pile Dynamics Incorporated, Cleveland, Ohio, USA
E Rausche
Goble Rausche Likins and Associates Incorporated, Cleveland, Ohio, USA
G.G.Goble
Goble Rausche Likins and Associates Incorporated, Boulder, Colo., USA

ABSTRACT: High strain PDA testing has become common for verification of capacity of both driven and
bored piles. It also investigates other aspects of the installation of driven piles such as hammer efficiency and
driving stresses. The high benefit of dynamic test results for the relatively low costs of the tests has resulted
in widespread use. An overview of typical practice in the United States is given including a review of codes
which guides the practical applications. The equipment hardware and soRware required to meet the current
needs and the vision of the fbture is detailed, and the expertise required is discussed.

1 INTRODUCTION hammer is greatly different than assumed.

As technology changes, the construction industry
adapts. Construction of increasingly larger
structures requires increased ultimate loads. Piles
therefore become larger and longer. To install these
piles, the pile driving equipment has changed
dramatically from simple drop hammers to air, diesel
or hydraulically powered hammers with greatly
increased energy rating. Increased complexity and
increasing loads, coupled with a need to maintain
economy, requires better quality control and/or lower
safety factors.
A century ago, dynamic formula, now widely
considered as unreliable, was the only available
means of assessing the pile capacity from a
“measurement” of set per blow. Test piles were
subjected to static tests for final proof. In the 1950’s,
the wave equation method of analysis using digital
computers was developed and gave more dependable
results since it was based on more accurate hammer
and pile models, although the soil model was
empirically developed. Personal computers have
made wave equation analysis a readily available tool
for estimating capacity and driveability.
Wave equation analysis requires assumptions
about hammer system performance and soil behavior
models. Unfortunately, these unknowns can cause
considerable variation in results when for example
the hammer efficiency from a poorly maintained
Fortunately, we can now simply measure the
hammer performance and driving stresses as a result
of research beginning in 1964 at Case Western
Reserve University in Cleveland Ohio under the
direction of Dr. G.G. Goble (Goble et al. 1975,
Goble et al. 1980). The initial goal of the Case
research project was to measure the pile capacity
using the pile hammer as the loading device.
Capacity was evaluated by both closed form
solutions and discrete numerical analysis and
correlated with hundreds of static tests to refine the
procedures and establish databases to assure the
necessary reliability. In addition to estimating
capacity, the equipment developed over many years
into a complete inspection of hammer performance,
driving stresses and pile integrity and is known today
as a “Pile Driving Analyzer@”,or simply the ‘‘PDA”.
2 CURRENT PROCEDURES FOR
MEASUREMENT
The Case research project resulted in both closed
form “Case Method” solutions (for capacity, energy
transfer, driving stresses and pile integrity) and
rigorous numerical modeling CAPWAP’ software.
Both required measurement of force and velocity of
the pile. These parameters are routinely obtained by
measuring strain and acceleration with the PDA
using reusable bolt-on sensors for both strain and

327

Figure 1. Windows PDA processing program.
acceleration quickly attached to any pile size and
pile type in any weather condition. Sensors for
underwater pile testing applications have proven
reliable (Harnar et al. 1996). Recent advances
include “smart sensors” which remember their
calibration information and transmit it to the PDA.
As an alternative to measuring strain and converting
to force, top transducers used in the early research
project which were sized to the pile dimensions and
placed between the hammer and pile to transmit and
measure force are now being seriously considered
again. The signals are processed and data acquired
by a PDA. Modern PDA systems are PC based with
large mass digital storage and high resolution
graphic screens. Data is acquired and analyzed by
the user friendly Windows PDA-W program as
shown in Figure 1.
The PDA is traditionally operated by a trained
engineer who travels to the job site with the
equipment, prepares the pile, and attaches the
sensors to the pile. In recent times, pile preparation
and especially sensor attachment are often done by
the pile driving crew. The time required to drill the
holes and attach the sensors is often only 5 to 15
minutes per pile tested (substantially less effort than
a static load test).
The engineer first evaluates the data for quality.
The PDA program checks for loose connections,
unstable results, various ratios including the known
theoretical proportionality between force and
velocity at impact. If problems are detected, a
warning is given, and corrective action can be taken.
Once the data quality is satisfactory, the engineer
interprets the data on-site using the PDA. He gives
advice or opinions and answers questions that
prompted the testing request. Following the on-site
test, the engineer may further analyze the PDA data
with CAPWAP, a signal matching computer
algorithm which extracts the soil model from the
measurements and provides a simulated static load
test result.
Being on-site gives the engineer a feel for the
project. He sees how the piles drive, the condition
of the hammer, and the care that the crew gives to
the installation process. However, this on-site
process requires a considerable cost including hours
of travel time and large travel expenses. The
engineer usually arrives prior to the first test pile
being driven. Unfortunately, this is often a time
which is at best a guess due to weather, pile delivery,
and assembly of the driving system. Further, the
contractor’s efforts to drive the first pile may be
delayed due to a variety of reasons or equipment
problems which are usually resolved during the test
pile program and prior to production pile driving. In
summary, during the test pile program considerable
time is spent simply waiting. The active time for the
PDA engineer on-site is often small.
Figure 2. PAL on site (left) with cell phone (right)
3 NEW METHOD FOR DYNAMIC PILE
TESTING
A new version Pile Driving Analyzer (PDA) called
PAL has many powefil features to revolutionize
dynamic pile testing. This smaller unit has a
touchscreen for user friendly data entry and a self
contained rechargeable battery sufficient for a full
day’s operation. Data is stored on a removable
memory card. This PDA can be connected by the
engineer on-site directly to a laptop running the
Windows PDA program to provide full on-site PDA
processing capability.
However, modern technology now allows for a
more cost effective approach to dynamic pile testing.
The PAL on-site can be remotely operated through
cell phone technology by an engineer in the office,
see Figure 2. The sensors are attached to the pile by
the crew on-site or a trained technician. The pile
sensors are connected to the PAL, which dials the
cell phone and connects to the office PDA engineer.
From this point on, the PDA engineer operating the
PDA-W program on his office PC controls the
remote PAL to collect the data and send it digitally
to the office. The office engineer sees the data from
the on-site remote PAL in real time as the test is in
progress. Site observations are communicated with
the office engineer with a cell phone or by the PAL’S
“message communicator” (which sends short
messages either preprogrammed or user
constructed). In turn the office engineer interprets
the PAL data and communicates his advice to the
sit: pr.ctic.!!y jnstantaneously.
Why test in this remote mode? Simply stated,
there are significant cost savings. In many cases, the
cost to the owner of the engineer’s travel time to site
and the travel expenses approaches the cost of the
test itself. Since the pile crew can attach the sensors
to the pile, the PDA engineer could operate the PDA
from his office and save this time and travel cost.
With reduced costs, the project owner can perform
more testing for the same total cost. The PDA
engineer does not waste time waiting on-site for the
test. But perhaps more importantly, the data analysis
can begin immediately after data collection resulting
in earlier availability of final results to speed up the
construction and decision process. The report can be
issued sooner. On large projects, the PDA engineer
may still travel to the site for the initial test piles.
However, subsequent production pile quality
assurance testing can be performed remotely. This
results in lower testing costs and easier scheduling of
routine periodic tests. Thus, more tests can be
performed at lower costs resulting in improved
quality assurance for the project.
4 DYNAMIC TESTING USES AND GOALS
Obviously, whether testing on-site or remotely, the
dynamic test must be performed with the correct
goals to achieve the maximum benefit. Objectives
of high strain dynamic pile monitoring during
installation include evaluation of the energy transfer
and driving stresses to assure the pile can be safely
driven to the desired depth without damage.
Capacity may be of interest during driving to
establish the bearing layer or establish the driving
criteria. However, for long term capacity evaluation,
restrike data is usually required due to soil strength
changes with time.
During driving, the PDA calculates energy transfer
from the integral of the product of force times
velocity. The energy transferred is then compared
with the hammer’s rated energy. When low hammer
performance is indicated, the hammer can be
serviced (cushion, compressor and hoses, air valves,
fuel pumps, or piston rings inspected or corrected,
etc.) or the cause of low performance determined
and corrected so that hammer performance and thus
329
pile installation productivity is improved. Maximum
driving stresses are investigated to reduce the
likelihood of pile damage. The PDA gives the direct
compressive maximum at the sensor location. Since
up to four strain sensors are attached to the pile and
monitored separately, bending and local contact
stresses can be assessed and hammer pile alignment
improved. The maximum tension stress below the
sensors for concrete piles is computed from the pile
top measurements. If tension is excessive, a lower
hammer stroke, or increased pile cushion thickness
may be required. The maximum compression stress
is estimated at the pile bottom for all piles driven to
hard bearing layers. These stresses need little
interpretation other than comparison with stress
limits imposed by code or specification. The PDA
can investigate the pile shaft for damage.
Evaluation of pile bearing capacity is more
complicated because the soil strength is often altered
by the installation process. The capacity during
driving is often less than the long term pile capacity
particularly for piles driven in fine grained soils
(clays, silts and even fine sands) due to excess
positive pore pressures generated during driving
which reduce the effective stresses. As these pore
pressures dissipate after driving ends, the pile shaft
resistance increases. Capacity reduction during
driving is also caused by lateral pile motions which
create an oversized hole; with time, the overburden
pressures reach equilibrium on the pile perimeter
and increase shaft resistance. This phenomena of
capacity gain with time is called soil setup.
Therefore, dynamic testing during restrike tests after
a sufficient wait period usually yield a better
indication of long term pile capacity than a test at the
end of pile driving. The wait time required is longer
as the soil grains become finer.
Although less common, relaxation (capacity
reduction with time) has been observed. Relaxation
can be a serious problem for piles driven into
weathered shale, and may take several days to h l l y
develop. These losses can be caused by exposing the
shale to water, due to fracturing of the rock or by
heave from driving adjacent piles. Pile capacity
estimates based upon initial driving can significantly
over predict long term pile capacity. Therefore, piles
driven into shale should be tested after a minimum
one week wait either statically or dynamically (with
particular emphasis on the first blows). Relaxation
has been observed for displacement piles driven into
dense saturated silts or fine sands due to a negative
pore pressure effect at the pile toe. Restrike tests
after a few days are usually sufficient, with emphasis
on early “high energy” blows. With sufficient
experience on a site or in some limited geographical
area, the engineer may eventually reduce the amount
of restrike tests and just apply a reduction factor to
the end of drive capacity to estimate the final
resistance available.
 Dynamic load testing indicates the activated or
mobilized pile capacity at the time of testing. At
very high blow counts (above about 10 blows per
inch, or less than 2.5 mm set per blow), dynamic test
methods tend to produce lower bound capacity
estimates as not all resistance (particularly at and
near the toe) is fully activated. This can occur when
piles are driven to rehsal or in cases where setup
increases are large and the hammer is then not
capable of moving the pile during testing. Several
solutions to overcome this under prediction dilemma
for refbsal conditions are available depending on the
site conditions and availability of equipment. One
solution is to apply a few blows at higher energy.
The higher energy can be generated from a higher
stroke, or a larger hammer or a big drop weight with
more energy resulting in a smaller blow count (less
than 10 blows per inch set; greater than 2.5 mm set
per blow), and thus mobilize the full capacity.
Another solution for closed end steel pipe piles at
refusal is to fill the pile with concrete and then test
as a composite section. The now stiffer pile will
have a higher applied force and thus overcome more
resistance. Another method is to add the end of
drive end bearing to the restrike shaft resistance by
superposition to estimate the service condition total
load (Hussein 2000); this should only be done if the
restrike is at rehsal conditions and where no toe
relaxation is possible. Alternately, if setup (and
relaxation) can be estimated, the pile need not be
driven to the full required ultimate capacity but
rather only to a usually reduced “target” capacity or
to refbsal.
A CAPWAP analysis usually is made to confirm
the PDA field capacity result. The soil is modeled
similar to wave equation methods. The wave
transmission is modeled by the method of
characteristics (CAPWAP Manual 1999). The
hammer model is replaced by the measured force
and velocity as a boundary condition. Since these
measurements are redundant, the soil model can be
iteratively investigated. If the measured velocity is
input into the CAPWAP analysis model, then the
force required to hold the system in dynamic
equilibrium can be computed and subsequently
compared with the measured force (usually the wave
down is input and the wave up is computed and
compared). The soil model is adjusted either
automatically by the program or by the engineer
until the computed and measured forces (or wave
up) are in agreement. The final soil model
(distribution and dynamic and static parameters)
then describes the soil behavior during the hammer
impact. CAPWAP has been proven to have good
agreement with static load test results (Likins et al.
1996). The CAPWAP pile and soil models can be
subjected to a simulated static loading to produce a
load movement curve comparable to static test.
CAPWAP also provides resistance distribution
330
results where negative friction (downdrag) may be a
concern.
Numerous factors are usually considered in pile
foundation design. Some of these considerations
include additional pile loading from downdrag or
negative skin friction, soil setup and relaxation
effects, cyclic loading performance, lateral and uplift
loading requirements, effective stress changes (due
to changes in water table, excavations, fills or other
changes in overburden), settlement from underlying
weak layers and pile group effects. These factors
merit consideration when considering the
interpretation of dynamic testing results. The
foundation engineer should determine if any of these
considerations apply to his design.
5 USAPRACTICE
The USA is interesting since practically every
possible pile-hammer-soil combination can be
encountered. Most PDA testing in the USA has been
on driven piles with the testing engineer on-site with
the PDA. A considerable experience has been
accumulated. Early tests from the research project at
Case included extensive correlations with static load
tests. Private consultants began in 1972 to apply the
method in their own private sector projects, either
testing themselves or specifling testing in project
documents (often resulting in the contractor hiring a
testing firm to provide the service). Based on
favorable experience and the results of the research
project, highway agencies specified dynamic testing.
With time and the support of the Federal Highway
Administration for this testing, more highway
agencies have included dynamic testing in their
practice, either by acquiring the equipment or by
hiring outside consultants. As a result of this varied
exposure, many contractors see a benefit and now
hire testing firms to assist them when piling
problems arise, or they often propose dynamic
testing as an alternate when static testing is
specified.
For small projects with only a few piles, a couple
piles per structure are tested. Since the time to
install all piles is relatively short, often the piles are
tested during driving or with relative short wait
times during restrike. For medium sized projects,
the first production piles often serve as dynamic test
piles and are distributed over the site to check site
variability. Usually some restrikes are included in
the test program. In a growing number of cases,
longer wait times before restriking are being
employed to take more advantage of the ususal
strength gains with time. For larger projects, the
amount of static testing is generally reduced after
establishing a correlation of dynamically and
statically tested piles, and then supplemented by
additional dynamic tests to increase the total
percentage of piles tested. This improves the overall
quality assurance while reducing testing costs.
Because of the large amount of dynamic testing
performed, various codes and specifications are now
in place for USA application. In 1986, the D4945
consensus standard was adopted by ASTM
(American Society of Testing and Materials) for
“High Strain Dynamic Testing of Piles.” Testing in
the USA follows these guidelines (ASTM D4945-96
is most recent revision).
Beginning in the 19803, the Federal Highway
Administration offered a demonstration project for
dynamic pile testing. They currently provide
workshops for State highway agencies which
includes dynamic methods as an important part of
their recommendations. Many State highway
agencies have developed their own specifications for
use. Florida DOT has probably the most extensive
code and PDA is an integral part of their quality
control procedures. AASHTO has general
specification T298 (similar to but improved version
of ASTM D4945). In additional they recognize
dynamic testing with its own reliability and benefit
when assigning safety factors to foundations in the
“AASHTO Standard Specifications for Highway
Bridges”. In general, higher levels of testing result
in a lower safety factor.
The US Army Corps of Engineers included
dynamic testing in their 1993 manual on “Design of
Pile Foundations”. The American Society of Civil
Figure 3. Drop weight test on augercast pile.
331
Engineers (ASCE) developed a consensus standard
for pile testing entitled “Standard Guidelines for the
Design and Installation of Pile Foundations”(ASCE
20-96) which includes dynamic pile testing. The
private sector work generally has followed the local
highway department practice and the three regional
building codes. A recent joint effort of these three
regional code authorities resulted in the International
Building Code (IBC) with application in April 2000.
IBC allows ASTM D4945 testing to evaluate static
capacity. Before IBC adoption, many consultants
employed dynamic testing on their private projects
for either capacity evaluation or monitoring hammer
performance and driving stresses
Two USA organizations (Deep poundations
Institute and Pile Driving Contractors Association)
have endorsed PDA testing. DFI produced a
consensus document “Inspector’s Manual for Driven
Pile Foundations (Second Edition 1997)” with
positive comments on dynamic testing. PDCA has
promoted a model code (dated 1999) in both
working stress and LRFD versions which allows
substitution of static tests by dynamic tests aRer a
proven correlation with considerable reductions in
safety factor and thus cost savings for increased
percentages of piles tested.
It is conservatively estimated that every year
several thousands of piling projects have dynamic
pile testing on driven piles. While occasional PDA
dynamic testing for drilled piles in North America
has been made starting in 1974, this procedure is
now routinely applied on drilled shafts and augercast
piles in many countries in Asia, Europe and South
America using drop weights (Rausche et al. 1985,
Hussein et a., 1996). Figure 3 shows high strain
acceleration and strain sensors attached to an
augercast pile. The pile was extended above the
ground surface with a thin liner. The lower section
of the liner was then removed and the sensors
attached to the resulting smooth concrete using
anchors in the same manner they are attached to
driven concrete piles. The pile top surface is usually
flat and relatively smooth and only needs some
minimal plywood cushion to distribute the impact
over the entire top surface. A steel plate is then
placed above the plywood as a striker plate for the
impact weight, If reinforcement protrudes from the
pile top, the pile can be built up above the
reinforcement and then removed after the test.
To perform the high strain test, the drilled shaft or
augercast pile is then subjected to an impact of an
impacting weight. In most cases a simple drop
weight is preferred. Figure 3 shows a two ton drop
weight (four H piles welded together) being
positioned to test an augercast pile. Other drop
weight designs include solid steel cylinders, concrete
filled steel pipes or heavily reinforced concrete
blocks. As a general guide, the weight should be at
least 1 to 1.5% of the desired ultimate capacity to be
proven (Hussein, 1996) to assure load activation at
reasonable stresses. Larger existing weights can be
used provided the weight and shaft diameters remain
about comparable. Regardless of size, shape or
composition, the drop weight is generally guided to
an axial impact by a short set of leads as in Figure 3,
and is raised by cable and dropped by releasing the
drum brake. An alternative and preferable drop
method involves raising and securing the weight and
then completely releasing it for a true free drop (e.g.
releasing hydraulic jaws such as used for vibratory
hammers used to grab steel piles, or by tripping a
simple mechanical release).
The test for a drilled shaft usually consists of a
few separate impacts. A low drop height is first
applied to assess signal quality and alignment of the
weight with the shaft. After each impact, the net
permanent displacement or “set per blow” is
carefdly measured to evaluate full capacity
activation. Compressive stresses are compared with
the concrete strength. Alignment adjustments are
made if necessary and a second higher drop height is
applied. The test continues with increasing drop
heights until either the set per blow exceeds a value
sufficient to insure the full capacity activation, or
until the indicated capacity is above the required
ultimate capacity, or until the stresses become too
large and the risk of pile damage is then too high.
Most tests are completed in less than five impacts.
If the pile top has been built up to accommodate the
dynamic test, the extra top section is removed to
facilitate completing the foundation.
The measured pile top strain and velocity data are
analyzed by CAPWAP to independently check the
total capacity mobilized for each blow. A CAPWAP
analysis can be performed in a short time on site
after each impact to determine if the set per blow is
low so that the full capacity has not yet been
activated and another larger impact is required.
Upon completing the CAPWAP analysis, a
simulated static load test is obtained.
6 WORLDWIDE PRACTICE
PDA use outside the USA revolves around
established local practice. In many locations, testing
drilled shafts as described above is the primary
application. In some locations, testing of driven
piles is predominate and follows procedures
common in the USA. Many contractors drive
precast segmental regularly reinforced concrete
piles. Often these contractors obtain the equipment
and perform the testing themselves. In many cases a
“design-build’ process is common and the contractor
is then encouraged to find better foundation
solutions and is fully responsible for the foundation
installation; in such cases many have found great
benefit in dynamic testing to assure quality and
332
economy. In some countries the ultimate pile
capacity has been greatly increased as a result of
more dynamic testing; for example, in Sweden the
allowable loads for the same identical piles have
approximately doubled as a result of codes requiring
higher percentages of piles to be tested by PDA.
There are numerous country specific codes detailing
application of dynamic PDA testing (Beim et al.
1998). Monitoring hammer performance for
offshore oil platform installation is regularly
specified by the oil companies; if driving is greatly
different than expected, then capacity is further
evaluated.
7 CONCLUSIONS
Dynamic pile testing with the PDA with subsequent
CAPWAP analysis has become a routine practice for
engineers and contractors worldwide. The methods
have been applied to driven piles during driving to
monitor the hammer and driving stresses. Because
pile capacity is a function of time due to changes in
soil strength due to effects of the driving process,
long term capacity is usually evaluated during
restrike several days after initial installation. In
many parts of the world, dynamic pile testing
methods have been successfully applied to drilled
shafts and augercast piles by applying an impact of a
large drop weight. To activate the full soil resistance
(and thus correlate best with a static test failure load)
for either driven or drilled piles, the energy input
must be sufficiently large to produce a 2.5 mm set
per blow or more. Numerous codes and
specifications now direct the proper application of
dynamic testing. New technology has been
introduced to allow the engineer in the office to
remotely monitor dynamic pile tests on site.
REFERENCES
Beim, J., Gravare, C.J., Klingmuller, O., Li, D.Q., &
Rausche, F. 1998. Standardization and codification of
dynamic pile testing, a worldwide review. Proceedings,
seventh international confirewe on piling and deep
.fo2tndation.r Deep Foundations Institute. Vienna,
Austria.
CAPWAP manual 1999. Goble Rausche Likins and
Associates, Inc. Cleveland Ohio USA.
Goble, G., Likins, G.: & Rausche, F. 1975. Bearing
capacity of piles from dynamic measurements, final
report. Case Western Reserve University, Cleveland
Ohio USA.
Goble, G.; Rausche, F.; & Likins, G. 1980. The analysis of
pile driving - a state of the art. Proceedings, first
international conference on the application of stress-
wave theory to piles, Stockholm, Sweden.
Harnar, N.& Likins, G. 1996. Underwater dynamic Testing
Experience. Proceedings, Fijih International Conference
on the Application of Stress-Wave Theory to Piles,
Orlando, Florida, USA.
Hussein, M., Likins, G., and Rausche, F. 1996. Selection of
a hammer for high-strain dynamic testing of cast-in-place
shafts.. Proceedings, F@li International Confirence on
the Application ofStre.s.7- W m e Theory to Piles, Orlando,
Florida, USA.
Hussein, M., Sharp, M.R., & Knight, W.F. 2000.
Superposition for evaluating pile capacity. Proceedings,
sixth international conjkrence on the application of
stress-wme theory to piles, Sao Paulo, Brazil.
Likins, G.: Rausche, F.: Thendean, G.>& Svinkin, M., 1996.
CAPWAP correlation studies. Proceedings, Jisth
international conference on the application of stress-
wave theory to piles, Orlando, Florida, USA.
Rausche: F., Goble. G. 8r Likins, G. 1985. Dynamic
determination of pile capacity. ASCE joiirnal of
geotechnical engineering, Vol. 11 1: March 1985.

Other Referenced Standards:

ASTM D4945-96. Standard test method for high strain
testing of piles.
American Society of Civil Engineers, 1997. Standard
guidelines for the design and installation of pile
foundations, ASCE 20-96.
US Anny Corps of Engineer~~1993.Design of pile
foundations, Engineering manual (EM 1 10-2-2906).
U.S. Department of Transportation, Federal Highway
Administration, 1996. Design and construction of driven
pile foundations. Publication No. FHWA-HI-96-033,
two volumes.
American Association of State Highway and Transportation
Officials, "AASHTO", 1993. Standard method of test for
high strain dynamic testing of piles. AASHTO
Designation T 298-93.
American Association of State Highway and Transportation
Officials "AASHTO", 1996. Standard specifications for
highway bridges, (Sixteenth edition).
Deep Foundations Institute, 1997. Inspector's manual for
driven pile foundations, (Second Edition).
Pile Driving Contractor's Association (PDCA), 1999.
Design specifications for driven bearing piles.
Canadian Foundation Engineering Manual, 3rd Edition,
1992.
International Building Codc, 2000 Section 1807.2.8.3.

333
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) (CJ 2000 Balkema, Rotterdam, ISBN 90 5809 150 3

Dynamic load testing and Statnamic load testing for acceptance and design
of driven piles in Japan
T. Matsumoto - Department of Civil Engineering, Kcrnazawu University,Japan
K. Fujita - Department of Civil Engineering, Science University of Tokyo, Noda, Japan
0.Kusakabe -Department of Civil Engineering, Tokyo Institute of Technology, Japun
M.Okahara - Public WorksResearch Institute, Miizistry of Coizstruction, Tsukuba, J ~ p t z
N. Kawabata -Nippon Steel Corporation, Tokyo,Japan
S.Nishimura - Fugro Geoscience Company Limited, Tokyo, Japan

ABSTRACT: This paper presents the applications of dynamic pile load testing and the Statnamic test in
Japan, in view of the acceptance and the design of driven piles. This paper covers 1) a historical review of
research activities concerning pile load tests, 2) definitions of the terms “dynamic load test” and “rapid load
test”, 3) aspects of dynamic load testing and 4) interpretation of the Statiiamic test. Emphasis is placed on the
use of soil investigations and soil tests in parallel with the dynamic load test and the Statnamic test to obtain a
more reliable load-displacement curve from the tests. Also discussed is tlie importance of conducting re-
driving tests to estimate the load-displacement curve after ‘set-up’, the increase in the pile capacity with
elapsed time after pile driving, for piles driven in saturated soils.

I INTRODUCTION This paper is coniposed of 4 parts: 1 ) a historical

Intensive research activities on the dynamic load test
(DLT) have been conducted during the past twoe two
decades. and research activities on the Statnaniic test
(STN) have been conducted during the past decade
in Japan. These research activities are accompanied
by a change in foundation design in Japan from tlie
conventional allowable stress design to the limit
states design and the performance based design. In
the new design concepts, accurate estimation of load-
displacement curve, as well as quality assurance of
pile foundations, will become issues of vital
importance (Kusakabe 1998). Actually, research
activities on the dynamic load test and the Statnamic
test in Japan have been focused on the estimation of
the static load-displacement curve for a pile from tlie
test signals.
This paper presents the application of the dynaniic
pile load test and the Statnamic test in Japan, in view
of the acceptance and the design of driven piles.
Emphasis is placed on the use of soil investigations
and soil tests in parallel with the dynamic load test
and the Statnaniic test signals to obtain a more
reliable load-displacement curve froin the tests. Also
discussed is the importance of conducting re-driving
tests to estimate the load-displacement curve after
‘set-up’, the increase in tlie pile capacity with elapsed
time after pile driving, for piles driven in saturated
soils.
review of pile load test research in Japan, 2)
definitions of dynamic and rapid loading, 3) aspects
of dynamic load testing, and 4) interpretation of the
Statnaniic test.
2 HISTORICAL REVIEW OF RESEARCH
ACTIVITIES CONCERNING PILE LOAD
TESTS IN JAPAN
Principal research activities in the area of foundation
engineering in Japan are listed in Table 1.
Research activity on dynamic load testing was
initiated in 198 I by the Japanese Geotechnical
Society (JGS). The Committee on Estimation of
Bearing Capacity of Steel Pipe Piles from Dynamic
Load Test was formed in the Japanese Association
for Steel Pipe Piles in 199 1 led by Prof. Kazuma Uto
of Toltai University. They conducted dynamic load
tests on steel pipe piles having various diameters and
lengths driven in various soils with various types of
hammers, to compare the bearing capacity and the
load-displacement curves derived from the dynamic
load tests with the results of the static load tests of
the piles during 1991 to 1995. They also conducted
their own test program in 1993, in which static,
dynamic and Statnaniic tests on steel pipe piles
driven in sand were conducted to investigate the ‘set-
up’ phenomena and the performances of the various

335
test methods. The results of these research activities
are to be presented in two papers at this Conference.
The first Statiiamic tests in Japan were performed
in 1992 on a cast-in-situ concrete pile (Chosokabe et
al. 1993) and on a steel pipe pile (Matsumoto et al.
1994). In 1993, a private joint research group
(Research Group on Rapid Pile Load Test Methods)
was formed led by Prof. Osamu Kusakabe of Tokyo
Institute of Technology, aiming to compile the
existing knowledge about rapid pile load tests and
examine basic characteristics and applicability of the
tests, as well as the scientific interpretations of the
Statnamic test results. The Research Group
coiiducted their own research program, in which
several factors which might influence the Statiiamic
test results were investigated in the field, such as
type of pile, the installation method, the loading
history, the curing period and the soil conditions
(Kusakabe & Matsuiiioto 1995; Kato et al. 1998).
Based on the accumulated research results and the
continuing efforts of the Research Group, JGS
agreed to form the Research Committee on Rapid
Load Test Methods in 1996 to critically review the
features of rapid load tests, including the Statnamic
test, in more general teims, and to prepare a draft
document of testing standards for rapid load tests. In
1998, the Committee on Standardization for Vertical
Load Test of Piles was formed in JGS for the
revision of the current standards for vertical load
tests of piles, with the scheduled enforcement of the
revised standards in the year 2000. The revised
standards will iiicludes dynamic load tests and rapid
load tests, as well as various types of static load tests.
The Research Group and the Canadian Embassy in
Japan co-organized the second International
Statiiamic Seminar in Tokyo in 1998. More than 40
articles were submitted to the Seminar from 11
countries, showing the large increase in the use of
rapid load test in the world. A technical session on
the Global Srctndardization of Statnamic Testing was
held during the Seminar, aiming to have a global
equivalency of standards for rapid load tests.

Table 1 , Research activities for pile foundations by Japanese Geotechnical Society (JGS) and other institutes in Japan.
Year Institute Event
1980 - 1982 JGS (JSSMFE) Committee 011Standardization of Lateral Load Test of Piles.
1980 - 1982 JGS (JSSMFE) Research Committee on Bearinrr, Capacity of Piles Installed by Construction Methods with
Low Vibration and Low Noise.
198 I - 1985 JGS (JSSMFE) Chairman ofTechnica1 Corninittee on f’enctrability and Drivability of Piles of ISSMFF
(Prof. K. Fu,iita)
1982 - 1989 JGS (JSSMFE) Sub-Committee on Wave Propagation Theory.
1984 JGS (JSSMFE) National Symposium on f’enetrability and Drivability of Piles, Tokyo.
1 986 - 1987 JGS (JSSMFE) Committee on Standardization of Tension Load Test of Piles.
1986 - 1989 JGS (JSSMFE) Chairman of Technical Committee on Pile Driving of ISSMFE.
;
1987 - 1991 JGS (JSSMFE) Committee 011 Standardization of Vertical Load Test of Piles.
1987 JGS (JSSMFE) 32nd JCS Symposium oill~oundationsSupported by Friction Piles.
1989 JGS (JSSMFE) National Symposium on Pile Drivability and Application of Stress-Wave Theory to Piles,
Tokvo.
199 1 JGS (JSSMFE) National Symposium on Vertical Load Test of Piles and Decision of Bearing Capacity.

1991 - 1995 JASPP Committee on Estimation of Bearing Capacity of Steel Pipe Piles from Dynamic Load Test.
1992 Takenaka Corp. First Statnamic test on cast-in-situ concrete pile in Japan.
Kanazawa Univ. First Statnamic test on steel pipe pile i n Japan.
1992 - 1995 JGS Research Committee on Limit State Design of Foundation Structures.
1993 JASPP Comparative study on static, dynamic and Statnamic load tests on steel .pipe . .piles driven in
sand at Masaki test yard of’ Sumitorno Metal Industry.
1993 - 2000 Research Group on Research Group 011Rapid Pile Load Test Methods.
Rapid Pile Load
Test Methods
1996 - 2000 JGS Research Committee on Rapid Pile Load Test Methods.
1996 - present Architectural Sub-Committee 011Seismic Design of Pile Foundations.
Institute of Japan
1997 - present JGS Research Committee on I’resent and Future of Japanese Foundation Design and Soil
Investigation in view of International Equivalency.
1998 - present JGS Committee on Standardization for Vertical Load Test of Piles.
1998. 10 Research Group on 2nd International Statnamic Seminar, Tokyo.
Rapid Pile Load
Test Methods
2000 JGS Publication of Standard for Vertical Load Test of Piles.
(Impact load test (dynamic load test) and rapid load test are included in this standard.)
Japanese Society for Soil Mechanics and Foundation Eng. (JSSMFE) was renamed Japanese Geotechnical Society (JCS) in 1997.
JASPP: Japanese Association for Steel Pipe Piles

336
Figure 1. Classification of pile load test inethods in t e r m of relative loading duration
(after JGS Research Committee on Rapid Pile Load Test Methods 1998).
Above. we have reviewed the history of research
concerning pile load testing in Japan. In the next
section, we turn to the theoretical definitions of
dynamic and rapid load testing used in Japan.
3 DEFINITIONS OF DYNAMIC LOAD TEST
AND RAPID LOAD TEST
When the dynamic load test and the Statnamic test
are used to derive a static load-displacement curve
for a pile, allowance for dynamic effects should be
made in the interpretation of the test signals. The
difference between the dynamic theory and the static
theory is whether inertial forces are considered or
not, in the sense of physics. However, in the field of
geotechnical engineering, 'dynamic effects' also
include such things as (1) the strain rate dependency
of soil strength and deformation characteristics.
including creep, (2) the viscous effect on interface
strength between the pile and the soil, arid (3)
changcs in the drainage condition of the soil, in a
broad sense. It does not seem that current pile
technology and geotechnical engineering fully cover
such 'dynamic effects' observed in dynamic load
testing, Statnainic testing, and even in static load
testing. The degree of influence of each dynamic
effect depends on the penetration rate, the
acceleration of the pile, and the loading duration.
Several interpretation methods are available for the
dynamic load test and the Statnainic test, such as
applications of finite element method (FEM), one-
dimensional stress-wave theory and single rigid inass
modeling of the pile. Of course, these interpretation
methods have limitations in their application in
practice, accordiiig to the purpose and the loading
condition, as well as the soil conditions. Therefore.
it may be necessary to classify pile load test methods
337
in terms of loading and soil conditions, in order to
select the appropriate interpretation method for each
pile load test.
The Research Committee on Rapid Pile Load Test
Methods (1 998) proposed a practical classification of
pile load test inethods shown in Figure 1 in terms of
relative duration, t,., based on the works of Holeymaii
(1993), Middendorp & Rielefeld ( I 995), Karkee et
al. (1997), Nishimura et al. (1998) and Matsurnoto
(1998). The relative duration, t,., is an index for the
loading duration relative to the traveling time for a
stress-wave up and down in the pile, and is defined
as follows (Karkee et al. 1997):
I,. = T 1 ( 2 L l c )
in which T is the loading duration as defined in
Figure 2, L is the pile length and c is the bar wave
velocity of the pile. The relative duration, I,, is
identical to the relative wave length. A, defined as
(IHoleyman 1992) :
Each bar in Figure 1 indicates the range of relative
duration for which the corresponding dynamic effect
is not negligible.
The range of stress wave phenomena in the pile
which are negligible was determined from the works
of Middendorp & Bielefeld (1 995) and Nishimura et
al. (1998). They carried out simulations of one-
dimensional stress-wave propagation in piles and
found the relative duration beyond which wave
propagation phenoinena are negligible. The range of
the influence of the generation of excess pore
pressure is based on Holeynian (1 992) which
suggested the relative wave length, A, necessary for
90% of excess pore pressure to dissipate for various
soil types. The range of viscous damping (the
influence of the relative velocity between the pile
0 30 60 90 120 150
Time (ms)
Figure 2. Definition of loading duration.
Table 2. Typical loading duration for the boundary between
dynamic and rapid loading, and for the boundary between rapid
loading and static loading.
Pile type Concrete pile Steel pile
Bar wave velocity, c 4000 ni/s 5 120 m/s
Typical range of pile length, L 10 - 50 n1 10 - 100 111
Loading duration for t,. = 5 25 - 125 ins 20 - 200 ins
L,oading duration for t,. = 500 2.5 - 12.5 s 2 - 20 s
and the soil on the increase in shaft resistance) was
roughly estimated based on the proposal by
Randolph & Deelts ( I 992).
'rhe boundary between dynaniic loading and rapid
loading (Statnamic loading) may be regarded as
equivalent to the boundary between the loading
durations where the stress wave propagation
phenomena in a pile are negligible, and those where
they are not. This Research Committee on Rapid
Pile Load Test Methods regards 1, = 5 as the relative
loading duration representing the boundary betwxn
dynamic loading and rapid loading, based on the
works by Middendorp et al. (1 995) and Nishimura et
al. ( 1 998). It should be mentioned that even if the
wave propagation phenomena in a pile are negligible
f o r I , 2 5, the inertial force of the pile as a
concentrated mass can not be neglected.
The boundary between rapid load testing and static
load testing can be regarded as equivalent to the
boundary between the loading durations where the
soil resistance dependent on the inertial force of the
soil and the rate dependency of the soil strength can
be virtually neglected, and those where they can 1701.
The Research Committee regards t, = 500 as the
relative loading duration representing the boundary
between rapid loading and static loading, based on
the proposal by Middendorp & Bielefeld ( I 995).
As a consequence, rapid loading is defined as a
load test having the relative loading, I,, of 50 5 I, I
500. Table 2 shows the typical loading duration for
the boundary between dynamic and rapid loading.
and [or the boundary between rapid and static
338
loading, for concrete and steel piles. Typical loading
dlirations of the Statnamic test fall in a range from
approximately 100 to 200 ms. Therefore, the
Statnamic test can be regarded as a rapid load test, in
most cases.
In section 4 below, dynamic load testing in Japan
will be discussed and in section 5, an example of
rapid load testing, the Statnamic test, will be taken
up.
4 DYNAMIC LOAD TESTING IN JAPAN
I n this section, test equipment for gathering
dynamic test data, methods used for interpretation of
this data, and the use of the results of the
interpretation for acceptance and design of piles will
be discussed.
4.1 Test equipment
Dynamic load test equipment commercially available
in Japan are the PDA system developed by GRL
Corporation and the FPDS system developed by
TNO that are widely used all over the world. In
these dynamic monitoring systems, changes with
time of accelerations and strains near the pile head
are measured, and they are converted to the time
variations of the velocity, v, and the force, F.
It is not uncommon that the pile is mounted with
strain gages to measure the pile forces in a static load
test. When the dynamic load test is conducted on
such a pile. the variations with time of the strains
measured at two levels of the pile near the top are
utilized to obtain the amplitudes of downward
traveling and upward traveling stress-waves at the
pile head (for example, Hayashi et al. 1994) from the
two-point strain measurement (Lundberg 1984;
Matsuinoto et al. 1992).
4.2 Interprettrtiori nzet1iod.s qf. tJynmTic load test
J ignnls
The measured dynamic signals are interpreted using
the CASE method (Goble et al. 1975) and/or the
wave matching analysis. It is seldom that the CASE
method alone is used to estimate the bearing capacity
of a pile, since the ./,-factor used in the CASE
method, which allows for the dynamic resistance. is
regarded to be too empirical, and the set-up
phenomenon is not considered at all if the pile
driving test is conducted during initial driving only.
The result from the CASE inethod is usually
calibrated with the result of the static load test of the
pile, and it is used for driving control of the other
piles at the site of the tested pile.
The wave matching analysis of the re-driving test
is regarded as more reliable and informative than the
CASE method, because the load-displacement curve
as well as the bearing capacity are estimated through
the wave matching analysis and the influence of the
set-up phenomenon is thought to be reflected in the
estimated load-displacement curve.
Foundation engineers in Japan often use
commercial computer programs such as CAPWAPC
and TNOWAVE in which the soil model proposed
by Smith (1960), shown in Figure 3, is usually
employed, although extensions of the Smith model
are available in the programs. Criticisms of the
commercial programs remain, which are summarized
as fbllows:
1) the soil model proposed by Smith is too
empirical,
2) the influence of soil inside an open-ended pipe
pile on wave propagation in the pile is not
accounted for, and
3) the difference in the drain condition of the soil
between pile driving and static loading is not
taken into account.

KWAVE, developed in Japan (Matsumoto &

Takei 199l), is a computer program which takes into
account the wave propagation in the soil inside thc
pipe pile and the influence of the internal shaft
resistance on the wave propagation in the pile. The
pile-soil system adopted in KWAVE is shown in
Figure 4. This pile-soil modeling follows Randolph
(1 987). In the KWAVE program, the improved soil
models for the shaft and base of the pile proposed by
Iiaiidolph & Simons (1986) and Deeks (1992)
respectively have been incorporated. Figure 5 shows
the shaft model and the base model respectively.
One of the rational soil models is that, from the
soil test data such as the shear modulus, Poisson's
ratio and the density, we can determine the values for
the elastic spring and the radiation damping constant
of the shaft model, as well as the elastic spring, the
dashpot and the soil mass of the base model.
A change in drainage condition between pilc
driving and static loading may be approximatelj
taken into account by using different Poisson's ratiox
for pile driving and static loading (Matsunioto et al.
1997, Hayashi et al. 2000).

4.3 U,c of cIj~nai?iic loud test i ~ s i / I t . s for- tlic

~icc'eplunce~ i n dcles ign of piles
Fujita & Kusakabe (1988) have pointed out the
following four unsolved problems in estimating the
static bearing capacity from the dynamic load test
signals:
1) the effect of the amount of elapsed time between
the end of initial pile driving and static load test
on the difference between the bearing capacities
estimated from the dynamic load test and the
static load test,
3) the effect of the method of the static load test on
the results of the static load test,

339
3) the effect of the method used for determining the
bearing capacity from the measured load-
displacement curve on the determined bearing
capacity, and
4) the effect of the behavior of pore water pressure
during and after pile driving on the load dis-
placement curve.
The problems 1) and 4) have been vital issues in the
use of the dyiiamic load test in Japan, because 63%
of the dynamic load tests on steel pipe piles in Japan
have been conducted offshore (Wakiya et al. 2000).
Research concerning the problems 1) and 4) are
discussed below. Figure 6. Increase in static bearing capacity of a steel pipe pile
It is widely thought that the set-up phenomenon is driven in a clayey ground (JASPP 1995).
largely related to the generation and dissipation of
excess pore pressures around the pile during and reduced, compared to the case in which a low
after the pile driving, which result in the changes in performance hammer is used. That is to say, the
the effective stresses in the surrounding soils and niagnitudes of accumulated excess pore pressures
acting on the pile shaft. In the authors' knowledge. generated at the end of initial pile driving depend on
the first theoretical attempts to estimate excess pore the performance of the hammer used, resulting in
pressures around a driven pile were made by Nishida different set-up ratios for the same pile driven in the
(1963) and by Ladanyi (1963). In Japan, field same ground with different hammers.
measurements of pore pressures around driven piles There are several papers which report that the
wcre done by Koizumi et al. (1967), Wakiya et al. bearing capacity of a pile derived froin the dynamic
( 1 992) and Matsumoto et al. (1 995) for clay soils and load test for a pile is comparable with the bearing
by JASPP in 1993 for a sandy ground (Shibata et al. capacity obtained from the static load test, if the
2000). These works clearly showed that the set-up dynamic load test and the static load test are
conducted a sufficient time interval after the end of
plienomenon is largely controlled by the generation
initial pile driving.
and dissipation of excess pore pressures around the
pile during and after the pile driving.
Figure 6 shows the increase in the static pile
capacity with elapsed time after the end of initial pile
driving (JASPP 1995). An open-ended steel pipc
pile. 35 in in length, 610 niin in outer diameter and
12 mm in wall thickness, was driven with a diesel
hammer in a clayey ground. The static pile capacity,
R,, was derived from the wave matching analysis
(CAPWAPC analysis) of the dynamic load test
signals. The set-up ratio, defined as the ratio of the
bearing capacity at re-driving test to the bearing
capacity at the end of initial driving, attained a value
of 5 about 10 days after the end of initial driving, and
the set-up ratio leveled off after that time. Hayashi et
al. (2000) also showed a similar result for steel pipe
piles driven in a mudstone ground.
It inay be still difficult to predict the increase in
the pile capacity with elapsed time after the end of
initial pile driving, although some theoretical
solutions have been proposed for the set-up
phenomenon (for example, Randolph & Wroth
1979). One reason for this may well be that haininer
performance is also a factor for the set-up ratio, in
Figure 7. Comparison of end and shaft capacities obtained
addition to the pile configuration and the soil from static load test with those estimated from dynamic Load
conditions. If a high performance haininer is used. testing and derived from various pile design codes for steel pipe
the time required for the pile driving process is pile driven in mudstone ground (Michi et al. 1996).

340
 Figure 7 shows a comparison of toe and shaft
capacities, Qrl and Q.$,obtained from the static load
test, tlie dynamic load test and various pile design
codes, for a steel pipe pile ( l l n i in length, 0.8m in
diameter, 12mm in wall thickness) driven in a
mudstone ground (Michi et al. 1996). The dynamic
load test and tlie static load test were conducted 65
hours and 29 days after tlie end of initial pile driving,
respectively.
Figure 8 shows similar comparison for a steel pipe
pile driven in a sandy ground (Shibata et al. 2000). It
can be seen from Figures 7 and 8 that the bearing
capacity and the Qp/Q5 ratio derived from the
dynamic load test are the most accurate, compared to
the empirical pile design formulas available in Japan
and other countries.
There are cases where conducting re-driving tests
is difficult due to practical considerations. In such
cases, calibration of the results of tlie initial dynamic
load test on a pile with the static load test at the site
is required to roughly estimate from the initial
dynamic load tests the bearing capacity of tlie other
piles after the completion of the set-up phenomena.
JHA: Japan Road Associatio~l JAC: Architectural Institute of Japaii
JR: Japan Railway JPH: Japan Port and I-larbor Association
Hiley ( I ) : Hiley's formula with use of potential hammer energy
Hiley (2): Hiley's formula with use of transfered hammer energy
Figure 8. Comparison of end and shaft capacities obtained
from static load test with those estimated from dynamic Load
testing and derived from various pile design codes for steel pipe
pile driven in sandy ground (Shibata et al. 2000).
5 STATNAMIC LOAD TESTING IN JAPAN
Here, we will discuss the interpretation of the
Statnaiiiic test signals to estimate the static load-
displacement curve of a pile
The Statnamic tests which have been conducted in
Japan from 1992 are summarized by Kato et al.
(1998) and Nishimura et al. (2000). Kato et al.
( 1 998) compared tlie pile head force-displacement
curves. FA/,,vs M', measured in the Statnamic tests on
34 1
various types of piles in various soil conditions with
various construction methods with the load-
displacement curves, P vs w, of the piles obtained
from the static load tests. Let the discrepancy
between Fsl,zvs w and P vs w be called the 'dynamic
effect' in the Statnaniic test, here. Based oil these
comparisons, they showed that the dynamic effect
depends on 1) the acceleration of tlie pile, 2) the
penetration rate of the pile, 3) the level of FIo2 against
the static yield load, PJ,, 4) the proportion of shaft
and toe resistances, 5 ) the stiffness of tlie pile shaft.
6) the soil type, and 7) the drainage condition. They
surnniarized the 'dynamic effect' in tlie Statnainic test
qualitatively as follows:
(1) the dynamic effect is small when F,,,, is below P,,,
(3)the dynamic effect is larger for the toe resistance
than for the shaft resistance,
(3) the dynamic effect is larger when the ground is
saturated, especially for the toe resistance, and
(4) the dynamic effect becomes larger as tlie pile
penetration rate increases.
And, they concluded that interpretation of tlie
Statnaniic test signals is required to correctly
estimate the static load-displacement curve from the
Statnanii c test .
Kato et al. (1 999) studied various interpretation
methods of the Statnamic test signals, including the
Unloading Point method, one-dimensional stress-
wave propagation analysis (wave matching analysis),
FEM, and systematized their applicability in practice.
The interpretation method mostly used in Japan is
the Unloading Point method (Kusakabe &
Matsumoto 1995), based on the single mass
rnodeling of. the pile proposed by Middendoi-p et al.
(1992). A weak point of this method is that any
change in the drainage conditions between tlie
Statnamic and static loadings is not taken into
account. Therefore, the static load-displacement
curve derived from the Unloading Point method is
thought to closely correspond with the load-
displacement curve obtained from the quick load
maintained load test. FEM analysis is regarded as a
promising interpretation method for the Statnaniic
test, because elastic-(visco)plastic soil behavior and
complicated boundary conditions are incorporated in
the analysis.
Although there still is room to improve tlie
existing interpretation methods of the Statnamic test,
the ineasurements of additional parameters such as
axial forces down the pile shaft and the movement
(or acceleration) of the pile toe are also useful to
make the interpretation of the Statnamic test more
re1i ab le.
6 CONCLUDING REMARKS
It seeins that the Statnamic test is increasingly being
viewed as an alternative to the conventional static
load test in Japan. It is not uncommon that the
dynamic load test results are calibrated with the
Statnamic test results. However, a large part of the
Statnamic tests as well as the dynamic load tests are
used only for the acceptance of the piles. It is
undeniable that pile designers in Japan stick to pile
design codes excessively.
The fact that standards for the dynamic load test
and the Statnamic test (rapid load test) have not been
available in Japan also is an obstacle to wide uses of
the new pile load test methods. Another aspect to be
considered is pile design codes. There is no
reduction in the safety factor in response to an
increase in the number of load tests beyond 1.
Furthermore, pile design is performed based on the
bearing capacity alone, since vertical deformation of
pile foundations is not specified explicitly in the
design codes. However, this situation is currently
improving. Actually, the JGS standards for vertical
pile load tests are to be revised this year to include
standards for rapid load testing and dynamic load
testing. Furthermore, the Research Committee on
Present and Future of Japanese Foundation Design
and Soil Investigation in view of International
Equivalency was formed in JGS in 1997, aiming to
change foundation design codes from the allowable
stress design to the framework of the limit states
design and/or the performance based design.
Quick, cheap and accurate pile load tests such as
the dynamic load test and the Statnamic test are
expected to play important roles in the trend of'
changing foundation codes in which accurate
estimation of the load-displacement curve as well as
quality control of pile foundations are vital issues.
REFERENCES
Chosokabe, M., Yamashita, K., Kakurai, M., Fukuhara, T. and
Yamada, T., 1993. A Statiiainic loading test applied for a
cast-in-situ concrete pile. Proc. Annual Meeting of'
Architectural Institute of ,Japan: 176 1- 1762 (in Japanese).
Decks, A.J. & Randolph, M.F., 1995. A simple model for
inelastic footing response to transient loading. /tit. </our.
iVzon Anal. Methods in Geoniech., Vol. 19: 307-329.
Fellenius, B.H., 1988. Variation of CAPWAP results as a
fiinctioii of the operator. Proc. 3rd lnt. Conj.. on the
Application of Stress- W a v e Theory to Piles, Ottawa,
Canada: 814-825.
Fujita, K. & Kusakabe, O., 1988. On the evaluation of static
bearing capacity. Proc. 3rd Int. Conf on the Application of
Stress- W a v e Theory to Piles, Ottawa, Canada: 525-534.
Goble, G.G., Likins, G.E., JI-. and Rausche, F., 1975. Bearing
capacity of piles from dynamic measurements. Dept. of
Civil Eng., Case Western Reserve University.
Hayashi, M., Okamoto, T., Kanai, F. & Matsuinoto, T., 1994.
Predictions of load-settlement relations of steel pipe piles
froin dynamic load testing. Proc. 5th /tit. Cot$ u17d
Exhibition on Piling ar7d Deep Foundations, Bruges: 3.9. I -
3.9.7.
Hayashi, M., Matsurnoto, T. & Suzuki, M., 2000. Dynamic
load testing on 102 steel pipe piles for bridge foundations
on mudstone. Proc. 5th Int. Conf: on tl7e Application ofthe
Stress- W a v e Theory to Piles, Sao Paulo (to be published).
Holeyman, A.E., 1992. Technology of pile dynamic testing.
342
Proc. of 3rd Int. Cot$ of Application ofStress- W a v e Theory
to Piles, The Hague, Netherlands: 195-2 15.
JASPP, 1995. Methods for prediction of bearing capacity of
steel pipe piles by dynamic load testing. Report of
Committee on Methods for Prediction of Bearing Capacity
(in Japanese).
Kato, K., Kusakabe, 0. & Matsumoto, T., 1998.
Characteristics of rapid load test appeared in various piles.
Jozw. qf Geotech. Eng., JSCE, No.589/111-42: 155-166.
Kato, K., Horikoshi, K., Matsumoto, T. & Kusakabe, O., 1999.
Interpretation methods of Statnainic load test results. Jour.
oj'Geotech. Eng., JSCE, No.624iIII-47: 267-282.
Karkee, M., Horiguchi, T. & Kishida, H., 1997. Static and
dynamic tests for evaluation of the vertical load bearing
capacity of piles. Proc. Deep Foundufions Institute 22titl
Annual Member 's Coi$erence, Toronto, Canada: 199-2 14.
Kusakabe, 0. & Matsumoto, T., 1995. Statnamic tests of
Shonan test program with review of signal interpretation,
Proc. First Int. Statnaniic Seminar, Vancouver: 1 13- 122.
Kusakabe, O., 1998. Changing foundation design code and the
role of Statnainic test, Proc. 2nd Int. Statnainic Seminar,
Tokyo, Japan. Balkeina (to be published in 2000).
Koizuini, Y. and Ito, K., 1967. Field tests with regard to pile
driving and bearing capacity of piled foundations. Soils and
Founckrtion, JSSMFE, V01.7, No.3: 30-53.
Ladanyi, B., 1963. Expansion of a cavity in a saturated clay
medium. ./our. Soil Mech. Found. Div., ASCE, vol.89,
No.4: 127- 16 1.
Lundberg, B., 1984. On the separation of overlapping elastic
waves from measured data, Proc. 2nd Int. Conj. on
Application qf Sfres.s-M/uve Theory to Piles, Stockholm:
44 1-452.
Matsuinoto, T. & Takei, M., 1991. Effects of soil plug on
behaviour of driven pipe piles, Soils and Foundutions,
V01.3 I , N0.2114-34.
Matsuinoto, T., Sekiguchi, H., Yoshida, H., & Kita, K., 1992.
Significance of Two-Point Strain Measurements in SPT.
Soils and Foundations, Vo1.32, No.2: 67-82.
Matsumoto, T., Tsuzuki, M. & Michi, Y . , 1994. Comparative
study of static loading test and Statnaniic on a steel pipe pile
driven in a soft rock. Proc. Int. Con( and Exhibition 017
Piling mid Deep Foundations, Bruge, Belgium: 5.3.1-5.3.7.
Matsuinoto, T., Michi, Y. & Hirano, T., 1995. Perforinance of
axially loaded steel pipe piles driven in soft rock. ./our. of'
Geofech. Etig., ASCE, Vol. 12 1, No.4: 305-3 15.
Matsumoto, T., Michi, Y. & Hayashi, M. (1997): Reliability of
dynamic load testing on steel pipe piles in soft rock, Proc.
14th ICSMFE, Hamburg, Vo1.2, pp. 1 185-1 188.
Michi, Y., Matsumoto, T. & Futatsuka, Y., 1996. Reliability of
dynamic load testing compared with soil parameters: A case
study on foundation piles of Noetsu bridge. Proc. 5th hit.
Coiif'. 011 Application of Stress-Wave Theory to Piles,
Orlando, USA: 465 - 479.
Middendorp, P., Bermingham, P. & Kuiper, B., 1992.
Statnamic loading testing of foundation piles. Proc. Of'dt/?
bit. Conf of Applicufion of Stress- W a v e Theory to Piles,
The Hague, Netherlands: 58 1-588.
Middendorp, P. & Bielefeld, M.W., 1995. Statnaiiiic load
testing and the influence of stress wave phenomena. Proc.
of 1st Int. Statnariiic Seminar, Vancouver, Canada: 207-220.
Nishida, Y., 1963. A basic calculation on the failure zone and
the initial pore pressure around a driven pile in clay. Proc.
2nd Asian Conf: Soil Mech. Found. Engrg., Vol. I ,
JSSMFE, Tokyo: 2 17-2 19.
Nishiniura, S., Matsumoto, T., Kusakabe, O., Nishiumi, T.,
Yoshizawa, Y., 2000. Case studies of Statnamic load
testing in Japan. Proc. 5th Int. Cot$ on the Applicu/ioti q/'
the S/i.e.s.s- Wave Theoiy to Piles (to be published).
Randolph, M.F., 1987. Modeliiig of the soil plug response
during pile driving. Proc. 8th S. E. Asian Ceotechiiicul
Cot$, Bangkok, Vo1.2: 6.1-6.14.
Randolph, M.F. & Simons, H.A., 1986. An improved soil
model for one-di~nensionalpile driving analysis. Proc. 3rd
Iiit. Con$ on Nunz. Meth. in Ojfihore Piling Nantes: I - 17.
Randolph, M.F. and Deeks, A.J., 1992. Dynamic and static soil
models for axial pile response. Proc. of3rd Int. Cot$ 017
Application qfStress- W u v e Theory to Piles,Hague: 3- 14.
Research Committee on Rapid Load Test Methods (formed in
Japanese Geotechnical Society), 1998. Research activities
toward the standardization of rapid pile load test in Japan.
Statnamic loading test '98, Proc. 2nd Int. Statnamic
Seminar, Tokyo, Japan. Balkema (to published).
Shibata, A., Kawabata, N., Wakiya, Y., Yoshizawa, Y.,
Hayashi, M. & Matsumoto, T., 2000. Comparative study of
static, dynamic and Statnamic load tests of steel pipe piles
driven in sand. Proc. 5th Int. Conz on the Application of
the Stress- Wave Theory to Piles (to be published).
Wakiya, T., Hashimoto, O., Fukuwaka, M., Oki, T., Shinomiya,
H. & Ozeki, F., 1992. Ability of dynamic testing and
evaluation of bearing capacity recovery from excess pore
poressure measured in the field. Proc. of 4th Int. Conj of
Application of Stress-Wave Theory to Piles, The Hague,
Netherlands: 665 - 670.
Wakiya, Y., K.Nishiumi, M.Hayashi, A.Shibata & S.Nishimura
2000. The case studies of dynamic load test in Japan.
Proc. 5th Int. Conf on the Application of the Stress-Wuve
Theoty to Piles, Sao Paulo, Brazil (to be published).

343
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Ethics and money. Are they compatible?

H-Goldernberg& J. J.Goldemberg
Geotecnica Cientec, Buenos Aires, Argentina

ABSTRACT: From the point of view of a testing house ... Is the final objective of the construction industry
the site itself or QA? New technologies appear and are offered to the market as the salvation- where in some
cases they try to substitute the engineer’s eyes and knowledge- to assist the designer to verify if prediction
matches performance. But is he exchanging money for a useh1 service or for a tick in the QA box? More
piles are been tested each year all over the world. But do we know more about them or are we just collecting
tons of data? Testing houses are getting more popular while their specialists write papers in every available
publication. Are these engineers enough educated to interpret results and diagnose? What is the role of ethics,
when the consultant is under pressure from every part involved and he is being paid by one of them?

A modern tendency is to supervise and assign a
qualification to the evolution of the building site
through the Quality Control, better known as QC.
But what does it actually mean? Is it, maybe, the fi-
nal objective of the site itself or the evolution of the
quality of works done in order to be checked against
the projected idea?
Nowadays, when the fiemy of the regional
economies requires low costs and time limits practi-
cally impossible to comply with, the QC concept
generates a dichotomy with the demands of the mar-
ket. Where this market, thanks to advertising and
marketing, offers their products and services en-
wrapped by the I S 0 9000 halo; conditioning the
building industry (producer of the container for these
products or services) to join up to this trend.
It is not the authors intention to disregard the ne-
cessity of quality in sites, but it is to call the atten-
tion saying that the QC, as well as the Quality As-
surance (QA), requires indispensables items such as:
an adequate project, organisation, methodology, apt-
ness, responsibility and honesty by those involved in
it. At first sight it seems that these ingredients have
opposite directions to low costs and short duration.
And us, in charged of the QA of deep foundations
are not unaware of this phenomenon.
In the cold world these words describe, the mer-
cantilist conception of life prunes a new branch of
what is human each morning and of our engineering
judgement.
In the last decades we have witnessed the evolu-
tion of science, technology, and knowledge.. .and
the ethics to the world of the ‘ant ethics’, where eve-
rything is the same and nothing is better.
This unsolved equation, where the independent
variables such us time, price and quality have been
taken, gradually, to the distortion of the QA concept
together with the proper behaviour of the Engineers,
to obtain as a result a form where the QA box is
ticked. That is to say interchanging money for a
simple signed paper instead of a useful service.
Notable contradiction since more piles are been
tested each year all over the world. New technolo-
gies are being developed while the old ones are be-
ing updated. Various are offered to us as the pana-
cea, the magic solution, the unique answer we were
waiting for. The ‘technology’ that does without the
engineering knowledge, feeling and judgement. This
creates the proper conditions to take us far away
from the purpose we were hired, which is to measure
and inform about the state of the foundations. But
even worse, it misrepresent the Engineer ideal and it
highlights how degraded the human ethics is...as we
all know that there are no magic solutions in our
professional scope.
But do we really know better our foundations or are
we just collecting data? Are we contributing with
vital information that will allow the Engineer clarify
the real behaviour of the pile-soil interaction? Can
we show the relation between prediction and per-
formance?
It is evident that the increasing tendencies all over
the world to start leaving aside the prediction- as the
last verdict summon- for the real measurement of the

345
behaviour, brought the increase in the testing houses
with it. But does this greater demand to know the re-
ality grew in the same way as the knowledge of
those responsible summiting the verdict? Because
technology did and what is most, it even superseded
the expectations.
Nowadays it is possible to find all kinds of
equipment to perform even the most unthinkable
measurements; any colour, size and origin. The
evolution of science and technology has invaded our
lives in such a way, that it is difficult to imagine our
daily routines with out them. And Engineering can-
not run away from this trend neither can our spe-
cialty .
We blindly believe that technology will solve
everything. It seems all sense of proportion has been
lost, for, even if technology can help with interesting
tools to the diagnosis, the main engine is still the en-
gineering knowledge based on experience. Even
though Engineering will hardly survive without
technology, it is dangerous to think that only it will
save the future of our profession.
If we favour the electronic connections and chips
over the brain connections, we are doomed to failure
as professionals and as homo sapiens. We are des-
tined to become animals capable of performing basic
and repetitive operations (the donkey can blow a
flute, but that does not make it a musician)
Some Engineers show a reverential posture as re-
gards machines which they judge mistakenly, very
difficult to use. They think that their employees,
who use them with odd familiarity, show as much
intellectual capacity as the researchers that devel-
oped the Stress Wave Theory. The fact of being able
to sit in fiont of a keyboard turns the employee into
a genius to his boss’ eyes, in a gifted being if the one
supervising is the company’s chairman.. .if continu-
ing to increase generation gap. In general terms, it is
believed that a technically complex environment en-
dows a person of a great intellectual capacity; it is as
stating that whoever sits in fiont of a TV. set will be
endowed with knowledge in electronics.
It seems that relying on the mere presence of
electronic means turns out to be enough stimulation
for the intellectual capacity of the engineers. The
simple fact of selecting and observing a screen is a
week substitute of the real mental activity. The users
end up learning how to press buttons and getting re-
wards, as laboratory animals, instead of a real com-
prehension of the cause-effect relation. If we keep
on going on this dangerous path, one day we will
find an exquisite brochure showing an equipment
which will seduce us with its ergonometric design
and low price; and it will have two little lights: the
green will indicate an undamaged pile while the red
one damaged!
It is necessary to admit that the Engineer’s work
is different fiom the repetitive tasks the collar em-
ployees or the industrial workers do, which has been
346
mechanized in order to increase their production. In
our case, not only, each event in the site can be
physically different, but we also have to analyse as
separate entities the boundary conditions that it is
subjected.
But should the manufacturers take all the blame? Of
course not. A good part of it resides in the transfer of
the knowledge and not of the technology. The pur-
pose of the ones involved in higher education should
be: ‘teach carpentry and not hammer’. We need to
teach the ‘why’ and not the ‘how’ of this speciality.
Tools come and go. Teaching engineers giving pri-
ority to tools limits their knowledge towards them,
and therefore their future.
Anybody who knows how to read and write and
has developed a reasonable abstraction capacity and
logical reasoning can learn a computer programme
in any moment of his life. That is why, it is impor-
tant that the education should exploit people’s ca-
pacities and skills, the ones which he would be able
to comply with the challenges and knowledge that
are constantly presented in real life.
The focusing on technology- in the computers,
for example- will lead us to have obsolete graduates.
The question is not, teaching instrumental prob-
lems of a technology that develops speedily, but
teaching the Engineers how to think. Giving them
the ability of thought built up on a concrete knowl-
edge basis, that will allow them to react intelligently
before the unceasing changes, even the technological
ones, that will appear inevitably in the decades to
come.
The role of the educational institutions is not that
of producing tailor made first class workers, but hu-
man beings capable of thinking and being creative,
that is what the most lucid businessmen are asking
for and that the politicians, blinded by the ‘high
technology’, do not have in mind.
There are no doubts that everybody needs to learn
how to use a computer and to be able to manage the
informatics field with certain ability, which it is get-
ting easier than imagined every day. We should al-
ways remember that computers are tools and not the
aim itself. They are not capable of replacing the in-
tellectual skills and knowledge the Engineer should
learn in order to be able to incorporate him to the
world economy.
In some cases it is being taught as if the study of
deep foundations could be understood through the
usage of some recipes or simple operations in order
to manipulate the equipment, obtaining results on
the LCD which the Engineer prefers not to argue be-
cause: “I got it fiom the equipment”. Having as a re-
sult prefabricated answers that don’t belong to him.
What happened with the power of knowing true
fiom false? And the answer is in the relationship
teacher-student and in its content.. .in the desire of
teaching and learning.
 What is the use of learning? The question is not
new. Almost 2300 years ago, a young geometry stu-
dent asked Euclid: ‘What will I get learning these
things’. The master called his slave and said: ‘Give
him some coins. It seems he needs to get something
in order to learn’.
Together with the world‘s educational problem, in
which our speciality is not exonerated, it is pro-
claimed that no previous knowledge is required
therefore it is being taught that knowledge is not hi-
erarchical, that is to say, you can start constructing a
building at any level, without foundations. By elimi-
nating the existence of a certain sequence and conti-
nuity in education, we are stating that sequence and
continuity have nothing to do with the thought itself.
We find ourselves those who state that in order to
interpret foundation dynamics no deep knowledge in
geotechnics is needed. In other words, it is the same
to test a pile which is imbedded in the soil that an
isolated column!
Euclid’s student is now the one sitting at the
schools desks!
The obsession of earning one’s living is the easi-
est way to waist it and it is not worthwhile. The pre-
sent educational system searches for the utility and
presumes the death of knowledge. Some of today’s
teaching methods are usage instructions to create ro-
bots operated by ignorants with ideological lumps.
The real knowledge is not on the Internet. but in the
brain.
a deep and long lasting satisfaction that may come
months or years later.
The students do not know that they do not know.
This ‘happy unconsciousness’ shows that the sys-
tem has not even been able to show sobre the an-
tagonism between true and false, knowledge and the
lack of it. This lack of information will generate, in
part of the new generations, a failure experience due
to the contradiction between the high expectations
and the insufficient knowledge. It can have an un-
thinkable impact on society as well, because of the
presence of a group of Engineers whose demands
will be incoherent compared with their skills.
All the community involved in the pile area must
keep an eye on this week balance between Manu-
facturer-Educator-User( equipment-kno wledge-use).
Non of these parts can afford the cost of becoming
the week link of this chain, otherwise it will break.
Each of us has the obligation of being better every
day and the noble mission of helping the other two
links meeting their commitment of being ethic and
having a proper behaviour. If this simple statement
can be performed, our job will last becoming a mar-
ket necessity. And as in every modern market, trans-
actions involves money, it is demonstrated the ax-
iom that ethics and money are compatible.

Large budgets are assigned to be invested on hard-

ware and the updating of software. Little or no at-
tention is paid to humanware, that is to say the users.

And this is the last member of the trilogy Manufac-

turer-Educator-User we will make reference. The
user, before becoming one, must be a student. In
other words, he must be eager to learn and willing to
be taught (The authors do not advice to follow
Euclid’s student’s example) and as a Chinese proverb
says: “The masters open the door, you must enter by
yourself ’.
‘Some months ago I thought that there was a
certain culture which was not given to us through
spontaneous experiences, as it is not given as a gift
the acquisition of a language or being able to play an
instrument. Learning is hard and requires discipline,
training and effort”, Luis Landero (Spanish writer)
Learning is a job itself, a complex task which
demands sacrifices. Learning assumes that each per-
son will take up an individual effort in order to
modify himself After more than 14 years, in the
business, we can only be sure in believing that
learning requires work, discipline, responsibility,
and commitment fiom the student as well as fiom
the teacher. There are no shortcuts for a quality edu-
cation and the reward is not a shallow exaltation, but

347
This Page Intentionally Left Blank
h strain dynamic testing of driven and cast in situ piles
narnic testing of large piles
This Page Intentionally Left Blank
 ApplicationofStress-Wave Theory to piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 1503

Keynote lecture: Pile acceptance criteria for large diameter and cast in situ
.-
piles
Robert E Stevens
Fugro-McClellund Marine Geosciences Incorporated, Houston, Tex., USA

ABSTRACT: Case histories are presented to illustrate the use of our pile acceptance criteria for large
diameter piles driven in very dense sand, hard clay, and rock. A fourth case history illustrates the use of
combined CAPWAP analyses to interpret the results of redrive tests in clay.

1. INTRODUCTION cost effective way to determine the static pile

It is essential that pile and hammer instrumenta-
tion be used to evaluate pile acceptance for large
diameter driven piles, especially should refusal
occur or if easier driving than predicted is
encountered. Pile instrumentation can also be
used to evaluate acceptance of cast in situ piles.
Pile acceptance at refusal is based on:
the static pile capacity computations per-
formed using soil parameters obtained fiom a
high quality geotechnical investigation,
the soil resistance to driving obtained from a
CAPWAP analysis,
the soil resistance to driving obtained from
the Case-Goble formulation, and
the soil resistance to driving determined from
wave equation results obtained using the
measured driving system performance data
and the field blow count.
ESTIMATE DESIGN PENETRATIONS
Design penetrations are estimated from the lower
bound static pile capacity curve computed for soil
conditions determined from a high quality site
specific geotechnical investigation. Due to limita-
tions in the pile capacity procedures, length
effects, and soil variability, the computed capacity
may be in error. Piles driven to shallow penetra-
tions in the Safaniya Field of the Arabian Gulf, for
example, had soil resistances to driving that were
as much as 10 times greater than the computed
static pile capacity. When discrepancies of this
magnitude are observed, pile load tests may be a
capacity or develop new methods to compute pile
capacity. A total of 26 pullout tests were
performed at eight offshore sites on 20- to 30-in.-
diameter open-ended pipe piles driven to
penetrations of 36 to 72 ft. Tensile capacities
were 2.5 to 6 times greater than the static pile
capacity computed using the API method. Load
tests were then performed on two 59-ft long,
24-in.-diameter open-ended pipe piles driven in
the very dense sands at Ras Tanajib. A new
procedure was developed to compute pile capacity
in very dense sands. Helfrich, Wiltsie, Cox, and
Al-Shafei (1985) and Al-Shafei, Cox, and Helfrich
(1994) describe these load tests in more detail.
The ultimate axial capacity is computed using
the static method of analysis. The lower bound
capacity is computed using the API RP 2A-WSD
(1993) method, and is usually that presented in the
soil boring report. The upper bound capacity may
be computed using the Ras Tanajib method for
very dense sands, an upper bound shear strength
profile for hard clays, and rock end bearing except
for weak rock of poor quality. Additional details
are given by Wiltsie, Stevens, and Vines (1984)
and Stevens and AI-Shafei (1996).
3. EVALUATE PILE DRIVABILITY
We recommend evaluating pile drivability using
procedures recommended by Stevens, Wiltsie, and
Turton (1982). Driving system parameters are
selected from our data base established from over
300 platform installations worldwide. The data
base includes hainmer efficiency for 40 different

35 1
pile driving hammers, cushion stiffness and
cushion coefficient of restitution for 18 different
cushion materials, and system efficiency for 100
different hammer-cushion configurations. Wave
equation analyses are performed for a range in
driving system performance parameters
determined from the mean and standard deviation
of values in our data base. Minimum blow counts
are computed for the lower bound coring case. If
our evaluation of pile drivability indicates
premature pile refusal or excessive driving
stresses, the hammer and/or wall thickness that
permits the piles to be driven to design penetration
is determined.
4. MONITOR PILE DRIVING
Our pile monitoring system records force and
velocity signals using strain transducers and
accelerometers attached on opposite sides of the
pile near the top. The impact stress and velocity,
the energy transmitted to the pile (or system
efficiency), and the soil resistance to driving are
computed for every hammer blow. We estimate
the soil resistance to driving using the Case-Goble
formulation assuming a damping coefficient that is
a function of soil type. The impedance or TNO
methods. as discussed by van Foeken, Daniels,
and Middendorp (1996), can also be used.
Estimates of the hammer efficiency and the
cushion stiffness and cushion coefficient of
restitution are obtained for selected hammer blows
using force-time characteristics obtained in a pre-
installation parametric study. We recommend
using controlled hard driving, as proposed by
Stevens, Wiltsie, and Middlebrooks (1984)
whenever the soil boring indicates the presence of
rock layers or cemented zones. Driving is con-
tinued if the end bearing is decreasing or erratic,
and the driving stresses at or just above the driving
shoe are acceptable. Driving is terminated if the
driving stresses are excessive, or if refusal is
reached and the end bearing is constant or
increasing with depth.
Our hammer monitoring system measures the
velocity of the ram just prior to impact using Hall
sensors and a small magnet attached to the ram.
Hammer efficiency is determined for every
hammer blow from the ram impact velocity. We
generally recommend that only ram velocity be
monitored when easy driving is expected in order
to evaluate pile acceptance should blow counts
lower than predicted occur.
A signal matching program, such as the CAse
Pile Wave Analysis Program (CAPWAP)
352
developed by Rausche (1970), or TNOWAVE,
discussed by Courage and Bielefeld (1992), is
used to estimate soil quake and damping
parameters, and the distribution of the soil
resistance to driving along the length and at the
toe of the pile. The pile is divided into continuous
segments and calculations are made using a
traveling wave algorithm. Either the measured
pile top force or velocity is used as a boundary
condition, and the complementary quantity is
computed and compared with the measured
quantity. The set of soil parameters is varied until
a best match between measured and computed pile
top force or velocity is obtained.
5. PILE ACCEPTANCE
Piles with adequate lateral capacity rehsing above
design penetration are accepted if:
the required tensile and compressive pile
capacities are obtained from the upper bound
static pile capacity curves.
o the required tensile and compressive
capacities are confirmed by the soil resistance
to driving obtained from a CAPWAP
analysis, and
8 the required compressive capacity is con-
firmed by the soil resistance to driving
obtained from the Case-Goble formulation
and wave equation analyses performed using
the measured driving system performance
data (i.e., the hammer efficiency, cushion
stiffness, and cushion coefficient of restitu-
tion) and the field blow count.
Piles can also be accepted if the required
tensile and compressive piles capacities are
obtained from a re-evaluation of pile capacity.
The soil resistance to driving profile obtained
from the Case-Goble bearing capacity is used to
modify the elevation and thickness of soil strata.
When pipe piles are driven open-ended,
relative movement between the pile and soil
occurs both on the outside and inside of the pile
wall. Skin friction is, therefore, mobilized on both
of these surfaces. The end bearing area in this
case is equal to the cross-sectional area of steel at
the toe. When a pile plugs, the soil plug moves
with the pile during driving. This friction is
mobilized only on the outer wall, and the end
bearing area is then the gross end area of the pile.
The pile initially plugs after each delay. Because
the distribution of the shaft resistance on the inside
and outside of the pile cannot be evaluated, we
recommend waiting a few minutes and restriking
the pile.
 Pile acceptance at refusal is simplified when
the pile driving hammers selected are large
enough to overcome the long-term static pile
capacity, and the CAPWAP and Case-Goble soil
resistance to driving can be considered a lower
bound because more resistance is mobilized at
larger pile displacements.
Typically, the pile driving hammer selected
for a particular installation may be large enough to
drive the piles to design penetration, but not large
enough to overcome the long-term static capacity.
In clay, the skin fiiction during driving is
generally much smaller than that mobilized under
static loading because large excess pore pressures
are generated during continuous driving.
CAPWAP analyses can be used to estimate the
distribution of the soil resistance to driving along
the length and at the toe of a pile during
continuous driving and after a set-up period. By
combining these results, it is possible to proof test
a pile without the expense of mobilizing a larger
hammer to the site.
High strain dynamic testing can also be used
to evaluate the capacity of cast in situ piles (also
known as drilled shafts or bored piles). Hussein,
Likins, and Rausch (1996) made recommendations
concerning the hammer weight, drop height, and
cushion thickness. The hammer weight should be
at least equal to 1.5 percent of the expected
capacity. The hammer drop height should be
approximately 8.5 percent of the pile length, with
a minimum value of 2 m. The thickness of the
plywood cushion is determined from:
t = L2/2D
where: t = cushion thickness, mm;
L = pile length, m; and
D = pile diameter, m.
The minimum cushion thickness is 100 mm, and
the cushion thickness should be increased 150 mm
when the pile length exceeds 30 m.
6. CASE HISTORIES
Case histories are presented to illustrate the use of
our pile acceptance criteria for very dense sand,
hard clay, and rock. A fourth case history
illustrates the use of combined CAPWAP
analyses.
6.1 In our fwst case history, piles were driven
to design penetrations of 35 to 84 m using Menck
MRBS 2500-S, 3000, 4600, and 5000 air/steam
hammers, and a Menck MHU 1000 hydraulic
hammer. Cone penetrometer tests indicated a
predominantly dense to very dense silty sand
353
profile with clay seams and layers. Drivability
studies indicated that a Menck MHU 1000
hammer could drive the 48-in.-diameter piles to
84-m penetration if the wall thickness was
increased to 2.00 inch. Blow counts with the
Menck MHU 1000 hammer ranged fiom 25 to 40
blows per 0.25 m at 60-m penetration, fiom 35 to
65 blows per 0.25 m at 70-rn penetration, and
fiom 45 to 85 blows per 0.25 m at 80-m
penetration.
With the exception of Pile C-3, all piles were
driven to design penetration. Pile C-3 was driven
to refusal at 56.8-m penetration, 1.2 m above
design penetration. Pile C-3 was accepted because
the required compressive capacity was obtained
fiom the upper bound static pile capacity curve
computed using Ras Tanajib pile design
parameters, and the required compressive capacity
was confirmed by the CAPWAP analyses, the
Case-Goble bearing capacity, and the hindcast
resistance to driving obtained fiom wave equation
analyses. The required compressive design load
for Pile C-3 was 26.1 MN. At fmal penetration,
soil resistance to driving determined by Case-
Goble (Fig. I), CAPWAP (Fig. 2), and wave
equation analyses were 28.7, 31.8, and 33.4 MN,
respectively. As shown in Fig. 3, the Case-Goble
bearing capacity was greater than the lower bound
compressive pile capacity.
6.2 In our second case history, piles were
driven to rehsal in a hard silty clay stratum
between penetrations of 118.9 and 134.3 ft with a
Vulcan 560 hammer. Piles F-5 and G-6 were
driven without interruption to penetrations of 120
and 134 ft, respectively. Pile G-5 had delays of 22
and 40 minutes that had little effect on the driving.
Pile F-6 had a delay of 3.1 hours that almost
resulted in pile refusal, as shown in Fig. 4.
The restart blow count for Pile F-6 was 730
bpf at 99-ft penetration. At this depth, the
required compressive capacity of 4523 kips was
not obtained fkom the upper bound static pile
capacity curve. The pile was driven to 118.9-R
penetration, which was deeper than the required
design penetration of I16 ft. The pile was
accepted because the minimum blow count was
obtained. Although not a requirement for pile
acceptance, the required compressive capacity was
confrmed by the CAPWAP capacity of 5878 kips
(Fig. 5), the Case-Goble bearing capacity of 5200
kips, and the soil resistance to driving back-
calculated from wave equation analyses of 6050
kips.
6.3 In our third case history, piles were
driven to practical refusal in a strong gypsum
 Blow Hammer Cushion Coefficient Transmitted System Resistance Impact Impact Reflected Maximum
Penetration Count Efficiency Stiffness of Energy Efficiency to Driving Stress Velocity Stress Stress
m b10.25m % M Nlm Restitution kN-m % MN MPa mlsec % MPa

53.50 78 339 38 28.6 144.0 3.52 27.1 174.9

53.75 82 363 41 29.9 156.6 3.85 24.7 185.2
54.00 72 379 43 29.6 152.7 3.84 23.9 182.8
54.25 74 385 44 29.7 151.8 3.84 24.1 183.2
54.50 62 389 44 29.7 149.0 3.79 25.3 182.3
54.75 72 385 44 29.3 146.0 3.73 25.3 179.7
55.00 81 386 44 29.1 144.3 3.69 24.5 178.2
55.25 77 390 44 29.1 144.1 3.68 23.9 177.6
55.50 79 -- -- -- 381 43 28.8 141.4 3.62 24.5 175.5
55.75 81 75 I5,026 0.60 379 43 28.4 139.0 3.56 24.0 172.8
58.00 88 75 14,186 0.60 374 42 28.1 137.3 3.53 24.4 171.8
58.25 81 75 14,518 0.61 380 43 28.2 137.3 3.52 23.4 172.3
58.50 105 75 11,646 0.61 385 44 28.8 136.2 3.52 25.0 173.6
58.75 313 75 11,156 0.60 374 42 29.3 131.4 3.44 29.0 174.6
58.80 5601.05 75 10,876 0.60 374 42 28.7 129.8 3.39 31.O 172.6

Average 75 12,901 0.60 378 43 29.0 142.9 3.6 25.3 177.1

Standard Deviation (0) (1,708) (0.01) (12) (1) (0.5) (7.3) (0.1) (2.1) (4.4)

Figure 1. Log of Driving System Performance (Pile C-3, Menck MRBS 5000 Hammer)

354
 Blow No. 2787

Final CAPWAPC Capacity: Ru 7154.1, Skin 6436.7, Toe 717.4 kips

Depth Depth Suffl of Ru Unit Resistance

Soil Below Below with Respect to Smith
Segment Gages Grade Ru UP Down Depth Area Damping Quake
NO. ft n kips kips kips kipslft kipslf2 Sift inch

7154.1
I 95.8 8.1 88.6 7065.5 88.6 13.42 1.07 ,053 .i00
2 102.4 14.7 48.0 7017.5 136.6 7.26 .58 ,053 .i00
3 109.0 21.3 5.2 7012.2 141.8 .79 .06 .053 ,100
4 115.6 27.9 4.8 7007.5 146.6 .72 .06 .053 ,100
5 122.2 34.5 35.6 6971.8 182.3 5.40 .43 ,053 ,100
6 128.8 41.1 65.6 6906.2 247.8 9.93 .79 ,053 .I
00
7 135.4 47.7 85.2 6821.O 333.1 12.90 1.03 ,053 00
.I
8 142.0 54.3 88.8 6732.2 421.8 13.44 1.07 ,053 .100
9 148.6 60.9 77.5 6654.7 499.4 11.74 .93 .053 ,100
10 155.2 67.5 62.3 6592.4 561.7 9.43 .75 ,053 .I 0 0
11 161.8 74.1 47.7 6544.7 609.3 7.22 .57 ,053 ,100
12 168.4 80.7 41 .I 6503.6 650.5 6.23 .50 .053 ,100
13 175.0 87.3 63.2 6440.4 713.7 9.57 .76 .053 .100
14 181.6 93.9 127.4 6313.0 841.I 19.29 1.53 ,053 ,100
15 188.2 100.5 206.6 6106.4 1047.7 31.28 2.49 ,053 ,100
16 194.8 107.1 243.2 5863.2 1290.9 36.83 2.93 ,053 ,100
17 201.4 113.7 217.8 5645.4 1508.7 32.98 2.62 .053 .I 0 0
18 208.1 120.4 194.9 5450.5 1703.6 29.51 2.35 .053 .I 0 0
19 214.7 127.0 266.8 5183.6 1970.5 40.40 3.21 ,053 ,100
20 221.3 133.6 452.5 4731. I 2423.0 68.51 5.45 ,053 .I 0 0
21 227.9 140.2 666.0 4065.1 3089.0 100.84 8.02 ,053 ,100
22 234.5 146.8 763.6 3301.5 3852.6 115.61 9.20 .053 ,100
23 241 . I 153.4 662.0 2639.5 4514.6 100.23 7.97 ,053 ,100
24 247.7 160.0 463.8 2175.7 4978.4 70.22 5.59 .053 .100
25 254.3 166.6 362.6 1813.1 5341.O 54.90 4.37 ,053 .100
26 260.9 173.2 384.1 1429.0 5725.1 58.15 4.63 ,053 .100
27 267.5 179.8 384.3 1044.7 6109.7 58.19 4.63 .053 .I 00
28 274.1 186.4 327.2 7174 6436.7 49.55 3.94 ,053 .100

Average Skin Values 229.9 34.53 2.77 .053 ,100

Toe 717.4 57.07 .351 . I 54

Soil Model Parameters/Extensions Skin Toe

Case Damping ,663 ,488

Unloading Level (% of Ru) 0
Soil Plug Weight (kip9 2.66

Figure 2a. Results of CAPWAP Analyses (Pile C-3)

stratum between penetrations of 58.3 and 62.2 A
with a Vulcan 560 hammer. Driving was
terminated because stresses exceeded 85 percent
of yield in the Grade B steel above the driving
shoe.
The piles were accepted because the required
penetration for adequate lateral capacity was
exceeded, the required tensile capacities were
obtained from the lower bound static pile capacity
curve, the required compressive capacities were
obtained from the upper bound static pile capacity
curve computed using rock end bearing, the
required tensile and compressive capacities were
confirmed by CAPWAP analyses, and the

355
Figure 2b. Results of CAPWAP Analyses (Pile C-3)
356
Figure 3. Soil Resistance to Driving (Pile C-3)
required compressive capacities were confirmed
by the Case-Goble bearing capacities and soil
resistances to driving obtained fiom wave
equation analyses performed using measured
driving system performance data and the field
blow count.
The driving system performance data for Pile
B-2, shown on Fig. 6, indicates a Case-Goble
bearing capacity of less than 1460 kips to 28-ft
penetration, and less than 2140 kips to 60-ft
penetration. The required compressive capacity of
3390 kips was obtained at 62-fi penetration, and
was confirmed by the CAPWAP capacity of 4823
kips (Fig. 7), the Case-Goble bearing capacity of
5072 kips, and the hindcast capacity of 6050 kips.
357
6.4 Our fourth case history is for a series of
redrive tests performed with a hammer that is
clearly not large enough to mobilize the full soil
resistance. The soil resistance mobilized during a
series of redrive tests performed on a 1.6-m-
diameter steel pipe pile driven to a penetration of
26 m in a very silty clay is presented in Fig. 8.
The four lower bound soil resistance profiles are
for continuous driving. The pile was redriven by
applying only two consecutive hammer blows
with a PMJ-400 hydraulic hammer after delays of
6 minutes, 15 minutes, 33 minutes, 2 hours, and
66 hours. The resistance generally increases with
time, but the resistance after a 33-minute delay
appears to be slightly smaller than the resistance
 Blow Hammer Cushion Coefficient Transmitted System Resistance Impact Impact Reflected Maximum
Penetration Count Efficiency Stiffness of Energy Efficiency to Driving Stress Velocity Stress Stress
- ft bf % kipshnch Restitution ft-kips % ki~s ksi Wsec % ksi

99 730 78 66,700 0.70 162 52 4340 19.1 10.6 42 28.8

100 335 77 162,100 0.81 184 59 4650 22.4 12.7 62 30.0

101 270 76 204,400 0.83 186 60 461 0 22.5 12.8 36 30.1

102 260 73 209,200 0.85 183 59 4530 22.3 12.7 39 29.7
103 256 73 209,200 0.85 184 59 4500 22.4 12.8 38 29.8
104 249 73 213,900 0.87 186 60 4470 22.8 13.0 34 29.6
105 352 67 216,300 0.88 174 56 4460 22.3 12.7 38 29.3

106 236 71 320,700 0.89 188 60 4560 23.4 13.3 35 30.1

107 207 74 423,100 0.88 195 62 4590 24.1 13.6 34 30.8
108 200 80 384,900 0.79 186 60 4740 23.4 13.6 40 31.3
109 193 72 423,100 0.88 187 60 4800 24.3 13.8 39 31.4
110 182 75 406.100 0.84 186 60 4750 24.3 13.8 39 31.5

111 158 77 414,600 0.86 196 63 4580 24.6 13.9 35 31.3

112 172 77 414,600 0.86 198 63 4450 23.7 13.5 39 31 .O
113 200 78 427,300 0.89 206 66 4460 22.8 12.9 42 30.0
114 195 77 258,500 0.78 177 57 4470 22.5 13.1 45 30.2
115 188 71 384,900 0.79 167 53 4780 23.6 13.9 47 32.4

116 204 73 380,600 0.78 168 54 5070 24.3 14.1 51 33.5

117 188 79 367,900 0.75 174 56 5030 24.7 14.1 48 33.1
118 222 73 384,900 0.79 173 55 5260 25.3 14.4 48 34.0
118.9 294 74 367,900 0.75 166 53 5200 25.1 14.2 45 33.7

Mean 75 316,200 0.82 182 58 42 31 .O

Standard Deviation ( 3) (107,600) (0.05) (11) ( 4) ( 7) ( 1.5)

Figure 4. Log of Driving System Performance (Pile F-6, Vulcan 560 Hammer)

358
 Blow No. 170

Final CAPWAPC Capacity: Ru 5877.5, Skin 5761.9, Toe 115.6 kips

Depth Depth Sum of Ru Unit Resistance

Soil Below Below with Respect to Smith
Segment Gages Grade Ru UP Down Depth Area Damping Quake
NO. ft ft kips kips kips kipslft kipslf2 slft inch

5877.5
1 205.4 6.4 13.4 5864.1 13.4 2.02 .I6 .068 ,080
2 212.0 13.0 15.9 5848.1 29.3 2.40 .I9 .068 ,080
3 21 8.6 19.6 23.1 5825.0 52.5 3.49 .28 ,068 ,080
4 225.2 26.2 35.2 5789.8 87.7 5.31 .42 ,068 .080
5 231.9 32.9 60.0 5729.8 147.7 9.06 .72 ,068 ,080
6 238.5 39.5 108.6 5621.2 256.3 16.40 1.30 .068 ,080
7 245.1 46.1 172.6 5448.6 428.9 26.05 2.07 .068 ,080
8 251.7 52.7 220.2 5228.4 649.1 33.24 2.64 ,068 ,080
9 258.4 59.4 224.9 5003.5 874.0 33.95 2.70 ,068 ,080
10 265.0 66.0 21 0.6 4792.9 1084.6 31.79 2.53 ,068 ,080
11 271.6 72.6 249.9 4543.0 1334.4 37.71 3.00 ,068 ,080
12 278.2 79.2 388.6 41 54.5 1723.0 58.66 4.67 ,068 ,080
13 284.9 85.9 583.6 3570.9 2306.6 88.08 7.01 ,068 ,080
14 291.5 92.5 740.7 2830.2 3047.2 111.80 8.89 ,068 ,080
15 298.1 99.1 806.5 2023.7 3853.8 121.74 9.68 ,068 ,080
16 304.7 105.7 772.2 1251.5 4626.0. 116.56 9.27 ,068 ,080
17 31 1.4 112.4 639.3 61 2.2 5265.3 96.50 7.68 .068 ,080
18 318.0 119.0 496.6 115.6 5761.9 74.96 5.96 .068 ,080

Average Skin Values 320.1 48.42 3.84 ,068 ,080

Toe 115.6 65.46 ,247 ,065

Soil Model Parameters/Extensions Skin Toe

Case Damping 1.004 .073

Unloading Level 36
Soil Plug Weight 1.72

Figure 5a. Results of CAPWAP Analysis (Pile F-6)

mobilized after a 15-minute delay, and the
resistance after a 66-hour delay is less than the
resistance mobilized after a 2-hour delay. The
maximum resistance overcome by the pile driving
hammer is about 3300 kips (14.7 MN). The
hammer is too small to mobilize the hll soil
resistance during the redrive tests. This is shown
very clearly for the redrive test performed after the
66-hour delay. Almost no soil resistance was
mobilized over the bottom quarter of the pile.
In our combined CAP WAP anaIyses, we have
assumed that the soil resistance mobilized during
continuous driving is a lower bound, i.e., the soil
resistance mobilized on a particular pile segment
is assumed to be the larger of the actual resistance
mobilized or the resistance mobilized during
continuous driving. The soil resistance mobilized
after a 66-hour delay in the combined CAPWAP
analysis is about 4400 kips (19.6 MN), as shown
in Fig. 9. In the standard CAPWAP analysis, the
resistance mobilized was only 3305 kips (14.7
MN).
To help put this in perspective, Fig. 10 shows
that the resistance mobilized by a Vulcan 060
hammer is about 3300 kips (14.7 MN), and the
resistance mobilized by a Vulcan 560 hammer is
about 4100 kips (18.2 MN), assuming the same
soil and pile parameters. Comparing the three
hammers shown in Fig. 10, we see that by
mobilizing a hammer having 50 percent more
rated energy results in only a 10 percent increase
in the maximum soil resistance overcome, and
mobilizing a hammer having 150 percent more
rated energy results in a 37 percent increase in the
maximum resistance overcome.

359
Figure 5b. Results of CAPWAP Analyses (Pile F-6)

360
 Blow Hammer Cushion Coefficient Transmitted System Resistance Impact Impact Reflected Maximum
Penetration Count Efficiency Stiffness of Energy Efficiency to Driving Stress Velocity Stress Stress
R mf % kipslinch Restitution ft-kips % kips J
s
-i Wsec % ksi

15 1 51 114,000 1.oo 152 49 980 22.4 14.7 -42 22.4

16 44 89,700 0.96 125 40 990 20.1 12.7 -41 20.1

17 48 92,500 1.oo 145 46 1177 21.6 13.0 -39 22.7
18 52 89,100 0.95 147 47 1170 21.9 13.4 -33 22.3
19 58 84,900 0.89 7 52 49 7190 21.0 12.0 -36 22.4
20 54 89,700 0.96 154 49 1150 22.4 13.9 -37 22.6

21 50 114,000 1.oo 162 52 1170 23.3 14.1 -45 23.6

22 56 114,000 1.oo 175 56 930 24.2 15.1 -50 24.2
23 71 108,100 0.94 198 63 1235 24.5 13.5 -45 24.6
24 56 103,900 1.oo 179 57 1185 24.4 14.8 47 24.6
25 72 97.600 0.93 200 64 1267 24.4 13.8 -40 24.7

26 10 78 94,000 0.89 204 65 1370 24.8 12.8 -42 24.9

27 10 79 92,200 0.87 204 65 1372 24.5 12.8 43 24.9
28 10 84 97,300 0.83 206 66 1458 24.8 12.8 -43 24.8
29 11 82 90,400 0.85 208 67 1515 25.0 12.9 -40 25.2
30 14 78 94,900 0.90 208 67 1613 25.1 13.0 -38 25.4

31 15 78 94,900 0.90 209 67 1623 25.2 13.1 -37 25.4

32 14 79 104,200 0.90 21 1 68 1563 25.4 13.1 -39 25.5
33 13 80 103,200 0.89 21 1 68 1548 25.3 13.2 -40 25.4
34 13 77 96,700 0.92 212 68 1560 25.4 13.1 -41 25.5
35 12 78 105.200 0.91 212 68 1615 25.4 13.2 -40 25.5

36 13 79 95,800 0.91 213 68 1685 25.5 13.2 -40 25.7

37 12 77 97,600 0.93 213 68 1646 25.5 13.2 -39 25.7
38 12 75 99,400 0.95 21 1 68 1616 25.2 13.2 -36 25.6
39 14 79 95,800 0.91 214 68 1779 25.4 13.2 -34 25.5
40 12 78 87,000 0.92 213 68 1823 25.3 13.2 -30 25.5

Figure Sa. Log of Driving System Performance (Pile B-2, Vulcan 560 Hammer)

361
 Blow Hammer Cushion Coefficient Transmitted System Resistance Impact Impact Reflected Maximum
Penetration Count Efficiency Stiffness of Energy Efficiency to Driving Stress Velocity Stress Stress
ft hf % kiwiinch Restitution ft-ki ps- O/O ki ps ksi Wsec % ksi

41 17 79 78,600 0.91 213 68 1986 25.2 13.2 -26 25.3

42 14 79 79,200 0.92 215 69 1829 25.4 13.3 -28 25.5
43 16 78 87,700 0.93 214 68 1763 25.3 13.3 -30 25.6
44 16 78 87,000 0.92 212 68 1894 25.3 13.2 -28 25.4
45 17 80 78,000 0.90 213 68 2139 25.4 13.2 -24 25.6

46 20 80 85,600 0.90 212 68 2133 25.2 13.1 -25 25.4

47 20 80 77,500 0.89 21 1 68 2094 25.1 13.0 -25 25.4
48 20 80 84,900 0.89 213 68 2077 25.1 13.1 -22 25.3
49 23 79 85,600 0.90 213 68 2055 25.2 13.2 -18 25.5
50 20 79 85,600 0.90 212 68 1927 25.3 13.1 -2 1 25.6

51 18 79 78,000 0.90 212 68 1851 25.3 13.1 -27 25.6

52 15 80 78,000 0.90 214 68 1868 25.4 13.2 -26 25.6
53 15 80 86,300 0.91 216 69 1850 25.4 13.2 -26 25.6
54 17 82 76,900 0.88 217 69 1781 25.7 13.2 -26 25.9
55 14 81 104,200 0.90 216 69 1727 25.9 13.3 -30 26.0

56 15 81 104,200 0.90 216 69 1758 25.9 13.3 -3 1 26.1

57 16 82 103,200 0.89 217 69 1837 25.9 13.3 -31 26.1
58 15 82 103,200 0.89 215 69 1841 25.9 13.3 -29 26.1
59 16 80 105,200 0.91 215 69 1816 25.8 13.3 -3 1 25.9
60 17 81 86,300 0.91 217 70 201 1 25.7 13.3 -24 25.9

61 27 79 95,800 0.91 212 68 3253 25.3 13.2 4 29.4

62 200 78 61,600 0.79 183 58 5125 22.3 11.4 62 34.3
62.2 708 81 45,700 0.70 169 54 5072 21 .o 10.6 69 32.1

Mean 74 91,900 0.91 200 64 1794 24.6 13.2 -29 25.4

Standard Deviation (11) (13,000) (0.05) ( 24) ( 8) ( 796) ( 1.5) ( 0.7) ( 22) ( 2.2)

Figure 6b. Log of Driving System Performance (Pile B-2, Vulcan 560 Hammer)

362
 Blow No. 739

Final CAPWAPC Capacity: Ru 4823.2, Skin 3676.8, Toe 1146.3 kips

Depth Depth Sum of Ru Unit Resistance

Soil Below Below with Respect to Smith
Segment Gages Grade Ru UP Down Depth Area Damping Quake
0. ft ft kips kips kips kipslft kipslfa! Slft inch

4823.2
1 167.5 9.5 99.8 4723.4 99.8 15.19 I .38 .I26 ,194
2 174.0 16.0 143.9 4579.5 243.6 21.91 1.99 .I26 .I 94
3 180.6 22.6 223.9 4355.7 467.5 34.09 3.10 .I26 ,194
4 187.2 29.2 353.6 4002.7 821 .I 53.85 4.90 .I26 .I 94
5 193.7 35.7 522.3 3479.8 1343.4 79.53 7.23 ,126 .I94
6 200.3 42.3 646.6 2833.2 1990.0 98.46 8.95 ,126 .I 94
7 206.9 48.9 655.0 21 78.2 2645.0 99.74 9.07 .I26 ,183
8 21 3.4 55.4 567.0 1611.2 3212.0 86.34 7.85 .I26 ,164
9 220.0 62.0 464.8 1146.3 3676.8 70.78 6.43 .I26 ,148

Average Skin Values 408.5 59.30 5.66 .I26 .I 82

Toe 1146.3 745.82 .301 ,148

Soil Model Parameters/Extensions Skin Toe

Case Damping 1.358 1.015

Unloading Level (% of Ru) 0
Soil Plug Weight WPS) 6.47

Figure 7a. Results of CAPWAP Analyses (Pile 9-2)

7. CONCLUSIONS required tensile and compressive pile capacities

Pile instrumentation records pile top force and
velocity. from which CAPWAP and Case-Goble
soil resistances are obtained. Soil resistances are
also back-calculated f?om wave equation analyses
performed using the measured hammer efficiency,
cushion stiffness, and cushion coefficient of
restitution. Hammer instrumentation measures
ram impact velocity, from which hammer
efficiency is obtained.
At refusal, we recommend accepting piles
when the required tensile and compressive pile
capacities are obtained from the upper bound
static pile capacity curves, the required tensile and
compressive capacities are confirmed by
CAP WAP analyses, and the required compressive
capacity is confirmed by the Case-Goble bearing
capacity and the soil resistance to driving obtained
fiom wave equation analyses performed using the
measured driving system performance data and the
field blow count.
We also recommend accepting piles when the
re-evaluated pile capacity indicates that the
are obtained. The soil resistance to driving profile
obtained f?om the Case-Goble bearing capacity is
used to modify the elevation and thickness of soil
strata.
Pile acceptance can also be based on
combined CAP WAP analyses. The soil resistance
mobilized on a particular pile segment is assumed
to be the larger of the actual resistance mobilized.
or the resistance mobilized during continuous
driving.
When piles are driven to design penetration,
we recommend accepting piles as long as the
minimum blow count is obtained. Minimum blow
counts are computed for the lower bound coring
case and a range of hammer efficiencies. Only
ram velocity is required to be monitored to
confirm hammer efficiencies.

363
Figure 7b. Results of CAPWAP Analyses (Piles B-2)

364
Figure 8. Soil Resistance Determined from CAPWAP Analyses

365
 Soil Resistancel Kips
1000 2000 3000 4000 5000

Figure 9. Soil Resistance Determined from Combined CAPWAP Analyses

366
 Figure 10. Driving Resistance-Blow Count Curves

REFERENCES American Petroleum Institute (1 993),

Al-Shafei, K.A., Cox, W.R., and Helfi-ich, S.C.
(1994), “Pile Load Tests in Dense Sand: Analysis
of Static Test Results,’’ Proceedings, 26th
Offshore Technology Conference, Houston,
Vol. I , pp. 83-103.
Recommended Practice for Planning, Designing,
and Constructing Fixed Offshore Platforms, API
Ihmnmended Practice 2A-WSD (w 2A), 20th
Ed., API, Washington, D.C.

367
Courage, W.M.G. and Bielefeld, M. W. (1992),
"TNO automatic signal matching," Proceedings of
the Fourth International Conference on the
Application of Stress-Wave Theory to Piles, The
Hague, The Netherlands, September 2 1-24, 1992.
Helfrich, S.C., Wiltsie, E.A., Cox, W.R., and
AI-Shafei, K.A. (1985), "Pile Load Tests in Dense
Sand: Planning, Instrumentation, and Results,"
Proceedings, 7th Offshore Technology
Conference, Vol. 1, pp. 55-64.
Hussein, M., Likins, G., and Rausche, F. (1996),
"Selection of a Hammer for High-Strain Dynamic
Testing of Cast-in-Place Shafts," Fifth
International Conference on the Application of
Stress-Wave Theory to Piles, Orlando, Florida,
September 1 1- 13.
van Foeken, R.J., Daniels, B., and Middendorp, P.
(1 996), "An improved method for the real time
calculation of soil resistance during driving," Fifth
International Conference on the Application of
Stress-Wave Theory to Piles, Orlando, Florida,
September 1 1- 13, 1996.
Wiltsie, E.A., Stevens, R.F., and Vines, W.R.
(1984), "Pile Installation Acceptance in Strong
Soil," Proceedings, 2nd International Conference
on the Application of Stress Wave Theory on
Piles, pp. 72-78.

Rausche, F. (1970), "Soil Response fi-om Dynamic

Analysis and Measurements on Piles," Ph.D.
Dissertation, Division of Solid Mechanics,
Structures, and Mechanical Design, Case Western
Reserve University, Cleveland, Ohio, 320 p.

Somehsa, P. and Stevens, R.F. (1989), "Pile

Driving Measurements for Assessment of
Hammer Performance and Pile Adequacy,"
Symposium on Underground Excavations in Soils
and Rocks," Asian Institute of Technology.

Stevens, R.F., Wiltsie, E.A., and Turton, T.H.

(1982), "Evaluating Pile Drivability for Hard
Clay, Very Dense Sands, and Rock," Proceedings,
14th Offshore Technology Conference, Houston,
Vol. 1, pp. 465-481.

Stevens, R.F., Wiltsie, E.A., and Middlebrooks,

J.R. (1 984), "Controlled Hard Driving,"
Proceedings, 2nd International Conference on the
Application of Stress Wave Theory on Piles,
Stockholm, pp. 162-169.

Stevens, R.F. and AI-Shafei, K.A. (I 996), "The

Applicability of the Ras Tanajib Pile Capacity
Method of Long Offshore Piles," Proceedings,
28th Offshore Technology Conference, Houston,
Vol. 1, pp. 171-180.

Stevens, R.F. (ZOOO), "Pile Acceptance Based on

Combined CAP WAP Analyses," Sixth
International Conference on the Application of
Stress-Wave Theory to Piles, Sao Paulo.
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Analyzing the bearing capacity mechanism of large diameter diving casing

cast-in-situ concrete piles by using high strain dynamic testing

Xi Liang
Fujian Academy of Building Research, Fuzhou, People’s Republic of China

ABSTRACT: In this paper, the bearing character of large diameter diving casing cast-in-situ concrete piles by
using high strain dynamic testing was discussed, and some problems in construction aspect were analyzed.
Advantages of bearing behavior and economy of this pile were indicated, and some suggestions were put
forward in the course of construction.
Key words: large diameter diving casing cast-in-situ concrete piles, high strain dynamic testing of piles, shafi
resistance, toe resistance

1 INTRODUCTION approximately 20-25mm. Diameter of pile former is

With the development of economy in Fujian
province, great changes have taken place in city’s
visage. Since great deals of high buildings are
constructed, the bearing capacity of piles is required
to improve with most economical method to the
greatest extent. At present, the main types of piles of
high buildings in Fujian province are pre-formed
concrete pile or bored piles and belled piles by hand
supporting on medium or slightly weathered rock. In
the course of engineering construction in Xiamen of
Fujian province, a rather novel type of pile-large
diameter diving casing cast-in-situ concrete pile has
been always encountered. By means of high strain
pile testing and analyzing with CAPWAPC program,
the author discovered that this pile has better
advantage in bearing capacity and economical
character comparing with that of other types of piles.
This type of pile fits for many areas in Fujian.
2 ‘THE CONSTRUCTION PROCESS OF LARGE
DIAMETER DIVING CASING CAST-IN-SITU
PILES
The outside former diameter of large diameter
diving casing cast-in situ piles is about 0 560 - concrete; ( 5 ) operating the vibrator and pulling the
700mm, mainly 0 700mm, @ 650nim, @ 6OOn1m
and 0 560mm, respectively the thickness of former is
mainly of@700mm and its thickness 22mm.The pile
hammer is OED that usually used in driven pre-
formed concrete pile. The model is mainly
MH72B, D62-22 and K80. The weigh of ram is 60-
801tN. Its impact force can reach more than 7300kN
and most impact energy can exceed 200k.I. On thc
top of the pile there is jacketing vibrator equippcci.
An exciting force of it is about 300-400kN. Whcn if
begin to cast concrete, vibrator is operated. It can 1101
only improve the ability of pulling the former bu1
also make the concrete be vibrated sufficiently . t 1
equipped with pile shoe beneath the pile tip. ’l’he
shape of pile shoe is similar to a cone. The rigidity
of pile shoe has been reinforced by installing
strengthening rib to improve the ability of
penetrating hardpan. The diameter of pile shoe is
larger than that of former as about 100mm. It make
the soil strata around pile skin not rebound entirely
when pulling the former, thus reducing resistance of
pulling former and ensuring construction
successfully.
The processes of construction are presented as
fo11ows :
(1) Placing the pile shoe and pile driver perching.
(2) driving the former; (3) examining pile tip with
lamp; (4) placing cage of reinforcement and casting
former; (6) pulling the former to the designed
elevation and finishing all records.

369
3 PROJECT EXAMPLE The geological column around the testing pile, Q-S
curve of static load experiment and CAPWAP result
27 piles in DH garden in Xiamen area are are shown in fig. 1, fig2, fig3, respectively.
conducted hinge strain dynamic testing. The typical
subsurface condition of this project is shown as
follows: (1)miscellaneous fill, its thickness is
about2.Om, (2)muck, its thickness about 18-23m,
(3)clay, its thickness about 2-6m, (4) residual sandy
clay, its thickness morn than 20m.
Designers initially intended to adopt 0 1OOOmm
bored cast-in-situ piles in this project and conducting
static testing of two bored piles with a length of 35m.
But they attained a capacity of only 3500kN and
4000kN. They considered adopting driven pre-
formed concrete piles afterwards. But because the
surface of residual sandy clay is strongly wavy, the
length of pile can’t be controlled accurately. Varying
pile length results in the cost of construction
increasing accordingly. So driven pre-formed pile Fig.2 Q-S curve of static load experiment
was refused. Thus for the pile foundation of this of testing pile in DH garden
project, large diameter diving casing cast-in-situ pile
with diameter Q 600mm and CD 700mm was The author has analyzed all the 23 piles with
adopted. Pile penetration is about 29-39m. The CAPWAP program. It follows from the analyzing
bearing stratum of piles is residual sandy clay. result that both the shaft resistance and toe resistance
8 piles with diameter Q600mm are tested by can be brought into a quite high level, so is it even
using high strain dynamic testing except 4 piles under the condition of very feeble subsurface. These
come across rock in the pile tip. The ultimate data are not only higher than that of small diameter
bearing capacity is about 4500-5800kN, the mean diving casing cast-in-situ piles or bored cast-in-situ
bearing capacity is equal to5150kN; 15 piles with piles, but also higher than that of driven pre-formed
piles. The ultimate toe resistance and shaft resistance
diameter 0 700mm are tested. the ultimate bearing
of this pile and values comparing with other piles arc
capacity is about 4600-6800kN, its mean value is
listed in table 1. (The data with a symbol “*” conic
equal to 5873kN. One pile among them is tested
from Technical Code for Biddii7g Pile Fozriw’cifioiis 1. A
with static and dynamic comparison experiment.
statistic result shown that the bearing capacity
This pile is carried out high strain dynamic testing in
provided by large diameter diving casing cast-in-situ
advance and determines its ultimate bearing capacity
pile with diameter Q 700mm and length 32m, is
of 6400kN; and then it is carried out static load
testing. Under the condition of load of 6100kN, the larger than that of bored cast-in-situ pile with
settlement of it is equal to36.9mm. The relative error diameter Q 1200mm and the same length or two pre-
between static testing and dynamic testing is 4.92%. formed concrete piles with a section 400mm X
400mm and the same length. It can be seen that this
type of pile has remarkable economical benefit.

Table 1 . Ultimate shaft resistance and toe resistance of large

diameter diving casing cast-in-situ piles in DH garden and
values cornoarinn with other tvoes of oiles
Ultimate Ultimate shaft resistance(kPa)
Pile toe ~ i ~ ~ ~ l l Resid iial
resistance aneous Muck Clay sandy clay
fill -
(1) 7150 46 19 85 I03
_ _
(2)’ 5900 24 14 66 82 _ _
(3)’ 1400 22 13 64 78
(4)’ 4200 19 11 52 63
(I) Large diameter diving casing cast-in-situ pile
(2) Pre-formed concrete pile
(3) Bored cast-in-situ pile
(4) Small diameter diving casing cast-in-situ pile

370
 The statistic data of static testing of 20 large
diameter diving casing cast-in-situ piles among 6
projects have been collected. The stress-strain
electric experiments of the pile shaft are conducted
for 5 piles among these piles, and the ultimate shaft
resistance in miscellaneous soil, clay and residual
sandy clay are obtained. The mean values of them
are respectively 57kPa, 79kPa and IOSkPa.
4 ANALYSIS OF BEARING CAPACITY
MECHANISM
The CAPWAP analyses of testing piles above have
shown that the shaft resistance and toe resistance of
this type of pile are all high. It follows from this
result that the mechanism of bearing capacity of this
type of pile is as follows: in the respect of toe
resistance, because of high penetration ability of this
pile equipment which can impact former to hardpan
even to weathered rock, the bearing stratum absorbs
great energy passing by steel pile shoe and is
compacted, and its modulus of deformation
increases greatly, which is similar to the driven pre-
formed concrete pile, but its bottom area is far
greater than driven pre-formed concrete pile, What
is the main reason why the toe resistance of this pile
can reach so high value.
In the respect of frictional resistance, also similar
to driven pre-formed pile, the pile former penetrate
into the soii strata by blowing of rani, soil around
former is extruded. Soil around pile rebounds and
combines close with concrete of pile after casting
concrete. But because the interface between pile and
soil around driven pre-formed pile is smooth, while
the interface between pile and soil around large
diameter diving casing cast-in-situ pile is rough,
meanwhile large diameter diving casing cast-in-situ
pile has not the mudcake problem of bored pile,
therefore, the shaft frictional resistance of this pile is
larger than that of pre-formed pile under the normal
stress.
Analyzing result of CAPWAP program of high
strain testing indicates that the end bearing of pile is
about 35-45% in all load of pile top, and its mean
value is 38.9%, which is larger than that of cominoii
bored cast-in-situ pile under the same geological
condition. The result also shows that this pile is
buoyant pile while its lengt~i-diameterratio (LiD) is
about 50. Because there are not sediment under the
pile tip and mudcake around the pile skin, the shaft
frictional resistance and toe resistance exert
simultaneously to resist the load of pile top, thus the
pile obtain higher bearing capacity. For the bored
cast-in-situ pile, because of the influence of
sediment, the effect of toe resistance may not bring
371
into full play even when the shaft frictional
resistance has been exerted to an extreme extent.
5 PROBAl3LE EXISTING PROBLEMS
After driving the former of large diameter diving
casing cast-in-situ piles has been finished, the pile
bottom is required to be examined with lamp.
Meanwhile, some effective technology measures
have been taken. There, some common faults of
quality in small diameter diving casing cast-in-situ
piles, such as pile tip mixing mud, etc., can be
avoided, and the qualified ratio of testing of large
diameter pile piles exceeds far that of the sinalf
diametcr piles.
However, when carrying out testing of a pile in SF
garden in Xiamen of Fujian province, the author also
discussed a quality problem. The diameter of this
pile with a bearing stratum of residual sandy-clay is
equal to Q> 600mm, and penetration length is 3 1Am.
By means of static load experiment, its ultimate
bearing capacity is only 2 150kN, which is only one
half of most entrusted load. After sending the pile
970mn1, then high strain dynamic testing is carried
out. Its ultimate bearing capacity was calculated
only 2620kN either. It follows from the curve that
the pile was damaged seriously on the location of
1 I m beneath the pile top. (The geological coiuinti
around pile, pile profile and pile model and
CAPWAP final results are shown respectivcly i n
fig.4, fig.5 and fig.6) This location happened to be
the boundary between muck and sandy clay, So w e
can infer the reason of causing this problem. Firstly,
when the pile penetrate densely into the saturated
soft soil during the process of construction, soil
strata around the pile is crushed to cause excess pore
water pressure. The pressure makes the ground
hunch and the pile bear a horizontal pushing force. If
the concrete intensity of adjacent piles is still very
low, the pile perhaps damaged by the horizontal
thrust force. Secondly, at the boundary between
hardpan and soft soil, the speed of pulling former is
so rapid that it causes the pile mixing muck.
In fact, on the other field with a deep muck such
as DH garden introduced above, many piles are
found some defects existing on the location of mitck
by testing. The reason may be the same as that of SF
garden.
4 CONCLUSIONS
1. The testing result of large diameter diving
casing cast-in-situ pile indicates: Its shaft frictional
resistance and toe resistance can resist together the
 Fig.3 CAPWAP final result of testing pile in DH garden

Figure 4. Geological column around Figure 5. Profile and model of testing

testing pile in SF garden pile in SF garden

load of pile top. The values of their shaft resistance other area in Fujian province, especially in the area
and toe resistance are far greater than that of with thicker residual soil and deeper weathered rock
common cast-in-situ piles. This pile has notable 2. Quality problem of large diameter diving casing
economical benefit and is worthy of generalizing to cast-in-situ piles has been reduced greatly

372
 F i g 6 CAPWAP final result of testing pile in SF garden

comparing with small diameter diving cast-in situ

piles. But when driving in the saturated sof’t soil
such as deep muck, the order of driving piles and the
speed of pulling the former must severely controlled,
and the distance between piles is properly widened.
3. High strain dynamic pile testing has playcd a i l
important role in engineering inspection. ‘l’hc
bearing capacity mechanism of pile can be analyzed
with this testing method and the location of pile
damage can be determined. It has obvious social and
economical benefits. With regard to large diameter
diving casing cast-in-situ piles, it should accumulate
constantly the data of comparison between static and
dynamic testing, especially that of contain the stress-
strain electric experiments of the pile shaft.

REFERENCE

373
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Analysis of dynamic load tests on steel rails piles

E M.A. Lima, J.C.A.Cintra & N.Aolu

University of SCo Puulo, SCo Carlos, Brazil

ABSTRACT: This paper presents an analysis of dynamic load tests performed on steel rails piles (TR-68) at
the Experimental Foundation Field of the Department of Geotechnics, Engineering School of S2o Carlos,
University of SBo Paulo, Brazil. The piles have segments of 12 and 3m, welded according to NBR 8800/86,
comprising a maximum driving length of 27m. The objective of this analysis is:To carry out a wide study of
steel track pile bearing capacity; To validate a tool predicts in NBR 6122/96: the rebound. Validate with the
control method of driving piles and verify the utilization of the steel rails piles with structural element of
foundation. The rebound measurements were performed with paper and pencil in eighteen piles and were
analyzed by applying increasing energy level. The energy was applied by the gravity hammer falling of a drop
heights of 0.2 - 0.4 - 0.6 - 0.8 - 1.0 - 1.2 and 1.5m. The bearing capacity of the piles, determined by
extrapolations of mobilized resistance - displacement curves is compared with two load tests employing the
Pile Driving Analyzer (PDA) and with one static load test on a representatives piles.

I INTRODUCTION
Dynamic load tests with increasing energy has
been used in Brazil for 10 years and allows the
bearing capacity determination of the soil-
foundation system from a load- displacement curve
defined by a certain number of points.
The rebound technique has also been spread in
Brazil as a driving control tool of all piles in a
construction site.
In practice, the measured rebound signal is the
wave equation solution. The rebound recording at
the top of the pile with pencil and paper can be used
to estimate the mobilized load for the corresponding
energy level. Assessment of the mobilized load -
displacement curve for several energy levels allows
the ultimate load estimation.
2 DYNANAMIC TESTS PERFORMED DURING
PILE DRIVING
drop heights of0.2, 0.4, 0.6, 0.8, 1.0, 1.2, and 1.5 m.
The elastic (rebound) and permanent (set)
displacements were recorded through paper and
pencil, placed at the top of the pile.
Based on the measured rebounds, the mobilized
load was calculated according to UTO et al. (1985)
and VELLOSO (1987) propositions.
- Displacenieri f parcels C2 arid C3 deierniirinfrori.
C2 and C3 displacement parcels are the first and
major variables in the load test assessment using the
rebound technique. Inadequate assumptions of these
variables can lead to misrepresentative mobilized
load values
Scatter results were obtained for the formerly
assumed values C3 = 2.5 mm and C = set. Thus, the
displacement parcels C2 and C3 were determined by
AOKI (1986) proposition

Dynamic tests with rebound measurement at the

end of the driving of the piles were performed Satisfactory results were achieved by evaluating
according to CHELLIS (195 1) proposition, C3 with this methodology. C3 mean value was 0.98
improved by AOKI (1989). Increasing energy levels mm, with a standard deviation of 0.34 mm,
were applied by the 20 kN gravity hammer, with corresponding to a coefficient of variation of 34.7

375
%. The maximum and minimum obtained values the stabilization is achieved even if the energy keeps
were 0.40 and 2.27, respectively. Figure 1 illustrates increasing;
C3 values normal distribution curve.

Fibwe 2. The plot of the inean value of C3 dsplaceinent versus

the liartuner potential energy

Figure 1. Illustrates C3 1,alues nonnal distribution cume The set displacements tend to a nearly zero value
for the lower energy levels, and increase as the
Table 1 presents an statistic analysis for the applied energy increases.
applied height blows, in an attempt to exemplify C3
variation at each applied energy level. -13enririg cnpncitj of the piles
Tabel 1. Statistic analysis for the applied height blows. The bearing capacity of the piles is evaluated
Blon. Potenc. Mean Stand. Coeff. Mar. Min. from the static mobilized load - dynamic
height Enerb? C3 Desv. of Var. Value Value displacement curve
(111) (kJ) (111111) (111111) (%)) (111111) (111111) For the mobilized load values calculated by
0.2 4.0 0.79 0.28 35.3 1.73 0.45 VELLOSO (1987) expression, the shape of the
0.4 8.0 0.91 0.34 37.4 2.08 0.40 mobilized load - displacement curves shows a clear
0.6 12.0 0.9') 0.35 35.4 2.21 0.64
0.8 16.0 1.03 0.33 32.0 2.21 0.73 failure in most of the cases, clearly defined by a
1.0 20.0 1.03 0.33 32.0 2.21 0.73 vertical asymptote
1.2 24.0 1.08 0.34 31.5 2.27 0.73 However, as UTO et a1 (1985) expression did not
1.5 30.0 1.05 0.34 32.4 2.27 0.79 allow a clear observation of the ultimate failure, an
extrapolation of the mobilized-displacement load
It can be observed that C3 values are affected by was then necessary
the maximum displacement and the permanent The extrapolation was performed through VAN
values of each blow. As these displacements depend DER VEEN (1 953) method Conventional failure
on the applied energy levels, one can affirm that C3 load methods by DAVISSON (1972) and the
value is an energy level dependent variable. Brazilian Standard NBR-6122/96 were also used
This statement is better understood by observing Table 2 and 3 show the mean bearing capacity
Figure 2, which presents a plot of the mean value of obtained by the above-mentioned methods
C3 displacement as a hnction of the hammer
potential energy. Table 2 - the niean bearing capacit? obtained b j mobilized load
The above results substantiate AOKI (1991) \alues calculated b\ VELLOSO (1987) elpression
statement' the tip quake (C3) is a variable that Velloso
depends on the applied energy level L Ma\ Veen Nl3R Davissoii
(in) Value (kN) (kN) (kN)
Dynamic load tests displacement parcel analysis (kN)
validate the results acquired by NIYAMA & AOKI Mean 263 1120 1267 1247 1201
(1 991): Stand Des\ 0 46 160.0 239.0 260.0 257.0
The rebound displacements increase with the Coeff ofVar 1 7 14.3 18.9 20.9 21.4
applied energy up to a maximum limit. Afterwards,

376
Table 3 The mean beanng capacih obtained by mobilized load
UTO et a1 (1987) expression
1 alues calculated b\
Uto
L Max Veen NBR Da.cisson
(in) Value (kN) (kN) (kN)
(kN)
Mean 26 3 926 1499 874 741
Stand Den 0 4 6 83 173,O 94.0 81.0
Coeff of Var 1 7 9.0 11.5 10.8 10.9

Tables 4 and 5 present the mobilized load values

for each applied energy level.
Table 4 The mobilized load values for each applied energ?
level calculated by VELLOSO (1 987)
Blon Pot R Stan Coef May Miii Mean
height Eiier Mean Den Var Value Value Displ
(m) (kJ) (kN) (nun) ('3)) (nun) (min) (inm)
0.2 4.0 745 181 24.3 1066 338 8.8
0.4 8.0 880 218 24.8 1285 295 11.3
0.6 12.0 1000 237 23-7 1489 605 14.0
0.8 16-0 1o-M 19-4 18.6 14% 786 15.8 Figure 3 The plot of set versus the ultimate resistance of the
1.0 20,O 1018 215 20.5 1592 756 17.2 soll-plle system
1.2 24.0 1132 258 22.8 1902 786 19.3
1.5 30,0 1056 158 15.0 1112 800 22.1
3 DYNAMIC LOAD TESTS PERFORMED
Table 5. The mobilized load values for each applied energy
level calculated by UTO et al. (1985). DURING PILE RESTRIKE
B1o.c.c Pot. R Stan. Coef Max. Min. Mean
llclglll Encr. Mca n Dcs! . Var. Valuc Valuc Displ. AOKI (1 989) proposition on using increasing
(111) (k J? (kN) (mm) ('%) (mm) (Inm) (ill111) energies to define the mobilized load-displacement
0.2 1.0 498 88 17.6 668 3 16 8.8 curve was also followed during this phase.
0-4 8.0 599 100 16,7 789 343 11.3 The dynamic load test with the PDA was carried
0.6 12.0 696 109 15.7 922 5 15 14.0
0.8 16.0 755 100 13,2 1000 563 l5,8
out during piles E-23 and E-10 restrike, both
1.0 20.0 793 105 13.2 1033 587 17.2 considered as representative of the set of piles.
1.2 24.0 873 113 12.9 1168 640 19.3 Potential energy levels of 4.8, 9.6, 14.4, 19.2, 24.0,
1.5 30.0 914 93 10,2 1092 758 22,l 28.8, and 36.6 kJ corresponding, respectively, to
hammer fall heights of 0.2, 0.4, 0.6, 0.8, 1.0, 1.2.
Results show that the mobilized load increases as and 1.5 m were applied in an attempt to compare the
the applied energy increases. results of these tests with the results by the rebound
Although VELLOSO (1987) method generated technique.
higher mobilized load values, the average ultimate Pile restrike was performed one year and eight
load value for the curve achieved by UTO et a1 months prior the ending of the driving, and the
(1985) method is higher than the average ultimate results were analyzed at each energy level by CASE
load corresponding to the curve obtained by and CAPWAP methods.
VELLOSO (1987) method. The rebound and set displacements were also
This substantiates the major role that the load- recorded through paper and pencil.
displacement curve shape plays on the definition of Signal trace analysis using the PDA and the
the ultimate failure load. rebound technique (concurrently measured) for each
The acquired failure loads follow the same trend applied energy level are exposed in Table 6 and 7.
of the mobilized load values when DAVISSON Signal traces measured by the PDA were assessed
(1972) and the NBR-6122/96 methods are through CASE and CAPWAP methods, while the
employed. This happens because the values by rebound signal traces were assessed through UTO et
VELLOSO (1987) are higher than the values by al. (1 985) and VELLOSO (1987) propositions.
UTO et al. (1985). The dumping constant J, was assumed to be 0.7
The set displacement does not define the ultimate for the CASE method analysis. This value is within
resistance of the soil-pile system, as shown in Figure the range proposed by RAUCHE et al. (1985) for
3 piles with tip embedded in a clayey silt soil.

377
Table 6 - The results of restrike arial!Led of E-10 pile with this dynamic load test. This statement is better
E-10 understood by observing Figure 4.
B l o ~ Energ! Case Capnap Rebound Rebound
Height Potencial (kN) (kN) Vclloso Uto Tabele 9 - The quake values obtained b) tlie CAPWAP
(111) (kJ) (kN) (kN) anal! sis
0.2 4.8 420 444 513 370 Golpc E-23 E- I 0
0.4 9.6 + + 784 510 H Energia Quake C3 Quake C3
0.6 14.4 790 1023 1046 647 Potencial Lateral Ponta Lateral Ponta
0.8 19.2 1060 1272 1136 693 (in) (kJ) (inin) (inni) (inm) (inin)
0.2 4.8 1.95 1.70 0.98 0.19
1.0 24.0 1120 1299 1316 810
1.2 28.8 1170 1344 1316 810 0.4 9.6 1.88 1.36 + +
1.5 36.0 1460 1490 1681 997 0.6 14.4 2.59 2,00 1.03 0.35
+ This blov did not shon n a good signal 0.8 19.2 2.05 0.30 1.23 0.32
1.0 24.0 0.99 1.25 1.22 0.35
Table 7 - The results of restrike analyzed of E-23 pile 1.2 28.8 0.94 0-60 1.50 0.40
E-23 1.5 36.0 + + 2.39 0.14
Blou Energy Case Capwap Rebound Rebound + This blon did riot slion ii a good signal
Height Potencial (kN) (kN) Velloso Uto

0.4 9.6 630 673 789 512

0.6 14.4 760 600 964 604
0.8 19.2 900 811 1149 651
1.0 24.0 960 1102 1234 768
1.2 28.8 990 1110 1418 862
1.5 36.0 + + 1418 862
+ This blon did not shon 11 a good sigiial

The lateral friction and tip resistance values, as

well as the obtained quake values through the
CAPWAP analysis are presented in tables 8 and 9.
Tabele 8 - The lateral friction and tip resistance values
obtained by the CAPWAP analysis.
Golpe E-10 E-23
H Energia Capwap Atrito Ponta Capwap Atrito Ponta
(in) Potencial (kN) (kN) (kN) (kN) (kN) (kN)
0.2 4.8 444 435 8 418 368 50
0.3 9.6 + + + 671 568 103
0,6 14,4 1023 831 173 600 440 160
0.8 19.2 1272 1094 178 811 428 383
1.0 24.0 1299 862 437 1102 862 240
1.2 28.8 1346 795 551 1110 896 214
1.5 36,O 1490 911 579 + + +
+ This bloli did not shown a good signal

Restrike results show that the structural element

ultimate load limits the bearing capacity of the piles.
The same conclusion can be observed either by the
PDA and the rebound techniques.
The mobilized and ultimate loads obtained by
VELLOSO (1987) proposition using the rebound
technique yielded close results with respect to the Figure 4 - the mobilized resistance-displacement curve in tlie
pile restrike.
load values generated by the dynamic tests with the
PDA and analyzed via CASE and CAPWAP
methods. However, load values obtained by UTO et Thus, it is demonstrated that the rebound
al. (1985) proposition did not show good agreement technique employment in pile driving control,

378
assessed by VELLOSO (1987) proposition, is a
highly reliable tool in comparison with the wave
equation based methods.
The signal trace assessment through the
CAPWAP method shows a virtually perfect
superposition of the measured and the calculated
load values.
A prevalence of the lateral friction load is shown
by the CAPWAP analysis, even with a substantial
part of the tip load.
Despite the scatter shown by some points, the
lateral resistance increases with the increase of the
applied energy level. Further study is needed to
understand this dispersion and the absence of
continuity on the resistance curves with depth.
However, some hypothesis can be drawn to explain
this:
A CAPWAP model incompatibility, which
assesses the mobilized load for an energy level
considering it to be the ultimate system load;
Figure 5 - the result of the static load test.
The soil residual tensions which dissipates at the
higher energy blows;
The tip resistance also increases with the increase
in the applied load. Specifically, the scattering on
pile E-23 data regarding the lateral resistance could
again be observed.
Although the mobilized load estimates showed
good results, comparisons between the PDA tip C3
and the rebound technique estimated C3 were poor.
Further explanations are needed, but a preliminary
hypothesis is that the assumed C3 in the dynamic
formulas consider the tip and friction parcels
together, while the C3 assessed by the PDA takes
into consideration the pile tip parcel only.

4 STATIC LOAD TEST

The static load test was performed on pile E-23,

after completion of the dynamic load tests. The
result of the test is shown in the Figure 5
Ten load stages of 120 kN were applied,
corresponding to 10 ?o' of the estimated ultimate
load. The test was terminated at a load of 1200 kN
and a maximum settlement of 17 mm, corresponding
to 10% of the larger pile dimension. The test was
interrupted because the applied load approached the
reaction system allowable resistance.
The static load test corroborates the trend of the
other tests, showing that the ultimate rail structural
resistance is probably the variable that limits the set
of piles. The pile-soil system response was elastic
because the pile tip lies in a very resistant layer.
Figure 6 compares the dynamic load test results and
the elastic pile compression.
Figure 6 - Comparisons between the dynamic load test results
and the elastic pile compression.
Comparisons between static and dynamic load
tests can be better evaluated through figure 7. All
dynamic test analysis showed appropriate results
with respect to the static test results. The dynamic
test with rebound measurements and assessed by
VELLOSO (1987) proposition yielded the closest
results.
5 DRIVING AND RESTRIKE COMPARISONS
Rebounds were measured during the driving and
restrike test phases. Figure 8 shows the mobilized
load - displacement curves during both phases.

379

Figwe 7 - Comparisons between static and dynamic load tests.
Figure 8 - the mobilized load - displacement curves during the
driving and restrike.
The mobilized load - displacement curves show a
considerable improvement of the soil resistance with
time, in a phenomenon named setup.
By analyzing the ultimate load of piles E-10 and
E-23, a set up between 30 and 100 % was found.
6 CONCLUSIONS
The dynamic load test with rebound measurement
is a very straightforward and, because it does not
require any type of equipment, almost costless
technique.
380
The assessment of C3 is the major drawback of
the dynamic rebound analysis. However, the method
developed by AOKI (1 986) generated excellent
results of the mobilized and ultimate loads.
The mobilized and ultimate load values
determined by VELLOSO ( 1 987) proposition and
the rebound technique showed satisfactory results in
comparison to the mobilized and ultimate load
values obtained by dynamic tests with the PDA,
assessed via CASE and CAPWAP. Comparisons
with the static load test show results that encourage
the employment of this methodology.
Rebound measurements are highly affected by the
applied energy level, increasing with the applied
energy growth up to a maximum limit. Afterwards,
stabilization is reached even if the energy keeps
increasing.
Good results are not obtained with the
assumption of a constant C3 value.
Because it considers a single point in the load -
displacement curve, the dynamic load test with the
PDA needs some adjusting regarding the used
models and the test procedure.
C3 values obtained by CAPWAP analysis
showed a great scatter, with no logical order with the
increasing in the energy.
The load-displacement curve ensures more
reliable results for the ultimate load evaluation.
Conservative analysis can be obtained if the load -
displacement curve is not used.
Time effect on the soil resistance recovery was
verified with the rebound technique, during driving
and restrike. An increase of 30 and 100 % was
observed in the initial resistance.
The steel rails piles has a excellent performed
when it is request for axial compression.
7 ACKNOWLEDGEMENT
The authors are grateful to SCAC Fundaqaes e
Estrutura Ltda for execution the dynamic load test
with PDA.
REFERENCES
ABNT/MB 3372 (199 1) Estacas - Prova de Carga Esthtica
A B N T N R 6 122 (1996) Projeto e ExecuCSo de FundaC6es
A B N T N R 8800 (1986) Projeto e Euecu@o de Estruturas de
AGOde Edificios (metodo dos estados limtes)
A B N T N R 13208 (1991) Estacas - Ensaio de Carregarnento
Dmhico
AOKI. N (1986) Controle iii sitir da capacidade de carga
”
de estacas pre-fabricadas via repique elastic0 da cravaGdo
PublicaCSo da ABMS/NRSP. ABEF E IE/SP, p 48
AOKI. N (1989) A new dynamic load test concept In XI1
INTEFWACIONAL CONFERENCE ON SOIL
MECHANICS AND FOUNDATION ENGINEERING. TC
 Pile Driving, RIO de Janeiro Proceedings for the
Discussion Session 11, T 1, p 1-1
AOKI. N (1991) Carga Admissive1 de Estacas Atraves de
Ensaio D i n h i c o In I1 SEMINAFUO DE ENGENHARIA
DE FUNDACOES ESPECIAIS. S2o Paulo, v 2, p 269-292
LIMA, F M A (1999) ”Analise de pro\ a de carga dinimca e
estaca metalica do tipo tnlho” Disserta@o de Mestrado
Escola de Engenhana S2o Carlos EESC - USP. 15Op
NIYAMA. S , AOKI, N (1991) CorrelaCBo Entre Provas de
Carga Dininiica e Estatica no Cainpo Exyenmental da
EPUSP/ABEF _In I1 SEMINAFUO DE ENGENHARIA
DE FUNDACOES ESPECIAIS. S2o Pado Anais. 1 1,
p 285-293
RAUSCHE, F, GOBLE. G G (1985) D!nainic Determination
of Pile Capacity Jourtial of Geotechiiical Ei?grneerri~g
ASCE, \ 111. n 3. p 367-383
UTO. K, FUYUKI. M, SAKURAI. M (1985) “An EquaQon
for the D j nainic Bearing Capacity of a Pile Based on Wave
Theon “Proceedings of the Iiiteriiatroiial S~i?lposiiri?ioil
Penetmbrlrtv a i d Drivabilitv of Piler San Francisco. v 2
VAN DER WEN. C (1953) “The Bearing Capacity of a
Pile” In 3rd INTERNACIONAL CONFERENCE ON
SOIL MECHANICS AND FOUNDATION
ENGINEERING, ZWch v 2 , p 81-90
VARGAS. M (1990) Provas de Carga em Estacas - Uma
‘I

apreciaGBo Historica Solos e Rochas Vol 13, Umco, p 3

I’

VELLOSO. D A (1991) “Capacidade de carga por meio do

SPT SEFE 11. SBo Paulo, p 293-3 12
VELLOSO. P P C (1987) Fuiida@es - &-lspectosgeoteciiicoc
PublicaC2o do Departainento de Engenharia Civil da PUC-
FU.\ 2.p 300C-300H

381
This Page Intentionally Left Blank
 A ~ ~ l ~ cofaStress-Wave
t~o~ Theory to Files, N i y a ~ a& Beim (eds)02000 Baikema, Rotterdam, lSBN 90 5809 1503

A discussion of penetration matching on high strain dynamic pile testing

Xu Dingliang
Provinc~alTesting Centre o f ~ r a fEnginee~.in~~,
~c Fujian Province, People’s ~ e ~of China
u ~ ~ ~ c
Wu Shiming
Tongji University,Shanghai, People’s Republic of China
Xiao Liping
The Inc of Foundution und Construction, Fuzhou, Fujian Province, People’s Republic of China

ABSTRACT: This paper points out the real connotation of penetration matching in High Strain Dynamic Pile
Testing. The relationship between some coefficients and CAPWAP result is set forth through the discussion of
two actual engineering examples. This paper may be useful to adopt the standard of penetration matching
correctly arid to judge the reliability of a CAPWAP result by analyzing the value of some main coefficients.

A perfect penetration matching in High Strain BCTq = 1.O I’ ( Utrn - qav ) (1)
Dynamic Pile Testing is always regarded as an
important rule to judge a pile’s capacity correctly. Where Utm is the largest displacernent at a pile’s
The material [I J point out “while using CAPWAP or bottom; qav is the weighted average of elastic-li~it
other similar program to analyze a pile’s capacity the of a pile section (include the pile bottom-section).
rules must be obeyed, they are: firstly, the accepted See Fig. 1. The calculating formula is as follows::
coefficients in program should be within a
reasonable range; secondly, the matching quality of qav = C(qi * Rui)/Ru (2)
the curve should reach a certain standard. For
concrete piles, the value of MQ should less than 10. Where qi is the elastic-limit of every pile’s section
Thirdly, the calculated penetration should include the pile bottom-section) , the Rui is the
approximate the tested value”. The state regulations static resistance of every pile’s section and RU is the
of High Strain Dynamic Pile Testing” also point out:pile’s total static capacity.
When a matching analyzing coming to an end, the The value Utm - Qav, or the reciprocal of BCTq
calculated curve should app~oxiniatethe tested curve in forrnula (l), approxi~~ates the pile bottom’s
and the calculated penetration should approxi~natetoremains displacement really. See the length AB in
tested penetration”. Then, what’s the real Fig. 1. We know the calculated penetration would
connotation of penetration matching? What condition approximate tested penetration on a pile’s top if the
does the standard of penetration matching suit? How curve matching reached a perfect quality. Therefore,
will a perfect penetration matching affect tlie control
the demand of “perfect curve matching” and
discreteness of analyziiig result? The paper is “perfect penetration matching?’ means that the
devoted to discuss these questions from the penetration on a pile’s top should be equal to the
definitions of penetration and analyzing two pile bottom’s one and the tested penetration while
examples of actual engineering. the CAPWAP analyzing come to an end.
It follows that so-called penetration ~ a t c l i i ~ i g
1 The real connotation of p e n ~ ~ a t i o nsuits only integrate piles which are rigid enough. If a
matching pile has a fatal defect in its body or it has a lower
rigidity body, the penetration on tlie pile’s top is
According to the section of BACKGROUD in really unequal to the bottom’s one; hence it is
material 121, the formula to calculate the times of obviously unreasonable to demand perfect Curve
hitting (The Reciprocal of the times of hitting is matching and perfcct penetration matching to be
Penetration) in CAPWAP is: satisfied at same time. Otherwise the accepted
mechanical coefficierits of the pile and the soil will
contain deviation

383
 P i l e bottom’s section, and the fissuring under a pile’s bottom etc.

t
damping f o r c e There are two examples below showing the possible
errors after the standard of a perfect penetration
matching and a perfect curve matching has reached.
Example 1: At ii worksite were constructed
prestressed pipe piles with a diameter of 500mm and
a length of 38111. After the piles were completed, we
Pile b o t t o m 7
testers used a 6.2t diesel hammer to excite. As
different CAPWAP coefficients were adopted,
difi’erent results were gotten as 1-A and 1-B:
Result 1-A. The analyzed result is as in
Table 1, where “Rut” eqrtals the pile’s judged static
capacity, “MQ” equals the matching quality value,
Fig. The displacement c u r v e
of a pile b o t t o m section “Dcw” equals the calculated penetration, and “Dms”
equals the tested penetration. ’I’he matching curves is
as in Fig 2, where “Fur Msd” equals the measured
force curve, “Vel Msd” equals the measured velocity
2 The possible range of deviation after a curve, and “For Cpt” equals the calculated force
perfect penetration matching is reached curve.

The PUTose Of adoPtillg the standard for good Table 1. The anajyzed result 1-A of CAPWAP
penetration matching is to control the discreteness of
the analyzed results in cooperation with the standard
Rut MQ Dcw Dms
for good curved matching, so as to obtain a better
analyzed result. Undoubtedly, the idea is correct. But,
besides the static capacity, there are a lot of 5789.IKN 4*25 2.79MM 2.80MM
coefficients in CAPWAP that may affect the
calculated penetration such as radiant soil damping
on pile’s circumference and bottom section, the Result 1-B. The analyzed result is as in Table 2.
elastic-limit on pile’s circumference and bottom The matching curve is as in Fig 3.

384
Table 2. The analyzed result 1-B of CAPWAP. different results were gotten as in Results 2-A and 2-
~

B:
Rut MQ Dcw Dms Result 2-A. The analyzed result is as in Table
3. The matching curve is as in Fig.4
4089.1KN 4.13 2.76MM 2.80MM
Table 3. The analyzed result 2-A of CAPWAP.
Example2: At a worksite were constructed Rut MQ Dcw Dms
prestressed pipe piles with a diameter of 5OOmin and
a length of 39.2m. After the piles were completed,
we testers used a 5.0t free falling haixnier to excite. 6123.9KN 3.25 0.955MM 1. OOMM
As different CAPWAP coefficients were adopted,

385
Result 2-B. The analyzed result is as in Table 4. penetration matching has been conducted. In
The matching curve is as in Fig 5. Examplel, compared to the average value, the range
of deviation of the judged static capacity is & 17.2%.
Table 4. The analyzed result 2-B of CAPWAP. In Example 2, compared to the average value, the
range of deviation of the judged static capacity is
Rut MQ Dcw Dins +-21.2%.
Analyzing the change of the coefficients in
3975.9KN 3.67 0.965MM 1. OOMM CAPWAP in the two examples above, we can find
that the bigger the adopted dynamic damping force is,
the smaller the pile’s static capacity is. The value of
The main coefficients in CAPWA‘E’in the above soil elastic-limit reflects the difficulty to penetrate a
two examples are as in Table 5. It shows a certain pile into the soil. The bigger the value of soil elastic-
range of deviaticn still exists even though all the limit is, the smaller the peiietration value is.
matching quality values are less than 5.0 and a good

Fig 5.

I The dynamic I Calculated

damping force Soil pile’s static
Example Damping force of radiant soil elaStiC-limit capacity Ru
the pile’s circum damping force
Result SK=2.2 Q2=0.18
1 1-A SS=0.189 MS=16 QT=0.20 5789.1KN
Result SK=O QS=O. 18
1-B SS=O.G9 MS=O QT=0.20 4089. IKN
Result SK=O QS=O.l8
2 2-A SS=0.25 MS=O QT=0.20 G 123.9KN
Result SK=O QS=O. 18
2-B SS=0.67 MS=O QT=0.20 3975.9KN

386
 In order to satisfy the standard of a perfect .c Dcw
I
matching, the radiant soil damping is adopted in
Result I-A ( SK=2.2, MS=16--indeed, the radiant
soil damping force under this status is less than that
under the status of SK=O and MS=O), the damping
force on pile’s Circumference is decreased (SS=O. 189)
and the static capacity is increased(Ru=5789.1 KN). QS, QT Ru
While in Result 1-B, the coefficients of the radiant
soil damping is set to zero (SK=O, MS=O, it means a
huge radiznt soil dampkg force is set), the damping Fig. 7
force on the pile’s circtimference is increased
(SS=0.690) and the static capacity (!<u=4089.1KN)
is decreased. Similarly, in order to satisfy the 3. CONCLUSIONS
standard of a perfect matching, in Example 2-A, the
lower elastic-limits of the pile’s circumference and We can get the four conclusions through the above
bottom section are adopted (QS=0.262, QT=0.264), discussion:
the damping force on the pile’s Circumference is A. The real connotation of penetration
matching is that the calculated penetration of the pile
should approximate the pile bottom’ s one and the
t R‘A
tested penetration.
B. The standard of the perfect penetration
matching does not suit all piles. It suits only the
integrate pile that is rigid enough. If a pile has a fatal
defzct in it’s body or it has a less rigid body, the
standard won’t be suitful.
C. It is necessary to adopt the standard of the
perfect penetration matching for an integrate pile that
is rigid enough, but the analyzing result may be
Fig. 6
discrete in a wide range. It is necessary to study
further the dynamic damping coefficients of soil such
as the QS, QT, SS, ST, SK and SM etc.
decreased (SS=0.25) and a higher calculated static
D. Fig.7 and Fig.8 may give a hint for judging
capacity (Ru=6 123.9KN) is accepted. While in the reliability of a CAPWAP result under the current
Example 2-B, the higher elastic-limits of the pile’s technical level.
circumference and bottom section are adopted
(QS=0.3, QT=0.35), the damping force on the pile’s
circumference is increased (SS=0.690) and a lower REFERENCES
calculated static capacity (€1 u=3975.9KN) is
accepted.
[ 11 The “Tentative Regulations of High-Strain
If we summarize the relationship among the
Dynamic Pile Testing” published by the State
radiant soil damping force (Rr), the damping force Supervision and Testing Center of Construction
on pile’s circumference and bottom (Rcb) and the Engineering in 1989.
pile’s calculated static capacity (Ru), we can draw
[2] “CAP WAP Manual” Goble Rausche Likins
the relationship as in Fig.6. and Associates Inc. in 1993
If we summarize tha relationship among the
elastic-limits of the pile’s circumference and bottom
section (QS and QT), the calculated static capacity of
the pile (Ru) and the value of calculated penetration
(Dcw), we can draw the relationship as in Fig.7.
According to Fig.6 and Fig.7 we can judge
whether a CAPWAP result tends to a safe one or not
by checking the main coefficients in CAPWAP.

387
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Experience gained and difficulties in performing dynamic load test

in composite piles made with steel rails
6.P. Bernardes
Department of Civil Engineering, UNES8 Guuratinguetb, Brazil
C. S.Andreo & C.GonGalves
Benaton Pile Foundation Company, Scio Paulo, Brazil

ABSTRACT: The results of dynamic instrumentation in two construction sites that used steel rails as pile
foundation are presented in this paper. The first dynamic load tests were executed in piles made by single
steel rails of the type TR32 and TR37. In the second group of dynamic tests, the piles were made by a
composition of two and three steel rails of the same type TR37. The difficulties in placing the sensors, the
effect of hammer blow eccentricity and the influence of the non-uniform welding along the pile length are
presented in detail and discussed.

1 INTRODUCTION In some regions in Brazil, steel old rails are

Piles with unusual cross section as steel rails,
represents a difficult task for engineers that carry out
dynamic instrumentation. The first problem is how
to place the transducers and the accelerometers when
the pile cross section does not permit that pairs of
sensors to be fixed diametrically opposite or, even
for one sensor, to be installed centrally in relation to
pile geometry and with the hammer blow. This
unusual position of the instruments certainly will
interfere in the signals and subsequence
interpretation.
No difficulties are encountered for placing the
pairs of sensors in pipe piles and any deviation in the
average signals of force and velocity in the
beginning might be related to hammer blow
eccentricity that is not difficult to adjust the hammer
system. For the case of "H" and "I" steel piles,
depending the size of the web and flange, both pair
of sensors can be placed in opposite side of the web
or diametrically opposite in the flanges (Paraiso
1990). However for this type of piles, force
induction is never completely centrally which might
result in bending and a great deviation of the signals.
One possibility to overcome this problem is to use
more pairs of sensors (Bergdahl and Moller 1984) in
order to select the best signals for further analysis.
More difficulties are encountered when doing
dynamic instrumentation in very slender piles and
sheet piles (Hartung et al. 1992). They strongly
emphases that; "piles with other geometry shapes
and consequently other specific characteristics than
normal piles, the usual test procedure and evaluation
method are only partially adequate".
available in abundance due to the process of
renewing the rails road system. This condition make
more attractive to use old rails, single and
composite, as pile foundation instead of others
conventional type of piles. Some requirements for
using old steel rails as foundation elements such as
the effective cross section area and the yield stress
are specified in the Brazilian Standard NBR-6122.
Here, the results of dynamic load tests in such piles
are presented and discussed.
2 INSTRUMENTATION OF SINGLE RAIL PILE
The first experience in performing dynamic
instrumentation in single steel rail was during the
foundation construction of a residential building in
SBo Paulo. The positions of the accelerometers and
the strain transducers in the first attempt are
presented in Figure la. For this test setup, there was
a great deviation in the signals and it was not
possible to do further analyses with the average
signal. After all the efforts, including the alignment
of hammer pile system, the best signal for one pair
of sensor is presented in Figure 2, test (a).
The second test setup is illustrated in Figure Ib.
One piece of rail web was cut and welded in order to
place the instruments diametrically opposite. Of
course there was an increasing in the pile impedance
in that area but, it was possible to evaluate the
quality of the signals in each pair of sensors and
storage the average signals with little deviation in
the beginning, Figure 2, test (b).

389
Figure 1. Test setups for single steel rail (A = accelerometer;
T = strain transducer)

Figure 3. Capwapc@signals match.

Table 1. Mobilized static soil resistance fiom CAPWAPC@

analyses.

(a> 5,7 500 375 125 14,6

(b) 5,1 645 495 150 13,2
Obs.: EMX = measured energy; R,,t = mobilized
total resistance; Rsk = skin resistance; Rt,, = toe
resistance; DMX = pile top displacement.

Figure 2. Force and velocity traces for different test setups steel
rail type TR 32.
The results of the CAF'WAPC@analyses base on
wave up matching are presented in Figure 3 for the
force (F) signal. The quality match for both cases is
very good with "Mqno" equal to 2.17 and 2.46,
respectively. However, the mobilized soil resistance
was considerable different as presented in Table 1.
For test (b), where force (F) and velocity
multiplied by pile impedance (ZV) signals are
congruent in the beginning, the mobilized soil
resistance is 29% higher than test (a) where great
deviation occured.
In this example, it is clear the necessity of a good
control of the test procedure and instrumentation set
up to carry out dynamic test in such piles otherwise,
their results should be use with caution.
3 INSTRUMENTATION OF COMPOSITE PILE
The difficulties and the experience with dynamic
instrumentation in composite steel rails were gained

390
 during the foundation construction for a shopping
center in the north of Brazil. In order to reduce the
number of piles and to increase the bearing capacity,
it was used composite piles made by two and three
steel rails welded according to Figure 4.
The sensors for the first test attempt were placed
in the web. However random signals make
completely impossible to do any hrther analyses.
One of the problems was related to the method used
to weld the two rail segments. To avoid bending of
the pile by high temperature, the welding was made
every other 40 cm along the entire length.
Unfortunately, the piles were driven without the
welding process complete these 40 cm gaps to
guarantee a uniform link along the pile length. It was
practically impossible to do any measurement when
placing the sensors in the segment that present this
gap between the rails.
The signals were recorded using the test setup of
Figure 4a. For the pile segment above the ground
surface the gaps were also welded to minimize the
high frequency observed in the first test setup
attempt.

Figure 5. (a) Force and Velocity traces for composed two steel
rails type TR 37 (b) CapwapcO signal match.

in the mobilized soil resistance about 52%.

Figure 4. Test setups for composite piles (A = accelerometer;
However, since no better measurement was possible
T = strain transducer). the results of the CAPWAPC@analyses for both
cases, should be use with caution as design criteria.

However the gaps still exist inside the ground and

the best signals measured are presented in Figure 5a.
The signals are congruent in the beginning and the
influence of the nonhomogeneous welding is clearly
seem. The force signal matching fiom CAPWAPC
analysis in Figure 5b gave a quality match (“Mqno”)
of 4.85. The mobilized soil resistance for plug
condition was 672 kN with 50% lateral resistance.
The pile geometry made with three steel rails and
the tests setup are presented in Figure 4b. Afker
many attempts, including placing the sensors in only
one rail, the best signals were measured for the
setup illustrated in Figure 4b. The stored signals and
the force matching are presented in Figure 6. The
great deviation in the signals and the gaps along the
welding resulted in a “Mqno” equal to 4.29.
For this situation the mobilized soil resistance
from CAPWAPC analysis (Fig. 6b) was 1020 kN
@
being 40% lateral resistance. Comparing the two
pile geometry fi-om Figure 4 there was an increasing
4 CONCLUSIONS
Here, the difficulties in placing the transducers and
@ accelerometers in steel rail to carry out dynamic
instrumentation were described. For single rail, when
placing one pair of sensors in the web and the other
in the flange, it is suggested to register the signals
separately. For fkther analysis, one must choose the
data that present signals congruent in the beginning.
Good signals were obtained when using one piece of
the web welded in the head rail, in order to place the
instrumentation diametrically opposite. For this
condition, where little deviation was observed, it was
possible to use the average values.
For composite rails, the welding must be
continuously along the pile to avoid multiple
reflection of the signals. For double rails it is
possible to place the transducers diametrically
opposite in the web. However, for this particular

391
 Hartung, M.; Meier, K. & Rodatz, W. 1992.
Dynamic Pile tests on Sheet Piles, Fourth Int. Con$
on the Application of Stress- Wave Theory on Piles,
The Netherlands. 259-264.

NBR-6122. 1996. Brazilian Pile Foundation

Standard.

Bernardes, G.P. 1989. Dynamic and Static Testing

of large Model Pile in Sand, Doctor Thesis.
Geotechnical Division. The Norwegian Institute of
Technology. Trondheim: Norway.

Benaton Pile Foundation Company 1998. Dynamic

Load test Results, Technical Report 001/98 (in
Portuguese) SBo Paulo.

Figure 6. (a) Force and Velocity traces for composed three steel
rails type TR 37 (b) Capwapc@signal match.

case, the best signal was for the sensors fixed in the
flange.
For pile geometry made with three steel rails,
welded according to Fig. 4b, the signals should also
be registered separately or more pairs of sensors are
necessary for hrther analysis.
One possibility to eliminate any eccentricity of
the ram during itnpact is to use the impact head of
the helmet spherically shaped at the top (Bernardes,
1989).

REFERENCES

Paraiso, S.G. 1990. Steel Piles- Second Part,

Technical Note, CTCEWINI, Construction
Magazine N.2217, SBo Paulo. 21-24 (in Portuguese).

Bergdahl, U. & Moller, B. 1984. Stress Wave

Measurement in Slender Steel Piles Driven by
Pneumatic Hammers and Drop Hammers, Second
Int. Con$ on the Application of Stress- Wave Theory
on Piles, Stockholm: Sweeden. 260-272.

392
 Applicationof Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3

The application of high strain dynamic pile testing to screwed steel piles

J.G.Cannon
Independent Geoscience Pty Limited, Melbourne, Vic., Austrulia

ABSTRACT: Screwed steel piles have been in commercial use in Australia for about 3 years and are becom-
ing a popular foundation support for small to medium commercial structures. Although new to Australia this
type of pile is described in the old British Foundations code CP4 -1954. In Australia they would previously
have been described as micro-piles as they were used for small loadings only but they are now becoming used
for more regular pile applications, with serviceability loads increasing to well over 1OOOkN. They comprise a
relatively small diameter steel tube with one or more relatively large diameter screw flights near the toe and
sometimes at intervals along the shaft. The structural and geotechnical design of these piles is still evolving.
This paper will describe several projects where dynamic pile testing has been successfully used to demon-
strate the capacity of these piles with a range of sizes, lengths and ground conditions. In one case a static load
test was conducted after the dynamic test and provided a close correlation with the dynamic test result in a
Grade A prediction. The general methodology used to conduct the successful tests with the PDNCAPWAP
system will be described

1 SCREWPILES required and, as commonly available equipment is

These piles comprise a steel tube of 325 or 350MPa
grade steel with one or more auger screw flights near
the toe or at intervals along the shaft. Originally the
screw flights were quite “steep” much like a “cfa”
auger piling rig, but more recent designs have used a
much “flatter” screw and it is much wider than
would be used for an auger rig intended for remov-
ing soil. Piles for some projects are galvanized for
corrosion protection. The piles observed by the
Author include a pointed vertical plate at the toe,
which prevents any substantial amount of soil en-
tering the steel tube, hence internal fiiction does not
need to be considered.
Piles have been excavated after installation to in-
spect disturbance of soils along the shaft. Anecdotal
reports fiom independent geotechnical engineers af-
ter inspecting excavated anchors in sandy clay sug-
gest that the auger flight does not significantly dis-
turb the surrounding ground mass. This is also
suggested by high shaft fiiction shown in dynamic
test results as presented in this paper.
The piles are screwed into the ground using a hy-
draulic torque motor attached to earthmoving ma-
chinery, normally being an excavator. No special-
ised purpose-built pile installation machinery is
used, reliability and ease of repair would appear to
the Author to be good.
Although the equipment that is used is commonly
available these piles still appear to be installed by
specialised screw piling contractors. We are aware
of at least 3 companies that specialise in manufac-
ture and installation of screw piles in Australia. As
this pile type is presently used at the low capacity
end of the piling “market,” price competition ap-
pears to be vigorous. A photo of a typical installa-
tion is shown in Figure 1.
Piles currently range in size fiom 76mm outside
diameter x 3.8mm wall thickness with a screw flight
diameter of 350mm and serviceability loads of 50-
1 O O k N to 273mmODx9.3mm wt with a screw flight
diameter of l m for serviceability loads of up to a
claimed 1800kN.
Connections for the pile to the hydraulic motor
and splices to other sections of pile currently com-
prise a bolted external flange butt joint. This has
advantages for dynamic pile testing as the horizontal
flange stiffens the tube against buckling at the top
during the impact of a drop weight. However it is
the Author’s view that this has a disadvantage for
the behaviour of the pile in that if this flange pene-
trates below ground level then an open annulus is

393

Figure 1 - Installation of screw pile
created around the pile that reduces lateral support.
However, in very soft soils it may be an advantage
in reducing negative skin friction. It is understood
that at least one company is developing a new
pressed fitting, which does not extend beyond the
outside diameter of the steel tube. The Author has
not dynamically tested a pile with the new joint.
Concerns have been raised by some clients that,
based on design calculations, the connection be-
tween the screw flight and the pile tube is a weak
point in the system as almost all pile resistance to
vertical load is supposed to be provided by the screw
flight. Concerns include the stiffness of the auger
flight and the strength of the welded connection to
the steel tube. Although these concerns are well
founded the Author’s experience is that CAP WAP
analyses suggest the shaft does supply considerable
resistance through shear fiiction with the soil. The
Author has also not observed any failure of the
screw flight connection during testing, which is gen-
erally to loads of 2 to 2.5 times the serviceability
load plus dynamic effects. Nevertheless, it is under-
stood that this concern is also being addressed with
development of what has been termed an “inter-
brace” connection to increase stiffness of the joint
and the length of weld at the joint. This should also
increase the stiffness of the piling system.
The screw pile contractors all attempt to relate in-
stallation torque to the load resistance of the pile and
adopt torque as the pile acceptance criteria. The
Author is not hlly convinced that the installation
torque can provide a quantitative measure of vertical
load resistance, which is mostly provided by end
394
bearing. It is accepted that the torque provides a
qualitative measure of resistance but the overburden
pressure, plasticity and moisture content of the soil
and the relative proportion of clay in the soil, par-
ticularly at the level of the screw flight, would ap-
pear to also affect torque resistance and thus only
when these conditions are also considered can the
installation torque provide a quantifiable measure of
vertical load resistance. With the present “state-of-
the-art” it is considered that the resistance should be
assessed using geotechnical calculations, with ap-
propriate uncertainty factors, termed “geotechnical
reduction factors” in the current Australian Piling
Code (Reference 2). If higher geotechnical reduc-
tion factors are to be used (ie lower “factors of
safety”) then the Code requires a proportion of piles
at each site to be tested in order to verify load resis-
tance. The Author agrees with this requirement. At
least one of the contractors has chosen to use dy-
namic testing by an independent consultant to pro-
vide this resistance verification.
2 DYNAMIC PILE TESTING
The Author uses the PAK Pile Driving Analyzer
from Pile Dynamics Inc. together with the associated
CAPWAP signal matching software. The method
and current “state-of-the-art” has been described in
Goble et a1 ( I 996).
A special cable drop hammer and frame was
made up by one of the Contractors specifically for
dynamic testing of his screw piles. See Figure 2.
The drop weights comprise 2tonne and 4tOnr1e solid
circular billets of steel of about 350mrn diameter and
2.6m and 5m long respectively. The guide frame
allows for a maximum hammer stroke of about 2m.
The frame is also supported laterally by 4 guy wires
Figure 2 - Test Rig
that are generally tied to adjacent screw piles, being Independent Gaoscience 04-Apr-97

Sydney International Airport, Project: 97037

either production piles or temporary anchors specifi- Pile: SF9 Blow: 1 Data: ReA Foundations
Collected: 03-Apr-97 Operator: Jon Cannon CAPWAP(R) Ver. 1996-2
cally for the test.
CAPWAP FINAL RESULTS
The large weight of the rams compared to the pile Total W A P Capacity: 482.0; along Shaft 182.0; at Toe 300.0 W
weights provides for a long-acting input pulse. This ==J==_PP-=sP_==F=SP--==-~===T=-===~=~==============*~=--=-=-==~~-===========

can be seen in the test results presented below, Soil Dist.

Sgmnt Below
Depth
Below
RU Force
in Pile
Sum
of Ru
Unit Resist.
W.
Smith
Respect to Damping
Quake

where the resistance also acts for a long time com- No. Gage5
m
Grade
m kN
at Ru
IrN kN
Depth
kN/m
Area Factor
kPa slm mm
pared to conventional dynamic tests on piles with 482.0
1 1.1 .6 .1 481.9 .1 .13 .47 .292 8.000
larger shafts. 2
3
2.1
3.2
1.6
2.7
.1
.2
481.8
481.6
.Z
.5
.08
.22
.30
.78
.292
.292
8.000
8.000
Testing has so far been conducted on the steel 4
5
4.3
5.3
3.8
4.8
.9
5.9
480.7
474.8
1.3
7.2
.81
5.56
2.91
19.86
.292
.292
8.000
8.000
tubes only prior to any concrete infill, where used. 6
7
6.4
7.4
5.9
6.9
81.1
4.5
393.7
389.2
88.3
92.8
76.30
4.24
272.49
15.14
.292
.292
8.000
8.000
Owing to the relatively small shaft diameter com- 8 8.5 8.0 89.2 300.0 182.0 83.99 299.97 .292 8.000

pared to the pile capacity it has been generally found Average Skin Values 22.8 22.76 76.49 .292 8.000

that testing is limited by shaft stresses rather than Toe 300.0 2439.02 .079 7.000

geotechnical capacity of the piles. However this is Soil Model Parmtera/Extensions Skin Toe
not always the case. Case Damping Factor .900 .400-Smith Type
unloading Quake ( a of loading quake) 50 100
unloading Level ( % of Ru) 10
Resistance Gap (included in Toe Quake) (mm) .500
Soil Plua Weioht ( kNl .15

3 CASE STUDIES
3.1 Sydney International Airport
This was the first project where the Author had en-
countered screw piles and the design appeared to be
under early development. Several configurations
were tested, being 89x5.5mm tubes with single
350mm screw flights near the toe (working load
1OOkN) and 168x7mm tubes with either two 700mm
screw flights (working load 500kN) or with 5 flights
(working load 600kN). All piles penetrated the
ground about 8m. Ground conditions were moder-
ately dense to dense medium sands.
Figure 3 -Plots for SP9
CAPWAF’ modelling of the piles should not in-
clude the screw flight as taking part in the travel of
stress waves along the pile ie a significant increase
in shaft impedance over a very short distance, of
only a few millimetres, at the screw flight is not re-
quired. As the flights are only a plate attached to the
Figure 4 -Summarv table for SP9
side of the pile the Author does not accept that it
truly is involved in the travel of the stress wave and
we almost never model the flights as anything other
than supplying a point resistance with low stiffness.
Owing to the small shaft diameter conventional
(2Fi2V) gauges were placed individually at 90” in-
tervals around the piles. Gauges were bolted to the
pile using drilled and tapped holes as with conven-
tional testing of larger steel tube piles. Testing was
started with small blows and then larger and larger
blows were used until shaft stresses reached or ap-
proached the rated yield stress of the steel. As the
piles are small diameter and therefore relatively
elastic it was found that a cushion was unnecessary
between the drop weight and the piles. If the piles
had been cut off then an external connection flange
was re-attached prior to testing in order to stiffen the
top of the pile shaft.
The purpose of the testing was proof testing of
working piles rather than as pure research tests. In
all cases the testing demonstrated more than 2 times
the working load. During the testing the piles were
“set” as with testing of conventional piles. The fact
that a sizeable permanent “set” could be achieved
suggested that the testing was mobilising a reasona-
bly high proportion of the available geotechnical re-
sistance. However, the testing was generally
stopped owing to high dynamic stresses in the shaft
rather than reaching the ultimate geotechnical resis-
tance of the piles with very high “sets.”
The piles with 5 flights could not be “set” at all
and they were probably significantly over-designed.
The Author has not seen this design used again.
We CAPWAPed data from all piles. During
CAPWAP analysis we used high “quakes” for both

395
shaft and toe resistance as the shaft resistance was Independent Geoscienoe 27-nay-98

really provided by high level “toes.” The CAPWAP Redcliff Hospital, Project: 98057
Pile: B-19 Blow: 11 Data: steel Foundations
Collected: 27-my-98 Operator: Jon Cannon CAPWAP(R) Ver. 1996-2
models show resistance concentrated at the flight lo- M A P FINAL RESULTS
cations, which seems reasonable. The CAPWAP Total CAPWAP Capacity: 1808.9: along Shaft 1108.5; at Toe 700.4 Ir)(
second toe option was not used to keep compatibility -_===pi ~

Soil Dist. Depth RU Force sun Unit Resist. Smith W a k e

with the multi-flight analyses although this might be Sgmnt Below Below in Pile of Ru W. Respect to Daaping
No. Gages Grade at RU Depth Area Factor
just as valid. The CAPWAP analysis was in other m m kN kN kN kN/m kPa s/m mm
respects normal and we would expect similar accu- 1 .9 .6 49.6
1808.9
1759.3 49.6 105.20 152.91 .146 2.616
racy to conventional analyses. The results show 2
3
1.4
1.9
1.1
1.6
49.6
49.6
1709.7
1660.1
99.2 105.20 152.91
148.8 105.20 152.91
.146
.146
2.616
2.616
considerable resistance and owing to the high ham- 4
5
2.4
2.8
2.1
2.5
900.6
59.2
759.6
700.4
1049.3 1910.27 2776.55
1108.5 125.49 182.40
.146
.146
2.616
2.616
mer/pile weight ratio and the elasticity of the system 6 3.3 3.0 .O 700.4 1108.5 .OO .OO .146 2.616

Average Skin Values 184.8 369.50 569.61 .146 2.616

the resistance “stays on” for a long time. This is
TOe 700.4 1297.10 .007 25.000
shown in the data plots and results table for piles
SP9 and SP23 in Figures 3, and 4. soil node1 Parameters/Bxtensions Skin Toe

Case Damping Factor .740 .022-Smith Type

Unloading Level ( % of Ru) 20
Soil Plug Weight ( kN) .25
3.2 Redcliff Hospital, Brisbane
Figure 6 - Summary table for B 19

which is considered reasonable. The fi-iction pre-

dicted along the shaft away from the flights is also
high at about 15OkPa, which suggests a Meyerhof
~ 0 . 2 5 which
, is also considered reasonable. The
piles were very short for dynamic testing and it was
necessary to use special start-up procedures in order
to by-pass CAP WAP default minimum length limits.
However, other than the abnormal initialisation a
standard CAP WAP modelling procedure was fol-
lowed. We again CAPWAPed data from all piles.
A static load test was conducted on this pile after
the dynamic test. The results of the static test were
obtained some months after the dynamic testing was
reported so the dynamic test was a “Grade A” pre-
Figure 5 - Plots for B19 diction. Both the static and the dynamic test passed
the initial elastic behaviour of the pile and the
Ground conditions comprised completely weathered “break“ in the dynamic test, which uses an elastic-
sandstone. which was essentially very stiff-hard clayey plastic model, appears to model the change in slope
sand or sandy clay with reported undrained shear strength of the static test with an acceptable accuracy. The
of up to 600-700kPa. This project was much more
recent and the piles were much larger and more
heavily loaded than previous tests. Piles were
2190DxSmm wt with 2 flights near the toe. The
flights were nominally 0.85m diameter but were ac-
tually made from square plates 0.76ni across with
rounded corners and thus the area would be slightly
less than a 0.85m circle. Note that the flight diame-
ters were about 4 times the shaft diameter and it is
questionable that they were stiff enough to truly be
working at the outer edge. The pile de-
sigrdconstructor is aware of this possibility and is
developing a stiffer flight as described earlier. Re-
quired working loads were 850 to lOOOkN and it was
necessary to demonstrate 2 times the working load.
The CAPWAP plots and summary table for pile
B19 are shown in Figures 5 and 6. Resistance was
again modelled with concentrations at the flights, Figure 7 Statiddynamic correlation for B19

396
correlation is shown in Figure 7. The static load test Independent Geoscience 06-Oct-99

Redlands Hater Aoapital, Project: 200011

was conducted in accordance with the Australian Pile: 109 extension Blow: 1 Data: Steel Foundations 2t
Collected: 05-0ct-99 operator: Jon cannon CAPWAP(R) Ver. 1996-2
Piling Code AS2 159- 1995 to 1.5 times serviceability CAPWAP FINAL RESULTS
load with measurement of deflection vs time at vari- Total CAPWAP Capacity: 425.0; along Shaft 222.0; at Toe 203.0 kN
===C=_==_P=======rJ=F======rFPL=li====IJ====-====
ous loads. These time dependent measurements
Soil Dist. Depth RU Force Sum Unit Resist. Smith Quake
have been removed from the static results presented Sgmnt
No.
Below
Gages
Below
Grade
in Pile
at RU
of RU w. Respect t o Damping
Depth Area Factor
here. The dynamic test mobilised considerably more m m kN kN kN kN/m kPa s/m mm

resistance and the test was again limited only by 1 1.9 1.6 32.5
425.0
392.5 32.5 34.85 97.35 ,185 1.200
2 2.8 2.5 34.6 357.9 67.1 37.04 103.47 .1851.200
stresses in the pile shaft. 3 3.7 3.4 46.0 311.9 113.1 49.32 137.76 .185 1.200
4 4.7 4.4 55.6 256.2 168.8 59.61 166.51 .I85 1.200
5 5.6 5.3 53.2 203.0 222.0 57.04 159.32 .185 1.200

Average Skin Values 44.4 41.89 132.88 ,185 1.200

3.3 Redlands Muier Hospital, Brisbane Toe 203.0 1276.73 .104 39.000

Soil Model Parameters/Extenaions Skin Toe

This was a recent project and testing had become Case Damping Factor .500 .258-Smith Type
Unloading Level ( 8 of nu) 60
reasonably procedural. Two pile sizes were tested. <nil D l n n Wainhr r kNI .lO

The smaller piles were 89x5.5mm with a single

auger flight that was basically a 350mm square with
chamfered corners. Larger piles 1 14x6mm were
also tested. These were also fitted with a single
flight that comprised a chamfered square but the
chamfers were larger such that the flight was essen-
tially a 450mm diameter circle. Pile penetrations
varied from 2.2 to 5ni. Ground conditions com-
prised stiff clays over stiff to hard gravely clays.
The prqject engineer required testing to demonstrate
at least 2.5 times the “working” or serviceability
load. This demonstrated load varied between 250
and 625kN. We tested 8 piles, which comprised
15% of the total piles. It is the Author’s opinion that
with this proportion of testing the required “factor of
safety” of 2.5 is unnecessarily high. In accordance
with the Australian Piling Code, which is a “partial
factor” Code. it would be possible to adopt a geo-
technical reduction factor of 0.8. For a typical pile
this would suggest an overall “factor of safety” over
serviceability of about 1.8.
Figure 9 - Summary Table for 109 extension
In this case we did not conduct CAPWAP analy-
sis of data from every test. CAPWAPs were initially
conducted on data fiom those piles with a mobilised
resistance close to the requirement, this being for
about half of the tests. A very close correlation was
found between the Case method field estimates and
the CAPWAP results, and because the other tests
had proved 120 to 170% of the required ultimate re-
sistance using the field “Case Method” hrther
analysis was considered unnecessary, especially
with regard for the already high “factor of safety.”
One of the most highly loaded test piles was
tested at 2 different penetrations with the pile being
extended by 1.5m between tests. The torque resis-
tance of this pile was quite high during installation at
both levels but during testing the measured set was
very high (32 and 37mndbl) and the mobilised re-
sistance was also correspondingly low at both pene-
trations. This is an example of why installation
torque does not give a good indication of pile capac-
ity. Eventually this pile was augmented with addi-
tional piles. CAPWAP results for this pile are
shown in figures 8 and 9.

4 CONCLUSIONS

Dynamic testing can be conducted quickly and accu-

Figure 8 - Plots for 109 extension
rately on this unusual new pile type. Testing can be
conducted using similar procedures to dynamic
testing of more common pile types with small modi-
fications to allow for the shorter pile length, if nec-
essary. Resistance concentrated at the flight loca-
tions provides a realistic model of the pile.
The use of testing to verify pile performance
should be conducted on every project, as the instal-
lation torque does not give a reliable quantitative es-
timate of pile resistance. Testing on each project is

397
already common practice for driven piles where
measurement of pile set appears to give a better
quantitative indication of pile resistance than the in-
stallation torque applied to screw piles.
A Grade A dynamic prediction has demonstrated
an excellent correlation of the dynamic test results
with conventional static load test results.

REFERENCES

G Goble + G Likins ( I 996) “On the Application of PDA

Dynamic Pile testing” “Proceedings of Fifth Interna-
tional Conference on the Application of Stress Wave
Theory to Piles” Orlando, Florida USA. September,
Townsend, Hussein, McVay Editors. pp263-273

AS 2 195-1995 “Piling - Design and Installation” Stan-

dards Association of Australia

398
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Case study on the application of high strain dynamic pile testing

to non-uniform bored piles

J.G.Cannon
Independent Geoscience Ply Limited, Melbourne, Vic., Australia

ABSTRACT: Dynamic pile testing is used frequently to prove the performance of driven pre-formed piles
in Australia. It is either specified by the designer or is offered as an alternative to conventional static testing
by piling contractors. However, it’s use on cast in place piles is much less frequent. This paper describes a
project at Noosa Heads in Queensland, Australia where an excellent “Grade A ’ correlation has demonstrated
that dynamic testing can provide a good prediction of the load vs displacement behaviour of cast in place piles
even when the pile shaft is not as designed.
The static test pile had the highest over-consumption of concrete at the site, with more than 2 times the
design quantity of concrete used during construction. A non-uniform shaft is known to make dynamic testing
more difficult and probably less accurate. Nevertheless a very good correlation was obtained between the
static test and the dynamic test results for both overall mobilized resistance and the stiffness response of the
pile.
The static test included measurements that demonstrate potential problems with static testing and these
will also be discussed.

1 THESITE constructed using a continuous flight auger (cfa) that

The site is on Hastings St, which is the coinmer-
cial/tourist centre of Noosa Heads in Queensland,
Australia. This is located only a few metres behind
a popular surf beach and is immediately adjacent to
a tidal inlet and small river. A geotechnical investi-
gation of the site had been conducted and 4 borehole
logs were provided that described subsoil conditions.
The logs suggest that the site is underlain by coastal
sand dune or “beach sand” material to a depth of
about 10m and this sand is underlain by very stiff to
hard clays with SPT results generally 35 but some as
low as 18. The sand is loose to moderately dense
with standard penetration test results generally in the
range 10 to 20 but with some higher and some lower
measurements. The SPT results in the sand did not
necessarily increase steadily with depth.
2 THEPILES
The foundation contract was let as a design and con-
struct package and the contractor adopted 6OO1nln
nominal diameter “cfa” cast in place concrete piles
founded at a depth of about loin (ie entirely within
the near surface sand with the toe being influenced
by the underlying hard clays). These piles were
is drilled into the ground to a pre-selected level or
depth and then a cement grout or, as in this case,
high slump concrete is injected down the hollow
core of the auger as the auger is withdrawn without
rotation. The rig used on this project allowed for
monitoring of concrete volume and pressure
throughout construction of each pile. This permits
the pile constructor to assess whether the pile is
“consuming” more or less concrete than would be
expected for the nominal shaft diameter. When the
pile concreting is completed a reinforcing cage is
lowered with vibration into the high slump concrete.
The piling contractor adopted a high “geotech-
nical reduction factor” (ie low factor of safety) i n the
design and in accordance with the Australian Piling
Code (see references) it was necessary to demon-
strate the load vs resistance behaviour of the piles.
He decided to adopt a single static load test and 4
high strain dynamic tests. One of the dynamic tests
was conducted on the static test pile to establish that
the dynamic testing would provide a good prediction
of a static load test.
Many of the piles for this project “consumed”
more concrete than would be expected for the nomi-
nal shaft diameter but the greatest “over-
consumption” was during construction of pile 63 and
this was selected for both static and dynamic testing.

399
The over-consumption on this pile was 105%, ie
more than 2 times the required volume of concrete
was used during construction. The contractor’s
equipment provided for measuring pressure and vol-
ume throughout construction of the pile so the pile
profile could be estimated. The contractors record
for pile 68 is shown in Figure 1 . Most of the extra
concrete consumption is shown as a cone between 5
and IOm depth.
3 STATIC PILE TESTING
A static load test had been conducted on pile 68
prior to the dynamic pile testing. The contractor was
careful to avoid the dynamic testing consultant be-
coming aware of the static test results before the dy-
namic test results were reported. The static test was
conducted with several cycles in accordance with the
Australian Piling Code. Applied load was measured
using the jack pressure only. There are shortcom-
ings to this system that are described below. Dis-
placement was measured using 3 dial gauges. A
check of pile displacement was also taken using a
level survey. Load was applied by jacking against a
reaction beam and displacement of this beam was
also measured by level survey.
Figure 1 - Construction record for dynanuc/static test pile
400
4 DYNAMIC PILE TESTING
The test piles were cast above ground level inside a
steel sleeve of about 4mm wall thickness and about
the same diameter as the pile for about 2 5 pile di-
ameters above the surrounding ground This was
done at the same time as casting the pile or as soon
as possible after casting the pile in order to keep the
concrete for the extension of the same strength and
age as the remainder of the shaft After the concrete
had hardened the bottom half of the steel sleeve was
removed This leaves a substantial steel collar at the
top to reinforce the pile top during impacts of a drop
weight and allows the test equipment to be attached
at a level where there is a regular smooth surface
with no additional impedance that might interfere
with strain measurements The location of the test
gauges had a diameter close to that of the pile shaft
and had similar reinforcing
The Author uses the PAK model Pile Driving
Analyzer from Pile Dynamics Inc together with the
associated CAPWAP signal matching software The
method and current “state-of-the-art” has been de-
scribed in Goble et al (1996) The option to test
with 4 strain gauges was not adopted for piles of this
size
 The contractor supplied a “Hydroquip” HQ5 hy-
draulic piling hammer to strike the piles. Some re-
building of the hammer’s hydraulic valving had been
conducted to increase energy transfer efficiency.
Highest energy transfer efficiency during this testing
was 76%’ which we consider to be high for a hy-
draulic hammer with a Stonne rain striking a solid
concrete pile of this diameter.
Testing generally commenced with one or two
small blows (0.5m stroke) to ensure haininer align-
ment was satisfactory and then two or three blows of
high energy (1.2m stroke) were applied to gather test
data for later analysis with CAPWAP.
During CAPWAP analysis the pile was modeled
using the construction record but some additional
enlargement of the shaft was required near the top.
The Author considers the additional pressure created
by the shaft extension after the contractors monitor-
ing record was completed justifies this. The model
pile volume in the CAPWAP model was very close
indeed to the recorded volume of 205% of the nomi-
nal design
5 RESULTS
The static results are summarized in Figure 2,
below together with the CAPWAP load vs deflection
prediction plotted on the same axes.
Figure 2 - Hastings St Load vs Deflcctioii
It can be seen that in the static test when the pile
reaches a displacement of slightly less than 151nni
the inferred load increases but there is no corre-
sponding deflection of the pile. The Author consid-
ers that this is impossible and that there must have
been an error in the test measurements The most
likely error was that the jack reached the end of its
travel or jammed such that although there was an in-
crease in jack pressure and hence inferred load on
the pile, in reality load did not increase and the pile
consequently did not deflect. The contractor’s per-
40 1
sonnel that conducted the test claim this was not the
case and consider that the jack did not reach the end
of it’s travel. However the measurements of the re-
action beam, which was also deflecting elastically,
show the same behaviour, with almost no deflection
during the last load application cycle. This is shown
in figure 3. The Author considers that the maximum
load applied to the pile did not exceed 1700kN.
Figure 3 - Hastings St Static Test Reaction Beam
As the static test was conducted On this pile prior
to the dynamic test it is relevant to plot the dynamic
test as an additional cycle to the static test The re-
sults of the other tests at this site show a lower
“break-point” in the load vs deflection behaviour
and the Author considers this to be related to the
loading history of the piles The “cfa” constiuction
is a non-displacement construction method Owing
to the stress relief that occurs durinz construction it
would appear that these piles deflect more during
initial loading Subsequent loading cycles appear to
behave with increased stiffness up to the point of
previous niaxiinuin loading This behaviour has also
been noted by the Author at other projects with
similar piles in sand ground conditions If deflection
is a critical acceptance criteria for- this pile type i n
sand ground conditions then it inay well be neces-
sary to “preload” the piles by “driving” them after
construction
If the inaxiinuin applied load during the static
load test was 1700kN this correlates well with the
“break point’ at about 1800kN in the CAPWAP load
vs deflection prediction for this pile The un-
load/reload stiffness shown i n each of the load cy-
cles of the static test is also of interest as this corre-
lates quite well with the initial loading stiffness
shown in the dynamic test results
The stiffness of the static test on initial loading
during each of the cycles also correlates reasonably
well with the dynamic test prediction after the
“break point.” The Author considers the dynamic
test was the first time the pile experienced sufficient
deflection to generate a resistance of more than
about 1700kN and so after this point the lower initial
loading stiffness is valid.
The load vs deflection behaviour shown in other
dynamic tests at the site were similar, with similar
initial loading stiffness and stiffness after the “break
point” but they showed a much lower “break point”
and a typical example is shown in Figure 4. The
Author considers this is because these other piles
Accuracy of the results for bored piles appears
similar to driven pre-formed piles provided the pile
shaft can be realistically modeled. This requires
some knowledge of the shape of the pile shaft.
The load-displacement behaviour of “cfa” piles in
sand appears to be related to previous load history
with initial loading to any level of load being less
stiff than reloading. Consequently dynamic testing
should be conducted with as few blows as possible if
it is necessary to predict initial load stiffness.
Designers should be aware of low initial load
stiffness. If displacement of this pile type is critical
then pre-loading either statically or by “driving”
should be considered.

7 REFERENCES
G Goble + G Likins (1996) “On the Application of PDA
Dynamic Pile testing” “Proceedings of Fifth Intenia-
tional Confereiicc 011 the Application of Stress Wave
Theory to Piles” Orlaiido, Florida USA September,
Townsend, Hussein, McVay Editors pp263-273

AS2 195-1995 “Piling - Desigii and Installation. Stan-

dards Association of Australia.

Figure 4 - Coiiiparisoii Pile 68 Static vs Pile 2 18 Dyiiaiiiic

have not experienced high loading and deflection be-

fore the test as did the static test pile. However the
initial stiffness in these other dynamic tests is still
maintained to higher loads than shown in the first
cycle of the static test. The Author considers the
small blows applied to the pile at the start of each
test cause this and owing to the previous loading by
the small blows the dynamic test results should be
plotted some distance to the right. The pile with
more preliminary blows prior to the “test” blow also
showed a higher “break point” in the prediction of
static load vs deflection, however there were insuffi-
cient tests to say that this behaviour has been
proven. Further analysis of the test data may pro-
vide more information on this behaviour. In par-
ticular it may be worthwhile analyzing several blows
from the one test to assess the change in “break
point.” It is suggested that if it is hoped to avoid this
behaviour that the number of small blows applied
before the hll test blows should be minimized. This
would appear to minimize the “preliminary” loading
of the pile and thus provide the best prediction of the
first loading deflection behaviour of a cast in place
pile.

6 CONCLUSIONS

Dynamic testing appears to be just as valid for

bored “cfa” piles as for driven pre-formed piles.

402
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

igh capacity dynamic load tests for bored piles in Sydney shale

David J. Klingberg & Phi1 Mackenzie

Wagstaff Piling Pty Limited, Brisbane, Qld., Austruliu

ABSTRACT: The installation of bored piles is undergoing some scrutiny with respect to acceptable
parameters to be used in the design for shaft resistance and end bearing. A dynamic test pile program was
designed for several shallow bored piles socketed into Sydney shale. A twenty (20) tonne drop hammer was
used for the dynamic testing. The results were compared to locally accepted design parameters with the test
results indicating a significant saving to be made with the adoption of the test pile results.

1 INTRODUCTION The remainder of this paper presents some of the

specifics of the test pile programme, summarises the
The design and installation of bored piles in and results of the dynamic testing and CAPWAP@
around Sydney, Australia is a highly competitive analyses completed for the test piles and provides a
market. The development of the highly refined comparison with some published design parameters
designs needed for such a competitive environment for bored piles in Sydney shale.
would prove difficult without the aid of test pile
programmes to allow for the confirmation of the
design parameters. With the knowledge and 2 SITE DESCRIPTION AND GEOLOGY
confidence that normal construction techniques
would provide pile capacities well above the The site is located in the north western suburbs of
‘acceptable’ values given from well-used publications Sydney, Australia. Reference to the Sydney
for the Sydney area, it was decided to instigate a test 1: 100,000 geological series sheet and fieldwork
pile program to more accurately determine the design results indicate that the site is underlain by stiff to
parameters to be allowed in the geotechnical design hard grey and brown clay over shale and fine
of the bored piles. sandstone-siltstone laminite.
The test pile program consisted of the installation The design values adopted on this project as
of three ( 3 ) No. test piles that were subjected to specified by the geotechnical report are given in
dynamic testing using a twenty (20) tonne, free-fall Table 1. The test piles were founded in the Class I1
drop hammer. Two (2) of the piles were constructed and I11 shales.
with a styrofoam base in order to determine only the The values in Table 1 were determined in
shaft resistance component of the pile capacity (ie. accordance with the recommendations of Pells et a1
there was little or low base resistance due to the (1978). This is typical of the majority of piling
presence of the styrofoam at the pile toe). In projects within the Sydney region.
addition, various construction techniques were used
to determine the difference (if any) in the
performance of the shaft resistance - one of the test Table 1. Geotechnical report recommended
piles was constructed using a grooving tool to parameters.
roughen the pile-soil interface along the shaft of the Rock Class Bearing Skin
pile, whilst the other two shafts remained ungrooved. Strength Pressure Friction
All test piles were sacrificial piles and were tested to (@a) (@a)
structural ‘failure’ of the pile head under the impact Verylow IV 1,000 100
of the testing hammer. Low I11 2,000 200
Medium I1 4,000 400

403
3 TEST PILE PROGRAMME Table 3. Dynamic testing results summary.
Bile Stroke Set TC EMX RMX
Test piles TP1, TP2 and TP3 were installed as No. (mm) (mm) (mm) (kNm) (kN)
sacrificial piles for the project. The piles were TP1 715 1.0 8.0 62.1 10,960
500mm diameter (nominal) bored piles with total TP2 320 0.0 6.1 32.1 9,280
lengths between 6.3 and 7.1 metres using 50MPa TP3 1,070 1.0 10.2 58.7 8,960
concrete. The installation details and techniques Set = pile set in mm/blow
used for each test pile are summarised in Table 2. TC = temporary compression in mm
EMX = max. energy transferred to pile head in kNm
Table 2. Test pile installation summary. RMX = PDA capacity estimate in kN
Pile Pen. Pile Grooved Pile
WO. (m) Size Shaft Toe
(mm) YiN Type”
Table 4. CAPWAP’ analysis summary.
TP1 5.3 500 Y Foam
TP2 5.5 500 N Foam Pile Shaft End Total
TP3 4.8 500 N Standard No. Resistance Resistance Resistance
* - The pile toe type indicates the method of (W (W (W
placement of concrete at the pile toe. ‘Foam’ TP1 11,846 1 11,847
TP2 10,484 11 10,495
indicates a 200mm thick styrofoam block was
TP3 6,804 1,201 8,005
inserted at the pile toe prior to the placement of
concrete to provide a pile with negligible toe
resistance. ‘Standard’ indicates the concrete was
placed in the usual manner at the pile toe to provide a Table 5 . CAPWAP’ Unit shaft resistance summary.
pile with ‘normal’ toe resistance. Depth CAPWAP@Unit Shaft Resistance
To enable the three test piles to be subjected to (m)
. _ (Ha)
dynamic testing, special heads were cast onto the TP 1 TP2 TP3
piles to contain the impact and bursting stresses from 1.o 585 685 100
the impact of the twenty (20) tonne hammer. The 2.0 1,150 1,100 450
typical dynamic test data is shown in Figure 1. The 3.0 1,400 1,275 825
results of the dynamic testing are summarised in 4.0 1,500 1,275 1,175
Table 3. 5 .O 2,100 1,275 2,100

PllETEST PDlPLE DRNNGANALSER

TP1-RS3 permanent pile displacements achieved during the
20T DROP M E R dynamic testing were very small and would not have
12500 been sufficient to mobilise the ultimate capacity of
the pile. The under mobilisation of capacity during a

kN ~

30mS
0;- ,2;
../ \ 4:-,

c?
.* -,,\, ,,,,I-‘
RMX 1 1 0 1 0 ~
EMXBlWWm
DMX 9 o m
WU2 3430W

_--__ 50 ;“s
dynamic test is normally related to the end resistance
of the pile, but in this instance it may also be
applicable to the shaft resistance and that the unit
shaft resistance values given in Table 5 may be
conservative.
All of the test piles were subjected to fiu-ther
analysis using the Case pile wave analysis program,
CAPWAP@. The results of these analyses are
summarised in Table 4. The unit shaft resistance
values determined from the CAPWAP@analyses are
also summarised in Table 5 .

4 COMPARISON WITH RECOMMENDED

Figure 1. Typical dynamic test data.
It should be noted that the data from Table 3
suggests that ultimate pile capacity has not been k l l y
mobilised during the dynamic testing. That is, the
DESIGN PARAMETERS
The design of bored piles in shale and sandstone in
the Sydney region has traditionally been in
accordance with the parameters as given by Pells et
a1 (1978). The parameters have recently been
updated @‘esl et al, 1998) to include ultimate design

404
values so that limit state theories can be adopted. parameters by a factor of approximately 4.0 to 5.5.
These values have been summarised in Table 6. The large difference in pile capacities may have very
significant cost ramifications for even the smallest
piled foundation.
Table 6. Design parameters summary.
Class Typical Ultimate' CApWAp@ Ratio
E Shaft Shaft 5 CONCLUSION
(MPa) Adhesion Adhesion
(Ha) (Ha) The test pile programme, using the techniques of
I >2000 1,000 nfa nfa dynamic testing methods, has shown that the use of
traditional bored pile design parameters for Sydney
I1 700 to 600 to 1,275to 2.0to shale may provide conservative designs by factors
2000 1,000 2,100 3.5 ranging from approximately 1.3 to 4.3. The benefits
I11 200 to 350 to 600 450 to 1.3 to of the test pile programme were highlighted further
1200 1,500 4.3 when compared to the design parameters
recommended in the geotechnical site investigation
IV 100 to 150 da nfa report where the factor ranged from 4.0 to 5.5.
500 These factors were achieved in shaR resistance only
V 50to300 50to 100 nfa nfa as the testing programme was not designed to
# - ultimate shaft adhesion after Pells et al, 1998. determine ultimate base resistances.
However, it should be noted that the results
achieved in this case study were the result of good
The data from Table 6 clearly shows that the construction control and installation technique and
results from the dynamic tests suggest the traditional that lesser values may be applicable when poor
values of unit shaft adhesion for the Sydney shales construction control and/or installation technique are
are conservative with factors ranging from evident. It is imperative that the designer take these
approximately 1.3 to 4.3. It is, however, noted that factors into consideration when attempting to employ
the values obtained are in part a result of good the indicated values from this study.
construction control and installation technique and The case study has also shown that, given good
that lesser values may be applicable when poor construction techniques, high shaft adhesion values
construction control or installation technique are are obtainable that exceed the 'normal' design
evident. parameters for the shales found in the Sydney region.
The difference between the dynamic test results However, it should also be noted that it is probable
and the recommended design parameters can be the parameters determined in this case study are
highlighted with a comparison of design capacities lower bound solutions to the ultimate capacities due
from each method. These have been provided in an to the inability of the test hammer to fully mobilise
approximate and comparative form in Table 7. the pile capacity without causing structural distress
It should be noted that the base capacities have not to the pile top.
been included for the comparison as given in Table 7
as the testing programme was not designed to
mobilise ultimate base resistance. REFERENCES
The data from table 7 indicates that the shaft
resistance values obtained fiom the testing CAPWAP Manual (1996). Pile Dynamics Inc.
programme exceeded the geotechnical recommended Cleveland, Ohio.
Pells, P.J.N., Douglas, D.J., Rodway, B., Thorne,
C.P. and McMahon, B.J. 1978. Design 'ioadings
Table 7. Design capacity summary. for foundations on shale and sandstone in the
Depth Geotechnical CAPWAP Sydney region. Australian Geornechanics
(m) Report Capacities" Journal, G8: 31-39.
Capacities* (W Walker, B.F. and Pells, P.J.N. 1998. The
(W construction of bored piles socketed into shale
TP1 TP2 TP3 and sandstone. Mini symposizim on receiif
1 155 400 415 125 developments ill piling practice in Sydney.
2 3 15 1,475 1,840 560 Australian Geomechanics Society. Aitpst I998.
3 3 15 1,895 1,840 1,360 Presented by Pro$ H.G. Poulos.
4 3 15 2,290 2,120 1,360
5 630 3,390 2,135 3,400
Total 1,730 9,450 8,350 6,805

405
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Predicting uplift deflection from dynamic pile testing

W.G.Chambers & D. J. Klingberg

Wagstaff Piling Pty Limited, Brisbane, Qld., Australia

ABSTRACT: Predicting pile uplift deflection from dynamic load testing on concrete and steel piles is very
difficult due to the variable nature of the soils encountered at any site. Two piles, one a 275 mm square
reinforced concrete pile and the other a 406 mm diameter (8.8 mm wall thick) steel pipe pile, were tested
dynamically with the Pile Driving Analyzer" (PDA) at the end of initial driving and on restrike. These piles
were then tested statically to determine pile uplift deflection.

For a successfd pile model behaviour response from CAPWAP@analysis, the calculated static deflection
from the dynamic test should match the field measured static deflection. This paper reviews techniques applied
during CAPWAP@analysis in an attempt to closely match the field measured static deflections in both piles.

1 INTRODUCTION High strain dynamic pile load testing utilising the

The test piles were to be subjected to static and
dynamic load testing to determine load carrying
capacity and load deflection characteristics of the
piles. Two test piles were installed to determine the
uplift capacity of piles driven for a project near
Cairns in northern Queensland, Australia. Test pile 1
(TP1) was a 275 mm square reinforced concrete pile,
12.0 metres long with full length reinforcement
comprising 4Y28 (i.e. 4 bars of 28 mm nominal
diameter steel of 400 MPa yield strength). Test pile
2 (TP2) was a 406 mm outside diameter steel pipe
pile (driven open-ended) with wall thickness of 8.8
rnm and a length of 12.8 metres.
Pile design methods (Poulos & Davis, 1980)
suggest various factors be applied to the shaft
resistance values from compression test results for
the determination of upliR capacity of piles. It is
generally accepted that a reduction factor between
0.5 and 1.0 be used on the shaft resistance for an
estimated pile uplift capacity. If a typical value of
0.8 is used on the shaft resistance to determine uplift
capacity, a 'Class A' prediction of deflection from
CAPWAP@ analysis can be made at various load
increments (i.e. deflection predictions completed
prior to any static tests being performed).
Pile Driving Analyzer" (PDA) is based on the
simplified Case Method for the determination of load
carrying capacity (Rausche et al. 1985). Field data
obtained is normally analysed by the CAse Pile Wave
Analysis Program, CAPWAP@(Rausche et al. 1972).
2 GEOLOGY
The site is located some 5 km south of the centre of
Cairns in Queensland, Australia. Reference to the
Queensland Geology (Scale 1:2,500,000) indicates
the site is underlain by Quaternary alluvials and
lacustrine deposits. These comprise interbedded
layers of sandy clay, sand and clay. The general
subsurface condition consists of stiff to hard sandy
clay and clay to 17.7 metres. A 1.0 metre thick sand
lens is located approximately 11.0 metres below
ground level.
3 HIGH STRAIN DYNAMIC TESTS
The piles were subjected to high strain testing using
the PDA at the end of initial driving and restrike one
day later. The piles were installed using a Banut
piling rig operating a six (6.0) tonne hydraulic drop
hammer.

407
3.1 Test Pile I
The PDA test results for TP1 are summarised in
Table 1 and the CAPWAP@'analyses results are
summarised in Table 2.
where: EMX - maximum energy transferred
RMX - PDA capacity estimate
From the restrike CAPWAP@analysis, a 'Class A'
prediction of deflection was determined assuming
80% of the shaft resistance was available for uplift
capacity. The predicted pile head deflections are
summarised in Table 3.
Table 3 - Summary of Predicted Pile Head
Deflections - TP1
3.2 Test Pile 2
The PDA test results for TP2 are summarised in
Table 4 and the CAPWAP@'analysis results are
summarised in Table 5.
where: EMX - maximum energy transferred
RMX - PDA capacity estimate
408
Total Resistance
As for TP1, if it is assumed that 80% of the shaft
resistance as determined from CAPWAP@'analysis
was available for uplift capacity, a 'Class A'
deflection prediction can be calculated. The
predicted pile head deflections are summarised in
Table 6.
Table 6 - Summary of Predicted Pile Head
Deflections - TP2
*It was noted that the CAPWAP@'analysis indicated
that TP2 would not attain the maximum static test
load if 80 % of the shaR resistance was used as the
estimated tension capacity.
4 STATIC LOAD TEST
The static load tests were conducted in accordance
with the Australian Standard@'Piling Code AS 2159 -
1995. The tests performed were incremental
sustained load tests in uplift. The tests were carried
out to a maximum test load of 150% of the Strength
Limit State Load, N* (Design Action Effect) which
was approximately equivalent to a maximum test
load of twice the pile 'working load'.
4.1 Test Pile I
The maximum static test load for TP1 was -951 kN.
This test was conducted some three (3) days after
initial pile installation and it was anticipated that the
effects of pile-soil set-up would influence the
comparison of the dynamic and static load test data.
The results, together with the 'Class A' predictions,
are summarised in Table 7.
4.2 Test Pile 2
The maximum static test load for TP2 was -702 kN.
This test was conducted some five (5) days after
Table 7 - Static Load Test Results - TP1 1
reduction in effective pile stiffness due to loss of
concrete tensile capacity). The maximum test load
A’ Prediction corresponded to a reinforcement stress of
approximately 385 Mpa, which is well in excess of
the stress level that would cause concrete cracking
-235 1.36 0.78 I (i.e. approximately 130 Mpa).

If the pile-soil set-up effects are ignored, hrther

analysis demonstrated that if a reduction in pile
modulus of 80 % is applied to the CAPWAP@model,
excellent agreement is obtained with the static load
uplift test results. A comparison of the revised and
‘Class A’ predictions with the static load test results
are presented in Table 9 and shown graphically in
initial pile installation and again it was expected that Figure 1.
pile-soil set-up effects would influence the
static-dynamic comparison. The results, together Table 9 - Summary of Static Load rest Results
with the ‘Class A’ predictions, are summarised in ncluding Revised CAPWAP@Deflectio s - TP1
Table 8. I CAPWAP@analvsis Results
Revised
able 8 - Static Load Test Results - TP2 “lass A’ Static Load
Deflection with Deflection
Load Predicted
reduced
(kN) Deflection Modulus (mm)
(mm) (mm) \ I

-70 I 0.12 I 0.59 I -23 5 0.78 2.21 1.36

-140 I 0.24 I 1.20 I -476 1.58
I
4.48 I
4.26
-234 I 0.44 I 2.00 I -571 I 1.89 1 5.53 5.36
-666 I 2.21 1
~ 6.66 6.46
-761 I 2.54 I 7.87 7.57
-856 I 2.88 I 9.19 8.68
-951 I 3.25 I 10.66 - 9.92 _i

*It was noted that the CAPWAp@analysis indicated

that TP2 would not attain the maximum static test Test Pile 1
load if 80 % of the shaft resistance was used as the Precast Concrete Pile - 275 rnm.
estimated tension capacity.
40
h

E
5 COMPARISON OF RESULTS -
E
30
,g ClassA Static
5.1 Test Pile I g 20 + +
Revised
a +
The static load uplift test results and the ‘Class A’ 1 I0
prediction of deflection from the CMWAP@analysis -
al
B
were not in close agreement although the static load 0
test confirmed that pile failure did not occur at the 0 -236 -476 -571 -666 -761 -856 -951
maximum test load. The maximum uplift pile stress Uplift Load (kN)
was very high and the test results indicated
reasonable agreement between the CAPWAP@’
prediction and the static load test up to Figure 1 - Chart of Static Load Test Results
approximately -200 kN. However, at higher loads including ‘Class A’ and Revised CAPWAP@
the correlation deteriorated. It is postulated that the Deflections - TP1
difference between the predicted and measured
deflection was due to the cracking of the concrete
section and effective loss of pile modulus (i.e.

409
5.2 Test Pile 2

The static load uplift test results show that the ‘Class
A’ predictions of deflection from CAPWAP@analysis
were conservative. This result was expected due to
the subsurface conditions encountered at this site and
the timing of the static load test. The static load
uplift test was performed some five (5) days after
initial pile installation and the dynamic tests were
performed at the end of driving and only one (1) day
after installation. The additional set-up time between
the static load uplift test and the initial driving is
likely to account for the difference in the settlement
predictions.
The static load uplift test for TP2 was carried out
to a maximum test load of -702 kN. If it is assumed
that 100 % of the shaft resistance was available as Figure 2 - Chart of Static Load Test Results
tension capacity, the predicted pile head deflections including ‘Class A’ and Revised CAPWAF@
at -562 and -702 kN were 4.80 and 6.11 mm, Deflections - TP2
respectively. A summary of the revised settlement
predictions are presented in Table 10 and graphically dynamic testing and CAPWAP@analyses provided a
in Figure 2. conservative estimate of the pile uplift capacities.
That is, the test piles did not fail at the predicted
maximum uplift capacities as indicated by the
Table 10 - Summary of Static Load Test Results CAPWAP@analyses. This was mostly attributed to
the time delay between the dynamic and static load
tests and the effects of pile-soil set-up.

The results also indicated that where the pile

impedance changes were not induced during the
static load test, CAPWAP@ analysis provided a
conservative prediction of pile head deflection under
uplift loads (i.e. steel pipe pile, TP2). However, it
was shown that for TP1 (reinforced concrete pile)
the effects of cracking under the uplift loads
significantly influenced the pile head deflections and
that a reduction in pile impedance of 80 % gave
excellent agreement with the static load uplift test
results.

7 REFERENCES

Poulos, H.G. & Davis, E.H. (1980) Pile Foundation

6 CONCLUSIONS
Test piles 1 and 2 (TPl and TP2) were subjected to a
static and dynamic load test programme to determine
the uplift and load deflection characteristics for the
two piles (i.e. reinforced concrete pile and steel pipe
pile). The test programme was also used to
determine the reduction factor to be applied to the
shaft resistance (as calculated from a dynamic
compression test) in estimating the uplift capacity.
The static load uplift tests indicated that the use of
Analysis and Design. John Wiley and Sons.
Brisbane.
Rausche, F., Goble, G.G. and Likins, G.E. (1985)
Dynamic Determination of Pile Capacity. Journal
of the Geotechnicai Engineering Division, ASCE,
March Vol. 111, No. 3, Pages 367 - 383.
Rausche, F., Moses, F. and Goble, G.G. (1972) Soil
Resistance Predictions from Pile Dynamics. Journal
of the Soil Mechanics and Foundation Division,
ASCE, Sept. Vol. 98, No. SM9, Pages 917 - 937.

41 0
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Applicability of dynamic load test on a toe improved pile

J. Sakimoto & N. Kita

Obuyashi Corporation, Tokyo,Japan
S. Nishimura & TTakeda
Fugro Geoscience Company Limited, Tokyo,Japan

ABSTRACT: Many dynamic load tests have been carried out to validate the bearing capacity of pile in Japan.
However, most of them are performed on the driven piles. The dynamic load test was also performed on the
pile having a grouted pile toe by cement in order to improve pile capacity recently. This type of pile was used
as the foundation of underground sub-way station structures and platforms
In this paper, the author presents an applicability of dynamic load test on the so-called “toe-improved pile”
around H-shaped cored steel pile. A result of signal matching analysis using added-mass model for the shaft
resistance on the pile clearly shows a reflection of stress waves from the pile toe and suggests that the
dynamic load test will be applicable to interpret the bearing capacity of the toe-improved pile.

1 INTRODUCTION viaduct where the pile driving faced installation

In Japan, many dynamic load tests have been carried
out for evaluating the bearing capacity of piles.
However, majority of them are performed on the
driven piles.
Recently many different kinds of pile are
introducing to the market. Those piles are aiming to
minimize an influence to the adjacent structures
particularly in the dense city area. Some of them
have an improved pile toe produced by the cement
grouting. In this paper, an example of the dynamic
load test result on an H-shaped pile that has the
improved-toe by the cement grouting is presented.
The pile was designed for the underground sub-way
station structures and platforms.
A dynamic load test was performed on the pile
for evaluating the bearing capacity. One of the
features in the signal interpretation stage was
adoption of the add-mass modeling in the signal
matching analysis to derive the bearing capacity of
the pile in static. The result of analysis suggests that
the added-mass modeling could be able to simulate
the behavior of cement grouting part of the toe-
difliculties due to less overhead room for the driving
work. Under such circumstance, the following pile
installation method was adopted.
Firstly, the ground was pre-drilled by the auger
with diameter of 600mm to the designed pile toe
depth of 9m from the ground level. Then cement
milk that was designed to provide compression
strength of 24Mpa was grouted into the bottom of
bore-hole through a grout pipe. Consequently, the
H-shaped steel pile with a length of 2.0m was
continuously installed into the bore-hole filled by the
cement milk. The H-shaped steels were successively
connected with the plates, which were welded to the
flange and the web, by the bolts for the installation.
Total length of the test pile was 10m and the toe-
improved length was 2.5m from the pile tip as
shown in Figure 1. Table 1 presents a specification
of the H-shaped pile.
Table 1. Speclficationof the H-shaped pile.
Total pilc Young’s
Cross
~ e n s i t y Section C1rcurn*ircntldi
Shape length length
(m) (MPa) (kgin1’) ’4rea
(mA (In)

improved pile by the dynamic load testing. 10.0

H-350 (2.0ws, 2 . 1 ~ 1 0 ~ 7850 171.9 2.054

2 INSTALLATION METHOD AND SOIL

PROPERTY
2.1 Iiwtallntioii method of the pile
The sub-way station was planned to built underneath
of the existing railway station and the railway
2.2 Soil property
Soil condition of the test site consists of the
cemented sandy silt up to 6.0m below GL and the
cemented silt bellow it as shown in Figure 1. Both
stratums were very stiff and N-value of SPT
indicated N=50. However, pre-drilling by the auger

41 1
 for the shear stress between the grouted cement and
its surrounding soil and 12.0MPa for the yield
compression stress of the soil below the pile toe.
Table 2 shows the calculated design capacities of the
pile. The table shows that the total bearing capacity
of Case 3 is largest, which is 4.1 1MN, and of Case
1 is smallest which is 1.76MN.

Figure 2. Design patterns of the bearing capacity.

Table 2. Design Capacities of the H-shaped pile.

Circum-
Stratum Stress ferential Length Capacity
(MPa) length (m) (MN

Disturbed soil 0.001 2.054 5.8 0.012

Grouted cement 0.300 2.054 2.5 1.541

Figure 1. Soil property and the location of the test pile. Case I Toe 12.00 (m2, 0.017 0.206
Shafi capacity 1.552
disturbed the soil condition around the pile so that ‘Total capacity 1.759
the pile toe was necessary to be improved by cement Disturbed soil 0.001 2.054 5.8 0.012
grouting. Figure 1 presents the soil property and the Grouted cement
0.300 0.7 2.5 0.525
location of the test pile. (Pile vs Conc.)
Grouted cement
1.000 0.7 2.5 1.750
Case 2 (Conc vs Conc )

3 DESIGN OF THE BEARING CAPACITY Toe 12.00 ‘Slm2,

. area 0.123 1.470

Three types of toe were designed for the load test. Shaft capacity 0.7 I2
Variety of the bearing pattern at the pile toe was Total camcity 2.182
evaluated in each design. Figure 2 presents the Disturbed soil 0.001 2.054 5.8 0.012
design patterns of the bearing capacity on each pile. Grouted cement 0.150 1.885 2.5 0.707
In the first case, the shaft fiiction would act on the Case 3 Toe
12.00 C.S. area
(m2) 0.283 3.392
interface between the grouted cement and the
surface of H-steel, and the toe capacity would be ShaA capacity 0.719
generated on a cross-sectional area of the H-shape Total capacity 4.1 I 1
steel. In the second case, the shaft friction and the
toe resistance were considered to be generated on
the circumference and on the cross-sectional area of 4 DYNAMIC LOAD TEST
the square formed by the size of H-shape steel,
350mm x 350mm, respectively. In the third case, 4.1 Test equipments
the shaft fiiction and the toe resistance were
A fiee fall steel hammer with the mass of 1.695ton
considered to be generated on the circumference and was adopted as the loading equipment. FPDS-5
the cross-sectional area of the grouted cement body system was used for the measurement of the
having 600mm in diameter respectively. dynamic behavior of the test pile during dynamic
The figures of design value were assumed to be loading. Figure 3 presents the set up of test
0.001MPa for the shaft stress of the disturbed soil, equipments.
0.3MPa for the bonding shear stress between the This test was the frst experience of the dynamic
grouted cement and the surface of H-steel, 0.15MPa load test to the H-shaped steel pile in Japan. The

41 2
authors expected that the quality of the signal would
be different according to the measuring points due to
non-uniform cross section of the H-shaped steel pile.
Transducers were therefore set on the flange and the
web as shown in Figure 4.

Figure 5 Measured signals (Blow No 1 )

Figure 5 shows the measured signals of Blow

No.1, which are time vs. force and time vs. velocity
multiplied by impedance.

5 SIGNAL MATCHING ANALYSIS

5 1 Snirth niodel niialysis

TNOWAVE program was used for the signal
matching analysis to estimate the static resistance
from the test result
At first, Smith model (Smith 1960) was applied as
the soil resistance model, which is the most popular
model in dynamic analyses of the driven pile
Figure 6 shows the measured and calculated up
ward traveling waves from the result of the analysis
4.2 Test yrocedirre and niensirred sig71al The values of soil resistance parameters determined
by the analysis are presented in Table 3 The total
After set-up of the testing system, the dynamic load static resistance calculated from the analysis closes
test was started. The hammer was dropped from the with the value of the design Case 1, but the shape of
height of 2.0m to the pile head that would generate up ward wave shows that the soil resistance was
the rated energy of 33.2kNm. Five blows were fully mobilized i n this test
carried out with the same rated energy. I n Figure 6, two signals are almost equivalent
The settlement of the pile head was about 1.Omm However the difference between the two signals is
per blow. The energy transferred to the pile was observed after the peaks of the signal At the next
approximetly 1OkNm. The driving resistance was
around 3.2MN.
Small differences about the quality of the signals
were observed due to the difference in measuring
locations. The signal measured at the flange had
slightly higher quality than measured at the web.
This difference was assumed due to the different
between flange and web. The thickness of the flange
is thicker than the web.
There was another worry about the quality of the
measured signals due to the influence of the pile
connecting plates welded and the bolts. Such
mechanical connection might be potential to
interrupt proper transmission of the stress wave
between the each segment. However the influence
of the joint was not appeared on the measured Figure 6. Result of signal matching analysis using Smith
signals. model.

413
 Table 3. The result of the analysis including the added mass W, = m g *a (1)
yield
Quake Damping Added where W, = the stress due to the added mass @Urn2),
stratum Thickness stress
(mn) ( ~ ~ s / m 3constant
)
(In) (Mpa) (kg/1n’) m g= the mass constant (kg/m2), and
Disturbedsoil 5.3 0.001 1.0 0.001 0.0 a = the generated acceleration (m/s2)
Disturbedsoil 0.5 0.310 2.0 0.350 0.0 The value of the mass constant was derived from
Grouted
2.5 0.310 2.0 0.350 289
the magnitude of the added mass, which was
cement calculated by the volume from the size of bored
Toe 12.00 2.0 5.00 0.0 hole, multiplied by the density of the grouted
Static shaft resistance 1.552 cement
Total Static resistance 1 759 Figure 8 shows the measured and the calculated
up ward traveling wave signals from the result of the
added mass model analysis. Table 3 also presents
the adopted values of the mass constant The
estimated static resistance did not changed in the
new analysis It is presumed that the pile settlement
was too small. However Figure 8 shows that the two
signals are more similar including the section after
the peak of the signals compared to the first analysis
(see Figure 6)
It is presumed that the motion of added mass,
which was the grouted cement actually, might
generate the force that pull the pile down to the
ground By this phenomenon, a dip of the up ward
traveling wave signals appeared As a conclusion,
the added mass analysis might be possible to
simulate the actual behavior of the grouted cement
as a lumped mass

6 CONCLUSIONS

A dynamic load test was performed on the H-shaped

Figure 8 Result of signal inatclung analysis using Added
model
step it was tried to make the two signals closer at
this part of the signal
5 2 Added mass model aricr&srs
Added mass model (TNO 1996) was adopted to
evaluate the behavior of the grouted cement as a
lumped mass Figure 7 shows the pile-soil model
adopted the added mass model The inertial stress
due to the mass, w.l,is defined as
pile which has improved toe section by the cement
grouting Following conclusions are drawn from the
test results and the consequent matching analyses
/ The added mass analysis is potential to simulate
the actual behavior of toe-improved pile with the
cement grouting by the dynamic load test
/ Minor difference of the qualities of the signal was
observed between the signals measured at
different locations But, the signal measured at
the flange had better quality than that at the web
/ The influences of the pile connecting plates were
not appeared in the measured signals
The test result suggests that the dynamic load test
11121SS can be applicable for the load test of the toe-
improved piles However, it is still necessary to
accumulate many test data applied on the similar
type of piles, and it is also expected that the added
mass analysis is possible to provide more reliable
estimation of static resistance of the pile
REFERENCES
Smith. E A L. 1960 Pile driring analysis by the rlare
equation. ,I Soil .\lecli Found, DIV , 1SW, I ‘0186,
t’o SW, 35-61
TNO report - 7’4’0-DLT Dwamrc Lond Testiiig Signal
A\latchiiig Lkers .Uaiiunl. 1985 - 1996
414
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Dynamic pile testing practice in Finland

H.Jokiniemi, J. Hartikainen & P: Korkeakoski
Geotechnical Laboratory, Tampere University of Technology,Finland

ABSTRACT: High Strain dynamic pile testing was introduced in Finland approximately 15 years ago. Today the
use of PDA is everyday practice at nearly every job site. The method has been proven to be an especially effective
and economical quality control tool for commonly used large diameter steel pipe piles. The Low Strain Pulse Echo
Method is also quite popular for impact driven reinforced precast concrete piles in Finland. The benefit of the
method is that as many as 300 piles per day can be tested meaning low cost per pile. This paper describes the use
of the both methods in Finnish piling practice.

1 INTRODUCTION research and development. Statnamic has not yet been

The soil deposits in Finland were mainly formed during
the last glaciation or thereailer as a result of various
geological processes. In southern Finland, where the
population is the greatest, the thickness of the soil
sedimentary deposits is on the order of magnitude of
about 10 metres but can also be as much as 60 metres.
The bearing layer beneath the soft soil layers is
typically very dense glacial till or very hard bedrock.
The glacial till usually contains stones and boulders
(Heinonen et al. 1997).
The above described soil conditions have had
remarkable effects on typical Finnish installation
methods and pile types. However, it must be admitted
that the Finnish concrete piling technology has mainly
been adopted from Sweden. As far as large diameter
steel pipe piles are concerned, Finland is the pioneer
among the Nordic countries.
The great efforts in the field of stress wave theory
and measurements in Sweden have probably also had
an d u e n c e on Finland. Today PDA is everyday
practice at nearly every job site and practically the only
way to verlfy the bearing capacities of the large
diameter steel pipe piles.
2 FINNISH DRIVEN PILING TECHNOLOGY
Due to the above described soil conditions most of the
piles in Finland are impact driven end bearing piles.
Static loading tests are performed very seldom, only a
few per year, and these tests are practically serving
Typical range of diameters are 356 to 1016 mm and
wall thicknesses 10 to 16 mm. Typical design loads
introduced in Finland.
Precast square reinforced concrete piles are the
most popular pile types in Finland. The most commonly
used cross-sections are 250x250 m2and 300x300
mm’. The concrete strength should be at least 45 or 50
MYa depending on the piling class. Minimum steel-area
ratio is 0,6%.
Typical design loads for the above mentioned pile
sizes are 437,5 to 810 kN. The piles are driven using
30 to 50 kN drop or hydraulic hammers. Design code
for driven piles includes instructions for the stroke and
required end of driving set. These instructions are
based on the method presented by Broms & H e h a n
(1972). Their formula is based on the stress wave
theory in which the shape of the initial stress wave is
assumed and the soil below the pile toe behaves ideal
plastic. Typical strokes at the end of driving are 0,25 to
0,50 m and permanent sets 7 to 25 mm per ten blows
depending on the hammer type and size, design load
and pile length.
In spite of the simplicity of the formula it usually
gives reasonable criteria for piles driven with drop
hammers. However, the piles driven with new
hydraulically accelerated Junttan HHK-A-hammers will
easily be overdriven. The reason for this is the great
efficiency of the latter new hammer type which the
formula can not take into account well enough.
The use of large diameter steel pipe piles has
increased considerably in Finland during the last 15
years. One reason for this was a thorough development
project between manufacturers, owners, design
engineers, contractors and the university.
vary between 1,5 to 7,O MN. Large diameter steel pipe
piles are provided nearly always with rock shoes. The

415
rock shoe consists of a dowel and stiffeners which are
joined to the thick bottom plate. The aim of the rock
shoe is to protect pile base from damages. In some
cases the rock shoe can be chiseled into the rock to
ensure proper seating. The piles are mostly driven with
hydraulic hammers or sometimes using the Franki-
method. The Finnish made Junttan hydraulic piling rigs
and hammers arc the most popular equipment among
contractors.
Due to the large variation in diameters, wall
thicknesses and hammers, part of the large diameter
steel pipe piles should always be tested dynamically.
Large diameter steel pipe piles have turned out to
be extremely effective in the construction of railway
underpass bridges. The bridge deck is fixed to the piles
after constructing it beside the railway track and then
placing it on the piles (Hartikainen et al. 1999). In
Finland the piling of almost 100 bridges has been
carried out during short traffic breaks, often during
weekend nights. It is obvious that the only way to
verify the bearing capacity is to use PDA.
Unfortunately the schedule is often so tight that the
PDA-engineer doesn’t have time to make a CAPWAP
but use only CASE-method. However, most of the
piles are driven to the very dense till or rock where the Figure 1. Junttan HHK-A hydraulic hammer
piles will have plenty of capacity. The lighter hammers
may not activate the required ultimate capacity, but the
expertise of the PDA-engineer together with careful 3 NATIONAL INSTRUCTIONS CONCERNING
review of the site investigations are of extreme DYNAMIC PILE TESTING IN FINLAND
importance.
Small diameter (D=60 to 323 mm) steel pipe piles The Finnish design codes for driven piles (LPO-87)
or driven steel micropiles have recently become more should be used merely to design precast concrete and
and more popular. They have replaced driven concrete wooden piles. These codes were updated in 1987,
piles at sites where soil layers include boulders, which means that only a few recommendations are
bedrock surface is inclined without supporting layers or given for dynamic pile testing.
where, soil displacement causes problems to nearby It is common practice in Finland to use factor of
structures. safety FS=2.0 for dynamic pile testing of concrete
Driveability studies using wave equation and in-situ piles. The amount of test piles varies depending on the
measurements have been demonstrated that it is often designer and owner. With railway projects 5-10% of
possible to drive these piles effectively without the concrete piles are subjected to dynamic pile testing.
excessive driving stresses with the same hammers as Up to 30 piles per day can be measured using restrike.
used with typical concrete piles. However, it is not unusual that similar piles with same
Figure 1 shows the most popular pile driving design loads used on house building projects are not
hammer in Finland, the Junttan HHK-A. The largest measured with PDA but are just driven according to
model available in Finland nowadays is HHK-9A and the criteria presented in the codes.
the product range of Junttan comprises models fkom The practice for low strain testing varies even more
HHK-4A up to HHK-18A. All models have a 1,2 m because nothing is said in the codes. Again the method
stroke and the number in the hammer type denotes the is most often used on the railway projects and it is not
weight of the ram in metric tons. The ram is slightly unusual that every pile is measured. In an 8 hour work
accelerated mainly to compensate friction losses during shift it is possible to measure more than 300 piles,
the drop. Therefore the velocity of the ram using full providing ladders need not be used to reach the pile
stroke just before impact is 4,9 to 5,O mls. This tops.
velocity and the kinetic energy of the ram can be Large diameter steel pipe piles are normally applied
measured with the energy measuring device made by the same factor of safety as for concrete piles although
Junttan for each blow. Jokiniemi (1997) analysed the it is possible to use FS= 1,8 on railway projects when
performance of these hammers. the piles are driven to the sound hard rock.

416
 With highway bridge projects normally at least one
steel pipe pile per support is measured. On railway
underpass bridge projects one pile per track per
support should be measured.
The CAPWAP-analysis is not required in any
Fiimish code. Therefore the choice of the proper
damping factor JC in a CASE-method is sometimes at
the discretion of the PDA-engineer and his experience.

4 DYNAMIC PILE TESTING SYSTEMS AND

THE AMOUNT OF TESTS IN FINLAND

Today there are three different companies or

organizations in Finland who conduct dynamic pile
testing.
Two of these companies use PDI’s system and the
third uses TNO’s system for high strain testing. For
low strain testing there are three different systems: PIT,
SIT and PET.
The total number of piles measured yearly in
Finland is not accurately known by the authors.
However, TUT Geotechnical Laboratory is using
PDI’s PAK and Piletestxom’s PET in addition to the
research also for construction control service at actual
job sites.
In 1999 more than 600 precast concrete piles and
about 300 steel piles were measured by the laboratory
using PDA. Low Strain Tests were made for more than
2500 piles.
5 SELECTED EXAMPLES
5.1 Large diameter steel pipe pile and buoyant force
during driving
As mentioned earlier, evaluation of dynamic pile testing
during construction of railway underpass bridges
during weekend nights requires sometimes f a t decision
making.
The existence of large boulders, which can ‘force’
the pile to lean or even prevent the pile from
penetrating to the ~ u or required
m depth, is
checked on every pile position by means of percussion
drilling.
However, the soil layers can also be very soft and
sensitive with high water content causing the closed
ended steel pipe pile to ‘float’ during driving. If the pile
does not penetrate deep into the hard bearing layer it
can even slightly rise &er the weight of the hammer is
removed from the top o f the pile. The phenomen makes
it difficult to activate the required capacity as the pile
may rise several c e n t ~ e t r eduring
s the upward stroke
of the ram. This “slack” generates a tension wave as
shown in the upper curve in Figure 2.
Normally under such circumstances the piles are
filled with water or loaded several hours with the dead-
weight of the hanuner after reaching the required
depth.
417
Figure 2. Empty steel pipe pile and the same pipe filled with
water.
Figure 2 illustrates two dserent stress waves which
are measured on the same pile. The diameter of the
closed ended pile is 8 13 mm and the wall thickness is
12,5 mm. The upper curve is measured when the pipe
is empty and the lower one when filled with water.
The impact is caused in both cases using 1 1500 kg
drop hammer and a 1,5 to 2,O m stroke.
The length ofthe pile is 21,O rn and above the very
hard bearing layer are very soft soil layers with a high
water content.
According to the upper curve the toe resistance is
low at 2L/c as the force is low and velocity high.
However, later the force increases to quite a high level
suggesting that the pile has risen a few centimetres
from the very hard bearing layer.
By filling the pipe with water the buoyant force
acting on the pile is compensated and the pile remains
at it’s lowest position.
Prediction of the above described behaviour is not
easy. On one railway underpass bridge project in
Fidand the closed ended steel pipe piles stopped totally
on soft clay/silt at a depth of 23 m due to a large
buoyant force. After removing the hammer one pile
rose slowly as much as 6 m. This can really be called a
false stop.
5.2 Small diameter steel pipe pile, friction joint
Smll diameter steel pipe piles installed by different
means are superior in underpinning projects.
Wave equation anaIysis, such as GRL‘UIEAP, and
actual measurements have demonstrated, that it is
possible to drive these piles with larger hammers
without damages as has earlier been supposed in
Finland. Therefore these piles are nowadays competing With the PDA this possible problem can be
more and more with impact driven concrete piles as the observed and the necessary procedure for the
contractor can use the same rig and hammer as with tightening of the splices can be established.
concrete piles. The benefit of these piles compared to
the square concrete piles is that each pile can also be
visually checked after the driving.
Earlier these piles were often driven by using light
drop hammers (M=250 ..SO0 kg) or pneumatic
hammers. Even very light hydraulic hammers, primarly
intended as a rock breaking tool, have nowadays been
considered. To establish a driving criteria for the light
hammers is a very difficult task and the reality and
simulation can differ considerably. The signal quality is
not often reasonable due to very high acceleration
caused by steel to steel impact. Even moderate sized
stones can cause a false stop.
In Finland, small diameter steel pipe piles are
usually jointed using friction splices. Two types are
available: external sleeve splice and internal double
conic splice.
Figure 3 demonstrates one possible problem
concerningjointing which can be detected by PDA.
The stress waves are measured from 36 m long
D=168,3 mm t=lO,O mm closed ended steel pipe pile.
The pile consists of three 12 m long elements and the
jointing has been made by using internal double conic
splices.
The pile has been driven to the required depth
several days previously using a 3000 kg drop hammer
and a 0,20...0,25 m stroke. This stroke causes a
downward travelling force in the magnitude of 500 kN
which is low compared to the required ultimate
capacity of 1200 kN.
Restrike tests were started with one low stroke
H=0,20 m blow. The two upper curves in Figure 3
represent this blow. The aim of this blow was without
breaking the setup to verifL the alignment between the
hammer and the pile etc. The locations of the splices
can be observed as local small compressive waves. This
Figure 3 . Loose mechanical splices
phenomen is according to the theory that the splice
location represents sudden increase of pile impedance.
The two lower curves are measured from the next
5.3 Precast reinforced concrete pile, wave velocify
blow using 0,60 m drop height. It is clearly visible that
using different systems
the upper splice is very loose and not tighten properly
as a large tension wave reflects fkom upper splice
In Finland the age of the precast reinforced concrete
location. However, during later blows the splices
pile should be at least 14 days before it is permitted to
gradually tightened.
be driven. Under certain circumstances, 7 days old
Although the 3000 kg drop hammer is suitable in
concrete piles are allowed to be driven.
magnitude for the above described sized piles, this The typical average wave speeds in Finnish
example demonstrates that the internal double conic
concrete piles are quite low. At the same site different
splices designed for high tension capacity require a few
piles can be stored, handled and driven in different
high energy blows to tighten sufficiently.
ways. This causes different amounts of hairline cracks
Figure 4 shows determination of the average wave to the piles and thus different average wave speeds.
speed for the same pile by using three different layer to hard rock or large boulder. The effective
systems: PDI’s PAK, Piletest.com’s PET and “NO’S length in PDA was 6,5 m.
SIT. The PDA testing (upper curve) was made by using
The pile is 7 m long 300 x 300 mm2 precast a 4000 kg drop hammer and 0,3 m drop height. The
reinforced concrete pile driven through a very soft soil. toe resistance is very high and shaft resistance very

418
low. Any damage or even small cracks can not be TUT Geotechical Laboratory during 1993-1997.
detected. The analysed data consisted of 326 different piles
The both low strain testing curves (middle curve from 77 different sites. Most of the piles were tested
using PET and lower curve using SIT) show no signs
of cracking but the toe is visible in both curves by using
moderate amplification.
The average wave speed with every system is
approximately the same c=3000...3 100 m/s. These
values are however smaller than average wave speeds
in Finnish concrete piles. On the other hand, wave
speeds of c=4000...4100 m/s are not impossible for
Finnish concrete piles.
The variation in interpreted wave speeds between
different piles at the same site can also part!) be
explained by the influence of the pick-up points. It is
probable that the pick-up points are structurally weak
locations due to early litling of the pile f?om the mould
or defective handling during lifting. As the driving
stresses may also be excessive at the location of the Figure 5. Extreme damage at the point ofthe lift.
lower pick-up point, the pile toe can easily be
interpreted to wrong location i.e., pick-up point by during end of driving. Transferred energy to the pile
using high wave speed. This concerns only the low EMX and maximum displacement of the pile DMX
strain method. were recorded by the PDA. The permanent set per
In Figure 5 is an extreme damage at the point of lift blow s was recorded manually and the ram stroke
presented. estimated as well as possible.
The form of the proposed dynamic piling formula is
5.4 Pile driving forrnula,for large diarneler steel pipe
piles
Viljakainen (1998) analysed dynamic load tests for OTE .Wr .H
P” =
large diameter steel pipe piles which were made by
s-%C

where
P, predicted static capacity
OTE Overall Transfer Efficiency ( E M W E )
PE Potential energy of the ram
W, weight of the ram
H stroke of the ram
s permanent set per blow
C elastic deformation of the pile and soil (DMX-s)

The Finnish hydraulic hammers were classified

according to the model and typical values of OTE were
established.
In figure 6.is comparison between Equation (1) and
Case-Goble capacity, hydraulic hammers

Figure 6 presents a comparison between Equation

(1) and Case-Goble capacity. According to the Figure
Figure 4. Wave speed determination by three different systems 6, the CO~elationis strong.

419

Figure 6. Comparison between Equation (1) and Case-
Goble capacity, hydraulic hammers
Equation (1) is actually the same as “true” Hiley
formula or Paikowsky’s (1 994) uncorrected Energy
Approach prediction. The correlation in Figure 6 may
be slightly improved by introducing an additional factor
for the soil below pile toe, which should consider the
dynamic resistance. It is probable that Equation (1)
overestimates the capacity in cohesive soil and
therefore the soil type should somehow be considered
However, most of the large diameter steel pipe
piles in Finland are driven to very dense glacial till or
rock. The decision whether the pile has reached the
rock is in many cases not hlly reliable without dynamic
load testing. Referring alone to the previous, it must be
admitted that the profit of the Equation (1) is only to
give the magnitude of the capacity. It is not
recommended to be used instead of dynamic load
testing.
REFERENCES
Broms, B.B. & Hellman, L. 1972. Methods used in Sweden to
evaluate the bearing capacity of end-bearing precast concrete
piles. IVA Pdlkomniissionen rapport 34: 27-30. Stockholm.
Hartikainen, J. & Koskinen, M. & Suomalainen J. 1999.
Effective construction of railway underpass bridges on steel
pipe piles. F.B.J. Barends et al. (ed). Geotechnical
Engineering for Transportation Infrastructure: I41 -1 46.
Rotterdam: Balkema.
Heinonen, J. & Hartikainen, J. & Kiiskila, A. 1997. Design of
axially loaded piles-Finnish practice. F. De Cock & C.
Legrand (ed.), Design of Axially Loaded Piles European
Practice: 133-160. Rotterdam: Balkema.
Jokiniemi, H. 1997. Egciency analysis of hydraulic Junttan pile
driving hammer. Proceedings of the XIVth International
Conference on Soil Mechanics and Foundation
Engineering: 1077-1080. Rotterdam: Balkema.
Paikowsky, S.G. & Regan, J.E. & McDonnell, J.J. 1994. A
simplified field method for capacity evaluation of driven piles.
Report No. FHWA-RD-94-042. Springfield, Virginia.
Viljakainen, J. 1998. Steel pipe pile’s dynamic load testing and
driving simulation. Master Thesis. Tampere University of
Technology. Tampere. (in Finnish).

6 CONCLUSIONS
The Finnish driven piling technology and dynamic load
testing practice is briefly presented. Some unusual
examples are also given.

ACJSNOWLEDGEMENTS
The pile research group at Tampere University of
Technology wish to thank Junttan Ltd, Rautaruukki
Ltd, Technology Development Center TEKES, VR-
Track Ltd, Finnish Road Administration and Finnish
piling contractors for long lasting co-operation and
financing of research projects.

420
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

riving behavior of large diameter steel pipe piles

Young-Nam Lee
Hyundui Engineering and Construction Company Limited, Seoul, Korea
Jong-Sub Lee
JungAng Engineering nnd Construction Company Limited, Seoul, Korea

ABSTRACT For the construction of a 4 8 km long Multi-Purpose Jamuna Bridge, 2 or 3 large diameter
open-ended steel pipe piles were used for the piers’ foundation A total of 123 piles were driven for 50 piers
and 2 test piles from the river bed through the normally-consolidated upper sand layer and rested on top of
gravel layer Two types of piles, having 3 15 or 2 50m diameter, were driven to depths of 69 to 74m with the
rake of 6 1 Dynamic pile tests were performed on 24 selected piles during pile driving and soil plug length in-
side the pile was also measured after driving
These piles were plugged about 75% of total length of pile driven Piles soon after driving showed a skin-
friction dominant pile behavior, that is, 92% of total capacity being developed by skin resistance Quakes val-
ues and Smith damping factors were almost constant regardless of pile diameters, reflecting the influence of
uniform soil condition at the site

1 INTRODUCTION and service crossing over the Jamuna River at Sira-

Large diameter steel pipe piles were used in Bangla-
desh to support Jamuna Multi Purpose Bridge The
Bridge was constructed across the Jamuna river, with
the surface consisting of normally consolidated loose
to medium dense micaceous sands and gravel layer
The used piles are either 3 15m or 2 50m diameter
steel pipes of variable wall thickness Pile driving was
done using a MHU 1700 Hammer(ram weight 102
tonnes)
This paper describes the dynamic load test results
performed on 24 selected piles during pile driving
and soil plug length inside the pile measured after
bottom and top section driving in all the driven piles.
The Plug length ratio, skin friction distribution during
driving, quakes values and Smith damping factors of
large diameter steel piles in the Jamuna bridge are
suggested.
jganj approximately l00km to the North-West of
Dhaka The proposed bridge consists of a total 48
span modules founded on 50 pile groups at each pier
location (Young-Nam Lee, and Jong-Sub Lee,
1997) Each span is approximately lOOm long The
bridge deck consisted of pre-stressed balanced canti-
lever units bearing onto concrete piers Each pier lo-
cation is founded on 1 6 absolute rake 2 and 3 steel
pipe pile groups Pile diameters were 3 15m for the
2-pile groups and 2 5m for the 3-pile groups(Figure
2) The bridge was linked to 120m long approach
viaducts at West and East Bank
The Jamuna Bridge was constructed to carry a
four lane highway, a 240kV electric power line, a
gauge railway, and a high pressure gas pipeline The
contractor was a joint venture led by Hyundai Engi-
neering & Construction Co., Ltd. of Korea and Mit-
subishi of Japan

2 PROJECT SITE CHARACTERISTICS
2.1 /+(?jt?cfdefciih
The Jamuna Multi Purpose Bridge(Figure l), Bang-
ladesh, consisted of a 4.8km long combined road, rail
The Jamuna river is as wide as 15km during flood
season and 5km during dry season. The main stream
of Jamuna river shifts each year so river training
work is needed.

42 1
2 2 Pile details
The piles for the 4 8km Jamuna Bridge are large
diameter open ended steel pipe piles There is a total
of 121 piles and all are driven on a 1H 6V inclina-
tion The Piles are either 3 15m or 2 50 diameter
steel pipes of variable wall thickness in the range of
40 to 60mm, and driven in two or three sections The
splice is accomplished by use of common offshore
“stabbing guides” which are internal pipes of slightly
smaller diameter with pointed end to assist in align-
ment of the two sections, the thickness of the stab-
bing guides is 25mm and the length of those is ap-
422
proximately 3m. The piles had the “Shear keys”
located neat the pile top and also near the pile bot-
tom(Figure 3 ) .
These keys were approximately 20mm thickness,
semicircular in shape, located in alternate planes per-
pendicular to the pile axis, and extended from about
3m below the pile top to about 13m below the pile
top. The shear keys at the bottom began one diame-
ter above the pile toe and extended about 13m from
the pile toe. Wall thickness for the 3.15m diameter
piles was 60mm at the top, reducing 55mm, then
5Omm, then40mm, before terminating at the toe with
 Figure 3. Stabbing Guide and Shear Key

a 6Omm drive shoe. Wall thickness for the 2.5m di-

ameter piles was 50mm at the top, reducing to
45mm, 40mm, and 60mm at the toe.

All piles consisted of two or three sections with

total lengths between 38-451n. The second or third
sections were equipped with stabbing guide to ensure
the pile alignment during stabbing and welding op-
erations. The stabbing guides were 3.4m long.

2 3 Haninier details
Pile driving was done using a MHU 1700 Ham-
mer(ram weight 102 tonnes) All piles extend above
the ground and water surface so that the underwater
capability of the hammer is not utilized The hammer
is nominally rated at 1700 kN-m for underwater op-
eration (1900 above water) The hammer energy
readout was monitored This energy is a computed
kinetic energy obtained from velocity measurement
by two proximity switches located above the impact
surface and thus underestimates the kinetic energy at
the impact surface due to both gravity and double
action of hammer.
The pile installation (Figure 4) was performed by
Hyundai Heavy Industry using the crane barge “HD acteristics are given below (Fugro-Engineers B.V.
1000” A so-called temporary jig jacket was installed 1996). CPT results were plotted as shown inFigure
at the piling area supported by four 0 61m diameter 5.
temporary anchor piles The pile sections were
stabbed through the jig jacket supported by a top
3.1 Upper Sand Layer
guide and bottom guide mounted on the jacket
These are normally consolidated loose to medium
dense micaceous sands. They are fairly uniform and
3 SOIL CONDITION extend to approximately PWD-64 to PWD-68m.
Their in-situ density is loose in their upper levels, be-
The soils encountered in this project area have been
coming medium dense depth. Below around PWD-
broadly classified into three Units( 1 ) Upper Sand, 2)
50m, they contain occasional fine gravel.
Gravel, 3) Lower Sand” ). Brief notes on their char-

423
 Based on the CPT data, the excess pore water
pressure remained essentially hydrostatic during CPT
penetration. Inclusions (lenses) of fine sand and silt at
random levels were also indicated. CPT results con-
sistently show a linear increasing tip bearing pressure
as a fhction of depth.
3.2 Gr.nvel Lnyer
The gravel layer consists of a mixture of both sand
and gravel. The gravel sizes range from fine to coarse
(up to 6Omm diameter) and occasional cobbles
(maximum diameter 80mm) The gravel at approxi-
mately PWD-70111 was found in all boreholes and the
majority of CPTs The layer is of variable grading,
both vertically and laterally The layer often exhibits
a “sandwich” layering, consisting of a thick layer of
gravel, underlain by sand, and thereafter gravel Oc-
casionally the complete layer is gravel. The gravel
layer is approximately 7m thick and dense
The excess pore pressure again remained essen-
tially hydrostatic during CPT penetration and no in-
clusions of silts or clay were found. Below the Upper
Sand layer is a sand and gravel layer (where pile tips
are founded) with gravel sizes up to 60mm diameter
All piles were driven to a depth of 74 to 76m (toe
elevation PWD-72m)
3.3 L O M W Snmi Lnyer
Micaceous sands encountered at elevations varying
between approximately PWD-7 1 m and PWD-80m,
and exended to depths beyond those investigated
(GL-I 20m) Their in-situ density is probably dense
Figure 5 . CPT Results
424
The CPT qc profiles are erratic, probably due to
the dense state of the sand, with high negative pore
water pressure readings (down to atmospheric) ob-
tained during the CPT pushes. These responses are
probably caused by a combination of dilatancy and a
lower permeability than the upper sand layer. From
the borehole data the sands appear to be uncemented,
and old (Pleistocene) in comparison to the overlying
Holocene “Upper Sand”. From both the borehole and
CPT data, the “Lower Sand” contains inclusions
(both layers and lenses) of very dense sandy silt and
very stiff to hard clay.
The “Lower Sand” is clearly a different layer than
the “Upper Sand’, both in terms of its geology (age)
and geotechnical (density) characteristics.
The detail soil characteristics is summarized in
Table 1
4 DYNAMIC BEHAVIOR
4 I P / I ~Lmgrh r < c m
Soil plug measurements were made to investigate soil
plug behavior during driving by using a so-called
plug follower, a ballast weight placed on top of the
soil plug This weight was connected to a wire
guided to the outside of the pile shaft to enable the
tneasurements of soil plug movements relative to the
pile head
The plugging can be indicated by the specific re-
covery ratio and plug length ratio A better indication
 Table 1. Soil Characteristics of Jamuna Site

Mineralogy Quartz 50-65 55-65 35-65

wolume Feldspar 10-20 15-20 10-20
Percentage) Mica 5-15 10 5-15
Finer(%) 2-10 0-10 8-20
0.06- 0.20- 0.02-
D1 mm)
()(

Size 0.20 5.00 0.06

0.20- 0.50- 0.08-
D5(1(1111n)
0.40 10.00 0.40
Coefficient
of Uniformit

of plugging can be provided by the specific (incre-

mental) recovery ratio, which is defined as the incre-
mental change in soil plug lengths with change in
pile penetration. The specific recovery ratio is the
first derivative of pile penetration depth with respect
t o soil plug length (Paikowsky et al, 1989).

In Jamuna project , the soil plug length could be

checked only two times over driving. The soil plug
length was measured after driving of bottom section
and after completion of pile driving(bottom and top
section) Because it is impossible to calculate the
specific recovery ratio, we used the Plug length ratio,
which is defined as the length of soil column inside
the pile over the total penetration.
The Plug length ratio of Jamuna Bridge pile was
plotted in Figure 6 Plug length ratio at the end of

Figure 7. Typical Force and Velocity x(EA/C)

Traces at a) 3Om, b) 50m, and c) 70m
Penetrations for a 3.15m Diameter Pile

bottom section driving for the 3.15m diameter piles is

about 84.8%, and its value for the 2.50m diameter
pile is about 78.5%. Plug length ratio at the end of
Figure 6. Plug Length Ratio top section driving of 3.15m diameter piles is about

425

Figure 8. Ultimate Capacity at the end of
Driving versus Skin Friction Ratio
76.1%, and its value of 2.50m diameter pile is about
72.1%. Final location of the interior soil plug was
about PWD-l 5m for 3, 5m piles and PWD-20m for
the 2.5m piles, respectively.
426
4.2 Dynamic Pi Ie Testing Results
Dynamic pile testing of the 24 selected piles during
driving was accomplished with use of Pile Driving
Analyzer, and the measured data analysis was per-
formed using CAPWAP(GRL, 1996). Typical dy-
namic pile testing result was plotted in Figure 7 at
the selected depths of 30m, 50m, and 70m. The pre-
diction of the driving resistance was also checked
using the simplified Case method(PD1, 1995). The
pile capacity calculated by Case method increase with
respect to pile penetration depth using case damping
Jc 0.5.
4.3 Pile Copacip
Capacity results obtained by the dynamic testing rep-
resent those present at time of testing. Generally data
presented was for the end of driving. A summary of
results is contained in the attached Figure 8. The av-
erage capacity at the end of drive of the 3.15m di-
ameter piles is around 70MN, and about 53MN for
the 2.5m diameter piles (this capacity is in excellent
agreement with the pile diameter). In both pile sizes,
shaft friction accounts for about 92% of the total.
Unit shaft friction is generally low (less than 100
kPa) down to a depth of about 60m and is developed
primarily along the outside skin friction of the pile.
Below this level (which roughly corresponds to the
top of the shear key location) the unit friction is
about 250 to 350 kPa, and probably represents the
sum of both inner and outer friction (Figure 9)
In addition, the inner soil plug will be to be re-
moved by water jetting and air lifting A concrete
plug will then be installed followed by base grouting
Thus final inner friction should be discounted en-
tirely, setup on the outside has not been filly
achieved at the end of driving, and the soil failure
mechanism at ultimate load for the pile toe will be
changed due to plug installation and base grouting
Thus the capacity of the soil and its dynamic re-
sponse are only vaguely related to driveability or in-
stallation of the piles themselves It is expected that
the long term service load these piles will very sig-
nificantly greater than that presented here, provided
that the base grouting is properly performed
4 4 Sorl Pammeler
Dynamic soil parameter results are plotted in Figure
10 Typically the shaft and toe quakes (QS and QT)
are 0 20 to 0 27cm and 0 27 to 0 4cm, respectively
The Smith damping factors for shafi and toe (SS and
ST) aree about 0 45 to 0 61s/m and 1 24 to 1 69s/m,
respectively The CAPWAP UN value of unloading
was typically 0 4
5 CONCLUSlON
The Jamuna Bridge pro-ject in Bangladesh provided
427
one of the most important case study about the one
of the largest piles driven in normally consolidated
sands. Dynamic pile testing was carried out to inves-
tigate dynamic pile behavior and soil plug length was
also checked.
Piles soon after driving showed a skin-friction
dominant pile behavior, that is, 92% of total capacity
being developed by skin resistance. Quakes values
and Smith damping factors were almost constant re-
gardless of pile diameters. Typically the shaft and toe
quakes (QS and QT) were 0.20 to 0.27cm and 0.27
to 0.4cm, respectively. The Smith damping factors
for shaft and toe (SS and ST) were about 0.45 to
0.61s/m and 1.24 to 1.69s/m, respectively.
6 REFERENCE
Fugro-Engineers B V ( 1996), " Interpretive Report,
Phase I Ground Investigation, Jamuna Bridge
Bangladesh, " Report No K-2380/117
GRL and Associates, Inc (1996), CAPWAP - CAse
Pie Wave Analysis Program, Manual, Cleveland,
Ohio
Paikowsky, Whitman, and Baligh ( 1989), A New
Look at the Phenomenon of Offshore Pile Plug-
ging, " Marine Geotechnology, Vo1.8, pp.2 13-
233.
PDI (1995), PDA - Pile Driving Analyzer : PAK
I'

User's Manual, Pile Dynamics Inc.

I'

Young-Nam Lee, and Jong-Sub Lee (1997), Behav-

ior Prediction and Capacity Estimation of The
Large Diameter Piles, Research Report, Hyundai
Institute of Construction Technology.
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

ynamic load test of cast in place pile using a free fall hammer

S. Niyama, G.C.de Campos & S.Navajas

Institute of Technological Research of SGo P a d o State, IPIZ: SGo Paulo, Brazil
S.C.Paraiso & C.M.C.Costa
Geomec, Engenheiros Consultores, Belo Horizonte, Brazil
G. E. Barbosa
Construtorcr Andrade Gutikrrez S.A., Brazil

ABSTRACT: The use of high strain dynamic test applied to cast in place piles is growing in the foundation
engineering practice in Brazil. This paper presents a single case of the use of this method to assess the bearing
capacity of large bored pile, part of the foundation of the new Silo Paulo subway line. The use of a special
hydraulic self-propulsion hammer delivering high energy, locally developed, allowed the verification of the
parameters required by the foundation design. The dynamic test was conducted according to the common
procedure, in Brazil, consisting of the application of variable energy.

1 INTRODUCTION costs. Within this limitation some alternatives of

This paper describes a case of the application of high
strain dynamic test applied to a short large diameter
bored pile, part of the foundation element of the new
superficial subway line in Silo Paulo city. A special
hydraulic hammer device was used in this test in
order to transfer the impact force to the pile top,
delivering sufficient energy to mobilize the total
shaft resistance along the pile. The test which was
carried out by applying five blows, with increasing
energies, made possible the evaluation of the bearing
capacity by the CAPWAP method.
The Franki type pile was selected as the
foundation element for this job, considering the
geological and geotechnical conditions, as well as
the economical aspects. Nevertheless, enviromental
restrictions in part of the job area, where one or two
floors old commercial buildings are located -
including a small medical center with very sensitive
equipment - demanded a review of the foundation
design. Theoretical study of the vibration and noise
level induced during the Franki type piling was done
in order to predict the possible effects in those
buildings and in medical equipments. The results of
the study indicated that dynamic piling should be
avoided.
The local geological formation, as presented in the
next item, limited the length of the foundation
element to approximately 11 m, depth of the
bedrock. Any use of piles below this depth would
demand techniques and equipment with vew high
foundation were studied, resulting in the indication
of single bored piles or single bored piles connected
to the rock mass through root piles executed in their
base.
Therefore, the test had as main objective, the
assessment of the capacity of the single bored piles,
particularly their tip resistance, to evaluate the need
of the complementary root piles.
The option of using high strain dynamic load test,
instead of a static load test, was due to economical
reasons and also to the time available. Even from the
technical point of view, there was a lot of difficulties
to perform a static test in the place, close to intense
and heavy traffic.
2 CHARACTERISTICS OF THE SOIL
The subsoil in the test area was composed of a
stratigraphy in which Quaternary sediments overlay
gnaissic residual soils and fractured rock at low
depth.
The Quaternary sediments from the Silo Paulo
basin are very heterogeneous due to their deposition
in a sequence of lacustrian regimens. The rocks that
underlay the Quaternary and Tertiary sediments are
of Upper Proterozoic age, partially constituted by a
metassedimentary rock and part of an igneous
looking rock. Wheathering of these rocks is also
extremely heterogeneous and the residual soils
originated from these rocks can show a micaceous

429
Figure 1. Soil profile and schematic of the hammer device
sandy silt, as well as a clayey silt matrix depending
on their rock of origin.
The dynamic testing was done on a production
pile. The area was investigated by SPT borings to
refusal, one located exactly where the test pile was
executed.
Figure 1 presents the soil profile from the SPT
boring, which presents the sequence of layers as
described below.
Presence of 1.9m superficial landfill layer, above
2.8 m of soft organic silty clay layer. Below the clay
layer there is a fine sand layer with gravels (0,95 m)
and, after this, there is a sandy silt residual soil layer.
At 11.26 rn depth appears, the top of the bedrock
constituted by granite and gnaisse. The water level in
this place is located 3.4 m below the land surface.
3 DESCRIPTION OF THE TESTING AND
IMPACT DEVICE
3.I Characteristics of the testedpile
The main characteristics of the tested pile are:
1.20 m diameter, total length of 10.50 m, and length
below the ground level of only 6.40 m.
430
The upper 3.0 m of the pile was specially cast in
place to allow the support of the hydraulic hammer.
This pile was dug using bentonite slurry up to the
bedrock depth, then the steel reinforcement cages
were introduced and finally the fresh concrete was
placed, prepared to have a characteristic compressive
strength of 20 MPa. The time interval between
casting and testing the pile was of 26 days, so
complete cure of the concrete could be assumed.
The combination of the vertical and horizontal
forces including bending moments, acting on the pile
top, resulted in the predicted working load of
2,260 kN for the particular tested pile. The tested
pile is one of a group of 8 piles, supporting a single
pillar of the elevated way of the metro line.
The value of 3,200 d s for the wave velocity was
adopted based in the field mesurements.
Considering 25 kN/m3 as the mass density of the
pile, a value of 26,100 MPa was adopted as the
elastic modulus of the pile.
3.2 Characteristics of the hammer
The device designed to apply high impact forces on
the head of the bored pile consists of a free fall self-
propulsion hammer with special features to test piles A plywood cushion thickness of t = L2/2D -
with high load capacity and variable diameters, where t (thickness) is expressed in mm, and L
varying from 700 mm to 2500 mm. (pile length) and D (pile diameter) are in meters -
Selection of a proper hammer size is essential for is suggested with a minimum value of 100 mm;
successful high-strain dynamic testing. Hammer an additional 150 mm should be added when pile
weight, drop height, and cushion details must be length exceeds 30 meters.
appropriately chosen so that hammer impact causes In this particular case, it was determined that a
sufficient pile movement in order to mobilize the ram of 140 kN, a maximum height free fall of 1.2 m,
required soil resistance, and to assure that dynamic and a plywood cushion thickness of 150 mm by
stresses in the shaft will not impair its structural 1.1 m diameter should be used to comply with the
integrity. recommendat ions.
The hammer apparatus also was constructed in a The diameter of the plywood cushion should be
way to facilitate mobility around job sites and assure about 80 to, at most, 90 percent of the shaft diameter
a uniform impact to the head of a pile. to be able to centralize the blow. Thickness of steel
The free fall hammer conceived by GEOMEC and striker plate to be placed above the top of the pile is
in use recently in Brazil consists of the main suggested to range from 50 mm to 100 mm, and its
components, shown in Figures 1 and 2: a steel striker minimum diameter should be equal to the cushion
plate, a cylindrical - 2 pieces - steel cage (both of diameter.
these components being referred to as hammer), a set The ram is ideally a simple drop weight which can
of round 20 kN steel parts allowing the building of a be raised to a variable height. To facilitate the
ram of up to 200 kN with a maximum drop height of assembly rams are made in segments which can be
3 m and a hydraulic system to lift the ram with an assembled at the job site. The ram to be used needs
automatic catcher, which activates and controls the to be "guided" to assure that the pile top and ram
hydraulic jacking equipment. bottom are perfectly parallel during impact, to assure
The steel striker plate consists of a round 2.1 m uniform contact stresses during impact.
diameter by 100 mm thick steel plate, with four
lateral latches that adjust themselves according to the
diameter of the pile to be tested. It also functions as
a hammer support on the pile top and as a helmet.
After positioning and centering the steel striker
plate on top of pile to be tested, the assembly of the
hammer is started in this sequence: the ram, the
cylindrical steel cage, the ram lifter's hydraulic
system, the hydraulic jacking equipment, and,
finally, the four diametrically opposed telescopic
support braces, which maintain the stability of the
hammer during the test. Figure 3 shows details of the
placement of the hammer on the pile top.
To apply each impact, the ram is lifted by the
hydraulic system up to the desired height, and the
automatic catcher releases the ram on a free fall
inside the steel case upon a hydraulic command,
hitting the plywood cushion that transmits the high
energy to the pile head. The drop height is controlled
by a milimetrically graded scale located outside the
hammer steel case.
To determine the hammer weight, drop height and
plywood cushion thickness to be used in this
particular test, the bibliographic recommendations of
Likins, 1994 and Hussein et al, 1996 were followed:
The hammer weight should be at least equal to
1.5% of the anticipated static test mobilized load,
The hammer drop height should be
approximately 8.5% of pile length, with a
minimum value of 2.0 m,
43 1
 Between the steel striker plate and the top of the
pile, a 20 mm thick by 1.0 m diameter steel wired
rubber pad was installed, in order to evenly
distribute compressive stresses on the pile head,
absorbing the dynamic impact, and preserving the
structural integrity of the pile top, above the
electronic sensors which will measure data to be
associated with the test.
3.3 Test procedure
Instrumentation for the dynamic test followed the
recommendations of ASTM D-4945/89, AASHTO
Designation T 298-93, and ABNT - NBR 13208
(Brazilian Standard).
In the present case, considering the 1.2 m diameter
of the pile, four strain transducers and four
accelerometers were used, located by pairs on each
90" along the pile perimeter. Sensors were installed
at 3.46 m from the top and for better adherence of
the sensors onto the shaft, its' surface was polished.
Field testing was done using the PDA (Pile
Driving Analyzer), PAK-95 model, manufactured by
Pile Dynamics Inc. The Case Method was used to
interpret the data in the field.
Figure 3. Detail for the fixation of the hammer to the
pile top
The test was done by a sequence of blows with
variable drop heights as a dynamic load test with
increasing energy. This type of test has been
discussed by Aoki, 1989; Niyama & Aoki, 1991;
Hussein et al, 1992; Aoki & Mello, 1992, Beim &
Aoki, 1996, Paraiso & Mello, 1998, among others.
The test was carried out with drop heights of
0.4 m, 0.4 m, 0.8 m, 1.2 m and 1.0 m, a total of
432
5 blows. The test was interrupted because the quality
of the signals was considered satisfactory to make
CAPWAP analysis later.
For each blow, set measurements were collected
from the shaft of the pile, in order to evaluate the
pile-soil interaction behavior. For the first blow,
15 mm of permanent set was observed and
practically no permanent set was measured in the
following blows.
4 RESULTS
Figure 4 presents the force and velocity x EA/c
curves as function of time, for the fifth and last
blow, showing a good proportionality. Table 1
presents the results of the maximum Case Resistance
(RMX) for each blow applied, for three values of Jc
(Case damping factor) and considering the wave
velocity of 3,200 m/s.
Table 2 presents the main parameter obtained by
CAPWAP analysis for the five applied blows.
During the analysis it was observed that the best
match of the curves was obtained when it was
considered a wave velocity of approximately
3,600 m/s, final value adopted for the interpretation
of the results.
Figure 5 shows the results obtained for the last
blow and Figure 6 presents the resistance
distribution for all blows.
The CASE method indicated values from 5,800 to
8,300 kN for the mobilized resistance, considering
Jc = 0.3, value close to that verified later by the
CAPWAP analysis. Nevertheless, CAPWAP
analysis presented smaller values varying from
5,730 kN to 7,471 kN. This difference could be the
consequence of the different values of wave velocity
adopted in the field for CASE method and in the
office for CAPWAP analysis.
The efficiency of the present hammer device
varied from 21% to 34% along five applied blows.
The CAPWAP analyses show that the maximum
RMX was obtained for the blow with maximum
energy (blow 4).
The skin friction resistance reached the maximum
value for the third blow; after this, despite of
increasing energy, the skin friction value was
smaller, suggesting that the saturation of the
available resistance was reached.
The toe resistance increased during the sequence
of blows, showing that its total mobilization only
occurs after the saturation of skin friction resistance.
Considering the maximum value of resistance
obtained by CAPWAP analysis it was determined
 H CSI CSX TSX EMX FMX DMX R M x (W)
(m) (MPa) (MPa) (MPa) (kJ) (kN) (mm) Jc=O,l Jc=0,5 Jc=0,7
0.40 8.9 4.6 0.7 12 5,300 3.0 6,500 5,200 5,000
0.40 12.9 6.3 0.8 19 7,200 4.1 7,400 6,200 6,100
0.80 17.2 9.6 1.1 38 10,900 5.9 10,200 7,500 7,400
1.20 20.3 9.9 0.7 49 11,200 7.3 9,700 7,900 7,600
1.oo 23.7 10.0 0.1 46 11,400 7.6 8,500 7,200 7,000
where:
H - drop height
CSI - maximum tension registered by one of the strain transducers
CSX - medium tension stress among four strain transducers
TSX - maximum tension stress
EMX - maximum energy transferred
FMX - maximum compression forces
DMX - maximum downward displacement
RMX - maximum resistance mobilized during the blow, by CASE method
Jc - CASE damping factor

H Ru RP Rs Qs Qt Js Jt
(m) (kN) (kN) (kN) (mm) (mm) CASE CASE
0.40 5,730 2,000 3,730 2.10 2.10 0.14 0.32
0.40 6,100 2,100 4,000 2.10 2.60 0.10 0.35
0.80 7,000 2,130 4,870 2.52 4.10 0.32 0.22
1.20 7,47 1 4,195 3,276 1S O 4.34 0.23 0.08
1.oo 6,540 4,640 1,900 1.20 6.40 0.10 0.26

433
that, in relation to predicted working load for the pile
(2,260 kN), the safety factor would be greather than
3. However, considering that no permanent set was
verified during the application of the blows, it is
possible to say that, probably, the total available
resistance was not mobilized by the blows.
Figure 6. Distribution of total resistance for the five
_ .
blows.
5 CONCLUSIONS
AASHTO T 298-93 - American Association of State
Highway Officials Standard method of test for high-
strain dynamic testing of piles.
ABNT NBR 13.208/94 - AssociaqBo Brasileira de
T6cnicas - Estacas de Carregamento
Dinamico (in Portuguese).
Aoki, N. A New Dynamic Load Test Concept. In:
ICSMFE, 12, TC Pile Driving, Rio de Janeiro.
Proceedings for the Discussion Session 14, v. 1, p. 1-
4, 1989.
Aoki, N. and Mello, V.F.B. - 1992 - Dynamic loading
test curves. Proc. Fourth International Conference on
the Application of Stress-Wave Theory to Piles.
Hague, Holland.
Beim, J.W., Aoki, n. - 1996 - Dynamic load test method
with variable energy. Proc. Fifth International
Conference on the Application of Stress-Wave
Theory to Piles. Orlando, USA.
Hussein, M., Rausche, F. and Likins, G. - 1992 -
Dynamics of pile driving as a function of ram drop
height. Proc. Fourth International Conference on the
Application of Stress-Wave Theory to Piles. Hague,
Holland.
Niyama, S. and Aoki, N.- 199 1 - CorrelaqBo entre provas
de carga diniimica e estatica no campo experimental
da EPUSP/ABEF. 2'd Seminario de Engenharia de
Fundaq8es Especiais e Geotecnia, SEFE, SBo Paulo,
Brazil (in Portuguese).
Paraiso, S. and Mello, L.G. - 1998 - Variable energy
dynamic load test on a 1.0 m diameter CFA pile.
BAP 111, November.

The present work showed that the use of high strain

dynamic testing for large diameter bored is an
economically feasible alternative, with advantages
upon the conventional static tests.
The development of the device designed to apply
high impact forces through a free fall self-propulsion
hammer made possible the execution of dynamic
testing on this kind of pile. The results obtained
using this technique are satisfactory and have given
important subsidies for the improvement of
foundation design criteria.

ACKNOWLEDGEMENTS

The authors would like to mention the valuable

support offered by Andrade Gutierrez General
Constructor for all the help in testing the pile.

REFERENCES

ASTM D 4945-89 - American Society for Testing and

Materials Standard test method for high-strain
dynamic testing of piles.

434
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 1503

ynamic testing of large auger pile using free fall loading system in a harbour
work
S. Niyama, S. Navajas & G.C.de Campos
Institute for TechnicalResearch of Siio Paulo Stute, IPT, Siio Paulo, Brazil

ABSTRACT: This paper presents results of high strain dynamic testing carried out on three concrete auger
piles, using a special free fall loading system, at Portocel Harbour Terminal enlargement job, in Espirito
Santo State. The geometrical characteristics of the piles are: 1.0 m dimeter, 34.71 m, 34.85 m and 33.48 m in
length, and maximum compression working load varying from 2,940 kN to 3,293 kN. Previous drivability
study was performed for the design of the impact force application device, which was manufactured by the
Contractor. Finally, the results obtained by the drivability study were compared with the data obtained during
the field tests.

1 INTRODUCTION 4 m of a superficial soft clay layer, followed by a

The dynamic testing of cast in place piles is still
relatively little used in Brazil, although a lot of
bibliography is now available worldwide. This paper
shows an experience with auger piles submitted to
dynamic testing, accomplished with success in the
Portocel Harbour Terminal amplification job, in
Espirito Santo State, located at southeast region of
Brazil.
The application of high strain testing instead of
the three conventional static load tests, initially
foreseen in that job, allowed a significant saving of
time and also of involved costs. The reaction system
in a static load test would demand the execution of
anchors in a place with water table varying from 4 to
8 meters, with inherent difficulties and costs. The
use of the dynamic method was possible because of
the recent revision of the Brazilian Standard for
Design and Execution of Foundations - NBR 6122 -
which has dynamic testing as one of the alternatives
accepted for evaluation of the bearing capacity of
pile foundations.
2 SOIL CHARACTERISTICS
The Portocel Terminal is located close to the city of
Barra do Riacho and of the industry Aracruz, about
80 km to the north of the capital Vitoria in Espirito
Santo State. The soil characteristics presents besides
the water table varying from 4 m to 8 m depth about
hard silty clay with presence of sand layers with SPT
values between 20 and 40 up to the 25 m depth, limit
of the existing borings.
3 DRIVEABILITY STUDY
A driveability study using the WEAP program was
conducted to predict the performance of the special
hammer system designed to allow the application of
the high strain dynamic load according to the
Brazilian Standard NBR 13.208 (1 994).
A free fall steel hammer was simulated with
80 kN weight, cylindrical form, and 2.0 m lenght.
An efficiency of 95% was adopted and the fall
height was varied between 1.0 m and 1.5 m. The
hammer cushion was considered as a wood disk with
89 mm thickness, cross section area of 2677 cm2 and
elastic modulus of 3,242 MPa. The weight of the
steel helmet was considered as 29 kN. For the pile
cushion plywood disks were simulated, havin
50 rnm thickness, cross section area of 6,362 cmLi
and Elastic Modulus of 207 MPa. The pile length, in
that study, was of 31 m, the upper portion of 16 m
length with 1.0 m diameter and lower portion of
15 m length with 0.9 m diameter. The concrete
elastic modulus was of 40,000 MPa and the specific
weight 24.0 kN/m3.
Total soil resistances of 3,000 kN, 6,000 kN and
9,000 kN were simulated with different
resistance distributions between friction and end

435
Figure 1. Summary of the results of the driveability study for triangular friction resistance distribution.

Bearing Maximum Maximum Permanent Energy

capacity compression tension stress Set
(kN) stress (MPa) (MPa) (m) (kJ)
Hammer weight 3000 13,5 625 3,9 36,l
=SO kN
Hammer stroke 6000 13,5 536 175 36,l
=1 .Om

bearing. Also, two forms of distribution of the 4 CHARACTERISTICS OF THE PILES

friction resistance were considered: a rectangular
form and a triangular one, both beginning at 18.6 m
depth down to the pile bottom. Smith damping used
for the soil along the shaft was 0.65 s/m, and a quake
of 2.5 mm. For the tip resistance, a Smith damping
of 0.50 s/m and a quake value of 8 mm were
considered.
Figure 1 shows a typical result obtained for the
case of triangular skin friction resistance distribution
representing 80% of the total static resistance.
This driveability study presented the results
shown on Table 1, for a total bearing capacity equal
to the working load, and for twice that value.
Those results showed the feasibility of
performing the dynamic testing, using the device
designed for application of the impact force.
The tested pile was installed by driving a steel pipe
pile to a hard soil layer and drilling the internal soil
plug up to a depth defined by the designers (below
the end of the pipe pile) and then inserting the
reinforcement cage and placing the concrete.
Three test piles have 1,0 m of concrete diameter
in the portion with steel pipe casing and 0.90 m
diameter in the portion without casing.
The relations between the total lengths and drilled
lenghts are 34.71 d 18.42 m; 34.85 m43.13 m; and
33.88 d 1 8 . 3 0 m.
Additional portion of the piles (from 3.3 m to
3.5 m) were executed above their tops, in order to
make possible the use of the impact application
device, as well as the sensors.
The average resistance of the concrete pile after

436
Figure 2. Schematic of the hammer device
28 days was of approximately 42 MPa. The concrete
of the piles was placed more than 30 days before the
dynamic tests.
5 IMPACT FORCE APPLICATION DEVICE
The impact force applicatjon device was developed
and built by GEOBRAS, responsible for the
execution of the piles in this job site, under general
orientation of IPT - Institute for Technological
Research of Silo Paulo, considering the results
obtained in the previous drivability study.
The device was composed of a steel tube having
1.0m diameter and 8.0 m total length? closed in the
bottom by two steel plates of 25mm thickness that
also served to support the system upon the pile top
and also as the pile helmet.
The impact was generated by dropping a
cylindrical steel mass of 8,600 kg weight, hoisted by
on hydraulic motor fastened in the top of this steel
tube, where an automatic de-hook system was
designed. The fall height was controlled visually
through spaced holes of 10 cm diameter drilled
along the steel tube.
The hammer cushion consisted of a wood disk
800 mm in diameter and with 100 mm thickness.
The pile cushion three plywood disks consisted of
800 mm in diameter and with 25 mm of thickness
each one.
437
Figure 3. Device for impact application
Figure 2 presents the schematic drawing of the
device used and in Figure 3 the same device is
observed already positioned in the top of the pile by
an ordinary crane available at the job site. In this
picture the piles located in the proximities can also
be observed, with part of the reiforcement cages
appearing from the pile heads.
6 INSTRUMENTATION
The instrumentation consisted of a pair of force
transdutores and accelerometers, used in high strain
dynamic test. They were fixed on the concrete
surface of the pile, through two windows done in
diametrically opposite positions on the steel casing
surface, located 2,5m from the pile top. It was
decided to install the sensors on the concrete surface
and not on the steel casing, because the perfect
adhesion of both materials could not be verified.
7 TEST DESCRIPTION
The high strain dynamic test was conducted
adopting the procedure widely used in Brazil, which
consists in the application of a series of blows with
 Pile d ;2" hfghts
series 3'd series
topographical leveling showed a displacement of
less than 1.O mm per blow.
(m> (m> (m>
1.oo 1.oo 1.oo
1.25 1.oo 1.25 8 RESULTS OF DYNAMIC TESTS
1S O 1.oo 1S O
1 1.oo 1.75 Figure 4 presents a typical result obtained by the
1.oo 2.00 instrumentation for different hammer drop heights,
2.00 with curves of force and velocity.
2.00 Because of the increasing drop heights, those
1.oo 1.oo 1.25 results show also growing values of maximum force.
1.25 1.25 1S O The peaks of those curves, however, are not
2 1.50 1S O 1.75 pronounced, showing that the impacts were
1.75 1.75 2.00 relatively well softened by the cushion system. In
2.00 2.00 fact, there was the concern not to damage the piles
1.25 1.oo as they were actual production piles, not test piles.
1S O 1.25 The summary of the results obtained for Piles 1, 2
3 1.75 1S O and 3, by the PDA (Pile Driving Analyzer), with
2.00 1.75 bearing capacities estimated by the Case Method, is
1.75 presented in Tables 3 to 5 .
2.00 The hammer device presented a low efficiency, from
13% to 26%, computed as the relation between
ENTRHU (transferred energy measured by PDA)
increasing drop heights (Aoki, 1989; Niyama and and potential energy.
Aoki, 1991). Table 2 shows the different heights Pile 1 , the first one tested, presented a lower
used on each pile during the three series of impact efficiency due to the significant energy loss at the
applications. automatic de-hook system. After the corrections to
In the case of Pile 1, the first two series were used minimize the loss, a significant efficiency
to verify the hammer behavior, allowing also certain improvement was verified.
accommodation of the cushion materials. Table 6 presents a summary of the results of the
The permanent set in all piles measured by CAPWAP analyses, where RU is the static

438
Table 3. Results o f t e CASE nethod 1
SERIES I BLOW DMX EMX RMX (kN)
H(m) (mm> (kN.m) J=O.O J=0.2 J=0.4 JzO.6 J=0,8
1st 1.oo 3.5 12 4,200 4,300 3,800 3,200 2,700 2,200
I 1.25 4.8 17 5,400 5,000 4,000
5.4 20 5,900 5,500 4,400
3.3 11 4,500 4,100 2,700
1.oo 3.1 12 4,900 4,400 2,600
1.oo 3.O 12 4,700 4,900 4,400 3,000
1.oo 3.3 12 4,700 5,100 4,600 3,100
1.oo 3.7 15 17 5,000 5,300 4,800 3,100
3rd 1.oo 2.5 10 12 4,500 4,800 4,200 3,000
1.25 4.2 16 5,700 5,200 4,500 4,000 3,500
1S O 4.2 19 6,100 5,500 4,900 4,300 3,700
1.75 5.6 23 5,800 6,700 6,200 5,700 5,200 4,600
2.00 4.5 26 6,300 7,400 6,800 6,200 5,500 5,000
2.00 5.2 27 7,700 7,100 6,400 5,800 5,200
2.00 5.8 28 6,300 5.100
DMX = maximum ( spiacement EMX = maximum energy
EFIC = eficiency (EMX/(weight x drop height)) FMX = maximum force
RMX = maximum resistance

Table 4. Results o f , e CASE iethod fc pile 2

SERIES I BLOW DMX EMX EFIC. FMX
(mm> (kN.m) (%) 0J=O.O J=0.2 J=0.4 J=0.6 J=0.8
2.8 17 20 6,600 6,900 6,300 5,700
1 5,200
1 4,600
1
1
1.25 3.8 23 21 7,200 7,800 7,100 61400 5,900 5,300
1.5 4.3 29 22 7,900 8,600 7,800 7,100 6,300 5,700
1.75 5.0 34 23 8,200 8,800 8,100 7,200 6,400 5,600
4.0 40 23 8,500 9,200 8,200 7,300 6,400
3.7 20 23 6,800 7,200 6,500 5,900 5,300

3rd
1
1.25
1.5
1.75
2.0
1.25
3.9
5.9
4.9
5.5
4.2
25
33
33
40
24
23
26
22
23
22
7,200
8,000
8,000
8,500
7,100
7,700
8,500
8,500
9,000
7,500
7,100
7,600
7,800
8,100
6,900
6,300
6,600
6,800
7,200
6,200
5,700
5,800
6,100
6,300
5,600
5,000
5,000
5,300

4.7 31 24 8,100 8,500 7,700 7,000 6,200 5,400

6.3 39 26 8,500 8,900 8,000 7,000 6,200 5,300
5.4 40 23 8,600 9,100 8,100 7,200 6,400

Table 5. lesults of the CASE method for the Dile 3

SERIES BLOW DMX EMX EFIC. FMX RMX (kN)
H(m) (mm) (kN.m) (%) (kN) J=O.O 1 J=0.2 I J=0.4 I J=0.6 J=0.8
21 1 20 I 5,900 I I I 1 5,500
~~

1st 1
1.L-J I 7,100 6,600 6,000 5,000
1.5
1.75
2.0
28
28
36
22
19
21
I6,500
6.700
7,800
I 8,000
8,300
9,400
1 7,300
7,600
8,500
6,600
6,900 1 6,000
6,200
5,300
5,600
6,200
211d 1.o 16 19 5,300 6,100 5,700 4,200
1.25 22 6,200 5,600 5,100
1.5 30 6,900 6,300 5,600
1.75" 26 6,500 6,000 5,400
1.75 33 7,800 7,200 6,500
2.0 32 7,100 6,300 5,600
+-defect ve blow

439
 PILE DROPPING “SET” RU RS RP JS JP Qs QP
HEIGHT (m) (mm) (kN) (rn) (kN) (mm) (mm)
E13 2.0 0a 1 7,806 7,266 539 0.915 0.022 2.97 2.70
E47 2.0 0a 1 8,195 5,186 3,009 0.478 0.667 1.63 2.76
E53 2.0 0a 1 8,322 5,578 2,743 0.655 0.246 1.95 1.47

resistance, Rs the skin resistance and Rp the tip Niyama, S. & Aoki, N. Correlation of Dynamic and
resistance. The CASE dampings and the quakes (Js, Static Loading Tests in the Experimental Field of
Jp, Qs and Q p ) corresponding to the skin and tip University of SBo Paulo. In: SEFE, 2, Silo Paulo.
resistances are also shown. According to the results Proceedings, 1991 (in Portuguese).
of CAPWAP analyses, the bearing capacity of the
piles were considered varying between 7800 KN to
8300 kN, which are satisfactory values according to
the design requirements.

9 CONCLUSIONS

The use of the high strain dynamic test method for

evaluation of the bearing capacity of large auger
piles was considered fully satisfactory by application
of impacts using a special device built for this job
site.
The use of a previous driveability study was
helpful for the design of the hammer device.
However, that experience showed that a better
evaluation of the effiency to be adopted in the study
is necessary. In this study, a 45% global efficiency
for the dropping height of 1.0 m was adopted and the
measured values in the field were 14%, 23% and
19%, respectively for the three piles. Probably the
main loss occurred in the hoisting system and also
due to the lack of parallelism between the face of the
mass ram and the hammer cushion during the
impact. The measured impact velocity was 0.5 lm/s,
when the theoretical value was 1.41 m/s for a drop
height of 1.O m.

ACKNOWLEDGEMENTS

The authors wish to thank G E O B R h for the

opportunity to perform this work and also IPT for
the incentive to prepare paper for this conference.

REFERENCES

ABNT Brazilian Standard for Design and Execution of

Foundations, NBR 6122. Rio de Janeiro, 1996 (in
Portuguese).
ABNT Piles - Dynamic Loading Test, NBR 13208. Rio
de Janeiro, 1994 (in Portuguese).
Aoki, N. A New Dynamic Load Test Concept., In:
ICSMFE, 12, TC Pile Driving, Rio de Janeiro.
Proceedings for the Discussion Session 14,v.1, p. 1-
4, 1989.

440
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Dynamic load test on high capacity pile socketed in basaltic rock

Skrgio C.Paraiso, Clhudia Maria C.Costa & Ecidinkia Pinto Soares

Geomec,Engenheiros Consultores, Belo Horizonte, Brazil

ABSTRACT: This paper is about the installation of a large diameter pile into weathered basaltic rock and
the interpretation of variable energy dynamic load test done with a specific free fall hydraulic hammer. This
allows modulation up to 200kN with a maximum drop of 3 meters. A comparison is done with bearing
capacity estimations, utilizing empirical methods. This article presents a case of a 1,30 meter diameter high
capacity pile socketed in basaltic rock with a working load up to 4400kN. It also consists of the foundation
and construction details, geological and geotechnical site characterization, the testing program, variable
energy load test, pile instrumentation, tests results, discussions and conclusions.

1 INTRODUCTION the BR-153, on the border between the states of

The use of the dynamic load test for the evaluation
of the load capacity on high load capacity piles with
large diameters also serves for the verification of the
integrity of the piles. Its employment has been
steadily increasing in Brazil with the use of the
hydraulic hammer with an available potential energy
of up to 600kN.m, and has been especially
developed for tests on this specific type of pile. The
dynamic load test has became an important tool for
the quality control of designed foundations. In this
article, the results of the executed dynamic load test
are presented along with a description of the
equipment used for the test and semi-empiric
analyses that were considered in the project phase
for the evaluation of the interaction between pile and
soil.
Santa Catarina and Rio Grande do Sul, Brazil.
2.1 Geological framework
The local of the implantation of the Ita hydroelectric
dam presents basaltic spill dominions of the Serra
Geral formation, with a thickness of approximately
400m, covering the sedimentary rocks (Botucatu
sandstone) where the Uruguay River forms a great
loop denominated Volta do Uva. The flat, sub-
horizontal layers of these spills are evident in the
hillsides of the river, where descending ruptures
occur in the form of steps.

2 GEOLOGICAL AND GEOTECHNICAL SITE

CHARACTERIZATION
The instrumentation and dynamic test were con-
ducted on pile T6lpillar P5, Figure 1, an integral
component of the foundation of the bridge on the
Uruguay River, located in the complex of the Ita
hydroelectric dam. This bridge has a total extension
of 6 3 0 q with a central span of 130% located on
Figure 1 - Cap Block - Pile location.
The following are identified as the main
geological structures present in the area:
DIABASIC DIKE - with an approximate thickness
of 20 meters, in a vertical N40W direction, and
diagonally cutting the forced bypass tunnels;

44 1
STRUCTURAL ALIGNMENTS - linear, sub-
vertical geological features, where differentiated
movements occur along the length; and
DESCONTINUITY J - structures present in the bed
of the river at an elevation located between 240 and
250 meters, a sub-horizontal zone which is
weathered and fractured with a high rate of
permeability (spill 9.
The area is also characterized geologically by an
excessive, thick mantle of soil, saprolito, and
weathered rock, with a depth varying from 10,O to
20,O meters in the hillsides, and that could surpass
30,O meters in other localized areas (Antunes
Sobrinho, 1999 and Infanti, 1999).
3 BEARING CAPACITY PREDICTIONS
Non-confined simple compression tests were
performed on core samples of basalt rock with a
22% 5 R.Q.D. 5100% whose results, presented in
Chart 1, were submitted to a probability analysis in
the concept of Gauss, allowing a 95% margin of
reliability (a characteristic value):
Chart 1 - Results of the compression tests.
I$ CS I Rupture Tension -OC (MPa)
1 1 89.0
2 1 52.5

2.2 Geotechnical characterization of the test area 3 138.8

4 89.0
Keeping in mind that we are dealing with a design of 5 1 149.7
great size, the implantation area was thoroughly
investigated, through rotary core borings along with 6 1 65.3
special tests in the laboratory and field. 7 1 164.0
Figure 2 presents the results of rotary boring - 8 1 63.6
SM-02, used as a reference for the tested pile. In
9 1 105.0
agreement with this log of boring, the underground
stratigraphy is composed of a superficial layer of
silty clay, not very sandy, with a medium thickness Maximum Value (o~M~J = 164MPa
of 10m and an average SPT of 10 along with basalt Minimum value (OcMin) = 52.5MPa
boulders -having an RQD (Rock Quality Designa- Mean value (X) = lUlMPa
Standard deviator (s) = 40MPa
tion) varying from 0 to 48%. It has a recovery
Characteristic value (x) = 35MPa
between 38 to 98% - with a medium thickness of I;, ~ = X - 1 . 6 5 ~
6m, interspersed by a layer of silty clay, not very
sandy, and terminating in sound basalt rock with According to Zhang (1998), the ultimate capacity
veins of quartz - an RQD and a recovery varying for a pile embedded in rock can be estimated as
from 91 to 100% - to the limit of the boring. such:
(QSp = Qal+ Qp

in that:
(QJP = ultimate bearing capacity;
Qal = Lateral skin friction; and
Qp = point bearing capacity.

a) Skin Friction
Broms (1988), and apud Horvath et al. (1983) have
proposed the following equation for weak rocks:
z
, = (0.20 to 0.30).,/0, MPa

ZMax = maximum unit socket shear; and

10.2) to (0.30) = emoirical factor.
Figure 2 - Geotechnical characterization.

442
 To obtain the oc value, the pile embedded in the L
weathered basaltic rock having an RQD of 0 was d=l+0.4LI3.4 (3.5)
considered. According to the local geology and bs
geomechanical classification presented in Chart 2 in that:
(Lima, 1983, Bieniawski, 1984 and Barton, 1974), a L, = socket length, and
RQD <25% average was adopted, stated as, b, = socket diameter.
cy slOMPa. The following values found for the
unitary lateral friction (zh.ia~)and resistance due to substituting the values, we have:
the lateral fiiction (Qa$, are presented in Chart 3.
3.2
d =1+0.4-=1.98 I 3.4 (3.6)
I .3
CLASSIFICA TION
The evaluation of the end bearing capacity,
RQD whose results are presented in Chart 4, considers the
100 Excellent > 200 characteristics of the support rock, signifying a
recovery of 73 and a 48% R.Q.D. It has a
90 - 100 Excellent > 200
characteristic resistance of 0, of 35MPa, according
75 -90 Good I00 - 200 to the results of the probability analyses obtained
Regular 50 - 100
through simple compression rock samples (Chart l),
50 - 75
values which are confirmed in Chart 2.
25 -50 Weak 25 - 50
< 25 Very Weak 1-10 Chart 4 - Toe resistance.

According to Canadian Foundations Engineering

Manual (1985) and Ladanyi and Roy (197 1):

in that:
4,, =o,.k,d

Q M ~= maximum toe bearing pressure;

X

cr, =
average unconfined compressive strength
of rock core;
K ~ , , = empirical factor; and
d' = depth factor.

Chart 3 - Skin friction.

(3.3)

: CT~;M~~

o,M,,

35.0
0.13

9.01
0.17 0.20 0.13 0.17 0.20
164.0 42.21 56.20 64.94 56,140 73,420 86.370
101.0 26.00 34.00 40.00 34,580 45,220 53.200
52.5 13.51 17.67 20.79 17,970 23,500 27,650

11.78 13.86 11,980 15,670 18,430

It was considered, in the design phase, that the

ultimate capacities of skin fiiction and toe resistance
would be the following:

QUi 3270kN 2 to oc = lMPa, 7Max = 0.25MPa

Qp = 11,980kN s to oc = 35MPa, qMax = 9.01MPa
TMax (ma) and ksp= 0.13
(Q,Jp 3270kN + 11,980kN = 15,250kN
0.20 0.25 .0.30 0.20 0.25 0.30
10 0.63 0.79 0.95 8230 10,320 12,420 It has been verified that the safety factor for a
work load of 4400kN is 3,4, and therefore
8 0.57 0.71 0.85 7450 9280 11,100 satisfactory.
6 0.49 0.61 0.73 6400 7970 9540
5 0.45 0.56 0.67 5880 7320 8760 4 HIGH DYNAMIC LOAD TEST
3 0.35 0.43 0.52 4570 5620 6800
4.I Pile and Instrumentation Features
2 0.28 0.35 0.42 3660 4570 5490
2610 3270 3920 The conditions of construction and the test, are
1 0.20 0.25 0.30
represented in Figure 3.
1 1
--I k, I- (3.4)
8 5

443
Figure 3 - High Dynamic Load Test - Typical setup.

The pile was cast in place with the concrete The instrumentation proceeded through the
fck=25MPa (characteristic compressive strength) installation of four strain transducers and four
and heavy longitudinal steel reinforcement accelerometers, positioned every 90 degrees, in a
associated with spiral reinforcement cage. manner to minimize the possible effects of bending
Instrumentation for the Dynamic Load Test (DLT) resulting from the dynamic impact. In the case of the
and the test itself followed the recommendations of accelerometers, two piezoeletrics and two
ASTM D-4945/89, AASHTO designation T-298/93 piezoresistives were used. The opening of four
and ABNT-NBR 13208/94 (Brazilian Standard). "windows" became necessary in the steel casing, for
Constructively, the shaft space incorporated the installation of the sensors in the concrete
between the confining soil and the top of the nucleus, as represented in Figure 4.
alluvium of basalt is constituted by a steel casing Chart 5, which follows, specifies the
with a 3/8" thickness and an internal diameter of characteristics of the tests.
1,20m driven with a casing oscillator rig. In order to
carry out the test, a supplementary 47cm shaft of Chart 5 - Test features.
concrete was executed above the steel casing, COLUMN = P5
where the conditions of longitudinal reinforcement BORED PILE = T6
were maintained according to project, altering the
spacing of the spirals to 8,Ocm. LENGTHS (m)
4
SOCKET
The excavation of the rock was accomplished
LENGTH MU.
with the use of explosives and a compressed air BELOW BELOW
ROCK
system, for the purpose of expulsing underground TOTAL
SENSORS GL
water which originates in the hydraulic load of the (4 (tD
Uruguay River.
After the excavation in rock, the pile was 120 130 3.20 440 16.29 15.01 14.01
concreted to a quota of 333.62, corresponding to _. .
Figure 3. *Maximum worklng J

444
4.2 Hammer Description
Dynamic Load Test on cast in place piles with high
load capacity have had routine application in the
USA, Europe, Asia etc. and in Brazil. The technical
literature available gives an account of several cases
of dynamic tests on cast in place piles, as discussed
by Rausche and Seidel (1984), Balthaus et al.
(1985)Townsend et al. (1991), Seidel et al. (1996),
Sawai et al. (1996) etc.
The option of using the high strain dynamic load
test (DLT) instead of the static load test, which is
routine in many places, comes from the
attractiveness of the economics and the scheduling
benefits. The cost of the high strain dynamic test in
Brazil is 25% less than the cost of a static load test
(SML-slow maintained load), and represents a
reduction of 85% in time consumption.
The device designed to apply high impact forces
on the head of the bored pile consists of the free fall
self-propulsion hammer with special features to test
piles with high load capacity and variable diameters,
varying from 70cm to 250cm.
The selection of the proper hammer size is
essential for successful high-strain dynamic testing
hammer weight, and drop height. The cushion
details must be appropriately chosen so that hammer
impact causes sufficient pile movement in order to
mobilize the required soil resistance, and to assure
that dynamic stresses in the shaft will not impair its
structural integrity.
The hammer apparatus must be constructed in a
way to facilitate mobility around job sites and assure
a uniform impact to the head of the pile (Hussein et
al., 1996).
The free fall hammer conceived and in use in
Brazil consists of the following main components,
shown in Figure 3 : the steel striker plate, the
cylindrical - 2 pieces - steel cage (both of these
components being referred to as the hammer), the
set of round 20kN steel rams allowing modulation
up to 200kN in maximum drop of 3rr4 and the
hydraulic system to lift the ram by an automatic
catcher, which activates and controls the hydraulic
jacking equipment.
The steel striker plate consists of a round 2.1m
diameter by 100mm thick steel plate, with four
lateral latches that adjust themselves according to
necessities in relation to the diameter of the pile to
be tested, and functions as a hammer support on the
pile top and as a helmet.
4.3 Technical Criteria
To determine the hammer weight, drop height, and
plywood cushion thickness to be used in this
p&icular test, the bibliographic recommendations of
Likins (1994) and Hussein et al. (1996) were
followed:
445
- The hammer weight should, at least, be equal to
1.5% of the anticipated static test mobilized load;
- The hammer drop height should be approximately
8.5% of the pile length, with a minimum value of
2.0m; and
- A plywood cushion thickness of t = L2 / 2 0 ,
where thickness ( t ) is expressed in and pile
length ( L ) and pile diameter ( D ) are in meters, a
minimum value of lOOmm is suggested, with an
additional 15Omm being added when the pile
length exceeds 30m.
In this particular case, it was postulated that a ram
of 120kN, with a maximum free fall height of 2 . 5 ~
and a plywood cushion thickness of 150mm by 1.l m
diameter would be used to comply with the
recommendations.
The diameter of the plywood cushion should be
about 80 to, at most, 90 percent of the shaft diameter
in order to be able to centralize the blow. The
thickness of the steel striker plate to be placed above
the top of the pile should range from 50mm to
100- its minimum diameter being equal to the
diameter of the cushion the ram, ideally, is a simple
drop weight which can be raised to a variable height.
In order to facilitate, rams can be made in segments
which can be assembled at the job site. The ram to
be used needs to be “guided” to assure that the pile
top and ram bottom are perfectly parallel during
impact, in order to assure uniform contact stresses
during impact.
Between the steel striker plate and the top of the
pile a 20mm thick steel wired rubber pad was
installed with a 1.0m diameter in order to evenly
distribute compressive stresses on the pile head,
absorbing the dynamic impact, and preserving the
structural integrity of the pile top above the
electronic sensors, which measures data to be
associated with the test.
Figure 4 - Details of the windows in the steel casing
for the installation of the accelerometers
and strain transducers
4.4 Case Method and thefield DLT
The test was done by a sequence of blows with
variable drop heights as a dynamic load test with
increasing energy. References can be seen by Aoki
(1989), Niyama and Aoki (1991), Hussein et al.
(1992), Aoki and Mello (1992), Mello and Aoki
(1993), Beim and Aoki (1996), Mello and Paraiso
(1998), among others. It’s important to note that the
sequence of blows was done specifically on a
concrete shaft inside in steel case.
The DLT with increasing energy can be seen as a
cyclic test similar to a static load test with increasing
load steps where each loading cycle corresponds to
an increasing energy impact. The test started with a
drop height of 0.5m up to 2.5m with increments of
0.5m.
The input data supplied to the PDA-PAK95
equipment for the beginning of the DLT were:
e SP (pile specific weight): 24kN/m3
e AR (pile area on the sensors level): 11,309.8cm2
EM (pile dynamic modulus): 3 12tf7cm2
e EA/C (pile impedance): 1,007.31tfXm
e WS (wave velocity): 350Ods
0 JC (Case Damping factor): 0.5 (medium value)
At the same time precise millimetrical set
measurements were collected from the shaft of the
pile for each blow.
Figure 5 presents the force and velocity x EA/C
traces as function of time for each blow, obtained
from the strain transducers and accelerometers
processed by Case Method and reprocessed by
Datpro Program. These traces exhibit good
proportionality between force and velocity, and the
peaks of force show an increase as the energy from
the dynamic impact rises, which is perfectly
consistent.
4.5 Cupwup Analysis
In the Capwap analysis the pile was initially
modeled considering it’s uniform cross section.
However, in the case of the pile tested, it was
necessary to model impedance variations along the
depth into the rock for improving the “match”
between the curves of force and velocity measured
and calculated. In this particular case, the analytic
procedure was extremely necessary.
Chart 6 summarizes the results of the Capwap
analyses for each blow applied on the top of the pile.
Chart 7 shows the typical results from the Capwap
Analysis with reference to all five blows. It
considers the values of skin friction, in the shaft
space, inserted in the alluvium of the basalt,.
between quotas 321.00 and 317.80, and toe
resistance disregarding the portion of skin friction
446
Figure 5 - Force and Velocity traces.
developed internally in the steel casing and between
the steel casing and clay soil, as following:
H - drop height;
FMX = the maximum compression force;
EMX = the maximum energy transmitted past
the transducer;
%EMX = efficiency;
SET = permanent SET;
DMX = the maximum downward displacement
at the transducer location;
total mobilized resistance;
toe resistance;
lateral skin friction;
lateral quake;
toe quake;
Damping factor;
lateral damping factor; and
toe damping factor.
5 CONCLUSIONS
Having the results presented by Capwap Analysis
and Case Method as a reference it is concluded that:
0 The traces of force and velocity x EA/C presented
in Figure 6, shows with clarity the increase of the
load mobilized with the applied growing energy.
0 The liquid energy transferred to the pile grew with
each applied blow, arriving at 29% in the fifth blow,
a low value for the hammer in fi-ee fall without the
interference of mechanical friction. The reason of
the low transfer of energy is in consequence of the
great loss of energy with the shock absorber system
used.
e It is verified that the maximum mobilized
resistance of friction and point are manifested
respectively when the fourth and fifth blows and
confrmed with the obtained values of Qs and Qt
according to Chart 6.
 Chart 6 - Capwap analysis - general data.
-
KS
Rs Rs-SKIN
SKIN up between level R p Point RU Total Re- RU
FMX EMX % SET DMX
BLOW 'I to 321.00 321,OO and Resistance sistance RU (Rs + Rp)
(kN) (kN.m) EMX (mm) (mm)
(m) Level 3 17.80 (kN) (kN) (kN)
(kN) (kN) (kN)
~

1 0.50 1840 1661.0 179.0 4025.0 4204.0 5865.0 4030.0 9.0 15.0 0 2.72
2 1.00 4572 3148.5 1423.5 5101.0 6524.5 9673.0 5750.0 20.0 16.7 0 3.80
3 1.50 6621 361 1.5 3009.5 5180.0 8189.5 11,801.0 9020.0 37.0 20.6 0 4.85
4 2.00 6365 2938.0 3427.0 7050.0 10,477.0 13,415.0 12,460.0 61.0 25.4 0 6.21
5 2.50 4114 1387.0 2727.0 11.129.0 13,856.0 15,243.0 16,420.0 86.0 29.0 0 7.45

Chart 7 - Capwap analysis - socketed in rock.

SMITH
H RU RP Ks Qs Qt J CASE
BLOW
(m) (kN) (kN) (kN) (mm) (mm) CAPWAP
Js (s/m) J, ( s h )

1 0.50 4204.0 4025.0 179.0 1.377 1.030 0.640 1.754 0.87

2 1.00 6524.5 5101.0 1423.5 2.280 1.750 1.ooo 1.165 0.25
3 1.50 8189.5 5180.0 3009.5 3.002 2.587 0.701 0.876 0.44
4 2.00 10,477.0 7050.0 3427.0 4.158 3.900 0.7 13 0.75 1 0.52
5 2.50 13,856.0 11,129.0 2727.0 4.000 4.682 1.136 0.426 0.41

It is observed that the measured value of zero set
in all the blows leads to the consideration that the
pile still presented an additional point load to be
mobilized, however, the value of the load mobilized
in the fifth blow indicated factor of safety larger
than 3, in relation to the predicted working load,
being opted by the paralyzation of the test.
8 It is considered that the ultimate load mobilized in
the dynamic test can be interpreted in the following
way:
- Ultimate Bearing Capacity (5th blow) = 13,856kN
- Ultimate Bearing Capacity, considering the friction
resistance (4th blow) added to the point resistance
th blow) = 14,556kN
The test showed the saturation of the skin friction
in the 4th blow and maximum point mobilized in the
5th blow, whose obtained values present a close
approximation with those considered in the semi-
rnpiric analyses in the project phase.
The dynamic test in piles of high load capacity
has become a reliable tool and has reduced costs and
shortened evaluation deadlines for the quality of this
type of deep foundation, as long as the hammer to be
used presents the available potential energy to the
load to be mobilized.
6 ACKNOWLEDGEMENTS
The authors wish to thank the support given by
the Company CBPO-Grupo Odebrecht,and Mello de
Azevedo contractor responsible for the construction
of the bridge on the Uruguay River and, especially,
to engineer Claudio Menin, project planner and
responsible for the structural calculations of the
bridge.
7 REFERENCES
AASHTO T 298/93 - American Association of State
Highway Officials Standard - Method of test for
high-strain dynamic testing of piles.
ABNT NBR 13208/94 - Associaggo Brasileira de
Normas TCcnicas - Estacas Ensaio de
Carregamento Dinilmico, in Portuguese.
ASTM D4945/89 - American Society for Testing
and Materials Standard - Test method for high-
strain dynamic testing of piles.
Antunes Sobrinho, J. et al. - 1999 - Criterios de
projeto na construqgo da barragem da UHE Ita.
XXIII Seminario Nacional de Grandes Barragens.
In Portuguese, Belo Horizonte, Brazil.
Aoki, N. - 1989 - A new dynamic load test concept.
Proc. Discussion Session on Pile Drivability, XI1
ICSMFE, Rio de Janeiro.
Aoki, N. & Mello, V. F. B. - 1992 - Dynamic
loading test curves. Proc. Fourth International
Conference on the Application of Stress-Wave
Theory to Piles. Hauge, Holland.
Balthaus, H. G. et al. - 1985 - Dynamic load test -

447
 German practice. Proc. 1lth Int. Cod. Soil Rausche, F.- 1988 - High strain testing of drilled
Mechanics and Foundation Engineering, San shafts. Seminar for dynamic testing of cast in situ
Francisco, USA. piles for capacity and integrity - Boulder, CO.
Barton N., Lien, R. & Lunde, J. - 1974 - Engineer- Seidel, J. P. et al. - 1996 - Dynamic testing of
ing classification of rock masses for the design of barretes for a cement silo project. Proc. Fifth
tunnel support. Rock Mechanics, 6 (4), 189-236. International Conference on the Application of
Beim, J. W. & Aoki, N. - 1996 - Dynamic load test Stress-Wave Theory to Piles. Orlando, USA.
method with variable energy. Proc. Fifth Townsend, F. et al. - 1991 - Dynamic load testing
International Conference on the Application of of drilled shafts - Final Report, Univ. Florida,
Stress-Wave Theory to Piles, Orlando, USA. Gainsville, USA.
Beim, J. W. and Paraiso, S. C.. - 1992 - Dynamic Zhang, L. & Einstein, H. H. - 1998 - End bearing
testing of enlarged base Franki piles. Proc. Fourth capacity of drilled shafts in rock. journal of
International Conference on the Application of geotechnical and geoenvironmental engineering,
Stress-Wave Theory to Piles, Hauge - vol. 124.
Netherlands.
Bieniawski, Z. T. - 1984 - Rock mechanics design
in mining and tunnelling. Balkema, Rotterdam.
Broms, B. B. et al. - 1988 - Bored in residual soil
and weathered rocks in Singapore - BAP I -
Ghent - Belgium.
Canadian Foundation Engineering Manual - 1985.
Capwap Users Manual - 1996 - Goble Rausche
Likins and Associates, Inc. - Cleveland, Ohio,
USA.
de Mello, L. G. & Paraiso, S. C. - 1998 - Variable
energy dynamic load test on a 1.0m diameter
CFA pile. Proc. Belgium - BAP 111.
de Mello, V. F. B. & Aoki, N. - 1993 - Updating
realism on large diameter bored piles. Deep
Foundations on Bored and Auger Piles, Ghent,
Belgium.
Hussein, M. et al. - 1992 - Dynamics of pile driving
as a function of ram drop height. Proc. Fourth
International Conference on the Application of
Stress-Wave Theory to Piles. Hauge, Holland.
Hussein, M. et al. - 1996 - Selection of a hammer
for high strain dynamic testing of cast in place
shafts. Proc. 5'" Int. Cod. on the Application of
Stress Wave Theory to Piles, Florida, USA.
Infanti, N. et al. - 1999 - Tensees residuais nas
obras subtenheas da UHE Ita. XXIII Seminario
Nacional de Grandes Barragens, Belo Horizonte,
Brazil.
Lima, M. J. C. P. A. - 1983 - Prospecqiio
Geotecnica do Subsolo - Livros Tecnicos e
Cientificos Editora S. A., Rio de Janeiro, Brazil.
Niyama, S. & Aoki, N. - 1991 - Correlaqilo entre
provas de carga diniimica e estatica no campo
experimental da EPUSP/ABEF. In Portuguese,
2rd -Seminario de Engenharia de FundaGdes
Especiais e Geotecnia, SEFE, Siio Paulo, Brazil.
Rausche, F. & Seidel, J. P. - 1984 - Design and
performance of dynamic tests of large diameter
drilled shafts. 2"d Int. Conf. on the Application of
Stress Wave Theory on Piles, Stockholm.
Rausche, F. et al. - 1985 - Dynamic determination
of pile capacity. Journal of Geotechnical
Engineering, Vol. 111 - Nc 3 - ASCE: 367 - 383.

448
7 SPT measurements and special field monitoring test
This Page Intentionally Left Blank
 Application of StreSS-WaVe Theory to Piles, Niyama & Beim {eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Keynote lecture: Frequency characteristics of stress wave and penetration

during SPT

K. Fujita
Department of Civil Engineering, Science University of Tokyo,Noda, Japan

ABSTRACT Recently wave equation analysis has been successfully applied to the SPT to determine the
energy transfer ratio for adjusting the N-value, but should not limited to this purpose The SPT N-value has
been applied to the design of foundations for more than SO years, however, the correlation between the N-
value and the design parameters have been estimated crudely based on experience, so that new methods or
constitutive equations should be established for determining the design parameters rationally. The frequency
characteristics of the stress wave and the behaviour of the penetration of the rod system are introduced,
because the relation between them could be keys to hrther studies

1 INTRODUCTION integrating the velocity, which was equal to the

As the ISSMFE was unable to standardize the
method and equipment to be used for the SPT
internationally, the SPT Working Party of the
ISSMFE Technical Committee on Penetration
Testing proposed the “International Reference Test
Procedures on SPT”, at ISOPT-I, Orland in 1988. A
method using stress wave equation analysis was
introduced in Document 4- Calibration Methods for
the adjustment of the N-values in the SPT to account
for different magnitudes of driving energy, as a
means of quality control for the SPT. The method
was established by referring to studies made by Uto
et a1 (1974), Schmertman et al. (1979) and others In
the author’s opinion, the calibration methods are
unable to account for a difference in the size of
equipment, however, this could be corrected by
calculations and the formula to be applied for the
adjustment of N-values should be revised.
Matsumoto et al. (1992) have applied the method
of two-point strain measurements proposed by
Lundberg & Henchoz (1977) to the measurement
and analysis of a SPT experiment carried out in the
field The displacement-time diagram, the
transmitted energy-time diagram and stress-time
diagram were verified, and the calculated data
corresponded well with the measured data
Dynamic measurements of the energy transfer
ratio have been carried out in practice in USA and
other countries with increasing frequency. Abou-
matar et al. ( 1 996) have presented a typical dynamic
SPT records including the measured force, adjusted
velocity and displacement with time computed by
observed final set as shown in Figure 1 The wave
equation soil constants, i e the quake, damping
constant and mass for use i n the wave equation
analysis, as well as the static soil resistance obtained
from the calculated toe force and motion were
determined reasonably well
Abou-matar & Goble ( 1997) have contributed a
valuable paper which broadly discusses the
problems associated with the dynamic measurements
and analyses concerned with the SPT They pointed
out that changes in the cross section of the drill rod
may cause substantial changes in the measured N-
value and presented method for detecting and
avoiding various problems based on one-
dimensional wave mechanics
Fujita & Ohno (2000) introduced studies
concerning the application of the stress wave theory
to the SPT in Japan, and included a figure that
indicates that the SPT hammer strikes the rod more
than three times afier the first impact They
suggested that analysis of the frequency
characteristics of the measured stress wave would
provide valuable information for the study of the
SPT concerned
The term = f i known as “wave propagation
speed’, etc is applied in one-dimensional wave
mechanics and assumes that the rod length is infinite
or the rod area (diameter) is negligibly small The
wave propagation speed in the anvil, rod connector
or hammer is about 60 to 70% of c So, the analysis
should be take into consideration the frequency of
the wave

451
 h

E
E
c
4

5
E
y
-cac
6
d
2 -44- - -4
0 P O I C(hN)
~ \'elocity.(LA,c) (IrN) --Divplaceinent (mtn)
CL
I I I I I I I
0 I r) 70 30 40
Time (ms)
Figure 1 Dl ii;uiuc SPT ineasureinent records - force, L elocih and displacement
(aficr Abou-matar et a1 1096)

\ netrdtiori vs time
l'enetratiori
Final penetration
II

.Maximum penetration

I
'50
10 20 30 40 50 GO I0
Time ( ms)
Figurc 2. Strcss and pcuctration L'S. tiinc diagram. (Fujita 2000).

In this paper, the penetration or vertical
movement of the rod, the frequency characteristics
of the stress wave and the relation between the rod
penetration and the frequency of the stress wave will
be introduced, while the study is still proceeding.
2 PENETRATION VS. TIME DIAGRAMS
Figure 2 is an example o f a diagram showing records
of the penetration of the rod with time and the stress
in the rod with time (stress wave) measured near the
top of drill rod during one blow of the SPT. The
figure shows that the hammer strikes the rod again
(second impact) about 14ms after the hammer blow
(first impact) and the penetration increases
accordingly. The penetration at the time of the
second impact is smaller than the maximum values
of the penetration which occurs between the first and
second impacts due to the rebound of the rod and/or
ground before the second impact. In this case, during
the completion of the SPT there are at least two
impacts of the hammer, and this evidence should be
considered in the wave equation analysis.
Figure 3 is a diagram summarizing the measured
penetrations of the rod as a finction of time during
the part of the SPTs at a depth of l l m . The
maximum penetration and the final penetration for
each blow are around 12 to 13mm and 4 to 7mm,
respectively, and the rebounds of the rod and/or
ground are between 6mm and 8mm. While the final
penetrations (set) per blow which contribute to the
N-value are scattered, the maximum penetration per
blow is almost identical in all cases.
Attention should be paid to this from the stress
wave analysis point of view, because the energy

452
 Figure 3 Penetration 1 s time diagram - at a depth of 1 l n i (Fujita 2000)
Figure 4 Relation betncen the iiia\iiiiuiii pcne-
tratioii and final penclration (Fiijita 2000)
transmitted to the rod is likely to correspond to the
masinium penetration, not only to the final
penetration So, dynamic measurements of the
penetration (vertical displacement) should be made
The strain or stress and the penetration of the rod
were measured about 2 to 2 5m aboveground, near
the top of the drill rod and below the anvil, in the
453
case shown in Figures 3-4, in ground which consists
mostly of sand and gravel layers of alluvial and
diluvium strata, and where the water table is 1 10m
below the ground surface and the N-value is smaller
than 50
Figure 4 shows the relation between the
maximum penetration and the fhal penetration of
the rod, (a) on a linear scale and (b) on a logarithmic
scale measured at various depths during the SPT,
without adjustment for the energy transfer ratio The
energy transfer ratio varies from blow to blow in this
case, mainly between 0 6 and 0 9, with an average of
74
Values of the maximum penetration smaller than
12mm were not observed during the SPT at this site,
except for the case when the energy transfer ratio is
small When the maximum penetration is about I 3 to
14mm,the final penetration will be less than about
1 Omni When the maximum penetration is larger
than about 22mn1, the difference between the
maximum penetration and the final penetration is
less than about 21nm The maximum penetration
occurs after the time (2L/c), in this case
3 FREQUENCY CHARACTERISTICS OF
STRESS WAVES
The straidstress with time measured by strain
gauges installed near the top of the drill rod as a
function of time is called the “stress wave” in this
paper, and an example was shown previously in
Figure 2 In the wave equation analysis of the SPT,
the hammer, guide rod, anvil, drill rod, coupler and
sampler are generally assumed to be made of the
same material and to have the same cross-section.
except for the hammer
The downward travelling wave arid upward
travelling wave caused bv the hammer blow are
propagate in the rod in the downward and upward
directions, respectively, at the wave propagation
speed c- = ,/A J>= . i i 111
~ for the material, based on
one-dimensional wave dynamics However, the
speed depends on the shape of the members, and
decreases when two-dimensional wave dynamics are
applied For example, the speed in the hammer or
anvil would be about 70% of the one-dimensional
wave propagation speed for the material
The waves are partially reflected at the ends of
the rod system and at the junctions between the
different cross sections so that various waves are
propagated in the rod following the hammer blow In
order to examine the frequency characteristics of the
stress waves, the method of Wavelet transformation
was chosen, with which it is possible to obtain a
thi-ee-dimensioiial diagram giving the frequency,
time and power spectra
Figure 5 shows two frequency vs power
diagrams obtained by Wavelet transformation of
stress waves, measured at (a) a depth of3m and (b) a
depth of 17m The natural frequencies of the rod
systems employed at (a) and (b) are 0 286kHz and
t I i I I The unit of power is ((N/mm2)s)2,where the .s is
I Pztietration 71nm
001 01 1 10
Frecluel1cy (hHz)
(b) at a depth of 171n
figiirc 5 Pciictration 1 s poucr spcclrum (Fqita 2000)
454
0 1 ISkHz, respectively, assuming the wave
propagation speed is = Jty. while the
predominant frequencies shown in Figure 5 (a), (b)
are calculated to be 0 19kHz and 0 074kHz,
respectively, which are about 65% of the natural
frequencies
In Figure 5 , more than four prominent
frequencies can be seen including the predominant
frequency, these peaks are referenced to as A, B, C,
D ..., from low to high frequency The natural
frequency is located between the frequencies of the
peaks B and C According to the time-power
diagram (not included in this paper), the peak B
occurs about the time of the second impact between
the hammer and the rod
As shown in Figure 6, the reciprocal of the time
between the first and the second impacts
corresponds well with the fiequency at the peak B
4 RELATION BETWEEN POWER AND
PENET RATION
Figure 7 indicates that the power at the peak A
corresponds to the final penetration of a hammer
blow of the SPT The frequency at the peak A is
estimated to be around 15 to 20Hz, which is
probably determined by the duration of the rod
movement of 50 to 70 ms, as shown in Figure 3
included in the term is the data sampling inteival,
5 s in this case The power is an index provided i n
the Wavelet transformation, which is probably
related to the penetration and the stress in the rod as
well as the dynamic and static soil parameters,
however, it has not yet been clearlv identified in the
study ofthe SPT
Figure 6 Reciprocal of the tiinc bctnceii 1 “ and 2’ld 1111-
pacts plottcd agaiiist fiequcnc! at peak B (Fujita 2000)
 nicnsiircnieiils on SPT. llriic. .>" Iiii. 'oil/. n i l , l ~ ~ p l ~ c o i i o ~
?%lV)i?' if, I'ilC.V, Orlniltl. 16.3- 175.
oj'.sfrf~.~.~-lCnl~f~
Abou-inarar. H.. X: G. Goblc. 1997. SPT dynamic analysis and
tnc:isiircincnIs. .1. o/( iGK . I
Fuiita. K. 2001). Strcss vavc i k o n iipplicilion 10 standard
pcnctr;itioii ICSI in japan. /+iic. c;" iiii. ( ' o i i t iiiJ .i/~p/icii/ioi7
~/.S/rc.v.v-~rmv /h<wcl'/oPilc,s, S m fmilo. (to be publislicd)
ISSMFE Tccllnical Coininittcc on Pcnctratioii Testing-SPT
\\.orking party. 1988. Standard Pcnclration Tcst (SPT).
Irircrt~:itionnl rcfcrciicc tcst proccdurc. Iri Dc Kuilcr (cd).
7?s/iilg I9HX. 0 r I ~ i I d: ( 1 ).7-26. Rotterdam :
/'c~I1~~ir(J/IOIl
B;i Ikcnia.
Lundbcrg. B.. X: A. Hcnchoz. 1977. Analysis o f el;islic w i ~ c s
froin two-point slmin mcasurcmcnl. .I. of' I*~.qurii~rei~/.
.\/<.ch. 17(6): 213-218.
Matsumolo. T.. H. Sckiguchi. H. Yoshida. X: K. Kiln. 1992.
SigiiBc;incc or two point strain nicasiirciiicnl in SPT. Siii1.s
nntl I'iiiiI~tlnikin.s.32(2) : 6 7 4 2
Uto. K.. M. Fuyiiki. H. Kondo. M. Morilwra. (G H. Mmniniira.
197-1. Considcmtion of .\-\.alrie nnd strcngth of ground
froiii a vicn point or w v c tlicon. I'roc.. t*hc.~il!\. of'
Liiginwriivg, 7bkni [ hiiw-si/,c. (in Japanesc).
Figtire 7. Penetration plotted against power st pcak A.
(Fujita 2000).

5 CONCLUDmG REMARKS

Since Terzaghi & Peck published their book "Soil

Mechanics in Engineering Practice", in 1948, the
SPT N-value has been widely applied to the practical
design of foundations, because correlations are given
between N-values and the relative density.
consistency, angle of internal friction, bearing
capacity of footings and earth pressure etc. although
these were estimated crudely based on experience
There are probleins involved in their proposals,
for example, the etyect of the overburden pressure on
the N-value is not accounted for, the energy transfer
ratio is not considered, the SPT equipment and
procedures are not standardized and there is no
instrumentation
The SPT N-value has been regarded as an index
of the shearing strength of the soil, rather than the
shearing strength of the soil, until now. However,
dynamic measurements and the analysis by applying
the one-dimensional wave equation to the SPT are
now available, and the energy transfer ratio and the
static resistance of the rod system can be estimated.
The time has come to obtain the shearing strength
of the soil directly from the SPT by incorporating
dynamic measurements and the analyses and to
determine the design paratneter rationally. The
frequency characteristics of the stress wave and the
behaviour of the penetration could be keys for
further studies, as well as for establishing
constitutive equations.

REFERENCES
Aboii-matar. H.. F. Rausche. G. Thendean. G. Linkins. Xt G.
Goblc. 1996, Wave equation soil constants froin dynamic

455
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3

The application of energy conservation Hamilton’s principle to the

determination of energy efficiency in SPT tests

N.Aoki & J.C.A.Cintra

University of S8o Puulo, S8o Carlos, Brazil

ABSTRACT: The Hamilton’s principle is applied to the analysis of variations in the kinetic and potential ener-
gies and in the work done by the non-conservative forces, during the dynamic SPT tests. This paper presents a
mechanically sound conceptual analysis of the energy transfer in the SPT hammer blow event different fiom
that of the energy content of the first compression wave, as proposed in the ASTM D4633-86.

1 INTRODUCTION where 6 = variation in the time interval (t2 - t,), T =

total kinetic energy in the system, V = potential en-
The equations of the Newtonian impact (de Mello ergy in the system, W,, = work done by non conser-
1971) and more recently the wave equation theory vative forces (including damping).
(Abou-matar & Goble 1997) have been traditionally Any reference system of coordinates can be used.
applied to the analysis of energy lost in the SPT tests. The first application of the Hamilton‘s principle for
The efficiency factor q of energy transfer is defined foundation engineering was the analysis of driven
by the relationship between the measured energy of piles under increasing energy dynamic loading test
the first compression 1ongitudin:l wave and the nor- (Aoki 1997).
malized SPT potential energy T = 474 J. This theo-
retical efficiency would be proportional to the rod
length (Schmertmann & Palacios 1979). 3 EFFICIENCY IN THE SPT TEST
The energy conservation Hamilton’s principle is
more generalized than Newton’s law and can be re- 3.1 Traditional definition of SPT efficiency.
garded as the fundamental law of dynamics (Lang-
haar 1962). According to this principle, during the The 623 N hammer weight impacts the anvil over the
sampler penetration time interval, the actual applied rod-sampler-soil system assemblage (Figure 1).
kinetic energy ENTHRU is transformed into elastic
deformation energy of the rod-sampler-soil system
and into work done by the non conservative forces
(includins soil damping) acting on the sampler.
It must be observed that the ENTHRU is actually
independent of the rod length although their evalua-
tion is done by the measurement of deformation en-
ergy of the rod-sampler-soil system. Indeed the SPT
blow efficiency will be related to the kinetic energy
that reaches the sampler-soil system. In this instance
it is showed that the actual efficiency is inversely Figure 1. Standard Penetration Test system
proportional to the rod length.
The impact height h = 0.76 m, the total length of
rod is C and the sampler length is s. The anvil is lo-
2 HAMILTON‘S PRINCIPLE cated at the origin of the vertical reference 0-x axis
and the top of the sampler is the section A. The ki-
The variational expression of Hamilton‘s principle netic energy of the first compression wave at the top
(Clough & Penzien 1975) is of the of rod section is defined by

T, = ta F(0, t ) . v(0, t ) dt (2)

457
where T, = energy content of the first compression
wave, F(0,t) and v(0,t) = force and velocity at the
section 0, and t, = cutoff time at the arrival of the
first tension pulse.
Figure 2 presents the corresponding force-time
diagram in the section x = 0 at the top of the rod.
Figure 2. Force-time diagram at the top of the rod.
The efficiency factor q (Schmertmann & Palacios
1979, ASTM 1986 & Belincanta 1998) is
-r
11 = -+.loo
T 7
where T, = eqergy content of the first compression
wave, and T = normalized SPT potential energy
(474 J)
Figure 3 presents a typical shape of the curve re-
lating the efficiency factor and the rod length.
0 20 40 60 80 100
Figure 3 Efficient! factor 7 and rod length relationship
(Sclimcrtiiiann & Palacios 1979)
This figure suggests that the efficiency factor q
tends towards zero in the case of short rod length,
and towards a maximum for rod length greater than
15 m. This is an important feature that will be dis-
cussed thereafter under the light of Hamilton's prin-
ciple.
3.2 ICedefinition of SP7' effrcrencj..
The maximum energy applied in the top of rod-
sampler-soil system is the well known ENTHRU
T,, = f" F (0. t ) v (0. t ) dt
0
where To = maximum energy applied to the rod-
sampler-soil system (ENTHRU), F(0,t) and v(0,t) =
force and velocity at the section 0.
The maximum energy transferred to the top of
sampler-soil system is expressed by
(5)
where T,\= maximum energy transferred to the top
of sampler-soil system, F(e,t) and v(t,t) = force and
velocity in the section A, and t, = elapsed time for
the compression wave to reach the section A.
Figure 4 presents the energy wave in the x = 0
section at the top of rod-sampler-soil system and in
the x = /1 section at the top of sampler-soil system.
kinetic energy
1'
(3) time (t)
Figure 4. Kinetic energ! naves at thc top of the rod aiid at
the top of the sampler sections
The maximum value of kinetic energy at section 0
is T,, and the maximum value of kinetic energy at
section A is T, Corresponding to the particular solu-
tions of stress wave equation for these sections.
Although the kinetic energy defined by either
equation (2) and (4) is referenced to the same section
x = 0 (top of rod-sampler-soil system), it is empha-
sized that the kinetic energy T, considers only the
time integration correspondent to the first tension
wave while T,,is the maximum applied kinetic energy
Considering that the kinetic energy expressed by
equation ( 5 ) is responsible for the deformations of
sampler-soil system below the point A, it is proposed
to redefine the SPT efficiency by
1
where q" = efficiency measured at section A, T,\ =
maximum energy transferred to the sampler-soil sys-
tem, and T"= normalized SPT potential energy (474
J)
to /he SPT
3 3 Ayplicnfron of HciniiItoti',sp7-1~1c1~3le
coiistnrit energy 4mnniic kondrrig test.
Let us consider the application of Hamilton's princi-
(4) ple to the time interval (t2 - tl), where tl is the time
when the compressive wave arrives the top of rod-
sampler-soil system, and t2 is the time when all the
458
kinetic energy is dissipated at the end of the hammer
impact.
Figure 5 presents the displacement - resistance
curve in the section x = 0 at the top of rod-sampler-
soil system.
Figure 5. Resistance - displacement curve at the top of
rod-sample-soil system.
According to figure 5 the maximum kinetic energy
reaching the rod-sampler-soil system at the time tl is
where T = total kinetic energy, and To = maximum
energy applied to the rod-sampler-soil system
(ENTHRU) .
The maximum downwards displacement (pmax)
corresponding to the maximum mobilized resistance
(R), occurs at the point B of the curve, when the
velocity becomes equal to zero and the kinetic energy
T is transformed in potential deformation energy ac-
cumulated in the rod-sampler-soil system, not taking
into account others energies losses,
where V = potential deformation energy accumulated
in the rod-sampler-soil system, equal to the area
OBCO in figure 5.
At the end of hammer impact, correspondent to
the point D of the curve, the potential deformation
energy is transformed in work done by non conser-
vative forces and in elastic recovery energy
(9)
where V = potential energy in the system, W,, =
work done by non conservative forces including
damping = area OBDO, and V, = elastic recovery
energy = area BCDB.
At the time t 2 after the unloading it is possible to
record the displacement components due to the elas-
tic and permanent deformations
459
where pmax = maximum downwards displacement, pe
= recovered elastic displacement, and pp= permanent
sampler penetration.
The recovered elastic displacement is correspon-
dent to the elastic energy of deformation accumu-
lated in the rod-sampler-soil system. Most of this en-
ergy is due to the energy accumulated in the rod
length t and in the soil beneath the sampler.
The permanent sampler penetration is directly
proportional to the work done in the rod-sampler-soil
system. More efficient is the hammer blow greater
will the permanent sampler penetration. This obser-
vation shows that the definition of the SPT blow effi-
ciency is best correlated to the work done and not to
the applied kinetic energy.
Now, let us consider the application of Hamilton's
principle to another time interval (t2 - tc), where t, is
the time when the compressive wave arrives the top
of the sampler-soil system, and t2 is the time when all
the kinetic energy is dissipated at the end of the
hammer impact.
At the moment of the maximum penetration of the
sampler into the soil, the kinetic energy TA is trans-
formed into deformation potential energy of the sam-
pler-soil system only. After unloading, this deforma-
tion energy is transformed into work done in the
permanent penetration of the sampler into the soil
and in recovered elastic energy accumulated in the
sampler-soil system.
The difference (To - TA) in figure 4 can be consid-
ered roughly equal to the elastic energy accumulated
in the rod length t.
When the rod length tends towards zero the ki-
netic energy TA tends towards To and the efficiency
measured at the top of the sampler-soil system be-
comes a maximum. Considering the expressions (2),
( 5 ) and (6) it can be concluded that in this case the
kinetic energy T, becomes zero and the efficiency
measured at the top of sampler is actually a maxi-
mum and not a minimum as indicated in figure 3.
Figure 6 presents :he probable relationship be-
tween the efficiency q measured at the top of sam-
pler-soil system and the rod length.
0 20 40 60 80 100
0
efficiency (Yo)
5
Figure 6. Efficiency factor q' and rod length relationship
 In the case of very short rod the blow efficiency de Mello, V.F.B. 1971. The standard penetration test. Proc.
wfi be independent of the rod length. For small rod PanameriCan c o n ! soil Mech. Found. Engrg.., 4,San Juan
Puerto Rico, June 1971, 1: 1-87.
length and’or in the Of and very low re- Langhaar, H.L. 1962. EnergV methods in applied mechanics.
sistant soil below the sampler, the applied kinetic en- New York: John wiley.
e r a To = TA Will be transformed into work done in Schmertmann, J.H. & A. Palacios 1979. Energy dynamics of
the sampler-soil system, during the permanent sam- SPT. J. Geotech. Engrg. Div., ASCE, 105(8), 909-926.
pler penetration, and Terzaghi, K. 1942. Discussion on pile driving formulas. Proc.
ASCE, 68(2), 3 1 1-323.
pmax = pp (1 1)

In this case the elastic soil deformation energy

can be neglected and the kinetic energy is almost all
transformed into work done during the permanent
penetration of the sampler into the soil, resulting

In the case of elastic soils the elastic recovered en-

ergy will be proportional and the work done in the
permanent penetration inversely proportional to the
compressibility of the sampler-soil system.
Finally does the failure load of sampler soil system
be always reached? Yes, for SPT in the range I < N
5 60, considering that the permanent penetration for
one blow is greater than 10% of sampler diameter, to
be in accordance with Terzaghi 1942 small pile rup-
ture definition. For N>60 and in the cases where
elastic recovered energy predominates N has no
meaning as soil parameter.

4 CONCLUSION
The SPT efficiency calculated by the expression (6)
and the energy conservation Hamilton’s principle
showed that: a) this efficiency is inversely
proportional to the rod length and, b) although not
proposed, the best definition of the SPT blow
efficiency would be related to the work done in
permanent penetration of the sampler and not to the
applied kinetic energy.
Finally it can be remembered that the energy
transferred by the SPT hammer impact not always
mobilizes the ultimate resistance of the soil.

REFERENCES
Abou-matar, H. & G.G. Goble 1997. SPT dynamic analysis
and measurements. J. Geotech. & Geoenviron. Engrg.,
ASCE, 123( 10), 92 1-928.
Aoki, N. 1997. Evaluation of the ultimate bearing capacity of
piles driven with increasing energy (in Portuguese). Doc-
torate thesis, University of Sao Puulo, Sao Carlos, Brazil.
ASTM 1986. Standard test method for stress wave energy
measurement for dynamic penetrometer testing systems;
D4633-86, West Conshohocken, Pa., 775-778.
Belincanta, A. 1998. Evaluation of intervening factors in the
penetration resistance of the SPT (in Portuguese). Doctoi’-
ate thesis, Universify of SuoPuulo, Sao Carlos, Bruzil.
Clough, R.W. & J. Penzien 1975. Dynamics of structures.
New York: McGraw-Hill.

460
 Application of Stress-Wave Theoryto Piles, Njyama & Beim (eds)02000 Balkema,Rotterdam, ISBN 90 5809 150 3

Correlative study of Smith damping coefficient and SPT blow count

Robert Y. Liang
Civil Engineering Department, University ofAkroiz, Ohio, USA

CT: The accuracy and consequently the usehlness of wave equation analysis in pile driving problem
are largely hinged upon accurate input of wave equation soil parameters that are representative of the soil
conditions at the site. Since Smith model parameters have been used in most wave equation analysis
techniques, emphasis is given to the development of algorithms to determine Smith model parameters from
static ultimate soil resistance R, (including shaft resistance f, and toe resistance R,), quake Q and damping J. In
this paper, a methodolozy for using SPT values to determine Smith model parameters is developed.
The approach used herein to develop usefbl correlations consists of a number of consecutive steps, where
the static soil resistance is first determined, soil quake Q evaluated for a variety of soil types, and finally, the
damping coefficient J is determined by computer simulation of the standard penetration test with the help of
GRLWEAP program.
The results obtained indicate the usehlness of the adopted method when compared to existing cases anaiyzed
by the CAPWAP.

1. DETERMINATION OF STATIC CAPACITY 2. DETERMINATION OF QUAKE Q FROM SPT

Liang (1999) described a method, and put down two 2.1 Measured Quake in Piles
sets of equations to estimate soil resistances from the
standard penetration blow count number SPT-N for Quake is defined as the maximum movement of the
different pile types and different soils. In the present pile relative to the soil in order to hlly mobilize
study, the adjustment factors m,q~ and ~1 are all ultimate resistance to pile movement (or moving
assumed to be unity for both shaft and toe pile). Terzaghi and Peck (1967) indicated that the
resistances. Thus, the shaft and toe resistances can be maximum shaft resistance would be developed in
calculated as follows. clays at a pile movement equal to about 1% of the
pile diameter. Coyle and Reese (1966) indicated that
a) Shaft Resistance (kPa) the maximum shaft resistance would be developed in
clays with approximately 6.3 mrn (0.25 in.) of pile
movement. Reese and O'Neill (1 989) provided
Clay 7I generalized results from numerous load tests. For
cohesive soils, they indicated that full development of
366.67,V Y
._ the shaft resistance occurs at a settlement of about
f' =-
" 110+.V
+3.5Ne j Sand J 0.5% of the shaft diameter, while the toe resistance
requires about 5% of the shaft diameter to be hlly
b) Toe Resistance ( H a ) mobilized. Cohesionless soils required settlements of
about 1% of the diameter of the shaft for a hll
development of shaft resistance, and in excess of
Clay 10% of the shaft diameter to hlly mobilize toe
resistance.
Sand Ho (1994) measured the t-z curve in compression
bored piles in Sinsapore old all-jvium.He graphically

46 1
showed the mobilized curves at pile shaft and toe. He
indicated that the shaft resistances increase with
increasing applied loads, with no signs of leveling off
in the mobilization curves. As for the pile toe, he
indicated that the trend sharply increases in a more
like curvilinear manner with settlement measured at
pile toe. A summary of soil quake values as deduced
from the SPT-N, and pile tests is presented in Table
1, below.
Table 1 Summary of Measured Quake Values as
Obtained from the SPT-N. and Pile Load Tests *
Author
(1967)
model parameters were also obtained via CAPWAP
match procedure. The database can firther be broken
down into two sets. One set of data contains 208 pile
cases (120 individual piles) that have both dynamic
testing and subsequent static load test results. The
second set of data contains a total of 403 pile cases
with only dynamic pile testing results.
To investigate the significance of soil type and
time of dynamic testing (i.e., end of driving- EOD,
and Beginning Of Restrike- BOR), a statistical
analysis of the Smith model parameters was
conducted on the Paikowsky’s database, particularly
the quake and the damping factor. Table 2 lists the
means and standard deviations of the Smith model
parameters for different soil types.

(1966) Table 2 Statistical Analysis of the Smith Model

Parameters Based
(1989). cohesive soils. Clay Sand iU1 Soils ’

(1989), Non-coh. Mean 0.30 0.-30 0.30

Soils. 0.15 0.15 0.15
Ho (1994) 0.25 (0.04-
0.46)
Mean 0.53 0.64 0.61
030 I 036 0.36 j

0.70 0.47 0.60

2.2 GRL’s Quake Values from CAPWAP Analysis
Std Dcv 0.60 0.40 0.53
One of the most frequently used methods of
Mean 0.77 0.57 0.63
determining the Smith model parameters is the back-
calculation of those parameters via the CAPWAP- s/m Std Dev 0 77 0.67 0.70
program (Goble, et al, 1988). Essentially, the
measured velocity-time records from PDA constitute
the input file to calculate the force-time histories. By Table 3 shows the statistical analysis results based
adjusting Smith model parameters iteratively, one can on the time of dynamic testing, regardless of soil
obtain a good match between the calculated and type Finally, Table 4 shows the combined statistical
measured force-time histories, from which the Smith analysis results, considering both soil type and time
model parameters can be identified. GRL performed of testing. From these analysis results, one may
the CAPWAP analysis of 82 piles in the FHWA
report (1988). Based on their work and data, an
average soil quake value of 0.10-in can be suggested Table 3 Statistical Analysis of Smith Model
at the pile shaft for all types of soils with a tolerance Parameters based on Time of Testing (EOD and
ranging from 15 to 27%. As for the quake at the pile BOR).
toe, sandy soils were found to have the largest values
and the least variation, that’s to say using an average
value of 0.39-in for piles in sand will produce an
error of no more than 5% of the actual value. For
orher types of soils (clays, silts, gravel with sand,
rock. or a mixture, a value 0.22 in is recommended
with a tolerance of no more than 25% of the real pile
toe quake.
0.50
2.3 Paikowsky’s Quake Value from CAPW-ZP 0.50 0.53
Method
0.43
J,. s/m
A data base consisting of 61 1 piles has been collected
0.13
by Paikowsky, et aL(1994). In their report, the smith

462
observe that the quake does not vary sigdicantly
with the soil type and the time of dynamic testing.
However, the damping factor seems to be affected
mostly by the time of testing, rather than the soil
type.

Table 4 Statistical Summary for Smith Model

’arame :rs at Different Testink times.
Soil Par. summary E.0.D B.0.R

Sand
Mean 0.53 0.67
J,, slm
Std. Dev. 0.53 0.53
Figure 1 Shaft quake in SPT test.
Mean 0.43 0.80
Jp. s/m
Std. Dev. 0.13 0.90
Me2 0 28 0.30
Qs, cm
Std. Dev. 0.13 0.15
Mean 0.61 0.53
Qp, cm
Clay Std. Dev. 0.48 0.20

Mean 0.43 0.73

J,, slm .
Std. Dev. 0.40 0.53

2.4 Quake Values from SPT Test

GRL, in the FHWA report proposed a method to
determine the Smith model parameters from the
modified SPT test in a manner similar to the PDA
testing. The modified SPT procedure requires the
measurement of the dynamic force and the velocity
near the top of the SPT drill rod during sampler
driving, as well as the N-value, and for every test
depth the recorded force and velocity signals at the
end of driving of the sampler are analyzed and
processed via GRLWEAP program, thus the Smith
model parameters, €&, Q, and J were back calculated
through an integrated iterative parametric
optimization process. Two types of analysis were
performed: the first utilized the ultimate resistance
k,and quake, Q, obtained from the static test while
solving for the damping factor J in the CAPWAP
analysis, where the ultimate resistance from static
pile test was determined from the load-displacement
curve based on Davisson’s failure criterion. While the
second analysis “dynamic” analysis assumes that all
three soil parameters &, Q, J) were unknown, and
Figure 2 Toe quake in SPT test
they were solved for by matching the calculated force
curve. The results of these two analyses are
graphically presented in Figures 1 and 2, showing the
variation of soil quake at pile shaft and toe,
respectively, as a function of pile installation depth,
and soil type. Based on Figures 1 and 2, and for
further analysis and calculation of Smith type
damping coefficient in this paper, soil quake values at
the pile shaR and toe will be taken as the averages of
the two figures, meaning that the quake values at the
shaft and toe will be equal to 0.5 and 0125 in,
respectively.

463
 P. Project Pile Pene.. Date Location
No Name ft
3 1 Determining Damping S using GWWErkP 1 LUC75081 412in 48.56 29-May-96 Toledo
Program
The wave equation analysis program- G
was used to simulate the SPT results To
model the driving process of SPT, the progiam
requires data input that feeds and describes the
hammer characreristic, the rod and sampler and the
soit.
The SPT drill rod is a hollow 0.7-in outer
diameter tube with a wall thickness of 3/8 in The
sampler must have the same imer diameter as that of
the rod, a greater outer diameter of 2 0-in, and 2 25-
eieas, the input data describing the
h a m e r properties are summarized in Table 5 The
elastic modulus of the drilling rod and the split-spoon
sampler is 30,000 ksi (207 -GPa), and the specific
weight is 492 pcf (77 7 W/m')
1
6
-~~~~~
7

11
I 175-1128
175-8.39
8
~
$12 in

1754.61

I75982
I750891
51
175-8.91
~
$12 in
28-Jm-97 Toledo

!I;i;
(612 in

$22in
(612 in
~
29

18
-
21-Jan-97

30-Jan-97
;24-F;-97
i

52
63
~

2 I -khy-97
28-X r-98
20-May-97
3-Jun-97
I
~

,
Toiedo

Toledo
Toledo
Toledo
Toledo
Toledo
Toledo
I
The shaft and toe resistances can be cdculated
from SPT-N values using Equations 1 and 2,
respectively Based on the dimensions of the SPT 13 LIC1308- 416 in 2.1 17-Jul-97 Colurnbus
sampler, the total shaft and toe resistances can be 80
calculated, from which the eotal soil resistance and 13 GRE-35- W12r53 27. 30 05-.4ug-97 Greene
the ratio of the shaft resistance to the total soil 2080
resistance C ~ be
R obtained This ratio is the required
soil resistance distribution in the
Program
As indicated earlier in this paper, the soil quake
values will be deduced Erorn Figures 1 and 2 that
assign 0 050 and 0 125 for soil quake at the pile shaft
and pile toe, respectively

I Stroke I 30 in I 762m 1 22 POR-59-

578
$12 in 33 27-Xpr-98 Portage

23 I\/IAH8317 $11 in 41 17-Jw-98 YoWgsto\Vn

The evaluation of the damping coefficient is
achieved through a number of consequent iterations
in a parametric optimization process, rasing the
developed software. The static resistance,
soil quake, along with an initial guess for damping
coefffcient constitutes the input dzta to the program,
so that the standard penetration number SPT-Nis
evaluated for that set of data. This process will be
repeated until the obtained SPT-N matches the
actual- recorded number
3.2 Results and Discussion
Du~ingthe course of this paper project, the research
team has perfomed series of dynamlJc pile tests at 23
ODOT piling project sites. From those field tests, a
total of 34 piles- listed in Table 6, have been tested
using the Pile Driving halyzer (PDA). Liang (1999)
provides a detailed description for those piles, and
their records as obtained from the PDA.
Tables 7 and 8, report the calculated damping
factor 3 together with the C M
tested piles, for 60%, and 70%
Note that column (I) in Table 8 lists the average pile
shaft damping factor 3 for
project site as obtained by G
will have the same value for all pile shaft segments
when operating in the automated CAF"
Column (2) lists the shaft-damping factor J calculated
via the method introduced in this paper. Usually, the
SPT-N value varies with depth, therefore several

464
 esuIts ofDamuinn J Ec = 0.7).
Pmj Shaft 3 (dft) Toe 3 ( s/ft )
NO. CMWAP 1Predxcted CAPWM 1 Predicted
(0) (1) (2) (3) (4) (5)=(4)xO.J2
1 0.104 0.161 1.672" 0.268 0.113
2 0.162 0.206 0.175 0.225 0.095
I 3 I 0.07 I 0.144 I 0.036 10.147 I 0.062 I
I 4 I 0.152 I 0.251 I
0.354 10.502 1 0.211 1 3
4
0.07
0.152
0.187 0.148
0.307 0.243
0.036 0.194 0.068
0.354 0.617 0.216
5 0.22 0.302 0.04 0.497 0.209
6 0.05 0.075 0.047 0.501 0.210 I 5 I 0.22 I 0.37 1 0.292 I 0.04 10.615 I 0.215 I
7 I 0.105 { 0.145 0.227 10.331 1 0.139
1 I 6 I 0.05 10.101 I 0.080 I 0.047 10.613 I 0.215 I
8 I 0.117 I 0.239 I 0.231 I 0.479 I 0.201 1 I 7 0.105 10.1841 0.145 I I
0.227 10.397 0.1391
I 9 I 0.12 I 0.155 1
0.255 10.295 I 0.124 1 8
9
0.117
0.12
0.297 0.235
0.197 0.156
0.231 0.593 0.208
0.255 0.358 0.125
10 0.177 0.129 0.136 0.24 0.101
11 0.032 0.2 1.278* 0.319 0.134
12 I
1 0.513 0.185 I 0.059 10.202 I 0.085
13 I 6.954* I 0.202 I 0.024 10.096 I 0.040
1 14 1 0.101 I 0.218 I0.06 1
0.28 I 0.118 1
15 0.165 0.095 0.061 0.233 0.098

17 10.239 1 0.134 1 0.133 10.148 I 0.062
18 1 0.37 I 0.293 I 0.077 I 0.36 I 0.151
19 0.524 0.123 0.027 0.25 0.105
20 0.116 0.189 0.018 0.212 0.089
21 0.078 0.224 0.251 0.182 0.076
22 0.06 0.141 0.03 0.232 0.097
23 I 0.305 I 0.151 I 0.036 10.318 I 0.134
A ~ ~0.177I I 0.181 I 0.111 l0.288 I 0.121
* Not included in the average value.
values for shaft damping factor J along the pile shaft
can be obtained. The predicted damping values in
Table 8 and in the subsequent Tables are all length-
weighted average values along the pile shaft. It can
be seen that there is a difference in the shaft damping
J between the C N W N results and the predicted
values. Nevertheless, their average values are very
close. As for project 13, the shaft damping J from
WAd) is too big, so the shaft damping factor J of
s project is not included in the average v
Columns (3) and (4) in Table 8 are the Clap
and the predicted values of the toe damping fa
ly. Similar to the shaft damping, the
values for toe damping shown in column
(3) are the average values. As for the toe damping in
column (4), if there is more than one pile with
different pile toe elevations, then average value is
used. Similar to the shaft damping, the values for toe
damping factor J from CAPWAP for projects 1 and
11 are too, therefore, they were not included in the
calculation of the average values. It can be seen that
the difference between the CBBWAP results and the
predicted values is rather significant. The author
recornmends that the calculated toe damping factor J
being modified by a reduction factor of 0.42. The
modified toe-damping factor J is shown in column
I 160.123 0.27 0.213 0.052 0.384 0.134
I 170.239 0.179 0.141 0.133 0.201 0.070
18 0.37 0.364 0.288 0.077 0.442 0.155
190.524 0.16 0.126 0.027 0.313 0.110
1201 0.116 (0.2471 0.195 1 0.018 10.275 1 0.096 1
21 0.078 0.274 0.216 0.251 0.233 0.082
22 0.06 0.18 0.142 0.03 0.293 0.103
23 0.305 0.188 0.149 0.036 0.396 0.139
AV 0.177 0.227 0.180 0.111 0.358 0.120
* Not included in the average value.
( 5 ) . In this way, the predicted average value (0.121)
is close to the C average value (0.1 11). The
comparison between C N W M and predicted values
are shown in Figures 3 and 4 for shaft and toe
damping, respectively. Figure 5 shows the sirmlar
comparison for the modified toe damping.
Figure 6 presents the correlation between the
predicted damping factor J and §PT N values, where
the damping values are shown for all SPT N values
from all projects. The damping values were too high,
so author recommends that the damping J be
modified with a modification factor of 0.75. Then the
correlations between the predicted damping J and the
SPT-W values are those shown in, in which, two
approximate fitting curves for clay and sand are
plotted. The equations for these two approximate
fitting curves are as follows.
0 .637 N
Clay
62 . 5 + N
J = 3
0 ,3375 N
Sand
62 . 5 + N

465
 F Efficiency = 0.6

0'5 I A v e . J = 0.181
t
0.4 1 //'

i
0.3
L

00
00
i..::-,-, 01 02
,

03
I

04
, I
Average J = 0 177
, I

05
I ,

06
N
Figure 6 Correlation between modified damping
and N values ( E c = 0.6)
value

CAPWAP shaft J (sift)

Figure 3 CAPWAP Vs Predicted Shaft Damping.
0.6 r-
1 AAer modification (0 79) ,,,\

DO -
L /
- 05
F Efficiencv 0.7 ,'

0.5 j-
1 a-
,' Efficiency =0 6 ,

t .

, I
/

Avenge J = 0 177 II
00 __-I____ii_L-I__J

00 0 1 0 2 02 0 4 0 5 06
r C.\P\VXP \ h a f t J (sift)
0 1 [-*
Figure 7 CAPWAP Vs Predicted Shaft Damping
!
0 -0 - L '
00 01 02 03 04 OS 00

CAPWAPtoe J (sift) Mod factor= 0 35 /'

Figure 4 CAPWAP Vs Piedicted Toe Damping Oh F- Efficiency = 07 /
Coefficient i Average J = 0 358 /
h O S L ,'
.=
5 1 /
0.1
1 ~

AAer modification (0.42) c-i 04; ,

0.6 4
3
Efficiency = 0.6 /' s i /

z"- . < 14
i
Ave. J = 0.121 /
0.5 /
/
0.4 7
t - 1 ,/
3 8
?
. 0.3 4 ,/'

{ j .
& 0.2 4- /' -
/'
on 0 1 I)? n3 0.1 ns 06 n-
CAPWAP toe J (sift)

Figure 8 CAPWAP Vs Predicted Toe Dampmg.

00 A^___-- _I___ -&-I- __I_A
_.

00 01 02 03 04 05 06 07
C.\PW,AP Toe-J (sift) is 0.6. For a hammer efficiency of 0.7, the
Figure 5 CAPWAP V s Predicted Toc D a m p m i n g correlations between the damping J and the SPT N
Coefficients ( After modification 1 values are presented in Figure 7. It can be seen that
the damping J for Ec = 0.7 is larger than the damping
The correlation results discussed above are based J for Ec = 0.6. This can be explained easily by the
on the assumption that the hammer efficiency of SPT fact that the static soil resistance correlations remain

466
the same regardless of the hammer efficiency. As a
result, if the hammer efficiency is high, the increased
energy is resisted by the dynamic soil resistance
indicated by an increased damping J. Table 7 along
with Figures 7 and 8 provide the results and
correlation between
predicted damping for h
may be seen that the predicted damping is higher
damping. As discussed before, a
modification factor of 0.6 is recommended if one
uses that Ec = 0.7.
Common Soil and Rock Tests”, Proc. Of the
Congress on Foundation Engineering: Current
Principles and Practices, Evanston illinois,
June, 1989, Vol. 2, 1026-10389.
7. Terzaghi, K. and Peck, R. B. (1967). “Soil
Mechanics in Engineering Practice”, John
Wiley and Sons, Inc., New York.

COWCLUSIONS

Based upon the field test results, data analysis, and

literature review, the author may draw the following
conclusions:
The equations relating pile static capacity, and
damping coefficient- both at pile shaft and toe, are
reasonably accepted from the engineering
background. Although there were some deviations
from the values obtained from the case method, the
average values were close, indicating the validity of
the approach. The bigger differences in the damping
coefficient may be attributed to the fact that, the
predicted damping values will include both the errors
in the static resistance, and the damping coefficient-
commutative
As a result, the use of SPT blow count in the
prediction of static resistance, and damping
coefficient may be considered to be satisfactorily
powerful.
The SPT provides a good predictive tool for soil
quake, only at pile toe.

1. . m., and Reese, L. C.(1966), “Load

Transfer for Axially Loaded Piles in Clay”,
JSMFD, Asce, Vol. 92, No. S W , pp. 1-26.
2. FHWA (1988). “Manual on Design and
Construction of Driven Pile Foundation”,
FHNA-DP-66- 1.
3
3. Goble Raudche and Associates, (1988),
CAPWAP Manual, Cleveland, OH.
4. Liang, R. Y. (1999), “Development and
Implementation of New Driven Pile
Technology”, Final Report submitted to the
ODQT, May, 1999, Report no. FHWMQh-
991008.
5. Paikowsky, S. g., Regan, J. E., and
McDonnell, J. J. (1994), “A Simplified Field
Method for Capacity Evaluation of Driven
Piles”, Publication no. FMWA-IQD-94-042.
4. Reese, L. C., and O’Neill, M. W. (1989),
“New Design Method for Drilled Shafts from

467
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

tress wave theory application to standar enetration test i

K. Fujita
Deparfmerit o j Civil Engineering,Science UrziversiQoJ’Tokyo,No&, Japan
M.Ohn0
HuzacIpm Technical Reseur-chInstitute, Tsukuba,Japan

ABSTRACT: Studies related to the application of stress wave theory to Standard Penetration Test(SPT) in
Japan are introduced together with a report on the discussions we have had on the dynamics and mechanism
of the SPT as well as aspects of the issues to be studied. Uto et al. were the first to apply the stress wave theo-
ry to the SPT in 1970’s. Matsumoto et al. have applied the method of two-point strain measurements to the
SPT, and presented the calculated stresses, penetration and transmitted energy as firnetions of time in dia-
grams which corresponded reasonably with measured data. Fujita et al. have proposed a fundamental equation
for the equilibrium between the driving and resisting energies considering the energy transfer ratio, an ad-
justment method for N-values, and Fujita’s diagram summarizing the SPT results, etc..

1 INTRODUCTION 2 STUDIES BY UTO ET AL.

The SPT Working Party of the ISSMFE Technical
Committee on Penetration Testing presented the
“International Reference Test Procedures on SPT”,
at ISOPT-1, Qrland in 1988. In Document 4-
Calibration Methods, a method using stress wave
equation analysis was introduced for the adjustment
of the N-values in the SPT to account for different
magnitude of driving energy, as a means of quality
control for the SPT. Many papers related to the ap-
plication of stress wave theory to the SPT have been
contributed and the measurements of the energy
transfer ratio have increasingly been made in actual
practice in the field in USA and other countries.
However, few high frequency measurements of the
movement or penetration per blow of the rod system
have been published or reported in connection with
the stress wave equation analysis of the SPT.
The application of the stress wave equation ana-
lysis to the SPT in Japan started after the study of
Uto et al. in the 1970‘s. Most of the studies carried
out include high quality measurements of the pene-
tration during the SPT in order to establish the
mechanics or dynamics of the SPT which are not
only for the purpose of quality assurance of N-values
but also for attempting to find the shearing strength
ox characteristics of the soil.
This paper will introduce a state-of-the art report
on studies concerned with the application of stress
wave equation analysis to the SPT or related matters
which have been done in Japan. These are mostly
written in Japanese and include unpublished data.
For applying wave equation analysis to the SPT, LJto
et al.( 1974) proposed a model using the generalized
solution of St. Venant assuming the reflection coef-
ficient to be the boundary condition at the bottom of
the rod and that the rod is supported on rigid ground
and struck by a hammer which is considered to be a
rigid body. A series of experiments were carried out
using an oscilloscope and strain gauges etc. for
monitoring the stress waves and a high-speed movie
camera for recording the movement of the top of the
rod. The calculated and measured maximum com-
pressive stresses in the rod were 154 M N m 2 and
about 84 MN/m’, respectively.
3 STUDIES BY MTSUMOTO ET AL.
Matsumoto et al. (1992) have applied the method of
two-point strain measurements proposed by Lund-
berg & Henchoz (1977) to a series of instrumented
SPTs that they carried out. The upper gauge was lo-
cated on the rod at a point of 225 rnrn below the an-
vil which had a height of 80 mm, and the distance
between the upper and lower gauges was 600 mm.
The total length between the top of anvil and the
bottom of the sampler was 3.39mYthe hammer was
operated by means of the rope and cathead method
and the energy transfer ratio was found to be be-
tween 0.6 and 0.9. The SPT was carried out based on
the Japanese Industry Standard (JIS) from a depth of
1.32m to 1.62111 and the number of blows was 12

469
 Time(nis) (N-12) through clay and sand layers.
The monitored stress in the rod was recorded at a
sampling rate of 5 ,U s for the wave equation analy-
sis, employing a DC preamplifier with a capability
of 250 kJ3z and an A/Dconverter. The displacement
of a target on the rod was measured by an electrical
device with a sensitivity of 0.01 111111.
Matsumoto et al. (1992) have proposed a proce-
dure for handling errors in the measured strains such
Calculated, o Measured
as noise, using two correlation factors for each of the
Figure 1 . Calculated and measured displacement-time upper and lower strain gauges in order to find the
diagram at a blow of SPT(after Matsumoto et al.1992).
true strains. According to case studies, if the value of
one correlation factor is taken to be unity, the other
correlation factor would be in the range between
-
E
500
0.985 and 1.015. Verifications have been made for
the displacement - time curves, the stress wave-
forms and the energy transmitted to the rod with
time as shown in Figures 1-3. The calculated data
compare well with the measured data so the two-
point strain measurements and the procedure for
handling errors in the measured strains can be re-
garded as very useful tools for the wave equation
analysis.
0 5 10 15 20
Ti me(ms)
--Calculated, - - - - Measured
4 SET-UP FOR THE LABORATORY SPT
Figure 2.Calculated and measured transmitted energy-time
diagram 31 ;I blow of SPT(after Matsumoto et al.1992). STUDY BY FUJITA ET AL.

For the laboratory experiments results to the SPT,

Fujita et al. have used the arrangement of testing
equipment, instruments etc. shown in Figure 4 and
Table1 since 1993. The SPT equipment conforms to
the JIS (Japanese Industrial Standard) specifications

- 1001
0 7 10 15 20
Time(ms)
(a) Calculated

5 10 1s 20
'I? me( ms)
(12) Measured

Figure 3. Calculated and measured stress-time diagrams Fi ure 4. Set-up for laboratory SPT.
at a blow of SPT (after Matsumoto et al. 1992). (&er Nagasaki et al. 1996)

470
Table 1. Specification of equipment and instruments.
Hammer -w:63.5kg, OD:19.5cm, ID:S.Ocm, h:33cm
Anvil -dia:8Scm, height6.5cm
Rod -OD:40.5mm, A:5.3cm: W:4.68kg/m
Sampler -OD:Slmm, ID:35mm, length:685mm
Steel tank -dia:59.4cm, depth:105cm
Pressure hag -SO, 100, 150, 200, 300, SOOkPa by air
Soils -Toyoura sand etc.
Strain gauge -Kyowa, semi-conductor, 1=21nm,
GF:-98, 120 d?
Amplifier -Kyowa, DC, 200kHz
A/D convertor-Sun system supply, ADH-12A
Displacement sensor --Izumi, Ldzer type, max.100mm,
40mm at l k H z

and a heavy steel tank, 0.60m in diameter by 1.05m

high was built to control the variable soil conditions
for the laboratory tests. A pressurized air bag is in-
stalled a load representing the overburden pressure
between the loading plate fitted to the top of the tank
and the surface of the test soil. The length of the
guide rod and the distance between the surface of the
anvil impacted by the hammer and the bottom of the
sampler are 1.50m and 1.80m, respectively, so that
the overall length is 3.30m.
As the method of two-point strain measurements
(TPSM) proposed by Lundberg and Honchez (1977)
is employed for the measurements and analysis, the
upper and lower sensors are installed on the rod
40cm and 60cm below the top of the anvil, respec- Figure 5. Diagrams summarizing a laboratory SPT.
tively, for measuring the strains and velocity of the (ai‘ter Fujila, 1997).
stress waves in the rod. Each sensor consists of four
semi-conductor strain gauges attached around the
circumference of the rod, at intervals of 90’. fiiction effects due to the guide rod.
A laser type displacement sensor, capable of Measurements were made at every hammer blow
monitoring a 1 kHz wave with an amplitude of after the excess pore water pressure had been dissi-
40mm is provided for measuring the movement of pated in the case of the tests using saturated sands.
each blow of the sampler at a certain point along the Analyses using Matsumoto’s program and digit-
connecting rod. ized data are carried out to obtain a stress (strain) -
Three high frequency DC amplifiers and an A/D time diagram, a penetration - time diagram, a veloc-
converter are employed for measuring the strain and ity of rod - time diagram, a transmitted energy -time
movement of the rod and the monitored signals are diagram and an energy transfer ratio (e).
digitized and recorded on a floppy disk on a com- One of the results of the test and analysis is
puter at a sampling rate of 5 U, s. shown in the set of diagrams in Figure 5 , which
The laboratory test ground provided in the tank is shows that a) the maximum compressive stress in-
made uniform aiming for a relative density either duced in the rod by a hammer blow is about 200
dry or saturated condition, and Toyoura sand is gen- MN/m2, b) the maximum and residual penetrations
erally used. Overburden pressures of 50, 100,200, are about 3.6mm and 3.4mm, respectively, and c)
300 or 5OOkPa can be applied by using the pressur- the maximum velocity of the rod movement is about
ized air bag. 9 d s . Figure 5 also indicates that the hammer strikes
The tests were carried out according to the provi- the rod more than three times after its first impacts.
sions given in the SPT procedure of the JIS, in
which the hammer fall height is specified to be
750mm (it will be changed to 760mm in 2000), so 5 FUNDAMENTAL EQUILIBRIUM EQUATION
that the kinetic energy is 467J, not 473J. AND FUJITA’S DIAGRAM
The hammer is hung by a wire on its vertical axis
and is released by cutting the wire, so that when the The following equation can be used to express the
rod system is held in a proper perpendicular posi- equilibrium between the kinetic energy given by the
tion, the hammer can be dropped in a free fall, i.e. no SPT hammer impact and the resistance due to work
done by the sampler, if the kinetic energy is trans-

47 1
rnitted perfectly to the connecting rod system or
sampler.

wh = R,.S

where W = weight of the hammer (622."); h = free

fall height of the hammer ( 7 6 0 m ) ; R, = dynamic
resistance of the sampler; S = penetration per blow
(mm); and wh = kinetic energy (473.35).
In practice, the energy transmitted to the con-
necting rod system is smaller than the kinetic energy
of the hammer and the energy transfer ratio e is de-
fined as,

Figure 6. Fujita's diagram for representing penctration,energy

transfer ratio and dynamic resistance at each blow o f SPT.
where e = energy transfer ratio; E, = energy trans- (after Fujita 1997).
mitted to the connecting rod system or sampler.
Fujita (I 997) has proposed a fundamental equilib-
rium equation (3) applicable to each hammer blow
of the SPT as follows.

In equation ( 3 ) , eWh is calculated by the wave

equation analysis of the measured stress with time at
the two points on the rod, and &5' is the measured
penetration per blow, then &Rd corresponds to the
dynamic resistance of the sampler. The reason why
&,S and &Rd are used in equation (3) will be dis-
cussed later.
Fujita (1997) has proposed a diagram with or-
thogonal axes as shown in Figure 6 for representing
equation (3), the x and y axes, using the same loga-
rithmic scale are the penetration per blow S and the
dynamic resistance R,, respectively, and the energy Figure 7. Penetration, energy transfer ratio and dynamic resistaricc
transfer ratio e are given by lines each with a slope at each blow of SPT in terms of rclativc density (Dr).
(after Fujita 1997).
of 45'. Figure 6 is the so called Fujita's diagram and
can summarized the data obtained at each blow of
the SPT and is useful for the interpretation and qual- smaller dynamic resistances. The range of each
ity assessment of the SPT and for the quality assur- group depends on the penetration depth of the sam-
ance of the N-value. pler which varies from 150 to 450mrn during the
Figure 7 shows an example of the results of the SPT. So, the penetrations of the sampler per blow
SPT laboratory experiments which were carried out are larger at shallow depths and become smaller
on dry Toyoura sand in a test tank prepared with an when the depth becomes larger, while the dynamics
uniform relative density of about 40%, 60% or 80%. resistances are correspondingly smaller and larger.
The measured penetration, calculated dynamic re-
sistance of the sampler and the calculated ratio of the
energy transferred to the rod system or sampler at 6 ADJUSTMENT OF THE N-VALUE USING
each hammer blow can be plotted as one point on THE SQUARE ROOT OF THE ENERGY
the coordinates of the diagram. TRANSFER RATIO
Figure 7 illustrates that all the data lie along a line
for which e is around e=0.8 or e=80%, and the plots According to the International Reference Test Pro-
for different relative densities are found to be locat- cedures, it is recommended that the N-value should
ed in their own groups. In the case of the relative be adjusted to the N-value corresponding to the en-
density of 40%, for example, all the plots belong to a ergy transfer ratio of 0.6 or 60% in the reference
group corresponding to larger penetrations and procedures, if required, assuming the N-value to be

472
inversely proportional to the energy transfer ratio
(4.
Figure 8 given by Nagasaki et al. (1998) shows
three sets of measured data with different energy
transfer ratio being plotted on a Fujita’s diagram.
Each was obtained by a blow of the SPT hammer at
the same depth in each test tank containing dry
Toyoura sand prepared with the same uniform rela-
tive density. The different energy ratios were pro-
duced intentionally by changing the free fall height
of the hammer.
The three plots in Figure 8 are located nearly on a
straight line which is orthogonal to the lines of the
energy transfer ratio. This means that the magnitude
of the penetration is proportional to the square root
of the energy transfer ratio e, not proportional to e.
As the N-value is defined as the number of blows
necessary to achieve a penetration of 300mm, and
the N-value has to be inversely proportional to the
square root of the energy transfer ratio, i.e. &, the
adjustment of the N-value should be made accord-
ingly.
Table 2 indicates the comparison of the N-values
which are adjusted in terms of the energy transfer
ratios given by Kovacs et al. and others, by the
method of the International Reference Test Proce-
dures assuming the N-value is inversely proportional
to the energy transfer ratio and the method proposed
by Fujita assuming the N-value is inversely propor-
tional to the square root of the energy transfer ratio.
7 RELATION BETWEEN THE ANGLE OF THE
HAMMER BLOW AND ENERGY TRANSFER
RATIO
The energy transfer ratio of the SPT represents the
efficiency with which energy is transferred from the
kinetic energy of the hammer to the energy trans-
mitted to the rod system, and it is generally under-
stood that its magnitude depends upon the method of
hammer release, e.g. larger for the trip method and

(A) Angled blow hetwecn bottom of hammer and top of anvil

Free fall height of hammer (cm)

(11) Free fall height of hammer vs penetration
Figure 8. Adjustment of penetration or N -value b y J e .
(after Nagasaki et al.1998).

Table 2.Adjustment of N-value by energy transfer ratio-

comparison between Int’l reference test procedures and
Fujita’s niethod.(after Fujita,l997).
Energy transfer ratioSymbols Adjusted N-value
for
e Referenccs N-value Int’l ref. Fujita’s method
1.00 NI00 E?? 18 23
__
0.92 Kovacs’s max Nne 20 19 24
0.67 K O V ~ C Save
’S Ntj7 27 22 28
0.60 Int’l reference 30 23 30
0.45 N45 40 27 35 (c) Free fall height of hammer vs energy transfer ratio
0.33 Kovacs’s min Nv 55 31 40
18,3:Assumed N-values, other N-values are caluculated Figure 9. Effects of angled blow to penetration and energy
by the respective adjustment methods. transfer ratio.(after Fujita et a1.1994).

473
smaller for the rope and cathead method.
Fujita et al. (1994) have frequently observed that
the bottom of the hammer contacts the top of the an-
vil at an angle during the hammer blow of the SPT,
since the rope of the hanging hammer, the centerline
of the hammer and the centerline of the rod are usu-
ally not in a straight line and usually are not com-
pletely vertical. The amount of the deviation is likely
to depend on the method of hammer release and
workmanship of the operator, and these many affect
the magnitude of the energy transfer ratio.
To examine the effects of angled contact during
the hammer blow of the SPT on the energy transfer
ratio by applying the wave equation analysis, a seri-
es of laboratory tests simulating the SPT were car-
ried out using a l m long instrumented steel bar with
a diameter of 25mm, an anvil with a surface inclined
at 0,1,2 or 3 degree and a steel donut type hammer
weighing 49N provided with an inside diameter
suitable for making contact at an angle. The results
show that when the slope of the surface of the an\d
is increased, the energy transfer ratio becomes
smaller as shown in Figure 9(a), and the penetration
of the rod per blow also becomes smaller as shown
in Figure 9(b). Figure 9 indicated that the SPT appa-
ratus should be set up accurately in a straight vertical
line in order to ensure the free fall of the hammer.
tween hammer blows should be determined prop-
erly.
9 INFLUENCE OF PORE WATER AND
OVERBURDEN PRESSURE ON THE N-
VALUE
Studies of the influence of water content and over-
burden pressure on the N-value are interesting issues
to be examined. Nagasaki et al. (1997) have contrib-
uted the results of a series of laboratory SPTs using
dry and saturated Toyoura sands with relative densi-
ties of about 40%, 60% and 80% prepared uniformly
in a heavy steel tank, employing two types of pres-
surized air bags for simulating overburden pressures
of 100kPa, 300kPa and 5OOkPa.
Figure 10 gives the results of the test, and shows
that the N-value increases with overburden pressure.
The N-value of the dry sand at relative density of
60% is approximately equal to that of saturated sand
and the N-values of dry sand at the densities of 40%
and 80% are smaller and larger than that of the satu-
rated sand, respectively. On the other hand the N-
values are larger and smaller respectively, when
another type of pressurized air bag is employed for
the saturated sand as shown in Figure 11.

8 PROBLEMS RELATED TO EXCESS PORE

WATER PRESSURE

Since the dynamic resistance of piles at the end of

the driving operation had been considered to be the
same as the static pile bearing capacity derived from
the load vs. settlement diagram using Davission’s
method. Fujita & Kusakabe (1988) have pointed out
that the number of days that have elapsed since the
pile installation is a very important factor for
evaluating the static bearing capacity of piles driven
below the ground water table. The recent introduc-
tion of a set-up ratio, around 2 to 5 in magnitude,
could be a solution for the said problem, if the ratio
could be determined properly.
When the SPT is carried out in ground below the
water table where excess pore water pressure is in-
duced by the hammering of the sampler, similar
problem may occur, In this case, a smaller N-value
would be obtained than if there had been no excess
pore water-pressure.
The magnitude of the excess pore water pressure
depends upon the type of soil, volume of displace-
ment, the interval and number of hammer blows etc.,
and dispersed with time at a rate proportional to the
pressure and permeability of the soil. In rare cases a
negative excess pore water pressure may occur lo-
cally. The excess pore water pressure affects the Figure 11. Effcts of pressure bag to relative density and
magnitude of the N-value and so the interval be- N-value of saturated sand.(after Nagasaki et al. 1997).

474
Table 3. Comparison of I'\;-valuc between dry and saturated
sands at relative dcnsity of 40,60 and 80%.(after Nagasaki-
et al. 19'97).
DJ 40%
~ Dr- 60% D r - 80%
Gibbs-l-loltz (1957) Nrr>N,. Nd> N , N(J>N,
Y anase (1963) N(s>NLV N ( J > N , Nct>N,br
Yaniacla et al. (1004) Ni>N,, N(s=N,. Nis<N,.
Fu,jita et al. (1996) N(i>N,, Ncs'N, Ncr<N,.
Nagasaki et a1.(1'997) Nci<Nw N(s-=Nw Nrl>N,.
A',; : .Y-\~alucof d r y sand. N , : N-value of saturated sand
These results are summarized in Table 3, which
also includes other contributions by Gibbs & Holtz
(1957), Yanase (1963) and Yamada et al. (1994).
According to Table 3, there are two cases in which
the N-value of dry sand is larger than that of saturat-
ed sand, and three other cases in which the N-value
of dry sand is approximately equal to that of saturat-
ed sand at relative density of 60%, while it is either
larger or smaller at the relative density of 40% and
smaller or larger at the density of 80%.
Table 3 suggests that either the N-value is not al-
ways proportional to the effective stress in the model
ground or the N-value of the saturated sand is sig-
nificantly affected by the type of air pressure bags
employed for the laboratory tests.
10 CONCLUSIONS
The application of stress wave equation analysis to
the SPT is very useful for the quality control of the
SPT as well as for the quality assurance of the N-
value. The laboratory experiments on the SPT and
modified tests using various types of uniform model
grounds contained in a steel tank which is provided
with an air pressure bag to simulate overburden
pressure loads are also very advantageous for
studying the mechanics and dynamics of the SPT,
because it is possible to produce various ground
conditions and to have repeatability. However, there
have not been enough studies of this kind in Japan.
With respect to the International Reference Test
Procedures, it is recommended that a method should
be introduced to adjust for the length of hammer,
because when the length of the hammer is longer,
the penetration per blow is larger, according to the
unpublished results of a recent modified test.
The analysis of the frequency characteristics of
the measured stress wave in the laboratory or on an
in-situ SPT would provide valuable information in
this field, according to the latest study now in prog-
ress. Further studies are recommended.
475
REFERENCES
Fujita, K., 0.Kusakabe. 1988. On the evaluation of static
bearing capacity. B.H.Fellenius(ed), Proc. 3"' Int.
Con$ on Application of Stress-wave Theory to Piles.
525-534. Ottawa:
Fujita, K., R.Tanaka, M.Umemura, & T.Une. 1994. An-
gled blow of hammer against head of piles. Proc. 49'"
Annual Con$ of JSCE. 3: 954-955. (in Japanese).
Fujita, K. 1997. Referable N-values of unstandardized
standard penetration tests. J. of JSCE. 82(12): 24-27.
(in Japanese).
Fujita, K. 1997. Mechanism of SPT, interpretation and
evaluation of N-values. Kisoko. 25( 12): 2-13. Tokyo:
Sogodoboku-kenkyusho. (in Japanese).
Matsumoto, T., H.Sekiguchi, H.Yoshida, & K.Kita. 1992.
Significance of two-point strain measurement in SPT.
Soils and Foundations, (32)2: 67-82.
Nagasaki, T., K.Fujita, T.Une. 1996. Dynamic resistance
of SPT in dry sand subjected to overburden pressure.
Proc. 51" Annual Conf of JSCE. 3A: 758-759. (in
Japanese).
Nagasaki, H., K.Fujita, Y .Imamura, K.Nohara &
K.Yamamoto. 1997. N-value, energy transfer ratio and
dynamic resistance in saturated Toyoura sands. Proc.
32"d Annual Con$ on Geotechnical Engineerng. JGS.
191-192. (in Japanese).
Nagasaki, H., K.Fujita, T.Yoshinaga & H.Yamaguchi.
1998. A rational adjustment of SPT N-value by square
root of energy transfer ratio. Proc. 53rdAnnual Con$
of JSCE. 3A: 762-763. (in Japanese).
Uto, K., M.Fuyuki, H.Kondo, M.Morihara & H.Matsu-
mura. 1974. Consideration of N-Value and strength of
ground from a view-point of wave theory. Proc. Fac-
ulty of Engineering, Tokai University. (in Japanese).
This Page Intentionally Left Blank
ra
This Page Intentionally Left Blank
 Application of Sfress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

ynote lecture: Vibratory driving analysis

Alain E. Holeyman
Universite‘Catholique de Louvain, Louvain-la-Neuve, Belgium

ABSTRACT : Engineering issues related to vibratory driving of piles and sheet-piles cover many facets in-
cluding the long-term bearing capacity of the installed pile, its vibratory penetration resistance, the performance
of vibrators, the degradation and liquefaction of the soil around the vibrated profile, and vibratory nuisance to
the environment. An argument is made that those issues will be adequately tackled once combined into a com-
prehensive framework of analysis where proper understanding of soil behavior is the key. The present paper
however focuses mainly on our current engineering ability to assess vibro-drivability, i.e. predicting the vibra-
tory penetration log of a given pile into a given soil profile using a given vibrator. Testing undertaken to pro-
vide insight into the pile-soil-vibrator interaction and its modelling is reviewed. Several available methods to
establish the vibratory performance of a pile from its vibratory capacity are discussed. A rational procedure to
model the dynamic nonlinear soil structure interaction during pile vibratory driving is discussed in more detail.
Degradation of the skin fiction upon cyclic shear stress is evaluated by applying elements of earthquake engi-
neering practice used to assess liquefaction potential. The present ability to assess the vibratory capacity of a
pile from the monitoring of its vibratory performance is critically reviewed. Finally, suggestions for hrther re-
search, design and practice are provided.

1 INTRODUCTION means of installing piles and sheet piles in appropriate

1.1 Scope
The main purpose of this key-note paper is to present
the author’s present view on engineering issues re-
lating to the drivability of piles and sheet-piles using
vibrators and the inverse problem, i.e. deriving the
pile resistance from its vibratory perf-ormance. It is
based on a survey of the relevant literature and origi-
nal research in the area.
1.2 Historical Development
The vibratory driving technique appears to date back
to the early 30’s when it was simultaneously devel-
oped in the former USSR and in Germany (Rodger
and Littlejohn, 1980). The observation by the Rus-
sian soil dynamics researcher Pavyluk that soil resis-
tance could be reduced thanks to vibrations led to the
industrial use of vibrators to drive piles, according to
Barkan, 1960. Extensive research on the effects of
vibration on soils was conducted in the 40’s and 50’s
by Barkan, while the vibratory driving technique was
gaining acceptance as an economical and effective
soil conditions. Major vibrators manufacturers are
now located in Germany, France, The Netherlands,
USA, the Former Soviet Union and Japan. Although
technological developments have been brought to
enhance the initial concept and extend commercial
application of vibrators, little is mastered by the en-
gineer when it comes to addressing soil related is-
sues. That limitation in the engineering knowledge is
viewed by many as an impediment for the vibratory
driving technique to enjoy its full potential.
1.3 Phenomena at P k y
Three major actors play a role in the mechanics of
the vibratory driving process, as illustrated in Fig. 1 :
(I) the pile to be driven, (2) the selected vibrator,
and ( 3 ) the imposed soil conditions. The pile can be
fully described by its material and geometry. The vi-
brator mechanical behavior can be assessed based on
its specifications and operational range, as discussed
in Section 2. Soil conditions are usually character-
ized by means of standard investigation tools such as
CPT soundings, borings and laboratory tests.

479

Figure I . Vibratory Driving : Players and Issues
Those investigation tools are geared towards an-
swering general design questions (mostly static) but
are not well suited to characterize soil behavior under
pile installation conditions, specially if the piles are
vibratory driven.
Because it is has been established for more than
half a century that soil resistance during vibratory
driving (likewise during impact driving) is lower than
the long-term bearing capacity, these two resistances
should be distinguished. As shown in Fig. 2, one can
estimate the vibratory capacity from the long-term
bearing capacity by taking soil degradation eEects
into account. Conversely, one can estimate the long-
term bearing capacity from the vibratory capacity if
soil set-up can be accounted for.
A hndamental understanding of soil behavior un-
der vibratory loading is required to establish the rela-
tionship between the pile vibratory resistance and its
long-term bearing capacity. Soil resistance degrades
upon cyclic shearing mainly because of fatigue of the
soil skeleton in cohesive soils (Vucetic, 1992), and
of effective stress reduction in granular soils (Ca-
sagrande, 1938). The effective stress can be ulti-
mately reduced to nearly zero, at which point the soil
behaves in a fluid-like manner. These phenomena will
be reviewed in more detail in Section 3.
480
Figure 2. Installation and design process of vibratory driven
piles
1.4 Engineering Issues
Engineering issues related to vibratory driving cover
niany facets, as iIlustrated in Figs. 1 and 2. They
prompt the following questions :
- What is the long-term bearing capacity of the in-
stalled pile? We h o w it depends on the pile ge-
ometry, on the soil parameters, but also on the vi-
bratory process.
- How are the soil’s long-tem strength parameters
influenced by the vibratory process? By how
much will the soil compact, and what is the mag-
nitude of the potentially induced settlement?
.- How are vibrations transmitted to the surround-
ing soil, and how much potential damage can
they cause to neighboring structures?
-. Will a given vibrator be able to drive the pile to
the required design depth? If so, at what speed?
Are there soil types that strongly limit vibratory
penetration depth?
- Are there ways to assess the vibratory capacity of
a pile from the monitoring of its vibratory per-
formance?
- Is there a vibratory testing technique and inter-
pretation leading to estimating the long-term pile
bearing capacity?
One can actually state that all issues will be prop-
erly tackled when combined into a comprehensive
framework of analysis where proper understanding of
soil behavior is the key. As clarified in Fig. 2, the
present paper will focus on our current engineering
ability to assess vibro-drivability, i.e. to predict the
vibratory penetration log of a given pile into a given
soil profile using a given vibrator. Testing undertaken
to provide insight into the pile-soil-vibrator interac-
tion and its modelling will be reviewed in Sections 4
and 5. We will then focus on some available methods
to establish the vibratory performance of a pile from
its vibratory capacity (Section 6), and look into the
potential to establish the reverse relationship (Section
7). FinalIy, Section 8 provides suggestions for fur-
ther research, design and practice.
Because of space hnitations, the paper does not
f ~ c u on
s other important engineering issues such as:
bearing capacity of vibro-driven piles derived from
soil characterization, vibrations transferred to the en-
vironment, and equipment specifications.
2 PILES AND VIBRATORY EQUIPMENT
2. I Vibrated Piles
Pile types or profiles mostly used in combination with
the vibratory driving technique include:
- sheet piles installed for temporary shoring, coffer-
dam and permanent retaining and containing
walls,
- H-piles vibro-driven as deep foundations or vi-
brated to help install underground hydraulic barri-
ers,
- Tubes to install cast-in-steel-shell (CISS) piles
- Precast prestressed concrete piles
- Steel profiles to vibro-compact granular soils at
depth.
Port, harbor, near-shore and offshore projects
very oRen take advantage of the vibratory penetra-
tion techniques, as the environment lends its self to
substantial tolerance of vibratory disturbances.
Steel and concrete profiles are generally cylindri-
cal or prismatic, and can be characterized by the fol-
lowing geometrical and mechanical properties :
A [m2]:profile section
L [m]: profile length
x [m]: profile perimeter
E [Mpa!: Material Young’s Modulus
p [kg/m 1: Material volumic mass
The profile section can be more fully characterized
by its shape, and inside and outside perimeters if
closed. This allows one to calculate the areas of the
profile in longitudinal and transversal contact with
the soil, once an embedment depth z [m] is assumed.
The mass of the profile M, equals PAL [kg] while the
longitudinal wave speed in the profile is given by
c = J E i p [Pnls].
Although they may at time play an important role,
transversal and flexural properties of the profile will
generally be ignored in the analysis that confines it-
self to the longitudinal behavior of the profile.
2.2 Mechanical action of a vibrator
The mechanical action of a vibrator onto a profile
consists of two part: a vibratory action and a station-
ary action.
48 I
Figure 3 . Mechanical action of a vibrator
The vibratory action imparted to the pile is pro-
duced by counter-rotating eccentric masses actuated
within an “exciter block”, as shown in Fig. 3a. The
centrifugal forces acting as a result of inertial effects
on an even number of symetrically moving nmses
combine into a sinusoidal vertical force :
F,(t) = me .a2sin (at ) = F,sin ( a t ) (1)
where
F, = maximum centrhgal force of the vibrator p]
me = eccentric moment of the vibrator [kg.m]
o = angular frequency of the vibrator [rad/s]
Alternative quantifications of the angular fie-
quency are the rotation speed R [rpm] and the fie-
quency v [Hz], with :
R = 6Ov=60-(a/h) (2)
The vibratory action can be therefore assessed
once both the eccentric moment and the operating
fi-equency are known. That action will be balanced by
reactive inertial effects of masses undergoing the im-
parted vibratory movement and by soil reactions op-
posing the profile movement. Provided the center of
gravity of the rotating masses belongs at all times to
the profile neutral axis, the exciter block is assumed
to exert a purely longitudinal force onto the profile.
The exciter block is connected to the profile via a
clamping device and is suspended to a carrier. The
suspension device includes a vibration isolator
mechanism consisting of a quasi-stationary heavy
mass directly suspended to the suspension hook and
an intervening spring, generally consisting of elas-
tometer pads.
The vibrator can be viewed as a two degrees of
freedom system moving in the longitudinal direction
(see Fig. 3b) : an exciter block of mass h&b and an
isolator block of mass Mlb., sometimes called bias
mass. Therefore Mvib = h& f Mlb. Those two
masses are interconnected via an isolation spring with
constant ki. In addition to the effort generated by that
spring, the mass mb is subjected to gravity (g) and
the sinusoidal force described by eq. (1) whereas the
mass Mib is subjected to gravity and the suspension
force T. The net quasi-stationary action on the pile
resulting from the carrier operation and vibrator is
the weight of the vibrator mass and its clamp Mcl de-
ducted by the suspension force:
2.3 VibratorMovemerit
The movement of the vibrated body will depend on
its so-called dynamic mass and the soil resistance.
Specifications of vibrators often list a “maximum
amplitude” S,. That number [generally expressed in
mm] corresponds to the total (i.e. double) amplitude
of movement for a free hanging vibrator, thus as-
suming a dynamic mass consisting of the exciter
block M,b and the clamping device ml:
Ssp = 2 so = 2 nze/(Meh + Mc,)
It should be noted that the double amplitude does
not depend on the operating frequency, as the center
of mass of the free mechanical system remains sta-
tionary, irrespective of the frequency. The amplitude
of the free hanging pile to be vibrated will always be
smaller than the specified amplitude, as can be de-
rived from eq. (4a), where the dynamic mass is in-
creased by that of the pile (PAL).
with s = actual (single) amplitude of the dynamic
mass.
A power for the vibrator is often listed in the
specifications. It generally corresponds to the nomi-
nal power of the motor actuating the eccentric
masses. It does not correspond to standardized op-
erational conditions of the vibrator in action. Power
consumption is indeed dependent upon testing con-
ditions. Barkan suggests that under pile vibratory
conditions, the power follows a squared velocity law:
W[kW] = M,, . n . ( s w ) ~=Pt.K
Experimental verification of that law shows the n
value to depend on soil type and pile type; a range of
15 to 50 Hz is observed.
Type Frequency range [rpm] Eccentric moment
[kg.ml
“Standard frequency” 1300-1800 up to 230
High fTequency 2000-2500
Variable eccentricity 2300 10 to 54
Excavator accessory 1800 to 3000
Resonant dnver 6000
O’Neill and Vipulanandan (1 989) provide an expres-
sion of the theoretical power required to maintain the
vibrating regime of a dynamic mass in the absence of
soil reaction but accounting for the presence of the
isolating spring and the bias mass Mib. That formula
is however of limited practical use as it provides very
low estimates of the power.
It is the author’s opinion that power limitation of
the equipment is neither sufficiently characterized,
nor (therefore?) properly accounted for in vibratory
driving analyses conducted to date.
2.4 Types of vibrators
Two main types of vibrators are commercially avail-
able: hydraulic and electrical. In both cases, the
motor is housed in the vibrator and powered through
a transmission line connected to a separate or carrier-
mounted diesel-hydraulic or diesel-electric power
pack (see Fig. 3a). Hydraulic vibrators are lighter
than their electrical counterparts, because of the
smaller size of the motor. The adjustment of the op-
erating frequency is more readily available on the hy-
draulic vibrators, which also explain why they are
more commonly used.
Five types of vibrators can be distinguished based
on operating frequency and eccentric moments, as
summarized in Table I .
It can be noted that initial improvements of the vi-
bratory driving technique targeted the speed of driv-
ing, whereas more recent improvements are at-
tempting to mitigate environmental impacts
associated with the technique. Noteworthy amongst
recent developments is the “variable vibrator”, which
can adjust on-the-fly its effective eccentricity by
shifting the phase angle between a multiple of 4
masses. The claimed advantage of such an adjust-
ment is to avoid “soil resonance”, a term coined after
the observation that vibration levels pass through a
peak upon vibration start-up and shut-down. This
phenomenon will however be shown later not to be
necessarily related to a particular fkequency.
Vibrator choice amongst practicionners is gener-
ally based on experience and field verification.
Rodger and Littlejohn (1980) have summarized that
body of experience into a table recommending fi-e-
quency and amplitude parameters for different piles
and soil types. Those recommendations are repro-
duced herein as Table 2.
Maximum centrifugal Free hanging double
force FNJ amplitude [mm]
up to 4,600 up to 30
6 to 45 400 to 2,700 13 to 22
600 to 3300 14 to 17
1 to 13 70 to 500 6 to 20
50 20,000 (in theory) Self destructing
482
 Cohesive soils Dense cohesionless soils
All cases Low point resistance
wlgh acceleration &gh acceleration
Low displacement amplitude
Predominant side resistancc Predominant side resistance. Predominant end resistance
Requires high acceleration for Iiequlres high acceleration Require? lugh displacement aniplitude and low Requires high acceleration
either sheanng or thixotropic for fluidization
transformation
3 SOIL BEHAVIOR UNDER VIBRATORY
LOADING
3.1 Fzn~dumentals
As the profile undergoes a vibratory vertical motion
of amplitude s, it communicates to the lateral neigh-
boring soil shear stresses and shear strains, as
sketched in Fig. 3b. It is also forcing normal and po-
tentially convective movement of soil below the pile
toe. As those mechanisms govern soil resistance along
the shaft and at the toe, the understanding of the
shear stresdshear strain relationship, i.e. T (y)? within
the soil becomes of paramount importance.
That aspect of soil behavior has been more exten-
sively studied within the field of earthquake engi-
neering, leading to the characterization of so-called
constitutive relationships, generally on the basis of
laboratory testing of soil samples (mainly triaxial
testing and simple shear testing). The constitutive
relationshipsthat represents the complex large-strain,
dynamic and cyclic shear stress-strain strength, be-
havior of the medium surrounding the vibrating pro-
file require the characterization of the following ele-
ments :
e Static stress-strain law expressing nonlinear be-
havior under monotonic loading and hysteresis
upon strain reversal,
Q Shear modulus at small strains and ultimate shear
strength,
Q Softening and increase of hysteretic damping
with increasing strain,
a Effect of strain rate on initial shear modulus and
ultimate strength,
o Degradation of properties resulting fi-om the ap-
plication of numerous cycles, and last but not
least,
e Generation of excess pore pressure leading sub-
stantial loss of resistance and possibly to lique-
faction.
The following paragraphs address key compo-
nents of the constitutive relationships and provide in-
sight on the intrinsic soil behavior in the vicinity of
1 Loosc cohesionless soils
Ihgh pomt resistance ] Heavy piles Light piles
Low frequency. Large displacement amplitude IIigh acceleration
Predominant side resistance.
frequency for maulmum impact to permit elasto- for fluid17ation
plastic penetration
the vibrating profile.
3 ,2 Static and Cyclic Stress-strain Behavior
A typical soil response to uniform cyclic strains
with amplitude yc is represented in Fig. 4, which
highlights the following hndamental parameters:
Gmax : initial (or tangent) shear modulus
zc : shear stress mobilized at yc
Gs: secant (or equivalent) shear modulus
h: hysteretic (or intrinsic) damping ratio;
A.= AW I 27cyczc (5)
with AW = Energy lost during a given cycle.
Both G, and h are strain-dependent parameters
that need to be described by specific laws within a
given cycle. T~~~ is the ultimate shear strength, re-
vealed at large strains. znlaVand G, are shown to
decrease with the number of cycles (cyclic degrada-
tion).
3.3 Initial Shear modulus and ultimate shear
strength (emu
and zmJ
Numerous studies have dealt with the initial shear
modulus to be used in earthquake engineering (e.g.
Drnevich et al., 1967). Most of them are supported
by parameters determined in the laboratory which are
generally not available when a vibratory penetration
issue arises. However, correlatins with CPT test re-
sults have been more recently developed (Seed and
De Alba, 1986, Robertson and Wride, 1998)
3.4 Secant Shear Modulus and Hysteretic Damping
(Gsand A.)
As can be observed in Fig. 4, G, decreases with the
shear strain during the initial monotonic loading. The
curve that represents the initial monotonic loading is
referred to as the initial "backbone" curve, because it
also serves as the basis to generate the
483

Figure 4. Soil Behavior under Constant Cyclic Shear Strain
Amplitude Loading (From Vusetic, 1993; 1994)
family of curves corresponding to unloading and re-
loading. Kondner's mathematical formulation ( 1963)
is frequently employed to describe the initial back-
bone curve in earthquake engineering. That hyper-
bolic law is best represented in terms of reduced
variables, q , the mobilization ratio and 6, the relative
shear :
q = zI z,, = 6 /(6+1)
and y, =zm /Gmax
Yr is called the reference strain. Two of the three
parameters Gmax, y,, and Tmax, are generally derived
from laboratory M~~~ extensive l a b -
ratory surveys by Robertson and Wride (1998) point
484
towards the upward curvature of the stresdstrain
curve at large cyclic strains.
From the point of maximum straining, the un-
loading curve is described by the following equation,
in accordance with Masing's rules 1 and 2 (Masing,
1926):
The energy dissipated within a loop depends for a
given soil on the amplitude of the cyclic strain. Em-
pirical data collected in laboratory tests indicates that
the damping ratio increases with yc as the soil under-
goes higher plastic deformations.
Dobry and Vucetic (1987, Vucetic and Dobry,
1991, and Vucetic,1993 and 1994) have suggested a
unifying approach to accommodate the influence of
the nature of the material characterized by the plas-
ticity index (PI), as indicated in Fig. 5
3.5 Strain Rate Effects
Although it is well known that undrained modulus
and shear strength increase with increasing strain rate
(P=ay /at ), experimental data generated using dif-
ferent apparatuses and loading conditions lead to
different conclusions. Viscosity mechanisms may
well provide a suitable framework to understand the
strain rate effect observed when comparing fast and
slow undrained monotonic stress-strain curves, as
well as to explain the roundness of the loop tips dur-
ing a sinusoidal strain-controlled cyclic test. Evidence
would point to the fact that sands and non plastic
silts have very small viscosity in that their stress-
strain loops exhibit sharp rather than rounded tips
(Dobry and Vucetic, 1987).
The mathematical fhctions proposed in the fit-
erature to represent the nonlinear viscosity also de-
pend on the type of experimental observations. A
power law is often adopted :
Figure 5. Soil stiffness degradation resulting from cyclic shear
(Vucetic, 1993)
'kin = Laboratory results conducted at constant cyclic
with Tkin = kinetic ultimate
shear strength [ P a l
xsta = "static" ultimate
shear strength [Wa]
= shear strain rate [s-]]
The advantage of that mathematical form is that re-
sistance does not vanish as the strain rate goes to-
wards zero. The power law also requires the strain
rate to vary by orders of magnitude to provide tangi-
ble increases in both the modulus and the ultimate
strength. The J coefficient and n exponent depend on
the nature of the soil. Based on pile driving data,
n=0.2 and J=0.3 have been suggested for plas-
tic soils. J should therefore essentially depend on the
plasticity of the soil and become quite limited for
granular materials.
3.6 Degradation Law
When subjected to undrained cyclic loading in-
volving a number N of large strain cycles, the soil
structure continuously deteriorates, the pore pressure
increases, and the secant shear modulus decreases
with N. This process known as cyclic stiffness deg-
radation can be best characterized on the basis of
strain controlled tests for the type of loading in-
volved with the vibratory penetration of piles. Typi-
cal results of strain-controlled tests are sketched in
Fig. 5 , where the degradation is clearly expressed by
the decrease of the amplitude of the peak stress mo-
bilized at successive cycles.
The quantification of the degradation process calls
for the introduction of the degradation index A, de-
fined by:
(9)
Figure 6. Effect of Plasticity Index (PI) on soil degradation
(Vucetic, 1993)
485
strain show that in many soils, the degradation index
after N cycles can be approximated by the following
relationship as suggested by Idriss et a1 (1 978):
A = A'-' (10)
The exponent t, called degradation parameter, de-
pends mainly on the amplitude of the cyclic strain and
the nature of the material (PI), as suggested by Do-
bry and Vucetic (1988) and as indicated in Fig. 6
Vucetic, 1993). It is noteworthy that the degradation
parameter assumes a zero value at strains smaller
than a cyclic "threshold" shear strain, y,, The thresh-
old strain increases with the plasticity of the soil, as
suggested in Fig. 6.
3.7 Soil liquefaction
Vibration induced compaction of saturated sands has
received attention not only from the earthquake en-
gineering community, but also from vibro-
compaction specialists.
Recent advances tend to indicate that build up of
pore pressures (eventually leading to liquefaction)
and volume reduction of cyclically loaded materials
are the expression of the same phenomenon, i.e. the
irreversible tendency for a particulate arrangement to
achieve a denser packing when sheared back and
forth.
Under drained conditions, the volume reduction is
immediate. Under undrained conditions, the tendency
for volume reduction is expressed by an increase in
the pore water pressure (see Fig. 7), such that the
effective stress is reduced to a value that may be
close to zero. It is then necessary to wait for the soil
to consolidate in order to see the volume reduction
take place.
The strain driven evaluation of the build up of
pore pressure as suggested by Dobry et al. (1 979) is
an approach that lends itself to a direct transposition
Figure 7. Build up of residual pore pressure in different sands
in undrained cyclic strain-controlled tests (Dobry et al., 1982)
to the problem of the vibrations induced by a verti-
cally vibrating pile. It also allows one to evaluate
potential changes of the void ratio based on a cyclic
strain rather than stress history, as supported by labo-
ratory drained tests conducted on sands by Youd
(1972). That framework of analysis entranced by the
threshold cyclic strain concept embodies in a single
model the intrinsic relationship between degradation
and pore pressure build-up, with the advantage that it
can be applied to general categories of soils (sands to
clays)
The excess pore pressure generated during cyclic
loading has been shown (see Fig. 7) to increase with
the shear strain and the number of cycles for a given
soil type. The damage parameter K approach (Finn,
1981) can be adopted to evaluate the excess pore
pressure 6u resulting from a particular strain history,
as characterized by the following equations :
d~ / d o = A / 4 In (1 + rd2 )
* (11)
with Relative Energy Loss given by Eq. 5, and
K =<ex’ (16) with C = 5 and
5 = length of strain path (12)
= 4 N yc, for constant amplitude cycles
4 PILE VIBRATORY DRIVING TESTING
The above discussion of soil behavior under cyclic
loading does not encompass the particular geometry
of the profile-soil interface, nor does it consider the
continuous penetration of the profile that leads to
successive exploration stages into “virgin” soil be-
havior. That is why a number of experiences have
been conducted to reveal soil-structure interaction
w i t h a vibratory framework. Based on the ambition
and complexity of the tested interface, one can cate-
gorize various experiences reported in the literature
as conceptual, interface, and both reduced and full-
scale testing.
4.1 Conceptual model testing
Tests have been conducted by several Russian re-
searchers to investigate the “vibro-viscous” resis-
tance of soils. In particular, Barkan (1963) reports
on the sphere test, shown in Fig. Sa, where a steel
ball is sunk into a vibrated soil vessel with the assis-
tance of a bias force. Penetration speed is shown to
obey Stokes sedimentation law (see Fig. 8b), allow-
ing one to determine an equivalent viscosity p . The
inverse of that equivalent kinematic viscosity
[cm.s/kg] was shown to vary linearly with the rela-
tive level of acceleration (a/g), passed a threshold
value of approximately 1.4 for a dry sand (see Fig.
8c). The influence of the water content on the “vibro-
viscosity factor l/p” of a sand vibrated at constant
a/g is also shown in Fig. 8d, highlighting the near
total loss of vibro-penetrability at optimal water
content.
4.2 Pile-soil interface testing
Soil shear strength resisting the pulling out of a vi-
brating steel plate against a normal stress controlled
medium sand (vibratory direct shear box) has been
investigated in the early days by Levchinsky and
Savtchencko (Barkan, 1963). The friction coefficient
(tan @ = do) was shown to decrease with cyclic am-
plitude and frequency. The ultimate relative reduc-
tion of the friction was also shown to increase with
the grain size within the investigated range shown in
Fig. 9. Fig 9 shows that the sand vibratory friction
angle can easily drop to !h to 1/5 of its static value.
4.3 Reduced scale tests
Testing of model profiles in soil tanks were initially
attempted by Bernhard (1968), Schmid and Hill
(1966), continued by Rodger and Littlejohn (1980),
Billet and SBert (1 985) and O’Neill et a1 (1990),
 and more recently by Viking (1 998) and Holeyman et
a1 (1999). The tests were generally conducted using a
lightweight vibrator acting on a heavily instrumented
profile. Monitoring included strain gauges, acceler-
ometers and displacement transducers. The soil used
was exclusively sand (dry, moist, or saturated),
placed at a controlled density, and in some cases,
confined at a controlled radial stress. Monitoring of
the soil response involved accelerometers, total stress
and pore pressure cells during installation as well as
compactlon and-in situ testing ager instalYation.
Insighthl observations relative to the vibratory toe
resistance have been reported by Schmid (1966),
who identified three regimes, depending on the mag-
nitude of the driving force :
a sinusoidal resistance domain, for a driving force
lower than the “resistance threshold”
an impact domain, when the upward force ex-
ceeds the soil uplift resistance; the toe of the pile
alternately separates from the soil and tamps it
a phase instability domain, when the downward
force exceeds the soil compressive resistance.
Rodger and Littlejohn (1980) call upon accelera-
tion amplitude to distinguish:
- the elastic state (a<0.6g)
- the trans-threshold state (0.6g<a<l.5g), wherein
most of the shear strength reduction takes place,
- the fluidized response state (a>l.5g).
Although their views are contradicted by some of
Barkan’s observations, these three different states are
stated to be confirmed by dynamic direct shear tests
performed by others.
Results of‘tarik experiments’have’ been reporte‘d’in
terms of friction reduction coefficients, potential op-
timal operation, and have shed some light on hnda-
mental soil behavior. Correlations have been estab-
lished between penetration speed and parameters in-
duced by the vibrator (amplitude, frequency) and by
the soil (grain size, relative density, and lateral
stress). Although conclusions of the tests conducted
under different conditions do not consistently agree,
those experiments generalJy identified that:
penetration speed increased when the relative
density decreased and the bias mass increased
friction was reduced to 30 to 50% of its static
value, while a more limited reduction was noted
for the toe resistance
optimum operation of the hammer required at
times that the frequency or eccentric moment be
reduced, while energy transfer was of the order
of 40% of the hll theoretical power produced by
the vibrator
a number of observations cannot be explained.
Although reduced scale models are of use, they
suffer from improper boundary conditions (at the
tank limits) that significantly prevent the vibration
energy from propagating away from its source.
487
4.4 Full scale tests
Because of inconsistencies in the conclusions derived
from reduced scale tests, research has been con-
ducted in several countries based on full-scale tests.
Early hll-scale programmes have been conducted by
Barkan (1963) and Davisson (1970). Other pro-
grammes have been conducted by manufacturers on
specific equipment, but lead to a limited diffusion of
their conclusions. More recently, collective European
programs have provided actual penetration speed,
but within soil conditions that cannot be controlled,
only characterized. Monitoring nowadays involvies
acceleration, strain, pore pressure, penetration speed,
making the tested profile a hlly instrumented probe.
Such programs have produced results that have not
been hlly analyzed (BBRI, 1994, Sipdis, 1997); oth-
ers are being presently conducted (KTH, 1999) or
planned (IREX, 2000). Publication of such research
results would be appreciated by the profession.
5 PILE-SOIL-VIBRATOR INTERACTION
MODELS
5.1 Types of models
Models that have been suggested by various authors
differ in the way they account for mechanical engi-
neering principles. We will review models purely
based on (1) force equilibrium, (2) momentum con-
servation, (3) energy conservation, and (4) integra-
tion of the laws of motion.
5.2 Force equilibrium models
The force models aim at predicting whether a vibra-
tor can or cannot overcome an estimated soil resis-
tance. They will not provide an estimate of the driv-
ing speed. Jonker (1987) and Warrington (1989)
have suggested, respectively:
force generated by the vibrator, per eq. (1)
inertia forces of dynamic mass, = Mdp . a
surcharge force, per eq. ( 3 )
empirical factor of shaft resistance outside
pipe pile,
soil resistance outside pile shaft,
empirical factor of shaft resistance inside
pipe pile,
soil resistance inside pile shaft,
soil resistance at Dile toe.
For sheet-piles Tunke; Company recommends to
replace x. z [m2] with 2.81 times the sheet-pile
width.
5.3 Energy based models
Energy based models assume the following general
form:
R.v, = p, .W, + ( F , +F').v, (1 4 4
leading to a direct estimate of the penetration speed :
vP=p, W,/ ( R -F,
* -F,) (14b)
With:
R = soil resistance,
v,, = average rate of penetration in m/s,
W, = theoretical power delivered to the system,
Fi = inertia forces of dynamic masses.
Davisson's formula (1 970) to estimate the bearing
capacity for the Bodine Resonant Driver suggests :
p, =I--v*s,. R/1000 w, (15)
where s,: is an empirically determined set [mm/cycle]
representing all energy losses.
Warrington (1989) has coined eq. (14b) as the
'Vibdrive' formula provided a value of 0. I is used for
p, and the power Wt is calculated according to his
procedure,
5.4 Momentum conservation models
Schmid ( I 968) has suggested a formula implying
that, for steady-state penetration, the momentum of
the total mass of the vibrator. additional bias mass
(MJ, and pile accrued by gravity over a vibration cy-
cie be balanced by the soil resistance impulse:
(Ads+hd,,,b+M,)g T= 1 6;Rdt=&T, (164
with T, = contact time between pile toe and soil
within a cycle and a = coefficient between 0.5 and 1,
generally assumed to be 213.
Conversely, the penetration speed follows a linear
trend passed the threshold acceleration amln,which
becomes a key parameter to successfdly apply the
niethod and estimate T, :
5.5 Integration of law.^ of'rnotion
Comprehensive accounting of the laws of mechanics
requires that movement be described at all times from
inertial equilibrium conditions. The simplest models
involve a single degree of freedom. I-D models al-
ready offer more detailed description of some form
of wave propagation, whereas 2-D models might
provide hture solutions that integrate all types of
wave propagation (compression, shear, Rayleigh,
etc..).
5.5.1 Single degree of freedom (SDOF)
Simplest models of the vibrator suggest that the dy-
namic mass be the focus of attention, thereby as-
suming that the pile behaves as a rigid body. New-
ton's second law can therefore be applied to the
dynamic mass :
mem2sin(wl)
a= (174
hrfIn,
where
U
,, =Meb +Md+ M , ( 17b)
Holeyman (1993) has suggested a method that
integrates the inertial effects of the excess force. That
excess force is defined as the difference between the
sinusoidal driving force and the opposing soil resis-
tance. A distinction is made between the skin ftiction,
which is reversible (Uplift resistance = Downward
resistance) and the toe resistance, which cannot pro-
duce uplift resistance. Attention is also paid to the
clutch resistance, which is combined with the skin
friction.
The soil degraded resistance at the toe and along
the shaft is estimated from CPT test results where the
friction ratio and acceleration ratio are used to assess
the severity of degradation. The method involves an
iterative procedure to identifL the coexisting accel-
eration and soil resistance (17b). The driving speed is
obtained by intuitively integrating the net downward
and upward accelerations over a complete cycle. The
method have been verified and liquefaction parame-
ters hrther refined through calibration with full-scale
tests (BBIPI, 1994)
Gonin (1998) has followed a similar approach that
analytically integrates the effects of an excess force,
as shown in Fig. 10. The integration is however per-
formed soleby on the toe resistance, while the skin
friction influence is accounted for in terms of damp-
ing of the driving force. In addition, the wave equa-
tion theory is used to estimate the displacement ac-
crued at the toe over the period of net force
exceedance.
Figure 10. Integration of excess toe force (after Gonin, 1998)

Figure 11. Resistance mobilization versus displacement for (a)
skin (b) toe compression (After Dierssen, 1994)
Dierssen (1 994) has used a numerical integration
scheme to closely follow the time dependence of the
skin and toe resistances. Figure 11 provides the
shape of the resistance mobilization versus displace-
ment for both skin and toe resistance. One can note
that separation of the pile from the soil at the toe is
explicitely accounted for.
5.5.2 Radial 1-D model
Holeyman (1993b) have suggested the use of a radial
discrete model to calculate the vertical shear waves
propagating away from the pile. The model, shown in
Fig. 12, consists in a succession of concentric cylin-
ders with a linearly increasing depth. The equations
of movement are integrated for each cylinder based
on their dynamic shear equilibrium in the vertical di-
rection, in a manner similar to that used by Smith
(1 960) in the longitudinal direction.
Figure 12. Radial I-D model (Holeyman, 1993b)
489
The model allows the constitutive relationships
described in Section 3 to be readily deployed. The
major advantage of that shear wave propagation
model is to closely follow the development of degra-
dation as more cycles are simulated. It can also pro-
vide insight into vibration levels in the vicinity of the
pile. Both features are illustrated by Fig. 13 which
provides the effective particle velocity calculated at
several distances away fiom a profile upon vibrator
start up. An apparent resonance is indicated,
whereas the model does not include a Iongitudinal or
radial dimension that could explain the frequency at
which the peak vibration is noted : why? Simply be-
cause the model most probably reproduces two soil-
pile interaction vibratory modes: the coupled mode
and the uncoupled mode.
In the coupled mode (similar to Schmid’s sinusoi-
dal domain), the soil remains in contact with the
slowly vibrating profile, and the transfer of energy
from the pile to the soil is nearly perfect. As the vi-
brator linearly accelerates (between 0 and 0.5 sec-
onds), vibration levels tend to increase with the
square of time since start up. However, as the soil
begins to degrade, its shear modulus decreases and
the specific shear impedance reduces, leading to loss
in the energy transfer. At that point, the couplmg
between soil and pile suffers some slippage, and
therefore time lag. After a sufficient number of cy-
cles, the soil has significantly degraded, and has (60
seconds ageing skipped in Fig. 13) entered into li-
quefaction.
Figure 13. Vibration levels and penetration state parameters
estimated upon vibrator startup and regime
 At the shear modulus of the soil in contact with
the profile is nearly zero, and very little energy can
pass through the fluidized surrounding zone. The soil
in the vicinity of the profile cannot anymore follow
the profile movement, from which it uncouples itsel,
resulting in a lower level of vibration. That example
demonstrates that apparent resonance of soil vibra-
tion may be no more than the transient combination
of increased rotation speed and soil degradation. The
model can also shed light on “damping” as it clearly
separates geometric damping from the energy losses
attributable to viscous and hysteretic behavior.
5.5.3 Longitudinal I-D models
Few authors have adapted Smith‘s (1960) classic
lumped parameters model to represent the longitudi-
nal behavior of a pile subjected to vibratory driving.
Gardner (1981) and Chua et al. (1981) have devel-
oped a wave-equation computer code where the vi-
brator is represented by a two-mass system, sepa-
rated by a sofi spring, while the exitor black is
subjected to a sinuso‘idal force (cfi-. eq. (I)); as
shown in Fig. 14. The soil behavior is represented by
spring-slider-dasplot systems, according to Smith’s
early suggestion.
Middendorp and J o k e r (1988), as well as Ligterink
et al. (1990) used the TNOWAVE computer pro-
gram to analyze the driveability of offshore vibratory
driven pipe piles, based on the methods of character-
istics. The authors identify the need for a soil model
able to describe the degradation of the soil resistance
as a function of the oscillation history, and warn that
soil parameters may depend on opera- ting fiequency
and pile movement amplitude.
Figure 14. Longitudinal I-D model
490
Moulai-Khatir et al. (1994) have developed to-
gether with the University of Houston, the so-called
VPDA computer program (for Vibratory Pile Driving
Analysis) wherein the action of the hammer is re-
placed by a static surcharge load and a sinuso‘idal
load. The soil model was modified from Smith‘s
original in that hyperbolic mobilization curves were
adapted for the shafi and toe resistance, as shown in
Fig. 15. A simple viscous damper was used to model
damping along the shaft, while no damping was
deemed necessary at the pile toe.
It should also be noted that the GRLWEAP pro-
gram has included in its latest versions (GRL,, 1998)
the capacibilty to model vibratory hammers.
6 VIBRQ-DRIVABILITY ANALYSIS
Most of the models discussed in the previous sec-
tions should be able to provide a reasonnable match
of calculations with relevant field observations pro-
vided the models parameters are properly calibrated.
The use of energy balance methods is discouraged
by the author, while force equilibrium methods are of
limited use because they do not provide vibratory
penetration speed. Momentum based methods m y
produce a penetration speed very similar to that ob-
tained through integration of the laws of notion of a
rigid body. Finally, wave equations methods should
not produce penetration speeds significantly different
form those obtained form a rigid body analysis, pro-
vided the vibrator speed is lower than the resonant
frequency of the pile, which is generally the case.
Figure 15. Resistance mobilization for (a) skin friction (b) toe
compression (After Moulai - Khatir et al., 1994)
Exceptions to that general case include the Bodine
Resonant driver and very long piles for offshore ap-
plications (LMO m).
In the author’s opinion, the most critical parameter
to assess in order to produce a resonable prediction
of vibro-drivability is the soil resistance to vibratory
driving.
That is unfortunately where pertinent information
and recent consistent experimental data is cruelly
missing. The reliability of the predicted vibro-
penetration log will strongly depend on the degrada-
tion parameters adopted to assess the vibratory
penetration resistance form the soil investigation re-
sults. The author’s experience leads him to use the
following crude ultimate degradations coefficients :
0.15 in sand, 0.4 in silt, and 0.65 in clay for skin fric-
tion; as well as 0.55 in sand, 0.7 in silt and 0.85 for
end bearikg.
A more involved assessment of the degradation
coefficient has been suggested (Holeyman, 19%)
based on CPT test results. In that method, the soil
driving resistance is obtained by interpolation be-
tween a static value and an ultimately degraded
value. The static base (qs) and shaft (zs) resistance
profiles derived from Cone Penetration (CPT) tests
results, i.e. from the cone resistance qc and local unit
skin friction f, (El cone).
The ultimately liquefied base (ql) and shaft (z1)
unit soil resistances are derived based on an expo-
nential law as expressed below :
rl=r,[(l-1/ A).e-l”;R+1/A] (1 8b)
where
= liquefied soil base resistance [Wa]
41
21 = liquefied soil shaft resistance [@a]
FR = friction ratio as measured in a CPT test with
E l cone (percentage of the mantle friction to
the cone resistance, i e FR = 100 f, /qc)
A = empirical liquefaction factor expressing the
loss of resistance attribuable to liquefaction
(A will be higher for saturated and loose
sands and is chosen in the range of 4 to 10)
The driving base (qd) and shaft (zd) unit resis-
tances are derived from the static and the “liquefied“
soil resistance depending on the vibration amplitude
following an exponential law as expressed below
qd=(q -ql ).e +qr (194
where
qd = driving base unit resistance
zd = driving shaA unit resistance
a = acceleration ratio (= dg)of the pile, as ob-
tained from Eq. (1 7a)
49 1
At each depth z the vibratory pile driving resis-
tance is calculated :
Rbase =qd
where $2is the pile section, x the pile perimeter and
D the pile penetration.
7 BEARING CAPACITY FROM
INSTALLATION MQNITOR1ING
Because soil resistance degradation is significant
during vibratory driving, one should expect it a chal-
lenge to estimate the static bearing capacity from the
end of penetration vibratory performance of a driven
profile.
In the impact driving practice, it is recognized that
end of driving (EQD) data generally provides a safe
estimate of the pile capacity; that is why beginning of
restrike (BQR) or “retap” data is strongly advised to
the owner who wishes to tap the value of letting the
soil set up. If the end of Vibratory driving (EOV)
data is used, methods to estimate the static capacity
should allow for recovery of soil degradation, as
highlighted in Fig. 2. However significant uncer-
tainty should be expected in the process because the
inverse of observed degradation coefficients may
range between 2 and 10.
That is why extreme caution is warranted when
applying so-called pile Vibratory driving (PVD),
formulae, even more so than already much detracted
(impact) pile driving formulae. A limited number of
such PVD formulae have been published; however
only one has been, to the author’s knowledge been
extensively field tested. The “Snipe” formula is
therefore the only one that will be discussed in this
paper.
The formula is a field-based method was devel-
oped in the former Soviet Union according to
Steffanof and Boshinov (1 977). The following em-
pirical formula is used to predict the static bearing
capacity Q,, :
where
QU = load capacity, in [kN];
w = power used by the vibrator to drive the pile,
in [kW]
Ft = total weight (force) of vibro-hammer and pile,
in [kld]; = (M,ib + MJg
1/p = empirical loss coefficient (in Soviet practice
143 is safety taken to be 5 in cohesionless
 soils) reflecting the influence of driving on
soil properties.
The bearing capacity in of vibro-driven pipe piles
has been verified by A monitoring the behavior of
the finished product the beginning of impa
strike. This of course requires that a specific
procedure be enforced a sufficient time aft
EOV installation, in order to allow pore pressures to
dissipate and accrue soil setup.
The monitoring of the installation of vibratory
driven piles is not at all as widely spread as for im-
pact driven piles. Recent improvements in the field
monitoring devices FaSA., “NO-System, etc) now
allows the geotec’nnical engi
celeration, stress and energy
equivalent interpretation of
formula or CAPWAP method, available for nearly 15
years in the impact driven products.
Field monitoring provides a tremendous advan-
tage in controlling the effective performance of the
vibratory hammer, as illustrated by the following case
history reported by Holeyman et al. (1996).
That case involved the installation of a 20.6 m
long tubular steel pile with a thickness of 9.5 mm and
a diameter of 1 m on a site in Kortrijk (
shows the subsoil profile as depicted by a CPT test
(cone M4) performed at the site. The water table
was encountered at a depth of -1.8 m. The upper
twelve meters consist of very soft river deposits; be-
low the underlying sandy layer was found a very stiff
tertiary clay layer in which the tube had to be driven.
A preliminary calculation using eqs. (1 7) through
(20) pointed out that the necessary time to install the
pile to a depth o f 2 0 m with a PTC 3 0 W V vibratory
hammer was 13% minutes. However, driving met
rehsal at a depth of 11 m.
The reason for the difficult driving and the differ-
ence between the predicted and the observed pene-
tration speed was explained by measurements taken
during the actual driving of the pile. The pile vibra-
tion amplitude was measured by means of a velocity
transducer placed at the pile head and a velocity
transducer (protected by a cover) at the pile toe.
Figure 17 shows the monitored amplitude of vibra-
tion at the pile top and at the pile base upon loss of
Figure 16. Subsoil profile site at Kortrijk
492
drivability. The observed frequency was 38 fi.
From the measurement results, one can observe
that :
- the vibration amplitude at the pile top (0.65 mm
zero to peak) is considerably less than the nominal
vibration amplitude which is,
me
-- -
me -
26000kg.mm
= 2.3mm
M +M (6500 + 4820)kg
- the amplitude at the pile base (0.45 mm zero to
peak) is smaller than the amplitude at the pile top
(0.65 mm)
It would appear that the pile base amplitude (0.45
mm) is not sufficient to allow the pile to penetrate as
the stress-strain behaviour for clayey soils is primarily
elastic for small amplitudes. ossible explanations
for that observation were considered :
- An important soil (i.e. clay) mass was sticking to
the vibrating pile, leading to a more important vi-
brating mass, leading to a smaller vibration am-
plitude.
- The vibratory hammer was unable to deliver the
required energy, and thus maintain its nominal
plitude or frequency. A characteristic of the
C variable eccentric hammers is that a lack of
power results in a reduction of vibration amplitude
(rather than a reduction of frequency (Houze,
1994)).
- A smaller amplitude at the pile base was obtained
due to the elasticity of the pile.
By applying the observed vibration amplitude to
the calculation model (Figure 181, a much better cor-
relation between the calculated and the observed
penetration time was obtained. The pile was placed
at the bottom of an excavation at -2.5 m and pene-
trated 4.5m under its own weight. As a result, ob-
served and calculated penetration rates are reported
starting at level - 7 m. Figure 18 evidences that the
difference for the predicted and observed penetration
times for the site in Kortrijk was not due to an incor-
rect estimation of the dynamic soil resistance but due
to an incorrect estimation of the vibration amplitude,
which happened to be limited by the nominal power
of the power pack. A more powefil power pack
Figure 17. Record from the vibration amplitude upon refusal

Figure 18. Predicted and Observed penetration log at Kortrijk
site compared with prehcted log using actual vibratory am-
plitude
was brought on site and the piles could be vibrated to
design depth using the same vibrator.
After reviewing the present state-of-the-art of vibra-
t o driving,
~ ~ the following suggestions for krther
consideration are offered '
soil mechanics research is needed in the area of
large cyclic deformation to better understand and
assess the effects of degradation and liquefaction
under those extreme conditions,
full scale vibratory driving tests, with extensive
field monitoring, will be required preferably to
reduced scale laboratory tests, which suffer from
improper energy dissipation boundary conditions,
potential and transferred power of vibrators need
to be better defined, as well as modeled for better
description of the mechanical behavior of vibra-
tois,
peak vibration of the soil surrounding a profile
upon vibrator start up does not necessarily imply
soil resonance; it can also result from the combi-
nation of increasing fi-equency and degrading soil
resistance,
monitoring of vibrated profiles is recommended
with the view to emulate the benefits accrued by
a similar practice for driven profiles, and
procedures for vibratory loading tests should be
developped
NCES
Barkan. D D . (1963). Methodes de vibration clans la con-
struction. Dunod. Fans, 302 p (Plreiich traidatioi? of
Originnl 117 1?uwa~i" Vibsometod v StroitelJSlVe. 1960)
BBRI (2994) High performance vlbralopv pile h v e r s base on
novel electromagnetic actuation systems and Iinproved UiI-
derstandng of soil dynamics, Progress reports of the
BR17I9ElJRAhl research contrucl CTYI-0.561, 19%
Ben&arcL R K . (1967), Fluichzation phenoniena ~n soils dur-
ing vibro-cornpaction and vibro-pile-hvmg and -prrllrng
Hanover. NH 1967. 58 pp CIS Alrnzv Cold Xeg1on.s Re-
~
search and Engineering
493
Billet, P . Siffert. J G . (1989) "Soil-sheet pile interaction in
\ ibro-piling" Journal of Ceotechnical Engineering, ASCE.
Vol 115, No8. pp 1085-1101
Chua, K M , Gardner. S . Lowery, L L . (1987) "Wave Equa-
tion Analysis of a Vibratory Hammer-Dnveii Pile". Proc
Oficrore Techm1og-y Coiif. VoI 4. pp 339-345
Davisson. M T . (I 970) "BRD Vibratory driving formula".
Foundatioii facts. Vol 1, No1. pp 9-11
Dierssen. Guillermo. ( I994), "Ein Bodenmechanrsches Modell
zur Beschreibung des Vibrationsrammens in komngen
Bodcn" Doctoral Thesis, University of Karlsnihe. Ger-
many
Dobry. R , Ladd. R S . Yokel. F V . Chrang. R M and Fowl!.
D (E982) "Prechction of Pore Water Pressure BuiPdup and
Liquefaction of Sands Dtrnng Earthquakes by the Cyclic
Strain Method" Uafional Bureau of Standards Building
Science Series 138, July 1982. 150 pp
Dobry. R and Swiger, W F (1979) "Threshold Strain and
Cyclic Beliavior of Coliesioiiless SO~IS" ~ r o c 3 1 ~
Specialiv Conference A ~ ~ s t i nTexas. pp
52 1-525,
Dobry. R and Vircetrc M (1987) "State-o€-the-Art Report
Dvnamic Properties and Response of Soft Clay Deposits"
Proceeding\ of the In11 Svmponurn on Geolechnicnl &,a
of.S'ofi Soils. Mexico City. Vol 2. pp 51-87
Drnevrch I J P Hall. J R . Jr , and kchart. F E Jr (1967)
"Effects o f Amplit~dcof Vibration on the Shear Modulus
of Sand Proceediiigs of the Iiifernatroi?al Svmposium on
Fave Propagatioii nnd Dvnat~iicPropertier of Earth hfu-
ferialr, Albtrquerque. N M . pp 189-199
Finn. Vd D L (1981) "Liquefaction Potential Developments
Since 1976". Proceedings, Intl Conf oii Recent Ad-
vaiices in Geofechnical Earthquake Eiigineering and wrl
l A ~ i i a ~ ? ~Sifc sLouis,
, Missouri. Vol I€, pp 655-681
Garclner SherriIl. (1987) "Analysis of vibratory driven pile"
~ r o cof 2'ld Inl ~ o n Of M ~ e e ~orni~atron,
p Luuenibourg,
5-7 May, pp 29-56
Gonin. J (1998) QueIques reflevions sur Ie vibrofonqage. Re-
vue Fraiqai re de Geoiechnique, No 83. 2'"Ie trimestre. pp
35-39
Harchn. B 0 and Black, W L (1968) "Vibration Modulus of
Normally Consolidated Clay mrnal of the Soil Ale-
chmnc I aim' Fowm'atio~rsllil n, ASCE. Vol 94. No
SM2. Proc Fapeer 5833. pp 353-369
Holevnan. A (1985) "Dynamic non-linear slun friction of
piles." Proceedings of the International Synposiurn on
Peiiefrdvlifv and Drrvabik fy of Piles, San Francisco. 10
August 1985. Vol 1. pp 173-176
Holeyman. A (1988) "Modeling of Pile Dvnainic Behavior at
the Pile Base during Driving," Proceedings of the 3rd Iii-
feriiafionnl Confereiice on the Applicnfion of Stress-Ffni e
Thenrv fo Piles. Otlataa. May 1988. pp 173-285
Holeyman. A (19932) "WUPERVIBI An analytical model-
based computer prograni to evaluate the penetration speed
of vibratory dnven sheet Piles". Research report prepared
for BBH. June. 23p
Holcymamr, A (1993b) "HYPERVIP3IIa. An detailed numenca!
model proposed for Future Computer Implementation to
evaluate the penetration speed of vibratory chrrven sheet
 Piles", Research report prepared for BBRI, September,
54p.
Holeyman, A. & Legrand C. (1994). Soil Modelling for pile
vibratory dnving, International Conference on Design and
Construction of Deep Foundations, Vol. 2, pp 1165-1178,
Orlando, U.S.A.. 1994.
Holeyman, A., Legrand, C., and Van Rompaey, D., (1996). A
Method to preQct the driveability of vibratory dnven piles,
Proceedings of the 3rd International Conference on the
Application of Stress-Wave Theory to Piles, pp 1101-1112,
Orlando, U.S.A.. 1996.
Howe, C. (1994). HFV Amplitude control vibratory ham-
mers : piling efficiency without the vibration inconven-
ience, in DFI 94, pp. 2.4.1 to 2.4.10, Proceedings of the
FiBh International Conference and Exhibition on Piling
aiid Deep Foundations, Bruges, Belgium, 1994.
Idnss, I.M., Dobry, R, and Singh. R.D. (1978). "Nonlinear
Behavior of Soft Clays during Cyclic Loadmg." J. Geo-
technical Engineering Div. ASCE, 104(GT12), pp. 1427-
1447.
IREX (1998), VibrofonGage des pieux et palplanches - Etude
exploratoire, (by Le Tirant, P., Borel, S., Gonin, H., Guil-
laume. D. and Longueval, A.)
Jonker, G., (1987). "Vibratory Pile Driving Hammers for Oil
Installation and Soil Improvement Projects". Proc. of
Alineteenth Annual Ofshore Technology Conf:, Dallas.
Texas, OTC 5422, pp. 549-560.
Kondner, R. L., (1963). "Hyperbolic Stress-Strain Response:
Cohesive Soils." .Journal of the Soil Mechanics and Foun-
dations Division, ASCE, Vol. 89; No. SMI, pp. 115-143.
Jan.
Ligterink, A., van Zandwijk, C., Wddentorp, P., (1990). "Ac-
curate vertical pile installation by using a hydraulic vibra-
tory hammer on the Arboath project". Proc. 2Jnd annual
Technology Conf: in Houston, Texas, May 7-10, pp. 315-
326.
Masing, G. ( 1926), "Eigenspannungen und Verfeistigung
beim Messing", Proceedings of Second International
Congress ofApplied Mechanics~pp. 332-33 5.
Mdendorp; P. and Jonker, G. (1988). Prediction of Vibratory
Hammer Performance by Stress wave Analysis, Preprint to
the 3'd Int. Cone on the Application of Stresswavre Theory
to Piles, Ottawa
Moulai-Khatir. Re&. ONeill, Michael W., Vipulanandan, C.:
(1994). "Program VPDA Wave Equation Analysis for Vi-
bratory Driving of Piles", Report to The U.S.A. Army
Corps of Engineers Waterways Experiments Station. Dept.
of Civil and Environmental Engineering, UHCE 94-1.
Univ. of Houston, Texas, August 1994, 187 pp.
Novak, M., Nogami, T., and Aboul-Ella, F. (1978). "Dynamic
Soil Reactions for Plane Strain Case", J. Engrg. Mech.
Div., ASCE, 104(4), 953-959.
NRC (1985). "Liquefaction of Soils During Earthquakes." Na-
tional Research Council Committee on Earthquake Eng-
neering, Report No. CETS-EE-00 1, Washington, D.C.
ONeill Michael W.: Vipulanandan, C., (1989). "Laboratory
evaluation of piles installed with vibratory dnvers". Na-
tional Cooperative Highway Research Program, Report no
494
316>National Research Council, Washington, DC. Vol. 1
pp. 1-51. ISBN 0-309-04613-0
ONeilI Michael W., Vipulanandan, C., Wong, D., (1990).
"Laboratory modelling of vibro-driven piles". Journal of
Geotechnical Engineering, ASCE, Vol. 116, No 8: pp.
1190-1209
Roberlson, P.K. and Wride, C.E. (1998). "Evaluating cyclic
liquefaction potential using the cone penetration test", Ca-
nadian Geotechnical Journal, No. 35, pp. 442-459.
Rodger, A.A. and Littlejohn, G.S. (1980). "A study of vibra-
tory h v i n g in granular soils". Geotechnique, Vol. 30, no
3, pp. 269-293.
Seed, H.B. and Idnss, I.M. (1970). "Soil Moduli and Damping
Factors for Dynamic Response Analyses." Earthquake En-
gineering Research Center, College of Engineering, Uni-
versity of California, Berkeley, Report No. EERC 70-10.
Seed, H.B. and De Alba, P. (1986) YJse of SPT and CPT
Tests for Evaluating the Liquefaction Resistance of Sands"
Proc. INSITU 86, V A 22 p.
Schmid, W.E., (1969). "Driving resistance and bearing ca-
pacity of vibrodriven model piles". American Society of
Testing and A4aterials Special Techn., Publ. 444, pp. 362-
375
Schmid, W.E. & Hill, H.T. (1967). "A rational dynamic equa-
tion for vibro driven piles in sand". Symp. Dynamic Prop-
erties of Earth Materials. New Mexico Universityy:p. 349.
Smith E.A.L., (1960). "Pile-dnving analysis by the wave
equation". Jourrzal of the Soil A4echanics and Foundatioris
Divisions, ASCE, Vol. 86, August 1960.
Vilung, K., (1997). "Vibratory dnven pile and sheet piles - a
literature survey", Report 3035, Div. of Soil and Rock Me-
chanics, Rqyal Institute of Technology, Sweden, 75 p.
Vucetic, M. and Dobry, R. (1988). "Degradation of Marine
Clays Under Cyclic Loading. ASCE Journal of Geotechni-
cal Engineering,Vol. 114, No.2,pp. 133-149.
Vucetic, M. and Dobry, R. (1991). "Effect of Soil Plasticity on
Cyclic Response." ASCE Journal of Geotechnical Eiigi-
iieering, Vol. 117, No. 1, pp. 89- 107.
Vucetic. M. (1993). "Cyclic Threshold Shear Strains of Sands
and Clays", Research Report, UCLA Dept. of Civil Engi-
neering, May 1993.
Vucetic, M. (1994). "Cyclic Threshold Shear Strains of Sands
and Clays", Paper in print, ASCE Journal of Geotechnical
Engineering
Wamngton, Don. C.: (1989). "Driveability of Piles by vibra-
tion". Deep Foundations Institute 141h Annual members
Con$, Baltimore, Maryland, USA pp. 139-154.
Westerberg. E.. Massarch, K. Rainer. Eriksson, K., (1995).
"Soil resistance during vibratory pile dnving". Proc. to hit.
Symposium on Cone Penetration Testing, Linkoping, Swe-
den, Vol. 3, Report 3.41, pp. 241-250.
You4 L.T. (1972). Compaction of sands by repeated shear
straining. Journal of the Soil Mechanics and Foundaiioiw
Division, 1972, Proc. ASCE, Vol. 98, SM7, pp. 709-725.
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) (c> 2000 Balkema, Rotterdam, ISBN 90 5809 150 3

Computation of ground waves due to piling

C. L. Ramshaw, A. R. Selby & P. Bettess
University cfDurhanz, U K

ABSTRACT: Pre-formed piles are normally installed into the ground using high energy impact or vibratory
hammers. Part of the energy is absorbed in advancing the pile, and much of the remaining energy is transmitted
into the ground in the form of outgoing vibrational waves. The current study is towards developing a
computational procedure to model the outgoing waves, to calibrate the waves against existing site
measurements, and then to expand the model by including simple structural forms so as to estimate structure
response, including dynamic soil-structure interaction. This paper reports the stage of the work which generates
ground vibrations and some calibration against site data. For impact hammers, the analysis has to be tackled in
three phases, of hammedpile impact, pileisoil interface effects, and finally the transient outgoing P, S and R
waves. For vibro-driving a different procedure is required. A two stage approach makes an estimate of rigid
body oscillation of the pile within a localised soil framework, and then the boundary disturbance is applied to
an analysis of the wider soil area for computation of the sinusoidal wave disturbances. Two case studies are
presented.

1. INTRODUCTION must first be calibrated against a substantial

Pre-formed piles, of steel, concrete or timber, are
widely used for transmitting building loads from
ground surface down through weak soils to more
competent soil or rock strata, while interlocking
sheet piles are used for temporary or permanent
retaining walls. Pile installation is achieved using
high energy vibratory or impact hammers. Some
energy is lost into the ground as outgoing P, S or R
waves. These vibrations may disturb neighbours
(Head & Jardine, 1992), and cause cosmetic
damage to nearby structures (Wiss, 1967, Selby,
1991, Todd, 1994). Analysis of risk from vibrations
has been attempted by empirical estimates of free
ground vibrations and arbitrary vibration thresholds
for building types (Attewell & Farmer, 1973,
Eurocode 3 part 5 , 1996) with some recognition of
frequency effects. Structural analysis for vibration
enhancement has been attempted in isolation.
An ongoing study is directed towards generation of
pile-soil models leading to representative ground
vibrations from driving, to be followed by pile-soil-
structure models which will incorporate dynamic
soil-structure interaction. However, the procedure
database of site records of “green-field” vibrations
(Uromeihy, 1990), which is the subject of this
paper.
The use of finitehfinite elements to generate
ground waves has been demonstrated (Ramshaw et
al, 1998), using the package ABAQUSO, and a
relatively simple energy transfer at the pile-soil
interface. This has now been extended, so as to
model the energy input from the hammer, the stress
waves in the pile, and the subsequent outgoing
ground waves. However, because of the complexity
of the system, and the need to use small nodal
spacings (x<h/lO) and very short time steps
(&<x/c), it has been found most efficient to use a
sub-model of the near-field to model the pile-soil
interaction and energy transfer, with the computed
edge response used as input to the wider soil model.
The pile-soil interface is modelled using a surface-
based contact simulation, with slip controlled by a
Coulomb friction model. In the wider soil field,
axisymmetric finite elements transmit the outgoing
P, S and R waves, and infinite elements are used to
prevent artificial wave reflection.

495
Calibration against real site data is difficult because
of the many physical parameters involved.
However, good agreement of both vertical and
radial surface vibrations, defined by velocity-time
traces, has been achieved for a number of site
records. These include bearing piles installed by
vibrodriver at the second Severn crossing and
impact driving on the M66 near Manchester.
1.1 General aims of paper
At present, guidance on vibrations resulting from
various types of pile driving is somewhat limited.
Various standards exist, but these are empirical and
in some cases are conflicting. While empirical
equations have been derived for estimation for peak
surface vibrations, detailed computation of- the
ground waves is lacking. This paper presents
simplified methods using finite element techniques
for predicting surface vibrations resulting from both
vibratory and impact pile driving. The methods
have been designed to be used as a preliminary
design tool for the prediction of ground surface
vibrations where site data are sparse. They are not
intended to be rigorous geotechnical analyses.
2. SIMULATION OF IMPACT PILING
The procedure developed here breaks the overall
problem into three stages, comprising:
A hammer impact model to simulate the force
imposed onto the pile head
A model of the propagation of the impact waves
down the pile shaft (soil response modelled by
springs and dashpots)
A model of the surrounding ground to simulate
the outgoing ground waves.
There are two advantages of breaking down the
computations in this way. Firstly, each parameter or
group of pmimeters can be isolated, and rapid
parametric studies can be used to ascribe values.
Secondly, more efficient computations are possible;
a finite element (FE) model of the pile requires very
small nodal spacings (x<h/lO) and very short time
steps (&<x/c, c=wave velocity), which is essentially
a one-dimensional system; an axisymmetric finite
elementhnfinite element (FE-E) mesh of the
surrounding soils can use larger elements and
longer time increments (subject to equivalent
constraints for soil properties).
The hammer impact model used here is that
presented by Deeks & Randolph (1993). The
496
conceptual modei includcs lumped rani an3 anvil
mcwes separated by a damped cushion which is
represented by a spring and dashpot in parallel.
These rest on the pile which is modelkd by a
dashpot, see Figure 1.
Figure 1. Hammer impact model.
in this stage it is sufficient to represent the pile by a
dashpot of impedance equal to that of the pile,
defined as
=-E,, .A,, (1)
P
where E, is the elastic modulus of the pile, A, is its
cross sectional area, and cp is the axial pile wave
velocity.
Deeks & Randolph showed that the behaviour of
the hammer impact model is governed by three
dimensionless parameters: the cushion stiffness,
kc", the anvil mass, ma", and the cushion damping,
cc", defined as follows:
* k , .m,
k , =?
z
(3)
c1 =c( (4)
c z
where k,is the stiffness of the cushion spring, n1,. is
the mass of the hammer ram, m, is the anvil mass
and cc is the damping in the cushion.
The second stage of the simulation is the
propagation of compression waves down the shaft
of the pile. This is done using a model, based upon
that developed by Deeks (1992), shown in Figure 2.
The pile is modelled by 8-noded axisymmetric
finite elements, the toe response is modelled by a
mass-spring-dashpot model proposed by Deeks &
Randolph (199.51, and the shaft/soil boundary is a
spring-dashpot system similar to that developed by
Decks (19921, for transmitting shear waves
independent of frequency.
The model for the toe resistance was developed by
Deeks from that proposed by Wolf (1988). The
complex stiffness of the Wolf model was matched
to the results of FE analyses for several values of
Poisson’s ratio, v, taken across a large range of
dimensionless frequencies.

The model parameters can be conveniently

expressed in dimensionless terms. The frequency of
the applied load, 0,is non-dimensionalised with
respect to the radius of the pile base R, and the
shear wave velocity of the soil, cs.
&.R
a,s = -
C,S
where c, = d(G/p).

The static spring stiffness, k, is taken as

k = -4.GR (6)
1-v

The non-dimensionalised mass and damping

parameters %, al, POand P 1 are defined as:
, \7 / \2
Figure 2. Pile model for axial waves and shear
transfer.

The third and final stage of the procedure is to

Further, Deeks proposed the following expressions
to accommodate variations in Poisson’s ratio, v, as:
a, = 0 . 6 3 - 3 . 6 ~ + 6 ~ ’ , (94
/j, = 1 . 5 8 - 1 0 . 3 ~ + 1 9 ~ ~ (9b)
The response of the pile shaft to the impact wave is
modelled using a new frequency independent
transmitting boundary for axisymmetric shear
waves derived by Deeks (1992). This boundary is
equivalent to viscous dashpots with a distributed
damping constant of p.cs (identical to a viscous
boundary) and a distributed spring constant of
G/2rb, where p is the density of the soil, cs is the
shear wave velocity in the soil, G is the shear
stiffness of the soil and rb is the pile radius.
The velocity-time function derived from the
hammer impact model is then imposed onto the
simplified pile model described above as a force on
the pile head (after multiplication by the pile
impedance). The output results are the
displacement-time functions at the shaft interface
nodes, and at the base.
impose the displacement-time functions onto a large
FE/IE axisymmetric mesh of the surrounding soils,
see Figure 3. The pile response is transferred to the
ground model by way of an unrestrained “false”
pile made up of axisymnietric finite elements.
The pile-soil interface is modelled using a surface-
based contact simulation, with slip controlled by a
Coulomb friction model with p=O.1 (Mabsout &
Tassoulas, 1994). In order to simulate the horizontal
stresses on the pile from the soil, the “false” pile is
moved laterally into the soil by a predetermined
distance. (This avoids the problem of rigid body
motion.) Once the horizontal stresses have
equilibrated, the vertical displacements computed
from the pile model are applied to the pile shaft
nodes of the “false” pile and the soil nodes
immediately under the pile toe in a dynamic
analysis with time steps of 0.001secs.
For simplicity, the soil is modelled as an elastic
medium as it has been shown that at a short distance
from the pile (about one pile radius) most of the
energy is propagated in the form of elastic waves
(Massarsch 1992). Parametric studies based on the
arrival times of ground waves at various distances
from the vibration source indicate that the dynamic

497
soil stiffness is much greater than the static
stiffness. The non-linear behaviour of soils with
very high stiffness at very small strains has often
been observed (Matthews et al, 1996). The infinite
elements around the periphery of the FE mesh are
required to avoid spurious wave reflections.

3. IMPACT PILE DRIVING AT THE M66

A valuable example of site data which include both

pile head records and ground surface vibrations was
obtained during dynamic testing of cast in situ
piling for bridge foundations on the M66 motorway
near Manchester.

Figure 3. Axisymmetric finite elemenvinfinite

element mesh of the surrounding ground.

1.0
[ A 0 --3 1
0.3

0.0

9 -0.5

d
i -1.0

Depth Soil type E V P 4

P
-1.9

(m) (Pa) (kgh') -2.0

0 - 10.5 Firdstiff 13 x 106 0.35 1970 -2.5

CLAY
-3.0
I I I I I I I I I
10.5>> Dense 2 4 x 106 0.30 2100 0. 5. 10. 15. 20. 25. 30. 35. 40. 03 90.

I
SAND TOTAL T I I E ( n ) [X l O -3 ]

A smalI'strain stifffiess oi-150-x lU^-%a was usedto Figure 4. Computed and measured (dashed line)
model the outgoing waves. displacement-time functions at the pile head.

In considering the impact of the hammer onto the

pile head, a parametric study was conducted to [do =]
0.4 I I 1 I

match values of k,*, m,* and cc*. The values which -----___-____-__
0.0---
gave the closest match for pile head response were
kc* = 0.96, ma* = 0.90 and cc* = 0.10. The normal 2 -0.4 -
ranges for these parameters are kc*= 1- 1000,
m,*=0.1-0.7 and cc* = 0.0-1.0. The values for the 6
2 I
-0.8
- I
spring stiffness and anvil mass are slightly outside i3 I
these ranges in this case due to the special pile cap ge -1.2 -
II
I
which is required for a SIMBAT analysis. 8
11
IJ
-1.6 - I'

I
The computed displacements and forces at the pile t , 1 I I
head show good agreement with the field data as -2.0
0.00 0 .oz 0.04 0 .OS 0.08 0.10

shown in Figures 4 and 5.

The force-time function computed from the hammer

impact model was then imposed onto the pile head

498
for the second stage of the computational procedure
in order to determine the displacement-time
functions on the pile shaft.
In the third stage, the generated displacement-time
functions were imposed onto the shaft nodes of the
“false” pile and onto the soil nodes under the pile
toe in the FE/IE mesh. The resulting radial and
vertical ppv’s at horizontal stand-off distances of
5.5m, 10m and 16.5m are compared with the
measured vibrations in Figures 6 and 7 respectively.
Figure 6. Radial ppv’s at 5Sm, 10m and 16.5m
respectively. (Measured = dashed line)
499
Taken overall, and recognising the ground
variability, the correlations between the equivalent
velocity-time traces are generally encouraging. The
main discrepancy appears to be due to a P-wave
from the pile shaft which may be caused by an
eccentric strike. This is not simulated in the model.
4 SIMULATION OF VIBRO-DRIVING OF PILES
The process of pile installation by vibratory driver
is fundamentally different from impact driving, and
Figure 7. Vertical ppv’s at 5Sm, 10m and 16.5rn
respectively. (Measured = dashed line).
so modelling of the process must also be different
from the above procedure for estimating vibrations
from impact hammers. In practice, the method is
best suited to saturated fine sands, and liquefaction
is induced by the vibration of the pile, which then
sinks into the ground under the combined weight of
driver plus pile.
Many site records have shown that ground
vibrations resulting from vibro-driving are
continuous and sinusoidal in character, in marked
contrast to the transient nature of vibrations caused
by impact piling.
The technique for modelling pile response to cyclic
excitation, and outgoing ground waves, is now
proposed as a two-stage procedure.
In stage one, the objective is to establish a model
for rigid body vertical oscillation of the pile in
response to the cyclic excitation of the vibro-driver.
This is done by the use of rigid axisymmetric
elements for the pile shaft, a limited axisymmetric
F E / E mesh representing the soil, and a mechanical
model for toe reaction based on a spring and a
dashpot in parallel, proposed by Lysmer & Richart
(1966), see Figure 8. The pile-soil interface
comprises a two-surface contact, with a Coulomb
friction model. A static computation is made for the
soil/shaft normal stresses due to geostatic stress. A
dynamic analysis is then conducted to set up steady
state response of the rigid pile to the cyclic
excitation from the hammer of
F(t) = m.e.w2. (10)
where m is the total mass at eccentricity e, rotating
at w rad/s.
Figure 8. Simplified model for vibro-driving
simulation.
500
The spring and dashpot constants are given by
k = -4.GR (1 1)
I-v
kR
~=0.85-
The time-stepping computation generally stabilises
into steady state after 1.0-1.5s with time increments
of 0.001s.
In the second stage, a €%/E mesh is established,
similar to that in Figure 3, and the sinusoidal
displacements of the shaft nodes and the base,
computed in stage one, are imposed in the form of
Fourier series. It is essential in an on-going analysis
that there is no spurious reflection from an artificial
boundary, and this is achieved by the infinite
elements around the periphery of the model.
Material damping is applied in the form of a
Rayleigh damping ratio given by
where U,, =m = undamped frequency of
vibration. A typical damping ratio of 5% has been
suggested by Massarsch (1992) for the elastic range
of soil deformations.
The output from the computation gives ground
surface vibrations as sinusoidal vertical and
horizontal particle velocities. These can then be
calibrated against site data.
5. VIBRODRIVING AT THE SECOND SEVERN
CROSSING
A detailed parametric study of vibro-driving
simulation has been undertaken using data obtained
from the installation of 1050mm diameter casings
for one of the approach bridges to the second
Severn crossing. Borehole records for the site
indicate soft to firm clay to a depth of 13.8m
overlying firm to stiff marl. The casings were
installed to a depth of 15.5m using a PTC 50H3
vibratory hammer with an eccentric moment of
50m.kg. The resulting vibrations were monitored by
geophones placed on the ground surface at distances
of 5.7m, 14.5m and 32.9m.
The two stage procedure was followed and
parameters were adapted to produce a match
between estimated surface particle velocities and
Figure 9. Radial ppv’s at 5.7m, 14.5m and 32.9m
respectively. (Measured = dashed line)
the measured site data. The influence of the various
parameters were found to be as follows:
The distribution of shaft friction with depth
affects the proportions of the ppv’s with relation
to each other at various stand-off distances. This
may be due to the interaction of the ground
waves propagated from the shaft and the toe as
observed by Attewell et al, 1991.
The magnitude of the dashpot constant affects
the magnitude of the ppv’s
Doubling the spring stiffness only causes a
small reduction in the magnitude of the ppv’s
501
Figure 10. Vertical ppv’s at 5.7m, 13.5m and 32.9m
respectively. (Measured = dashed line)
Increasing the material damping from 5% to 8%
only causes a small reduction in the ppv’s
The computed ppv’s at the ground surface for a
casing installed to a depth of 1l m at a frequency of
15.8 Hertz and a penetration speed of about 0.18mds
are shown in Figures 9 and 10. These computations
assumed a small strain stiffness of 50 x 106 Pa for
the soft/firm clay with v = 0.4 and p = 1750 kg/m3.
The computed ppv’s show good agreement with
both the radial and vertical ppv’s measured on site
over a considerable stand-off distance.
6. CONCLUSIONS
Simplified procedures have been developed for the
computation of ground surface vibrations caused by
impact driving and vibratory driving of pre-formed
piles.
The procedures have been shown to be viable by
comparisons with site measurements at two
locations, provided that appropriate values for
certain parameters can be identified. Further pile
driving simulations are currently being undertaken
for a large number of sites in the U.K. (Uromeihy,
1990) in order to refine the various parameters, and
ultimately, to develop general guidance on the most
appropriate parameters for each set of
circumstances.
The procedures do not require detailed knowledge
of site conditions and are therefore particularly
useful as a preliminary design tool.
The methods are now established for a further phase
of the work, which will be to add typical building
forms to the finitdinfinite element meshes, for
estimation of building response with full soil-
structure interaction.
ACKNOWLEDGEMENTS.
The funding of this work by EPSRC through Grant
No. GNLl8679 is gratefully acknowledged.
REFERENCES.
Attewell, P.B. & I.W. Farmer 1973. Attenuation of
ground vibrations from pile driving. Ground
Engineering. 6.4:26-29.
Attewell, P.B., A.R. Selby & A. Uromeihy. 1991
Non-monotonical decay of ground surface
vibrations caused by pile driving. Earthquake,
Blast and Impact ed SECED. London
Elsevier:463-48 1
Deeks, A.J. 1992. Numerical analysis of pile
driving dynamics. PhD thesis. Uni. of W.
Australia
Deeks, A.J. & M.F. Randolph 1993. Analytical
modelling of hammer impact for pile driving.
Int. J. f o r Num. and Analytical Methods in
Geomechanics. 17:279-302
Deeks, A.J. & M.F. Randolph 1995. A simple
model for inelastic footing response to transient
502
loading. Int. J. for Num. arzd Analytical
Methods in Geomechanics. 19:307-329
Eurocode 3 1996 Design of steel structures, Pt 5
Piling CENflC 250/SC3/PT5
Head, J.M. & F.M. Jardine 1992. Grouizd-borne
vibrations arising from piling. TN 142
London: CIRIA
Lysmer, J. & F.E.Richart 1966. Dynamic response
offootings to vertical loading. J. of the S.M &F
Div. Proc. A.S.C.E 92; SM1 65-91
Mabsout, M.E. & J.L. Tassoulas 1994. A finite
element model for the simulation of pile
driving. Int. J. for Nunz. Methods in Eizg.
37:257-278
Massarsch, K.R. 1992. Static and dynamic soil
displacements caused by pile driving.
Application of Stress- Wave Theory to Piles, ed
Barends, F.B.J. Balkema Rotterdam : 15-24
Matthews, M.C., V.S. Hope & C.R.I. Clayton 1996.
The use of surface waves in the determination
of ground stiffness profiles. Proc. Instrz. Civ.
Erzgrs Geotech. Eizgng, 1 19:84-95.
Ramshaw, C.L., A.R. Selby & P. Bettess 1998.
Computation of the transmission of waves from
pile driving. Ground dynamics and ninn made
processes. ed Skipp, B. 0. T.Telford Publ.
London : 1 15- 128.
Selby A.R. 1991 Ground vibrations caused by pile
installation Proc of 4th Int Conf on Piling arzd
Deep Foundations. Stresa DFUTespa :497-502
Stain, R.T. 1992. SIMBAT - a dynamic load test
for bored piles. Piling: European practice and
worldwide trends. T Telford London: 198-205
Todd, A. 1994. Berwick river walls - an unique
scheme to safeguard an historic part of Berwick
upon Tweed. ICE papers cornpetition. London
Uromeihy, A. 1990. Ground vibration
measurements with special reference to pile
driving. PhD thesis. Uni. of Durham. UK
Wolf, J.P. 1988. Soil-Structure-Interaction Analysis
in Time Domain. Prentice-Hall, New Jersey.
Wiss J.F. 1967. Damage effects of pile driving
vibrations. Highways Res B. 155 : 14-20
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

The effect of pile impedance on energy transfer to pile and ground vibrations

M.R.Svinkin
VibraConsult, Cleveland, Ohio, USA
B.C. Roth
Munico Consultants, San Francisco, Cali$, USA
W. R. Hannen
WJE,Northbrook, Ill., USA

ABSTRACT: Pile impedance effects energy transfer to the pile and surrounding soil in different ways.
Increasing pile impedance increases force and decreases velocity at the pile head and might insignificantly
affect energy transferred to the pile. Displacement piles affect predominantly ground vibrations in the
proximity of pile driving. Effects of displacement piles are more substantial on amplification of ground
vibrations than those from piles with pile impedance reduction.

1 INTRODUCTION and pile driveability on induced ground vibrations.

Pile driving operations are powerful sources of

construction vibrations which may adversely affect
nearby buildings, harmfully affect sensitive
electronics and disturb people. Therefore, certain
concerns frequently arise about possible vibrations
which might be generated from pile driving
activities, especially impact pile driving.
The effect of pile driving on ground vibrations
depends on the energy transferred to a pile, pile
type, soil conditions and the distance from the
source.
Pile driving vibrations can be a problem with all
kinds of pile drivers and all types of driven piles
(Woods 1997). However, installation of various
piles at a construction site generates different ground
vibrations at the same distance from driven piles.
Substitution of certain pile types for others can be
used as partial mitigation of potentially high levels
of ground vibrations.
The effect of pile type on induced ground
vibrations is at least in part attributable to pile
impedance. Considerable data have been collected,
analyzed and published with respect to the pile
impedance effect on force-energy transfer to piles
and ground vibrations generated during pile driving,
e.g. Parola (1970), Peck et al. (1974), Heckman &
Hagerty (1978), Massarsch (1992), Abe & Thendean
2 FORCE TRANSFER VERSUS PILE
IMPEDANCE
Pile impedance is a significant factor in the transfer
of dynamic longitudinal force into the pile and from
the pile into the surrounding soil. The reason for
this is because, independently of how much energy
may be applied to the pile head, the force that can
be transmitted down the pile is limited by the
impedance (Peck et al. 1974). Pile impedance
characterizes the pile ability to overcome the soil
resistance to pile penetration and develop required
capacity.
Pile impedance, Z, is defined as
Z = EA/c (1)
where E = modulus of elasticity of pile material; A
= pile cross-sectional area; c = longitudinal stress
wave velocity.
The maximum force, F, measured at the pile head
during driving can be expressed as (Svinkin 1994)
F =
1: 2-W, (2)

(1 996), Woods (1997), Svinkin (1999), Brettmann & where L = pile length; W, = energy transferred to
Cotton (1999) and others. the pile; remaining parameters are the same as
This paper presents further consideration of energy above,
transfer to piles, assessment of the effect of pile type According to equation (2), force transmitted to the

503
Figure 1. Effect of hammer (a) and pile (b) cushion stiffness on Figure 2. Effect of steel pile (a) and concrete pile (b) impedance on
ETR (after Abe & Thendean 1996) ETR and refusal capacity (after Abe & Thendean 1996)

504
pile is proportional to the square root of pile Table 1. Energy transferred to concrete and steel piles
impedance and also depends on pile length, velocity
Pile Hammer Energy transfer efficiency
of stress wave propagation in the pile and energy
transferred to the pile. Type Type Average Standard ETR
Wave equation analysis simulation (Abe and ETR Deviation Ratio
% %
Thendean 1996) confirmed that the mobilized pile
capacity increases with increasing pile impedance Concrete Air 40.4 12.0 0.79
(Figure 2).
Hvdraulic 5 1.2 15.9 1.oo

Diesel 24.8 6.8 0.48

3 VELOCITY TRANSFER VERSUS PILE
IMPEDANCE
Svinkin (1996) derived an equation to calculate the
peak particle velocity (PPV) at the pile head, V, in
advance of pile driving
v =.
J 2-w,
z"L
(3)
where all parameters are the same as above.
Equation (3) is similar to equation (2), but the
peak particle velocity of the source is inversely
proportional to the square root of the pile
impedance.
4 ENERGY TRANSFER VERSUS PILE
IMPEDANCE
The energy transferred to the pile head is computed
by integrating the product of force and velocity
records (Hannigan 1990). According to equations
(2) and (3), pile impedance effects force and velocity
in opposite ways. Obviously, it can be expected that
pile impedance has negligible impact on the energy
transferred to the pile.
The performance of a pile driving system or
hammer performance, is important to the proper
execution of pile driving. The measure of hammer
performance obtained from dynamic pile monitoring
is the ratio of the transferred energy to the rated
hammer energy. This ratio is known as the energy
transfer efficiency, ETR. Intensive study of hammer
performance was made by Abe & Thendean (1996).
They summarized statistical results of ETR compiled
from Pile Driver Analyzer outcomes at numerous
projects and provided wave equation parameter
studies.
Six combinations were analyzed for single acting
air, hydraulic, and diesel hammers on steel and
concrete piles. ETR mean and standard deviation
for considered six hammer-pile systems are shown
in Table 1. It can be seen that average ETR is 20-
Steel Air 50.2 11.5 0.61
Hvdraulic 82.0 15.1 1.oo
Diesel 34.3 34.3 0.42
34% lower for concrete piles than those for steel
piles, but ETR ratios between different hammers for
concrete piles are similar to those for steel piles.
This is an indicator of almost adequate hammer
conditions for driving steel and concrete piles but
with lower transferred energy to concrete piles on
account of low pile cushion stiffness.
The obtained effects of hammer and pile cushion
stiffness and pile impedance on ETR are displayed
in Figures 1 and 2. For steel and concrete piles, the
hammer cushion stiffness does not affect ETR in the
commonly used stiffness range above 1,000 kN/mm
(Figure la). Therefore, Figure 2a demonstrates the
reasonable effect of steel pile impedance on ETR
without implication of hammer cushion influence.
For steel piles in the commonly used impedance
range of 100-3000 kNs/m, pile impedance affects
negligibly on ETR for the air and hydraulic
hammers but decreases ETR about 30 % for the
diesel hammers in the impedance range of 100-
1000 kNs/m.
The pile cushion stiffness significantly effects ETR
for all hammers at lower stiffness from 100 to
1000 kN/mm (Figure lb). Probably a share of pile
cushion stiffness influence was not completely
excluded from the effect of concrete pile impedance
on ETR in Figure 2b where considerable ETR
variation is observed at the pile impedance range of
300-3000 kNs/m. In this impedance range, ETR
decreased 20, 30, and 50 % for air, hydraulic, and
diesel hammers, respectively. These results are
different from those obtained for steel piles though
similar results for both pile types in the same pile
impedance range could be expected. It seems that
the effect of concrete pile impedance on ETR ought
to refine from pile cushion influence with
normalization of pile cushion stiffness.
Thus, steel pile impedance does not actually affect

505
ETR for air and hydraulic hammers and results in Table 2. Data for tested piles
30 % reduction of ETR for diesel hammers. The Designation Description Impedance Toe Area Depth
effect of concrete pile impedance on ETR is not kNs/m crn2 rn
clear so far.
TP-1 HP 362~1.71 910 222 13.41
rnrn, kN/m
~~

TP-2 HP 362~1.71 910 222 14.94

5 GROUND VIBRATIONS VERSUS PILE TYPES m, kN/m

TP-3 HP362~1.71 910 222 15.54

The effects of some factors on ground vibrations are mrn. Id\l/rn
obvious. For example, increasing the energy
TP-4 356 x 13 mm 561 994 7.01
transferred to the pile increases ground vibration Pipe Pile Shell
intensities, and increasing the distance from a driven w/conical
DOint
pile results in reduction of ground vibration
intensities. Influence of other factors on ground TP-5 356 x 13 I I
II
I
I 561 994 7.62
Pipe Pile Shell
vibrations may not be so apparent. w/conical
It is valuable to know how a change of pile type point
affects ground vibrations during pile installation. TP-6 406 x 406 mrn 1618 1652 7.01
Pile impedance should be considered as a factor Precast
affecting the intensity of ground vibrations induced Prestressed
Concrete
by pile driving.
Heckman and Hagerty (1978) and Massarsch TP-7 406 x406 rnrn 1618 1652 6.01
Precast
(1992) pointed out the important effect of the pile Prestressed
impedance on the peak ground velocity and ~
Concrete
developed a graph for the amplification factor of TP-8 365 x 4.5 mm 204 -500 5.18
ground velocity as a function of pile impedance. Fluted Pile
Shell
The lower the pile impedance, the higher the
induced ground vibrations. By way of illustration,
they showed that a reduction of the pile impedance
from 2000 to 500 kNs/m could increase the peak
ground velocity by a factor of 8. According to
equation ( 3 ) , for the referenced impedance range,
the expected amplification of the peak pile velocity
and the peak ground velocity can only be 2.
In the following case histories velocity 01ground
vibrations were measured from driven piles with
different impedances and results obtained were used
for assessment of the role of pile impedance in
amplification of ground vibrations.

5.1 Case 1

Eight different piles were tested within an Advance

Test Pile Program performed at a site in Phoenix,
Arizona, in connection with the design and
construction of the 1-10 West Papago/Inner Loop
Freeway (Hannen & Linehan 1984). A pile
description is presented in Table 2. A pile and
geophone layout is depicted in Figure 3. The soil
consisted of about 1.4 m of clayey sand followed by
about 2.4 m of sandy clay followed by about 5.6 m
of dense clayey sand underlain by about 2.8 m of
sandy clay deposited on 12.2 m of sand, gravel and
cobble. The water table was not encountered during
Figure 3. Layout of TP-1 - TP-8 piles and geophones
soil boring at the site. A MKT DE-70B diesel
hammer was employed for initial driving and
restrikes. A hammer stroke changed between 1.5
and 2.6 m. The measured PPV of ground vibrations
were adjusted to the maximum hammer stroke.
Ground vibrations were measured at distances of
9.1, 30.5, 152.4 m from group 1 of TP-1, TP-3,
TP-5, TP-7 piles and at distances of 7.0, 28.3,

506
Figure 4. Maximum PPV of ground vibrations at distances
of 9.1, 30.5 and 152.4 m from TP-1, TP-3, TP-5, and TP-7
driven piles
150.2 m from group 2 of TP-2, TP-4, TP-6, TP-8
piles. Maximum PPV of radial and vertical ground
vibrations are depicted in Figures 4 and 5 . Radial
ground vibrations had higher intensity in comparison
with other components of ground vibrations.
Transverse horizontal ground vibrations were not
considered in the study because their level was 3-5
times smaller than level of radial horizontal ground
vibrations.
Four different pile types were tested. H-piles
were non-displacement piles and the rest of piles
were displacement ones.
The analysis of soil vibration records, measured at
the same distances from each group of piles with
different impedances and driven by the same
hammer, revealed some interesting observations.
e In general, piles being driven as displacement
piles generated the greatest ground vibrations, but
there effects were varied. In pile group 1, the effect
of concrete TP-7 pile was dominant for radial
component at distances of 30.5 and 152.4 m and for
vertical component at all distances from the source.
Figure 5 . Maximum PPV of ground vibrations at distances
of 7.0, 28.3 and 150.2 m from TP-2, TP-4, TP-6 and TP-8
driven piles
In pile group 2, the effects of TP-4 pipe pile shell
and TP-8 fluted pile shell were major for radial and
vertical components at all distances from driven
piles.
@ Driving non-displacement piles induced lower
ground vibrations, but the effects of TP-1 and TP-3
H-piles were dominant for radial and vertical
components at distance of 9.1 and 30.5 m from
driven piles at the pile penetration length between
6.4-7.3 m with the pile toe in dense clayey sand.
@ Effects of displacement piles were more
substantial on amplification of ground vibrations
than those from piles with impedance reduction.
Ground vibrations increased due to pile
penetration depth of 6.4-7.3 m and then started to
decrease. This effect is related to soil conditions
rather than to pile length.
@ Restrikes were made for TP-2 - TP-6 piles.
Restrikes did not actually increase the level of
ground vibrations.
Figure 6 . Maximum PPV of ground vibrations at various
distances from driven precast pile and H-pile
5.2 Case 2
Pile driving of four piles and measurements of
ground vibrations were made at a site in San
Francisco, California, to determine the various
seismic effect of installation of two pile types on
existing buildings (Municon Conschants 1999). The
test piles included two 305x305 mrn precast concrete
piles uith lzngth of 18.3 111, Z=910 kNs/m, and two
HP 3 0 8 ~ 1 . 0 8(mm, kN/m) with length of 21.0 m,
2=577 kNs/m. The soil consisted of about 11.3 m
of loose to dense sand followed by about 1.5 m of
sandy clay underlain by a bearing layer of dense to
very dense silty sand. The water table was at the
depth of 6.1 m. For all piles tested, a predrill was
made to the depth of 4.6 m. A Delmag D30-32
diesel hammer was used for pile driving. The
hammer stroke was 3.2 m. The soil resistance to
pile penetration was mostly in the range of 5-
30 bU0.3 m except the end of driving for concrete
piles with the soil resistance between 60-
100 bY0.3 m.
Ground vibrations were measured at various
distances from driven piles. The maximum PPV of
ground vibrations measured during driving of two
pile types are shown in Figure 6. Installation of
displacement concrete piles generated larger ground
vibrations at distances up to 20 m from the piles
than those induced from driving H-piles. A PPV
ratio of compared ground vibrations was three at a
distance of 3 m from driven piles. Even though the
impedance of concrete piles was 1.6 times the
impedance of H-piles, the seismic effect of
displacement pile installation was more significant in
the proximity of pile driving.
508
Figure 7. Maximum PPV of ground vibrations at disranxs
of 7.0 and 9.1 rn versus soil resistance to pile driving
6 GROUND VIBRATIONS VERSUS SOIL
RESISTANCE TO PILE PENETRATION
The soil resistance to pile penetration, measured in
blow count per 0.3 ni, changes during pile driving.
The higher blow count corresponds with the greater
soil resistance. The blow count greater than 100
b110.3 m is considered as high blow count.
For various penetration resistance, maximum PPV
of horizontal ground vibrations at distances of 7.0
and 9.1 m from driven piles are shown in Figure 7
(Case 1). It can be seen that the high blow count
resulted in increased ground vibrations at the pile
penetration depth between 4.27 and 7.62 m but did
not affect ground vibrations at the greater
penetration depth. Analogous results were observed
at distances of 28.3 and 30.5 m from driven piles.
Holfoway et al. (1980) reported similar phenomena
that driving to refusal of concrete, steel shell and
H piles at depth about '7.0-7.6 m increased radial
and vertical particle velocities by factors 1.2-3.9 ar
distances of less than 10 m from driven piles.
7 DISCUSSION OF WSULTS
7.1 Pile Impedance
The greater the pile impedance, the greater the pile
capacity and the greater the dynamic force that can
be transferred to the ground.
Pile impedance affects the intensity of ground
vibrations in two ways at the same time. On the one
hand, according to equation (2), increasing pile
impedance increases the force transmitted to the pile
and suimxmding ground, but on the other hand.
according t ~equation
! (31, increasing pile impedance
decreases PPV of pile and ground vibrations. An
incress of h m m e r energj- magnifies ground
vibrations until the pile impedance allows to increase
the force transmitted to the pile and the surrounding
soil. The impedance affects in opposite ways in
both equations, and the pile impedance effect on the
intensity of ground vibrations is not o b l h s .
Presented case histories revealed that pile
impedance cannot he always used as a predictor of
intensit? of ground vibrations. In Case 1, precasc
concrete piles and pipe pile shells ~ i t closed
h ends
generated similar ground T ibrduons though the latter
had pile impedance 2.9 rimes less. In Case 2, che
dri\.ing of precdst concrete piles induced PP\’s of
ground \ lbrarions IZCX rhe prics three times more
tlian the driving oc the H-piles eke11chough the latter
had pile impedance 1.6 times smaller.
7 . 2 Displacernmr piles
Piles being driven as displacement piles transfer
p a t e r force to the surrounding ground. The more
pronounced effects of displacement piles on ground
\ ibrations occurs predominantly at distances less
t h m 10 m from pile driving.
7 . 3 Pile-soil load transfer
The dynamic force transferred to the pile at each
hammer blow is then transferred to surrounding
ground through the pile toe as a concentrated load
and through the pile shaft as a non-uniformly
distributed load.
During driving. the various soil resistances to pile
penetration are developing as the pile penetration
depth is increasing. Generally speaking, dynamic
loads transferred from the pile shaft and toe to the
ground should be increasing with increasing pile
penetration depth. However, this does not always
occur. In Case 1. the intensity of ground vibrations
was independent of the pile penetration depth and
depended on soil properties. Maximum PPV of
ground vibrations were recorded during driving all
piles through dense clayey sand at depth 6.4 to
7.6 ni.
Pile-soil load transfer is realized by means of both
concentrated loads from the pile toe and distributed
loads generated along the pile shaft. However, for
many practical purposes, a pile can be deemed as a
509
point source of vibrations. This assumption might
be correct at dismce. D, from driven pile (Svinkin
1996)
Where c, = velocity of surface wave propagation in
the ground; remaining parameters are the same as
above.
8 CONCLUSIONS
File impedance affects force and velocity at the pile
head in opposite ways. Therefore it can be expected
that pile impedance does not significantly affect
energy transferred to the pile. Analysis of published
experimental data and wave equation simulation
cunfiims that the impedance of steel piles has a
negligible effect on ETR for air and hydraulic
hammers. Additional research IS needed to clarify
the impedance effect of concrete piies on ETR for
all hammer modek.
An increase of hammer energy magnifies ground
vibrations until the pile impedance allows the force
transmitted to the pile and the surrounding sod to
increase. Since the impedance affects force and
velocity in opposite ways at the same time, the p i e
impedance effect on the intensity of grourici
vibrations is not obvious.
The more pronounced effects of displacement piles
on ground vibrations occurs predominantly at
distances less than 10 m from pile driving.
In many situations, a driven pile. can be
considered as a point source of vibrations.
ACKNOWLEDGEMENT
The writers are thankful to Mr. Anatol Longinow,
Project Manager at Wiss, Janney, Elstner
Associates, Inc. for valuable comments.
REFERENCES
Abe, S. & G. Thendean 1996. Hammer performance
evaluation. In F. Townsend, M.Hussein &
M. McVay (eds), Proceedings of the Fifth
International Conference on the Application of
Stress Wave Theory to Piles, Orlando, Florida:
9 12-927.
Brettmann,T. & B. Cotton 1999. Case history: Woods R.D. 1997. Dynamic efsects of pile
estimating ground vibrations caused by pile installation on adjacent structures. Synthesis
driving. DFI 24th Annual Members ’ Conference, Report, National Cooperative Highway Research
Decades of Technology - Advancing into the Program NCHRP Synthesis 253, Washington,
Future: 161-171. D. C. : National Academy Press.
Hannigan, P. J. 1990. Dynamic monitoring and
analysis of pile foundation installations. Sparta,
New Jersey: DFI.
Hannen, W.R. & P.W. Linehan 1984. Vibration
monitoring during the I-10 West Papagohner
Loop Advance Test Pile Program. WJE Report
830535, WJE Engineers, Northbrook, Illinois.
Heckman, W.S. & D.J. Hagerty 1978. Vibrations
associated with pile driving. ASCE Journal of the
Construction Division, 104(C04): 385-394.
Holloway, D.M., Y. Moriwaki, E. Demsly,
B.H. Moore, & J.Y. Perez 1980. Field study of
pile driving efsects on nearby structures, Special
Technical Publication, Minimizing Detrimental
Construction Vibrations, Preprint 80-175: 63-100.
New York: ASCE.
Massarsch, K.R. 1992. Keynote lecture: Static and
dynamic soil displacements caused by pile driving.
In F. Barends (ed.), Proceedings of the Fourth
International Conference on the Application of
Stress- Wave Theory to Piles: 15-24, Rotterdam:
Balkema .
Municon Consultants 1999. Pile driving vibration
energy attenuation survey - Reports, San
Francisco, California.
Peck, R.B., W.E. Hanson & T.H. Thornburn.
1974. Foundation Engineering, 2nd ed., New
York: John Wiley & Sons.
Parola, J. F. 1970, Mechanics of impact pile driving,
Ph.D. thesis, University of Illinois, Urbana,
Illinois.
Svinkin, M.R. 1994. Influence of pile parameters on
pile driveability. Proceedings of International
Conference on Design and Construction of Deep
Foundations, Orlando, Florida,II: 1150-1164:
FHWA.
Svinkin, M.R. 1996. Velocity-impedance-energy
relationships for driven piles. In F. Townsend,
M .Hussein & M. McVay (eds), Proceedings of the
Fifth International Conference on the Application
of Stress Wave Theory to Piles, Orlando, Florida:
870-890.
Svinkin, M.R. 1999. Prediction and calculation of
construction vibrations. DFI 24th Annual
Members’ Conference, Decades of Technology -
Advancing into the Future: 53-69.

510
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

etemination of modulus of subgrade reaction in a pile with a vibrating

apparatus
M.Hilmi Acar
Department of Civil Engineering, University of S e l p k , Konya, Turkey

ABSTRACT: The pile foundations are mostly under the influence of the horizontal forces due to the earth-
quake, wind and wave forces. The soil properties is considerably different in this loaded medium. It is neces-
sary to know the value of modulus of subgrade reaction which has an important place in the designing and
calculation of the pile foundations. This study is aiming to present by the numerical model the suitability of
the value of modulus of subgrade reaction, determined experimentally through a vibrating apparatus placed on
a free headed concrete filled steel pile.

1 INTRODUCTION
The pile foundation inay be under a horizontal load
as well as under an axial load. The pile problems,
which are under the effect of the horizontal force,
appears frequently at quay structures, petrol produc-
tion plants and marine pilings. The horizontally
loaded pile is closely related with the beam prob-
lems, based on the elastic foundation, and is an ap-
plication to head of pile in the pile-soil system of the
external forces and moments. According to the
beam theory based on the elastic foundation:
In this equation
$4) : The horizontally displacement of the pile
x : The depth in soil
E l : Pile stiffness
12 : Soil reaction.
The soil reaction “p” is a function of the pile
characteristics and (1.1) the solubility of the equa-
tion is depending on the fbnction, chosen for the soil
reaction “f. If we accept that there is a direct rela-
tion between the soil function “f and displacement
of the pile “ w ” , the soil reaction which effect to the
unit length of the pile will be found by the following
equation:
~=-K.w ( 1.2)
In this equation, K represents the ratio of soil re-
action on a point of pile to the displacement at the
same point. If the equations (1.1) and (1.2) are
united:
d‘w
EI 7 + KW 10 K=k.d (1.3)
dx
/F Coefficient of soil reaction, d= Diameter of the
pile, the equation number (1.3) is obtained. The “k”
value is given in the literature approximately for the
soil classes (Bowles 1988). The aim of the study is;
first, to determine in testing manner of the “I? value
through mounting a vibrating apparatus on the head
of the pile, then, through adding the technical terms,
taking into account tlie dynamic effects establish-
ment of a numerical model and its evaluation. The
new equation obtained is:
a‘w a2w
El- + ,U---- + k-&v + Kw = 0 (1.4)
ax‘ at2 at
Theoretically, the pile head flexibility, i.e. tlie
horizontal displacement curves, which are the coun-
terpart for a unit force, are obtained for differcnt val-
ues of coefficient of soil reaction. The “k” value of
the curve is the coefficient of soil reaction, with
which curves experiinentally obtained flexibility
values fits.
2 EXPERIMENTAL STUDY
The experimental study has been carried out one of
the composite pile of the quay at Clierbourg
(France). The composite piles have been formed
through driving of steel tube piles of l n i of diameter
and then filling them with concrete without rein-
forcement. The experiment has been carried out by
casting after 2 months of the vibrating apparatus on a
pile figure 1.

511
 of the system as a whole is provided. Before beginning
of the experiment all the control units x e checked.
Then a sinusoidal fmce is given in horizontal. 1vay to
the head of pile in any choseii frequency iyalue. The
force, which will be provided by the vibrating appa-
ratus is calculated as below:

f= frequency, Hz (2.1)
Figure 1. Fixing of the apparatus on the pile
T n this state by dividing the force to& term, the
effectir-e force is taken into account. The displace-
ment at head of pile, which is the equivalent to this
force is found as a proportion of the speed value (V),
determined by the speed detector to the “o”. By- di-
vision o f this so found value to the applied fori.e is
c,alculatedthe flexibility value of head of pile.

(2.’)

B!. suhsrituting this obtained espcrinientall>-

fleribilitj- value in the equation number 2.3. wliicli
neglects the basic arid dyiamic effects (Briard 1V S ) .
the “k“ value m a y be calculated.
Figure 2. Interior view ofthe apparatus For thc corupleteiy For the p i t i d l y
mbedded piles embedded piles
2. 1 The Vibrating Apparatus and Its Chmaiwristics S=------I S=-
1+2utx
.
,
(2.3)
The vibrating apparatus is fomied of two blocks, ro- 2Ela.’ &?Ela -?
tating in reverse direction in same angular velocity
around a vertical axis, driven by an electrical engine Here: Q = -I is the characteristic length
figure 2. The apparatus is producing a sinusoidnl LE1
horizontal force at (cycle/sec.) frequency hnction.
equal to the rotating velocity. The amplitude of this a: the part of the pile, which is 011 the soil level.
force is proportional to the square of the fiequency. However the aim o f the study is the determination
With an aim to measure the horizontal displacement of the “k” by a numerical model, including also the
at the head of pile, again on the head of the pile a dynamic effects.
speed detector is adhered in a manner, in which its
direction is parallel to the direction of the vibration.
The horizontal displacement of the head of pile is 3 NUMERICAL MODEL
given by the measured speed value (V) with the pro-
portion to the wvalue. The amplitude of the sinusoidal horizontal force
The mass of the apparatus: 50 kg, Diameter: 510 F = F ( u ) s i n ut , which is applied to the head of
mm, height: 200 mm, frequency interval 1 - 14 Hz.
The engine is Mavilor type M-600-6OOW. pile, embedded in an elastic soil, is:
F(w) = h W 2
2.2 Fixation of the vibrating apparatus on the pilc “h” is a constant of the apparatus. The cross sec-
head and measuring of the flexibility (Horizontal tion of the pile is small compared to its length. The
displacement, counterpart of the unit.foi~e)
pile is subjected to a simple bending under this
An aluminum plate of the vibrating apparatus (500 force. The “x” abscise has been accepted from the
mm x 500 mm x 25 mm) is bound by bolts to the base of the pile to its upper part. In the cases when
pile head in absolutely non-moving manner. Then the operational frequency of the vibrating apparatus
the vibrating apparatus is also bound to this plate in “ U ” does not fit with the natural frequency of the
a strong manner by bolts. By this way the behavior

512
 r . = E .I +E I
I a a b'b
The coefficient of soil reaction value is:
k, = d, .K, (Nm-* ) (3-7)
In this formulations
d , ,i = 1,...n The pile diameter in each J , pile space
K , ,i = 1,...n The soil module in each J, pile space
If some part of the pile is up of the soil level,
K, =o, k, = 0, p , = 0 (3.8)

3.2 The limit conditions

The pile end behavior (PO abscise) has an effect on
the ations ions of the examined system. The w(x,t) are
the values of the rotating, bending moment and shear
force, depending on the displacement e(x,t), M(x,t),
T(x,t). At F O abscise, two of these four are always
zero.
On the embedded pile: w(0,t ) = 0,8(0, t ) = 0 V t .

3.3 :Mosement equations

Figure 3. Divided pile spaces It is taken into account that an embedded in the soil
pile is exposed to external load as shown on Figure
3. If the solution is demonstrated by an S vector:
pile-soil system, a horizontal displacement will oc- General casc:
cur. This displacement is :
w(x,t , U ) = W ( X ; w>sin(wt + $) (3.2)
The amplitude of the movement of pile head in awiax=e
the next calculations will be found as bellow:
(3.9)
A ( @ )= W(H;(01 (3.3)

3. I Meckanical characteristics ($the soil and the

pile
Sinusoidal case:
?'he pile is formed of fixed sectioned elzments.
However the soil is a superposition of the homoge- s =' fi~x),G(x),M(x),- T(x))Sinot
nous layers Figure 3. The elements [O,H], situated

1
along of the pile: dZ/dx=8
xo = O l x , <x, <...<x, l...<x, = H (3.4) d g i L i x = a / r , ateachJ~interva1 (3.10)

Each pile, situated in the space J,, and the rne- d ai d X = -T

chanical character of the soil is constant.
d T / d x = (ki- p iw2)w
J , = [x,-,,x,], i = 1,...y1 (3.5) dW
Due to the sufficient energy the term k'- has
at
I, , i = 1,. ..n The length of each space
been neglected.
p,, = I ,...n ,The mass of the pile in the space J,
i Limit conditions: Limit conditions:
r, , i = I , ...n The rigidity of the pile in the syacc -
w(0) = 0
w(O,t) = 01 .
J, , and due to the fact that the pile section is com- Ptle end (3.1 1) Pile end(3.12)
tI(0.t) = 0 j
I

8(0)=0
posite:

513
M (,H . t ) = O
, >
1 By forming of a global transfer matrix the solu-
External loads (3.13) tion of the system is reached through:
T ( H , t ) = Fsinmt]
S ( H ) = T S(o)
M(H) =0
(3.14)
T ( H )= F 4 EVALUATION OF THE CALCULATIONS
The coefficient of subgrade reaction value at J, interval. AND EXPERIMENT CONCLUSIONS

k I( ” j = kl -p,w’ (3.15) The experiment has been done on a partially embed-

The equations number (3.10)-(3.14) has to be
solved.
3.4 Solution oftlie system in the interval [0,H]
The vector “S” is separated to two sub-vectors,
which provides the limit conditions at the pile end.
s’=/(wl,el,M ’ J ) ,
ded in the soil, having fixed sections composite pile
D=l m. The part of the pile, situated in the soil, is
J , , and its situated on the soil part is J , .
The part, situated in the soil:
I , = 8 . 2 5 m , p I = 3 . 1 x I 0 3 k g n - ’ , r , =5x10’ NI,?’
The part, situated on the soil:

S”=’( w 2 , Q 2 , M 2 , - - T 2 ) ,

The pile end limit conditions; embedded pile The “ k , ” value in the calculation has been chosen
-I in a manner, permitting the appearing of the effects
w (0) = O,G‘(O) = O,G’(O)
= l,-?(O) = 0, (3.17) of the characteristics of the soil against the pile
-1
w-(o)= O,%’(O) = O,M2(0)= O,-T2(0) = I, movements.

Each solution of the equations number (3.10)- k~ = z x l o i j 2.5x107~ 3x10‘1 5x10’ N117.2

(3.14) is a linear combination.

Flexibility - frequency curve:
a’s’ i- a’s2 (3.18) s = A / F( F = l )
The a’ and a’ constants will be the solution of
the system. Horizontally displacement-frequency curve:
A = 1000’S 1 4 ~ ’
(3. 8, Speed-frequency curve:

For the “a”values, which do not make “0” the ’ =10003S/4n’

determinant of this system: Acceleration - frequency curve:
r =1 0 0 d / 4 ~ ~
These values are drawn in Figures 4, 5 , 6, 7. The
initial slope in the flexibility curve is horizontal,
however, this slope in the displacement curve is
Through the solution of the equations number high. This state of the flexibility curve is demon-
-
(3.10) (3.14) the amplitude of movement at the strating only that the forces, which are the counter-
head of pile is calculated as: part of the small frequency values of the used vi-
brating apparatus may be neglected. It may be said
-
A = w(H)= F
-- w2
-,
( H ) M I ( N )+ W ’(If)%? ( H ) (3.20) that the speed is constant on the speed curve between
- T - ( H ) M ’ ( H )+ T I ( E . I ) M 2 ( H ) 5-10 Hz, and that this fact is depending to the choseii
mechanical characteristics.
The calculations are made by transfer matrices. At the change of frequency in value of 10 Hz, the
At the first stage at each Jl = [ x ~ - ~ , xspace
~ ] the flexibility, displacement, speed and acceleration val-
z,elementary transfer matrix is set up, ues are changed in counter way according to the cho-
sen coefficient of soil reaction. This fact shows that
S. = 2, .Si-l, i = 1,...n is calculated. how high is the resistance of the soil, so important is
Then, the movement of the pile head.
T = z,?zJI-,

514

Figure 4. Flexibility curve (k=2500;5000 kdaNm-')
Figure 5. Displacement curve of the pile head
(k=25 00 ;5000 kdaNm-')
I I ",I1
Figure 6. Speed curve of the pile head
(k=2000;2500;3000;5000 kdaNm-')
Figure 7. Acceleration curve of the pile head
(k=2000;2500;3000;5000 kdaNm-')
515
FLEXIBILITY
LEGEND
v Experitnental values
Unstable measunngs
- Theoretical vzlues
Figure 8. Comparison of the theoretical and experimental re-
sults.
The flexibility frequency curves, shown at the
figure 8 have been drawn using the
k=1000,2500,5000 IcdaNnz-' (@2) values. The ob-
tained experimentally flexibility values have been
placed on these curves. As it may be seen at the fig-
ure 8, the values 2500 kdaNni-2 obtained as a result
of the experiment, are shown on the curve. As a
consequence the coefficient of soil reaction value of
the soil is 2500 ldaNnz-2(tf / m2)
As a conclusion, the numerical model and the ex-
periment result fit between 1 Hz and 1,5 I-Iz. The
fact that the experiment results between 1,5 Hz and
4,5 Hz are spread is caused due to tlie lack of the
resonance at 3,5 Hz. The hai-tnony between the
model and the experiment result after 5 Hz obligate
to take into account the dynamic effects in the nu-
merical model. Thus it is observed that the iiumeri-
cal model which is out of the resonance space is in
agreement with the experiment results.
5 CONCLUSION
The study is an evaluation of the fitting of tlie coeffi-
cient of soil reaction value, obtained experimentally
by a vibrating apparatus, mounted at a composite
pile head, embedded partially in the soil, with a nu-
merical model. The vibrating apparatus, fixed at the
pile head, is applying during the experiment a meas-
urable siiiusoidal horizontal force to the head of pile.
The horizontal displacement, which is the equal of
this force is measured, too. The horizontal displacc-
inent flexibility values, which are the equal to the
unit force, are easily calculated in place. Using these
values, in the calculations, which do not include the
dynamic effects, the not-sure coefficient of soil re-
action values are obtained.
The aim is to determine the actual coefficient of
soil reaction value, by evaluation through a model,
including the dynamic effects through a more realis-
tic approach of these flexibility values, obtained ex-
perimentally. By this way taking into account the
dynamic effects in the iiuinerical model the flexibil-
ity values are drawn by giving different coefficient
of soil reaction values. The coefficient of soil reac-
tion value of the soil, which is the coefficient of soil
reaction value, used in the curve, drawn in the nu-
merical model, with which curve the flexibility val-
ues fit, obtained in the experiment.
Through a prepared computer program and in the
cases \Then the pile is embedded or articulated and at
all kind of soil terms the results of the experiments,
made by a vibrating apparatus, may be evaluated.
Thus a more realistic coefficient of soil reaction
value is obtained.

REFERENCES

Acar. M.11. 1953. Analyse theorique et experimentale du coin-

porteinent des pieux. These de docteur-ingenieur, Univer-
site Pierre et Mark Curie, Paris.
Baguelin, F. & Jezequel J.F. 1972. Etude expirimentale du
comporrement des pieux sollicites horizontalement, Annu1e.s
I. T.8.1’.P.11‘’ 297, Paris.
Bowles, J.E. I9SS. Foundation Analysis and Design, Mc Graw-
H i 1I, Ne\v-Ynrk.
Rriard, A. 1978. Mesure du coefficieiit de raideur horizontale
d‘un pie11 ancre dans le sol a I’aide d’un excicateur a Ba-
lourd, Contrnt DBTPCICEBTP, Paris.
Chtti, L. & PoLtlos I l.G. 1993. Analysis of pile-soil interaction
iirider lcrci~al iociding using infinite and finite elements,
i 'amp m d Gcorc‘chnics,Oxford. England.

516
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Salkema, Rotterdam, ISBN 905809 150 3

Analysis of crosswise vibration of pile

Chen Renpeng, Zhu Bin & Chen Yunming
Geotechnicul Engineering Institute, Zhejiang University,Hmgzhou, People's Republic of Chirzu

ABSTRAT: A rational mechanical model of foundation of precast concrete pile on layered eiastic base
when driving is established in this paper. Through dynamic anal!isis. the nurneric solurion of forced
reaction of the foundation of single pile which eccentric dynamic \erticill load XIS on pile top is
provided. And effects of excitation intensity. impact duration of pile lmiimer, diamcter or length of side
of pile. eccentricit), of pile hammer, foundation modulus and coefficient of viscosity of soil to the
dynamic response of pile are analyzed. The primary analytic method of studying the pile response LI hich
intense axial excitation and moment excitation act on. which is very important when studying the
vibration of excited precast concrete pile.

pile, and thcq haic not farther discuseJ thc

1 INTRODUCTION dynamic response of tlie pile ~ h c heccenrric
As for the design and the construction of pile, it dynamic vertical load acts on.
is very important to study the dynamic response 'Thc niomcnt of pile h m n i c r ncting on pile
of single pile on the condition of intense top is an intciisi'c7c impact loxi whc11 <ii i\ ing
excitation. Based on the theory that the pile-soil This iinpact load j s subjcctcd 10 deviate becdusc
system is linear. there are four models which are of deflection of (a171 guide and pile top, and the
discrete modcl, continuous medium model, finite allowance deviation of the location of p i ecaqt
element model and boundary element model. pile reaches lOOnim. This eccentric loxl leLrdsto
For example, the research that aiialyzing the
response of pile which axial dynamic load and
crosswise dynamic load respective acting on. and
the pile is supported OII rigid soil layer (Shisheng
LU & Yachao 131 1987). On the other hand,
there are two main niodels considering tlie pile-
soil system is non-linear, one is the model
consisting of discrete springs, damper and
frictional element (Matlock, H.: Foo, H.C. LCt
Bryant, L.M. 1978), the other mode1 assurries
that there is a cirque-column weakencd rrrca and
the mechanical character of soil in the area is
different from that of soil out the area (Novak,
M. & Sheta, M. 1980, Novak. M. & Han, Y.C.
19S8). These models have provided new
approaches to study the dynamic characteristic of
the pile, however, these researches only
emphasized on the single excitation acting on the Figtire 1. zjlna15rlicai moLjcl

51 7
additional moment OII the pile top, which in which E =modulus of elasticity of pile,
thereafter causes the damage of pile in EI =bending rigidity of the pile, y =horizontal
probability. With the increase of this kind of
displacement of the pile, m,]=mass of the pile
project, the research of this problem becomes
per unit length, c =coefficient of viscosity of the
more important. In this paper, dynamic property
of precast pile under the eccentric load is studied, soil, foundation modulus and coefficient of
and effects of various factors on dynamic viscosity of soil, m =constant of foundation
response of pile is analyzed to improve the modulus which is proportional to deep,
acquirement on dynamic property of pile when b =length of side or diameter of the pile,
driving eccentrically, which subsequently
x = distance from pile top.
provided warrant for engineering design and
Numerical method can solve this differential
construct.
equation, the partial derivatives that the
displacement of the pile to x is dispersed by
2 CROSSWISE DISPLACEMENT AND IN- finite-difference method, and the partial
TERNAL FORCE ANALYSIS OF PILE derivatives that the displacement of the pile to t
is dispersed by Newmark method. The division of
Based on Winkler model, the soil around pile is the different unit of the pile is shown in Fig 3.
substituted by a series of distributed springs an3
dampers. Pile is assumed as equivalent section,
elastic, and the pile top is free. After the point of
pile enter into bearing stratum, the penetration
of pile is so little that we can assume that the
point of pile is fixed in the bearing stratum, and
the length of pile on the ground is EI,, the
length of pile under the ground is N 2 (Figure 1).
The moment caused by eccentric impact force
acting on pile top is :
M ( t ) = N o siiiwt . e (1)
in which t =time, N o =maximal value of the
impact force, e =:eccentricity of the impact force,
05 wt < 7r ,w = 7r I I , .
where t , =duration of the impact force (Figure
2).
When eccentric vertical load is exciting pile
top, besides dynamic moment and dynamic axial
force, there are pile's dynamic inertial force,
counter-force and viscosity damping of the soil
acting on the pile. Based on assumptions above,
deduced by elastic mechanical, the crosswise
vibration equation of the pile excited by
eccentric vertical dynamic load on the pile top is:
Assuming that i is the number of the unit of
time, and j is the number of the unit of pile.
[KXYI = bfI (6)
After the partial derivatives of equation (1) are in which, [ K ] is a coefficients matrix with the
replaced with center different equations and
order of n + 1 :
Ncwrnark equations, equation (1) can be written
as: [KI =
A . Y,,/-2 + B . Yl,/-l + Cl,/ . Y l , , 0 1 - 2 1 0 o...o 0 0 0 0
- 1 2 - 2 1
0 O ' . . O 0 0 0 0
A B C, D E 0 ... 0 0 0 0 0
0 A B C , , D E . . . O 0 0 0 0
in which , j = 2 - r? -2, ,
y k =displacement of ... ...
0 0 0 0 0 0 ... A B C,,,?D E
the pile unit which the number is 1 at the time 0 0 0 0 0 0 . ' . 0 0 0 1 0
unit which the number is k , 0 0 0 0 O O " ' 0 0 1 0 - 1
A = F = - El the column vector of crosswise displacement in
11' '
center of every pile unit is:

bI = {Y,0 YI,,> Y , , 2 . . Y ,
7 ' 11-2 ,
Yl ) I - , > Y,,,, I' (7)
the column vector of load in center of every pile
unit is:

the program based on the method mentioned

above is provided in this paper, which can
present the displaceinent response of every pile
unit at any time, and so do the shear force and
the moment.

(there is no algebra function of in and c in

3 EFFECTS OF CROSSWISE VIBRATION
equation (4) and ( 5 ) of the pile unit above
OF PILE
gro u 11 d )
where h = step length of center different
When pile haininer impacting on pile top, the
H
= ( H=pile length), s = step length of time, pile-soil system is vibrated by the excitation of
n-3
pile hammer to pile, and the vibration is
a , 6 is constants: a 3 0.25(0.5 + 6)' , 6 3 0 . 5 ,
transient and nonperiodic. Commonly, the time
s, =(j-2).h, I, =i..r, when 0 <t < t,, of energy translation is short than cycle period of
auto-oscillation of the pile, but the duration of
hi, = N o .sinat, ; when 1, <I <f2, NI =0 . response can be several times of cycle period of
Additionally, there are four equations auto-oscillation of the pile. There are 5 main
deduced with boundary conditions, thus, as for factors effecting on the crosswise vibration of
every time unit, there is one equation group pile, which are excitation intensity, impact
consist of 17 + 1 equations, which can be written duration of pile hammer, eccentricity of pile
as: hammer, foundation modulus and coefficient of

519
viscosity of soil. The effects of them on the
crosswise vibration of pile can be analyzed by 5
examples following, where the crosswise section
of the precast concrete pile is square, and the
parameters H,HI,H , , E are invariable, that is
N = 2 4 m , H, = 2 m , H,= 3 2 m , E = 3 x 1 0 7 K P a .
It is found that the maximal moments of the pile
are all at pile top through computing. and their
values are N o . e , so this paper only analyzed the
maximal crosswise displacement and maximal
she ar .
(1) Effects of excitation intensity on crosswise
vibration of pile. Assuming that t , = 0.02s,
L' = 300 KN . s / m , n7 = 500OKhi / n?-', e = 0.1m .
From Figure 4 and Figure 5 , it is found that the
maximal displacement and shear are increasing
ivith excitation intensity of the pile hammer, the
main reason of which is that the moment at pile
top is proportional to the excitation intensity of
pile hammer when the eccentricity of pile
hammer is stated. The niaximal displacement
and shear are affected mostly by length of side or
diameter of pile. They decrease while length of
side or diameter of pile increases. So we can
adopt the lighter pile hammer or descend it in
order to reduce the additional internal force of
the pile caused by the eccentricity of the pile
ham mer .
(2) Effect of impact duration 01 pile hammer on
crosswise vibration of pile. Assuming that
N o =5000KjV, c=300KN.slm, e=O.lm,
ni = 500OK.A'! m'. From Figure 6 and Figure 7, it
can be found that the crosswise vibration of pile
is mostly affected by the duration of impact, and
the maximal displacement is increasing with Figurc 6. Curve of I he maximal displacement changcd
duration of impact, but the maximal shear with impact duration
presents decreasing trend while duration of
impact increases. The transient vibratioywhose driving, accordingly, the internal force of the pile
impact duration is very short increases the will decrease effectively. But the heavier pile
internal force of the pile, while reduces the hammer will increasing the excitation intensity,
capability of pile to endure dynamic load, and so we should choose the appropriate pile
therefore destroy the pile. We usually adopt the hammer v, hcn dri\iing.
heavier pile hammer and thicken the buffer (3) Effect of eccentricity of pile hammer on
cushion to increase the duration of impact when crosswise vibration of pile. Assuming that

520

Figure 9. Curve of the maximal shear changed with
eccentricity of pile hammer
N o = 5000KN ,c = 3O O K N . s / n? , = 5000KN / W I ' ,
I , = 0.02s. From Figure 8 and Figure 9, it can be
found that the maximal displacement and shear
is proportional to the eccentricity of pile hammer
when the impact force is invariable. Actually,
what controls the crosswise vibration of the pile
when driving eccentrically is the shear wave
excited by the moment at pile top. Thus it can be
seen that we must adopt a series of methods to
521
Figure 11. Curve of the maximal shear changed with
foundation modulus
control the eccentricity of the pile hammer. For
example, enforcing the management during
constructing in order to ensure the pile hammer,
the pile cap and the pile in one axis; keeping the
thickness of the buffer cushion equal.
(4) Effect of foundation modulus on crosswise
vibration of pile. Assuming that N o = 5 0 0 0 K N ,
From
e = O . l m , c = 3 0 0 K N . s / n i , t , =0.02.~.
Figure 10 and Figure 11, it can be found that the
crosswise vibration is less affected by the
foundation modulus.
(5) Effect of coefficient of viscosity of soil on
crosswise vibration of pile. Assuming that
t , = 0.02~, N o = 5000KN, m = 5000KN/m4,
e = 0 . 1 ~ From
. Figure 12 and Figure 13, it can
be found that the crosswise vibration is less
affected by the coefficient of viscosity of soil
when the part of pile head is on the ground. But
when the part of pile head is under the ground,
the damping of soil to pile decreases the maximal
Figure 13. Curve of the maximal shear changed with
coefficient of viscosity of soil
displacement of pile while increases the shear
being. Analyzing with example (4), we can
conclude that the effect of soil to pile is little
when the part of pile head on the ground, and
the length of the part of pile on the ground
hardly affects the crosswise vibration of pile. This
can be seen on the other hand, the effect of soil
to pile is obvious only at the part of pile head.
4 ANALYSIS OF ENGINEERING EXAMPLE
A certain precast concrete square pile (Figure
14), H = 24n1, H ,= 2m H z= 2 2 m , b = 0.5m,
)
E = 3.5 x 10‘ KPa . adopting the cylinder diesel
hammer weighs 4 tons when driving. Assuming
that the inipact force is N o =45OOKN, the
eccentricity of the impact force is e = 0.08~1,the
impact duration of pile hammer is I , =0.02s.
The lognitudinal placing of steel bars are 016,
and the placing of hooping is (D6@50, the
522
Figure 15. Cur-vc ol displaceriient of pile at different
time
grade of concrete is C40, the pile is in three soil
layers, the parameters of the soil is in Figure 14.
This paper plots the curves based on the
conclusion of computing at the time of t , / 4 ;
I , /2 and 31, /4 (Figure 15-17).
From Figure 15-17, it can be found that the
maximal displacement and moment of pile are at
the pile top, but the maximal shear is 3 meters
away from the pile top. The maximal moment is
3 6 0 m . ? I ? , and the maximal shear is lOOKN .
The dynamic response in 10 meters away from
pile top is rather great, but it is almost zero out
of 15 meters away from pile top. Through the
further checking computation of pile intensity, it
can be found that the pile was destroyed by the
eccentric impact force of 4500KN. The maximal
axial load that this pile can bear is 4.88 x 10’ KN ,
but the maximal eccentric load that this pile can
bear is only 3 . 8 5 x l O 3 K N , so the vertical
bearing capacity of the pile was decreased

Figure 17. Curve of moment of pile at different ime
approximately 21%. In a word, the impact
capacity that the precast concrete pile cai bear is
great decreased because the pile hammer is
eccentric.
5 CONCLUSIONS
In this paper, the dynamic response of precast
concrete pile which eccentric dynamic vertical
load acts on pile top has studied, and effects of
excitation intensity, impact duration of pilc
hammer, diameter or length of side of pile,
eccentricity of pile hammer, foundation modulus
and coefficient of viscosity of soil on the dynamic
response of pile has analyzed. Through analyzing
in theory and calculation, it can be concluded as
follows:
1. The effect of the soil on the crosswise vibr-
ation of pile is obvious only when the length of
the part of pile on the ground is shorter than 6
meters. When the part of pile head is under the
ground, the maximal displacement of pile
decreases while the coefficient of viscosity of soil
increases, but the maximal shear increases
accordingly. Commonly, the coefficient of
523
viscosity of the hard clay is greater than that of
the soft clay, so the internal force of the pile is
greater when eccentric driving in the hard clay
than in the soft clay.
2. The maximal displacement increases while
duration of impact developing, but the maximal
shear presents decreasing trend while duration of
impact increases. For adopting the heavier pile
hammer or thickening the buffer cushion will
increase the duration of impact when driving, the
internal force of the pile can be decreased
effectively by these measures.
3. As for the flexible long pile is concerned,
the dynamic response of the pile is little affected
by the length of pile. The dynamic response of
the pile is almost invariable regardless of length
of the part of the pile on the ground when the
pile head is on the ground, the main reason of
which is that the pile is not restricted by the soil
and the dynamic response of the part under the
pile head is little.
REFERENCES
Lu, S.S. & Lin, Y.C. 1987. The calculation and
arzabsis of pile foundation. People traffic
publishing House: 304-318.
Matlock, H. et al. 1978. Simulation of lateral pile
behaviour under ear-thquake motion. Pro-
ceedings ASCE Specially conference on
Eart hqua ke Engin - eering and Soil Dyizamics .
Pasadena, Calif: 600-619.
Novak, M. & Sheta, M. 1980. Approximate app-
roach to contact effects of piles, Proc. ofDy-
izanzic response of Pile Fourzdation. An-
alytical aspects, ASCE, Florida: 53-79.
Novak, M. & Han, Y. C. 1988 Impedance of soil
layer with boundary zone. The University of
Western, Ontario research Report, GEOT:
16-85.
This Page Intentionally Left Blank
 Application of Stress-Wave Theory fo Piles, Niyama SC Beim (eds) C) 2000 Balkema, Rotterdam, ISBN 90 5809 1503

onitorirag and c0ne;rsl of ynamic effects of pile i ~ S ~ a ~ prior

l ~ t itoo ~
driving

M. R. Svinkin
VibmCoiisidt, Clevelaiid, Ohio, USu

A4BSTRACT:Pile driving induced vibrations may be harmful to adjacent and distant structures, sensitive
instruments and people. It is important to assess intolerable vibrations before the beginning of construction
activities. Guidelines for preconstruction survey are presented. Analysis of existing methods shows
potentialities of various approaches to predict ground and structure vibrations. This paper demonstrate the
application of the impulse response function prediction method to solve the geotechnical problem of predicting
ground and structure vibrations before pile installation. Control of the predicted vibrations is important step
in preventing intolerable-vibrations.

1 INTRODUCTION Staalduinen and Waarts (1992), Svinkin (1992) and

Pile driving generates elastic waves in soil which
may adversely affect surrounding buildings. Their
effects range from serious disturbance of working
conditions for sensitive devices and people, to
visible structural damage.
The dynamic effect of pile installation on adjacent
and distant structures depends on soil deposits at a
site and susceptibility ratings of structures. It is
likely that intolerable structure vibrations may be
induced in close proximity of the driven piles, but
permanent deformation resulting from vibratory
densification of loose granular soils may occur at
various distances from the source.
It is important to assess the dynamic effect before
the beginning of pile driving. Therefore various
approaches are used to predict soil and structure
vibrations.
Monitoring construction vibrations have to be
started prior to pile driving to provide the safety and
serviceability of sound and vulnerable structures and
facilities.
Specific guidelines for the monitoring and control
of construction vibrations in particular generated by
pile driving were prepared by Dowding (1996) and
Woods (1997). Separate questions of the dynamic
effects of pile installation on surrounding structures
were considered, e.g. Heckman & Hagerty (1978),
Wiss (1981), Lacy & Gould (1985), Broers &
Dieterman (1992), Massarsch (1992), van
others.
This paper presents some guidelines for
preconstruction survey, prediction and control of soil
and structure vibrations generated by pile driving
before the start of construction activities at the site.
2 PRECONSTRUCTION SURVEY
A predriving survey is the first step in the control of
construction vibrations to ensure safety and
serviceability of adjacent structures and/or distant
structures with sensitive equipment like hospital
facilities and offices with precision instruments.
According to pile driving practice, construction
vibrations may cause direct damage to structures at
a distance about of one pile length from the driven
pile. However, there is no single opinion regarding
the maximum radius of a preconstruction survey
area with buildings surrounding a construction site.
Dowding (1996) suggested a radius of 120 m of a
construction activities, or out to a distance at which
vibrations of 2 mm/s occurs. Woods (1997)
considered distances of as much as 400 m to be
surveyed to identify settlement damage hazards.
Obviously, a radius of the area of preconstruction
survey is various and depends on building condition
and utilization.
The predriving survey includes a few steps of site
investigations.

525
 @ Inspect present conditions of surrounding
buildings.
Perform damage susceptibility study to establish
vibration control limits.
Measure vibration background at the area under
investigation.
Assess problems such as cracking of building or
foundation failure.
First of all a preliminary desk study of a layout of
the area for a preconstruction inspection and a site
walk-over observation should be made. Existing
cracks found in buildings have to be marked. It is
necessary to distinguish cosmetic and structural
cracks. Most attention should be paid to cracks in
the structures themselves. The width of cracks
should be measured with a proper ruler.
Determining the cause of cracking is important to
predict lengthening and dilatation of old cracks
under the vibration effect of pile driving.
For assessment of the dynamic effect on
surrounding structures it is necessary to take into
account the thresholds of damage, cracking and
perception. Buildings inspected under the contract
requirements are commonly classified depending on
structure susceptibility to cracking and proximity of
structures to pile driving operations.
Structure susceptibility is usually related to the
threshold of cosmetic cracking (Dowding 1996) and
depends on a degree of degradation of the building
structural and nonstructural systems. Apparently, it
makes sense to broader consider this terminology as
susceptibility of the building-soil system depending
on degradation of building systems, utilization of
buildings and soil conditions. It is important for
certain cases. For example, a building, located in
the proximity of the driven piles, identified as
having low susceptibility and built on liquefiable
soils might have substantially larger deformations
than a building identified as having high
susceptibility but erected on non-liquefiable soils at
long distance from the driven piles. Certainly, for
some sites only the threshold of cosmetic cracking
could be sufficient.
Measurement of a vibration background at soil and
buildings with sensitive equipment and/or
computerized technology should be a part of the
preconstruction survey. The vibration measurement
might reveal microvibrations and vibrations induced
by industrial machinery located nearby. A waveform
recorder has to be used for these measurements.
Assessment of the dynamic effect of pile
installation on surrounding structures can be made
on the basis of prediction of soil and structures
vibrations from pile driving.
526
3 VIBRATION ENVIRONMENT
Pile driving operations generate different kinds of
waves in the ground, but Rayleigh waves have the
largest practical interest. Rayleigh waves become
predominant over other wave types at distances
smaller than a pile length. Impact hammers induce
transient ground vibrations with the dominant
frequency between 7-50 Hz (Svinkin 1999), and
vibratory hammers induce steady-state ground
vibrations in the frequency range of 5-40 Hz
(Warrington 1992).
The intensity of ground vibrations at the surface
and the certain depth are comparable. Linehan et al.
(1992) studied the effect of pile driving in the
proximity of the gas line with a pipe diameter of
1 m. The sheet and H piles were driven in sandy
soils with vibratory and diesel hammers,
respectively. For both types of driving, particle
velocities on the pipe were less than those measured
at the surface at a distance of 1.5 m but became
similar at distances about 10-12 m from driving.
Barkan (1962) analyzed the variation with depth of
vertical displacements in loessial clay induced by a
pile driver and concluded that at small depths about
of 0.2 to 0.5 wavelength changes in vibration
displacements are relatively small.
3.1 Deterministic approach
Statistical analysis is used for evaluation of the test
data from blasting (Dowding 1996). Such an
approach for assessment of other construction
vibrations should be carefully used because ground
vibrations generated by the dynamic forces applied
to the soil directly or transferred through driven
piles have deterministic nature.
The effect of large plastic soil deformations at the
contact area under falling mass on ground vibrations
was studied with falling mass of 15.0 tonnes at a site
where soil deposits were mostly fine sands with
natural moisture content (Svinkin 1996a). The drop
height was 10 m. Many impacts were performed at
the same spot; consequently, large plastic soil
deformations occurred at the point of impact.
Records of ground vibration displacements at various
distances from the place of impact on the ground are
depicted in Fig. 1.
Comparison was made for records obtained for
two equal impacts with different degrees of plastic
soil deformations at the contact area. In particular,
vibrations were measured at distance of 43 m for the
first and ninth impacts, and at a distance of 57 m for

r = Z rn
r=33 rn
Figure 1. Comparison of two different displacement records
of ground vibrations in fine sands for identical impacts on
ground by falling mass of 15.0 tonnes
the first and seventeenth impacts. For the first
impact, the falling mass dropped on a flat ground
surface, but for the seventeenth impact, it dropped
into a pit deeper than 1 m. In spite of considerable
soil deformations at the contact area, each pair of
ground surface vibrations are similar at locations of
measurements. The results demonstrate that at any
location on the ground, except probably a zone at
close proximity to the source, soil vibration
displacements measured simultaneously with impact
on the ground are stable, have well-defined shapes,
and are independent of the intensity of soil
deformations at the contact area. The differences
between displacement amplitudes measured during
various impacts are within the limits of error of the
measurement system. This confirms the necessity of
using a deterministic perspective for prediction and
analysis of construction and industrial vibrations.
3.2 Attenuation of surface waves
The waves travel outward from the construction
source and attenuate in the results of geometrical
spreading and material damping.
The following equation is known to calculate the
527
amplitude reduction of Rayleigh waves, generated by
an earthquake, between two points at distance rl and
r2 from the source as (Golitsin 1912)
Where A, = amplitude of vibrations at a distance rl
from the source, A, = amplitude of vibrations at a
distance r, from the source, y = coefficient of
attenuation. The term ( ~ J I - , ) ~indicates
,~ the
radiation or geometric damping and the term exp[-
y(r2-r1)] indicates the material damping of wave
attenuation between two points.
Equation (1) is highly popular in prediction of the
peak particle velocity (PPV) of vertical ground
vibrations from construction sources, for example,
Massarsch (1992), van Staalduinen & Waarts
(1992), Gerasch (1998) and others.
Nevertheless, there are obstacles for the
application of equation (1) to predict maximum PPV
of vertical ground vibrations (Svinkin 1999). Test
data along the ground surface shows that for various
pairs of widely separated points on the ground
surface, values of y can differ by more than an
order of magnitude and even change sign. Thus, the
coefficient, y,acceptable for small distances may be
inadequate for long distances. On account of wave
refraction and reflection from boundaries of diverse
soil layers, an arbitrary arrangement of geophones at
a site can yield incoherent results of ground
vibration measurements because waveforms
measured at arbitrary locations at the site might
represent different soil layers. Moreover, an
amplitude of ground vibrations is unknown before
the beginning of pile driving.
So, problems with uncertainty in assignment or
determination of the coefficient, y , and unknown
referenced amplitude show that the application of
equation (1) to predict ground vibrations is
questionable.
Brettmann & Cotton (1999) suggested other
equation to predict the peak particle velocity, v, of
ground vibrations
v = K(E/r)'.' e-7'
where K = soil constant; E = energy transferred to
soil; r = distance from source; y is the same as
above.
The term K(E/r)o.5 represents the geometric
damping and there is some uncertainty in
determination of constant K. The term exp[-yr]
represents the material damping of the ground and
there are problems with determination of coefficient
y described above. So, equation (2) has
disadvantages similar to the equation (1).
Wiss (1981) found the ground velocity-distance-
energy relationship, so-called scaled-distance
approach, to calculated the peak ground velocity at
surface distance, D, from a source normalized with
energy as
v = k[D/fi]-" (3)
Where W, = energy of source or rated energy of
impact hammer, k = value of velocity at one unit of
distance. The value of 'n' yields a slope in a log-
log plot between 1 and 2. It was an important
finding because a slope of amplitude attenuation for
all tested soils was in the narrow range of 1 to 2. It
turned out that the scaled-distance approach was
very useful in the assessment of construction
vibrations,
On the basis of the actual range of energy
transferred to piles and the range of the measured
peak particle pile velocity at the top of steel,
concrete and timber piles, the results of Woods and
Jedele (1985) were adapted by Svinkin (1996b;
1999) to calculate the peak ground velocity prior to
the beginning of pile driving. The graphic
relationship between PPV of ground vibrations and
scaled distance from driven piles is used. The
results obtained from the graphs depends on PPV of
the source. The peak particle velocity at the pile
head can be calculated in advance as (Svinkin,
1996b)
v = 12-
Cw.
(4)
ZL
where Z = ES/c is pile impedance, E = modulus of
elasticity of pile material, S = pile cross-sectional
area, c = velocity of wave propagation in pile, W,
= energy transferred to the pile.
This new development of the scaled-distance
approach eliminates the need to know in advance the
factor, k, and enhances accuracy of predicted upper
limits of ground vibrations before pile installation.
The hammer energy is important source property
which effects intensity of surface waves, e.g.
equations (2) and (3). However the energy of
source is not always the dominant factor in
determination of the intensity of ground vibrations.
Dowding (1996) made comparison of ground
vibrations induced by pile driving with a diesel
hammer and by Franki pile driving and revealed that
the latter developed two times more energy but
induced smaller ground vibrations at the same
528
distances from the sources. Perhaps the cause of
this phenomenon was a higher attenuation of surface
waves in the loose soil deposits where Franki piles
were driven. This observation underlines the
significance of soil contribution to the formation of
ground vibrations.
4 STRUCTURE RESPONSE
4.1 Predicting structure vibrations
The impulse response function prediction method
(IRFP) is based on the utilization of the impulse
response function technique for predicting complete
vibration records on existing soils, buildings and
equipment prior to installation of construction and
industrial vibration sources (Svinkin, 1996a; 1999).
The impulse response function is an output signal of
the system based on a single instantaneous impulse
input. Impulse response functions are applied in the
analysis of any complicated linear dynamic system
with unknown internal structure for which its
mathematical description is very difficult.
The following is a general outline of the method
for predicting vibrations at a distance from an
impact source.
1. At the place in the field for installation of the
impact source, impacts of known magnitude are
applied on the ground. The impact can be created
using a rigid steel sphere or pear-shaped mass falling
from a bridge or mobile crane, or a hammer blow
on the tested pile. At the moment of impact on the
ground, oscillations are measured and recorded at
the points of interest, for example, at the locations
of devices sensitive to vibrations. These oscillations
are the impulse response functions of the considered
system which automatically take into account
complicated soil conditions.
2. Various ways are used to determine the
dynamic loads on the ground from different
vibration sources. For pile driving, dynamic loads
are computed by the wave equation analysis. In the
case of the operation of machines on foundations,
these loads can be found using existing foundation
dynamics theories. For dynamic compaction sites,
loads from the source are easily calculated with
known falling weights and heights.
3. Duhamel's integral (Smith and Domney, 1968)
is used to compute predicted vibrations which will
arise from operating construction impact source.
For each single output point, the considered input
- soil medium - output system is a one degree of
freedom system and predicted displacements can be
written as follows
where I, = impulse force transmitted from machine
to foundation, kNs; fI1z = circular natural frequency
of vertical vibrations of foundation, rad/s; f n d =
circular natural frequency of vertical damped
vibrations of foundation, rad/s; CY = effective
damping constant, rad/s; \(t-7) = impulse response
function at the output point under consideration; 7 =
variable of integration.
Examples of predicted results are shown in Fig. 2,
and 3. Measurement and prediction of vertical and
horizontal ground surface displacements were made
at diverse distances from the foundation under a
sizeable drop hammer with a falling weight of
147.2 kN and a maximum drop height of 30.0 m.
The soil at the site consisted of about 1.6 m of loose
sand followed by about 6.8 m of medium density
sand and 1 m of sandy clay underlain by about 10 m
of slightly moist sand. The water table was about
6 m below the ground surface. The Rayleigh wave
velocity was 270 m/sec. A layout of the machine
foundation, the place of impact on the ground and
geophones is displayed in Fig. 2.
Predicted and measured vertical and horizontal
components of ground surface vibrations at eight
locations are shown in Fig. 3. It can be seen that
good agreement is matched in time domain vibration
records, except for horizontal vibrations at two
locations close to the foundation. This can be
explained by the different wave paths from the
foundation under the operating machine and the
place of impact on the ground. The distance
between these two sources was 18.7 m. Lack of
coincidence of the two dynamic sources slightly
affected ground vibrations at a distance from the
machine foundation. Agreement of predicted and
measured vibration displacements is quite
satisfactory. The differences between the peak
predicted and measured vibration amplitudes are less
than 30 % at distances larger than 43.0 m from the
foundation. For some individual points amplitudes
actually coincide.
Spectrum analysis of predicted and measured time
histories revealed that both records have similar
frequency domain curves with the same dominant
frequency. Moreover, predicted records are slightly
dependent on the dynamic parameters of the source
(Svinkin, 1999).
529
Figure 2. Layout of machine foundation, place of impact
loads on ground and geophones
5 CONTROL OF VIBRATIONS
There are three ways of how ground vibrations from
construction sources can effect adjacent and distant
structures. First, excitation frequencies from ground
vibrations do not match structural natural
frequencies. Second, ground vibrations may
generate the conditions of resonance in the building.
Third, foundation settlements result from relatively
small ground vibrations. For all situations, sensitive
equipment in buildings will be exposed to vibrations.
The tolerable vibration level can be verified before
beginning of construction activities at a site.
For structures which natural frequencies do not
match excitation frequencies of ground vibrations,
structural responses would be lower than the ground
vibrations. Prediction of the maximum peak particle
velocity independent of frequency could be used.
However, it is necessary to point out that low
structure vibrations cannot prevent high vibrations of
equipment installed in the building. Dowding (1996)
demonstrated a case where PPV on the ground and
the second floor were 2.03 and 1.52 mm,
respectively, but the table on the floor had PPV of
10.5 mm.
Levin (1969) found that 3-4 cycles of ground
vibrations could generate conditions of resonance in
the building. Such conditions can be evaluated on
the basis of the prediction of complete time domain
records.
Significant foundation settlements caused by pile
driving in vulnerable sands can result from PPV of
ground vibrations as low as 2.5 to 5.1 mm/s (Lacy
& Gould 1985). Such velocity values are much less
than 51 mm/s which is considered as the tolerable
safety limit for buildings. Therefore at sites with a
predominant granular soil deposits, the maximum
peak particle velocity that varies with frequency
should be predicted at the area where intensity of
vibration is expected higher than 2.5 m / s .
 Vertical
a b
0
r = 2 5 rn
m

r = 3 3 in
0
m
N

r = 4 3 rn

r = 1 3 2 rn
0

r = 2 0 0 rn
0

r=266 in
m

0.2 s
U
Time (s) Time (s)

Horizontal

0
r = 2 5 rn
m

r = 3 3 rn
.--
0

0 r = 4 3 rn
r

r=57 m
0
r

r = 2 0 0 rn
0
l-

r=266 m
Lo
r-4

0.2 s
U
Time (s) Time (s)

Figure 3. Vertical and horizontal soil vibrations from operating drop hammer, measured (a) and predicted (b) vibrations

530
6 CONCLUSIONS
Pile driving is a wide-spread source of construction
vibrations which may harmfully affect surrounding
buildings. It is important to accurately predict
vibrations of ground, structures, and sensitive
devices prior to the beginning of construction
activities to avoid the undesirable effect of generated
vibrations.
The predriving survey should be made to ensure
safety and serviceability of adjacent structures and/or
distant structures with sensitive equipment like
hospital facilities and offices with precision
instruments.
Empirical equations provide only calculation of a
vertical amplitude of ground vibrations and not
always with sufficient accuracy. For pile driving,
the scaled distance approach with calculated peak
particle velocity of the source is probably the most
appropriate method for predicting upper limits of the
peak particle velocity of ground vibrations.
The impulse response function prediction method
(IRFP) has obvious advantages in predicting time
domain ground and structure vibrations prior to the
beginning of construction activities.
The IRFP method provides opportunity to control
the tolerable vibration level of soil, structure and
sensitive equipment in advance. This is an
important step in preventing intolerable vibrations
from pile installation.
REFERENCES
Barkan, D.D. 1962. Dynamics of bases and
foundations. New York: McGraw Hill Co.
Brettmann,T. & B. Cotton 1999. Case history:
estimating ground vibrations caused by pile
driving. DFI 24th Annual Members ’ Conference,
Decades of Technology - Advancing into the
Future: 161-171.
Broers, F. & H.A. Dieterman 1992. Environmental
impact of pile-driving. In F. Barends (ed.),
Proceedings of the Fourth International
Conference on the Application of Stress-Wave
Theory to Piles: 61-68, Rotterdam: Balkema.
Dowding, C.H., 1996. Construction Vibrations.
Upper Saddle River: Prentice Hall.
Gerasch, W.J. 1998. A way to predict vibrations.
Seventh International Conference and Exhibition
on Piling and Deep Foundations: 3.3.1-3.3.13.
Rickmansworth: Westrade Group Ltd.
Golitsin B.B., 1912. On dispersion and attenuation
of seismic surface waves (in German). Russian
53 1
Academy of Science News, Vol. 6, No. 2,
Heckman, W.S. & D.J. Hagerty 1978. Vibrations
associated with pile driving. Journal of the
Construction Division, A X E , 104(C04): 385-394.
Lacy, H.S. & J.P. Gould 1985. Settlement from
pile driving in sands. Proceedings of ASCE
Symposium on Vibration Problems in Geotechnical
Engineering, Detroit, Michigan, G. Gazetas and
E.T. Selig, Editors, pp. 152-173.
Levin, G.E. 1969. Dynamic influence of drop
hammer foundations on surrounding structures (in
Russian). Proceedings of the Second Conference
on Dynamics of Bases and Foundations: 147-152,
Moscow: NIIOSP
Linehan, P. W., A. Longinow & Dowding, C .H.
1992. Pipeline response to pipe driving and
adjacent excavation. Journal of Geotechnical
Division, ASCE, 118(2): 300-316.
Massarsch, K.R. 1992. Keynote lecture: Static and
dynamic soil displacements caused by pile driving.
In F. Barends (ed.), Proceedings of the Fourth
International Conference on the Application of
Stress- Wave Theory to Piles: 15-24. Rotterdam:
Balkema.
Smith, G.M. & G.L. Downey 1968. Advanced
engineering dynamics. Scranton: International
Textbook Company.
Svinkin, M.R., 1992. Pile driving induced
vibrations as a source of industrial seismology. In
F. Barends (ed.), Proceedings of the Fourth
International Conference on the Application of
Stress-Wave Theory to Piles: 167-174. Rotterdam:
Balkema.
Svinkin, M.R. 1996a. Overcoming soil uncertainty
in prediction of construction and industrial
vibrations. In C.D. Shackelford, P. Nelson, and
M.J.S. Roth (eds.), A X E , Proceedings of
Uncertainty in the Geologic Environment: From
theory to Practice, Geotechnical Special
Publications No. 58. 2: 1178-1194.
Svinkin M. R. 1996b. Velocity-impedance-energy
relationships for driven piles. In F. Townsend,
M. Hussein and M. McVay (eds.), Proceedings of
the Fifh International Conference on the
Application of Stress-Wave Theory to Piles: 870-
890.
Svinkin, M.R. 1999. Prediction and calculation of
construction vibrations. DFI 24th Annual
Members ’ Conference, Decades of Technology -
Advancing into the Future: 53-69.
van Staalduinen & P.H.Waarts 1992. Prediction of
vibrations due to pile driving. In F. Barends (ed.),
Proceedings of the Fourth International
Conference on the Application of Stress-Wave
 Theory to Piles: 159-166, Rotterdam: Balkema
Warrington, D.C. 1992. Vibratory and impact-
vibration pile driving equipment. Pile Buck, Inc. ,
Second October Issue: 2A-28A.
Wiss, J.F., 1981. Construction vibrations: State-of-
the-Art. Journal of Geotechnical Engineering,
ASCE, Vol. 107, NO. GT2, pp. 167-181.
Woods R.D. 1997. Dynamic eflects of pile
installation on adjacent structures. Synthesis
Report, National Cooperative Highway Research
Program NCHRP Synthesis 253 , Washington,
D. C. : National Academy Press.

532
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Full-scale field-test study of dynamic soil resistance of vibratory driven sheet

piles

K.Viking
Deparonent of Civil a i d EnvironnzentulEngineering, Royal Institute c.f Technology (KTH),Stockholm, Swederi

ABSTRACT: The vibratory characterization of full-scale sheet piles vibratory driven into the ground has so
far not been performed systematically in detail. This paper presents the preliminary results of a series of full-
scale field-tests where the driveability has been continuously monitored. In the light of the presented results,
the paper briefly discusses how the complexity of driveability prediction can be subdivided into three parts:
vibro-, sheet pile-, and soil-related parameters, all of which significantly affect the driveability.

1 INTRODUCTION used instrumentation and main preliminary results of

Modern variable vibratory drivers to drive sheet
piles into the ground are extensively used all over
the world. It is not always obvious which available
driver that should be chosen to insure driveability.
Amongst the uncertainties attached to driveability
formulas, the key one is the assessment of the soil
resistance during the installation phase. Another one
is whether the assumption of pile rigidity, which is
reasonable in laboratory tests, can be applied to long
sheet piles in the field. The third obstacle is the dif-
ficulty in estimating the influence of clutch friction
on driveability, an unknown parameter, but believed
to be of significant importance.
This paper briefly discusses these uncertainties
amongst other important parameters significantly in-
fluencing the driveability, together with some of the
preliminary results of a series of well-documented
full-scale field tests.
1.1 Objectives
The overall objective of the ongoing research project
at KTH is to improve the understanding of the per-
formance of vibro-driven sheet piles and to develop
a better capability of predicting the driveability.
From the information collected during the litera-
ture studies, (Viking 1997), and conducted labora-
tory tests, (Viking 1998), a driveability monitory
system was designed, built, and appropriately
equipped in order to dacument the dynamic soil re-
sistance of full-scale vibratory driven sheet piles.
The literature studies reviled that publications as-
sociated with full-scale test are not generally avail-
able. However, in the foll.owing paragraphs are the
the performed tests presented and discussed.
2 MECHANICS OF PENETRATION INITIATED
BY VIBRO-DRIVERS
A first step toward predicting the driveability of vi-
bratory driven sheet piles is to define driveability,
thereafter determine the magnitude of forces acting
on the system, and identify the key factors influenc-
ing the driveability.
In order to do so, it is the author’s opinion that
the complexity of driveability prediction of vibratory
driven sheet piles can be subdivided into three main
parts: vibro-, soil-, and sheet pile related parameters
significantly affecting the driveability. Another issue
is the fact that driveability estimations must come
out of a trade-off between the accuracy of the drive-
ability model and ease of application to real prob-
lems.
2.1 Driveability
The driveability of a vibro-driven sheet pile can be
defined either in terms of the rate of penetration or
by the forces generated and acting on the vibrohheet
pile/soil system. The influences of various parame-
ters on the three main parts are dealt with in the sub-
sequent subsections.
2.1.1 Rate ofpenetration
Driveability can be visualized in two ways: i.) as the
relationship between driving time t and penetration
depth z, (Figure la) and ii.) as the plotted relation-
ship between the global rate of penetration v,, and the
penetration depth, (Figure lb).

-,I<>,,
, .-- /
/
....,...... ..... ,/
a) b)
igure la-b. Two schematic plots of the driveability curve.
The slope of the t-z curve, described in Figure la,
represents the global rate of penetration vp. The en-
largement of the t-z curve (Figure la) illustrates the
principle of the penetrative motion of the sheet pile,
i.e. the different up- and downward parts of the dis-
placement amplitude.
The pattern of the recorded v,,-z curve can be used
as a tool to study the change in dynamic soil resis-
tance. Rao (1993), involved in the extensive research
work at University of Texas at Houston, stated that
the penetration speed can be subdivided into three
states, ranging from slow to high penetration speeds,
(Table 1).
Table 1. States of penetration speed values vp in [rnrn/s], after
Rao (1993).
Slow Fairly high High
Vp<32 22< v,, <60 60> v,,
2.1.2 Forces
Holeyman et al. (1996) defined driveability using
forces instead of penetration speed. Driveability is
defined as long as the driving force F, generated by
the vibro-equipment exceeds the three dynamic re-
sistance forces generated along the shaft R,, toe R,
and in the clutch R,, represented by the following
expression:
Fg, = (Fo+ FL,)
2 X, + R, + R,
This definition of driveability is rather simple but
at the same time also powerful. However, the key is-
sue is the difficulty in assessing the dynamic soil re-
sistance along the shaft and toe during the installa-
tion phase. The forces in Equation (1) acting on the
vibro/sheet pile system are schematically described
in Figures 2 and 3.
2.2 Vibrator parameters
The choice of optimum vibrator parameters con-
cerning driveability has not yet been established.
However, choice of driving frequency seems to give
rise to two driving states: fast and slow vibratory
driving. The two states are mainly due to the fact
that the soil beneath the toe moves out of phase rela-
534
tivc to the penetrating toe. The dynamic shaft resis-
tance has been found to be virtually independent of
chosen frequency, indicating that the choice of
driving frequency is more important in respect to the
dynamic toe resistance.
It has also not yet been established how the static
surcharge force F,, should be determined on relation
to the unbalanced force F,, O’Neill et al. (1989).
However, it is well known that increasing the bias
weight increases the global rate of penetration.
The eccentric moment M , determines the ampli-
tude of both the displacement and the unbalanced
force, but it has not yet been established how the
best choice of static moment should be determined
concerning the driveability. Several authors, (e.g.
Rodger et al. 1980) have stated that the displacement
amplitude should be chosen (high or small) to drive
the pile efficiently at the chosen driving frequency
but none gives the underlying assumptions or why it
should be so.
The conducted field tests reviled that the effi-
<
ciency of the vibro-equipment is another essential
parameter. The peak compressive force measured at
the sheet pile head during the field test should be
theoretically equal to the sum of the applied static
surcharge force, maximum unbalanced force and de-
celeration of the vibrator, expressed according to
Equation (2).
Explanations of parameters in Equation (2) are
found in Figures 2 and 3. However, the recorded
peak forces in the sheet pile head were less than the
theoretical value. These results are believed to be
related to energy loss: in the hydraulic system, in the
connection between sheet pile and clamp, and also in
propagation of flexural and torsional energy, not
sensed by the sensors that only record axial load.
Recommendations concerning choice of the effi-
ciency factor are provided in Moulai-Khatir et al.
(1994). The efficiency for field conditions could be
estimated in the range of 20 < < < 25 percent ac-
cording to the authorsSheet pile parameters
It is well known that the influence of clutch friction
R, on both driveability and the vibrations in the envi-
ronment is of great importance. At the moment,
good research on the influence of the clutch on
driveability is being conducted amongst the Euro-
pean sheet pile manufacturers. However, knowledge
regarding clutch friction remains limited.
Another sheet pile related uncertainty that exists
in applying a driveability formula is pile rigidity,
which is a reasonable assumption in laboratory tests,
can be assumed for sheet piles in field tests.
The natural frequency of a longitudinal vibrating
rod can be calculated according to Equation (3).
When applying Equation (3) on the instrumented
sheet pile, with a bar velocity c, = 5,100 m/s, L = 14
m, the first mode (n=l, half wave mode), a longitu-
dinal resonance frequency of fi 182 Hz can be cal- It would be possible to treat the vibro/sheet pile
culated. system as a rigid body, if the displacement of the
entire sheet pile is in phase with the motion of the
f,,= il *
Cl,
__ (72 = 0,1,2, ...) (3) vibro-driver. Such an assumption would greatly
2L simplify the problem to a single-degree-of-freedom
The point to be made here is that longitudinal vibrating system, where the inertia force and driving
resonance phenomena generally occur at very high force are in phase with each other.
frequencies compared to the driving frequencies In laboratory tests by Viking (1998), a rule of
used in today’s vibratory equipment. thumb was used to treat the system as a rigid body if
the criterion t, 2 2, was fulfilled, where t, is the time
it takes the unbalanced force F,, to vary from zero to
maximum, and t, is the time it takes a stress wave to
travel the length equivalent to four times the length
L of the pile. A criterion allowing the model pile to
be treated as a rigid body if the driving frequency
was chosen to be equal to or less that 12.5 percent of
the resonant frequency, i.e. f,ls 0.125f,,.
In O’Neil1 et al. (1989) it was stated that a full-
sized pile could essentially be treated as a rigid body
if the driving frequency fll was chosen to be equal to
or less that 10 percent of the resonant frequency f,,of
the vibro/pile system. However, from the conducted
full-scale field tests it was noted that the vibrohheet
pile system used did essentially behave as a rigid
body when the driving frequency iiwas as high as
25 percent of the resonant frequency of a free vi-
brating 14 m long steel rod.
Viking et al. (2000) related another important
sheet pile related parameter to the lateral flexibility
of singularly vibro-driven sheet piles. The flexibility
is expiained by the fact that the driving force FCien-
ters the sheet pile web at an eccentric distance e
from the neutral axis of the sheet pile. The bending
moment M = e-F, induced at the sheet pile head re-
sults in a lateral sinusoidal motion U , of the sheet pile
during the installation phase. The flexibility of the
sheet pile was monitored by a laterally mounted ac-
ceIerometer at mid-position of the instrumented
sheet pile, (see Figure 6).

2.3 Soil parameters

The vibratory technique is known to perform best in
loose granular soils. It may not work as well in mod-
erately stiff, saturated clays, dense sand or gravel
since the dynamic soil resistance docs not decrease
as favorably as it does in loose sand during the in-
stallation phase.
Results from earlier investigations and research
by O’Neill et al. (1989), clearly reveal that the three
main soil related parameters that significantly affect
the driveability are the initial relative density, degree
of saturation, and layers with different density.
The primary soil mechanisms behind the favora-
bly reduction of penetrative resistance are the local
sand liquefaction, the induced cyclic motion of the
soil particles during driving, and the relation be-
tween initial void ratio and its critical state value.
The degree of saturation of sand has been found
to be closely related to the recorded global rate of
Figurc 2. Forccs acting on the vibro/soil/shcet pile system.

535
penetration. The rate of penetration has been ob-
served to be higher in saturated sand than in dry
sand, but the difference scems to be more pro-
nounced in the case of low displacement pilcs. Re-
corded pore pressure readings near the pile shaft (in
the sand) have been observed to be higher near the
pile shaft than those away from the oscillating pile
shaft. The pore pressure development during driving
is closely related to the driveability. However, stud-
ies by O’Neill et al. (1989) could not relate all of the
shear strength loss to the pore pressure build-up
during driving. This could be explained either by the
fact that liquefaction occurs near the oscillating sheet
pile, or by shear strength reduction due to the in-
duced cyclic motion of the soil particles during
driving, or a combination of both mechanisms.
The relative density is the soil parameter that has
been reported to have the most significant influence
on the global rate of penetration. The rate of pene-
tration has been found to decrease with increasing
relative density. The induced volume changes, (den-
sification or dilation) in the soil are believed to give
rise to a characteristic shape of the driveability vp-z
curve, see Figure lb.
The content of fines (greater than 12%) is a ne-
glected soil parameter that usually dictates if the use
of vibro-drivers will be successful or not, a fact that
has been experienced by several authors, but none
gives the underlying exp!anation.
Figure 3. Schematic descriptions of the relationship between driving force, penetrative motion, shaft- and toe resistance.
536
2.4 Pei i elra 1ioii 171CCIINIZi st i t s of lli c system
The difficulty in applying a driveability formula
comes from the fact that a great deal of trade-off
must be made between the accuracy of the candidate
formula and the practicability of applying it to real
problems. This means that the key issue is to access
the magnitudes of both the dynamic soil resistance
and the friction force R, in the sheet pile clutch.
However, the magnitude of the theoretical forcing
function of the vibro-driver Fd, i.e. sum of static sur-
charge force F, and unbalanced force F,, is easier to
estimate. It varies sinusoidally in time with driving
frequency and eccentric moment according to Fig-
ure 3a.
The motion of the sheet pile u(t) is described by a
downward directed sinusoidal displacement, corre-
lated in time with the driving force Fd, see Figure 3b.
The dynamic shaft resistance R, (Figure 3c) varies
between positive and negative, correlated with the
up- and downward motion of the sheet pile.
The dynamic toe resistance R, varies between zero
and maximum, also correlated with the motion u(t)
of the sheet pile, and reaches it’s peak value at the
lower end of the up- and downward penetrative mo-
tion u(t) of the sheet pile, see Figure 3d.
No theoretical explanation has been presented for
how the generated clutch friction R, varies during
the penetrative motion. However, it is well known
that the dynamic clutch friction R, is of great signifi-
cance.
3 FIELDTEST

The first test site was situated beside the steel-girder

bridges over the Fittja Straits at Virby, situated 20
kilometers south-west of Stockholm. The test site
was chosen for the driveability study for two main
reasons, firstly, due to the relatively homogeneous
soil conditions, and secondly, the high probability of
keeping the sensors in place.
All of the vibro-driven sheet piles were installed
in co-operation with the Swedish foundation com-
pany Stabilator AB,using a leader mounted ABI-
vibro, model MRZ 800V, mounted on ABI-leader
system TM 14/17L.
The driveability tests were executed on a 14 m
long instrumented sheet pile described in Figure 6.
Three uninstrumented sheet piles was initially driven
I
into the ground with only environmental effects and Figure 4. Evaluation of the CPT-1 results.
penetration speed monitored. The instrumented sheet
pile was then driven three times by itself and two
times into the clutch of the three initially installed found approximately 2.1 m below ground level, and
sheet piles, Green & Nilsson (2000). the soil, which is relatively well graded, varies be-
tween silty sand and gravelly sand. The relative den-
3. I Soil conditions sity is estimated to lie between 30 < D,. < 50 % from
the CPT using relationship from Jamilkowski et al.
The soil investigations consisted of soil sampling at (1985). The sand is considered to be normally con-
six levels, three CPT tests and 21 dynamic probing solidated with respect to its geological history. The
tests. Moreover, the pore pressure was measured mineral composition of the sand shows that it mainly
using one piezometer and one open standpipe. Apart consists of hard minerals, such as quartz and feld-
from the top two-meter thick clay layer, the soil spars. The results from the CP1 probing are pre-
consists of more than 40 m of loose to medium sented in Figure 4, showing a soil profile that is
dense glacial sand. The ground water table was relatively homogenous with depth.

Figure 5. Schematic of vibro-, soil-, and sheet pile instrumentation.

537
3.2 Itislrritn~iiluliol?
A
The instrumcntation system employed to record the
driveability gcncrated environmental effects con-
sisted of three parts: vibro-, sheet pile-, and soil-
instrumentation, (see Figure 6).
The instrumentation on the vibro-equipment con-
sisted of three sensors, monitoring the applied static
surcharge force (via oil-pressure), penetrative mo-
tion of the sheet pile, and the position of eccentric
weights (e.g. the varied static moment Me).
The instrumentation on the sheet piles consisted
of ten Weatstone mounted strain gauge bridges and
three accelerometers with position illustrated in Fig-
ure 6. The three low-g accelerometers, Analog De-
vices Inc., were mounted on specially developed
printed circuit boards.
Prior to the tests, the instruments were calibrated.
During the strain gauge calibration, the instrumented
sheet pile was loaded at one end by a hydraulic cyl-
inder and held between end plates welded on a 22 m
long I-beam. Prior to the driveability tests, the strain
gauges were zeroed when the sheet pile was free
hanging in the hydraulic clamp before driving it into
the ground.
The readings from the calibration procedures
were compared to standard known calibrations, de-
scribed in Green & Nilsson (2000).
The data acquisition system consisted of a DAT-
tape recorder Sony PC216Ax together with Sony
extension unit PCCX32Ax. The recorded dynamic
data obtained from the various instruments were
later anal yzed, using the acquisition software pack-
age PCscan MK IITM,further discussed in Sec-
tion 3.3.

3.3 Methods of analysis of data collected

Driving records of the results recorded on the DAT-
recorder were analyzed using the computer program
PCscan MK IITM.Short time windows of the digit-
ized time histories, at any depth of interest, were ex- $
P
ported to MS ExcelTMin order to present and graph
the different readings against each other. The choice
of window size should contain a sufficient number
of cycles of the recorded signal, and the sheet pile
should penetrate with a uniform velocity in the con-
sidered time window. The development of the pre-
sented time window of displacement U ( [ ) of the sheet
pile is obtained through double integration of the
pile head (or near the toe) acceleration records. Then
a correction constant c was applied to the recorded
acceleration a(t) to ensure that zero average accel-
eration exist for the considered time window dt = tll-
t, of interest.
The dynamic force-displacement curves were de-
termined from the time histories of forces and dou-
ble integration of corrected acceleration records. The
corrected acceleration a,(t) was first integrated ac-
cording to Equation (4) in order to obtain the veloc- Figure 6. Instrumented sheet pile with location of the sensors
ity change v(t) in time, (Figure 8a). mounted on the 14 m long PU16 profile, S240 GP.

538
 I,, kgm, f' = 41 Hz,in, = 1.670 kg, a,, = 130 m/s' to
V(t) = p; dt
1,
(I) (4) about = 515 kN. The actual delivered unbalanced
force amplitude to the sheet pile head is given by
Figure 9a to %-(236+105) = 170 kN, which gives an
The harmonic penetrative motion of the sheet efficiency of 6 = 100.(170/515) = 33 percent.
pile, (Figure 8b), is finally obtained by once more
integrating the sum of change in velocity, i.e. Equa-
tion (4), and the uniform velocity v,,, at the begin-
ning of the time window, I = I , , expressed according
Equation (5). ---
I so
U(t) = j("(+')
1,
v,,)dl (5) -100

4 RESULTS
-200 I
0,65 0,70 0,75 0,so
fl) trme / I 1
Observation of the time histories of the sheet pile- Iiead acceleration ot sheet pile I31
- - - -- .- --__ - -
head and sheet pile-toe forces, acceleration and dis-
placement of full-scale field tests have provided a
unique opportunity to develop a better understanding
of the penetrative mechanisms earlier addressed in
Section 2.4.
A few preliminary records of documented force
and acceleration time histories of a sheet pile with 0,65 0,70 0,75 0,80
14
no clutch friction, at 3.1 m penetration depth are pre- time [ s 1
sented in this section in order to discuss and point
out some of the significant aspects of the penetrative Toe acceleration ot sheet pile B1
behavior of vibratory driven sheet piles in clean
-200 r------- _____

sand. -100 -
---
s-=. 0..
4.1 Typical time histories CZ loo.
The sign convention for the signals is as follows:
positive acceleration occurs during the downstroke, 0,65 0,70 0,75 0,so
positive velocity corresponds to downward move- c) rime [ s I
ment of the sheet pile, positive force corresponds to
compression of the sheet pile. I p r e 7a-c. Typical time histories of acccleration signals.
Relationships of both the lateral acceleration a, at
together with the axial acceleration of the head a,,
and toe a,have been graphed in Figure 7a-c, at 3.1 m
penetration depth.
From the preliminary results it appears that the
phase difference between sheet pile head and toe ac-
celerations is small. The recorded lateral acceleration
of the sheet pile, which also appears to be in phase
with the axial, shows a lateral displacement ampli-
tude of about fi, = &Jw'= 5 0 / ( 2 . ~ - 4 1 )=~ 1 mm. At ~~ 065 0,70 0,75
rbue [ s 1
first glance, 5, = 1 mm does not appear significant,
but it is in fact about 40 % of the axial displacement *, 1 . Displaccmnt of shcct pile toe U1
amplitude.
Figure 8a-b graphs the double integration proce-
dure described in Section 3.3, resulting in the pene-
trative motion u(t) of the sheet pile toe.
The time history of the driving force FCiat sheet
pile head and the dynamic resistance R, near the toe
has been graphed in Figure 9a-b. The theoretically b) 0,65 0,m
0,70 t i m I sI 0,75
generated unbalanced force amplitude is calculated
according to Equation (2), with 77 = 50 %, Me = 12 Igure 8d-b. Intcgratcd velocity and displacement histories.

539
 Ilriviiig h r c r ~its1ic.etpile 131 3.) The sheet pile with head and toe acceleration
- 1 10
sensors behaved essentially as a rigid body during
- 60
the vibratory installation phase.
.
t
-10
40
4.) The description of the penetration mechanisms
< visualized in Figure 3, shows several similarities
.
,-i CX)
1.40 with the recorded behavior of full-scale sheet piles
1'93 under field conditions.
2% 5.) The preliminary results of the ratio of unbalanced
(1) 0.65 0,70 0,75 sheet pile head force to unbalanced force produced
time [ s ] by the vibrator were approximately 33 percent. It
0,65 0,70 0.75 0.80 should be noted, all sheet piles came directly from
45 the manufacturer, the test site provided favorable
soil conditions, and the vibro-equipment was new.
. 25

L 6 ACKNOWLEDGEMENTS
.
+
-15 The writer would like to express his sincere gratitude
-35 to The Development Fund of the Swedish Construc-
b, Dynamic force in sheet pile webb 731 near the toe tion Industry (SBUF) together with Stabilator AB
for funding the ongoing research project.
igure 9a-b. Typical time histories of generated forces.
7 REFERENCES
4.3 Force versus displacement
Green, J., Nilsson, C-O., (2000),Drivbarhets- och omgivning-
Further insight into the penetration mechanisms of spiverkansstudier av spontdrivning med vibroutrustning.,
sheet piles during vibratory driving can be obtained MSc. Thesis, Dept. of Civil and Environmental Engineer-
by observing the load transfer relations (RI-Uand ing, Royal Institute of Technology, (in press).
RA-ucurves. Holeyman, Alain E., Legrand. Christian., Van Rompaey, Dirk.,
The load transfer relation of the dynamic toe re- (1996), A method to predict the driveability of vibratory
sistance in the webb at 3.1 m penetration depth has driven piles., Proc. of 5th Int. Conf. On the application of
Stress-Wave theory to piles., September 11-13, Orlando
been graphed in Figure 10. FL., pp. 1101-1112.
Dynamic force in sheet pile webb B1 near the toe Jamilkowski, M., Ladd, C.C., Germaine, J.T., and Lancelotta,
43
R., (1985), New developments in field and laboratory
3
testing of soils., Theme Lecture, Illh Int. Conf. on Soil
Mech. & Found. Engng, San Francisco.
2 Moulai-Khatir, Reda., O'Neill, Michael W., Vipulanandan, C.,
5 5
(1994), Program VPDA Wave Equation Analysis for Vi-
bratory Driving of Piles., Report to The U.S.A. Army
.
Y

-15
-35
0,11 0.15 0.15
1'
igure 10. Load transfer curves of sheet pile B1 near the toe.
The negative toe resistance noticed in Figures 9b-
10, is assumed to be related to inertia effects in
combination with a small portion of shaft resistance
on the 600 mm long part of the sheet pile below the
position of the toe sensors.
5 CONCLUSIONS
Based on this study and the range of variables in-
vestigated, the following primary conclusions can be
advanced:
1.)The presented instrumentation system has proven
to successfully register the dynamic soil resistance.
2.) The position of applied sensors made it possible
to document field related properties of vibro-
capacity and penetrative resistance.
Corps of Engineers Waterways Experimental Station.,
UHCE 94-1, Univ. of Houston, Texas, 187 pp.
O'Neill Michael W., Vipulanandan, C., (1989), Laboratory
0,16 0,16 0.17
Iml evaluation of piles installed with vibratory drivers. National
Cooperative Highway Research Program, Report No. 316,
National Research Council, Washington, DC. Vol. 1. pp. 1-
51. ISBN 0-309-04613-0.
Rao, Pramod M., (1993), Effect of pile geometry and soil satu-
ration in the behavior of non displacement piles installed by
vibration., M.Sc. Thesis, Dept. of Civil and Environmental
Engineering, University of Houston, Texas.
Rodger, A.A., Littlejohn, G.S., (1980), A study of uibrarory
driving in granular soils. Geotechnique, Vol. 30, No. 3, pp.
269-293.
Viking, K., (1997), Vibratory driven piles and sheet piles -a lit-
erature survey., Report 3035, Dept. of Civil and Environ-
mental Engineering, Royal Institute of Technology.
Viking, K., (1998), Driveability studies of vibro-driven model
piles in non-cohesive soils -laboratory simulations., Licen-
tiate Thesis 2029, Dept. of Civil and Environmental Engi-
neering, Royal Institute of Technology.
Viking, K., Green, J., Nilsson, C-O., (2000), Uppmatta
markvibrationer av vibroinstallerad spont., 13th Nordic
Geotechnical Conf. on Geotechnics and Sustainable Devel-
opment., Helsinki, June 5 - 7, 2000 (in press).

540
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 0 2000 Balkema, Rotterdam, ISBN 90 5809 1503

Accelerations of a driven pile and the surrounding soil

Edward L. Hajduk & Samuel G. Paikowsky

Geotechnicul Engiiieering Research Laboratop, Deparfment of Civil und Environmental Engineering, University
of Massachusetts, Lowell, Mass., USA
Paul Holscher & Frans B. J. Barends
GeoDelft, Netherlunds

ABSTRACT The traditional mathematical formulation of pile driving (e g Smith, 1960) assumes that the
soil surrounding the penetrating obiect remains stationary while one-dimensional wave propagation takes place
in the pile Paikowsky and Chernauskas (1996) examined large data sets of driven piles and suggested that the
soil's inertia greatly affects the actual physical phenomenon controlling pile resistance during driving and hence
the accuracy of the dynamic methods Due to the fact that soil inertia is not accounted for by most traditional
pile driving models, a new, practical model needs to be developed that accounts for this factor in a routine
manner
The development of a reliable pile-driving model incorporating soil inertia requires an understanding of
the relationship between pile penetration and soil motion Holscher (1995) investigated these relationships in a
laboratory and field study To enhance the understanding of the problem, the presented research examines the
measurements of pile and soil accelerations during the installation of a closed ended steel pipe pile at a bridge
reconstruction site in Newbury, Massachusetts The steel pipe pile, part of a test pile cluster investigating time
dependent pile capacity in the Boston Area, was instruniented with accelerometers and strain gages at the pile
top, middle, and tip Ground accelerometers, piezometers, and total pressure cells were installed in a silty sand
layer about 19 5m from the ground surface at various distances from the pile
The layout and installation of the pile and soil instrumentation is provided Measurements showing the
effect of pile installation as the pile tip passes by the soil instrumentation are shown Analysis of the pile and
soil instrumentation measurements showed (i) such measurements are possible and provide insight into the be-
havior of the soil during pile penetration and (ii) the majority of soil disturbance occurred at or near the pile tip

To develop such a model, actual soil acceleration

I . BACKGROUND
Typical pile driving modeling used in current engi-
neering practice is based on Smith (1960) and does
not take into account movement of the soil during
the installation process The soil-pile interaction is
merely being presented as stationary forces activated
through the pile's displacement
Previous research that examined large data sets of
driven piles has suggested that soil inertia greatly af-
fects the physical phenomenon that controls the dy-
namic resistance of the pile. The analysis of these
data sets indicated that under high soil inertia condi-
tions (i e low driving resistance of large displace-
ment piles), the performance of wave matching tech-
niques is extremely poor. For updated data analysis.
see Paikowsky and Stenersen (2000) The need for a
reliable, physically based soil inertia model is there-
fore evident (Paikowsky and Chernauskas, 1996,).
measurements need to be obtained to provide a
framework for understanding the soil motion mecha-
nism. Such measurements can also be used to cali-
brate the models once they have been developed.
As part of a joint research effort into this field, the
University of Massachusetts Lowell's Geotechnical
Engineering Research Laboratory (UML) and Geo-
Delft (Delft) of Delft, the Netherlands recorded pile
and soil accelerations during impact driving of a
closed ended steel pipe pile, designated Test Pile #1.
Complimentary measurements of pile forces, pore
water pressures, and total soil pressures were also
recorded. These measurements were taken at the
UML Geotechnical Engineering Research Site lo-
cated in Newbury, Massachusetts. This paper de-
scribes the site, the equipment and its installation,
and presents measurements of the pile and soil in-
strumentation as the pile tip passes near the soil in-
strumentation.

54 I
2. SITE AND INSTRUMENTATION Prior to the installation of Test Pile #I, a ground
piezometer field, two test piles, and twelve reaction
2. I The Newhirry Research Sire piles had been installed at the Newbury Site.
Paikowsky and Hajduk (1999) describe in detail the
The UML Geotechnical Engineering Research layout of the site and the various test pile cluster
(Newbury) Site was located at a bridge reconstruc- components. Figure 2 shows the layout of the New-
tion site along US Route 1 on the New- bury Site prior to the installation of Test Pile ff 1.
bury/Newburyport Massachusetts border. An inten-
sive field and laboratory study of the site was
2.2 Test Pile ff I
conducted to determine the characteristics and engi-
neering parameters of the subsurface (Paikowsky and Test Pile #I was a 32.4cm (12.75in) diameter x
Chen, 1998). 1.3cm (0.5in) wall thickness x 31.4m (103A) long
The Newbury Site soil stratigraphy is typical of the closed ended steel pipe end-bearing pile. The pile
conditions found in the Boston area. The general soil was heavily instrumented for monitoring soil-
profile at the pile testing location (from ground sur- structure interaction over a lengthy time period. In-
face downward) consists of the soil strata listed in strumentation installed within the pile relevant to the
Table 1. The soil profile is shown in Figure 1. pile-soil acceleration research included piezo-
resistive accelerometers and electrical resistance
strain gages installed near the pile midpoint and pile
Table 1. Newbury site soil profile. tip. These instruments are listed in Table 2 and are
Depth Soil Layer shown relative to the soil profile in Figure 1.
(m) The piezo-resistive accelerometers were standard
0-2.4 Granular Fill (Cased)
2.4-2.7 Organic silt and Peat dynamic gages assembled and distributed by Pile Dy-
2.1-5.4 O.C. Clay namics, Inc. (PDI) of Cleveland, Ohio. Only one in-
5.4-1 1.5 Soft N.C. Clay terior accelerometer was installed at each pile loca-
11.5-16.4 N.C. Clay
16.4-19.3 Interbedded Silt. Sand. and Clay tion. The electrical resistance strain gages were
19.3-21.6 Silty Sand Model LWK-06-W250B-350, manufactured by
2 1.6-23.9 Interbedded Silt. Sand. and Clay Measurements Group, Raleigh, North Carolina. Two
23.9-26.3 Fine to Mehuni Sand strain gages were installed at each pile location.
26.3-30.5 Till
N.5+ Bedrock

Figure 2 . Newbury site layout.

Table 2. Test Pile #1 Instrumentation.

Gage Type Gage Depth from
Designation Pile Top
Piezo-resistive 1-8-APCB-A 14.5m
Accelerometers (APCB) 1-16-APCB-A 30.2m
Electrical Resistance 1-8-ERSG-A&B 14.5m
Strain Gages (ERSG) 1-S-ERSG-A&B 30.2m

542
 A Pile Driving Analyzer (PDA), also manufactured corded the gages at a frequency of 1 reading /30 sec-
by PDI, recorded measurements of the interior accel- onds throughout the pile installation Paikowsky and
erometers. Each accelerometer was recorded at a Hajduk (1999) provide details of the ground pie-
frequency of IOkHz for a span of 0 1023 seconds A zometers and their installation
CR9000 measurement and control system, manufac- The ground accelerations in the deep soil were
tured by Campbell Scientific of Logan, Utah, re- measured using s e r o accelerometers manufactured
corded the interior strain gages at a frequency of by Sundstrand Data Control Inc of Redmond,
lOkHz over a time span of 05 seconds Washington. These instruments have a resonance
In addition to the instrumentation installed within frequency of 600Hz, and are therefore filtered by an
Test Pile #1, dynamic transducers consisting of analog 3'd order filter with a -3dB point at 150 Hz
piezo-electric accelerometers and strain gages were The amplification from 0 to lOOHz is 1 Each
attached to the pile top during driving These gages ground accelerometer (GA) instrument consisted of a
were manufactured by Pile Dynamics, Incorporated cluster of three accelerometers perpendicular to each
of Cleveland, Ohio and were recorded at the same other placed in one special cone which allowed for
frequency as the interior accelerometers with the measurements along three orthogonal axes
same Pile Driving Analyzer The total pressure cells (TPC) in the soil were
Test Pile # I was installed with a Delmag D30-32 composed of PR-transducers manufactured by Druck
single ended diesel hammer on May 29, 1997 Limited of Leicester, England The signals are ampli-
fied by using HBM amplifiers
Measurements of the ground accelerometers and
2.3 Iiiili'imiet7fntroi~ .for Motirtoi.iiig the SOI/'.I total pressure cells were stored after A/D conversion
Re rp0t7se within a portable computer Data from each GA and
A variety of transducers were installed around the TPC was sampled at a frequency of I M z Total sig-
planned location of TP#l to record pore pressure, nal length for each gage was 2 4 sec, including the
total pressure, and ground accelerations within the pre-trigger time of 0 5 sec
soil To allow comparison with previous soil accel- Due to the crowded nature of the test site, the
eration measurements conducted in cohesionless soil ground accelerometers and total pressure cells were
by Delft (Holscher, 1995), it was decided to install installed outside of the reaction piles to (i) prevent
the ground accelerometers and total pressure cells disturbance to the installed instrumentation and (ii)
within the silty sand layer (depth 19 3 - 21 6m) allow room for a drill rig to install the equipment
Three vibrating wire ground piezometers were al- Figure 2 shows the locations of the casings and gages
ready installed within this layer as part of the test pile relative to the site layout
cluster at various distances from the planned location Each ground accelerometer set and total pressure
of Test Pile # I A summary of the soil instrumenta- cell was installed into the soil at an approximate an-
tion around Test Pile #1 in the silty sand layer is gle of 10 degrees from the vertical Installation con-
listed in Table 3 sisted of placing 1Ocm (4in) OD by 13 7m (45fi) long
steel casing through the clay layer using standard
boring techniques Accurate location of the gages
Table 3 Soil iiistruiiieiitahon v ithiii the silt! sand la!er was achieved using inclinometers to measure vertical
Gage Gilge Distance Elev. Depth and horizontal deviation of the casing relative to the
T] pe Name fromTP#l (m) (m) planned location of Test Pile # I and detailed survey-
(m) r/R ing measurements The gages were then pushed into
Vibrating Glid PZ-8 0.61 3 7 5 -1566 21.24
place at the end of standard AW drill rods To insure
Wirc Gild PZ-9 2.19 13.53 -14.0s 19.66
Piciomctcr Grid PZ- 1 0 2.22 1.3.68 -14.75 20.33
that the rods were centered within the casing, 3 metal
Ground GA-1 1 40 8.64 -13.92 19.50 guides were welded to the rods at 120' intervals
Acccler- GA-2 2 30 14 17 -13 76 1934 around the rod diameter at 1 5ni (Sft) increments A
oiiietcrs GA-3 2 93 1809 -13 76 I934 typical installation cross-section is shown in Figure 3
TPC- 1 124 764 -1407 1965 The interior of each casing was sealed with bento-
Total Prcs- 11 75 -1398 1956 nite slurry to prevent the boreholes from acting as
TPC-2 1 90
siirc Cells -1388 I946 drain holes, which would affect adjacent pore pres-
TPC-3 2 97 1832
KC! sure dissipation measurements
I = Distaiicc froin the pile centcr R = Pile radius Final elevations and distances from TP#l for the
various soil gages were determined based on the
measurements made during installation and the final
The ground piezometers were Model 4500 vibrat- surveyed position of TP#l after driving and are
ing wire pressure transducers manufactured by summarized in Table 3
Geokon Incorporated of Lebanon, New Hampshire
Each piezometer was connected to Micro-10 Data-
loggers, also manufactured by Geokon. which re-

543

Figure 3 TJ pica1 ground accelerometer (GA) arid ~olalpres-
w e cell (TPC) iiistalla~ioncross-~ectioii
3 . MEAS U REMENTS
Analysis of the pile and soil gage data focused on
rneasurements taken during passage of the pile tip
approximately 3-5m above and below the soil in-
strumentation locations This corresponded to a
range of soil starting at the interbedded layer (depth
16 3 lm) and ending within the fine to medium sand
layer (depth 25 58111)
4. PILE RESPONSE
Problems experienced during Test Pile # 1 installation
limited pile acceleration measurements at both the
pile iniddle (gage 1 -8-APCB-A) and pile tip (gage 1 -
16-APCB-A). Gage 1 - 16-APCB-A experienced
equipment failure at pile tip depth of 18.90m, directly
above the silty sand layer Data acquisition system
problems with the PDA resulted in a gap in pile mid-
dle acceleration measurements between depths 20.42
and 23.48m. Measurernetits of the dynamic gages at
the pile top and the interior electrical resistance strain
gages were continuous throughout the driving of
Test Pile,# 1 near the soil instnmentation.
Permanent set for each blow was determined
through a detailed analysis of a video taken during
the installation of Test Pile # 1 , Figure 4 shows these
measurements converted into driving resistance for.
pile penetration around the silty sand layer.
544
Typical measurements of the pile forces and accel-
erations and the corresponding velocities and dis-
placements with time for a single blow are shown in
Figure 5. Figure 5 shows the data for Blow 107
(pile tip depth of 18.901-11) while the pile driving re-
sistance was 1 .Oblows per I Ocni (B 1 Ocm).
Figure 5 shows the pile tip experiencing significant
acceleration with corresponding high velocity and
displacement during driving near the soil instr-umen-
tation. The measurements of high acceleration and
displacement at the pile tip, coupled with the low
driving resistance of the pile through the siltp sand
layer as shown in Figwe 4, make this pile ideal for
studying the efTects of soil inertia.
Examination of the pile measurements during
penetration at and near the silty sand layer showed a
consistent drop between the maximum pile top, pile
middle, and pile tip accelerations. Figure 6 s h o w
the maximuin pile middle and tip accelerations nor-
malized to maximum pile top acceleration with
depth. The ratio between pile middle and pile top
*
accelerations was 0.85 0.07, while the ratio be-
tween pile tip and pile top was 0.58 7t 0.05 (mean Sr 1
s.d.). This consistent range suggests that a reason-
able estimation of pile tip acceleration can be made
fi-on1the measured values at the pile top. The ratios
presented in Figure 6 are not necessarily representa-
tive of standard piles.
5 . SOIL RESPONSE
5.1 Soil Veiocip Meastiremenfs
Acceleration measurements taken by the ground ac-
celerometers (GA) were processed in the following
manner:
o Signals fiom the three individual accelerometers
within each GA were rotated to an axis system
with a pure vertical axis. This produced trans-
formed axes labeled as vertical (parallel to the
pile), horizontal (perpendicular to the vertical
axis of the pile), and tangential (perpendicular to
the plane through the pile and transducer loca-
tion). The transformed axes are shown in Figure
7. The known slope of the transducers, measured
during gage installation, allows for this transfor-
mation.
e The signals are rotated around the vertical axis to
reduce tangential accelerations. This rotation is
kept independent of pile depth.
e The signals were inteyrated with respect to time.
Figurc 6. Nomializcd maximum pilc accclcratioii measurc-
inents with depth.
Figure 5. Tl;pical Tcst Pile #1 measurements (blow 107. dcpth
18.9111).
545
'i
--
I .-_- I
I Vertical
I -I Horizontal
Axis
1 -
PILE
Figure 7 Transfomied ground accelerometer axes
A sign convention was established to identify ab-
solute vertical, horizontal and tangential measure-
ments. The sign convention is shown in Figure 7 .
Insight into the behavior of the soil around the pile
tip during driving was achieved by examining the
velocity vs. time records for individual blows as the
pile tip passes the ground accelerometers. Figure 8
presents vertical velocities vs. time for several se-
lected blows. The tinie for each blow in Figure 8
was adjusted to allow for the start of the velocity
changes to be located at approximately the same
time
The data of Figure 8 shows that the location of
GAI is outside the major deformation of soil rnove-
ment (plastic flow) and therefore the measurements
are in the shape of a disturbing wave The upward
(neyative) velocities are larger as the pile tip is above
the point of measurement. As the pile tip approaches
and passes the gage elevation, the absolute peak ve-
locities changes relation with regards t o the sign,
with the upward velocities being larger, indicating
the soil is moving primarily downward (positive).
Figure 9 was prepared based on the data shown in
Figure S. Vertical velocit_v with time measurements for se-
lected blows of CA1
Figure 9. Maximum and minimum measured velocities with
depth for GA1.
546
Figure 8. The maximum and minimum vertical, hori-
zontal, and tangential velocities are presented as
fUnCtiOnS Of the Vertical distance from CA1 to the
pile's tip and depth. Figure 10 presents the meas-
urement results from GA2 in an identical format. In-
spection of Figure 9 and 10 reveals that the maxi-
mum and minimum tangential components remain
nearly constant over the examined region, which
shows that the applied rotation of the axes was sat-
isfactory.
Inspection of the variation of the vertical and hori-
zontal velocities of GA1 and GA2 with the pile
penetration needs to consider the location of the ac-
celerometer relative to the pile's tip. As GA1 and
GA2 are 8.64 and 14.17 radii from the pile's skin, it
seems that the start of the major influence can be
identified when the pile tip corner is approximately
9.4 radii (1.52m) diagonally from GA1. At this point
the vertical component of the velocity is 8.80 m d s e c
downward and the horizontal component is 3 .80
m d s e c , approximately equal to the tangential veloc-
ity. The velocities at GA2 are insignificant at that
point, suggesting that the outside boundary of ,the
moving soil in a plane below or normal to the pile tip,
is between 9 to 14 radii away from the shaft The
peak velocities of GA 1 and GA2 take place when the
pile tip passes by about 1 to 2 radii the horizontal
plane on which the accelerometers are located. At
this point, the maximum soil velocity at GA1 is 41.20
m d s e c (directed downward) and the corresponding
outward horizontal velocity is 24.40 m d s e c .
5.2 A ccwacy of the Soil Acceleration Measurements
Analysis of the soil acceleration measurements
showed that the frequency content observed in the
Newbury testing was much higher than the previous
Figure 10. Maximum and minimum measued velocities with
depth for GA2.
testing conducted in Del&, Holland (Holscher and
Barends, 1996). This observation raised questions
concerning the ability of the ground accelerometers
to accurately record the soil accelerations generated
by the pile penetration. As stated previously, a 3rd
order analog filter was used to attenuate frequencies
greater than 1OOHz in the ground accelerometers. If
the ground acceleration frequencies were greater
than the filter frequency, the recorded accelerations
and subsequent velocity and displacement integra-
tions would be inaccurate.
To determine if the frequencies in the soil were
greater than 1 OOHz, the accelerations in the pile were
examined. Figure I 1 presents the accelerations at the
pile top and tip for blow 93 (vertical distance to GA1
of 1.85m). A Fourier transform analysis was per-
formed on these two acceleration measurements to
determine the frequencies present within the wave-
forms. The results of the Fourier transforms are pre-
sented in Figure 1 1.
The results presented in Figure 11 show that at the
pile top, the acceleration waveform largely consists
of a range of frequencies between 1 to 3000Hz. At
the pile tip the acceleration wave is primarily com-
prised of frequencies between 1 to 600Hz. This sub-
stantial reduction suggests that the soil attenuates the
acceleration wave as it travels down the pile. Al-
though these observations do not reveal if high fie-
quency signals are present within the ground accel-
eration measurements, they suggest that a reduction
Figure 1 1. Typical test pile acceleration measurements and
corresponding Fourier transform analysis.
547
Figure 12. Pore and total pressure measurements \\4th depth
for Gnd PZ-8 and TPC I.
in frequency content would occur within the accel-
eration wave in the soil mass. If these frequency re-
ductions did occur, then the ground acceleration
measurements should have accurately recorded the
behavior of the soil mass due to pile penetration.
5.3 Total Sod and Pore Water Pressure Measure-
ments.
The increase of pore water and total soil pressure
with respect to pile tip penetration for gages Gnd
PZ-8 and TPC- 1 are shown in Figure 12
The data presented in Figure 12 shows that the in-
creases in pore and total pressures in the soil are
caused by the passing of the pile tip AAer the tip
has passed, any excess pore water and total pressures
generated begin to dissipate These measurements
reinforce the observations of the ground acceler-
ometer measurements that the soil surrounding the
pile is most affected between an approximate range
of Im below to 2m above the pile tip (approximately
3 to 6 pile radii)
The ground piezometers measurements were af-
fected by (i) disturbance of the vibrating wire within
the gages caused by pile penetration, which led to
extreme gage readings, (ii) the fast dissipation of ex-
cess pore water pressure in the granular material
relative to the slow measurement frequency of the
vibrating wire gages, and (iii) pauses in pile driving
Extreme gage readings were edited from the data ap-
pear as gaps in Figure 12 Pile driving operations
were stopped at various intervals to remove pile pie-
zometer caps along the pile and account for the mul-
tiple readings at various depths The relatively low
measurement frequency of the ground piezometers,
caused by limitations of the gage type and data ac-
quisition system, did not allow for detailed pore pres-
sure measurements per blow
CONCLUSIONS
Based on the measurements presented in this paper,
the following conclusions are made:
Measurements of pile and soil forces, accelera-
tions, pore and total pressures during pile driving
are possible and provide insight into the behavior
of the soil during pile penetration.
Over the range of pile penetration examined, the
majority of soil disturbance occurred at or near
the pile tip.
Changes in total and pore water pressure around
a driven pile are caused primarily by movement
of the soil around the pile tip.
The accuracy of the ground acceleration meas-
urements and corresponding velocity and dis-
placement integrations at the Newbury Site may
have been affected by the filtering of high fre-
quencies. To avoid this problem, ground accel-
erometers in fbture research into soil accelera-
tions around driven piles should be capable of
recording data at the same range of frequencies
generated within the pile.
Examination of the pile measurements showed
that the frequency content of the acceleration
wave decreased as the wave traveled through the
pile. This attenuation suggests that the frequen-
cies within the soil mass are also attenuated in
relation to those generated within the pile.
Additional, detailed examination of the Newbury
Site measurements is suggested in order to assist in
planning fbture research to develop a reliable, physi-
cally based soil inertia model to incorporate into pile
driving formulations.
ACKNOWLEDGEMENTS
The presented research was a joint venture between
the University of Massachusetts Lowell Geotechcal
Engineering Research Laboratory, Lowell, Massa-
chusetts and GeoDelR, DelR, the Netherlands as an
additional element to an ongoing research project.
Funding for this project was provided by the Massa-
chusetts Highway Department (Test Pile #1 and
ground piezometer field installation and measure-
ments, site support and maintenance) and GeoDelR
(ground accelerometer and total pressure cell instal-
lations, measurements, and removal). The authors
would like to thank the contributions of Fokke
Wijnstra, Tom van der Poel, and Dr. Frans Barends
of GeoDelR, Nabil Hourani and John Pettis of the
Massachusetts Highway Department, Car1 Ealy and
Al DiMillio of the Federal Highway Administration,
and Gary Howe of the University of Massachusetts
Lowell.
548
REFERENCES
Holscher, P. 1995. Dynamical response of satiwated and drv
soils. Delft: Delft University Press.
Holscher, P. and Barends, F.B.J. 1996. In-situ Measurement
of Soil-motion near the Toe of a dynamically Loaded Pile.
In F.C. Townsend, M. Hussein. & M.C. McVay (4).Pro-
ceedings3 ShInternationaI Coq/erence 011 tlie Applicatioii
of Stress-Wave Theory to Piles, 11-13 September, 1996:
26-36. Orlando, Florida.
Paikowsky. S.G. and Chen. Y.L. 1998. Field and Laboratory
Testing of the Physical Characteristics of the Subsurface at
the Newbury Bridge Site. Research Report subirtitted to
the Massachusetts Highwav Department, January 1998.
Boston, Massachusetts.
Paikowsky, S.G. and Chernauskas. L. 1996. Soil Inertia and
Use of Pseudo Viscous Damping Parameters. In F.C.
Townsend, M. Hussein, & M.C. McVay (ed), Proceedings,
5IhInternational Colf i r m c e on the Application qf Stress-
Wave Theorv to Piles, 11-13 September, 1996: 203-216.
Orlando. Florida.
Paikowsky, S.G. and Hajduk. E.L. 1999. Design and Con-
struction of an Instrumented Test Pile Cluster. Research
Report submitted to the A fassachusetis Highwqv Depart-
ment, September 1999. Boston, Massachusetts.
Paikowsky, S.G. and Stenersen. K. 2000. The performance of
the dynamic methods, their controlling parameters and
AASHTO 2001 Deep Foundation Specifications. Pro-
ceeditigs, dhInternational Cotfirence on the .Ipplication
of Stress-Wave Theory to Piles, 11-13 September, 2000.
S I i Pado City. Brazil.
Smith. E. 1960. Pile Driving Analysis by the Wave Equation.
Journal of Soil Mechanics and Foundations, Aniericaii So-
ciety of Civil Engineers: 3 5-6 1. A irgust 1960.
9 Statnamic and other similar techniques
This Page Intentionally Left Blank
 Application ofStress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Keynote lecture: Statnamic the engineering of art

€? Middendorp
TNO Psofouiid, Rijswijk, Netherlands

ABSTRACT: In present standard engineering practice there seems to be a big gap between engineering
and art. It is the experience of the author that engineering and art can exist as an excellent combination. This
will be illustrated by examples from history and the author’s personal experience. One example will be
treated extensively: the continuous development of the Statnamic load testing method as a marvelous combi-
nation of engineering and art. The start of the Statnamic concept is described as an interaction between an art-
ist and engineers together with developments on the theoretical approaches and technical applications. Further
the present Statnamic state of the art will be discussed briefly.

1 INTRODUCTION ham, the inventor of Statnamic and the nowadays

President of the Berminghammer company.
Art was of minor interest to the author when he was
starting his engineering study at the Technical Uni- During his career the author was impressed by the
versity. Within a short time his interest got a strong many creative solutions of engineers all over the
impulse when he met his present wife, a painter and
artist, at that time studying at the Royal Academy of
Art in The Hague. After assisting here with some
projects he experienced that the “logical” thinking of
engineers is in no way superior to the “associative
and intuitive” thinking of artist in finding practical
solutions, but that both are complimentary and when
combined into an artist-engineer as a person or a
team, can result in marvels.

In the 15‘hcentury the “artist-engineer”, was a so-

cially prominent and respected figure, commissioned
by powerful and wealthy patrons, well paid and of-
ten regarded as one of the brightest ornaments in
sovereign courts. The most famous example of
course is Leonardo da Vinci (P.Galluzi, 1996).

Because of cultural changes and specialization a

gap has been generated between engineering and arts
and most engineers are not aware nowadays of their
artist-engineer forefathers.

Still a strong interest of artists for engineering can

be observed in modern art, for example Panamer-
enko (1996). The artist-engineers are still among us
and it was the privilege of the author to cooperate Figure 1. Leonardo da Vinci. Automatic
for long period with one of them: Patrick Beming- file-making machine

551
world. Many engineers are not aware that their crea-
tive solutions can be considered art and that they act
Figure 2. Early sketch of a Statnamic piston and
cylinder arrangement by Patrick Bermingham
as “engineer-artist”. In this paper the author wants to
illustrate his view by using the Statnamic develop-
ment as an example in which one “artist-engineer’’
and many “engineer-artist” contributed with creative
ideas and solutions.
The Statnamic development will be described by
introducing the start of Statnamic together with
some milestones, theoretical approaches and techni-
cal applications. Further the present research and
developments will be mentioned and problems that
still have to be solved.
2 THE STATNAMIC CONCEPT
Patrick Bermingham (1998) got the first idea about
the Statnamic concept in Hamilton in 1985 while
watching a static load test with kentledge, when he
first thought about utilizing the inertia of the
kentledge (Fig. 2).
According to Fellenius (1995) the idea for the Stat-
namic concept was born in 1987 when he asked Pat-
rick Bermingham to design a drop hammer for im-
pacting a pile to perform dynamic load tests. The
552
idea of Fellenius at that time was that dynamic pile
testing would become independent of the piling con-
tractor, the pile driving rig, and in many instances
contractor’s unionized labor, as well make him and
others free to perform dynamic load testing after a
good and long set up time. This idea was not original
because several such devices were already around,
but Fellenius just needed a local practical tool.
At that time Bermingham had just finished a pro-
fessional education as a sculpturer in London and
worked also for the Bermingham pile driving and
hammer manufacturing company. From his child-
hood on Patrick was interested in both engineering
and art and supplied, for example, several ideas for
improvement to the Berminghammer pile driving
hammers.
It was not such a strange idea that the Berming-
hammer Company chose Patrick to come up with a
design for a drop hammer. Patrick contacted Fel-
lenius and suggested a pile loading system design
from a fully different viewpoint compared to stan-
dard engineers.
“Why do engineers want to drop the weight, why
do they not send it up into the air?”
Intuitively he converted Newton’s Law from
Force = Mass times Acceleration
To
Mass times Acceleration = Force (Load)
This concept needed a few months to evolve and
Figure 3. First Statnamic device with acceler-
ometer and early catch mechanism.
Bermingham made a first prototype (Figure 3) and
started experiments by shooting masses upwards in
Hamilton, Ontario in April 1988. He determined the
feasibility of accelerating a mass upwards from the
top of the foundation rather than dropping a mass
onto the foundation
Figure 4. First Statnamic trial tests.
Bermingham also tried out several concepts of
catching the reaction mass when falling back from
launching and also here creative ideas from the “art-
Figure 5 . Successive stages of Statnamic
Testing.
553
ist-engineer” can be observed, especially the gravel
catching mechanism based on the reliable gravity of
the earth (Fig.4). With the gravel mechanism a
gravel container is placed around the reaction mass
and the space between gravel container and reaction
mass is filled with gravel. During testing four suc-
cessive stages can be distinguished. In stage 1 the
Statnamic device is ready for launching. In stage 2
the reaction mass is launched upwards by high-
pressure gases. During this stage the pile is loaded
and a Statnamic test performed. Because of the mo-
mentum the reaction mass will remain moving up-
ward in stage 3 and the gravel will flow under the
reaction mass and over the pile head because of
gravity. In stage 4 the reaction mass will fall back
and will be caught by the gravel inside the container
and the impact load will not be transferred to the pile
head but to the subsoil. This creative, simple and
elegant principle is still applied as one of the meth-
ods in absorbing the energy from the falling reaction
mass.
Bermingham presented his results and ideas to
several parties and also to the author at the OTC
(Offshore Technology Conference) at Houston in
1988. Based on the combination of his engineering
background and experience with ideas of artists, the
author immediately recognized the beauty and
power of Bermingham’s idea for pile testing applica-
tions.
Berminghammer and TNO agreed to start a joint
development and decided to do the first prototype
testing immediately after the Third Stress Wave
Conference at Ottawa in 1988. With the help of
Fokke Reiding and Matthew Janes they realized that
the long duration feature of the load allowed a fully
different approach in instrumentation and analysis
compared to dynamic load testing.
It was decided to base the load measurement on a
calibrated load cell, to make the measurement inde-
pendent from pile material properties and to measure
displacement directly by the use of an electronic
theodolite. The electronic theodolite was a rather
expensive instrument and a new tool for measuring
displacement was developed based on a laser and a
laser sensor, which is still in use. So the basis of
measurements became load-time signals and dis-
placement-time signals similar to static load testing.
3 FIRST DEVELOPMENTS
In May of 1988 the first model tests where per-
formed with instrumentation provided by T’NO.
These first two days of testing confirmed the ability
of the very small Statnamic device to produce loads
of up to 5 tons with duration of up to 30 ms. From
this point onwards the direction of Statnamic was
upwards. A second model was built and sent to
TNO in Holland where the instrumentation would be
developed. In the laboratory at TNO, Statnamic
tests were performed using a calibrated load cell and
a new laser measuring system developed specifically
for the Statnamic test. Both measuring systems
worked very well the first day and they have re-
mained virtually unchanged to this date.
The next step was to build a Statnamic tester,
which incorporated the instrumentation and was
large enough to test a real pile in the ground. A 0.6
MN tester was built, which was one tenth of full
scale, but still able to test small piles driven into real
soil. This load-testing device was first used to test
piles in the Berminghammer yard in Hamilton,
McMaster University, and Ashbridges Bay. Since
that time it has performed tests in Europe, Japan, and
the United States. The primary objective of this
equipment was to prove the durability of the instru-
mentation in all weather conditions, and to prove the
practicality of the equipment in the field. This
equipment was also used to make the first compari-
sons between conventional static load tests and the
new load test method. The 0.6 MN device proved
that Statnamic testing could be performed in all
types of adverse weather including rain and snow. It
Figure 6. Patrick Bermingharn launching a
0.6MN device
554
also proved the simplicity and practicality of the sys-
tem in the field and the first load test comparisons
proved an unexpectedly close agreement with con-
ventional static load tests. The success of this first
prototype enabled Berminghammer to manufacture
of a full-scale 5MN tester.
Figure 7. Set up of a 5MN Statnamic device with
gravel catch mechanism
Statnamic was first called Inertial load testing
(Bermingham, P., et all., 1989. The author gave the
method its present name Statnamic, realizing that
the method was positioned between Static load test-
From the very beginning Statnamic was an inter-
national development rather than a regional or na-
tional development. Testing of driven and cast in-
situ piles was carried out in Canada, Holland,
Germany and the United States during the first two
years.
At this time all of the testing was conducted with
the aim of gaining a better understanding of the be-
havior of piles subjected to very quick loading cy-
cles. Statnamic and static load tests were conducted
side by side as well as on the same pile in as many
different soil types as possible. Every effort was
made to collect as much data as possible and to
avoid making predictions about the static behavior
of a foundation until we could collect a wide range
of test results. Today many companies and universi-
ties are still collecting and expanding this worldwide
database.
The first two years of research revealed a great
deal about pile behavior when subjected to a Stat-
namic load of 120ms duration. It was observed that
in the elastic range there was a very good agreement
between static load deflection and Statnamic load
deflection, it was observed that in very soft soils and
clays it was possible to apply a much larger load
than a static load prior to plunging the pile. In stiff
non-cohesive soils and rock sockets it was observed
that there was very close agreement between Stat-
namic load test results and static load tests per-
formed along side. It was also observed that the se-
quence of loading a foundation had a great effect on
the perceived similarity of test results and this had to
be taken into account. It was also observed that dur-
ing a typical Statnamic test the pile would reach
maximum displacement at some time after peak
force had been applied. In other words the pile
would continue to move downwards while the ap-
plied load at the pile top was decreasing. At the
point where the pile reached maximum displacement
the velocity of the pile was zero and then the pile
would begin to rebound as the load was further de-
creased. This observation which was present in
nearly all test results except very stiff piles on rock,
lead to the development of the Unloading Point
Method (UPM) by the author (P. Middendorp,
1992).
4 METHODS OF ANALYSIS
From the very start of development there has
been a determined effort to make Statnamic a means
of measuring rather than a prediction method. This
has meant putting a very strong emphasis on using
accurate measuring equipment and recording equip-
ment. The measured data will then be more reliable
and may then be examined more closely. From the
beginning we have been observing the behavior of
foundations subjected to very rapid loading with a
view to being able to better understand the mecha-
nism of failure during a Statnamic test. In the end it
is hoped that Statnamic testing will stand alone as a
rapid test with a distinct method of analysis, which
will measure the load deflection behavior and de-
termine the factor of safety of the foundation.
Initially no attempt was made to convert the re-
sults of Statnamic load testing into quasi-static load
test results, because they would loose integrity in the
process. What was recognized was that every Stat-
namic test result was unique and that very small dif-
ferences in the relative stiffness of two different
foundations could be measured accurately. Much
like the dynamic resistance of a driven pile, it is very
useful even though there is no direct correlation to
static resistance.
The Statnamic test has been described as applying
a controlled strain while monitoring corresponding
deflection. When a test is performed, a predeter-
mined load is applied and the resulting deflection is
measured.
555
The first approach to analyzing failure was to
look to the displacement curve and to analyze the
rate of change of displacement, or velocity of the
pile. Normalizing the load and plotting load vs. ve-
locity was examined in an effort to pinpoint the load
at which the velocity begins to increase. This only
worked well when the foundation experienced a
plunging failure, and it did not work well when the
pile was in a cohesive material.
Statnamic test results were also evaluated with a
simple 2.5mm offset method, which was analogous
to the Davisson failure criterion but much more con-
servative. All three of these methods of determining
the point of failure were far too subjective to be of
any great value.
In January of 1993, while reviewing the results of
pile 7 at Texas A&M the author noticed that during
the unloading of the test the velocity of the shaft
reached zero at a load, which corresponded closely
to the ultimate static resistance. The foundation be-
gan to rebound as the load was further decreased.
PDA users had observed the significance of the
point of zero velocity in the 1970's and some at-
tempts were made to make use of it. However, dur-
ing pile driving the point of zero velocity at the pile
head does not correspond to zero velocity anywhere
else in the pile unless the pile is very short and rigid.
The author's observation provided both a practi-
cal means of determining a significant point on the
static load displacement curve and also a means of
estimating the damping coefficient directly from the
test results rather than from a soil boring. This
Unloading Point Method (UPM) assumed that the
damping was a constant, which was zero when the
velocity was zero, and that the pile was behaving as
an elastic body, which could be treated as a lumped
Figure 8. Stress wave influences as function of
wave number N,
mass and a spring. Subsequent research has concen-
trated on testing these assumptions and determining
the limit of their validity. The unloading point
method has provided a very simple universal method
of analysis and was first published at the 4ti1Interna-
tional Conference on Stress Waves in 1992. The ba-
sic principles of the method have been presented in
the Appendix.
The UPM method is based on the assumption that
stress wave phenomena can be neglected. The author
studied the validity of the method with the stress
wave program TNOWAVE (1996) by varying the
pile length with constant load duration. To quantify
the stress wave influence he assumed a wave num-
ber constant N,,, = D/L, in which D = c.T and T the
duration of the load, c the stress wave velocity and L
the pile length. In this way it was possible to indi-
cate when stress wave phenomena could be ne-
glected and when they should be taken into account.
A valuable extension to the UPM method is the
"Modified UPM" (M-UPM) by Justason ( 1997). The
method simply involves the averaging of the top and
toe velocity and acceleration for calculating the iner-
tia and damping. The method can be applied to any
length of pile, but becomes more necessary as the
pile becomes longer (low N, numbers). The stan-
dard UPM method assumes that pile top velocity and
pile toe velocity are in the same range. The M-UPM
method is particularly useful when the pile top and
pile toe velocity are not in the same range (elastic
pile, high toe resistance). Averaging the pile top and
pile toe velocities and accelerations yields more ac-
curate inertia and damping forces. The method
yields the best results when used in conjunction with
an embedded toe accelerometer.
Prof. Gray Mullins of the University of South
Florida made an additional improvement to the M-
UPM method, the "Segmental Unloading Point" S-
UPM. This method uses measured strain gage data
to separate the pile into "segments" and perform an
M-IJPM on each segment. The data for each seg-
ment are added together to produce a total "derived
static" load-displacement for the top of the pile. The
S-UPM can be applied to any pile, so long as the
pile has strain gages distributed over the pile shaft.
The first application was the Taipei Financial Center
in Taiwan - 1999.
The S-UPM method is briefly described below.
The Segmental Unloading Point Method extends the
applicability of M-UPM to long piles. All assump-
tions of the Unloading Point Method remain valid.
The Segmental Method assumes each segment of a
pile behaves as a single degree of freedom system.
The method requires embedded strain gauge data. A
556
measure of toe displacement is desirable. All results
are based on measured quantities.
F =E,E,A,
where F; is the measured force at level i, E; is the
measured strain at level i (typically an average of all
strain gages at level i), E, is the calculated (or as-
sumed) elastic modulus at level i, and A; is the calcu-
lated (or assumed) area at level i
where U ; is the calculated displacement at level i,
L; is the length of the pile segment between levels i
and i+l
-
wherev, is the first derivative with respect to time
of the average displacement for the pile segment be-
tween levels i and i+l
__
- dv.
a.- I
I -
dt
-
where a, is the first derivative with respect to
time of J
The Unloading Point method is performed on
each pile segment using the following equation:
F - F - , -5'; -civ, =mja,
where S; is the equivalent static force for the
segment between F, and F,.)
S, represents the friction forces on the each pile
segment, with the exception of the bottom pile seg-
ment, which also has some component of end bear-
ing. in; is the mass of the pile segment between i
and i- 1.
The cumulative derived static force at each level
can be calculated by the following equation:
1=I
where n is the pile level number, and FSTAT,,is the
cumulative derived static force at each level.
 The derived static load displacement curve can be
drawn at each strain gage elevation using FSTAT,~and
U,.
Each of the above variables represents an entire
data set measured over time.
The S-UPM was first used for the Taipei Finan-
cial Center for 80m piles in 1999
5 THE HYDRAULIC CATCH MECHANISM
Figure 9. Statnamic device with hydraulic
catch mechanism (4 MN)
The foundation industry not only wanted to perform
larger Statnamic tests but also more of them and at a
higher frequency. Also in this case Patrick Berming-
ham and design engineers came up with a creative
solution. In 1995, the hydraulic catching mechanism
was built to provide a means of testing, without us-
ing the conventional gravel container, or gravel.
This simple piece of equipment makes it possible to
test up to ten individual piles and to perform multi-
ple load-cycles. The catch mechanism provides the
luxury of multiple load cycles within a matter of
minutes, the ability to inspect the ignition circuit
without disassembly, the benefit of single truck mo-
bilization, and its avoids the environmental problems
with gravel retrieval with testing over water.
557
Hydraulic catching systems eliminate the need for
gravel and gravel structure since the upward moving
reaction masses are caught at the top of their flight
by four hydraulic actuators (or rams). These 3.2m
stroke rams are activated by four low pressure (100
bar) nitrogen accumulators, which store compressed
nitrogen gas over hydraulic oil. As the weight on the
rams is released during a test, the compressed nitro-
gen quickly expands to force hydraulic oil into the
rams causing them to chase the reaction masses to
the apex of their flight. The hydraulic oil is routed
(in series) through one-way valves at the base of
each ram, which restricts reverse flow, and thus the
downward movement of the masses. Each of the
four rams is independent of the others providing re-
dundancy and safety. The masses remain at this
position until the user redirects the additional fluid
in the rams back into the accumulators. At which
time, a subsequent load cycle can be performed.
By transferring the initial weight of the masses to
the rams at the onset of the test it is possible to per-
form Statnamic testing without a pre-load condition.
Additionally, hydraulic catching systems have no
minimum required jumg-height for the silencer-
reaction mass assembly, which is a concern for
gravel catching systems. By removing this restric-
tion, low load tests can be performed with much
greater than 5% reaction mass. Such tests can pro-
duce long duration load pulses greater than 0.5 sec-
onds, thus reducing inertial and damping forces for
large portions of the test.
Although the set-up time for a 4MN gravel or hy-
draulic catching systems is comparable, multiple cy-
cles can be performed in a matter of minutes when
using the latter. Further, the breakdown typically
takes less time. In using gravel-catching systems,
great care is exercised in the preparation of the igni-
tion circuitry. An inadequate igniter connection
could cost a project as much as a day of delay time.
This of little concern when using the hydraulic
catching system due to the ability to raise the entire
stack of reaction masses with the hydraulic rams so
as to access the fuel basket.
A substantial portion of all Statnamic testing costs
stems from the mobilization of equipment. Typi-
cally, a 4 MN test requires two tractor-trailers to
ship the combined weight of the equipment and re-
action masses (27,000 kg total) where only 20,000
kg is permitted per truck in the USA (30,000 kg in
Europe). The device can be equipped with two reac-
tion mass options: (1) an entire set of six concrete-
filled steel masses, which requires two trucks to
ship, or (2) a set of six empty, structurally reinforced
steel cans. The empty cans option allows single
truck mobilization to distant sites with a total
shipped mass of 19,000 kg. Once at the site the cans
can be filled with sand, gravel, water or any combi-
nation to attain the required mass

3MN and 4MN hydraulic catching systems are

now in use in the United Kingdom, the USA and the
Netherlands for 3MN and 4MN systems. In 2000 a
mechanical catch mechanism will be constructed for
a 16 MN device.

6 BATTER PILE TESTING

Drop hammers and dead weight static load tests

are fully dependent on gravity. One of the big ad-
vantages of Statnamic is its independence of gravity Figure 11. Lateral load testing preparations for a
because generating the load it is based on inertia 7.5MN test.
forces. This means that the test can be performed in
any direction: under batter, lateral and even allows Major pioneering developments have been per-
to perform a tension test on a pile. In Figure 9 an ex- formed by Dan Brown (1998) of Auburn University
ample of the application of batter pile testing is pre- USA in the analysis of lateral Statnamic tests.
sented. It is almost impossible to perform static load
test in this over water pile testing situation. The
flight of the reaction mass is guided by a support 8 WATER REACTION MASS TESTING
beam
A most significant development is the use of wa-
ter as a reaction mass when testing piles over-water
7 LATERAL LOAD TESTING or near-water. By being able to mobilize the inertia
of the ocean or a river, very large tests may be per-
Lateral STN testing is becoming popular in the formed with testing equipment weighing only 1 % of
USA and was strongly encouraged by Barry Berko-
witz of the FI-IWA. The first lateral test with a large
device was at the Salt Lake City, Utah Airport in a
research project with Kyle Rollins of Brigham
Young University USA and a 14MN device was
used. Mike Muchard and Don Robertson of Applied
Foundation Testing perfected lateral load testing.
They developed a "jig" for holding the piston and a
"sled" for holding the silencer and masses.

A very significant job was for the Mississippi

DOT in 1998 to simulate a ship impact of 7.5MN by
the company AFT. Loads of up to 10 MN have now
been applied in lateral load testing.

Figure 12, Set up of Statnamic water reac-

Figure 10. Over water Statnamic testing on a tion mass testing
batter pile

558

Figure 13. Water reaction mass containers
the test load. The necessary 5% orlO% reaction mass
would be provided by water confined within a vessel
and submerged below the surface of the water. This
weightless reaction mass makes it possible to per-
form very large tests of longer duration than are
practical today. The use of water reaction will also
make it possible to drive piles underwater with a
tool, which will be virtually weightless.
Figure 14. Water reaction mass testing
559
Figure 15. Artist’s impression of a water re-
action pile-driving tool.
Statnamic testing using water reaction mass was first
done by Berminghammer in 1998 in Hamilton Har-
bour. These tests went to 600kN. The first field-
testing was performed for the Port of Lake Charles
in Louisiana in 1999 by Applied Foundation Testing
assisted by Benninghammer. The loads were up to
5 MN.
9 PILE DRIVING WITH WATER REACTION
Geert Jonker of IHC Foundation Equipment envi-
sioned the idea to extend the use the water reaction
mass into a pile-driving tool. Berminghammer, IHC
and TNO are now working together to build an un-
derwater Statnamic hammer, which will consist of a
large inertial mass, made of water and a Statnamic
tool capable producing multiple loading pulses.
This tool will be used to push an anchor pile into the
seabed and measure its capacity at the same time.
In the coming years, we will see driving small
onshore and offshore piles according to the Stat-
namic principle.
10 EVENTS
To share the experience among pile engineers and
to improve the Statnamic test technology the first In-
ternational Statnamic Seminar was held in Vancou-
ver in 1995 with 25 papers and 54 participants. The
Japanese research group on Rapid Pile Load Test
Methods organized the Second International Stat-
namic Seminar in 1998 with 48 papers and 132 par-
ticipants. The third Statnamic seminar is planned in
the Netherlands in 2002.
In March of 2000, the Japanese Geotechnical So-
ciety (1998) published a standard for “Method for
Rapid Load Test of Single Piles”
 An ASTM standard for Statnamic type pile test-
ing is in progress.
1 1 RESEARCH AND DEVELOPMENTS
Companies and universities have accumulated up
to date more than 700 case histories throughout the
world. The total number of contract Statnamic tests
worldwide has exceeded 1000, with testing now oc-
curring at a frequency of more than one each day.
The largest volume of testing to date has occurred in
Malaysia, with the number of contract tests exceed-
ing 300. Similar numbers of contract tests have now
been performed in the United States. The UK and
Japan are close behind in their numbers of tests.
An important contribution to the development of
Statnamic was supplied by the Japanese Geotechni-
cal Society, which established the research group of
Rapid Pile Load Test Methods in 1993. Professor
Osamu Kusakabe of Tokyo Institute of Technology
chaired the group. A strong promoter and initiator of
Statnamic in Japan was Makoto Tsuzuki of Fugro
Japan. The research group included 30 private insti-
tutes and companies as members. The activities of
the Research Group aimed at cataloging the existing
knowledge about rapid load tests, examining the ba-
sic characteristics and the applicability of the test,
and producing scientific interpretations of the Stat-
namic test results.
The University of South Florida (1998) con-
ducted over 150 Statnamic tests in conjunction with
privately and federally funded test programs. The
tests programs have included: (1) axial load tests on
piles and shafts in sands, clays, or rock-sockets, (2)
lateral load tests on pile groups and shafts, and (3)
plate load tests on sands and full-scale spread foot-
ings on sands and vibro-compacted soils (stone col-
umns).
The application of Statnamic produces excellent
results in stiff andor granular soils, although loading
rate effects have to be taken into account. The influ-
ence of soil viscosity alongside buildup of pore wa-
ter pressure in fine grained soils requires further de-
velopment of analysis tools and experience.
(E.L.Hajuk et. al. 1998).
The soil viscosity shows up in two different
ways:
- Creeping, this means continuing settlements
under constant pile load
- Velocity dependent soil behavior
Creeping phenomena cannot be determined with
Statnamic, dynamic load testing and in many cases
not even with static load testing.
560
The velocity dependent soil behavior can be split
up in soil damping phenomena and strain rate de-
pendency. Soil damping phenomena can be derived
straightforward from a Statnamic test. Strain rate
dependency for fine-grained soils is still subject to
study, for example by the University of Sheffield
UK (A.F.L. Hyde. Et all, 1998). Well-documented
data from pile load test projects is becoming avail-
able to support the insight in strain rate effects
(Holyman, et al., 2000).
12 CONCLUSIONS
The success of Statnamic stems for a significant
part from the concepts and ideas generated by “art-
ist-engineers” and “engineer-artists”.
According to Brand1 (2000): an excellent engi-
neer requires not only a firm theoretical knowledge
but also comprehensive experience as well as engi-
neering feeling and intuition in equal parts. The au-
thor would like to add: the ability to be creative and
think in unconventional ways.
The success of Statnamic can be further explained
by the high degree of international cooperation and
research, which has brought the technology to the
forefront .
The Statnamic community originated from the
stress wave community and the author is sure that
they will remain in close contact. Both have a com-
mon interest in the research of dynamic phenomena
of soils and the development of tools for the load
testing of piles. The incorporation of the ideas of
“artist-engineers” and “engineer artists” will guaran-
tee more marvels in the development of pile testing
applications and other fields of engineering.
13 REFERENCE:
Bermingham, P., Janes, M., ”An innovative ap-
proach to load testing of high capacity piles”, Pro-
ceedings of the International Conference on Piling
and Deep Foundations, London, 1989
Middendorp, P. Bermingham, B Kuiper, Statnamic
load testing of foundation piles. 4th International
Conference on Stress Waves, The Hague, Balkema,
1992
Galluzi, P., Mechanical Marvels, Invention in the
age of Leonardo, ISBN 88-09-20959-1, Instituto e
Museo di Storia della Scienza, Florence, 1996.
Baudson, M., Panamarenko, Paris 1996
Fellenius, B., Welcome from the Chairman, First In-
ternational Statnamic Seminar, Vancouver, 1995

Middendorp, P., Daniels, B., The Influence of Stress

Wave Phenomena during Statnamic Load Testing,
5th International Conference on the Application of
Stress-Wave Theory To Piles Orlando, Florida,
1996

Bermingham, P.D., Statnamic the first ten years,

Proceedings of the Znd International Statnamic
Seminar, Tokyo, 1998

Hajduk, E.L., Paikowsky, S.G., Mullins, G., Lewis

C., Ealy, C.D., Hourani. N.M., The behavior of piles
in clay during Statnamic and different static load test
procedures. Proceedings of the 2"d International
Statnamic Seminar, Tokyo, 1998

Mullins, G., Garbin, E.J., Jr., Statnamic testing:

University of South Florida Research, Proceedings
of the 2"d International Statnamic Seminar, Tokyo,
1998

Brandl, H., Civil and Geotechnical engineering in

society - Ethical aspects and future prospects, Pro-
ceedings of the First International Conference on
Geotechnical Engineering Education and Training,
Sinaia, Romania, 2000

Brown, D.A., Statnamic IateraI load response of two

deep foundations, Proceedings of the 2"d Interna-
tional Statnamic Seminar, Tokyo, 1998

Hyde A.F.L., Anderson W.E., Robinson S.A., Rate

Effects in clay soil and their relevance to Statnamic
pile testing, Proceedings of the Znd International
Statnamic Seminar, Tokyo, 1998

Justason, M. D.; Janes, M. C.; Middendorp, P.; Mul-

lins, A. G. Statnamic load testing using water as re-
action mass, The 6th International Conference on the
application of stress wave theory to piles, Sao Paulo,
Brazil 2000.
Holeyman, A., Maertens, J., Huybrechts, N., Le-
grand, C., Preparation of an international pile dy-
namic prediction event. The 6th International Con-
ference on the application of stress wave theory to
piles, Sao Paulo, Brazil 2000

561
14 APPENDIX

The Unloading Point Method (UPM)

Step I) Determination of static resistance from

unloading point

Assumption:
The long duration Statnamic load F,,, allows modelling of the pile as a
zoncentrated mass (M) and springs

F,,,, = Statnamic force (measured 'and known)

U =displacement (measured and known)

v = du\dt = velocity (known)

a = d'u/dt'= acceleration (known)

F,,,,i = F, + F,
F,, = static resistance (unknown)
F, = C.v = damping force (unknown)
C = damping factor (unknown)
F, = M.a (known)

Eq U i I i bri uiii :
F,,,, = F,,,,I+ FA
= F,
F%!" + F. + Fd
F,, = F,,, - C v - M.a

At maximum displacement (Unloading Point)

v = 0 + U = inax~rnurn,t =tun,,,
F,,,I = F,,,(t,,,,) . aiini = a(tunIAx)

Static resistance F. is known at uulli

F,(t,,,,,) , uUlilis a static point

Step 11) Construction of static load-displacement diagram

Assumption:
The soil is yielding over range Fw,,,,) to FUni
So F, = FU,,i

Over this range the following equation is valid

F, = F,,, - F,,,i - FJ
with F, = C.v
C = (F.,,>- F,,,I- F,) / v
Calculate mean damping factor C,,,,. for above range
Now static resistance Fu can be calculared at all points

E, = F.m - Cmw, .V- Fd

D r w static load-displacement diagram with F,, and U

562
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 1503

ynote lecture: Three-dimensional finite element analysis of statnarnic load

test

Tirawat Boonyatee & Makoto Kimura

Kyoto University,Japan
Feng Zhang
GifLl Universiq,Japan

ABSTRACT In order to stud! the inechaiiisin of pile-soil interaction during statnamic loading. a FEM program is de-
veloped Since it n i l 1 also be applied to the statnamic test of piles in lateral direction. batter piles. group pile. and pile
raft foundation, the anal! sis S J stem is formulated i n three-dimensional space Elasto-plastic model is used for the
ground material Joint element is also used for slip at the pile-soil interface The analysis program uses the direct inte-
gration scheine to solve the elasto-plastic anal!sis in time domain Laboratory tests of single piles are analjzed to iii-
vestigate the validit! of the developed program Firstly. the simulatioiis of static load test are done and compared \i7ith
the nieasurcd data from the tests Then. the simulatioiis of piles uiider statnamic are conducted uiider the same grouiid
coiidition Finall!. thc comparisons betu cen the static and statnamic responses of piles are discussed

1 INTRODUCTION
At present, the primary approach to investigating
pile behavior under statnamic loading is to make a
comparison between the results of full-scale stat-
namic and static load tests. Unfortunately, this cali-
bration method has an inherent problem in its testing
consistency, as was addressed by Amir & Amir
(1995). Under the uncertainty and complexity of the
target ground, the situation whereby any model can
be validated by calibrating statnamic load tests
against corresponding static load tests can be dem-
onstrated as follows:
1 If a comparison is made of the results of tests on
the same pile, the tests have to be done in se-
quence. Therefore, the quality of the later tests
invariably deteriorates due to the induced residual
stresses from the former tests.
2 If a comparison is made between the results of
tests on different piles, significant samples should
be tested in order to achieve some degree of reli-
ability over the ground uncertainty.
lnstead of studying the data from the field tests, the
mechanism of the piles during statnamic loading can
be thoroughly investigated by the numerical analysis
approaches. Applying the fact that the system of in-
terest is axis-symmetry, Matsumoto (1998) applied a
2D-FEM program to analyze the behavior of a single
pile under vertical loading. Although a two-
dimensional analysis is simpler and the solution can
be found faster than with a three dimensional analy-
sis, it cannot be applied to a complex system such as
a pile-raft foundation or batter piles Moreover, it is
obvious for a horizontal loading case that the analy-
sis should be done in three-dimensional space
In order to develop a unified tool that can be ap-
plied for various types of problems, a three-
dimensional finite element analysis program called
DYSC (Dynamic and Static systems analysis Code)
is originally developed No-tension criteria and a
simple elasto-plastic model based on the Drucker-
Prager theory are applied as the yield functions for
the model ground An interface layer element is in-
serted at the interface between the pile and the soil
For the pile body, a linear elastic relation is used
since the applied force is lower than its yielding
point
It is thought that before applying the DYSC to the
simulation of piles, its fidelity should be checked by
a physical model For calibration purposes, a com-
parison of the analytical results to the laboratory test
results is preferred to the results from the field tests
The reasons are due to the complexity of the ground
and the qualities of the geotechnical parameters re-
quired for the analysis
In our previous study, a small-scale statnamic
loading device (3SLD-Mkl) was used to conduct
laboratory tests of piles under vertical loading (Ki-
mura et a1 1998) This device is used to conduct ex-
periments under a controllable environment Conse-
quently, guaranteeing that the series of tests are done
under the same conditions is possible Accompanied
with the data from those experiments, a comparison

563
between the simulation and the experiment is made
and reported in this paper.

2 DETAILS OF ANALYSIS
2.1 Inyzrt yaranieters
To determine the necessary parameters used in the
simulations of the statnamic load test, trial calcula-
tions are made. As shown in Figure 1, an analysis is
conducted in the half area of the pile-soil system.
0 1 2 3 3 5 6
Young's modulus of the pile is back calculated from
the flexural test results. Poisson's ratios for the pile D i sp1acein ent (inin)
and the soil material are assumed to be 0.2 and Figure 2. Load-displacement relations from simulation and
0.333, respectively. A parametric study has been experiment.
made to determine the Young's modulus of the
model ground. From various trial calculations and
comparisons with data from static load tests, a Table 1. Prouerties of materials
Young's modulus of 4.9 MPa and a frictional angle
Sand Pile
of 32 are selected as rational quantities for the
ground materials. The frictional angle of 32 for soil
Density (kg/m3) 1467 2150
gives the calculated ultimate load at more or less the
same level as the measured data. A Young's modulus
Young's modulus (MPa) 4.90 5.00s103
of 4.9 MPa controls the shape of the load-
displacement of the pile before failure. Using these
values, a satisfactory approximation for pile re- Poisson's ratio 0.333 0.200
sponse can be obtained. The load-displacement plot
from calculations and measured data is shown in Friction angle ( ) 32
Figure 2.
Dilatmcy angle () 9

The properties of each material determined from an

inverse analysis are summarized in Table 1.

2.2 Pile-soil inteiface model

The Mohr-Coulomb yield criterion is used for the
interaction between the pile and the soil. It states
that failure will take place if the magnitude of shear
stress (z) on the failure plane is equal to the value
given by the following relationship:

in which 1 1 denotes the absolute value, 0, is the

normal stress on the failure plane, and $0 and CO are
material constants for the pile-soil interface. In this
study, adhesion, or CO, is assumed to be zero. The
frictional angle of the pile-soil interface is assumed
to be 0.9 times the frictional angle of sand. Equation
(1) can be written in the form of yield fbnction F as

F = jzl- CT, tan$o

Figure 1. Analysis mesh.

564
If the material is sheared to the yield surface and the
associated flow rule is adopted, the rates of plastic
normal strain d&; and shear strain 7' are given by
which implies
(4)
Increments in shear displacement along the plane are
accompanied by increments in normal displacement.
The dilation of the shear plane will go unbound un-
der yielding. To avoid this unfavorable behavior, the
non-associated flow rule is adopted for the pile-soil
interface. By introducing dilatancy angle y, the
plastic potential function can be written as
In this study, no dilation is assumed for the pile-soil
interface, i.e., w-+ 0.
3 SIMUL,ATION ON STATNAMIC LOAD TEST
After necessary parameters were back calculated
from the previous section, a dynamic FEM analysis
of a pile under Statnamic loading is done and com-
pared with the experiment results. The loading rate
dependency of the ground response is represented by
a constant damping parameter. In this simulation,
velocity and acceleration are calculated from the
displacement by the Newmark method. The stiffness
matrix (IQ the damping matrix (C), and the mass
matrix (A4) are calculated from the following equa-
tions:
K = JB'DBdd
1.01
(7)
565
0.0 0.2 0.4 0.6 0.8 1.0 1.2
Di sp1acein eiit (m in )
Figure 3. Load-displacement relations from statiiaiiuc simula-
tion and experiment.
where N is a so-called shape function or displace-
ment interpolation function, R denotes a displace-
ment to the strain transformation fbnction, and 11 is a
strain to stress-strain transformation function. Pa-
rameter ,U is used to represent the damping constant
per volume of interested material in the same man-
ner as the density (p) is applied in the mass matrix.
Based on trial calculations, a damping constant of
3.9 MN/(m/s)/m3 (,U as defined in Equation (7)) is
used as the material constant for statnamic simula-
tions. The estimated load-displacement relation and
the test data are shown in Figure 3 . Note that the
initial displacement in statnamic load tests does not
conform to that in the static load tests. The initial
displacement in the statnamic load tests is a little bit
smaller than the corresponding value in the Static
load tests. This may contribute to the loading rate ef-
fect during the equipment installation process. To
correct this inconsistency, the load-displacement re-
lation (from the experiment) of the statnamic load
test is shifted to the right in order to match the 'true'
displacement in the static load tests.
4 COMPARISON BETWEEN COMPUTED
STATIC AND STATNAMIC LOADING
RESULTS
The distributions of displacement along a pile under
static loading are shown by Figure 4. At the peak
load, the elastic contraction represented by the dif-
ference in settlement between the pile head and the
toe is about 0.05 mm. When compared to the overall
settlement, this contraction is about 5% of the pile
head settlement. The distribution of axial force along
the pile is shown in Figure 5 . The application of load
is sustained almost totally by friction resistance,
with only a small amount of force being transmitted
to the pile tip. The ratio of the end bearing resistance
to the shaft resistance is about 1 :4.5.
 St3t.k s tatnarfl k:
150 N - 0 - 150 N
+250 N - 0 - 250 1\1
+400N - 0 -400N
+ 450 N - H - 450 N
~"'"1'""1"""'"I""'I""'I''"'I'""I'

() 0 02 0 -4 00 08 10
D 1s p lacem en t ( m m)
Figure 4 Distnbdion of displaccnient along pile under static
loading

The pile velocitv versus time relation is shown in

Figure 6. At the maximum displacement, the veloc-
ity of the pile head is equal to zero. At the same
time, the velocity of the pile tip is almost identical to
that of the pile head. This implies that the pile be-
haves as a rigid body at the unloading point. Axial
force distributions of the pile during statnamic
loading as well as a comparison with those of the
static load test are shown in Figure 7. At the same
0 50 100 150 200 250 300 3% -100
applied load level, the axial force distributions of the
two cases are almost identical. There are no sub-
Load (N) stantial changes in load sustained by the pile tip. In-
Figurc 5. Distribution of axial forcc along pilc undcr slalic creases in capacity mainly contributed to the shaft
loading. resistance. The effect of a stress wave is not ob-
served in this analysis.

5 CONCLUSIOK

When compared to field tests that have the inherent

problem of ground uncertainty, the model tests pre-
sented here show the possibility for conducting Stat-
namic load tests under one uniform condition. Only
by this approach can the data be interpreted on ab-
solutely the same basis. Based on the results pre-
sented here, the following conclusions can be sum-
marized:
1 The finite element analysis (DYSC) shows that
the elastic contraction of the pile is relatively
small. This value represents about 5% of the total
Figurc 6 Vanation III vclocitj of pilc licad <andtoe along tiiiic displacement. From the load distribution plot, the
applied load is mainly supported by shaft friction.
The model pile is thought to be a friction pile.

566
 The ratio of shaft resistance to end bearing resis-
tance is about 1:4.5.
2 Velocities of the pile head and the pile tip are al-
most identical and equal zero at the unloading
point. This supports the assumption that a pile
moves as a rigid body at the unloading point.
3 The axial force distribution from the statnamic
load test is almost similar to that from the static
load test. The stress wave effect, as in dynamic
load tests, is not observed in the present. This im-
plies that in the statnamic load test, the pile is
loaded in the same manner as in the static load
test.
At present, the horizontal load tests on group piles
have been conducted and reported elsewhere (Ki-
mura et al., 1999). The improvement on DYSC in
order to simulate the piles under lateral loading is
under operating.

REFERENCES

Amir. J M and Aniir. E I 1995 A Lumped-Parameter Model

for Statnanvc Testing ProL of the 1" Iiit 'I ,Stntiianiic & i ~ i i -
iiar, I i7iicoiiver pp 221-230
Matsuiiioto. T 1998 FEM anal!sis of Statnainic test on open-
elided steel pipe pile pro^ oftlze T dIiit 'I StatiiamiL Seiiii-
I?W, Ibkvo, pre-printed volunie Balkerna
Kiniura. M . Boonjatee, T & Yoshida. A 1998 Expenmental
stud! of Statnainic load test b! air-pressure based loading
apparatus ~ r o cof the 2 1 ' ~~ i i t StntmniiiL
' ~ Seiviiiar, 70Iil0,
pre-priiited volutiie Balkenia
Kimura. M . Boonlatee, T & Yoshida. A 1999 Evperinicntal
Stud! of Lateral Statnaiiuc Load tests on Group Piles I ' i m
oftlie J'" Iiit 'I Colif oii Deep Foiiiidatroii I+actice 111 or-
prating PILEX-ILK '99,Singapore pp 263-27 1

567
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 1503

Lateral statnamic load testing of model piles

Makoto Kimura & Tirawat Boonyatee

Kyoto Univei-sify,Japan

ABSTRACT: In order to generate the statnamic loading for laboratory tests, a small-scale loading device is
developed. It is driven by the compressed air instead of gas from the explosion. By applying our developed
device, statnamic lateral load tests of a single pile in sand are performed under free head condition. The di-
ameter and length of model pile are 24 mm and 1225 mm respectively. The test results show that the unload-
ing point method can also gives a good estimation for the lateral load tests of single piles.

1 INRODUCTION haved as the short, or rigid, piles. Therefore, an at-

Up to present, the potential of the statnamic load test
to estimate the bearing capacity of pile foundation is
well recognized. However, since there is little un-
derstanding of pile-soil interaction during lateral
statnamic loading, the rational interpretations for the
lateral statnamic load tests have not yet been estab-
lished. Generally, it has not been unusual to conduct
studies based on the data from the field tests, i.e. by
comparing the results of statnamic load tests with
those of the static load tests on the same piles. One
example is the work done by Rollins et al. (1998).
With these data, the consistent interpretation is diffi-
cult to achieve due to the uncertainty and the com-
plexity of ground.
Since consistent systems for interpretation and
comparison can be provided under laboratory test
conditions, it is thought that the model tests are an
indispensable tool for the study of pile-soil interac-
tion during Statnamic loading. In our previous study
(Kimura et al. 1998), a small-scale statnamic loading
device (3SLD-Mkl) that can perform the statnamic
load test under one constant condition has been de-
veloped. By applying pre-compressed air instead of
gas from an explosion, a loading similar to the stat-
namic loading can be generated. This loading device
was used to perform vertical statnamic load tests.
For lateral load tests, a loading equipment named
3 SLD-Mk2 (Small-scale Statnamic Loading De-
vice-Mark 2) is used to apply statnamic loading to
the model piles in a lateral direction.
Using 3SLD-Mk2, the tests of single piles and
group piles have been successklly conducted in the
previous work (Kimura et al. 1999). However, it was
found that the model piles used in that study be-
tempt to make the tests on the flexible piles has been
made. In this paper, the results of the tests on the
single, flexible piles are reported.
The procedure of the tests is the same as the pre-
vious study. Firstly, the static load tests of single
pile are done in order to establish the benchmarks
for the pile resistance. Then, the model ground is
completely reproduced and the statnamic load tests
are suubsequently carried out. The unloading point
method, originally used for vertical statnamic load-
ing, is also used to estimate the static response of
piles in this study. Although there are some assump-
tions that take a certain amount of soil, which sticks
to the shaR, into consideration, only the mass of the
piles is used in this computation.
2 LOADING APPARATUS
2.1 Static load test
In the static loading tests, a mechanical jack is used
to load the target piles. By winding the handle of the
jack, the piles can be loaded by the displacement-
controlled method. Noted that a steel plate is used to
transmit the applied force from the jack to the piles
since their elevations are not the same.
The applied loads to the piles are measured by a
load cell that is attached to the steel plate. The point
of loading is 1 cm above the ground surface. For the
displacement of piles, the displacement at pile head
is measured by a laser displacement gauge. This
gauge is installed on an additional stand to isolate
any effects from the target system. For the static
load tests of a single pile, strain gauges attached at

569
 eleven elevations are used to measure the bending
moments along the shaft

2.2 Stattmmic load test

A general view of the apparatus is shown by Figure
1. Statnamic loads are generated by the quick deliv-
ering of compressed air into the cylinder. The main
part of the device, which is the cylinder, can be
moved smoothly by attaching it to the slide rails.
This slide rail also restrains the moving direction of
the device to the horizontal axis. The flow of air is
controlled by a magnetic valve that is shut in the
normal state and open when an electric current is re-
ceived. By applying an electric timer, an adjustable
pulse of electric current can be given to the magnetic
valve. The mechanism of 3SLD-Mk2 can be shown
schematically as in Figure 2. The operation of load-
ing by 3SLD-Mk2 can be divided into the following
five stages:
Initial stage (Figure 2a.).
Air loading stage (Figure 2b.). In this stage, pre-
compressed air is allowed to flow into the cylin-
der. The air piston will start to load the pile head.
Acting stage (Figure 2c.). While the piston is
pressing on the pile head, the air in the near-pile
side of the cylinder will flow out through the si-
lencer. At the same time, the main body of the
device will move away from the target pile due to
the reaction force from the pile head.
Termination of air supply (Figure 2d.). As men-
tioned before, the magnetic valve will be trig-
gered by the electric timer and the path of air will
be shut off. At this stage, the main body of the
device will still be moving for a while because of
inertia force
Termination of test (Figure 2e.). The motion of
Figure 1. The statnamic loading apparatus. the main part is gradually terminated due to the
friction forces of slide rails.

Figwre 2. Loading mecllanism of 3SLD-Mk2.

570
Table 1. Propertics of iiiodel piles The steel plates are also used to transmit the force
Thickness (nun) 1.o from the piston to the pile heads. The piles are
loaded 1 cm above the ground surface. To ensure
LCil~$Il(cm) 122 5 that the transmission plates contact with the pile
heads and no impact force is induced, a small initial
Weight (8) 591 0 load is applied to the pile's head on the beginning.
Diameter (nun) 24 0
The measured parameters are the applied loads
and the displacement of piles. For statnamic cases,
Young's inodulus (MPa) 8 58\10' these data are recorded every 0.2 ms.
Bending stiffness (Niii') 4 09x10'
3 TEST CONDITIONS AND TEST PATTERNS

Hollow brass pipes are used as the model piles. The

Table 2. Properties of sand and iiiodel ground.
diameter and the thickness of the piles are 24 mm
Specific weight 2.6.3 and I mm, respectively. The bending stiffness is
Maximum void ratio 1.03 determined from bending tests. The properties of the
model piles are shown in Table 1.
Miiiiniuni void ratio 0.64 In this study, No.6 silica sand is used for the
ground material. The model ground is prepared by
Moisture ratio (%) 0.3
vibro-compaction method. At first, the model pile is
D60 (W) 310 placed in the center of the soil chamber. After the
model pile is set up, the sand is filled into the soil
Diu (w) 120
chamber, and then, it is compacted evenly through-
Density (kg/m3) 1167 out the twenty-four locations by injecting a vibration
rod into the ground. Due to the length of vibrator rod
Relative density (%) 59.6 is limited, the ground preparation is divided into two
Frictional angle ( ) 36 steps. The first step is for the lower and the second
step is for the upper layer. The thickness of the
lower and upper layer are 60 cm and 57.5 cm, re-
spectively. The relative density of both upper and
Table 3. Test patterns.
lower layers are about 60%. The overall density of
Pattern Pile type Loading type Length (cm) the model ground is 1467 kg/m'. The properties of
the sand and prepared model ground are shown in
TS1 Single Static 122.5
Table 2.
TS2 Single Static 122.5 In this paper, two types of tests are reported, as
shown in Table 3. For each pattern, the tests are
TD1 Single Statnainic 122.5

TD2 Single Statnamic 122.5

0 1 2 3 4 5
Displacement (mm)
Figurc 3. Load-Displacemcnt relations from static load tests. Flguc 1 Moment dlstnbutlon along thc pilcs.

571
conducted twice in order to ensure that there is no 2 0.3 1 ' ' ' ' ' I ' ' , ' ' I ' ' ' ' ' I ' ' ' ' ',' ' ' ' ' ' ' ' ' ~ ' i 1
disparity in the records. Addition tests will be done
if noticeable differences are observed.

4 EXPERIMENT RESULTS AND

INTERPRETATIONS

4 I Static load tests

By winding the handle, the jack will elongate and
thrust the piles, therefore, the piles are loaded by the
constraint displacement method A loading speed of
0.6 mm/min is applied The piles are loaded up to 1 s iinic
Figrlrc 6 Vclocit! and accclcrai~o~~
135 N, which is approximately the same level as the
maximum force measured from the statnamic load
test The relation between the load and the dis- 43 C bi?iparisonhetit'eeii stntimmic mu' static load
placement are shown in Figure 3. The moment dis- test 1.e.slilt 5
tributions calculated from the strain along the shaft
are shown in Figure 4. Since the induced bending When compared to the data from the static load test,
moments reduce to negligible proportions below 72 the load-displacement plot from the statnamic load
cm from the ground surface, it can be thought that test shows a stiffer relation As shown in Figure 7,
the model piles behave as the flexible piles the load-displacement curves from the static load
test and the statnamic load test intersect near the
unloading point under an applied load of 125 N
4.2 Sfaiiinniicload tests
-4pressure of 0.3 MPa is used to generate the Stat- 4 4 Estii~iatioiiby the ~niloadirig
poiiit method
namic load. Note that this value has not been com-
puted on a theoretical basis, but has come from trial The unloading point method, originally used for
tests. The load and the displacement of the pile head vertical statnamic loading, is also used to estimate
are shown in Figure 5 . From the tests, a maximum the static response of piles in this study. The outline
Statnamic load of 140 N and a maximum displace- for the calculations can be found in the works of
ment of about 4.0 mm are observed. As was also ob- Rollins et a1 (1998) In this method, the inertia can
served in the vertical Statnamic load tests, the be computed by multiplying the lumped mass of
maximurn displacement occurs a short time after the system with the acceleration. Although, it is rational
peak load due to the presence of damping forces. to include a certain amount of soil that moves to-
The duration of loading is about 70 ms. Figure 6 gether with the shaft, only the mass of the piles is
shows the history of the velocity and the accelera- used for inertia force Computation
tion of the piles during statnamic loading. The The predicted static load-displacement relations
maximum velocity and acceleration are 0.24 m/s and compared with the results from the static load tests
25 m/s2, respectively are shown by Figure 8. Although these plots are

150 150 t ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' j
h

100
-20 5 0
-1

h
0

, . . . . . I . . . . . I . . . . .I . . . . .
0 1 2 3 4
.e
a 0 20 40 60 80 100 120
Time (nis) Di splaceinent (inin)
Fibwre 5. Load and displacement vs time. Figtrrc 7. Comparison of load-displaccmcnt rclation.

572
 500
125

100 400
n
5:
W 75 300
20 50
200
1-3
25
1 00
0
t . . . . . l . , . . . l . . . , . l , , , , . l , , . . . i

0 1 2 3 1 5
0
0 1 7
I 3 3 5
Displacement (inin) Di splacein en t (mm)
Figure 8. Estimated load-displacement relations Figurc 9 Dainping cocfficicnt 1 s displaccmcnt

slightly different, the estimated curves show a good 5 CONCLUSIONS

estimation of static response of the model piles.
Based on the presented results and the previous In this study, statnamic load tests of flexible piles
study on the relatively short piles (Kimura et al. have been conducted The data collected from the
1999), it can be concluded that the unloading point statnamic load tests are interpreted by the unloading
method also gives a good estimation for the stat- pint method and compared with the measured static
namic load tests in lateral direction. responses Conclusions from the study can be sum-
marized as follows
1 Load-displacement plots from the SLT and the
4 5 Daniprig coeficierif of the imlondirig point
niethod
STN meet at about the unloading point This evi-
dence agrees well with the assumption of the un-
Accompanying the data from the static load tests, loading point method (Horvath et a1 , 1993),
damping coefficient C, used in the unloading point which states that dynamic forces at an unloading
method, can be back calculated Firstly, damping point are very small and can be disregarded
force I*;{ is calculated by subtracting the statnamic 2 By the calculations that only the mass of pile is
loading with static resistance and inertia force taken into consideration, the unloading point
(F',it and Mn, respectively) as shown in the follow- method can give a good estimation for the tests of
ing relationship flexible piles
3 Based on back-calculated damping parameter (C),
it can be concluded that a good estimation of
piles can be obtained even C is assumed constant
When compared with the prediction on the piles un-
Note that all the three items on the right hand side of der vertical loading, the unloading point method also
above equation can be obtained directly from the shows a good prediction for piles under lateral
experiments. To obtain the damping coefficient, the loading However, it is obvious for a horizontal
damping force is, then, divided by pile velocity L: as loading case that the analysis should be done in
in the following relation: three-dimensional space Therefore, an attempt to
make a versatile analytical tool that can cope with
several loading conditions in three-dimensional
space has been made (Boonyatee & Kimura 2000)

REFERENCES
The variations in calculated C with pile displace-
ments are shown in Figure 9 BoonJatee. T & Kimura. M 2000 Three-dimensional finite
Since the velocity of piles is close to zero, C element analysis of statnamic load tests Proc dh Iiit 'I
changes rapidly at the beginning and around the un- Coiif oil tlie .Ipplication of Strew Iiave 'Ilieon to Piles,
loading point. However, it changes gradually and Sao Paul(,, Brazil Balkenia
Horkath, R G . Beniiinghaiii. P & Middciidorp P 1993 The
undulates within a small range during the intermedi- Equilibrium Point Method of Anal! sis of the Statnainic
ate stage From this evidence, it can be concluded Loadiiig Test \\ ith Supporting Case Histories Proc of the
that for the interpretation of piles under lateral stat IRfh 1iinunl Ueetirig of the Deep Founciatioii Iiistituie.
namic loading, the parameter C can be assumed l'itttburg pp 61-80
constant

573
Kimura. M . Boonlatee. T & Yoshida. A 1998 Expenmental
study of statnanic load test by air-pressure based loading
apparatus Proc Znd hit 'I Statnamic Seiniiiar, Tokvo, Japaii
Balkema
Kimura. M . Boonyatee, T & Yoshida, A 1999 Expenmental
Study of Lateral Statnartuc Load tests on Group Piles Proc
of the Jih Iiit'l Coif on Deep Foundatioi? Practice 111 Cor-
poratiiig PILETiILE: '99, Siiigapore pp 263-27 1
Rollms. K W . Weaver. T J & Peterson, K T 1998 Lateral
statnanuc load testing of a pile group Proc 2'ld hit 'I Stat-
iiatiiic Seminar, Tokvo, Japan Balkem

APPENDIX

To make a rough comparison between the lateral

loading case with the vertical loading case, the re-
sults from the model tests of single piles under verti-
cal loading is provided here. The load-displacements
from the static and statnamic load tests of a model
pile are shown in Figure A l .
1 ~ " ~
-- _------
' ~ ' ' " ~ ~ " " ' ~ ' ' ' ~ '

0 ~ ~ ~ ~ ~ ~ ~ ~ ~ " ~ ~ ' ~ ~ ~ ~ ~ ~ "

00 02 04 06 08 10 I 2
Displacement (mm )
Flgurc A 1 Load-displaccincnt rclallons of a inodcl pilc undcr
1 crlical loading

Then, the result from statnamic load test is used to

estimate the static load-displacement relation by the
unloading point method. Note that this computation
do not include the mass of soil that may be stick to
the shaft during loading. The estimated curve and
the relation from the static load test are shown in
Figure A2.
I ' " " I " ' ' ~ " ' ' ~ ~ " ' " ' 1

400

300

x0
200

100

00 02 04 06 08 I 0 12
Displacement (inin)
F i p r e A2 Estimated 1o;iddisplacemciit relations

574
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Application of the Stress Wave method to automatic signal matching

and to statnamic predictions

G. Esposito, W. M.G.Courage & R. J.van Foeken

TNO Building und Construction Research, Department of Structurul Dynamics, Rijswijk, Netherlunds

ABSTRACT: The Statnamic method is an increasingly popular technique to cany out loading tests on cast in-
situ piles. The method has proved to be a cost-effective alternative to a static loading test. As-sociated to
Unloading Point Method (UPM) and to automatic signal matching, the Statnamic testing technique provides an
estimation of the static load-settlement curve of foundation piles. For long piles with wave numbers (Nw) less
than 12, however, the method becomes less accu-rate because of stress wave phenomena (concrete piles longer
than 35 m or steel piles longer than 40 m). In these cases, the signal matching stress wave method can be used
to estimate the static load-settlement curve. The TNO Stress-Wave method is a convenient way to carry out
automatic matching on Statnamic signals. The use of TNO Stress-Wave method and of auto-matic signal
matching for a Statnamic prediction is investigated in this paper. A case study is introduced and the results of
a signal matching are presented. Further, the case study is mod-elled in TNO Stress-Wave and the results of
the model are compared to the results of the field Statnamic test.

1 INTRODUCTION the differential equation for longitudinal stress waves

Where a static load test is expensive and time con-
suming, a Statnamic load test can be much cheaper
and faster. Together with a dynamic load test, Stat-
namic is a cost-effective alternative to a static load
test. Dynamic load tests perform well on pre-cast
concrete piles and steel piles. On cast-in-situ piles
(auger, bored, etc.), the Statnamic method is pre-
ferred, because the pile shape is not known and can
vary. To match soil parameters, the cross sectional
changes are also considered as soil variations (Mid-
dendorp 1998).
The Unloading Point Method (UPM) is an ac-
cepted method of analysis for Statnamic tests to de-
termine static bearing capacity of foundation piles
having wave number greater than 12. The wave
number, N,, is defined as the ratio of the product of
the duration of the blow and the wave velocity re-
spect to the length of the pile. To interpret Statnamic
load tests on long piles with wave numbers less than
12, the unloading point method becomes less accu-
rate because of stress wave phenomena (Middendorp
1995). The pile does not move as a rigid mass with a
uniform velocity. The movement of the pile head and
pile toe is not synchronized and, thus, the pile veloc-
ity and acceleration can not be approximated by the
observed pile head displacement.
There are several programs which can describe the
stress wave phenomena in a pile and are all based on
in a pile. The first stress wave program was devel-
oped by Smith in 1960 and solved the differential
equation by means of the fmite difference method
(Smith 1960). The TNOWAVE stress wave package
was developed in the early 80s and is based on the
method of characteristics (Voitus 1974), where the
pile is defined and modelled by elements with a char-
acteristic length. At the interfaces of the elements, a
soil model is defined. Between each interface, the
pile remains frictionless and a wave travels undis-
turbed as if in a fiee pile. The soil response is a fimc-
tion of the relative dis-placement between pile and
soil, the pile velocity, and the pile acceleration. At
each interface, part of the wave is reflected and part
is transmitted, depending on the properties of the
elements and soil resistance. For signal matching, the
measured load on the pile head is used as a boundary
condition at the pile top. Statnamic predictions can
also be made with TNOWAVE (Bielefeld 1995), al-
though the he1 combustion must be modelled.
The objective of this study is to demonstrate the
application of automatic signal matching for a Stat-
namic prediction. The case study herein presented
refers to a bored pile having wave number lesser than
12 where the UPM proves to be not adequate. The
same Statnamic test is fbrther modelled by means of
the TNO Stress-Wave Method.

575
2 THEORETICAL BACKGROUND
2.1 Unloading Point Method (UPM)
The long duration of loading (N, > 12) allows the
pile to be modelled as a rigid mass on which the fol-
lowing forces act:
Applied load or Statnamic force, FSatn;
Inertia force of the pile mass, Fa(t) = Mplle a(t),
where Mplle is the total mass of the pile and a is
the actual acceleration;
8 Soil resistance along the shaft of the pile and at
the toe of the pile, F,, which is composed of a
static resistance, water pore pressure, and damp-
ing force from the soil;
Static resistance force, FU-SlatlC;
Water pore pressure, Fp;
Damping force from soil, Fv(t) = C - v(t), where
C is the mean damping value and v is the actual
velocity.
The displacement of the pile top equals the dis-
placement of the pile relative to the soil. For simplic-
ity, the pore pressure resistance is included in the
damping. Figure 1 shows the Statnamic load-
displacement with dynamic components.
Displacement U Load displacement diagram
-
Figure 1 : Statnamic load-displacement diagram
1. The Statnamic reaction mass is placed on the pile
top. The load-displacement behaviour is fully elastic.
2. The reaction mass is launched and Statnamic
loading starts. The soil resistance is elastic and inertia
and damping forces act on the pile.
3. The static soil resistance reaches ultimate strength
and yields. Velocity and inertia forces increase pro-
gressively. At the end of this area, the maximum
Statnamic load is reached.
4. Statnamic load decreases. Due to pile inertia, dis-
placement is increasing. The soil is yielding and pile
velocity reduces to zero. The pile displacement
reaches a maximum value and damping forces and
velocity become zero. The Statnamic load minus the
inertia force equals the static soil resistance at time
tu-max.
576
5. The unloading force of the static soil resistance
overcomes the opposing forces and pushes the pile
upwards. At the end, final settlement can be ob-
served.
The static load behaviour can be derived as fol-
lows:
or
At each point, the inertia force and the damping
force can be calculated, resulting in the following
static resistance:
All terms of equation (2) can be determined from
the Statnamic loading data, except the damping fac-
tor Cmean. In the unloading point method, at maxi-
mum displacement the velocity and therefore damp-
ing force is zero. The damping factor is calculated
fi-om the average of
obtained from the time range between FStat,, maxi-
mum value and the unloading point with the condi-
tion that v(ti) is greater than 25% of the maximum
velocity contributing to the average.
With the damping factor, C,,,,, known, the static
load curve can be calculated from the Statnamic load
curve at any time. A smooth curve can be con-
structed from the calculated values using appropriate
curve fitting.
2.2 TNOWA VE - Stress Wave Method
TNO Building and Construction Research developed
the stress wave package TNOWAVE to simulate
stress wave propagation in piles. The first application
was signal matching for dynamic load testing. The
method also applies to signal matching for Statnamic
load testing. Other applications include sonic integ-
rity testing simulation and signal matching (Midden-
dorp 1988), pile driving prediction for impact ham-
mers and vibratory hammers (Bielefeld 1994 and
1992) and Statnamic simulation (Bielefeld 1995).
The stress wave method for wave numbers less
than 12 is much more reliable than those with the
unloading point method (Middendorp 1995). In the
stress wave method, the soil layers and classification
and the actual pile shape are defined. Since signal
matching with soil parameters is time consuming and
requires an experienced engineer, TNO developed an
'automatic' matching routine to help the engineer in
the matching process Early automatic matching rou-
tines in 1984 calculated pile shapes from measured
signals (Ypey 1984) and determined the pile shape of
several model piles measured in research conditions
without soil As soil influences were not considered,
however, the accuracy of practical results was v e q
low I n 1 989, more advanced research programs be-
gan (Staalduinen 1989) I n 1992, a routine for auto-
matically calculating measured signals was developed
(Courage 1994)
The routine used in automatic signal matching of soil
or pile parameters has a dif-f-'erentbasis now instead
of directlg. calculating the pile shapes and soil models,
TNOWAVE matches signals as an engineer would,
with a "search strategy" to find the best fit between
calculated and measured signals Superbision by an
engineer, with experience in signal matching, geo-
technical Anowledge, and practical experience in
foundation engineering, is still 1-equired Automatic
signal matching is not replacing the engineer, but of-
fers support to perform matches more easily and
faster
The p remi ses for deve1opi ng autoin at i c signa 1
matching in TNOWAVE were
A mathematical model \n/hich accurately describes
the behaviour of foundation piles in soil layers,
with an appropriate paraineterization of the pile
properties and pile-soil interaction In
TNOWAVE, the mathematical model is based on
the method of characteristics, developed by Voi-
tus van Haniine ( 1 974, 1977), Middendorp
(1 966), and others
* An accurate and efficient computational algo-
rithm to solve the boundary value problem de-
scribed by the tnathematical model Such an algo-
rithm is available in TNOWAVE and is referred
to as the "farward" model.
0 Reliable measurement data of excitation and re-
sponse of the pile on the site Using digital proc-
essing techniques, the Foundation Pile Diagnostic
System (FPDS) is capable of obtaining high
quality signals
A method was sought that would determine un-
known or uncertain parameter values in the model
giving good agreement between the forward model
and the measured data The method should incorpo-
rate n yrrori ltnowledge about probable values and
uncertainties using n p~ioricovariance Apart from
parameter values, the method should also give co-
variance using a Bayesian approach (Norton 1986)
577
Usually, time signals are obtained froin measure-
ments. revealing information of lower pile parts and
soil layers up to the reflection of tlie pile toe with in-
creasing time It is therefore useful that the method
considers new data points iteratively, updating the
parameter values and their reliabilities
To satisfi the above properties, tlie Kalman fil-
tering method was chosen The I<alman filter was
originally developed to identie the behaviour of dy-
namical systems and has proben to be a reliable tool
in identification problems When applied to paranie-
ter estimation, its performance has been denion-
strated more recently For a complete derivation of
the I<alman filter equations, refer to the extensibe lit-
erature (Kalinan 1960, Jazwinski 1970, Norton
1986) For parameter estimation, more information
can be found in Hendrilts et al (1990) and Courage
et a1 (1994)
The method is schematically depicted in Figure 2
With the impact signal and initial pal-ameter Lalues
given by tlie user, TNOWAVE calculates the re-
sponse of the system by the forward model This re-
sponse together v. ith the measured response is then
supplied to tlie Kalman algorithm as implemented
nithin TNOW 4VE automatic signal matching It is
based on the difyerence betmeen both responses, the
partial derivatives of the responses with respect to
the parameters as numerically derived by a peiturba-
tion metliod, and CO\ ariance concerning the nieas-
ured er1 01s, modelling errors, and parameter values
Proper gain is calculated to update the actual pa-
rameter \ alues Besides iie\b parameter values,
po,/erior covariance for the parameters are also cal-
culated
The procedui-e is repeated until a convergence
criterion is met, for instance small residuals in calcu-
lated and measured response, small updates for the
parameters, or a maximum number of iterations As
described above, the data can be considered itera-
tively, updating values from new data points All
features are incorporated in TNOWAVE by means of
a user-friendly menu-structure for input data and
control parameters as well as the output and presen-
tation of results calculated
Figurc 2 Schematic flou chart for automatic signal matching
For Statnamic load test signal niatching, the soil,
measured impact signal. and the pile must be mod-
elled The starting values of the soil parameters for
each layer are estimated from soil investigation re-
sults, such as CPT-values or SPT-values The layer
definition can not be changed. but the soil properties
(yield stresses, quakes, and damping factors) are up-
dated until a good match is made In dynamic signal
matching, the match is done on the upward travelling
wave For Statnamic signal matching, it is better to
match on the velocity signal at the pile top The dis-
tinction between measured and calculated signals are
more pronounced i n the velocity signal comparison Figure i Force \crs~istime signal c m be split into riitcnals
The measured signal is split into 5 intervals for
which each allowable errors can be set (see Figure
3) Also a certain updating scheme is used First the
match is concentrated on the shaft soil model and
than on the toe soil model Than a second update is
made oii both The iteration is finished with a re-run
011 shaft, toe and the combination The best match
during the inatcliing process is stored for further
use This is in most cases the match at the end of the
update scheme
The soil model used for Statnamic signal matching Figure 1 Soil models at shali m d toe
is composed of asymmetric springs and linear damp-
ers along the shaft and an asymmetric spring with
gap and linear damper at the toe as shown in Figure
4 Velocity along pile axis

3 100
E
0 75
3 CASE HISTORY STATNAMIC TEST IN 0 :0
TA1Lli A N
0 25

3 1 7 N O WA I 7< Sigrid Mcrkhiirg 0 00

A 16 MN test was performed in Taiwan on a 51 ni 0 25

000 1000 2003 3000 4000 5000 E000
long bored pile with a diameter of 1 0 in The pile Pile axis [m]
toe is situated in a sand layer of I7 ni On top of this
layer is a 20 in thick clay layer The upper soil is Figure 5 Velocitj along pile axis at three stages
composed of sand, silt, and clay layers
The duration of the Statnamic load was about
100 nis giving a wave number of about N,,= S and,
Table 1 : Soil match data
thus, the direct application of the UPM was not
Layer Thickness Yield Stress Quake Damping
possible The velocity and acceleration observed at
the pile head were not representative of the overall Lml [MPaI [mml C
velocity and acceleration of the pile as shown in [MNs/m]
Figure 5 To obtain a static load-displacement 1 2.5 0.01 5 0.03
curve, a signal match must be made The soil model 2 5 0.0 1 5 0.048
along the shaft and toe must be chosen in a way 3 7 0.027 5 0.038
such that the calculated velocity signal at the top of 4 20 0.0 I 5 0.076
the pile matches the measured velocity After a good 5 7 0.03 1 5 0.06 1
velocity match has been made. the measured dis- 6 9.5 0.098 513 0.071
placement should also match the calculated dis- toe 8.677 25 0.00 1
placement The results of the signal match are
shown below for velocity, displacement, and up-
ward travelling wave at the top of the pile in Figure
6 The soil match data is listed in Table 1 The yield 1 .O(last layer 2.0; toe is 0.0). The soil damper is as-
factor (F\2/F,1, see figure 4) for the shaft springs is sumed linear.

578
 [mml -MoblllsedLoad Displ Dragram, Pile Head 3
0.0

8.0

16.0
,
18 I
4
i '\
0.6
I . d'
-0.6
00 100 200 20.0I
Calculated and - - measured Time [ms]
_ _ I
nowAMIm I"I*)_."
[MNI

Figure 7: Static load-displacement diagram

28

20

12 I

4 1

4
0 100 200
-Calc. and --measured Time [ms]

Figurc S: Statnamic iiiodcl

-08
0 100
-Calculated and _ _ measured
(cl
Figure 6: Results of signal match for displacement(a), velocity
(b), and upward travelling wave at top of pile(c).
The sum of the damping values over the whole
shaft and toe give a damping value of 14 MNs/m3.
From the UPM, a damping value of 9.6 MNs/m3 is
obtained. Both values are related to the toe cross
section. From the best soil match, the static load-
displacement curve can be calculated, by simulating a
static load test on this model using only the dis-
placement dependent part of the soil model as shown
in Figure 7. The static resistance from the signal
match is 10.2 MN. The UPM gives a static resistance
of 14 MN. The latter value is not correct and con-
firms that a signal match is required for wave num-
bers less than 12.
200
Time [ms]
3.2 TNOWAVE - Stress Wave Method
A Statnamic test is a complex interaction between
the Statnamic device, pile, and soil. As with a diesel
hammer, the pile-soil s t f i e s s has an influence on the
peak load and jump height of the reaction mass. The
modelling of the load at the pile top from the burning
of the propellant is important to properly simulate a
Statnamic test. The loading of the Statnamic device
depends on the initial chamber volume, type, and
amount of fuel and reaction mass, vent distance, and
vent height. As well, the shape and dimensions of the
vent are important (Bermingham 1995). The Stat-
namic loading in TNOWAVE includes the burning of
the fuel according to pyrotechnics and depends on
the fuel chemistry, burning coefficients, and geome-
try of the fuel (Bielefeld 1995).
The signal match case history is used to make a
prediction. The 16MN device is modelled on top of
the pile as shown in Figure 8. A fuel loading of 11.2
kg was used with a reaction mass of 0.8 MN. The

579

Figurc 9: Results of Statnamic prediction for displacement (a)
and force at top of pile (b).
following parameters have been used to model the
Statnamic device: combustion chamber area 0.2082
m2; initial chamber volume 0.12426 m’; vent diame-
ter 0.15875 m; vent length 0.1016 m; vent distance
0.1524; plenum volume 1.24542 m3; plenum vent
area 0.02 194 m.
The Statnamic simulation starts with the pre-
loading of the pile by the reaction mass. After 20 ms,
he1 combustion in the Statnamic chamber begins.
The pressure in the chamber increases rapidly and as
a result the mass is accelerated and launched. The re-
sulting reaction on the pile top introduces forces
causing stress waves to travel in the pile. Shaft fric-
tion and toe resistance will be mobilised. The dis-
placement of the pile will have its influence on the
burning fuel. The calculated displacement and force
as hnction of time is shown in Figure 9.
The predicted jump height of the reaction mass is
4.46 m, a conservative result since friction forces
from gravel are not considered. The calculated force
and displacement are about 10% to high. A good
comparison can be obtained by using 10% less hel,
but can not be currently confirmed by other back
analysis.
4 CONCLUSIONS
For wave numbers greater than 12, the UPM (un-
loading point method) can be straightforwardly ap-
580
plied. The inertia force calculated from the total pile
mass should be considered to predict the proper
static resistance. For wave numbers less than 12, only
part of the pile mass will contribute to the inertia
force.
Signal matching with TNOWAVE offers a good
method to determine the static capacity of a Stat-
namic load tested pile when the wave number is less
than 12.
Statnamic stress wave predictions give a conservative
result. More back analysis is required, especially to
predict the launch (jump) height.
REFERENCES
Bermingham. P.. Janes, M.. 1989. An innovative approach to
load tcsting of high capacity piles. Proceedings of the In-
ternational Conference on Piling and Deep Foundations.
Lol1don. p.409-4 13.
Bermingham. P., White. J.. 1995. @roteclinics and the accu-
rate prediction of Statnamic peak laoding and fuel charge
size. First In/erna/ional S/atnniiiic Senrinnr. I ancoiives,
British Colunrhia, Canada. September 27-10. 1995.
Bielefeld. M. W., Middendorp. P.. 1995. Statnamic Simulation.
bi’rst International Statnainic Seminar, I ancoutter, Lh-itish
Colimbia, Canada. September 27-30.
Bielefeld. M.W.. 1991. Prediction of Installation of Sheet Piles
using Vibratory Hammers. Fl’jih lnlernational C’onfesence
and Exhibition on Piling and Foundations. DFI. 13- 15
June. Belgium.
Bielefeld. M.W. and Middendorp. P.. 1992. Improved PiIe
Driving PreQction for Impact Hammers and Vibratory
Hanuners. Fourth Inrerna~ionalConference on the .4ppli-
cation of Stress Wave Theorv t o Piles. The Hague. 21 -24
September.
Courage. W.M.G. and Bielefeld. M.W.. 1992. TNOWAVE
automatic signal matching. Application of S/re.s.s-J b l w
Theory to Piles, F.B.J. Barends (ed.) Balkema. Rotterdam.
ISBN 9054100826.
Courage, W.M.G. and Foeken, R.J. van, 1992. TNOWAVE
Automatic Signal Matching for Dy-namic Load Testing.
Application of 5’tre.w-Wave Theoiy to Piles. F.B.J. Barends
(ed.) Balkema, Rotterdam, ISBN 9054100826.
Foeken, van R.J., Daniels, B. and Middendorp, P., 1996. An
improved method for the real time calculation of soil re-
sistance during h v i n g . Application of Stress- Wave Theory
to Piles, F.B.J.Barends (ed.) Balkema, Rotterdam.
Hendriks, M.A.N., Oomens, C.W.J., Jans, H.W.J, Janssen,
J.D. and Kok, J.J., 1990, A Numerical Experimental Ap-
proach for the Mechanical Characterization of Composites.
In V. Askegaard (ed), Proceedings of the ninth Interna-
tional Conference on Experimental Mechanics, Aaby Tryk.
Copenhagen.
Horvath R.G., 1990, Statnamic, an accurate and innovative
load test method for high capacity deep foundations. Foun-
dation Drilling, Volume XXVIII, No. 11,
Jazwinski, A.H. 1970, Stochastic Processes and Filtering
Theoty. Academic Press, New York and London.
Kalman, R.E., 1961. A new Approach to Linear Filtering and
PreQction. Trans. ASME.
Koten, H. van, Middendorp, P. & Brederode, P. van.,1980. An
analysis of dissipative wave propa-gation in a pile. Intl.
Seminar on the Application of Stress-Wave Theory on
Piles, Stockholm, 4-5 June.
Middendorp, P., Foeken, R.J. van, 1998. When to apply Dy-
namic Load Testing or Statnamic Load Testing, Second
 International Statnamic Seminar, Tokyo, Japan, October
28-30.
Middendorp, P., 1993. First experiences with statnamic load
testing of foundation piles in Europe. 2nd International
Geotechnical Seminar “Deep Foundations on Bored and
Auger Piles”, Gent, Belgium, 1-4 June.
Middendorp, P., Bielefeld, M.W.,1995. Statnamic and the In-
fluence of Stress Wave Phenomena. First International
Statnamic Seminar, Vancouver, British Columbia, Canada,
September 27-30.
Middendorp, P. and Weele, van A.F.,1986. Application of
Charateristic Stress Wave Method in Offshore Practice.
Proceedings of the third International Conference on the
Numerical Methods in Offshore Piling, Nantes.
Middendorp, P. and Zandwijk, van C.,1985. Accuracy and
Reliability of Dynamic Pile Testing Techniques. Proceed-
ings of the fourth International Conference on the Behav-
iour of Offshore Structures. pp 601-609.
Middendorp, P. and Reidmg F.J.,1988. Determination of Dis-
continuities in Piles by TNO Integrity Testing and Signal
Matching Techniques. Third International Conference on
the Application of Stress Wave Theory to Piles. Ottawa,
Canada, 25-27 May.
Norton, J.P., 1986. An Introduction to Identification, Aca-
demic press, new York and London.
Ser. D.J.,1960.Basic Engineering, Vol. 82, pp 95-109.
Smith, E.A.L. 1960. Pile Driving Analysis by wave equation.
Journal of the Soil Mechanics and Foundation Division,
ASCE vol 80. SM4.
Staalduinen. P.C. van, Bielefeld, M.W. and Middendorp, P.,
A.,1989. Direct Technique for Discon-tinuities in Piles.
The second International Conference on Foundations and
Tunnels. London.
TNO report, 1985-1996. TNO-LILT Dyncmzic Load Testing
Signal Matching, Users Manual.
Ypey. L.P., 1984. De deconvolutie rnethode toegepast op het
sonisch doortneten van fundatiepalen. Master’s Thesis,
Delft University of Technology. The Netherlands (in
Dutch).
Voitus van Hamme. G.E.J.S.L, 1977. Bepalen van heiweer-
standen uit meetresultaten. (In Ducth). Presentation on
13-eedenhriightiag.
Voitus van Hamme. G.E.J. S.L, Jansz J.W., Bomer. H., and
Arentsen, 1974. D., Hydroblok an improved pile driving
analysis. De Ingenieur. Vol. 86. no. 8. pp 344-352.

581
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) Cl 2000 Balkema, Rotterdam, ISBN 90 5809 150 3

A comparative study of static, dynamic and statnamic load tests of steel pipe
piles driven in sand
A. Shibata - Kubotu Corporation, Ichikawu, Japan
N.Kawabata -Nippon Steel Corporation, Tokyo, Japan
YWakiya - Kawasaki Steel Corporation, Tokyo,Japan
YYoshizawa - Sumitorno Metal Industries Limited, Tokyo,Japan
M. Hayashi - NKK Corporation, Kawasaki, Japan
T. Matsumoto - Department of Civil Engineering, Kanazawa University,Japan

ABSTRACT Comparative static, dynamic and the Statnamic load tests on three steel pipe piles driven in
sandy ground were conducted. In this test program, dynamic load tests were performed at various time
intervals after the initial driving to investigate the variation of the pile capacity with different hammer driving
energies, and the increase in the pile capacity with elapsed time after the initial driving Two different
computer programs were employed for the wave matching analyses to estimate the static load-displacement
curves for the piles, which were compared with the static load test results The Statnamic test was also
performed on one of the test piles that had underwent the static load test. This paper will discuss the uses of
the dynamic and the Statnamic load tests as alternatives to the static load test

1 INTRODUCTION 3. To investigate the differences between the

More than 20 comparative cases of the dynamic and
the static load tests on driven steel pipe piles were
collected by the Japanese Associatioa for Steel Pipe
Piles (JASPP) to examine the use of the dynamic
load testing to estimate the static load-displacement
curve for steel pipe piles. The data were collected
from the field tests with various pile configurations,
various soil conditions, various driving hammers
used, and various rest periods after the initial driving
for the dynamic load test (Wakiya et al. 2000).
The collected data suggested that the reliability of
the static load-displacement curve estimated through
the wave matching analysis of the dynamic load test
signals depends on hammer energy, rest period
before the re-driving test, the computer program
used for the wave matching analysis and soil test
data available for the site.
Therefore, JASPP conducted their own test
program of various load tests on three open-ended
steel pipe piles in a relatively uniform sandy ground
at Hasaki (the test ground of Sumitorno Metal
Industries), Japan, in 1993, to evaluate the use of the
dynamic load test and the Statnamic load test for
piles in sandy soils as an alternative to the static load
test. The emphasis was placed on the following
goals in this particular test program:
1. To evaluate the influence of different hammer
driving energies on the estimated pile capacity.
2. To measure the increase in the bearing capacity
of the pile with elapsed time after initial pile
driving, (the so-called "set-up" phenomena).
results of various computer programs used for
the wave matching.
2 TEST DESCRIPTION
2 1 The f e s fsife md f e s fpr1e.r
Figure 1 shows the soil profile and the results of site
investigations at the test site The Standard
Penetration Test (SPT) and the Cone Penetration
Test (CPT) were conducted immediately before and
7 weeks after the pile installation The test ground
consisted of fine to gravel sands from the ground
surface to a depth of 20 m The SPT N-value
measured prior to the pile installation was relatively
high and uniform at depths greater than 7 m The
ground water level existed at a depth of 5 m from the
ground surface The variation with depth of the tip
resistance, qc, from the CPT before the pile
installation seems to be similar to the variation of the
SPT N-values The sleeve friction, A, tends to
increase linearly with depth to a depth of 9m and
level off for larger depths
The layout of the test piles, the location of site
investigations and a piezometer are shown in Figure
2 PThree test piles, designated piles D for dynamic
load testing, S for static load testing and M for soil
measurements, were prepared The test pile
specifications are listed in Table 1 Pile S and pile D
were individually prepared, so that the dynamic and
static load tests could be carried out at the same
elapsed time after initial pile driving to minimize the

583
Figwe 1. Soil profile and results of site investigations at the test site.

M also, so that piles D and M had the same

configuration as pile S, although piles D and M were
not instrumented with strain gages. The steel
channels increased the net cross-sectional area of the
test piles to 0.017m2.
Pile M was prepared for the purpose of
investigating the change in the soil conditions
around pile M before and after the pile driving. The
CPTs and the SPTs were conducted at a distance of
0.4m from the center of pile M immediately before
and 7 weeks after the pile driving. There was little
increase in N-values and qc-values, while there was a
clear increase in&-values after pile driving to depths
deeper than 8m. The excess pore pressures
measured by the pre and post CPTs were very small
Figure 2 Lavout of the test piles, the location of site
inwtigations and a piezometer indicating the relatively high permeability of the
sand.
Table 1 Specifications of test piles An electric pore pressure transducer (piezometer)
Propert) Value was placed in the ground at a depth of 11.5 m from
Length L ni 13 0 the ground surface prior to the initial pile driving.
Outer diaineter U inin 100 0 The horizontal distance between the center of pile D
Wall tluckness t,, Illill 12 0
Cross-sectional area ,A ni2 0 017
and the piezometer was 0.4m, which was equal to
Young’s modules E MNlin’ 206x10’ the outer diameter of the test pile.
Mass densih p t/in’ 7 85
WaL e 1 elocity c m/s 5120
Mass .U ton 1.78 2 2 Test sequences and procedures
Cross-sectional area- ;1, includes the cross-sectional area of A series of dynamic load tests was performed on pile
steel channels for protection of strain gages.
D after 8 different time intervals, 0 min, 5 min, 15
influence of the loading history for each pile. Pile S min, 30 min, 1 hour, 3 hours, 20 hours, 6 days and
was instrumented with strain gages at a total of 6 30 days, after the initial driving. In the dynamic
levels. Steel channels were welded to the outside of load tests, the hydraulic hammer (HK65), which
the pile shaft for the protection of the strain gages. allows the control of the hammer driving energy,
Note that steel channels were attached to piles I)and was used. Pore pressures were measured during the

584
pile driving. The dynamic signals (time variations
of accelerations and strains at the pile head) were
monitored by a PDA (Pile Driving Analyzer) and an
FPDS (Foundation Pile Diagnostic System), which
are widely used in the world on a commercial basis.
Two different computer programs, CAPWAPC and
TNOWAVE, were employed for the wave matching
analyses of the dynamic load test signals to estimate
the static load-displacement curve for the pile.
A series of static load tests were conducted on pile
S, separately. The first test was carried out 6 days
after the initial piIe driving and the last test was
conducted 30 days after the initial pile driving. The
measured load-disp~acementcurves were compared
with those estimated from the dynamic load tests on
pile D to examine the applicability of the dynamic
load test to the estimation of the static load-
displacement curve for a pile.
The Statnamic test was also carried out on pile S
after the completion of the series of static load tests,
52 days after the initial pile driving (Nishimura et al.
1995).
of the shaft resistance along the top 3 5m of the pile
at 6 days after the pile driving may be attributed to
the fact that a gap between the pile shaft and the
surrounding soil was generated by lateral
movements of the pile near the ground surface
during the pile driving. The shaft resistance along
the top 3.5m of the pile recovered after a rest period
of 30 days, although it is difficult to comment on
this mechanism.
4 DYNAMIC LOAD TEST R E S a T S
4.1 The resirifsdirring the irzitiai driwig
Figure 5 shows the change in the total resistance, Rt,
with penetration depth of the test piles during the
initial driving. The total resistance, Rt, was
estimated by the Case method (Goble et al. 1975).
The total resistance of each test pile increases almost
Load on pile head (MN)

3 STATIC LOAD TEST RESULTS

A total of 5 static load tests on pile S were carried

out 6, 7 and 30 days after the initial driving as
shown in Figure 3. The maintained step load test
method was employed for the first and the last (5th)
test Each load step was maintained for 15 min. The
number of load steps was 5. The quick maintained
load test method was employed for the Znd, 3rd and
4th tests, in which the pile head load was increased
to the ~ a x i m u mload in about 20 minutes. It can be
seen fiom Figure 3 that the curvatures of the load-
displacement curves after the yield loads obtained Figure 3 Load 4splacenient c m e s obtained from static load
from the first and the last tests are similar and they tests (pile S).
are more moderate than the other tests in which the
quick ~aintainedloading method was used. It is Shaft resistance. z-(kN/m3)
also seen from the comparison of the 4th test and the 0 50 100 150 200 250 300
last (5th) test that the yield load in the last test is
smaller than that obtained from the 4th test, althoug~ 6 daj s after EOID
the rest period for the last test is longer than the 4th
test. This fact indicates that the influences of the
loading rate andior the Ioading duration are not
"I '
- 4
negligible even for sandy ground when it is E
W
saturated.
Figure 4 shows the distributions of the shaft
resistance, 2, obtained from the first static load test
and the last static load test. The shaft resistance, 2,
was estimated from the measured axial strains of I I
pile S when it penetrated a distance equal to 10
percent of the pile diameter in each test. There is a
slight change in the distributions of the shaft
resistance after the rest periods of 6 days and 30
days, except for depths shallower than 3.5m.
Accordingly, it is thought that set-up has completed Figure 4. Distributions of the shaft resistance obtaincd from
within 6 days after the initial pile driving. The loss the static load tests conducted 6 and 30 days afrer initial
driving (for pile S).

585
linearly with the increasing penetration depth and
attains about 2 MN at the end of initial driving.
Therefore, the bearing characteristics of all the test
piles are comparable with each other.
4.2 The ii$'iierrce of hammer driving energy
The static pile capacity, R,,estimated from the
dynamic load test may vary with the change in the
driving energy actually transmitted to the pile To
investigate this aspect in detail, the dynamic load
tests were carried out on pile D changing the
hammer driving energy after different time intervals
from the initial pile driving
Figure 6 shows the change in the static resistance,
R,,estimated from the wave matching analysis of the
dynamic load test signals using CAPWAPC, as a
function of the nominal hammer driving energy, EH.
The measured values of set per blow, S, are also
shown in Figure 6 The static resistance and the set
per blow increase gradually with increasing hammer
driving energy to 75 kNm. The static resistance
attains it's peak value of 2 MN when the hammer
energy is greater than 80 kNm On the contrary, the
set per blow increases steeply for hammer driving
energies greater than 75 kNm, indicating that the
pile capacity is fully mobilized by hammer driving
energies greater than approximately 75 kNm.
586
It is recommended that the relations (EHvs R,and
~ 5') as shown in Figure 6 be obtained prior to
E I vs
the actual pile driving, in order to assure that the pile
capacity is fully mobilized by the selected driving
hammer. If EH versus R, alone is measured, it is
difficult to judge that the peak value of R, is the
actual pile capacity or it is a limitation due to the
insufficient hammer driving energy. The relations
shown in Figure 6 are useful for the selection of an
appropriate hammer for each site.
4.3 Set-irp phenomerza
Dynamic load tests were performed after 8 different
time intervals measured from the end of initial
driving, in order to investigate the set-up phenomena
of a steel pipe pile (pile D) driven in sandy ground.
Figure 7 shows the change in the total resistance, Rt,
which was estimated using the Case method, with
elapsed time after the pile driving. Two dynamic
monitoring systems, PDA and FPDS, were used to
record the dynamic signals. The total resistance, Rt,
reaches it's peak 60 min after the initial pile driving,
and remains almost constant after that time. The
peak total resistance is 1.2 times R, measured at the
end of initial driving, that is to say, the "set-up ratio"
for the total resistance is 1.2 in this case.
Figure 7 Increase 111 total resistance. R,. nit11 elapsed tiine after
inihal driving (pile D)
Figure 8 shows the change in excess pore
pressures with elapsed time after each blow for the
different depths of the pile tip, during the initial
driving of pile D Note again that the piezorneter
was placed at a depth of 11.5m from the ground
level and at a horizontal distance of 0.2rn from the
outside of the pile shaft When the pile tip level is
above the level of the piezometer, the magnitude of
the first peak of the positive excess pore pressure,
Au, increases, and the time instant of the peak of Air
becomes earlier, as the pile tip approaches the
piezometer level. When the pile tip level is below
the piezometer level, the value of the peak positive
excess pore pressure, Au, decreases as the pile
 initial driving of pile D were conducted to estimate
the static load-displacement curve for the pile for
E
.-.
-0.1 each time interval after the initial pile driving.
z Depth of pile tip = 10.0 in Figure 9 shows the measured wave signals of the
re-driving tests. Figure 10 shows the results of the
wave matching analyses. The wave matching
analyses were conducted using CAPWAPC and

>
mn Depth of pile tip = 10.5 in TNOWAVE programs. The calculated and
2 0.1 measured upward traveling forces are compared in
2 0.0 Figure 10.
$
+I
-0.1
a Depth of pile tip = 1 1.0 ni
0.1
6 da\ss after EOID.

m
U,
Depth of pile tip = 1 1.5 in

1$
(I)
3 0.1
w 0.0
-0.1
Depth of pile tip = 12.0 iii
O l t - " " * " ' ~ 4
01
-0 ° ~
000 001 002 003 004 005 h

-(b) 6 days after EOID

Time (s)
Figure 8 Changes in elcess pore pressures uith elapsed time
-
E
2
& 4 -- 30 days after EOID

dunng each blou at different depths of the pile tip (pile D) -: 32:
'5

1-

penetration proceeds It is interesting that negative

5 0.

excess pore pressure is generated after the peak

8 -1:
" " " '
-
- - - - -___....
~

positive pressure, although the magnitude of the

negative excess pore pressure decreases as the pile
penetration proceeds
Although the measured excess pore pressures
have not been fully interpreted, the influence of the
excess pore pressures on the pile penetration 20
resistance may not be negligible even for sandy
ground One of the interesting features of the - 15
measured pore pressure is the fact that the pore
pressure dissipated almost completely within 30 ms
Ta) 0l 50

As indicated later, the duration of each dynamic load g 00

test was about 30ms also These facts show that the LL -05
excess pore pressure is not accumulated during -1 0 (a)TNOWAVE \\as use
successive pile driving, resulting in the small set-up -1 5
ratio of 1 2 for this site 0 10 20 30 40
The set-up ratio for steel pipe piles driven in clay Timc. t (ins)
grounds is in the range of 3 to 5 (Wakiya et a1 1992, 20
Matsumoto et a1 1995) In these cases, the 15
h

existence of the accumulated excess pore pressures z 10

was confirmed It may be thought that the set-up is
completed just after the end of each driving for
E 05
2 00
sandy ground r - ~-0 5
-1 0
4.4 Wave matching ctimlyses to estmiate the static -1 5
load-d s p l n c e ~t ~cim'es
l 0 10 20 30 40

Wave matching analyses of the re-driving test Time, t (ms)

signals recorded at 6 days and 30 days after the Figure 10 Coinpanson betu eeii ineasured and calculated
upv ard tral eling forces in pile D (6 daj s after EOID)

587
 Shaft resistance, z (kN/niL) Shaft resistance, z (kN/m’)
oO 100 200 300 0 100 200 300
(a) 6 days after EOID (b) 30 days after EOID
Figure 11. Distributions of shaft resistance estimated from re-
driving tests using TNOWAVE. together with static load test
results (piles D and Sj.
Figure 12 Load -displacement cunes obtained from static
load tests. and denved froin d!naniic load test using
CAPWAPC and TNOWAVE (pile Dj
Figure 11 shows the distributions of the shaft
resistance, z, estimated from the wave matching
analyses using TNOWAVE and measured in the
static load tests. Although the total shaft resistance
estimated from the wave matching analysis is
comparable with the static load test results, the
588
calculated distribution of the shaft resistance along
the pile shaft does not coincide with the static load
test results.
The distributions of the shaft resistance, z,
estimated from the wave matching analyses using
CAPWAPC were similar to Figure 11. In these
wave matching analyses, the traditional empirical
soil model proposed by Smith (1960) was used. The
maximum soil resistance, the spring value and the
damping constant for the Smith model were
estimated only from the agreement between the
calculated and measured signals. These factors may
well be the cause of the difference between the
estimated and the measured soil resistance as shown
in Figure 11.
The static load-displacement curves estimated
using the wave matching analyses are compared
with the load-displacement curves measured in the
static load tests in Figure 12. The estimated load-
displacement curves are comparable with the
measured curves. Especially, the estimated initial
pile head stiffness is fairly coincident with the
measured values. This fact is usefbl in the limit
states design and the performance based design of
pile foundations, in which the estimation of the load-
deformation relation will be a vital issue (Kusakabe
1998).
5 STATNAMIC TEST RESULTS
A loading device having a loading capacity of 8 MN
was used in the Statnamic test conducted on pile S
52 days after the initial pile driving
The measured variations with time of force, F,[,,
displacement, w ,velocity, 11, and acceleration, a, at
the pile head during the Statnamic test, are shown in
Figure 13 The loading duration is about lOOms
The peak ofF,,, is 3 68 MN, the maximum pile head
displacement is 37 mm, and the residual
displacement is 16 mm which corresponds to 4% of
the pile diameter The maximum downward
velocity is 1 m/s The maximum downward
acceleration is 60 m/s2 while the maximum upward
acceleration attains 120 m/s2 The FAln- w curve is
shown in Figure 14
The pore pressures were measured during the
Statnamic test also The magnitude of the pore
pressure was very low, o 0 1MN/m2 in maximum
The Unloading Point method analysis (Kusakabe
& Matsumoto, 1995) and the wave matching
analysis of these Statnamic test signals were
conducted to derive a static load-displacement curve
for the pile The KWAVE program developed by
Matsumoto & Takei (1991) was used for the wave
matching analysis Figure 15 compares the derived
load-displacement curves and the load-displacement
curve obtained from the static load test The curves
compare well for practical purposes
 Pile h e a d load (MN)

0 50 100 150 200 250

i

-
m
ro
0 50 100 150 200 250 73
(d 30
.
.5i' 1.0
E
v
0.5
I
:
-ei-
a,
"- 0.0 40
0' -0.5
0
_.
Figure 15 Load 4isplaceineiit cun es obtaiiied froiii the static
a, -1.0 load test, derir ed from Statnanic load test using Uilloadiiig
>
0 50 100 150 200 250 Point method (ULPM) and derived from the wave matching
anal! sis

capacity, Q, from the static load test, but

overestimate the end bearing capacity, Qp, and
underestimate the total shaft capacity, Qy. It is seen
0 50 100 150 200 250
Time, t(ms) that the dynamic load test predicts well the total pile
capacity as well as the proportions of the total shaft
Figure 13. Statnarnic test signals of pile S. capacity and the end capacity.

Pile head force, Fll,,(MN) 7 CONCLUSIONS

0 1 2 3 4
The results of a comparative study of the static load
test, the dynamic load test and the Statnamic test on
three open-ended steel pipe piles driven in relatively
uniform sandy ground have been presented, and the
uses of the dynamic load test and the Statnamic load
test in sandy ground as an alternative to the static
load test has been discussed in this paper. The
following conclusions were derived from this study:

The hammer driving energy should be enough,

40' " " ' "
Figure 11. Frl,,versus II'
6 RELIABILITY OF DYNAMIC LOAD TEST
The ultimate bearing capacity of the test pile
obtained from the static load test is compared with
the bearing capacity derived from the dynamic load
test and various pile design codes in Figure 16. OIn
Figure 16, Q is the ultimate capacity which is the
sum of the ultimate end bearing capacity, Qp, and the
ultimate shaft capacity, 0,.
Four Japanese design codes, which are based on the
SPT N-values, predict well the total pile
so that the pile capacity is fully mobilized. A
method to select an appropriate hammer driving
energy was proposed, in which the mobilized
static resistance as well as the measured set per
blow are utilized to determine the minimum
required driving energy (see Figure 6).
Excess pore pressures are generated during pile
driving even in sandy ground. However, the
generated excess pore pressures dissipate within
the duration of each driving without
accumulation of residual excess pore pressures.
Therefore, the 'set-up ratio' is relatively low in
this case study. The set-up phenomena was
completed within 60 min after the end of the
initial driving process.

589
 0.0 0.5 1.0 15 20 2.5 3 0 35 4 0 4.5
Measured b! SLT (6 da! s)
Measured by DLT (6 da) s) I//
JHA S P T l V
JAC SPTl / 1
JR SPT
JPH SPT
CPT method
AASHTO S P T l I ]
~

AASHTO CPT
1
Hilc! I\ ith poicniial hamiiicr cncrg! 1 L.,'+C),
Hilc! v, 1111 iiicasurcd liainincr cncrg\- Q +<I, I * I 1 1 1 1 1 1 1 1 1 , , , 1 , , , ,

00 05 10 15 20 1 5 3 0 35 40 45
Ultimate End and Shaft Capacities (MN)
JHA Japan Road Association J A C Arch~tecturalInstitute of Japan
JR Japan Railna! JPH Japan Port and Harbor Association
Figurc 16 Coiiipanson of end and sliafl capacitics obtained from the static load test n i l h Iliosc cstmialcd from d! naiiiic load test
and dcm cd froni 1;irious p ~ l cdcslgn codes

The load-displacement curve derived from the REFERENCES

dynamic load test is comparable with the curve
obtained from the static load test, if the dynamic
load test is performed after the same time
intervals as the static load test, measured from
the initial pile driving. However, the
distribution of the shaft resistance and the toe
resistance derived from the dynamic load test
did not coincide with the static load test results.
Estimation of soil parameters and selection of
the soil model used in the wave matching
analysis need hrther study.
Goble. G G . G E Jr Liluns. & F Rausche 1975 Bearing
Capacih of Piles from D! iiaiiiic Measurements Final
Report, Dept of Civil Eng Case nestern Rescne Unn . .
Cle! eland. Ohio
Kusakabe. 0 1998 Changing foundation design code and the
role of Statiiaimc test, Statnaiiuc loadmg test '98. Proc b 7 d
h r S'rntlinim Seniiiinr, Tok! 0. Japan Balkema (to be
published)
Kusakabe. 0 & T Matsumoto 1995 Statnamic tests of Shonan
test program nith re\ ieir of signal interpretation. Proc 1Y f
hit Stntiiainic Seiiiiiinr 113-122 Vancom er
Matsuinoto, T & M Takei 1991 Effects of soil plug on
beha\ lour of dn\ en steel m e mlcs Soils K- Fouiidatioiis. ' I 1

The load-displacement curve derived from the JSSMFE, 3 l(2): 14-34

Unloading Point method analysis of the
Statnamic test signals is also close to the curve
obtained from the static load test.
The ultimate bearing capacity and the Qp/Qy
ratio derived from the dynamic load test are the
most accurate, compared to empirical pile
design formulas available in Japan and the other
countries for this case study.
This study is an activity of the Committee on
Methods for Pile Capacity Prediction (formed in the
Japanese Association for Steel Pipe Piles from 1991
to 1995 led by Professor Kazuma Uto, Tokai
University) and the Research Group on Rapid Pile
Load Test Methods (formed in Japan in 1993 led by
Professor Osamu Kusakabe, Tokyo Institute of
Technology). The members of these groups and
Sumitomo Metal Industries Corporation are greatly
appreciated for their valuable discussions,
suggestions and support for this study.
Matsumoto, T., Y.Michi, & T.Hirano 1995. Performance of
axially loaded steel pipe piles dri\,en in a soft rock. .low: of
Geotecli. Eiig., ASCE. V01.121. No.4: 305-315.
Nishimura, S. & T.Matsumoto 1995. Wave propagation
analysis during Statnamic loading of a steel pipe pile. Proc.
1st Int. Stntnnniic Seminar: 23-3 3 . Vancouver.
Smith. E.A.L. 1960. Pile dnving analysis by the nave
equation. Jour. Soil iifecli. Foiiiid. Div.. ASCE. Vol.86.
No.SM4: 35-61.
Wakiya, Y.. 0.Hasluinoto. M.Fukuwaka. T.Oki &
H.Sliiiioiniya 1992. Abiliq of dynamic testing and
evaluation of bearing capacity recovery from exess pore
pressure measured in tlie field. Proc. of 4th Int. Cot$ oii the
.4pplicatioii of Stress-Wave Theory to Piles (B.J Barends
Eds.): 665-670. The Hague: Balkenia
Wakiya, Y., M.Hayashi, T.Katayama & Y.Kobayashi 1994.
Experimental inspection on accuracy of pile capacip
prediction in dynamic loading test. Proc. 29th Japnnese
.Aiinunl Meeting of Soil Mecli. B Found. Eiig.: 1431-1434
(in Japanese).
Waluya. Y.. K.Nisliunu, M.Hayashi. A.Shibata & S.Nis1iinura
2000. The case studles of dynamic load test in Japan. Proc.
6th Itit. Coiif: OM the .Application of tlie Stress- W m e Tlieoy
to Piles, Sao Paulo. Brazil (to be published).
Yoshizawa. K., N.Kawabata. T.Oki & A.Shibata 1994. Vertical
loading test of steel pipe pile dnven into sand focusing on
set-up characteristics. Proc. 29th .Japanese ilnnual Meeting
ofSoil-Wech. B Fhunrl. Eiig.: 1429-1430 (in Japanese)

590
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Case studies of statnamic load testing in Japan

S. Nishimura - Fugro Geoscience Company Limited, Tokyo,Japan

T. Matsumoto - Department of Civil Engineering, Kanazawa Universi~, Japan
0.Kusakabe -Department of Civil Engineering, Tokyo Institute of Technology,Japan
K. Nishiumi -Nippon Steel Corporation, Tokyo,Japan
Y.Yoshizawa - SuniitonzoMetal Industries Linzited, Tokyo,Japan

ABSTRACT: More than 70 Statnamic tests have been carried out on various types of piles in various soil
conditions with various construction methods, since the Statnamic test’s first application in Japan in 1992.
This paper reviews the Statnamic tests in Japan, presenting their chronology and statistics. This paper also
discusses the a plicability of the Statnamic test to the estimation of the static load-displacement curve for a
pile, based on tEe accumulated results of the Statnamic tests in Japan.

1 INTRODUCTION in Japan due to the simplicity and speed of the

Foundation design in Japan is beginning to change
from the conventional allowable stress design to the
limit state design, or the performance based design,
following the global trend in foundation design. In
the new design methods, accurate estimation of the
load-deformation curve as well as quality control of
pile foundations will become an issue of vital
importance (Kusakabe 1998). Hence, quick, cheap
and yet accurate pile load test methods have been
sought in Japan. The dynamic load test and the
Statnamic test may be regarded as two of the newest
developments in construction technology in this
area.
More than 70 Statnamic tests have been carried
out on various types of piles in various soil
conditions with various construction methods, since
Statnamic test’s first application in Japan in 1992.
This paper reviews over the Statnamic tests in Japan,
presenting their chronology and statistics. This
paper also discusses the applicability of the
Statnamic test to the estimation of the static load-
displacement curve for a pile, based on the
accumulated results of the Statnamic tests in Japan.
testing procedure. Statnamic test devices of 8, 16,
30 MN loading capacities are currently available in
Japan. They can cover a range of loading force from
3 to 30 MN. The time required for pre-test
preparation of the Statnamic test is shorter, while
time requited for decomposition of the loading
device is longer, because gravel is used to catch the
launched reaction mass. Introduction of a system for
catching the launched mass such as the hydraulic
catch mechanisms (Bermingham 1995) would be
desirable to shorten the total time required for the
Statnamic test.
Statnamic loading has been also applied to lateral
loading of two steel pipe piles (Tada et al. 1997).
3 CHRONOLOGY AND STATISTICS OF
STATNAMIC TESTS IN JAPAN
Table 1 summarizes the Statnamic test conditions of
the total of 42 sites in Japan. Kusakabe (1998) has
reviewed statistics concerning these Statnamic tests
using the database listed in Table 1. Some statistics
are added in this paper.

3.1 Nuniber. of Stalnamic tests and test sites

2 STATNAMIC TEST DEVICES AVAILABLE IN
JAPAN
The first use of the Statnamic loading method
(Bermingham & Janes 1989) in Japan was the load
test of a cast-in-situ concrete pile in 1992
(Chosokabe et al. 1993). The Statnamic load testing
method has attracted the interest of piling engineers
Figure 1 shows the chronological variation of the
number of Statnamic tests conducted over the past 7
years, indicating a sharp increase in 1994 and a
steady increase of 6 to 8 tests per year after that. A
total of 74 Statnamic tests have been carried out at a
total of 42 test sites.

591
Table 1. Achievement of Statnamic tests in Japan
Date Nuin- Num. Pile Pile 'Ianned Measured
No (year- Pile tyye ber of of test diam.0 length Soil Type Notes
111011t11) tests piles (in) O(m)
5,0 Sand I Sandy Friction cut: 9m
1 92-5 CISC 4 3 1.3 14.5 5.0
mavelE
v
DLT
2 92-11 Dri1,en SPP 2 2 0.8 11.0 5.0 5.7 Soft rock SLT. DLT
2 1.4 13.5 80 Reclamation / Sand Friction cut: 13.5m
3 93-5 CISC 1 1.2 13.5 6.0 6.3 C
Silt sandy gravel Friction cut: 39.5m
3 93-6 BoredPHC 1 1 0.8 54.0 5.7 j,7
SLT
5 93-10 Driven SPP 1 1 0.4 15.0 4.5 3.7 Sand SLT.DLT
6 93-1 DrivenSPP 2 1 1.4 25.5 2.51
0.11
Mudstone Ld = 13m
94-1 DrivenPHC 5 2 0.3 7.0 0.300.6
Clay / Sand SLT
BoredPHC 6 3 0.3 7.0 0.300.6
8 94-2 CISC 1 1 1.5 10.0 13.0 12.1 Silt/Sandygravel SLT
9 94-3 DSPP 1 1 1.4 29.5 5.0 Mudstone Ld = 131n

21 96-4 CISC 1 1 0.8 47.5 10.0 10.4 Sandy/ Silw Frictionpile

1 1 1.6 29.0 10.0 11.5 Shale
22 96-6 Driven SPP 1 1.6 30.0 0.3 Sandstone
Shale Lateral STN
1 1.6 29.0 0.3
23 96-9 CISC 1 1 1.0 12.5 8.0 8.0 Sandy soil Recycled pile
24 97-2 CISC 1 1 1.5 9.0 13.0 13.7 Aeolian rock
25 97-3 CISC 2 1 1.0 13.0 7.0 6.3136.9 Soft rock
26 97-4 BoredSPP 1 1 0.9 18.0 13.0 Sandy gravel
1 1.2 19.5 15.0 15.4
27 97-5 Driven SPP Mudstone DLT, Lateral SLT
1 1.2 21.5 15.0 15.3
28 97-5 Bored SPP 1 1 0.6 20.0 8.0 Sand
29 97-5 CISC 1 1 1.2 13.4 16.0 16.0 Sandygravel SLT
30 97-6 DrivenPHC 1 1 0.6 11.0 3.5 3.7 Sandstone SLT, Lateral SLT
1 1 1.4 70.0 16 15.5 Sandy gravel DLT
31 97-7 DrivenSPP 1.4 36,0 12
97-9 12.3 Slate DLT
~

32 97-10 CISC 1 1 1.2 6.0 13 13.0 Sandstone I Shale SITGSLL

33 97-10 BoredPHC 1 1 0.6 28.0 4.5 3.8 Cemented silt 0
34 97-12 DrivenPHC 2
-- 12.0
12,0
5 5.2
5,0 Sand / Hard clay Recycled pile
35 98-1 Micropile 1 1 0.18 21.5 1.9 Fine sand SLT. DLT
1 1.5 46.6 24.5 23.8
36 98-3 DrivenSPP 3 1 1.5 46.6 24.5 27.2 Gravel sand DLT
1 1.5 46.7 24.5 22.8
37 98-7 DrivenSPP 1 1 1.0 36.5 7 8.0 Calcareous rock DLT
38 98-8 CISC 1 1 1.2 13.5 16 15.2 Sandy gravel
39 98-8 CISC 1 1 1.2 14.5 16 15.3 Sandy gravel
40 98-10 CISC 1 1 1.2 8.0 12 Soft rock I
Weathered rock
41 99-2 CISC 1 1 1.2 20.8 20 Sandy gravel
42 99-3 BoredPHC 1 1 0.8 19.0 2
Cemented silt
BoredPHC(SC) 1 1 0.6 22.0 2
CISC: cast-in-situ concrete pile SPP: steel pipe pile PHC: precast concrete pile SC: steel-concrete composite pile
DLT: dynamic load test SLT: static load test SIT: sonic integrity test Ld:embedded length

592
3 . 2 Test objectives
Figure 2 presents a breakdown by objective of
Statnamic tests at the test site. The Statnamic tests
have been conducted to estimate the bearing
capacity and load-displacement curves of newly
constructed piles in 57% of the test sites, and of
existing piles in 12% of the test sites. The Statnamic
test was employed in 14% of the test sites for the
development of new pile construction methods. The
remaining were research projects, concerning
aspects of the Statnamic test itself such as loading
characteristics, dynamic effects, applicability to the
estimation of static load-displacement curves and so
on.
Figure 2 Breakdown of the site number nith respect to test
objective.
3.3 Pile ~ p e s
The breakdown of the number of sites with respect
to the kind of pile is given in Figure 3, where steel
pipe piles have been tested at 43% of the test sites,
cast-in-situ concrete piles at 38%, and precast
concrete piles at 17% of the test sites, suggesting a
rather even distribution over each kind of pile.
There was one site where a micro pile having a
diameter of 0.8m was tested. Out of 18 test sites
with steel pipe piles, 10 sites were offshore piling
projects where execution of a static load test would
have been rather difficult.
593
3.4 Pile length
Figure 4 presents a breakdown of the number of
sites with respect to pile length and pile type. Large
shares of precast concrete piles having a length less
than 10m and cast-in-situ concrete piles having a
length less than 20 m are seen, whereas most steel
pipe piles range in length from 10 to 40 m, 70 m in
maximum. The longer lengths of the steel pipe piles
can be attributed to the fact that most steel pipe piles
are offshore piles, while all the concrete piles are
constructed on land.
Figure 4. Breakdown of the nuniber of tests with respect to pile
length and pile type.
3.5 Soil types
Figure 5 shows the breakdown of the number of
tests with respect to the soil type of the bearing
stratum and the kind of pile type. Most piles are
seated on firm soil layers such as rock, gravel, sand,
soft rock or indurate silt. This is because of the fact
that pile design codes in Japan recommend that the
pile toe be seated on stiff soil layers. The bearing
capacity of piles seated on rock tends to be
underestimated in the pile design codes, so loading
tests are often conducted on piles seated on rock to
obtain realistic pile bearing capacities for design.

Figure 5. Breakdown of the test iiuinber with respect to soil
type and pile hpe.
3 6 Load magnrtirdes
Figure 6 presents a breakdown of the number of
tests with respect to maximum test load and pile
type Maximum loads are less than 16 MN for most
piles in each pile type Note especially that the
maximum loads are less than 8 MN for all precast
concrete piles, because these piles were installed by
means of a pre-auger method and had relatively low
bearing capacities However, there is a clear trend
of increasing testing load in recent years, an average
of 4 88MN in 1996, 9 44MN in 1997, 16MN in
1998 and 13 3MN in 1999
594
A breakdown of the site number with respect to
planned maximum load is given in Figure 7. Most
tests were conducted to confirm whether piles had
design ultimate capacities or not. A merit of
conducting a load test is that a safety factor of 3 can
be reduced to 2.5 if a load test is conducted.
Statnamic tests were conducted at only 3 sites to
obtain bearing capacities, indicating that pile
designers in Japan stick to pile design codes
excessively.
3.7 Loading diu-atron
A typical loading pattern of the Statnamic test is
shown in Figure 8, with the definition of loading
duration, tr, The relative loading duration, 7;, is
defined as follows (Karkee et a1 1997).
T, = t L l ( 2 L l c )
in which L is the pile length and c is the bar wave
velocity
The relative loading duration, T,., is a measure for
the influence of stress-wave propagation phenomena
in the pile during the Statnamic loading. The
Research Committee on Rapid Pile Load Test
Methods formed in the Japanese Geotechnical
Society (JGS) in 1996 regards the rapid pile load test
as a test in which stress-wave propagation
phenomena are negligible, although other dynamic
effects such as the inertial force of the pile as a
mass, radiation damping and viscous damping of the
soil are not. They proposed T,. = 10 as a boundary
between dynamic loading and rapid loading based
on the work of Nishimura et al. (1998) (Research
Coiiimittee on Rapid Pile Load Test Methods 1998). In
the draft of “Method for Rapid Load Test of Single
Piles (JGS 1815 - 200X)”, the rapid load test is
defined as a load test with 5 < T,. < 500. A load test
with 7;.greater than 500 is regarded as a static load
test (JGS Committee on Standardization of Pile
Loading Tests 1998)
As seen from Figure 9, most Statnamic tests
excluding 3 tests, had the relative loading duration,
Tr, greater than 5 , satisfying the requirement for
rapid loading
z
z-
-U
CCI
-
0
-0
-a,
.-
Q
Q
Q
0 30 60 90 120 150
Time (ms)
Figure 8. Definition of loading duration.

Figure 9. Breakdown of the test number with respect to
relative loading duration and pile type.
The Unloading Point Method (Midendorp et a1
1992, Kusakabe & Matsumoto 1995) is usually
employed to estimate a static load-displacement
curve for the pile in the case of Tr > 5 . In cases of
Tr < 5 , the wave-propagation analysis (wave
matching analysis) is usually conducted to estimate
the pile capacity and the corresponding static load-
displacement curve (for example, Ochiai et a1 1996).
3.8 Pcrr.allel test nielhods
The types of load tests other than the Statnamic test
conducted at the test site are given in Figure 10 in
the form of a breakdown by the number of sites
The Statnamic test (STN) was carried out alone at
56% of the test sites, whereas dynamic load test
(DLT) and/or static load test (SLT) were conducted
in parallel at the remaining sites. At most test sites
where the Statnamic test alone was conducted, it
was confirmed that the pile had bearing capacity
greater than the design load that was derived from
pile design codes.
On the other hand, Statnamic test results were used
for calibration of a dynamic load test at the test sites
where both load tests were carried out for quality
assessment and driving control of the other piles at
Figire 10 Breakdonn of site number uith respect to load test
methods
595
the sites (for example, Terada et al. 1998). Static
load tests were conducted mainly for comparative
research of static load tests and the Statnamic tests.
4 LOAD-DISPLACEMENT BEHAVIOR OF
STATNAMIC TESTS
4 1 Load-dispIacemerif curves
Load-displacement curves from 5 1 Statnamic tests
out of the 72 tests listed in Table 1 are overviewed
to grasp the responses of the piles and the ground
during the Statnamic tests with a wide range of
loading forces
The relationships between the pile head force,
Fstn, and the pile head displacement, w, are shown in
Figures 11, 12 and 13 for the cast-in-situ concrete
piles (CISC), the steel pipe piles (SPP) and the
precast concrete piles (PHC), respectively In these
figures, Fsttl is normalized by the maximum pile
head force, I ~ ~ t ~ land
( zw( m ~~),
is normalized by the pile
diameter, d The F&v curves may be classified
into two groups In one group, the F\t,,-w curves
have a characteristic 'tear drop' shape in which the
pile displacement, w,tends to increase and decrease
with increasing and decreasing the pile head force,
FqtI1, regardless of the kind of pile In the other
group, the Fstn-~i' curves have shapes far from the
'tear drop' shape In the Statnamic tests classified
into the latter group, the pile displacement continues
to increase after the start of unloading of I i m ,
although the recovery of the displacement occurs at
a later unloading stage of F,t,, Such behavior of the
J;stn-wcurve is notable for cast-in-situ concrete piles
which have relatively large masses compared to
steel pipe piles and precast concrete piles
According to single mass modeling of a pile
during Statnamic loading (Middendorp et a1 1992),
the total soil resistance, Fsoli,is given as
I;'. -1-
soil - 'stn - M a
in which A4 is the mass of the pile, and a is the
measured acceleration at the pile head. The total
soil resistance, Fsoil, is thought to be the sum of the
static soil resistance, F,,, and the dynamic soil
resistance, Fv, which is dependent on the pile
penetration rate.
The Fsoil - 1.1, curves of the Statnamic tests are
shown in Figures 14, 15 and 16. Note that F'soil is
normalized by the yield load of I'soil, FSoil(y1, that is
defined as Fsoilat the maximum curvature of Fsoil - w
curve. For most piles, the pile head displacement
increases and decreases following the increase and
decrease in Fsoil. It can be seen that the influence of
the correction of the measured pile head force by the
pile inertia, Ma, is predominant for the cast-in-situ
concrete piles with large residual displacements.
The Fsoil - IQ curves are again classifiable into two
groups: the group with elastic responses in which the
I”soli- it’ curves have the characteristic ‘tear drop’
shape, and the group with plastic responses in which
plunging loads are clearly detected in the F s o -~ w~
curves. It may be judged that the pile exceeded the
‘static’ yield load in the Statnamic tests classified
into the group with plastic responses, although the
dynamic soil resistance, F,,, which depends on the
pile penetration rate, is not corrected for.
If the classification o f Statnamic test results into the
596
elastic or the plastic response type is possible from
the measured pile head displacement immediately
aRer the completion o f the ~tatnamictest, this may
be useful An overview of the measured pile head
displace~entsis provided below for this purpose
The maximum and residual pile head displacements
are shown in Figures 17, 18 and 19 for cast-in-situ
concrete piles, steel pipe piles and precast concrete
piles, respectively. In these figures, the Statnamic
tests which were categorized into the group of
plastic responses are marked with a star symbol. It
can be seen that the Statnamic tests classified into
 displacenients are small with ~ ! /typically
d less than
0 5%
As the number of the Statnamic tests on
prestressed concrete piles at actual sites is still low
(Tests 7-1 to 7-10 were performed by the Research
Group on Rapid Pile Load Test Methods at the
Shonan test site Kusakabe & Matsumoto 1995), it
is difficult to determine the relationship between pile
displacement and the classification of the Statnamic
test for precast concrete piles at the present stage

3 3 Rehtiorishrp heh,rveri niaxiniuin F,,,, ~?~cixiiniini

I.,o,[ utrd ut?/oudulgYolrlt fi?fW,I T i LP.
Figure 20 presents the relationships between
measured maximum loading force, F,t,,c,,,,,), and
maximum total soil resistance, F,o,~(llld\) Definitions
of f*~111(,,18~)and J*qoll(lnn,)
are presented in Figure 23
The maximum total soil resistances, I*col~(,lla~), are
almost equal to the maximum loading forces,
l~rtn(nln~),
for the wide range of I ~ ~ ~ t nIt~means
l , l a \that
)
the amplitude of f ~ ~ ~ l ~is( transferred
l l l n ~ ) to the ground
effectively
Figure 2 1 presents the relationships between
and the Unloading Point force, fit ~ p which
l~,o,~~l,la\~ ,
is defined as I,\ollat the maximum displacement on
f*<ollvs w(see Figure 23) The Unloading Point force
is thought to be the maximum static resistance of the

the plastic response group tend to have larger

maximum head displacements as well as larger
residual displacements than the Statnamic tests of
the elastic response group for all kinds of pile.
For the Statnarnic tests of the plastic response
group, the amplitudes of the elastic displacement
(difference of the maximum disp. and the residual
disp.) seem to be similar in each kind of pile.
The Statnamic tests on the cast-in-situ concrete
piles are clearly divided into two groups: the
Statnamic tests with large maximum displacements
and the Statnamic tests with very small maximurn
displacements. This fact may reflect the soil
conditions of the bottom of the boreholes prior to
concreting.
For the steel pipe piles, 11 piles out of the 19 piles
were classified into the plastic response group,
although the amplitudes of the residual

597

______------
pile obtained in the Statnamic test For the
Statnamic tests classified into the elastic response
group, F,o,l(md,) almost equals FI I P for most cases
On the contrary, for the Statnamic tests classified
into the plastic response group, t(\ol~(l,,d\)exceeds
FITp by 5 to 30% of F1.1 p, indicating that sufficient
load was applied to obtain the static pile capacity
The difference between F\ol~(,nrl\) and FIIIP may be
regarded as the dynamic soil resistance which
depends on the pile penetration rate
Figure 22 presents the relationships between
and k i ~ pThis figure indicates that the
k~~l,~llla,)
planned maximum force could be determined to be
I I to 1 3 times the expected static pile capacity
5 CONCLUDING REMARKS
An overview of the case studies of the Statnamic
tests conducted in Japan has been presented in this
paper, showing their chronology and statistics. The
results of the Statnamic tests also have been
reviewed from the view point of whether the
ultimate capacity of the piles was obtained from the
Statnamic test or not. It was shown that the ultimate
pile capacity was obtained in a large number of the
Statnamic tests, encouraging the use of the
Statnamic test to estimate the design pile capacity.
598
This paper owes much to the activities of the
members of the Research Committee on Rapid Pile
Loading Test Methods in JGS, and permission to use
the Statnamic test results which was given by the
users of the Statnamic test in Japan. The authors
wish to express their sincere gratitude to them here.
REFERENCES
Bcriniiigliain. P & M Jams. 1989 An innoLatii c approach
for load tcstlng or high capacltj piles PIW 117r (‘017/ 017
Piling arid Deep Foundaliont London 109-413
Bcnningliaiii P . 1995 Futurc applications of tlic Statnamic
method of testing foundations, l’roc 1ct 1/11 Smrnrrtnic
Senirnnr, Vaiicom er, Canada 255-258
Chosokabe, M . K Yanasluta, M Kakurai. T F ~ i k u h a& T
Yanxida, I993 A Statnanuc loading test applied for a
cast-in-situ concrete pile, l’roc .lnniinl .Ueetii7g of
~ lrchiteciurnl I n criiiire of Japnii I76 1-1762
Karkcc. M T Horiguclii & H Kisliida 1997 Static and
dynainic icsts for c\dluahoii of tlic \crlical load bcanrig
capacit! of pilcs. Proc Deep Foii17tlatio17r I I ~ W U 2nd I~
1nniial 1Ieniher ’ 5 ( ‘014ere17ce.Toronto. Canada 199-214
Kusakabc. 0 . 1998 Changing roundation dcsign codc and ihc
role of Slatnainic tcsi. Statnainic loading test Proc ,?lid
hit Sinmnmic Seiniiinr, T o h 0. Japan Balkeina (to be
published)
Kusakabc. 0 & T Matswiioto. 1995 Statnainic tests of
Slionan tcst prograin L\ 1121 rck icw of signal intcrprctation.
Proc Fir$/lilt \tntnannc Setmnnr Vancow cr 1 13-1 22
Middcndorp. P P Bcnningham & B Kuipcr 1992 Statiiamic
load tcstirig of fot1ndatlon pilcs. P i w J t h 1171 (‘oi7J on
/he 1ppliCntiori o/ 5ti-et+Hki e T17eon to Pile5 (B J
Baicnds Ed ). Tlic Hague Thc Nctlicrlaiids 581-588
Rot t crdaiii. Ba lkcnia
Nisluinura. S . A Shibata, K Yaiiwshita, N K ~ t aN Ogita &
M lsluda. 1998 One &ineiisional stress wile simulahon
malj sis of Stahiamc load tests, Statnanuc loading test,
Pmc 2 ~ dhit Srntmmrc Senrincw T o h 0, Japan
Balkeina (to be pnnted)
Rcscarch Connnittcc on Rapid Pilc Load Tcst Mcthods. 1998
Rcscarcli Actii itics tovard Standardmtion of Rapid Pilc
Load Tcsl Mctliods i n Japan. Statriunic loading tcst
l’roc 2nd lnr Sfntnniiiic Seiirrnar. Tolq 0. Japan Balkeiiia
(to be published)
JGS (Japanese Geotecluucal Socieg ) Conuiiittee on
Standardization of Pile Loaduig Tests. I998 Draft of
“Method for Rapid Load Test of Single Piles (JGS 1815 -
2000)”. Statnrunic loadmg test. l’roc 2nd Int S’tntnrnrr~c
Senrmar. Tokyo. Japan Balkcma (10 bc publishcd)
O c h m H . 0 Kusakabe. K Sum, T Matsuinoto & S
Nisluniura. 1996 Statnanuc Tests on Offshore Steel Pipe
Piles for Foundations of Access Bridge for New f i t a -
Kjushu Airport, Proc 4th 1171 Conf on the. fpplimnfion of
Sfress-Ifirve Iheorv to Piles (B J Barends Ed ), The
Hape. The Netherlands 997- 1014 Rotteddain Balkeina
Tada. K . M Ohnishi. Y Cliono. S Nishiniura & S Sakaiiioto.
I997 An cffcct of loading Lclocity r a m at latcral
loading test for a steel pipe pde. l’roc 321id -lni?iinl
Meerrng, Japanese Geotecluucal Societj . Kuinanioto
1601-1602 (in Japanese)
Terada, M . R Nakatsuka, K Olmiwa, N Akisato, K
Honkoshi & S Nishiinura, 1998 Applicafion of the
Statnamc loadng test for large-scale pier construcfion
Statnarmc loadmg test. Proc 2nd Int Statnainrc Seminar.
Tokyo, Japan Balkema (to be published)
 Application of Stress-Wave Theory to Piles, Niyama & 5eim (eds)02000 5alkema, Rotterdam, lS5N 90 5809 1503

Statnarnic and dynamic load tests for large diameter steel pipe piles supported
by a thin bearing layer at Nagoya port in Japan

Y. Kikuchi
Port and Harbour Research Institute, Yokosuku, Japan
S. Nishimura
Fugro Geoscience Company Limited, Tokyo,Japan
M.Tat suta
Nippon Steel Corporation, Tokyo,Japan

.-\BS'TR,ICT: Nagoya port is located in center of Japan and knows as one of the most active ports. Steel pipe
piles were adopted for the development o f the wharf at the west area of Nagoya port. According to the
expericncc o f Great Hanshin Earthquake. the seismic safety factors were considered !or the structural design.
To save the construction cost, it was planned to install large diameter steel pipe piles into the thin gravelly
sand stratum. Series of Statnamic and Dynamic load tests were carried out to estimate the bearing capacity of
!he piles in order to establish a site-specific driving thrmula used for driving control of production piles at the
site. I'his paper discusses the parameters !'or the design o!. [he toe capacity of piles that had seated on the thin
gravelly sand stratum as well as shalt fiiction o f upper alluvial clayey stratum. And interpretations o f
Statnaniic and Dynamic tests with its results are described.

i INTRODUCTION All tcst piles had a length of50m and diameter of

1.5ni. 'The wall thickness o f the pile was 17mm at
At Nagoya Nishi-5 wharf (- 14m). locating near the pile head and 15mm at the toe. The main
Nagoya City in Nonshu Island in Japan. the steel purpose of the test was to confirm the design
pipe piles were adopted for founding the wharf capacity of the pile as well as to establish the site-
structures. To achieve economical construction. the specific pile driving control procedure to be applied
piles were designed to drive -40.0 m below N.P. for the driving ofother production piles in the site.
levei so that the piles would be seated on the
gracclly dense sand stratum in spite of' penetrating
into the deeper diluvial sand stratum existing the 'Table I ~
- . General_information
_ on soil layers. -~
depth of' -68m. According to soil investigation IIIc'iation Soil l'liick- Remarks
rcsults. the soil stratification can be classified as 111) t\ne nzss
shown in Table 1 . A thickness o f the gravelly sand
stratum was relatively thin. varying from 4.5 to
h.5m. as a bearing stratum against expected toe
rcs i stance.
To compare the toe resistances of' the t w )
dif'fcrent t y e s of' open-ended pile. which had just
seated on the then bearing stratums. Statnaniic and
Dynamic tests were c'arried out on the three test
piles. One of the pile toe designs was simple open
ended and other installed the cross-rib plates at the
toe The instrumentation along the pile shaft was
arranged in accordance with the soil stratification as
shown in Figure I . 'Two test piles, V-1 and V-3.
were simple open-ended pile and V-2 had the cross-
rib plates at the toe. The V-1 and V-2 piles were
seated on the gravelly sand stratum having 4.5 ni
thickness in order to evaluate increases of bearing
capacity o f toe resistance due to the cross-rib plates.
-.
I he V-3 pile was installed on the stratum hacing 6.5
m thickness t o evaluate the effect of the thickness of
bearing stratum to the toe resistance.

599
Figure 1. Schematic diagram of test condition.

2 ESTIMATING DESIGNED VALUES OF PILE Two cases were considered for the assessment of
BEARING CAPACITY shaft fiiction act on the pile. Because, soil behaviors
might be change after the pile load tests. The upper
The bearing capacity of the piles was derived by the soil deposit at the construction site had to be
following methods with some assumptions. removed till the depth of -1 5.0m and the production
piles to be driven in this site would be in the same
2.1 Designed vulue of sha$,friction condition including the test piles. Table 2 shows the
Equation 1 was established to determine cohesion calculated values of the design shaft fiiction.
variation from unconfined compression test results
in the upper alluvial clay stratum. The equation has Table 2. Design value of shaft friction.
been used for calculating the design value of the Thickness of layer (m) Design total
shaR fiiction, C. Case Alluvial clay
,aver Bearing layer
shaft friction
(KN)
C = 9.0 -t 3.02 (5 100 ) (kN/m2) Test 32.5 1.5
12240
pile (-8.0t0-40.0) (-40.0t0-41.5)
Product- 24.0 I .5
Where, Z= depth from N.P.=O (m). 10820
ion pile (-15.0to-40.0) (-4O.Oto-41.5)

600
2.2 Designed value of toe resistance Table 4. Designed pile toe resistance considering the effect of
thin bearing stratum.
In this design, the effects of soil plugging and of the
thin bearing stratum were evaluated and taking into Thickness of bearing Bearing capacity of pile toe (KN)
the pile toe resistance. Increase of toe resistance due stratum (m) Method (a) Method (b)
6.5 6170 3700
to the soil plugging effect was calculated by the
4.5 4760 3280
following two methods.

(a) The design code for steel jacket wharfs, in Japan

2.4 Effect of the cross-rib plate on pile toe resistance
a = 2 X . ( 5 +X)
Increase of pile toe resistance due to the effect of the
(b) Specifications for highway bridges, in Japan cross-rib plates were evaluated by using Yamahara’s
a = XI5 method (Yamahara, 1963). According to this
method, when the rib plate length is 1.5m, the
where resistance was found 1.5 times greater than that for
a = Coefficient of soil plugging effect; the pile without cross-rib plate
X = a ratio of an embedded length into bearing
layer / pile diameter, D (Xg5). 2.5 Comparison of designed and required capacities
Table 3 presents the values of toe resistance
calculated f?om the two methods. The method (a) Table 5 summarizes the designed values of pile
yielded larger value than the method (b). capacities calculated by the methods described
above. The total bearing capacity is shown in Table
5.
Table 3. Designed pile toe resistance.
Method fa) Method fb’l
Coefficient Table 5. Designed pile capacity.
0.333 0.200
of plugging effect 4.5 m thick 6.5m thick
Bearing capacity Shaft friction
6170 3700 for working piles 10820 I0820
of pile toe (KN)
fKN1
Toe resistance Open end Cross-rib Open end
(KN) 3280 4760 6170
2.3 Eflect of thin bearing stratum to pile toe Total
bearing capacity 14 100 1 5580 16990
resistance (KN)
Pile toe resistance decreases when a pile toe was Test pile v-1 v-2 v-3
seated on the thin bearing stratum. Toe resistance of
the pile in this case was derived by the method,
proposed by Hanshin Highway Department in Japan. The required bearing capacities as common and
In this method, the pile toe capacity is defined by the seismiclwind vertical loads for the production piles
Equation (4). have been given in Table 6 that shows the values of
designed capacities had enough allowance to the
(4) required capacities based on the designated vertical
loads.
where,
q d = Pile toe resistance at the thin bearing stratum; Table 6. Required bearing capacity.
q,= Pile toe resistance when the bearing stratum is Loading con- Designed Factor Required
continuous below the toe level; dition vertical load of capacitv
qd = Bearing capacity of weak layer existing below (KN) safety TKNj
Common 3500 2.5 8750
the bottom stratum; Seismiclwind 4500 1.5 6750
kf = Effective thickness of bearing stratum;
D = Diameter of pile.
Table 4 presents the toe resistance taking the 3 LOAD TESTING PROCEDURE
effects of the thin bearing stratums of 4.5m and 6.5m
thickness into the account. The table suggests that
the calculated result for 6.5m thickness did not 3.1 Load test sequence
reduce regardless of applying the Equation 4 in Statnamic and Dynamic load tests were carried out
comparison with the result of Table 3. on the three test piles to verify the bearing capacities
and to establish the site-specific driving formula for

60 1
installation control of other production piles by 3.2 Loud testing equipment
driving. Table 7 presents the specifications of the To meet the loading condition of the test piles,
test piles. Table 8 shows the load test sequence. Statnamic test was carried out by a 30MN loading
As shown in Table 8, Dynamic load test was capacity device manufactured by Berminghammer
carried out five times, three times and twice on the Foundation Equipment. A laser sensor was used to
V-I, V-2 and V-3 piles respectively. Statnamic test monitor displacement of the pile top during the
was carried out on each test pile 42 to 51 days after loading. The strain gauges and accelerometers (50G)
the end of driving. Pore water pressure was as shown in Figure 1 were used to monitor forces
measured just before and after the load test and accelerations of the test piles.
performed. Dynamic load test was performed by using IHC-
S90 hammer with ram weight of 70KN
manufactured by IHC that generates impact energy
Table 7. SDecifications of the test tiles.
of 90 KNm. However a fiee fall hammer with ram
Test pile v-1 v-2 v-3
weight of 50KN was used instead of the IHC
Layer thickness 4.5m 4.5m 6.5m
Pile toe Open Cross-rib Open hammer for re-driving after 23 days, 25 days and 28
Size $1 500mm x t 17/15mm(23/27m) x L5Om days to the V-1, V-2 and V-3 piles respectively.
Material Steel (Tension strength = 490 N/mm2) Force and acceleration responses during dynamic
load test were monitored by FPDS system, which
was developed by TNO.
Table 8. Load tests seauence.
Test d e V- 1 v-2 v-3
Laver thickness 4.5m 4.5m 6.5m 4 TESTRESULTS
Pile toe Open Cross-rib Open
Elapsed time after E.O.D. (days) 4.1 Dynamic load tests results
E.O.D. E.O.D. E.O.D.
2 days 1 days 23
As example of results from the dynamic load tests,
the measured force and velocity multiplied by
Dynamic load test 3 days 25 days
impedance curves were presented in Figure 2 and
4 davs
the measured and calculated upward traveling force
28 davs
wave curves in a wave matching analysis were
Statnamic test 51 days 49 days 42days presented Figure 3. They were obtained from the
test on the V-1 pile at the end of driving. Table 9,
10 and 11 present the results of signal matching
analyses of the dynamic load tests on the V-1, V-2
and V-3 piles respectively. Those tables show the
following analysis results.
At the end of continuous driving, toe resistance of
the V-3 pile with the cross-rib plates was greater
than that of the V-1 and V-3 piles without the cross-
rib plates. During elapsed time of 23 to 28 days, the
total capacity of those piles had increased about two
times of those at the end of initial driving. However

Photograph I . 30MN Statnamic loading device. Figure 2. Measured force and velocity (Vl-03, E.O.D.)

602
 Table 11. Results of signal matching analyses (V-3).
V3-03 V3-R 1
E.O.D. 23days
Sandy U.E. 1 .o 2.0
( I 1 .Om) I,.I:. 2.0 6.0
Clayey U.E. 4.0 18.0
(20.0m) L.E. 4.0 18.0
Yield Clayey U.E. 6.0 35.0
stress
(9.0m) L.E. 6.0 35.0
(kN/m?
’ Clayey 1J.E. 11.0 56.9
(4.34 L.E. 11.0 56.9
Sand U.E. 10.0 15.0
(2.31~11 L E . 10.0 15.0
Figure 3. Measured and calculated upward force (VI-03, Sand Toe 18992 23990
E.O.D.)
Total shaft resistance(KN) 1044 A6QA
Toe resistance (KN) 2702 2813
Table 9. Results of signal matching analyses (V-1). Total capacity (KN) 3746 7507
V1-03 VI-RI V1-R2 VI-R3 VI-R4
E.O.D. 2 days 3 days 4 days
Sandy lJ.E. 1.0 1.0 1.0 1.0 2.0 toe resistance of those piles were assumed not hlly
( 1 1.0m) L.E. 2.0 2.0 2.0 2.0 6.0 mobilized by dynamic loading due to the lack of
Clayey U.E. 5.0 5.0 5.0 5.0 20.0 impact energy by the hammer applied.
_ _...
(20.0m) L.E. 5.0 5.0 10.0 10.0 20.0
‘leld Clayey (I.E. 5.0 8.0 9.0 12.0 35.0
stress 4.2 Statnumic loud tests results
(9.0m) L.E. 5.0 8.0 9.0 12.0 35.0
(kN/m’)
Clayey U.I. 10.0 18.0 20.0 20.0 55.0 Statnamic load test was performed on the V-1, V-2
(4.3m) L.E. 10.0 18.0 20.0 20.0 55.0 and V-3 piles with elapsed time of 51, 49 and 42
Sand U.E. 8.0 12.0 14.0 14.0 17.0 days after initial driving respectively. Each
(2.31~11 L.E. 8.0 12.0 14.0 14.0 17.0 Statnamic test result is summarized in Table 12 to
Sand Toe 21791 21991 24990 29988 20491 14. Figure 4 to 6 present the load-displacement
curves of each test.
Total shaft resistance(KN) 1052 1386 1727 1854 4865 Shape of the load-displacement signals were
Toe resistance (KN) 2455 2477 2815 3378 2308 similar each other and suggests that the test piles
Total capacity (KN) 3506 3862 4542 5232 7174 were still in elastic region during Statnamic loading.
Measured axial forces transferred to the pile toes
were about 5000KN.
Table1 0. Results of signal matching analyses (V-2). To estimate static capacities of the piles fiom the
V2-03 V2-Rl V2-R2 Statnamic test results, the unloading point method
E.O.D. 1 day 25days named ULP method (Middendorp et al. 1992,
Sandy U.E. 1.o 2.0 2.0 Kusakabe & Matsumoto 1995) and the signal
(1 1.0rn)
matching analysis was adopted. The ULP method
L.E. 2.0 2.0 10.0 was performed using two different acceleration
Clayey U.E. 4.0 8.0 25.0 values; one was measured at the pile head and
(20.0m) L.E. 4.0 8.0 25.0 another was averaged value along the pile axis. The
Clayey 1J.E. 8.0 10.0 40.0 reason of adopting averaged values was due to the
(9.0m) L.E. 8.0 10.0 40.0 long length of the test pile that had a relative loading
Clayey U.E. 8.0 12.0 48.0 duration, proposed by Karkee (I 998j, T,, of about 5.
(4.31111 L.E. 8.0 12.0 48.0 T,=5 is regarded as to be a boundary between
Sand U.E. 9.0 15.0 16.0 dynamic and rapid loading (Japanese Geotechnical
( 2 . 3 4 L.E. 9.0 15.0 16.0 Society 2000) such as Statnamic loading. The signal
Sand Toe 18992 24990 8397 matching analyses was performed. Smith model
Total shaft resistance(KN) 1056 1695 5493
(Smith 1960) was adopted as the soil resistance
Toe resistance (KN) 3359 442 1 1486
model in the analyses. Figure 4 to 6 also present the
force-displacement curves obtained from the results
Total caoacitv (KN) 4417 61 15 6979
of analysis by the above three methods. Table 15

603
Table 12 Statnamic load test result (V-1). Table 14 Statnamic load test result (V-3).
Initial load (KN) 1,653 Initial load (KN) 1.653
Initial displacement (mm) 2.5 Initial displacement (mmj 2:5
Max. loading force (KN) 23,861 Max. loading force (KN) 22,841
Max. pile head displacement (mm) 54.5 Max. pile head displacement (mm) 5 1.O
Max. pile head velocity (m/s> 1.47 Max. pile head velocity (m/s) 1.27
Max. pile head acceleration (m/s) 48.92 Max. pile head acceleration (m/s) 40.71
Residual pile head displacement (mm) 12 Residual pile head displacement (mm) 6.0
Max. pile toe axial force (KN) 4820 Max. pile toe axial force (KN) 4995

Figure 6. Load-displacement curve (V-3).

Figure 4. Load-displacement curve (V- 1).

Table 15 Static capacities calculated from Statnamic test

Table 13 Statnamic load test result (V-2). results.
Initial load (KN) 1,653 * Signal matching analysis
Initial displacement (mm) 2.5 v- 1 V-2 V-3
Max. loading force (KN) 27,173
Max. pile head displacement (mm) 53.9 Sandy U.E. 20.0 20.0 20.0
Max. pile head velocity (m/s> 1.83 (11.0m) L.E. 20.0 20.0 20.0
Max. pile head acceleration (m/s) 59.25 Clayey U.E. 80.0 80.0 80.0
Residual pile head displacement (mm) 11.0 Yield (20.0m) L.E. 80.0 80.0 80.0
Max. pile toe axial force (KN) 5273 Clayey U.E. 80.0 80.0 80.0
(9.0m) L.E. 80.0 80.0 80.0
(kN/m2) Clayey U.E. 100.0 100.0 100.0
(4.3m) L.E. 100.0 100.0 100.0
Sand IJ.E. 120.0 150.0 150.0
(2.3m) L.E. 120.0 150.0 150.0
Sand Toe 45000 40000 40000
Total shaft resistance
15288 15613 15613
(KN)
Toe resistance (KN) 5069 7076 5069
‘Total capacity (KN) 20357 22689 20682
* ULP method
v-1 V-3 V-l
Total capacity
(Pile head Acc.’l (KN) 21391 24132 20509
Total capacity
20081 23292 19956
(Averaged Acc.) (KN)
Figure 5. Load-displacement curve (V-2).

Table 15. Each value was greater than the

summarizes the static capacities calculated by the
three methods.
The total static capacities derived from the results
of signal matching analyses were estimated to be
20357KN for the V-1 pile, 22689KN for the V-2
pile and 20682KN for the V-3 pile as shown in
designated bearing capacity. Figure 7 presents the
comparisons between measured and calculated time-
pile head displacement curves as the results of signal
matching analyses ofstatnamic test.

604
 Figure 8. Comparison of distributions of shaft friction.

Figure 7. Results from signal matching analyses for Statnamic

tests (Force - displacement relationships).

5 DISCUSSION

5.1 Shaft resistance

Figure 8 presents the Statnamic and Dynamic test
results refer to shaft resistance. Figure 8 shows the
Statnamic test results where shaft resistances are
varying with depth. In the Figure 8, the designed
values estimated from the unconfined compression
strength (qu/2) that is represented by black lines.
Black dots indicates the values measured by a
fkiction meter, which is able to measure friction
stress simply by turning a steel pipe with a diameter
of l O O m r n installed into designated depth in the
ground.
From Figure 8, it is obvious that the distributions
of shaft resistances estimated from Statnamic tests
were close to the designed values and also the values
measured by the friction meter.
In the Figure 9, elapsed time-static total shaft
resistances obtained from Dynamic and Statnamic
tests were shown on the lower diagram. This figure
shows that the total ;haft resistances had been Figure 9. Variety Of pore water pressure and shaft fricti0n:V-1.

605
increased with elapsed time. During elapsed time of
42 to 51 days (1008 to 1270 hours), the total shaft
resistances from the Statnamic test was greater than
15 to 20 times of the end of initial driving.
From this result, followings can be concluded.
The alluvial clayey soil around the pile was fully
disturbed during the initial driving, hence the
friction value is small. However after the set-up
period, large set-up characteristics in the alluvial
clayey soil is indicated.
The upper diagram in Figure 9 shows recovery of
pore pressures around the surface of the V-1 pile
with time after the end of initial driving. Pore
pressure gages were installed facing to the alluvial
clayey at the levels of -32m and -42m from the pile
head as shown in Figure 1. The upper diagram in
Figure 9 shows that the measured pore water
pressures at the both levels had been settling down
in 1100 hours (46 days) and dynamic tests and
Statnamic test had generated large pore water
pressures around the pile.
Figure 9 shows obvious correlation between the
recovery of the pore pressure and the set up
phenomena of the shaft resistances in elapsed time
after the end of initial driving.
5.2 Toe resistance
Toe resistances obtained fi-om the load tests were
compared with the design values estimated. Those
values are shown in Table 16.
Table 16. Comparison of toe resistances.
r - - - Statnamic (KN) Dynamic(KN) Designed value
Lilbf:
(Matching) (Matching) (KN)
v -1 5069 3378 4760
v-2 7076 442 1 7140
v-3 5069 2813 6170
The toe resistance was not fully mobilized in
every case; therefore the effect of the thin bearing
stratum (gravel with sand; 4.5m to 6.5m thickness)
could not identify from the test results.
Figure 10 shows the relationship between force
and displacement at the head of test piles during
Statnamic test. Stiffness of load-displacement curve
of the V-2 pile with the cross-rib plates at the pile
toe is larger than those of the V-1 and V-3 piles
without the cross-rib plates. This certainly suggests
that increase in load capacity of pile due to the
presence of the cross-rib plates at pile toe. However,
the piles were not loaded up to yield values,
therefore it is difficult to verify an effect of the
cross-rib plates on allowable resistance of the pile
toe.
Figure 10. Comparison of load-displacement curves.
6 CONCLUSIONS
Safety of the bearing capacities for steel pipe piles
seated on the thin bearing stratum were confirmed
by the result of Statnamic and Dynamic load tests. In
addition, the site-specific driving formula to control
driving of the production piles was developed based
on the test results.
The enforcement of the load tests at the wharf in
Nagoya port materialized economical and reliable
pile installation in such tricky soil conditions.
Following interesting result was obtained from the
series of Dynamic and Statnamc tests.
Shaft resistance of the alluvial clayey stratum
shows large recovery in its set-up behavior.
Pore water pressures around the pile in the alluvial
clayey stratum had increased at the time of
Dynamic and Statnamic tests and settled down
natural pressure within 1100 hours (46 days).
The effects to the toe resistance due to the thin
bearing stratum could not be observed from the
test results. However, the pile with the cross-rib
plates at the pile toe indicates large stiffness than
that of without the cross-rib plates in the load-
displacement curves.
REFERENCES
Japanese Geotechnical Society, 1998. Draft of “Method for
Rapid Load Test of Single Piles (JGS 1815 - 2000)”,
Proc. 2nd Int. Statnamic Seminar, Tokyo, Japan.
Balkema (to be published)
Karkee, M., T. Horiguchi and H. Kishida, 1997. Static and
dynamic tests for evaluation of the vertical load bearing
capacity of piles, Proc. 22nd Annual Member’s
Conference, Deep Foundations Institute : 199-214.
Kusakabe, 0 & T. Matsumoto, 1995. Statnamic tests of
Shonan test program with review of signal interpretation,
Middendorp, Proc. 1st Int. Statnamic Seminar,
Vancouver, Canada, 113-122.

606
Middendorp, P., P. Bermingham & B. Kuiper, 1992. Statnamic
load testing of foundation piles, Proc. 4th Int. Conf: on
the Application of Stress-Wave Theory to Piles (B. J
Barends Ed.), The Hague, The Netherlands: 581-588.
Rotterdam, Balkema
Smith, E.A.L, 1960. Pile driving analysis by the wave
equation, J. Soil Mech. Found, Div., ASCE, Vo1.86,
No.shd4, pp.35-61.
Yamahara, H., 1963. Plugging effects and bearing mechanism
for steel piles, Summaries of technical papers, Vol, 96,
97, Architectural Institute of Japan, (in Japanese).

607
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 1503

Statnarnic load testing using water as reaction mass

M. D. Justason - Berminghammer Foundation Equipment, Hamilton, Ont., Canada

M.C. Janes -Deacon Industrial Design, Vancouver,B. C., Canada
P, Middendorp - TNO Profound, Deljt, Netherlands
A.G. Mullins - University of South Florida, Tampa, Flu., USA

ABSTRACT: The following paper describes experiments performed in Hamilton, Canada, using water as re-
action mass for Statnamic testing. These trial tests represented the first such tests performed in the world.
These tests were performed in Lake Ontario, on a 323 mm diameter pipe pile driven in 8 in of water. The
tests were performed using a 0.6 MN Statnamic device and hydraulic catching mechanism with a specially
designed hanging structure from which a submerged steel container was used to contain a mass of water. The
goal of the testing was to investigate the potential for using water to replace concrete and steel for use as reac-
tion mass for Statnamic tests. This paper also describes Statnamic test results froin the first two Statnamic
testing contracts for the Port of Lake Charles, Lake Charles, Louisiana, USA. These jobs were performed by
Applied Foundation Testing, Inc. of Green Cove Springs, Florida, USA. These tests were performed on
square concrete piles, with test loads ranging from 4.0 MN to 5.0 MN. This paper also summarizes the theo-
retical research conducted by Middendorp and Courage in 1995, which influenced the design of the actual wa-
ter reaction mass assemblies for the experimental work of 1998.

1 INTRODUCTION preceding theoretical work. This paper also de-

Statnamic load testing is frequently performed on
foundations for bridge piers and port structures.
Since these locations are over-water, a support struc-
ture is required for the Statnamic apparatus. The
typical support structure consists of four temporary
piles and a steel platform. Although this setup is not
difficult, it does require extra time and cost to install.
Alternatively, it is possible to construct a plat-
form supported only by the test pile using a collar or
sleeve that slides over the test pile. This method has
been popular in Japan and has some advantages and
disadvantages compared to the four temporary sup-
port piles.
In both of the methods described above, the Stat-
namic reaction masses, base frame and gravel con-
tainer must be mobilized to the jobsite, in addition to
materials for the support structure.
The idea of using the water surrounding the foun-
dation to replace the conventional use of concrete
and/or steel as reaction mass has been discussed for
several years. The advantages of this idea would be:
mass would not need to be mobilized to the jobsite;
a support frame would not be required; and the time
for setup and tear-down could be decreased.
This paper describes the research and experiments
performed in 1998 surrounding the use of water as
reaction mass for Statnamic testing, as well as the
scribes the test results for the first contract Statnamic
tests performed using water as reaction mass in
1999.
2 RESEARCH
2.1 Theory
In a typical Statnamic test, a mass (usually concrete
or steel) is used as reaction for an upward thrust
produced by expanding gases within the Statnamic
device. The masses are placed on top of the Stat-
namic device and are typically accelerated at about
20 times the acceleration of gravity in the upward di-
rection. The result is a downward force on the
foundation of 20 times the weight of the reaction
masses. Although the applied force on the founda-
tion cannot be sustained (force durations of lOOms
are typical), the magnitude of the force is large in
relation to the amount of mass needed.
Even though the amount of mass required to per-
form a Statnamic test is small in relation to the mass
required for a static load test, there is still cost in-
volved in the mobilization of this mass. It had been
proposed in the early 1990's, that for foundations
situated in a marine environment it should be possi-
ble to use the readily available quantity of water to

609
provide the mass needed to perform a Statnamic test,
thus eliminating the mobilization cost of the con-
crete or steel masses.
Middendorp and Courage of ” N O Building and
Construction Research performed one of the first
theoretical studies of water as Statnamic reaction
mass in 1995. In this analysis three configurations
of the underwater container were examined.
The first mathematical model described a com-
pletely closed steel container, filled with water and
submerged. The steel container was assumed to re-
main entirely below the water surface throughout the
test. After the initial upward acceleration of the
Statnamic test, it was calculated that the continued
upward movement of the container would be very
large due to the upward momentum of such a large
volume of water. The deceleration force of gravity
could only act on the mass of the steel container and
not on the mass of water, which was weightless
while submerged, and moving upward very quickly.
The main advantage of this first model was that a
relatively small amount of mass would need to be
’caught‘ by whatever type of catching mechanism
was to be devised. The disadvantage was that an un-
realistically large jumping height of the mass would
need to be accommodated.
c, drag coefficient water in container
=
c, drag coefficient containers
= valve at the bottom of the containers wi!l open
A, = cross section water
A, = cross section containers
FA = inertia force = M.a
M = M,+ M,
Mw= mass water
Mc = mass steel containers
G = MF*g - V c . p
g = gravity
Figure 1 Force Diagram of Middendorp and Courage mathe-
matical model of water reaction mass
The second mathematical niodel examined the
simple case where the container described in the first
model was allowed to rise above the surface of the
water. In this case the deceleration force of gravity
610
was able to act on the mass of water as well as the
mass of the steel container. As expected. the jump-
ing height of the mass was greatly reduced. Unfor-
tunately, this configuration would place greater
physical demands on the catching mechanism.
The third model presented the most attractive so-
lution. In this model the top of the steel container
was left open, and valves or ‘trap-doors‘ were intro-
duced to the bottom of the container. This cdnfigu-
ration allowed for the same mass of water to be
‘contained’ as in the first two models. thus providing
equal inertial reaction for the Statnamic loading
event. During the upward acceleration of the Stat-
namic device the valves in the container bottom re-
mained closed. During the deceleration phase of the
event the contained moving mass of water was al-
lowed to flow through the container rather than ‘lift’
the container along with it.
Modest jumping heights were observed in this
container configuration, even when the container
remaining completely submerged throughout the
event. This was the most favorable model for the
design of an over-water catching mechanism.
A schematic of this third model is shown in Fig-
ure 1 . Figure I also contains the definitions of the
variables used in the following equations.
From equilibrium:
FSTN= Fa + F,, + F, +G
Equation 1 can be rewritten as:
d’x d’x
t = Mw. -+ M,. -+ ...
. sin a
Fb,AX
dt’ dt’
As soon as the Statnamic force becomes zero, the
steel containers will lag behind the accelerated water
mass since they are decelerated by gravity. The
automatically and the moving water will flow up-
wards through the container. In this condition the
equation becomes:
Additional theoretical work was performed by
Baddour in 1998, just prior to the construction of the
first prototype underwater reaction mass container.
The main issues addressed were: the additional mo-
bilized mass of water ‘above’ the container, drag
forces above and below the container, and drag
forces on the sides of the container. Calculations
generally agreed with the previous work by Midden-
dorp and Courage. However, the work by Baddour
focused more on the forces that would govern the
structural design of the steel containers.
 Most of the remaining questions centered on the
physical concerns of assembling such a test. The
only remaining task was to undertake a physical ex-
periment.
2.2 Experiment
In the spring of 1998, a series of 12 Statnamic load
tests were performed in Hamilton, Canada. These
tests were conducted using a 0.6 MN Statnamic de-
vice and a prototype submerged reaction mass con-
tainer.
2.2.1 Apparatus
Testing was conducted using a 0.6 MN Statnamic
device. A hydraulic catching mechanism was also
used to facilitate repetitive testing. The testing appa-
ratus, along with the fabricated reaction mass as-
sembly is shown in Figure 2.
Figure 2 Test apparatus consisting of 0.6 MN Statnamic de-
vice, hydraulic catching mechanism, platform, and underwater
reaction mass assembly (shown above water).
The test pile was located in a corner of the dock
structure at the Berminghammer Facility, thus it was
possible to span the corner with long beams to pro-
vide a temporary platform for the Statnamic appara-
tus. A view of the test setup can be seen in Figure 3.
The fabricated steel reaction mass assembly con-
sisted of a 1.07 m diameter steel casing 2.13 m in
length. On the inside of the casing, another steel
pipe was used to centralize the assembly over the
323 mm diameter pipe pile that was used as the test
61 1
pile. The top of the container was constructed with a
flat top with six 250 mm holes. The container was
fabricated with a flat bottom with six 250 mm holes.
The purpose of the bottom holes was to allow the
container to fill with water as it was lowered over
the test pile. Each 250 mm bottom hole was also
equipped with a hinged door that opened in the up-
ward direction. These doors were designed to re-
main closed during the upward acceleration of the
container to prevent the water from simply flowing
through in the downward direction. The doors did
allow water to flow through the container in the up-
ward direction after the completion of a test, to pre-
vent the creation of a low-pressure area at the bot-
tom of the container due to the upward momentum
of the contained water and to act as the valves pro-
posed by the Middendorp and Courage model. A
photo of the submerged reaction mass assembly is
shown in Figure 4.
Figure 3 View of test setup clearly showing the support plat-
form spanning the corner of the dock.
Figure 4 View of submerged reaction mass assembly
High tensile strength anchor rods provided the
load transfer from the Statnamic device to the top
and bottom of the reaction mass assembly. Two sets
of three rods were used. The first set was 8 m long
while the second set were 4 m long. The two sets
were coupled together to provide the required length
to perform one series of tests with the top of the re-
action mass assembly 3.3 m below the water surface
(masses stayed below the water surface during test-
ing), and another set of tests with only 0.3 m of
submergence (masses jumped out of the water dur-
ing testing).
The test pile was a 20 m long closed end steel
pipe pile 323 mm in diameter with a wall thickness
of 8.0 mm. The pile was installed using an MIST
V16 vibratory hammer with a peak driving force of
1.1 MN.
The soil at the test location consisted of 7.8 m of
water, overlying soft marine silt sediments to a depth
of 33 m. The purpose of the foundation was to pro-
vide a consistent elastic load deflection behavior for
all tests with the loading capacity of the 0.6 MN
Statnam i c device.
After the first seven load tests using the above re-
action mass configuration, a 250 mm wide flange
was welded to the outer edge of the water container.
This flange (added to the bottom outer edge), essen-
tially doubled the plan area of the reaction mass as-
sembly. This was done to investigate the effects of
drag on the behavior of the underwater reaction
mass assembly.
The total reaction mass for this testing was 3350
kg, with 1490 kg (44%) provided by the Statnamic
cylinder, silencer and reaction mass assembly. The
remaining 1860 kg (56%) of the reaction mass was
provided by the contained water.
2.2.2 lnstrume ntation
Accelerometers were placed on the reaction mass
flange of the Statnamic device. These measure-
ments were used to determine the acceleration of the
under water reaction mass assembly. The jumping
height of the Statnamic device was measured after
each load cycle.
Pressure transducers were mounted in various lo-
cations, both inside and outside the water reaction
mass assembly. These measurements were made to
help investigate the behavior of the water both inside
and outside the steel container.
In addition to the above special instrumentation,
the typical Statnamic load cell, laser displacement
sensor, and pile accelerometer was also used.
2.2.3 Typical Results
For all of the loading cycles performed in this ex-
periment, the test pile was not displaced beyond its
elastic range, providing an essentially rigid test
foundation. Peak displacements of 7-8 mm were
typical for the test pile, with net displacements of
zero.
612
For discussion purposes, the results of three typi-
cal loading cycles will be discussed.
2.2.3.1 Series 1, Test 4 - Results
The first test to be discussed was performed dur-
ing the first series of tests prior to the addition of the
‘drag flange’. The peak Statnamic load was 480 kN,
with a measured upward acceleration of the reaction
mass assembly of 10 g. Typical upward accelera-
tions for the same test, performed with a typical
amount of concrete reaction mass would have been
approximately 15 g. The drag of the water provided
a reduction in the upward acceleration of approxi-
mately 50%.
On land, with conventional reaction masses, the
jumping height of the masses would have been ap-
proximately 2.3 m, while the measured jumping
height of the water reaction masses was only 1.7 m -
a reduction of 30%.
The pressure transducers mounted on the reaction
mass assembly suggested peak drag forces on the
underside of the reaction mass assembly of 35 kN.
The peak drag force on the top of the assembly was
58 kN. These forces occurred at the same time as
the peak Statnamic load. Together the drag forces
accounted for 93 kN, or about 19% of the applied
load. Significant changes in the lateral pressure on
the container were not observed suggesting minimal
drag along the sides of the reaction mass assembly.
2.2.3.2 Series 2, Test 10 - Results
This series of tests was performed with the addi-
tional drag flange welded to the bottom of the reac-
tion mass assembly.
The peak Statnamic load for this test was 496 kN.
The load was very near to the previously discussed
test, however the 3% increase in the peak force was
achieved without increasing the amount of Stat-
namic fuel. This was likely achieved by a combina-
tion of the increased mass of the drag flange ( 1 20
kg), as well as the increased drag at the peak load.
The peak upward acceleration of the reaction
mass assembly was approximately 9 g rather than 10
g. And the jumping height of the masses was only
1.4 m, down from the previous 1.7 m, a reduction of
18%.
The pressure transducers indicated similar pres-
sures to the previous testing, both on the top and bot-
tom of the reaction mass assembly. The increased
area on the bottom of the container translated into an
estimated peak drag force of 85 kN.
The addition of the drag flange increased the total
drag resistance to approximately 140 kN or 28% of
the peak Statnamic load.
2.2.3.3 Series 2, Test I ? - Results
Test 12 provided data on one of the highest peak
loads of all the tests performed. This test was also
 performed with the drag flange included on the reac-
tion mass assembly.
The peak Statnamic load for this test was 573 kN,
very close to the device capacity of 600 kN.
For an equivalent land test using the same mass
of concrete or steel, the peak upward acceleration
would have been approximately 17 g. The peak
measured acceleration was 11.5 g, a reduction of
32%.
Pressure transducer measurements implied a peak
drag force beneath the container of I I5 kN and peak
drag forces on top of the container of 40 kN. In this
case the drag forces accounted for a total resistive
force equal to 27% of the applied load.
Compared to Test 10, the drag forces provided a
slightly lower percentage to the peak applied force,
however the magnitude of the forces did increase
with the increased applied load. This suggested that
the drag forces provided additional resistance ‘for
free’ under additional applied load. In other words,
the total ‘effective’ reaction mass was increased Figure 5 Plume of water visible shortly after each loading
without increasing the actual mass. event

2.2.4 Discussion More detailed results and analysis of this series of

From the data presented here, it was observed that
the drag forces on the top and bottom of the reaction
mass assembly both contributed to the overall resis-
tance provided to the Statnamic loading event. More
generally, it was observed that the combination of
water mass and drag forces provided an effective re-
action mass.
The valves proposed by Middendorp and Courage
appeared to function as intended, although an addi-
tional series of tests were also performed with the
valves welded shut. These tests showed only small
increases to the jumping height of the masses. This
was not readily explained by the theory. For tests
with the valves open and with the valves shut, there
was a visible plume of upwardly moving water that
appeared at the water surface about 2 seconds after
the Statnamic test (see Figure 5).
Tests were also performed in which the sub-
merged reaction mass assembly was permitted to
exit the water surface. In these tests, the reduction
in the jumping height was significant and, although
spectacular, the tests it did produce an increased
strain on the catching mechanism.
Use of the hydraulic catching mechanism pro-
vided a convenient means of performing this re-
search, however, the contribution of the hydraulic
cylinders to the jumping height of the reaction mass
was not quantified. One small test was performed
without the use of the catching cylinders simply to
observe the rate of decent of the masses afier the
test. This test was not repeated due to the rapid rate
of decent that was observed and due to the possible
risk of damage to the equipment. This highlighted
the need for a catching device.
tests are provided in Janes (1 998).
3 FULL SCALE TESTING
Given the positive results of the small-scale testing
in 1998, it was decided that the water reaction mass
technology and theory was sound enough for use in
the commercial sector. The American testing com-
pany, Applied Foundation Testing, Inc. in conjunc-
tion with Berminghammer Foundation Equipment of
Canada collaborated on a project performed for the
Lake Charles Harbor and Terminal District.
Two testing contracts were undertaken for the
Harbor. The first contract was performed in May of
1999, and involved the testing of two 600 mm
square concrete piles. The test load for these piles
was 4.0 MN.
The second contract for the Harbor was per-
formed in June of 1999 and involved the testing of
three 750 mm square concrete piles and three 600
mm square concrete piles. The test load for these
piles was 5.0 MN.
3.1 Apparatus
Figure 6 shows an elevation schematic of the
apparatus used for testing in Lake Charles. For this
project it was not possible to use a supporting
platform as for the experiments performed in 1998.
Without the supporting platform, it was necessary to
devise a catching frame that was solely supported by
the test pile. The resulting design consisted of two 5
m steel truss towers, connected at the top and

613

Figure 6 Elevation view of the Lake Charles testing apparatus
truss towers, connected at the top and bottom, with
enough space in between for the Statnamic device.
The catching mechanism was a mechanical latch
system running on vertically mounted racks of
‘teeth’. Due to the reduced quantity of mass that
was actually ‘caught’, it was believed that a simple
mechanical system could provide the required catch-
ing capacity. This was in contrast to the hydraulic
catching systems that are normally required for land
based tests in which the entire reaction mass must be
‘caught’.
The water container consisted of six, 1.2 m di-
ameter steel casings with closed bottoms. The
length of the containers was 2.0 m. The structural
support for the casings was provided by two, 1 m
deep steel I-beams, mounted below the casings. The
upward Statnamic force was transferred to the
containers through high-tensile rods connecting the
I-beams with the Statnamic device, similar to the
1998 experimental testing.
A photo of the system is shown in Figure 7, in
which the entire reaction mass assembly js visible.
614
Figure 7 Statnamic equipment being placed on a test pile
Figure 7 also demonstrates the means by which the
system was moved from one test pile to the next.
The convenience of a large crane allowed the entire
system to be moved from pile to pile without disas-
sembly. On the second testing contract this allowed
for as many as three piles to be tested in one day
with the possibility for more
Unlike the prototype equipment of 1998, the reac-
tion mass container was not equipped with ‘valves‘
in the bottom of the steel casings. The trap-door
type valves of the MiddendorplCourage model were
replaced by two manually operated gate valves. The
superior watertight sealing of these valves allowed
the reaction mass containers to actually float for sev-
eral hours while the above-water equipment was
assembled on the test pile. Allowing the containers
to float made the logistics of the test setup much
more straightforward. The gate valves were opened
just before the each test to allow the containers to fill
with water and to sink to their starting elevation be-
low the water. During testing the gate valves were
left open. After the testing, the open gate valves al-
lowed the water to drain from the containers as they
were extracted from the water.
In comparison to the 1998 testing, the Lake
Charles under-water equipment contained a larger
percentage of actual contained water mass. The
mass of water was 13,600 kg, while the mass of the
containers and associated structures was only 9,400
kg. However, including the mass of the Statnamic
silencer, high-tensile rods, and reaction beams
mounted on the Statnamic device, the total non-
water mass was very close to the mass of the con-
tained water. In general, the testing system was ex-
pected to behave much like the 1998 prototype.
Perhaps the largest uncertainty in this test pro-
gram was the performance of the new ‘mechanical’
catching mechanism.
3.2 Test Results
Unlike the 1998 experimental testing, the main focus
during the Lake Charles testing was the actual
movement of the pile, rather than the movement of
the reaction masses. Unfortunately, data was not
collected on the upward acceleration of the reaction
mass assembl>i. From the test data the force-time
curves give the most information about the behavior
of the testing apparatus. Jumping heights were re-
corded, but are not presented here.
Figure 8, shows the force-time curve for one of
the first 4MN tests on the 600 mm piles. Figure 9.
shows a typical force-time curve for a conventional
Statnamic testing using concrete and steel masses.
The similarity between the two graphs shows the ef-
fectiveness of the water reaction mass assembly.
Closer examination of Figure 8 shows a loading
event of just over 100 ms duration. Figure 9 is at
least 10 ms shorter in duration. This was the ex-
pected result since the land test was performed with
a total reaction mass of 22,000 kg, whereas the water
test used closer to 25,000 kg. Although the results
were not surprising they do offer direct evidence of
the effectiveness of the contained water as reaction
mass for the Statnamic test.
Also of ncte was the successful operation of the
mechanical catching mechanism. As the Statnamic
silencer reached the high-point of its jump it was
noted that the subsequent downward movement was
abrupt - similar to the 1998 experimental testing
Figure 8 Load vs Time usin$ water as reaction mass
615
Figure 9 Load vs Time using concrete and steel as reaction
when the catching cylinders were deactivated. As a
consequence, the descending silencer and reaction
masses produced a sharp impact load on the catching
frame. Several of the steel teeth as well as the me-
chanical latches showed noticeable wear. The catch-
ing device functioned well nonetheless, and proved
the idea of a mechanical catching mechanism to be
sound.
4 SUMMARY AND CONCLUSIONS
The main conclusion drawn from the work described
in this paper was that water could be used success-
fully as reaction mass for Statnamic testing.
The theoretical work by Middendorp and Cour-
age was useful in the design of the water containers.
In particular, the use of a valve at the base of the wa-
ter containers was highlighted in their work as an
important element. In the 1998 experimental testing
the mass of the steel containers was high in relation
to the mass of contained water, hence the Midden-
dorp and Courage valves had only a moderate effect.
The containers used in the Lake Charles contract
testing omitted the valves entirely. As the percent-
age of contained water increases in relation to the
overall reaction mass (this is likely for larger capac-
ity tests), it is theorized that the valves will become
increasingly important.
The experimental work of 1998 was the first test-
ing of its kind, and proved the general concept of us-
ing water as reaction mass. Pressure transducers
contained in the submerged reaction masses indi-
cated that the drag forces of the container provided
an additional resistance to the Statnamic event of 30-
30%. The moderate jumping heights observed in
this testing also demonstrated the usefulness of drag
forces.
The final proof of the concept of using water as
reaction mass was the contract testing in Lake
Charles, Louisiana. This testing proved that a low-
capacity mechanical catching mechanism could be
used to catch the water reaction masses. It also
proved that such a test could be assembled
freestanding on the test pile, without the need for a
support frame or special support structure.
The success of the 1998 experiments as well as
the Lake Charles testing point toward the future use
of water as reaction mass for Statnamic testing in
marine environments.

REFERENCES
Applied Foundation Testing, Inc. Phase I1 - Report of Stat-
namic and Dynamic Load Testing, New Ship Berth at Con-
traband Bayou - Contract “A“. Prepared for The Lake
Charles Harbor and Terminal District. c/o CBK Soils Engi-
neering. Inc.. September 14, 1999.
Applied Foundation Testing. Inc. Phase I - Report of Static,
Statnarnic and Dynamic Load Testing, New Ship Berth at
Contraband Bayou - Transit Shed. Prepared for The Lake
Charles Harbor and Terminal District, c/o CBK Soils Engi-
neering. Inc.. May 28. 1999.
Baddour. R., 1998. Calculations modeling the reaction mass of
hydraulic dynamic effect: Report fiom the tiniversigi of
Western Ontario, Depr. of Civil Engineering, London, On-
tario, Canada, I998
Janes. M.C. 1998. Statnamic testing using submerged reaction
mass, Second International Starnamic Seminar; preliminary
proceedings, Tokyo, October 29-3 I 1998. Tokyo: Japanese
Geotechnical Society.
Jonker. G. 1999. Use of a ballistic hammer for installation of
anchor piles for off-shore structures. Conrinuous advances
in Moorings Ce Anchors; proceedings, Aberdeen, 26-2 7
M ~1999.
J London: IBC UK Conferences Limited.
Middendorp. P.. and W.M. Courage, “STN above-water testing
with undenvater containers“, Reporr ,from TNO Building
and Construction Research, Delfi, Netherlands, 1995.

616
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Introducing statnamic load testing in Europe: Case studies in the Netherlands

G.J. J.van Ginneken & l? Middendorp

irTV0 Profound, Delft, Netherlands

ABSTRACT: A short introduction is given of the principles of the statnamic (STN) pile load testing. Exam-
ples are given of the applications of STN in The Netherlands and some European initiatives. Practical experi-
ences with the use of the 4MN STN device with catching mechanism are described. Concluding economic
considerations will be given when applying STN in practice.

1 INTRODUCTION The duration and loading rate of a STN test can

be controlled with the volume of the burning cham-
Statnamic started in the early nineties to satisfy ber and the shape of the cylinder and piston, the
construction Industry’s demand for cost effective amount and type of fuel and the amount of reaction
and accurate means of testing high capacity founda- mass launched. As a result, the load can be intro-
tions. Berminghammer Foundation Equipment of duced more gradually and for a much longer dura-
Canada joined with TNO Building and Construction
Research of The Netherlands to develop STN. At
this moment over one thousand STN load tests have
been performed in Canada, United States, Japan,
Malaysia, United Kingdom, The Netherlands and
many other countries in the world. The equipment
available at present can perform STN load tests
starting from 0.1 MN up to 30 MN. Design and con-
struction of even higher capacity STN devices (60
MN or more) can be expected in the coming years.

The principle of STN is based on the launching of

a reaction mass from the pile head. Launching takes
place by generating high pressures in a cylinder,
caused by the burning of a special fuel. As a reaction
on the launching the pile is gently pushed into the
soil. The load exerted on the pile head is measured
by means of a load cell. The displacement of the pile
head is registered by means of a special developed
laser sensor. Load cell and laser sensor are inte-
grated components of the STN loading device. No in-
strumentation has to be installed on the pile shaft.
The required reaction mass is 5% of the force gener-
ated. The high capacity STN devices are not re-
stricted to test single piles, but also allow the testing
of pile groups and structural elements such as bridge
piers and spread footings. Because the principle of
STN is based on the acceleration of masses, piles
can be tested in any direction, also horizontal and
under batter. Fig. 1. Force versus time diagrams

617
tion compared to dynamic load testing.
As an example the force time diagrams of dy-
namic load test, a STN load test and a static load test
have been presented in Fig. I . The long duration of
the STN loading keeps the pile under constant pres-
sure and tension stresses cannot develop. The central
location of the STN loading device on the pile top
guaranties an axial introduction of the load onto the
pile.
The long duration of the STN loading causes for
all pile levels a similar displacement behaviour as
can be observed with static load testing. This justi-
fies a simple modelling of pile and soil, in which
stress wave phenomena do not have to be taken into
account. The pile is considered to be a mass on
which the STN force, the inertia force and soil resis-
tance are acting (Fig. 2).
Fig. 2. Forces acting on the pile during STN
loading
2 HISTORY OF STATNAMIC DEVELOP-
MENTS
Worldwide during the last decade over one thou-
sand statnamic tests have been carried out. The first
developments of STN started in Canada, USA and
the Far East where alternative ways of pile testing
were introduced to accommodate for the high de-
mand for pile testing in strong developing econo-
mies. There was much to gain, especially in the lar-
ger construction works where pile testing appeared
to be cumbersome and expensive. Numerous com-
parisons between static and STN tests on driven and
cast-in-situ pile have been made to convince authori-
ties that STN proves to be good alternative for static
load tests. Acceptance was acquired when authori-
ties had seen that the results obtained from the STN
618
tests indeed showed the same or similar results as
static tests.
During the first years the desire to use STN for
higher loads quickly rose. Statnamic devices to
measure up to 30 MN were developed and are still
used in larger projects. From the beginning the eco-
nomic gains in these larger projects were evident.
After first setting ground in Asia in mega con-
struction projects STN is now quickly developing a
wide acceptance in the USA as well. Testing is
steadily growing, having performed up to present
some 300 STN tests in the USA alone. The way
STN was introduced in the USA was identical to its
acceptance in the Far East. First comparative studies
were carried out between static and STN load tests
to confirm the reliability and accuracy of STN be-
fore the transportation authorities adopted the
method.
Several case histories in the Netherlands and
Germany during the early nineties showed satisfac-
tory and good agreement between static load behav-
iour from STN results and static load test results. It
was concluded that:
During STN load testing, pile behaviour
can be modelled as a mass on which iner-
tia forces and soil resistance are acting.
This allows the simple calculation of the
static load behaviour.
Statnamic load testing can be performed
on piles situated in soils with a strong dy-
namic response.
The point of unloading (maximum dis-
placement) in the STN load displacement
diagram allows the direct calculation of
the maximum static soil resistance during
testing.
Strain rate sensitivity has to be included
for clayley soils.
Europe has been somewhat reluctant to accept
STN as the alternative testing method. Reasons for
this are the existing building codes based on a tradi-
tionally strong preference for static load testing for
well defined and safe cases. Moreover, for driven
piles over the years a market for dynamic load test-
ing had been developed and is now well established.
The incentive to adopt alternative methods has
not been large because existing regulations did and
in many case still do not mention STN as an accept-
able method for pile load testing and the market has
the perception that STN testing is expensive. The
question whether to perform pile testing or not in
many cases depends on the funds available and the
judgement of the experts on site. Money appears to
be dominant over quality assurance and as a result
often a lot of construction works go without regular
testing of foundation piles. The reason is that static
load tests are considered too expensive, time con-
suming and therefore cumbersome. Dynamic load
testing is not always recommended because of its
limitations.
For STN to be attractive applications are to be
found in the lower pile loads testing range up to 2.5-
3.5 MN and competition with established low cost
static and dynamic testing methods is strong.
In addition, regulations in most European countries
are not yet adjusted in considering STN as an ac-
ceptable means for pile Ioad testing replacing or in
conjunction with static load tests.
3 INTRODUCING STATNAMIC IN THE
NETHERLANDS AND EUROPE
In response to the increased interest in statnamic
in Europe TNO has decided to set up TNO Profound
(Professional Foundation Diagnostics) responsible to
introduce STN further in Europe. In combination
with the knowledge base in foundation technology
present at TNO Building and Construction Research
a strong basis has made to make STN a success fol-
lowing the examples in the Far East and the USA.
3.1 Stntnarnic load testing with a 4MN device
From its start in March 1999, TNO Profound has
a 4 MN STN device at its disposal at a central loca-
tion in The Netherlands for testing on the European
continent. This has opened up the market for STN in
Europe. Steps are taken to position a second 16 MN
device in the Central European market with the aim
to have a STN testing facility available on local
markets to conduct testing services at commercially
attractive prices.
In 1999 TNO Profound has been engaged in several
research projects in Europe in which the results of
Static and Dynamic Load Testing have been com-
pared with STN test results on the same and/or simi-
lar piles. Projects have been carried out with Jacbo
(The Netherlands), WTCB (Belgium), the Technical
University of Budapest (Hungary). Preparations are
being made to initiate similar projects in Germany,
although Germany has a strong tradition in static
load testing the willingness to apply STN in Ger-
many as well is rising now STN tests become eco-
nomically more attractive. TNO Building and Con-
struction Research is participating in a research
project with the University of Sheffield to develop a
619
model for the strain rate sensitivity of clayley soils
under STN loading conditions.
Demonstrations have been trend-setting in adopt-
ing this new technology into the market. The method
used by TNO Profound to introduce this new con-
cept is to organise seminars and training courses for
users and potential users to acquaint themselves with
this new technology. This introduction involves a
theoretical part in which the geo-technical back-
ground of STN is being explained and a practical
part in which the actual testing is demonstrated.
Collaboration exists with the leading Universities
and Research Institutes in the countries involved se-
curing adequate technical support, acceptance and
being able to open steps towards introduction of
STN into the national regulations. This means train-
ing, educating and continuously introducing techni-
cians into the world of STN.
New countries covered are Spain and Turkey,
where they are very interested to get STN intro-
duced. Preparations are made for a first STN demon-
stration project in Barcelona. In Poland and Hungary
progress is made to get STN into the regulations and
perform static-STN demonstrations in 2000. Infra-
structural EU supported investments for the pre-
accession countries are the driving force for the in-
troduction of STN.
3.2 Major considerations for the introduction of
Statnanzic on the European continent
Regulations
The tradition in most European countries is to
conduct static load tests on specially made test piles.
Loads to be applied range between 1.5 - 2.0 times
the design load. These static load tests require a lot
of preparation work and piles are to be carefully se-
lected and prepared for testing. This is a costly and
cumbersome activity. For this reason static load tests
are virtually abandoned in The Netherlands for proof
testing and are becoming less popular in other coun-
tries counties like Belgium as well. Time and con-
struction costs are increasingly under pressure
squeezing cumbersome testing practices to the bare
minimum.
In Europe most countries the use of alternative
load testing methods is not foreseen in the national
regulations. This hampers the wider introduction and
acceptance of other innovative and more convenient
methods. Mainly in Germany dynamic load testing
has gained interest because of its convenience to use.
A pile driving hammer or drop hammer can be used.
Piles to be tested can be selected randomly. The
same counts for STN where the reaction mass re-
quired is only 5% of maximum load to be applied.
Construction engineers of local organisations may
decide to accept DLT or STN when judged suffi-
ciently suitable for its purpose. Their application de-
pends on the availability of a testing capacity and is
mostly used on a case-by-case basis, where conven-
ience, time pressure and economics play a major
role. For cast-in-situ piles STN is the preferred
method (P.Middendorp, 2000). TNO Profound is
promoting to consider also alternative testing meth-
ods like STN into the national regulations of EU
countries. This is believed to be an essential prereq-
uisite to get STN widely accepted in Europe. STN
demonstrations with the involvement of the acade-
mia and regulation authorities are required to get de-
cision makers acquainted with the method.
Statrzanzic as the only feasible alternative for test-
iiig
The major applications of STN in The Nether-
lands have been cases where the alternative testing
methods where not seriously considered because of
technical limitations (DLT on cast-in-situ piles
would not generate satisfactory results), time pres-
sure or undesired excessive costs. STN was not re-
quested to replace routine static load testing, but was
considered to identify the load bearing capacity of
existing piles for re-use or in situations of doubt
about the performance of disturbed soil profiles
where conventional calculation and testing methods
could not be applied. Groundwater seepage has been
one of the major causes. STN would also be ideal in
situations where a lot of piles are to be tested in a
short time frame. For such larger projects STN is
still lacking sufficient EU based reference projects.
Ecorzoinic considerations
The price per pile for static testing varies consid-
erably depending on the load to be applied and the
specific site situation in which the tests are per-
formed. When more piles are to be tested economies
of scale cannot be made since each pile needs to be
individually prepared. When one compares static
load testing with STN, the testing costs of STN are
usually 30-60% lower. The real economic gain is not
so much in the testing costs as such, but in the con-
siderable time saving when applying STN. For STN
the preparation time in minimal, no reaction piles are
required, piles for testing can be selected randomly
and testing results are immediately available. This
allows for instant decisions on additional testing re-
quirements. For calibration purposes of STN it is
always recommended to conduct one comparison
test with a static load test in unknown soil condi-
tions.
620
The price of STN and dynamic load testing is largely
influenced by the costs of mobilizing the testing
equipment. When the testing equipment is used for
testing of one pile only the price remains relatively
high. When more piles are tested on one project site
the price per pile drops considerably. When
comparing STN with DLT, pile preparation for STN
is easier, but more time is required for the
installation of the testing device. Four to 6 piles can
be tested a day with a STN device with catching
mechanism.
Transportation costs in Europe are considerably
higher than in the USA and allowable transportation
weights per truck are maximized. Therefore in
Europe it is preferred to move empty containers to
be filled on the site with load material, whereas in
the USA it’s easier to transport concrete/steel reac-
tion masses.
4 CASE STUDIES OF STATNAMIC IN THE
NETHERLANDS
4.1 Statizamic pile testing in Rotterdam
Project Data
Location : Rotterdam Harbour
: Wilton-Feijenoord
Time : June 1999
Client : Van Hattum & Blankevoort
Testing : Prefabricated piles 420x420
4 piles, 23 m long (new)
2 piles, 24 m long (existing)
Load applied : 2 -4 MN
Project Description
At one of the piers in the Rotterdam harbour a
large construction crane was present for the con-
struction of offshore vessels. It was intended to ex-
tend the moving range of the crane to an old recon-
structed pier by using the existing foundation piles.
It was questioned whether the existing piles under
the pier would be suitable to bear this additional
load. TNO Profound was requested to apply maxi-
mum loads in the order of 3.5 MN to check the load
bearing capacity of the piles.
Project Characteristics
The piles had to be tested over water under diffi-
cult site conditions, limited availability of crane ca-
pacity and considerable time pressure (sea vessel to
be loaded). static load testing and DLT not feasible
because of shoreline conditions and accuracy. The 4
MN STN device was moved from pile to pile in one
 Fig. 4. STN Testing at Eindhoven

4.2 Stutnamic pile testing in Eindhoven

Project Data
Location : Eindhoven Railway Station
: Large Shopping Complex
Time : October 1999
Client : Inpijn-Blokpoel Consultants
Testing : Jacbo Avegaar piles (cast-in-situ)
Fig.3 .4MN STN Device in Rotterdam. 15 piles, 2 1 m long (new)
Load applied : 2,5 - 4 MN

piece. Careful and accurate installation of the STN

device above water appeared a time consuming Project Description
process. Two piledday were tested with one- two A large shopping complex was to be built. Con-
testing cycles per pile. struction works were suffering delays and trouble
with the foundations would negatively affect the
timely progress of works. The piles were located in a
Results 6-7 meter deep construction pit and testing had to be
The following loads were applied to the piles: carried out on selected piles. After placing the foun-
dation piles considerable groundwater seepage had
Max load Displacement (in mm) occurred and maintained to grow during the con-
(static) in kN (total) (permanent) struction works. The seepage was stopped through
injecting the soil, but the soil stability around the
Pile 1: 4190 20,s 395 piles was believed to be disturbed. Doubts arose
Pile 2: 3980 19,5 377 whether the piles under the main structure would
Pile 3: 3960 18,s 530 perform according the calculations based on soil in-
Pile 4 (step 1): 2750 12,l 2,o vestigation results and their capacity had to be de-
Pile 4 (step 2): 4340 17,6 1,5 termined.

Conclusion Project Characteristics

The conclusion was that the existing piles per-
formed above expected calculated bearing load and
could be safely used for the pier extension. After the
first tests the client wanted the tests to go further up
to failure to determine ultimate load capacity. The
tests showed that the performance of existing piles
may well exceed calculated capacities.
First a reference test was carried out on a pile in-
stalled in an area which had not been affected by
seepage. The results of the STN test on the reference
pile were compared with the test results of the seep-
age affected piles. It was also indicated that the
amount of piles to be tested would depend on the
testing results. The piles selected for testing were in-
dicated from the ones most severely affected in con-
centric circles going outwards towards less seepage
affected piles.

621
 Results Project Characteristics
The aim was to minimize the time spent on test- This project is typical for a common situation in
ing. During the first tests it appeared that cyclic test- the Netherlands where the standard procedure is not
ing would have a positive influence on the perform- to conduct static load tests on a regular basis, but to
ance of the piles in working conditions. It was rely on the results of CPT tests to approve piling
decided to conduct 2-3 cyclic loads per pile on the jobs. In case of poor CPT results construction engi-
piles most seriously affected. neers are only inclined to approve the foundations
when the piling contractor can prove that the piles
The following loads were applied to selected installed perform within the limitations. TNO Pro-
piles: found was requested to test and provide an inde-
pendent judgement on the load bearing capacity of
Max load Displacement in mm the Omega piles. Cyclic STN tests on a reference
(static) in kN (total) (permanent) pile and on a questionable pile were compared to as-
sess the difference between the two and to judge the
Pile 1 : 2220 11,l 790 acceptability of the tested piles. On both piles 4 cy-
Pile 2: 2130 8,1 4,l clic load tests were carried out up to failure capacity
Pile 3: 3460 27,9 21,5 to fully satisfy the demands of the construction en-
Pile 4 (step 1):2450 9,1 2,4 gineer.
Pile 4 (step 2): 31 10 18,5 3s

Conclusion
It was concluded that the piles proved to be just
within the margin of acceptance. Signal matching
with TNOSTN revealed that the soil skin friction in
the upper 4 m layer was reduced to only 35% to of
the total due to the effect of seepage. The total pile
capacity was seriously affected by this, but the lower
soil layers compensated for the loss occurred and
would still have sufficient capacity to provide the
necessary support for these pile foundations.

4.3 Statnamic pile testiizg in Utrecht

Project Data
Location : Utrecht Fig 5. 4MN Statnamic device in Utrecht
: Jaarbeurs Fair Complex
Time : December 1999
Client : Jacbo/HBG Results
Testing : Jacbo Omega piles (Auger) The aim was to minimize the costs and therefore
2 piles, 18 m long (new) the amount of time spent on testing. The results were
Load applied : 2,5 - 3,5 MN satisfactory and the foundation piles were accepted
by the construction engineer.
Project Description The following loads were applied to the piles:
At the Jaarbeurs fair grounds of the Municipality of
Utrecht, new high rise office buildings were planned Max load Displacement in mm
for construction. For the Omega pile type it was a (static) in kN (total) (permanent)
routine procedure to compare the CPT results de-
rived from tests before and after the pile construc- pile 1 pile 2 pile 1 pile 2 pile 1 pile 2
tion works. These values are then being compared to
step 1 1690 1850 16,3 19,2 12,O 10,O
check for a reduction of CPT values caused by the
step 2 2170 2570 16,O 15,5 13,O 12,O
installation of the piles. In one of the a corners of the
step 3 2190 2730 9,9 36,l 6,O 28,O
building site the CPT results after piling showed step4 3110 3140 18,5 38,5 9,O 31,O
values lower than the set limits. STN tests were car-
ried out to verify that the capacity and settlements of
the piles would be within the allowable limits.

622
 Conclusion
Statnamic proved to be a simple and effective
way to test the performance of cast-in-situ founda-
tion piles in cases where CPT tests would give
doubtful results. STN testing confirmed that in some
cases the interpretation of CPT tests alone could lead
to questionable conclusions. With STN an effective
tool was provided to verify independently the actual
performance of the pile tested.
Middendorp, P., Foeken, R.J. van, 1998, When to
Apply Dynamic Load Testing and Statnamic Test-
ing, 2nd Statnamic Seminar, Tokyo
Middendorp, P, Ginneken, G.J.J. van, Foeken,
R.J. van, 2000, The advantages and disadvantages of
x
D namic Load Testing and Statnamic Load Testing.
6' International Conference on the Application of
Stress Wave Theory to Piles, Sao Paulo, Brazil

5 CONCLUDING REMARKS

When considering the STN testing worldwide it

can be concluded that nowadays every other day a
STN load test is carried out somewhere in the world.
USA, Asia and Japan are taking the lead. Europe is a
region where STN has a large potential to grow.
TNO Profound has actively started to promote STN
in Europe and aims to motivate centres of expertise
in European countries to initiate STN activities. STN
can replace static load testing, however, when deriv-
ing design rules, static load testing will remain the
preferred testing method.

6 REFERENCES

Bermingham P., Janes M., 1989, An innovative

approach to load testing of high capacity piles, Pro-
ceedings of the International Conference on Piling
and Deep Foundations, London, p.409-413.

Middendorp, P., Bermingham P., Kuiper B.,

1992, Statnamic load testing of foundation piles. 4th
International Conference on Stress Waves, The
Hague, Balkema

Middendorp, P, 1993, First Experiences with

Statnamic Load Testing of Foundation piles in
Europe, Proceedings 2nd International geotechnical
seminar on Deep Foundations on Bored and Auger
Piles, Gent, p .265-272, Balkema

Brown, D.A., 1994, Evaluation of Static Capacity

of Deep Foundations from Statnamic Testing. Geo-
technical Testing Journal, Vol 17, No.4, American
Society for Testing and Materials

Matsumoto, T., Tsuzuki, M., 1994, Statnamic

Tests on Steel Pipe Piles Driven in a Soft Rock. In-
ternational Conference on Design and Construction
of Deep Foundations, Orlando, U.S. Federal High-
way Administration

Middendorp, P, Bielefeld, M.W., 1995, Statnamic

Load Testing and the Influence of Stress Wave Phe-
nomena, First International Statnamic Seminar,
Vancouver

623
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 750 3

The advantages and disadvantages of dynamic load testing and statnamic load
testing

l? Middendorp & G.J. J.van Ginneken

TNO Profound, Delft, Netherlands
R. J.van Foeken
TNO Building and Construction Research, Department of Structural Dynamics, Rijswijk, Netherlands

ABSTRACT: Pile capacity testing by high strain dynamic loading methods is widely applied because of its
economy and efficiency compared to static load testing methods (SLT). Frequently applied dynamic loading
methods are dynamic load testing (DLT) and statnamic testing (STN). The paper will deal with the very often
raised question in practice: What are the advantages and disadvantages and when to apply DLT and STN
when pile type and soil conditions are known. Special attention is given to DLT on cast in situ piles, and
complicating factors like limited knowledge of concrete material properties and pile shape. The suitability of
DLT and STN is discussed for cast in situ piles and driven pre-cast piles by the evaluation of reliability, econ-
omy, mobilization of capacity and the chance on pile damage.

1 INTRODUCTION

Pile capacity testing by high strain dynamic load-

ing methods is widely applied because of its econ-
omy and efficiency compared to static load testing
methods (SLT). The most popular dynamic loading
methods are dynamic load testing (DLT) by an im-
pact hammer and statnamic testing (STN) by launch-
ing a reaction mass from the pile head.

DLT introduces a short duration shock pulse into

the pile. STN generates a relative long duration push
load onto the pile head. Extensive descriptions of
load testing methods and comparisons are published
by Holeyman (1992) and Karkee et a1 (1997). How-
ever these papers do not deal with the very often
raised question from practice: What are the advan-
tages and disadvantages and when to apply DLT and
STN when pile type and soil conditions are known.
The answer to this question will be treated in the
next paragraphs .

Special attention is given to DLT on cast in situ

piles, because the calculation of the pile load is
based on signals from strain transducers mounted on
the pile shaft. So for DLT the pile load calculation
depends strongly on pile material and cross section
properties and factors complicating the analysis like
limited knowledge of concrete material properties
and pile shape are discussed. Fig. 1 statnamic test on a cast in situ pile

625
 - economy
- mobilization of capacity
- chance on pile damage

2 THE APPLICATION OF DLT AND STN ON

CAST IN SITU PILES

For cast in situ piles both DLT and STN are per-
formed a certain period after pile production, to al-
low the piles to reach the required compressive
strength to withstand the test loads. For DLT strain
and acceleration transducers are mounted on the pile
shaft near the pile head. The load displacement be-
havior is calculated by signal matching. For STN the
load displacement behavior is calculated in most
cases by the Unloading Point Method (UPM), how-
ever signal matching techniques are also applied.

2.1 Accuracy in load measurement for STN

With STN the load is accurately measured by a

calibrated load cell placed on the pile head. The
measured load is not dependent on the pile proper-
ties. The load measurement error is less than 0.1% of
the maximum range of the load cell.

2.2 Material properties and accuracy in load

Fig, 2. Dynamic load test on a cast in situ pile. measurement for DLT

With dynamic load testing strain transducers are

mounted on the shaft near the pile head.
The load (F) on the pile head is calculated by
multiplying the measured strain (E) with the modulus

Fig. 3. Statnamic piston with built in load

cell and laser displacement sensor placed on
a cast in situ pile

Finally the suitability of both DLT and STN will

be evaluated by taking into account the following
points:
Fig. 4. Strain transducer mounted on the shaft of
- accuracy of the load measurements a cast in situ pile
- reliability

626
of elasticity (E) of the concrete and the pile cross
section (A).
F=E. A.& (1)
The accurate determination of the properties E
and A for bored piles is difficult in many cases.
To calculate the force from the measured strain in
a pile during DLT we need to know the cross section
and the modulus of elasticity of the concrete at the
measuring level. For piles with homogeneous mate-
rial the stress wave velocity (c) is used to calculate
the E-modulus with
E = c2. p
c = 2L/T (3)
Knowing or estimating the stress wave velocity c
we can calculate the pile load at the measuring level
with the formula
F= c’.p.A. (4)
So the derived stress wave velocity has a strong
influence on the value of the load measured in the
pile. An error in the measured load will result in an
error for the pile capacity prediction.
Fig.5. Calculation of stress wave velocity c from toe reflection
627
The stress wave velocity is calculated from the
time (T) it takes for a stress wave to travel over the
pile length (L) from the pile head to the pile toe and
back to the pile head (Fig. 5). For this method it is
required that the reflection coming from the pile toe
is clearly visible in the signals. In Fig. 6 the force
and velocity times impedance signals of two dy-
namic load test are presented. The first case shows a
clear toe reflection and the stress wave velocity can
be calculated accurately. If the toe reflection is not
visible one has to estimate the toe reflection time.
However an error in the estimated toe reflection time
(T) and stress wave velocity (c) will result in a con-
siderable error in the calculation of the E-modulus.
For example a 5% error in the stress wave velocity
will result in a 10% error for the E-modulus and cor-
responding load in the pile. Another option in this
case is to rely on an estimate for the E-modulus from
the pile material properties.
Making an estimate on the E-modulus is difficult
because it is not a constant value but depends on the
age and the quality of the concrete (Franklin,
1971)(Fig. 7), the loading rate (Sparks et all, 1973 ),
and even the temperature of the concrete (Abbasi
1990). For example, for static load testing the
modulus of elasticity for concrete is in the range of
28 GPa to 32 GPa while for dynamic load testing it
is in the range of 32 GPa to 52 GPa
Fig. 6 Well visible and no visible toe reflection.
Fig. 7 Relations between dynamic modulus of elasticity and age for concretes made with various aggregates
Another complicatjng factor in determing the
stress wave velocity c tor cast in situ piles is the fact
that the concrete is not homogeneous. The concrete
quality will vary over the cross section and over the
pile axis. The concrete in contact with the soil will
be of lesser quality than the concrete in the center of
the pile and the shaft area that has been in contact
with the soil might be the location where the strain
transducers are mounted. The concrete quality dif-
ference over the pile length is caused by the pouring
procedure and the difference in concrete pressure
628
during construction. The quality of the concrete near
the toe will in general be better than the quality of
the concrete near the pile head. This also means that
the stress wave velocity will vary with the pile
length. So the stress wave velocity calculated with
c=2L/T is a mean value for the whole pile. The
modulus of elasticity calculated from the toe reflec-
tion represents a mean value for the pile and there
can be a considerable difference with the modulus of
elasticity at measuring level.

Fig 8. Cast in situ pile with bulb.
2.3 Injluence of pile cross section variations on
DLT capacity prediction
To predict capacity from DLT results, signal
matching techniques are the most frequent applied
methods, (TNOWAVE, CAPWAPTM).Based on a
wave equation computer program calculated signals
are matched with measured signals by adjusting the
computer soil model and pile model in an iterative
way. When signals match it is assumed that the com-
puter soil model represents the real soil behavior and
the static pile capacity is calculated from it.
Pile discontinuities like necking, bulbs, and ma-
terial changes introduce stress wave reflections,
which can influence the calculated signals strongly.
Reflections from bulbs yield an almost similar wave
equation result as a local stiff soil layer and a neck-
ing similarly results as a local soft layer. When pile
discontinuities are not properly taken into account,
either a proper match cannot be obtained or the ca-
pacity prediction will not be reliable. Soil properties
can be confused with pile discontinuities.
2.4 Reliability for testing on cast in situ piles
Because of the many unknowns that have to be
solved to perform a proper DLT signal matching
analysis on cast in situ piles, there is considerable
chance of errors in pile capacity predictions.
The load measurement for STN is similar as for
static load testing and unknown pile properties of
cast in situ piles will not influence the load meas-
629
urement results. During STN the load duration is
long enough that all pile parts move in the same ve-
locity range. Under these conditions the pile can be
considered to act as one mass with a pile rigidity be-
havior similar to static load testing (Middendorp
1995 ) For this reason pile behavior during STN is
closer to static load testing than DLT.
2.5 Economy
For DLT on cast in situ piles a drop hammer with
a guiding system has to be mobilized. The required
ram mass is as rule of thumb 2% of the maximum
load that has to be applied. A crane is required to
move the drop hammer over the building site. The
pile head has to be prepared to prevent damage from
impact loading. An epoxy or grout cement is used to
smooth the pile head surface to prevent stress con-
centrations during impact loading. The location of
the transducers has to be at least 2 pile diameters
from the pile head. When the pile head is located at
ground level this requires an extension of the pile
head for a similar length or the excavation of the pile
head. For small capacity piles multiple piles can be
tested in one day. For loads above lOMN the testing
rate is normally in the range of two piles per day.
For STN a loading device with a reaction mass
catching system has to be mobilized. The required
reaction mass is as rule of thumb 5% of the maxi-
mum load that has to be applied. Local available ma-
terial can be used as reaction mass to reduce trans-
port costs. A crane or a crawler system is required to
move the STN device over the building site. For
loads up to 4MN a STN device with a hydraulic
catch mechanism can be applied. For higher loads
STN requires a gravel catch system. Testing can take
between 0.5 and 2 days per pile depending on the
pile capacity. However for piles with a capacity less
than 4 MN, a loading device with hydraulic catch
mechanism can be applied and the number of piles
tested in one day are in the same range as with DLT.
STN can be even more efficient when the loading
device and hydraulic catch mechanism are placed on
crawlers. An epoxy or grout cement is used to
smoothen the pile head surface to prevent stress
concentrations during push loading.
2.6 Chance of pile damage
With DLT the load on the pile head is introduced
by an impacting ram. When the ram is not properly
guided and hits the pile in an eccentric way, bending
stresses will occur and result in excessive compres-
sion and/or tension stresses which can damage the
pile.
Most cast in situ piles need considerably more
displacement to mobilize the ultimate capacity than
driven piles. This softer response will easily gener-

Fig 9. Dynamic load test on a pre-cast pile
ate tension waves. Cast in situ piles are not designed
to withstand high tension stresses. As soon as allow-
able tension stress levels are reached the impact en-
ergy has to be reduced to prevent pile damage. As a
result, DLT has to be stopped at a stage where full
capacity has not yet been mobilized.
With STN the duration of the loading is long
enough to keep the pile is under constant compres-
sion and tension stresses will not occur. To prevent
bending stresses the piston of the statnamic device is
installed exactly on or near the center of the pile
head cross section. The launching of the reaction
mass, and the resulting push load starts from the
center of the pile.
3 THE APPLICATION OF DLT AND STN ON
PRECAST DRIVEN PILES
For pre-cast driven piles both DLT and STN are
performed after a setup period after pile installation.
This allows the soil to recover from driving induced
disturbances like pore water pressure. In most cases
the soil will regain strength during the setup period.
For DLT strain and acceleration transducers are
mounted on the pile shaft near the pile head. The
load displacement behavior is calculated by signal
matching.
630
For STN the load displacement behavior is calcu-
lated in most cases by the Unloading Point Method,
however signal matching techniques are also ap-
plied.
3.1 Accuracy
With DLT on pre-cast driven piles, the load in the
pile is measured by strain transducers mounted on
the pile shaft. Pre-cast piles are considered to be of
homogenous material and with the method described
in section 2.2 and based on the determination of the
stress wave the E-modulus can be determined accu-
rately. The toe reflection will be visible at several
stages of driving and the stress wave velocity can be
determined easily. Only when the pile head is heav-
ily reinforced will the E-modulus at the pile head be
different from the E-modulus calculated by the stress
wave velocity.
With STN the load is accurately measured by a
calibrated load cell placed on the pile head. The
measured load is not dependent on the pile proper-
ties. The load measurement error is less than 0.1% of
the maximum range of the load cell.
3.2 Reliability
The capacity of driven piles is mobilized at rela-
tive small displacements.
Both DLT and STN are performed after a set up
period. For DLT the pile load displacement behavior
is calculated by a signal matching technique (CAP-
WAPTM,TNOWAVE) in most cases.
For STN the pile load displacement behavior is
determined by a direct method, the Unloading Point
Method (UPM) and in some cases by signal match-
ing.
3.3 Economy
DLT has the advantage that the pile driving ham-
mer used for pile installation can also be used for re-
driving the piles after a set-up period. However,
when the pile driving hammer has to be used for
constant production, an additional pile driving ham-
mer or drop hammer has to be mobilized. When the
mobilization of the full pile capacity is requested,
the production hammer might not be sufficient to
mobilize pile capacity after the set up period and an
additional heavier hammer has to be mobilized. For
STN the same economical conditions are applicable
as mentioned in paragraph 2.5 (Ginneken, van G.J.J.,
2000)
 Table 1. Preferences for DLT or STN with respect to
economy for driven piles
Driven piles
Soil set up I DLT I STN Preferred
dium
medium to ** ***** STN
high
3.4 Mobilization of capacity
Set up phenomena can increase the soil resistance
considerably. The pile driving hammer used for pile
installation might not be able to mobilize the full
pile capacity in such a case.
Another reason that capacity can not be mobilized
with DLT is that the load cannot be increased be-
cause compression or tension stresses becoming too
high.
To mobilize the pile capacity a STN device will
be sent to the building site with at least a corre-
sponding loading capacity. Only when the piles are
over-designed will the full bearing capacity not be
mobilized.
3.5 Chance of pile damage
For DLT there are some cases with a chance of
pile damage. In the case of low friction and a soft
toe response tension waves will be generated during
DLT. When the maximum allowable tension stresses
are reached the load on the pile cannot be increased
because this will generate higher tension stresses and
the pile will experience damage. In the case of a pile
with a hard toe response, for example pile toe on
rock, the compressive stresses at the pile toe can
theoretically be two times higher than the maximum
compression stress at the pile head. This is caused
by the superposition of compression stress waves at
the pile toe. So, if during DLT the compression
stresses at the pile head are higher than half the
compressive strength of the pile material, collapse of
the pile material at the pile toe will occur. In this
case piles can only be tested up to half the compres-
sive strength of the pile material, which may not cor-
respond with the capacity of the pile.
For STN the pile is kept under constant compres-
sion and tension waves are suppressed. Superposi-
tion of compression waves at the pile will not occur.
As with SLT piles can be tested near to the compres-
sive strength of the shaft.
631
4 CONCLUSIONS
For bored concrete piles, auger piles and caissons
the dynamic load testing method has some disadvan-
tages and is less suitable and statnamic load testing
is the preferred method. The most important reasons
for the preference of statnamic load testing in the
case of cast in situ piles are:
1. Accuracy in load measurement
STN is not dependent on pile material and cross
section properties
2. No influence of cross sectional variations
STN results are not influenced by cross sectional
variations over the pile length
3. No tension during compressive testing
STN long duration loading will keep pile under
constant compressive pressure
4. Concentric loading
Easy placement of STN loading device in center
of the pile
5. Pile and soil response closer to static
With STN the pile moves as one unit, similar to
static load tests. Stress wave phenomena can be ne-
glected resulting in a simple method of analysis
For driven piles both DLT and STN methods can
be applied reliably and each has its advantages and
disadvantages. A big economic advantage for DLT
can be the use of the production rig for testing. A big
advantage for STN is the fact that maximum avail-
able energy can be used to mobilize capacity and
that that testing does not have to be stopped when
tensional stresses become too high like with DLT.
5 REFERENCES
Abbasi, A.F., Al-Tayyib, 1990. Effect of hot
weather on pulse velocity and modulus of elasticity
of concrete. Materials and Structures, 1990, 23,
~~334-340
Ginneken, van G.J.J., 2000, Introduction Stat-
namic Load Testing in Europe, Case Studies in the
Netherlands. Proceedings of the Sixth International
Conference on the Application of Stress-wave The-
ory to Piles, Sao Paulo.
Holeyman, A.E., 1992. Keynote Lecture: Tech-
nology of Pile Dynamic Testing. Proceedings of the
Fourth International Conference on the Application
of Stress-wave Theory to Piles, The Hague, F.B.J.
Barends, Editor, A.A. Balkema Publishers, ppl95-
215.
Franklin, R.E., King, T.M.J. 1971, Relations be-
tween compressive and indirect-tensile strength of
concrete, Road Research Laboratory, RRL Report
LR 412

Madan B. Karkee, Takashi Horiguchi, Hideaki Ki-

shida. Static and Dynamic Tests for Evaluation of
the Vertical Load Bearing Capacity of Piles. , 22nd
DFI Annual Member's Conference, Toronto, Can-
ada, 1997, pp199-214

Sparks, P.R., Menzies, J.B., 1973. The effect of

rate of loading upon the static and fatique strength of
plain concrete in compression. Magazine of Con-
crete Research, Vol25/ 1973, No. 83, pp 73-80.

632
10 Case histories, pile set-up and correlations between test methods -
Prediction reliability
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Keynote lecture: Improving the reliability of pile bearing capacity prediction

by the dynamic increasing energy test (DIET)
Nelson Aoki
University of SLio Paulo, SLio Curlos, Brazil

ABSTRACT: The dynamic increasing energy test (DIET) allows the measurement of the dynamic resistance -
displacement curve and the prediction of the static resistance - displacement curve of the pile-soil system. The
variations of energy and work done are evaluated by Hamilton’s principle of energy conservation, and this
evaluation allows the prediction of the dynamic failure load of the system. The usefulness of this approach is
confirmed by the good correlation between the results of the DIET and of the static load test.

1 INTRODUCTION static approaches and the definition of the ultimate

The main purpose of a dynamic pile test is to predict
the static response and it is essential that any transfer
model can be applied to both static and dynamic
loading conditions (Randolph & Deeks 1992). This
is fimdamental for reliable predictions of behavior
under static and dynamic conditions (Aoki 1989a).
The evaluation of the static resistance by the ap-
plication of the stress wave theory to one blow of
hammer is questionable and some discussion points
are: (a) the choice of one blow among several im-
pacts of nearly the same energy; (b) the effects of the
elapsed time, from the end of driving to the start of
the test; (c) the magnitude of deformations around
the pile toe and shaft; (d) the comparison of the dy-
namic penetration with the static movement; (e) the
methodology of the static load test and the evalua-
tion of the pile bearing capacity (Fujita & Kusakabe
1988).
The need for the standardization of the pile - soil
interaction model, the establishment of soil parame-
ters, the statistical treatment of discrepancies be-
tween the model and reality and derivation of design
values for the load bearing capacity, which can be
used within the modern Ioad and resistance partial
factors design approach, are essential for the total
foundation quality assurance (Vrouwenvelder &
Dieterman 1992). The average resistance and the
standard deviation allow verifying the actual char-
acteristic load of the piling and the quality assurance
depends on how and who takes the decision to stop
the driving in the field. In contrast, the safety in the
allowable load philosophy is simply a matter of per-
sonal past experience. In any case, the reliability of
the pile bearing capacity prediction, by dynamic or
resistance, are of foremost importance.
This paper will deals with: (a) the ultimate resis-
tance definition in the conventional static load test
and in the constant energy dynamic test, accounting
for the energy in the system; (b) the Hamilton’s prin-
ciple of energy conservation (Clough & Penzien
1975); (c) the analysis of the dynamic resistance -
displacement curve of the dynamic increasing energy
test and (d) some case histories.
The static pile bearing capacity, in the traditional
dynamic loading test, is predicted from the analysis
of one blow, among several blows of same energy.
This approach can be called a constant energy test.
In that case, depending on the magnitude of the ob-
served permanent displacement (set), the Smith’s
model static resistance (Smith 1960) could satisfy
either the lower or the upper bound theorems of the
limit analysis (Chen & Liu 1990).
The dynamic increasing energy test (DIET) in-
volves the measurement of the total resistance to
penetration or simply dynamic resistance, the dis-
placement and the kinetic energy applied to the sys-
tem at increasing energy hammer impacts (Aoki
1989b). The result is a dynamic resistance - dis-
placement curve, from which the corresponding
static resistance - displacement curve (Smith’s
model) is generated. The dynamic resistance - dis-
placement curve is then examined and the variations
of energy and work done are analyzed by using the
Hamilton‘s principle of energy conservation.
Under the action of the failure load the system
displaces as a rigid body. Considering the similarity
of the failure under static and dynamic loading con-
ditions, since both events involve energy and work
done variations, it is possible to evaluate the ultimate

635
bearing capacity with the dynamic increasing energy
test (Aoki 1997).
The DIET has been largely used in Brazil as an
alternative to the traditional static load test and the
constant energy dynamic test (Aoki & Mello 1992,
Beim & Aoki 1996). The static and dynamic results
compare reasonably well (Aoki 1989c, Niyama &
Aoki 1991). The DIET can also be used to analyze
time effects on pile behavior “set up” after short time
intervals.
2 RESISTANCE - DISPLACEMENT CURVE
AND THE ULTIMATE PILE RESISTANCE
The goal of a loading test is the determination of the
ultimate resistance or bearing capacity of an isolated
foundation pile. The isolated foundation pile is a
system constituted by the structural pile element and
the soil layers along the shaft and under the pile toe.
This system’s response to the action of an axial
load applied to the pile head is the displacement of
this point, which is a complex function of many in-
dependent variables. Among them, the most impor-
tant are: (a) time history of the applied force; (b) ge-
ometry of the pile and the soil layers; (c) physical
properties of the pile and the soils; (d) peculiarities
of pile installation that change material’s properties;
and (e) elapsed time after the pile installation.
During the static loading the applied force and the
vertical movement are measured at each loading
stage. In this case the applied force is equal to the
mobilized resistance and the response of the pile -
soil system to the static loading is the well-known
load - settlement curve. The slow maintained loading
(SML), the quick maintained loading (QML), the
constant rate of penetration (CRP) and the cyclic
loading are examples of static test procedures (Fel-
lenius 1980, Whitaker & Cooke 1961).
Nowadays, the definition of the static pile bearing
capacity or static ultimate resistance, which is the
maximum reaction capacity of the pile - soil system,
is a matter of discussion in the pile foundation engi-
neering practice. Nevertheless, it is recognized that
the shape of the load - settlement curve should be in-
cluded in this definition (Reese 1972, Fellenius
1980) and this can be done by the consideration of
the deformation energy of the pile - soil system.
For a given curve there are many interpretations,
each one leading to a particular ultimate resistance.
The absolute or relative displacement values and the
relationship between the elastic and permanent com-
ponents of the displacement after the unloading or
the time dependent behavior are some of the criteria
for the establishment of the static bearing capacity
(Vesic’ 1975, de Beer 1988). The absolute displace-
ment equal to 10% of the pile dimension is a typical
definition of the ultimate resistance (Terzaghi 1942).
636
The upper bound and lower bound theorems are
used to determine the static failure load from energy
transformations and equilibrium of forces consid-
erations in the theoretical soil mechanics limit analy-
sis (Chen & Liu 1990). By the upper bound theorem,
the load is not less than that of the actual failure
load, determined by equating the external rate of
work to the internal rate of energy dissipation in an
assumed velocity field which satisfies: (a) the veloc-
ity boundary conditions and (b) the strain and veloc-
ity compatibility conditions. Note that the stress dis-
tribution in the system does not need to be in
equilibrium. Under this limit load the system loses
its capability to store more elastic recoverable energy
of deformation and starts to displace as a rigid body.
By the lower bound theorem, the stress field must
satisfy the stress boundary conditions and must no-
where violate the materials yield condition.
The most popular dynamic loading test is the im-
pact of a hammer on the cushion - pile - soil system.
In this case the pile bearing capacity can be predicted
by the old pile dynamic formulae or by the applica-
tion of the stress wave theory to the hammer impact
(Whitaker & Bullen 1981, Goble et al. 1975). The
statnamic test has been developed as an alternative
dynamic test (Middendorp 1993).
During the hammer impact the time histories of
deformation and acceleration are recorded at the in-
strumented section and the velocities and displace-
ments are calculated by integrating the measured ac-
celerations. The force history is obtained from the
deformation and the applied kinetic energy is evalu-
ated by integrating the force and the displacement.
The dynamic resistance or total resistance to the
pile penetration is evaluated from the analysis of
forces and velocities. The correspondent static re-
sistance is evaluated by using the classical Smith‘s
model (Smith 1960) or the derived model of CASE
and CAPWAP methods (Goble et al. 1981, Rausche
et al. 1985, Goble & Likins 1996). The ultimate re-
sistance from one single hammer impact is defined
by the Davisson‘s limit load criterion (Davisson
1972). By this criterion the ultimate resistance is de-
fined by the elastic toe displacement equal to (0.15 +
B/120) inches, where B is the toe dimension.
Finally it can not be forgotten that the soil resis-
tance changes with the time and neither the static nor
the dynamic loading can predict the future behavior
of the pile - soil system, from the measured values at
the time of the test.
It can be concluded that the definition of the ulti-
mate resistance is not clear in both static and dy-
namic loading tests.
2.1 Static force - displacement curve response.
The loading is called static when the force (Q) is
reached in stages of infinitesimal increments (dQ)
and each stage is of infinite time duration. Thus, the
frequency of static loading is equal to zero. How-
ever, in practice, there is no static loading once the
increment of force (AQ) and the time duration (At)
have finite values.
Figure 1 presents the static force - displacement
curve, resulting from the force increment (AQ) and
displacement increment (Ap) and the consequent in-
crement of the deformation energy (AV).
Figure 1. Static force - displacement curve.
The rate of internal energy of deformation corre-
sponding to the force increment (AQ) is the power
g=AV I At (1)
where 9=
power or rate of the internal energy of de-
formation increment, AV = increment of energy de-
formation, and At = increment of time loading.
The average power of the actual static or dynamic
loading is variable from some watts to various kW.
Under the action of the static load (Q), the dis-
placement of the pile head (point D) is expressed by
P = Pe + Pp (2)
where p = pile head displacement, pe = elastic dis-
placement component, and pp = permanent dis-
placement component.
The nature of the deformations can not be known
at the end of the loading stage. The potential energy
of deformation (V) at the end of the loading is the
hatched area ODFO. The complementary energy
(Langhaar 1962) can be calculated by the difference
Vc= Q . P - V (3)
where V, = complementary energy (area OGDO), Q
= applied force, p = total displacement of the pile
head, and V= potential energy of deformation.
The static equilibrium of forces is expressed by
R=Q (4)
where R = total mobilized resistant forces (resistance
or reaction), and Q = applied force (load or action).
According to expression (4), the system response
under static loading can be indifferently expressed
by the relationship between the force - displacement
curve or the resistance - displacement curve.
The system response to unloading is called static
637
if the zero load condition is reached in the inverse
path described for the static loading. In the actual
example of a building construction, the unloading
path does not exist while the load-unloading cycle is
observed only for the live load.
The potential energy of deformation (V) stored in
the system after the static loading can be elastic re-
coverable or permanent, depending on the rheologi-
cal characteristics of the materials.
This fact can be only known after the system un-
loading, when the potential energy of deformation
(V) is transformed into recovered elastic energy (V,)
plus the work done in the system (W,,) in such a
way that
v = v, w,,
f (5)
where V = stored potential energy of deformation,
V, = recovered elastic energy, and W,, = work done
by the non-conservative forces.
The reliability of the ultimate resistance predicted
from the static load - settlement or the resistance -
displacement curves is discussed thereafter.
2.2 Shape of the static load - settlement curve.
Figure 2 presents the typical static load -settlement
curve in the case where the ultimate resistance is
well defined by the final shape of the curve. This
curve correlates the settlement (p) to the load (Q)
where it is implicit that the static reaction (R) is al-
ways equal to the static load (Q).
The failure is characterized when the pile head
displaces from point D to point A under the action of
the constant limit load (Q,), equal to the ultimate
system resistance (Ru).
0 QU
Figure 2. Load - Settlement curve: defined ultimate re-
sistance.
When the limit load (Q = Q,) corresponding to
the point D of the loading curve is applied, the limit
total displacement is (p = pu), the limit potential en-
ergy of deformation is (V = V,), and the limit com-
plementary energy is equal to (V, = V,,).
In the case where all materials are perfectly rigid-
plastic, the unloading path is the horizontal line DF.
In that case the deformation energy is completely
transformed into work done by the resistant forces
(V = Wn,). There is no elastic recovery (Pe = 0) and where Q = applied force, Q, = physical ultimate re-
the total settlement OB is permanent (p = pp). For sistance, and p = settlement.
perfectly elastic materials the loading path is the Expression (6) can also be written
curve OD and the unloading path is DO resulting in
a minimum work done (Wnc= 0) and (V = Ve). lim Q = Q , (7)
V+m
In actual cases some elastic and plastic deforma-
tions take place and the unloading path is DE. The where V potential energy of deformation.
=I

recovered elastic energy (Ve) is represented by the In this case the load - settlement curve can be de-
area DFED and the area ODE0 is the work done scribed by the well-known expression (Veen 1953)
(W,,) in the system.
For any point on the vertical asymptote line be-
tween points D and A, all the additional energy of
deformation is fully transformed into work done, if where Q = applied force, Qu= physical ultimate re-
the elastic recovery energy becomes constant after sistance, a = curve shape coefficient, and p = settle-
point D (DE line being parallel to AC line). ment.
It is not possible to know beforehand if this de- This expression can also be rewritten as
fined failure load is equal to the upper or to the
lower bound of the failure load, unless more data
concerning the external rate of work and the internal (9)
rate of energy dissipation could be known.
Figure 3 presents the typical shape of load- The potential energy of deformation is equal to
settlement curve corresponding to the case where the
resistance reaches a maximum equal to Q m a x at point
F and thereafter it drops up to a constant residual ul- (10)
timate resistance, after point D. The limit load is L J

again the well defined value Qu on the vertical as- where V = stored potential energy of deformation,
ymptote DA to the load-settlement curve. Qu= physical ultimate resistance, p = settlement, and
a = curve shape coefficient.
0
0 Q"
--P load
+-'

)
-s5
2

ID
C

B -----
D
I

Figure 3. Load - settlement curve: defined residual ulti-

mate resistance.
This shape of curve is observed in the cases of
floating piles where, after the peak of resistance
(Qmax) the system resistance drops up to the defined
residual ultimate resistance (QL,).The unloading path
is dependent on the pile-soil system characteristics
as explained in the case of figure 2.
Figure 4 presents the case where the ultimate re-
sistance is not defined. The system displacement in-
creases with increasing loading and the vertical as-
ymptote to the curve appears to be at the infinity.
In this particular case the ultimate resistance can
be taken as the physical ultimate resistance defined
by the condition
lim Q = Q,
P+a
Figure 4. Load - settlement curve: physical ultimate
resistance.
Considering expression (3), the complementary
energy can be estimated by the expression
where V, = complementary energy of deformation,
Q, = physical ultimate resistance, p = settlement,
and CI = curve shape coefficient.
The limit complementary energy corresponding to
the physical ultimate resistance would be
lim V,= V,, (12)
P+W
where V,, = limit complementary energy.

638
 When the curve is expressed by Veen's expres-
sion (8), the limit complementary energy becomes
constant and its value is expressed by
v,, = -Q "
o!
Therefore the physical ultimate resistance can be
defined by the limit complementary energy of ex-
pression (13). This criteria has the advantage of in-
volving both the curve shape and the ultimate resis-
tance in the definition of the failure behavior of the
pile - soil system (Aoki 1997).
Figure 5 presents the case where the OD path of
the curve is similar to that presented in figures 1 to
4. The special feature in this case is that the settle-
ment increases linearly with increasing loading up to
the maximum load test value Qmax,without the ulti-
mate resistance characterization.
A
Figure 5 . Load - settlement curve: structural ultimate re-
sistance
From point D up to point A, the measured dis-
placement increases linearly under increasing load-
ing level and the corresponding potential energy of
deformation increases linearly with loading.
This is the case where all the materials of the pile
- soil system present strain hardening behavior. As
explained previously the unloading path is depend-
ent on the rheological properties of pile and soil
materials. In this case both deformation energy (V)
and complementary energy (V,) increase linearly
with the increasing applied load up to a limit load
corresponding to the ultimate resistance of the weak-
est link in the pile - soil system.
In this analysis the following variables must be
considered: (a) the ultimate soil resistance, (b) the
structural pile strength and, (c) the limit displace-
ment from where the maximum elastic potential en-
ergy of deformation is reached and the system starts
to behave as a rigid body.
Although it is not remarked, in all the cases pre-
sented in figures 2 to 4, the applied force (Q) was
previously assumed to be always smaller than the
soil material ultimate resistance and, on the other
hand, the soil ultimate resistance was smaller than
639
the structural pile material strength. This is ex-
pressed by the following condition:
QSQu<R, (14)
where Q = applied force, Q, = ultimate soil resis-
tance, and R, = structural pile strength.
In some cases, where the structural pile strength is
reached before the ultimate soil resistance has been
mobilized, the system fails by the weaker structural
link and the following condition prevails:
Q u > Re (15)
where Q, = ultimate soil resistance, and Re = struc-
tural pile strength.
This is the case where the pile toe is placed either
on very hard and rigid soil material or in very long
floating piles.
The analysis of the limit maximum elastic poten-
tial energy of deformation of the pile - soil system
requires the prediction of the maximum elastic dis-
placement of the pile head.
The elastic displacement component can be pre-
dicted by the expression (Vesic' 1975)
where pe = elastic displacement of the pile head,
pep = displacement component due to the elastic
shortening of the structural pile shaft, and pes= dis-
placement component due to the elastic deforma-
tions of the soil layers under the pile toe.
The elastic displacement component due to the
elastic deformations of soil layers under the pile toe
(elastic pile toe displacement) is equal to
Pes = Pes,p + Pes.f
where pes.p = component due to the action of the
load transferred to the soil by the pile toe, and pes.f=
component due to the action of the load transferred
to the soil by the pile shaft.
When the load transferred to the soil by the pile
shaft is equal to zero, the elastic displacement of the
pile head corresponding to the compression axial
load (Q) can be predicted by
e , = Q ~ -------+-
* ~B i s p] (18)
where pe = elastic pile head displacement, Q = ap-
plied force, L = length of pile; E, = elastic modulus
of the pile material, B = diameter of pile toe, and E,
= representative elastic modulus of the soil layers
under pile toe.
Considering expression (1 4) for materials of per-
fectly elastic - plastic behavior, the limit elastic pile
head displacement corresponding to the load equaI to
the pile toe ultimate soil resistance will be:

where Peu,s = limit elastic pile displacement, and Qu
= pile toe ultimate soil resistance.
Considering expression (1 5 ) for materials of
perfectly elastic - plastic behavior, the limit elastic
pile head displacement just before the structural fail-
ure of the pile will be:
/ \
where Peu,p = limit elastic pile displacement, just be-
fore the structural failure of the pile.
The limit elastic energy of deformation, when the
ultimate soil resistance is smaller than the structural
pile strength, can be evaluated by
where Ves,m, = limit elastic energy of deformation,
when the soil is the weakest link in the pile - soil
system.
The limit elastic energy of deformation, when the
structural pile strength is smaller than the ultimate
soil resistance, can be evaluated by
where Vee.m, = maximum elastic energy of defor-
mation, when the pile is the weakest link in the pile -
soil system.
The elastic energy of deformation of expression
(21) is a component of the limit potential energy of
deformation (V,) described in figure 2.
Figure 6 presents the idealized load -settlement
curve in the case of a pile - soil system constituted
by elastic perfectly plastic materials. This model is
currently used to describe the actual curves pre-
sented in figures 2 to 5.
QU
load
Figure 6. Idealized load - settlement curve, for elastic
perfectly plastic materials.
In this case the ultimate resistance is equal to the
load (Q,) in the vertical line DA of the ideal curve.
Along the loading path OD the system response is
linear and the energy of deformation is reversible
640
until point D, where the limit elastic displacement
(Peu,s) is reached under the action of the load (Q,)
corresponding to the ultimate resistance of the sys-
tem. Under the action of this constant load, the pile
head displaces from point D to any point A on the
vertical line. Observe that, in this loading path, all
the deformations are permanent.
The potential energy of deformation (V) stored in
the system at point A is expressed by
where V = potential energy of deformation, Q, =
ideal ultimate resistance, p = total displacement, plv
= permanent displacement, and Peu,s = limit elastic
pile displacement.
The unloading path is line AC. The limit recov-
ered elastic energy of deformation is expressed by
where V,, = limit recovered elastic energy (area
ABCA).
The work done in the system is expressed by
where W,, = work done in the system (area
ODACO).
It could be interpreted that any plastic deforma-
tion in the linear loading path OD was 'concentrated'
in the final DA path of the curve.
The soil response is not always instantaneous and
it takes a more or less variable time, to fully mobi-
lize all the resistance in the system.
Most soils change their properties after the exe-
cution by freezing (soil setup) or relaxing (soil re-
laxation). In this case it is important to observe that
the load-settlement curve represents the system be-
havior only for the period of time when the test is
under progress.
The comparison between two pile load tests is re-
liable if their results represent the behavior under the
same time test conditions. The elapsed time is an in-
dependent variable that can affect the results of both
static and dynamic loading tests.
2.3 Shape of the dynamic resistance - displacement
curve.
The loading is called dynamic when the applied
force (Q) varies with time (Clough & Penzien 1975).
The impact of a hammer on the pile-soil system is a
typical example of a dynamic loading.
By analogy, with the resistance to the penetration
of piles under static load, it can be assumed that the
dynamic resistance - penetration curve is similar to
the traditional static load - settlement curve repre-
sented in figures 2 and 3. The resistance against
penetration under a single blow of the hammer
changes with increasing penetration. However the
expression ’dynamic resistance against penetration’
has no definite meaning, unless this term is applied
to the final resistance represented by the abscissa
(QJ of the vertical asymptote to the penetration
curve (Terzaghi 1943). This observation shows the
importance of the shape of the resistance - displace-
ment curve and its link with the ultimate resistance
in the static loading test. Terzaghi’s final resistance
(QJ of the vertical asymptote to the penetration
curve is the physical ultimate resistance of expres-
sion (6). According to Terzaghi, the question is: is
the kinetic energy of the hammer blow enough to
mobilize the ultimate resistance of pile- soil system?
Figure 7 presents the idealized dynamic toe re-
sistance - pile toe displacement curve corresponding
to the idealized load - settlement curve for elastic
perfectly plastic materials of figure (6).
where R,-, = dynamic shaft resistance, n = number of
sections, Rf ,m = maximum local shaft static resis-
tance, J, = local Smith’s damping factor, and v(m,t)
= velocity in the shaft section (m) at time (t).
The total dynamic resistance (static + dynamic) or
simply dynamic resistance is not mobilized at the
same time at all pile sections and is computed by the
following expression:
where Rt = dynamic resistance, R,, = dynamic toe re-
sistance, and Rf, = dynamic shaft resistance.
From expressions (26), (27) and (28) it is possi-
ble to conclude that the corresponding static resis-
tance is
n
R = R, + E R f m
1

o toe resistance Rp Rpt where R = static resistance (Smith‘s model), R, =

static toe resistance component, and Rfm = static
4-
shaft resistance component.
a,

E
cd
Besides the simplified rheological behavior, a
fundamental feature of Smith’s model is that the dis-
_1

4 placement under static loading is assumed to be

.- equal to the displacement under dynamic loading.
-U
0
a,
+*
In the case of a single hammer impact, the load-
.-
2 c ing time history starts at time (tl) when the force be-
a gins to be applied on the pile head. When the pile
B head displacement is maximum, the maximum ki-
A netic energy (T) is the well-known ENTHRU value
Figure 7. Idealized dynamic toe resistance - pile toe dis- (Goble et al. 1975). The resulting total dynamic re-
placement curve for elastic perfectly plastic materials. sistance (static + dynamic) or simply dynamic resis-
tance - pile head displacement curve is presented in
This ideal curve (Smith 1960) correlates the dy- figure 8.
namic toe displacement (Dp) and the total toe resis-
tance (Rpt).The total resistance is usually considered
to have two components: the ’static’ one, which is dynamic resistance : RI
dependent on the displacement and the ’dynamic’ I\, >
one, which is dependent on the velocity. For the sake ? \
of simplicity the total dynamic toe resistance will be
referred simply as dynamic toe resistance. In this ? \
case the dynamic toe resistance is related to the static ?
toe resistance by the expression
Rpt = Rp (1 + Js . Vp,rnax) (26)
where Rpt= dynamic toe resistance, R, = maximum
static toe resistance, J, = Smith‘s toe damping, and
v , , =~ maximum
~ velocity at the pile toe.
The Smith’s model for the resistance at a given Figure 8. Dynamic resistance - pile head displacement
section (m) of the pile shaft is similar to figure 7 and curve, in the case of a single hammer impact.
the difference comes from the fact that the shaft can
displace downwards and upwards. The shape of the corresponding static loading
The total (dynamic + static) shaft resistance or curve is under discussion as pointed out by Terzaghi
simply dynamic shaft resistance is equal to (1943). At the end of the dynamic loading (point A
n of the curve) the maximum pile head displacement is
ft = c R~~ [I + J m - ~ ( m , t ) ] (27 1 (D) and the mobilised dynamic resistance is (RJ.
1

64 1
 At this point the resulting kinetic energy (T) was
stored as potential energy of deformation (V) in the
pile - soil system. The potential energy of deforma-
tion (V) and the maximum kinetic energy (T) are
represented by the area OABO.
The measurements of the maxiqum total (dy-
namic + static) resistance (RJ, the maximum dy-
namic displacement (D) and the maximum kinetic
energy (T) can be done, for example, by monitoring
the hammer impact with PDATMsystem.
The dynamic loading cycle is unique and it is not
possible to separate the loading and unloading stages
as in the static loading case. The applied forces and
the corresponding mobilized resistances vary during
the impact time and only the maximum values at
each pile section are of importance.
The dynamic unloading path starts soon after the
maximum displacement (D) is reached and ends at
the time (t2) when the final pile penetration is equal
to the permanent displacement or set (S). This un-
loading path is represented by curve AC in figure 8.
The area ABCA is the potential recoverable elastic
energy of deformation (V,) and the area OACO is
the work done (W,,,) by the non-conservative forces.
The maximum pile head displacement under dy-
namic loading measured by either the old cardboard
and pencil system or the double integration of meas-
ured acceleration is equal to
D=K+S (30)
where D = maximum pile head displacement under
dynamic loading, K = rebound or elastic displace-
ment, and S = set or permanent displacement.
The elastic displacement component is equal to
where C2 = elastic shortening of the pile, and C3 =
elastic displacement due to deformations of the soil
layers under the pile toe (pile toe quake).
In figure 8 the maximum displacement (D) is
equal to segment OB, the elastic displacement (K) is
equal to segment BC and the set (S) is equal to seg-
ment OC. The set (S) is usually measured by the
pencil and cardboard technique.
Although fundamental, there is not much interest
in measuring the actual dynamic resistance - pile
head displacement curve from one hammer impact.
The ideal shape indicated in figure 7 for the toe dis-
placement is based on the ideal static response of
figure 6. The actual static curves of figures 2 to 5 are
being simulated by the ideal curve of figure 6.
The CAPWAPB analysis simulates the shape of
the static load - settlement curve based on the stress
wave theory applied to one single hammer impact
measurements, by zeroing the velocity in figure 8. In
this curve the toe ultimate resistance is defined by
Davisson's limit load criteria.
642
Despite such difficulties, the reliability of the pile
bearing capacity prediction can be improved by the
dynamic increasing energy testing (DIET), based on
Hamilton's principle of energy conservation which
will be discussed thereafter.
3 HAMILTON'S PRINCIPLE OF ENERGY
CONSERVATION.
The ultimate pile resistance has been predicted by
the dynamic pile driving formulas and by the appli-
cation of the stress wave theory to model the ham-
mer impact. In both cases the repeatability of the
displacements under impacts of same energy shows
that the dynamic loading is a reliable procedure.
The fundamental problem to evaluate the energy
lost in the pile driving system was clarified when the
ENTHRU concept was introduced in practice (Goble
et al.1975). From this time, the actual kinetic energy
available in the pile - soil system has been currently
measured during the pile driving, allowing the
measurement of the hammer impact efficiency.
Apart from this use, not much has been made to pre-
dict the static resistance from the direct energy
measurement (Broms & Choo 1988).
This keynote lecture brings back the old Hamil-
ton's principle of energy conservation that has not
been yet mentioned in the traditional pile dynamic
literature (Aoki 1997). It is a valuable tool to better
understand the energy transformations in the pile -
soil system, subjected to the action of a dynamic
loading, in the case of the ultimate resistance analy-
sis.
Hamilton's principle of energy conservation
states that the variation of the kinetic and potential
energy, plus the variation of the work done by the
non-conservative forces acting in any material sys-
tem, including the work done by damping and any
arbitrary external forces at any time interval tl to t2,
is equal to zero. In variational form, Hamilton's
principle of energy conservation is expressed by
ft': S (T - V ) dt + S,:S ( Wnc ) dt = 0 (32)
where T = system total kinetic energy, V = system
potential energy, including the energy of deforma-
tion and the potential energy of any external conser-
vative forces, W,, = work done by the non-
conservative forces acting in the time interval, in-
cluding the damping forces and any arbitrary exter-
nal loading, and 6 = variation during the time inter-
val (t2 - tl).
According to Hamilton's principle, during the dy-
namic loading corresponding to one impact of the
hammer, it follows that: (a) at the time (t = tl), the
kinetic energy applied to the pile - soil system is
equal to (T); (b) at the time when the displacement
reaches the maximum value (D), all kinetic energy is
transformed into potential energy of deformation The average soil data are presented in table 1.
(V=T); and (c) at the time (t = t2) after the unloading,
this energy of deformation (V) is partially recovered Table 1 : AveraEe soil laver characteristics.
as elastic energy of deformation (Ve) and partially Layer depth Soil description
transformed into work done (Wnc)by the non - con-
servative forces acting in the pile - soil system. 0 to 1.9 Brown medium silty clay fill, N,sI17.=5/30 cm.
In the case of a static loading, the kinetic energy 1.9 to 3.3 Brown medium organic clay, N,s/,7,= 5/30 cm.
T is equal to zero and the Hamilton's principle of 3.3 to 10.3 Pink and yellow sandy silt residual soil; veins
energy conservation is reduced to the well-known of fine gravel, medium, Ns/.l'7.=10/30 cm.
principle of minimum potential energy 10.3 to 16.1 Yellow and grey weathered sandy very dense
rock, N.ypr= 43/30 cm
6(-V) + 6(W",) = 0 (33) Soil under the Very dense sand resulting of weathered
pile toe 37/30 cm.
gneiss, N.~/>.I.=
In this case it is said that the variation of the po-
tential energy plus the variation of the work done by The test results are presented in table 2.
the non-conservative static forces is equal to zero. It
follows that, after the system unloading, the energy Table 2: Dynamic increasing energy test results.
of deformation (V) is partially recovered as elastic H Rt D S T=V T=V
energy of deformation (Ve) and partially transformed (field) (PDA) (PDA) (field) (PDA) (curve)
in work done (W,,) in the system, as expressed by (m) (kN) (m) (In) (kJ) (kJ)
equations ( 5 ) and (33). 0.2 850 0.0035 0 1.9 1.5
0.4 1010 0.0051 0 3.0 3.0
0.6 1140 0.0054 0 3.9 3.3
4 DYNAMIC INCREASING ENERGY TEST 0.8 1330 0.0072 0 5.8 5.5
1.0 1530 0.0082 0 7.3 7.0
The static ultimate resistance can be obtained ei- 1.2 1730 0.0111 0 11.8 11.5
ther in a single cycle of static loading or in a single 1.4 1930 0.0132 0.002 15.9 15.5
hammer blow impact as discussed in chapter 2.2. 1.6 2010 0.0147 0.003 18.6 18.5
The increasing energy impacts has been used to 1.8 2060 0.0158 0.004 21.1 20.7
analyze the behavior of model piles (Koten at al. 2.0 2080 0.0160 0.005 21.8 21.2
1988, Bernardes 1989) or to define some parameters 2.5 2110 0.0178 0.006 24.7 24.9
of the traditional pile driving formulas (Whitaker & 3.0 2180 0.0203 0.009 29.9 30.3
Bullen 1981).
In the test procedure called dynamic increasing For the increasing hammer impacts of height (H),
energy test (DIET) the pile - soil system is subjected the total measured resistance (RJ, the maximum dis-
to the action of several increasing energy blows and placement (D) and the set (S) measured in the field
each blow is instrumented and interpreted, for ex- according to the pencil and cardboard method
ample, by the PDATMsystem. The difference is that (Chellis 1951) are presented in table 2. The last two
the resulting set of dynamic resistances and dis- columns present the kinetic energy T (PDA) meas-
placements are analyzed as a whole by plotting each ured by the PDA system and the kinetic energy
couple of data from the same origin, although they T(curve) calculated from the dynamic resistance -
have been obtained in a cyclic loading (Aoki 1997). displacement curve using the single expression

4.1 Dynamic resistance - displacement curve.

A typical example of dynamic resistance - dis-
placement curve determination carried out with the
PDATMsystem is presented and discussed in this
item.
This is the case of a pre-cast reinforced concrete
pile of external diameter of 42 cm, wall thickness of
8 cm and 12 m long. The length below the sensors
was 10.9 m and the driven length was 10.1 m. The
average structural strength of the pile was 3000 kN.
The age of the pile at the time of the test was 6 days.
The impacts were applied with a 30 kN free-fall
hammer. The shock absorbing system consisted of a
steel helmet of 1 kN of weight, a hardwood cushion
of 30 cm of height with 6 cm thick plywood cushion.
where T(curve) = kinetic energy calculated from the
measured curve, n = number of points, Rt,(i)/ R t,(i-l)
= dynamic resistances at blows (i) and (i-1), and D(i)
/ D(i-1)= displacements at blows (i) and (i-1).
Figure 9 presents the dynamic resistance - dis-
placement curve obtained in this dynamic increasing
energy test. It is assumed that the initial conditions
of any blow do not change with the increasing ap-
plied energy. This is the reason why the points of the
curve are plotted from the same origin, although they
have been obtained in a cyclic load test.
The reliability of this procedure can be verified by
the comparison of the last two columns of table 2,
where the kinetic energy calculated from the curve

643
 0 D (displacement) S (permanent set) A K (elastic rebound)
Figure 9. Dynamic resistance - displacement curve
plot is in good agreement with the direct measure-
ment with the PDA system, within the measurements
accuracy. If the initial condition of the pile-soil sys-
tem has changed after each increasing loading stage,
the measured potential energy of deformation would
not fit the calculated value from the dynamic resis-
tance - displacement curve.
The maximum displacement at point A corre-
sponding to the last applied blow of 3.0 m of height
was 20.3 mm and the mobilized dynamic resistance
was 2 180 kN. At this point the stored potential en-
ergy deformation was equal to 29.9 kJ.
After unloading, the permanent displacement was
9.0 mm. The area ABCA is equal to the recovered
elastic energy of deformation and the area OACO is
equal to the work done by the non - conservative
forces including the damping forces. According to
the Hamilton’s principle of energy conservation, the
following equation can be applied at the end of any
impact of increasing energy:
T = V = V, -t W,, (35)
where V, = recoverable elastic energy of energy of
deformation, and W,, = work done by the non - con-
servative forces.
Assuming that the unloading path AC is linear,
the elastic energy of deformation will be
V,- R,(W2) (36)
where Rt = dynamic resistance, and K = rebound or
elastic displacement.
In that case the work done is
W,CZR~(W~)-T (37)
where Wn, = work done by non - conservative forces.
According to the measurements the wave takes 16
millisecond after the impact to reach point A, re-
sulting an average test power equal to 1868 kW.
644
Figure 9 also shows the Permanent (set) and elas-
tic displacement (rebound) components of the dis-
placement at each increasing energy impact. The last
point where the set is equal to zero is point E on the
curve. In the next impact (H = 1.4m) the system
starts to present permanent deformations ( S O ) . It
can be observed that, after this impact, the elastic re-
bound presents variations, but remains nearly con-
stant while the set increases under increasing applied
energy up to point A where S=9 mm. The occurrence
of the first permanent deformation fulfills the state-
ment of the lower bound theorem for the determina-
tion of the system static failure load. Therefore, it
can be inferred that 1930 kN is the dynamic resis-
tance that could be taken as the dynamic lower
bound of the system failure load (H = 1.40m).
At point F of the dynamic resistance - displace-
ment curve (H = 1.8 m), the displacement be,’*ins to
increase linearly with the applied energy. This
change in the system response involves energy and
work done variations under increasing loading,
which will be analyzed under the light of Hamilton’s
principle of energy conservation.
The subsequent search for the ultimate resistance
in the DIET procedure is a similar problem to the
prediction of the failure load in the static loading.
Considering that the energy and the work done under
increasing energy rates can be measured, it is natural
to analyze their variations in order to check the ex-
istence of a dynamic upper bound of the failure load
similar to the case of the static loading.
For static loading, the upper bound theorem de-
termines the limit load by equating the rate of inter-
nal energy dissipation to the rate of the work done by
the external forces in the system. This theorem states
that the variation of the internal potential energy is
equal to the variation of the work done by the exter-
nal forces, if the rates of variation are referred to the
same time interval.
The load and the analysis are called “limit” be-
cause it is not possible to apply a load greater than
this limit value. This limit is equal to the ultimate re-
sistance of pile-soil system.
In the dynamic increasing energy test it is possi-
ble to measure the variation of the internal potential
energy and the work done at any dynamic loading
level. Similarly, the limit dynamic resistance can be
determined by the upper bound theorem, as follows:
(a) choose an energy of deformation (VU) and
suppose that it fulfills the limit upper bound theo-
rem;
(b) calculate the variation of the energy of defor-
mation for the impacts of energy greater than the
value under analysis by the expression
AV=V-Vu ( 3 8)
where AV = energy of deformation variation,
V = current energy of deformation, and VU = limit
energy of deformation under analysis.
 (c) calculate the work done variation by The corresponding upper bound limit dynamic re-
sistance of system is equal to 2060 kN. Theoreti-
AW Wnc - Wnc,u (39) cally, the elastic rebound would be constant from
where AW= work done variation correspondent to point F, where the system becomes theoretically
AV, Wnc= current work done, and Wnc.u= work done rigid. Nevertheless the average elastic rebound ap-
at the limit energy of deformation under analysis. pears to become nearly constant before this point
The failure dynamic resistance is reached when (see figure 9). This fact can be explained by the re-
the energy of deformation variation is equal to the sidual stresses stored in the system. The elastic de-
work done variation at the same time interval. This formations that take place between points E and F
condition can be expressed as are not recovered as elastic rebounds after unloading
AV - AW = 0 and remain in the system as residual stresses, being
(40)
incorporated into the permanent displacement S after
Table 3 presents the values of the potential energy each blow. After point F, the displacement increases
of deformation, elastic recovered energy and the linearly with increasing kinetic energy satisfying
work done in the present example. equation (40) condition. This is also true for any
Table 3. Elastic recovered e n e r m and work done. value exceeding the limit energy 21.1 kJ, as the case
Rt K where the potential energy is equal to 24.7 kJ.
H V=T Wnc Ve
(field) (PDA)
Therefore, the application of Hamilton’s principle
(D-S) (PDA) (PDA) (PDA)
of energy conservation to the dynamic resistance -
(m) (kN) (m) (kJ) (kJ) (kJ)
0.2 850 1.488
displacement curve allows the prediction of the
0.0035 1,9 0.412
0.4 1010 2.576
lower and upper bound of the dynamic limit resis-
0.0051 3,O 0.424
0.6 1140 3.078
tance, from the dynamic resistance - displacement
0.0054 3,9 0.822
0.8 1330 5,s 1.012 4.788
curve, in a similar way the limit analysis defines the
0.0072
1.0 1530 7,3 1.027 6.273
static failure load by the lower and upper bound
0.0082
1.2 1730 0.0111 11,s 2.198 9.602
theorems.
1.4 1930 0.0112 15,9 5.092 10.808
1.6 2010 0.0117 18,6 6.841 11.759 4.2 Dynamic post -.failure behavior
1.8 2060 0.0118 21,l 8.946 12.154
2.0 2080 0.0110 21,s 10.360 11.440 When the failure load is smaller than the structural
2.5 2110 0.0118 24,7 12.251 12.449 pile strength, the applied kinetic energy can be
3.0 2180 0.0113 29,9 17.583 12.317 greater than the limit energy corresponding to the
Figure 10 presents the results of this research in upper bound of the ultimate resistance, configuring
graphical form. The difference between AV and AW, the post - failure behavior of the system (Aoki &
represented by the shaded area, diminishes until the
Cintra 1997). In this case the kinetic energy (AT) in
value VU= 21.1 kJ has been reached. For this value excess to value (T,) of the upper bound limit is equal
the shaded area is equal to zero, fulfilling the dy- to the excess of the work done (AW) by the forces.
namic upper bound condition. Observe that the maximum applied energy can not
be greater than the limit elastic energy of deforma-
tion presented in equation (22).
Table 4 shows the values of the excess of kinetic
energy, the recovered elastic energy and the excess
of work done by the non-conservative forces. con-
sidering that the limit dynamic resistance corre-
sponds to the blow of height equal to 1.8 m.

Table 4a. Post dvnamic failure behavior Darameters

V=T R S Wnc V, AT=AV AW
PDA PDA field (PDA) (I’DA)
(kJ) (W (m) (kJ) (kJ) (kJ) (kJ)
21.1 1790 0.004 8.946 12.154 0 0
21.8 1800 0.005 10.360 11.440 0.7 1.41
24.7 1790 0.006 12.251 12.449 3.6 3.30
29.9 1810 0.009 17.583 12.317 8.8 8.64

Figure 11 presents the curve of the excess of en-

ergy and excess of work done, considering the limit
values VU = 21.1 kJ and Wnc,u = 8.946 kJ, corre-
Figure 10. Research on the limit dynamic resistance sponding to the point F of the figure 9.

645
 AW (kJ) 17 AV (kJ)
Figure 1 1. Dynamic loading: post failure behavior.
As showed by figures 10 and 11, in the post fail-
ure the excess of energy is equal to the excess of
work done in the system.
From the practical point of view, this means that,
after the dynamic failure load, all the excess of ki-
netic energy is totally transformed into excess of
work done in the system.
the pile toe. Moreover, it is considered that the
maximum toe displacement under dynamic loading
is equal to the toe displacement under static loading.
Under these basic considerations the static load - top
movement curve obtained for last 3.0 m height blow,
by the conventional CAPWAPB analysis simulation,
is presented in figure 12. Point G on the curve corre-
sponds to the maximum simulated static movement.
An alternative to this conventional CAPWAPB
simulation is the direct transformation of the dy-
namic resistance - displacement curve into a static
resistance - displacement curve. In this case the
static resistance is predicted from the dynamic re-
sistance by the application of expression (29) for
each increasing energy impact of the DIET.
Figure 13 presents the Smith's static resistance -
dynamic displacement curve corresponding to the
dynamic resistance - displacement curve of figure 9.

4.3 Static resistance - movenzent curve p o m

dynamic tests.
The displacement of the pile head under static
loading can be predicted by equations (2) and (16).
It is remarkable that the elastic pile toe displacement
expressed by equation (17) is dependent on the load
transferred to the soil by the pile toe and the pile
However the Smith's mddel toe Figure 13. Slnith's model static resistance - displacelne11t
(figure 7) is dependent only on the pile toe resis-
curve from (DIET) dynamic increasing energy test.
tance, not taking into account the effect of the load
transferred by the pile shaft to the soil layers under
The static parameters calculated from the data of
table 2 are presented in table 4.

- the DIET.
Table 4b. Static parameters resulting from
H R D V, Vs,c W, V,,,
(field) (PDA) (PDA) (PDA) (Chellis) (PDA) (PDA)
(m) (kN) (m) (kJ) (kJ) (kJ) (kJ)
0.2 810 0.0035 1.418 1.418 0.000 1.417
0.4 950 0.0051 2.826 2.422 0.404 2.019
0.6 1070 0.0054 3.129 2.889 0.240 2.649
0.8 1240 0.0072 5.208 4.464 0.744 3.720
1.0 1430 0.0082 6.543 5.863 0.680 5.183
1.2 1610 0.0111 10.951 8.935 2.016 6.920
1.4 1730 0.0132 14.458 9.688 4.770 8.378
1.6 1770 0.0147 17.083 10.354 6.729 8.936
1.8 1790 0.0158 19.041 10.561 8.480 9.241
2.0 1800 0.0160 19.400 9.900 9.500 9.400
2.5 1790 0.0178 22.631 10.561 12.070 9.23 1
3.0 1810 0.0203 27.131 10.226 16.905 9.612
Figure 12. CAPWAP'S static load - top movement curve Ohs.: 'S ' denotes parameters ullder static loading.
for H= 3.0 in blow of the dynamic increasing energy test.

646
 The analysis of figure 13 and the data in table 4 Table 6: Static and dynamic loading test procedures.
show that the static resistance R is practically con- Measured response Static Dynamic Dynamic
stant from point D to point A of the curve. In this of the pile - soil loading constant increasing
path the corresponding static resistance varies be- system energy energy
tween 1790 and 1810 kN with an average value of Shape of the curve available available available
1800 kN. Considering that the shape of this curve Energy & work done not available available available
fits well the static load - settlement curve of figure 2, Failure lower bound ? ? available
failure upper bound ? ? available
it can be concluded that the static ultimate resistance
was mobilized at point D and its value is nearly
equal to 1800 kN. require a very sophisticated interpretation. Such in-
Table 4 shows that, for the hammer fall height of terpretation can only be correctly performed by
1.8 m, the ultimate static resistance is R, = 1800 kN, highly qualified professionals.
the limit displacement is p,=15.8 mm, the limit po-
tential energy is VS,, = 19.04 kJ and the limit com-
plementary energy is the nearly constant value equal 5 CASE HISTORIES OF DIET APPLICATIONS.
to VC,,z 9.4 kJ. Point D of figures 13 and 2 corre-
sponds to point F of figure 9. 5.1 Correlation between static (SML) tests and the
The comparison between the CAPWAP’s static dynamic increasing energy test (DIET) results.
load - top movement curve and the Smith’s static re- Brazilian experience has showed that the Smith’s
sistance - displacement curve is presented in figure static resistance - displacement curve predicted from
12. Point G corresponds to the CAPWAP’s simula- the dynamic resistance - displacement curve meas-
tion of the maximum static movement of the pile ured in the dynamic increasing energy test and per-
head for the last 3 m height blow. Point A corre- formed on prefabricated reinforced concrete piles
sponds to the last point of the Smith’s static resis- has presented a good correlation with the load - set-
tance - displacement curve from the DIET. The dy- tlement curve executed in the same pile, in the tradi-
namic increasing energy test resulting from this tional slow maintained static (SML) load test.
example is summarized in table 5. Figure 14 presents the correlation between the
static resistance - displacement curve from DIET
Table 5. Dynamic increasing energy test (DIET) results performed at the end of the driving, and the static
H Dynamic Smith’s static Observation SML load - settlement curve obtained 25 days later.
(m) resistance (kN) resistance (kN) In this example the concrete pile of 11.5m of length,
1.40 1930 1730 lower bound diameter of 26 cm and 6 cm of wall thickness was
1.80 2060 z 1 800 upper bound driven with a hammer of 22.4 kN of weight, in 9 m
thick soft clay layer overlying a very hard silty clay
In this example it can be concluded that: (a) the at the pile base, in the tertiary Taubat6 Formation in
lower bound of the dynamic failure load is equal to Sgo Paulo State.
1930 kN; (b) the upper bound of the dynamic failure The maximum mobilized load was 600 kN in the
load is equal to 2060 kN; (c) the upper bound of the static and 870 kN in the dynamic test.
static failure load is equal to 1800 kN; and (d) the
lower bound of the static failure load is equal to
1730 kN (static load of Smith’s model).
Observe that the dynamic resistance continues to
increase after the dynamic resistance upper bound
has been reached, but the same is not true for the
static collapse load.
The comparison between the results obtained in
the traditional static and dynamic loading tests pro-
cedures with the results of the dynamic increasing
energy test is presented in table 6.
This table do not intend to suggest that the con-
ventional static load test is less reliable than the dy-
namic load test, either conventional or with increas-
ing energy. It was presented only to show that the
dynamic increasing energy test (DIET) provides
more information on the behavior of the pile - soil
system than the other conventional static and dy-
namic loading test procedures.
The interpretation of the static load test is rela-
tively simple, as opposed to the dynamic tests which Figure 14. Correlation DIET and SML test curves.

647
 Figure 15 presents the correlation between the
static resistance - displacement curve from DIET
performed one day after the driving and the load -
settlement curve obtained in the static SML test pe-
formed 45 days later. In this case the concrete pile of
diameter of 50 cm, 10 cm of wall thickness and 16.2
m of length, was driven with a 50 kN hammer
Figure 15. Correlation DIET and SML test curves.
weight, in varied sandy soft to medium sedimentary
soil, overlying the dense silty sand (20 <N< 50)
gnaissic residual soil in the pile base, in Belo Hori-
zonte, Minas Gerais State. The maximum mobilized
resistance was 2550 kN in the SML test and 3310
kN in the DIET. The static test shows a better
performance probably due to the set up effect.
Figure 16 presents the pile test results of a con-
crete pile, of 15.1 m of length, external diameter of
50 cm and 9 cm of wall thickness, driven with a 50
648
kN weight hammer. The sedimentary soil layers are
constituted by 12 m of soft clayey silt soft clay layer,
overlying 4.0 m of hard micaceous sandy silt and
12.0m of very hard micaceous clayey silt, in the
Tertiary Itambe Formation, Parana State. The
maximum static load was 2600 and the maximum
mobilized Smith's static resistance was 2520 kN.
Figure 17 presents the correlation between SML
static test and DIET static results performed on the
concrete pile of 50 cm of diameter and 9 cm of wall
width, 24.0 m of length and hammer weight of 55
kN, driven on shoreline soil formation, through 22.8
m thickness very soft sedimentary silty clay, under-
lying a 8.7 m thickness of compact silty sand
(23<N<56 blows /30 cm). The maximum mobilized
load was 1950 kN in the SML and 2940 kN in the
DIET test.
Figure 17. Correlation DIET and SML test curves
Figure 18 presents the results of SML and DIET
tests performed in three concrete piles, of 80 cm of
external diameter, wall width of 15 cm and 48 m of
length. These piles were driven in homogeneous ma-
rine lacustrine sedimentary soil formation, with 30 m
thickness soft silty clay (N=2) layer, overlaying 10 m
of medium over consolidated clay (N=5) and 12 m
of dense to very dense (20 <N< 50) fine silty sand.
The pile driving was done with a 200 KN weight
hammer and the pile stops on the very dense fine
sand layer. The shape of the load x settlement curve
can be better visualized in a dimensionless form.
The average structural pile strength (Re) and the dis-
placement equal to 10% of the diameter (B) were
chosen as the reference parameters for force and dis-
placements plot. In that case k = 10760 KN and B=
80 cm. Pile A was subjected to both the conventional
static slow maintained load (SML) test and the
maximum static applied load was 47% of the struc-
tural pile strength. The resulting displacement was 2
% of the pile diameter and the permanent settlement
was less than 0.7 '70of the pile diameter. The maxi-
mum mobilized static resistance in the DIET was 62
YOof the structural pile strength corresponding to the
maximum displacement of 3 YOof the pile diameter.
Figure 18. Correlation (4) DIET‘S and (2) SML test
curves on large diameter piles with the same length
Piles I and I1 were subjected only to the dynamic
increasing energy test. The set of the points shows
that the curve trends are in good agreement with the
static results of pile A. The shape of the curves are
typical of piles working by the point. The general
comparison between the DIET and SML test curves
(figures 14 to 18), shows that the load - settlement
curve of the conventional SML static loading and the
Smith‘s static resistance - displacement curve of the
dynamic increasing energy test show a reasonable
agreement for engineering purpose.
It can be concluded that the dynamic increasing
energy test (DIET), as currently used in Brazilian
foundation practice, improves the reliability of the
pile bearing capacity prediction.
5.2 Setup measured in increasing energy test
In coastal sedimentary soft clay formations it is
well established that the soil setup takes place very
quickly. But how quick is the setup effect? The
elapsed time between the end of the driving and the
start of the test can be hours or days, depending on
the pile test assemblage time. An advantage of the
dynamic increasing energy test is that the loading
can be performed at time intervals as short as de-
sired.
Figure 19 presents the results of a typical investi-
gation of the soil setup in the case of a concrete pile,
of 26.6 m of length, external diameter of 42 cm and
8 cm of wall thickness, driven with a free fall ham-
mer of 42 kN of weight, in the sedimentary marine
coastal of the Baixada Santista soil Formation in Siio
Paulo State.
649
EOD + t = 10’ t = 20’
= 40’ t = 80’ o t = 13days
Figure 19. Initial setup in very soft clay layers.
The soil was constituted by 5 m thickness soft
clayey sand layer (N=2) and 9 m of very soft marine
sandy silty clay (N=l), overlaying 12 m thickness
soft to medium fine clayey sand (1 <N<7).
These sedimentary layers rest on 2 m thickness
very dense, fine to medium silty sand young residual
gnaissic soil (25 <N<37).
In this example the measurements are done just at
the end of the driving (EOD) and after 10 minutes,
20 minutes, 40 minutes, 80 minutes and 13 days af-
ter the EOD. For comparison 16 points correspond-
ing to the last 35 cm of driving were registered.
By taking the referential line corresponding to 15
mm of displacement, it can be seen that the setup ef-
fect was 70% in 13 days, but half of this value was
developed in the first 80 minutes.
This measurement confirms the local practice that
it is enough to stop the driving for some minutes to
get smaller sets due to the soil freezing. In the con-
ventional static load test it was not possible to per-
form this kind of research.
6 CONCLUSIONS
The dynamic increasing energy test (DIET) is a
promising procedure to evaluate the behavior of the
pile - soil system. Its potentialities have not been
fully explored and more experience is necessary be-
fore its reliability on the prediction of the pile bear-
ing capacity from measured dynamic pile behavior
can be confirmed.
The authors express their gratitude to SCAC Fun-
daqiies e Estmturas Ltda, by the support given in the
last 18 years of research on the behavior of driven
reinforced concrete piles under dynamic loading
conditions.
REFERENCES
Aoki, N. 1989a. Prediction of the Behavior of Vertical Driven
Pile under Static and Dynamic Conditions. Drivability of
Piles (2): 55-61; Proc.Disciission Session 14, Int. Con$
Soil Mech. and Fozind. Eng. (12), Rio de Janeiro, August
I9S9. Japan : Japanese Soc. Soil Mech. and Found. Eng.
Aoki, N. 1989b. A new dynamic load test concept. DrivabiliOJ
of Piles (I): 1-4; Proc.Discussion Session I4, Int. Con$
Soil Mech. and Foui7d. Eng. (IZ), Rio de Janeiro, August
1989. Japan : Japanese Soc. Soil Mech. and Found. Eng.
Aoki, N. 1 9 8 9 ~Discussion
. on Driveability of Piles. Proc. Int.
Con$ Soil Mech. Found. Eng. (I2), Rio de Janeiro, Discus-
sion Session I j : ( 5 ) : 2977-2978. Rotterdam: Balkema.
Aoki, N. & de Mello, V.F.B. 1992. Dynamic loading test
curves. In Barends (ed.), Proc. Iift. Coif on Application of
Stress- Wave Theory to Piles (4); 525-530. Rotterdam:
Balkema.
Aoki, N. 1997. Evaluation of the ultimate bearing capacity of
piles driven with increasing energy (in Portuguese). Doctor
of Engineering thesis, School of Engineering of Sao Car-
l o ~ University
, of Sao Paiilo: 1 10.
Aoki, N. & Cintra, J.C.A. 1997. New interpretations of the dy-
namic loading curves for driven piles based on the energy
approach. In M. Almeida (ed.), Recent Developments in
Soil and Pavement Mechanics: 467-472. Rotterdam:
Balkema.
Beer, E.E. 1988. Different behavior of bored and driven piles.
In W.F.van Impe (ed.), Proc. Int. Geotechnical Seminar:
Deep Foundations on Bored and Auger Piles: 47-82. Rot-
terdam: Balkema.
Beim, J.W. & Aoki, N. 1996. Dynamic load test method with
variable energy. In F.C. Townsend & M.Hussein & McVay,
Proc. Int. Con$ on Application of Stress-Wave Theory to
Piles (5): 274-28 1. Orlando.
Bernardes, G.P. 1989. Dynamic and static testing of large
model piles in sand. Doctor of Engineering thesis, Depart-
ment of Civil Engineering, Norwegian Imtitute of Technol-
ogy, Trondheim.
Broms, B.B. & Choo, L.P. 1988. A simple pile driving formula
based on stress-wave measurements. Proc. Int. Con$ on
Application of Stress-Wave Theoiy to Piles (3): 59 1-600.
Ottawa: Canada.
ChelIis, R.D. 1951. Pile foundations. New York: McGraw-Hill.
Chen, W.F. & Liu, X.L. 1990. Limit analysis in soil mechanics.
Developments in Geotechnical Engineering (52). Amster-
dam: Elsevier.
Clough, R.W. & J. Penzien 1975. Dynamics ofstrzxtures. New
York: McGraw-Hill.
Davisson, M.T. 1972. High Capacity Piles. Proc. of Lectitre
Series. Innovations it7 Foundation Construction. Soil Mech.
and Found. Div. Illinois Section. ASCE and Dept. of Civil
Eng. Illinois Inst. of Technology. Chicago.
Fellenius, B.T. 1980. The analysis of results from routine pile
load tests. Ground Engineering v. I3 n. 6, September: 19-3 1.
Fujita, K. & Kusakabe, 0. 1988. On the evaluation of static
bearing capacity. Proc. Int. Conf on Application of Stress-
Wave Theory to Piles (3): 525-534. Ottawa: Canada.
Goble, G.G.; Likins Jr., G.E. & Rausche, F. 1975. Bearing ca-
pacity of piles from dynamic measurements. Final Report.
Depart. of Civil Eng. Case Western Reserve University,
Cleveland
Goble, G.G.; Rausche, F. & Likins Jr., G.E. 198 1. The analysis
of pile driving - A state-of-the-art. In H. Bredenberg (ed.),
650
Proc. Int. Seminar on Application of Stress-Wave Theory
on Piles: I3 I - 16 1. Rotterdam: Balkema.
Goble, G. & Likins, G. 1996. On the application of PDA dy-
namic pile testing. In F.C. Townsend & M.Hussein &
McVay, Proc. Int. Conf on Application of Stress-Wave
Theory to Piles (5): 263-273. Orlando.
Koten, H.; Sandbergen, R. & Starink, T.W. 1988. Model test
on static and dynamic bearing capacity of piles. Proc. Int.
Con$ on Application of Stress-Wave Theory to Piles (3):
368-376. Ottawa: Canada.
Langhaar, H.L. 1962. Energy methods in applied mechai7ics.
New York: John Wiley.
Middendorp, P. 1993. First experiences with statnamic load
testing of foundation piles in Europe. In W.F.van Impe
(ed.), Proc. of the 2'"' Int. Geotechnical Seminar on Deep
Foundations on Bored and Auger Piles: 265- 272. Rotter-
dam: Balkema.
Niyama, S. & Aoki, N. 1991.Correlation between dynamic and
static loading tests in Experimental Research of EPUSP-
Campus. (in Portuguese). Proc. Seininario Engenharia de
FundaGGes Especiais - SEFE 11, SE0 Paiilo (1): 285-293.
Poskit, T.J. 1992. Keynote lecture: Problems of reconciling
stress wave measurements with theory. In B.J. Barends
(ed.), Proc. lilt. Conf on Application of Stress- Wave The-
oty to Piles (4): 495-507. Rotterdam: Balkema
Randolph M.F. & Deeks A.J. 1992. Keynote lecture: Dynamic
and static soil models for axial response. In B.J. Barends
(ed.), Proc. Int. Conf on Application of Stress- Wave The-
oty to Piles (4): 3- 14. Rotterdam: Balkema.
Rausche, F.; Goble, G.G. & Likins Jr., G.E. 1985. Dynamic
determination of pile capacity. Journal of Geotechnical
Engineering ASCE: v.111 n.3: 367-383.
Reese, L.C. 1972. Axial capacity. Proc. Specialty Conference
on Petformance on Earth and Earth-supported Structures.
Purdue Ut7iversity - ASCE, June.v.111: 20 1207.
Smith, E.A.L. 1960. Pile driving analysis by the wave equation.
Journal of Soil Mech. and Found Division, ASCE, v.86,
SM4: 35-61.
Terzaghi, K. 1942. Discussions. Pile-driving formula. Progress
report of the Committee on the bearing value of pile foun-
dations. Discussion. Proc. American Society of Civil Engi-
neers, v.68 n.2: 31 1-323.
Terzaghi, K. 1943. Theoretical soil mechanics: 137- 143. New
York: John Wiley.
Vesic', A.S. 1975. Principles of pile foundation design. Soil
Mechanics Series No. 38. Duke University, School of Engi-
neering.
Vrouwenvelder, A. & Dieterman, H.A. 1992. General report 4:
Reliability of predictions. In B.J. Barends (ed.), Proc. Int.
Conf: on Application of Stress-Wave Theory to Piles (4):
509- 5 17. Rotterdam: Balkema.
Veen, C. 1953. The bearing capacity of a pile. Proc. Int. Conf
on Soil Mech. and Found. Eng. (3). ( 2 ) :84-90. Zurich.
Weele, A.F. 1993. Quality assesment foundation piles after in-
stallation. In W.J.van Impe (ed.), Proc. Int. Geotechnical
Seminar: Deep Foundations on Bored and Auger Piles (2)
BAP 11: 459-467. Rotterdam: Balkema.
Whitaker, T. & Bullen, F.R. 198 1. Pile testing and pile-driving
formulae. In F.E. Young (ed.), Piles and Foundations
12 1-134. London: Thomas Telford.
Whitaker, T. & Cooke, R.W. 1961. A new approach to pile
testing. Proc. Int. Con$ on Soil Mech. and Found,. Eng.
(4). (2): 17 1-1 76. Paris: Dunod.
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Correlation analyses of dynamic and static loading tests for nine piles
Y.M.Zheng, J.M.Zheng & B.Chen
Shunde Supervision Stutionfor Quality ancl Sufety of Construction Engineering

ABSTRACT: Can the accuracy of dynamic pile testing meet the demands of engineering? Engineers are al-
ways concerned with this problem as well as fall into arguments on it. In this paper, the correlation results of
dynamic and static loading tests for nine prestressed-concrete pipe piles are studied. The two testing methods
are compared in three aspects including the capacity, the settlement ratio and the load-movement curve. In
addition, causes of diversity are discussed. As a tool for the acceptance test, the dynamic testing method is
evaluated based on comparisons and analyses.

1 INTRODUCTION dynamic tests, correlation can be studied in three as-

The dynamic method mentioned in this paper is the
hi h strain testing in which the Pile Driving Analyz-
e?" manufactured by PDI is employed and meas-
ured curves are matched with the CAPWAPC"'
program Dynamic testing is performed in accor-
dance with Specification for High Strain Dynamic
Testing of Piles (JGJ 106-97 1997), which is a ge-
neric standard of the People's Republic of China
The static loading test is carried out by the main-
tained load test in which reaction load system is
used The standards of the static loading test are
Code for Building Foundation Design (GBJ 7-89
1989) and Technical Code for Building Pile Foun-
dations (JGJ 94-94 1994), which are a national stan-
dard and a generic standard of China respectively
As for the theory of the high strain testing and the
detail specifications of the static loading test, please
refer to the standards mentioned above and Applica-
tion of PDA dynamic-testing method to bored piles
(Zheng & Ma 1996)
Dynamic tests were performed on nine test piles
first, then the pile predicted capacities were used as
the basis of static load increments, so that the error
due to rough increments might be reduced Pile in-
stallation parameters, soil conditions and test results
are illustrated with tables and figures in the paper
Correlation between the dynamic and static tests
is investigated by two quantitative analyses and one
qualitative analysis They are the bearing capacity of
single pile, the settlement under the ultimate load
(Sd in dynamic test, Ss in static test) and the shape
of the load-movement curve For piles tested to fail-
ure in static tests as well as mobilized sufficiently in
pects. However, for piles neither tested to failure
statically nor full capacity mobilized dynamically,
only latter two aspects, the settlement and the load-
movement curve, can be studied.
2 PILE INSTALLATION AND SOLL CONDITIONS
The nine test piles are chosen from five sites in
Shunde district, Guangdong province, China Four
of them were installed by silent piling in two sites,
the others were driven with a diesel hammer. One is
0500mm, four are 0400mm and others are
0300mm The shortest pile is about 7m long, the
longest one is 38m, the others are from 20m to 30m,
(Details see Table 1)
The boring log of the hole nearest to one test pile
is referred as the soil conditions of this pile Among
them, since pile #81 of Meidi Shore Garden is far
from any boring hole, its penetration depth is less
than the bearing stratum depth according to its near-
est hole, no boring data can be used for analysis
Thus, corresponding stratum description of only
eight tested piles is given in Figure 1 (The ground-
water levels of all sites are less than 1 m below the
ground surface)
3 TESTING RESULTS AND ANALYSES
The correlation results of dynamic and static loading
tests for nine test piles are shown in Table 2. Figure
2( 1)-2(9) present the comparisons of the static load-
movement curves and dynamic simulated ones for

65
 I
kpth Pile #33 of Pile #76 of Pile #85 of Pile #27 of Pile#113 of Pile #27 of Pile #429 of Pile #709 ol

(m) Gaozan Gaozaii Gaozan Gigui Yujing Meid Huarun Huarun

1 fill plain fil
2 plain fil plain f i U vlain fill silty clay
3 i f h i d plastic) muckv cla. dredge
1 fine sand fine sand fine sand silt Sand
5 (loose (loose (loose) (little-dense) fine sanc
6 (lower p: LS siltv sand

7 mucky silty mucky silty 1llUCky s1lty trongweathcrec muck siltv sanc
8 salid (loose sand (loose) sand (loose) siltstonc coarsesill(
9 fine silt silw clar
10 sand(1oose)
11
12
13 fine sail fine saiil fine saic gravel sand
11 (loos (loose (loose' p v e l sand
15
16
17 sandv clav
18 sandy c l q
19 muck
20 strong \v&m
21 IilIlCk?' Silt! muck! silt! muclq Silt!
22 sand (loose sand (loose salid (loose1
23
21
25 silt!, cla! silt?. c1a.i silo claj mi1g \\eatllcIx muck
26 (fluid plastic (fluid dastic (fluid plastic ilt iiiterlaycr fluid plastic
27 mucky silt!
28 salid (loose
29 muck? silt? muck! silt!
30 sand (loose sand (loose)
31 fine sand
32 puddtiig stoiir puddtng stoiic medium-densit!, )
33
34 oarse g r x el
7 -
13

36 mllg \wltllc
37 inudstoiic
38
39
40

Figure 1. Conespoiidtng stratuin descriptioii of eight test piles

652
Table 1 Installation Parameters of tlie iiiiic prestressed pipc piles
Project Pile Pile Em- Date of In- Design Installation Rani Weight Drop Total Blou Count Set of the last
Name No Dia bednient stallation Bearing Method / Ma\iiiiuni Height B l o ~ of the last 30 blom s
(mm) (in) Cap& PllC CoIn- (cm I 10
ity pressing blows)
(kN) Force
W)
Gaomn 33 300 29.0 Mar. 25.99 800 1700
Market 76 300 3 1.0 Mar. 3 1.99 800 silent piling 1700
85 300 31.0 Apr 3. 99 800 1700 ----
Bigui 27 400 7.2 Apr. 9. 99 1200 Dicscl hanimer 65 1.1. 1.1. 1.1
Garden
Y uj i ng 113 400 24.5 Apr. 4. 99 1200 Static press 2300
Garden
Meidi 27 400 38.2 Jun. 3. 99 1400 Diesel hammer 62 2.0 1506 113 1.5. 1.5. 1.5
Shore (0.2111)
Garden 81 400 29.0 Jun. 3 1. 99 1400 62 6.2 396 308 1.2. 1.2. 1.1
Huaruii 429 400 24 8 May 23, 99 1500 Diesel hanuner 50 2.0 1369 219 1.2. 1.2. 1.1
(0.8111)
Fiz 709 300 19.6 May 23. 99 1000 50 1.5 848 256
(0 6111)
1.5. 1.5. 1.5

Table 2 Correlation results of dyiatnic and static loadmg tests for nine prestressed-concrete pipe piles
Project Pile Pile EnibedDate of Design Dq iiaiiuc Testing Results Static
Name No Dia nieiit Installation Bearing Bearing Rd St" Sd* Ss* sf
(111111) (111) Capaeitj, Beaniig Shaft Re- Toe Re- Capacit!
(kN) CapaciQ sistaiice sistarice Rs
is (111111) (mm) (mm)
ss
Rd (kN) (lw OtN)
(kN)
Gao7a1i 33 300 2 9 0 Mar 25.99 800 1863 1341 522 1900 098 6535 3 3 6 3 1 3 107
Market 76 300 31 0 Mar 31.99 800 1881 1417 464 1900 0 99 63 30 32 2 3 7 2 0 87
85 300 31 0 Apr 3,99 800 2188 1432 756 1980 111 7224 35 3 4 1 8 083
Bigui 27 ** 400 7 2 Apr 9 99 1200 4007 431 3576 r306O ---- ---- 21 1 1 6 4 129
Garden
Yujiiig 113 400 2 4 5 Apr 4.99 1200 2051 655 1396 2160 095 5207 22 1 21 9 101
Garden
Meidi 27 ** 400 38 2 Jun 4. 99 1400 4710 4110 600 2 5100 ---- ---- 43 7 3 9 9 1 10
Shore
Garden 81 ** 400 2 9 0 Jun 31.99 1400 4453 2619 1834 24700 ---- ---- 364 306 119
Huarun 429 **300 24 8 Maj 23. 99 1500 4734 3605 1129 2 5225 ---- ---- 2 6 4 21 2 124
Dope
Plant 709 ""300 19 6 May 2.3, 99 1000 2943 1990 953 2 3000 ---- ---- 25 1 195 123
* St is the settlcineiit under the static failure load.
Sd is the settleinent under tlie dynanuc bearing capacity or the largest testing load.
Ss IS tlie settlenient mider the static beaniig capacity or the largest testing load
Neither the static loadtiig test nor the dynaniic testing reached the ultmate bearing capacity

each pile. Due to limited space, only two CAPWAP
matching results are given in Figure 3 and Figyre 4.
Comparing both the capacities and the load-
movement curves, we can see that:
1. The dynamic result of each test pile is a little
less than its static value, except pile #85 of Gaozan
Market.
2. For piles mainly depending on the friction re-
sistance, such as: three piles of Gaozan Market and
pile #113 of Yujing Garden, which were installed by
silent piling, failure loads were obtained in static
tests and soil resistance sufficiently mobilized in dy-
namic tests, the capacity errors between the static
and dynamic are not more than 11%, settlement ra-
tios(Sd/Ss) =l, the load-movement curves of two
methods are highly close with the second knee
points almost identical.
3. For pile #27 of Bigui Garden, two piles of
Meidi Shore Garden and two of Huarun Dope Plant,
which were driven with a diesel hammer, the main
resistance is at or near the pile toe and the static
loading test curves are belong to "slight slope" with-
out steeply downward sections. The total settlement
couldn't achieve the failure criterion according to
Code, therefore the ultimate bearing capacity
couldn't be determined. Sd is a little more than Ss
for this type, the ratios are from 1.1 to 1.3. As to the
load-movement curves of two methods, three of five
piles are analogous. While pile #429 and pile #709
are much the same as others at the

653
beginning and separate obviously from the middle
part, and that the load-movement curves from dy-
namic tests are below the static ones.
We have given Some preliminary analyses to
above correlation results:
Transient impact load is applied to mobilize the
soil resistance around pile in the dynamic testing,
654
which is different from the static loading test. The
pile capacity is gained by slowly loading in static
tests. When the dynamic test is being performed,
impact is applied on pile top for many times with a
heavy hammer. In general, the upper resistance
around the pile shaft is mainly mobilized by former
blows. While the lower shaft resistance and end-
Figure 4. CAF'WAF' results of Pile #429 of Huarun
bearing capacity act gradually under later blows, at
that time, pile has possibly lost small part of top
friction resistance. This is the reason why the dy-
namic results are sometimes conservative.
Besides, according to transfer mechanism of ver-
tical load for pile foundations, adequate mobilization
of shaft and toe resistance needs a certain displace-
ment of the pile, which is related to the pile diameter
655
and soil around the pile. Commonly mobilization of
shaft resistance requires 5-1 Omm displacement and
10-20mm is needed for sandy soil while for mobili-
zation of toe resistance, even more displacement
may be required.
The displacement may come to 20-30mm in the
dynamic testing. Such a displacement is enough for
friction piles to be mobilized adequately. That is
why three piles of Gaozan Market and pile #113 of
Yujing Garden have good correlation Nevertheless
if lower soil layers have good conditions or the pile
toe is built-in firmly, it will be difficult to have all
resistance mobilized For example, in Huarun site,
the middle soil layer around the pile is gravel sand
and the pile toe is built in intensely weathered rock
The penetration of the end driving was only
1 2-1 5cmllOblows (normally 2 0-5 OcdlOblows)
The dynamic testing was performed one week after
driving Thus, it should be expected that soil resis-
tance of piles is hard to be hlly mobilized in this
site
It should be explained that although the dynamic
results are more conservative than the static ones for
the two piles of Huarun See Table 2, #429 is a
400mm dia pile with a design capacity of 1400kN,
and the soil resistance mobilized in dynamic testing
is up to 4734kN, which is more than three times of
the design value, #709, 300mm dia , design capacity
is lOOOkN, the dynamic capacity is 2943kN, near
three times more than the design value Both of re-
sults are more than twice of the design vertical load
capacity of pipe pile defined in Technical Specifica-
tion for Prestressed Concrete Pipe Pile Foundations
(DBJ/T 15-22-98 1998), which is depend on the ma-
terial strength of pile shaft It can be calculated by
following formu1a
R1, = 0.3 (fee - o p c ) A (1)
Where R, = the design vertical load capacity of
the pile shaft; f,, = the compression strength of the
centrihgal concrete pipe pile; oPc= the effective
prestressing force of the concrete pile; A = the sec-
tion area of the pipe pile.
This capacity for 400mm dia. pile is about
4200kN and for 3OOmm dia. pile is about 2400kN
Hence, it is insignificant for design engineers to in-
crease impact force in order to gain more resistance
in dynamic testing, In addition, the ratios between
the static and dynamic settlements of pile top are
from 1.O to 1.1 under the design load. Therefore, for
practice, the difference between the results of the
dynamic testing and static loading test are not much
at all.
CONCLUSION
1. If the soil resistance has an adequate mobiliza-
tion, results of the dynamic testing and the static
loading test are quite close for most of prestressed
pipe piles. That is to say, the measurement accuracy
of dynamic testing method may be ensured when the
correct testing procedure is carried out. Therefore, in
general, dynamic pile testing can be used for the ac-
ceptance of piles instead of the traditional static
loading test.
656
However, for the short (less than 1Om) or extra
long (more than 40m) piles where the resistance ac-
cording to the safety requirement of design is hard to
be mobilized, the static test is recommended. For
pipe pile foundations of the first grade building, in
order to ensure safety, it is suggested that testing
piles should be driven in the design stage and dy-
namic testing should be performed during the driv-
ing and restriking. And then, static loading test
should be carried out. After the comparison of two
testing results, dynamic method may serve as an ac-
ceptance test for working piles.
Today in Shunde district, a great number of pipe
pile foundations and square pile foundations are
tested and accepted by the dynamic testing (More
than seven thousand prestressed pipe piles have been
tested in recent seven years). Considerable testing
cost has been saved from this. As for the advantage
in enlarging the testing quantity, ensuring safety and
quickening the tempo of construction, it is much
more than economical.
2. In Shunde, the design capacity of the
prestressed pipe piles is always on the low side. For
300mm dia. piles, design capacity is less than
lOOOkN, 400mm dia. less than 1500kN and 500mm
dia. less than 2000kN. The measured capacities are
close to or even more than three times of the design
values
(such as the five of nine piles in the cases chosen in
this paper) The annual amount of the working pipe
piles is up to tens of thousands, it will cause enor-
mous waste and losses. We suggest that the design
vertical load capacity defined in Technical Specifi-
cation for Prestressed Concrete Pipe Pile Founda-
tions, which is depend on the material strength of
pile shaft, should be applied for designing, if the pile
length and soil conditions are similar to the Huarun
and Meidi site. In this way, both safety and economy
would be achieved.
We are gratehl to Prof. De-qing Li for his sug-
gestion on this paper.
REFERENCES
DBJ/T 15-22-98 1998. Teclinical Specjfication for Prestressed
,
Coiicrete Pipe Pile Fouidations: 8. A standard of Guang-
dong Province.
GBJ 7-89 1989. Code Sc)r Building Foundafions Desig17: 104-
105. A national standard of the People’s Rcpublic of Cluna.
Publishing House of Construction Industry of China.
JGJ 94-91 1993. Technical Code for Buililing Pile Fouiida-
tions:130-136. A generic standard o f the People’s Republic
o f China. Publishiiig House of Construction Industry o f
China.
JGJ 106-97 1997. Specification for High Strain Dvnaniic Test-
i17g of Piles. A generic standard of the People’s Republic of
China. Publishing House of Construction Industry of China.
2%ig Y.M. & C. Ma 1996. Application of PDA dynamnic-
&sting method to bored piles. Proceedings of PDMPIT Us-
ers: 83-94.
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkerna, Rotterdam, ISBN 90 5809 150 3

Back-analyses of steel pile driving records for quality assurance

B. R. Danziger
Department of Civil Engineering, Fluminense Federal University,Nitero'i,Brazil
J. S.Ferreira
Coupe, Rio de Janeiro Federal University,Brazil

ABSTRACT: Steel piles are driven, in many cases, deeper than the SPT borings. To improve confidence in
bearing capacity estimation for piles with penetration exceeding boring depth, steel pile driving records were
back-analysed by the wave equation and the Danish formula. A comparison of the results by both methods is
presented. Only those data including piles driven to penetrations shorter than SPT boring depth and located no
farther than 5 metres from the SPT were selected. A good correlation between the driving resistance evaluated
by the wave equation and the Danish formula was found. The authors then assured the confidence of the
Danish formula for steel piles with characteristics within the range of the data base.

1 INTRODUCTION selected for the analyses, as only this small part

contained complete and confident data set, including
Regardless of all the criticism involved in pile an SPT boring not farther than 5 metres away from
driving formulas, they are still used in many the pile and extending below pile tip.
construction sites as a simple method available to Most discarded driving records were related to
control driving operation in the field. Although it is piles driven deeper than the SPT borings. The
well known that the wave equation model can more excessive driving length is mainly due to the high
faithfblly describe pile behaviour during driving, the impedance of steel piles and the close range of pile
driving formula is easier and less time consuming set specifications. The soil nature around the pile
for practical applications. Furthermore, pile driving length below SPT boring, in those cases, is
formula calculation can be extended to the whole unknown, posing difficulties for the determination of
piling, providing an efficient tool for quality pile bearing capacity.
assurance of the foundation job.
Pile driving formulas are based on the
conservation of momentum and energy (Chellis 2 THE DANISH FORMULA
1961).
The Brazilian experience incorporated the Danish The Danish formula (DGI 1978) is written as:
formula for application to steel piles on land. They
are used mainly as a means to assure uniformity
control in construction for a given pile job.
Smith (1960) introduced the wave equation
method that solves the problem of wave
transmission along pile length, modelling pile and
soil behaviour during driving in a more adequate where
form. The wave equation allows the introduction of
the real boundary conditions provided by the
embedded soil.
The present paper analyses the driving data from
steel piles using both the Danish driving formula and Q,, - driving resistance
the wave equation. In spite of the numerous driving 7 - efficiency factor
records available, nearly 10% of the records were G - hammer weight

657
H - hammer drop
L - pile length
A - pile section area
E - modulus of elasticity of pile
s - pile penetration
SO - elastic pile displacement
The DGI (1978) bulletin presents some
recommendations concerning mean value of pile
length to be considered for piles shorter than 20
times the width of the pile. All the piles of the data
base are longer than that limit, so the actual pile
length was considered in the analysis. It is also
mentioned in the DGI (1978) bulletin that in case of
no direct measurements, q should be assumed as 0.7.
In the present analysis a q value of 0.7 was adopted,
as it also corresponds to the usual procedure in the
application of the Danish formula in Brazil.
The driving resistance will be evaluated in this
paper for the two records usually considered in
Brazil: (1) the blow counts for the last 50 cm
penetration and (2) the pile set at final penetration,
related to the 10 last successive blows. For the
whole data base, both records (1) and (2) were taken
with a drop hammer falling from one metre height.
Table 1 presents the data base containing the
complete set of information related to the 19 driving
records, including SPT borings extending below pile
toe. The columns include, from left to right, the job
number, a description of a simplified soil profile, the
pile designation and type, the pile section area, the
pile length, the hammer weight, the blow counts for
the last 50 cm penetration (record number (1)) and
the pile penetration for 10 blows (record number
(2)).
3 THE WAVE EQUATION ANALYSES
Two wave equation programs were used to estimate
pile resistance during driving for the analyses
described in this section: the Weap program (Goble
et a1 1980) and the Dinexp program (Costa et a1
1988). Details concerning the use of the programs
are discussed elsewhere, GRLWEAP (1988), Costa
et a1 (1 988) and Danziger (1 99 1).
The data needed to run the program were
available in the pile driving bulletin, except soil
nature and the soil resistance distribution along pile
shaft and toe. The soil nature was obtained by means
of the SPT boring profile close to the pile. The
estimation of soil resistance during driving is
necessary to make a proper choice of toe resistance
percentage and side resistance distribution along pile
shaft.
The driving of steel piles on land is usually
performed with the hammer striking directly the top
of the pile. Since no cap was used, and the Weap
program can not be run without a cap, a solution was
found by modelling the pile 0,lO m shorter and
considering the pile top as the cap. The same
procedure was introduced in the Dinexp program.
The data preparation and analysis of each pile
followed the sequential procedure:
i) Establishment of soil profile (layering and
estimation of soil parameters).
ii) Estimation of soil resistance distribution during
driving for the embedded pile length.
iii) Establishment of driving data (hammer, pile).
iv) Performance of driving analysis.
v) Drawing the driving curve.
vi) Introduction of the blow counts from the pile
driving bulletin in the driving curve.
vii) Determination of the driving resistance.
The estimation of soil resistance distribution
during driving (step ii above) has been performed
according to Semple and Gemeinhardt (1981) and
Stevens et all (1982).
The soil damping at the pile point (Jp) and at the
pile shaft (Js) introduced in the wave equation
analyses were (Aas, 1993):
Sand layers: J, = 0.1 s/m
J, = 0.5 s/m
Clay layers: J, = 0.65 s/m
J, = 0.15 s/m
The quake values were assumed as 2.54 mm
(Smith 1960).
The driving resistance was obtained from the
driving curve, both for the blow count related to the
last 50 cm penetration of continuous driving and
also for the 10 successive blows.
The soil resistance during driving calculated by
the Danish formula and the Wave equation (Weap
and Dinexp program) are indicated for the two
records in Table 2.
4 STATISTICAL ANALYSES
Two statistical analyses were carried out. A one
population analysis which included the whole data
base and a two population analysis, in which the
data were separated according to soil type mostly
present along the pile length. Unfortunately, the data
base included predominantly clayey soils. Piles
embedded mainly in sandy soils included only five
cases, which were not sufficient for a statistical
analysis.
Table 1. Pile Data Base.
Job Simplified Pile Section ~ ~ ~ gHaininer
t h Blow Set 10
Number Soil Profile Weight counts (1) blows (2)
Depth (m) Type of soil Designation Type (cm’) (m) (kN) (I 5Ocin) (inm)
1 0-9.2 sandy clay B- Bloc0 2 I
9.2-14.0 silty sand 12” 5 1/- 77 3 14.50 91 192 0
14.0-19.0 sandy silt
1 0-9.2 sandy clay B- Bloc0 1 I
9.2-14.0 silty sand 1277 5 w 77 3 14.50 91 194
14.0-19.0 sandy silt
2 0-10 silty sand M I1
10-20 sandy clay 10 10” 15/8” 96 2 22.00 12 2 204
20-25 clayey sand
3 0-8 clayey sand 16 0
8-10 sandy clay F3 1 120u120sll 181 8 26.65 13 1 80
10-18 clayey sand imn
18-30 sand! cla!
r
3 0-3 sandy cla! 16
3-6 silh sand D12 120s120sll 184 8 29.50 43 1 93 18
6-18 sandy silt iniii
18-30 sandy clay
3 0-3 sand! clay 36 L

3 -6 silh sand D13 120u120\11 181 8 28.60 13 1 62 25

6-18 sandy silt inin
18-30 sandy claj
3 0-3 sand! clay 16 U
3 -6 silty sand D9 420u120ull 184 8 29.18 43 1 208 25
6-18 sandy silt inin
18-30 sand! cla!
3 0-3 sand! cla! 16 L1
3-6 silt! sand D11 120u120ull 181 8 29.19 13 1 156
6-18 sand! silt inm
18-30 sand! cla!
3 0-3 sand! cla: 16
3-6 silt? sand D 10 120\420\11 181 8 29.70 13 1 175
6-18 sand! silt llllll
18-30 sand! cla!
1 0-3 sand! cla\ 16
3-6 silo sand D5 120u120.;ll 181 8 25.00 13 1 71
6-18 sandy silt mni
18-30 sand! claj
3 0-3 sandy cla! 16 -1
3-6 silo sand E1 12Ox120ull 181 8 29.25 13 1 156 23
6-18 sand! silt 111111
18-30 sand! claj
3 0-3 sand! clay 16
3-6 silh sand E5 120u120u11 181 8 29.00 13 1 85 25
6-18 sand! silt inm
18-30 sand! cla!
3 0-3 sand! cla! 16
3-6 silty sand E6 420\120\11 18-18 29.50 13 1 85 25
6-18 sand! silt inin
18-30 sandy clay

659
Table 1. Pile Data Base (cont.).
Job Simplified Pile Section Length Hammer Blow Set 10
Number Soil Profile Weig11t counts (1) blows (2)
Depth (m) Type of soil Designation Type (cm') (in) (kN) (I 50cm) (mm)
3 0-3 sandy clay 46 c
3-6 silty sand E3 420u420x11 184.8 30.50 13.1 80 20
6-18 sandy silt mm
18-30 sandy clay
7 0-3 sandy clay 46 U
3-6 silty sand E2 120u420ull 184.8 3 1.OO 43.1 72 25
6-18 sandy silt inm
18-30 sandy clay
4 0-8 clayey silt P I1
8-10 silty clay 130L 10'' 4 518" 96.2 12.00 12.2 191
10-13 sand
1 0- 12 clayey silt P I1
12-17 silty clay 7A 10" 4 518" 96.2 11.00 12.2 150
17-18 clayey sand
5 0- 1 silty sand P I
1-1 sandy clay 11B 12" 6 9/32" 78.0 18.50 7.5 250
1-27 silty clay
6 0-10 sandy clay P I
10-12 sand M20 10" 4 518" 18.1 14.70 6.8 219 11
12-18 sandv silt
(1) blow counts for the last 50 cm penetration.
(2) the pile penetration for 10 successive blows.

Table 2. Soil Resistance during Driving. Danish Formula and Wave Equation Programs (Weap and Dinexp).
Job Simplified Pile SRD (kN) SRD (kN) Weap SRD (kN) Dinesp
Number Soil Profile Type Danish Formula
Depth (in) Type of soil (1) (2) (1) (2) (1) (2)
1 0-9.2 sand! cla! I
9.2-11.0 silh sand 12" 5 'A'* 802.9 1194.3 813.1 952.0
11.0-19.0 sandy silt
1 0-9.2 sandy cla! I
9.2-11.0 silo sand 12'' 5 %" 804.9 1191.3 838.5 955.5
14.0-19.0 sandy silt
2 0-10 silh sand I1
10-20 sand) cla! 10" 4 518" 921.3 1065.0 1290.5 1559.2 1858.0 2750.0
20-25 clale) sand
r
3 0-8 cla!e! sand L!

8-10 sandy clay 420s420\rll 1836.5 2477 4 1613.6 2263.1 1756.0 2590.0
10-18 cla~e! sand mm
18-30 sand! c l a ~
-
3 0-3 sand! cla) -
3-6 s i l ~sand 420\;120xll 1876.4 2112.0 1570.4 2201.8 1867.7 2585.0
6-18 sand! silt min
18-30 sand! c l a ~
1
7 0-3 sand! cla!
3-6 s i l sand
~ 420s120sll 1627.0 2313.0 1321.0 2014.8 1510.0 2338.5
6-18 sand! silt I11111

18-30 sandy clay

660
 Table 2. Soil Resistance during Driving. Danish Formula and Wave Equation Programs (Weap and Dinexp)
(cont.).
Job Simplified Pile SRD (kN) SRD (kN) Weap SRD (kN) Dinesp
Number Soil Profile TYPe Danish Formula
Depth (in) Type of soil (1) (2) (1) (2) (1) (2)
3 0-3 sandy clay -
3-6 silty sand 120.1120sll 2311.9 2291.3 2294.3 2011.8 2366.0 2140.0
6-18 sandy silt inm
18-30 sandy clay
1
0-3 sandy clay 3
3-6 silty sand 120.1120sll 2172.0 2377.1 1923.5 2191.0 2170.5 2360.0
6-18 sandy silt inin
18-30 sandy clay
3 0-3 sandy clay -
1

3-6 silty sand 12Ox12Osll 2217.6 2312.8 1935.0 2037.7 2335.5 2175.0
6-18 sandy silt mm
18-30 sandy clay
3 0-3 sandy clay 0
3-6 silty sand 120x420.111 1785.1 2513.3 1179.4 2265.5 1721.6 2755.0
6-18 sandy silt niin
18-30 sandy clay
3 0-3 sandy clay C
3-6 silty sand 120s120sll 2176.3 2327.1 1975.3 2129.5 2270.0 2199.0
6-18 sandy silt inm
18-30 sandy clay
3 0-3 sandy clay 0
3-6 silty sand 120s420ull 1828.8 2300.0 1577.3 2079.3 1750.0 2110.0
6-18 sandy silt inin
18-30 sandy clay
3 0-3 sandy clay C
3-6 silty sand 120s120sll 1818.7 2281.2 1587.0 2086.2 1810.0 2156.0
6-18 sandy silt inin
18-30 sandy clay
3 0-3 sandy clay I
-
3-6 silty sand 120.1120sll 1760.1 2340.9 1506.8 2111.9 1719.5 2550.0
6-18 sandy silt nim
18-30 sandy clay
3 0-3 sandy clay i
3-6 silty sand 420.1420sl1 1683.9 2238.6 1128.1 2013.4 1611.0 2370.0
6-18 sandy silt inin
18-30 sandy clay
1 0-8 clayey silt I1
8-10 silty clay 10”1518’- 11 15.4 1391 7 1100.0 1411.5 1245.0 1633.0
10-13 sand
1 0-12 clayey silt I1
12-17 silty clay 10“4 518” 1017.6 1091.8 1200.0 1209.6 11 19.0 1170.0
17-18 clayep sand
5 0-1 silty sand I
1-4 sandy clay 12”6 9/32’’ 705.2 827.4 602.3 756.9 790.0 1300.0
1-27 silty clay
6 0-10 sandy clay I
10-12 sand 10’’4 518” 583.0 681.4 459.5 556.0 918.0 1305.0
12-18 sandy silt
(1) blow counts for the last 50 cm penetration.
(2) the pile penetration for 10 successive blows.

66 1
4 1 A one poyzrlatron atialysrs
Values from Table 2 suggest that a strong
correlation can be found between the driving
resistance evaluated by the Danish formula and by
the wave equation Such correlation should pass
through the origin
A fbnction of the type y = B x (SRDIVE= B
SRDD,tnlsh) was searched, with the coefficient B and
the correlation coefficient given by Bussab (1 988)
SRDWEis the soil resistance during driving obtained
by the wave equation analysis and S R D D a n l s h is that
evaluated by the Danish formula
Two cases were analysed one related to the data
obtained at the end of continuous driving, “the last
50 cm penetration record’, labelled as (1) in Tables
1 and 2, and the other obtained after complete
Figure 1. Correlation between soil resistance during
driving, “1 0 successive blows penetration record”, driving (in kN) from Danish (1) formula and Wave
labelled as (2) in the same Tables Equation Dinexp (I).
For each case, two fbnctions were established
one from the results of Weap program, SRD\ijq, and
the other including the results from Dinexp program, 4.2 A two yoyu Iatiori analysis
S R h n s , , , (Table 3) The data from Tables 1 and 2 were now separated
into two groups: the first group including the piles
Table 3. Soil Resistance during Driving. Statistical embedded in predominantly clayey soil and the
Results. One Population. second group those piles embedded in sand. In view
X Variable Y Variable B Correlation of the very few cases included in the second group,
Coefficient Coefficient only the first group was analysed. The results are
Danish (1) Weap (1) 0.90 0.9946 shown in Table 4.
Danish (1) Dinexp ( 1 ) 1.03 0.9897
Danish (2) Weap (2) 0.92 0,9964 Table 4. Soil Resistance during Driving. Statistical
Danish (2) Dinexp (2) 1.10 0.9829 Results. Piles predominantly in Clays.
Weap (1) Dinexp (1) I.I 5 0.9958 ~~

X Variable Y Variable B Correlation

Weap (2) Dinexp (2) 1.21 0.9917 Coefficient Coefficient
Danish (1) Weap (1) 0 89 0.9975
A good correlation was found, with a correlation Danish (1) Dinexp (1) 1.OO 0 9990
coefficient around 0.99 for all correlations, Danish (2) Weap (2) 0.72 0 9840
considering both the driving records for the last 50 Danish (2) Dinexp (2) 1.07 0 9988
cm penetration (1) and also for the additional 10 Weap (1) Dinexp (1) 1.13 0 9987
successive blows (2). Weap (2) Dinexp (2) 1.18 0.9982
Similar results were found in the comparison
i s ~ ~SRDW (Danish formula and
between S R D D a ~ ~and The same good correlation was found, with a
wave equation analysis) and in the comparison of correlation coefficient even higher now.
the wave equation performed by two distinct
programs (Weap and Dinexp).
Figure 1 illustrates one of the linear fknctions 5 TIME EFFECT
passing through the origin. It can be seen that the
individual points are really very close to the curve. As soon as the penetration for the additional 10
The surprisingly good correlations between the successive blows occur after the pile is driven to its
Danish formula and the wave equation assures the final depth, that is, after the end of continuous
confidence of the Danish formula in evaluating the driving, the comparison of the estimates of driving
soil resistance during driving for the steel piles of results from both records can express a time effect.
the data base. It should be emphasised, however, that the time lag
separating both records is very short in practice,

662
lasting just some minutes. Redriving the pile days
later would probably show much stronger effect.
The analyses are only related to piles in clay.
Piles in sand included very few cases for which
time effect is not expected to occur. The results are
indicated in Table 5.
Table 5 . Soil Resistance during Driving. Time effect
from end of continuous driving and additional 10
sucessive blows.
X Variable Y Variable B Correlation
Coefficient Coefficient
Danish (1) Danish (2) 1.20 0.9925
Weap (1) Weap (2) 1.03 0.9897
Dinexp (1) Dinexp (2) 0.92 0.9964
According to what was expected, time effect was not
significant within the time lag separating both
records.
6 CONCLUSIONS
The paper analyses the driving data from steel piles
using both the Danish driving formula and the wave
equation.
Good correlations were found between soil
resistance during driving determined by the Danish
formula and two different wave equation programs.
High correlation coefficients were obtained when
the data base was grouped as one population and
also when the data was separated according to soil
nature mostly present along pile shaft.
Similar results were found when comparing both
the Danish formula and the wave equation solution
and two distinct wave equation programs. Such
results assure the confidence of the Danish driving
formula in evaluating the soil resistance during
driving for steel piles with characteristics similar to
those of the data base.
The conclusion above is particularly usehl when
the pile penetration is deeper than the SPT boring,
which represents a very common situation in the
engineering practice in Brazil. For those cases, the
driving formula may be the only tool available to
verify piling performance.
ACKNOWLEDGEhENTS
The authors are gratehl to Dr. Alvaro Maia da
Costa, from PETROBRAS (the Brazilian State Oil
Company), for allowing the analyses with Dinexp
Program.
663
REFERENCES
Aas, P M 1993 Personal communication, from
Norwegian Geotechnical Institute Experience
Bussab, W 1988 Analise de F'nridncia e de
RegressLio. SLio Paid0 Atual Editora LTDA
Chellis, R D 1961 Pile Foundatioris. McGraw-Hill
Book Company New York Second Edition
Costa, A M , L F R Moreira, N F F Ebecken,
A L G A Coutinho, L Landau & J L D Alves
1988 Recent Application of Computer Methods
for Drivability Analysis of Offshore Piles in
Brazil Proceedings of the Iriteriiatronal
Conference on Computer Modellitig r n Oceari
Str1~tici.e.665-672 Veneza, Italy
Danziger, B R 1991 Dynaniic arm'ysis of pile
drising D Sc Thesis COPPE, Federal
University of Rio de Janeiro Brazil (in
Portuguese)
DGI 1978 Code of Practice for Fouridation
E~gi~eeririgBulletin I T' 32 The Danish
Geotecnical Institute
Goble, G G , F Rausche & G E Linkins 1980 The
analysis of pile driving - A state of the art
report Proceedings of the 2nd Iriteniatiorial
Corference on the Applicatrori of Stress Wave
Theory or?Piles, pp 131-161 Stockholm
GRLWEAP 1988 Mamal for Wave Epntiori
Aiialysrs of Pile Driving Goble Rausche Likins
and Associates, Inc
Semple, R M & P Gemeinhardt 1981 Stress
History Approach to Analysis of Soil Resistance
to Pile Driving Proceedrrigs of the 131h Offshore
Technology Coiiference. Vol 1, 465-48 1
Houston, Texas, USA
Smith, E A L 1960 Pile Driving Analysis by the
Wave equation. ,Joimal of the Sorl Mecharircs
arid Foinidatroris Divisioti ASCE Vol 127, part
I, 1145-1193
Stevens, R , A Wiltsie & T H Turton 1982
Evaluating pile drivability for hard clay, very
dense sand and rock Proceedirigs of the I 4 rh
Ofshore Techriology Coifererice Paper No
4205, Vol 1, pp 465-481 Houston, Texas,
USA
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Evaluation of pile set-up from penetration per blow

Gary Axelsson
Skunsku Teknik AB, Royal Institute of Technology,Stockholm, Sweden
Staffan Hintze
Skunsku Teknik AB, Stockholm, Sweden

ABSTRACT: Case histories, published in recent years, show that the long-term increase in bearing capacity
of driven friction piles in non-cohesive soil is often large. The present paper describes two cost-effective
methods, in which the penetration (set) per blow is used to evaluate the relative set-up. The reliability of the
methods was assessed in a comprehensive field-test involving both concrete piles and steel rods, driven in
glacial sand. Dynamic loading tests were performed at different points in time, between 0-216 days after end
of driving. Both an increase in bearing capacity and a decrease in set per blow were observed over time. The
set-up was roughly 40% per log cycle of time for both the piles and the rods. Both methods showed that it was
possible to determine the relative increase in bearing capacity with very good accuracy from the set per blow.

1 INTRODUCTION set-up. In particular, it is the relative increase in

The long-term increase in bearing capacity (i.e. set-
up) of driven friction piles in non-cohesive soil is
often large. Published case histories, summarised by
Chow et al. (1998), suggest that the capacity is, on
average, doubled over a period of 1- 100 days. A few
studies (Chow et al., 1998 and Axelsson, 2000) also
indicate that the set-up process continues for several
months and possibly even years. Furthermore, the
set-up is approximately linear with the logarithm of
time. However, capacities older than one week are
very seldom used in practice. The reasons for this
are often lack of available waiting time, and also that
it is relatively costly to perform loading tests at sev-
eral points in time.
The present paper describes two cost-effective
methods to evaluate the relative set-up at a specific
site by simply measuring the penetration (set) per
blow at different points in time. The methods were
assessed in a comprehensive field-test involving
both concrete piles and steel rods, driven in glacial
sand. The methods are believed to be very useful for
an early estimation of the pile set-up at a site and for
production control purposes.
2 PROPOSED METHODS
bearing capacity that is the main item of interest, not
the absolute capacity. Hence, the aim was to see if a
general relationship exists between the decrease in
set, due to pile set-up, and the related increase in ca-
pacity.
2.2 Method I : Empirical con-elation
The authors’ have noticed from a number of piling
projects, where piles were observed to have exhib-
ited set-up, that the measured capacity (0) as a
function of set per blow (S), can be approximated by
the following type of relationship:
Q=C7.S-b
where a and b are constants.
It was firther noted that calculated results from
various dynamic formulas and WEAP analysis can
be also approximated with very good accuracy by
this equation.
The empirical method, derived from Equation (l),
is based on the following simple equation, which
relates the relative decrease in set per blow (SySJ to
a corresponding relative increase in bearing capacity
(WQ1):

2.1 Principal coricept

The main idea with this study was to see if the To establish the constant, b, in the equation the
penetration per blow could be used to estimate pile set per blow and capacity have to be determined at

665
two points in time. Further, in order to estimate the distribution of shafthoe capacity and increased
set-up at a third point in time only the set is required. damping)
Alternatively, b could be chosen from well- 4. A second WEAP analysis is performed. The
established correlations for a certain combination of penetration, 5'2, and the corresponding capacity,
pile-hammer-soil conditions. 0 2 , is then evaluated at another point in time.
Advisedly, the method should be limited to data 5 . Finally, (I? / Q I ,gives the relative increase in ca-
from piles where capacity has been hlly mobilised. pacity
A permanent penetration greater than 3-5 mm is of-
ten an indication that the ultimate capacity has been Table 1. The effect of time .on the dynamic soil parameters
reached. However, using data from pile testing with (typical trends).
small sets will in any case produce conservative pre- Shaft I Toe
dictions, or in other words low b-values. Damping Quake Damping Quake
The equation could also be applied when the Increases Constant No trend Decreases
maximum transferred energy (E,,,,) varies between over time found over time
blows. In this case the penetration per blow is nor-
malised with respect to the transferred energy by
simply dividing S by E,,,ax,i.e: 3 TEST SITE AND TESTING PROCEDURE

The test site is situated on the Fittja straits, 20 kilo-

metres south-west of Stockholm. Extensive field
testing has previously been undertaken at the site on
Results from WEAP analyses, for various pile instrumented concrete piles and steel rods, the main
types and soil models, show that for a certain capac- aim being to investigate the nature and the mecha-
ity, the ratio, S/Emax,is almost constant when the set nisms of the set-up phenomenon. The results of
is large (i.e. when the capacity is clearly mobilised). these tests are presented in Axelsson (1998a,b,c),
Axelsson (2000) and Axelsson & Westin (2000).
2.3 Method 2: back-cnlczrlntion iising WEilP
In the second method WEAP-analysis is used to 3 1 Soil conditroris
back-calculate the capacity from pile penetration Figure 1 shows a plan of the test site with the loca-
data. The soil model is either based on dynamic tion of the soil investigation, piles and rods The soil
testing and C APWAP-analysis, or on experience investigation consisted of soil sampling at six levels,
from similar piling conditions. The method makes it CPT and dynamic probing. Furthermore, the pore
possible to take into account both different trans- pressures were measured using one piezometer and
ferred energy and time-dependent changes in the dy- one oper, stand-pipe
namic soil parameters. The soil consists of more than 40 m of loose to
Several studies by Svinkin (1997) and others in- medium dense glacial sand. The results from the
dicate that the most significant long-term change in CPT, presented in Table 2, show a soil that is rela-
the soil parameters is an increase in shaft damping tively homogenous with respect to the penetration
over time. Axelsson (2000) also showed that set-up
in non-cohesive soils is closely related to the shaft,
and suggests that the set-up can, for practical pur-
poses, be assumed to take place entirely along the
shaft. In Table 1 are some typical trends summarised
regarding time-dependent changes in the dynamic
soil parameters. They are based on a review of the
relative limited number of published cases that is
available on the subject (Hunt & Baker, 1988,
Svinkin 1997, Castelli & Hussein, 1998), as well as
results from the dynamic testing in this study.
The following procedure is suggested for esti-
mating the set-up from pile penetration data:
Asoil model related to the first testing occasion
is created.
A WEAP analysis is performed, and the capac-
ity, 01, corresponding to the measured penetra-
tion, 5'1,is evaluated.
Appropriate changes in the soil model are made
to take into account the effect of set-up (e.g. re- Figure 1. Plan of (lie test site.

666
resistance. The groundwater table lies approximately
2.0 m below ground level, and the soil being rela-
tively well-graded varies between a silty sand and a
gravelly sand. The relative density (Dr) is estimated
from the CPT to lie between 35-50 % using a rela-
tionship by Jamiolkowski et al. (1985). The sand is
considered to be normally consolidated with respect
to its geological history. A mineralogical inspection
of the sand showed that it mainly consists of hard
minerals, such as quartz and feldspars.
Table 2. Evaluation of CPT results
Depth Soil type 9c
[ml [MPa)
0-2 Clay
2-1 Silty sand 2-3
1-7 Sand 3-6
7-13 Silty sand 3-5
13-19 Sand 5 -9
19-21 Silt? sand 3 -8
21-23 Sand 7-9
23- Sand
3.2 Irisfallatiori of piles and rods
Figure 1 shows the location of the piles and rods.
The rods (Hl-H21) were installed with the Swedish
standard equipment (type “Borros”) used for per-
forming dynamic penetration tests (dynamic prob-
ing). The equipment is fairly widespread outside
Scandinavia and is classified between a DPH and a
DPSH. The weight of the hammer is 63.5 kg and the
drop height is 50 cm (E,,, = 63.5.g.0.5 = 311 Nm),
with an efficiency of approximately 80%. The rod is
a 32 mm diameter solid steel rod, normally equipped
with an enlarged 45 mm diameter cone shaped toe to
minimise shaft friction. However, to resemble the
behaviour of normal pile driving, the dynamic
probing in this study was performed without an en-
larged toe. The rods were driven to a depth of 19.1
m below ground.
Three 235 mm-square concrete piles (Pile A-C)
were installed after completing the dynamic testing
of the rods. The piles were dynamic load tested for
set-up determination. In addition, two of the piles
were instrumented with earth pressure cells on the
shaft for measuring the long-term stress relaxation
The piles were installed under easy driving with a 4-
tonne hydraulic free-fall hammer (ECH) and a drop
height of 20 cm. Piles A and B were driven to a
depth of 19.1 m and, after concluding the test pro-
gram, Pile C was driven to the same depth in be-
tween the two piles. A piezometer, placed at a depth
of 7.1 m and at a distance of one meter from Pile A,
indicated that the excess pore pressure induced from
the pile driving dissipated within a few minutes after
the passage of the toe.
667
3.3 DyFianiic f e s f i q
Dynamic testing was performed on the three piles
and the 21 rods at different points in time from the
end of driving, up to a maximum of 72, 143 and 2 16
days respectively for the piles, and up to a maximum
of 69 days for some of the rods. The tests were per-
formed using a Pile Driving Analyzer (PDA), to-
gether with two strain gauges and two accelerome-
ters attached close to the top of the pile (or rod).
To determine the static capacity and its approxi-
mate distribution along the shaft and on the toe,
CAPWAP analyses were performed on the measured
signals for seven of the rods and all the piles. In ad-
dition, the static capacity was also calculated for all
the tests using the Case Method (RMAX) with
damping factors of Jc=0.7 for the piles and Jc= 0.3
for the rods. These values were chosen as they pro-
duced capacities that best agreed with the CAPWAP
results
The drop height was held constant at 0.5 m for all
the dynamic testing on the rods, while the drop
height for the piles was gradually increased over
time, from 0.2 m to 0.8 m, in order to mobilise the
ultimate capacity. The penetration aRer each blow
was measured by simply marking the pilehod be-
tween blows with a sharp object. The accuracy was
estimated to be approximately a half-millimetre.
4 RESULTS AND ANALYSES
4.1 Bearing capacity versiis time
Figure 2 shows the relative increase in bearing ca-
pacity (Q.’Ql) versus the logarithm of time, where
(21 is the capacity one day after end of driving
(EOD). As can be seen, the capacity increased
roughly linear with the logarithm of time. In par-
ticular, the increase in bearing capacity was ob-
served to be on average approximately 40% per log
cycle of time for both the piles and the rods. Moreo-
Figure 2. Increase in bearing capacity of piles and rods.
ver, the CAPWAP analyses indicated that the toe
capacity was roughly constant over time for both the
piles and rods, 335 kN (6.1 MPa) and 4 kN (5.3
MPa) respectively, signifying that the set-up took
place along the shaft. Although the scatter was fairly
large, the CAPWAP analyses also showed that there
was a tendency for the pile toe damping and the rod
shaft damping to increase over time. Further, there
was also a tendency for the toe quake to decrease
over time. More detailed information of these tests is
provided in Axelsson (1998a,b).

4.2 Appraisal of the empirical method

Figure 3 shows the bearing capacity versus set per
blow for the rods. Furthermore, Equation (1) is best-
fit to the data using regression analysis (least-square
method). As can be seen good correlation (R2 =
0.88) is obtained for a wide range of set values (be-
tween 1 - 16 mm). However, it is questionable if the Figure 4. Bearing capacity as a fhction of normalised set for
three highest capacities were hlly mobilised. If they the piles.
are omitted, b will increase from 0.64 to 0.71.
Depicted in Figure 4 is the bearing capacity (Case Table 3. Results from the dynamic testing of the piles.
and CAPWAP) versus the normalised set per blow Time Capacity (kN) Set per E,,
for the piles. The evaluated capacity, the transferred after blow
energy and the set per blow are also presented in EOD I Case I CAPWAPl (mm) I (Wm)
Table 3. In Figure 4, Equation (1) is best-fit to the EOD 481 560 -7 8.7
data using regression analysis (least-square method). 1 hour 630 703 6 10.5
As can be seen, very good correlation (R2 = 0.93 and Pile I day 740 890 7 15.6
0.95) was obtained for the relatively large range of A 6days 1285 1247 4 17.4
37days 1505 1354 4 23.1
capacities and energies. Further, the CAPWAP re- 143days 1533 1441 3.5 19.9
gression curve is not as steep as the Case regression EOD 460 529 7 9.0
curve, providing an indication that an increase in soil 40min. 585 678 5 7.3
damping took place over time. Pile 1 day 830 1006 5 16.1
It is interesting to note that the constant, b, which B 6 days 1445 1402 4 16.8
determines the curvature of the relationship, is ap- 37 days 1644 1677 -- 18.9
143days 1668 1710 3 20.3
proximately the same for the piles and rods, 0.68 216days 1728 1774 4 20.8
(CAPWAP) and 0.71 respectively. Pile 1 dav 81 1 744 6.5 16.7
C 72 dais 1436 1441 6 23.8

4.3 Appraisal of the WEAP based method

Figure 3. Bearing capacity as a function of set for the rods.
As with the empirical method, the WEAP based
method was also compared with the results from the
dynamic testing on the piles and the rods.
Table 4 shows the dynamic soil parameters that were
used in the WEAP analyses for the piles and the
rods. These are the generally-used values (“average
values”) and recommended according to the WEAP
manual by Goble et al. (1998). In the WEAP analy-
ses the toe capacity was assumed to be constant, as
indicated by the CAPWAP analyses, while the shaft
capacity was assigned as a variable (and increased
over time). An analysis was also performed in order
to examine the effect of increasing shaft damping
(J.) over time. In this case the damping was gradu-
ally increased, from 0.10 s/m for the capacities at

668
EOD, up to 0.30 s/m for the capacities after a two-
week set-up period.
Normalised set, S / E,,
Capacity h = 0.2 I h = 0.3 I h = 0.4 I h = 0.5 I h = 0.6 I h = 0.7
Table 4. Soil model used in the WEAP analyses
Soil parameters
Smith damping,
Shaft
0.16 s/m, and
Toe
0.5 dm
500kN
600 kN 1.08
91:
1.10
2:1
1.07
m

1.04
' m
I 1.19 1.17 I 1.15
1.01
m

0.99
0.10-0.30 slm 700 kN 0.86 0.94 0.94 0.92 0.90 0.88
800kN 0.64 0.78 0.82 0.82 0.81 0.79
d' I 120, 900 kN (0.45) 0.64 0.71 0.72 0.72 0.72
Quake 2.5 mm (However, 1000 kN (0.30) 0.52 0.61 0.64 0.65 0.65
> l m ) i i o o k ~ (0.17j (0.39) 0.5 0.55 0.57 0.58
Constant 1300 kN - (0.19) (0.32) 0.40 0.44 0.46
Capacity Variable (Rod: 4 kN, 1500kN - (0.05) (0.17) (0.26) 0.31 0.36
1700kN - - (0.05) (0.14) 0.21 0.25

* For square piles: d = 2b I dz

The results from the WEAP analysis are pre-

sented in Figure 5 , together with the measured data
for the rods. As can be seen, very good correlation is
obtained between measured and calculated capacity
for the whole range of data. However, the best cor-
relation was achieved in the case with increasing J,
over time, as can also be seen by comparison with
the regression results in Figure 3.
The results from the WEAP analyses for the piles
are presented in Table 5 as S/El,,,, for the different
drop heights. In the analyses, EInmwas varied be-
tween approximately 6.5 - 23 kNm for drop heights
between 0.2-0.7 m. The values where the set per
blow was less than approximately four millimetres -
for which it could be assumed that the ultimate ca-
pacity had not been completely mobilised - are dis-
played in parentheses. Presented in bold are the val-
ues that should be compared with the measured
values, as the transferred energies are similar. The
results show that the normalised set is approximately
constant for a fixed capacity, under the condition
that the capacity is hlly mobilised. Moreover, the
results also clearly support the normalisation, S/El,,,
made in Equation ( 3 ) .
The results from the WEAP analyses are pre-
sented in Figure 6, together with the measured val-
ues. Further, a regression analysis was performed on
the WEAP results using Equation (l), i.e. the em-
pirical method. As can be seen, the regression curve
gives a very good fit (X2= 0.96) to the WEAP data.
Although the WEAP analyses could not match the
measured data very well, it can be seen that the rela-
tive change in capacity as a hnction of set, i.e. the
curvature, is fittingly matched, which is the main
interest here. This is also revealed by the constant, b,
determined by regression analysis, being almost of
the same value, 0.68 and 0.72 for the measured data
and the WEAP data respectively.

669
4.4 Comparison between the methods
In the previous sections the empirical method and
the WEAP-based method were compared against
measured field data. Since both methods showed
very good correlation with measured data, it could
be instructive to compare the methods with one an-
other. In Figure 5 and 6 it is shown that Equation (1)
can be best-fit to the WEN analyses with very good
Correlation.
A sensitivity analysis was performed to see if this
good correlation could be also achieved for a wide
spectrum of soil models. In this case 14 WEAP
analyses were performed for the same pile and
hammer as above. The drop height was chosen to 0.5
m, the pile length was 20 m, and the shaft distribu-
tion chosen as constant with depth.
The results from the sensitivity study are pre-
sented in Table 6 in terms of the correlation factor,
RZ,and the constants, a and b, that produced the best
curve-fit using Equation (1). The range for the set
per blow that was used in the regression analyses
was between roughly 5 and 25 millimetres.
The most interesting result fi-om these analyses
was that the variation in the constant b was relatively
small, ranging fi-om 0.73 to 0.87, where low values
were linked to small damping factors, and corre-
spondingly, high values were linked to high damp-
ing factors. On the other hand, the toe quake was
shown to have very little influence on b. Some of the
results are also illustrated in Figure 7.
The influence of pile length was also examined. It
was found that by shortening the pile length, h will
increase. For example, for a pile length of 5 my the
ST2 model (see Table 6) produced a b of 0.84 in-
stead 0.79, for the same range of set. Furthermore,
the correlation was also improved, resulting in RZ =
0.99.
A similar sensitivity study was also performed for
the rods. For a wide variety of soil models, the re-
sults produced values of h within approximately
0.80 to 0.85, for a range of set per blow of roughly 2
- 20 mm. It was also noted that a decrease in the
shaft quake to 1 m produced slightly higher values
of b. The correlation factor was as good as 0.99 for
all the regression curves.

4.5 Factors injluencing the constant b

Figure 7. Some results from the sensitivity study on the piles.
As was revealed when applying the empirical equa-
tion to both measured data and WEAP results, the
constant b showed a tendency to vary within a rela-
tively narrow range of values, approximately 0.6-
0.9. Before this can be applied in practice it has to be
examined under a variety of different conditions.

Table 6. Results from the sensitivity study using WEAP.

Shaft Quake (mm) Damping (mls) a b R2 Set Ident.
capacity
("/I Shaft Toe Shaft Toe
0 - dl120 0.25 5098 0.77 0.97 5.2-25 T1
0 - d/120 0.5 5190 0.86 0.98 5.5-25 T2
0 - 2dl120 0.25 4536 0.74 0.98 5.3-24 T3
0 - 2d/120 0.5 4673 0.85 0.98 5.6-24 T4
50 2.5 d/120 0.16 0.25 5551 0.77 0.97 5.7-26 ST 1
50 2.5 d120 0.16 0.5 5159 0.79 0.97 5.6-23 ST2
50 2.5 dl120 0.65 0.25 5080 0.82 0.97 5.6-23 ST3
50 2.5 d120 0.65 0.5 4714 0.83 0.97 5.4-24 ST4
50 2.5 2dl120 0.16 0.25 5034 0.75 0.97 5.3-26 ST5
50 2.5 2dl120 0.16 0.5 5176 0.81 0.98 5.3-26 ST6
50 2.5 2dl120 0.65 0.25 4740 0.81 0.98 5.8-23 ST7
50 2.5 2d/120 0.65 0.5 4694 0.84 0.98 5.6-24 ST8
100 2.5 0.16 5538 0.73 0.96 5.7-23 s1
100 2.5 0.65 5207 0.87 0.97 5.6-24 s2

670
 Nevertheless, it would be of great benefit to be able
to obtain a rough estimate of the relative set-up at a
site by just striking the piles at different points in
time. According to Equation (2), a halving of the set,
for instance, would indicate a relative increase in ca-
pacity of 63 % for b = 0.7. If b was instead 0.6, the
increase in capacity would be 52%, the difference
only being 11 % in the estimated set-up. This ex-
ample demonstrates that the variability in the rela-
tive set-up as a function of set is small, if b lies
within the mentioned range.
The authors would, however, like to emphasise
that the two methods presented in this paper are by
no means intended to replace pile testing for set-up
prediction. Their main purpose is to provide an esti-
mate of the set-up in, for instance, an early testing
stage, or during production control.
If we consider the following dynamic formula
(the Energy Approach) derived from the energy
equation by Paikowsky & Chernauskas (1992), and
and D,,,, are measured quantities:
where E,:,,,a,
It can be seen that if D,,,,, N" S, i.e. the total quake
(pile and soil) is very small in relation to the set, the
following expression is derived:
s (5)
In fact the expression is a conventional dynamic
formula (initially presented by Sanders in 1851) in
its simplest form, thus disregarding the energy
losses It also very similar to Equation (1) Never-
theless, it implies that the maximum value b can ac-
quire is 1 0 Hence, high values of b should be ex-
pected for small soil quakes and large permanent
sets, and further, for short piles with high stiffness
(impedance) This was also shown in the WEAP
sensitivity study However, a value less than 0 9, but
probably in most cases less than 0 8, is more realis-
tic, as near to perfect conditions (a ideal plastic soil
and an infinitely rigid pile) are not very likely to be
encountered in the field Changes in soil damping
over time will also have an influence on h, in as
much as an increase in soil damping over time will
produce a smaller value, than if the damping was
constant
It is also possible to reason that for a normal
hammer to pile impedance ratio and for non-extreme
soil conditions, a value of h greater than about 0 6 is
assumed reasonable Again under the circumstances
that the ultimate capacity is mobilised
At the moment, however, site specific determina-
tion of the constant h, by dynamic load testing and
set measurements, is highly recommended. at least
until further calibration of the constant b has been
performed for a variety of field cases
671
5 CONCLUSIONS
Both methods, the empirical and the WEAP based,
showed that it is possible to evaluate the relative in-
crease in bearing capacity with very good accuracy
from pile penetration data.
The particular advantage with the empirical
method is its simplicity, making it ideal for produc-
tion control purposes. Furthermore, it was noted that
the constant, b, can be expected to range between a
relatively narrow span, approximately 0.6-0.8, which
increases the value of the method even more.
The WEAP based method, however, is suggested
for use in more detailed evaluations of the relative
set-up, as it is possible to account for time-
dependent changes in the soil parameters.
The results also showed that there is very good
correlation between the empirical equation and the
WEAP analyses for a variety of possible soil mod-
els. An interesting observation was that the choice of
dynamic soil parameters had little influence on the
b-value.
ACKNOWLEDGEMENTS
The research was funded by the Development Fund
of the Swedish Construction Industry, the Swedish
Council for Building Research, Skanska AB. Fur-
thermore, important contributions were made by the
Commission on Pile Research, Stabilator AB, the
Swedish National Railway Administration and the
Swedish National Highway Administration. This is
acknowledged with gratitude.
REFERENCES
A\elsson G (1998a). "Long-term set-up of dn\ en piles in non-
cohesn e soils e\ aluated froin dpnaiiuc tests on penetration
rods" Proc I '' hit Co~f on Site Chnracterizntioii, .Itlantn
Axelsson G (1998b). "Long-term set-up of dn\ en piles in non-
cohesive soils". Liceritiafe thesis, Kovnl Itistirute of Tech-
nology, Stochliolni
Ayelsson G (1998c)."Long-tenn increase in shaft capacih of
dmen piles in sand" Proc JthInr Coif on Case Histories
117 Geotech Eiigng , St Louis, l\fissouri, paper I 25
Alelsson G (2000). "Set-up of dmen piles in sand - The ef-
fect of constrained dilatancj on the sliafi belia~iourdunng
loading". Proc 1171 Coiif oii Geotechnical R- Geological
Eiig , GeoEng2000, .\feIboume, -4ustrnlin, (in press)
Alelsson G & Westin A (2000). "Torque tests on dmen rods
for predichon of pile set-up ' , Proc Iiit Colif, Geo-Dewer
,7000, Denver, Colorado, (IM press)
Castclli R J & Hussein M (1 998). "Pile foundation construc-
tion for the Buckinan Bridge. Jacksoni ille. Florida". Proc
4th Itit Cotif 0 1 1 Cave Ifirtorier 117 GeotecJi Eiig17g Si
I,OUIJ,.I fi s souri',paper 11 01
.
Chov F C Jardine R J . Naro! J F and Brucj F (1998) "Ef-
fects of time on the capacih of pipe piles in dense inaniie
sand". -J GeotecJ?Engug , ASCE 1'01 121(3),p p 251-261
Goble G.G.. Rausche F.. L k n s G. & Associates Inc. (1998),
“GRLWEAF’ manual“, P’ersion 1998-1.
Hunt S.W. & Baker C.N. (1988), ”Use of stress-wave nieas-
urements to evaluate piles in high set-up conditions” Proc.
3”’ Iiit. Coi$ App. Stress-wave Theory to Piles, Ottawa, p p
689- 705.
Jamiolkowsky M.. Ladd C.C.. Gennaine J.T. and Lancelotta R.
{ 1983, ”New developments in field and laboratory testing
of soils” Theme Lecture, 1 lthInt. Coic O M Soil Alech. B
Found. Eiigiig, San Francisco.
Paikowsky S.G. & Chernauskas L.R. (1992),” Energy Ap-
proach for capacity evaluation of driven piles”, Proc. drh
Int. Conf App. Stress-wave Theoty to Piles, The Hague,
The iVetherlai7ds,p p 595-601.
Sanders J. (185 I), “Rule for calculating the weight that can be
safely trusted upon a pile which is driven for the foundation
of a heavy structure”. J. Franklin Inst. XYII, p.304.
Svinkin. M.R. (1997). ”Soil damping in wave equation analysis
of pile capacity”, Proc. 5”’hit. Con$ .4pp. Stress-wave The-
or?; to Piles, Orlanclo, Florida, p p 128-143.

672
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Comparative analysis of dynamic and static test of foundation pile

Zhou Guoran & Wu Jiaduo

Shanghai Harbor Engineering Design and Research Institute, People’s Republic of China

ABSTRACT: Since the non-ultimate bearing capacity of pile dynamic test can’t compare with the result of
static test, this paper presents a comparison method between the P-S curves of static tests and the P-S curves
of CAPWAPC matching based on experimental data obtained fiom 8 piles. At the same time, this paper also
presents an assessment way of the quantitative analysis for the CAPWAPC result.

1 INTRODUCTION 2 COMPARATIVE ANALYSIS

Designer, contractor, supervisor and constructor pay 2.1 CAPWAPC Method

close attention to the determination of the bearing
capacity of the pile in pile foundation engineering.
Early of this century, the mechanism of the trav-
eling of the stress wave along the impacted rod had
been recognized. Smith(1960) presented the numeri-
cal solution of partial differential equations for trav-
eling of the stress wave along a one-dimensional rod:
and the solution was called wave equation method.
Beginning in 1964, Dr. G.Goble(1975), professor
of Case Western Reserve University of U.S.A, stud-
ied the closed form solution of wave equation and
presented Case and CAPWAP analytical methods
from measured strain and acceleration of the pile top.
These methods have been confirmed all over world.
It is a cheap, fast and reliable method for the deter-
mination of the bearing capacity of a pile.
In some countries, this pile dynamic test method
has already become a standard for replacing or sup-
plementing the static load test. Many cases(Stephen
S.M. Cheng, Shaheen A. Ahman, 1988) showed
there existed little discrepancy between the results of
dynamic and static loading test, but the consistency
of the two load tests have to be verified if the dy-
namic loading test is to be accepted widely. In the
past, the comparative analysis between dynamic and
static test was performed at the ultimate state. How-
ever, it isn’t always possible that both dynamic and
static test reach the ultimate state simultaneously.
This paper presents other comparative analysis
method between the P-S curves of static tests and
the P-S curves of CAPWAPC based on experimen-
tal data obtained from 8 piles, and explains this sci-
entific method.
CAPWAPC[7] method is an analytic method that
calculates the bearing capacity of pile by measured
force and acceleration with wave equation. Stress
wave is computed fiom the force or velocity meas-
urements. Improved Smith elastic-plastic model was
adopted for soil. Soil resistance was modeled as a
function of the moving pile, and in two components,
namely: one is an elastic-plastic force of static state,
the other is a viscous damping force of dynamic
state.
To start a CAPWAPC matching, a certain curve
chosen fiom measured axial force of pile shaft or
velocity with time change during impact of the ham-
mer is taken as the initial input of matching. Then the
parameters (soil resistance, Quake and damping) are
adjusted continuously until computed and measured
signals agree.
After various parameters of pile and soil are con-
firmed, CAPWAPC calculates the axial forces and
the settlement values of each cross sections of pile
and determines the axial force and the settlement of
the pile top, and the loading - displacement curve is
drawn. This method is called the simulated static
loading test.
2.2 Static Load Test
Static load test is a testing method where a static ac-
tion force is applied at pile top, and the settlement
(displacement) is measured at pile top. Generally the
action force is applied at pile top in increments. Test
stopping criterion is that the displacement of pile top
can’t stop or reach a certain limit value during a cer-
tain stage loading.

673
 For reasons such as lack of sufficient reaction
load, limit of hydraulic jack, limiting strength of re-
action beams, the ultimate bearing capacity of the
pile can't be obtained during test.
Static loading test was carried out according to
"Specifications For Port Engineering Technique"
((1987) JTJ222-87) of the Ministry of Communica-
tions of the People's Republic of China( 1987).
When settlements are same, sd~~]=&r-]and Sdr=SSr,
the absolute error of maximum loading between both
P-S curves is:
E of matching P-S curve relating to P-S curve of
static is:

2.3 Contrast between Dynamic Test & Static Test

The comparative analysis is needed between the re-
sults of dynamic and static load test. In the past, the
results of both tests could not be compared if the E reflects the accuracy and reliability of CAP-
bearing capacities were not hlly mobilized. Similarly WAPC matching. E provides an assessment way of
the past comparative method was not so comprehen- the quantitative analysis for reliability of CAPWAPC
sive although the bearing capacities were in ultimate matching.
state. Settlement is a very important data in design.
Even though both bearing capacities are at ultimate
state for dynamic and static test, and settlement can't 3 THECASES
be the same, the comparative result between dynamic
and static test is difficult to evaluate scientifically. The results of CAPWAPC matching and static load
This paper presents a new comparative method test of 8 piles had been collected in this paper. AAer
that considers the comparison of both bearing capac- submitting the results of CAPWAPC matching of 8
ity and settlement. See figure 1. piles to supervisor and owner, then static load test
was performed for 8 piles. The size of 8 piles were
PHC(Pre-stressed Reinforced Concrete Spun Pile)
Figure 1.
pipe pile, steel pipe pile and square concrete pile.
Calcula-
Testing location is at Marco and Shanghai in P.R. of
tion
China. The details are shown on table 1.
method of
the relative
Table 1. Size and soecification of testing tiles and olaces
error of
Name Type Spec. Length Toe Level Testing
two curves
(mm) (m) (m) place
A-6 PHC 0800tl10 50.50 -44.43 Macao
T-2 Steel pipe 0914t16 50.50 -44.49 Macao
S is B-3 PHC 0800t110 48.50 -37.71 Macao
the P-S B-3a SC 600x600 48.50 -38.68 Macao
curve of B-3c PHC 0800t110 43.50 -35.00 Macao
static B-3b PHC 0800tl10 41.50 -32.59 Macao
load test. B-2 PHC 0800t110 45.50 -38.26 Macao
D is the P-S curve of CAPWAPC matching. When E-6 Steel pipe 0900t19 57.87 -51.97 Shanghai
both AP and AS reach an acceptable range simulta-
neously, the comparative result will describe that
CPAWAPC matching is right. E is assumed to be the
relative error of the bearing capacity obtained fiom
CAPWAPC matching relating to the bearing capacity
obtained fiom static test. There is:

E=(IAd-Asl)/AsX 100% (1)

where Ad is the enclosure area of D curve,
As is the enclosure area of S curve.

6 is defined as the absolute error of both P-S

curves. Between the loading of stage i and stage i+l,
the area of their absolute error is following:

674
 only according to the previous comparative REFERENCES
method. The results of both dynamic and static load
test of 8 piles are as follow: Goble, G.G. , Likins, G. E.1975. Bearing Capacity of Pile
from Dynamic Measurements: Final Report for Ohio De-
partment of Transport, Department of Civil Engineering,
Table Results of d iamic and st ic test Case Western Reserve University.
CAP WAPC Redriving Static Soil state for Hussein, M.& Likens, G. 1991. Static Pile Capacity by Dy-
Name capacity set capacity static namic Methods, First Geotechnical Engineering Confer-
(mm/blow) (kNI test
(kN) ence, Cairo University.
A-6 7248.4 1 .o 7326 Non-ultimate Hussein, M., Rauche, F., 1991, Bearing Capacity of Deep
T-2 8303.2 0. I 10567 Ultimate foundation from Dynamic Measurements and Static Test ...
B-3 7269.6 0.6 7890 Non-ultimate Ten Correlation Cases, Malaysia.
B-3a 800 1.7 1.o 8453 Ultimate Ministry of Communications of the People’s Republic of
B-3c 6729.6 1 .o 6199 Non-ultimate China, 1987. Specifications For Port Engineering Tech-
B-3b 5448.6 2.0 5636 Ultimate nique” (( 1987) JTJ222-87)
B-2 6703.4 0.5 6199 Non-ultimate Pile Dynamic, Inc. 1992. Pile Driving Manual.
E-6 12784.7 0.4 12768.1 Non-ultimate Smith, E. A. L.1960. Pile Driving Analysis by the Wave
Equation: Proceedings of American Society of Civil Engi-
The P-S curve obtained &om CApWApC neering, vol. 86, No. SM4, pp35-61
matching is drawn on the p-s curve obtained from Stephen S.M. Cheng, Shaheen A. Ahman, 1988. Dynamic
static loading test (see figure2-9), then the relative Testing Versus Static Load Tests: Five Case Histories: Pro-
error of maximum loading values of both curves at ceedings of Second International Conference on Case Histo-
the same settlement is calculated, respectively. The ries in Geotechnical Engineering.
details are shown in table 3.

error between P. curves

The error be- The relative er-
tween ror between
comparison of ultimate ca- max loadings at
P-S curve pacities the same set-
(“/.I (”/I tlement(%)
A-6 16.19 Can’t compare 15.36
Can’t compare 0.62
Can’t compare 2.09
B-3a 29.16 Can’t compare 32.72
B-3c 6.89 3.32 2.70
B-3b 1.42 Can’t compare 1.05
12.47 Can’t compare 3.88
13.35 Can’t compare 4.76

4 CONCLUSIONS

4.1 This paper presents a comparative method

between dynamic and static test. This comparative
method may be carried out, no matter whether the
bearing capacities of dynamic and static tests reach
the ultimate state. This comparative method provides
a scientific assessment of the quantitative analysis for
CAPWAPC matching.

4.2 There exists a good correlation between the

results of dynamic and static test. The discrepancy
between static load test and CAPWAPC matching
has been determined.

4.3 To mobilize the soil resistance hlly, the set per

blow during dynamic test should not less than 2 mm.

676
 Applicationof Stress-Wave Theory to Piles, Niyama & Beim (eds}02000 Balkema, Rotterdam, ISBN 90 5809 1503

Dynamic load test and elastic rebound analysis for estimation of the bearing
capacity of piles in residual soil
P.J. R.de Albuquerque
Departamento de Estruturas e Fundapjes, Escola Politkcnica da U S 8 S6o Paulo, Brazil
D.de Carvalho
Departamento de Construg6esRurais, Universidade Estadual de Campinas, Brazil

ABSTRACT: In this paper we will present the final bearing capacity results from a static load test and the
control by the elastic rebound. These tests have been carried out by driving 0.18m diameter and 14.0m long
precast concrete pile, at the Unicamp Campus, in a place where the subsoil is formed of residual diabasic soil,
common in Campinas and other regions of the country. These results are compared to the ones obtained
afterwards through an instrumental static load testy carried out on the same pile. The results presented refer to
of two more dynamic load tests, which were performed close to the pile, where the static load test was
performed and also the results of the field tests performed on the site.

1 INTRODUCTION region, basic intrusive rocks occur from Serra Geral

(diabase) formation. The local subsoil is
In order to know the bearing capacity of a pile, characterized by two types of soil: silt-sandy clay (0
empirical prevision methods are commonly used, to 6m) and clay-sandy silt (6 to 18m). Water level
since load tests are often economically unfeasible. was reached only at 18.0m. The soil of the first layer
New bearing verification methods have been has collapsible behavior (Monacci 1995). The
developed to reduce costs. Among them is the average results of the field tests are shown in Table
dynamic load test, already known in the technical 1.
area, often providing point and lateral load values
very close to the real ones. There is also the
Reboundmeter System, developed by IPT/SP 3 PILE
(Machado 1995), to check bearing capacity in a
simple and inexpensive. The piles are 14.0m long, the first segment is 6.0m
Three precast concrete piles were driven (L=14.Om long, and the second is 8.0m long, connected
and $=0.18m) in a subsoil formed by residual through a metal ring bound by a weld cord. The
diabasic soil. cross section is circular, with a 0.18m nominal
The results obtained through the dynamic load test diameter. The concrete used has fck=35 MPa and the
were analyzed through Case, CAPWAPC and IPT-
Case methods. Specifically for the Reboundmeter
(Machado 1995), the Uto et al. (1985) method and Table 1 Alerage results of field tests
Chellis (1951) method were used, modified by Parainctcr 1'I La! er 2'ld L ~crJ
Velloso (1987). The load test was the slow type NVT 30 73
(SML) and counted on the reading of deformations 9c 920 8 P d 2473 9 kPa
fc 44.2 kPa 214 1 kPa
along the shaft, which comes from previously e 172 152
installed instrumentation, thus obtaining loads n 63 1% 60 0%
absorbed by the point and shaft. IF 23 8% 30 3%
Y 13 6 kN/in3 15 5 kN/in3
C 27 7 kPa 58 7kPa
2 SUBSOIL 6 70" 22"
Remark NspT=nuInber of blows of SPT. ql=polnt resistance
(CPT). f,=skin frichon (CPT). e=r old rabo. n=porosio.
The tests have were conducted in an experimental 15 =moisture content. y=specific natural n eight. c=cohesion
area at Unicamp Campus, in CarnpinadSP. In the (total tension). $=friction angle (total tension)

677
steel has fyk=1500 MPa. In instrumentation CA-50
steel bars (L=0.60m and +=12.5mm) were used, in
which strain-gages were installed, bound in
complete bridge. These bars were installed at depths
of 0.60m, 5.Om, 10.0m and 14.0m, inside an iron
sheath left inside the pile while it was being made.
Cement slurry was injected, so elements could be
connected (Albuquerque 1996).
Driving system used is a free-falling hammer, with a
ram of 1,650 kg of mass, falling from a height of
approximately 0.5m.
4. IPT-CASE SYSTEM
Developed by IPT ( S ~ OPaul0 State Institute of
Technological Research) this is a computer-based
data acquisition and analysis system to be used to
monitor dynamic pile driving. It digitizes, analyses
and records strength and speed signals produced by
the instruments (acceleration meters, specific
deformation transducers) placed in a section near
the pile head while it is being driven (Figure 1).
By analyzing these signals, several important values
are obtained in sifu, such as an estimation of pile
bearing by the Case Method, maximum strength and
speed, maximum transferred energy, maximum and
final displacement, and evaluation of pile structure
integrity.
The purpose of developing this system is to supply
the technical area with a domestic and potentially
less expensive alternative to PDA
5. STATIC LOAD TEST
The directions established by NBR 12131/91 were
followed, adopting the slow load (SML). The loads
were made in consecutive stages, with load increases
of 40kN, until the load in which the settlements
indicated that the rupture of the connection pile-soil
was reached. The unloading was made in
678
consecutive stages, with load reductions equal to
25% of the total load achieved in the test. The soil
was excavated from the head of the pile up to the
depth of 0.60m7 keeping this area as reference
section to determine its Elasticity Modulus.
6. DYNAMIC LOAD TEST AND
REBOUNDMETER SYSTEM
A basic set of instruments and other commonly used
equipment (force transducers, accelerometer etc)
were used (PDA) for dynamic load tests. These tests
allowed checking integrity and bearing capacity of
the piles (Albuquerque 1996). While the piles were
being driven, measurements of elastic rebound took
place were performed using the Reboundmeter
System developed by IPT, which was in testing
period (Machado 1995)
The Reboundmeter System is a technique to assess
soil mobilized resistance as a pile is being driven,
through the elastic rebound value observed. This
technique is an alternative to other methods such as
conventional load test and dynamic driving
monitoring.
The value of elastic rebound (K) is obtained by
measuring the displacement caused by the hammer
blow. The displacement curve along time is
measured for a section near the pile head and, using
measured values of maximum (M x F) and final (S)
displacements of the section observed, elastic
rebound (K) is calculated. This value expresses the
sum of numbers of maximum elastic compression of
the pile shaft (KO) and of soil below the pile point
(W.
The Reboundmeter developed by IPT records and
processes signals coming from a sensor. These
signals describe the displacement of an
instrumented section of the pile along time, as a
result of the driving hammer blows.
The system has a software that controls data
acquisition and makes analyses, and a measuring
device (Figure 2) comprising a displacement sensor,
a mechanical device and a PC interface card
(electronic circuit).
The sensor is provided with a small rubber wheel
When it rotates, this wheel transmits the rotating
movement to the sensor, which converts this
movement into displacement signals. The sensor is
mounted on a mechanical device, which keeps the
wheel pressed against the pile wall by means of a
spring to ensure adherence between both surfaces.
The displacement of this pile section, as a result of
hammer blows, causes a corresponding rotation on
the wheel, thus allowing the sensor to record its
displacement curve along time
 Table 5. Results obtained from the static load tcst
Pilc QiZi (kN) Q,,li(W Ql, (W
2 219 43 262

According to the results obtained from the static

load test, it was possible to calculate the values of a
and K in this particular case, for the first and second
layers. These parameters are used in the Aoki &
7. RESULTS Velloso (1975) formula to predict the ultimate loads
of the piles (Table 6).
Results obtained from the tests conducted, namely:
dynamic load test (analysis: CAPWAPC, Case and
IPT-Case) (Table 2), analysis of elastic rebound Table 6. Values for a and K calculated henceforth the results
(Reboundmeter System) (Table 4) and static load froin the static load test.
test (Table 5 and Figures 3 and 4) are shown below. Layer a K
First (0 to 6in) 3.56% 389 kPa
Second (6 to lam) 8.58% 356 kPa
Table 2. Results obtained froin the static load tests
Pile Analysis QI"(kN) QPU ( W (kN)
Qil

1 172 36 208
2 CAPWAPC 20 1 15 216
3 178 62 240
1 213"
2 Case 248"
9 7<+*
3 ,522

1 200"
2 IPT-Case 243w
3 224*
*Numbers obtaincd from thc a\ cragc of thc last bloM s apphcd
to the pile J=O 4 was selected for the analysis by Case because
it represents the average value obtained by CAPWAPC. in thc
order of 0 2 for the meakenmg of tlie edge. added 0 2
(according to the TPT (1991) that indicates it for tlie cases m
which RMX values are used)

The dynamic tests were interrupted before the

lOmm/lOblows, usual for this kind of pile, had been
reached, since the piles had pre-determined lengths
(Table 3).

Table 3. Results of the end rate of penetration, obtained from

the dnving of the three piles.
Pilc 1 Pilc 2 Pilc 3
3 3 1 d 1 0 blows 1lOitun/IO blows 1251m/10 blows

Table 4. Results for the ultimate load. obtained froin the

Reboundmeter for the piles 1 and 3
Pile Method Qii (W

1 Uto* 346
3 319 8. ANALYZING THE RESULTS
1 Chellis-Velloso 186
3 175
*used for piles with point in sand. The values for the ultimate load (Q,,), obtained from
the analysis CAPWAPC (208, 216 and 240kN),
Case (213, 248 and 255kN) and IPT-Case (200, 243

679
and 224kN), from piles 1, 2 and 3, respectively,
were analyzed individually for each pile. The results
obtained from the three analysis for pile 1 were very
close. For piles 2 and 3 the results were a bit farther
than for pile 1.
It was verified that the values of QI, obtained by the
Rebound Meter are very different for the two piles
analyzed (1 and 3) among the analyzed formulas
(Table 3).
Comparing these results with the ones obtained from
the static load test, it is possible to observe,
therefore, that the Uto et al. (1985) formula offered
superior values, around 330kN. Regarding the
Chellis (195 1) method, modified by Velloso (1985),
this has presented values of a lower level, around
180kN. Nevertheless, in this case the Chellis (195 1)-
Velloso (1985) formula must be used, because the
Uto et al. (1985) formula is valid for piles with point
in compact sand.
Based on the data shown in Table 2, it is possible to
observe that, related to point load, skin friction and
ultimate load, the results obtained from static and
dynamic load tests were very close. However the
same cannot be said about the methods obtained
through the rebound, because they present large
variation.
9. CONCLUSIONS
The dynamic load test through the CAPWAPC, Case
and IPT-Case analyses, has shown reliability as a
means to determine pile side and point loads. The
value of ultimate load obtained from the dynamic
load test was placed 10% below the one obtained
from the static load test.
Although the Rebound Meter System did not
indicate results comparable to the ones from analysis
CAPWAPC, Case and IPT-Case, shows great
potential for utilization, mostly due to its simplicity
and low cost. However, an improvement in its
calculation formulas is required. It is important to
emphasize that the Uto et al. (1985) formula is only
valid for piles in sand.
REFERENCES
AssociagBo Brasileira de Normas Tecnicas 1991
MB - 3472191 Esfacas - Provas de Carga
Estbtica. Rio de Janeiro, AssociagBo Brasileira
de Normas Tecnicas, 4p. Rio de Janeiro:
A.B.N.T.
680
Albuquerque, P J R 1996 A~idhse do
Contportanwito de Estaca Pre Moldadu de
Peqrreiio Dicintetro, Insfruvneizfnda em Solo
Residual de Diabasio da Regifio de Canipmas
DissertagBo de Mestrado 154p Campinas
Feagri-Unicamp
Aoki, N & Alonso, U R 1989 Correlation
Between Different Evaluation Procedures of
Static and Dynamic Load Tests and Rebound
In XI1 I C S M F E, Rio de Janeiro
Proceedrngs...,v 2, p 1 1 15- 1 1 16
Aoki, N & Velloso, D A 1975 Urn Metodo
Aproximado para a Estimativa da Capacidade
de Carga de Estacas In Cotigresso
Parianterrcaiio de Mec&ica dos Solos e
Eqqenharm de Fimda@s, V, Buenos Aires,
Arms. .., p 367-376
Aoki, N 1986 Co’ontroleIrA’itu da Capacidode de
C’arga de Estacas Pre-Fabricadas L’la Repiqire
Elbstico da Crava@o SZo Paulo ABMS,
ABEF e IE-SP
Aoki, N 1989 A New Dynamic Load Test Concept
In XI1 I C S M F E, Discussion Session 14,
Technical Committee on Pile Driving,
Drivability of Piles, 1989, Rio de Janeiro
Proceedings..., v i, p 1-4
Aoki, N 1991 Carga Admissive1 de Estacas Atraves
de Ensaios Dinimicos In Seminario de
Engenharia de Fundagaes Especiais I1 - SEFE
11, SBo Paulo, Anais ..., v 2, p 269-292, SP
Aoki, N , Alonso, U R & Trindade, 0 A 1990
Aplicaqiio de Registrador de Deslocamento
Diniimico (RDD) na Avaliagiio da Carga
Mobilizada em Estacas Cravadas In Simposio
sobre InstrumentagBo Geotecnica de Campo
SINGE0 90, Rio de Janeiro Anais..., p 45-51
Chellis, R D 1951 Pile FourzdatronA. Theory-
Desrgti-Pratrce New York McGraw-Hill Book
Company lnc
IPT - Instituto de Pesquisa Tecnologica do Estado de
Siio Paulo SJA 1994 Ensnios de Carregamento
Dinbniico iia Faculdade de ErzgeiTharra
Agricola da Uriicantp, em Estacas Pre
Moldadas de Concreto Relatorio no 32 075
Siio Paulo IPTDCC - AMSFEICC
Machado, J R A 1995 A Avaha@o da Capacidode
de Carga de Estacas, com Base 110 Repique
Elastico Medrdo no Final da Crnvap?o
DissertagBo de Mestrado 265p SBo Paulo
EPUSP
Monacci, M D 1995 Estudo da Colapsibilrdade de
unz Solo do ramp0 Expertnwztal da Facirldade
de Eigenharia Agricola - Urircamp.
 Dissertaqiio de Mestrado. 114p. Campinas:
Feagri-Unicamp.
Uto, K., Fuyuki, M., Sakurai, M. 1985. An Equation
for the Dynamic Bearing Capacity of a Pile
Based on Wave Theory. In: International
Symposium on Penetrability and Drivibility of
Piles. San Francisco-USA. Proceedirigs..., v. 2,
p. 95-100.
Velloso, P. P. 1987. Fiinda@es - Aspectos
Geotecnicos. 5a ed., NA 01/82. Rio de Janeiro:
DECPUC -RJ.

681
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 1503

Strain dynamic testing on pressure-grouted piles

Liu Xi-An & Zhang Yao-Nian

Fujian Academy of Building Research, Fuzhou, People's Republic of China

ABSTRACT Pressure grouting can greatly increase bearing capacity of bored pile and improve shaft
integrity. The grouting effectiveness is evident by performing high strain and low strain dynamic tests. Results
of some actual cases are presented in the paper.

1 INSTRUCTION In the first stage, both Ta and Tb Mere l m in

The bearing capacity can be increased by
injection of, cement mix with bored pile a
water/cement ratio of 1/1 to 0.5/1 at the bearing end.
The grouting pressure is normally 2.5MPa to SMpa,
and in some cases can be as high as 10 MPa. In
order to control the solidify time and grouting area,
some admixture is added to the ceineiit mix. Both
the end bearing capacity and shaft resistance can be
increased significantly after grouting treatment. The
total pile bearing capacity can be increased by more
than 30 percent in most cases, and the largest
increment observed was 200 percent.
The grouting can be used to treat the pile faults
such as pile end slime and shaft concrete segregation.
In such cases, high-pressure water is used to clean
the faulty parts first. Then the grouting of cement
mix is performed. The treatment can lead to
satisfactory result for moderate extent of faults. The
load transfer behavior of grouted pile can be
determined by high strain dynamic test and by
further analysis of CAPWAP. The result of grouting
on pile faults can be understood using low strain
dynamic test, which identifies the pile faults and
their change due to grouting.
diameter and 58m in length. The bearing stratum of
the piles was medium weathered granite. The results
of static loading tests led to the conclusion that the
bearing capacity of Ta and Tb were 8000kN and
9000kN with corresponding displacement of
12.73mm and 14.71mm, respectively.
In the second stage, five piles of 800mm in
diameter and of length between 35.0m and 36.lm,
which were treated by pressure grouting. The
bearing stratum of the piles was gravel miued with
clayey soils. All of five piles were loaded 10 90001<N,
and the displacement at pile top were measured to be
between 13.29mm and 15.01mm.
Piles T2 and T4 in the second stage mere elected
to perform high strain dynamic tests furtherly. 'rhe
results of dynamic tests and static loading tests were
compared as follows.

2 HIGH STRAIN DYNAMIC TEST

A case of pile foundation project in Fuzhou,

China is introduced. The building is 30 storeys with
two basements. The height of the building is 99.8m
above the ground level. The total building area is
47000in2. Large diameter bored piles were used for
the foundation. The static loading tests of the piles
were carried out in two stages. Fig.1 .The static test load-settlement curves.
 Fig.2 The distribution of axial forces measured by Fig.3 The distribution of axial forces measured by
static tests for T2. The total load of pile top was static tests for T4. The total load of pile top was
9000kN, and the total toe resistance was 2310kN, 9000kN, and the total toe resistance was 2080kN,
and the total shaft resistance was 6690kN. and the total shaft resistance was 6920kN.

CAPWAP FINAL RESULTS

Total CAPWAP Capacity: 12175.5; along Shaft 2056.2; at Toe 3119.4 IcN
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Soil Depth Depth Ru Force Sum Unit Resist. Smith <Xiake

i n Pile of Ru w. Respect to D a m p i r i y
"%E E$: m
EES:m kN
at R u
ICN kN
D e p t h
kN/m
Area
kN/m2
Factor
s/m mm

12175.5
1 2 . 0 2.0 37.4 12138.1 37.4 18.65 6.98 .547 3 . 0 0 0
2 4.0 4.0 150.2 11987.9 187.6 74.88 28.01 -547 3 . 0 0 0
3 6.0 6 . 0 348.0 11639.9 535.6 173.51 64.91 -547 -3.000
4 8.0 8.0 397-8 11242.1 933.4 198.29 74.18 -547 3.000
5 10.0 10.0 417.6 10824.5 1351.0 208.21 77.89 -547 3 . 0 0 0
6 12.0 12.0 412.0 10412.5 1763.1 205.41 76.85 .547 2 . 8 6 6
7 14.0 14.0 321.8 10020.6 2084.9 160.45 60.03 .S47 2.550
-
R 1G 0 16.0
~- ~ 454.4 9636.3 2539.3 226.51 84.74 -547 :-!.355
9 18.1 18.1 578.8 9057 4 3118.1 288.56 107.95 .547 149
-
~

10 20.1 20.1 633 - 7 8423 8 3751 8 ~ 315.91 118 - 1 9 .547 1 970

11 22.1 22.1 647.3 7776 - 4 4399 1 322.71 120.73 .547 1 818
-
~

12 24.1 24.1 679.1 7097.3 5078 - 2 338.56 126 66 .547 1 684

13 26.1 26.1 705 - 3 6392.0 5783 - 6 351.64 131.55 .547 1 561
14 28.1 28.1 781 - 8 5610.2 6565 - 3 389.73 145 - 8 0 .547 L 445
15 30.1 30.1 787 - 5 4822.7 7352.8 392 .59 146.87 -547 1 206
16 32.1 32.1 835 - 2 3987 - 5 8188.1 416.40 155.78 .547 1 094
17 34.1 34.1 868 - 1 3119 - 4 9056.2 432.78 161.91 -547 1 000

~ v e r a g eSkin Values 532.7 265.58 99.36 - 547 1 87f3

Toe 3 119.4 5482.17 .721 I 200

Fig.4 The CAPWAP final results of T2.

684
 CAPWAP FINAL RESULTS
Total CAPWAP Capacity: 11838.7; along Shaft 8804.4; at Toe 3034.3 I ~ N
. ._.- -. -. -.- ._ ._.- ._.- -. _. _
_ . _
. ._._._. _. = = _ _
- -_- _
- -_= = = _ _ _ _ _ _ _ _ -===___-__________________:===_

Soil Depth Depth Ru Force Sum Unit Resist. Smith Quake

Sgmnt Below Below i n Pile of Ru w . Respect t o Damping
NO. Gages Grade at Ru Depth Area FactOL'
m m IcN kN IcN kN/m lcN/m2 s /m mrn

1x838 7
-
~

1 3.0 3.0 1 8 - 8 11819.9 18 a 9.38 3.50 -577 2 -500

2 5.0 5 - 0 75.0 11744.9 93.8 37.52 14.00 -577 2 -500
3 7.0 7 - 0 248.6 11496.3 342 - 4 124.29 46-38 -57-7 a .so0
4 9.0 9.0 400.7 11095.6 743.1 200.36 74.76 -577 2.500
5 11.0 11.0 3 3 3 9 1 0 7 6 1- 7
~ 1077.0 166.96 62.30 -577 a -500
6 13.0 13.0 3 9 5 . 4 10366.3 1472.4 197.70 73.77 -577 2.500
7 15.0 15.0 391.8 9974.5 1864.2 195.88 73.09 -577 2 -500
8 17.0 17.0 480.4 9494.1 2344 6 ~ 240.22 89.63 -577 2.500
9 19.0 19.0 626.4 8867.6 2971.1 313.22 116.87 -577 2 . 5 0 0
1 0 21.0 21.0 750.4 8117.2 3721 5 ~ 375.21 140.01 -577 2 -500
11 23.0 23.0 781-7 7335.5 4503.2 390.85 145.84 -577 2 .so0
12 25.0 25.0 813.0 6522.6 5316.2 406.48 x.51.67 -577 2.500
13 27.0 27.0 844.2 5678.3 6160 - 4 422.12 157.51 -577 2.500
14 29.0 29.0 876 1 ~ 4.802 - 2 7036.5 438.06 163.45 -577 2 .so0
15 31.0 31.0 829.9 3972.3 7866.4 414.94 154.83 -57-7 2 . 5 0 0
16 33.0 33.0 938.0 3034.3 a804 4 - 469.02 175.03. -577 2.500
Average Skin Values 550 - 3 266.80 102.66 -5.77 2 .so0
Toe 3034 - 3 5323.32 -815 1 .so0

Fig.5 The CAPWAP final results of T4

comparative results of low strain dynamic test
before and after pressure grouting are as follows.
The fitting results of CAPWAP were shown as Fig.
4 and Fig. 5, and reasonably closed to the these of
static loading tests. The shaft resistance near the pile 3.1 The dynamic test for shaft.
end was comparatively high, which characterized
the positive effect of pressure grouting. It was noted
that the toe resistance analyzed by CAPWAP were
larger than that by static loading tests. This was
simply due to the fact that in the static loading tests
the piles were not loaded to the ultimate state, so the
toe resistance capacity was not mobilized
completely.

3 LOW STRAIN DYNAMIC TEST

Pressure grouting not only can strengthen the
bearing stratum but also improve the integrity of pile.
Using PIT tester produced by PDI Corporation, the
locations of the defects, the end slime of pile or the
strength of bearing stratum can be determined
reasonably accurate. The results of the tester clearly
indicated the effectiveness of pressure grouting. The
Fig.6 The curve of low strain dynamic tests for
pile L74-4 indicated the break at the location of
5.5m from pile top (shown in first curve). The defect
was verified by boring sample test with the break at
the location of 5.6m to 6.8m, which was the mixture
of gravel and slurry mortar. The second curve
showed the improvement of pressure grouting.

685
 Fig.7 The PITWAP final results of pile L74-4
indicated the break at the location of 5.5m from pile
top (shown in first curve). The second curve showed
the improvement of pressure grouting.
Fig.8 The curve of low strain dynamic tests for
pile L80-1 indicated the severe segregation at the
location of 2.5m from pile top (shown in first curve).
The defect was verified by boring sample test at the
location of 2.5m with the concrete clipped with
clayey slurry mortar. The clayey slurry mortar was
substituted by pressure grouting shown as the
second curve.
3.2 The dynamic test curves for pile end.
The pressure grouting not onIy apply to the soil
surround the pile end but also to the soils along the
pile shaft, which combines the pile and the soil as a
whole mass. The velocity curves of grouted piles
collected by PIT tester show the improvement
around pile shaft and at the end of pile.
686
The comparative curves of piles before and after
pressure grouting can be seen as follows.
Fig.9 The curve of low strain dynamic tests for
pile L2 1-2 indicated that the pile was set in soft soil
(shown in first curve). The situation was verified by
boring sample test that the end of pile was placed in
the residual clay of 7.33m thickness. not in the
medium weathered granite required by the design.
Pressure grouting (shown as the second curve)
obviously strengthened the bearing stratum.
Fig.10 The curve of low strain dynamic tests for
pile L43-4 indicated that the bearing stratum had
low strength obviously (shown in first curve). The
situation was verified by boring sample test that the
end of pile was placed in the residual clay with
thickness of 8cm and with thickness of 32cm
strongly weathered granite to the medium weathered
granite. Pressure grouting (shown as the second
curve) obviously strengthened the bearing stratum.
4. CONCLUSIONS
The pressure grouting to pile foundation is a new
technology. It enhances the axial bearing capacity of
a pile significantly. The effectiveness of pressure
grouting can be analyzed by the results of static
loading test, the combination of high strain dynamic
test and low strain dynamic test.

REFERENCES

Xi-an Liu, 1998, “The Pile Testing in F L I Z ~ O U

Fig.11 The curve of low strain dynamic tests for
pile L49-1 indicated that the bearing stratum had
low strength obviously (shown in first curve). The
situation was verified by boring sample test that the
end of pile was placed in the residual clay with
thickless of 9.2gm instead of medium
granite according to design. Pressure grouting
(sl1own as the second curve) obviously strengthened
the bearing stratum.
International Airport”, Proceedings of the THIRD
INTERNATIONAL GEOTECHNICAL SEMINAR
ON DEEP FOUNDATIONS ON BORED AND
AUGER PILES, Ghent, Balkema, Rotterdam,
pp28 1-284.
si-an Liu, Ping-Hui Zhuallg and Do11g-bo Zhong,
1996, “Distinguish Rock-Socketed Status of pile Toe
’Irain Testing”, Proceedillgs of the FIFTH
INTERNATIONAL CONFERENCE ON THE
APPLICATION OF STRESS-WAVE -rmow TO
PILES, Orlando, Florida, pp.607-611.

Fig.12 The curve of low strain dynamic tests for

pile L49-4 indicated that the bearing stratum had
low strength obviously (shown in first curve). The
situation was verified by boring sample test that the
thickness of slime of pile end and residual clay was
39cm. Pressure grouting (shown as the second curve)
obviously strengthened the bearing stratum.

All of the above cases were chosen from a pile

foundation of Changle international airport in
Fuzhou, China. The piles are 800mm in diameter
and socketed into medium weathered bedrock 1.5m
or 0.5m into slightly weathered bedrock. The
working bearing capacity of the pile is designed to
be 4MN. The low strain dynamic tests were carved
out most defects of the piles. Core sampling was
conducted for some moderate and all the severely
defective piles. Pressure grouting was applies to
these defective piles before pile cap construction.
The total settlement of the building observed to this
date has been not more than lOmm until now.

687
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 1503

Assessment of the interface between dynamic and rapid loading tests

Madan B. Karkee
Department of Architecture and Environment System, Akita Prefectural University,Honjo, Japan
Yoshihiro Sugimura
Department of Architecture and Building Science, Tohoku University,Sendai, Japan
Takashi Horiguchi
GEOTOP Corporation, Tokyo,Japan

ABSTRACT: This paper is concerned with the investigation and identification of indicators representing the
characteristics of rapid and dynamic loading tests. Attempt is made to evaluate the range of values of such
indicators from actual loading test observations so as to depict the interface between rapid and dynamic tests.
To provide certain consistency, the data from the rapid and dynamic tests utilized in the investigation were
low pass filtered to remove components in the frequency range higher than 1OOHz. Three of the parameters
that may be obtained directly from the measurements during tests are found to be effective for representation
of the interface between the two types of test. The concept of normalized wave length is introduced and
utilized for the purpose of investigation. Comparison of the load transfer characteristics in static, rapid and
dynamic tests indicate relatively closer resemblance of rapid loading test to static behavior. In contrast the
dynamic test tends to be distinctly different. In addition, the extent of resemblance between rapid and static
loading is adequately reflected in the three simple parameters proposed. The results of this research are
expected to find useh1 practical application in the implementation of rapid loading test in practice.

1 INTRODUCTION (Middendorp et a1 1993) and pseudostatic test

The rationality of and the continuing need for
relying on loading test results while taking critical
design decisions concerning pile foundations is
evident from the recent studies devoted to
investigation of test results (e.g. Bruno and
Randolph 1999, Danziger et al. 1999 etc.). It is usual
to employ different types and categories of loading
test methods (Karkee et al. 1997) in practice.
Irrespective of the method employed, evaluation of
the static load bearing behavior of the pile is the
desired outcome in practice. When the loading itself
is not static, the measured response requires fbrther
analytical processing to estimate the static response.
Consequently, the choice of loading method
generally involves a tradeoff between higher cost of
direct measurement by static test and a measure of
ambiguity that may be unavoidable in the indirect
estimation through dynamic or rapid loading tests.
The obvious alternative to static test used to be
the dynamic test, which was originally developed as
an extension of the pile driving operation (Smith
1960). With the predominance of installation
methods other than driving (Karkee 1999), the
dynamic test is not limited to driven piles alone any
more. In addition, methods of load application
intended to lie in a domain intermediate between
static and dynamic, such as the Statnamic test
(Schellingerhout & Revoorte 1996), have also been
developed These methods are categorized as rapid
loading tests (ASTM, 1999) The availability of the
different categories of loading test provides for
prudent selection of a method suitable for a given
situation However, it also means proper distinction
between such categories based on the differences
involved Considering the close resemblance, the
distinction between dynamic and rapid loading tests
is particularly important Disregard of the stress
wave phenomenon in case of rapid loading test
while estimating the static load transfer behavior of
the pile may be justified based on such distinctions
In other words, clarification of the interface between
dynamic and rapid loading tests is crucial for usefbl
implementation of rapid loading test in practice The
rapid loading considered here is the Statnamic test
Considering the similarity in the nature of load
application, it is not surprising that some degree of
vagueness exists regarding the interfaces between
dynamic and rapid loading test methods
Identification of clearly defined indicators of how
the rapid loading condition differs from that of
dynamic loading condition may be usefully utilized
for defining such interface One such indicator
proposed is the relative duration of loading, T,
(Karkee & Kishida, 1999) The validity of T, is
further investigated in this paper based on the nature

689
of frequency contents in rapid and dynamic loading
tests on bored precast piles. The concept of the
normalized wave length is introduced and the
applicable range of the dimensionless frequency a0
is also investigated. The extent of the dynamic effect
in load transfer through shaft resistance is compared
based on the data from static, dynamic and
Statnamic tests. The results may be usefbl in
defining the basis for disregarding the stress wave
phenomenon in rapid loading tests.
2 INDICATORS OF THE INERFACE
It is evident that the interface between dynamic and
rapid loading tests has to be evaluated depending on
how the pile body and the pile soil interface respond
to the respective loading methods. Basically, the
nature of loading method employed is reflected in
the dynamic characteristics of the resulting load and
velocity time histories recorded during the test.
Prospective indicators considered for investigation
are also derived from the time histories, and include:
(a) relative duration of loading T,, (b) average rate of
loading re, (c) normalized wave length ;1’2L~,(d)
dimensionless frequency an, and (e) time lag
between load and movement peaks t,,,.Here, 3, is the
wave length and LP is the pile length. Discussions
are based on actual observations from loading tests.
Fibwe 1. Definition of the duration of loadmg and the rise time
Here, T,. and re depend on the values of the
duration of loading t d and the rise time t e
respectively, which are defined as illustrated in
Figure 1. The idea of the factor a (cl.0) for
estimation of t d was introduced by Karkee and
Kishida (1999). In practice, the value of a may be
considered to lie in a range such as 0.7 to 0.9. The
relations for T,. and re are given by:
T
)*
=Ctd
2Lp
y =-aF
te
where, c is the stress wave velocity of the Pile
material and aF is the load increment in time fe.
Other details of the different indicators are discussed
subsequently in the respective sections that follow,
where attempt is also made to evaluate their
suitability in representing the interface between
dynamic and rapid loading tests.
3 TEST SITES AND STATIC LOAD TRANSFER
Three types of loading test results, including static,
rapid (Statnamic) and dynamic, are utilized. Both the
static and dynamic tests were conducted at the site
shown in Figure 2, where the ground condition, steel
pipe pile profile and its instrumentation details are
depicted. Circled numbers indicate sections along
the pile where strain is measured. The shear wave
velocity VS profile in Figure 2 is estimated from the
correlation with standard penetration test N-values
(Japan Road Association, 1990) given by:
80N1I3 (Sand)
vs ={
1 0 0 ~ (clay)
~ 1 ~
The relation of the static load POapplied at the
pile head and the strain measurement at level 0 is
given in Figure 3. Approximately a straight line
relationship may be noted except at very high strain.
The Young’s modulus of the steel pipe material ES is
found to be about 2 . 1 5 105~ MN/m2 based on the
straight line approximate shown in Figure 3. The
value of Es is quite similar to what may be expected
for steel, indicating validity of the observations from
strain measurement.
(1) Figure 2. Static and dynamic loading test on a steel pipe pile
(2) 3.1 Load transfer relation IH static loading
Strain was measured at three levels as shown in
Figure 2, and the corresponding shaft resistance z,
and local vertical movement w, relations for the
690
segments ill-0and 0-0are given in Figure 4. The (wp).The corresponding movements from the load
z,-w, load transfer relation is approximated by transfer analysis are compared with the measured
bilinear representation shown by the solid lines in values in Figure 6. Fairly good agreement may be
Figure 4. Similarly, the load transfer hnction noted. The value of ES obtained from Figure 3 was
relating the pile toe resistance PP and pile toe utilized in the analysis.
movement w p is shown to be approximated by a
hyperbolic curve in Figure 5 . It may be noted that
the maximum pile toe movement in the static
loading test is relatively small. Consequently, the
mobilized toe resistance is not substantial and the
data points in Figure 5 seem to lie in a straight line.
Thus the hyperbolic curve approximation also serves
to extrapolate the load movement behavior at the
pile toe. In contrast, the pile shaft load transfer
seems to be h l l y mobilized in Figure 4.

3.2 Cotfifirnmtionof the load transfer.relntioris

The load transfer relations of Figures 4 and 5 appear
to be quite approximate in nature To confirm the
adequacy of such approximation, load transfer
analysis (Hirayama, 1990) was carried out to
compute the vertical movements at different load
levels at the pile head. As may be noted in Figure 2
that the vertical movements were directly measured
at the head (WO), midpoint of pile (wm) and pile toe

69 1
3.3 Rnpid (Statnnmic) loading test sites time history to compute the frequency content. To
provide a certain consistency in the comparison
The rapid loading test data utilized are from the
between different tests, components with frequency
Shonan test site in Japan (Karkee et al. 1995). A
higher than lOOHz were removed in all the data by
number of tests were carried out in 1995 on the 7m
low pass filter.
long bored PHC pile shown in Figure 7. Details of
the steel pipe pile installed by driving are shown in
Figure 8. Statnamic test was conducted on this pile 4.1 Norninlized wnve length
in 1995. Both the piles were at the same site, as may
The Fourier transforms of the load and velocity time
be noted from the ground condition. Again, the VS
history recorded near the pile head provide the
profiles in Figures 7 and 8 are obtained from
nature of the frequency components involved with
Equation 3. Altogether, the results of 3 repeat tests
rapid and dynamic tests. Since the wave length A of
on the pile in Figure 7, and one test on the pile in
these components is more reflective of the nature of
Figure 8, are utilized in this investigation.
stress wave transmission along the pile length, it is
logical to express /z in terms of c as.

where, f is the frequency. Nature of transmission

of the component of a certain wave length ;1 along
the pile length is basically determined by how large
3, is in relation to the pile length L p . Thus it would be
useful to normalize /z by L p . Since 2Lp already
appears in Equation 1 in consideration of the
importance of the time taken for the stress wave to
travel down and up the pile, it is logical to define the
normalized wave length as A 2Lp. The Fourier
spectra components can also be normalized by the
peak spectra. Figure 9 shows the plot of the
normalized velocity spectra S I - S , in~ ~ ~~~
terms of~
/z 2Lp for the dynamic and Statnamic tests described
above. It may be noted that the spectral velocity
components of dynamic and Statnamic tests tend to
form distinct groups, with normalized wave length
of about 10 to 20 forming as the interface

3 4 I)L’tmniic tesi oil piles and site conditions

One of the dynamic test data utilized in the analysis
is from the steel pipe pile in Figure 2. Three other
test data from 39m long bored PHC piles at a site in
Tokyo area, where the F<Y of the near surface soil
was about 150m/s, are also included. The later three
data are from different piles at the same site.

4 ASSESSMENT OF THE INTERFACE

Based on the information available from the three

types of loading test discussed above, attempt is
Figure 9. Nomiallzed spectra and nornialized nave length
made here to delineate the interface between rapid
and dynamic loading situation through the
comparison of the range of values of the selected It may be noted in Figure 9 that most of the
indicators. The investigation basically involves dominant components in the Statnamic test have the
Fourier transform of the measured load and velocity normalized wave length of 20 to 200. In comparison,

692
the dominant components in the dynamic test have value of the shear wave velocity VSat the upper part
normalized wave lengths in the range of 0.5 to 10, of the site where the tests were conducted, the value
indicating a fairly clear distinction. of a0 can be easily computed fr0m.f:

4 2 Relative dirratiori arid average rate of loading

Figure 9 clearly shows that a certain range of
normalized wave lengths are involved in the velocity
time histories from the Statnamic as well as the
dynamic tests. The normalized wave length A,,, of the
peak component S I ~ can,, , ~ however,
~ be determined
easily from Figure 9. Here, A,,,is the normalized
wave length corresponding to the peak spectral
velocity, and is referred as the relative wave length

Figure 11. Dimensionless frequency content in loading tests

The spectra of the load time history SP

normalized by the peak spectra Sp,,, is plotted with
respect to no in Figure 11. Again a distinct difference
in the range of no may be noted between dynamic
and Statnamic loading tests. Effectively, apO.2 for
Figure 10. Relative duration & average rate of loading Statnamic tests and a0<:3.Ofor dynamic tests seems
to be a reasonable range based on the loading test
Figure 10 shows the variation of T,- computed data under consideration, indicating a difference of
with alO.75 in Figure 1 with the values of PL,,, the order of 10 in the applicable range.
obtained from Figure 9. The data points for T,
denoted by open marks tend to fall on a straight line, 4.4 Time lag betweer?load arid movement
indicating that T,. and AL,,, tend to have a proportional
relationship. It is apparent from the tendency that the One of the effects of the using dynamic and rapid
relative duration of loading T,. serves as a suitable load application in loading tests is the development
indicator of whether a loading test falls into rapid or of the lag time between load and movement. That is,
dynamic category. This is advantageous because the movement peak occurs after a certain time after
unlike L,,,T,. can be determined quite easily from the the load peak. This is illustrated in Figure 12 for
time history record. typical dynamic and Statnamic loading cases.
The average rate of loading re is also computed
based on the value of rise time te with a=O.75 in
Figure 1, and is plotted in Figure 10. Compared to
the very high re (>800MN/s) in dynamic tests, its
value is quite small (7 to 28MN/s) in Statnamic
tests, indicating the difference of an order of
magnitude

4.3 ~imerisionlessJi.eqz~eiicy
Another important parameter indicative of the nature
of load application in rapid and dynamic loading
tests is the dimensionless frequency a0 given as:
Figure 12. Illustration of time lag between load and movement
wrQ 27r f r o
a0 =-=-
vs vs It may be noted that the movement in dynamic
tests is obtained from the acceleration by integrating
where, r0 is the outer radius of the pile and o is twice and the accuracy may not be very good. The
the circular frequency. Considering the average

693
movement is, however, directly measured in
Statnamic tests. The values of the time lag t,,.
between load and movement are given in Table 1
along with other relevant parameters. The difference
in between dynamic and Statnamic tests is not
evident. That is, t,,. is not found to be usefbl for
distinction between rapid dynamic loading
conditions based on the loading test data considered
in this study. Further study would be needed for
more conclusive outcome.
5 LOAD TRANSFER RELATIONS
Another important aspect to be considered in
investigating the different loading conditions is the
difference in load transfer fbnctions. The load
transfer fbnction under loading conditions in which
the velocity is not negligible has been investigated
by Randolph (1990) and Danziger et al. (1999). The
relation for unit shaft resistance zhas been given in
terms of the soil stiffness k,sand damping coefficient
cs as:
mu
Z = k,Tv+C,- (6)
fit
k,s = -
2'9G and c ,
2nro
= JG'P =Vsp (7)
where G is the soil shear modulus, ro is the pile
radius and p is mass density of soil. In what follows,
some investigations are made concerning the
difference that may be expected in the behavior of
load transfer by shaft resistance.
5.1 Load transfer in dynamic and static tests
As mentioned above, static as well as dynamic
loading tests were conducted on the steel pipe pile
shown in Figure 2. The unit shaft resistance obtained
from the static test is given in Figure 4, where the
bilinear load transfer relation is found to represent
well the observed behavior. It is also seen that k, is
about l.7kN/mZ/ntmat the upper part of the ground
(section E-Cl) where VSis about I30mis in Figure 2.
Considering G = ~ ( V Sat) ~low strain level? the value
of k, works out to be about 42.5kN/m-/mmfrom
Equation 7, much larger than the value obtained
694
from actual loading test This is because the soil
shear modulus G degrades at higher strain level
Conversely, for k,=I. 7kiVm' nini in Equation 7, the
effective yalue of G works out ,to be about
I IO5kN ~ i i -compared to 2 756JkN ni- at low strain
level (a difference of about 25 times)
Similarly, for the section 6-1of the test pile in
Figure 2, where L's is again about 1 3 0 d s on the
average, the effective value of the soil ,shear
modulus G ,works our to be 6500kN/ni- (for
k,=IO.OkNm- nini in Figure 4). Low strain shear
modulus in this case is about 4 times the effective
value, compared to 25 times in section 17-5 as
mentioned above. Both the damping coefficient cs
and the soil stiffness k, are derived directly from the
effective shear modulus G as given by Equation 7
Consequently, the selection of a suitable value of
shear modulus, applicable to a given loading test,
dynamic or rapid, becomes a critical decision in the
analysis of the load transfer behavior
Figure 13. Load transfer in shaft resistance in dynamic loading
Assuming that the velocity measured near the pile
head in the dynamic test on the pile in Figure 2 acts
directly to mobilize the shaft resistance at the upper
part of the pile, the unit shaft resistance may be
obtained from Equation 6. For this, the values of k,y
based on the value of G estimated from the
and c , ~
static test can be utilized. The shaft resistance thus
obtained is compared with that of the static loading
case in Figure 7. Large difference in the shaft
resistance behavior between dynamic and static
loading, in terms of the initial slope as well as in
terms of the magnitude, may be noted.
In Figure 4 the static shaft resistance tends to
increase up to a movement of more than 20mm. In
contrast, the maximum movement affected in the
dynamic test is only about 8mm. However, there
appears to have been a clear permanent movement
of the shaft at the end of the dynamic loading cycle
in Figure 13, in spite of the relatively small
maximum movement. The result has very important
implication, because the maximum shaft movement
at which a permanent movement may be expected in
a dynamic test may not be coincident with the
movement at which the shaft resistance is fully
mobilized in static test. This aspect needs to be
adequately reflected in defining the soil model in the
analysis for dynamic test.
5.2 Load trau.fer in Statnaniic tests
Following the interrelation between k, and cs noted
above in Equation 7, it may be noted that the
dynamic and static load transfer functions are
basically determined by the effective value of G
applicable to a given loading test situation. The
movement w at the pile head is directly measured in
the Statnamic tests depicted in Figures 7 and 8. The
velocity response dvldt is obtained from it by
differentiation. Now, knowing that the low strain
shear modulus close to the surface a; the test site
works out to be about 36697kN/ni-, it may be
reckoned from the static test discussed above that
the effective value of G would be a certain fraction
of this value. Once this fraction is determined, the
initial slope of the static load transfer function as
well as the Statnamic load transfer fbnction may be
defined by Equation 7. Assuming the fraction, noted
above for the static loading test depicted in Figure 2,
to be applicable here as well, the load transfer
relations for the Statnamic tests (no. 1 and 4 in Table
1) may be obtained as shown in Figures 14 and 15.
Figure 11. Shaft resistance in Statnarnic test on PHC pile
Figure 14 gives the relative comparison of the
initial slope of load transfer function in static test
with that in the Statnamic test on the pile shown in
Figure 7 (test no. 1 in Table 1). It can be noted that
the difference in magnitude as well as in the trend of
the initial slope between the Statnamic and static
tests is much smaller when compared to that
between dynamic and static test shown in Figure 13.
This is a clear indication that the load transfer in
Statnamic test tends to be much closer to the static
trend when compared to the conventional dynamic
test.
695
Similarly, Figure 15 compares the trend of load
transfer relation in Statnamic and static tests for the
test on steel pipe pile in Figure 8 (test no. 4 in Table
1). The two load transfer relations are noted to be
even more similar to one other when compared to
those in Figure 14.
Figure 15. Sl& resistance in Statnamic test on Steel pipe pile
It would be interesting to see if the indicators in
Table 1 reflect the better resemblance of Statnamic
test with the static load transfer trend in Figure 15.
Table 1 shows that test no. 1 (Figure 14) has smaller
values of A,,, (25.5) and T,. (12.96) and higher rate of
loading re (13.3MWs), compared to the
corresponding values of A,,, (38.6), T,. (26.48) and I’,
(9.1MN/) for the test no. 4 in Table 1 (Figure 15).
This provides further verification that the indicators,
proposed here for representation of the interface
between dynamic and rapid loading tests, logically
depict the trend towards resemblance or otherwise to
the static loading. That is, the three simple
parameters may be suitably utilized to see if a
loading condition is closer to conventional dynamic
test or to the rapid loading test with relatively closer
resemblance to the static loading condition.
6 CONCLUSIONS
Attempt has been made in this paper to propose
suitable indicators representing the interface
between rapid (Statnamic) and dynamic loading
tests. The range of the values of such indicators are
evaluated based on the actual loading test
observations. To provide certain consistency in the
comparison between several Statnamic and dynamic
tests, low pass filter is used to remove components
in the frequency range higher than 1OOHz. The
concept of normalized wave length is introduced and
utilized in the investigation. The load transfer
characteristics in static, rapid and dynamic tests have
been compared and investigated. Following
conclusions are drawn from the study:
i. The relative duration of loading Tr tends to have
a proportional relation with the normalized
wave length A,,, at the peak spectral velocity
component. Alrj is referred to as the relative
wave length in this paper.
11. Three of the indicators investigated, namely the
relative duration of loading T., the average rate
of loading re, and the relative wave length A,,j,
are found to be well suited to depict the
distinction between dynamic and rapid loading
tests. Certain range of values of these
parameters may be defined to serve as the
interface between dynamic and rapid loading
tests.
iii. Based on the loading test data analyzed in this
study, the applicable range of the dimensionless
frequency a0 is noted to be less than about 0.2
for rapid loading tests. In comparison, dynamic
tests seem to involve a0 of up to 3, and may be
more.
11'. The pile shaft load transfer relation in static
loading test can be represented by bilinear
approximation quite satisfactorily, as evidenced
by the comparison between measured and
computed load movement behavior at different
points along the pile length.
1). The load transfer by shaft resistance in case of
rapid loading test tends to be relatively closer to
the static behavior when compared to the case
of a dynamic test.
VI. Different extent of resemblance in load transfer
behavior between rapid and static tests can be
expected in practice, indicating the need to look
at the individual cases in detail before deciding
to adopt a certain method of analysis for
estimating the static load bearing behavior.
1'11. The extent of resemblance between rapid and
static loading tests in relation to the load
transfer knction is adequately reflected in the
values of the indicators proposed in this paper.
Closer resemblance with static loading may be
expected in case of rapid loading test with larger
values of relative duration of loading Tr, and the
relative wave length A,,,, and a smaller value of
the average rate of loading re.
ACKNOWLEDGEMENT
The Statnamic data utilized here constitute part of
the series of tests in 1994/95 organized by the
'Research Group on Rapid Load Test Methods'. of
which the first author was a member. The authors
wish to acknowledge the contribution of all the
members of the research group to the research
initiative that has helped substantially in developing
better insight about the rapid loading test method.
696
REFERENCES
ASTM 1999. Standard test method for piles wider rapid axial
cotnpressive load (Designation: D 0000-99)
Bruno. D. & M. F. Randolf 1999. Dynamic and static load
testing of model piles dnven into dense sand. J. Geotecli. 61.
Geoenv. Eizgrg., ASCE. 125(11): 988-998.
Danziger. B. R.. A. M. Costa, F. R. Lopes & M. P. Pacheco
1999. Back analysis of offshore pile dnving with an
improved soil model. Geotechnique, 49(6): 777-799.
Hirayama, H. 1990. Load-settlement analysis for bored piles
using liperbolic transfer functions; Soils & Foundations.
VO1.30. No.1: 55-64.
Japan Road Association 1990. Specifications for highway
bridges, Part IV, Earthquake resistant design (in Japanese).
Karkee, M. B. & H. Kishida 1999. Dynamic, intermediate and
static loading tests and application of test results in design
practice. Keynote paper: Proc. 3Ih Int. Col$ Deep
Foundation Practice incorporating PILET.4LK '99.
Singapore, 29-30 July 1999: 13-26.
Karkee, M. B. 1999. Developments in low noise and low
vibration methods of pile installation in Japan. Proc. 1Ith
Asian Regional Col$ Soil Mechanics & Geotechnical
Engineering, Seoul, Korea, Vol. I1
Karkee, M. B., H. Horiguclii & H. Kislida 1999. Limit state
formulation for the vertical resistance of bored PHC
nodular piles based on field load test results. Proc. I I "
.4siaii Regional CO@ Soil Mechanics & Geotechnical
Engineering, Seoul, Korea, 1999: 237-240.
Karkee. M. B., H. Horiguchi & H. Kishida 1997. Static and
dynamic tests for evaluation of vertical bearing capacity of
piles. Deep Foundation Institute, 2?ld Aiinual Metriber's
Conference, Toronto, Canada. 1997: 199-214.
Karkee. M. B., 1. Kojima & Y. Slunoda 1995. Influence of
different test conditions on the results of Statnaniic load test
at Shonan test site. Proc. I"' Int. Statnaiiiic Seminar,
P'aiicouver, Canada. 1995:137-147.
Matsunioto. T. & S. Nisliiniura 1995. Wave propagation
plienomena in Statnamic test of a steel pipe pile. Proc. 5'"
Int. Conf. Application of Stress-wave Theor?: to Piles.
Orlando. Florida. 1996: 1015-1030
Middendorp. P.. P. Bermingham and B. Kuiper 1993. Statanic
testing of foundation pile. Proc. -!Ih Int. Con$ Application of
Stress-wave Theoty to Piles, the Hague, 1993: 585-588.
Randolph. M. F. 1990. Analysis of the dynamic pile driving.
Developments in soil mechanics IV: .-lnlxmcet/
geotechnical analvsis. Elsevier Applied Science Publishers.
Poulos. H. G. 1998. Pile testing - from the designer's point of
view. Preliminary Proc. 2"" Int. Statnamic Seminar. Tokyo.
Schellingerhout, A. & E. Revoorte 1996. Pseudo static pile
load tester. Proc. 5'" Int. Conf. Application of Stress-wave
Theon- to Piles, Orlando. Florida. 1996: 103 1-1037.
Smith. E. A. L. 1960. Pile driving analysis by the wave
equation. J. SoilMecli. Found. Div., ..ISCE, 86(1): 35-61.
 Application of Stress-Wave Theory to Piles, Niyama & Beirn (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Dynamic load testing on 102 steel pipe piles for bridge foundations
on mudstone
M. Hayashi
Applied Research Center, NKK Corporation, Kawasaki, Japan
T Matsumoto
Department of Civil Engineering, Kaizazawa Universig,Japan
M. Suzuki
Izurni Research Institute, Shirnizu Corporation, Tokyo,Japan

ABSTRACT: A total of 102 steel pipe piles were driven into diatomaceous mudstone on Noto Peninsula,
Japan, for the foundations of a highway bridge. In this pile construction, dynamic load testing was carried out
on all of the 102 piles. The results of the dynamic load testing showed that the bearing capacity of all the piles
increased rapidly with elapsed time after the end of pile driving, and exceeded the design ultimate bearing
capacity one day after the end of pile driving. The sufficient number of dynamic load tests to be conducted at a
site to determine the distribution of pile capacity is also discussed in this paper. A non-parametric bootstrap
method was employed with the use of the measured bearing capacity of the piles derived 6om the dynamic
load tests. It was found that the execution of 10 or more dynamic load tests gives us the distribution of pile
capacity that has a acceptable probability of failure which may be obtained from the dynamic load tests on all
the piles for this particular site.

1 INTRODUCTION The sufficient number of dynamic load tests to be

The construction of a new highway route was
completed in 1998 on Noto Peninsula in Japan. Steel
pipe piles were used for the foundations of the
highway bridge named Noetsu bridge No.3. The
construction site consists of a thick deposit of a soft
rock called diatomaceous mudstone underlain by a
sand rock. A total of 102 steel pipe piles were driven
into the diatomaceous mudstone. Pile driving of
foundation piles for an abutment (AI) and a pier (PI)
was carried out in 1995 (Matsumoto et a1 1997).
Another abutment (A2) and three other piers (P2, P3
and P4) were constructed in 1996. For these pile
constructions, dynamic load tests were conducted on
each of the 102 piles at pile installation (initial
driving) and after a rest period from the end of initial
pile driving (re-driving) in order to estimate the static
load-displacement curve for the piles after the
completion set-up phenomena.
Rational soil models that have been proposed by
Randolph & Simons (1986) and Deeks (1992) were
used in the signal matching analysis of each dynamic
load test. The soil model parameters except the shaft
resistance and the toe resistance were estimated 6om
the soil test results. The shaft resistance and the toe
resistance alone were estimated through signal
matching analysis. The validity of this wave
matching analysis procedure was examined through
the comparison of the results of the dynamic load test
and the static load test on a test steel pipe pile.
conducted at the site to determine the distribution of
pile capacity is also discussed in this paper. A non-
parametric bootstrap method was employed with the
use of the bearing capacities of the piles estimated
from the dynamic load tests.
2 SITE CONDITIONS
Figure 1 shows the soil profile along the axis of
Noetsu Bridge No.3. The construction site is
characterized as a thick deposit of diatomaceous
mudstone. The dynamic load testing and the static
load testing of a steel pipe pile were carried out in
1991 near the construction site of the bridge (Figure
2), to obtain the design parameters for the steel pipe
pile to be driven in the diatomaceous mudstone for
the bridge (Matsumoto et a1 1995).
Figure 3 shows the soil test results at the
construction site and the test site. The variations
with depth of the natural water content, w,,,the
density, p, the unconfined compression strength, q,,
and the secant modulus, E50, are shown in Figure 3.
It can be seen that the construction site ground is
relatively uniform in plane and in depth.
Figure 4 shows the distributions of the SPT N-
values and the shear wave velocities, Y,, measured by
the seismic cone penetration tests. The shear wave
velocities which are indicated by the dashed lines were
measured at the test site, pier PI and abutment A2.

697
Figure 4. SPT N-values and shear wave velocities measured by seismic cone penetration tests at the construction site of Noetsu
Bridge No.3 and the test site.

698
3 TESTPILING
3.1 Test pile
The test open-ended steel pipe pile was driven at the
test site with a diesel hammer having a ram weight of
40.2 kN. Table 1 lists the specifications of the test
pile. The test pile was instrumented with foil strain
gages at a total of 10 levels in order to measure axial
forces during the static load test. Steel channels
were welded inside the pile for protection of the
strain gages, which increased the net cross-sectional
area, A , of the pile to 0.041 m2. The increase in the
cross-sectional area was taken into account in the
wave matching analysis by increasing the wall
thickness (by decreasing the inner radius) of the pile.
For dynamic load tests, two sets of a strain gage
and an accelerometer were mounted 180' apart at a
distance of 1 m from the pile head. The dynamic
load tests were carried out at the end of initial driving
(EOID) and 66 hours after the end of initial driving,
and the static load test was performed 26 days after
the re-driving test.

Table 1. Specifications of the test pile.

Length L (m> 11.0
Embedded length Ld (m> 8.3
Wall thickness t, (mm> 12.1
Outer diameter 43 (mm> 800.0
Inner diameter 4 (mm> 775.8
Cross-sectional area A (m') 0.04 1
Young's modulus E (MN/m') 2.06 X IO5
Bar wave velocity c (m/s> 5120

3.2 Procedure of the wave matching analysis

Wave matching analyses of the dynamic load test
signals obtained at the end of initial driving and the
re-driving were conducted using a computer program
KWAVE to estimate the static responses of the test
pile. The computer program KWAVE (Matsumoto
& Takei 1991) is based on the characteristic
solutions of the wave equation and the pile-soil
system illustrated in Figure 5.
The spring value of the internal soil (soil plug) is
estimated fiom the one-dimensional modulus, Eo, 6), the slider represents the maximum shaft
since the radial strain of the internal soil may be resistance, f , the spring represents the elastic
assumed to be negligible: deformation of the surrounding soil, k,, the dashpot
in the upper potion represents the viscous damping,
2(1- v)G
E, =- cv,and the dashpot in the lower potion represents the
1-2v
radiation damping, cr, respectively. According to
where G and v are the shear modulus and Poisson's Novak et a1 (1978), the spring value, k, and the
ratio of the soil. radiation damping, er, of the outer shaft resistance
Rational soil models that have been proposed by were estimated as follows:
Randolph & Simons (1986) and by Deeks (1992)
were used for the shaft and the base resistance in the
wave matching analysis. In the shaft model (Figure

699
C, = &G= pV, =G/VS (3)
Although the shaft model shown in Figure 6 was
used also for the internal shaft resistance, the
radiation damping was set to be zero since the
inward radiation does not occur in the soil plug.
In the base model (Figure 7), the slider represents
the maximum base (toe) resistance, q b , the spring
represents the elastic deformation of the soil below
the pile base, kb, the dashpot expresses the damping,
cb, and the mass is the lumped soil mass, Mb,
respectively. The soil parameters, kb, c b and Mb are
given as follows for soil plug base and for the annular
pile base (Deeks 1992):
for soil plug base;
respectively, and n is the porosity (n = 0.25 for the
mudstone).
The soil parameters except for the maximum shaft
resistance, f, and the maximum base resistance, q b ,
were determined by means of Equations (2) through
(9) with v = ves.The values off and q b alone were
established by the wave matching. The static load-
displacement curve was calculated using the
identified f and q b , and the drained value of the base
spring, kb, and the shaft spring, under static
loading. The ks(stat,c) was estimated according to the
following equations (Randolph 1991):

3.3 Comparison of the dynamic and static load tests

3.2pV5 - 3.2G
cb =- - The validity of the above procedure of the wave
n(1- v) n(1- v>y5 matching analysis was examined through the
0.1 - v 4 comparison of the results of the dynamic load test
M b =2d?p- and the static load test on the test pile.
1-v
for annular pile base;
8G
kb =
~ ( 1 v- ) ( d , + d;)
3.2pVs __- 3.2G
"b =-
n(1- v) n(1- v)V,
3 O.l-V4
M b = 2(d; - di )p-
1-v
where p is the soil density.
The mudstone is hlly saturated (degree of
saturation S = 0.995) and has a low permeability.
Hence, the undrained condition is thought to be
maintained during pile driving. In order to take
account of this situation, the equivalent Poisson's
ratio, v,,, which has been proposed by Verruijt
(1969), was used, instead of the drained Poisson's
ratio, v, of the mudstone ( v = 0.15), to estimate E0
and the base model parameters such as kb, cb and Mb:

where

K,, = K, + Kf / n (1 1)

where K,, Kf, K,., and K, are the bulk modulus of the
soil skeleton, the pore fluid, the saturated water (K, Figure 9. The results of the save matching analysis the re-
= 2000 MN/m2) and the air (K, = 2000 kN/m2), driving test.

700
 Figure 8 shows the results of the wave matching h Load on pile head, P (MN)
analysis of the driving test performed at the end of . g - 0 1 2 3 4 5
s u
v
initial driving. The upper figure (a) shows force vs 3
time and the lower figure (b) shows velocity vs time. G 5
10
Similarly, the results of the wave matching analysis of @

the re-driving test are shown in Figure 9. The 0E

-
15
maximum base resistance, q b , was identified to be a
.-U) 20
1500 kN/m2 for both the initial and the re-driving U
25
tests. \ (66 hours after EOID) ',
Figure 10 shows the distributions of the shaft 30 i -Static Load Test I -

resistance, J estimated from the initial driving and

the re-driving tests and the shaft resistance measured
_.

a 35 } , , : , ! , ! , , , I
(29 days after EOID)

in the static load test. Comparison of the Figure 1 1. Load-displacement curves estimated from initial
distributions of ffrom the two driving tests clearly driving and re-driving tests, together with static load test
result.

w
...... End of initial driving (EOID)
-Redriving (66 hours after EOID)
Static Load Test (29 days after EOID)
showed that about 80% of the ultimate capacity is
Shaft resistance, f (kN/rn*)
the shaft capacity, which is comparable with the
0 100 200 300E.L.1.44m static load test result.
1 : ' "
: I

-8
Figure 10. Distributions of shaft resistance estimated from
initial driving and re-driving tests, together with static load
test results.
indicates that a large set-up of the shaft resistance
occurred during the rest period of 66 hours. The
results from the re-driving test are comparable with
the static load test results.
The load-displacement curve derived from the
wave matching analysis of the re-driving test is fairly
comparable with the static load test result (Figure
11). The result of the wave matching analysis
4 ACTUAL PILINGS
4.1 Pile specifications and test sequence
A total of 32 piles were driven at abutment AI and
pier PI in March, 1995. Note that 2 piles out of the
20 piles at abutment AI and 2 piles out of the 12 piles
at pier PI had been driven 5 months before this pile
construction for drivability assessment. A total of 70
piles were driven at abutment A2, piers P2, P3 and Pq
in April, 1996.
The geometrical and mechanical properties of the
piles are listed in Table 2. Two sets of a strain gage
and an accelerometer were mounted 180' apart at a
distance of 1 m from the pile head for the dynamic
load tests.
In the pile construction work in 1995, 28 initial
driving tests and 33 re-driving tests were conducted
within 4 days without a delay of the usual pace of the
pile construction work. The initial driving tests were
performed on all the piles except the four piles which
had been driven prior to this construction work.

Table 2. Properties and design ultimate bearing capacities of the foundation piles of Noetsu Bridge No.3.
Design ultimate bearing
Pile properties
Abutments, Number capacities Qu(kN)
Piers of piles Ordinary Seismic Embedded Outer diameter Wall thickness
condition condition Length (m) length L,(m) do (mm> 4\ (mm)
9.0 upper 12 I
AI 20 3161 8.0 600
3706 (upper 5.0 +lower 4.0) lower 9
Pi 12 1598 2942 10.0 9.0 600 9
11.5 upper 12 I
p2 16 3787 3604 (upper 4.5+lower 7.0) 11.0 600 lower 9
16.5 upper 14 I
p3 20 1585 2556 (upper 5.5+lower I 1.o) 16.0 600 lower 9
-~
14.5 upper 12 I
p4 1200 16 14.0 600 lower 9
2624 (upper 4.5+1ower 10.0)
A2 18 2050 328 1 8.5 8.0 600 9
(for all piles) Young's modulus E = 2.06 X 10' MN/m2 , Density p = 7.86 ton/m3 , Bar wave velocity c = 5120 m/s

70 1

Figure 12. An example of wave matching for the re-driving
test of a pile at abutment A,.
Several piles underwent multiple re-driving tests. In
the pile construction work in 1996, 70 initial driving
tests (one test for each pile) and 116 re-driving tests
(several piles underwent multiple re-driving tests)
were conducted within 13 days.
4.2 Wave matching results
The procedure of the wave matching analysis
described in 3.2 was used for the wave matching
analysis of the dynamic load tests.
Figure 12 shows an example of wave matching of a
Figure 13. Load-displacement curves derived via wave matching for all piles at each abutment and pier.
702
pile at abutment AI. The estimated load-
displacement curves of all the piles at abutments AI,
A2 and at piers PI through P4 are shown in Figure 13.
Although rest periods of the piles between the end of
initial driving and the re-driving are different, the
ultimate pile capacity at re-driving is quite a bit larger
than at the end of initial driving. The estimated shaft
capacity amounted to about 80% of the estimated
ultimate capacity in all the piles.
The increase in the estimated static resistance, Qu,
of the piles at abutments AI, A2 and piers PI through
P4 as a function of elapsed time after the end of initial
driving are shown in Figure 14. It is seen that the
set-up ceases at about 100 minutes for abutment AI,
at 200 minutes for pier PI, at 1000 minutes for
abutment A2 and pier Pz. The estimated static
resistance after these time intervals exceeds the
required ultimate capacity. However, for the pile at
pier P3 and pier Pq, it was judged that the ultimate
capacity was not attained, because the set per blow
was very small, less than 1 mm.
The frequency distributions of the estimated
ultimate static resistance, Qu, of the piles obtained
after the elapsed time of 100 minutes in abutment AI,
200 minutes in pier PI, and 1000 minutes in abutment
A2 and pier P2 are shown in Figure 15, with the
statistical properties such as the average and the
coefficient of variance, COV. All the frequency
distributions seem to take the form of normal
distribution.
5 SUFFICIENT NUMBER OF DYNAMIC LOAD
TESTS
Let us discuss the sufficient number of load tests to
be conducted at a site to determine the distribution of
pile capacity. The design value, Qd, is assumed to be
given by
in which Qk is the characteristic value and YQ is the
partial safety factor.
If the bearing capacity of the piles takes the form
of normal distribution as shown in Figure 16, Q d and
Q k are expressed as (Murakami et a1 1988):
in which ,qis the average of Q, VQis the coefficient
of variance and k is a factor that depends on the
required safety level. If we take the probability of
Figure 15. Frequency distributions of estimated static resistance, Qu,of the piles at abutments A, and A', and piers PI and P2.
703
failure to be 5% for the characteristic value and 0.1%
for the design value, the value of k is calculated as
1.64 for the characteristic value and 3.0 for the
design value.
Using the results for pier P:! and abutment A2
shown in Figure 15, the relationships between the
number of load tests and the design and characteristic
values are examined, applying the non-parametric
bootstrap method. In this method, the number of
tests, n, was hypothetically varied f?om 2 to the total
number of piles (16 for pier PI and 18 for abutment
A2). For each test number, n, n data were selected
from the measured bearing capacities in random
fashion. Then, the values of ,UQ and VQ were
calculated from the selected data to obtain the values
of O d and Q k by means of Equation (14). The pair of
such obtained is called 'bootstrap sample'. A total of
1000 pairs of the bootstrap samples were obtained
for each test number, from which the values for the
average and the standard deviation of Q d and Q k
were calculated for each test number, n.

Figure 17. Variations of average and standard deviation of Qd
and Qkwith of number of load tests at Pier P?.
The variation of Qd and Qkwith the test number, n,
are shown in Figure 17 and in Figure 18, together
with the range of the standard deviation. It can be
seen that the average value converges when the
number of tests reaches 3 or 5 , whereas the standard
deviation tends to converge when the number of
tests reaches 10. These results suggest that the
execution of 10 or more load tests give us the design
pile capacity that has an acceptable probability of
failure which may be obtained from the load tests on
all piles for this particular site.
6 CONCLUSIONS
The results of the wave propagation analyses of the
dynamic load tests of a total of 102 steel pipe piles
for the foundations of Noetsu Bridge No.3 have been
presented in this paper. Rational soil models
(Randolph & Simons 1986; Deeks 1992) were used
in the wave matching analyses. A wave matching
704
Figure 18. Variations of average and standard deviation of Qd
and Qkwith of number of load tests at Abutment A?.
analysis procedure based on the rational soil model to
derive the load-displacement curve for a pile was
demonstrated in this paper. In the proposed wave
matching analysis procedure, soil investigation and
soil test data are effectively utilized to estimate the
soil parameters except the shaft resistance and the
toe resistance, and change in drain conditions
between dynamic and static loading are taken into
account .
The results of dynamic load tests showed:
I . The bearing capacity of all the piles exceeded the
design ultimate bearing capacity.
2. Set-up ceased within a short period after the end
of initial driving, 100 minutes to 1000 minutes,
depending on pile length.
Further, the sufficient number of load tests to be
conducted at a site to determine the distribution of
pile capacity was discussed from the view point of a
probabilistic design method, based on the results of
the dynamic load tests. It was found that the
execution of 10 or more dynamic load tests gives us
the design pile capacity that has an acceptable
probability of failure which may be obtained from the
dynamic load tests on all the piles for this particular
site.
REFERENCES
Matsumoto, T., Y . Michi & M. Hayashi 1997. Reliability of
dynamic load testing on steel pipe piles in soft rock. PTOC.
14th ICSMFE, Hamburg, v01.2 : 1185-1 188.
Randolph, M. F. & H.A. Simons 1986. An improved soil
model for one-dimensional pile driving analysis. Proc. 3rd
Int. Con$ on Num. Methods in Offshore Piling 1- 17.
Deeks, A.J. 1992. Numerical analysis of pile driving
dynamics. PhD Thesis, The University of Western
Australia.
Matsumoto, T., Y. Michi & T. Hirano 1995. Performance of
axially loaded steel pipe piles driven in soft rock. Jour. of
Geotech. Eng., ASCE, 121(4) : 305-315.
Matsumoto, T. & M. Takei 1991. Effects of soil plug on
behavior of driven pipe piles. Soils h Foundations,
JSSMFE, VOI.3 1 . No.2 : 14-34.
Randolph, M.F. & A.J. Deeks 1992. Dynamic and static soil
models for axial pile response. Proc. 4th int. Conf: on the
Appl. of Stress- Wave Theory to Piles: 3- 14.
Novak, M., T. Nogami,& F. Aboul-Ella 1978. Dynamic soil
reactions for plane strain case. J. Mech. Eng. Div., ASCE,
104(EM4): 953-959.
Verruijt, A. 1969. Elastic storage of aquifers.R.J.M. De Wiesl
(ed.), Flow through Porous Media, New York : Academic.
Randolph, M.F. 1991. Analysis of the dynamic pile driving,
Chapter in Developments in Soil Mechanics - IF Advanced
Geotechnical Advances, P.K. Banerjee & R. Butterfield
(ed.): Elsevier Applied Science.

705
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Behavior of short CFA piles in an overconsolidated clay based on static

and dynamic load tests

A.C. M. Kormann & €! R.Chamecki - Structures and Materials Laboratory, LAME, Federal University
of Paranci, Curitiba, Brazil
L. RUSSO Net0 - PonthiJcial Catholic Universig of PararicilIn Situ Geotecnia, Curitiba, Brazil
L.Antoniutti Net0 -In Situ Geotecnia, Curitiba, Brazil
G.€! Bernardes -Department of Civil Engineering, UNESR Campus of Guaratinguetci, Brazil

ABSTRACT: A comprehensive investigation of the behavior of two short continuous flight auger (CFA) piles
tested at the experimentation site of the Federal University of Paranh is presented. The piles were submitted
both to static and dynamic load tests. After the tests, a visual shaft inspection and complementary SPT bor-
ings were performed. Samples of the pile material were tested in order to evaluate the concrete specific
weight. The analysis shows the importance of the correct assessment of the pile concrete properties. For a
proper comparison between static and dynamic test results performed at the same drilled shaft, the preloading
effects should be accounted. Cast-in-place piles may develop increasing point resistance from large displace-
ments resulted from accumulative blows. The CAPWAP static simulations, for blows with energy increasing,
are compared with the results of the static load tests. A remarkable agreement was found in the load-
displacement curves obtained from both test methods.

1 INTRODUCTION 2 THE GEOTECHNICAL PROPERTIES OF THE

The use of dynamic load tests in both precast and
bored piles is increasing in Brazil. In some major
jobs where continuous ilight auger (CFA) piles were
used, dynamic load tests have been carried out for
foundation control.
There is a number of papers in the literature cov-
ering the dynamic high strain testing of drilled shafts
(e.g. Seitz 1984, Seidel & Rausche 1984, Likins &
Hussein 1995, Hussein et al. 1996, Liu et al. 1996,
de Mello & Paraiso 1998). However, the current ref-
erences are mainly related to the performance of the
tests. Pile geometry, concrete properties and soil in-
ertia effects require a proper evaluation for a correct
assessment of the dynamic behavior of bored piles.
Comparisons between static and dynamic tests at the
same pile should account for the preloading effects
mainly the accumulative pile displacement.
In order to contribute to a better understanding of
high strain dynamic testing of drilled shafts, this pa-
per describes the results of the research about the
behavior of CFA piles which is in progress at the
experimentation site of Federal University of ParanB
- UFPR. This is part of the studies about the founda-
tion behavior in the Guabirotuba Geological forma-
tion at ParanB State, south of Brazil. The information
here presented are also discussed by Kormann et
al. (2000).
EXPERIMENTATION SITE OF UFPR AND
THE CFA PILES CHARACTERISTICS
The Geotechnical Experimentation Site of the UFPR
is located at the “Centro Polit6cnico” Campus in Cu-
ritiba, Brazil (Chamecki et al. 1998).
The natural soil belongs to the tertiary Guabiro-
tuba Geological Formation. Overconsolidated silty
clays and clayey silts, with high plasticity, are com-
monly present soils in this scdimcntary formation.
Polished and shiny surfaces are commonly seen in
the clayey soil mass. These slickensides follow a
pattern of difficult identification. Well defined tec-
tonic structures are also prcsent. Lenses of sands,
rich in feldspar, frequently occur inside the clayey
mass. Some conglomerate and carbonate deposits
may occur at specific sites. A detailed study about
the geology of the Guabirotuba Formation is pre-
sented by Salamuni (1998) and the geotechnical as-
pects of these soils are discussed by Kormann
(1999). Figure 1 presents representative data of the
geotechnical profile and the pattern of the two CFA
piles at the time of dynamic testing.
At the site of the CFA piles, the subsoil of the
Geotechnical Experimentation Site consists basically
of high consistency gray and brown silty clay.
Lenses of sandy silt and silty sand are also present.
The subsoil shows high resistance to the SPT blow
count, with N values over 20 blows since the first
meter. The CPT test shows stiff clay subsoil, with

707
Figure 1. Geotechnical profile data and the pattern of the piles at the time of dynamic load tests.

point resistance as high as 12 MPa (Fig. lb). The 1. Employing some extrapolation methods (c. g.
friction ratio decreases with depth, changing from Van der Veen), it was suggested the ultimate loads
4 9% more superficially down to about 3 % close to of 1006 kN and 1473 kN for the CFA-1 and the
the test refusal at 6 m of penetration. A high degree CFA-2 piles respectively.
of saturation occurs as result of the water present at 2. The extrapolated values are near the maximum
the discontinuities of the soil mass. load applied during the load tests. These loads were
The CFA piles were installed in May/1997. Ta- limited by the structural integrity of the pile caps. It
ble 1 presents the main features of the two piles and is important to note that the applicability of the ex-
some information obtained by a Turacord system trapolation methods for bored piles have some limi-
during the installation. tations, especially when the maximum displace-
ments applied during the static tests were
insufficient.
Table 1. Pile geometry and installation data. 3. Although the lengths of the two tested piles
Properties CFA- 1 CFA-2 differed only by one meter a markedly distinct be-
Nominal diameter (m) 0.35 0.35 havior was observed in their load-displacement be-
Installed pile length (m) 6.0 7.0
Average injection pressure 100 > 200 havior. The CFA-2 pile presents more stiff and
at the auger head (kPa) higher bearing capacity.
Average auger withdrawal 4.7 5.6 4. Such unexpected behavior does not seem to be
velocity (cm/s) explained only by subsoil characteristics. The grout
pressure might have affected the performance of the
piles.
3 STATIC LOAD TESTS In order to improve the understanding about this
distinct behavior of the two piles further investiga-
Two slow maintained load (SML) tests were carried tion was planned, which includes:
out in Aug-Sep/98 according the Brazilian code 1. The test of the piles with low strain tests (PIT).
NBR-12131. The penetrations were 6.0 and 7.0 m 2. The reinforcement of the pile tops and the per-
for the CFA-1 pile and the CFA-2 pile respectively. forming of high strain dynamic tests.
The measured load-displacement curves are pre- 3. A visual shaft inspection of the piles and the
sented in Figure 2. The tests procedures, instrumen- surrounding soil.
tation setups and interpretation are described by 4. The collection of samples of pile material in
Kormann et al. (1999). The following points were order to determine the concrete specific weight.
outlined by these authors:

708

Figure 2. Static load tests results.
5 . The performing of additional SPT borings just
beside the piles.
These investigations and the results of all the
available data are presented and discussed in the
following items.
4 LOW STRAIN TESTS
Since the penetration of the piles was well con-
trolled, the low strain tests allowed to calculate the
wave speed, resulting 3550 m/s and 3800 m / s for the
CFA-1 and the CFA-2 piles respectively. These
wave velocities are somewhat reduced.
Although the concrete used in both piles was es-
sentially the same, the higher wave speed of the
CFA-2 pile might be associated to a material having
a slightly better performance (Kormann et al., 1999).
5 HIGH STRAIN DYNAMIC TESTING
In order to perform the dynamic testing the pile tops
needed to be reinforced. The reinforcement was car-
ried out using longitudinal steel bars and a cylindri-
cal steel ring mounted at the top of the piles
(Fig. la). It is important to note that the reinforce-
ment reduced the pile embedded length by approxi-
mately one meter which resulted in some loss of
shaft friction near the top of the piles. The dynamic
load tests were conducted in Dez/99.
For a proper selection of the hammer to be used
for dynamic testing of cast-in-place piles, Hussein et
al. (1996) suggest a ram weight about 1.5% of the
pile static resistance to be verified. A 30 kN free fall
hammer was selected for the high strain dynamic
tests. One cushion made by wood plate was used to
reduce the impact maximum stress.
Two piezoeletric accelerometers and two strain
transducers were attached to the piles. A Pile Driv-
ing Analyzer, PAL model, was employed
(PDI, 1999). In order to mobilize as much as possi-
ble the soil resistance, the tests employed increasing
hammer energies (e.g. Aoki 1989, Bernardes 1989,
Beim & Aoki 1996). The drop height was increased
in 20 cm increments until the compressive stresses
709
caused failure of the pile material. The set and re-
bound of the pile top were recorded for all blows.
Table 2 summarizes the dynamic load test data
for the CFA-1 and the CFA-2 piles. The first pile
tested was the CFA-2 pile. At the beginning of the
test the new cushion with low stiffness and low coef-
ficient of restitution generated a low driving system
efficiency - a reduced energy (EMX) was trans-
ferred. The pile was initially stroked six times with
40 cm fall height until the set became perceptible.
During the test, the driving system efficiency in-
creased with the succession of the blows. An effi-
ciency ranging from 12.2 to 38.6% was observed.
The dynamic load test comprised seven blows with
drop heights ranging from 60 to 170 cm. The maxi-
mum tension stresses below sensors (TSX) were
small for all blows. The maximum compressive
stress (CSX) of 28.5 MPa generated by the seventh
blow caused the pile top material to fail just below
the steel reinforcement.
The CFA-1 pile received six blows with drop
heights comprised between 40 and 140 cm. The
driving system efficiency increased from 1 1.7 to
32.1% during the test. The tension stresses were
slightly higher than i n the CFA-2 pile, but as ex-
pected the successive increase in compressive
stresses caused the failure of the pile top immedi-
ately below the steel ring. The compressive stresses
that controlled the end of the test (15.0 MPa at the
gages) were lower than the CFA-2 ones. It can be
related to bending effects due to eccentric blows or
concrete properties.
Figure 3 shows the records of force and velocity
multiplied by the impedance for the two piles. For
the CFA-1 pile, a wave speed of 3000 m/s is suit-
able. The CFA-2 pile shows a wave speed of
3250 d s . These findings are in full agreement with
the wave speeds measured in the PIT tests, which
due to the level of deformation usually present
higher values. The ratio between the wave speed
from both tests for the CFA-2 pile to the CFA-1 one
was the same regardless the type of test (1.07- 1.OS).
Due to the reduced length of the piles, the width of
the pulse impact is considerably higher than 2L/c.
Despite that, it will be demonstrated that good re-
sults can be achieved.
The traces of the force and the velocity of the
CFA-1 pile suggest a high skin friction (Fig. 3a).
The succession of the increasing energy blows
causes the force signal to be close to the velocity
signal after 4L/c. It can be related to increased pile
point displacements. A considerable pile set (9,4
mm) was measured at the last blow (Table 2). It is
expected a high mobilization of the soil resistance.
The first record of the CFA-2 pile (Fig. 3b) shows
an increase in the velocity just after 2L/c. It is re-
lated to a hammer interference. The subsequent rec-
ords do not present such feature.
 Table 2. Dynamic load test data.
CFA- 1 CFA-2
Blow Drop EMX CSX TSX Set Rebound Blow Drop EMX CSX TSX Set Rebound
number height (kNm) (MPa) (MPa) (mm) (mm) number height (kNm) (MPa) (MPa) (mm) (mm)
(cm) (cin)
1 40 1.4 9.4 0.4 0.1 1.9 1 60 2.2 14.1 0.0 0.5 1.6
2 60 3.4 11.2 0.4 1.8 2.2 2 80 5.0 19.0 0.0 0.6 2.9
3 80 5.1 12.5 0.3 3.0 2.2 3 100 8.2 21.5 0.0 1.4 3.8
4 100 8.2 13.8 0.5 5.2 2.3 4 120 12.2 23.9 0.5 2.2 5.1
5 120 11.8 15.1 0.1 8.0 2.0 5 140 14.9 25.1 0.3 2.8 4.2
6 140 13.5 15.0 0.7 9.4 2.1 6 160 18.1 27.4 0.0 4.3 3.7

BN - Blow number
H - Hammer drop height
0- Pile ( ~ u c )

1500 00 2.70 4 1s
BN=l H=40cm

-7 -7
3 40 rns L _.--
+]
1500.00
B N = 2 H=60cm

I / .-
\ '1
4
-\

L / I I ' , A 2 0 5m5
- 20 5m5
3 20 rns
3 20 rns I \ .//

Figure 3. Records of force and velocity multiplied by the impedance for CFA-1 (a) and CFA-2 (b) piles.

710
The force signal is considerably higher than the ve-
locity record. It can be due to compressive upward
waves that are related to a high skin friction.
The comparison between the measured sets and
rebounds of the two piles also reflects the distinct
behaviors observed both in the static load tests and
in the dynamic records. Despite the higher energies
applied to the CFA-2 pile, its sets are lower than a
half of the CFA-1 ones. The measured rebounds of
the CFA-2 pile were as high as 5.5 mm, while in the
CFA- 1 pile the maximum rebounds were 2.3 mm.
All blows presented in Figure 3 were submitted to
CAPWAP analysis. The results will be presented in
the item 9.
6 VISUAL SHAFT INSPECTION
After the dynamic tests, the soil around the piles was
excavated approximately 3.2 m below the ground
level.
The visual inspection of the soil allows the fol-
lowing remarks to be made:
1. As expected, the silty clay around the piles has
polished and shiny surfaces closely spaced. Slicken-
sides following a preferential orientation also were
found. The clayey soil around the two piles seems to
be similar.
2. A hard stratum of sandy silt is present from the
ground level to 1.3 m deep at the CFA-2 pile (see
Fig. la). It is important to note that skin friction of
this layer acted only on the static test, since the soil
around the pile top was removed for the dynamic
testing.
The following comments can be outlined about
the inspection of the piles geometry and material:
1. The surface of the two piles was very smooth
and regular. There was no sudden changes in the pile
diameter which was expected for this type of pile
construction in stiff clayey soils.
2. The actual diameter of the two piles is lower
than the nominal diameter (35 cm). The measured
diameter ranged between 33 and 34 cm.
3. The reports of the installation data of the two
piles indicated excessive amount of concrete used -
47-59% more volume than the nominal volumes of
the piles. The direct measurements of the diameter
of the piles show that there is no reason for the high
concrete overconsumption and a better control dur-
ing pile installation is required.
4. The CFA-1 pile shows some small soil intru-
sions along the shaft. They are present at the exter-
nal pile surface and also approximately over to 4 cm
inside the concrete mass. The CFA-2 pile does not
exhibit soil intrusions. It may explain the lower
wave speed (3000 d s ) measured in the CFA-1 pile.
711
7 CONCRETE TESTING
In order to evaluate the concrete density, a total of
17 samples of the material of the two piles were
collected just below the steel reinforcement and at
the bottom of the excavation.
Table 3 presents the mean results of the tests. It is
important to note that the concrete specific wei ht of
the piles - in the range of 19.13-19.52 kN/m - is 5
low. It is due to the high porosity of the material.
These findings are in agreement with the low wave
speeds measured. It can be noted that despite the low
concrete density and probably the low resistance the
piles supported well the loads both of the static and
the dynamic tests.
Table 3. Average pile concrete properties.
Pile Concrete specific Voluine of voids / To-
weiglit (W/m’) tal volume (9%)
CFA- 1 19.52 25.07
CFA-2 19.13 25.56
It should be noted that the concrete specific weight
is currently used in wave equation analysis to cal-
culate the dynamic elastic modulus of the pile. It is a
common practice to adopt to the specific weight the
value of approximately 24 kN/m3. However, if the
cast-in-place shaft exhibits a reduced specific weight
the analysis would produce overestimated results.
In order to illustrate the influence of the concrete
specific weight, Figure 4 shows a simplified analysis
for the CFA-2 pile. The pile capacity was evaluated
with the Case Method. It is observed that for a spe-
cific weight range of 20-24 kN/m3 the pile capacity
can change about 20%.
Figure 4. The influence of the concrete specific weight in the
pile capacity evaluation.
8 ADDITTIONAL SPT BORING
Since the excavations around the piles were not suf-
ficient to explain the distinct behavior between the
two piles, complementary SPT borings were per-
formed just beside the CFA-1 and the CFA-2 piles.
The distance between the borings and the piles was
approximately 50 cm. Obviously, the excavation
around the piles implies in a stress relief that
 Table 4. Complementary SPT boring data.
CFA- 1 CFA-2
Depth Blows130 cni Blows/30 cm Blows130 cm Blows130 cm
below (first and (second and (first and (second and
ground second 15 third 15 cm) second 15 third 15 cm)
level cm) NSI’T cm) NSPT
(m)
4.0 16 27 16 27
5.0 21 30128cm 22 30 1 2 5 cni
6.0 28 30123cm 19 30
7.0 30129cni * 17 19
changes the conditions of the soil. In effect, even the
pile installation changes the natural state of stresses.
However, since the conditions of the ground around
the two piles were similar at the time of the com-
plementary SPT boring, a qualitative comparison
can be made among them. Table 4 presents the re-
sults.
The following remarks can be made:
1. The complementary SPT borings data are in
agreement with the SPTT-1 information.
2. The new borings revealed a lens of silty sand
just above the point of the CFA-2 pile and 60 cm
below the CFA-1 pile. These thin layer is approxi-
mately 40 cm thick and it probably appears in the
soil profile as presented in Figure 1 a.
3. The N s p ~values along the shaft of the CFA-2
pile are not higher than the CFA-1 ones. At the point
of the CFA-2 pile the N s p ~values are lower than the
CFA-1 ones.
4. On the basis of the known limitations of the
SPT testing, the blow count suggests that the higher
resistance of the CFA-2 pile could not be explained
only by a stronger soil around the shaft or at the pile
point. The hypothesis of the influence of the grout
pressure in the performance of the pile cannot be
disregarded. Also, since the sand lens intercepts the
CFA-2 pile just above the tip, a distinct response of
this soil to the installation procedure - such as an
slight enlargement - might affect the pile end bear-
ing.
9 CAPWAP ANALISYS: COMPARISONS
WITH STATIC LOAD TESTS
The records of both the CFA-1 and CFA-2 piles
were submitted to conventional CAPWAP analysis.
In order to perform a suitable analysis eight pile and
soil segments were employed. A concrete specific
weight of 20.0 kN/m3 was adopted. The main results
are presented in Table 5.
The skin quake values of the two piles lies in a
narrow range (0.97- 1.622 mm) and they do not de-
pend on the energy level. On the other hand, the toe
quakes exhibit a trend to increase with the energy
transferred to the pile. It can be related to a growing
mobilization of the toe resistance. High toe quakes
712
were found, in special for the CFA-1 pile. For the
two piles the ratio between the toe quake and the
measured set is approximately constant - a repre-
sentative range is 1.O- 1.7.
The Case Method damping JC values of Table 5
does not exhibit a trend and should be disregarded.
The shaft and toe Smith damping values of the two
piles showed dependent on the energy level. At the
first blows the toe damping was too high. It de-
creases with the increasing of the drop height. Such
behavior is an agreement with the findings of Aoki
& de Mello (1992). These authors show that the
damping and quake of the Smith model are not con-
stant soil parameters, but are dependent on the en-
ergy level.
The use of the radiation damping - that may ac-
count for the soil motion associated with the pile
movcment in drilled shafts - did not lead to im-
proved results. It seems in agreement with the
smooth surface of the piles revealed by the visual in-
spection and the stiff characteristics of the soil.
Rausche et al. (1996) discusses the applicability
of the Multiple Blow Analysis (MBA). They rec-
ommend such analysis for variable energy blows and
provide an example for a drilled shaft. However, for
the piles here studied such analysis did not seem
suitable. The MBA analysis keeps constant the end
bearing. That is not in agreement with a expected in-
creased mobilization of the toe resistance due in-
creasing hammer energies. In this study, the Resid-
ual Stress Analysis (RSA) option also did not
improve the results.
The mobilized capacities of the two piles are
compatible with the results of the static load tests. At
the end of the dynamic tests, for the CFA-1 and the
CFA-2 piles respectively, the maximum mobilized
capacities were 10.7% and 48.1% higher than the
static capacities reported by Korrnann et al. (1999).
The high CAPWAP resistance of the CFA-2 pile
suggests that the limited settlements of the CFA-2
pile in the static tests may have conducted to an un-
derestimation of the extrapolated capacity.
It is important to note that the CFA-1 and the
CFA-2 piles were statically loaded before the dy-
namic test. These preloading leads to a set of resid-
ual stresses both in the pile and in the soil (e.g Mas-
sad 1992, Maiorano et al. 1996). The comparison
between static and dynamic load tests performed on
the same pile should account for the residual
stresses. The set of residual stresses does not change
the ultimate load. However, the form of static load-
ing curves can be markedly affected. The residual
loads generated by a cycle of static or dynamic
loading increase the pile head stiffness.
In addition, as discussed by Dkcourt (1998), the
meaning of the physical failure cannot be applied to
bored piles. Cast-in-place piles may develop in-
creasing point resistance over to large displace-
ments.
 Table 5 . Main CAPWAP results.
CFA- 1 CFA-2
Blow number 1 2 3 4 5 6 1 2 3 4 5 6 7
Mean Shaft Quake (mm) 1.298 0.996 1.001 1.001 1.001 1.000 1.361 0.9 7 1.445 1.526 1.595 1.622 1.000
Toe Quake (mm) 0.764 2.682 4.080 6.413 8.557 11.05 0.757 1.004 2.109 3.422 3.573 5.866 5.028
Shaft Damping (s/m) 1.298 0.512 0.324 0.208 0.091 0.099 1.183 0.713 0.533 0.39 0.138 0.1 17 0.1 17
Toe Damping (dm) 1.419 2.682 1.411 0.677 0.663 0.13 4.936 2.526 2.055 1.301 1.371 0.974 0.702
Jc 0.36 0.68 0.7 1 0.42 0.41 0.07 0.00 0.18 1.oo 1 .oo 0.84 0.63 0.59
MQno 1.02 1.32 1.37 2.34 1.96 2.21 1.28 0.86 0.93 1.58 1.14 3.26 3.12
RS (kN) 62 1 590 620 719 760 800 1296 1592 1570 1470 1574 1776 1876
RT (kN) 24 8 296 280 300 300 314 148 167 180 310 300 309 306
RU (kN) 869 886 900 1019 1060 1114 1444 1759 1750 1780 1874 2085 2182
RS - Skin mobilized resistance RT - Toe mobilized resistance RU- Total mobilized resistance

If the soil resistance degradation effects are reduced,
it is expected that the succession of static or dynamic
cycles of loading will lead to increased pile capaci-
ties. For instance, Liu et al. (1996) points out that
when the dynamic test is performed after the static
loading, in average a 10% increased pile capacity
can be found.
On the basis of the discussion above, it seems
suitable to compare static and dynamic tests per-
formed at the same pile as a succession of cycles of
loading. In order to perform such analysis, the per-
manent displacement at the end of a cycle of loading
can be added to the next one (e.g. Seitz 1984, Ni-
yama & Aoki 1991).
Figure 5 shows the static load tests and the
CAPWAP static simulations superimposed as cyclic
loading. In order to plot the calculated curves, the
measured sets were employed. The range of 0.928-
I . 1 15 is representative for the ratio among the meas-
ured sets and the CAPWAP ones. The unloading of
the blows number 4 to 7 of the CFA-2 pile was ad-
justed as a straight line between the CAPWAP
maximum displacement (DMX) and the measured
set.
The simulations plotted in Figure 5 shows a re-
markable agreement with the static load tests. The
CAPWAP inferred static behavior follows within
reasonable limits the trend of the static load tests.
For the CFA-I pile simulations, it can be ob-
served that the succession of blows increases the end
bearing mobilization and reduces the toe stiffness.
The gradient of the last straight portion of the load-
ing cycles increases with the succession of the
blows. In effect, the ratio RT / Toe Quake decreases
when the drop height increases. It can be related to a
nonlinear stress-strain behavior of the soil. Despite
the lower displacements, the same trend can be ob-
served in the CFA-2 pile data.
10 CONCLUSIONS
At the current stage of the research here presented,
the following conclusions can be outlined:
1. The visual shaft inspection of the two CFA
piles revealed a very smooth and regular surface. It
is consistent with the stiff and hard clayey soil
around the piles. The actual diameter of the two
piles, ranging between 33 and 34 cm, is lower than
the nominal diameter (35 a n ) . It shows that there is
no reason for the high concrete overconsumption re-
ported by the installation monitoring system. The
CFA-I pile shows some small soil intrusions over to
4 cm inside the concrete mass. The wave speed of
the CFA-1 pile (3000 m/s) was lower than the wave
speed of the CFA-2 pile (3250 m/s).
2. The concrete testing of the CFA-1 and the
CFA-2 piles showed a reduced specific weight
(19.52-19.13 kN/m3) and a high porosity. Despite
that, the structural behavior of the piles was satis-
factory both in the static and dynamic load tests.
However, it should be noted that the concrete spe-
cific weight is currently employed in wave equation
analysis to calculate the dynamic elastic modulus of
the pile. The use of an overestimated concrete spe-
cific weight would produce an overestimated com-
putation of the cast-in-place pile capacity. In addi-
tion, the low specific weight suggests that the effects
in the wave equation results of a possible nonlinear
stress-strain relationship of the pile material should
be investigated.
3. The distinct behavior observed between the
two piles in the static load tests was confirmed by
the dynamic load tests. Despite their similarity, the
pile-soil interaction of the two piles is markedly dis-
tinct. The CAPWAP analysis suggest that the skin
friction of the CFA-2 pile is higher than the CFA-1
one. On the basis of the current soil investigation
performed, the higher capacity of the CFA-2 pile
docs not scem to be explained only by a stronger soil
around the shaft or at the pile point. The hypothesis
of the influence of the grout pressure on the pile-soil
interaction cannot be disregarded. Also, since a sand
lens intercepts the CFA-2 pilc just above the tip, a
distinct response of this soil to the installation pro-
cedure might affect the pile end bearing.
4. For the two piles, the maximum mobilized ca-
pacities in the dynamic tests ( 1 1 14 and 21 82 kN) are
in reasonable agreement with the results of the static
load tests (1006 and 1473 kN). The higher
CAPWAP capacities are due to the higher displace-
ments achieved in the dynamic tests. A comparison
between static and dynamic load tests performed on
the same pile should account for the preloading ef-
fects. Cast-in-place piles may develop increasing

713

Figure 5. Static load tests and CAPWAP static simula-
tions for the CFA-1 (a) and the CFA-2 (b) piles.
point resistance over to large displacements. If the
soil resistance degradation is reduced, it is expected
that the succession of static or dynamic cycles of
loading will lead to increased pile capacities.
5 . In order to compare the results of the static and
dynamic load tests, the permanent displacement at
the end of a cycle of loading should be added to the
next one. As shown in Figure 5 , the CAPWAP static
simulations shows a remarkable agreement with the
trend of the static load tests.
6. The results here presented demonstrate the po-
tential of the high strain dynamic testing as a valu-
able tool for the assessment of the load-settlement
behavior of drilled shafts.
ACKNOWLEDGMENTS
The authors would like to express their gratitude
to Eng. Alexandre Chwist (Estacas Premold), Eng.
Janies Barossi (Sondar) and Norberto E. Calliari for
their support in performing the tests here reported.
REFERENCES
Aoki, N. 1989. A new dynamic load test concept. Proc. Dis-
cussion Session 14, TC Pile DiYvirig, X I 1 lilt. Coi$ Soil
Mech. Fouiidcition Erig., Rio de Janeiro: 1-4.
714
Aoki, N. & V.F.B. de Mello 1992. Dynamic loading test
curves. Proc. 4th Itit. Cot$ Applic. of Stress- Wuve Theory
to Piles: 525-530.
Beim, J. & N.Aoki 1996. Dynamic load test method with vari-
able energy. Proc. 5th lilt. Conf: Applic. of Stress- Wave
Tlieoiy to Piles: 274-28 1.
Bernardes, G. P. 1989. Dyiiciiiiic mid stcitic testbig of large
model piles in sciiid. D.Sc. Thesis, Norwegian Institute of
Technology, Trondheim.
Chamecki, P.R., A.C.M. Kormann, N.A. Nascimento & A.S.
Dyniinski 1998. Sitio experimental de geotecnia da UFPR -
objetivos e dados preliminares. Proc. of X I COBRAMSEG,
ABMS, Bmsilia, Brazil: 8 19-826.
de Mello, L.G. & S. Paraiso 1998. Variable energy dynamic
load test on 1.0 m diameter CFA pile. BAP 111, Belgiuii:
32 1-334.
Dkcourt, L. 1998. Ruptura de fundaG6es e coeficientes de segu-
r a n p a luz do conceit0 de rigidez. Proc. o f X I COBRAM-
SEC, ABMS, Brcisilia, Brazil: 1599-1606.
Hussein, M., G. Likins & F. Rausche 1996. Selection of a
hammer for high-strain dynamic testing of cast-in-place-
shafts. Proc. 5th 1nt. Coi$ Applic. of Stress- Wave Tlzeory to
Piles: 759-772.
Kormann, A.C.M. 1999. Comportamento de argilas rijas: as-
pectos geotkcnicos da Forma@io Guabirotuba. A m i s du
M e w Redoiidci Ccimcteristiceis Geote'cniccis du Fosrii~ip7o
Gu(ibirotubLi,ABMS/UFPR, Curitiba, Briizil: 1 19- 128.
Korniann, A.C.M., P.R. Chamecki, N.A. Nascimento & A.S.
Dyminski 1999. Load tests on continuous flight auger piles
in the Guabirotuba Formation. Proc. X I Punciiiiericciti Cot$
on Soil Mecli. iirid Geotech. Eiig. : 1537- 1544.
Kormann, A.C.M., P.R. Chamecki, L. Russo Neto, L. Antoni-
utti Neto & G.P. Bernardes 2000. Estacas hklice continua
em argila sobreadensada: comportamento em provas de
carga est6ticas e dinbmicas. Proc. SEFE l V , S3o Paulo,
Brazil.
Likins, G.E. & M.H. Hussein 1995. High strain dynamic test-
ing of drilled shafts and cast-in-place piles. Deep Foulid.
liist., 20th Aiiriuul Menibera Cot$ Meet., Cliurlestoii.
Liu, C., Q. Lin & F. Shi 1996. Determining the bearing capac-
ity of large-diameter bored cast-in-situ piles by high strain
dynamic pile-testing. Proc. 5th lrit. Coi$ Applic. of Stress-
Wuve Tlieoiy to Piles: 797-804.
Maiorano, R.M.S., C. Viggiani & M. Randolph 1996. Residual
stress system arising from different methods of pile instal-
lation. Proc. 5th lilt. Coilf: Applic, of'Stress-Wuve Theory to
Piles: 5 18-528.
Massad, F. 1992. Sobre a interpretaq5,o de provas de carga em
estacas, considerando as cargas residuais na ponta e a re-
vers3o do atrito lateral. Park I: solos relativamente ho-
mogEneos. Solos e Roclzas 15 (2): 103-115.
Niyania, S. & N. Aoki 1991. Correla@o entre provas dc carga
dinbmica e esthtica no campo experimental da
EPUSP/ABEF. 2'. Seni. de Erig. de FuridugGes EspeciLiis,
SCO Ptiiilo, B t - ~ i ~ i285-293.
l:
PDI 1999. Pile Dt-iviiig Aiidyzer, PAL Model, Users Mariual.
Cleveland: Pile Dynamics.
Rausche, F., B. Richardson & G. Likins 1996. Multiple blow
CAPWAP analysis of pile dynamic records. PI-oc. 5th bit.
Cor!f:Applic. of Stress-Wcive Theor? to Piles: 435-446.
Salamuni, E. 1998. Tectdriicci dci bucia Sediiiientar de Curitibu
(PI?). P1i.D. Thesis, Instituto de GeociEncias e CiEncias
Exatas - UNESP. Rio Claro. Brazil.
Seidel, J. & F. Rausche 1984. Correlation of static and dynamic
pile tests on large diameter drilled shafts. Proc. 2nd Iiit.
Conf;Applic. of Stress- Wcive Theory to Piles: 3 13-3 18.
Seitz, J.M. 1984. Dynamic testing of bored piles i n non-
cohesive soils. Proc. 211d lrit. Cot$ Applic. of Stress- Wuve
Tlieorj to Piles: 201 -209.
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Static and dynamic testing of the 'Campile' - A displacement, cast-in-situ

pile

David J. Klingberg & Phi1 Mackenzie

Wagstaff Piling Pty Limited, Brisbane, Qld, Australia

ABSTRACT: Throughout the world, developments are continually being made in the foundation industry to
overcome construction and environmental difficulties. The need for a low noise and vibration method for
installing piles led to the development of such a system called the 'Campile'. The 'Campile' was developed in
Australia by Wagstaff Piling as a means of installing piles on sites that are sensitive to noise and vibration,
whilst maintaining the advantages of a displacement pile. This paper outlines the results of a dynamic and
static test program conducted on a 'Campile', with the static load test taken to failure. The test results indicate
a good correlation between the two test methods.

1 INTRODUCTION the probe is replaced by concrete or grout as the

Within the foundation construction industry,
environmental considerations with respect to noise
and vibration levels have become of more concern in
recent years. Modern piling contractors continually
look for new ways to provide cost-effective and
technically sound solutions to these environmentally
challenging projects. The Campile was developed by
Wagstaff Piling, Australia as the result of such a
need.
The Campile is a displacement, cast-in-situ pile that
is installed under low noise and vibration conditions.
As a displacement pile, it combines the advantages of
non-removal of soil, but maintains the low noise and
vibration of traditional non-displacement, drilled
piling methods.
This paper presents the outline of the test pile
programme completed on a Campile, summarises the
results of the 'Class A' dynamic test data and
provides a comparison of these results with the static
test data. Some recommendations are also included
in this paper with respect to methods to be adopted
for wave equation analysis for displacement,
cast-in-situ piles.
probe is removed from the ground. The finished pile
may be reinforced either partially or full-length,
depending on the design requirements.
The Campiles installed for this project were at a
developmental stage and were constructed as a
straight-shafted, nominal 450mm diameter pile.
Further advances have since been made with the
design and installation of the Campile which allows
for other sizes and profiles to be constructed,
including a 'screw-shaped' pile which provides
increased shaft resistance and greater overall pile
capacity.
3 SITE GEOLOGY
The project for the installation of the Campiles was
located at an industrial site near Sydney, Australia.
The site was underlain by loose to medium dense
sand and silty sand layers to depths of approximately
4 to 5 metres over a dense sand layer to 11 metres.
The target penetration for the Campiles ranged
from 5.5 to 7.5 metres where the base resistance was
expected to be high in the dense sand layer.

2 DESCRIPTION OF THE CAMPILE 4 TEST PILE PROGRAMME

The Campile is a proprietary pile of Wagstaff Piling 4. I Dynamic test results

and is constructed by screwing a specially designed
probe into the ground, which displaces the soil as the Test pile TP was installed as a sacrificial pile for the
probe penetrates the ground. The void created by project. The pile was a 450mm diameter (nominal)

715
Campile with a total length of 7.0 metres and a Table 2. ‘Class A’ CAPWAP@analysis summary.
drilled depth of 5.8 metres. The Campile was Pile ShaR End Total
constructed with grout using a 40MPa design No. Resistance Resistance Resistance
strength. Test cubes were crushed prior to the (W (W (W
dynamic test being performed and indicated a TP 370 1,427 1,797
strength of approximately 32MPa at 7 days. The
dynamic test was conducted when the pile was 8 4.2 Static test results
days old.
To enable the test piles to be subjected to dynamic The test pile was subjected to an incremental
and static testing, a special 40 MPa concrete head sustained load (ISL) compression test in accordance
was cast onto the pile to contain the impact and with the requirements of the Australian Standard@
bursting stresses from the impact of the test hammer. Piling Code - AS2159 (1995). The load test was
The test hammer was a five (5) tonne drop hammer, conducted to a load level where plunging failure of
with drop heights up to 13OOmm used during the the pile occurred.
testing. The results of the ISL load test have been
The dynamic testing and CAPWAP@analysis were summarised in Table 3 and shown graphically in
completed as ‘Class A’ predictions (ie. without Figure 2.
knowledge of the static test results). The typical As dynamic load tests cannot predict any time
dynamic test data is shown in Figure 1, where the dependent deflections and to enable a realistic
difference in impedance between the pile head and comparison to be made, all deflections from the ISL
pile shaft (as expected) is clearly seen. The results of test have also been corrected for creep. The creep
the dynamic testing are summarised in Table 1, with corrected values are given in Table 3.
the results of the CAPWAP@ analysis included in
Table 2. Table 3. ISL compression test results summary.
Load Pile head Creep Creep
PILETEST PD(PLEDRMV,UULIZER
(kN) Deflection Correction Corrected
TP-RST
(mm) (mm) Deflection
51 DROP W E R
(mm)
..................... Lrz 225 0.52 0.00 0.52
RMX 1890 kN
450 1.63 0.00 1.63
EMX 43.30 kKm 675 3.06 0.03 3.03
DMX 22.8mn
WU2 100kN 900 4.75 0.20 4.55
LE 6.10111
50 m s 1,125 8.15 0.27 7.88
...... .................. ....... ..... .........
1,350 17.70 5.89 11.81
3.2mS 0 1,575 25.60 6.41 19.19
1,800 50.00 10.27 39.73
.................. F 1‘
4.3 Comparison of static and dynamic test results

The results of the static and dynamic tests have been

kN ~ l ..,. p.. .x . l
I & ; s
............................................................ plotted in Figure 3. It can be seen that the
..
.
. I ..I
CAPWN@ analysis provided a conservative
... ,.
I load-deflection behaviour compared to the static test
data. Afker the results of the static test were known
and compared to the ‘Class A’ predictions, the
Figure 1. Dynamic test data for test pile TP CAPWAP@analyses were re-analysed to provide a
predicted deflection behaviour that more closely
modelled the static behaviour. The results of the
Table 1. Dynamic testing results summary. re-analysis are given in Table 4. It is noted that the
Pile Stroke Set TC EMX RMX revised CAPWAP@analysis did not alter the ultimate
No. (mrn) (mm) (mm) (kNm) (kN) capacity of the pile, but adopted a different
TP 1,300 16.0 6.8 43.3 1,890 distribution of capacity to achieve the ‘better match’
to the ISL test results. The comparison between the
Set = pile set in mm/blow revised CAPWAP@analysis and the static test data is
TC = temporary compression in mm given in Figure 4 and Table 5 . The revised
EMX = max. energy transferred to pile head in kNm CAPWAP@ analysis showed excellent agreement
RMX = PDA capacity estimate in kN with the ISL test results.

716
 Figure 4. ‘Revised’ CAPWAP@ and ISL
compression test results.

4.4 Discussion on results

The re-analysis of the dynamic test data and the

corresponding better match with the ISL test results
indicated that the shaft resistance on the pile was
significantly higher than initially obtained from the
‘Class A’ CAPWAP@ analysis. The revised
CAPWAP@showed approximately twice the shaR
resistance of the ‘Class A’ analysis. The re-analysis
was achieved with significantly lower skin damping
and quake parameters, with only minor changes to
the toe damping and quake parameters.
It should be noted that shaft radiation damping
Figure 3 Comparison of ISL compression test and parameters were used in both the ‘Class A’ and
‘Class A’ dynamic test results. ‘revised’ CAPWAP@analyses (CAPWAP@Manual,
1996). Without the use of this modelling tool, the
predicted ultimate capacity from the CAPWAF@
Table 4. ‘Revised’ CAPWAP@analysis summary. analyses may have been significantly lower than the
Pile Shaft End Total static test value (approx. 15% to 20% lower).
No. Resistance Resistance Resistance A comparison of the shaft resistance values from
(W (lm (W the ‘Class A’ and ‘revised’ CAPWAP@analyses was
also made with standard pile design calculations
TP 776 1,021 1,797
using the borehole information for the site (Poulos
and Davis, 1980). Three separate design calculations
Table 5. Revised CAPWAP@and ISL compression were performed on the borehole data: (a) bored pile
test results summary. design; (b) grout-injected pile design; and (c)
Load ‘Class A’ Revised ISL Creep driveddisplacement pile design. A summary of these
design calculations is given in Table 6.
(kN) CAPWAP@ CAPWAP@ Corrected
Deflection Deflection It can be seen from Table 6 that the ‘Class A’
Deflection
(mm) (mm) (mm) values from CAPWAP@correlated with bored pile
design parameters, whereas the revised CAPWAP@
225 1.78 0.48 0.52 values were slightly higher than those from a
450 3.55 0.96 1.63
driveddisplacement pile design. This confirmed that
675 6.34 1.50 3.03 the Campile was performing in the manner in which it
900 10.59 4.39 4.55 had been designed - that is, as a displacement pile
1,125 14.89 10.67 7.88
and not as a non-displacement pile.
1,350 19.20 16.95 11.81
1,575 26.23 23.22 19.19
1,800 43.79 43.85 39.73

717
Table 6. Design calculations summary.
Depth Stratum Bored Pile GIP Driven CAPWM@
(m) Design Design Design (W
(kN) (kN) (kN) ‘Class A’ ‘Revised’
2.5 L/MD Sand 107 193 214 126 253
3.2 Loose Sand 29 52 58 45 84
4.5 MD Sand 73 132 147 61 185
5.8 Dense Sand 199 22 1 23 8 138 254
Total 408 598 657 370 776

5 CONCLUSION REFERENCES

The Campile is a proprietary pile type developed by Australian Standard Piling Code. AS 159 (1995).
Wagstaff Piling, Australia as a low noise, low Standards Australia.
vibration, cast-in-situ, displacement pile that CAPWAP@ Manual (1996). Pile Dynamics Inc.
produces limited spoil. The Campile was developed Cleveland, Ohio.
as part of an ongoing research programme to provide Poulos, H.G. and Davis, E.H. 1980. Pile
piling solutions for environmentally sensitive sites, foundation analysis and design. John Wiley and
particularly inner city developments where noise and Sons.
vibration considerations are paramount.
The test pile programme for the Campile installed
for this case study showed that dynamic testing
methods could be used to predict the behaviour of
the Campile under static loading conditions.
The ‘Class A’ CAPWAP@analysis provided a
conservative prediction of the pile behaviour. A
‘revised’ CAPWAP@ completed after the static
results were known showed excellent agreement
between the dynamic and static methods. It was
noted that the revised analysis provided a
significantly larger shaft resistance than the initial
‘Class A‘ analysis and adopted significantly lower
shaft damping and quake parameters.
Comparison of the shaft parameters obtained from
the CAPWAP@analyses with calculations from the
borehole information for the site indicated that the
Campile was performing as a displacement pile. That
is, the shaft parameters obtained from the test
programme on the Campile were similar to those
expected for a driven, displacement pile.
It was also noted that the Campile installed for this
case study was a development stage of the pile type
and was installed as a straight-shafted pile. Further
advances in the pile installation process have since
been made with the Campile which allows for various
sizes and profiles to be constructed, including a
‘screw-shaped’ pile which provides increased shaft
resistance and greater overall pile capacity.
During the installation of the Campile, torque
measurements were also recorded but have not been
included in this paper. However, it is noted that the
torque readings show good correlation with the CPT
data collected for the site.

71a
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 1503

LT the final word? Correlation between DLT and SLT

H.Goldemberg & J. J.Goldemberg

Geotecnica Cientec, Buenos A i m , Argentina

ABSTRACT: Due to the increase in service loads of foundations, a world wide trend is being followed by
geotechnical engineers and designers to measure the real behaviour of the pile-soil interaction. Historically,
this was defined by the Load-Displacement Curve, obtained from a Static Load Test, and the concept of
bearing capacity.. .if those involved in the project arrived to an agreement on its definition. Technology de-
veloped and the concept of Dynamic Load Test (DLT) arrived, bringing with it a much economic and faster
way of measuring the response of piles to imposed loads. But is it this what the geotechnical engineer re-
quires?
A comparison between DLT and SLT is analysed throughout the eyes of a geotechnical engineer and not
from the perspective of a testing house. Different types of piles were tested with both methods, seeking for
correlation, side effects and installation influence in results, aiming to know the real performance vs. predic-
tions.

1 INTRODUCTION It is the objective of this Paper to present correla-

In the first half of the ‘SOS, it was introduced in Ar-
gentina the pile testing speciality based on the Stress
Wave Theory; at the beginning throughout the Sonic
Integrity Test (SIT), as a parameter for control qual-
ity of foundations and, afterwards, with the Dynamic
Load Test (DLT) in order to measure the pile-soil
behaviour.
But up to then, how was determined the bearing
capacity of a piled foundation? The answer is sim-
ple, in the same way than in the rest of the world,
that is, with a Static Load Test (SLT). As we all
know, in spite the familiarity SLT has within the en-
gineering community.. .that, if we first agree on
what type of SLT, due that the ASTM D 1143-81
Standard has five different methodologies.. .it is
time and money consuming.
During those years, where Quality Assurance
(QA) started to play an important role in the Argen-
tinean construction market, as well as the increment
in service loads, leaded to an increase in the demand
for verifying piling works. As well, a reactivation of
the local economy, the availability of new technol-
ogy in construction methods and tighter schedules
for finishing site works, allowed the DLT to be in-
troduced as an alternative to the cumbersome static
test.
tions between Dynamic and Static Load Tests per-
formed on the same pile, including all the informa-
tion that the Geotechnical Consultant had at the
moment of analysing the results and highlighting
how it was arrived to those conclusions. It is the
moment to mention that in all cases the Dynamic
Load Tests were performed using the FPDS-3
equipment, developed by TNO Building & Con-
struction Research (the Netherlands), while the SLTs
were conducted under ASTM Standard. When the
Static test was carried out, the transfer load structure
assembled consisted in a beam and reaction piles an-
chored well below the influence area of the test pile.
2 CASE HISTORIES.
2.1 Buildings,
The site was a group of buildings constructed in the
City of Buenos Aires, Argentina.
The ground conditions are described in the fol-
lowing geotechnical profile (Figure l), and the pile
type was of substitution, reinforced concrete and
casted in-situ under bentonite mud (pile characteris-
tics are presented on Table 1).

719
Table 1. Pile characteristics. ‘slow method’ and then 7% of the foundation popu-
Pile Bentonitemud Diameter Length lation (29 piles) were going to be tested with the
Type (m) (m) Dynamic Load Test (the tests information is pre-
I J 0.35 14.0
0.40 15.0
sented on Table 2)
I1 J
111 J 0.50 15.0 Conclusions: The results of the homologated tests
are presented on Figure 2, where it is clearly seen
the good correlation between DLT and SLT. The
quality of this match depends dramatically on the
quality of the geothecnical investigation, the knowl-
edge of the mechanical properties of the concrete
and the concrete consumption, in order to check
variations in the cross section of the piles.

2.2 Industrial.facility.
The site consisted in the expansion of an industrial
facility in the South Area of Gran Buenos Aires,
where 250 precast reinforced concrete piles were
going to be driven with a diesel hammer. Pile data is
available on Table 3 .
In order to determine the pile length in different
areas of the building site, Pile Driving Analysis
(PDA) was performed, monitoring Driving Resis-
tances, Enthru Energy, Stresses, etc. The 100% of
the pile population was tested with SIT searching for
cracked piles, 10% of the foundation (20 piles) was
tested dynamically and only 1 ‘slow’ SLT was per-
formed to have a correlation between both methods.
The summary of these investigations is presented on
Table 4.

Figure 1. Geotechnical profile.

Table 2. Test programme*

Test Pile Quantity Service Max. Test Maximum
Type Type Load Load Settlement
(MN) (MN) (mm)
DLT I 10 0.6 1.4 2.6
I1 10 0.7 1.9 3 .O
I11 9 0.9 2.4 3.9
SLT I 1 0.6 1.2 2.2
I1 1 0.7 1.4 2.1
Ill 1 0.9 1.8 2.6 Figure 2. DLT and SLT load-settlement curves.
Note *: information is presented only for those piles where
DLT and SLT were performed.
Table 3. Pile characteristics.
Pile Cross section Length
Type (mxm) ( 4
The foundation project consisted in 408 piles Concrete precast 0.4x0.4 19.0
placed under caps in groups of one, two or three
units each. All the installed piles had to be con-
trolled through the Sonic Integrity Test (SIT) in or- Table 4. Test programme.
der to verify that they were fiee of damages and to Test Quantity Service Max. Test Maximum
Type Load Load Settlement
select which ones were to be subject for DLT and (MN) (MN) (mm)
SLT. PDA 40
It was decided, by the Engineer, that the first DLT 20 1.2 2.9 4.8
three piles were going to be tested statically with the SI.T 1 1.2 2.4 3.8

720
 It is interesting to discuss what happened with
this correlation. Due to the unreal short period of
time allowed to finish all the driving works, the En-
gineer decided to perform the Dynamic Load Test
immediately after driving- although GEOTECNICA
CIENTEC intended to persuade him not to do so be-
cause a less total resistance was going to be obtained
due to the influence of pore pressure.
The geotechnical information, described on Fig-
ure 3, shows the stratigraphic profile throughout
SPT and CPT tests. Because discrepancies were de-
tected during the original borings in different areas
of the site, Dutch Cone was used. It was also very
useful the Pile Driving Prediction (PDP), performed
before the first pile was driven in order to select the
most suitable diesel hammer.
On Figure 4, the results of the load tests per-
formed on pile No 135 are plotted, where the first
DLT carried out did not fidfilled the expectations-
due to what was mentioned above. After the results
were available, the Engineer decided to perform the
SLT two weeks later, having in mind that the test
could not be taken to failure because it was a service
pile. Five days after finishing the Static test, the pil-
ing rig was used to redrive pile No 135 to monitor its
Dynamic behaviour.
As it can be clearly observed, the differences
between the first DLT and the SLT as well as among
the Dynamic ones was caused because it was not
taken into account that in cohesive saturated soils,
the driving effects on remoulding soil layers and the
slow dissipating pore pressure cause a reduction of
the bearing capacity.
Figure 3. Geotechnical profile.
721
Conclusions: Dynamic Load Tests are affected by
local geotechnical characteristics as well as by soil
conditions at the moment of carrying out the meas-
urements. This is the reason why the frst DLT did
not provide reliable information, because the be-
haviour of the pile-soil interaction during testing was
different than that at the design stage.
It is the authors’ opinion that in order to perform
a better geotechnical design of piles, more efficient
and foundation works less expensive, it must be to
carry out CPTs tests during the soil investigation
programme. In this way better Signal Matches will
be obtained between calculated and measured Upper
Travelling Waves, because more exact soil models
can be evaluated.
Figure 4. DLT and SLT load-settlement curves.
2.3 Bridge.
In the Province of Santa Fe, Argentina, due to yearly
over floods of the Parana River- caused mainly by
heavy rains during the rainy season, many bridges in
the area of the lowlands had to be rebuilt and some
others expanded.
In this case, the project consisted in raising the
level of the road as well as increasing the length of
the bridge fi-om one span 10m long to a deck of two
spans 15m each.
The geotechnical design for the foundation was
based on the soil strata illustrated on Figure 5 ,
adopting bored piles casted under bentonite mud.
The pile details are included on Table 5.
The North, South and Intermediate piers had
three piles each, which were tested with SIT, aiming
to detect defects such us cracks, soil inclusions,
contaminated concrete and their real lengths. The
testing programme carried out is shown on Table 6.
Table 5. Pile characteristics.
Pile Bentonite mud Diameter Length
Type (m) (m>
Bored J 1.20 21.5
 Figure 6. DLT and SLT load-settlement curves.

Due to the local geological conditions, mainly

sandy soil, a ‘quick’ Static Load Test was going to
be performed because the creep effect was not criti-
cal for the bridge behaviour. The results are plotted
on Figure 6 jointly with its dynamic counterpart.
Conclusions: As it is clearly presented, the Dy-
namic Load Test can have an accurate static correla-
tion, inclusive for high loads- 7MN as in this case,
As in the previous cases, of the building and the
industrial facility projects, it is an irrevocable condi-
tion - in order to obtain a good match or correlation
between dynamic and static measurements of a same
phenomenon pile-soil interaction - to have an ex-
haustive geotechnical investigation, knowledge of
the parameters of the materials used for the con-
struction on the foundation and the pile shape. Al-
Figure 5. Geotechnical profile. though some skepticism, that still over flies on some
Engineer’s minds, Dynamic Load Testing is reliable
Table 6. Test programme. if the above conditions are fulfilled, inclusive for
Test Quantity Service Max. Test Maximum high loads in cast in-situ piles.
Type Load Load Settlement
(MN) (MN) (mm>
DLT 1* 4.0 6.5 7.3 3 GENERAL CONCLUSIONS.
SLT 1 4.0 7.2 8.6
Note *: Originally 1 SLT and 1 DLT were projected, but as one
pile was damaged it was agreed to perform a second DLT. In the above case histories, where piles varied in di-
mensions and construction methods- fi-om driven
precast concrete to bored piles under slurry,
From the interpretation of the Sonic Integrity throughout Franki type; installed in clay, lime and
Tests signals, it was concluded that pile NO4 had an sand affecting the pile-soil interaction in different
impedance reduction. ways under small to big loads, they all have a com-
Afier a detailed analysis using the Signal Match- mo n den0 minator.
ing technique with TNOWAVE, it was calculated That common denominator is, leaving aside the
that the abnormality consisted in a reduction of cross fact that all tests were properly performed, recorded
section from 1.20m of nominal diameter to 1.12m and post-processed, the quality of the basic inforrna-
extended over a length of 0.5m and at a depth of tion. In other words, the geotechnical information,
14.6m. With this result, it was decided to go further dimensions and pile shape, mechanical properties of
on in the investigations, in order to determine the materials used and the installation procedure.
real behaviour of the pile-soil interaction and meas- For both types of test, Dynamic (DLT) and Static
ure if the foundation element was suitable of trans- (SLT). that information is vital. Notable contradic-
ferring loads to the surrounding stratigraphy. A Dy- tion, because it is required for both designing the test
namic and a Static Load Tests were going to be and interpretation of results. It is useless to obtain
conducted on pile N04.
722
huge amount of data, in white paper in case of a SLT
or a computer file for a DLT, in order to be plotted
as the ‘Load-Settlement Curve’ and handed in to the
Client as a great achievement certifying that ‘the pile
will stand the load’ without the geotechnical-
structural diagnosis. In other words, the interpreta-
tion!
If not, effort, time and money would have been
wrongly spent.. .three not rehndable goods in any
aspect of life.
But making an abstraction and travelling to the
Contractors’ Paradise, where no testing is required
and QA has not been invented- not even as a con-
cept. Nevertheless, would not be necessary to know
the geotechnical investigation, material properties
and selected pile in order to construct the project?
In almost all the above case histories, a good cor-
relation between Dynamic and Static Load Tests was
obtained, satisfying the expectations for homology
of both methods. It is a fact the acceptance for DLT
fiom the local market and its growing reliability, due
that it provides similar results than the Static Test
with almost no delays on site and reduced budget.

4 REFERENCES.

ASTM, D 1143-81 (Reapproved 1994). “Standard

Test Method for Piles Under Static Axial Com-
pressive Load”. ASTM
ASTM, D 4945-89. “Standard Test Method for
High-Strain Dynamic Testing of Piles”. ASTM
C. R. Mullins (1 992). ‘‘Large diameter pile test proj-
ect”. Proceedings of the Conference on Recent
Larfe Scale Fully Instrumented Piles in Clay,
London, June 1992
Carol1 L. Crowther (1988). “Load Testing of Deep
Foundations”.
German Society for Static and Dynamic Pile Tests
(1997). “Recommendations for Static and Dy-
namic Pile Tests”

723
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Results of an international pile dynamic testing prediction event

A. Holeyman
Universitk Catolique de Louvain, Louvain-la-Neuve, Belgium
J. Maertens
Katholieke UniversiteitLeuven, Leuven, Belgium
N.Huybrechts & C. Legrand
Belgian Building Research Institute, Brussels, Belgium

ABSTRACT: An international prediction event was carried out within the framework of a 30-pile testing pro-
gram organized in Belgium. That program called upon several testing methods: static load tests, Statnamic
testing, and dynamic testing. This paper provides a summary of received predictions and results obtained from
the static pile load tests, which were carried up to failure of the instrumented piles. The comparison between
predictions is made using load-settlement curves with reference to the results of the static load tests. A com-
panion paper reports on the project background information that was required to prepare the prediction event,
including a description of the pile types and results of an extensive soil investigation program.

1 INTRODUCTION al, 1999b). Predictions could be established based on

1.1 Program Background
A national research project has been organized by
the Belgian Building Research Institute in order to
establish the performance of different types of cast-
in-place ground displacement screwed piles. The
program included the installation and testing of 30
test piles that allowed the organization of a predic-
tion event. That prediction effort was undertaken
with the hope to document the profession's ability to
estimate these new piles behavior based on standard
investigation means as well on dynamic testing.
Six different types of ground displacement piles
were installed (five of each) and tested: one prefab
and five cast-in-place screwed types: Atlas, De
Waal, Fundex, Olivier, and Omega. An extensive
soil investigation was performed as part of the re-
search project, including in situ tests (CPT, PMT,
SPT, DMT, SASW) and laboratory tests on undis-
turbed samples.
1.2 Prediction preparation
A project synopsis had been prepared to invite inter-
ested parties to make those predictions (Holeyman et
al, 1999a). It included a description of the pile
types, site characterization, the static load test proce-
dure, the dynamic load test procedure and the format
of the prediction.
Interested predictors were sent the complete in-
formation, available as laboratory and in situ inves-
tigation and dynamic load test results (Holeyman et
the dynamic load tests, the geotechnical investiga-
tion, experience, or a combination of the above.
The piles types, layout, and the site investigation
are described in a companion paper (Holeyman et al,
2000). The present paper focuses on load tests re-
sults and their comparison with results of the static
load tests.
2 DYNAMIC LOAD TESTS
2.1 Procedure for dynamic load tests
The loading device used to impact the 12 piles in-
stalled for that purpose was a 4 tons drop hammer
operated by a crane. A sequence of several blows
was applied to each pile. The drop height sequence
most often applied was as follows: 0.40m, 0.80ni,
1.2m, 0.8m, and 1.2 m.
Dynamic measurements of strain and acceleration
were acquired for all 12 piles using a TNO FPDSS
system. In addition 6 piles were also monitored us-
ing a PDI PDA-PAK system (Prefab, Fundex, and
De Waal piles) and one Omega pile was monitored
using a PDI PAL system. A 0.4m-diameter head was
cast on July 6'' on top of the 10 cast-in-place piles.
The transducers were attached generally 0.8 m from
the top of the approximately 1.5 m high head. Dis-
placements were acquired using a laser system.
2.2 Distribution of dynamic load test measurements
The results of dynamic load test measurements (pile
head force, velocity and displacement) were made

725
Figure l a : Long Piles Load Settlement Curves - Prefab, Fundex and De Waal

726
Figure 1b : Long Piles Load Settlement Curves - Olivier, Omega and Atlas

727
available under a digital format to parties that had
expressed an interest to make a prediction on that
basis. Also characteristics needed to interpret the
measurements were provided in part in Table 1 of
the companion paper (in particular dimensions and
properties of the pile heads extension) and in part in
the files containing the measurements. Additional
characteristics (wave propagation speed, density,
pile impedance, etc.) needed to further analyze the
measurements were distributed together with the
measurements.
Interested parties obtained the digital files of the
events by e-mail, which required the structuring of a
vast amount of information, totaling more than 8
Mbytes of digital records. For each pile type a di-
rectory was established (for example the Prefab di-
rectory), containing subdirectories according for
each pile of that specific pile type. The ‘pile number’
subdirectory (for example subdirectory pile A7 in
the directory Prefab) contained a word file: (e.g.
‘A7info.doc’) and was further subdivided into the
following subdirectories: TNO files, ASCII files,
PDA files, and Displacement files. The word
file ‘A7info.doc’ gave supplementary information
about the dynamic load test on the pile A7 (the blow
numbers, the drop height, and field notes). It was
thus possible for the predictors to reprocess the raw
signals using adjusted pile parameters, number of
samples, etc.
2.3 Meas urements nomin a I interpretat ion
The choice of the relevant moduli and sections is
often considered as part of the predictors’ art and was
purposely left open to some degree, as is usually the
case for cast-in place piles.
It was emphasized that all files had been uni-
formly acquired using a nominal modulus of ap-
proximately 40,000 MPa and a nominal wave speed
of 4,000 m/s at the measurement section (i.e. in the
concrete of the cast head). The pile heads were cyl-
inders with a diameter of 0.4 m, except for the Pre-
fab pile where the current 0.35x0.35m section pre-
vailed. It was the predictor’s responsibility to assess
the measurement section modulus adequate for his
prediction. It was also emphasized that the concrete
of the tested pile below the added head had a differ-
ent modulus and a different section. It was also the
predictor’s responsibility to assess the appropriate
section and modulus for the shaft.
Peripheral information allowing the predictor to
perform that important assessment included Table 1
of the companion paper, the digital signal them-
selves (e.g. impedance match or 2L/c check), integ-
rity tests of the piles. Strength and ultrasonic wave
speed measurements on concrete samples cast at the
time of installation of the piles and concreting of the
pile heads were made available finally. Low strain
testing had been performed on all piles both the
728
BBRI and CEBTP and results provided to predic-
tors.
3 RECIEVED PREDICTIONS
3.1 Reporting format
The predictions are reported herein under an anony-
mous format in order not to stigmatize those with
less accurate predictions. Each prediction is however
labeled with a code corresponding to the prediction
type. Each predictor is thus enabled to position his
own prediction within the cluster of results and en-
couraged publishing his prediction procedure, using
the present paper as a reference.
3.2 Predictions types
Results from 10 predictors had been received on
November S’’ 1999, the ultimate submittal date.
Contractors had also predicted the ultimate bearing
capacity of their own piles (they were not asked to
supply the load-settlement curves.)
According to the reference data used to cast those
predictions, the following labels have been used:
- “CPT” for predictors using the CPT results.
- “PMT” for predictors using the PMT results.
- “LAB” for predictors using the laboratory results.
- “DLT” for predictors using the Dynamic Load Test
results.
- “STN” for predictors using the Statnamic Test re-
sults.
The CPT predictors used different methods, in-
cluding ultimate state design as well as load transfer
curves. All the contractors’ predictions were made
using CPT results and De Beer’s 1974 method. The
PMT predictors used the pressiometric approach that
provides stress-displacement relationships for the
shaft and the base. The LAB predictor used a load-
transfer functions method based on plasticity indi-
ces.
The DLT predictors’ methods included either
CAPWAP or SIMBAT: the soil parameters in a
model are adjusted to get the best match between the
measured and the predicted signals of a Dynamic
Load Test. SIMBAT is an empirical method con-
verting the dynamic reaction to a static reaction.
The STN predictor used the Unloading Point
Method (UPM) to predict the static load test. It was
mentioned by the predictor that, due to strain rate
sensitivity of clayey soils, a 30% reduction coeff-
cient had to be applied on the usual UPM method. A
hyperbolic approximation of that reduced function
was then calculated. This is the reason why those
predictions are labeled as “0.7 STN”.
It should be noted that the STN predictor was not
provided with the results of the dynamic load tests,
and that no TNO-WAVE prediction was submitted.
3.3 Predictions classes
Predictions had to be made before static pile load
tests were performed in order to qualify as Class A
type predictions. If predictions were made after the
static pile load tests, they qualified as C type predic-
tions, according to accepted definitions of predic-
tions classes (Lambe, 1973).
Each prediction for each pile can be classified ac-
cording to its submittal date relative to the date of
static loading. Table 1 shows that most of the pre-
dictions are Class C. The only Class A predictors
were CPT 1 (except for piles A1 & A4) and Con-
tractors Atlas and Fundex. Other Class A predictions
were those of predictor DLT2 for A3 pile and of
predictor PMTl for C2 to C4 piles.
4 STATIC LOAD TESTS
4.1 Procedurefor static load tests
The static pile load tests were to comply with the
following loading guidelines, referring to Q,,,, the
maximum anticipated test load, chosen with the hope
to cause bearing failure:
- A pre-load stage of maximum 5% of Q,,, was
applied in order to check the measurement
equipment and the centricity of the applied force,
- 10 maintained load steps with equal AQ until Q
reaches Q,,,
- No intermediate unloading cycles
- Duration of maintained load step of 60 minutes
- Load test was performed until a pile head settle-
ment 2 15% Obase was reached
- When the pile head settlement has reached a
value of 25 mm, subsequent load steps can be
applied using a smaller increment (AQ/2), in or-
der to refine the pile load-settlement curve as it
approaches failure,
- Unloading in 5 steps of 10 min. each, except for
final unloading (30 min at least of monitoring).
The system used to apply the maintained loads on
the piles called upon a sophisticated hydraulic regu-
lation that guaranteed a tolerance of 5 kN. That sys-
tem had just been developed by the BBRI. The 3
MN reaction was provided by a kentledge consisting
of concrete blocks. Besides load and settlement
monitoring, extensometers provided longitudinal
strains along 5 to 7 shaft segments along the pile
length. The results provided by those more detailed
measurements are to be reported elsewhere.
Such a procedure requires a value for the ultimate
capacity R, of each pile. Those capacities were esti-
mated by the BBRI and the national experts using
De Beer’s method based on the CPT tests results (De
Beer, 1974). The load increments A Q were actually:
- R,,/8 for Atlas, Fundex, Prefab and Olivier piles.
- RJ0 for De Waal and Omega piles.
729
The ultimate capacity was considered reached
when the pile head settlement was equal to 10%
Obase.It should be noted that a maximum Constant
Rate of Penetration ( C W ) of 0.6 m d m i n was en-
forced towards the end of the loading procedure for
all piles (except for piles A1 and A4).
4.2 Results
Figures 1 and 2 show the various load (Q) - settle-
ment (s) curves for the long and the short piles, re-
spectively.
“SLT” refers to the Static Loading Test. The pre-
dictors’ curves are also identified using the labels
discussed in Section 2.2. “Contractor” refers to the
ultimate capacity predicted by the Contractor. This
value is drawn for 30mm < s < 50mm with a bold
line. “Target SLT” refers to the ultimate capacity
estimated by the BBRI. It is a “box” corresponding
to: - 8AQ<Q<lOAQ
- 9.75 % a b a s e s < 10.25 % OZlbase
5 DISCUSSION
The load-settlement curves resulting from the static
load tests show a good proportionality between load
and settlement up to 5 mm settlement. Beyond that
point, the curves deviate from their initial linear
trend. After evidencing a peak resistance, the pile
settles under a slowly decreasing load beyond set-
tlements exceeding 10 to 20 mm.
Pile C3 exhibits an unusually high peak, with the
load increasing very rapidly up to about 2300 kN
within a 6 to 20 mm settlement range. This was due
to a problem in the pressure regulation of the hy-
draulic jack. After reaching a settlement of 6 mni,
the pile started to settle at a rate of penetration of ap-
proximately 48 m d m i n . This explains that the peak
capacity of the SLT (about 2300 kN) is an overesti-
mate of the pile ultimate static bearing capacity. In
the absence of that regulation problem, the peak
would have been reached under a load of approxi-
mately 1500 kN. This test evidences the influence of
the rate of penetration on the assessment of piles ul-
timate bearing capacities.
PMT and LAB predictions are consistently on the
safe side of the SLT curve. PMTl and PMT2 curves
are very similar, which tends to show that the pressi-
onietric approach is consistently applied.
BBRI predictions (targets “windows”) were good
except for Piles B3, B4, C1 and C4 where the ulti-
mate capacity had been overestimated.
CPTl predictions were extremely accurate for,
A1 to A4 but overestimated the results for B and C
piles, probably because of an overestimated installa-
tion coefficient.
Concerning the DLT and STN predictions, it can
be observed that maximum transient pile displace-
ments rarely exceed 15 mm That might explain why
Figure 2a : Short Piles Load Settlement Curves - Prefab, Fundex and De Waal

730
Figure 2b : Short Piles Load Settlement Curves - Olivier, Omega and Atlas

731
Table 1. Predictions classes
Pile Long Long Long Long Long Long slim m Short slim Short Short
’refab ‘unde: )e Wai Olivier 3megz Atlas Prefab -unde: )e Wa; Olivier 3megi Atlas
Static Test Date 3018 1519 1719 2319 5110 12110 2l9 719 1519 2119 3019 7110

Submittal Date
-- ___I _s_ -- -- -- - -
5th Sept 99 C A A A A A C A A A A A
9th Sept 99 C C
10th Sept 99 C A C C
1st Oct 99 C C
4th Oct 99 C C A A C C C A
27th Oct 99 C C C C C C C C C C C C
28th Oct 99 C C C C C C C C C C C C
29th Oct 99 C C C C
29th Oct 99 C C C C
5th Nov 99 C C C C C C C C C C C C
5th Nov 99 C C C C C C

‘Contractor
- __I -- -- - _I - __.__ -
Fundex 12th Aug 99 A A
Franki 26th Aug 99 A A
Socofonda 3rd Nov 99 C C
Olivier 4th Nov 99 C C
De Waal 5th Nov 99
-
C
--
C C C
-

those methods encounter more difficulties in pre-
dicting pile behavior under large displacements.
The predictions (DLT1 to 4 and 0.7 STN) are quite
good within the service load range of the load-
settlement curves. DLTl and DLT 3 are very close
to the SLT ultimate capacities even though they
did not predict the pile behavior for settlements
greater than 20 mm. DLT2 (Capwap method),
DLT4 (Simbat) and 0.7 STN overestimates of the
ultimate capacities would warrant the following
approximate reductions:
- DLT2by25%,
- DLT4by50%.
- 0.7 STN by 25 % (which means the 30 % re-
duction coefficient initially taken by that pre-
dictor should have been 50 %).
If such reductions were applied to these predic-
tions, they would however not fit as well the initial
part of the SLT curves. The reduction of dynamic
soil resistance to its static value still needs to be
clarifi ed .
6 ACKNOWLEDGMENTS
The authors would like to acknowledge the Bel-
gian Ministry of Economic Affairs for its financial
support of the testing program (Convention CC
CIF - 562). The ‘De Nayer’ Institute at Sint-
Katelijne-Waver made the test site available. Con-
tractors installed their own piles free of charge.
Profound performed the Statnamic tests and
CEBTP the integrity tests also free of charge. Dr.
Klingmuller, of GSP, Mannheim loaned a PDA
while TNO provided an FPDS5 system free of
charge. Bachir Dali, Paris, using an in-house laser
system he developed, acquired displacements on a
voluntary basis. The authors would like to thank
Mr. J. -M. Couvreur, Research Assistant at UCL,
for his management of communication with the
predictors and of the information, including prepa-
ration of comparative figures. Finally this exercise
would not have been possible without countless
hours accrued amongst the several prediction team
that bravely took up the challenge. Many thanks to
those heroes; some of who will hopefully identify
themselves through follow-up publications.
7 REFERENCES
Lambe T.W. (1973), Predictions in soil engineering,
Giotechnique XXIII, n02, p 149-202.
Holeyman, A. et al., Design of axially loaded piles - Bel-
gian practice in Design of axially loaded piles - European
Practice, edited by De Cock and Legrand, Balkema, Rotter-
dam, 1997, pp.57-82
Holeyman et a1 (1999a), Prediction Invitation, Belgian
Research Project “Ground Displacement Screwed Piles at
Sint-Katelijne Waver”, August 20‘” 1999.
Holeyman et a1 (1999b), Additional Reference Document
for Predictors, Belgian Research Project “Ground Displace-
ment Screwed Piles at Sint-Katelijne Waver”, August 29“’
1999.
Holeyman et a1 (2000), Preparation of an International
Pile Dynamic Testing Prediction Event, Proceedings of the
VIth SWC, Sao Paulo.

732
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Preparation of an international pile dynamic testing prediction event

A. Holeyman
Universiti Catholique de Louvain, Louvain-la-Neuve, Belgium
J. Maertens
Kutholieke Unviversiteit Leuven, Leuven, Belgium
N.Huybrechts & C. Legrand
Belgian Building Research Institute, Brussels, Belgium

ABSTRACT: An international prediction event could be carried out on the basis of an extensive pile testing
program organized in Belgium. Five units of six pile types were installed, allowing various testing methods to
be applied: instrumented static load tests carried up to failure, Statnamic testing, and dynamic testing. This
paper provides the project background information that was required to prepare the prediction event. The pile
types are fully described while the results of the extensive soil investigation program are summarized. A com-
panion paper reports on the received predictions and results obtained from the static pile load tests.

1 INTRODUCTION 2 PILETYPES

A national research project promoted by the Belgian
Building Research Institute (BBRI) has been con-
ducted in order to establish the performance of dif-
ferent types of cast-in-place ground displacement
screwed piles. A national advisory committee under
chairmanship of the first two authors directed the
program, which included installation and testing of
30 test piles. The program has also been followed by
an international panel of experts selected amongst
members of ITC 18 (International Committee on
Pile Foundations, of the International Society of Soil
Mechanics and Geotechnical Engineering).
A total of 30 loading tests could be performed at a
site located in Sint-Katelijne-Waver, some 20 km
north of Brussels, according to the following sched-
ule:
- 6 Statnamic tests and 12 dynamic tests took place
within the first and second week of August 1999;
- 12 Static pile tests were performed between Sep-
tember Znd and October 12* 1999.
That timing allowed the organization of a Class-A
type prediction event (Lambe, 1973), with the view
to document the profession’s ability to estimate
those new piles behavior based on standard investi-
gation means as well on dynamic testing.
An extensive soil investigation was performed as
part of the research project, including in situ tests
(CPT, PMT, SPT, DMT, SASW, ...) and laboratory
tests on undisturbed samples. Whereas the present
paper provides a description of the pile types and
subsurface conditions, its companion (Holeyman et
al, 2000) focuses on load tests results and their com-
parison.
Six different types of ground displacement piles
were installed and tested: one prefab and five cast-
in-place screwed types:
- Atlas pile, installed by Franki Co.
- De Waal pile, installed by De Waal CO
- Fundex pile, installed by Fundex Co.
- Olivier pile, installed by Olivier Co.
- Omega pile, installed by Socofonda Co.
Figures 1 through 5 illustrate the installation pro-
cess of the five ground displacement screwed piles
and define the pile toe level relative to the geometry
of the augedscrew tip.
Five piles of each type have been installed on the
test site to accommodate the following conditions
for each pile type:
- One short pile for static load testing
- One long pile for static load testing
- One short pile for dynamic load testing
- One long pile for dynamic load testing
- One long pile for Statnamic testing
The short and long piles had an approximate
depth of 7.5 m and 1 1.7 m, respectively.
The different pile types, their testing destination,
their nominal shaft and base diameters for geotech-
nical bearing capacity calculations, and their meas-
ured pile base depths are listed in Table 1.
A total number of 30 piles were thus installed ac-
cording to the pile layout shown on figure 6. The
following load tests were to be performed on the
following piles referenced according to their grid
line locations :

733
- 12 static load tests on piles A l , A2, A3, A4, B1, - 6 Statnamic load tests on piles D1, D2, D3, D4,
B2, B3, B4, C1, C2, C3, and C4 D5, and D6.
- 12 dynamic load tests on piles A5, A6, A7, A8,
B5, B6, B7, B8, C5, C6, C7, and C8; and

734
 tests and laboratory tests, the locations of which are
shown on figure 7:
In-situ tests
- 30 CPT(Cone Penetration Test)-E with electric
cone in the axis of each test pile
- 27 CPT-M1 with mechanical M1 cone (standard
discontinuous penetration)
- 3 CPT-Ml with mechanical M1 cone (alternate
continuous penetration)
- 4 CPT-M4 with mechanical M4 cone
- 4 DMT (Dilatometer test)
- 2 borings with PMT (Pressuremeter test) tests at
1 m intervals
- 2 borings with SPT (Standard Penetration Test)
1 2 3 4 5 tests at 1.5 m intervals
1. Setting up rig. - 1 boring for undisturbed soil samples
2. Screwing in displacement auger tip, optimally with ver- - SASW (Spectral Analysis of Surface Waves)
tical pressure. tests
3. Bringing in reinforcement (eventually after concreting). - Seismic Refraction Tests
4. Injecting concrete and pulling out auger head, while - Seismic Cone tests
pursuing clockwise rotation. Lost point at pile base. Laboratorv tests at several depths:
5 . Finished pile. - Grain size distribution
Toe level determined by the level of the top lost bottom - Atterberg Limits
point - CU - Triaxial tests (consolidated, undrained)
Figure 5. - Installation process of Omega Pile - UU - Triaxial tests (unconsolidated, undrained)
- Triaxial tests with Bender Elements

3.2 Subsurface geology and properties

3 SOIL INVESTIGATION
Borings B1, SPTl and SPT2 revealed the following
3.1 Overview succession of soil layers :
The extensive soil investigation performed as part 0 - 0.40 m : rubble
of the research project included the following in situ 0.40 - 0.65 m : Quaternary loamy sand
0.65 - 13.90 m : Tertiary O.C.Boom Clay

Figure 6 - Test Piles Layout

735
Table 1 - Features of installed piles
Pile TEST Pile type - Nominal Shaft Base Pile base Excavation End level Top level
dimensions [cm] Diameter ( I ) Diameter depth ( I ) level ( 2 ) pile head (’) pile head ( 2 )
( 4 (m) (4’ [ml [ml [ml
A1 Static Prefab. 35x35 0.395 0.395 -7.39 -1.04 -0.84 +0.22
A2 Static Fundex 38/45 0.380 0.450 -7.38 -1.07 -0.87 +0.27
A3 Static Fundex 38/45 0.380 0.450 -11.50 -1.02 -0.82 +0.29
A4 Static Prefab. 35x35 0.395 0.395 -11.58 -0.98 -0.78 +0.36
A5 Dynamic Fundex 3 8/45 0.380 0.450 -7.39 -0.8 1 -0.60 +0.9 1
A6 Dynamic Fundex 3 8/45 0.380 0.450 -11.56 -0.75 -0.53 +0.97
A7 Dynamic Prefab. 35x35 0.395 0.395 -11.63 -0.75 Continuous +1.37
A8 Dynamic Prefab. 35x35 0.395 0.395 -7.44 -0.71 Continuous +056
B1 Static De Waal4 1/4 1 0.410 0.410 -7.53 -1.07 -0.87 +0.25
B2 Static De Waal4 1141 0.410 0.410 -11.73 - 1.07 -0.87 +0.28
B3 Static Olivier 3615 1 0.510 0.510 -11.68 -0.97 -0.77 +0.31
B4 Static Olivier 36/5 1 0.510 0.510 -7.43 -0.90 -0.70 +0.39
B5 Dynamic De Waal4 1141 0.410 0.410 -7.48 -0.76 -0.58 +0.92
B6 Dynamic De Waal4 1141 0.4 10 0.410 -11.74 -0.76 -0.52 +0.94
B7 Dynamic Olivier 36/5 1 0.510 0.510 -11.77 -0.79 -0.46 +1.06
B8 Dynamic Olivier 3615 1 0.510 0.510 -7.90 -0.75 -0.58 +0.94
c1 Static Omega 4 1/4 1 0.410 0.410 -7.67 -1.08 -0.88 +0.32
c2 Static Omega 4 1141 0.410 0.410 -11.83 -1.02 -0.82 +0.35
c3 Static Atlas 36/5 1 0.510 0.510 -11.76 -0.95 -0.75 +0.37
c4 Static Atlas 3615 1 0.510 0.510 -7.72 -0.96 -0.76 +0.39
c5 Dynamic Omega 4 1/4 1 0.410 0.410 -7.53 -0.84 -0.57 +0.95
C6 Dynamic Omega 4 1141 0.410 0.410 -11.78 -0.84 -0.53 +0.96
c7 Dynamic Atlas 36/5 1 0.510 0.510 -11.61 -0.74 -0.35 +1.01
C8 Dynamic Atlas 3615 1 0.510 0.510 -7.68 -0.82 -0.48 +0.99
D1 Statnamic Fundex 3 8145 0.380 0.450 -1 1.54 -0.84 -0.66 +0.30
D2 Statnamic Prefab. 35x35 0.395 0.395 -11.67 -0.87 -0.57 1-0.38
03 Statnamic De Waal41/41 0.410 0.410 -11.58 -0.75 -0.49 +0.43
D4 Statnamic Olivier 36/5 1 0.510 0.510 -11.55 -0.80 -0.66 +0.38
D5 Statnamic Omega 4 1141 0.410 0.410 -11.71 -0.78 -0.52 +0.48
D6 Statnamic Atlas 3615 1 0.510 0.510 -11.68 -0.76 -0.48 +0.44
(1) Diameter governing soil failure along the shaft
(2) Measured pile base depth relative to original soil surface level and according to the definition of the pile base level
(figures 1 to 5 )

Figure 7 - Site Investigation Plan

@ CPT M1-disc. 0 CPT M4-dis.
0 Pile
DMT
* CPTM1-cont.
A SPTeBoring
X CPTE ++
SASW
PMT 0---o Seis. Refract

736
 Figure 11 - PMT profiles of creep and limit pres-
Figure 9 - Typical CPT-E Log (EB5 Location) sures and pressiometric modulus (PMT2 Location)

The properties of Boom clay, a stiff fissured and Liquid limit: wL= 84 %
stratified clay belonging to the Oligocene, are well Plastic Limit: w,=37%
documented in the vicinity (De Beer et al, 1977): Clay fraction 55 %
Natural water content: w = 22 to 3 1% Permeability 1O-l0 m / s

737
 UU triaxial tests performed for the research pro-
gram confirmed the local variability of the properties
resulting from the layered and fissured nature of
Boom clay: C, varied between 80 kPa at 4.7m depth
to approximately 150 kPa in the 8.5 to 13.9 m depth
range. Those results confirmed the trend established
by De Beer et a1 (1 9977) at the near-by Kontich site:
C,[kPa] = 8 4 + 6 . 5 z[m]
More discrepancy was found between the two
sites regarding effective strength parameters derived
from consolidated undrained triaxial tests conducted
with pore pressure measurements: 4' = 27" and
c'= 30 kPa for this testing program versus $, = 18"
and c'= 11 kPa at the Kontich site.

3.3 In sitii testing

The bulk of the investigation effort was directed to-
wards in situ geotechnical testing, which are gener-
ally recognized as the "ad hoc" testing for pile de-
sign in Belgium (Holeyman et al, 1997). The site
was investigated using the several testing tools
available to the profession, with a view to accom-
modate various geotechnical design cultures around
the world. The site is herein characterized from the
following angles:
CPT-M1, as shown in Fig. 8
CPT-E, as shown in Fig. 9
SPT N-values profiles, as shown in Fig. 10
PMT, as shown in Fig. 11
DMT, as shown in Fig. 12
SASW, as shown in Fig. 13

4 PREDICTION ORGANIZATION

4.1 Prediction preparation

Figure 13 - Typical SASW profile
(SASW D Trace)
According to geologists, Boom clay was covered,
prior to the Continental Pleistocene erosion, by a
layer of Neogene sand with a thickness of approxi-
mately 40 meters. That layer has been completely
eroded at Sint-Kathelijne-Waver.
On August 20th 1999, a reference document was dis-
tributed internationally among interested parties, and
in particular to members of ITC 18 (International
Committee on Pile Foundations of ISSMGE) and
APTLY (Association of Pile Testing Laboratory) to
enable them to:
- predict the load-bearing behavior of the piles based
on the results of the dynamic pile load tests, and
- predict the static ultimate pile bearing capacity and
the load-bearing behavior of the piles by means of
the ground investigation results.
A project synopsis had been prepared to invite
interested parties to make those predictions (Holey-
man et al, 1999a). It included a description of the
pile types, site characterization, the static load test
procedure, the dynamic load test procedure and the
format of the prediction.
Interested predictors were asked to fill in an invi-
tation document to accept the information release
conditions associated with this prediction event. In
particular any publication using part of the data
herein and public release of any of the research

730
measurements warranted the prior permission of the
BBRI. Once this was done, they were sent the com-
plete information, available as laboratory and in situ
investigation and dynamic load test results (Holey-
man et al, 1999b).
4.2 Prediction Format
It was requested that the prediction submittal in-
clude:
- A description of the used model(s), with a list of
governing parameters,
- The type of soil investigation method on which
the calculations were based,
- A detailed calculation methodology, with spe-
cific references (data provided, standards, publi-
cations, ...) and derivation of governing pa-
rameters,
- A separation between pile base resistance and
shaft resistance;
- A criterion for the ultimate pile bearing capacity
The predicted static load-settlement behavior of the
piles was to be summarized into a table providing
the loads corresponding to the following settlements:
1, 2, 4,6, 8, 10, 15, 20, 25, 30, 35, 40,45, 50, 60, 70,
80,90,100, 150, and 200 mm.
Predictions could be established based on dynamic
load tests, detailed geotechnical information (bor-
ing+lab tests, CPT, PMT, ...), experience, or a com-
bination of the above.
7 REFERENCES
De Beer, E., Lousberg, E., Wallays, M., Carpentier, R., De
Jaeger, J., & Paquay, J. 1977. Bearing capacity of displace-
ment piles in stiff fissured clays. I.R.S.I.A. - I.W.O.N.L.,
Comptes rendus de Recherches - Verslagen over Navorsingen,
No 39, Brussels
Lambe T.W. ( 1973), Predictions in soil engineering,
Geotechnique XXIII, n02, p 149-202.
Holeyman et a1 (1997) Belgian practice in Design of axially
loaded piles - European Practice, edited by De Cock and Le-
grand, Balkema, Rotterdam, 1997, pp.57-82
Holeyman et a1 (1999a), Prediction Invitation, Belgian Re-
search Project “Ground Displacement Screwed Piles at Sint-
Katelijne Waver”, August 20‘h 1999.
Holeyman et a1 (1 999b), Additional Reference Document
for Predictors, Belgian Research Project “Ground Displace-
ment Screwed Piles at Sint-Katelijne Waver”, August 29Ih
1999.
Holyeman et a1 (2000), Results of an International Pile Dy-
namic Testing Prediction Event, Proceedings of the VIth SWC,
Sao Paulo

5 CONCLUSION

The soil investigation performed as part of the

program was extensive and included enough ele-
ments to allow most geotechnical engineering cul-
tures to have a fair chance in the prediction event.
Pile types were varied. Although ground displace-
ment screwed piles are not widely known, the pro-
gram included a prefabricated concrete pile as a
more widely known reference pile. Conditions were
set to assess the bearing capacity of those pile types
and compare them with international predictions.

6 ACKNOWLEDGMENTS

The authors would like to acknowledge the Belgian

Ministry of Economic Affairs for its financial sup-
port of the testing program (Convention CC CIF -
562). The testing site was made available by the ‘De
Nayer’ Institute at Sint-Katelijne-Waver. Contractors
installation their own piles free of charge, is also
very much appreciated.

739
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 1503

Case studies of dynamic load testing in Japan

Yasushi W&ya - Kawasaki Steel Corporation, Tokyo,Japan
Kenji Nishiumi -Nippon Steel Corporatioiz,Tokyo,Japan
Masahiro Hayashi - NKK Corporation, Kawasaki, Japan
Atsushi Shibata - Kubota Corporation, Ichikuwa, Japan
Shinji Nishimura - Fugro Geoscience Incorporated, Tokyo,Japan
Tatsunori Matsumoto - Department of Civil Engineering, Kunazawa Universig, Japan

ABSTRACT: This paper reviews 69 cases of dynamic load testing performed in Japan on a commercial basis
since 1983, and points out some subjects, such as 'set-up' phenomena and hammer driving energy, to be
considered in the interpretation of the dynamic load test signals, in order to make the dynamic load test a more
useful tool in the design, quality assurance and driving control of piles. The Japanese Association for Steel Pile
Piles (JASPP) conducted their own dynamic load tests on a total of 13 steel pipe piles at different site
conditions, the results of which are compared with the static load test results. This paper also presents the
results of these comparative tests, emphasizing that adequate conditions for the dynamic load test are required
to derive accurate estimations of the static load-displacement curve, as well as ultimate pile capacity, from the
dynamic load test.

1 INTRODUCTION driving control of piles. This paper also presents the

The first dynamic pile load testing in the field in
Japan goes back to 1983, in which dynamic signals
were recorded for the monitoring of stresses in a pile
to prevent excessive impact. After that, research was
focused on the estimation of the total resistance of a
pile during driving from which only static pile
capacity was estimated empirically. Next, researchers
focused on the estimation of the static load-
displacement curve for a pile as well as the ultimate
pile capacity. The number of the dynamic load tests
performed in a year in Japan has been increasing
since the early 1990s.
Foundation design in Japan is currently changing
from the allowable stress design to the framework of
the limit states design and/or the performance based
design. In the new design methods, confirmation of
the performance of each constructed foundation
structure is of vital importance. Therefore, engineers
have to select an appropriate test method to confirm
the performance of the foundation, considering the
balance of accuracy and cost. The dynamic pile load
test is one of the promising test methods in this
respect, because of its quickness and the relatively
low cost of testing.
This paper reviews 69 cases of the dynamic load
test performed in Japan on a commercial basis since
1983, and points out some subjects, such as 'set-up'
phenomena and hammer driving energy, to be
considered in the interpretation of the dynamic load
test signals, in order to make the dynamic load test a
more useful tool in design, quality assurance, and
results of comparative dynamic and static load tests
on a total of 13 steel pipe piles at different site
conditions, emphasizing that adequate conditions for
the dynamic load test, such as proper hammer driving
energy and a sufficient time interval after the end of
initial pile driving, are required to derive an accurate
estimation of the static load-displacement curve as
well as the ultimate pile capacity from the dynamic
load test.
2 CHRONOLOGIY AND STATISTICS OF
DYNAMIC LOAD TESTS I
N JAPAN
The application of the dynamic load test at actual
construction sites on a commercial basis is increasing
in Japan. A chronological review and statistics
concerning the dynamic load tests are presented,
using the data measured by Fugro Geoscience Inc.,
Japan.
Figure 1 shows the chronological variation of the
number of test sites and the accumulated number of
the test sites in which dynamic load testing was
performed. In the first half of the 1980s, dynamic
signals were recorded for monitoring of stresses in
piles to prevent excessive impact. The use of the
dynamic load test rapidly increased after 1990, and
the total number of the test sites reached 69 in 1999,
with 6 sites per year in average.
Figure 2 shows the chronological variation of the
average number of test piles per site, indicating a
steady increase in the number of test piles per site

741
 after 1990 The large number found in 1997 is
associated with big projects involving highways and
power plants
Figure 3 shows a breakdown of the number of
piles with respect to type of tested pile, indicating a
large share of steel pipe piles, 89% of the total
number
Figure 4 shows a breakdown of the number of piles
with respect to type of hammer used Hydraulic
hammers occupy the largest share, 90% of the total
pile number Diesel hammers were not used after
1994, because of problems of noise, vibration and
pollution A merit of the use of hydraulic hammers is
that the driving energy of hydraulic hammers can be
controlled Figure 4 Breakdonn of pile number nith respect to the type
of haininer used

Table 1. Type of superstructure.

Location of
Super structure Nuinber test pile
Ratio
Onland Offshore
Pier 63 35% 3 60
Higliway bridge 19 27% 19 30
Quay wall 37 20% 15 22
Plant 16 9?4" 16 0
Building 12 7% 12 0
Breakwater etc 5 3% 3 2
Total 182 68 114
100% 37% 63%

Table 1 lists the number of superstructures with

respect to the type of superstructure. The number of
each type of superstructure is hrther divided
according to the location of the test pile, onland or
offshore. The number of offshore locations of the
test piles exceeds 60%. The dynamic load tests are
mainly conducted on piles supporting port and
harbor structures such as piers, revetments, and
highway bridges.
Figure 5 shows a breakdown of pile number with
respect to the diameter of piles. About 80% of the
piles have diameters greater than 1000mm, with a
share of 53% by piles having diameters from 1200 to
1400mm.
Figure 6 shows test (pile) number with respect to
pile length. The number of tested piles having
lengths from 40 to 50m is the largest. The maximum
pile length until now is 91.5m.
As mentioned already, most dynamic load tests
have been conducted on offshore steel pipe piles
having large diameters and large lengths. In such
cases, and execution of static load tests is difficult,
because it is considerably time and cost consuming.
On the contrary, the merits of the dynamic load test
become pronounced in such cases, because piles are
installed mainly by driving, and the noise problem is
Figure 3. Bre&down of pile number \&h respect to type of not SO Critical in offshore pile driving conditions.
pile.
742

Figure 7 Breakdown of the pile nuniber vith respect to the
soil t?pc of the pile toe bearing stratum
Figure 7 presents a breakdown of the pile number
with respect to the soil type of the pile toe bearing
stratum. Sand and gravel are the major soil types for
the bearing stratum. The piles driven in sand stratum
make up about half of the total number of piles, and
the piles driven in gravel occupy a quarter of the
total number of piles.
Figure 8 shows a breakdown of the pile number
with respect to the elapsed time at the re-driving test
after the end of initial pile driving. The piles tested 1
week or more after the end of initial pile driving
make up 73% of the total number of piles. Most
dynamic load tests (re-driving tests) could be
743
Figure 8. Breakdonm of the pile nuniber with respect to the
elapsed time at the dynamic load test after the end of initial
pile driving
conducted after a sufficient elapsed time, because
these dynamic load tests were performed at large-
scale construction projects, where testing long after
the initial pile driving was practical
Figure 9 shows the relationship between the
driving energy (ENTHRU), EH, transferred to the
pile head and the product of the total resistance, RI,
of pile during driving and set per blow, S The total
resistance, R,, was estimated by the well-known
CASE method Theoretically, R , + Sshould be less
than EH Most data satisfies this relation, although
several data violate this condition due to inaccurate
measurements of set per blow, S It may be seen that
the ratio, el = (R,.S)/EH, tends to increase,
approaching 1, as the set per plow, S, increases For
piles having S greater than lOmm, the ratio, er, is
near 1 and greater than 0 6 at the least In such
cases, it is thought that the soil resistance is fully
mobilized by the pile driving On the other hand, the
soil resistance does not seem to be hlly mobilized
for piles with small S resulting in small values of er
Small e,- would mean that a large proportion of the
transferred driving energy is transformed to
recoverable energy such as elastic deformations of
the pile and the surrounding soil and is not used for
the permanent penetration of the pile, because the
soil resistance is not fully mobilized
Figure 10 shows the relationship between the set
per blow and the ratio, R,IR,. The static pile
capacities, R,:, were estimated through the wave
matching analyses of the dynamic load test signals.
The RJRt ratio tends to decrease with increasing S
until S reaches about 10 mm, and then to level off
after that. As mentioned above, R, is not necessarily
the maximum total soil resistance against the pile
during driving in the cases of small values of S. In
such cases, R, may not reach the yield load of the pile
during driving. Hence, higher values of RJR, for
small values of S mean that the dynamic component
of the total soil resistance is relatively small due to
the elastic response of the soil,

Figurc 10 Thc rclationshlp bctvccn tlic scI pcr blon. S, and
tlic ratio U , I<,
whereas the dynamic component becomes larger if
the pile reaches the yield load during driving, and
slippage of the pile shaft occurs, resulting in high
values of S greater than about lOmm It may be said
that the static component ranges from 40 to 60% of
the total soil resistance if the pile reaches the yield
load
Based on the results of Figures 9 and 10, S greater
than lOmm or more is thought to be an indication of
the fit11 mobilization of the soil resistance, although
the authors are aware that it is only an approximate
value, because the data in Figure 9 were obtained
from the dynamic load tests on steel pipe piles having
various lengths, diameters driven in various bearing
strata. If the static load test results to be compared
with the dynamic load test results are not available at
the site, the relation shown in Figure 10 is useful in
practice to judge roughly during driving whether- the
pile has the required pile capacity or not
Figure 11 shows the relationship between the
pile length and the ‘set-up’ ratio. Here, the set-up
744
ratio is defined as the ratio of the static pile capacit),
R,, derived from re-driving test to R, derived fr-oni
the dynamic load test at the end of initial driving
The plots in Figure 11 were selected from the
dynamic load tests in which re-driving tests were
conducted after sufficient rest periods from the end
of initial driving for a full recovery of the pile
capacity The set-up ratio tends to increase with
increasing the pile length for pile lengths below about
40m, and level off after that length This result
would indicate that the set-up ratio for the shaft
resistance is larger than for the toe resistance The
set-up ratio attains 4 in piles having length less than
6Om, whereas the set-up ratio ranges from I to 2 for
piles having length over 6Om In actual piling sites,
re-driving tests are usually performed with the same
driving hammer as is used in the initial driving I n
usual, a minimuni hammer that is capable to
penetrate the pile to the design depth is selected
Hence, the soil resistance is not fblly mobilized in
cases of re-driving tests due to insufficient dribing
energy, resulting in small set-up ratios apparently
especially for longer piles
Figurc I 1 Tlic rclationsliip bctwccn sct up ratio and pilc
Icngt h
3 COMPARATIVE DYNAMIC AND STATIC
LOAD TESTS ON 13 STEEL PIPE PILES
The Japanese Association for Steel Pipe Piles
(JASPP) conducted their own dynamic load tests at
13 different test sites from 1991 to 1995, while other
engineers conducted static load tests Main purpose
of these comparative dynamic and static load tests
was to examine the applicability of the dynamic load
test to the estimation of the load-displacement curve
for a steel pipe pile as well as it’s bearing capacity
Table 2 lists the test conditions, such as the hammer
used, the test pile specifications and time intervals
between the dynamic load test and the static load
test, and summarizes the main results of the dynamic
and static load tests. The ultimate bearing capacity,
Q, the toe capacity, Qll, and the shaft capacity, Qr,.
Table 2 Main results of comparative tests of dynamc load test and static load tests
Pile@ Total Pile capacity Pile capacity Tram Pile Re- Elap Pile capacity at SLT Elap
dia pile at end of drivlng at re-driving energy set bound time (kN) time
No Haininer (inin> length (kN) (kN) (kNm) (mm> (mm)
(In)
Shaft Toe Total Shaft Toe Total Shaft Toe Total
1 STM500 2000 62 0 13651 5468 19120 19424 5978 25402 321 0 43h 29067 6213 35280 >14d
02 250
2 DSLSO 800 7 3 0 2724 411 3165 7213 196 7409 683 4711 8722 1998 13720 >14d
02 220
3 DSL15 800 1 1 5 20 141 960 19 1009 426 Ilh 4292 118 4110 >11d
12 1 17 100
4 DSL4S 609 37.5 465 186 651 3679 1214 1923 32.5 1.6 10.0 115h 4059 596 4655 31tl
5 DSLIS 609 3 9 0 1067 206 1273 1861 570 2131 15 3 1111 6217 515 6762 >14d
250 120
6 DSL45 609 35.0 284 1019 3925 918 1843 64.6 2.0 16.0 105h 4240 1245 5485 27tl
735
7 DSL10 800 11.0 1259 61 1320 3850 754 4604 28.2 23.0 5.0 66h 4200 158 4658 29d
8 DSL72 1000 7 6 0 8118 2656 11071 10672 2274 12916 121 2 0 1 220 66h 12279 1215 over 90d
13524
9 HYD100 1600 52 0 11662 1312 15971 14501 3822 18326 295 3 17 21 0 15mi 22510 3920 26160 70d
n
10 NYDl00 700 51.0 1735 2862 4596 7860 960 8820 92.8 1.2 27.0 67211 6105 2127 8232 l4cl
11 DSL45 508 32.0 1480 323 1803 3410 823 4234 90.6 1.5 24.0 138h L] - 3822 l1d
12 HYD100 800 82.0 1823 5958 7781 11731 4067 15798 86.3 1.6 15.0 15h 10780 3920 14700 lltl
13 HYD65 400 18.0 1098 892 1989 1401 774 2176 60.6 12.0 13.0 720h 1421 735 2156 90d
STM Steam hammer DSL Diesel hammer HYD Hydraulic hammer

were estimated through the wave matching analysis
of the dynamic load test signals In the cases of the
static load tests, Q, and Cl, were separately estimated
when axial forces of the pile were measured,
otherwise 0 alone was measured Note that Q is
defined as the load corresponding to a pile toe
displacement of 10% of the pile diameter
The load-displacement curve estimated from the
wave matching analysis of the re-driving test signals
is compared with the load-displacement curve
obtained from the static load test in Figures 13 to 19
for 7 tests out of the 13 comparative tests For the 7
selected cases, set per blow, S, was greater than
1 2mm In cases 1 and 2, S was almost equal to 0
indicating insufficient driving energy to estimate the
ultimate pile capacity In fact, the ultimate capacity
derived from the re-driving tests for cases 1 and 2
are notably smaller than the ultimate capacity
obtained from the static load tests
In cases 3, 5 and 9, the time intervals between the
end of initial pile driving and the re-driving tests
were short (14 hours in cases 3 and 5 , 15 min in
case 9), not allowing the completion of the set-up
phenomena of the piles In these cases, the ultimate
capacity derived from the re-driving tests is clearly
smaller than those obtained from the static load tests
It is interesting to note that the same hammer
(DSL45) was used in cases 3, 4 and 5 and an
ultimate capacity of 4 9MN was derived in case 4
Therefore, the small values for the ultimate capacity
derived from the re-driving tests for cases 3 and 5
can be attributed to the fact that the set-up
phenomena have not been completed at the time
instants of the re-driving tests
In case 8, the ultimate capacity was not obtained in
the static load test due to insufficient load capacity of
the loading device Therefore. case 8 was excluded
from the comparison of the load-displacement curves
obtained from the dynamic and static load tests
Figure 12 shows the change with time of the set-
up ratio of each pile after the end of initial pile
driving The values of set-up ratio are in a wide
range from 1 1 to 10, although the set-up ratio is
lower than 4 if the re-driving test is conducted within
3 or 4 days after the end of initial pile driving It
may be very difficult to predict the set-up ratio
accurately at the current stage of the pile technology
Figures 13 to 19 show the load-displacement
curves obtained from the static load test and derived
from the re-driving test, together with the profiles of
the soil layers and the SPT N-values, for piles 4, 6, 7
and 10 to 13 where re-driving tests were conducted
at time instants sufficiently after the end of initial
driving to permit a full recovery of the pile capacity
Various commercial computer programs such as
CAPWAPC and TNOWAVE, and KWAVE
(developed by Matsumoto & Takei (1991)) were

745
Figure 14. Load- displacement cunes of case No.6
used for the wave matching analyses to derive the
static load-displacement curves. However, the
empirical soil model developed by Smith (1 960) was
used in all of the computer programs. Even with such
conditions, the derived load-displacement curve is in
good agreement with the static load test results for
each case. While some discrepancy between the
derived ultimate bearing
746
capacity and the measured bearing capacity is seen in
some cases, the initial portion of the derived load-
displacement curve is fairly coincident with the
measured curve, suggesting negligible influence of
the computer program and the operator
Table 3 shows a comparison of the calculated yield
load and the yield load obtained in the static load test
of each pile. The yield load is defined as the load
corresponding to the first rapid increase in the pile
displacement on the load-displacement curve. The
 Table 3. Yield pile capacity derived from the dynamic and
static load tests.
Static yield capacit! (kN)
Static load Matching
Case No. error
Test Analwis
4 4635 4924 5.9%
6 4998 4843 -3 2 %
7 429 1 3150 -3 4%
10 5880 8397 30.8Y"
11 3822 4367 12.5%
12 1 1760 12576 -10 8%
13 2940 2653 6.5%

Table 4 Equations to estimate soil paraiiieters for the rational

soil models
for shaft iiiodel k , = 2 7% l ( d ). c, = G/175
8(; 3 2G
for base model kl, = ____ . Ch =-
'

7T( 1 - v)d ?r(l- c,)I ->

\c Poisson's ratio of soil. p, soil dciisitj, cJ pile diameter

(; =pl :2Shear modulus of soil.
T ', shear u a\ e 1clocit> of soil

Figure 20 Ratioiial soil models (after Raiidolph & Siiiions

1986. Raiidolph & Decks 19'92)

difference between the derived yield load and the

measured yield is within t10%, except for pile 10
where KWAVE program was employed for the wave
matching analysis
An error of I-tlOO/b may be acceptable, if the safety
factor of 3 is used as is in Japan Even if the
accuracy of the dynamic load test is less than the
accuracy of the static load test, the increase in the
number of tests performed at a site allows an
increase of the design value for the bearing capacity
(Hayashi et a1 2000)
Figure 21 W a ~ cmatching results n i t h the use or Smith
iiiodel
Pile No 10 was re-analyzed by means of KWAVE
with the use of the rational soil models developed by
Randolph 8L Simon (1986) and Randolph & Deeks
(1992) The rational soil models have been
incorporated in the KWAVE program In the
KWAVE program, the wave propagation in the soil

747

Figure 22. Wave matching results with the use of the rational
soil models.
Shaft resisitance, z (kN/m2)
-from Static load test
_ I - Wave matching analysis with rational soil models
- - -Wave matching analysis with the smith soil mode
Figure 23. Distributions oft,, derived from wave matching
analyses using the S m i t h model and the rational soil models.
inside the open-ended pipe pile and the mobilization
of the internal shaft resistance are taken into account,
according to the modeling of the internal soil (soil
plug) proposed by Randolph (1987). The soil
parameters except for the maximum shaft resistance,
zma,, and the maximum toe resistance, q b , were
estimated from the equations listed in Table 4. The
shear wave velocity, 6,of the soil was estimated
using the following empirical equation (Imai 1977);
The soil density, p, was assumed to be 1800 kg/m3
for all the soils.
In the wave matching analyses, the distribution of
zm, and q b alone were assumed. The final matching
results using the Smith model and the rational soil
models are shown in Figure 21 and Figure 22,
respectively. A better matching was obtained in the
wave matching analysis using the rational soil
models. The distributions of rmx derived in both the
wave matching analyses are shown and compared
with the static load test results in Figure 23. The
corresponding static load-displacement curves are
shown in Figure 24, together with the static load test
results. It can been seen fiom these figures that the
wave matching analysis using the rational soil models
results in better predictions of the static load test
results.
4 CONCLUSIONS
This paper reviewed dynamic load tests performed in
Japan, showing their chronology and statistics. It
was shown that most dynamic load tests performed
in Japan have been conducted on steel pipe piles in
offshore conditions for the acceptance of the tested
piles. A total of 13 comparative cases of dynamic
load testing and static load testing, which were
conducted by JASPP, have been reviewed in detail.
The following findings were presented:
1. It is difficult to predict the set-up ratio of a pile
accurately. Hence, conducting a re-driving test of
the pile is required to estimate the bearing
capacity as well as the load-displacement curve
for the pile.
2. The load-displacement curve derived fi-om the
dynamic load test is in fairly good agreement with
Figure 24. Static load-displacement curves derived from the
wave matching analyses and obtained from the static load test.
748
 that obtained from the static load test until the
yield load is reached, if both the tests are
conducted at time points sufficiently after the end
of initial pile driving.

REFERENCES
Hajaslii. M . Matsuiiioto. T & Suzuki. M . 2000 Dliiaiiiic
load testing on 102 steel pipe piles for bridge fouiidations
on inudstone Proc 5th Int Conf 011 the .4pplicatioli of
rhe Stre,r-Iim’e Theor1 to Pilec. Sao Paulo (to be
published)
Iiiiai. T 1977 P and S n a \ e lelocities of the ground iii
Japan. Proc. 9th I C Y W E . T o b o . 1977
Matsuiiioto. T & Takei. M . 1991 Effects of soil plug on
beha1 lour of drn en pipe piles. Soils and Foundations.
V0l3 1. NO 2 11-31
Raiidolph. M F . 1987 Modeliiig of the soil plug respoiise
duriiig pile d m iiig Proc 8th ,C E .-lsrnti Geotechtiicnl
C’onf. Bangkok. Vol 2 6 1-6 1 1
Randolph. M F & Siiiioiis. H A . 1986 An iiiiproved soil
inodel for one-diiiiensioiial pile drir ing anal! sis Proc 3rd
1111Coiif on \’zii?? .\Jet11 I I I Oj,?ihore Pilitig, Nantes 1-1 7
Raiidolph. M F and Decks. A J , 1992 Djnaiiiic and static
soil inodels for alial pile response Proc of 3rd Iiit C ’ m f
oii .4pplrcntion of Stress-11nve Tlieort to Piles. Hague 3-
11

749
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3

Case studies of high capacity CFA pile testing in Australia

S. Baycan
Pile Test International,Melbourne, Vic.,Australia

ABSTRACT: Continuous Flight Auger (CFA) piling continues to be a cost effective solution employed
throughout Australia to solve many foundation engineering problems. This paper aims to describe two exam-
ples of dynamic load testing on recent CFA projects, including a comparison with a static load test, A brief
description of pile construction, dynamic and static load testing for piles constructed at the new Sydney Inter-
national Airport and the new Adelaide Convention Centre will be outlined.

1 INTRODUCTION 2 GEOTECHNICAL PROFILE

1.1 Sydney International Airport 2.1 Sydney International Airport

As part of the upgrading of Facilities at the Sydney The idealised geotechnical profile within the area of
International Airport in preparation for the 2000 piling works at Sydney International Airport was
Olympic Games, Vibro-pile Aust Pty Ltd was com- characterised by fill, overlying variable relative den-
missioned by Transfield Constructions to design and sity sand with interbedded layers of soft clay and
construct foundations for the proposed extension sands of medium density. Very stiff sandy clay was
works. The project required constructing founda- found to overly the sandstone bedrock. CPT testing
tions into widely variable ground conditions over- was used extensively to investigate the ground con-
lying a weathered rock profile. Due to limitations on ditions and a summary of ground conditions is
vibration and disturbance to underground services shown in Table 1.
approximately 500 number CFA piles varying from
0.6 to 0.9m diameter were constructed for the proj-
ect. The test program consisted of a series of low Depth Geotechnical Profile
strain integrity, dynamic and static load testing. 0-4m Med. Dense - Dense SAND
Geotechnical information, pile construction records 4-5m Soft CLAY
and dynamic test results were used to determine a 5-8m Med. Dense SAND
CAPWAPmodel. Model results compared well with
the static load test result.

1.2 Adelaide Convention Centre

A combination of ‘V’ (a Vibropile proprietory
displacement pile) and CFA piles were adopted by
Vibro-pile Aust Pty Ltd for the foundations at the
Adelaide Convention Centre project. This project
required careful planning due to the structure strad-
dling a roadway and railway lines. A total number of
approximately 300 piles were constructed. An ex- 2.2 Adelaide Convention Centre
ample of a dynamic test result on a 500mm diameter The geotechnical profile at the Adelaide Convention
CFA pile and the corresponding CAPWAPmodel for Centre site consisted of a variable layer of fill up to
such a test is described. 3m in depth, overlying very stiff silty clays to ap-
proximately 14m, overlying very dense sand and
gravel. Residual siltstone - hard clays underly the

75 1
sanddgravels below 18m depth. Site investigation Excess 36 t Rig M R Site WHUENIION, BDElilIDE 2182
consisted of borehole drilling and the use of SPT Diaieter 588 NI Total Vol 4.551 n3 Borelin 16 nin Start 87% Pile P.TEST
testing. The shaft of pile T3 was predominantly in length 16\58 n Precharge 8.284 w3 Canclin 9 Rin Stop M:56 Ilate 19/82/88
very stiff - hard clays with the base at 16.5m in very Bore Penetrate Toque Extract Pressure Oversupply Profile
dense sandy gravel. (dmin) (wdrev) ($1 (Wnin) (bar) (11 (RR)
8 1 2 3 8 288 8 58 8 2 4 6 8 1 2 3 8 48 148 258 R 258

3 PILE DESIGN AND CONSTRUCTION

Careful planning and site specific constraints re-

quired consideration from the Vibro-pile designers
for the Sydney Airport project. These included pil-
ing within 50m of Boeing 747 aircraft and minimis-
ing affects on underground fuel lines and other
services. CPT investigations were used extensively
to establish indicative pile design parameters. Indi-
vidual piles were designed on a combination of shaft
and end bearing. Design parameters up to 45%
greater than those recommended in the geotechnical
report were adopted. Elastic pile design using PIG-
LET was used to initially design pile groups. Pile
testing was planned early on in the project to justifi
seemingly ambitious design assumptions. Pile test-
ing was critical as results were used to back-analyse
pile global settlement calculations. Pile construction
was digitally monitored and relayed to the office - in
real time to provide added assurance of pile con-
struction quality.
Vibro-pile used a similar on-board computer 1

system PL20 to monitor pile construction for the

Adelaide Convention Centre piles. The computer
system is able to display construction parameters
such as depth, penetration rate, torque, injection
pressure and volumetric oversupply. This data is
used to obtain a better CAPWAPpile model. An ex-
ample printout of the PL20 system is shown in Fig-
ure 1.
4 PILE TESTING PROGRAMME
4.1 Sydney International Airport
Every pile constructed for the new Sydney Interna-
tional Airport upgrade was integrity tested using low
strain, frequency response and sonic echo methods.
Based on integrity test data, all piles were found to
be satisfactorily constructed. A total of eight piles
were dynamically load tested, however for brevity
the results for pile T5 only will be reviewed here.
The test system for the static load test comprised of
jacking against steel beams with a lOOOtonne capac-
ity hydraulic actuator. Tension piles were used as re-
action. Pile movement was monitored by string PO-
tentiometers, with dial gauges and dumpy level
readings used as backup to verify data. The loading
cycle was applied in accordance with AS2159 . The
load-unload and reload sequence comprised of
loading in additional increments of 15% up tc 2
Figure 1 Adelaide Convention Centre PL20 pile record T3
value of 150'?Aos* (where S* is the factored design
load in ultimate limit state analysis, which is usually
1.35 x working load).
The test acceptance criteria consisted of:
(a) pile top settlement 4 0 m m at design
working load,
(b) Net pile settlement <5mm upon re-
moval of test load,
(c) Pile top settlement < 25mm at test
load = 2xdesign working load.
The specification included an allowance for elas-
tic shortening of the slender test piles, so that the
limit of gross settlement could be increased by the
amount of calculated elastic pile compression ex-
ceeding 1Omm.
All piles tested were constructed using concrete
having a 28 day characteristic strength f c 2 40MPa.
Pile construction, load requirements and test details
are listed in Table 2.
4.2 Adelaide Convention Centre
As with the Sydney tests, a purpose built 12tonne air
hammer with a seating frame was used to apply the
test load. Hammer drops were varied between 0.3

752
and 1.2m. All dynamic testing was conducted using
PDA equipment designed and manufactured by PDI
-USA.

Pile Length Date

Adelaide 500m 16.5111 19/2/00 12/3/00 2300

5 TEST RESULTS AND ANALYSIS

5.1 Sydney international Airport - Static &

Dynamic Load Tests
The static load test on pile no T5 was carried out be-
fore dynamic load testing in order to be able to cali-
brate results for application to other piles tested on
the project. A maximum load of almost 8900kN was
recorded at a load of 150% of S*, which is equiva-
lent to twice the working load. Measured gross set-
tlement at this load was 25mm. As described, the ac-
ceptance criterion for the static test included an
allowance for elastic pile shortening (19mm for
32.9m pile). Consequently, gross settlement could
not exceed (19-10+25)= 34mm i.e. almost 5% of
pile diameter.
PDA test results were signal matched using
CAPWAPand results suggested a mobilised pile re-
sistance of 8000kN, with over 60% of the load car-
ried in shaft resistance.
A comparison of the static load test and the
CAPWAP prediction is shown in Figure 2.
Subsequent to the results, designers were able to
be confident in their initial assumptions regarding
pile parameters and to further refine their pile de-
sign.

5.2 Adelaide Convention Centre Dynamic Load

Test
This pile was selected to reveal some of the pa-
rameters which affect the interpretation of PDA data
from a cast-insitu CFA pile test. To enable reliable
data to be obtained, four sets of strain gauge and ac-
celerometers were attached to the pile for dynamic
load testing. The 500mm CFA pile described here
was testing with four sets of PDA strain and acceler-
ometers attached to the extended section of the pile.
Other dynamic tests on the site were conducted with
gauges attached to the actual pile section with 0.5m
of ground around the pile being excavated. That ap-
proach was taken where the pile extension section
was cast at a different time to that of the actual pile.
In this instance, the interpretation of the wave speed
was made simpler by using casting the top section of
pile T3 at the same time as the actual pile. A wave
Figure 3 Adelaide Convention Centre PDA test result CFA
500 pile T3
speed of 3950ds was deduced, with a maximuni
transferred energy of 85kJ corresponding to a ham-
mer system efficiency of 60%. A typical set of PDA
results for the blow described is shown in Figure 3.
It is of interest to note the lack of forcehelocity pro-
portionality at peak. This was concluded to be due
primarily to the increase of impedance from a
400mm diameter extension above ground to the ac-
tual pile 500mm nominal diameter.
The maximum compressive stresses was greater
than SOMPa, for a transferred energy of 85kJ applied

753
 I 1
out on T3 was not able to mobilise the available base
resistance.
2 Current

6 DISCUSSION
2 Modified
The Sydney International Airport static load test
proved that an effective, early planned pile-testing
z Suggest program was able to achieve efficient outcomes for a
CFA piling contract. By using a static load tested to
failure, calibration of dynamic load test results was
Figure 4 Adelaide Convention Centre CAPWAP Imped-
undertaken. This lead to added confidence in the de-
ance Profile Pile T3
sign assumptions. Back analyses of data allowed de-
sign assumptions to be verified.
Numerous pile testing programs for socketed piles in
Load in KN rock have shown that unit resistances employed can
000 6000 8000 be significantly greater than those normally adopted
0. in design. Published examples include Tchepak
(1998), Baycan (1996) and Seidel et a1 (1 998).
5.
Ru = 6018.6 KN
10. Rs = 3912.1 kN
7 CONCLUSION
Rb = 2i06.5 KN
Dynamic load testing of piles combined with a static
15. Dy = 15.0 mm test such as that employed on the Sydney Interna-
Dmx = 16.0 mm tional Airport project, can allow a correlation to be
used effectively to save time and money on piling
Displacement in mm contracts.
The interpretation of a dynamic load test on a
Shaft Resistance non-uniform CFA needs to take into account the
KN/m
800 1 Distribution geotechnical profile, concrete properties of pile and
extension, pile construction records and any other
effects of the test setup. By making use of such data,
Pile Forces a reliable static model of the pile is determined.
at Rut
10000
kN
8 ACKNOWLEDGEMENTS
Figure 5 Adelaide Convention Centre - CAPWAP result pile T3
The author gratefully acknowledges the support and
advice of Dr Julian Seidel.
with a 1.2m drop of the 12 tonne hammer. A mobi-
lised resistance of 5900kN using an assumed Case
Damping Factor of Jc=0.7 was achieved. The re- 9 REFERENCES
corded set for the blow was 2.6mm.
Due to the added complication of a variation in AS2159, 1995. Australian Standard, Piling - Design
impedance and pile concrete properties (and there- and Installation Published by Standard Australia,
fore wavespeed), for CFA and for all cast in situ ISBN 0 72629884 0
piles, a review of all available piling records and Baycan, S. 1996. Field Performance of Expansive
geotechnical profile becomes critical. The assess- Anchors and Piles in Rock. PhD Thesis, Department
ment and building of a CAPWAPmodel for pile T3 of Civil Engineering, Monash University, Australia.
required such an assessment.
Using the PL20 record as guidance, the non- Seidel, J.P. and Haberfield, C.M., and Baycan, S.
uniformity of the CFA pile was modeled in CAPWAP 1998. Load displacement performance of bored piles
and the adopted profile is shown in Figure 4. in weak rock. Deep foundations on bored and auger
Using the adopted parameters, a reasonable signal pile, Van Impe (ed), Balkema, Rotterdam.
match was achieved in CAPWAP and results are
summarised in Figure 5. Tchepak, S. 1998. The design and performance of
It is well known that in order to mobilise the full bored piles in shales for the Australia Stadium proj-
available base resistance of a pile, sufficient move- ect. Deep foundations on bored and auger piles,
ment is required. It is expected that the test carried Vam Impe (ed), Balkema, Rotterdam.

754
11 Supplement
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama & Beim (eds)0 2000 Balkema, Rotterdam, ISBN 90 5809 150 3

Keynote lecture: trospective of Sonic Integrity Tests - Its application

to the control quality on piles

Juan J.Goldemberg
Geotecnica Cientic, Buenos Aires, Argentina

ABSTRACT: A retrospective of the past twenty years since the First Stress Wave Conference of Stockholm
is made, narrating the evolution of Sonic Integrity Testing since its origins until today from the two major
equipment manufacturers. As well, three case histories are presented, highlighting the necessity of introduc-
ing SIT tests as a regular quality control in all piled foundation sites.

1 INTRODUCTION. systems, and the market conditions that encouraged

the development and current applications of this
The purpose of this paper is to look at the technical technology. In addition, some cases histories are
data and documentation on Sonic Integrity Tests described of Deep Foundation Pathologies, detected
(SIT) and the major breakthroughs that have taken through the performance of Sonic Integrity Tests
place over the last twenty years since the First and simulation techniques for defect quantification.
Conference in Stockholm, Sweden, by the initia-
tive of a few pioneers in Stress-Wave Theory and it
Applications to Foundations Piles in 1980. Since 2 CONFERENCES ON THE STRESS-WAVE
then, the number of participants as well as the vol- THEORY
ume and content of papers submitted to each Con-
ference, even from the so-called “peripheral” coun- Conferences on Stress-Wave Theory continuc to be
tries has been on the increase. held successfully every four years since 1989 as
Undoubtedly, this reflects the marked interest witnessed by the increasing participation of the in-
and concern shown by engineers across the world to ternational engineering community and a larger
be able to determine what happens to deep founda- number of papers being submitted at each Confer-
tions once they are installed- irrespective of the ence [ 11, [2],[3], [4], [ 5 ] ,[6]. Table 1 shows an in-
procedure employed. creasing interest in Quality Control on deep foun-
A great number of applications have been de- dations as a result of a greater knowledge of the
rived from the Strcss Wave Theory, among them: available methods, equipment and signal evaluation
Pile Driving Analysis (PDA) systems.
Dynamic Load Test (DLT)
Rapid Load Test STATNAMIC (RLT-STN)
Vibration Monitoring (VM) Table 1. SIT Papers presented at Stress Wave Conferences
Monitoring Hammers (MH) Conference Papers on Total
Pile Driving Prediction with Impact Hammers SIT PaDers
(PDP-IH) 1980- Stockholm (Sweden) 5 24
Pile Driving Prediction with Vibratory Ham- 1984- Stockholm (Sweden) 5 54
mers (PDP-VH) 1988- Ottawa (Canada) 13 88
Soil-pile Interaction Modelling and Signal 1992- La Naya (Holland) 16 107
Matching Technique (TNOWAVE, CAPWAP, 1996- Orlando (U.S.A.) 21 101
etc.). 2000- San Pablo (Brazil) 14 100
Sonic Integrity Test (SIT)
This paperlodks at the basic documentation on As a result, Sonic Integrity Tests- as they are
Sonic Integrity Tests underlying a rational inter- known today- have been in most in-
pretation and evaluation, the fundamental contribu- dustrialised countries and are being done in several
tion of data processing and compilation means and countries where SIT^ are beginning to be applied.

757
Another result of these International Conferences
are the Alliance of Pile Testing Laboratory Engi-
neers (APTLY), an association recently set up dur-
ing 1995 by users and manufacturers of these tech-
nology, as well as the periodic Seminars (among
them, Control Quality on Piles: Uses & Abuses) on
the occasion of the annual meetings of the Deep
Foundation Institute (DFI).
That is how nowadays, the Quality Control
through Non Destructive Tests (QC-NDT), has be-
gun to be included in the Technical Specifications
for the construction of piled foundations in most
Civil Engineering works, by private companies or
governments from those countries that have
adopted this technology. The availability of this
technology has resulted in an early detection of
defects likely to jeopardize the stability of deep
foundations and the superstructure resting on them
and reduced the occurrence of accidents likely to
lead to the loss of both lives and assets.
3 QUALITY CONTROL ON PILES
In the late ’70s and early ’ ~ O S the
, situation of most
industrialised countries and some “developing”
countries can be summarised as follows:
3.1 Growth of Construction.
This growth was characterised by a strong flow of
capital to the construction and/or upgrading of in-
frastructure projects, such as roads and bridges,
thermal, nuclear and hydro power stations and as-
sociated transformer stations and transmission
lines, industrial and petrochemical plants, airports,
port facilities and housing developments. In many
cases, these projects required the construction of
piled foundations with high design loads. This re-
quirement led to the introduction of new technology
for large diameter cast-in place concrete piles under
bentonite mud.
3.2 Inadequate and/or insufficient Quality Control
A long-standing general concern in the construction
community has been the safety of foundation
works, in particular those associated with excava-
tions and piles. This concern arose from the very
nature of these structures, the imperfect knowledge
of the soil profile, wide range of building methods
and customs, general lack of effective quality con-
trol methods, lack of tests and measurements on thc
behaviour of the soil-structure interaction and a
certain blend of pseudo-confidence and fatalistic
resignation (or hope ?) on “...pile contractors surely
know what they are doing ...”
In addition to the above conditions, there has
bcen a marked absence of involvement from the
758
Geotechnical Engineer, particularly in South
America, during the Preliminary Project, Project
Preparation and Construction stages. In most cases,
Geotechnical Engineers participated only in the
“soil investigation” and failed to be included in the
preparation of a Geotechnical Research Plan neces-
sary for proper subsoil identification based on the
available data, subsoil characteristics and site spe-
cific requirements.
Upon completing of the “soil investigation”-
with the scope and vision of a merely formal re-
quirement in most cases- specialists are kept away
from the tasks within their specific field of exper-
tise, such as the analysis of foundation alternatives,
determination of load capacity of foundations and
prediction of foundation behaviour, geotechnical
and structural design (which is beyond a straight-
forward determination of the “allowable stress” as
is mistakenly believed), selection of the construc-
tion method (or alternatives) better suited to local
conditions, design of containment and anchor
structures, excavation procedures, groundwater
abatement project, drainage network and relocation
of digging material, soil compaction from mechani-
cal, chemical or thermal causes, soil filling project
(including sanitary fillings and embankments).
Last but not least, the necessary Quality Control on
soil movements and/or upgrading, contention and
anchor, underground and open pit excavations, and
foundations, among other things.
Worse still, some of these tasks are sometimes
carried out by the Client without the support of
properly equipped laboratories or personnel lacking
the necessary experience or post-graduate training
and, in some cases, personnel alien to the disci-
pline or lacking the necessary skills. It can then be
concluded that Foundation and Contention Project
Design and Construction and consequent Safety and
Economics are, at least, “unstable”. This is unac-
ceptable for any structure designed under the Static
Principles.
Because the usual design criteria have failed to
catch up with the current building practices in de-
veloped countries and the quality precepts associ-
ated with safety and economics are slow in being
adopted, deep foundations and contention structures
continue to be, theoretically, over- (or hyper ?) di-
mensioned with the resulting loss or poor use of,
among other things, intellectual resources, materi-
als, time and labour ... in other words, money.
The above circumstances and the many cases of
structural and foundation pathologies we were in-
volved in our region, paved the way for a study of
the way in which these problems were approached
and solved in developed countries. Our study al-
lowed us to be in contact with the European and
U.S. experience, scope and limitations of Non-
Destructive Testing technology derived from the
Stress Wave Theory and to decide on its introduc-
tion to the services provided by our Company.
Thus, in 1986, the first tests were performed by
us using equipment developed and manufactured by
TNO - Building and Construction Research (DeIft,
The Netherlands) [7].
3.3 Market Demand: Quality Assurance (QA).
The current trend in most industries is to attain the
necessary targets to reassure the industry, its cus-
tomers and the community at large, and that all
quality standards or criteria have been met at each
stage. This involves systematic compliance with
each and every procedure leading to the achieve-
ment of planned objectives within each area in the
organisation of the purposes of meeting the Quality
Policy pursued by the Company, the Owner and/or
the Consumer (or the Community itself), within the
scope of adcquate conditions in terms of safety and
economics.
This approach, which is not opposed to the old
concept of Quality Control (somehow associated
with an “autopsy”, “death certificate” or “post-
mortem” investigation) is helpful on the basis of a
set of tools and modern techniques based, in most
cases, on rapid Non-Destructive Tests (NDTs) suf-
ficiently known all over the industrialised world,
numeric analysis methods and simulations, and
control procedures developed for rapid accurate di-
agnosis of the detected anomaly.
The fitness of these tools for their intended pur-
pose is dependent, prior to the implementation of a
successful Quality Control System, on a strict per-
formance of quality standards associated with the
scope of the Architectural and Engineering Prelimi-
nary Project, study of the available data, project
feasibility study, essential construction service
availability study, regulatory and environmental
provisions, basic geotechnical studies (including
feasibility of using local materials for fillings and
construction, and final disposal of digging materials
and waste generated by the works), Executive Proj-
cct, analysis of alternatives, as well as geotechnical
and structural foundation, excavation and conten-
tion and anchor design and, if needed, abatement,
drainage and/or digging materials relocation.
Successful performance of these tasks will en-
sure an efficient construction in terms of quality,
safety and cost-efficiency. This requires specialists
in each trade to meet the Quality Assurance (QA)
standards for the project. The increasingly accepted
principle of having a geotechnical specialist during
each stage of project preparation and implcmenta-
tion has led the most developed countries to stage
Technical Meetings, Seminars, and Congresses to
analyse this issue in greater detail [8].
Figure 1 (quoted by R.P.Thompson during the
Seminar on “The Value of Geotechnics in Con-
759
struction” [9] organised by the Institution of Civil
Engineers- ICE- in London on November 4, 1998)
clearly show the incidence of the possibility of cost
optimisation consistent with the Project Develop-
ment Stage and Costs of the sometimes unavoidable
and sometimes unforeseeable -Scope Changes de-
pending on the same parameter. In these cases,
while the immediate increased costs are paid by the
Owner or Concession Operator, the final actual
costs are paid by the Community at large.
Figure 1. Changcs in “Reduction Cost Potential” and
“Change Order Costs” relative to the time schedule.
3.4 Needs of Industrialized Countries
World concern for the energy crisis generated by
geo-political instability in the Middle East attrib-
uted great priority to the search, exploration and
operation of new oil and gas fields, in particular
those located in the North Sea and the Mexican
Gulf as a result of their geographical proximity to
the natural high demand markets. As an immediate
result of the above, there was the urgent need to
build “off-shore” and “near-shore’’ platforms
which, as usual in most cases, are supported by
piles.
This drove to rapid reliable solutions to predict
and check the bearing capacity of piles under ex-
treme performance requirements and environmental
conditions calling for the traditional Static Load
Tests (SLT) to check the design assumptions
adopted. For this reason, a number of International
Conferences and Seminars have been held since the
‘60s to present new developments in Dynamic
Tests and Numerical Analysis Methods on pile
driving and load capacity associated with off-shore
structures.
When evidence was obtained of the actual suit-
ability of certain tests and analysis of results, Sonic
Integrity Tests (SIT) emerged, almost naturally, as
the “younger brother” of other prior tests with
similar performance assumptions relative to the
Stress Wave Theory.
3.5 Scientijk Support Necessary for Technology a) Free end (F = 0, equivalent to a pile toe on soft
Development soil).
Most likely, the first references to the propagation b) Fixed end (v = 0, equivalent to a pile toe on
of lengthwise vibrations along elastic bars in- rock).
spected during our undergraduate training must c) Toe with some end bearing (F = Fp, equivalent to
have included the works by Love, A.E.H. (IO), Ti- a intermediate condition).
moshenko, S. & Goodier, J.N. (11) and Ti- d) Change of impedance (Z1 > 2 2 or Z1 < 22,
moshenko, S. & Young, D.H. (12). If the cross sec- equivalent to an anomaly detected with SIT).
tion of the bar remained flat during deformation e) Impact (equivalent to the driving condition dur-
(Navier-Stokes assumption), all these works pro- ing PDA or redrive during DLT).
posed a general equation for movement of the fol-
lowing type:
4 REVISION OF THE MAIN TECHNOLOGY
DEVELOPMENTS
a2U/o't2= c2 a2U/aX2
in which c = (E/p)o's He have considered gathering the information de-
livered by the two main sources that manufacture
where: equipment, including hardware and software:
U: displacement of the section being studied in di-
P The European, represented by TNO Building &
Construction Research, Delft, (Holland) and,
rec tion x (longitudinal 1y).
c: velocity of the stress wave. P The American, represented by PDI- Pile Dy-
namics, Inc. Cleveland, Ohio (USA).
E: dynamic elastic module.
A: cross section of bar.
p: density of material. 4.1 The main developments of TNO
The equipment produced by this R&D Institute is
The general solution of the equation (1) leads to:
characterised by the generic denomination of
+ f,(x-ct)
U = f(x+ct) (3 1 FPDS- Foundation Pile Diagnostic System and it is
and using the notation proposed by Voitus van described as follows (1 4):
Hamme et al. (1 3), we obtain:
P Equipment FPDS-0.
u = f (x-ct) + f (x+ct) (4)
Likewise, the force applied on the section is
given by:
F = -EA du/dx
F = -EA df /d(x-ct) - EA df /d(x+ct) (5)
F=F +F (6)
In turn, the velocity is expressed by:
v = duldt =
v = - c df /d(x-ct) + c df /d(c+ct) (7)
v = v +VI (8)
'These two velocities are linked to their respec-
tive f ~ r c e by
s the following expressions:
v =Fi/Z (9) Photograph 1. Equipment FPDS-0 system with an oscillo-
v = - F /Z (10) scope and Polaroid camera

where Z = EA/c
Z : bar impedance.
The application of the above equations to the
different situations arising daily in the pile industry-
coupled to the powerful new "high tech" develop-
ments- have paved the way for a rational analysis of
the signals obtained from SITS, PDAs and DLT
tests, as a result of which it has been able to differ-
entiate the following cases:
P Equipment FPDS-1.
In the mid 80s, the FPDS-I was developed, based
on a UNIX workstation and a MC68000 (1 6 bit mi-
cro-processor chip), with built-in keyboard, disk
drive, graphical display, and printedplotter. The
signal was digitised (12 bit) at an early stage after
retrieving the signal to ensure that all subsequent
processing was of the highest possible standard.
The system had certain automatic functions, but the

76@
skilled operator could also override, magnify, or
clarify a result to facilitate interpretation. The
FPDS-1 system is still used today by clients for
SIT.
Photograph 2. Equipment FPDS-I .
9 Equipment FPDS-2.
With FPDS-2, the computer and signal conditioner
were separated, a philosophy continued with subse-
quent systems. The FPDS-2 system, introduced in
1986, was based on a portable IBM-PC/AT com-
patible computer running under MS-DOS, equipped
with a 20 MB hard disk for program and data stor-
age and a floating point co-processor. For each
FPDS application, specific sensors, signal condi-
tioning subsystem, and software were available,
such as Pile Driving Analysis and Dynamic Load
Testing of foundation piles (PDNDLT). The
FFDS-2 system did not require knowledge of elec-
tronics, requiring only a limited number of compo-
nents and no oscilloscope or tape recorder. Real
time presentation of signals, automatic controls,
warnings, and error messages were a big improve-
ment on earlier systems.
Photograph 3. Equipment FPDS-2.
9 Equipment FPDS-3.
At the beginning of the 90s, the FPDS-3 system
was developed, based on a GRiD computer (80386
761
processor) and a subsystem with an analogue-to-
digital (A/D) converter board and a specially de-
signed board for integrity testing. As with FPDS-2,
the subsystem could accommodate different FPDS
applications, such as PDA/DLT, the Statnamic load
test (STN), vibration measurements (VIBRA), and
others. Each option consisted of specially designed
boards, sensors, software, and accessories, an inno-
vation begun with the FPDS-2 system. Automatic
signal conditioning freed the operator from bridge
balancing and scale selection. To reduce weight
while conducting multiple SIT tests, the signal con-
ditioner was built into the GRiD computer (called
the Consultant System).
With FPDS-3, SIT was also extended to include
an optional instrumented hammer (an accelerometer
fitted in a hammer head) to record the applied im-
pact blow. The design of the hammer head, impact
cap, accelerometer sensitivity, and hammer shaft
stiffness produces a load pulse signal perfectly rep-
resentative of the impact amplitude and phase.
Since force is proportional to velocity at the pile
head, the shape of the force and velocity traces are
identical until a reflection is recorded. By compar-
ing the force signal to the velocity signal, defects in
the first 2 meters of the pile are revealed. The force
is equivalent to the pile head impedance ’ velocity.
Photograph 4. Equipment FPDS-3
P Equipment FPDS-4.
In 1993, FPDS-4 was developed as a very robust
and economically designed lunch box type system.
The FPDS-4 system includes a 80486 processor,
8°-200 MB hard disk, and a built-in graphical
screen and key board, both sealed to protect against
rain and dust.
P Equipment FPDS-5.
Recently, the FPDS-5 system was developed using
a notebook computer and separate signal condi-
tioning and A/D card subsystem. Data communica-
tion from subsystem to notebook is via a PCMCIA
card and cable. The FPDS-5 system is suited for
PDNDLT and STN, but can also be used for SIT.
Photograph 6. Equipment FPDS-5, Notebook with SIT
PCMCIA card.
P Equipment FPDS-6.
As well, TNO's latest FPDS system, the FPDS-6,
represents a new approach to foundation testing
equipment. The FPDS-6 is a hardware/software
package for SIT, designed to work with a wide
range of IBM compatible PC computers equipped
with a PCMCIA card slot. The SIT hardware con-
sists of a hammer, sensor, cable, and PCMCIA
card. The PCMCIA card, measuring just 8.5 cm by
5.5 cm by 5 mm, incorporates all data acquisition
(300 kHz, 14 bit) and signal conditioning electron-
ics. Besides the enormous space and weight sav-
ings, this innovation dramatically increases overall
system reliability and service life. The user can use
his own notebook computer or a hand-held Hard-
body PC.
All input is via an easy to use touch screen.
Knowledge of electronics is unnecessary-all elec-
tronic functions are controlled by the computer.
Advice and warnings are generated by the system.
Signals and data are stored automatically and can
be easily recalled. Automatic reporting capabilities
greatly reduce reporting time.
4.2 The main developments of PDI
The equipment manufactured by this company and
its respective innovations are described in the fol-
lowing list (1 5):
762
Photograph 7. Equipment FPDS-6, Field computer with SIT
PCMCIA card.
P Years 1960/1979.
Hardware: Development of analog instrumenta-
tion and computation technology.
Software: Digital integration in the laboratory.
Beta Method published.
P Years 1983.
Hardware: a) Data acquisition: For low strain,
highly sensitive accelerometer and power supply;
digital data acquisition and integration with Pile
Driving Analyzer (PDA); digital display on oscillo-
scope. b) Kecording: Oscilloscope, photos of os-
cilloscope traces, analog FM magnetic tape, X-Y
plots.
Software: Analysis in time domain
k Year 1985.
Hardware: IBM PC or compatible 8087 math co-
processor, 5 12 kRam, A/D board (acceleration rec-
ords integrated and digitised) , signal amplifier.
Software: m e g r i t y Program, exponential am-
plification, averaging of records, smoothing.
P Year 1989.
Hardware: PIT test equipment: PC + signal con-
ditioning; PIT processor and plotter or graphics
printer in field; permanent storage on disk.
Software: Analysis in frequency domain.
P Year 1990.
Hardware: PIT-SC: battery powered, signal con-
ditioning and printer built in a 80286-based PC
computer. Instrumented hammer.
 Software: Built in time and frequency domain
software. Impedance profile. PITWAP (calculates
pile shape given a soil model, based on wave equa-
tion).
> Year 1992.
Hardware: PIT Collector, touch screen, 16 bit
A/D converter, high and low pass data filtering;
signal amplification; interface to plotter/printer; full
day battery; storage of 350 piles.
Software: Built in analysis software in time do-
main. Post processing in frequency domain.
> Year 1996.
Hardware: PIT Collector memory expanded; im-
proved display capabilities.
Software: PITSTOP (frequency domain, 2 ve-
locity, multiple plots, impedance profile).
P Year2000.
Hardware: Improved memory (1000 piles) and
processing speed.
Software: PIT-W (Windows)
5 COMMON PATOLOGIES IN DEEP
FOUNDATIONS.
In order to describe the principle causes of pathol-
ogy detected in the foundation structures, it is con-
venient to list them into:
5. I Problems originated in its construction.
5.2 Problems derivedjkom its use during lif. time
of the structure.
5.3 Problems due to geotechnical- environmental
phenomenon.
In this paper a reference to the problems generated
during the construction of cast-in-place piles will be
made for being the most common in our deep foun-
dation market and the ones which have more possi-
bilities of showing anomalies, according with the
authors experience. In general the problems that
can turn out of the construction of different piles
and deep foundations, had been described in a
number of publications: texts, seminars, confer-
ences, etc. (16), (17), (18). In order to identify those
causes of damage, problems can be classified in the
following way:
5.1. I Problems associated to cast-in-place
piles.
763
5.1.2 Problems associated to drived piles
5.1.3. Problems associated to external causes in
the manufacturing process.
A detailed description of the problems and the
methods available for their detection has been
clearly exposed by M.J. Turner (19) and by J.J.
Goldemberg & H. Goldemberg (20). As typical ex-
amples of the anomalies detected in foundations
built with cast-in-place piles the following deserve
to be highlighted:
> Case History I.
Corresponds to a sector belonging to a group of
Towers of 1 1 stories destined to economic housing,
built in the south zone of Buenos Aires (Argentina)
and where, due to contractual reasons, two of the
mentioned buildings remained for a long time with
their foundations (piles, caps and foundation
beams) partially built without their superstructure
completed.
The geotechnical profile of the building site (see
Figure 2) shows a first layer of low plasticity clays
(CL) up to 4.00/4.50m depth with N value (SPT)
between 3 and 16, laying on a non plastic silty
strata (ML) with similar respond to the SPT test, up
to a depth of 6.00m. From here onwards, the soils
prcsented a strong preconsolidated characteristic
with N>40, up to well below 10.00m from zero
level. At the moment of the investigations, the wa-
ter table level was 2.00/32.50m below ground level.
The foundation system used consisted of dis-
placement piles by driving a recoverable close
ended caisson and casting the concrete while the
mould was pulled out with the help of a vibratory
hammers. Their nominal diameters were: 300/350
and 400mm and length of 4 to 6m. When the con-
tinuation of the building was decided, by the Main
Building Contractor’s Structural Advisor, the
Quality Control upon the total number of the piles
Figure 2. Existent geotechnical profile in the building site.
of both buildings through the performance of the Table 2. Summary of tests on piles.
Sonic Integrity Tests was arranged. TOWER PILES
In accordance with it, the 108 piles of each No Total w/Anomalies
building were tested, upon preparing its in order to A 108+6* 72+6*
eliminate all the material “that could be loose”, B 108 26
concrete contaminated by the soils, cracks caused Note: * Piles replaced by order of the Main Building Con-
by the trimming work due to use of inadequate tractor before SIT testing.
techniques, etc. That could originate false signals.
Given that some of the piles had the cap already
built, they were tested under two different condi-
tions: a) from the upper face of the cap and b) from 5 7m
a concrete brick expressly attached to the upper
sector of the shaft - beneath the lower face of the
c a p by means of special epoxy resin of quick cur-
ing due to the groundwater proximity.
Given that, the equipment FPDS-3 was used as
A .
f
57rr I
I
well as its corresponding software developed by
TNO - Building & Construction Research; the rec-
ords obtained were adequately “filtered” to take 0. Icm/ 5’6m
0 1 2 3 4 5 6 7 8 9 : G v2 a
into consideration the presence of the caps, in the Pile 65 4000P/s f 5 sr
indicated cases. The preliminary analysis of the 29 Jan 94 e x p 70
signals revealed a real alarming picture, as they
previously showed a number of piles with anoma- Figure 3. Pile #68; anomaly close to the 3m.
lies which easily exceeded- the usual expected per-
centages for this type of piles.
In the view of the possibility that the diagnosis
were influenced by a mistaken interpretation of the
geotechnical data provided by the Customer, the
performance of new reliable Geotechnical Studies
were arranged, as to being able to apply- in a sec-
ond stage- the Signal Matching Technique (21).
Likewise, and giving the worrying situation (be- 4 7m
cause it could lay open to discussion the diagnosis
by the use of technology that was rather unknown I v2 0
in our environment at that time), the following a 1 2 3 4 5 6 7 a 9 1 0
P i l e 12 4000m/s f. a sr
measures were adopted: 26 Jan 94 exp. 5
__
a) Execution of a new Geotechnical Study, as it
is described above. Figure 4. Pile #12; without anomalies.
b) Take new signals in the total number of piles.
c) Apply a Signal-Matching Technique on some iii.- From the number of Simulation Analysis
of the observed piles. done, on the behalf of the TNO specialists as well
d) Ask for a second independent opinion to as the author and his colleagues- in which the signal
TNO- Building & Construction Research, in accor- matching techniques were applied through the
dance with the agreement we kept with the men- TNOWAVE program option SIT-SM -,we arrived
tioned Institute. to the conclusion that most of the “suspicious” piles
As a consequence of the previous, we arrived to presented anomalies that were incompatible with
the following conclusions:
their security (stability and life time). The respec-
i.- The original geotechnical study contained no
tive graphics are included (Figures 5 and 6).
mistakes that would have led to erroneous conclu-
sions when modelling the soil to apply the iv.- Based on the conclusions that were obtained,
TNOWAVE program option Signal Match. we came to agreement with the main building con-
ii.- The repetition of the Sonic Integrity Tests tractor: to carry out the excavation and the drain of
showed that the preliminary diagnosis, derived the staircase and elevator sector of Tower A in or-
from the first series of signals, kept its force; it is der to be able to observe the condition of the piles
summarised as Table 2 shows. To illustrate the ex- involved. In spite of the difficult work, they were
ample, some typical signals corresponding to “piles performed efficiently, allowing the personal obser-
with anomalies” (Figure 3) and “normal pile” (Fig- vation- on the behalf of all the people involved in
ure 4) corresponding to Tower A are enclosed the construction- of the anomalies which their pre-
diction, based on the Sonic Integrity Tests and the

764
 Figure 5 . Signal Match of pile #68 (anomalous).

Figure 6. Signal Match of pile # I 2 (normal).

adequate application of the Signal Matching Tech-

niques, allowed the early detection of serious
problems derived from the lack of geometric-
structural integrity of a large number of piles in
both buildings (see photographs 8 and 9).
v.- In order to solve the structural problems, the
replacement of the defective piles through the in-
stallation of additional ones using the following
construction methodology was determined:
a) Installation of a temporary open ended caisson
by means of a vibratory hammer up to the indicated
depth by the complementary studies.
b) Inside cleaning.
c) Installation of the reinforcements.
d) Casting concrete with tremie pipe.
e) Retrieving the caisson with vibratory hammer.
f ) Control of the substitute piles through the
Sonic Integrity Tests (SIT).
vi.- The substitute piles were exposed to the
Quality Control Tests by means of the indicated
technique, and the obtained results are showed in
the table:
Photograph 9. Characteristics of the damages.
Table 3. Summary of the additional piles.
TOWER PILES k Case History I1
N" Total w/Anomalies
A 49 0 Corresponds to a group of three luxurious buildings
B 23 0 of thirty stories high built in the north of Buenos

765
Aires city (Argentina); the Architecture Project
foresaw the building of a basement that held the
totality of the place (a block). Taking into account
the bad quality of the upper soil layers (see Figure 4.61~
7 ) it was decided that the foundation should be with
displacement cast-in-situ piles with extended bulb 0.I c d
(Franki type). 4.681
For building reasons, the piles were installed
from ground level and casted up to the basement 0.lcn/s
level, which was located beneath the level of gen- 0 1 2 3 4 5 6 7 8 9 1 0 v2.0
P i l e 581 4000a/s f: I sr
eral excavation and close to the groundwater level. 21 dug 94 exp: 20
Because of budget reasons, excavations for the caps
and foundation beams construction were done by Figure 8. Pile #55:anomaly at 1.5m
reduced sectors- without a general groundwater
abatement.
for the already mentioned reasons, was taken as
model pile (Figure 9).

O.lCrn/ 3’7m I
0 1 2 3 4 5 6 7 8 9 1 0 v2.0
P I ] @ 38 4000m/s I: sr
17 ADP 94 exa: 2

Figure 9. Pile #38: without anomalies.

Figure 7. Existent Geotechnical Profile in the building site In accordance with the usual confrontation of
opinions (the Engineer-Main Building Contractor-
Foundations Subcontractor-Geotechnical Consult-
That brought the impossibility of an appropriate ant-Structural Advisor-etc.) when a customer is in-
coordination that allowed carrying out and efficient formed that in his site a pile with anomalies has
quality control of the totality of the piles because of been detected.
which, the Engineer and the Construction Company The conclusion obtained was similar to the one
determined that the work was done in a partial and obtained previously, as Figures 10 and 11 show.
random way as the progress of the work plan of the
building will allow, thing that gave place to a sta-
tistical pseudo-control.
In accordance to the previously mentioned, a
small amount of Sonic Integrity Tests were done
through the usage of the already mentioned FPDS-3
equipment. Nevertheless the limited quantity of
available signals, made the adoption of the statisti-
cal standard signal corresponding to the family of
records representative of the construction (or the
sector) problems, anomalies in several of the regis-
tered signals were detected. The main one is
showed next (Figure 8).
The Simulation program TNOWAVE option
SIT-SM (Signal Matching Technique) was applied
on that signal and the pile NO38 that is located near, Figure 10. Signal Match of pile #55 (anomalous).

766
 Not-withstanding the previous, and in the view
of the disbelief showed by the Engineer, we pro-
posed the excavation of the pile and it could be
seen a series of anomalies (deep craks, loss of
coating, etc.) which are indicated in Photograph 10.
quality of the soil in the building site (see Figure
12) the planners decided that the foundation should
be piled. Franki type piles were adopted.

Figure 1 1. Signal Match of pile #38 (normal).

Figure 12. Geotechnical profile

Figure 13. Pile #1: anomaly at 5.501~1

+--@--
+ I
Photograph 10. Pile #55. View of the observed damages.

P Case History 111 er

This construction site, which took place in the north
zone of the Gran Buenos Aires (Argentina), con-
stituted by a number of 10 storey towers destined to Figure 14. ano'nalyat 8.10m.
economic housing. Taking into account the bad

767
 -
Figure 15. Pile #44: without anomalies.
I
The total number of the piles of the construction
(420) was tested as the Foundation Company was
completing the works corresponding to each of the
towers. They had previously done the trimming
work and cleaning after seven days of casted, in or-
der to obtain an adequate resistance and to be apt to
transmit the stress wave (SW) generated by the im-
pact of the hammer. Only two piles, No 1 of the
Tower C2 and the No 154 corresponding to Tower
A, thrown anomalous signals so a more detailed
analysis through the application of the Signal
Matching Technique had to be done. Due to the
proximity of the observed piles, Pile 44 from the
Tower C2 was adopted as "model pile"; the respec-
tive signals are showed next: (see Figures 13 to 15).
As a consequence of the application of the
TNOWAVE program SIT-SM option, the graphics
that shows figures 16 to 20 were obtained.
The detailed analysis allowed us arriving to the
following conclusions:
a) Pile 1 (Figs. 13 and 16). This is a complicated
signal because the original pile, as a consequence of
having problems with its installation, had to be re-
placed by redriving the caisson in the same place
with the concrete of the first one still fresh; as a
-eoo1 , . , , I , , , I I consequence, the pile shape turned to be highly ir-
0 0 20 1 0 6 0 6 0 100 120 260 160 I80 200
regular showing increases and reductions in its di-
----- CDlLUldlCB an0 - - - m?aI"PE(I "ClOCillCS sec I 10-3
.~
ameter. The signal match determined a variable in-
Figure 16. Signal Match of pile # I .
crease in the pile diameter- from 350mm to 450mm
in relation to the depths between 2.5 and 5.2m-
followed by an abrupt reduction to the nominal di-
ameter starting at 5.6m. The corresponding reflec-
tion of the pile top is clearly marked at an approxi-
mately depth of 13.0, while the signal that is
detected at a depth of 11.5m corresponds to the
repetition of the previous signal of increase and the
subsequent section reduction.
b) Pile No 154 (Figs. 14 and 17). The Signal-
-1 2 t -1 Match (SM) established the presence of a "larger
anomaly" to an approximately 8.1m corresponding
-2.01 , , , , , , , J
to a discontinuity, reason why the stress wave that
00 20 4 0 69 8 0 100 12.0 140 160 IS0 ZWO was generated could not continue its way further
.___. - .-
C*,C"llrfCd an0 P P 0 ~ " P C ~"CIOCIIIES 5rC I 10-3
than that depth, giving place to the "reflection cor-
Figure 17. Signal Match of pile #154.
responding to the end"; the rest of the diagram cor-
responds to the first repetition of the signal.
From the Sonic point of view and consequently
from its capability to transmit shear stresses, is as if
the pile were 8.lm long even how, probably, it was
driven to its nominal depth, but that discontinuity
could have been a crack or an important soil inclu-
sion as it happened in the Case History I. The
problem was solved when the Engineer arranged
the replacement of the pile by other two of smaller
dimensions and the modification of the pile cap and
foundation beams.
-2 0
0 0 2 0 4 0 8.0 0 0 10.0 I20 140 160 100 a 0

Figure 18. Signal Match of pile #44

768
6 CONCLUSIONS. neering. Design and practice guide”. (Thomas Telford,
London, 1996).
The results obtained in the Quality Control done (9).- Institution of Civil Engineers: “The Value of Geotech-
since the beginning of the application of this tech- nics in Construction”. (Construction Research Communi-
nology, in Argentina as in the rest of the world, al- cations, London, 1998).
low establishing- without doubt- that the Sonic In- (10) Love, A.E.H.,: “A Treatise on the Mathematical Theory
tegrity Tests are an essential tool to know the of Elasticity”. (4th. Edition, New York, Dover Publica-
condition of a pile, new ones and old ones (case of tions, 1944).
pre-existent old foundations recycling or in cases of (1 1) Timoshenko, S. & Goodier, J.N.: “Theory of Elasticity”.
Forensic Engineering) (2nd. Edition, McGraw-Hill Book Company, Inc. 195 1).
As in every indirect investigation technique, (12) Timoshenko, S. & Young, D.H.: “Vibration Problems in
where the diagnosis is obtained through analyzing a Engineering”. (3rd. Edition, D. Van Nostrand Conpany,
measurements with sensitive instruments or image Inc. 1955).
(13) Voitus van Hamme, G.E.J.S.L.; Jansz, J.W.; Bomer, H.
analysis of signals corresponding to physic-
& Arentsen, D.: “Hydroblock and improved piledriving
mechanic phenomenon, the unavoidable require-
ments are: anilisis”. (De Ingenieur, no 18, ~01.86,1974).
> The availability of appropriate equipment, with 14) Middendorp, P. Private communication, June 2000.
(1 5 ) Beim, G. K., Private communication, June 2000.
proved and reliable hardware and software.
(1 6) Thorburn, S. & Thorburn, J.Q.: “Review of the problems
P Skilled operators
associated with the construction of cast-in-place concrete
P Periodic calibration of sensors.
piles”. (DOE and CIRIA Piling Development Group, Re-
P An exhaustive and reliable Soil Investigation.
port PC2, London, 1977).
P An adequate control of the execution of the
( I 7) Healy, P.R. & Weltman, A.J.: “Survey of problems asso-
foundations.
ciated with the installation of displacement piles”. (DOE
P Full knowledge of the Foundation Engineering,
and CIRIA Piling Development Group, Report PG8, Lon-
Geotechnical Engineering and Foundation Dynam-
don, 1980).
ics by the Engineer in charge of the Integrity Tests.
(18) Fleming, W.G.K.; Weltman, A.J.; Randolph, M.F. & EI-
P Independent criteria and ethical integrity on its
son, W.K.: “Piling Engineering”. (Surrey University
behalf.
PresdBlackie and Son Ltd., London, 1985).
(1 9) Turner, M.J.: “Integrity testing in piling practice”
(CIRIA Report 144, London, 1997).
7 BIBLIOGRAPHIC REFERENCES.
(20) Coldcmberg, J.J. & Goldemberg, H.; “Patologia en
fundaciones profundas. Origen y clasificacion”. (Memorias
( I ) Bredenberg, H. (Editor); “Application of Strcss-Wave
Theory on Piles” (Proceedings of the International Seminar del V Congreso Iberoamericano de Patologia de las
on the Application of Stress-Wave Theory on Piles, Stock- Constnicciones, VlI Congreso de Control de Calidad -
holm, Sweden, 1980). CONPAT 99, Montevideo, Uruguay, 1999).
(2) Holm, G.; Bredenberg, H. & Gravare, C.J. (Editors); “Ap- 6) (21) Goldemberg, J.J. & Goldemberg, H.; “Confiabilidad
plication of Stress-Wave Theory on Piles” (Proceedings of de 10s ensayos de integridad sobre pilotes por mCtodo
the Second International Conference on the Application of sonico - SIT”. (Memorias del I Congreso Paraguay0 de
Stress-Wave Theory on Piles, Stockholm, Sweden, 1984). Ingenieria Geotecnica, IV Jornadas Gcotkcnicas
(3) Fellenius, B.H. (Editor); “Application of Stress-Wave
Estnicturales, 1 ra. Reunion de Ingenieria Geotecnica del
Theory to Piles” (Proceedings of the Third International
MERCOSUR - 1er. COPAINGE, Asuncion, Paraguay,
Conference on the Application of Stress-Wave Theory to
Piles, Ottawa, Canada, 1988). 1997).
(4) Barends, F.B.J. (Editor); “Application of Stress-Wave
Theory to Piles” (Proceedings of the Fourth International
Conference on the Application of Stress-Wave Theory to
Piles, The Hague, The Netherlands, 1992).
5 ) Townsend, F.C.; Hussein, M.H. & McVay, M.C. (Editors);
“Stresswave ‘96” (Proceedings of
the Fifth International Conference on the Application of
Stress-Wave Theory to Piles, Orlando, Florida, USA,
1996).
(6) PEDIR INFORMACION A SUSUMU. Stresswave
‘2000” (Proceedings of the Sixth International Conference
on the Application of Stress-Wave Theory to Piles, San
Pablo, Brasil, 2000).
(7) - van Koten, H. & Middendorp, P.: “Testing of Founda-
tion Piles” (Delft University of Technology, Delft, The
Netherlands, HERON, vol. 26, No 4, 198 1).
(X).- Institution of Civil Engineers: “Creating value in Engi-

769
This Page Intentionally Left Blank
 Application of Stress-Wave Theory to Piles, Niyama 8 Beim (eds)02000 Balkema, Rotterdam, lSBN 90 5809 150 3

Author index

Abreu, L. 201 Dyminsky, A.S. 127 Justason, M. D. 609

Amir, J.M. 313
Andreo, C.S. 261,389 Ealy, C.D. 223 Kalinowski, M. 267
Antoniutti Neto, L. 707 Esposito, G. 575 Karkee, M. B. 689
Aoki, N.241,375,457,635 Kawabata, N. 335,583
Axelsson, G. 665 Fellenius, B. H. 3 13 Kikuchi, Y. 599
Ayasrah, I. M. 233 Ferreira, J.S. 657 Kimura, M. 563,569
Aye,Z.Z. 163, 171 FOB,S.B. 157 Kirsch, E 249
Fujita, K.335,451,469 Kita, N. 41 1
Balech, J. 241 Klingberg, D.J. 403,407,715
Barbosa, 6 . E. 429 Goble, G.G. 3,305,327 Korkeakoski, F? 415
Barends, E B. J. 541 Goldemberg, H. 345,7 19 Kormann, A.C. M. 707
Baycan, S. 75 1 Goldemberg, J.J. 345,719,757 Kusakabe, 0.335,591
Beim, J.W. 127 Gonqalves, C. 261,389
Benamar, A. 1 17,255 Gutikrrez, A. 201
Liang, R.Y. 121
Bernardes, G.P. 261,389,707
Hajduk, E.L. 541 Lee, J.-S. 421
Bettess, F? 495
Hannen, W.R. 503 Lee, M.W. 41
Boonyatee, T. 563,569
Bnino, D. 47 Hart, L.J. 223 Lee, S.-B. 99
Hartikainen, J. 415 Lee, W.-J. 99
Cai, L.B. 11 Hasard, D. 201 Lee, Y. -N. 42 1
Camapum de Carvalho, J. 157 Hayashi, M. 583,697,741 Legrand, C. 725,733
Cannon, J.G. 393,399 Hilmi Acar, M. 51 1 Liang, L. 53
Chambers, W.G. 407 Hintze, S. 665 Liang, R.Y. 461
Chamecki, E! R. 707 Hoffmann, Ch. 201 Likins, G.E. 205,211,327
Chen, B. 65 1 Holeyman, A.E. 479,725,733 Lima, E M.A. 375
Chen, R.P. 29,517 Holscher, P. 541 Liu Xi-An 683
Chen, Y. M. 29,5 17 Horiguchi, T. 689 Lucieer. W. J. 65
Chernauskas, L.R. 223 Huch, T. 249
Cho, C.W. 41,47 Husein Malkawi, A. I. 233 Mackenzie, P. 403,7 15
Cintra, J.C.A. 375,457 Hussein, M. 91 Maertens, J. 725,733
Costa, C. M.C. 429,441 Huybrechts, N. 725,733 Matsuda, Y. 187
Courage, W. M.G. 575 Matsumoto, T. 179, 187,335,583,
Cunha, R. I? 157 Imada, K. 179 59 1,697,741
Iskandarani, W. M. 91 Maung, A.W. 163,171
Danziger, B. R. 657 Michi, Y. 187
de Albuquerque, F? J. R. 677 Janes, M.C. 609 Middendorp, E! 55 1,609,617,625
de Campos, G.C. 429,435 Joer, H.A. 47 Morgano, C. M. 205
de Carvalho, D. 677 Jokiniemi, H. 415 Mukaddam, M.A. 9 I
DiMillio, A. E 223 Jonker, G. 135 Mullins, A.G. 609

77 1
Nakata, Y. 179 Rausche, E 53,59,75,205, Tatsuta, M. 599
Navajas, S. 429,435 21 1,327 Thasnanipan, N.163, 171
Navaneethan, T. 171 Restrepo, C. 219
Nawari, N.O. 121 Robinson, B. 53,59 van Foeken, R. J. 135,575,625
Nishimura, S. 335,411,591, Rodatz, W. 249 van Ginneken, G.J. J. 6 17,625
599,741 Romanel, C. 127 Viking, K. 533
Nishiumi, K. 591,741 Romell, J. 145
Niyama, S. 429,435 Roth, B.C. 503 Wakiya, Y. 583,741
Russo Neto, L. 707 Wu, J. 673
Oh, J.-H. 99 Wu, S. 383
Ohno, M. 469 Salumoto, J. 41 1
Okahara, M. 335 Seidel, J.P. 59, 193,267, 319 Xi Liang 369
Selby, A. R. 495 Xiao, L. 383
Padcowsky, S.G. 223,28 1,541 Shibata, A. 583,741 Xu, D. 383
Paraiso, S.C. 429,441 Skov, R. 107
Parente-Ribeiro, E. 127 Stenersen, K.L. 281 Yang Wu 153
Pereira, J. H. E 157 Stevens, R.E 17,351 Yoshizawa, Y. 583,591
Pinto Soares, E. 441 Sugimura, Y. 689
Pinto, EL. 275 Suzuki, M. 697 Zhang, F. 563
Piscsalko, G. 205 Svmkin, M.R. 35,107,113, Zhang, Y.-N. 683
PlaBmann, B. 249 503,525 Zheng, J. M. 65 1
Zheng, Y. M. 65 1
Ramshaw, C. L. 495 Takeda, T. 41 1 Zhou, G. 673
Randolph, M.E 41,47 Tanseng, I? 163 Zhu, B. 5 17

772