Professional Documents
Culture Documents
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
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
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
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 Gonplves, Cliiudio
Aoki, Nelson Mello, Jayme Ricardo de
Bernardes, George Kormann, Alessander Morales
Campos, Gisleine Coelho Merighi, Valmir
Cintra, Jos6 Carlos Navajas, S6rgio
Danziger, Bernadete Paraiso, S6rgio C.
Falconi, Frederico E Valverde, Skrgio
INTERNATIONAL,ADVISORY COMMITTEE
Altaee, Ameir - Canada Likins, Garland - USA
Amir, Joram M. - Israel Matsumoto, Tatsunori -Japan
Burbano, Germiin J. - Spain Mendiguren, Eugenio - Argentina
Fellenius, Bengt H. - Canada Middendorp, Peter - Netherlands
Fujita, Keiichi -Japan Paikowsky, Samuel G. - USA
Goble, George G. - USA Pinto, Paul0 - Portugal
Goldemberg, Juan J. - Argentina Randolph, Mark - Australia
Gravare, Carl-John - Sweden Rausche, Frank - USA
Gutierrez, Alvaro - Uruguay Sakai, Tomoaki - Japan
Holeyman, Alain E. - USA Seidel, Julian F! - Australia
Holloway, D. Michael - USA Towsend, Frank C. - USA
Hussein, Mohamad H. - USA Van Impe, William - Belgium
Kusakabe, Osamu - Japan Van Weele, A. E - Netherlands
Liang, Xu Ding - China 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:
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
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.
Today it can be said, with confidence, that the use of Probably the most remarkable work in developing
one dimensional wave propagation theory is gener- quantitative solutions was done by E. A. L. Smith,
ally accepted and widely used in the pile driving in- then Chief Engineer of the Raymond Company, at
dustry. The hallmark of these applications is that that time, the largest pile driving company in the
they have centered on usefil applications of the the- world. Before 1950, he published notes on the de-
ory. In this paper, two topics will be discussed, one velopment of a numerical solution to the wave
dealing with the closed form solution of the wave propagation problem applied to pile driving (Smith,
equation and the other, the discrete solution. Useh1 1950). Shortly aRer, a proprietary program was op-
results will be presented in each case. erating on an electronic digital computer. Smith
called this program the “Wave Equation” and it was
The one dimensional wave equation was derived and widely publicized (for example; Smith, 1957; Smith,
solved in the nineteenth century. Probably the most 1960). This program may have been the first appli-
important contributions to the use of the closed form cation of electronic digital computers to a civilian
solution were made by Professor Fischer of Upsula engineering problem.
University beginning in the early 1950’s. He devel-
oped many useh1 applications of the solution in- The implementation of wave equation analysis has
cluding several graphical approaches. A summary of been painstakingly slow. A public domain program
this work was presented at the Second Conference was developed in the early 1960’s at Texas A&M
on the Application of Stress Wave Theory to Piles University and it was widely used in the offshore in-
(Fischer 1984). This work laid the basis for the dis- dustry (Samson, et a1 1963). Today, the
3
GRLWEAPTMprogram (Goble & Rausche 1976,
GRLWEAPTM,1998) is widely used in the United i3x c dt
States and that usage is increasing rapidly. It is be-
coming standard practice to perform driveability where A is the cross sectional area of the pile, E is its
analyses both during the foundation design stage and elastic modulus, c is the velocity of wave propaga-
at the beginning of construction. However, outside tion, o is the stress at the top of the pile, e is the
the United States usage is much less common. 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-
2 PILE CUSHION SELECTION pression for the driving force, F, in Equation (1). It
is then written in terms of the displacement of the
A direct application 'of a closed form solution to the pile top and the resulting expression is rearranged
problem of selection of a pile cushion for driving a and expressed in terms of the pile top velocity ,v.
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 k k
v+-++-v =0 (4)
the problem shown in Figure 1. The pile driving Z m
system is modeled by three elements, a rigid ram, a
linear spring (cushion) and an elastic pile. The solu- In Equation (4), the weight of the ram has been ne-
tion provides the maximum delivered impact velocity glected and Z is the pile impedance, ENC. This fa-
induced at the pile top during the first wave passage miliar equation form has the solution
(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 v= (A sin o t + B coso t) (5)
some selected value.
The variables k/m and k/Z in Equation (4) have been
replaced by a2and 205, respectively, and
F=mg-muh (1)
4
Maximum values of v/vh were determined for a range used quantitatively so contours for four values of Z
of values of ram mass, cushion stiffness, and pile im- are given in Figures 3 through 6.
pedance.
In order to use the curves, trial cushion stiffness and
A parameter study was made for a 300 mm square ram mass is selected, and they are divided by the pile
pile with a full range of practical values of ram mass, area. Likewise, the pile impedance divided by the
cushion thickness, and concrete modulus. Values of pile area is also determined. Using the appropriate
the ratio of the pile top velocity to the ram impact figure, the pile velocity-ram velocity ratio is deter-
velocity, dvh, were determined for a range of values mined from the curve. This quantity is multiplied by
of ram mass and cushion stiffness, and four values of the anticipated ram impact velocity to obtain the
pile impedance. An.example of the results is shown maximum particle velocity transmitted to the pile top.
in Figure 2 where the surface of v/vh is given for a The particle velocity times the pile impedance gives
range of values of k and m with a specific value of the maximum impact force transmitted to the pile.
pile impedance, 2. Such a representation cannot be
5
Figure 7. Wave equation model
It should be noted that the model does not include There can be little hope of ever accurately knowing
the hammer cushion or the helmet mass. Further- the actual hammer efficiency prior to going to the
more, the result has not been tested against field ex- field, Even beginning driving the hammer efi-
perience and such testing must be done to assure the ciency will only be known when measurements are
usehlness of the results. available.
6
Damping constants have been selected based on tra- evaluated as dynamic but it is slower than the tradi-
dition using the original recommendations of Smith tional dynamic test.
(1960). They are selected based on soil type and
they show a wide range of accuracy based on signal The Statnamic test was modeled using a mass of the
matching analyses (Rausche et al 1994). This vari- typically used size as a “ram” in GRLWEAPTM. It
ability can strongly affect the results of wave equa- was dropped on the top of the pile with a cushion
tion analysis. Rausche et a1 (1994) suggested a dif- having a very soft spring. The spring constant was
ferent model that is based on the work of Coyle and selected by trial-and-error so that a force pulse of
Gibson (1970). Coyle’s research, fbrther proven by about 100 ms was generated. The magnitude of the
Herema (1979), indicates that the dynamic resistance peak force was controlled by the ram drop height.
has a strongly nonlinear relationship to the velocity of The test pile used in the study was a closed-end steel
pile motion. The study by Rausche et a1 indicated pipe with a length of 24 meters and an ultimate ca-
that the direct application of the Coyle research pacity of 4000 kN. An example of the force-time
caused numerical problems. A suggestion by and velocity-time record for one of the tests is shown
Rausche avoided those problems. in Figure 8.
R, = R,(1+ j,v)
R , = R,(1+ j g v N ) (9)
where j, is the Gibson damping constant and N is an Figure 8. Statnamic Test - Force and Displacement vs.
exponent, typically less than 1.0. The difficulty with Time
this representation is that numerical problems arise
when the velocity changes sign. Rausche suggested Four cases were studied, each with a range of applied
that damping be represented by forces. They are (a) Smith damping for sand, (b)
Rausche damping for sand, (c) Smith damping for
v clay, and (d) Rausche damping for clay. The results
R, = R,[1+ j R v r -- are given in Figures 9 through 12.
v, R,
where j, is the Rausche damping constant, vx is the In Figure 9 the force-displacement results for the
maximum velocity achieved up to a particular time case of the pipe pile in sand is shown. The wave
during the blow, and R, is the maximum static resis- equation results include the top force in the pile as a
tance actuated prior to the time under consideration. hnction of time together with the pile top displace-
Both the Smith and the Rausche model are available ment. The maximum applied force was 5200 kN. In
in GRLWEAPm. this case a maximum displacement of 55 mm was
achieved. The method that has been recommended
The Statnamic method (Janes et a1 1994) uses a for use in determining the static capacity was to use
charge of slow burning explosive to generate a force the force at the time of zero velocity and adjust this
between a large mass and the pile top. As the mass is force by the mass of the pile times the acceleration at
accelerated upward a downward acting force is in- that same time. This method was first suggested by
duced against the pile top. This force has a slower Nara (1970) and was used by the Case Research
rate of increase than the typical ram impact used in project (Goble & Rausche, 1970). It can be seen in
dynamic testing. Force pulses of the length of 80- Figure 9 that this method gives quite good results
100 ms are common. The resulting test must still be when compared with the specified static capacity of
4000 Kn. The applied forces were 5200 kN, 4840
7
kn, and 4180 kN. In all three cases the capacity at in a clay soil. The usual Smith damping constant
maximum displacement was about 4000 kN. The result is given in Figure 11. The predicted static ca-
adjustment for the acceleration at zero velocity (iner- pacity for all three Statnamic load levels are all quite
tia force) would be small due to the low mass of the close to the known static capacity.
pile.
Figure 12 shows the result for Rausche damping in
Rausche damping was applied to the same sand ex- clay for applied forces of 8000 kN, 7150 kN, 6070
ample and the results are shown in Figure 10. Forces kN, and 5450 kN. The associated predicted static
of 6400 kN, 5960 kN, and 5060 kN were applied. capacities are 6400 kN, 6400 kN, 5830 kN, and 5240
The results for the two larger forces, given by the kN, respectively. All of these values are quite high
capacity at zero velocity, are reasonably close to the with the largest 60 percent too large.
static capacity. However, the case of the smallest
applied force gives a predicted static capacity of This study showed that it was possible to induce
about 4800 kN, about 20 percent larger than the Statnamic-like forces using a wave equation analysis
given capacity. by GRLWEAPTMwithout modification for this par-
ticular application. The force time record is of the
Examples of the modeling of Statnamic tests in clay appropriate length and a similar shape. The force-
for the same pile that was analyzed above are given displacement curve has the appropriate shape and
in Figures 11 and 12. Figure 11 shows the results for appearance.
three Statnamic tests of different applied force levels
8
It is well-known that the Smith damping representa- Fischer, H. C. 1984. Stress wave theory for pile driving appli-
tion does not produce results that agree with experi- cation. lecture at the Second International Conference 011
ence. (Rausche et a1 1994) has shown that the Smith the Application of Stress Wave Theory on Piles, Stock-
damping constant does not agree with measurements holm, Sweden.
Goble, G. & F. Rausche 1970. Pile load test by impact driving.
for the assumptions used. Some results reported for Highway Research Record, No. 333.
Statnamic (James et a1 1994) show a character of re- Goble, G. G. & F. Rausche, 1976 Wave equation analysis of
sult that agree with the wave equation analysis. In pile driving - WEAP program. Volumes 1 through 4,
particular, the case presented with Rausche damping FHWA #IP-76-14.1 Through #IP-76-14.4.
for sand has the observed result. As the applied GRLWEAPTM- Wave equation analysis of pile driving 1998.
Goble Rausche Likins and Associates, Inc., Cleveland,
Statnamic force is increased the agreement between
Ohio.
the evaluation method used and the actual capacity is James, M., A. Sy, & R. Campanella 1994. A comparison of
improved. The conclusion has been reached that a statnamic and static load tests on steel pipe piles in the
substantial permanent set should be achieved. Fraser Delta. Proceeding of the Annual Svniposium on
Deep Foundations, Vancouver Geotechnical Society, Van-
The results presented here support the idea that the couver, B.C., Canada.
Heerema, E. P. 1979. Relationships between wall friction dis-
Rausche damping constant should be studied hrther
placement velocity and horizontal stress in clay and in sand
and that recommendations should be developed for for piled driveability analysis. Ground Engineering.
the required damping constants. 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
CONCLUSIONS Construction of Deep Foundations, Orlando, Florida, De-
cember.
Curves are presented to assist in obtaining cushion Rausche, F., G. Thendean, H. Abou-matar, and G. Goble
1994. Investigation of dynamic and static pile behavior
stiffness requirements for concrete piles to limit the from modified standard penetration tests, Federal Highway
induced compression stress at the top of the pile. Administration Report.
Impact stresses can be determined for a selected set Samson, C. H., T. J. Hirschjr., & L. L. Lowery, 1963. Com-
of ram mass, cushion stiffness, and pile impedance. puter study for dynamic behavior of piling. Journal ofthe
These values have been found for reasonable ranges Structural Division, ASCE, Volume 89, No. ST4.
Smith, E. A. L. 1950. Fundamentals of electronic calculation -
of the variables using a closed form solution of the
pile driving impact. International Business Machines.
one dimensional wave equation. They have not been Smith E. A. L. 1957. What happens when liammer hits pile.
checked against field measurements and such checks Engineering News-Record 159, September, 5.
should be performed. It should also be noted that the Smith, E. A. L. 1960. Pile driving analysis with the wave
analysis does not include the pile cushion and the equation. Journal of the Soil Mechanics and Foundation
helmet. Engineering Divisioii, ASCE, No. 86, August.
REFERENCES
9
This Page Intentionally Left Blank
Applicationof Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3
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.
11
A radiation damping dynamic model is built to However, the construction of pile influences the
resolve the energy dissipating in pile-soil inter- roughness. Two cases are taken into account: case
faces, where the energy is radiated rather than (I), for hand-dug belled pile, explosion is used to
consumed for soil shearing, shown as Figure 1. 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.
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;
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
15
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyarna & Beirn (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
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.
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 soil resistance, as shown in Figs. 3 and 4, was 295
continuous driving, as shown in Figs. 1 and 2, was kips (1.31 MN) for Pile 271 and 337 kips (1.50
about 91 kips (0.40 MN) for Pile 271, and 102 kips MN) for Pile 480, resulting in a set-up factor of
(0.45 MN) for Pile 480. Also shown are the about 3.3 in less than 5 days. Also of interest is
maximum compressive ( C S X ) and tensile (TSX) the shape of the soil resistance versus penetration
driving stresses. The lower compressive stress curve during redriving. The soil resistance
plotted is the average of two strain transducers, and decreases about 40 percent as the pile is redriven
the higher compressive stress is the maximum 3 feet, but is about twice the soil resistance
value obtained from either transducer. The piles measured at the end of continuous driving.
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
B5-2 A5-4 A5-2 A5-1 A1-4
3. CASE HISTORY NO. 2 enough to mobilize the full soil resistance. The
soil resistance mobilized during a series of redrive
Our second case history is for a series of redrive tests performed on a 1.6-m-diameter steel pipe pile
tests performed with a hammer that was not quite driven to a penetration of 26 m in a very silty clay
large enough to mobilize the full soil resistance. is presented in Fig. 7. The four lower bound soil
Sixteen 72-in.-diameter open-ended pipe piles resistance profiles are for continuous driving. The
were driven to 383-R (117-m) penetration with a pile was redriven by applying only two
Menck MRBS 8000 hammer. Soil conditions consecutive hammer blows with a PMJ-400
consisted of alternating strata of silty clayey sand hydraulic hammer after delays of 6 minutes, 15
or sandy clayey silt and stiff to very stiff sandy minutes, 33 minutes, 2 hours, and 66 hours. The
silty clay. resistance generally increases with time, but the
The distribution of the soil resistance was resistance after a 33-minute delay appears to be
determined from CAPWAP analyses performed for slightly smaller than the resistance mobilized after
a series of redrive tests. Results are presented for a 15-minute delay, and the resistance after a 66-
continuous driving at the end of driving for four hour delay is less than the resistance mobilized
different piles driven to final blow counts of only after a 2-hour delay. The maximum resistance
16 to 20 bpf. Results are also presented for the overcome by the pile driving hammer is about
beginning of redrive are for five different piles, 3300 kips (14.7 MN). The hammer is too small to
each driven about 3 inches. The soil resistance mobilize the full soil resistance during the redrive
along the length of the pile is presented in Fig. 5 tests. This is shown very clearly for the redrive
for continuous driving and after set-up times test performed after the 66-hour delay. Almost no
ranging fi-om 0.7 to 13.4 days. This plot shows soil resistance was mobilized over the bottom
that the full soil resistance was not mobilized for quarter of the pile.
the redrive tests performed after the two longest In our combined CAPWAP analyses, we
delays, as indicated by a change in slope in the soil have assumed that the soil resistance mobilized
resistance along the bottom 15 ft (5 m) of the pile. during continuous driving is a lower bound, i.e.,
A summary of the driving system performance the soil resistance mobilized on a particular pile
data is presented in Fig. 6. segment is assumed to be the larger of the actual
resistance mobilized or the resistance mobilized
during continuous driving. The soil resistance
4. CASE HISTORY NO. 3 mobilized after a 66-hour delay in the combined
CAPWAP analysis is about 4400 kips (19.6 MN),
Our third case history is for a series of redrive tests as shown in Fig. 8. In the standard CAPWAP
performed with a hammer that is clearly not large 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
To help put this in perspective, Fig. 9 energy results in only a 10 percent increase in the
shows that the resistance mobilized by a Vulcan maximum soil resistance overcome, and
060 hammer is about 3300 kips (14.7 MN), and the mobilizing a hammer having 150 percent more
resistance mobilized by a Vulcan 560 hammer is rated energy results in a 37 percent increase in the
about 4100 kips (18.2MN), assuming the same maximum resistance overcome.
soil and pile parameters. Comparing the three
hammers shown in Fig. 9, we see that by
mobilizing a hammer having 50 percent more rated
26
5. CONCLUSIONS
REFERENCES
27
This Page Intentionally Left Blank
Applicationof Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam,ISBN 90 5809 1503
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.
29
2 THEORETIC ANALYSIS
2.1 Theoretic model ofpile driving
(a) pile driving model (b) analysis model presented in shaft. is pile length and is the stress wave
by Chen (1996). No paper, shaft resistalce is velocity, which is described as c = 4~I p , where E
shaft resistance considered considered. is the elastic modulus and p is the density of pile
material. The reflection wave at pile-top is not
Figure I . Pile Driving Model considered. Up-going velocity is supposed to reach
pile-top without attenuation.
Figure 2. Point Resistance vs. Depth Estimated by The11 EqL1atiorr (2) 1s rewitten as
the Method (Chen 1996) [C+ z]V(t) + K [ U ( f ) [ I p ( t ) ]= 2Fd (I)
- (4)
or
As shown in Figure l(b), soil at pile shaft is [(’ + 2 1 9+ K [ l / ( t )-U,(r)]
= 2 F , (f) (5)
postulated to be rigid-plastic, and soil at pile-tip is dt
postulated to be elastoplastic and modeled by Smith The movement of pile-tip U(t) (or soil at pile-tip)
soil model (Smith, E.A.L., 1960). The soil at pile can be divided into three stages: elastic movement,
shaft is homogenous, and the shaft resistance plastic movement and rebound. When the movement
distributes evenly around the pile. The driving of pile-tip U ( f )is less than the maximum elastical
model is shown in Figure l(b). The driving force is movement of soil at pile-tip (denoted as Qp),there is
simplified to be a triangular impact force, as shown only elastic movement of soil. The static soil
in Figure 3. In Figure 3, F/Jis the amplitude of the resistance increases linearly with the elastic
driving force. tl is the loading time. In Figure I(b), movement of pile-tip until it reaches at the
R, is the static shaft resistance and R, is the static maximum static resistance R,. Equation ( 5 ) is
point resistance. C, is the damping coefficient at pile 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
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
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
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
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.
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), Figure 1. Pile capacity versus time for prestressed
Thompson & Thompson (1985), Thorburn & Rigden concrete piles in clayey soil, after Svinkin & Woods
(1980), Tomlinson (1971), Wardle et al. (1992), (1998a)
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 4 WAVE EQUATION ANALYSIS
well as dynamic testing (DT) yields the pile capacity
at the time of testing, Svinkin (1997, 1998b). By The main goal in using the wave equation method is
way of illustration, results of DT and SLT are to provide a better prediction of the pile capacity, as
shown in Figure 1 for two identical cylindrical, a function of pile penetration resistance, than can be
1372 mm x 127 mm, prestressed concrete piles, TP1 obtained from classical dynamic formulas.
and TP2. The depth of penetration of each pile was The wave equation method was originally
approximately 24.4 m. The soil consisted of about suggested by Smith (1960) to compute the pile
25.6 m of mainly gray clays followed by a bearing capacity at the end of driving. This method is also
layer of silty sand. The water table was at the used for prediction of pile capacity at restrike
ground surface. A Delmag D 46-13 hammer was (RSTR) performed at any time after EOID. By
employed for initial driving and restrikes. Each of adjusting wave equation analysis input with results
the piles TP1 and TP2 was tested 2, 9 and 22 days of dynamic measurements, some researchers , for
after the end of initial driving. The difference was example, Hunt and Baker (1988), York et al. (1994)
that three restrikes were made for TP1 and three have obtained good correlation between computed
SLTs were made for TP2. Pile capacity from three and observed pile capacities. However, in most
SLTs was a function of time as was the pile capacity other cases, computed pile capacity differs
obtained from DT. substantially from results of static or dynamic tests.
Existing dynamic models of the pile-soil system
mainly use a velocity-dependent approach for
3 DYNAMIC FORMULAS calculation of the dynamic resistance as a damping
component of the total resistance during pile driving.
The well-known dynamic formulas have been There are various linear and nonlinear relationships
criticized in many publications. Tested data in between the damping component and the velocity.
Figure 1 help to explain the causes of unsatisfactory For certain pile capacity, the dynamic resistance
prediction in pile capacity by dynamic formulas. depends only on pile velocity and the damping
Dynamic formulas using maximum energy, pile set coefficient.
and maximum displacement from DT do not take Statistical analysis indicates no correlation between
into account the time between SLT and DT. In the the pile penetration resistance and velocity values.
case of a few SLTs made on one pile, like three The pile-soil system changes with time after the
SLTs performed on pile TP2, what would be the completion of driving, but the pile velocity is only
reliability of pile capacity prediction by the energy a pile property and remains in the same range for
approach methods? Which SLT should be taken for EOID and RSTRs. The largest values of pile
comparison? Currently, there are no answers to velocity measured at the upper end of the pile and
these questions. In principle, dynamic formulas calculated along a pile shaft depend only on pile
cannot predict time-dependent pile capacity (Svinkin parameters and energy transferred to the pile and
1998b). cannot reflect regain in soil strength and pile-soil
36
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
38
equation analysis can be performed by the use of Theory to Piles, Ottawa, 25-27 May: 689-705,
variable damping as a function of time. Variable Vancouver: BiTech Publisher.
damping is the key parameter to enhance accuracy of Long, J.H., J. Kerrigan & M.H. Wysockey 1999.
wave equation solutions because this damping takes Measured time effects for axial capacity of driven
into consideration soil consolidation after pile piling. Proc. I999 Transportation Research Board,
installation. I 3 January, Washington, D.C.
The main criterion for accurate assessment of pile Randolph, M.F., J.P. Carter & C.P. Wroth 1979.
capacity prediction based on dynamic measurements Driven piles in clay - the effect of installation and
of force and velocity at the pile head during driving subsequent consolidation. Geotechnique, 29(4):
is the ratio of capacities obtained by dynamic and 361-393.
static tests. Such a ratio, taken for arbitrary time Rausche, F., G.G. Goble & G. Likins 1985.
between compared tests, in not a verification of Dynamic determination of pile capacity. Journal
dynamic testing results. of Geotechnical Engineering, ASCE, 1985,
Dynamic testing and analysis yield the real, not 11l(3): 367-383.
predicted, static capacity of piles at the time of Scov, R. & H. Denver 1988. Time-dependence of
testing. The static capacity from a static loading test bearing capacity of piles. In B. Fellenius (ed),
is not a unique standard for assessment of dynamic Proc. Third Inter. Conf. on the Application of
testing results. Both static loading test and dynamic Stress- Wave Theory to Piles, Ottawa, 25-27 May:
testing yields the pile capacity at the time of testing. 879-888, Vancouver: BiTech Publisher.
In soils with time-dependent properties, Seed, H.B. & L.C. Reese 1955. The action of soft
comparison of static loading test and dynamic testing clay along friction piles. Transactions, ASCE,
must be made only for tests performed immediately, 122: 731-754.
in short succession. Smith E.A.L. 1960. Pile driving analyses by the
Dynamic testing and analysis is a great tool to wave equation. Journal of the Soil Mechanics and
determine the time-dependent pile capacity after pile Foundation Division, ASCE, 1960, 86: 35-61.
installation. 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
REFERENCES Exhibition on Piling and Deep Foundations,
Bruges, 13-15 June: 1.11.1-1.11.8, Rotterdam:
Axelsson, G. 1998. Long-term increase in shaft Balkema.
capacity of driven piles in sand. Proc. Fourth Svinkin, M.R. 1995a. Pile-soil dynamic system with
Inter. Con. on Case Histories in Geotechnical variable damping. Proc. 13th International Modal
Engineering, St. Louis, 8-15 March, Paper 1.25. Analysis Conference, IMAC-XII, Beyond the
Briaud, J.L. and L.M. Tucker 1988. Measured and Modal Analysis, Nashville, 13-16 February, 1:
predicted axial response of 98 piles. Journal of 240-247, Bethel, Connecticut: SEM.
Geotechnical Engineering, ASCE, 114(9): 984- Svinkin, M.R. 1995b. Soil damping in saturated
1001. sandy soils for determining capacity of piles by
Chow, F.C., R.J. Jardine, F. Brucy & J.F. Nauroy wave equation analysis. Proc. DFI Annual
1988. Effects of time on capacity of pipe piles in Member’s Conference, Charleston, South
dense marine sand. Journal of Geotechnical and Carolina, 16-18 October: 199-216, Englewood
Geoenvironmental Engineering, ASCE, 124(3): Cliffs: DFI.
254-264. Svinkin, M.R. 1996a. Discussion of ’Setup and
Fellenius, B.H., R.E. Rker, A.J. O’Brien & G.R. relaxation in glacial sand’ by York et al., Journal
Tracy 1989. Dynamic and static testing in soil of Geotechnical Engineering, ASCE, 122(4): 3 19-
exhibiting set-up. Journal of Geotechnical 321.
Engineering, ASCE, 115(7): 984-1001. Svinkin, M.R. 1996b. Soil damping in wave
CRL and Associates, Inc. 1997. GRLWEAP - Wave equation analysis of pile capacity. In
Equation Analysis of Pile Driving, Manual, F. Townsend, M. Hussein & M. McVay (eds.),
Cleveland, Ohio. Proc. Fifh Inter. Con. on the Application of
Hunt, S.W. & C.N. Baker 1988. Use of stress-wave Stress- Wave Theory to Piles, Orlando, 11- I 3
measurements to evaluate piles in high set-up September: 128-143, Gainesville: University of
conditions. In B. Fellenius (ed.), Proc. Third Florida.
Inter. Con. on the Application of Stress-Wave Svinkin, M. R. 1996c. Velocity-impedance-energy
39
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
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.
41
shown in Fig. 1 were obtained for different pile and GRL,1996) analyses of dynamic measurements at
soil conditions. Thus dynamic measurements are EOID and restrike. Only the datum at the final re-
needed for the particular site, but, WEAP is mostly strike was used in cases where multiple restrike tests
used in preliminary design where dynamic data are were undertaken. Since many of the restrike tests
unavailable. Consequently it is desirable to suggest were not able to fully mobilize the pile capacity due
new soil input parameters to account for time de- to limitations of either hammer capacity or material
pendent characteristics. To do so, it is first necessary strength of the piles, the set-up factors were deter-
to analyze and appraise the soil parameters measured mined from the larger of (i) the ratio of failure load
for the same piles at both EOID and at any time after (interpreted from the static response computed using
driving. 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.
Figure 1. Change in damping coefficients with time CAPWAP is a trial and error method of signal
(after Svinkin and Woods, 1998) matching, which is known to lead to subjective, non-
unique, sets of parameters. As such, it is possible to
Statistical analyses of dynamic measurements fail to discover any distinct trends in the parameters,
(quake, damping coefficients, set-up factor) are pre- since each parameter determined from CAPWAP is
sented in this paper, which recommends new pa- influenced by the other parameters. In order to
rameters to consider time effects in wave equation minimize operator-dependent trends in the statistical
analysis and confirms the reliability of the values analyses, only data with high matching quality (MQ
recommended. Although time effects may be di- in CAPWAP) were analyzed.
vided into two types, set-up and relaxation, only the In order to confirm the validity of the parameters
former will be dealt with in this paper. 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
2 DATA AND ANALYSIS METHOD equation analyses were performed for a total of 24
pile tests at 6 sites (4 tests per site).
Dynamic pile loading tests for the same pile at dif- In wave equation analysis, adjustments of the
ferent times after driving were carried out at 28 sites. maxima of force (FMX) and the transferred energy
To avoid bias in the statistical analysis, a maximum (EMX) on the basis of the value analyzed by CAP-
of only 2 data points were used from any one site. WAP were made in order to increase the accuracy of
For the determination of the set-up factor, the ratio the compared data. The adjustment was considered
of bearing capacity at restrike to bearing capacity at sufficient when the calculated and measured EMX
EOID, 54 piles from 28 sites were investigated and values agreed to within 10 %, following the guide-
analyzed, while to estimate time effects on quake lines in FHWA (1996). Other input parameters (ex-
and damping parameters, 46 piles were analyzed due cept the soil parameters) were based on the CAP-
to the lack of some data. The piles comprised 18 WAP data and test conditions in the field. New soil
steel pipe piles, 23 concrete piles and 5 H-piles. The parameters were then used as the input data for
diameters of piles ranged from 3 0 mm to 600 mm. WEAP to match the measured drivability.
Time lags between EOID and restrike ranged from For the reliability analysis of the results of wave
1 day to 30 days (averaging about 8 days). Further equation analysis, using the new parameters drawn
details of the data are given by Cho (1998). from the statistical data, a log-normal probability
Data for statistical analysis were drawn from function was used. The log-normal function was
CAPWAP (Case Pile Wave Analysis Program, chosen to allow consideration of the non-
symmetrical Characteristics and distribution of the
42
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.
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.
All data
44
The dynamic parameters used in WEAP are not The ratio of the values calculated by WEAP to the
inherent properties of the soil but changeable ones measured ones using CAPWAP were then obtained.
with time or test conditions. Consequently, it is in- Finally the ratios were used for the reliability analy-
evitable that suggested values will not represent ac- sis using a log-normal probability function.
curately a given set of conditions in the field. Previ- The results of the reliability analysis using a log-
ous suggestions for the dynamic parameters in normal probability function are shown in Fig. 7 to
WEAP have been given by Reese et al (1964) and Fig. 10. In those figures, each initial stands for as
Coyle et a1 (1970). From this viewpoint, the recom- follows:
mended values in Table 4,which are based on reli- RI : mean,
able data from various sites, should prove reasonable SD : standard deviation,
for application in WEAP. However, it should be CV : coefficient of variance (=SD/M),
emphasized that the recommended values are at best ES : WEAP calculation with the parameters
average values for the sites in question and are suggested in WEAP at EOID,
strictly applicable only over the relatively limited RS : WEAP calculation with the parameters
range of pile diameters (76 % between 400 mm and suggested in WEAP at restrike,
500 mm). ER : WEAP calculation with the parameters
The set-up factors in Table 4 are the averages of recommended in this study at EOID,
the data excluding data beyond 1 standard deviation RR : WEAP calculation with the parameters
from the mean in the statistical analyses. Consider- recommended in this study at restrike.
ing that bearing capacities at restrike were mostly Function value : probability density distribution
determined from dynamic tests where the full pile Fig. 7 and Fig. 8 show that the capacity ratios
capacity was not mobilised, the set-up factors are re- based on the recommended parameters in this study
garded as conservative values. are closely grouped around unity, while those based
It is interesting that the suggested values in WEAP on parameters suggested in WEAP show a much
are more similar to those recommended at restrike in broader range. This suggests that the bearing ca-
this study rather than ones at EOID, although WEAP pacities predicted using the recommended parame-
was originally developed for drivability analysis at ters are more reliable than those using the suggested
EOID. The recommended values in Table 4 can be ones in WEAP.
appropriately adjusted with site conditions, and are From Fig. 9 and Fig 10, it is found that there are
particularly useful for the determination of bearing some differences in the predicted and measured
resistance at restrike. maximum force, FMX, using the suggested pa-
To analyze the reliability of the recommended pa- rameters in WEAP, while the new parameters pro-
rameters, a comparison was performed between key posed here give improved accuracy. Either set of
predictions (such as pile capacity) calculated by parameters lead to good prediction of maximum en-
WEAP with the parameters recommended in this ergy, EMX.
study and those calculated by WEAP using parame-
ters suggested in the WEAP manual.
2.0
Soil with
2.0
~
0.65
0.65 ~
DnO
D/80 ~ 1:; ~
( .O)*
* : (1
Soil
I 2.0 I 0.65 I D/60 I 0.25 2.0 0.65 D/90 0.5 1.5
45
5 CONCLUSIONS
REFERENCES
46
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) K3 2000 Balkema, Rotterdam, ISBN 90 5809 150 3
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.
47
Figure 2 Conical tips for closed-ended piles
48
and 22 - 36 m (10 %) for Sample 1, and 0 - 11 m The model piles were driven continuousPy be-
(4%), 11 - 1 4 m ( l O % ) , 1 4 - 2 4 m ( 6 % ) a n d 2 4 - tween one test penetration and another. The blow
34 m (10 %) for Sample 2. Four cone penetration count for each pile was recorded by monitoring the
tests were carried out ‘in-flight’ in each sample, us- number of times the hammer was fired between each
ing a 7 mm diameter cone. These showed excellent sampling point. Two ram masses were used in the
repeatability, and gave the average cone profiles centrifuge tests. In Sample 1, the first two open-
shown in Figure 3. As may be seen, the ranges of ended piles were installed with a 37 g mass; this was
cone resistances for the different cement contents are later replaced with a 54 g mass, which was used
13 to 18 MPa, 21 to 30 MPa and 35 to 60 MPa re- throughout the remainder of the testing programme.
spectively (although it appears that there may have In the centrifuge model, the scaling law for mass
been some shortcomings in preparing the base layer is given by:
for Sample 1).
49
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
50
Figure 8 Driving record for over-sized closed-ended pile, Figure 9 Profiles of dynamic and static resistance, Sample 1
Sample 2
51
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
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
7 CONCLUSIONS ‘94,Singapore, 473-478.
Randolph, M.F., Jewell, R.J., Stone, K.J.L. & Brown, T.A.
This paper has described a study of the drivability of 1991. Establishing a new centrifuge facility. Proc. Itit. CO$
open and closed-ended piles into cemented calcare- Cetirrifuge 91, Boulder, Colorado, 3-9.
ous soils. The study was undertaken with a view to Rickman, J.P. & Barthelemy H.C. 1988. Offshore construction
of grouted driven pile foundations. Proc. Int. Cot$ on Cal-
assessing the feasibilityof drivingclosed-ended piles careous Seditiietits, Perth, 1 , 3 13-319.
in such soils, with or without post-installation
52
Application of Stress-Wave Theory to Piles, Niyama Ei Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
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.
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 automated CAPWAP procedure produces capacity
calculation is repeated with a corrected set of soil results that are very similar to those obtained by
model parameters. Obviously, the more realistic the experienced engineers working interactively on a
soil model, the better its capability t o match the computer. On occasion, however, the automatic
measured quantities. On the other hand, a very method calculates an unsatisfactory resistance
sophisticated soil model may have too many distribution near the pile toe. An experienced
unknowns and may not be uniquely defined by the analyst must therefore always check the solution by
matching process. For that reason, the relatively means of additional trial analyses. For the
simple Smith soil model (Smith, 1960) has been simplification of the interactive matching task the
most successfully employed for pile dynamic signal CAPWAP program does provide difference
matching, minimization routines for individual soil resistance
The traditional iterative matching procedure can parameters.
be summarized as follows:
1. Data Input: select a record with appropriate
energy and data quality 3 MATHEMATICAL MODEL
2. Data Check and adjustment (normally
automatic) The pile is modeled as a series of uniform, elastic
3. Build pile model (normally automatic) segments, typically 1 m long, of equal stress wave
4. Check and change resistance distribution travel time. Calculations involve tracking the
5 . Recheck data adjustment upward and downward traveling waves and their
6. Check damping parameters reflections where segment properties change or soil
7. Check quakes and unloading parameters resistance effects exist. The simplicity and strictly
8. Find absolutely best match quality elastic nature of this model is, unfortunately, a
9. Produce output disadvantage when modeling non-linear or non-
An important part of the matching procedure is the elastic situations such as cracks in concrete piles or
evaluation of the match quality, i.e. quantiQing the certain types of mechanical pile splices.
difference between measured and computed The effect of the soil, resisting the pile motion, is
quantity. In CAPWAP the match quality is the modeled as a series of N lumped forces at intervals
normalized, weighted sum of the absolute values of not greater than 2 m which depend on pile velocities
the differences between computed and measured and displacements. The parameters of this Smith
values of all analyzed time steps. Normalization is soil model are the unknown quantities that
achieved with respect to both maximum pile top CAPWAP must determine. In the standard analysis
force and the number of data points. The match over situation, the displacement dependent (static)
a 3 ms time period, following the first return of the resistance forces are represented by both a stiffness
stress wave from the pile toe, is given a double and a capacity value. The velocity dependent
weight because of its importance for total capacity (dynamic) resistance forces are calculated using a
determination. damping factor. For the resistance forces acting on
CAPWAP can either be used in an interactive the shaft, soil stiffness and damping parameters are
mode or automatically. The automatic procedure chosen proportional to the static capacity values and
searches for a best match using a step by step in this way, the number of shaft unknowns is kept to
procedure that is also recommended to the analyst a manageable N + 2 values. For the toe an
for interactive signal matching. In other words, the additional 3 unknowns have to be determined. To
automatic CAPWAP is not a standard minimization produce a good signal match over a long record time
software which would search in a more or less period several additional parameters had to be
random manner for a set of soil parameters that defined. The most important ones allow for a
produces a minimum difference between computed modification of the static soil stiffness and plastic
and measured pile top variable. Experience has limit (upward directed capacity) for the rebound
shown that such a relatively mindless procedure may phase of the pile and therefore have little or no effect
lead to unsatisfactory results. On the other hand, the on the calculated total static pile capacity.
54
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).
In earlier versions of CAPWAP, the analyzed record consecutive blows, this condition requires that all
length was generally set to 25 ms after the first points along the pile achieve the same final set equal
return of the impact wave from the pile toe. This to that at the pile top. It is reasonable to require that
relatively short record length saved computer time the CAPWAP signal matching process also produces
but did not always allow for an accurate calculation a match of calculated with observed pile top set; in
of the final set. Today modern personal computers other words, the average of all the sets of all
and more sophisticated operating systems provide segments equal the observed set. For lack of
the analyst with high computational speeds and huge computing power this requirement had not been
memory space at a low cost. It has therefore become imposed on previous analyses. (It should be added
possible to economically analyze dynamic pile that the most accurate method of calculating the final
records over longer time periods and to perform set of the pile is a residual stress analysis (RSA),
many more trial analyses for more reliable results. which repeats the analysis several times thereby
The longer analysis time period assures that the calculating the stresses locked into pile and soil.
calculation can be carried out until the pile stops This analysis method is available as an option in
moving, i.e. until the pile velocity becomes zero and CAPWAP. Regrettably, RSA is infrequently used
the displacement has reached the final set. To be because of its complexity.)
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 6 PROGRAM PERFORMANCE
top force, velocity and displacement record whch
includes major vibrations after 100 ms. The record The enhanced analysis procedure is available in the
also indicates a final displacement value that Windows based CAPWAP. To check whether or
matches the pile set or the inverse of blow count. not the blow count matching (BCM) procedure
Ideally, the pile top set per blow is verified by would help improve the program performance, most
independent observation so that the double of the data presented by Likins et al. (1996) was
integration of the acceleration can be checked and, if reanalyzed. The reanalysis involved adjustment of
necessary, corrected. This two point data check and the acceleration values for both a zero final velocity
adjustment process is automatically performed in and a final displacement matching the observed
CAPWAP by slightly shifting the zero line of the blow count at a later record time. The soil model
acceleration. This zero line shift starts at impact. has parameters were then re-initialized to reset all soil
a somewhat higher magnitude for a few milliseconds parameters to the CAPWAP start up conditions.
and a lower magnitude over the remainder of the Any modifications to the pile profile were removed,
record. The magnitude of this zero line shift is only the analysis count was reset to 1 and the initialized
a small fraction of the maximum measured pile top data file was saved. These steps assured the same
acceleration. results would be obtained as if the analysis had been
started from the very beginning. With the
reinitialized data file the automatic procedure was
5 NEW DEVELOPMENTS then run first with blow count matching and then
again without blow count matching. The analyst did
With a record that correctly double integrates to final not interfere with the automatic analysis process in
set and with computational speeds allowing for an any way. Capacity, computed final set and match
analysis to final zero velocity at the end of the quality results were then subjected to a variety of
record, the final pile set or its inverse, the blow correlation studies. As in the 1996 study,
count, can be calculated for all segments of the pile. correlations were done using the ratio of CAPWAP
As long as the final set is the same for several divided by static load test capacity. Also, the time
55
Table 1. Calculated CAPWAP capacity divided by static load capacity at
different time ratios
8 MATCH QUALITY
56
Figure 4. Computed to observed set comparison with and without blow count matching
0.28
Mean
57
factor to account for dynamic resistance losses. A in the measurement of restrike blow count. It is
justification for this approach is the difficulty of believed that lack of accurate blow count
separating static from dynamic components by measurement limited the improvement over the
signal matching when the pile displacements are traditional CAPWAP approach. On the other hand,
small. The static, displacement dependent the signal matching process itself already
components then differ littfe from the damping, incorporates blow count matching to a certain
velocity dependent components which easily degree, since the measured velocity and therefore the
introduces errors in the calculation. Worse yet, in a top displacement are imposed as top boundary
hard or very dense soil or in a rock, the static toe conditions. The improvement achieved with the new
resistance components sometimes appear to be CAPWAP program should therefore be primariiy
velocity dependent and could therefore be attributed to a more accurate analysis over a longer
misinterpreted as damping resistance by the time period.
traditional CAPWAP approach, leading to an As in earlier correlation studies, the time factor
underprediction of static capacity. To check for (time between load testing and installation divided
possible improvements in capacity prediction, by time between restriking and installation) proved
various methods of i~terpretationof the total shaft to have the greatest effect on the accuracy of the
and/or toe resistance values were explored. Table 3 CAPWAP prediction. Obviously waiting times
shows the most promising method which adds the comparable to those of the static test assures the best
calculated static shaft resistance to the total, factored possible prediction of long term bearing capacity by
toe resistance. Sorting the results by dominant soil CAPWAP.
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 REFEENCES
would be 1.02 with a COV of 0.24. Further
exploration of this method is warranted. However, Balthaus, H., 1986. Zw bestirnmung der tragfaehigkeit von
pfaehlen mit dynamischen pfahlpruefmethoden.
at this time too little experience is available (only 14 Dissertation, TU Braunschweig, Germany.
cases for the sands) and the time factors should also GobIe Rausche Likins and Associates, Inc., 1999,
be considered in future studies. CAPWAP background report, Cleveland, OH.
Hannigan. P.> and Webster, S. W., 1987. Comparison of
Table 3 : CAPWAP static shaft resistance and total static load test and dynamic pile testing results, Second
toe resistance divided by static load test capacity for Int. Symposium, Deep Foundations Institute, May.
different soil types 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.
Clay 6 1 1.30 0.50 I 0.38 Klingmiiller, O., 1984. Dynamische pfahlpruefung also
SandvClav
4 2
4 1 1.23
t
0.22 I 0.18 nichtIineare systemidentifikation: dynamische probleme -
modellierug und wirklichkeit. tagungsband. Curt-Rsch-
SiltyClay t 8 1 1.08 1 0.20 0.18 Institut, TU Hannover, Oktober.
Rock I
6 1 1.15 I 0.18
1 1
0.15 L k n s , G., Rausche, F., Thendean, G., and Svinkin. M.,
14 1 1.02 0.25 0.24 1996. CAPWAP correlation studies. Proceedings of the
8 1 1.03 0.24 0.24 I jthhnt. Cod. on the application of stress wave theor). to
1
piles, F. Townsend ed., Orlando, Fiorida.
8 1 1.12 0.24 0.21
Mure, J.N., Kightiey, M.L.. Gravare.C.J.t and Hermansson,
4 1 1.93 1.51 0.78 -
1983. CAPWAP an economic and comprehensive
I 5 1 1.071 0.15 0.14 alternative to traditional methods of Ioad testing of piles,
Paper 16, Piling and ground treatment for foundations,
Thomas Telford, pp 167- 174.
11 CONCLUSIONS Rausche, F.: Goble, G.G., and Moses, F., 1972, Soil
resistance predictions from pile dynamics, JozirnaI ox the
Soil Mechanics and Foundation Division, ASCE.
The correlation between CAPWAP predicted pile Reiding, F. J., Middendorp, P., Schoenmaker. R.P.,
bearing capacity and static load test capacity can be Middeldorp, F.M., and Bielefeld, M.W., 1988. FPDS-2:
improved if not only the difference between A new generation of foundation pile diagnostic
computed and calculated pile top quantity but also equipment, Proceedings, Third Int. Cod. on the appl. of
the difference between calculated and observed blow stress wave theory to piles, Bengt Felienius, ed., Ottawa,
count is included in the match quality evaluation. Canada.
The improvement over the traditional method, which Smith, E.A.L., 1960. Pile dnving analysis by the wave
equation, Journal of Soil Mechanics and Foundations,
ignored the calculated blow count, is significant and ASCE, 86>August 1960,35-61.
since, with modern computers, the additional
computational effort is minor, blow count matching
should always be done. Great care should be taken
58
Application of Stress-Wave Theory to Piles, Niyama & Beim [eds) (9 2000 Balkema, Rotterdam, ISBN 90 5809 150 3
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..
59
for File type selection.. Instead, GRLWE.AP's 1. To evaluate the ability of a given pile driving
analysis is an additional effort in the preparation of hammer to drive the pile to the nominated capacity
the driveability analysis. and/or penetration;
2. To estimate the final blow count;
3. To evaluate the stresses which will be induced
2 G E N E R U CONSIDERATIONS in the pile during the installation process.
In such an analysis, a conservative choice of soil
Driven pile analysis is generally performed in two strength parameters has an ambiguous meaning. For
steps (Hannigan et al. 1996): an initial static analvsis instance, when assessing tension stresses or the
which is followed by a dynamic or wave equation required penetration, an underestimate of soil
analysis. Static analysis can be based on one of the resistances would usually produce a conservative
many proposed approaches, whch calculate shaft analysis, whereas in assessing, compression stresses,
resistance and end bearing for a particular pile an overestimate of soil resistances would generally
penetration depth. Depending on the quality of the be conservative. Assessment of the bearing capacity
available soil strength data, the variability of the soil which can be achieved by a given hammer may not
properties over the site, and the realism of the be significantly affected by the assumed distribution
calculation procedure, the reliability of the results of resistance. Because of this ambiguity, and
obtained from such analyses varies significantly. because the consequences of an incorrect assessment
Statically calculated capacities are indirect estimates of soil resistance distribution does not have the same
based on site soil parameters estimated from the implications as traditional static design with regard
foundation investigation, not based on the additional to the safety of the structure, it is most appropriate to
information provided by the installation process. undertake driveability analysis using the most
Because of this inherent shortcoming, engineers realistic assessment of soil strength parameters.
often prefer the use of a dynamic formula andior In the GRLWEAP static calculation method, it
wave equation analysis which provide a more direct was therefore attempted to avoid a static resistance
method of capacity evaluation for each individual bias as much as possible. Again, the assessment of
pile. In these approaches, blow counts from pile the pile's bearing capacity should always be done
installation observations, plus a calculated with prudent limits or reductions as dictated by
relationship between blow count and beanng experience.
capacity (the bearing graph) is used to make a pile Having calculated the static soil resistance based
capacity prediction. The driveability approach can on in-situ soil strength parameters, an estimate has to
also be used prior to pile installation to predict rather be made of the static soil resistance that is actually
than confirm the installation process. In this case, present during pile driving. For example, pore water
field observations of the installation process are not pressure changes in the g o u n d during pile
available to improve the reliability of the analysis installation tend to change the effective stress regime
procedure. and therefore the resistance acting on the pile.
In traditional methods of static pile analysis. the Unfortunately, no matter how accurately the soil
requirement that the analysis provide a safe design is exploration and static analysis was conducted,
paramount. Efficiency of design is something for estimating the Soil Resistance to Driving (SRD)
whch all engineers should strive, however. it is of based on its static capacity may generate significant
less importance than providing a safe structure. errors and although experience values are available,
Conservatism in the design is provided in several the selection of so-called gain-loss or capacity
ways: reduction factors is one that should always be
1. In conservative interpretations by the engineer carehlly reviewed. These uncertainties are greater
of the site soil strength parameters; for sensitive soils such as marine clays and it is
2. In upper limits on strength values imposed by common practice to estimate both a lower and an
the design method; upper bound SRD and calculate lower and upper
3 . In a conservative appraisal of the data on Lvhich bound blow counts by the wave equation.
the design method is premised; When pedorming a dynamic analysis using
4. In the application of safety factors or strength GRLWEAP, it is not only necessary to calculate the
reduction factors. static resistance and its distribution, additional
The last of these factors affects only the maximum dynamic soil resistance parameters, damping and
structural load which can be allowed on the quake, both at shaft and toe have to be estimated.
foundation element during its service life. However, Actually, GRLWEAP recommendations (Goble et
the first three of these factors ensure that most al. 1999) are rather simple for standard analyses
design methods applied in practice will where little is known about the soil. Based on these
underestimate ultimate pile capacity. commonly accepted rules, unless dynamic test
Driveability analyses are conducted as part of the results indicate otherwise, only the shaft damping
design process for a number of reasons: factor is a function of soil type. Toe damping is
60
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.
61
pile analysis. However, it contains information cases for which static pile load test was available in
necessary to expand the analysis methods in the a variety of soil and pile types. The proposed
future. GRLWEAP static capacity method was used. It is
The unit toe resistance in kPa is estimated noted that the predictions are automated, and thus
according to Meyerhof, 1976 using the simple operator independent. The ratio of predicted static
empirical expression: failure load to actual capacity as evaluated by the
Davisson offset method averaged 0.92, with a
cot = min(200N; 12,000)
coefficient of variation of 0.31 (see Table 1). By
where N is the SPT N value in the strata at the comparison, the same data was also subjected to
bottom of the pile. Of course, it is possible to analysis by the SPT97 approach which is popular in
improve this very simple method in future updaIes the South-Eastern United States (Townsend et al.
and as experience is gathered with this approach. 1997), which yielded a mean of 0.31 and a
Soil data should be entered where layers change, coefficient of variation of 0.39. Figure 2 shows the
where the water table is encountered, or where major capacity ratios of GRLWEM and SPT97,
changes in N-values occur; also, data must be demonstrating that the two methods show similar
entered when the soil type changes, and at minimum tendencies of underprediction or overprediction.
intervals of 3.5 m. This may be an indication of the failure of the basic
As discussed, the N-value and soil type is used to soil data, primarily the SPT -value, to provide
estimate unit weight, friction angle (effective hc:ion correct and sufficient soil strengqh information.
angle for clays) and horizontal stress based on Clearly, for the particular set of correlation cases
correlation studies by Kulhawy et al. (1939 and investigated, GRL’ltrEAP’s static method is slightly
1991), Schmertmann (1975 and 1973): and less conservative than SPT97 and also less scattered
Robertson and Campanella ( 1933). The lateral earth (Coefficient of Variation 0.3 1 .vs 0.39).
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. GRLWE,i\P SPT-97
Both calculated unit shaft resistance and end
bearing values are then input into the G R L W L V I Minimum I 35 I 33 I
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 I Standard Deviation I 29 1 32 I
these may be the plugged areas), which can be Coefficient of
immediately converted to toe quake. As discussed, Variation
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 5 Sb%&MY . & CONCLUSIONS
I
calculation method proposed by Fleming ( 1992).
This additional result requires estimation of the A static analysis method has been presented that has
settlement characteristics of the soil. Again. soil been incorporated in the dynamic wave equation
type and SPT -values are converted to provide the analysis program GRLWEAP. The method is based
necessary input. Obviously, calculation of on SPT -values plus an assessment of soil type. An
settlements is an even more challenging task than effort has been made to make this method:
bearing capacity prediction, particularly when only 1. hlly automated and therefore independent of
SPT information is available. These results must user judgement. Covering the complete range of
therefore be used with appropriate caution. X possible N-values and soil types.
correlation study on this result had not been 2. Unbiased, i.e. neither conservative nor non-
completed by the time t h s paper was finished. conservative for capacity, stress and blow count
predictions.
3. Generate automatically the dynamic soil
4 CORRELATION WITH STATIC LOAD TESTS
resistance parameters necessary for both bearing
graph and G R L W E A P whole-installation
Using data from the database presented by Liluns et driveability analysis.
al. (1996), a correlation study was conducted on 53
62
Figure 3. Comparison of G R L W ~ A Pwith SPT97 correlation.
4. Capable of calculating a prediction of pile-top Hannigan. P.J.. G.G. Goble. G. Thendean. G . E . . L h s . & F.
load-movement for a specific pile embedment. Rausche. 1996. Design and construction of dnven pile
Results from this analysis method compare foundations. U.S. DOT Federal Highuay .4timrnisiranon
Reporr No. FHW.4-41-96-03;,
favorably with similar methods. However. the user Kulhany. F.H.. C.W. Jackson. & P W. Ma\-ne. 1989. First
of this approach is advised of the shortcomings order estimation of KO in sands and clays. Forindanon
inherent in any static formula approach. Engineering: Ciirreni Principles anti Practices. Vol. 1 ~
63
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
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.
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
66
Equation 3 gives the maximum force F, of the force wave from element 1 arrives at the virtual
downward traveling impact wave originated in the plane. Even so the force with size F?, of the upward
plane between ram and pile. From pile driving meas- traveling wave from element 2 arrives.
urements of steel to steel contact between ram, pile Those two forces are the known forces. From those
cap (no cushion !!) and pile it is known that this two forces the two departing forces could be calcu-
equation is very close to reality. lated with equations 6 and 7. In these equations there
What could this equation say ? First, the force F, are two terms after the sign of equality, which are
is linear related to the impact velocity of the ram. contributing to the downward traveling wave (equa-
Second, with a chosen hammer, a particular ram tion 6) or to the upward traveling wave (equation 7).
with a particular maximum impact velocity v,, That is the way the wave comes through the virtual
maximum force F, is depending on the impedance of plane (refraction) and the way the wave reflects on
the pile Z,. ( =Practical statement). Because in prac- (reflection) the virtual plane. Equation 6 shows first
tice the rate of impedance of hammer and the im- a refraction term and second a reflection term. In
pedance of the chosen pile varies from say 10:1 to equation 7 the order of the terms is reversed.
4:l. Therefore the impedance of the pile and hence For the simplified process only the downward
the (minimum) cross section area of the pile is very traveling wave is of interest. So if there is no upward
dominant to determine the size of the maximum wave then in equation 6 only remains the refraction
force introduced in the pile. term (first term).
t
in which n(t)= integer ( 2 delta-TJ
Hr
delta-Tr = 7 ; Hr = Height of ram ,
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. Figure 3. Force - time graphs at top of pile for down-
Suppose a virtual plane with zero thickness, ward force wave in case of simplification (I), approxi-
which is situated between element 1 and 2. At a cer- mation ( 2 ) and according method of characteristics ( 3 )
tain time the force with size F.J~of the downward
67
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.
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
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.
E
SRDmax = 0.7 q c vt ; Ap Friction + Point (36)
With A, = cross section area close to pile tip.
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
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 At the end of the 191h century, the mechanical
hammer was born. Steam pressure pushed against a
In the beginning there was the drop hammer. Pulled a piston attached to the ram. Then, after tripping a
up by humans (Fig. I), horses or other creatures it valve, the steam pressure was vented and this
was suddenly released and then pulled up again. allowed the ram to coast up to its highest point and
Guides or leads for the piles were always an then fall with relatively little energy loss. Even
important part of the driving equipment. today steam is still being used as a power source for
The nineteenth century saw the general adaptation large hammers with ram weights up to almost 200
of steam engines to a variety of tasks, including pile tonnes. However, more convenient air compressors
driving. A first development the replacement of gradually replaced the steam boilers on hammers
human or animal power by steam power to pull the with typically less than 15 or 20 tonnes ram weight.
rams up with a winch. Today, there are still many Following the developments in general machine
winch driven drop hammers in use, typically with industries, it was obvious that double acting steam
ram weights between 1 and 5 tonnes. However, hammers should be built. They increased the
steam power has generally been replaced by electric operating speed of the hammer, but also wasted a lot
motors. diesel or gasoline engines. 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
Figure 1 . Artist's concept of an early pile driver. hammers suffered pre-admission problems, meaning
the motive fluid was admitted into the chamber
75
before the hammer impacted, thereby self cushioning latest development allows for a gradual increase of
the ram. the eccentric moment which is advantageous during
The first successfbl hydraulic hammers were the start-up of the hammer.
developed in Scandinavia in the 1960s for the A French and later US development was the
installation of regularly reinforced piles. The best Bodine hammer, a diesel engine driven Sonic Pile
machine that met this objective had a heavy ram and Driver with 60 and higher VPS. Such frequencies
a short stroke. This was accomplished with a simple created resonance in piles with length of 20m or
hydraulic jack for a short upward ram motion more. Resonance in turn would greatly improve the
(typically at most 1 meter) and then a free fall of the speed of pile penetration. Because of mechanical
ram after quickly pulling the jack downward. These problems this concept is still somewhat like a
machines have ram weights of nearly 20 tonnes. utopian dream.
They are being built by Banut, Uddcomp and others. For the sake of completeness the Silent Pile Driver
A parallel hammer development was engineered in which was developed in the last decades of the 20th
Holland and Germany for the offshore hammer century should also be mentioned. It is really a
market. Their innovative design approach created a static machine which uses hydraulic pressure to push
computer controlled double acting hydraulic piles into the ground. The required reaction forces
hammer. The computer is provided with sensor must be provided by previously installed
information of ram position and ram speed and neighboring piles. The system is therefore
accordingly adjusts the hydraulic pressure for an advantageous to sheet pile installation.
operator selected ram impact energy. The so-called driving system includes a hammer
Self-monitored hydraulic hammers have been built cushion, helmet with inserts (cap) and pile cushion.
with ram weights reaching 150 tonnes by IHC and While softwood and straw have been practically
Menck. Being hlly enclosed, they can be used abandoned for hammer cushions and pile cushions
underwater which requires solving a variety of respectively, little other progress has been made to
additional problems as described below. improve driving systems. Exception is made to the
A variety of hydraulic hammers which either systems that have eliminated hammer cushions
incorporate features of both the drop hammer and the entirely.
computer controlled type or which cannot be strictly The guiding systems also have undergone some
put in either of those two categories have also been important development. Dedicated systems prior to
developed. Some of these hammers are only World War I1 were often barge or skid mounted and
optionally fitted with sensors for computer control or had little freedom for adjustment. European
they are drop hammers with some downward developments combined carrier and leads into a
pressure assistance. Manufacturers of these dedicated system that requires little set-up time.
hydraulic hammers are BSP, HPSI, ICE, Junttan and This was particularly advantageous where pile sizes
others. and lengths did not vary much, such as was the case
Diesel hammers replaced the external power with the European segmental reinforced piles. In the
source winch, steam, air or hydraulic fluid with an US, a greater variety of pile and hammer types made
internal combustion underneath the ram. Starting in the combination of simple pile guide with a general
the 1920s in Germany, the first diesel powered rams purpose crane (often rented) more feasible. In the
typically had a mass of '/2 or 1 Mg. Today, ram nearshore and offshore environment, free riding
masses of single acting hammers range from 0.2 to leads are used almost exclusively to align hammer
20 Mg. and pile while a template or jacket provide support
A double acting diesel hammer that limited the for pile and hammer.
high diesel hammer strokes and improved hammer Throughout the history of pile driving, a great deal
speed by creating a passive air pressure chamber at of creativity was expended on the development of
the top of the hammer was also developed in the driving aids. The objective was the saving of pile
United States after World War 11. Today closed end material while improving driveability. Mandrels,
diesel hammers are built by ICE with ram weights jets, spuds, pre-augers, friction reducers and other
up to 5 tonnes. means were developed, improved, optimized and
The vibratory hammer was developed in the occasionally abandoned.
Soviet Union prior to and during World War 11.
These hammers were first built with relatively low
frequencies (less than 20 vibrations per 2 MECHANICS OF IMPACT PILE DRIVJNG
second[WS]) and were electrically powered. (After
the war US, French and other companies hrther Pile driving hammers are generally classified by
developed the low frequency hammers which today their potential energy prior to the ram beginning its
generate centrifbgal forces of up to 4000 kN and downward movement. The kinetic energy available
now use hydraulic power almost exclusively.) The to do work on driving system, pile and soil is a
76
certain fraction of that potential energy. The ratio of pile toe a reflection occurs that depends on the
kinetic to potential energy is the hammer efficiency. resistance that the material at the pile toe offers. If
This concept of rating the hammer by a potential there is no resistance at all the pile toe will move
energy and calculating the available kinetic energy twice the distance that the top moved during impact
using a reduction factor has been replaced by a and it will then pull the upper pile particles
measured kinetic energy for those hammers that downward. This creates a tension force which is
monitor ram position and speed immediately prior to potentially damaging to concrete piles. If there is
impact. an extremely high soil resistance such that the pile
Energy is not the sole quantity that defines the toe cannot move then the force at the pile toe will be
ability of a hammer to move the pile into the ground. twice the force that the pile top experienced during
An alternate measure would be its momentum or the the impact event. Theoretically, therefore, the
product of ram mass and impact velocity. While this hammer can overcome a force which is twice the
physical quantity would better indicate the impact force of Equation 1. In practice, because of
maximum force that the ram can generate in the pile the need for all materials to compress before they
and therefore the soil resistance force that it can can exhlbit resistance, the limit of the soil resistance
overcome, it is a measure that is only useh1 for is about 1.4 for piles with predominant end bearing
comparing hammer capability in very hard driving piles and 1.O for shaft resistance piles.
cases. In the case of the high toe resistance a strong
Impact velocity is, however, one of the most compressive wave reflection occurs which pushes
important parameters of the impact event. In the the upper pile particles upwards. For large ram
absence of a cushion and helmet mass between ram weights, the ram will still have downward
and pile top, the particles of the pile assume the momentum when the reflected wave reaches the pile
same velocity, vi, as the ram during the first instant top. In the case of the hard driving pile it is possible
of the impact event. The corresponding force can be that another compressive reflection occurs at the pile
calculated from top which may cause the pile top force to grow
above its initial peak. (In addition, a new wave will
F = viz (1) cause downward pile motions.) The benefit of a
where Z is the pile impedance Z given by large ram is therefore a longer lasting downward
motion of the pile, a fbrther increase of the pile toe
Z=EA/c force and the chance for additional permanent pile
with E being the pile's elastic modulus, A its cross penetration into the ground.
sectional area and c its wave speed given by Since it is uneconomical to use very large ram
weights, we normally do not see cases where the soil
c = (E / p)" ( 31 resistance is much greater than twice the impact
force. However, damage at the pile toe when it is on
with p being the mass density of the pile material rock can easily happen when either the impact force
is very high or the ram mass is large.
The presence of a cushion reduces the peak force Calculation of pile penetrations into the ground
over time and spreads it over a greater time. A and stresses along the pile are difficult and
helmet mass has a similar effect (Rausche et al. inaccurate when considering simplified systems and
1972). Even so, the maximum force in the pile at the employing wave propagation theory. It is better to
beginning of the impact event is primarily dependent use a wave equation computer program, such as
o n the impact velocity. GRLWEAP (GRL 1999) to realistically model
The greater the mass of the hammer, the greater its hammer, driving system, pile and soil, and
ability to maintain the pile top force near the mathematically simulate the pile driving process.
maximum impact force. Therefore, a small hammer Obviously, such models require assumptions
will only generate a short force pulse which may be
regarding the performance of the individual
ineffective in maintaining a sustained downward
components of the system and it is therefore best to
movement and therefore pile penetration. On the
measure forces and velocities at the pile top and
other hand, a slowly falling ram does not generate
calculate the desired quantities based on these
the force necessary to overcome soil resistance.
measurements using CAPWAP (GRL 1996).
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 3 HAMMERPERFORMANCE
after the ram has impacted, only the top of the pile
moves downward. It takes a certain time, the wave MEASUREMENTS
travel time L/c (L is the pile length) before other pile Pile Driving AnaZyzedC
particles firther down the pile start to move. At the
The basis for the results calculated by the PDA are
pile top strain and acceleration measurements which
77
are converted to force and velocity records, A modification of the Saximeter is the Saximeter-
respectively. The PDA conditions, calibrates and E which accepts timing signals from two proximity
displays these signals and immediately computes switches installed in the hammer such that the time
average pile force, F(t), and velocity, v(t), thereby between these two signals is a measure of the ram
eliminating bending effects. The PDA calculates a velocity immediately preceding impact. Again, this
variety of results for each hammer blow, including measurement can be simply converted to hammer
pile bearing capacity, pile stresses, pile integrity and kinetic energy. The signals from the proximity
other important pile quality parameters using closed switches are sent to the Saximeter-E using telemetry.
form Case Method solutions based on the one-
dimensional linear wave equation. Pile Installation Recorder nwfor Driveii Piles
The PDA also calculates the energy transferred to (PIR-0)
the pile top from:
The P E - D accepts signals from a variety of sensors,
E(t) = ,,jtF(z)v(z) dz (44 among them time of occurrence of hammer blow,
time when ram passes two positions just before
The maximum of the E(t) curve is the most
impact (as for Saximeter-E), pile depth of
important information for an overall evaluation of
penetration, bounce chamber pressure. These
the performance of a hammer and driving system.
signals are evaluated and plotted vs depth as follows:
This EMX value, also called ENTHRU, allows for a
classification of the hammer's performance when
stroke for open end diesels,
presented as the rated transfer efficiency, eT, also
impact velocity of all types of hammers, and
called energy transfer ratio (ETR), system efficiency
bounce chamber pressure for closed end
or global efficiency:
diesel hammers as an indirect measure of
= EIW(/ER (4b) stroke.
where ER is the manufacturer's rated energy value.
4 TECHNICAL DETAILS OF IMPACT
Hammer Pwfix-niance Ana@er TM HAMMERS
The ram velocity may be directly obtained using
radar technology in the Hammer Performance Cable siisperided drop hammers
Analyzer ( W A ) For the HPA to work the ram must The system generally uses a crane mounted winch
be visible. The impact velocity results can be for ram raising. Releasing the brake on the winch
automatically processed with a PC or recorded on a causes the ram to drop. Of course, the falling ram
strip chart. The result of the HPA is the velocity has to pull the winch cable and therefore accelerate
immediately preceding impact. The kinetic energy the winch which costs energy. Immediately prior to
can be calculated directly from this measuremen! impact, the crane driver is often tempted to reapply
(see appendix). the brake to prevent a spilling of the winch cable.
This action cushions the impact. It therefore must be
Smimeter nw expected that approximately 50% of the potential
hammer energy is lost. Up to 3 m drops are not
For open end diesel hammers, the time between two uncommon. In general, these hammers apply
impacts indicates the magnitude of the ram fall impacts of inconsistent magnitudes. Considering
height or stroke. This information is not only nominal fall height times ram weight the full energy,
measured and calculated by the PDA but also by the then energy losses are due to
convenient, hand-held Saximeter.
Both Saximeter and PDA calculate the stroke inaccurate drop height,
(STK) of an open end diesel hammer using fiction,
STK = (g/8) TB2- h, a winch drum inertia,
early catching of the ram, and
where misalignment.
70
not a practical solution for normal pile driving tasks. motive fluid to enter the chamber too early would
In fact it is advisable to transfer the support of the effectively cushion the impact of the ram prior to
ram from the crane to a stiff device such as the leads impact (preadmission). This effect is equivalent to
prior to releasing the ram to avoid damage to the the operator of a drop hammer applying the break
crane. In this way a precise hammer fall height can too early. If all settings are carefblly adjusted the
be achieved. Of course, the hammer efficiency losses would be small, however, losses due to
would be expected to be higher than that of the preadmission are also possible if the cushion block
winch brake released hammer. The energy losses are were too thin allowing for a longer than design drop
then only due to aRer the valves had been switched. On the other
hand, a very thick cushion block would reduce the
0 guide friction and stroke and, in the worst case, prevent switching of
0 misalignment. the valves. Then the hammer would not run at all. It
should also be mentioned that a sufficient aidsteam
The hammer efficiency therefore may be assumed pressure is important such that the ram coasts up to
to be quite high, e.g. 95%. the nominal fall height after the valves have vented
For those involved in dynamic load testing, the the chamber. In fact, in easy driving the upward ram
development of free release devices has frequently motion is sometimes so low that the hammer short
been of interest. The following solutions have been strokes (but in that case it is not a problem.).
encountered: There are two different arrangements of the steam
engine. One type attaches the piston directly to the
0 camhook, ram (Vulcan, Conmaco, MKT) and makes the
Spread hooks, and cylinder stationary. The other one uses a hollow ram
0 cut-off (torch, hydraulic shear) support cable. making it a movable cylinder while the piston is
stationary (Menck MRBS). There is another
Depending on the height, angle and friction factor difference between these two hammer types: The
of the cam, the force to release the weight will be a former maintains constant pressure under the piston
small percentage of the drop weight. Up to certain until the chamber is vented; the latter works with a
loads such devices are commercially available. constant volume of steam charge which is then
The tests described by Seidel et al. (1982) utilized allowed to expand. Obviously, the former wastes
a double hook device from which a 20 tonnes load more energy but is faster, while the latter is more
could be released by hand. economical.
As mentioned earlier, whenever a heavy ram is Single acting air hammers are not much different
suddenly released, care must be taken that the energy from steam hammers except that steam is hot and
stored in the supporting carrier is carefblly released. compressed air is colder. The different temperatures
In a winch release this is automatically the case of these two motive fluids require that
because the weight is only gradually released. manufacturing of piston and cylinder is done with
However, a sudden free release should either be different tolerances because of the differences in
done after the support of the weight is transferred thermal expansion.
from crane to ram guide or by utilizing a ballast It is important that aidsteam hammers are
between release device and crane hook. The inertia lubricated both internally with a lubricant injected
force of the ballast slows the release of the energy into the motive fluid and externally along the ram
stored in the crane system. guides. Energy losses primarily occur because of
79
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
Figure 2 . Transfer eficiencies of single acting strongly, the ram obtains a high upward velocity
aidsteam hammers on steel piles at the end of which necessitates reduction of the operating
driving. 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:
Figure 3. Transfer efficiencies of single acting Normally, for wave equation analyses a hammer
efficiency of 0.5 is therefore reasonable. Transferred
aidsteam hammers on concrete and timber piles at
energies average 35% on steel piles and 30% on
the end of driving. concrete piles (Figs. 4-5).
These hammers run vary rapidly and have been
slide bar which is attached to the moving ram. made very small for many purposes, such as jack
Several of these hammers have been equipped with a hammering. For pile driving their ram weights vary
variable stroke attachment which allows for typically between 2 kN and 90 kN.
alternating between two different strokes. This is
particularly useh1 for concrete pile driving where DieseI hammers
tender care is required in easy driving and high
It is not In the 1920s, DELMAG built the first open end
energies are needed in hard driving.
diesel hammer which utilized diesel combustion
uncommon that the hammer overstrokes on the short
underneath the ram to provide the energy for the
stroke and therefore operates at apparently higher
efficiency. It is therefore reasonable to increase the upward ram motion. Thus, the ram serves as the
piston of a 2-cycle diesel engine. The advantage of
hammer efficiency (e.g. to 0.75 from 0.67) of short
this hammer is a relatively light weight, sturdy
stroked hammers in a wave equation analysis.
construction and no need for an additional external
Double and differential acting aidsteam hammers power supply.
In order to generate hammer blows in quicker The hammer consists of four parts: a cylinder, a
succession it is desirable to shorten the stroke and ram or piston, an impact block that closes the bottom
accelerate the ram on its downward motion by of the cylinder and a fbel pump that injects he1 into
pushing downward with active pressure. This has the cylinder prior to impact. There are four phases
been accomplished in double or differential acting that this hammer type undergoes prior to, during and
hammers. The difference between these two engine
80
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
Figure 4. Transfer eficiencies of double acting much greater extent than airlsteam hammers the ram
aidsteam hammers on steel piles from end of drive drop height of diesel hammers is dependent on the
measurements. 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
Figure 5. Transfer eficiencies of double acting energy is between 20 and 30% of the potential
air/steam hammers on concrete piles from end of hammer energy. It may be much higher if the
hammer pre-ignites, which may happen if the
drive measurements. hammer is poorly maintained or overheats to such a
degree that either the fuel oil or the lubrication oil
after the hammer impact. During the first phase the combusts prior to impact. The ram then falls against
ram first rises and then falls only subject to gravity. a very strongly increasing pressure which reduces
When on the downstroke the ram closes the exhaust ram velocity plus kinetic energy. Note however, that
ports. The air underneath the ram is being some energy already gets transferred to pile and soil
compressed, and thus raises its temperature. This is during the pre-compression phase.
the second or compression phase. Impact occurs and Pre-compression is only possible if the pile can
injected fuel combusts in the third, the impact and offer a resistance that matches the compression
combustion phase. Finally, the ram is driven forces. For this reason, the hammer will not run if
upwards during the expansion phase that ends when the soil resistance is very small.
the excess pressure is vented through the exhaust In diesel hammers, energy losses are typically
ports and new air is sucked into the cylinder during caused by:
the ram’s upward motion which - being only subject pre-compression,
t o gravity - is part of the first phase of the next friction between ram and cylinder,
hammer blow. a impact and inertia losses of the impact block,
Open end diesel hammers are made or distributed low stroke,
by APE, Berminghammer, DELMAG, and ICE. pre-ignition, and
misalignment.
Considering pre-compression a design loss that The first successfhl hydraulic hammers were
can be calculated, the remaining losses are typically simply drop hammers: a jack would pick up the ram
less than 20% of the potential energy. The appendix and then retract quicker than the falling ram. Since
contains a numerical example for the calculation of these hammers were at first built with low drop
pre-compression energy. heights, energy losses were extremely small.
Ram weights range from 2 kN to 200 kN, rated Examples are the Banut, ICE, Junttan, Menck MHF
energies from 0.4 to 600 kJ. The hammer efficiency series, Uddcomb and other hammers. Ram weights
typically averages 0.8. Transferred energies have are as high as 150 kN and rated energies reach 200
medians of 3 1 56 on steel piles and 25 % on concrete kJ.
piles (Figs. 6-7). Energy losses are caused by:
Other hj*drmiichammers
There are a variety of other hydraulic hammers
available today. For example, BSP makes hammers
82
that are operated either with or without active pile stresses. Indeed, diesel hammers whose rams
downward pressure. They may or may not include a strike the impact block directly and which have been
monitoring equipment. The variety of hammers in in existence for decades without damage to ram or
this category preclude general statements of energy impact block have proven that well fitting impact
losses and recommended efficiencies. Indeed, surfaces generate tolerable stresses. These hammer
Figure 8 shows that the combination of all hydraulic manufacturers therefore devised a cushionless
hammers tested by the author’s firm generate an driving system which features a well fitting and
efficiency diagram with two distinctly different machined impact surface for the ram such that high
groups of performance: a lower area around 55% and contact stresses are avoided. This idea was first
a higher one with an 85% mean. Additional data is incorporated in the LHC hydraulic hammers.
needed to produce a clearer classification of these
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
Figure 8. Transfer efficiencies for a variety of limit can be a problem when driving for an extended
hydraulic hammers on steel piles. time period with a follower under water where a
cushion exchange is not possible. Using thick
5 DRIVING SYSTEMS cushion blocks to lower the impact force by
spreading it over time leads to a rather wastehl use
Driving systems are defined as the accessories of plywood. A combination of thin plywood
necessary to transfer the impact forces safely to the sandwiching a somewhat more resilient material can
pile top. They include a cushion for the protection of then be more economical. Hamortex, a man-made
the hammer, a helmet with adaptor inserts and a material consisting of rolled-up, aluminum coated
cushion to protect concrete piles. paper has been successfully employed for this
purpose.
Pile cushions are also made of a variety of other
Hammer cushions materials. For example, in Latin America sisal and
With a few exceptions, hammer impact is usually hemp rope have been successfully used though with
cushioned by a relatively resilient material. The greater energy losses than plywood.
reason is the need of protecting the ram from
excessive fatigue stresses. Often man made He Ime ts
materials are used and they require little exchange as
long as the cushion is protected by a well fitting steel Helmets must align hammer and pile and securely
striker plate and is sandwiched between aluminum spread the impact forces over the pile top.
sheets for heat extraction. If they are made of a stiff Depending on the type of pile, they must be fitted
material (softwood is not acceptable as a hammer with inserts that match the pile profile. Helmets also
cushion), hammer cushions store and dissipate accommodate the hammer cushion. It is important
relatively little energy. that the hammer only allows for a small lateral
tolerance (e.g. 25 mm) on the pile top for good
hammer pile alignment.
Driving systems without hammer cushions
Certain hammer manufacturers have decided that it
is unnecessary to protect the hammer against high
83
6 PILEGUIDES Supplying high hydraulic pressures at high rates
of volume through long hoses pose another
Guiding of the pile is an important part of the pile challenge. One solution would be a remote hydraulic
driving equipment. Historically, dedicated systems power inside the guiding system of the hammer.
with a stiff leg and a skid have fulfilled this role and That means that the power pack also has to fit
these systems are still used for some very large through the guides and that electric power is
projects. However, these systems are not very easy supplied over a long distance.
to mo\re and therefore reduce productivity. 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
Swinging leads
and care has to be taken that air passages are large
A very simple system is the swinging lead which can enough so as not to impede the ram movement. Of
be cable supported from a multi-purpose crane and course, the hammers' sensors will be able to indicate
which is stabbed in the ground for a second point of the actual kinetic energy.
support. Obviously, such a system is relatively Finally, driving against a water column inside the
flexible, a feature that sometimes protects piles from pile would dissipate valuable energy. Thus, excess
breakage when the pile encounters an obstruction water must be allowed to escape through holes in the
and wants to move laterally. pile near its top and an air cushion should be
maintained beneath the helmet.
Fixed leads
Fixed leads are supported at the boom tip of the 8 VIBRATORYHAMMERS
crane with a hinge and near their bottom by a brace.
Normally, this brace can be extended or contracted An impact hammer generates a short duration
hydraulically to set the batter. Fixed leads are forcehelocity pulse in the driven pile. Between
probably the most common support system and hammer blows the pile and soil are practically at rest
many dedicated lead-hammer systems that use a for a time (typically 1 second) that is large compared
dedicated carrier also achieve mobility and control to the impact event itself (typically 20 to 50 ms).
with such a lead type. Often these systems can adapt The impact hammer therefore generates a complete
to a compound batter. cycle of loading and unloading during the short
impact event
First built in the Soviet Union in the mid 20th
Freeridiilg leads
century, vibratory hammers became the installation
Over water, piles are often driven through a template device of choice for non-bearing piles, because of
which assures the correct location and direction of their superior speed of pile installation in many soil
the pile. In such a case leads are merely aligning the types.
hammer with the pile. These "caison-type" or Vibratory hammers consist of pairs of rotors with
freeriding leads are supported at their top by the eccentric weights attached. Always two rotors are
crane. A disadvantage is for battered piles that the synchronized such that their horizontal centrifbgal
transverse component of hammer and lead weight force components cancel while their vertical ones
has to be supported by the pile, thereby introducing add. The vibratory machine therefore obtains an up
potentially high bending stresses in the pile which and down sine motion with a frequency equal to the
are superimposed to the dynamic stresses from pile speed of rotation Obviously, the maximum vertical
driving. 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
7 UNDERWATER PILE DRIVING moment of all rotors in the machine is ME and the
frequency. f. then the downward maximum dynamic
Under water pile driving in water depths of 1000 m force is
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 Originall\,. iibratory hammers had a relatively low
guiding system. The hammer outside diameter frequency of less than 20 Hz (1200 FWM), were
therefore should not be greater than the pile electrically powered, and of moderate size. Today
diameter. In fact, it is sometimes necessary that a these standard machines are powered with hydraulic
pipe extension is added to the hammer top for aggregates reaching 600 kW (800 HP) with
additional guiding once the pile has been driven centrihgal forces of up to 4000 kN (800 kips) and
below the lowest guide
84
frequencies typically up to 30 Hz. Larger outputs Predrillir ig
may be achie\,ed with two machines running in
tandem. 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
Resonant pile drivers percent of the pile size is drilled through the
The vibratory pile driver’s dream has always been to potentially difficult overburden. It is important that
achieve even better penetration rates with a machine the hole is not drilled too deep. Strongly reducing
that would generate resonance in the hammer-pile shaft resistance on a long pile and then driving
system. Resonance would generate forces much against a high end bearing may actually cause
larger than the centrihgal force of the hammer. damaging pile stresses and make the pile harder to
Such a high frequency machine has been built, used, drive.
improved and retried, generally with the result that
pile driving indeed became very quick, however, the Jetting
machine would suffer and frequently break down.
The Boudine pile driver is equipped with two diesel Either on the side of a pile, or sometimes through a
engines driving the eccenters directly at the pile top. center tube of a concrete pile, water may be used to
It is capable of frequencies up to 80 Hz which would loosen and remove soil from around the pile tip. In
generate resonance in a pile of 15 to 30 m length, softer soils large cavities may accidentally be opened
depending on the relative weights of pile and up and that would, of course, detrimentaIly affect the
machine and on the resistance magnitude and bearing behavior of the soil.
distribution.
Spudding
Variable moment hammers Prior to driving large concrete displacement piles
An interesting aspect of vibratory pile driving is the (e.g. a 900 mm square) a hole can be driven with a
response of pile and soil to various frequencies. spud which is typically a steel section of the same
Typically, for somewhere above 3 and below 10 Hz size as the concrete pile. The spud may be so heavy
a resonance between pile and soil occurs which that it can be repeatedly lifted and dropped to form a
generates strong soil motions. Nearby buildings can hole in which the concrete pile is then inserted. The
sustain damage due to these vibrations. When concrete pile only needs enough hammer blows to
shutting the machine off, the standard vibratory seat it into soil undisturbed by the spud.
hammer again generates these lower damaging
frequencies. In addition, while higher frequencies Externalfric 1ion reducers
tend to densifL the soil, low resonant frequencies
may actually cause a loosening of denser materials. For pipe piles that have to penetrate competent soils
It is therefore highly desirable to make a hammer for deep penetrations it is advantageous to attach
that does not cause a low frequency resonance and shallow rings to the outer pile circumference. These
since no motor can suddenly run at a high speed the rings reduce the friction temporarily during pile
only solution is to increase the eccentric moment driving by creating a heavily disturbed zone in the
aRer the desirable hammer speed has been reached. area above. Design details such as size of and
Such variable moment machines are now being distance between the rings are either proprietary or
produced in Europe. They are also effectively used non-existent .
for soil densification (Massarsch 1992).
Internalfric fioii rediicers
9 INSTALLATION AIDS For pipe piles. a so-called internal driving shoe, i.e. a
section of pipe with increased wall thickness but
It is often not wise to force a pile through harder identical outside diameter is attached to the pile
layers into softer materials or to give the pile a much bottom. Above this section the soil penetrating into
higher strength than necessary merely for the pipe during pile driving experiences reduced
driveability. Instead it may be possible to use lateral confinement and therefore reduced friction.
installation aids such as jetting or drilling. However,
such installation aids must be used wisely. They Electric potential
may actually cause harm to pile and soil if not
properly executed . For offshore steel pipes that have to be driven into
clay, an electric current can produce a temporary
85
-- HHK 5 on Concrete Pile
-
Banui 7 tonnes on Concrete Pile GRLWEAP (TM) Version 1998.:
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
'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.
-
20onSteelfi~e GRLWEAP (TM) Vemion 199F3-i
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 Figure 9 shows that the heavier hammer, even
dramatically reduce the installation time and effort. though it has a lower energy, requires approximately
However, this system apparently has not been 20% less blow count to accomplish the same bearing
proven economical or safe enough for genera! capacity. The compressive and tensile stresses were
applications. considerably lower. Obviously, the lower stresses
are a result of the lower impact velocity of the
Steamfor permafrost heavier hammer. The somewhat surprising result of
lower blow counts for the lower hammer energy is
Pile driving in permafrost poses special problems. caused by a relatively large quake (10 mm) at the
The frozen ground behaves under hammer impact pile toe where 50% of the total resistance were
like a soft rock. Resonant vibratory pile driving is a assumed to act. This large quake means that the toe
actually a possible solution only limited by the lack resistance has a high flexibility which can best be
of reliable machines. For regular pile driving steam overcome with a force pulse in the pile that extended
injection thaws the permafrost temporarily and over a long time. The heavier ram does this b,etter
therefore allows the pile to penetrate under normal than the lighter one with the stroke.
hammer impacts. 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
10 EQUIPMENT SELECTION a diesel hammer with 4.6 Mg ram and an aidsteam
hammer with 9 Mg ram is shown in Figure 10.
Selecting the proper equipment for pile installation Obviously the lighter hammer is performing better in
can make and break a project's technical and hard driving. This is not too surprising if rated
economic success. An underpowered equipment energies are compared which are 153 and 81 kJ,
may cause excessive numbers of hammer blows and respectively. In other words, the heavier hammer is
pile fatigue and installation may not even be rated with a much smaller stroke (.9 m) than the
possible. Selecting very powerfid equipment also diesel (3.4 m). Obviously, an air/steam hammer
may cause damage and/or require unnecessary with a heavier ram and a higher stroke for
capital expense. comparable energy ratings would make the aidsteam
While many "rules of thumb" have been tried and hammer perform better than the diesel.
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 CONCLUSIONS
reliable tool to size hammer and driving system for
reasonable blow counts and safe stresses. Very Equipment for driven pile installations has
briefly: for friction piles, blow counts should be less undergone slow changes. Depending on where in
than 80 blows for .25 m, for end bearing piles where the world piles are being installed, equipment may
hard driving is expected to be of a relatively short range from a simple winch driven drop hammer to a
duration, this criterion can be relaxed to 200 blows sophisticated computer controlled hydraulic
for .25 m. According to Hannigan et al, 1996, hammer. The hammers and piles may be guided by
compressive concrete stresses should be less than an old skid rig or a dedicated fixed lead system.
85% of the concrete strength (after any subtracting Depending on the type of hammer and driving
prestress). Tensile concrete stresses should amount system, great variations of the system efficiency
t o less than 70% of the equivalent yield strength of must be expected. These eficiencies can be
regular reinforcement or prestress plus 50% of the accurately assessed in the field using a Pile Driving
concrete tensile strength. For steel piles, stresses Analyzer which calculates the so called transfer
should be less than 90% of yield. energy from pile top force and velocity
A comparative analysis was made as a measurements.
demonstration of the value of the basic wave Predictions of pile blow counts, stresses or
equation output, the bearing graph. As shown in capacities should be based on a wave equation
Figure 8, a 20 m long concrete pile, 400x400 mm analysis which incorporates realistic models for
square was anaiyzed under two somewhat different hammer, driving system, piles and soils. These
hydraulic drop hammers. One of these hammers, the predictions should be verified by measurements.
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
REFERENCES This is the ram impact velocity in the absence of
any fiction or other losses. The work that a friction
GRL, Goble Rausche Likins and Associates, Inc. 1996. force R would be doing while the ram is falling is
CAPWAP Background report and users manual. 4535
Renaissance Parkway, Cleveland, OH, 44128. EF=Rh (A4)
GRL, Goble Rausche Likins and Associates, Inc. 1999.
GKLWMP Background report and helpful data. 4535 practically reducing the available energy fiom WR
Renaissance Parkway, Cleveland, OH, 44128. h to (WR - R)h. Let us assume that the friction force
Hannigan, P., G.G. Goble, G. Likins, G. Thendean, & F. is a certain percentage, p, of WRYthen we can write
Rausche 1996. Design and construction control of driven the reduced potential energy as
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. Where q = (1 - p) is the so-called efficiency of the
Rausche, F., & G.G. Goble 1972. Performance of pile hammer while p indicates the fraction of the energy
driving hammers, ASCE, Journal of the Construction that has been converted to heat rather than speed. Of
Division, Vol. 98, No. C02, Proc. Paper 9188, 201-218,
September 1972. course, the impact velocity of the ram is now
Seidel, J. & F. Rausche 1984. Design and performance of
dynamic tests of large diameter drilled shafts. Pd Strass vi = r ( 2 g h q ) (A6)
Wave Conference, Proceedings of the PdInternational
Conference on the Appiication of Stress Wave Theory to In addition to fiiction, other losses also occur
Piles., 9-16, Balkema, Rotterdam. during the descent of a hammer. For example, a
diesel hammer compresses air which requires the
following energy:
APPENDIX ED= I (P dV)
Energy considerations
=A 1 ( pIN[V, /VIexp)ds
= [A pIN/(exp - I)] [VN/A+ hJeXP A (A71
This appendix summarizes the various forms of
energy in hammer, driving system and pile. It is with
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 In this formula, V, is the initial volume of air that
to a certain speed, compress a spring, etc. is present in the diesel hammer chamber when
Mathematically energy is defined as the product of compression starts, A is the inside area of the
force and distance over which the force acts. For cylinder, pINis the initial pressure (atmospheric), s is
example, the weight of a ram, WR, suspended a the distance that the ram has traveled after the
distance h above a reference datum, has the ability to compression started, V is the associated volume, &
do work equal to its potential energy is the compressive stroke, and exp is the exponent of
adiabatic compression, typically 1.4 for clean air.
Ep=WR h (AI) The chamber volume, Vc , i.e. the volume left over
Since energy is "indestructible" (in the worst case during impact is the product of compressive stroke
it is converted to heat, sound or other forms of and area A equal the initial volume. The
energy) after falling through a vertical distance h, the compression ratio is therefore Vm / V,. Note that the
ram now contains a kinetic energy E, while its atmospheric pressure also does positive work on the
potential energy has become zero. top surface of the ram. This energy is merely the
product of atmospheric pressure times ram top area
EK= !h.V;wR/ g = Ep WR h (A21 times compressive stroke.
where g is the gravitational acceleration. As an example, let us use a few realistic numbers
Therefore representing a typical diesel hammer with 2 Mg ram
mass. Assume = 394 mm, Vc = 2,584,240 mm3,
vi = T ( 2 g h ) (A31 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.
88
Entering these numbers in the above equation and stored is 3.3 kJ of which one half may be converted
considering the work of the atmospheric pressure to heat (if it is a wooden block).
yields a pre-compression energy of 1 1.4 kJ. Let us now calculate how much energy it takes to
If this hammer has indeed a ram weight of 20 kN accelerate the helmet to its highest velocity which is
and if its maximum stroke is 3 m, then it would be approximately equal to the maximum velocity that
rated with 60 kJ. The pre-compression loss is we see at the pile top and which may reach 3 d s .
therefore 19% of the rated energy. Again assuming a helmet weight of 10 kN or a mass
of roughly 1 Mg yields a helmet kinetic energy of
Energy in //it. driviiig sysfem Em = % m v: = % 1(3)2 = 4.5 kJ (A1 1)
During impact the ram also compresses one or two It may be assumed that this kinetic energy is still
cushions. If the spring constant of the cushion is k doing useful work on the pile after the helmet has
and the maximum compression x and the maximum started to slow down.
force F = k x, then the energy stored in the cushion is 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
E,=%Fx (As)
cushion stores approximately 3.3 kJ, the helmet 4.5
Let us compare a 20 mm thick plywood cushion kJ, the pile cushion between 0.5 and 5 kJ. For a
and a 200 mm thick plywood cushion with cross single acting air/ steam hammer we would expect an
sectional area 0.1 m and normal plywood modulus energy loss of 33% of 60 W or 20 kJ due to friction,
(200 MPa). The corresponding cushion stiffness misalignment, pre-admission etc. For a diesel
values are k = (0.1)(200) / 20 = 1000 kN/mm and hammer we would normally expect additional losses
100 kN/mm, respectively. A mass falling onto these (friction, misalignment) of 20% of (60 - 10) = 10 kJ.
springs and producing a force of 1000 kN in the pile For the diesel hammer there would also be kinetic
underneath will produce the following respective energy temporarily stored in the impact block. Its
cushion compression values magnitude may be comparable to the energy stored
in the helmet when it moves. However, most of this
x = 1000/k = 1 and 10 mm kinetic energy is probably transferred to the pile at a
and therefore store energies of time when it still can do useful work on the pile and
soil. The energy transferred to a concrete pile may
Ec = % 1000 (1) or % 1000 (10) = 0.5 or 5 kJ (A10) therefore be approximately 30 kJ for a single acting
aidsteam hammer and 25 kJ for a diesel hammer.
(Actually, the force underneath the softer cushion For steel piles which have no pile cushion the
is probably lower than the force under the thinner transferred energy would be about 5 kJ higher. The
and stiffer cushion; this would make the effect transfer efficiencies would then be 0.5 and 0.45 for
demonstrated here a little less extreme.) For a the single acting hammers on steel and concrete
concrete pile and cushion of this size it would be piles; for diesel hammers these numbers would be
reasonable to assume that the hammer rated energy 0.45 and 0.4. These results are in reasonable
ranges between 20 and 60 kJ. The thick cushion agreement with the measured transfer efficiency
therefore stores a considerable amount of energy shown in Figures 2 through 7 .
during the compression phase. The thinner cushion, Energy losses are generally lower for hammers
o n the other hand generates rather insignificant with low impact velocity, obviously because inertia
energy losses. Assuming a coefficient of restitution and cushion forces would be lower and thus the
of 0.5, which is reasonable for a plywood cushion, associated cushion strain energy and mass kinetic
half of the compression energy would be converted energy losses. However, experience, measurements
into heat. The other half is returned to the ram and wave equation analyses all indicate that high
during its upward movement. velocity hammers do rather well when driving gets
Hammer cushions are subjected to higher forces hard.
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
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 cation approximately one to two pile diameters be-
EQUIPMENT AND ANALYTICAL METHODS low its head. Data analysis is based on field meas-
ured dynamic records, sophisticated modelling and
Static loading tests have traditionally been per- numerical techniques, and one-dimensional elastic
formed to determine the characteristics of pile load- wave propagation principles.
movement relationship. Due to the expense involved Field testing is done according to Case method
and time required, this type of testing is performed procedures as applied with the Pile Driving Analyzer
on only selected piles on a very limited basis. Lower (PDA). The Case method is a closed form solution
cost and speed of testing make dynamic pile testing based on a few simplifying assumptions such as
according to Case method and CAPWAP analysis ideal plastic soil behavior and an ideally elastic and
particularly attractive (Hussein et al. 1988). uniform pile (Linkins & Hussein 1988). Assump-
The dynamic testing method was developed at tions of a soil damping constant must be made at the
Case Institute of Technology in Cleveland, Ohio in time of testing. The PDA is user friendly field data
the late 60’s (Rausche et al. 1985). This method pro- acquisition system and computer that provides
vides an economical, easily portable system which power supply, signal conditioning, processing and
bases pile bearing capacity predictions on electronic evaluation of the measured dynamic data. Generally,
measurements of pile top force and acceleration. The testing results obtained in the field, include results
following are the main objectives of dynamic cast- for evaluating hammer performance, dynamic pile
in-situ pile testing: (a) evaluation of pile static ca- stresses, pile structural integrity, and pile resistance
pacity (shaft friction and toe bearing), (b) determi- and static bearing capacity. High strain dynamic pile
nation of dynamic piles stresses under hammer im- testing is part of many standards and specifications
pacts, (c) assessment of pile structural integrity. such as ASTM - 1989, AASHTO - 1993 and the
High-strain dynamic pile testing is based on the British CIRIA guide.
iiieasurement of pile strain and acceleration under Rigorous analysis on the field collected PDA data
impacts of a relatively large weight. Pile strains are is performed using the CAPWAP method (CAse Pile
measured with strain transducers. Sensors are Wave Analysis Program) (Rausche at al. 1985). This
moiuited around pile circumference, ideally at a lo- 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.
92
Consider K, = A, E / L, ; i = Segment 1, 2 or 3. Wave equation analysis input information have
By using tlie Kirchoff s Law, tlie effective ram stiff- included a hammer efficiency of 30 %, skin aiid toe
ness shall be 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
K’,= KIK2 / (Kl + K2) (1) 5.0 mm respectively for 60 cm pile diameter, and
soil damping factors of 0.25 and 0.50 s/m for skin
S = A I A2 L’, / ( A I L2 + A2 L I )
T ~ UA’, (2 1 and toe resistance. The GRLWEAPR analysis re-
sults, called ‘inspector’s chart’, showed a relation-
where L’, = LI 4- L2 (3) ship between tlie stroke and the blow count for a
constant capacity option, aiid the compression and
Again K, = K’r I<i / (K’r + K3) tension dynamic stresses for each hammer drop
height (Fig. 2)
u s= A’, A3 L, / (A’r L3 + A3 L’r)
T ~ ~ A, For pile 6 having 50 cm diameter and hammer
stroke ranging between 10 and 100 cm, the constant
where L, = L’, + Lj capacity analysis showed the necessary driving re-
sistance at each hammer stroke for the input capacity
And finally D, = (4 A, / pi)”2 (4) of 1325 KN. Based on this analysis, a 30 cm ham-
mer stroke was found to be reasonable to test 50 cin
Based on the above, a GRLWEAP data file re- diameter shafts since it shows a reasonable driving
lated to our hammer system was made available. resistance and keeps the driving compressive stress
Such data are given in Table 2 below. 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-
Material Plywood
quired pile capacity has been reached (Haiinigan et
Pile Cross section cm2 3 847 al. 1996).
cushion TIlickness imn 18 During the field dynamic measurements coil-
WEAP Modulus MPa 207 ducted 011 pile 6, the CASE method has showed a
Coeflicieiit of resti(iitioti 0.5 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
4 CASE STUDY (1) cm (Fig. 3). The point of intersection of the stroke
and driving resistance is falling above the inspec-
For tlie proposed buildirig at Nabbaa area in Sharjah tor’s chart curve, hence tlie desired pile capacity has
Emirate, low capacity piles were drilled to a depth of been obtained aiid thus field measurements have
about 15.5 m below the site reference level. The in- confirmed our initial analysis.
vestigated soil consisted in general silty sand occa- The same applies to pile 24 where a stroke of 50
sionally cemeiited and slielly within the top 9.8 m, cm has showed a proof load of about 2750 kN, a
with the presence of sandstone formation at depths generated driving compressive stress of about 9.1
beyond 14.4 m. Static capacity analysis conducted MPa and a driving resistance of about 2000
by the soil investigation firm concluded a design blows/meter, for which tlie intersection point falls
working load of 883 kN (324 kN for skin friction) above the curve of the WEAP inspector’s chart (Fig.
and 1226 kN (390 kN for skin friction) for pile hav- 4).
ing diameters of 50 and 60 cm respectively. About
50 to 55 % skin friction of the desired load is coil- 5 CASE STUDY (2)
s idered for wave equation analysis distributed more
or less uniforiiily over the pile shaft according to the The second case study consisted of the proposed
soil characteristics as per the investigated soil layers. Building at Noor AI Aiiz East Area in Dubai Einir-
93
Figure 2. GIUWEAP analysis for case I
94
Figure 5. GIUWEAP analysis for case 2.
95
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
Figure 9. GRLWEAP analysis for case 3 (35 % hammer effi- about 4220 kN, a maximum driving stress of about
ciency).
9.5 MPa and a measured set of about 0.5 inm for
which the intersection point with the stroke falls al-
MPa for an observed blow count of about 1000 most on the curve of the inspector’s chart. Hence,
blowdmeter (Fig. 8). As far as the inspector’s chart our initial analysis has been confirmed.
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
7 RECOMMENDATIONS ON WAVE properly, wave equation analysis shall lead to a SUC-
EQUATION ANALYSIS cessful dynamic test.
8 CONCLUSION
97
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
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.
99
the core. This layer, typically 125p,11in diameter, is 2.3 Fiber- Br-agg Gmting sensor(FBG sensov)
glass similar to the core, but with a slightly different
composition. Surrounding the cladding is a buffer FBG sensor is made by exposing tlie fiber of about
layer that protects the inner glass. The buffer lcin to the interference pattern of two coherent UV
diameter is 900/,,,, made of ceramic or plastic. 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
2.2 Tjpes of Fiber. Optic Seiisors of the glass. The resulting spatial modulation of
Fiber optic sensors are divided into two broad refractive index produces the Bragg grating and acts
categories,intensionietric and interferometric sensor. as a narrow band in-line wavelength notch filter.
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 When the light signal from broad band sourcc is
application, is tlie Fiber Bragg Grating. transmitted, tlie grating reflects strongly at tlie
The comparison of different types of optical fiber wavelength, ~~,
for which the Bragg resonance is
sensors is summarized in Table 1. The higher the satisfied, as shown in Figure 2. This Bragg
number, tlic better in its functional category. resonance is determined by the condition:
Table 1. Comparison of Sensors
I I I I I 1 1
100
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,
Figure 3. Multiplexing of FBG sensor 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,
of FBG has distinct advantages over other sensing
mirror and phase mask is shown in Figure 4.
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.
101
3.3 Compression test on cement mortar specinien bar with a length of 50cm for loading test was
inserted into the ceiiter of model pile. Model pile
To observe the performance of FBG sensor
was cured for 7 days and the loading test was
embedded in the concrete structure, a stress-strain
performed.
relationship was measured for a cement mortar
specimen. Two steel bars attaching a FBG or a strain 3.4.3 Static load test on moclel piles
gauge in the mid-length were embedded in mortar
specimen in the direction to the applied load. The Two tension tests and one compression test were
mortar specimen was a rectangular hexahedron performed, following a procedure of "quick
(1 5cm" 15cm"20cm). The steel bars were first maintained load test". The interval of maintained
iiiserted in a mold and then a fluid mortar was load at each loading stage was 5 minutes and the
poured. upward displacement was checked at 0, 2.5, 4 and 5
Compression loading was applied on the mortar minutes. The schematic drawing of tensile loading
specimen by universal testing machine and the test system is shown in Figure 6. Load applied by an
strains measured by FBG and strain gage were oil jack of 5 ton maximum capacity was measured
compared. by a proving ring which can be loaded up to 3 tons,
3.4 Model Pile test and displacements at pile butt was measured by two
dial gauges with accuracy of 1/200mm.
3.4.1 Sciiizple pvepamtion
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
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.
104
4.3.2 Distribution of axial pile loads
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
106
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
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.
107
some time elapsed after initial driving results in After simplification formula (2) becomes
developing soil set-up. Quantity, t, is a time elapsed
from the end of initial driving and capacity RRSTR-1 is R"(t>
- 1 = B rlog,,(t) + 11 (3)
determined for t=t,. A factor, A, is dependent on RE,,,
soil conditions.
Consolidation of cohesive soils around the pile Formula (3) is similar to formula (1). However,
after pile installation requires much more time in proposed formula (3) has certain advantages. This
comparison with other soil types like sand and formula (a) uses the traditional set-up formulation,
gravel to regain in soil strength and pile-soil (b) takes into account the actual time in days passed
adhesion after EOID. For this reason the existing after pile installation, (c) provides determination of
formula is pertinent for clay and cohesive soils. the soil set-up independently of the time of the first
Case histories presented by Svinkin et al. (1994) restrike.
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 4 CASE HISTORIES
factor, A, depends not only on the soil but also on
the pile type and size. For verification of formula (3), in the following case
The application of formula (1) usually requires to histories the existing and proposed expressions were
make the first restrike in 1-2 days after pile driving. used to calculate the pile capacity as a function of
It might be especially convenient for dynamic testing time after EOID.
at construction sites where many piles should be
tested by restrike in a short period of time.
Nevertheless, the existing formula yields relative 4.1 Case I
set-up versus relative time for assessment of the pile
capacity after the first restrike. This is a Three piles were considered in this study from total
contradiction to the set-up approach commonly used number of seven prestressed concrete piles tested for
in geotechnical practice and inconvenient for certain a bridge approach (GRL Report 1987, Svinkin et al.
construction sites to restrike piles on 1-2 days after 1994). A pile description is presented in Table 1.
pile installation. 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
3 PROPOSED FORMULA layer of silty sand. Water table was at the ground
surface. A Delmag D 46-13 hammer was employed
The main goal to derive a new formula for for initial driving and restrikes. For each pile, 3 to
evaluation of the set-up effect of cohesive soils in 4 dynamic testings were performed after pile
pile capacity is to take into consideration the pile installation. For piles TP3 and TP4 static loading
capacity at EOID and the actual time elapsed after tests (SLT) were made as well. The elapsed time
pile installation. after EOID, the penetration resistance and the time
For a soil set-up straight line passing through two dependent ultimate capacity of piles tested are shown
points corresponding pile capacity at EOID, REOID, in Table 1. Pile capacities from dynamic testings
and pile capacity at any time after pile installation, were determined by CAPWAP analysis and the
R,(t), a formula with a logarithmic time scale can be Davisson criterion of failure load was used for static
written 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
Time is calculated in days after pile installation. calculated in accordance with the existing formula
The time for EOID is taken 0.1 (2.4 hours) that have been plotted in ordinates Ru(t)/RR,,R.l and
negligibly affects increasing in the pile capacity at loglo(t/tl)as broken and straight lines, respectively,
EOID but gives an opportunity to use the in Figure 1. Note that the set-up coefficients from
logarithmic time scale. the field tests generally match well to those obtained
from the existing formula.
For piles under consideration, the factor, B, has
108
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
~~
been found on the basis of back calculations using earth fill followed by about 19.8 m of soft to
formula (3). This factor ranges from 1.6 to 3.5 medium stiff compressible post-glacial silty clay and
(Table 1). Scattering of the factor, B, is the same clayey silt underlain by about 27.4 m of glacial
for the existing and proposed formulae. The set-up material deposited on dolomite bedrock. The water
coefficients derived from both the field tests and table was about 2.5 m below grade. The H-pile was
from formula (3) are given in Table 1 and depicted founded in the glacial material.
in ordinates Ru(t)/REOIDand Time after EOID in days Five restrikes were made for pile tested. Pile
(logarithmic scale) as broken and straight lines, capacities from dynamic testings were determined by
respectively, in Figure 1. Good agreement is found CAPWAP analysis (GRL Manual 1993). The static
between calculated and measured values of pile loading test for this pile did not show a plunging
capacity as a function of time after EOID. 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
4.2 Case 2 capacity from the static loading test was evaluated
from the Davisson criterion. The results of dynamic
Initial data for this case were taken after Fellenius et and static tests are shown in Table 1.
al. (1989). An H-pile 310x94 (mm, kg/m) with Pile capacity calculation according to formula (1)
length of 47.2 m was driven and restruck by a was made by Svinkin et al. (1994) and shown in
Vulcan 010 h a m e r with a nominal energy of 44 kJ. Figure 1. It can be seen the calculated set-up line
Restrike No. 4 of this pile was performed by drop averages measured values of pile capacity.
hammer with nominal energy of 65 kJ. The soil at Measured soil set-up and calculated set-up in
the site consisted of about 6.1 m of miscellaneous 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, soils depends on soil conditions, pile material and
the proposed formula provides better fit to results dimensions. However, the same equation with a
tested than the existing formula. different factor, B, can be applied for assessment of
soil set-up in cohesive soils during relatively long
elapsed time after pile installation.
5 SUMMARY
The existing and proposed formulae demonstrate
similar tendency of the set-up effect of cohesive soil
The application of the proposed formula shows that
in pile capacity. However, there are substantial
the magnitude of gain in pile capacity in cohesive
differences between two approaches.
110
The existing formula yields relative set-up versus REFERENCES
relative time for assessment of the pile capacity after
the first restrike. This is a contradiction to the set- Butler, H.D. & H.E. Hoy 1977. Users manual for
up approach commonly used in geotechnical the Texas quick-load method for foundation load
practice. Also, it is inconvenient for certain testing. FHWA, Office of Development,
construction sites to restrike piles on 1-2 days after Washington.
pile installation. Moreover there is no standard time Davisson, M.T. 1972. High capacity piles. Proc.,
for the first restrike. If this time is different for Lecture Series, Innovations in Foundation
various piles, the existing formula yields different Construction, ASCE, Illinois Section.
assessment of soil set-up at the same site and Fellenius, B.H., R.E. Riker, A.J. O’Brien, & G.R.
obtained results of increasing in pile capacity cannot Tracy 1989. Dynamic and static testing in soil
be compared. exhibiting set-up. Journal of Geotechnical
The proposed formula uses the traditional set-up Engineering, 115(7): 984-1001.
formulation, calculates the pile capacity at the actual Fuller, R.M. & H.E. Hoy 1970. Pile load tests
time after EOID, and provides determination of the including quick -load test method, conventional
soil set-up independently of the time of the first methods and interpretations. HRB 333: 76-86.
restrike. GRL and Associates, Inc. 1993. CAPWAP - Case
The proposed formula provides determination of Pile Wave Analysis Program, Manual, Cleveland,
pile capacity as a function of time after pile Ohio.
installation using pile capacity values obtained GRL and Associates, Inc. 1987. Dynamic pile tests
through dynamic testing at EOID and one restrike or pe@ormed during June and July, 1987, Advance
one static loading test. This approach is identical Pile Test Program, Louisiana DOT, Project No.
for any construction site. Obtained information 450-36-06, Cleveland, Ohio.
would be beneficial to choose sensible pile Randolph, M.F., J.P. Carter & C.P. Wroth 1979.
penetration depth at EOID and search the proper Driven piles in clay - the effect of installation and
times in days after EOID to verify set-up in cohesive subsequent consolidation. Geotechnique, 29(4) :
soils with an additional dynamic testing or the static 36 1-393.
loading test. Skov, R. & H. Denver 1988. Time-dependence of
bearing capacity of piles. In B. Fellenius (ed),
Proc. Third Inter. Con. on the Application of
6 CONCLUSIONS Stress- Wave Theory to Piles, Ottawa, 25-27 May:
879-888, Vancouver: BiTech Publisher.
Determination of long term pile capacity is Seed, H.B. & L.C. Reese 1955. The action of soft
important for proper design and construction of pile clay along friction piles. Transactions, ASCE,
foundations in cohesive soils. 122: 731-754.
A new relationship like a linear equation in a Svinkin, M.R., C.M. Morgano & M. Morvant
logarithmic time scale has been derived to predict an 1994. Pile capacity as a function of time in clayey
increase in pile capacity with time after pile and sandy soils. Proc. Fifth Inter. Con$ and
installation. Soil set-up for several piles in cohesive Exhibition on Piling and Deep Foundations,
soils was verified on the basis of the existing and Bruges, 13-15 June: 1.11.1-1.11.8, Rotterdam:
proposed formulae. The latter has certain Balkema.
advantages. Obtained results showed that changes of Thorburn, S. & W.J. Rigden 1980. A practical
pile capacity with time in cohesive soil may be study of pile behavior. Proc. 12th Annual OfSshore
predicted well. Technology Con., Houston.
Results presented in this paper certainly do not Tomlinson, M.J. 1971. Some effects of pile driving
mean that pile capacity will change with time exactly on skin friction behavior of piles. Proc. Institution
like shown above. However, the demonstrated pile of Civil Engineers: 107-114, London.
capacity versus time relationships can be used as Wardle, I.F., G. Price & T.J. Freeman 1992. Effect
guide for assessment of pile capacity with respect to of time and maintained load on the ultimate
time. Presented results give a chance to safe capacity of piles in stiff clay. Piling: European
significant amount of time, energy and materials practice and worldwide trends, Proc. Institution of
taking into account the gain of pile capacity from Civil Engineers: 92-99, London: Telford.
soil set-up. Derived relationship can also be useful
in choosing representative times for both static
loading test and dynamic restrike testing.
111
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
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.
113
2.2 Unequal numbers of layers
Relationship between equivalent stiffness and where all parameter are as defined previously.
equivalent modulus of elasticity is After transformation and simplification,
equation (8) becomes equal to equation (6).
According to equation ( 6 ) , equivalent stiffness of the After transformation and simplification, equation (9)
composite hammer cushion depends on the modulus takes the form
of elasticity, the layer thickness, and the cross-
section of a soft material.
After analogous simplification, equation ( 5 ) takes
the form
According to equation (lO), equivalent stiffness of
(7) the composite hammer cushion filled with three
materials depends on the modulus of elasticity of
where a! = tstf/tsoft. two soft materials, the thickness of a soft material
It can be seen that equivalent modulus of elasticity layer and the cross-section of cushion materials.
Relationship between equivalent stiffness and
of the composite hammer cushion depends on the
modulus of elasticity of a soft material and a ratio of equivalent modulus of elasticity is
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 %
Equating the right parts of equations (10) and ( l l ) ,
in comparison with results of equations (3) and (5).
and, assume Esoft=2Esoftestfor simplicity, we obtain
114
(12) 4.2 Three cushion materials
115
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 1503
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.
117
3 . LABORATORY DRIVING TESTS hammer system is much lower, directional stability
during driving is improved. Subsequently tip deflec-
The driving tests are performed in a tank of 1.50 m tion, torsion and bending pile stress can be avoided
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 4 STRESS WAVES PROPAGATION IN THE
L’Down-the-hole” piling tests are performed without PILE
using the device allowing uncoupled penetration of
pile tip and pile shaft. The hammer strike simply in The typical time-force signals recorded during down-
the bottom hole of pile (figure 1). The pile model the-hole driving at top of pile is shown on figures 2 .
made of steel, is 1.50 in long and 0.42 m of diameter. Unlike the top piling where the initial stress wave is
The conical pile tip with an angle of 40 degrees is of compression and propagate from the pile head to
welded to pile shaft. The driving system is composed the tip, in “Down-the-hole’’ driving the initial wave
of a 2 in long tube surrounding the top of the model- at pile tip is of tension and propagate upwards to the
pile, and guiding a free falling mass which impacts on top. The tension wave TI (figure 2) is then reflected
top or inside the hole pile, without any cushion. The at pile top as a compressive wave C I . After the first
test pile is instrumented with strain gauges in order to up and down travel of the stress wave in the pile, its
collect the stresses near the head and the toe of the amplitude is quite reduced by shaft and tip interaction
pile. A displacement sensor is placed at the pile head. 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
Figure 1 : Laboratory Hole-Pile-Driving device
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
The “Down-the-hole” Piling technique is tested
to the pile top (figure 4). The energy transfer from
and compared with top pile driving by the hammer
the pile to the soil is made through the tension wave
involving the same energy In driving piles it is essen-
transmitted and reflected at the pile tip.
tial to control the hammer energy and the way it is
The stress waves measured on the pile allow to
transmitted to the pile and the soil. The energy in-
identify the incident and resistant waves using the
volved in the two types of driving tests is ranging
stress wave theory in an unidimensional elastic media.
from 13.5 to 34.6 J. In top piling the hammer pushes
Stress wave measured in top driving show that the
the pile into the soil, allowing a compression stress
wave duration increase after the successive reflections
wave, while in “Down-the-hole” driving the hammer
at the pile extremities (Benamar 1999b). This effect is
pulls the pile down, creating a tension stress wave
caused by the high amplitude of shaft resistance. But
The hammer used is enough long to develop a S U E -
in “Down-the-hole” driving the initial tension wave
cient blow duration to mobilize the total shaft resis-
is weakly influenced by friction waves during
tance Because the center of gravity of the pile and
propagation along the pile.
118
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.
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
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
REFERENCES
120
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 1503
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.
121
pile test data made it possible to investigate the use (memorizing the training pattern rather than
of neural network for estimating the soil resistance generalizing the prediction). On the other hand,
using the same parameters as the simplified dynamic using fewer hidden neurons often increased the
test data (High Stress Testing, (HST)). 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.
2. NEURAL NETWORK APPROACH One rough guideline for choosing the number of
hidden neurons in many problems is the geometric
2.1 General pyramid rule. It states that, for many practical
networks, the number of neurons follows a pyramid
Recently some investigators suggested artificial shape, with the number decreasing from the input
neural networks as a model for the prediction of the toward output. The number of neurons in each layer
behavior of axially loaded piles, using Hiley formula follows a geometric progression. Other investigators
(Chan et al., 1995; Teh et.al. 1997). No attempts (Berke and Hajela, 1991) suggested that the nodes
have been made so far to investigate the application on the hidden layer should be between the average
and the sum of the nodes on the input and output
of artificial neural networks in the prediction of the
layers. These are only rough approximations to the
soil resistance under dynamic loading using a ideal hidden layer size. The best approach, found by
combination of HST data and soil properties from the authors, to find the optimal number of hidden
simple soil tests. neurons is to start with a few numbers of neurons,
The main objective is to develop optimal models then slightly increase the number of hidden neurons,
using only simple input data. These data include the until no significant improvement is noted.
geometry of the pile, soil properties in terms of type A variety of different transfer hnctions were
and SPT-N values and HST data. The models investigated to achieve best performance in training
involved are backpropagation neural networks. 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
2.2 Neural network structure and output layers respectively.
122
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
124
CONCLUSIONS Master, T. (1 993), “Practical Neural Network
Recipes in C++”, Academic Press Inc., New
Simplified backpropagaton neural network model is York.
introduced to estimate the soil resistance under Rausche, F., Moses, F. and Goble, G.G. (1972),
dynamic loading conditions. The approximation of “Soil resistance predictions from pile dynamics”,
the physical behavior of the pile in these models is Journal of soil mechanics and foundation
based only on basic data, like pile length, pile cross division, ASCE, 98, pp.917-937.
sectional area, circumference and HST data along Smith, E.A. L. (1960), “Pile driving analysis by the
with soil characteristics. This combination of data wave equation”, Journal of soil mechanics and
seems to enhance the prediction accuracy of the foundation division, ASCE, 86, pp. 35-61
neural network models. The presented neural Teh, C.I., Wong, K.S., A.T.C. Goh, and S.
network structures provide promising results that Jaritngam, ( I 997), “Prediction of Pile Capacity
will be beneficial in the practice. using neural networks”, J. Comp. In Civ. Engrg.,
The chief preference of the neural network London, pp. 1-8 .
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.
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
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.
127
based on the simulation of the problem using exist- 4 MODEL PARAMETERS
ing experimental data, and would allow for bearing
capacity predictions to be made in the field, simulta- It is known that several factors can influence the
neous with the collecting of data. bearing capacity of a pile and, consequently, the re-
Other authors have already used neural networks sult of a dynamic load test obtained by the CAP-
to determine the bearing capacity of piles using data WAP method. However, it is impossible to precisely
from load tests. For example Teh et. al. (1997) used assess the relative importance of the several inter-
digitized samples of the force and velocity curves as vening factors, due to the complexity of the dynamic
inputs to the network, and Kiefa (1998), utilized phenomenon.
general regression neural networks to obtain the Thus, several variables were selected to be used
shaft and tip resistance of driven piles in cohesion- as inputs to the neural networks, in an attempt to
less soils. A different and innovative approach is study the importance of each one in the deterniina-
used in this work, as described in the following ses- tion of the total mobilized resistance by the CAP-
sions. WAP analysis. Those variables are listed on Table 1,
together with their code names.
128
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
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.
130
Figure la. Training results for network " 3 Figure Id. Test results for network "6
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
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
Figure 3a. Test results for the revised (10 neurons) NN3 2. Neural networks with a higher number of neurons
network using data set TEST02 in the generalization process 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
Figure 3b. Test results for the revised (10 neurons) NN6 net.- REFERENCES
work using data set TEST02 in the generalization process
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).
133
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.
134
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
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.
135
accelerated more than the ram because of the soft For soft pile cushions the peak helmet velocity is
pile cushion spring. The helmet can move 1 to 20 higher than the impact velocity and the pile head ve-
mm before a reaction force is activated in the pile locity. During driving the pile cushion is compressed
cushion. Because of this high velocity, there will be a permanently and the cushion thickness is reduced
gap between the hammer cushion and the helmet just continuously till it is completely burned. During the
after impact. After some milliseconds the gap is
closed and a second impact will occur.
136
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.
137
Figure 5 : Measurements made in the field
For each blow a measurement as shown above has During these test no damage of the pile head has
been made. The tangent stiffness of the pile cushion been noticed when using plywood cushions. The
at 5 , 10, 15 and 20 MPa stress level is shown as maximum tensile stresses in the pile were about the
hnction of blow number in figure 6. After about 950 same as with softwood.
blows the pile cushion was replaced with a new one.
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.
139
Behavior at 95 " (e=0.70; E25 = 900 MPa)
Behavior at 50 " (e=O.76; E25 = 1700 MPa)
kept constant for a very short period and the load is Presently IHC is using a nylon based material,
removed over a period of 0.3 seconds. A lot of alter- variation on the MC-901. For this materials the ob-
native hammer cushion materials have been investi- served behaviour is shown at different temperatures
gated. The main selection parameter was the coeffi- (20", 50" '80" and 95" Celcius)
cient of restitution measured during the tests and the The measured stress-strain curve for loading and
Young's modulus at different stress levels. unloading of the cushion material are fitted with the
following function:
140
shown in figure 10. Under impact conditions the
stiflkess of the material will increase approximately
The third degree fbnction gives the best results by 30 %.
and is required when describing cushion behavior.
This material model is incorporated in the
TNOWAVE-R stress wave simulation program. For 5 PILE DRIVING PREDICTION
unloading a linear behavior is assumed.
During the driving of a concrete pile with a SC-75
The tangent stifiess at a stress level of 25 MPa hammer, the behavior of the drive setup was meas-
and the coefficient of restitution as hnction of tem- ured. With the TNOWAVE-R program a prediction
perature is shown in figure 9 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.
141
The PDA measurement was done about 3000 Shen, H.R., Yu, Z.C, Tang, N.C., Studies on Pile
blows on the hammer cushion. This means that the Driving Cushion MaterailS. Proceedings ofthe Int.
temperature of the hammer cushion will be about 80 Conf. On Deep Foundations, Beijing, 1-5 Septem-
"C This has been adopted in the stress wave simula- ber 1986.
tion.
6 CONCLUSIONS
REFERENCES
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.
145
2 INVESTIGATION PHASE 2.4 Impuct-Echo somdings
2.1 General 2.4.1 General
The underbase grout situation under the West Pylon Non Destructive Tests were performed in 665 loca-
Caisson was investigated intensively. Several differ- tions covering all 42 compartments. The Impact
ent methods were used to clarify the situation under Echo Method was used for the NDT investigations.
the base slab prior to repair works and to determine The purposes of the NDT tests were:
the best method for performing the repair of the de- 0 To increase the knowledge about the situation
fective underbase grout layer. existing underneath the base slab prior to the repair
The investigations clarified the following two main works. To clarify the presence and spatial distribu-
subjects: tion of voids and sand in the underbase grout layer.
1. Detection of voids in the interfaces: bottom a To use the status prior to repair works for later
slab - underbase grout, underbase grout - limestone verification of the repair works by comparative
and inside the underbase grout layer. studies.
2. Verification of the structural capacity of the The Impact-Echo method is based on stresswave
entrapped sand and the existing underbase grout analysis of an impulse response. The chosen system,
layer. PIT-Collector (Pile Integrity Tester) from GRL is
To investigate the situation, the following tests and basically used for integrity tests on piles. The
investigations methods were used: stresswave from an impulse of a small sledge-
0 Core drillings in 7 drilling stages. hammer of steel is registered with an accelerometer.
0 Endoscope investigations. The acceleration is integrated to velocity. The ve-
0 Impact-Echo measurements. locity record is studied in the time as well as in the
frequency domain. A simple sketch of the execution
2.2 Cove drillings of the measurements is shown in Figure I and Equa-
tion 2. l below.
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.
146
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
147
using Fast-Fourier-Transformation (FFT). This re- The interpretations that can be done are presented
sulting curve represents the velocity response for the in Table 2 below. The degrees of connection were
impulse in the frequency domain. In this curve the evaluated by studying the Sansalone and Streett ref-
peaks are studied. The important details are to find erence.
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- Table 2. Interpretation of bond between concrete and
lying material is high i.e. reflection against air, water grout.
or very loose sand. A damping material such as sand Degrees of connection Classification
or grout causes a more rounded and broader peak. If between slab and grout Peak
the grout has very high acoustic impedance the 100 % A
stresswave tends to travel through the grout with 80 - 100 % B
only small amount of reflection. 25-75 % C
In hnpuct-Echo. Non-destwcti\*e Evuluution of 0-25 % D
Concrete m d Mmonry, Sansalone and Streett 0% E
( 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 The results of the measurements were classified in
case be compared with different degrees of connec- five main groups A-E where A represents full con-
tact and E represents no contact. Eight subgroups
tion between the concrete slab and the underlymg
were also established. All thirteen groups were in-
grout. The investigations showed that voids and terpreted further in order to establish three areas.
pockets between the concrete and the grout most @ Full or close to full contact between slab and
ol'ten caused the problem. Findings of washed in
sand were very limited. grout [A, ( B N , ( A B B, (GIB].
Partial contact between slab and grout [(B)C,
To compare theory with practice, five curves with
different shapes of the peaks were chosen. Such c , (DKI.
curves from the West Pylon are presented below in @ No or low contact between slab and grout
Figure 4. A short description of significant features [(CID, D, ( W , (D)E, El.
of each type of peak is given below. The NDT-measurements were correlated to core
Peak A. The response had a nice rounded shape. drillings and endoscope investigations, as the NDT-
The peak could often not be distinguished. measurements solely could not tell the actual physi-
Peak B. The response was broad and had a simi- cal reason for a reduced degree of connection.
lar look before and after the peak. The peak could be
hard to distinguish. Often it was possible to notice a 2.5 ~otrc1usion.sofflie iizvestigutionphase
response from the interface grout-limestone (the first
vertical line in the Figure) Using the information available from all investiga-
Peak C. The appearance of the peak was more tion results the situation under the Pylon base was
difficult to distinguish if just looking before the described by establishing three separate categories:
peak. The fall of the peak was though quite obvious. 1. Area with good contact.
The response before the peak was more "filled up" 2. Areas with small gaps 0 - 2 mm.
than after. 3. Areas with gaps > 2 mm.
Peak D. The peak was more rounded than in peak The average size of the voids under the baseslab
E. The fall of the peak was steeper than the rise. The observed by the endoscope was < 10 nim but in cer-
response around the peak became broader. tain areas the voids were o f 3 0 - 120 mni thickness.
Peak E. Both the rise and the fall of the peak was Regarding the interpretation of the low strain meas-
very steep. The response before and after the peak urements most importance were attached to the ve-
had offen low amplitude. The peak was very sharp. locity-spectrum curve. The structural capacity of the
148
existing underbase groutfill was verified by per- 0 Coverage of the area to be repaired.
forming unconfined compression strength tests on 0 Structural requirements and foundation de-
grout cores. Grout specimens were taken from vari- mands regarding support.
ous depths and locations from the underbase grout The core drilling results were evaluated espe-
fill of the West Pylon Caisson. The average com- cially with regards to:
pressive strength was sufficient compared to the 0 Loss of core (comparison of core recovery
requirements. The load carrying capacity of the en- with drilling depth).
trapped sand was tested with respect to the static and
0 Evidence of injection grout.
cyclic loading conditions. Based on the test results
the designers demonstrated that entrapped sand in its 0 Filling of the voids under the Pylon Caisson
in-situ state provided adequate bearing capacity and baseslab with injection grout.
thus no further remedial measures on entrapped sand 0 Bond of injection grout or existing underbase
was required i.e. soil fracture grouting was not grout to the baseslab.
needed.
4.2 Endoscopy Iizvestigatiom
3 FOUNDATION REPAIR The boreholes of the verification drillings were en-
doscoped and video recorded subsequent to the core
The foundation repair works, injection works, were drilling. The following subjects were especially con-
performed between 20th November 1997 and 2 1st sidered when endoscoping the verification drillings
February 1998. The injection works for filling the in addition to the main criteria established in the
voids under Pylon baseslab were executed according investigation phase:
to the following five principal steps: 0 Bond between the bottom slab and injection
1. Sealing of limestone fissures. grout.
2. Hydraulic connectivity test (HCT). Prior to 0 Bond between injection grout and existing un-
injection works tests of the hydraulic connectivity derbase grout.
between adjacent boreholes checked the connections 0 Any evidence of filling of voids, cracks and
of void areas. fissure by injection grout.
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- 4.3 Non Destructive Testing
cess pressure to fill voids and fissures. Later on the Non-destructive tests were carried out continuously
pressure in this step was increased to 6 bar. during the injection works. The measurements were
5 . 2nd Injection Step, i.e. control of filling injec- used, first to control the grouting works as an active
tion. The 1st filling injection was controlled by an design, second, to verify the filling injection works
additional injection step at an excess pressure of up after completion. The tests were performed in sev-
to 10 bar. The 2nd injection step was repeated in eral steps as the grouting works were executed in
most of the areas pending intermediate NDT-test
results. nine separate zones. The number of measuring posi-
3 3 tions taken in the zones subsequently to the comnple-
In total, almost 15 m of injection grout, 10.1 m
tion of the repair works were 595 positions covering
in the filling injection step 4 and 4.8 m’ in the repe- all areas where injection works were performed.
tition injection steps 5, were introduced under the The results of the measurements were classified
baseslab of the West Pylon Caisson. according to the investigation phase in five main
groups A-E and eight subgroups where A represents
full contact and E represents no contact.
4 VERIFICATION PHASE The voids under the bottom slab were filled by
4.1 Core drillings injection grout during the repair works. The impact
Verification drillings were performed after comple- echo method was used to verify the foundation con-
tion of the injection works. The main purpose of ditions of the Pylon caisson during and after com-
pletion of the repair works.
these drillings was verification of the injection
works by endoscopy. However core drilling using a Examples of improvements after the repair works
geotechnical drilling technique was executed to are presented in Figures 5a-c. The result was inter
achieve a good core recovery on the verification preted as (E)D curve evaluated prior to repair works
drillings. For verification purposes in total 11 drill- improving to a (C)D after the first treatment result-
ings were performed. The location of the drillings ing into a B after the repeated treatment.
were selected considering the following criteria: The measurements were performed when the ini-
0 Size of the voids being filled by the injection tial part of the curing process in a treated area was
works. completed. The interpretations of the curves were
8 NDT-test results prior and post injection compared to the previous measurements taken prior
works. to the repair works at the same location.
149
Figure 5a. Velocity spectrum 9708 14. Evaluated as
(E)D. i.e. area with no/little contact.
150
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
REFERENCES
151
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
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.
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)
~
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.
155
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 1503
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 of resonance frequency - normalized by ASTM
Engineering is the quality control, and the attendance C215/60 - and with the determination of the speed of
to the specification of structural elements in service. propagation of sonic and ultra-sonic pulses in the
At the present time, it is observed that both the safety concrete, not yet normalized);
and reliability of any geothecnical work are strongly Integrity of the structural element;
influenced by the executive process. In this context, 0 Maintenance of the compressive resistance.
it is also advocated Mello (1981) an isolated factor The Pulse Echo Method or Pile Integrity Test
of safety for the execution of the engineering work. (PIT) is one of the tests most widely used for the
The field tests in structural elements can, in a post installation inspection of the integrity of
general way, be divided into destructive and non concrete piles, drilled shafts, augered piles (CFA’s)
destructive tests. The destructive test demands the and, eventually, timber piles. This test is also known
molding of special test samples, which don’t by its easy execution, with great reliability allied to a
necessarily supply direct information of the concrete low cost. It appears, then, to be a potential tool to
from the real structure. In this case, it becomes more the effective cost of quality control of deep founda-
interesting to use extracted samples of the structure tions.
in service. However, this procedure can damage the The PIT test has been commonly used in deep
structure, even if in a localized manner. To avoid this foundations to evaluate the integrity of piles with
problem, non destructive tests have been developed suspicion of constructive defects, as fissures, cracks
to be used directly onto the structure. They are the or variation of the shaft cross sectional area. The PIT
most widely used methods in the engineering and one might also be used to evaluate the pile length in cases
of their great advantages is the accomplishment of were the wave propagation speed can be assume
multiple tests in the same place. These methods can This speed is a function of the density and the quality
also be performed with time, in order to evaluate the of the concrete, hence, it can be correlated with the
variation of the characteristics of the structure during characteristic resistance of the pile. It can also be
its useh1 life. monitored with time, thus allowing the determination
One of the most used types of non-destructive test of the quality of the foundation along its normal life.
is the sonic (or ultra-sonic) test. They have been The PIT test assumes that the stress (or acoustic)
continuously and increasingly used in the last years. wave propagates in one dimension. This will require
They supply the knowledge of the concrete fissures a foundation with a minimum ratio in term of
plus the material’s deformation. 1engtMdiameter about 5 (Pile Dynamics, 1998) to
Some of the internal characteristics of the 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 Modulus and inversely proportional to the square
The testing program consisted of PIT tests in order root of the pile’s material density. This is due to the
to determine the length and the integrity of some of fact that the elastic parameters supply the restoring
the piles. Due to time restrictions only two piles forces of the displaced regions, and the density is
foundations were tested. This was done because no related to the inertia of the system.
(post-construction) information could be obtained The test consists of measuring the response of a
low strain impact caused by a small hand-held
from the original blueprints.
hammer. Each blow generates a compression wave
Two blocks in particular were selected, due to its
that spreads from the pile top in direction to its
location and load capacity within the complex
bottom. This wave can be reflected during its
Blocks PI6 and P41 (related to pillars 16 and 41) are
journey at each cross sectional variation of the
those studied herein, and can be easily found in
shaft, thus generating second order waves
Figure 2.
propagating with the opposite direction.
The traces of the acoustic wave signals are
3 TEST METHODOLOGY captured by an accelerometer, which is attached to
the pile head. This signal is hrther converted into
The PIT test is based on the acoustic wave an electric signal, which is amplified and sent to a
propagation at a low deformation level. These conditioning center where is transformed from the
waves are measured by using a PEM (Pulse-Echo analog to the digital format. All this information is
Method) equipment. The waves are longitudinal stored into a data acquisition system which is built
and they induce displacements in the same in the PIT unit system. Once this stage is ac-
direction of their propagation. They generate complished, the data can be transferred to a PC for
dilatation and compression effects in the material, hrther processing of the signal (interpretation
when passing to all sort of material types (solids, stage).
liquids or gaseous). Their propagation speed is The verification of the pile integrity is made
proportional to the square root of the Young’s through the interpretation of the shape of the
159
obtained (reflected) acoustic wave recorded at the accelerometer to be fixed in the vertical position.
pile head From the tests, it is possible to obtain Figure 3 depicts, schematically, the position of the
graphs of velocity and force versus time (which can lateral windows in each studied pile of the pile
be later translated into pile length) or frequency group.
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 4 CRITERIA FOR SIGNAL EVALUATION
shape of the velocity/force measured curve Those
changes, and the curve itself, allow for the The PIT test should be used in precast concrete
establishment of conclusions regarding the condi- piles, which are driven into place with the use of
tions of the pile, and the location of any eventual standard crawling cranes. This is so because this
structural damage, bulging, etc along its shaft type of pile has, in general, an almost perfect cross
sectional geometry, thus reducing the effect of the
3 1 Pi.eparatron of the pile top reflected “secondary” acoustic waves. Never-
One of the most important steps for any PIT test is theless, more recently, some authors started using
the preparation of the pile top surface It is well the PIT in the evaluation of cast in place piles
known that the accelerometer (and the hammer (Arias, 1977, Rausche & Seitz, 1983, Berger &
impact blow) should be placed onto a horizontal Cotton, 1990, Cunha & Costa, 1998 and others).
and free of dirt surface Hence, the accelerometer With this objective, Cunha & Costa, 1998 have
can be attached only after a horizontal and proposed a simple alteration to the evaluation crite-
reasonably free of dirt top surface is prepared ria (grading letters) presented by Rausche et al.
Generally this attachment is made with the use of a (1 988) with the PIT acoustic wave results. Cunha
thin layer of a soft paste (vaseline, etc ), allowing a & Costa (1998) modified this grading criteria to in-
good recording of the impact and post impact clude an “extra” letter C’, as follows, to be specifi-
wave An imperfect installation will undoubtedly cally used in the case of cast in place piles:
hamper the quality of the signal (which will have
low amplitude, noise, etc ) A- No significant defect is apparent in the
The generated acoustic wave will be of clear and records. It shows a perfect reflection in the
sharp magnitude solely when an intense impact, of pile toe, and no other significant reflection
short duration, is produced at the pile head prior to this point. The pile is considered as
Therefore, it shall be again emphasized that the physically sound;
correct preparation of the “impact surface” is a
fundamental step for the success of this test B- No significant defect is apparent in the
In order to carry out the tests at the INEP records, however the pile toe reflection was
building, it was necessary to excavate the capped not observed, thus indicating that the full pile
block down to its base (at the block/pile interface) length wasn’t tested. The pile is considered
This was done in order to open a “window”, or as partially tested;
lateral niche, where the hammer blow could be
applied This small recess was flattened out by C- The record showed significant impedance
using a thin layer of mortar, thus allowing the changes, but the pile toe reflection was
clearly observed. The pile is considered of
limited value;
160
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 Rausche, F. e Seitz . J. 1983. bitegrip testing of sliaff.s arid
caissons. Specialv Session on Shafts and Caissons.
the existing pile groups, which will be subjected to ASCE Annual Convention,
new superstructure loads Rausche, F.. Likins, G.E. e Hussein, M. 1988. Pile Ziitegriry
In the foundations tested herein, the PIT tests b,v loiv a i d high straiii impacts. Seminar on the Applica-
detected sound piles with a reasonably good quality tion of Stress Wave Theory on Piles.
concrete. The piles had a length of around 8 m, Rausche, F.. Beim, J.. Beim, G. e Morgano M. C.. 1999.
and a shaR of primarily bulges ( in a random Custom tailoring qualily assurance ,for augercast piles.
XI Panamerican Conference on Soil Mechanics and
manner). These results, although few for the site, Geotechnical Engineering . Volume 3 Brazil. p. 1466-
were extrapolated to all other existing piles, given 1370.
the fact that fbrther PIT tests could not be per- Sales. M.M.. Cunha, R.P. & Jardim. N.A. 1999. Analysis of
formed (due to time constraints). Piled Footing Tests 011 a Tropical Porous Clgv. XI
The very simple example presented in this paper, Panamerican Conference on Soil Mechanics and Geo-
therefore, is another indication that the PIT has a technical Engineering. Volume 3 Brazil. p. 1395-1402.
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.
ACKNOWLEDGMENTS
REFERENCES
162
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
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.
163
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.
164
Stiff Clay boundary is present. Figure 2 illustrates Basically, following formulae are used in analysis of
the typical features of external forces, which cause a sonic integrity test.
crack in the piles. Pile damage associated with sheet
pife wall excavation is presented in this paper. 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
Sonic integrity testing also called as low-strain in- Pile toe reflection (for pile length verification)
0
165
investigation (if applicable excavate down to level of movement as well as settlement of ground including
detected crack for visual inspection) may require be- tension cracks of more than 300mm in width were
fore final conclusion is made. observed at some locations. Minor damages were
As a main focus of this paper, the typical features also investigated at the adjacent properties. Tempo-
associated with the pile damage (discontinuity/ rary raking struts were immediately installed in re-
cracks) detected by sonic integrity test is discussed sponse to these excessive ground movements caused
in detail with the data obtained from three cases as by sheet pile failure. A typical scheme showing ex-
presented in the following sections. cavation works of Case I is presented in Figure 3.
7 CASE STUDIES Figure 3. A picture showing slope excavation with sheet pile
7.1 Case I - Sheet pile excavation supported by soil support - Case I
berm
7.1.4 Sonic integrity test andpile damage assess-
7.1.1 Subsoil condition
ment
A 12111 thick soft clay layer occurs beneath fill mate-
rial and weathered crust of 1.Oto 2.5m thick. Uncon- More than 50% of piles were tested with sonic integ-
fined shear strength (Su) of soft clay layer ranges rity test. Test results were thoroughly reviewed and
“site signature”, most common features of the ma-
lt/m’ at top and 2.5t/ni2 at bottom whereas the natu-
jority of piles, is established. It is to be noted that
ral water content is about 6040%. Medium clay
layer of 1 to 2m in thickness, having shear strength test signals from piles of less potential for damage
of2.5t/ni2 to 3.5t/1112 underlies soft clay layer. Be- (least effect from excavation) are selected in estab-
lishment of “site signature” or reference signal of
low medium clay, at the depth between 18 to 27m,
stiff clay with traces of fine sand having SPT N good piles. Based on the “site signature” of good
value of 15 to 37 is observed. Thick layer of fine piles, defect piles were identified and assessed. Ta-
ble 2 shows the summary of pile damage assessment
sand interbedded with clayey sand occurs under-
neath stiff clay layer. by sonic integrity test.
Table 2. Summary of pile damage assessment by sonic integ-
7.1.2 Pile foundation rity test - Case I
Pile foundation of eight high rise buildings in this
f’ilc No. of piles with defect Defcct depth
project comprises 904 cast-in-place bored piles of
Dia. Imst Less p r o ,ni below ground
diameter 1.0m and 1.5m, founded at 59111 below
(~n) promincnt prominent crack level (in)
ground level. Pile tips are embedded in dense sand
layer and the entire length of piles are reinforced crack crack
with steel reinforcement of 0.72% of pile sectional I .o 4 8 7 6.0 - 13.8
area for the top part and reduced gradually to 0.23% 1.2 10 2 6. I - 7.6
for the bottom section. To support the underground 1.5 1 9.9
water tanks, 88 pre-cast concrete piles (I-30cm)
having 21m length were also driven down to stiff The typical sonic integrity test results of undamaged
clay layer. pile and damaged pile classified as least prominent
crack, less prominent crack and prominent crack are
7.1.3 Excavation method and observed failure shown in Figure 4. Majority of piles with crack de-
Excavation was initially carried out by using sheet tected were located in the periphery of the excava-
pile (Type FSP 111) with soil berm support or slope. tion zone. Piles, particularly with prominent cracks,
Before reaching to -8.lm maximum depth of exca- were observed to be deviated from their original lo-
vation. sheet pile failure and associate severe lateral cation up to 6OOn1m towards excavation.
166
7.2 Case 11- Pile damage caused by sheet pile
excavation with one level temporary support
Figure 6. Sonic integrity test results of good pile and pile with
prominent crack - Case I1
167
Table 3. Suinmary of pile damage assessment by sonic integ-
rity test - Case I1
7.3 Case 111- Pile damage caused by sheet pile 7.3.4 Sonic integrity tests and pile damage assess-
excavation with two level temporary support ment
Though sonic integrity test was employed as a part
7.3.1 Subsoil condition of quality control regime in the original plan, it was
Underneath the top 2m of weathered crust, Soft Clay eventually applied for a retrospective investigation
layer can be observed up to 15m below the ground after encountering many piles with defect. Hence all
surface. Medium Clay layer underlies Soft Clay piles were undergone sonic integrity test. According
with the thickness of 2m. Undrained shear strength to the test results, 84% of piles located within the
of Soft Clay ranges between lt/m2 and 3t/m2 excavation zone were identified to be with crack.
whereas it is about 50t/m2 in Medium Clay. Beneath Sonic integrity test results of good pile, pile with
Medium Clay, Stiff Clay layer is found up to 25m less prominent crack, and pile with prominent crack
where the boundary of Hard Clay is investigated. are illustrated in Figure 8. As can be seen in Figure
Within Stiff Clay stratum softer zone of Medium 8, for pile with prominent crack, a distinct aiid sharp
Clay is identified at the depth between 19 and 2 l m reflection is evident at depth about 5.5m below pile
characterized by reduction in SPT values which ap- top. Repetition of reflection at multiples of this
pear consistent with increases in iiioisture content. depth can also be observed for this pile. These fea-
tures are comparatively not distinct for the pile with
7.3.2 Pile joundation less prominent crack. An excessive basal heave, in-
A total of 143 bored piles of 0.6m diameter were in- sufficient reinforcement to resist the tensile stress
stalled with pile tip at about 26m below the ground and relatively small pile size are considered to be the
to support a residential apartment. Dry process was main reasons causing the cracks in this case. Table
employed to construct the piles. Cut-off level of 4 shows a summary of pile damage assessment by
bored piles ranges from 0.85m to 5.95m depth. En- sonic integrity test.
tire lengths of all piles are fully reinforced with steel
bars grade SD40 of 0.35% of pile sectional area.
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$- 9 CORRELATION BETWEEN RESULTS FROM
cution FINITE ELEMENT AND SONIC INTEGRITY
Coring was carried out up to the depth below the TEST
detected crack to verify the pile damage and then it
was grouted with non-shrink cement. To justify the The analysis results indicate that bending moment in
capacity of damage piles, high strain dynamic load piles induced by excessive ground movement due to
test was carried out on 25 piles in which 24 piles excavation works are higher than that of cracking
with detected cracks and 1 pile without crack detec- capacity of pile in all cases. Table 6 presents the
tion. Summary of dynamic load test is presented in summary of the analyses.
Table 5 .
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
169
bending moment induced by excavation exceeds the Observation of actual construction works on sites
ultimate moment capacity of pile . Sonic integrity suggests that over excavation prior to installation of
test results of some piles indicate the presence of support and inadequate retaining system are the
more than one crack at some depths, suggesting piles mail1 causes leading to excessive ground movement,
would have experienced both bending and tensile a primary reason of pile damage. Excavation con-
stress due to the excavation. 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.
I 0 CONCLUSIONS
170
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
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.
171
550m deep (Balasubramaniam, 1991). Existing 5 THE SELECTION OF SUITABLE METHOD
pore water pressure conditions in upper part of BSC
are hydrostatic from nearly l m below the ground In general, non-destructive tests are conducted to
level. Then the hydrostatic conditions change to pie- verify the quality of the pile. Turner (1997) high-
zometric draw down near bottom level of BSC. lighted the factors in selection of pile testing in
CIRIA Report 144 as follows;
The perceived nature of possible features or de-
3 COMMON CONSTRUCTION METHOD
fects within the pile
The common practice of bored pile construction The ability of the test method to detect the fea-
method in Bangkok can be divided into wet and dry ture or defect under investigation
processes. The cost of testing and examination
For the wet process, bored piles are constructed The ease of use and interpretation
using bentonite or polymer slurry as drilling fluid
over the full length of the piles, to ensure the stabil- Like other indirect measurement of pile integrity,
ity of the excavation at all stages. Top 15m-thick sonic integrity and sonic logging tests are only capa-
soft clay was temporarily cased. Drilling for the ble of identifying structurally significant features.
piles was done with a conventional drilling auger or Rejection or acceptance of individual piles should
auger bucket. To clean the sediments at the borehole not rely only on the results of these tests. Further in-
base the airlift technique or cleaning bucket is ap- vestigation, engineering evaluation and judgement
plied. After lowering the rebar cage, concrete is are highly recommended to confirm the defect de-
poured with tremie method. tected by these indirect techniques. Both methods
Generally the dry process is employed for small however are very useful for cost-effective screening
piles (0.35m to 0.6m in diameter and 25m in length). test to identify piles with potential defect.
This method uses a steel temporary casing of about Sonic integrity test is usually selected for both
151~1 in length to protect the soft clay froin caving in. quality check (control test) and retrospective investi-
Boring is carried out by rotary drilling or percussion gation. It is the cheapest in terms of cost and the
tripod rig. simplest in terms of testing process. The main ad-
Bored piles are commonly constructed with rein- vantage of this test is that since no pai-ticular meas-
forcement of 0.5-1.2% of sectional area of pile. 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
4 COMMON DEFECTS CAUSED IN BORED integrity testing needs considerable experience and
PILES knowledge in testing, subsoil condition and con-
struction method.
Common pile defects such as size reductiodnecking, Sonic logging test is relatively expensive. It is
discontinuity, soil/slurry inclusions, etc. may be
caused during pile construction. Most of the cracks mainly employed as a pre-planed site quality control
in bored piles caused at post construction stage are testing. The major advantage of this method is that
usually induced by construction activities associated test can be carried out shortly after the pile
with basement excavation and improper trimming of construction. Hence, rectification measures can be
pile head to design cut-off level. Causes of defects implemented while the foundation contractor is on
in bored piles during construction are; site. However, this method is generally not applica-
ble if pile integrity is in question due to post con-
Inadequate length to protect the soft clay layer struction activities as access tubes are usually
Delayed feeding of slurry and improper slurry grouted after completion of the test.
level maintenance for wet process High strain dynamic load test is usually selected to
Long time duration of maintaining open borehole verify the load carrying capacity of piles. Pile integ-
Non continuous concrete pouring or disruption in
concreting rity can also be determined by high strain dynamic
Lifting up of insufficient workable concrete or load test. The major advantages of this test in com-
hardened concrete upon extracting the temporary parison with static pile load test are those of cost,
casing time and space requirement. In some projects, dy-
Contamination of concrete with drilling slurry or namic load test was applied to justify the capacity of
soil pile after rectifying its defect detected by other integ-
Soft pile toe due to improper base cleaning rity tests. Table 1 shows the suitability of test
method for different type of common defects.
172
Table 1. The suitability of test method for different type of de- attenuation of the sonic signal by shaft friction is
fects. relatively low.
Type of defects Test method and suitability
for defect detection
Sonic Sonic High strain
integrity logging dynamic
~
load test
Crack Suitable Possible Suitable
Pile size Suitable Possible Suitable (if
change severe defect) Figure 2. Sonic signal showing a toe reflection of bored pile at
Poor quality con- Possible Suitable Suitable (if 52.0m tested few days after concreting.
Crete or Contami- severe defect)
nated concrete Use of temporary casing usually causes larger size at
Soft pile toe - Not usually Suitable 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
6 SONIC INTEGRITY TEST 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
6.1 Overview of the test in Bangkok nominal diameter formed in the borehole made by
Sonic integrity test, also known seismic test is the drilling tool (Figure 3). Test signal shown in Figure
most common method of integrity testing for both 3 indicates a clear reflection at approximate 1 l m
from pile top (about 15m below ground level) with a
driven and bored piles in Bangkok. In many projects
it is a part of the contractual requirement to conduct repeat or multiple reflection at about 2 times of that
sonic integrity test. Minimum 10 % to maximum distance (approximate 22m from pile top). As the
100 % of production piles are commonly tested. It is polarity of the reflection is negative, the first inter-
pretation is that this pile has decrease in cross sec-
also a reasonably acceptable method for applying as
a retrospective investigation in determining integrity tion or crack at 15m below ground level. However,
of the pile. 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
6.2 Interpretation of the test results 15m long temporary casing. The sonic signals with
As has been reported in various literatures, it is to be anomaly acquired on those piles are often misinter-
preted as indicative of defect piles, leading to an ar-
emphasized that sonic integrity test is highly de-
pendent on the experience in both field testing and gument.
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.
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.
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
175
It is to be noted from the author’s suggestion based fect by other integrity tests. Figure 5 shows the
on the above factors that only variations of transmit photo of dynamic load test carried out on barrette of
time more than 15 to 20% of the norm should be re- size (1 .Ox2.7m) with toe depth 48.94m located at the
garded as warranting further investigation. bank of water supply canal. A 20 ton hammer was
used to activate the 2050 ton test load in this project.
176
ameter piles respectively. As can be seen in the fig- ACKNOWLEDGEMENT
ures, load-settlement characteristics obtained from
dynamic load test agree well with those of static load The authors wish to express their appreciation to
test for both large and small diameter bored piles. 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
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.
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
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
181
The calculated and measured pile head velocities might be due to loose contacts between the pile and
for Cases-5 and 6 are shown in Figure 7. The the surrounding soils.
assumed profiles of VS of the model grounds are
shown in Figure 8, together with the measured VS.
2.3 FEM analyses of the static load tests
The profiles of VS estimated f?om the sonic integrity
tests seems to represent the measured results well, FEM analyses were conducted to estimate the load-
although the estimated VS underestimate the displacement curves for the model pile in Cases-5 and
measured VS of shallower depths down to 350 mrn. 6.
The underestimation of V . for the shallower depths 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 9. Axi-symmetric FEM mesh for model ground and Figure 10. Comparison of estimated and measured load-
pile. displacement curves.
182
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.
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
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
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.
187
Figure 1. Soil profile and SPT N-values at the construction site, together with estimated soil parameters.
2
E, = P C C
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 .
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:
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.
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
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.
193
(1 iic nc 1 do ilia i n . Gen era11y , proprietary systems d i s- more common. The magnitude of tlie response will
plal. data in only one or tlie other of these forms. depend not only on the end-condition, but also on
Testers tend to form their judgements based only on tlie attenuation of the incident and reflected waves.
that particular view of the data. A third form of dis- The toe response will be a half-cycle rather than the
pla) of the data will be presented in the latter sec- full cycle resulting from discrete impedance reduc-
tions of this paper. tions along the shaft.
If an amplification routine is applied to the time
domain signal, it is not possible to make an absolute
2.1 Time doiiiciin
determination on the toe condition from the ampli-
’fhe time domain display plots pile-top velocity fied response. However, the toe condition may be
as the vertical axis against elapsed time as the hori- judged relative to like piles on a project fi-om a com-
zontal axis (see Figure 1). Records are interpreted parison of responses as long as the piles are tested in
particularly by focussing on the response during the the same manner and transformed using the same
first passage of tlie input wave down and then up the amplification level.
pile (i.e. to a time of 2L/c, where L is the pile length Although the indications of the impedance reduc-
and c is the speed of the compression wave in the tion and the toe response shown in Figure 1 are
pilc). This is not to say that the response beyond clear, it is not always possible to locate these fea-
2L/c is not used to assist in interpretation. 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.
Figure 1. Typical response in time domain for pile with dis- There is another reason which can cause difficul-
Crete impedance reduction ties in interpretation. When the hammer strikes the
pile-top, the energy that enters the pile is composed
In Figure 1, the initial impact is indicated. Sub- of a spectrum of individual frequencies. The com-
sequent features are all reflections from the p,ile posite of these frequencies results in the summed in-
shal’t and the surrounding soil. From one diinen- put waveform which is indicated by the measured
sional mave mechanics theory, it can be established velocity response.
that a discrete reduction in pile impedance will gen- Each of these individual frequency components
erate a tensile reflection from the con~n~encement of travels down the pile. Elasticity theory actually
the reduction. A compression reflection will follow shows that the transmission speeds for the different
where the pile returns to its normal dimension. This frequencies are very marginally different. More im-
can be seen as a characteristic positivehegative cy- portantly. because the high frequency components
cle i n the time domain. The amplitude of this cycle contain relatively little energy, these components are
will be a fhction of the relative reduction in pile attenuated faster than the low frequency components
impedance; of the length of the impedance reduc- which contain the bulk of the input energy.
tion, and the wavelength of the input wave. The consequence is that the response received
By tlie same logic, a discrete increase in pile im- from the pile has a different frequency spectrum
pedance will cause a negative/positive cycle in the than the incident wave, being more dominated by the
response viewed in the time domain. low frequency components. The composite of the
The toe response is also a key feature which is individual frequency reflections will be a different
always sought in the time domain. If the toe re- shape to the incident wave, complicating interpreta-
sponse can be identified, it is possible to establish tion.
whether or not the pile is free of acoustic reflectors. A shortcoming of the time domain method there-
Assuming that energy reaches the pile toe dif- fore is that it does not allow discrimination of the
ferent reflections can be expected. If the pile toe ap- frequency response.
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 2.2 Freqirency domain
reflection, and a velocity decrease. Depending on As noted above, the hammer impact generates a
the stiffness of the toe condition, therefore, a spec- spectrum of waves with different frequencies and
trum of responses from large tensile to large com- amplitudes or energy contents. In order to view the
pressive could theoretically be generated. In prac- response of the pile in the frequency domain, it is
tice. at such low strain levels, a tensile response is customary to measure not only the pile-head veloc-
194
ity. but also the pile-top force generated by the As discussed earlier. this method of display is
hammer. Both the force and velocity responses are dependent on the determination of the frequency
deconstructed to establish the frequency spectrum of spectrum of both the velocity and force records.
the hammer input and the pile/soil response respec- This is normally done by a technique called Fast
tively. Depending primarily on the pile geometry, Fourier Transformation (FFT). It is implicit in this
the pile Lvill tend to resonate at particular frequen- technique that the recorded transient signal is actu-
cies when the reflection from the pile toe or an in- ally part of an infinitely recurring response. I n this
termediate feature is "in phase" with the input wave. way, the signal can be transformed into an equiva-
The result will be a heightened velocity response at lent set of infinite sinusoidal functions, each with a
such frequencies. The ratio of peak velocity re- defined amplitude and phase shift.
sponse to peak force (called the mobility) at these As the analysis is then based on infinite series,
part i c 111ar i 11 put frequencies w i 11 be a inay i mum. the frequency response analysis does not contain any
1 he frequency domain display plots pile-top mo- temporal (time) information. To highlight this, Fig-
~~
bility as the vertical axis against applied frequency ure 3 shows two signals - the first consists of the su-
as the horizontal axis (see Figure 3). 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.
195
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.
196
tion can be performed. This is then referred to as a Depending on the purpose of the time-frequency
discrete wavelet transform. For the discrete trans- analysis, it may not be necessary to impose the strict
form, the time steps are expanded in relation to the mathematical requirements of wavelet analysis.
dilation of the wavelet. 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-
4 WAVELET ANALYSIS OF LOW STRAIN correlation technique.
INTEGRITY TEST SIGNALS In cross-correlation techniques, the correlating
function should have a similar characteristic to the
4.1 Previozu investigations feature to be extracted from the signal. Such a
choice will lead to the highest correlation values.
The application of wavelet theory to low-strain in-
For low-strain pile integrity testing, the input wave-
tegrity testing has been pioneered by researchers at
form is reasonably well approximated by a half-
Napier University, Edinburgh (Addison et al., 1997; cycle of a sine wave function.
Adclison and Watson, 1997, Watson et al., 1998).
It is evidently that a half-cycle sine wave does not
In Addison et al. (1997), the authors present fi-
meet all the criteria required for wavelets.
nite-element generated responses for imaginary piles
1. The energy may not be unity. However, if an
10m in length, with and without a 20% impedance
appropriate amplitude of the “mother” sine
increase at mid-length. They indicate that the 20%
wave is chosen, a unit energy could result.
iinpedance cliange is difficult to spot in the time
Furthermore, if the resulting integrands are all
domain, whereas the defect can be observed clearly
normalized prior to plotting, the energy in the
in tile wavelet transformation. The author believes
mother sine wave is irrelevant. It is only nec-
that in practice a 20% impedance change would be
essary to ensure that the sine wave is appro-
clearly discernible in the time domain if a toe re- priately scaled during contraction or dilation
sponse as large as that indicated were obtained. It is
so that energy remains constant under all
assumed that the poor reproduction of the paper does transformations.
not allow their contention that the change is clearly 2. The half-cycle sine wave does provide com-
evident in the wavelet plot to be confirmed. pact support and fast decay, thus satisfying
At the time of these papers, Addison and his co- the requirement of time localization.
workers appeared to favour use of the Daubechies 3. As it has only positive components, the half-
wavelet, possibly because its shape provides high cycle sine wave does not meet the criterion of
resolution. The resulting wavelet transformation has a zero integral in the interval -us to +us. The
a somewhat chaotic appearance, which is somewhat author does not believe this to be critical for
disturbing to the uninitiated. Their data is typically the purpose of analysis of the low-strain sig-
presented as one-dimensional transforms plotted
nals.
against time, although time-frequency fields are also As suggested before, the advantage of applying
presented. It appears that at that time, a preference
the dilational and translation techniques of wavelet
for coiitinuous or discrete wavelet transformations analysis to a sine wave form rather than a true
was not clear. The discrete wavelet approach has
wavelet, is in providing an optimum opportunity for
significant computational advantages and does not correlation with the measured signal. It is believed
senerate redundant data. By comparison, the con-
that the fact that all the criteria for wavelets are not
tinuous form does not impose restrictions on dilation satisfied is not relevant for a relative assessment of
incizments, but generates significant data redun- time-frequency localization.
dancy, and hence takes longer to perform. If, however, the purpose of the analysis is to al-
The team at Napier University gives every indi- low deconstruction, modification and reconstruction
cation of continuing with their interesting research in of the signal, as wavelet analysis allows, the use of
this field. the sine-wave form could not be assumed to satisfy
these requirements.
4.2 Current Investigations
The author has also experimented with applica- 5 RESULTS
tion of wavelet analysis, exclusively to actual data
collected from field measurements.
In the following discussions, the term wavelet analy-
The similarity between the wavelet technique and sis is used advisedly, taking into account the differ-
cross-correlation technique, in terms of passing a de- ences with classical wavelet analysis already re-
fined wave across the signal to be analyzed are evi-
ferred to. The outcomes of this study, W P W
deni. In wavelet analysis, however, the waveform is
presented in one of two ways.
dilated or contracted and an appropriate scaling fac-
tor applied to ensure constant energy.
197
5.1 Time-doinciin presentation 5.2 Time-fi-eqiiencydomciin preseimlion
As the result of time-stepping using a spectrum of
As the result of time-stepping a wavelet of a single wavelet frequencies varying from approxiinately 300
frequency through the original signal. A modified Hz to 15 kHz. The wavelet integrands are normal-
response curve, being the normalized integrand val- ized to a peak maximum value of 1 for PVJ(h,r). The
ues of the signal and the wavelet results. The results array of results are then presented as a contour dia-
are presented in the time domain, and have a similar gram of the wavelet integrands with frequency as the
form to the original data. This is shown in Figure 7, vertical axis and time as the horizontal axis. Exam-
where a set of data, and the wavelet-modified data ples of this presentation, also compared with the
are compared. It can be seen that the original data original data in the time domain are shown in Fig-
has been smoothened by the wavelet transforrna- ures 8, 9 and 10.
tion.. because the frequencj of the chosen wavelet
was lower than the high frequency perturbations in 5.2.1 Pile with dcimnge
the data. It is, if used in this manner, a useful filter- Figure 8 shows the time-frequency response plot
ing technique. As always, filtering of data must be for a pile with a significant defect close to tlie pile
undertaken with care to eiisure that significant fea- mid-length. The horizontal axis is time, at the same
tures are not removed. This algorithm has been in- scale as the time-domain plot also shown. The ver-
corporated in coinniercially available software tical axis is frequency in Hz, ranging from approxi-
(Likins and Rausclie, 2000). 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
Figtire 7. Comparison of original signal arid modified signal wavelength of approximately 0.7m. Another ridge
after wavelet anaiysis using a single fi-equency can be seen at approximately 3.00ms. This corre-
Figure S. Comparison of time-domain data and \vavelet analysis in tlie time-frequency domain for pile with inid-
length defect
198
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 Another feature can be seen emanating from the
case, the peak frequency occurs at approximately base corresponding to the high frequency compo-
3600 Hz. This downward frequency shift is charac- nents at about 1.9ms. This is the damage indicator.
teristic of this type of plot. As noted previously, the As stated by Ellway (1987), structural defects in the
higli frequency components of the incident wave are pile preferentially reflect high frequency compo-
iiiore quicltly absorbed, and the reflected wave is nents of the incident wave. Low frequency compo-
expected to be more dominated by low frequency nents pass through with little reflection. The dam-
components. age indication cannot be seen in the low-frequency
The 2L/c time is determined by measuring from range of the plot.
ridge to ridge, giving a wave return time of 2.1 1ms. It can be seen that pile damage indications will be
199
rcstricted to the high frequency range, whereas soil Wavelet analysis is a mathematical technique
resi btance, which generates damping responses will which allows the data to be viewed in the time-
dominate the low frequency range. frequency domain. Little research or investigation
of the application of wavelets to pile integrity testing
5.2.2 Pile slruck with difji;r.ent haiiuners has been undertaken to date. However, it appears
Figures 9 and 10 show both time domain and that this approach may provide a more complete w y!
time-frequency domain graphs for the same pile of viewing the data. Significant further work 1s
struck lvitli a 0.9kg (2 pound) and 5.5 kg (12 pound) needed to refine our understanding of the application
han: mer, respectively. A comparison of the two of wavelets to low strain test data, and to its inter-
tests is interesting. The time domain results are first pretation.
coin pare d. The author has undertaken “hybrid” wavelet
The pile was initially tested with the lighter analysis using some of the principles of wavelet ap-
liammer. Figure 9 indicates the possibility of some plication using a half-cycle sine wave function. The
impedance reduction above the pile toe. The reflec- time-frequency responses presented provide deeper
tions are not strong, and in the time-domain there is insight into the pile-head response, and demonstrate
overlap between this “damage” response and the some physical principles previously established by
later response presumed to be the pile toe. This others. At this stage, it is a qualitatively useful
makes it more difficult to make a definitive determi- method of presenting low-strain data. It has poten-
nation of the pile wavespeed. tial to provide useful quantitative analysis with fur-
In order to confirm (or otherwise) tlie damage, the ther research.
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 REFERENCES
toe response, and fk-thermore, the toe response has a
very long rise time. Again, determination of the pile Addison, P.S., Sibbald, A. and Watson, J.N. (1997). Wavelet
wavespeed is unclear, and possibly in conflict with analysis : a matliematical microscope with civil engineering
applications. Insight Vol 39 No. 7, July, 1997 : 493 - 497.
the data from the smaller hammer, if tlie rise-to-rise Addison, P.S. and Watson, J.N. (1997). Wavelet analysis for
method is adopted. low strain integrity testing of foundation piles. 5“’ Intl.
The time-frequency responses are now examined. Conf. On Inspection, Appraisal, Repairs, Maintenance of
First. it is clear that the input frequency of the signal Buildings and Structures. 15-16 Mat, 1997, Singapore.
generated by the larger hammer (about 300 Hz) is Davis, A.G. and D i m , C.S. (1974). From theory to field expe-
rience with the non-destructive vibration testing of piles.
significantly lower than that generated by the Proc. Institution of Civil Engineers, Vol. 57, Part 2, De-
smaller hammer (about 800 Hz). The lower fre- cember : 57 1-593
quency content of the returning wave is also evident, Ellway, I<. (1987). Practical guidance on the use of integrity
particularly for tlie smaller hammer which is subject tests for the quality control of cast-in-situ piles. Ground
to greater degradation. Engineering, Vol. 20, No. 7, October : 5-13,
As noted before, pile defects are more likely to be I<umar, P. and Foufoula-Georgiou, E. (I 994). Wavelet analy-
sis in Geophysics : an introduction. Wavelets in Geophys-
detected with high frequency waves. The ridge at ics, Foufoula-Georgiou and Kunar (eds.) Academic Press
about 7.3ms in the high frequency range of Figure 9 Inc. ISBN 0-12-262850-0 : 1-43
for the smaller hammer is the damage indication. Lee, D.T.L. and Yamamoto, A. (1994). Wavelet analysis :
No such ridge can be seen in Figure 10, because of theory and applications. Hewlett-Packard Journal, Decem-
the limited high frequency content of this signal. ber, 1994 : 44 - 54.
Likins, G.E.L. and Rausche, F. (2000). Recent advances and
The pile toe is seen clearly and definitively in proper use of PDI low strain pile integrity testing. 6“’ Intl.
both plots, and the 2L/c time can be firmly estab- Conf. on the application of stress-wave theory to piles., Sao
lished as 8.1 ms for both tests, thus overcoming the Paulo, Brazil, Sept., 2000 (not yet published).
confusion from the time domain plots. 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
6 CONCLUSION wavelet transform for pile integrity testing. Proc. 12‘” En-
gineering Mechanics Conf., La Jolla, California, May 17-
Lo\\ -strain pile integrity testing data is normally 20, 199s. Mttrakami, H. and Luco, J.E. (eds). 0 ASCE. :
plotted in either the time-domain or the frequency 546 - 549.
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
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 1503
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 .
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.
202
way the company could use 12m bars easily avail- 3 GRAIN UNLOADING WHARF EXTENSION -
able on the market by just cutting them in half with- COLONIA
out the necessity of connections. Where the bars
ended, they were closed off by bending the bars, During the piling phase of a grain unloading wharf
forming a “nest” that, because of the quantity of in the Nueva Palmira port, a loaded, 20,000 ton
bars, left very little space for a proper concrete pas- barge accidentally hit a group of piles constructed
sage. It is important to keep in mind that the con- only 10 days before, reacting elastically. SIT studies
crete used for these types of piles is a dry concrete were done to see if the piles involved had suffered
that contains thick aggregates. As mentioned above, structural damage.
the zone where the narrow section appeared, In agreement with the information given by the con-
corresponded to the passage fiom a higher lateral tractor, the piles constructed were drilled piles made
ground kiction to another with less resistance. For of reinforced concrete, all about 27m in length, 0.8m
this reason, if during the concreting the case was not in diameter, and fixed in about 4m of a high resis-
carefully raised, and was done abruptly, the pile can tance ground layer, belonging to the Fray Bentos
be cut off with the invasion of saturated ground. formation. The piles have steel casings in their up-
The high number of defective piles found at this per section of 20m.
depth can then be explained by the difficulties re- The signals obtained from the integrity testing pre-
sulting iiom the passage of the thick aggregate con- sented a clear hammer impact, and a good reflection
crete through the folded bars, and also due to the of the pile toes. The lengths obtained corresponded
transition from resistant soil to a lower lateral iiic- reasonably with the ones given by the construction
tion during the case raising. company. All signals indicated a lateral fiiction in-
The foundation was reinforced by building several crease below the depth of 20m, that corresponded
new piles, following the proposal from the planner, generally with the end of the steel casing and the di-
all made with continuous reinforcing bars up to the rect contact between the ground and the concrete.
pile toe. See Figure 5.
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.
0 To evaluate the possible damages on pile number 11,
its signal was compared with the average signal
0 0 0 taken from the adjacent and unaffected piles (9,10,
0 0 0 0 O0
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
0 0 0 damages.
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 6. The signal of affected pile number 11, and
the comparison with the average signal
Figure 5. Arrangement of the foundation piles of
tank 720.
203
The tests allowed the integrity control of the affected shortly. Although there are still no official standards
piles and the comparison of their signals with the for the tests, many state organizations like the
signals of the nearest piles. In this way the theory of Transportation and Public Works Ministry (MTOP)
structural damage to the piles could be rejected. and the Petroleum National Company (ANCAP), re-
quire these controls in the sites under their admini-
stration.
4 MACHINE FOUNDATION IN THE TECHNO-
LOGICAL LABORATORY OF URUGUAY
(LATU)
6 FUTURE PROSPECTS
204
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3
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 even all piles after the grout is hardened. There are
some limitations for long or jointed piles.
The quality of augercast piles often depends on the Bearing capacity can be evaluated on any pile on
contractor’s skill. The most critical is the control of site using High Strain Dynamic Pile Testing (Likins
auger withdrawal during grout placement (Roberts 2000a). Test Piles can be selected after installation
1998). However, for a visual inspector, it is difficult based on their site location (to assess soil variations),
to accurately assess grout volume and auger depth on specific pile installation records, or at random.
simultaneously. Automated measurement systems The capability of installation monitoring and
can accurately monitor volume pumped as a fimction subsequent low and high strain dynamic testing
of depth (Likins 1998). This information can guide allows the engineer to completely investigate the
the operator during installation to produce a shaft foundation.
exceeding any minimum requirements.
Low Strain Pile Integrity Testing of selected piles
verifies the installation monitoring effectiveness. 2 CURRENT INSTALLATION PROBLEMS
This requires striking the pile top with a small MONITORING AUGERED PILES
hammer and measuring both input and reflections
from pile non-uniformities or the pile end (Rausche Visual inspection counts pump strokes as a fhction
1992). Pile Integrity Testing is applied to any pile or of estimated auger depth to estimate volume versus
205
depth. Due to a large volume of information, there the grout line. During augering, the auger torque is
are many possibilities for deficiencies or errors in measured.
recording all the necessary information. Further, the In practice, after the operator inputs the pile name,
volume per pump stroke can be easily vary by as the PIR-A Control Unit handles virtually everything
much as 20% (Likins et al. 1998). Further, pump else. During drilling, the operator observes the
strokes are usually manually counted per five foot auger's current depth and torque. Higher torque
interval; this interval has marginal precision and (just below the crane stall torque) makes drilling
even the visual determination of the depth may have more efficient and reduces the spoils brought to the
one foot errors. Alternative methods of more surface. Geotechnically, the engineer can assess the
accurate inspection are desirable. torque at auger rehsal to distinguish strong soils or
low torque.
206
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.
207
The Pile Integrity Tester's records of all the shafts record along the shaft is relatively flat and thus
in Pier B18 indicated a characteristic increase of indicates a relatively uniform shaft and no defects.
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 Figure 6. PIT data for pile 349.
cross sectional area or modulus at approximately
14.6 m (48 R). The PIT toe reflection in the rock The PIR-A result is given in Figure 7. The
socket resulted in a negative velocity (compression) volumes printed are for 610 mm (2 R) depth
at 17.7 m (58 R). This is typical for shaft sections increments. This depth increment was selected to
with either a high soil resistance or an increased pile provide sufficient precision taking into consideration
impedance. 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
Figure 5. PIT data (case 1). 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
The PIR-A result for the same Pile 18A showed segments were above the volume required. The low
that between 15.2 and 14.6 m (50 and 48 R) below volumes pumped near the pile top (bottom two lines
the top that the grout volume per 0.6 m (2 A) in Figure 7) are acceptable since the grout return was
increment decreased from an oversized average of already observed. This early grout return is due to
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 depth
volume and adjacent increments were higher than 44 7-88 22 0
required; some redistribution of the fluid grout 42 4.a 7 0
s10 63 8 0
between neighboring increments is likely. 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 used to judge pile consistency and acceptance. With
withdraw al. more accurate information available, augercast piles
are more readily specified and accepted by designing
engineers. If automated PIR-A installation records
5 HIGH STRAIN DYNAMIC TESTING of grout versus depth indicate a good shaft, then this
may reduce the need for hrther PIT tests. Thus, PIT
Monitoring during installation does not assure testing can then be restricted to shafts with
sufficient pile bearing capacity. Thus, either questionable PIR-A records or shafts with problems
additional static testing or high strain dynamic pile observed after installation, during installation of
testing (Hussein et al. 1996, Likins et al. 2000a) are subsequent piles or during excavation. In addition, a
required. High strain dynamic pile testing with the small percentage of randomly selected production
Pile Driving Analyzer@(PDA) has also been applied piles may also be tested by PIT for quality control.
since 1974 to drilled shafts and augercast piles with Capacity of augercast test piles can be determined
increasing frequency using drop weights to generate by a static load test. However, dynamic pile tests
the loading. In some countries extensive dynamic followed by CAPWAP analysis is a well proven
pile testing is now routinely performed on drilled alternative for augercast piles when minimum
shafts and augercast piles. installation time is critical or if multiple tests are
The pile top may be temporarily extended above desirable to evaluate site variability. Dynamic tests
the ground surface t o encase protruding on augercast piles usually require some pile
reinforcement or to facilitate attaching the strain and preparation and a drop weight to apply the impact.
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 REFERENCES
impact over the entire top surface. The drilled shaft
or augercast pile is then impacted by a simple drop Hussein, M., Likins, G. & Rausche, F. 1996. Selection of a
weight guided by a short set of leads. A low drop hanmer for high-strain dynamic testing of cast-in-place
height is first applied to assess signal quality and shafts. Proceedings,j$h international conference on the
application of stress-wave theory to piles. Orlando, FL,
alignment of the weight with the shaft. After each USA.
impact, the net permanent displacement or "set per Likins, G., Rausche, F. & Goble, G. 2000a. High strain
blow" is carehlly measured. The drop height is then dynamic testing, equipment and practice. Proceedings,
increased until either the set per blow exceeds 2 or 3 sixth international confirence on the application of
mm (to activate the h l l capacity), or until the stress-wave theory to piles. Sao Paulo, Brazil.
indicated capacity exceeds the required ultimate Likins, G., Rausche, F., Morgano, C.M., & Piscsalko, G.,
2000b. Advances in PDI low strain integrity test methods
capacity, The PDA monitors stresses directly and and equipment. Proceedings, sixth international
additional cushion material can be inserted to avoid conference on the ayplication of stress-wave theoiy to
overstressing. The measured strain and velocity data piles. Sao Paulo, Brazil.
can be analyzed on site by the signal matching Likins, G., Piscsalko, G. & Cole, C. 1998. Pile installation
software CAPWAP@ to independently check if the recorder tests for ACIPJCFA piles. Proceedings, 7ke
activated capacity exceeds the desired test load. 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
6 CONCLUSIONS testing and analysis. Proceedings. &iirth international
conference on the application ?f stress-wave theory to
L o w strain Pile Integrity Testing (PIT) can detect piles. The Hague, the Netherlands.
major defects in the pile shaft at low cost and with Roberts, T. 199X. Quality control for augered cast-in-place
little effort. However, particularly for very long piles in Texas and Louisiana. Proceecling.y the deep
piles, PIT results may be difficult to interpret and fixinclations institute '.c nrtgeretl cusf-in place piles
should not be the only means to verify the quality of specialty serninar: aiigered cast-in-place piles. Houston,
TX, USA.
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
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.
Structural engineers are often faced with the question The pulse echo method (Rausche et al 1992) uses a
of pile integrity. Difficulties during installation, hand held hammer to impact the pile top and generate
excavation procedures, slope failures or lateral a compr-essive stress wave in the pile Figure 1
movements due to accidental impact can create doubt shows a Pile Integrity Tester (PIT) for pulse echo
about the integrity of any pile. In some cases, doubt testing. The small battery powered main unit is
also arises due to lack of information about existing highly portable from pile to pile and extrcmel!
piles when a new use or increased design loads are rugged. It uses a touch screen for input rather than ii
required. In some cases lack of inspection on current computer keyboard. Experience has proven this to
installations leads the structural or geotechnical be a highly reliable method and extremely user
engineer to question the foundation adequacy. friendly. The screen displays intuitive menus to guide
the user and shows the graphical signals for field
If access tubes have not been installed in the interpretation. Stress wave inputs and reflection?;
questioned piles (required preplanning and extra (from non-uniformities or the pile toe) are measured
cost) to allow for Cross Hole Sonic (CSL) testing, as a hnction of time by an accelerometer attached to
then the costs for evaluation can quickly become the pile top. The acceleration is integrated to
large. Static loading for every pile is obviously velocity by PIT, and then interpreted by the test
prohibitively expensive. Coring every pile is likewise engineer.
not feasible, and even if it were it would only provide The pile top surface is prepared by removing the
a small sample that could easily miss a local defect at upper concrete if it has been contaminated with soil,
some depth. Therefore, the only practical alternative bentonite slurry or other foreign materials during
is often PIT. Since PIT is simple and low cost. it i s construction, and finding or making a smooth
conceivable to inspect every pile on a site if required location when a driven pile top may have been rough
Often a statistical sample is made and if no problems cut. An accelerometer is then attached to the smooth
are found, then testing may end. If defective piles ar-e pile top surface as shown in Figure 1 with a thin layer
located, then additional tests of neighboring piles ar-e of a soft paste like vaseline, petro wax. etc.
then justified. The selection of which piles to test can Accelerations from several hammer blows are
be based on various observations. or made normalized, integrated, averaged and displayed as
considering redundancy of the pile, or at random velocities. Further data processing includes
21 1
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
212
not important. Therefore, a high pass filter (I-IPF) is vibration of reinforcement bars protruding fi-or11the
applied which removes low frequency events usually pile top cracked pile top concrete or other- surtice
caused by soil resistance. In the case of the extreme effects. Since the “noise” frequency is much hishei-
example in Figure 4 (most data is generally near zero than the input frequency, if we select a wa\,elet
even after magnification), the data is subjected to mother function with similar frequency content to the
high pass filtering which removes components with input pulse, unwanted frequency content will be
less than a defined filter frequency. In general, care suppressed as shown by the wavelet filtered Figur-e 7
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
213
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
214
magnification factor is only about 2.5 while the factor frequency domain, then it is even more apparent
4 is only achieved at a depth of 47 fi. Thus the use when presented in the standard time record The
of similar magnification factors for different length major advantage for the frequency aiialj,sis i n our
piles does not result in “consistent” niagnification opinion is to obtain a quantification of shaft qualit),
functions. by calculating the relative pile stiffriess (piles v, it11
lower stiffness are more likely to be defective.)
In Figure 12 a magnification of 15 has been applied One other tool associated with PIT is to obtain an
t o the longer pile resulting in a factor of 4 at about impedance profile. This makes a shape visualization
31 fi so this then is a consistent magnification as a as in Figure 13. However, experience has shown that
function of depth. The toe response is now of similar this is not often required and piles can be classified
without this extra effort for most piles.
215
Category A piles are good piles within the accuracy concern where there are localized defects of small
limits of the method. For example, a it is generally extent.
agreed that a defect that affects less than 20’10 01‘ the
pile cross section cannot be detected with certaint!, The method works particularly well on solid pile
(actually this a strength of the method since it does cross sections, such as concrete or timber piles. It
not create questions where only minor problems has only limited applicability to steel piles. Of
exist.) In this category pile integrity is satisfactoiy course, concrete filled pipes can be tested.
and unless there is some othei- reason to suspect the
shafts, they are generally acceptable Of course, this In cases where there are mechanical joints or full
assumes that also the pile length indication i h section cracks, the wave cannot usually cross the
satisfactory. resulting gap. Therefore only the portion above the
gap is really tested.
Category B piles are somehow defective and some
contingency plan must be used Extra tests could be Defects which are very short compared with the
made. If the defect is near the ground sur-t’ace. input pulse width and do not cover the entire section
excavation to expose and repair the dettctile pile can be misleading due to superposition of the waves
portion is generally possible. The shaft could a l s , ~be generated by both initial reduction and following
cored and the defective are pressure grouted, a increase. A similar problem exists for gradual
subsequent PIT could confirm the effectiveness of‘ changes which create relatively slow changes in the
any repaii- measure The pile could be abandoned and record. Normal processing of PIT data then removes
simply replaced The course of action inav depend the reflections from gradual changes since they are
on the cost of a new pile vei-sus the cost to repaii- similar in frequency content as the soil resistance
effects. Therefore if a pile gradually increases and
Category C piles may also be assigned a reduccd then quickly returns to the nominal cross section as it
capacity. Also, other pile tests may be considered. or enters a new strong soil layer, the sudden reduction
excavation if the defect is located near the y-ound inay produce a record that indicates a defect. I t is,
surface. If the pile is a friction pile and the det’ect is therefore, important to compare any PIT result with
located far down the pile. the upper s o i l resistailcc the general soil profile to assess the likelihood of
above the defect may render- the h i - c e in thc p i l e 21; these situations.
the defect as acceptable structur-ally and the ciclilii
inay in some cases be not serious There is a general rule of thumb that the test
effectiveness is limited to piles with lengths less than
Category D piles may have poor data dcic to pooi. 30 times their diameter (L/D ratio) This is really
concrete quality at the pile top. Tr-imming the pile to applicable to the length embedded in soil since there
a lowei- elevation where good quality concrete IS is little signal degradation for shafts acting as
assured and retesting the pile is an alternatiie Ollici columns in water until below the mudline. In some
reasons for inconclusive results such as a long oi cases it might not be possible to “see” even this far
irregular pile are discussed in the next section I t is down the pile. However, with recent improvements
reasonable to accept a certain percentage of piles in electronics including lower noise and higher
with inconclusive records at a site where soil and pile precision 16 bit sampling (which gives 8 times more
conditions produce complex recoi-ds. resolution to older 12 bit samplers), PIT now
regularly tests piles with L/D ratios far in excess of
30.
LlMITATIONS
The pile length effectively investigated by PIT
PIT is a very useful test for many sitiiatioiis I n ~ i i i e depends not only on soil damping, but also on the
cases it is the only practical alternative to doins non-uniformity of the pile, because any change in pile
nothing which may not be acceptable i z hci-e quality or shape produces reflections. The first non-
questions arise during constr-uction I t is certaInl\ ;i uniformity detected will be more reliably analyzed
simple and relatively inexpensive test and thus i s than additional sources of reflection farther down the
widely used. It is intended to locate major defects shaft. Highly non-uniform piles produce complex
and to that put-pose it generally is adeqiratc l r i h ii records which are dif’ficult to analvze I n such c a w h
strength of the method that it does not generate the records collected under comparable
circumstances can be compared Records thitt dill;,i
216
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.
218
Application of Stress-Wave Theory to Piles, Niyama 13Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
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
219
2.3.1 Heterogenity between probes within a con- 3 RESULTS
crete strength
Given the curves obtained for every ages within a
Includes variations introduced at the casting mo- concrete or mortar strength vs wave propagation ve-
ment by the fluid concrete mix vibrating of the locity two intervals can be clearly differenced. The
probes. To measure this error the coefficient of first includes mixtures of ages between 1 to 7 days
variation of the velocities obtained using PITPC were an strong dependancy between age and veloc-
software. ity.
All four probes data velocity vs mix age at early
ages (1 to 7 days) were averaged and regression
REF. RANGE AVERAGE STAT DISTRIBU- analysis were performed to modeled the age-velocity
TION OCURRENCE PROBABILITY (note Nol) relationship during early.
Whole sample 1.0% - 3.0% 2 CHI SQUARE The following chart resumes the regression
65% analysis performed and the results obtained
at ages from 1 to 7 days.
4 CONCLUSIONS
2.3-2 Heterogeneity during PIT curves acquisition The possible errors within the testing program
correspond to normal conditions during the construc-
tion or PIT performance and they were reduced to
Heterogeneity introduced by variations of the PIT acceptable ranges.
technician criteria, accelerometer disposition etc. The wave velocity is directly proportional to the
Such error was statically analyzed in terms of meas- concrete‘s mix age during the first 7 days of curing
ured pile length based on the results obtained with process.
PITPLOT software. The variation of length error’s According to R2 obtained the regression analysis
magnitude obtained is as follows: performed characterizes accurately the stress wave
velocity at early ages; for ages older than 7 days the
Table.3 stress wave velocity reduces its dependancy of age
and becomes a property related to the compressive
strength of the concrete or mortar.
220
22 1
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 1503
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
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.
Figure 3 Presentations of CSL test results in the form of Figure 4. Typical SSL testing set-up showing transmitter and
threshold time and energy with depth. receiver at different depths.
225
to high coring costs, a single hole is advanced (of- zone is detected in this stage and the tomographv
ten down the middle) to the bottom of the shaft or option is enabled, the probes are lowered and
slightly below the depth where a defect is antici- raised relative to each other around the suspect
pated. It may also be desirable to perform SSL zone, to fkrther investigate and delineate the area.
during CSL testing to isolate the location of a de- The signals can be examined and adjusted by
fect at a certain depth (i.e. distinguishing whether manually picking the points or using preset algo-
the defect identified using CSL is adjacent to the rithms to automatically determine the first arrival
tube or in between the tubes). Recently SSL has time (FAT) as shown in Figure 7.
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. I Gene~crl
The PISA (Pile Integrity Sonic Analyzer) is a
modular system allowing for adoption, upgrade
and incorporation of additional integrity testing Figure 5 . PISA system components.
technologies The current integrity testing options
available in the PISA include cross-hole sonic log-
ging (CSL) and single-hole sonic logging (SSL) 4 CASE HISTORY
using CHUM (Cross-Hole Ultra Sonic Module)
and sonic echo (a k a small strain propagation) 4 I Rnckgrormd
using PET (Pile Echo Tester) module Additional
modules are currently under development Four drilled shafts were constructed at the Auburn
In addition to its modularity, two advantages of University in Auburn, Alabama as part of a re-
the PISA integrity testing system over other sys- search study The shafts are 914mm (36in) d]-
tems include its software and portability The PISA ameter and approximately 11 Om (36fi) long Each
is the only Windows 95/98 based system and is shaft was equipped with 4 access tubes and various
also compatible with Word 2000 The software is defects were installed during construction The de-
updated periodically to incorporate new develop- fects were constructed as soil inclusions formed by
ments and algorithms that make data collection, sand bags made of a tough material and tied to the
interpretation, and report preparation easier and ef- rebar cage The cross sectional area of the defect
ficient The PISA is lightweight (only 42 3 N (9- was based on the measured perimeter of the bags
1/2 Ib)) and self powered, hence can be easily car- following their installation The concrete was
ried around from shaft to shaft or site to site This poured into a dry hole which was cased the full
feature is also beneficial for air travel The system length of the shaft
can be also used as a standard laptop, saving the Over a year after construction, Geosciences Test-
cost and space required for an additional personal ing and Research, Inc (GTR) personnel tested the
computer (PC) when using a dedicated CSL testing shafts using the PISA system and submitted the re-
system sults to Prof Dan Brown of Auburn University
Figure 5 presents a photograph of the PISA The actual "manufactured" defects were then re-
system, including computer and sensors As a vealed and comparisons were held between the
scale, the width of the computer screen is 23cm (9 predicted and actual defects
inches) Figure 6 presents the layout of the pile
screen, where one can enter the pile information 4.2 Detected w. Planned Defects
and select the tube orientatiodlocations Selection Figures 8 and 9 summarize the defects as detected
of the desired tube combinations is accomplished by the PISA testing (on the right hand side) versus
by drawing a line between any two tubes Real- those planned/manufactured during construction
time graphical presentation of the concrete integ- (on the left hand side).
rity is provided during data collection If a suspect
226
Figure 6. Layout of the pile 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 trated in two zones which may be the case
were for example: "Shaft 4 - soil inclusion at 11 to following the casting The lower defect was iden-
13ft below the ground surface on the north side tified in the right location, however was marked as
(towards shaft 9), covers 20% of the cross- approximately 25 6% of the cross-sectional area
section." As such, the defects outlined in Figures 8 versus the manufactured defect planned as 10% of
and 9 are a reasonable approximation of the de- the cross-sectional area.
scriptions. Shaft 9: Two zones of defects were identified in
Shaft 4: Two zones of defects were identified in shaft 9 (Figure 9). The upper defect was identified
shaft 4 (Figure 8). The center of the upper defect as the right size but at a location approximately
was suggested to be about 0.6m (2ft) above the 0.6m (2ft) above the center of the actual location.
center of the actual defect with overlapping mar- The lower defect was identified at the right loca-
gins. The area of the defects was correctly identi- tion and with the correct size
fied though the test suggested that it is concen- Shaft 2: A soft bottom in part of the cross-
228
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 5 SUMMARY AND CONCLUSIONS
intentional defect and may have resulted from the
regular construction process Four relatively small size sand inclusions were in-
Shaft 7 A weaker zone was identified between stalled in two shafts out of four constructed shafts.
1 5 to 3 7m (5 to 12fi) along one segment of the All the defects were identified in the tests con-
shaft (tube 2) No intended defect was installed in ducted over a year after construction. Three out of
this shaft The shaft, however, was loaded to fail- four defects were identified in their approximately
ure in bendins and extensive tension cracks were correct size, the fourth defect was assumed to be
expected to be developed on the south side around about 3 times the size of the actual defect. Two of
3 7m (12ft) This information, like all other infor- the defects were also identified within approxi-
mation, was provided after the tests results were mately O.Gm (2ft) of their actual locations.
submitted Overall, the test results of Class 'A' prediction
provided accurate and reliable evaluation. The to-
229
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
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
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
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.
233
difference in damping period, that the over-damped and then plotted over the experimental signal to
motion endures more than the critically damped as check for the matching between the derived and
shown in Figure (1). experimental time-displacement signal.
As the value of damping constant, c is being verified,
and with the static pile capacity, R, in hands, the
2.1 Under-damped Motion:
viscous damping- Smith damping, coefficient (J) can
For such cases two methods can be used to calculate be computed as:
the damping ratio, the first utilizes Hilbert transform
techniques either derived by Thrane (1 984), 8
Mohammad and Tomlinson (1988), or by Husein
Malkawi and Mohammad (1996), while the second is
2.2 Csitically Damped Motion
the logarithmic decremental approach. The analytical
procedure for the signals in this research will be For the critical damping condition the solution
based on the logarithmic decrement due to its algorithm is iterative; given the initial conditions (yo,
simplicity that it requires little calculations. The v,), the time displacement hnction becomes:
logarithmic decrement (6) is defined as:
Y ) 9
6= In( 1 4
Y2 The natural frequency (a,) is adjusted in eq.(9)
hence, the damping ratio is calculated as. until it fits best with the experimental signal, then for
the selected fitting equation:
6
5
where;yl and y2 are the amplitudes of the first and 3 TEST MODEL AND MATERIAL
second peaks of the oscillation, respectively.
Measuring the damping period- from the recorded 3.1 Model Pile
signal, between the first two successive peaks (zd), The test pile is a steel, close ended, smooth pipe pile
then:
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
now the damping constant (C) is computed as:
pile cap, the cap has a semi-spherical notch on its
7 upper face where a steel ball (ball bearing) is placed
c=cc = 217143
Cl' 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
Figure 1 Comparison between critically damped and internal friction varied in the range (26 - 30")
over-damped motions. depending on the level of the effective stress
Whereas, in the densest, the dry density of the sand
The values computed above along with the initial is (1 72 g/cc) and the corresponding friction angle
conditions (yo, v,) are now used to obtain the laid in the range (35-40") Those friction angles
parameters of the under-damped equation of motion,
234
were obtained from direct shear test conducted on to a maximum of (100-cm) because the resulting
loose and dense sand specimens at different dust may seriously affect the instrumentation of the
representative normal stress levels laboratory. Consequently, for the case of (70 %)
relative density, gentle vibrations are necessary to
achieve such high density.
Medium 50 1.63 80
* Vibration employed.
Figure 2 Grain size distribution curve
4 TESTPROGRAM
3.3 Sand Container A general view for the test elements, instrumentation
The sand container is a square (100 x 100 cm2) and setup is shown in Figure (3). The first step in all
sectioned, (125 cm) high steel box, with (5mm) thick pile’s tests is the preparation of sand bed at the
wall, and an empty weight of 2.73 kN (278.37 kg). desired test density (or relative density). The sand is
Those dimensions were chosen in such away to filled in the silo which is then elevated to an
prevent the interaction of pile driving stresses with elevation corresponding to the desired test relative
the container’s either wall. This container was density obtained from Table ( l ) , and as a double
manufactured in the engineering workshops of check, tank full weight is measured by the dartec
Jordan University of Science and Technology. To machine, and knowing the depth of sand in the
obtain precise measurements of the in tank sand container the density is insured.
depths the tank was marked at corners and walls
mid-lengths with a metric graduation to the nearest
(0.5 cm)
235
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)
Figure 4 Pile driving input, square load-time wave. 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
To measure pile’s displacement with time a Linear depths and sand initial density are shown in Table
Variable Differential Transducer (LVDT) was placed (2)
under the pile cap thus allowed to simulate pile’s Damping constants and damping coefficients are
actual movement, the (LVDT) is connected to an then studied with different test parameters, pile
analogue storage oscilloscope from which the final installation depth, sand’s initial density, weight of
Time-Displacement signal was obtained. The soil column at the level of pile’s toe, and ultimate
(LVDT) is one primary, two secondary coils and a static soil resistance
magnet, the motion of the magnet induces an The under-damped motion is clear since it is
electrical current, this current is transformed into characterized by the oscillation (wave like motion),
displacement by the (LVDT) control panel, and the while critically and over damped motions can only
signal is shown on the screen of the oscilloscope differentiated by their damping period, as mentioned
The pile is forced to move for a definite forced earlier Below are detailed examples for signal
displacement depending on the strike of the machine matching processes along with parametric evaluation
ram, then continues to move freely for a little for two damping cases, under-damped, and critically
displacement upon which analytic works are later damped or near Each of which is analyzed
conducted For the case in hands, forced separately
displacement can be easily measured as that
corresponding to constant Time-Displacement (t-d) Table 2 Static pile bearing capacity in (kN) as
slope, i.e , for the initial linear part of the (t-d) signal obtained from CRP tests.
where the slope is a constant value Even when Depth 30 45 60
running the machine in the dynamic mode, the ram D,%
will move down as fast as specified by the frequency 35 1.5 2.4 3.25
input until it reaches a maximum forced displacement so 2.2 4.5 56
at a high constant jacking rate, then rebounds up to 70 4.3 6.1 8.7
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- 5 1 I I I i ~ e r . - ~ ~ ~ ~ i y e u (Refer
~ ( j t i otoi iFigure 5 )
Y adoptei- (Pen Recorder) was employed instead of The first step is to plot the initial tangent for the
the LVDT to indicate pile loads and displacements curve starting from (0 0,O 0), this tangent match with
plotted on a millimeteric paper. Pile capacity is taken the initial linear portion that corresponds to constant
as that corresponding to a displacement of 10% of pile velocity, thus indicating the region of forced
pile’s diameter motion The last point on the linear initial part of the
As for the expected experimental errors, they may curve, is the first in the free movement region
be comparatively insignificant The machine is The axes is now transformed to the point (t‘=O 0,
capable of measuring within 0 5 N (0 05 kg), sand x‘=O 0) in the free motion part This point
level in the container is precisely achieved by corresponds to (t=75, x=6 4) The initial linear part
applying gentle vibrations to the container’s walls, of curve is extracted form Figure ( S ) , the remaining
and adjusted via a levelmeter allowing the part represents pile’s free motion and is re-plotted on
achievement of the desired sand depth with a the new axes ordinates (t’, x‘) as shown in Figure (6)
tolerance of few millimeters, and the oscilloscope below The solid line represents the recorded
expected errors are probably the least. 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
5 DISCUSSION yl=O S inin, y2=0 22 mm, at z,1=36 ms
The logarithmic decrement, 6 = In yl/y2= 0 82 1
Measured time-displacement signals are to be
analyzed utilizing the “Logarithmic Decrement”,
Damping ratio, = 0 1296 <
Damped frequency, atl = 0 174 rad /ms
Natural fi-equency, a,, = 0 175 rad h i s
236
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.
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
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.
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
REFERENCES
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.
241
Pile driving system Pile type Pile type
_A_ R
I
Concrete fck
Steel fvk 500 500
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.
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
244
Figure 18. Local friction diagram along depth: pile B1
245
greater than the point resistance in all sequence of The so-called quake of the point (elastic displace-
growing applied energy. ment resulting from the elastic deformations of the
Finally in pile B1 for all blows of growing energy, soil under the base of the pile) is directly dependent
the point resistance predominates. on the amount of energy reaching the soil under the
pile point. It can not be considered a soil parameter.
4 CORRELATION BETWEEN STATIC AND The ultimate load resulting from the static load is
DYNAMIC PILE LOADING RESPONSE in good agreement with the ultimate load obtained in
the gradually increasing energy dynamic load test.
All the equivalent static load transfer diagrams ob- The load transfer diagrams obtained in dynamic
tained in this site, from dynamic loading test per- load test show a good resemblance to those obtained
formed with gradual increasing energy, show a gen- in instrumented static load tests. In the soft clay layer
eral trend that is in good resemblance to those the lateral skin friction grows up to a maximum and
obtained with instrumented piles subjected to static thereafter diminishes to a residual value when the pile
load test. The reliability on the application of a dy- is gradually subjected to increasing applied energy.
namic loading test procedure to obtain information
about the pile behavior under static loading has
grown with the increasing utilization of this type of 6 REFERENCES
test all around the world. Otherwise the correlation
Aoki, N. 1989. A new dynamic load test concept. In Japanese
between static and dynamic load test results in the Society of Soil Mechanics and Foundation Engineering
same pile can be made and, in doing so, it would be (eds): International conference on Soil nzechanics and
possible to improve our understanding of when dif- foundation engineering, 12. Drivability of piles, 2: 1-4. Rio
ferent soil conditions occur. Such correlation was de Janeiro, Brazil
done in this site and the results are presented in Aoki, N. 1997. The evaluation of the ultimate bearing capacity
of driven piles by using increasing energy dynamic load
figure 3 1 . 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.
Figure 3 1. Resistance versus displacements curves: pile B 1 Van der Veen, C. 1953. The bearing capacity of a pile. Inter-
national Conference on Soil Mechanics and Foundation
The dynamic test was performed in 1 day and the Engineering, 3. v. 2: 84-90. Switzerland.
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
248
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.
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
250
soil. Afterwards the steel profiles are driven by a The tests at the driven steel profiles and tubular
15 tons Menck MHF 10-15 hammer for a hrther five steel piles respectively were done at the end of the
meters into the subsoil to ensure sufficient bearing driving operation and also as a restrike test three
capacity. For the driving of the anchor piles and the weeks after installation. Up to now a total of 23 tests
cast-in-place concrete piles hydraulic hammers of ap- with CAPWAP evaluation were carried out. Figure 6
propriate capacity are used. shows an example.
25 1
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).
252
The simulation of the construction process is done the horizontal displacements of the soil towards the
by a step-by-step analysis beginning with the in-situ slurry filled trench before pile installation It is appar-
stresses, then modelling trench and pile installation ent that the pile is located in the area of influence
and finally the stepwise loading of the pile The load Figure 1 1 shows the vertical displacement of the
was applied in steps of 200 kN up to failure soil around the axially loaded pile in the state of fail-
Two different situations were investigated and the ure
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 The gained by finite ele-
trench ment analyses as well as obtained by the dynamic pile
tests before and after the slurry trench installation are
given in Figure 12
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
5 CONCLUSIONS
253
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
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.
255
Table1:Soil characteristics (oC= 500 Wa) During driving tests the pile-soil interface can be or
not drained at the top and bottom sample. The drain-
I Plasticity limit I 31 % 1 age allows seepage of water induced by excess pore
Plasticity index 23 pressure along the steel rod.
Compression index 0.70
I Water content I 39 % I 3. EFFECT OF PILE DRIVING IN CLAY
I Shear strength 65 W a
3.1. Displacement of soil due to pile driving
Before each driving tests series ten hammer blows Driving pile through a soil layer is always combined
are performed in order to avoid the "healing" effect with deformation of the soil adjacent to the pile. In
in the pile-soil interface and making the soil under non-cohesive soils voids decrease due to compaction
current driving conditions. The piling is operated at effects of driving, while the volume of cohesive soils
a steady blow rate of one blow per minute approxi- remains constant due to lower permeability and the
mately. Several clay samples are tested in various dynamic nature of solicitation. Near the pile partially
conditions. At the end tests, water content and high excess pore water pressure arises. It decreases
undrained shear strength are measured using for the after the end of driving process due to reconsolida-
second parameter the laboratory vane test. The evo- tion of soil. According to Steenfelt et al. (1981),
lution of this characteristic with water content is il- during pile installation significant horizontal move-
lustrated on figure 2. This evolution is quite linear ment of soil occurs even in a distance up to ten pile
for the homogeneous reconstituted clay: the radius f?om the pile axis. At this distance move-
undrained shear strength decreases linearly when the ments of the magnitude of 0.05 times the pile radius
water content increases. 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.
256
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
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
"We 3: Of Pore Pressure and pile-soi] interface. There is a large decrease of ex-
pile displacement during driving. cess pore pressure at the first blow before reaching a
steady state of constant excess pore pressure. But the
It is evident that in piling the displacement in- unit shaft friction does not follow the same evolu-
creases with hammer energy. Using driving tests re- tion, showing a low variation. When the excess pore
suits involving several hammer energies, the evolu- pressure decreases the effective stress in the soil
tion of excess pore pressure and pile displacement must increase. The measured unit shaft friction does
with driving energy is investigated (figure 4). 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.
258
4. EFFECT OF SURCONSOLIDATION ON IN- 5. CONCLUSIONS
DUCED PORE PRESSURE DURING DRIVING
The experimental set up in laboratory allowed an in-
The clay reconstituted sample was consolidated un- vestigation of the influence of pile driving in in-
der a confrning pressure of 500 kPa and driving tests duced pore pressure. Such experiments are difficult
were performed under an isotropic stress of 300 kPa. to perform in field piling. The dimensions of driving
The variation of shaft resistance with the overcon- laboratory model can affect the interpretation of in-
solidation ratio (OCR) was investigated (Benamar et duced pore pressure measures, but the general con-
al. 1993) and an increase of 30 % in unit shaft resis- clusions concerning the influence of various pa-
tance is noted for a value 1.67 of OCR. Driving piles rameters seems to agree. The variation of pore water
is than more hard in overconsolidated soils. The in- pressure in the soil during driving do not induce the
fluence of overconsolidation on the induced pore same variation on the dynamic shaft resistance. The
pressure in soil is carried out (figure 6). For an OCR experimental results show that very small excess
value of 1.67, the excess pore pressure is higher in pore pressures are developed due to dynamic shear-
normally consolidated soil. This difference is much ing of the pile (pore pressures less than 2 % of the
more marked at the first blow, excess pore pressure consolidation stress). The pile driveability is not
being four times more important in normally con- strongly affected by the excess pore pressure which
solidated soil. In the rest of driving sequence the dif- partially dissipate in the pile-soil interface. The
ference of induced pressure is less important, the ra- measured pile displacements seem very low by
tio being approximately equal to 2. The driving tests comparison with the applied dynamic energy, im-
are performed in each case after a wait, leading con- plying high shaft fiiction. Development of excess
solidation to be made. The surconsolidation makes pore pressure depends on the initial state of stress in
the soil sample in low stresses conditions and wave the soil before driving. Parametric studies show that
stress is not sufficient to reach stress statelevel as in the most significant effects associated with pore
normally consolidated soil. These results explain water pressure are the reduction of the maximum
perfectly the conclusions concerning the shaft fiic- skin fiiction allowed at the interface and the altera-
tion mentioned above. Nevertheless the increase of tion of the pile response due to slippage. The shaft
unit shaft resistance does not reache the same pro- resistance can be fully mobilized easily without pore
portions as excess pore pressure. The such parameter water pressure than with it.
can not be directly related to the shaft resistance. An
effective stress method of analysis for estimating
dynamic pile skin fiiction taking into consideration REFERENCES
the induced pore pressure can not be adjusted, using
effective stress theory. 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
Figure 6: Induced pore pressure in normally consoli- sands. Proc. 4th Int, Con$ on the Application o,
dated and overconsolidated clay Stress-wave Theory to piles, The Hague, Thc
Netherlands, Sept 1992., pp. 117-120. Balkerna.
259
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
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
261
pile in clay shales formed in Cenozoic age, Taubate depth between 3.80 to 11.00 meters it was
Basin, S.P. (Bernardes et a1.1999) it was possible to horizontally bedded clay shale formation, plastic,
v e r that
~ the bearing capacity prediction was too dark gray to graphite. Underneath, the soil layer
conservative while the prediction for steel piles in consist of stiff clay shale with no sign of lamination,
clay shales from "Ilhas" Formation in Bahia high plasticity, green, until the end of SPT boring at
(Velloso and Hammers 1982), showed exactly the 19.00 meters deep. The ground water table was
opposite situation. Therefore, it is recommend to measured at one meter below the ground surface.
propose factors and constants to be used in this The bearing capacity predictions for the clay shale
semi-empirical method related to a specific layer based on DCcourt & Quaresma Method
geological formation. (Fig.1b), gives 54.0 kPa and 4720.0 kPa for average
In this paper, the influence of the pile diameter in unit shaft and point resistance respectively.
the bearing capacity is investigated. The predictions
using DCcourt & Quaresma Method is compared
with the dynamic load test results executed in 3 DYNAMIC TEST PILE PROGRAM
precast concrete pile in clay shale.
Three pile with different diameters (23, 26 and
33 cm) were driven down to 9.00 meters deep. These
2 GEOLOGICAL AND GEOTECHNICAL test piles were precast concrete piles having a yield
CONDITION OF THE SITE stress of 35.0 MPa, Elastic Young Modulus of
30.15 GPa and speed wave propagation of
The area of study is located in Taubat6 City founded 3510.0 m/s. The driving equipment was a free fall
over Terciary sediments formed in Cenozoic age drop hammer with 20 kN ram weight for the 23 and
known as Taubat6 Basin. These sediments cover an 26 cm pile diameters and a 28 kN ram weight for the
area of approximately 170 km in length and 20 km 33 crn pile diameter. A capblock of hard wood and a
wide forming the Paraiba River Valley. The cushion made with Masonite fiber plate were used
representative soil profile of the sediments near the together with a steel helmet to reduce the high
test area is presented in Figure la. 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.
Figure 1. Soil characteristics and bearing capacity predictions For all three piles, it can be seen from Figure 3
with depth. 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 upper layer of sandy clay fill with dirt is 2.70 The results of the CApWApC@analysis made
m thick overlying a 1.10 m thick layer of very soft with the signals from Figure 3 are presented in
organic clay, black, with fragments of roots. For Table 2.
262
Figure 2. ( b x ) versus (Dmx) from Case@Method with J=0,4
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.
(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
264
Engenharia de FundaqGes, 1 :45-54. Rio de Janeiro:
Rio de Janeiro.
265
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
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.
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- were the soil conditions in the founding Werribee
ray Street Gravels, MSG - the lowest Qua- Formation. A borelog close to the initial piling lo-
ternary stratum) cation at Pier 4 is shown in part in Figure 2 at the
e 3m of dense sands (Newport Formation, NF depth of the Werribee formation. Note in particular
-the highest Tertiary stratum) that the SPT N values were 50 or greater throughout
e 13m of dense to very dense sands and sandy the layer. An automatic trip hammer device was
gravels (Werribee Formation, WF) used. It is also emphasized that the formation at this
e bedrock - Melbourne mudstone of Silurian location is predominantly sand, with evidence noted
Age (Silurian, S) of some (<5%) silt and traces (4%) of clay. Occa-
The sequence at Pier 3 was generally similar, al- sional layers of slightly clayey sand were found.
though the upper layers are predominantly absent Piles were driven with a purpose-built 16 tonne
because of the dredged river channel. A thin, ex- Juntann hydraulic drop hammer. Preliminary Wave
tremely weathered Tertiary Age Basalt layer (TOV) Equation analysis had indicated that this hammer,
was also intersected at Pier 3. Estimates of with a maximum stroke of 1 . 2 would ~ be required
undrained shear strength and the results of SPT to reach the target founding depths. A hammer with
testing at Piers 3 and 4 are also shown in Figure 1. a larger energy rating would be required to demon-
All piers are supported on large groups of driven strate the required ultimate capacity.
open-ended steel tube piles. The piles are 1200mm Testing at the site consisted of monitoring 15% of
diameter with a wall thickness of 20mm. No addi- piles using a Pile Driving Analyzer@and 3 static pile
tional reinforcing plate was used at the pile toe. De- load tests.
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 2 TRIAL PILE DRIVING
bank of the river. In both cases the target founding
stratum was the Werribee Formation comprising Driving of the first test pile commenced at Pier 4.
dense to very dense sands. An ultimate capacity of This pile, denoted TPl penetrated very easily to the
25MN was computed for piles at both piers. Allow- target depth. A hammer stroke of only 400 to
ance had been made for downdrag in the design. 500mm was required - significantly less than the
maximum stroke of 1200mm. This was unexpected,
given previous wave equation predictions. The
37 CLAYEY GRAVELLY SAND (SC-GC) very dense, grey-
green, fine to coarse sand, subanbdar gravel N>50 (33.45, hammer operated very well, with energy transfers
5135mm)
consistently in excess of 90%. Table 1 shows the
38 Increasing sand content
driving record for the last 3m of driving of TPl to
39 N>50 ( I 7,28,
target depth. Depth intervals are 250mm. The en-
221100mrn)
ergy figures are those determined from proximity
40 SAND (SM-SP) very dense, green-my, some silt, trace switches built into the hammer.
clay; gravel medium to m a n e grained
46
Grey-brow1 clayey sand layer to 46.4m
(SP) Grey, coarse-grained, slightly gravelly
47
48
Light grey, trace of fines
Of particular importance, both to the pile design The pile was allowed to set-up at this depth for a
and to the dynamic pile test data discussed hereafter 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.
269
The different hammer characteristics are evident
from comparing the input waveform of the fxst rec-
ord against all the other records.
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
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- the "cookie-cutter" effect. As a result, the dynamic
deflection prediction is, as stated, very good. 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.
. tan 6
z = ort1
4 DISCUSSION
The project is notable for a number of reasons: Consider the driving of Pile 301, and in particular
the inferred shaft resistance of approximately 13 kPa
1 It provides additional documented 'Class A' at a depth of 43m. Assuming an average bulk den-
comparisons between dynamic and static load sity of 19 kN/m3 to this depth, and a water table at
testing, in particular for open-ended steel tube ground surface, the vertical effective stress at this
piles. depth would be approximately 400 kPa. According
2 The shaft resistances of the tube piles during in- to Kulhawy (1 984), the horizontal effective stress
stallation was significantly lower than anticipated for this type of low-displacement pile should be ap-
in design, or than would be considered reasonable proximately equal (i.e. K = 1.O, and (rrll= 400 H a ) .
by conventional geotechnical analysis. The pile-soil fiiction angle, 6, is a hnction of the
3 The project provides some evidence, albeit indi- soil fiiction angle, $, and the pile material, in this
rect, regarding the proportion of internal and ex- case steel. Kulhawy and others recommend values
ternal shaft resistance of steel tube piles of 6/+ of approximately 0.7. For the dense to very
4 The shaft resistance of piles at this project in- dense sands at this depth, a Eriction angle of at least
creased dramatically (but inconsistently) over 40" should be expected. From which 6 = 28" is es-
time, even in deposits that only contained a small timated.
fiaction of fines 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
4.1 Load test comparisons depth is accepted, the expected shaft resistance
Predicting the behavior of open-ended steel tube would be at least 90 kPa at this depth. it is noted
piles can be complicated because of physical differ- that the value of 210 kPa actually agrees reasonably
ences in pile response during dynamic and static with the shaft resistance values determined for the
loading. In particular the behavior of the soil col- third restrike of Pile 301.
umn, as plugged or unplugged can be quite different This suggests that the long-term response of these
during static and dynamic loading. During a dy- piles can be justified by geotechnical theory. The
namic event, the pile will tend to move past the in- key question is what caused the severe reduction (to
ternal soil column because of the inertia of the soil 13 kPa) during pile installation.
column. The pile therefore acts as an unplugged or The possibility of a reduction in interface fi-iction
partially plugged pile. This has been referred to as angle is discounted. Such changes have not been re-
ported. The presence of mica is known to reduce
272
interface friction angle, but this effect would be It is interesting to observe that the characteristic
permanent; would not cause a reduction of this mag- resistance distribution is linked to the pile penetra-
nitude; and in any case, mica was not reported in tion rather than the adjacent soil, suggesting a
these sands. structurally-induced geotechnical phenomenon. It is
Temporary increases in pore water pressure can proposed that the pile was flexing in a Mode 2 Euler
reduce the effective horizontal stress in fine-grained buckling shape. With this type of flexure, a local
soils, resulting in substantially lower installation re- minimum in movement (and maximum in shaft re-
sistance, however, in this case, the sands are too sistance) could be expected at mid-length. Lateral
coarse to sustain elevated pore pressures. Note the movement would be greatest at the !A and % lengths,
capacity of test pile 1 only recovered marginally causing the greatest tendency for formation of a
over a period of 4 days. cavity, and hence least shaft resistance.
Another mechanism for reduction of lateral stress Further research is required to substantiate this
on the pile must be found. Tomlinson and Poskitt hypothesis.
have both previously observed the phenomenon of
pile whipping, which can create a cavity around a
4.3 Internal shaft resistance
pile. This cavity can only be sustained if the soil
possesses some amount of binder which gives the As noted earlier, the evidence at this site suggests
soil a temporary or permanent cohesion. A dense that internal shaft resistance was a small percentage
soil in particular may be capable of self-support and of the shaft resistance measured during dynamic
arching f?om the pile surface. testing. The results of removing the internal soil
For the want of a more plausible explanation, the column were obscured by other effects which caused
phenomenon of pile-whip is believed to have caused the external shaft resistance to increase in the period
low contact stresses between the pile and the soil, between the restrikes prior to and after the soil re-
and hence the low observed driving resistances. moval.
Given that the pile hammer and assembly had a Construction programming prevented restrike test-
combined weight of 28 tonne; that the hammer was ing immediately before and immediately after re-
not restrained by any leader system; and that free moval of the soil column. Such a procedure would
pile lengths above ground of up to 20m were ob- have minimized other effects and made the compari-
served during driving, sway and pile whip could be sons of greater value. Further such testing is rec-
feasible. ommended in the fkture to provide some reliable
Support for this conjecture may also be found in international data on internal shaft resistance for
the driving shaft resistance distribution. The resis- open-ended steel tube piles.
tance distributions for test-pile 1 shown in Figure 3
at the completion of driving and for various stages of
4.4 Shaft resistance increases
restrike show a local maximum at mid-pile length
(depth 20m) and local minima at the !A and % posi- The shaft resistance increases measured at this
tions (8m and 32m respectively), especially at the site would appear to be linked to the very low initial
end of drive. shaft resistances measured during installation and
This resistance distribution was characteristic of discussed at length in Section 4.2.
the initial driving behavior of most of the piles In that discussion generation of locally elevated
tested, and this characteristic also persisted in some pore pressures was discounted as a mechanism for
of the restrike data. Figure 7 shows the resistance the low shaft resistances measured in the sands.
distributions for Pile 301 at the end of driving, and Equalization of pore pressures must logically there-
also at a penetration of 43.7m, 7m above final toe fore be discounted as a mechanism for the observed
level. 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
Figure 7. Shaft resistance distributions for Pile 301 during in- increased horizontal stress at the pile-soil interface.
stallation It is quite possible that these stress levels would de-
273
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
REFERENCES
274
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
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.
275
where M is the mass matrix, C is the damping and K The pile-soil interface was modeled as impervious
the stiffness matrices. Each array is composed by the by slaving the horizontal displacements of the fluid
following sub-matrices: 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.
276
5 RESULTS used has an associated flow law, with the consequent
dilative behavior of the plastified points in the soil.
Results for the quasi-static load are presented and This behavior causes higher confining stress around
discussed first. The results show that for the elastic the shaft, providing higher skin friction and limiting
solution, without interface elements, the skin friction the force reaching the pile tip. Also, as the soil di-
provides most of the resistance to the applied load lates, the pore pressure is reduced. The combined ef-
and only a small percentage of the load reaches the fects explain the small departure of the plastic curve
tip (close to 3.5%). This load generates pore pres- from the elastic solution. It is expected that a more
sure in phase with it. Figure 3 depicts the excess of complete model, with non-associated flow law and a
pressure, normalized by the hydrostatic pore pres- cap to control the volumetric plastic strains would
sure, for soil with high and low permeability. As perform better.
should be expected, the less permeable soil has the On a second set of runs, the same problem was
highest pore pressure increase. 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.
Figure 5. Pore pressure ratio at the pile tip for the clayey soil
277
The results presented so far relate to a quasi-static 6 CONCLUSIONS
load. It is of great interest to study the importance of
the loading rate on the pile-soil response. For the The purpose of the current research is to evaluate
following set of analysis, the pile was loaded by a the potential of a FEM coupled dynamic formulation
single pulse, with the duration of 6.3 msec and the to predict pore pressure generation on piles installed
response was followed for 25 msec. This scenario is in saturated soil under dynamic axial loads. A small
closer to a typical dynamic load test. study to determine the influence of permeability,
On the case of the sandy soil, the Drucker-Prager soil’s reologic models and loading rate is presented.
model was used with the same properties as before. 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
Figure 6. Pore pressure response for dynamic loading condi-
tions on sandy soil.
model the load transfer through skin friction.
Reologic models capable of modeling cyclic and dy-
The influence of the loading rate is captured on namic soil behavior are also of interest.
figure 6. The excess pore pressure raises to close to Most importantly, calibration with values meas-
80% of the hydrostatic value, a significant increase ured in field and laboratory tests is required.
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- REFERENCES
sults.
Similar findings are shown on Figure 7. Here the Biot, M.A. 1955. Theory of Elasticity and Consolidation for a
Porous Anisotropic Solid, J. of Applied Physics, 26, 182-
clayey soil was used along with the CamClay model. 185.
A significant increase of pressure is computed but Biot, M A . 1956. Thcory of Propagation of Elastic Waves i n a
now it is quickly damped. 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-
Figure 7. Pore pressure response for dynamic loading condi- gineering, ASCE, 11 1 (GT3), March, 367-383.
tions on clayey soil.
278
5 Codes
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 1503
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.
28 1
2 STATE OF STRESS DESIGN Limit States Design (LSD). The two types of limit
METHODOLOGIES states are the Ultimate Limit State (ULS) and the
Serviceability Limit State (SLS) both of which must
2.1’ Working stress design be satisfied when using the LSD method. Regardless
Working Stress Design (WSD) method, also called of the complexity of the analysis and calculation, all
the Allowable Stress Design (ASD) method, has limit states designs are carried out to satisfy the fol-
been used traditionally in Civil Engineering since it lowing criteria for ULS and SLS, respectively,
was introduced in the early 1800s. The design loads,
Q, consist of the actual forces estimated to be ap- Factored resistance 2 Factored load effects (2)
plied directly to the structure.
Deformation 2 Tolerable deformation to remain ser-
viceable (3)
282
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
Figure 1 . An illustration of probability density functions for LRFD rules for ship structures that were developed
load effect and resistance.
for the US Navy. The AASHTO LRFD Bridge De-
sign and Construction Specifications (1994) resulted
from work in NCHRP Project 12-33 (Nowak 1993,
In LRFD, partial safety factors are separately ap- 1999) provide design guidance for girders.
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) 3.3 LRFD performance and advantages
load effects. In this format, the strength is reduced Experience has shown that adopting a probability-
and the load effects are increased, by multiplying the based design code results in significant savings
corresponding characteristic (noniinal) values with and/or efficient use of materials. Reliability im-
factors, which are called strength (resistance) aiid provements are still under evaluation even though
load factors, respectively. The nominal values (e.g. the new codes are designed for reliability to be equal
the nominal strength, R,,) are those calculated by the to or better than the older codes. Experiences are not
specific calibrated design method and are not neces- yet well documented, but designers have commented
that, relative to the conventional working stress
sarily the mean strength or resistance, 3
and (see code, the new AISC-LRFD requirements are saving
Figure 1). For example, while and the distribution from 5% to 30% stecl weight, with a 10% typical
represent the mean and performance of signal savings (Ayyub 1999). This may or may not be the
matching analyses predictions from many case histo- case for other industries. Specific benefits in pile de-
ries, R,, is the predicted value for the specific ana- sign include the following:
lyzed pile. 1 . A more efficiently balanced design results in cost
Based on considerations ranging from analyzed savings and/or improved reliability.
2. Uncertainties in the design are treated more ra-
case histories to existing design practices, a pre-
tionally and rigorously.
scribed value is chosen for the probability of failure. 3. An improved perspective of the overall design
The probability for failure of the designed pile fol- and construction processes (sub- and super-
lowing the application of the resistance and load fac- structures), and the development of probability-
tors should be smaller than the prescribed value (the based design procedures can stimulate advances
overlapping area in Figure 1). Practically, the resis- in pile analysis aiid design.
tance factors are calculated to provide the prescribed 4. The codes become a living document that can be
probability of failure. easily revised to include new iiiformatioii reflect-
ing statistical data on design factors.
5 . The partial safety factor format used herein also
3.2 Background information
provides a framework for extrapolating existing
The concept of using the probability of failure as a design practice to new foundation concepts and
criterion for structural design was presented in the materials where experience is limited.
283
3.4 LRFD in geotechnical engineering the pile databases alone. Knowledge factors need
therefore to be developed and combined with the re-
Early use of limit state design for geotechnical ap-
liability-theory-based calibration of the database re-
plications was examined by the Danish geotechnical cords to achieve a meaningful LRFD approach.
Institute (Hansen 1953, 1956) and was later formu-
lated into code (Hansen, 1966). Independent load
and resistance factors were used, with the rcsistance 3.5 LRFD forpiles
factors applied directly to the soil properties rather 3.5.1 1999 AASHTO LRFD Bridge Design Spec$-
than to the nominal resistance. cations for Foundations
Considerable effort has been directed over the past The ultimate resistance (I?,,) multiplied by a resis-
decade to the application of LRFD in geoteclinical tance factor (4, which is equal to the factored resis-
engineering. LRFD approaches have been developed tance (R,.), must be greater than or equal to the sum-
in offshore engineering (e.g. Tang 1993, Hamilton & mation of the loads (Q,) multiplied by a modifier (71,)
Murff 1992), in general foundation design (e.g. Kul- and load factor (x)for the strength limits states:
hawy et al. 1988, 1994, 1996), and in pile design for
transportation structures (Barker et al. 199 1, O’Neill (4)
1995).
In geotechnical practice, uncertainties in resis- where:
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- 11, = factors to account for effects of ductility ( q ~ ) ,
struction effects. Uncertainties in extzrnal loads are redundancy ( q ~ and
) , operational importance (71).
small compared with uncertainties in soil and water The 1999 AASHTO LRFD Bridge Design Speci-
loads and the strength-deformation bchaviors of fications for Foundations provides the following
soils. The approach for selecting load and resistance equations for the determination of the factored bear-
factors developed in structural practice, though a ing resistance of piles, QR:
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. for which:
Philosophically, the selection of load and resis-
tance factors need not be made probabilistically, al- Q,>= q,A, (7)
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 where $q = resistance factor for the bearing resis-
subjective factors unique to geotechnical practice. tance of a single pile specified for methods that do
An expanded approach is needed if the full benefits not distinguish between total resistance and the indi-
of LRFD are to be achieved for foundation design. vidual contributions of tip resistance and shaft resis-
The National Research Council reported, “this.. . tance; Q,,[,= bearing resistance of a single pile (F);
subjective approach reflects the general lack of ro- Q,= pile tip resistance (F); Qs = pile shaft resistance
bust data sources from which a more objective set of (F); qr, = unit tip resistance of pile (F/L’); qs = unit
factors can be derived” (NRC 1995). The report shaft resistance of pile (F/L’); A, = surface area of
continues, “. . .realistically, because of the tremen- pile shaft (L2); A, = area of pile tip (L2); and #q)qp,#qs
dous range of property values and site conditions = resistance factor for tip and shaft resistance, re-
that one niay encounter, it is unlikely that com- spectively, for those methods that separate the resis-
pletely objective factors can be developed in the tance of a pile into contributions from tip resistance
foreseeable future.” and shaft resistance,
Today, the situation has changed somewhat, but The resistance factors for use in the above equa-
not entirely. The research team of NCHRP 24-17 tions are presented in Table 10.5.5-2 of the specifi-
gathered robust data sources on pile capacity from cations. The table incorporates a factor h,, to ac-
which a more objective calibration of resistance fac- count for the level of field capacity verification. As
tors can be made. Nonetheless, there remain uncer- an example, if a wave equation analysis is used to
tainties associated with (1) site conditions, (2) soil determine the pile capacity with an assumed driving
behavior and interpretation of soil parameters, and resistance, a resistance factor of 0.65 is recotn-
(3) construction methods and their quality. These mended. If pile driving formulas, e.g. ENR equa-
factors are difficult or impossible to understand from tion, are used to verify the pile capacity without
stress wave measurements during driving, a h, factor
284
of 0.80 is recommended. This results in an actual The current AASHTO specifications encounter
resistance factor of 0.52 to be used in the above additional difficulty due to the irrational multiplica-
analysis/verification sequence (0.65 x 0.80). tion of the resistance factor by the modifier h,. This
procedure requires the interaction of two independ-
3.5.2 Worldwide LRFD for dynamic methods ent evaluations (e.g. static analysis and dynamic
Review of the development of LRFD applications methods) and results in unnecessary and confusing
for Geotechnical Engineering is presented by Goble conservatism. A clear separation of the resistance
(1999). A short description of LRFD codes that in- factors on the basis of design and construction is re-
clude dynamic methods follow. quired.
The Australian Standard for Piling-Design and
Installation ( 1995) provides detailed recommenda-
tions for resistance factors to be used with the dy- 4 PD/LT2000 DATABASE
namic methods. The recommended resistance fac-
tors range betwecn 0.45 to 0.65 for methods without The database PD/LT2000 contains information re-
dynamic measurements (including WEAP) and be- lated to 210 driven piles that have been statically
tween 0.50 to 0.85 when utilizing dynamic ineas- load tested to failure and dynamically monitored
urements with signal matching analysis. Selection of during driving and/or restrike (389 analyzed meas-
the appropriate resistance factor depends on driving urements). PD/LT2000 is comprised of the integra-
conditions, geotechnical factors (e.g. extent of site tion of databases PD/LT (Pailtowsky et al. 1994) and
investigation), and extent of testing (e.g. low range PD/LT2 (Paikowsky & LaBelle 1994) with expan-
for ~ 3 of % the pile tested and high range for >15%). sion by an additional 57 pile cases (Stenersen 2000).
In traditional structural design specifications, a The data of PD/LT2000 were carefully examined
nominal value is given and the value used is based and analyzed following procedures described by
primarily on engineering judgment and cannot ex- Paikowsky et al. (1994), resulting in detailed static
ceed the nominal value. The Australian Standard is and dynamic pile capacity evaluations. Table 7 pre-
therefore unique by providing a guide for choosing sents a summary of the data contained in
the appropriate resistance factor. Interestingly, no PDiLT2000 broken down according to site location
distinction is made regarding either soil type or time and soil type, pile type and capacity, driving behav-
of driving (i.e. EOD, BOR) when referring to the ior and time of driving.
signal matching based on dynamic ineasureinents.
The method by which the resistance factors were
generated is not provided in the code. 5 REFERENCE STATIC CAPACITY
The AUSTROADS Bridge Design Code (1992) 5.1 Overview
provides resistance factors for four categories of dy-
namic methods. The range of resistance factors is The dynamic methods are assessed by comparing the
quite large and there is no explanation as to how the pile capacity of the evaluated method with a refer-
resistance factors were obtained. Goble (1 999) pos- ence static capacity of the pile. The determination
tulates that the resistance factors were calibrated via of the pile’s static capacity based on load- displace-
the working stress design method. ment relations is not unique. The test results depend
The Ontario Bridge Code (1 992) recommends on the load testing procedures and the applied inter-
relatively low resistance factors with no differentia- pretation method, both of which are subjective. The
tion between the dynamic analyses. No information following sections examine each of these factors and
is provided on how the resistance factors were ob- its influence on the reference static capacity.
tained.
The Canadian Bridge Code (1992) is brief in its 5.2 Failure Criterion
design requireinents for deep foundations. For dy-
namic load testing, resistance factors of 0.4 and 0.5 Past work (Paikowsky et al. 1994) have resorted to a
are recommended for routine testing and analyses “representative” static pile capacity based on the as-
based 011 dynamic nieasurements, respectively. sessment of five interpretation methods; Davisson’s
Criterion (Davisson 1972), Shape of Curve (similar
3.5.3 Dificu1tie.s of the Existing LRFD Codes to the procedure proposed by Butler and Hoy 1977),
All existing codes suffer from two major difficulties. Limiting Total Settlement to 25.4 mm and to 0.1B
One is the application of LRFD to geotechnical (Terzaghi 1942), and the DeBeer log-log niethod
problems as described in section 3.4 (e.g. site vari- (DeBeer 1970). A single representative capacity
ability, construction effects, etc.). The other problem value was then calculated for the analyzed case as
is lack of data. None of the reviewed codes and as- the average of the methods considered relevant (i.e.
sociated resistance factors were developed based on provided reasonable value). The development of a
databases enabling the calculation of resistance fac- framework for future modifications requires that the
tors from case histories. evaluated resistance factors be based on an objec-
285
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 diameter of‘ 610nini (examined through 30 pile
static capacity of each pile in database PD/LT2000 cases). The data presented in Figure 2 demonstrates,
was evaluated utilizing all five aforementioned crite- however, that though small, a bias exists in the static
ria and a representative capacity was assigned for capacity used as a reference for the evaluation of the
each pile. A statistical analysis was then carried out dynamic methods’ performance.
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- 5.3 Load Test Procedure
gram and calculated distributions (normal and log- An additional factor to examine is the influence of
normal) for Davisson’s failure criterion. Davisson’s the static load testing procedure (loading rate) on the
criterion was found to perform the best overall and designated pile capacity. This influence was exam-
was therefore chosen as the single method to be used ined in two ways.
when analyzing load-displacement curves. The Two detailed case histories from a research site
method provides an objective failure criterion and in Newburyport, Massachusetts were evaluated. A
was also found to perform well for piles exceeding a pipe and pre-stressed concrete heavily instrumented
286
Figure 3. Comparison between pile capacity based on Davis-
Figure 2. I-listograni and frequency distributions of KsD for son’s criterion for slow maintained load tests and static cyclic
186 PDiLT2000 pile-cases in all types of soils. load test capacity for 75 piles, (Paikowsky et al. 1999).
friction piles were tested over a lengthy period of These evaluations led to the conclusions that
time at a bridge reconstruction site. Both piIes were Davisson’s pile failure criterion can be used as a
tested using three types of static load testing proce- method to determine the reference pile capacity irre-
dures: slow maintained (testing duration of about 45 spective of the static load-testing procedure.
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- 6 DRIVEN PILES AND DYNAMIC METHODS -
sented by Paikowsky and Hajduk (1999, 2000) and STATE OF PRACTICE IN USA HIGHWAY
Paikowsky et al. (1999). Thc interpretation of the CONSTRUCTION
load-displacement relationships in both cases sug-
gested that the test type had an insignificant influ- To address the state of practice in the USA highway
ence on the pile capacity, (referring to a failure crite- construction, a questionnaire was sent to the De-
rion irrespective of the displacement). partment of Transportation offices of all fifty states,
The effect of the test type was further investigated requesting information on the use, design and con-
utilizing a databasc containing information related to struction of deep foundations. The main results
75 piles tested under slow maintained arid static- from 45 states and 2 Federal Highway officials re-
cyclic load testing procedures. In the static-cyclic garding driven piles are:
procedure, the piles were loaded to failure using a 0 Driven Piles constitute 75% of all bridge founda-
high loading rate and then unloaded. The process tions. From the people who responded, 64% prefer
was repeated for four cycles. The testing procedure the use of driven piles to any other foundation type.
and its interpretation method are presented by Pai- In the driven pile categories: 21% PPC (Pre-cast
kowsky et al. (1 999). A comparison between the pile Pre-stressed Concrete), 52% H-Piles, 25% CEP
capacity based on Davisson’s failure criterion for the (Closed Ended Pipe) and 2% OEP (Open Ended
slow maintained tests and the static-cyclic capacity Pipe).
is presented in Figure 3. The obtained relations and 0 Dynamic Equations usage: 45% ENR (Engineer-
the associated statistical information suggest that ing News Record, Wellington 1892) and 16% Gates
there is no significant influence on the static pile ca- (Gates 1957).
pacity based on the applied static load rate. The 0 WEAP (Wave Equation Analysis Program) is used
288
Table 3. Summary of the available dynamic methods.
Categoiy
I Metlrod
WEAP
Advai7 rages Disadvnt 1rages Cotntiiell/
~ ~ - 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
289
procedure of signal matching was first suggested by that during high driving resistance (i.e. refusal) there
Goble et al. (1970), utilizing the computer program is not sufficient pile penetration to mobilize the full
CAPWAP. Others developed similar analyses, (e.g. pile capacity (e.g. Chellis 1961). As such, the dy-
Paikowsky 1982, Paikowsky & Whitman 1990) namic methods are deficient under high driving re-
utilizing the computer code TEPWAP. The TNO sistance, categorized as equal or above 12BPI
program was developed by Middendrop and van (Blows Per Inch) or approximately 5BPcm (Blows
Wee1 (1986), which led to improvements and the Per cm), (MHD 1988).
CAPWAPC program, which is in coininon use to The soil type is also believed to constitute a major
date. The signal matching technique is often refei-red factor as soil damping parameters are commonly
to as post driving analysis or office method. With employed to represent viscous resistance in the
the availability of faster, portable computers, it modeling of the soil’s dynamic behavior. This vis-
became reasonably simple to conduct the analysis in cosity is assumed to be soil type dependent, and as-
the field as well. In difference from the ficld method sociated with intrinsic soil properties. As such, high
however, the signal matching analyses cannot be viscosity values are expected for cohesive soils and
carried out for each blow during driving. low viscosity values are expected for cohesionlcss
soils. Naturally, under a given velocity, high viscous
values are associated with higher dynamic resis-
8 THE CONTROLLING PARAMETERS tance, and logically should prove more difficult to
accurately define the static resistance.
8.1 Overview The effect of time is well recognized but poorly
Preliminary examination of the parameters control- quantified. Piles undergo a decrease or increase of
ling the performance of the dynamic analyses is car- capacity with time, also known as relaxation and set-
ried out prior to a final detailed evaluation of these up, respectively. While the resistance during driving
methods, leading to resistance factors. Such exami- and its static component represent the conditions en-
nation iiifluences the sub categorization of the dy- countered during penetration, the major interest re-
namic methods (according to the important control- mains the long-term ability of the pile to carry load
ling parameters), hence, directing the user to utilize during its service life. The examination of the dy-
the appropiate resistance factor according to the namic method predictions with static load tests (of-
relevant conditions of the employed method. For ex- ten carried out long after the driving) therefore re-
ample, if soil type is a controlling factor and the ac- mains valid. The predictions can be assessed in
curacy of the signal matching method is largely af- relation to the time in which the driving and/or the
fected by soil type, evaluation of the method for dynamic data have been obtained (e.g. EOD or
different soil types will result in the development of BOR).
resistance factors, depending on the soil type. Con- The sections below examine the importance of
versely, if soil type does not control the accuracy of each of the above three assumed controlling parame-
the spccific dynamic method, categorization based ters. The results are used to evaluate additional pos-
on soil type is neither desired nor perused. sible controlling factors, laying down the framework
The following sections outline the logic used for the detailed evaluation of the dynamic methods.
for the preliminary examination of the controlling
parameters, the analyses and the results. The ration- 8.2The Eflect oJ’Soil Type
ale presented in this section follow previous studies
by Paikowsky et al. (1994), Paikowsky (1995), Pai- The effect of soil type is examined in two ways: (i)
kowsky et al. (1995), and Paikowsky and Cher- the correlation between the parameters assumed to
nauskas (1 996). be soil type dependent and soil type, i.e. damping
The evaluation of static capacity from pile driving parameters and (ii) the accuracy of the predictive
is based on the concept that the driving operation in- methods relative to the soil type.
duces failure in the pile-soil system, (i.e. a very fast Figures 4 and 5 present the relationship between
load test is carried out under each blow). Dynamic soil type and Smith damping parameters used in ap-
analyses encounter three fundamental difficulties: (i) proximately 370 CAPWAP analyses, from
correct formulation of the penetration process (e.g. PD/LT2000, for the tip and side pile resistances, re-
soil motion, soil plugging etc.), (ii) evaluation of the spectively. Figure 6 presents the back-calculated
static resistance out of the total resistance encoun- case damping factors for 290 cases, from the PD/LT
tered during penetration, and (iii) time dependent database (Paikowsky et al. 1994), required to obtain
pile capacity, (Paikowsky 1995). The parameters a match between the predicted capacity and the
controlling the accuracy of the dynamic predictions measured static capacity. All three figures clearly
therefore reflect the ability of each method to ad- indicate that no unique relationship exist between
dress the above difficulties. soil type and damping parameters, suggesting that
Based on the concept of a pile loading to failure other mechanism control the value required for a
under each blow, it has been traditionally assumed damping factor rather than the soil type.
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).
291
Table 4. CAPWAP analysis based on soil type categorization.
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**
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 Stdev. 0.1 19 226.3 0.068 7.9 0.106 237.9
compression and disturbance, resulting in soil resis-
tance during driving that differs from the long-term * closed-ended piles only
pile capacity. Although factors such as thixotropy
*‘F
excluding the case from Denmark
and aging contribute to this phenomenon, the most
significant cause for gain of capacity with time is as- on log scale) is similar based on static and dynamic
sociated with the migration of pore water. Meas- measurements (Cg, = 0.389, Cgtd= 0.348) the associ-
urements carried out on a model (Paikowsky & Hart ated time for achieving 75% of the maximum capac-
1998) and full-scale piles (Paikowsky & Hajduk ity (normalized for all piles to 254mm (lft) diame-
1999, 2000) show that pore pressure at magnitudes ter) is about 20 times greater. In other words,
similar to the total soil pressures create in clays dynamic testing (namely CAPWAP) while following
around the pile’s shaft zones of about zero effective the physical behavior of capacity gain, exhibit this
stresses, resulting in almost a complete loss of fric- gain much faster than the actual gain monitored by
tional resistance. Pailtowsky et al. (1995, 1996) ex- the static load test results. The ramifications of
amined the static and dynamic gain of capacity with these conclusions are: (i) actual gain of capacity is
time based on radial consolidation; a normalization much slower than that exhibited by the dynamic
process was followed, allowing for a comparison be- methods, (ii) scheduling of construction or testing
tween different pile sizes. Table 5 presents a sum- based on capacity gain should consider the reason
mary of parameters describing the pile capacity gain for time evaluation (i.e. actual loading in construc-
with time based on static and dynamic testing, The tion or dynamic testing as part of quality control),
presented data shows that while the rate of capacity and (iii) at present, the dynamic methods evaluation
gain (normalized to the maximum capacity vs. time 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 .
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
9 TIHE PERFORMANCE OF THE DYNAMIC FIHWA version of the Gates equation. The cases
METIHODS CATEGORIZED ACCORDING TO with dynamic measurements evaluates both, CAP-
THE CONTROLLING PARAMETERS WAP and Energy Approach methods. The dynamic
methods are brolam down into subsets based on time
9.1 The Amlysed Cases of driving, driving resistance, and area ratios. Judg-
The time of driving, driving resistance, aiid area ra- ment arid statistics guidelines were used for the in-
tio proved to be the major controlling parameters of clusion or exclusion of a few cases. For example, ex-
the dynamic niethods. To facilitate the codes' sepa- treme CAPWAP under-predictions (beyond 2 S.D.)
ration between design and construction, the database were observed at the EOD in one site. All cases 111-
was organized into these categories, followed by chided easy driving and large area ratios and if in-
subcategories of methods that use aiid do not use cluded in the general population of data, the EOD
dy ii aiii1c in eas uremeii t s, with subsets fo 11o w i 11g the statistics would have become 1.861 i 1.483
controlling parameters. Figure 10 presents for the (niean+lS.D.). This site is included only in the sub-
aiialyzed subsets, the number of cases in the set aiid category of blow count < 16 BP 1 Ocin and AR < 350.
the normal distribution mean aiid standard deviation.
WEAP is utilized in the design stage. The analy-
9.2 The critical cases
sis is carried out for driving stress evaluation, lead-
ing to a load fs-ctor, not included i n this paper. The The statistical analyses presented in Figure 10 al-
use of the method for the evaluation of pile capacity lows for the identification of the critical cases that
was examined through WEAP results for default in- require calibration aiid developnient into resistance
put values and the blow count at the EOD compared factors. For exaiiiple, the CAPWAP cases include
to the static load test results. The presented data was (i) all data, (ii) EOD, (iii) BOR, and (iv) the worst
provided by GRL Inc. (Haiinigan et al., 1996). combination of soil motion effect (Blow count < 16
For the construction category, tlie cases without BPlOciii aiid AR < 350).
dynamic nieasureiiients evaluate the dynamic equa- Table 6 presents a summary of the major catego-
tions, specifically tlie ENR, the Gates, and the ries of the dynamic methods that are identified from
295
I
I
Dynamic, Analysis II
Construction Design
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
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 18. Iiistogram and Frequency Distributions for EOD
default value GRLWEAP pile-cases, (99),data provided by
GRL (see Mannigaii et al. 1996).
298
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:
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:
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%
301
calculations of the presented resistance factors. Drs. Ayyub, B., Assakkaf, I. & Atua, K. 2000. Reliability-Based
Mile McVay and Bjorii Birgissoii of the University Load and Resistance Factor Design (LRFD) of Hull Gird-
ers for Surface Ships. N u i d Engineers Joui.1~~1, ASNE,
of Florida and Dr. Ching Kuo of GeoStructures con- May 2000.
tributed to a questionairre resulting in the informa- Ayyub, B., Assakkaf, I., Atua, K., Engle, A., Iiess, P.,
tion presented in section 6. Dr. Frank Rausche of Icaraszewski, Z., Kihl, D., Melton, W., Sielski, R., Sieve,
Goble, Rausche, Likins (GRL) and Associates pro- M., Waldman, J . , 61 White, G. 1998b. Reliability-based
vided the data pertaining to the evaluation of Design of Ship Structures: Current Practice and Emerging
GRLWEAP. 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.
DISCLAIMER NCHRP Report 343. Transportation Research Board, Na-
The research presented here is part of an ongoing tional Research Council, Washington, DC.
project, which is aimed at rewriting the AASHTO 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
Deep Foundation Specifications for the year 200 1. Highway Administration, Office of Development, Report
The presented parameters are preliminary and do not No. FHWA-IP-77-8, Washington, DC.
reflect in aiiyn'ay the recoiiiiiiended or approved pa- Canadian Geotechnical Society. 1992. Caiiadiuti Founckiliot?
rameters. The opinions and conclusions expressed Etigi~eer-ingA4~irziia1,3'd Edition. Bi-Tech publishers, Ltd.,
or implied in the paper are those of the eiiti- Richmond, British Columbia, Canada.
CIRIA 63. 1977. Rationalization of Safety and Serviceability
ties/iiidividuals performing the research and are not Factors in Structural Codes. Construction Industry Re-
necessarily those of the Transportation Research search and Information Association, SWIP 3AU, Report
Board, the National Research Council, tlie Federal 63, London.
IHighway Administration, tlie American Association Chellis, R.D. 1961. Pile Foundations, McGraw M i l l 2'ld ed. NY.
of State Highway and Transportation Officials, or Davisson, M.T. 1972. Iiigh Capacity Piles. P~oceeditigs,Soil
the individual states participating in the National A4ecliutiics Lecficre Series 011 Itiiiovatioris it7 Fourzdafiotz
Coristrucfioii. American Society of Civil Engineers, llli-
Cooperative Highway Research Program. nois Section, Chicago, 8 1- 1 12.
DeBeer, E.E. 1970. I'roefondervindellijke bijdrage tot de studie
van het grandsdraagvermogeii van zand onder funderinger
IGFERENCES op staal. English version, Geofech~ique,Vol. 20, No. 4,
AASHTO, 1994. LRFD Bridge ntid Consrruclion Sjiec(/icir- 387-41 1 .
tio17.s. American Association of State Highway and Trans- Ellingwood, B., Galambos, T., MacGregor, J., & Cornell C.
portat i o n 0ffic i a 1s, Washington, DC. 1980. Developnient of a Probability-Based Load Criterion
AASHTO, 1 997. Sfaiirlwci Specif;cutioti.s.fbl- Higlz~vuyBridges. for American National A58. National Bureau of Standards
American Association of State Highway and Transporta- Publication 577, Washington, DC.
tion Officials, Washington, DC., 16"' Edition (1996 with Ellingwood, B., Galambos, T., MacGregor, J., & Coniell, C.
1997 interims). 1982a. Probability Based Load Criteria - Assessment of
AASHTO, 1999. LRFD Bridge atid Cotisr~uclio~i Speci/icn- Cui-rent Design Practices. Jour-nal of [lie S r t u c t i d Divi-
tioizs. American Association of State Highway and Trans- sion, ASCE, Vol. 108, No. ST5, 959-977.
portation Officials, Washington, DC. Ellingwood, B., Galambos, T., MacGregor, J., & Cornell, C.
ACI, 1995, Builriiizg Code Requireiiients ,for Reiufol-ced Coti- 1982b. Probability Based Load Criteria - Load Factors and
crefe. American Concrete Institute, Detroit. Load Combinations. Jourmd 03''the Stnictut.al Division,
AISC, 1994. Load and Resistance Factor Design. Muiiuul of' ASCE, Vol. 108, ST5, 978-997.
Steel Cotistt*iicfioii, American Institute of Steel Construc- FIHWA, 1988. FHWA Guide Specifications for Driven Piles.
tion, Chicago, IL. Federal Highway Administration.
API, 1989. Draft Recommended Practice for Planning, Design- Freudenthal, A.M. 1947. Safety of Structures. T~ansactionsof'
ing and Constructing Fixed Offshore Platforms - Load and the ASCE. Vol. 1 12, 125-1 80.
Resistance Factor Design. API RP2A-LRFD, American Galambos, T. & Ravindra, M. 1978. Properties of Steel for use
Petroleum Institute, Dallas, TX. i n LRFD. Joul-tin1 U / S t ~ u c t w a Eiigiiieei.ing.
l ASCE, Vol.
ASCE, 1993. Minimum Design Loads for Buildings and Other 104, NO. 9, 1459-3468.
Structures. ASCE 7-93 (formerly ANSI A58.1). Gates 1957. Empirical Formula for Predicting Pile Bearing Ca-
AUSTROADS, 1992. A USTROADS Bridge Design Code, Na- pacity. Civil Eiigiiiewirig, Vol. 27, No. 3, 65-66.
tional Office, AUSTROADS, Surry Hills, NSW, Australia. Goble, G . 1999. Geotechnical Related Development and 1111-
Ayyub, B.M. 1999. Cost-Benefit Analysis of Risk-Infornied plementation of Load and Resistance Factor Design
Design of Structural Systems. American Society of Me- (LRFD) Methods. NCHRP Synthesis 276, Transportation
chanical Engineers, the Research Committee on Risk Research Board, National Research Council, National
Technology, Washington, DC. Academy Press, Washington, DC.
Ayyub, B.& Assakkaf, I. 1999. LRFD Rules for Naval Surface Goble, G . , Likens, G.,& Rausche, F. 1970. Dynamic Studies
Ship Structures: Reliability-Based Load and Resistance on the Bearing Capacity of Piles - Phase 111, Report No.
Factor Design Rules. Naval Surface Warfare Center, 48. Division of Solid A4echunic.s, Sfruclures, atid Mechani-
Carderock Division, U.S. Navy. cal Desigiz. Case Western Reserve University.
Ayyub, B., Assakkaf, I., & Atua, K. 199Sa. Development of Goble, G. & Rausche, F. 1976. Wave Equation Analysis of Pile
LRFD Rules for Naval Surface Ship Structures: Reliabil- Driving-WEAP Program. Vol. 1-4, FHWA #IP-76- 14.1
ity-Based Load and Resistance Factor Design Rules, Part through #IP-76- 14.4.
Ill - Stiffened and Gross Panels. Naval Surface Warfare GRL. 1999. Pile-Driviizg Ana/yzet., PAK Users A4unual. Goble,
Center, Carderock Division, US.Navy. Rauschc, Likins and Associates, Inc.
302
GTR. 1997. Dynamic Pile Testing Report, Central Ar- Nowak, A. 1993. Calibration of LRFD Bridge Design Code.
tery/Tunnel Project C07D2, I-90/Airport Ititerchange Ar- Department of Civil and Environmental Engineering Re-
rivals Tunnel - Phase I, East Boston, Massachusetts. Geo- port UMCE 92-25. University of Michigan, NCHRP 12-
sciences Testing and Research, Inc. North Chelmsford, 33.
MA. Nowak, A. 1999. Calibration of LRFD Bridge Design Code.
GTR. 1998. Dynamic Pile Testing Report, Central Ar- Department of Civil and Environmental Engineering Re-
teiy/Tunnel Project C07D2, I-9O/Airport Intel-clznnge Toll port UMCE 92-25. University of Michigan, NCHRP 12-
Plaza, East Boston, Massachusetts. Geosciences Testing 33.
and Research, Inc., North Chelmsford, MA. Olsen, R. & Flaate, K. 1967. Pile Driving Formulas for Friction
Hajduk, E., Paikowsky, S., Holscher, P., & Barends, F. 2000. Piles in Sand. ASCE JSMFC, Vol. 93, SM6, Nov. 1967,
Accelerations of a Driven Pile and the Surrounding Soil. 279-297.
Proceedings, 6'" International Conjerence on the Applica- Ontario Ministry of Transportation and Communication. 1992.
tions of Stress- Wave Theory to Piles, September 11-13, Ontario Highway Bridge Design Code and Comnientaiy,
2000. S2n Paulo City, Brazil. 3'" ed.
Hamilton, J.M. & Murff, J. 1992. Selection of LRFD resistance Paikowsky, S. 1982. Use of Dynamic Measurements to Predict
factors for pile foundation design. Proceedings Structures Pile Capacity Under Local Conditions. M.Sc. Thesis,
Congress '92. American Society of Civil Engineers, San Dept. of Civil Engineering Technion-Israel Institute of
Antonio, ASCE, NY. Technology.
Iiannigan, R., Goble, G., Thendean, G., Likins, G., & Rausche, Paikowsky, S. 1984. Use of Dynamic Measurements for Pile
F. 1996. Design and Construction of Driven Piles Founda- Analysis. Including PDAP-Pile-Driving Analysis Program,
tions - Volume I1 (Working Draft). U.S. DOT Federal GZA Inc., Newton, Massachusetts.
Highway Administration. Washington, DC. Paikowsky, S. 1995. Using Dynamic Measurements for the
Hansen, B. 1953. Earth Pressure Calculation. Danish Technical Capacity Evaluation of Driven Piles. Civil Engineering
Press. Copenhagen. Practice, Journal of the Boston Society of Civil Engineers
Hansen, B. 1956. Limit Design and Safety Factors in Soil Me- SectiodASCE. Vol. 10, No. 2, 61-76.
chanics. Bulletin No. 1, Danish Geotechnical Institute, Paikowsky, S. & Chernauskas, L. 1992. Energy Approach for
Copenhagen. Capacity Evaluation o f Driven Piles. 4'" International
Iiansen, B. 1966. Code of Practice for Foundation Engineering. Coizference 011 the Application of Stress- Wave Theoty to
Bulletin No. 22, Danish Geotechnical Institute, Copenha- Piles. The Hague, Netherlands. 595-601.
gen. Paikowsky, S. & Chernauskas, L. 1996. Soil Inertia and the
Holscher, P. 1995. Dynamical Response ojSaturated and Diy Use of Pseudo Viscous Damping Parameters. 5'" Interna-
Soils. Delft: Delft University Press. tional Coiiference on the Application of Stress- Wave The-
Holscher, P. and Barends, F. 1996. In-sihi Measurement of ory to Piles. Orlando, FL. 203-2 16.
Soil-Motion near the Toe of a Dynainically Loaded Pile. Paikowsky, S. & Hajduk, E. 1999. Design and Construction of
In F.C. Townsend, M. Iiussein & McVay, M. (ed), Pro- an lnstruiiiented Test Pile Cluster. Research Report sub-
ceedings, 5"' In tern n tion a / Coizfeimce of the AQp/icalion mitted to the Massachusetts Highway Department, Sep-
of Stress- Wave Theory to Piles, September 11-13, 1996. tember, 1999. Boston, Massachusetts.
_nrla.1ld5, F1 76-36 Paikowsky, S. & Hajduk, E. 2000. Theoretical Evaluation and
Kulhawy, F.H. et al. 1988-1994. Reliablilty-Bascd Foundation Full Scale Field Testing Examination of Pile Capacity
Design of Transmission Line Structures. Report EL-5507. Gain with Time. Reseat-ch R e p o i ~to be submitted to the
Electrical Power Research Institute, Palo Alto. Massachusefts Highway Department, December, 2000.
Kulhawy, F.H. et al. 1996. Proceedings, Uncertainty in the Boston, Massachusetts.
Geological Environnzent. Madison, ASCE, NY. Paikowsky, S. & Hart, L. 2000. Development and Field Testing
Long, J., Bozkurt, D. & Kerrigan, J. 1998. Value of Methods of Multiple Deployment Model Pile (MDMP). Research
for Predicting Axial Pile Capacity. Ti~ansportation Re- Report to be submitted to the Federal Highway Admini-
search Board Paper No. 99-1333, 76'" Annual Meeting, stration, April, 2000. FHWA-RD-99- 194 Washington,
Januury, 1999. Washington, DC. DC.
McVay, M., Birgisson, Bjorn., Zhang, L., Pcrez., A & Putcha Paikowsky, S. & LaBelle, V. 1994. Examination of the Energy
S. 2000. Load and Resistance Factor Design (LRFD) for Approach for Capacity Evaluation of Driven Piles. US
Driven Piles Using Dynamic Methods - A Florida Per- FHWA International Conference on Design and Consfruc-
spective. Geotectinicul Testing Joio-nul, ASTM, Vol. 23, tioii of Deep Foundations, December 6-8, 1994. Orlando,
NO. 1, 55-66. FL. Vol. 11, 1133-1 149.
MHD. 1988. Standard Specifications for Highways and Paikowsky, S., LaBelle, V. & Hourani, N. 1996. Dynamic
Bridges. Massachusetts Highway Department. Analyses and Time Dependent Pile Capacity. 5'" Inlerna-
Middendoip, P. & van Weel, P. 1986. Application of Charac- fional Confer.eiice on the Application of Stress- Wave The-
teristic Stress-Wave Method in Offshore Practice. Pro- ory to Piles. Orlando, FL. 325-339.
ceeclings oftlie 3'" Interriationu1 Coiljierence on Numerical Paikowsky, S., LaBelle, V. & Mynampaty, R. 1995. Static and
A4ethods in Oflshore Piling. Dynamic Tinie Dependent Pile Behavior. Research Report
Moses, F. 1985. Implementation of a Reliability-Based API submitted to the Massachusetts Highwqy Department, No-
RP2A Format, Final Report. API PRAC 83-22. American vember, 1995. Boston, Massachusetts.
Petroleum Institute. Paikowsky, S., Operstein, V., and Bachand, M. 1999. Express
Moses, F. 1986. Development of Prelimiilary Load and Resis- Method of Pile Testing by Static Cyclic Loading. Research
tance Factor Design Document for Fixed Offshore Plat- Report subriiilted to the Mussachusetts Highwaji Depart-
forms, Final Report. API-PRAC 95-22, American Petro- nient, October., 1999. Boston, Massachusetts.
leum Institute. Paikowsky, S., Regan, J., and McDonnell, J. 1994. A Simpli-
National Research Cot~ncil. 1995. Probabilistic Methods i n fied Field Method for Capacity Evaluation of Driven Piles.
Geotechnical Engineering. Washington. FHWA Report No. FHWA-RD-94-042, September, 1994.
National Research Council of Canada. 1977. National Building Paikowsky, S. & Whitman, R. 1990. The Effect of Plugging on
Code of Canada. Ottawa. Pile Performance and Design. Canadian Geotechnical
Journal, Vol. 27, No. 4,429-440.
303
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
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
307
The concept of having the resistance factor re- of elasticity of the pile material, C is the velocity of
lated to the static analysis method is inconsistent stress wave propagation and the other two terms are
with current practice. not defined. The source of this approach is not
given.
3.2 Ontario Bridge Code (1992)
The Ontario Highway Bridge Code defines the limit 3.4 Florida DOT (1997)
states for strength and specifies resistance factors for The resistance factors recently developed by the
deep foundations subjected to axial compression Florida DOT in a calibration study are given in Ta-
load as given in Table 5. The resistance factors ble 7. This presentation is clear and presents a good
seem to be quite small when compared with the representation of the Florida practice, one of the
LRFD Bridge Code. There is no discussion of the
most modern in the United States.
methods used to obtain these values.
Table 5. Ontario Bridge Code resistance factors for Table 7. Florida DOT (1 997) resistance factors
deep foundations Design methodology Resistance factor
Axial load Factor SPT97 0.65
Static analysis 0.4 PDA 0.65
Static test 0.6 Wave equation analysis 0.35
Dynamic analysis 0.4 Static load testing 0.75
PDA test 0.5
The resistance factor for wave equation analysis
seems to be rather small when compared with the
3.3 Canadian Bridge Code (I 988) value for SPT 97 a static analysis procedure. How-
The design requirements given in the Canadian ever, the Florida DOT follows a design process
Bridge Code are quite brief. The soil limit state is where the static analysis is used only for determina-
dealt with in three separate categories. For load tests tion of estimated pile length for bidding. It is inter-
the resistance factor is specified as 0.5 for routine esting that a resistance factor is used in this applica-
testing and 0.6 for high-level testing and the two tion. This is the first presentation of such an
levels of testing are defined. However, the defini- approach and it merits firther study. If data is col-
tions are quite qualitative. If dynamic testing is used lected on the error in the length estimate on each job
the resistance factor is specified as 0.4 for routine it would be possible to adjust the resistance factor to
analysis and 0.5 for analysis based on parameters arrive at the best possible length prediction.
obtained from dynamic field measurements.
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
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 .
313
held in Ghent, Belgium in 1987 (De Jaeger et al. 3.3 Texas 1990
1987). The Belgian Society for Soil Mechanics and
foundation engineering organized this event. The Texas test program was a continuation of the
Altogether, the twenty test piles were constructed by FHWA project (Baker et al. 1993). Altogether, nine
four different methods, five piles each: bored piles were constructed, seven of which had
1. Atlas helicoidal (sawtooth profile) piles, known irregularities. All piles had a nominal
diameter 430/530 mm diameter of 915 mm, with lengths varying between
2. De Waal precast concrete driven piles, 320x320 11 m and 24 m (L/D ratios between 12 and 26).
mm The irregularities were of different character and
3. Socofonda CFA piles, diameter 460 mm magnitude. Four of the piles had a single planned
4. Fundex piles, bored with rotated casing and cast nccking at a dcpth of betwcen 3 and 18 m, the
in situ , diameter 390 rnm reduction in cross section being between 12 and 50
The five testing firms that participated were given percent. Three other piles had both increased and
the pile diameters, and told that the pile lengths decreased cross sections at various depths. In
range between 11 m and 16 m. No defects were addition to the planned defects, some unplanned
knowingly produced, and the task in hand was to defects occurred and were recorded during
determine the correct length. construction.
The best overall length agreements were obtained Five testing firms took part in this competition.
in the Socofonda and Fundex piles. The lengths The testing firms received fbll information regarding
obtained for the Socofonda piles were 93% and the subsurface conditions, as well as the lengths of
100% of the correct lengths. For the Fundex piles, the two reference piles. No m h e r data about the
the spread was between 94% and 102%. Such results existence of defects was divulged.
are perfectly acceptable. On the other hand, poor The tests were conducted in two stages: In stage
results were reported for the precast piles (82% to one, only surface (sonic) methods were used. In
120%) and the Atlas piles (101% to 125% for three stage two, the contestants were allowed to lower
of the piles, with no results at all for the other two). testing equipment into access tubes which were
The testing of the Atlas piles also produced poor prepared beforehand.
results. Although the Atlas piles had the lowest L/D The results of the Texas program may be
ratio (26 to 30), all testers reported lengths that were summarized as follows:
too high. 1. At depths smaller than 7 m below the top of the
The conclusions from this exercise are as follows: pile head, 80 percent or more of the testing firms
1. The precast driven piles were difficult to test managed to identify all important defects in the
both due to higher shaft resistancc and to having cross section.
the highest L/D ratio (41 to 53). 2. The success rate dropped to 60 percent at a depth
2. The testers of the Atlas piles probably neglected of 9 m.
the fact that a helicoidal pile exhibits a wave 3. All participants failed when the defect (necking)
velocity that appears to be much lower than that was located at a depth of 18 m.
of a straight-shafted pile (Vyncke and van 4. Even at shallow depths, participants failed when
Nieuwenburg 1987). the reduction in cross section area was merely 12
3. The CFA piles and the Fundex piles were the percent.
easiest to test. This is probably due to their lower 5 . In general, an enlarged cross-section was more
L/D ratios (32 to 37) as opposed to the precast difficult to fmd than a reduced one.
piles. 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
3.2 California 1990 near the toe of a long pile.
The California test program was carried out in the 7. As expected, no participant could identify a “soft
fiamework of a research project for the Federal bottom” condition.
Highway Administration (Baker et al. 1993). It took
place on two sites: Cupertino, with dry gravelly and 3.4 Delj? 1992
sandy soil and San Jose, with clayey soil under
groundwater. The piles had a nominal diameter of The Delft competition took place in conjunction
915 mm and lengths ofbetween 7.6 m and 18.9 m. with the Fourth International Conference on the
‘l’here were five participants in the program, Application of Stress-Wave Theory to Piles (Smits
applying four testing methods: Sonic echo (time 1996). Twelve laboratories, employing six different
domain), transient dynamic response (fiequency types of instruments, participated in the event. The
domain), cross-hole (ultrasonic) and radioactive testing objects in this case were rather uncommon:
(gamma-gamma). All the participants were provided All of the ten test piles were made fiom precast
in advance with the lengths and shapes of the piles concrete and installed in the following way: First,
that they were to test.
314
closed-end steel tubes were driven to a 7% of the tests confirmed the integrity. In the
predetermined depth. Then, a thin bed of sand was anomalous piles, 82 % of the defects were found.
placed at the bottom of the tubes and the precast These rates improved somewhat in the fmal reports,
piles were lowered into the empty tubes and placed to 25% and 83%, respectively. The success rate for
on the sand-bed. The space around the piles was the defective piles may seem impressive, but a
then filled with a bentonite-cement mixture deeper look into the matter is far less encouraging:
supposed to represent the local soil stiffness. Of all the reported defects, only 36% managed to fit
The test piles had a nominal cross-section 250- the depth of the defects within +20 percent. Of
mm square. Two piles were straight shafied, with these, only 37% provided the size of the defects
respective lengths of 17 m and 18 m. Of the rest, six within -t20 percentage points. On the other hand, the
piles were produced with the cross-section along a participants reported on the average 1.3 “phantom”,
given length either enlarged to 300 mm. square or i.e. nonexistent, defects per pile. None of the
reduced to 200 mm square, or both. The two contestants, or the testing instruments used,
remaining piles had a sawed notch, 10 mm. thick demonstrated a markedly outstanding performance.
and occupying one half of the cross section. Similar
notches were also produced in two of the piles with
enlarged section. 4 EVALUATION
The testing circumstances were also noteworthy,
in two important respects: First, the participants 4.1 Ghent 1987
were not allowed to approach the piles, and the
notary public’s clerk was mobilized to hit the piles The organizers declared that the piles were between
with the hammer. Second, the participants were 11 m and 16 m long. Thus, a tester who would have
given beforehand the exact shapes of all the piles, reported a uniform length of 13.5 m (the mean of the
and their task was to decided which of these shapes above limits) for all the piles would be accurate to
best fits each of the reflectograms they obtained. within f.10% in 90% of the piles.
All the participants managed to achieve was a While highlighting the influence of construction
correct fit for between three and seven piles, with a method on testability, the Ghent competition totally
mean success rate of 44%. The scores for the neglected the most important purpose of sonic
individual piles varied between zero (straight shaft, testing -- finding defects!
L = 18m) to 100% (straight shafi, L= 17m). Publication of the test results was very complete,
and included a special seminar where the results
were presented and discussed.
3.5 Texas 1996
This competition (Samman and O’Neill 1997) took 4.2 California 1990
place on the campus of the University of Houston,
Texas. Altogether, twenty-two piles were tested. Since the organizers of the California testing
Eleven of the piles had a diameter of 460 mm and program gave the participants full details regarding
were bored to a depth of 4.6 m. The other eleven the planned defects, success rates have no meaning.
piles were 760 mm in diameter and 7 m long in the It is therefore questionable whether the California
ground. Some of the piles were constructed with test may qualify as a competition and therefore the
polymer slurry while the rest were cast in the dry. case is not pertinent to the present paper. (The
Six of the piles were regular piles, while sixteen results are of course interesting in other contexts).
piles had planned built-in defects. These defects
were produced from 25 mm thick soft rubber mats, 4.3 Texas 1990
laid horizontally. The mats were placed at different
depths, but not more than one per pile. Each While in California, the participants were given
occupied between 10 and 50 percent of the total information beyond that normally provided to
cross section of the pile. testers, in Texas they got too little. When a pile
Eight laboratories took part: Two were from testing firm is invited to a construction site, it is
government agencies, five were commercial and one customary and necessary to provide it with all
academic. The contestants were asked to report for relevant information, such as soil data, pile
each pile whether it is sound or defective, and, in the construction records and piling logs with the as-
latter case, specify the depth and severity of the made length and details of any irregular events that
defects. The participants were to submit two reports: may have happened. Testing under the “Texas
A preliminary report on the same day, and a final rules“, with no a-priori knowledge of the pile
report within five days. length, is therefore the exception and detracts fi-om
The results, as can be expected, were far from the effectiveness of the results.
satisfactory: In the piles intended to be sound, only In other respects, the planning and execution of
the Texas tests was very effective. The tasks were
315
well graded from easy through difficult to area, it took a lot of good luck to discover any of
impossible. Thus, the performance of contemporary them. The fact that some defects were placed as
systems was well defined. This competition proved close as 300 mm to the top only made things worse.
convincingly that the sonic method is a viable Moreover, a rubber sheet, such as that used to model
technique for investigating pile integrity. It showed cracks in the piles, has a low stiffness when
that sonic equipment is able to identifj most unstressed. In contrast, a rubber sheet stressed by
important defects where they matter most, that is in the weight of the concrete has a considerable larger
the upper part of the pile. On the other hand, it stiffness that does not deviate enough from the
demonstrated that the sonic method is unable to stiffness of the concrete. It would be have been very
distinguish features that are relatively small or hard to detect a reflection from the rubber sheet.
located deep down the pile. Since most competitors were keen to find defects,
and the nature of the defects made them practically
4.4 Deyt 1992 undetectable, the competitors found defects even in
perfectly good piles. This is the main explanation for
As expected, the Delft 1992 event triggered a the participants discovering defects also in the good
lengthy debate in Ground Engineering magazine piles (plain coin tossing would do markedly better!).
(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 5 DOWNHOLE TESTING COMPETITIONS
piles, with an inherent variability of both
concrete quality and shaft resistance soil friction Admittedly, the sonic method has a few basic flaws:
and a rough top surface. Precast piles in an First, the wavelength used is about 3 m, which
artificial “soil” with smooth tops thus cannot do provides rather poor resolution and second, both
not represent real-world life conditions. input (hammer blow) and output (accelerometer
2. The unusually high L/D ratio (72) is generally signal) are remote from potential defects. To
considered to be beyond normal testing limits. overcome these drawbacks, the industry has
3. Most of the anomalies, and especially the saw developed instrumentation that is lowered into the
cuts, were outside the theoretical performance pile through-access tubes+
envelope of the sonic method. Historically, access tubes were first used for
4. Actual testing was performed by inexperienced testing piles with radioactive isotopes. This method
people, not familiar with fine points of the test. is now fast disappearing due to its limited range,
environmental limitations, and regulatory
In all important respects, the Delft competition requirements. It was largely replaced by ultrasonic
did little to advance the state of the art, and in fact instrumentation, using wavelengths in the range of
was a large backward step from the Ghent affair. 50 mm to 100 mm. Modern ultrasonic equipment
With the whole setup being detached from the real (Amir and Amir 1998) combines long range (-3m)
testing world, it had only reinforced the (erroneous) with high resolution. With a suitable setup, it can
belief that sonic testing is not to be taken seriously, also perform tomographical imaging and produce
being based on little more than guesswork. two-dimensional vertical sections. Another
technique, still experimental (Samman and O’Neill
4.5 Texas 1996 1997), utilizes clear plastic tubes and a downhole
video camera.
In view of the poor results obtained in Houston In view of their obvious advantages, downhole
1996, the organizers declared that sonic testing “may testing of piles has become the preferred method in
not be reliable enough to be regarded as a stand- certain sectors such as bridges and high rise
alone measure of the assurance of drilled shafts”. buildings. The time is ripe to organize suitably
Could it be that not the sonic method was to blame, designed competitions which would greatly benefit
but the organization of the competition? In principle, the piling industry.
the Houston competition had the correct ingredients
to simulate a realistic testing assignment. The main
factor that detracted fiom the success of this 6 RULES FOR FUTURE COMPETITIONS
competition lay in the nature of the “defects”: To be
applicable, the sonic method utilizes a wavelength To be effective, competitions must satisfy certain
that is large in comparison with the pile diameter. A minimum criteria. Based on the analysis of five such
defect with a vertical dimension of 25 mm is competitions, The following rules are therefore
therefore well beyond the capability of the sonic suggested:
method unless it occupies all (or almost all) of the 1. The test program should be based on sound
cross section of the pile. Since the defects in theoretical foundations - participants must not
Houston occupied at most one half of the total pile be asked to perform the impossible.
316
2. The tests should be carried out on real piles, REFERENCES
conventionally constructed in real soil. Amir, E.l and Amir J.M (1998) Recent Advances In Ultrasonic
3. The piles should have different lengths and Pile Testing, Proc. 3rd Intl Geotechnical Seminar On Deep
length-to-diameter ratios. Foundation On Bored And Auger Piles, Ghent ; 18 1-185
4. As a rule, organizers should create no more than Baker, C.N., Parikh, G., Briaud, J.L., Drumwright, E.E. and
one anomaly per pile. Mensah, F. (1993): Drilled Shafts for Bridge Foundations,
5. Anomalies should be carefully designed and FHWA, McLean, Virginia
De Jaeger, J., Carpentier, R. and v.d. Broeck, M. (1987):
constructed to resemble, as far as possible, Integrity Tests, Ch. IV in Seminar on Pile Dynamic Testing
anomalies that are actually encountered in (Integrity And Bearing Capacity), Brussels
practice. This includes soil pockets and zones of Ellway, K. (1993): Objectives Of Competition Are Unclear,
weak, honeycombed concrete. Ground Engineering, June; 8
6. The anomalies should be of different Samman, M. M. and O’Neill, M. W. (1997): An Exercise In
Seismic Testing Of Drilled Shafts For Structural Defects,
magnitudes, with an importance ranging between ADSC Foundation Drilling, Dec. - Jan; 1 1- 17
minor irregularities to complete discontinuities. Smits, M. Th. J. H. (1966): Ch. 3 - Pile Integrity Tests,
7. Tests should be carried out by experienced Application of Stress Wave Theory to Piles: Test Results,
personnel, familiar with the testing systems Balkema, Rotterdam;25-54
8. Participants should be provided with normal Stain, R (1993): Test’s Integrity Is Questionable, Ground
Engineering JanuaqdFebruary, p. 7
testing conditions. Stain, R (1 993a): Competition Was Not Applicable, Ground
9. Pile heads should consist of reasonably good- Engineering, April; 15
quality concrete. Testers who desire to improve Steinbach, J. and Vey, E. (1975): Caisson Valuation By Stress-
the surface must be given an opportunity to do Wave Propagation Method, .J. Geotech. Div. ASCE, Vol
--.
qc) 101 No. GT4, April ; 361-378
10. Participants should be provided with sufficient Turner, M. J (1993): Integrity Test Usefulness Is Not The
Issue, Ability is, Ground Engineering, July/August pp. 27-
soil data (borehole logs) and pile data in the 28
manner and to the extent usually provided to Turner, M. J. (1997): Integrity Testing In Piling Practice,
testers on actual construction projects. This CIRIA Report 144, London
includes the as-made lengths and any special Vyncke, J. and van Nieuwenburg, D.(1987): Ch. I1 - Theorie
events observed during construction. Van De Dynamische Proeven: A. Integriteit, Seminar On
11.Participants should not get any data regarding Pile Dynamic Testing Integrity and Bearing Capacity,
Brussels.
the special features installed in the piles. Samman, M.M. and McNeill, M.W. (1997): Fiber-optic
12.In addition to the piles specially prepared for inspection of drilled shafts, Foundation Drilling, Sept.-Oct;
testing, the competitors must be given an 16-19
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.
7 SUMMARY
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.
319
aware of conditions that will potentially damage the on self-regulation and the professional integrity of
pile during installation. The driller must prevent each player in the process, are open to criticism on
collapse of the shaft, and must be sensitive to the this account.
drilling equipment to ensure that a suitable founding The potential disadvantages of this approach are :
layer has been attained. These are all aspects of the e it is an additional cost to the project;
process which are based on knowledge, experience, the liability for any problem may be partially
subjectivity, good practice, and even intuition, which transferred from the contractor to the super-
cannot be defined in a contract specification, or vising body;
measured for compliance. and are therefore outside the effectiveness of this process is dependent
the ambit of a QA System. on the knowledge and experience of the par-
This is not to say that QA Systems do not have a ticular superintendent.
role, or should not be an element of the engineering Particularly for foundation engineering, where
process. Rather, it is important that the limitations many proprietary systems are used, and where the
of these systems be understood and acknowledged in skills of foundation construction are very special-
the formulation of a more global approach. ized, the ability, knowledge and experience of the
What must be avoided is a belief that by virtue of person providing oversight will generally be inferior
implementation of a QA System, that the end prod- to that of the specialist foreman and crew undertak-
uct is necessarily of high quality. In the author’s ex- ing the work. In this case, the supervision is not an
perience, there is a real danger that the application of effective approach for quality assurance.
such techniques can actually reduce the quality of In the particular case of dynamic pile testing, it is
construction. The QA System requires such a pleth- unlikely that a construction supervisor would be able
ora of documentation that time which should be to effectively oversee the testing process.
spent “at the coal face” ensuring real quality is spent
filling in forms to document compliance.
1.3 Quality Assurance - professional review
Dynamic pile testing can, and often is an integral
part of QA systems for driven piles, usually in the Quality assurance can also be applied to the engi-
aspect of confirmatory testing, which will be dis- neering design and construction processes. This can
cussed subsequently. It is also possible and desir- be in the form of either a design check, review or in-
able, that the wider capabilities of dynamic pile dependent design. The independent design would
testing be utilized in setting parameters or guidelines normally be performed by an external organization,
for the installation process, i.e. as an integral part of but checks and reviews could be undertaken both in-
the construction control process. However, this use house or by an independent external engineer. The
is generally underutilized, as it is not as easy to pre- use of professional reviews is common for larger or
scribe this type of application in the System. more complex projects.
It is also noted that dynamic pile testing is also It is implied in this process that the design or con-
applied to only a limited percentage of the piles in- struction records be transparent and verifiable, and
stalled on a contract (usually 15% or less). An in- that independent reviewers with the necessary skills
ference must be made that the quality of the test to interpret this information are available. Further
sample is similar to the quality of the remaining comments on this aspect with regard to dynamic pile
piles installed. testing will be made later.
320
rees. To .the knowledge of the author, this approach load factors and strength reduction factors under the
is not widespread for dynamic pile testing. It is also Limit State approach.
applied to organizations rather than individual test- Nevertheless, there is an important implied as-
ers. sumption that the capacity estimate provided is the
best, unbiased and ‘correct’ estimate available
within the constraints of that technique.
1.5 Quality Asstiva~ce- Direct Testing
A key element of quality assurance in civil engi-
neering projects is the testing of constituents, indi- 2 QUALITY ASSURANCE -THE TESTING
vidual elements or in rare cases, large or complete PROCESS
systems. With specific reference to foundations, the
quality of concrete is routinely tested by standard As noted, direct testing is a key element of quality
compression testing, or chemical testing. Weld assurance in many civil engineering projects. It is a
quality is also evaluated by inspection and ultrasonic particularly common technique in foundation engi-
testing. There are either absolute or statistical stan- neering due to the uncertainties introduced by the
dards with which the tests must comply. Such rou- natural stratigraphy which is not known perfectly in-
tine testing of constituents fits well within Quality advance of any foundation contract.
Assurance Systems or more traditional contractual As verification is a key element of the overall
arrangements, as compliance or non-compliance is construction process, it would be ilIogical if the
easily demonstrated. testing were not itself subject to the principles of
The testing of individual piles is also commonly quality assurance. This section considers quality as-
stipulated as part of a project specification. In this surance of the testing process in the particular con-
case, a variety of techniques can be used. For driven text of static and dynamic pile load testing.
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- 2.1 Static Load Test@
mula such as Janbu, Hiley, ENR or Gates. Although Standards and Specifications which set minimum
these methods are applied to the entire population of requirements for compliance exist as part of the
piles, they are known to be simplistic estimates of framework of quality assurance for testing. For
capacity, and hence subject to high levels of uncer- static load testing, ASTM D 1 143-81(1 994)el
tainty. (ASTM, 1994) is the U.S. National Code of Practice
In many countries, for all but the smallest proj- which sets standards for the equipment, calibration,
ects, more sophisticated, and hence more reliable procedures and records required for testing piles un-
testing is performed in some combination of static der axial static compressive loads. Equivalent
load testing, dynamic pile testing, StatnamicB pile Codes of Practice for static pile load testing are
testing, Osterberg testing and various pile integrity mandated in many other countries.
testing techniques. The particular testing strategy If tests are performed in accordance with the ap-
developed for each site is a function of economics, propriate standard, it is generally assumed that the
available technologies, pile type, site conditions and load-settlement response is a true representation of
historical precedent. the pile load-settlement response. This may not,
Testing of piles as individual components of the however, be entirely correct.
final system also fits well within the framework of a Fellenius (1 984) describes the large errors which
Quality Assurance system in which quantitative out- can be introduced by using a manometer attached to
comes can be measured against the required specifi- the hydraulic jack which is simultaneously a load
cation. application device and a load measurement device.
Engineers generally recognize that despite con- It is more appropriate to use a separate load cell so
ducting a targeted pile testing program, some un- that the load estimate is true and unbiased. AS2 159
certainty still exists for the following reasons: (Standards Australia, 1995) mandates the use of a
e Only a fraction of the contract piles are tested load cell for static pile load testing.
using high-level techniques. The capacity of The complexities of interaction effects between
untested piles must be extrapolated or other- the test-pile and reaction pile, anchors or reaction
wise inferred; weights are usually ignored, and will not be dealt
a Every testing method provides an estimate of with here, other than to note that both the inferred
the (axial) pile capacity at the time of testing, capacity and settlement characteristics can be af-
and each method has an associated uncer- fected. Engineering analysis is required to correct
tainty. these physical effects.
These uncertainties are accommodated by speci- It should also be noted that there are many alter-
fication of ultimate capacities which incorporate an native procedures for static pile load testing. Stan-
appropriate factor of safety or through prescribed dard methods are known variously as Maintained
321
Load Tests, Quick Maintained Load Test and Con- correct interpretation requires specialist geotechnical
stant Rate of Penetration Test (other names are knowledge. However, the following important dif-
used). Within these broad categories, an infinite ferences exist:
number of specific test regimes are possible. Due to The direct output of a dynamic test is not a
the different loading paths. any pile subjected to the load-settlement response, but usually pile-
various tests will exhibit a different load-settlement head strain-time and acceleration-time re-
response. The significance of the different re- sponses;
sponses, the separation of elastic, plastic and creep * The measured test response is a dynamic re-
components, and the extrapolation to service be- sponse, and the static behavior which is to be
havior is a matter for engineering analysis and inter- determined must be extracted from the test
pretation. using either simplistic or more complex ana-
Furthermore, Fellenius (1 980) notes that the in- lytical or numerical techniques.
terpretation of ultimate capacity from a static load- The fact that the direct test outcome bears no resem-
movement curve is not unique. Application of the blance to the load-settlement response is significant.
many constructions proposed (e.g. Davisson Offset This means that most structural or geotechnical en-
Limit , Brinch-Hanson, Chin-Kondner) result in sig- gineers are unable to interpret the test result. To this
nificantly different estimates of ultimate capacity. extent the results are not transparent, and the tech-
It can be appreciated, therefore, that correct inter- nique is therefore considered “black box” technol-
pretation of the simple static load test is more com- ogy - sometimes with the attendant negative conno-
plex than it would first appear. Analysis of the test tations.
data would generally be performed by a specialist Interpretation of these pile-head time records is a
geotechnical engineer using accepted methods that specialized technique, which is generally known or
are in the public domain. understood by the small number of practitioners who
In general. however, the assumptions made in the are providing dynamic pile testing services. This is
analysis of the load test data are transparent (it is, not to say that information on how to interpret dy-
after all only a correction which is applied to the namic pile testing records is not well published and
measured response), and verifiable. The fraternity available in the public domain. However, the reality
of local geotechnical engineers is usually large is that the technique is so specialized that those not
enough to enable a professional review to be made, directly involved simply “leave it to the experts”.
if required. Local professional review is also not generally
Furthermore, as noted previously, in the majority available, as the only potential reviewers are likely
of cases, the static load test response is taken to be a to be testing or construction competitors, and com-
true representation of the pile load-settlement re- mercial and professional sensitivities about releasing
sponse. For contractual purposes, it is often neces- data have even resulted in claims of intellectual
sary only to ensure compliance with the specifica- property over test records. This results in a problem
tion of the peak applied load and the settlement at with verification of results.
one or more defined loads. Further analysis is not Dynamic pile testing suffers from the problem
undertaken, and the test outcome is accessible and that if the process is not typically transparent or veri-
open to direct and immediate interpretation by both fiable, the client is not in a position to independently
structural and geotechnical engineers. 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,
2.2 Dynamic Pile Load Testing given the constraints of the technology.
Just as for static load testing, Standards exist It appears to be a universal experience, judged by
which prescribe the requirements for dynamic pile the author’s personal communications with col-
testing methods. ASTM D4945-96 (ASTM, 1996) leagues around the world, that not all practitioners
sets standards for the equipment, calibration, proce- providing dynamic pile testing services have ade-
dures and records required for testing piles using quate skills. The author is aware of cases where dy-
dynamic impacts. AS2 159 (Standards Australia, namic pile testers have given gross errors in advice
1995) stipulates requirements for the use and inter- due either to poor data quality which is undetected
pretation of dynamic pile testing. The Institution of or ignored, or due to misinterpretation or incorrect
Civil Engineers (ICE, 1996) have published specifi- analysis of the test records. For obvious reasons,
cations m d good practice guidelines on inter alia these cases are not detailed here. The author’s expe-
dynamic pile testing. Other similar documents with riences are not unique.
national authority exist, although fewer than would These errors in advice (if detected by the client)
exist for static load testing. can not only affect the client’s confidence in the
The interpretation of dynamic load tests has both practitioner, but also their view of the reliability of
similarities and differences to the interpretation of the dynamic testing method in general. This has an
static load tests. They are similar to the extent that
322
unfortunate and undeserved flow-on effect to all * Stage 4 : Unconsciously competent. This per-
practitioners. son has become an expert; has an intimate
This is to say nothing about our responsibility as knowledge of the subject area which they can
professionals to provide professional advice, and our apply without effort and can apply their
moral obligation to ensure the integrity and safety of knowledge to areas beyond their direct expe-
the structure and people who could be affected by a rience.
collapse. In developing strategies for quality assurance in
We must collectively address the challenge of the dynamic pile testing industry, all four stages of
quality assurance within the dynamic pile testing in- competency must be catered for and addressed.
dustry world-wide. 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
3 QUALITY ASSURANCE IN THE DYNAMIC the previous descriptions that the Unconsciously In-
PILE TESTING INDUSTRY competent person - unaware of their own limitations
is the person that requires the greatest attention.
The previous section has addressed quality assur- The key to competency is knowledge. Knowl-
ance of the testing process in the context of the edge is typically acquired by one of three methods -
foundation industry and with regard to dynamic pile by formal or informal training, education, mentoring
load testing in particular. or reading; through experience; and finally by mak-
The following section discusses some possible ing mistakes. All are powerful methods of learning.
strategies to ensuring that dynamic pile testing serv- Obviously, in a contractual environment, the last
ices are provided in accordance with the principles method is undesirable, and should be avoided by one
of quality assurance. This discussion is placed in the of the first two methods. Experience, unfortunately
context of current practice. can only be acquired over time. Training is the tra-
ditional way to quickly develop competency.
There are no current standards or guidelines for
3.1 Training training within the industry, and no standards or cer-
A fundamental requirement for ensuring compliance tification which can demonstrate competency. Be-
with the principles of quality assurance is that the cause of the highly specialized nature of this field,
individual providers of advice are suitably compe- the subject matter is not generally covered in an un-
tent. dergraduate engineering degree. Because of the
Four stages have been identified in the general limited interest and again because of the scarcity of
development of competence. These can be applied people qualified to teach in this area, there are no
to the specific case of dynamic pile testing: formal post-graduate courses available in pile dy-
0 Stage 1 : Consciotisly incompetent. The per- namics (to the author’s knowledge). There is, there-
son is not able to perform a task in a compe- fore, no formal qualification or certification which
tent manner, and is aware of their inability. can be obtained which indicates competency in this
This person is not a danger, because he/she field.
will generally be prudent enough not to pro- There is no requirement to either hold a degree or
vide advice. equivalent undergraduate qualification in Civil En-
e Stage 2 : Unconsciously incompetent. The gineering, or in Engineering in general. This is not
person is able to perform tasks at a basic to say that such a qualification is necessary, although
level, but is unaware of what they don’t such a degree provides obvious background to some
know, and the implications of their advice. of the principles employed in dynamic pile testing.
This person is a danger, because he/she will Guided experience, self-help and training can pro-
provide advice without the necessary skills to vide an equivalent level of competency over time.
assess whether this is provided on a sound ba- An individual or organization new to dynamic
sis. pile testing could expect one of the following train-
e Stage 3 : Consciotrsly competent. This person ing experiences:
has reached a stage where they have achieved * An on-site training program of 1 to 3 days,
a basic to advanced understanding. They are generally provided in association with pur-
also aware of what they don’t know and the chase of new equipment. Training is con-
possible implications of their lack of knowl- ducted by the equipment manufacturer, or a
edge. In general, this person will provide ad- designated agent. This training would only be
vice within the limits of their knowledge, and an introduction to use of the equipment and
seek assistance in areas outside their compe- underlying theory. At the conclusion of this
tency. They would only be dangerous when training, the user would be at either Stage 1 or
they do not seek appropriate advice outside Stage 2 competency in the above model
the limit of their expertise.
323
0 An informal training program conducted on- review from a tester with either advanced or
the-job and in-house by a tester from the same expert status.
organization. The trainer in this case may
have been given the original training by the 3.2.1 Evaltiation of competency
equipment manufacturer, or may have been Assessment of competency could be formulated
given similar informal in-house training. In in different ways. Some of these proposals may be
this way, 2'Id or 3rdgeneration training is quite difficult to develop in practice:
common. There are no controls on the quality By a mentoring scheme in which assessors
of such training. It would typically concen- (accredited by an independent industry group)
trate more on the particular types of projects would evaluate the competency of an appli-
encountered, and practical aspects of testing, cant over a period of time. Both field and
and would give less broad overview, princi- analysis skills could be assessed.;
ples and theory than the manufacturer's By an independent review panel (comprised
training program. Manuals will generally be of acknowledged industry experts) that could
made available to supplement the training. assess a submission from an applicant based
The quality of the training will only be as on examples of the applicant's work; a list of
good as the competency of the trainer. Any projects completed; details of training under-
misconceptions of the 2'Id or 3rd generation taken and referee reports. An interview proc-
trainer will be passed on to the new trainee. ess could also be part of such a scheme;
This.current ad hoc method of training is not an By standard examination. This could cover
ideal, and is not consistent with a quality assurance both data acquisition and data interpretation
approach. skills, but could not assess practical skills on
In order to comply with quality assurance princi- site. A multiple-choice format would give a
ples in training, it would be necessary to provide a most objective assessment of capability
consistent minimum standard of training to all new Any method used to assess competency should be
practitioners. If the training is to be industry-wide, it able to effectively distinguish three levels of com-
would be desirable to provide this through an inde- petency - basic, advanced and expert. All ap-
pendent industry organization or educational institu- proaches should be capable of giving feedback to
tion. The practicalities of providing uniform train- applicants on areas of weakness so that targeted
ing to a dispersed world-wide market would tend to training and improvement is encouraged.
suggest the need for the material to be delivered in As skills in dynamic pile testing need to be rein-
distance education mode. Concentrated workshops forced by regular practice, it may be necessary to in-
could be an alternative model. 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.
3.2 Testing Conipetency
The purpose of training should be to establish basic
3.3 Review
competency in the providers of testing advice so that
at the completion of the training period they are Having established a system of assessing compe-
moving into Stage3 - Consciously Competent. Of tency, it is a logical consequence to ensure that those
course, this stage will not be fully reached without a practitioners assessed to have at most basic compe-
period of field experience to reinforce the principles tency be required to obtain review from others with
learned during training. either advanced or expert status.
With further experience, discussion with peers, The ideal situation would be for such partnering
attendance at industry seminars and conferences, to be undertaken in-house with someone with higher
there would be an expectation that the tester would competency. This person would be required to "sign
progress to advanced and eventually expert status. off' on all testing and analysis undertaken by their
Under a quality assurance philosophy. there junior.
would be a need not only to provide and undertake Where a person of higher qualification was not
training, but to assess competency both of those that available within the organization, arrangements
undertake training, and those that are already pro- would be required for review by any eligible re-
viding dynamic pile testing services viewer. The arrangements for such review, includ-
e It needs to be established that testers have ing commercial and legal aspects, would be a matter
achieved and maintain a basic level of com- for joint agreement. It is anticipated that standing
petency, and arrangements would be made with a particular re-
0 Those testers moving from Stage 2 to Stage 3 viewer to streamline the process. This would lead to
(hence from a basic level to an advanced an effective mentoring arrangement and encourage
level) should receive an appropriate level of transfer of knowledge.
3.4 Preqzral~fic~rtion The principle of proportionality; pile imped-
ance;
The relationship between mass density,
As noted earlier, prequalification is a quality as-
modulus and wavespeed;
surance technique which is used for contracts re-
Material properties;
quiring high levels of expertise or resources. Some
Allowable stresses;
statutory authorities already require prequalification
Principles of one-dimensional wave mechan-
for providers of dynamic pile testing services.
ics;
It is suggested that this system could be more
Computation of pile wavespeed in easy and
widely adopted, and that acceptance be based on the
hard driving conditions;
levels of competency assessed by the evaluation
Recognition of easy, moderate and hard driv-
process. The requirements for review, as suggested
ing conditions;
in the previous section could be formalized in the
Identification of high compression stress lev-
conditions for prequalification. It would be impor-
els at the pile head and at the pile toe;
tant that providers not be excluded from providing
Estimation of pile toe stress;
services, but rather that their provision of services
Identification of high tension stresses before
should be accepted conditional on demonstration of
and after the 3L/c time;
effective and timely review arrangements.
Identification of damage intensity and loca-
tion;
Recognition of a broken pile, pile joints and
4 BASIC SKILLS
changes in section;
The continuing development of dynamic pile testing Recognition of response from end-bearing
equipment has enabled the operator to have a vast piles, and piles with small and large shaft re-
array of information available in real-time. Algo- sistance;
rithms have been developed which provide critical Estimation of end bearing, shaft resistance,
feedback to the operator at critical stages of testing and shaft resistance distribution;
alerting the operator to potential hazards (such as Recognizing high and low cushion stiffness;
development of damage, excessive stresses, exces- Recognizing bending and poor hammer-pile
sive bending etc.) Many calculations which were alignment;
previously undertaken manually are now automated, Understanding the effect of resistance, or im-
freeing the operator to concentrate on more critical pedance changes close to the transducers;
observations. Principles of valid data adjustment;
Although these developments are positive, there Pile compression and tension capacity;
is a danger that the operator will lose the ability to Estimating Case damping factor;
critically evaluate the results which are provided by Basis of the Case method of capacity deter-
the equipment. No algorithm is flawless, and no mination, and factors influencing capacity es-
equipment, however smart the software, can repli- timates;
cate the abilities of an expert to critically evaluate The implications and conditions for pile un-
the pile responses and make appropriate judgements loading;
and decisions. Without a detailed understanding of Pile set-up and relaxation;
the basis for the computations, the limitations of Mobilization and under-mobilization of ca-
these computations, and the implications of changes pacity;
in parameters or assumptions cannot be known. Hammer performance and transfer efficiency;
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 5 CONCLUSIONS
skills to make the necessary independent critical
judgements. Only by developing these skills can he This paper argues the need for dynamic pile test-
progress to advanced and then expert level. ing to be subjected to a rigorous industry-wide qual-
Following is a tentative list of the types of skills ity assurance scheme, compatible with the founda-
necessary for effective dynamic pile testing. Many tion industry which the testing serves. The scheme
of these relate to recognition of conditions from the should be one which delivers real, fundamental
forceivelocity and upward/downward wave re- quality, and not one which merely produces docu-
sponses: mentation to satisfy administrative requirements for
o Recognition of valid and invalid data in a va- an audit trail.
riety of scenarios; Ideas for a possible comprehensive model have
8 Field measures required to rectify poor data; been suggested. The challenge to adopt such an ap-
proach should be taken up by all players in the field
325
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
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.
327
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
Figure 1. Windows PDA processing program. summary, during the test pile program considerable
time is spent simply waiting. The active time for the
PDA engineer on-site is often small.
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 Figure 2. PAL on site (left) with cell phone (right)
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 3 NEW METHOD FOR DYNAMIC PILE
and especially sensor attachment are often done by TESTING
the pile driving crew. The time required to drill the
holes and attach the sensors is often only 5 to 15 A new version Pile Driving Analyzer (PDA) called
minutes per pile tested (substantially less effort than PAL has many powefil features to revolutionize
a static load test). dynamic pile testing. This smaller unit has a
The engineer first evaluates the data for quality. touchscreen for user friendly data entry and a self
The PDA program checks for loose connections, contained rechargeable battery sufficient for a full
unstable results, various ratios including the known day’s operation. Data is stored on a removable
theoretical proportionality between force and memory card. This PDA can be connected by the
velocity at impact. If problems are detected, a engineer on-site directly to a laptop running the
warning is given, and corrective action can be taken. Windows PDA program to provide full on-site PDA
Once the data quality is satisfactory, the engineer processing capability.
interprets the data on-site using the PDA. He gives However, modern technology now allows for a
advice or opinions and answers questions that more cost effective approach to dynamic pile testing.
prompted the testing request. Following the on-site The PAL on-site can be remotely operated through
test, the engineer may further analyze the PDA data cell phone technology by an engineer in the office,
with CAPWAP, a signal matching computer see Figure 2. The sensors are attached to the pile by
algorithm which extracts the soil model from the the crew on-site or a trained technician. The pile
measurements and provides a simulated static load sensors are connected to the PAL, which dials the
test result. cell phone and connects to the office PDA engineer.
From this point on, the PDA engineer operating the pile installation productivity is improved. Maximum
PDA-W program on his office PC controls the driving stresses are investigated to reduce the
remote PAL to collect the data and send it digitally likelihood of pile damage. The PDA gives the direct
to the office. The office engineer sees the data from compressive maximum at the sensor location. Since
the on-site remote PAL in real time as the test is in up to four strain sensors are attached to the pile and
progress. Site observations are communicated with monitored separately, bending and local contact
the office engineer with a cell phone or by the PAL’S stresses can be assessed and hammer pile alignment
“message communicator” (which sends short improved. The maximum tension stress below the
messages either preprogrammed or user sensors for concrete piles is computed from the pile
constructed). In turn the office engineer interprets top measurements. If tension is excessive, a lower
the PAL data and communicates his advice to the hammer stroke, or increased pile cushion thickness
sit: pr.ctic.!!y jnstantaneously. may be required. The maximum compression stress
Why test in this remote mode? Simply stated, is estimated at the pile bottom for all piles driven to
there are significant cost savings. In many cases, the hard bearing layers. These stresses need little
cost to the owner of the engineer’s travel time to site interpretation other than comparison with stress
and the travel expenses approaches the cost of the limits imposed by code or specification. The PDA
test itself. Since the pile crew can attach the sensors can investigate the pile shaft for damage.
to the pile, the PDA engineer could operate the PDA Evaluation of pile bearing capacity is more
from his office and save this time and travel cost. complicated because the soil strength is often altered
With reduced costs, the project owner can perform by the installation process. The capacity during
more testing for the same total cost. The PDA driving is often less than the long term pile capacity
engineer does not waste time waiting on-site for the particularly for piles driven in fine grained soils
test. But perhaps more importantly, the data analysis (clays, silts and even fine sands) due to excess
can begin immediately after data collection resulting positive pore pressures generated during driving
in earlier availability of final results to speed up the which reduce the effective stresses. As these pore
construction and decision process. The report can be pressures dissipate after driving ends, the pile shaft
issued sooner. On large projects, the PDA engineer resistance increases. Capacity reduction during
may still travel to the site for the initial test piles. driving is also caused by lateral pile motions which
However, subsequent production pile quality create an oversized hole; with time, the overburden
assurance testing can be performed remotely. This pressures reach equilibrium on the pile perimeter
results in lower testing costs and easier scheduling of and increase shaft resistance. This phenomena of
routine periodic tests. Thus, more tests can be capacity gain with time is called soil setup.
performed at lower costs resulting in improved Therefore, dynamic testing during restrike tests after
quality assurance for the project. 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
4 DYNAMIC TESTING USES AND GOALS as the soil grains become finer.
Although less common, relaxation (capacity
Obviously, whether testing on-site or remotely, the reduction with time) has been observed. Relaxation
dynamic test must be performed with the correct can be a serious problem for piles driven into
goals to achieve the maximum benefit. Objectives weathered shale, and may take several days to h l l y
of high strain dynamic pile monitoring during develop. These losses can be caused by exposing the
installation include evaluation of the energy transfer shale to water, due to fracturing of the rock or by
and driving stresses to assure the pile can be safely heave from driving adjacent piles. Pile capacity
driven to the desired depth without damage. estimates based upon initial driving can significantly
Capacity may be of interest during driving to over predict long term pile capacity. Therefore, piles
establish the bearing layer or establish the driving driven into shale should be tested after a minimum
criteria. However, for long term capacity evaluation, one week wait either statically or dynamically (with
restrike data is usually required due to soil strength particular emphasis on the first blows). Relaxation
changes with time. has been observed for displacement piles driven into
During driving, the PDA calculates energy transfer dense saturated silts or fine sands due to a negative
from the integral of the product of force times pore pressure effect at the pile toe. Restrike tests
velocity. The energy transferred is then compared after a few days are usually sufficient, with emphasis
with the hammer’s rated energy. When low hammer on early “high energy” blows. With sufficient
performance is indicated, the hammer can be experience on a site or in some limited geographical
serviced (cushion, compressor and hoses, air valves, area, the engineer may eventually reduce the amount
fuel pumps, or piston rings inspected or corrected, of restrike tests and just apply a reduction factor to
etc.) or the cause of low performance determined the end of drive capacity to estimate the final
and corrected so that hammer performance and thus resistance available.
329
Dynamic load testing indicates the activated or results where negative friction (downdrag) may be a
mobilized pile capacity at the time of testing. At concern.
very high blow counts (above about 10 blows per Numerous factors are usually considered in pile
inch, or less than 2.5 mm set per blow), dynamic test foundation design. Some of these considerations
methods tend to produce lower bound capacity include additional pile loading from downdrag or
estimates as not all resistance (particularly at and negative skin friction, soil setup and relaxation
near the toe) is fully activated. This can occur when effects, cyclic loading performance, lateral and uplift
piles are driven to rehsal or in cases where setup loading requirements, effective stress changes (due
increases are large and the hammer is then not to changes in water table, excavations, fills or other
capable of moving the pile during testing. Several changes in overburden), settlement from underlying
solutions to overcome this under prediction dilemma weak layers and pile group effects. These factors
for refbsal conditions are available depending on the merit consideration when considering the
site conditions and availability of equipment. One interpretation of dynamic testing results. The
solution is to apply a few blows at higher energy. foundation engineer should determine if any of these
The higher energy can be generated from a higher considerations apply to his design.
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 5 USAPRACTICE
per blow), and thus mobilize the full capacity.
Another solution for closed end steel pipe piles at The USA is interesting since practically every
refusal is to fill the pile with concrete and then test possible pile-hammer-soil combination can be
as a composite section. The now stiffer pile will encountered. Most PDA testing in the USA has been
have a higher applied force and thus overcome more on driven piles with the testing engineer on-site with
resistance. Another method is to add the end of the PDA. A considerable experience has been
drive end bearing to the restrike shaft resistance by accumulated. Early tests from the research project at
superposition to estimate the service condition total Case included extensive correlations with static load
load (Hussein 2000); this should only be done if the tests. Private consultants began in 1972 to apply the
restrike is at rehsal conditions and where no toe method in their own private sector projects, either
relaxation is possible. Alternately, if setup (and testing themselves or specifling testing in project
relaxation) can be estimated, the pile need not be documents (often resulting in the contractor hiring a
driven to the full required ultimate capacity but testing firm to provide the service). Based on
rather only to a usually reduced “target” capacity or favorable experience and the results of the research
to refbsal. project, highway agencies specified dynamic testing.
A CAPWAP analysis usually is made to confirm With time and the support of the Federal Highway
the PDA field capacity result. The soil is modeled Administration for this testing, more highway
similar to wave equation methods. The wave agencies have included dynamic testing in their
transmission is modeled by the method of practice, either by acquiring the equipment or by
characteristics (CAPWAP Manual 1999). The hiring outside consultants. As a result of this varied
hammer model is replaced by the measured force exposure, many contractors see a benefit and now
and velocity as a boundary condition. Since these hire testing firms to assist them when piling
measurements are redundant, the soil model can be problems arise, or they often propose dynamic
iteratively investigated. If the measured velocity is testing as an alternate when static testing is
input into the CAPWAP analysis model, then the specified.
force required to hold the system in dynamic For small projects with only a few piles, a couple
equilibrium can be computed and subsequently piles per structure are tested. Since the time to
compared with the measured force (usually the wave install all piles is relatively short, often the piles are
down is input and the wave up is computed and tested during driving or with relative short wait
compared). The soil model is adjusted either times during restrike. For medium sized projects,
automatically by the program or by the engineer the first production piles often serve as dynamic test
until the computed and measured forces (or wave piles and are distributed over the site to check site
up) are in agreement. The final soil model variability. Usually some restrikes are included in
(distribution and dynamic and static parameters) the test program. In a growing number of cases,
then describes the soil behavior during the hammer longer wait times before restriking are being
impact. CAPWAP has been proven to have good employed to take more advantage of the ususal
agreement with static load test results (Likins et al. strength gains with time. For larger projects, the
1996). The CAPWAP pile and soil models can be amount of static testing is generally reduced after
subjected to a simulated static loading to produce a establishing a correlation of dynamically and
load movement curve comparable to static test. statically tested piles, and then supplemented by
CAPWAP also provides resistance distribution additional dynamic tests to increase the total
330
percentage of piles tested. This improves the overall Engineers (ASCE) developed a consensus standard
quality assurance while reducing testing costs. for pile testing entitled “Standard Guidelines for the
Because of the large amount of dynamic testing Design and Installation of Pile Foundations”(ASCE
performed, various codes and specifications are now 20-96) which includes dynamic pile testing. The
in place for USA application. In 1986, the D4945 private sector work generally has followed the local
consensus standard was adopted by ASTM highway department practice and the three regional
(American Society of Testing and Materials) for building codes. A recent joint effort of these three
“High Strain Dynamic Testing of Piles.” Testing in regional code authorities resulted in the International
the USA follows these guidelines (ASTM D4945-96 Building Code (IBC) with application in April 2000.
is most recent revision). IBC allows ASTM D4945 testing to evaluate static
Beginning in the 19803, the Federal Highway capacity. Before IBC adoption, many consultants
Administration offered a demonstration project for employed dynamic testing on their private projects
dynamic pile testing. They currently provide for either capacity evaluation or monitoring hammer
workshops for State highway agencies which performance and driving stresses
includes dynamic methods as an important part of Two USA organizations (Deep poundations
their recommendations. Many State highway Institute and Pile Driving Contractors Association)
agencies have developed their own specifications for have endorsed PDA testing. DFI produced a
use. Florida DOT has probably the most extensive consensus document “Inspector’s Manual for Driven
code and PDA is an integral part of their quality Pile Foundations (Second Edition 1997)” with
control procedures. AASHTO has general positive comments on dynamic testing. PDCA has
specification T298 (similar to but improved version promoted a model code (dated 1999) in both
of ASTM D4945). In additional they recognize working stress and LRFD versions which allows
dynamic testing with its own reliability and benefit substitution of static tests by dynamic tests aRer a
when assigning safety factors to foundations in the proven correlation with considerable reductions in
“AASHTO Standard Specifications for Highway safety factor and thus cost savings for increased
Bridges”. In general, higher levels of testing result percentages of piles tested.
in a lower safety factor. It is conservatively estimated that every year
The US Army Corps of Engineers included several thousands of piling projects have dynamic
dynamic testing in their 1993 manual on “Design of pile testing on driven piles. While occasional PDA
Pile Foundations”. The American Society of Civil 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
Figure 3. Drop weight test on augercast pile. least 1 to 1.5% of the desired ultimate capacity to be
331
proven (Hussein, 1996) to assure load activation at economy. In some countries the ultimate pile
reasonable stresses. Larger existing weights can be capacity has been greatly increased as a result of
used provided the weight and shaft diameters remain more dynamic testing; for example, in Sweden the
about comparable. Regardless of size, shape or allowable loads for the same identical piles have
composition, the drop weight is generally guided to approximately doubled as a result of codes requiring
an axial impact by a short set of leads as in Figure 3, higher percentages of piles to be tested by PDA.
and is raised by cable and dropped by releasing the There are numerous country specific codes detailing
drum brake. An alternative and preferable drop application of dynamic PDA testing (Beim et al.
method involves raising and securing the weight and 1998). Monitoring hammer performance for
then completely releasing it for a true free drop (e.g. offshore oil platform installation is regularly
releasing hydraulic jaws such as used for vibratory specified by the oil companies; if driving is greatly
hammers used to grab steel piles, or by tripping a different than expected, then capacity is further
simple mechanical release). evaluated.
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 7 CONCLUSIONS
weight with the shaft. After each impact, the net
permanent displacement or “set per blow” is Dynamic pile testing with the PDA with subsequent
carefdly measured to evaluate full capacity CAPWAP analysis has become a routine practice for
activation. Compressive stresses are compared with engineers and contractors worldwide. The methods
the concrete strength. Alignment adjustments are have been applied to driven piles during driving to
made if necessary and a second higher drop height is monitor the hammer and driving stresses. Because
applied. The test continues with increasing drop pile capacity is a function of time due to changes in
heights until either the set per blow exceeds a value soil strength due to effects of the driving process,
sufficient to insure the full capacity activation, or long term capacity is usually evaluated during
until the indicated capacity is above the required restrike several days after initial installation. In
ultimate capacity, or until the stresses become too many parts of the world, dynamic pile testing
large and the risk of pile damage is then too high. methods have been successfully applied to drilled
Most tests are completed in less than five impacts. shafts and augercast piles by applying an impact of a
If the pile top has been built up to accommodate the large drop weight. To activate the full soil resistance
dynamic test, the extra top section is removed to (and thus correlate best with a static test failure load)
facilitate completing the foundation. for either driven or drilled piles, the energy input
The measured pile top strain and velocity data are must be sufficiently large to produce a 2.5 mm set
analyzed by CAPWAP to independently check the per blow or more. Numerous codes and
total capacity mobilized for each blow. A CAPWAP specifications now direct the proper application of
analysis can be performed in a short time on site dynamic testing. New technology has been
after each impact to determine if the set per blow is introduced to allow the engineer in the office to
low so that the full capacity has not yet been remotely monitor dynamic pile tests on site.
activated and another larger impact is required.
Upon completing the CAPWAP analysis, a
simulated static load test is obtained. REFERENCES
332
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.
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.
335
test methods. The results of these research activities agreed to form the Research Committee on Rapid
are to be presented in two papers at this Conference. Load Test Methods in 1996 to critically review the
The first Statiiamic tests in Japan were performed features of rapid load tests, including the Statnamic
in 1992 on a cast-in-situ concrete pile (Chosokabe et test, in more general teims, and to prepare a draft
al. 1993) and on a steel pipe pile (Matsumoto et al. document of testing standards for rapid load tests. In
1994). In 1993, a private joint research group 1998, the Committee on Standardization for Vertical
(Research Group on Rapid Pile Load Test Methods) Load Test of Piles was formed in JGS for the
was formed led by Prof. Osamu Kusakabe of Tokyo revision of the current standards for vertical load
Institute of Technology, aiming to compile the tests of piles, with the scheduled enforcement of the
existing knowledge about rapid pile load tests and revised standards in the year 2000. The revised
examine basic characteristics and applicability of the standards will iiicludes dynamic load tests and rapid
tests, as well as the scientific interpretations of the load tests, as well as various types of static load tests.
Statnamic test results. The Research Group The Research Group and the Canadian Embassy in
coiiducted their own research program, in which Japan co-organized the second International
several factors which might influence the Statiiamic Statiiamic Seminar in Tokyo in 1998. More than 40
test results were investigated in the field, such as articles were submitted to the Seminar from 11
type of pile, the installation method, the loading countries, showing the large increase in the use of
history, the curing period and the soil conditions rapid load test in the world. A technical session on
(Kusakabe & Matsuiiioto 1995; Kato et al. 1998). the Global Srctndardization of Statnamic Testing was
Based on the accumulated research results and the held during the Seminar, aiming to have a global
continuing efforts of the Research Group, JGS 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 in terms of loading and soil conditions, in order to
concerning pile load testing in Japan. In the next select the appropriate interpretation method for each
section, we turn to the theoretical definitions of pile load test.
dynamic and rapid load testing used in Japan. 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
3 DEFINITIONS OF DYNAMIC LOAD TEST relative duration, t,., based on the works of Holeymaii
AND RAPID LOAD TEST (1993), Middendorp & Rielefeld ( I 995), Karkee et
al. (1997), Nishimura et al. (1998) and Matsurnoto
When the dynamic load test and the Statnamic test (1998). The relative duration, t,., is an index for the
are used to derive a static load-displacement curve loading duration relative to the traveling time for a
for a pile, allowance for dynamic effects should be stress-wave up and down in the pile, and is defined
made in the interpretation of the test signals. The as follows (Karkee et al. 1997):
difference between the dynamic theory and the static
theory is whether inertial forces are considered or I,. = T 1 ( 2 L l c )
not, in the sense of physics. However, in the field of
geotechnical engineering, 'dynamic effects' also in which T is the loading duration as defined in
include such things as (1) the strain rate dependency Figure 2, L is the pile length and c is the bar wave
of soil strength and deformation characteristics. velocity of the pile. The relative duration, I,, is
including creep, (2) the viscous effect on interface identical to the relative wave length. A, defined as
strength between the pile and the soil, arid (3) (IHoleyman 1992) :
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 Each bar in Figure 1 indicates the range of relative
testing, Statnainic testing, and even in static load duration for which the corresponding dynamic effect
testing. The degree of influence of each dynamic is not negligible.
effect depends on the penetration rate, the The range of stress wave phenomena in the pile
acceleration of the pile, and the loading duration. which are negligible was determined from the works
Several interpretation methods are available for the of Middendorp & Bielefeld (1 995) and Nishimura et
dynamic load test and the Statnainic test, such as al. (1998). They carried out simulations of one-
applications of finite element method (FEM), one- dimensional stress-wave propagation in piles and
dimensional stress-wave theory and single rigid inass found the relative duration beyond which wave
modeling of the pile. Of course, these interpretation propagation phenoinena are negligible. The range of
methods have limitations in their application in the influence of the generation of excess pore
practice, accordiiig to the purpose and the loading pressure is based on Holeynian (1 992) which
condition, as well as the soil conditions. Therefore. suggested the relative wave length, A, necessary for
it may be necessary to classify pile load test methods 90% of excess pore pressure to dissipate for various
337
soil types. The range of viscous damping (the loading, for concrete and steel piles. Typical loading
influence of the relative velocity between the pile 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.
338
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.
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 various types of piles in various soil conditions with
capacities, Qrl and Q.$,obtained from the static load various construction methods with the load-
test, tlie dynamic load test and various pile design displacement curves, P vs w, of the piles obtained
codes, for a steel pipe pile ( l l n i in length, 0.8m in from the static load tests. Let the discrepancy
diameter, 12mm in wall thickness) driven in a between Fsl,zvs w and P vs w be called the 'dynamic
mudstone ground (Michi et al. 1996). The dynamic effect' in the Statnaniic test, here. Based oil these
load test and tlie static load test were conducted 65 comparisons, they showed that the dynamic effect
hours and 29 days after tlie end of initial pile driving, depends on 1) the acceleration of tlie pile, 2) the
respectively. penetration rate of the pile, 3) the level of FIo2 against
Figure 8 shows similar comparison for a steel pipe the static yield load, PJ,, 4) the proportion of shaft
pile driven in a sandy ground (Shibata et al. 2000). It and toe resistances, 5 ) the stiffness of tlie pile shaft.
can be seen from Figures 7 and 8 that the bearing 6) the soil type, and 7) the drainage condition. They
capacity and the Qp/Q5 ratio derived from the surnniarized the 'dynamic effect' in tlie Statnainic test
dynamic load test are the most accurate, compared to qualitatively as follows:
the empirical pile design formulas available in Japan (1) the dynamic effect is small when F,,,, is below P,,,
and other countries. (3)the dynamic effect is larger for the toe resistance
There are cases where conducting re-driving tests than for the shaft resistance,
is difficult due to practical considerations. In such (3) the dynamic effect is larger when the ground is
cases, calibration of the results of tlie initial dynamic saturated, especially for the toe resistance, and
load test on a pile with the static load test at the site (4) the dynamic effect becomes larger as tlie pile
is required to roughly estimate from the initial penetration rate increases.
dynamic load tests the bearing capacity of tlie other And, they concluded that interpretation of tlie
piles after the completion of the set-up phenomena. 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
JHA: Japan Road Associatio~l JAC: Architectural Institute of Japaii
curve derived from the Unloading Point method is
JR: Japan Railway JPH: Japan Port and I-larbor Association
thought to closely correspond with the load-
displacement curve obtained from the quick load
Hiley ( I ) : Hiley's formula with use of potential hammer energy
Hiley (2): Hiley's formula with use of transfered hammer energy
maintained load test. FEM analysis is regarded as a
promising interpretation method for the Statnaniic
Figure 8. Comparison of end and shaft capacities obtained
test, because elastic-(visco)plastic soil behavior and
from static load test with those estimated from dynamic Load complicated boundary conditions are incorporated in
testing and derived from various pile design codes for steel pipe the analysis.
pile driven in sandy ground (Shibata et al. 2000). Although there still is room to improve tlie
existing interpretation methods of the Statnamic test,
the ineasurements of additional parameters such as
5 STATNAMIC LOAD TESTING IN JAPAN axial forces down the pile shaft and the movement
(or acceleration) of the pile toe are also useful to
Here, we will discuss the interpretation of the make the interpretation of the Statnamic test more
Statnaiiiic test signals to estimate the static load- re1i ab le.
displacement curve of a pile
The Statnamic tests which have been conducted in
Japan from 1992 are summarized by Kato et al. 6 CONCLUDING REMARKS
(1998) and Nishimura et al. (2000). Kato et al.
( 1 998) compared tlie pile head force-displacement It seeins that the Statnamic test is increasingly being
curves. FA/,,vs M', measured in the Statnamic tests on viewed as an alternative to the conventional static
34 1
load test in Japan. It is not uncommon that the Proc. of 3rd Int. Cot$ of Application ofStress- W a v e Theory
dynamic load test results are calibrated with the to Piles, The Hague, Netherlands: 195-2 15.
JASPP, 1995. Methods for prediction of bearing capacity of
Statnamic test results. However, a large part of the steel pipe piles by dynamic load testing. Report of
Statnamic tests as well as the dynamic load tests are Committee on Methods for Prediction of Bearing Capacity
used only for the acceptance of the piles. It is (in Japanese).
Kato, K., Kusakabe, 0. & Matsumoto, T., 1998.
undeniable that pile designers in Japan stick to pile Characteristics of rapid load test appeared in various piles.
design codes excessively. Jozw. qf Geotech. Eng., JSCE, No.589/111-42: 155-166.
The fact that standards for the dynamic load test Kato, K., Horikoshi, K., Matsumoto, T. & Kusakabe, O., 1999.
Interpretation methods of Statnainic load test results. Jour.
and the Statnamic test (rapid load test) have not been oj'Geotech. Eng., JSCE, No.624iIII-47: 267-282.
available in Japan also is an obstacle to wide uses of Karkee, M., Horiguchi, T. & Kishida, H., 1997. Static and
the new pile load test methods. Another aspect to be dynamic tests for evaluation of the vertical load bearing
considered is pile design codes. There is no capacity of piles. Proc. Deep Foundufions Institute 22titl
Annual Member 's Coi$erence, Toronto, Canada: 199-2 14.
reduction in the safety factor in response to an Kusakabe, 0. & Matsumoto, T., 1995. Statnamic tests of
increase in the number of load tests beyond 1. Shonan test program with review of signal interpretation,
Furthermore, pile design is performed based on the Proc. First Int. Statnaniic Seminar, Vancouver: 1 13- 122.
bearing capacity alone, since vertical deformation of Kusakabe, O., 1998. Changing foundation design code and the
role of Statnainic test, Proc. 2nd Int. Statnainic Seminar,
pile foundations is not specified explicitly in the Tokyo, Japan. Balkeina (to be published in 2000).
design codes. However, this situation is currently Koizuini, Y. and Ito, K., 1967. Field tests with regard to pile
improving. Actually, the JGS standards for vertical driving and bearing capacity of piled foundations. Soils and
Founckrtion, JSSMFE, V01.7, No.3: 30-53.
pile load tests are to be revised this year to include Ladanyi, B., 1963. Expansion of a cavity in a saturated clay
standards for rapid load testing and dynamic load medium. ./our. Soil Mech. Found. Div., ASCE, vol.89,
testing. Furthermore, the Research Committee on No.4: 127- 16 1.
Lundberg, B., 1984. On the separation of overlapping elastic
Present and Future of Japanese Foundation Design waves from measured data, Proc. 2nd Int. Conj. on
and Soil Investigation in view of International Application qf Sfres.s-M/uve Theory to Piles, Stockholm:
Equivalency was formed in JGS in 1997, aiming to 44 1-452.
change foundation design codes from the allowable Matsuinoto, T. & Takei, M., 1991. Effects of soil plug on
behaviour of driven pipe piles, Soils and Foundutions,
stress design to the framework of the limit states V01.3 I , N0.2114-34.
design and/or the performance based design. Matsuinoto, T., Sekiguchi, H., Yoshida, H., & Kita, K., 1992.
Quick, cheap and accurate pile load tests such as Significance of Two-Point Strain Measurements in SPT.
Soils and Foundations, Vo1.32, No.2: 67-82.
the dynamic load test and the Statnamic test are Matsumoto, T., Tsuzuki, M. & Michi, Y . , 1994. Comparative
expected to play important roles in the trend of' study of static loading test and Statnaniic on a steel pipe pile
changing foundation codes in which accurate driven in a soft rock. Proc. Int. Con( and Exhibition 017
Piling mid Deep Foundations, Bruge, Belgium: 5.3.1-5.3.7.
estimation of the load-displacement curve as well as Matsuinoto, T., Michi, Y. & Hirano, T., 1995. Perforinance of
quality control of pile foundations are vital issues. 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.
REFERENCES 14th ICSMFE, Hamburg, Vo1.2, pp. 1 185-1 188.
Michi, Y., Matsumoto, T. & Futatsuka, Y., 1996. Reliability of
Chosokabe, M., Yamashita, K., Kakurai, M., Fukuhara, T. and dynamic load testing compared with soil parameters: A case
Yamada, T., 1993. A Statiiainic loading test applied for a study on foundation piles of Noetsu bridge. Proc. 5th hit.
cast-in-situ concrete pile. Proc. Annual Meeting of' Coiif'. 011 Application of Stress-Wave Theory to Piles,
Architectural Institute of ,Japan: 176 1- 1762 (in Japanese). Orlando, USA: 465 - 479.
Decks, A.J. & Randolph, M.F., 1995. A simple model for Middendorp, P., Bermingham, P. & Kuiper, B., 1992.
inelastic footing response to transient loading. /tit. </our. Statnamic loading testing of foundation piles. Proc. Of'dt/?
iVzon Anal. Methods in Geoniech., Vol. 19: 307-329. bit. Conf of Applicufion of Stress- W a v e Theory to Piles,
Fellenius, B.H., 1988. Variation of CAPWAP results as a The Hague, Netherlands: 58 1-588.
fiinctioii of the operator. Proc. 3rd lnt. Conj.. on the Middendorp, P. & Bielefeld, M.W., 1995. Statnaiiiic load
Application of Stress- W a v e Theory to Piles, Ottawa, testing and the influence of stress wave phenomena. Proc.
Canada: 814-825. of 1st Int. Statnariiic Seminar, Vancouver, Canada: 207-220.
Fujita, K. & Kusakabe, O., 1988. On the evaluation of static Nishida, Y., 1963. A basic calculation on the failure zone and
bearing capacity. Proc. 3rd Int. Conf on the Application of the initial pore pressure around a driven pile in clay. Proc.
Stress- W a v e Theory to Piles, Ottawa, Canada: 525-534. 2nd Asian Conf: Soil Mech. Found. Engrg., Vol. I ,
Goble, G.G., Likins, G.E., JI-. and Rausche, F., 1975. Bearing JSSMFE, Tokyo: 2 17-2 19.
capacity of piles from dynamic measurements. Dept. of Nishiniura, S., Matsumoto, T., Kusakabe, O., Nishiumi, T.,
Civil Eng., Case Western Reserve University. Yoshizawa, Y., 2000. Case studies of Statnamic load
Hayashi, M., Okamoto, T., Kanai, F. & Matsuinoto, T., 1994. testing in Japan. Proc. 5th Int. Cot$ on the Applicu/ioti q/'
Predictions of load-settlement relations of steel pipe piles the S/i.e.s.s- Wave Theoiy to Piles (to be published).
froin dynamic load testing. Proc. 5th /tit. Cot$ u17d Randolph, M.F., 1987. Modeliiig of the soil plug response
Exhibition on Piling ar7d Deep Foundations, Bruges: 3.9. I - during pile driving. Proc. 8th S. E. Asian Ceotechiiicul
3.9.7. Cot$, Bangkok, Vo1.2: 6.1-6.14.
Hayashi, M., Matsurnoto, T. & Suzuki, M., 2000. Dynamic Randolph, M.F. & Simons, H.A., 1986. An improved soil
load testing on 102 steel pipe piles for bridge foundations model for one-di~nensionalpile driving analysis. Proc. 3rd
on mudstone. Proc. 5th Int. Conf: on tl7e Application ofthe Iiit. Con$ on Nunz. Meth. in Ojfihore Piling Nantes: I - 17.
Stress- W a v e Theory to Piles, Sao Paulo (to be published). Randolph, M.F. and Deeks, A.J., 1992. Dynamic and static soil
Holeyman, A.E., 1992. Technology of pile dynamic testing. models for axial pile response. Proc. of3rd Int. Cot$ 017
Application qfStress- W u v e Theory to Piles,Hague: 3- 14.
342
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
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 the ethics to the world of the ‘ant ethics’, where eve-
qualification to the evolution of the building site rything is the same and nothing is better.
through the Quality Control, better known as QC. This unsolved equation, where the independent
But what does it actually mean? Is it, maybe, the fi- variables such us time, price and quality have been
nal objective of the site itself or the evolution of the taken, gradually, to the distortion of the QA concept
quality of works done in order to be checked against together with the proper behaviour of the Engineers,
the projected idea? to obtain as a result a form where the QA box is
Nowadays, when the fiemy of the regional ticked. That is to say interchanging money for a
economies requires low costs and time limits practi- simple signed paper instead of a useful service.
cally impossible to comply with, the QC concept Notable contradiction since more piles are been
generates a dichotomy with the demands of the mar- tested each year all over the world. New technolo-
ket. Where this market, thanks to advertising and gies are being developed while the old ones are be-
marketing, offers their products and services en- ing updated. Various are offered to us as the pana-
wrapped by the I S 0 9000 halo; conditioning the cea, the magic solution, the unique answer we were
building industry (producer of the container for these waiting for. The ‘technology’ that does without the
products or services) to join up to this trend. engineering knowledge, feeling and judgement. This
It is not the authors intention to disregard the ne- creates the proper conditions to take us far away
cessity of quality in sites, but it is to call the atten- from the purpose we were hired, which is to measure
tion saying that the QC, as well as the Quality As- and inform about the state of the foundations. But
surance (QA), requires indispensables items such as: even worse, it misrepresent the Engineer ideal and it
an adequate project, organisation, methodology, apt- highlights how degraded the human ethics is...as we
ness, responsibility and honesty by those involved in all know that there are no magic solutions in our
it. At first sight it seems that these ingredients have professional scope.
opposite directions to low costs and short duration.
And us, in charged of the QA of deep foundations But do we really know better our foundations or are
are not unaware of this phenomenon. we just collecting data? Are we contributing with
In the cold world these words describe, the mer- vital information that will allow the Engineer clarify
cantilist conception of life prunes a new branch of the real behaviour of the pile-soil interaction? Can
what is human each morning and of our engineering we show the relation between prediction and per-
judgement. formance?
In the last decades we have witnessed the evolu- It is evident that the increasing tendencies all over
tion of science, technology, and knowledge.. .and 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 mechanized in order to increase their production. In
with it. But does this greater demand to know the re- our case, not only, each event in the site can be
ality grew in the same way as the knowledge of physically different, but we also have to analyse as
those responsible summiting the verdict? Because separate entities the boundary conditions that it is
technology did and what is most, it even superseded subjected.
the expectations.
Nowadays it is possible to find all kinds of But should the manufacturers take all the blame? Of
equipment to perform even the most unthinkable course not. A good part of it resides in the transfer of
measurements; any colour, size and origin. The the knowledge and not of the technology. The pur-
evolution of science and technology has invaded our pose of the ones involved in higher education should
lives in such a way, that it is difficult to imagine our be: ‘teach carpentry and not hammer’. We need to
daily routines with out them. And Engineering can- teach the ‘why’ and not the ‘how’ of this speciality.
not run away from this trend neither can our spe- Tools come and go. Teaching engineers giving pri-
cialty . ority to tools limits their knowledge towards them,
We blindly believe that technology will solve and therefore their future.
everything. It seems all sense of proportion has been Anybody who knows how to read and write and
lost, for, even if technology can help with interesting has developed a reasonable abstraction capacity and
tools to the diagnosis, the main engine is still the en- logical reasoning can learn a computer programme
gineering knowledge based on experience. Even in any moment of his life. That is why, it is impor-
though Engineering will hardly survive without tant that the education should exploit people’s ca-
technology, it is dangerous to think that only it will pacities and skills, the ones which he would be able
save the future of our profession. to comply with the challenges and knowledge that
If we favour the electronic connections and chips are constantly presented in real life.
over the brain connections, we are doomed to failure The focusing on technology- in the computers,
as professionals and as homo sapiens. We are des- for example- will lead us to have obsolete graduates.
tined to become animals capable of performing basic The question is not, teaching instrumental prob-
and repetitive operations (the donkey can blow a lems of a technology that develops speedily, but
flute, but that does not make it a musician) teaching the Engineers how to think. Giving them
Some Engineers show a reverential posture as re- the ability of thought built up on a concrete knowl-
gards machines which they judge mistakenly, very edge basis, that will allow them to react intelligently
difficult to use. They think that their employees, before the unceasing changes, even the technological
who use them with odd familiarity, show as much ones, that will appear inevitably in the decades to
intellectual capacity as the researchers that devel- come.
oped the Stress Wave Theory. The fact of being able The role of the educational institutions is not that
to sit in fiont of a keyboard turns the employee into of producing tailor made first class workers, but hu-
a genius to his boss’ eyes, in a gifted being if the one man beings capable of thinking and being creative,
supervising is the company’s chairman.. .if continu- that is what the most lucid businessmen are asking
ing to increase generation gap. In general terms, it is for and that the politicians, blinded by the ‘high
believed that a technically complex environment en- technology’, do not have in mind.
dows a person of a great intellectual capacity; it is as There are no doubts that everybody needs to learn
stating that whoever sits in fiont of a TV. set will be how to use a computer and to be able to manage the
endowed with knowledge in electronics. informatics field with certain ability, which it is get-
It seems that relying on the mere presence of ting easier than imagined every day. We should al-
electronic means turns out to be enough stimulation ways remember that computers are tools and not the
for the intellectual capacity of the engineers. The aim itself. They are not capable of replacing the in-
simple fact of selecting and observing a screen is a tellectual skills and knowledge the Engineer should
week substitute of the real mental activity. The users learn in order to be able to incorporate him to the
end up learning how to press buttons and getting re- world economy.
wards, as laboratory animals, instead of a real com- In some cases it is being taught as if the study of
prehension of the cause-effect relation. If we keep deep foundations could be understood through the
on going on this dangerous path, one day we will usage of some recipes or simple operations in order
find an exquisite brochure showing an equipment to manipulate the equipment, obtaining results on
which will seduce us with its ergonometric design the LCD which the Engineer prefers not to argue be-
and low price; and it will have two little lights: the cause: “I got it fiom the equipment”. Having as a re-
green will indicate an undamaged pile while the red sult prefabricated answers that don’t belong to him.
one damaged! What happened with the power of knowing true
It is necessary to admit that the Engineer’s work fiom false? And the answer is in the relationship
is different fiom the repetitive tasks the collar em- teacher-student and in its content.. .in the desire of
ployees or the industrial workers do, which has been teaching and learning.
346
What is the use of learning? The question is not a deep and long lasting satisfaction that may come
new. Almost 2300 years ago, a young geometry stu- months or years later.
dent asked Euclid: ‘What will I get learning these The students do not know that they do not know.
things’. The master called his slave and said: ‘Give This ‘happy unconsciousness’ shows that the sys-
him some coins. It seems he needs to get something tem has not even been able to show sobre the an-
in order to learn’. tagonism between true and false, knowledge and the
Together with the world‘s educational problem, in lack of it. This lack of information will generate, in
which our speciality is not exonerated, it is pro- part of the new generations, a failure experience due
claimed that no previous knowledge is required to the contradiction between the high expectations
therefore it is being taught that knowledge is not hi- and the insufficient knowledge. It can have an un-
erarchical, that is to say, you can start constructing a thinkable impact on society as well, because of the
building at any level, without foundations. By elimi- presence of a group of Engineers whose demands
nating the existence of a certain sequence and conti- will be incoherent compared with their skills.
nuity in education, we are stating that sequence and
continuity have nothing to do with the thought itself. All the community involved in the pile area must
We find ourselves those who state that in order to keep an eye on this week balance between Manu-
interpret foundation dynamics no deep knowledge in facturer-Educator-User( equipment-kno wledge-use).
geotechnics is needed. In other words, it is the same Non of these parts can afford the cost of becoming
to test a pile which is imbedded in the soil that an the week link of this chain, otherwise it will break.
isolated column! Each of us has the obligation of being better every
Euclid’s student is now the one sitting at the day and the noble mission of helping the other two
schools desks! links meeting their commitment of being ethic and
The obsession of earning one’s living is the easi- having a proper behaviour. If this simple statement
est way to waist it and it is not worthwhile. The pre- can be performed, our job will last becoming a mar-
sent educational system searches for the utility and ket necessity. And as in every modern market, trans-
presumes the death of knowledge. Some of today’s actions involves money, it is demonstrated the ax-
teaching methods are usage instructions to create ro- iom that ethics and money are compatible.
bots operated by ignorants with ideological lumps.
The real knowledge is not on the Internet. but in the
brain.
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.
35 1
pile driving hammers, cushion stiffness and developed by Rausche (1970), or TNOWAVE,
cushion coefficient of restitution for 18 different discussed by Courage and Bielefeld (1992), is
cushion materials, and system efficiency for 100 used to estimate soil quake and damping
different hammer-cushion configurations. Wave parameters, and the distribution of the soil
equation analyses are performed for a range in resistance to driving along the length and at the
driving system performance parameters toe of the pile. The pile is divided into continuous
determined from the mean and standard deviation segments and calculations are made using a
of values in our data base. Minimum blow counts traveling wave algorithm. Either the measured
are computed for the lower bound coring case. If pile top force or velocity is used as a boundary
our evaluation of pile drivability indicates condition, and the complementary quantity is
premature pile refusal or excessive driving computed and compared with the measured
stresses, the hammer and/or wall thickness that quantity. The set of soil parameters is varied until
permits the piles to be driven to design penetration a best match between measured and computed pile
is determined. top force or velocity is obtained.
Our pile monitoring system records force and Piles with adequate lateral capacity rehsing above
velocity signals using strain transducers and design penetration are accepted if:
accelerometers attached on opposite sides of the the required tensile and compressive pile
pile near the top. The impact stress and velocity, capacities are obtained from the upper bound
the energy transmitted to the pile (or system static pile capacity curves.
efficiency), and the soil resistance to driving are o the required tensile and compressive
computed for every hammer blow. We estimate capacities are confirmed by the soil resistance
the soil resistance to driving using the Case-Goble to driving obtained from a CAPWAP
formulation assuming a damping coefficient that is analysis, and
a function of soil type. The impedance or TNO 8 the required compressive capacity is con-
methods. as discussed by van Foeken, Daniels, firmed by the soil resistance to driving
and Middendorp (1996), can also be used. obtained from the Case-Goble formulation
Estimates of the hammer efficiency and the and wave equation analyses performed using
cushion stiffness and cushion coefficient of the measured driving system performance
restitution are obtained for selected hammer blows data (i.e., the hammer efficiency, cushion
using force-time characteristics obtained in a pre- stiffness, and cushion coefficient of restitu-
installation parametric study. We recommend tion) and the field blow count.
using controlled hard driving, as proposed by Piles can also be accepted if the required
Stevens, Wiltsie, and Middlebrooks (1984) tensile and compressive piles capacities are
whenever the soil boring indicates the presence of obtained from a re-evaluation of pile capacity.
rock layers or cemented zones. Driving is con- The soil resistance to driving profile obtained
tinued if the end bearing is decreasing or erratic, from the Case-Goble bearing capacity is used to
and the driving stresses at or just above the driving modify the elevation and thickness of soil strata.
shoe are acceptable. Driving is terminated if the When pipe piles are driven open-ended,
driving stresses are excessive, or if refusal is relative movement between the pile and soil
reached and the end bearing is constant or occurs both on the outside and inside of the pile
increasing with depth. wall. Skin friction is, therefore, mobilized on both
Our hammer monitoring system measures the of these surfaces. The end bearing area in this
velocity of the ram just prior to impact using Hall case is equal to the cross-sectional area of steel at
sensors and a small magnet attached to the ram. the toe. When a pile plugs, the soil plug moves
Hammer efficiency is determined for every with the pile during driving. This friction is
hammer blow from the ram impact velocity. We mobilized only on the outer wall, and the end
generally recommend that only ram velocity be bearing area is then the gross end area of the pile.
monitored when easy driving is expected in order The pile initially plugs after each delay. Because
to evaluate pile acceptance should blow counts the distribution of the shaft resistance on the inside
lower than predicted occur. and outside of the pile cannot be evaluated, we
A signal matching program, such as the CAse recommend waiting a few minutes and restriking
Pile Wave Analysis Program (CAPWAP) the pile.
352
Pile acceptance at refusal is simplified when profile with clay seams and layers. Drivability
the pile driving hammers selected are large studies indicated that a Menck MHU 1000
enough to overcome the long-term static pile hammer could drive the 48-in.-diameter piles to
capacity, and the CAPWAP and Case-Goble soil 84-m penetration if the wall thickness was
resistance to driving can be considered a lower increased to 2.00 inch. Blow counts with the
bound because more resistance is mobilized at Menck MHU 1000 hammer ranged fiom 25 to 40
larger pile displacements. blows per 0.25 m at 60-m penetration, fiom 35 to
Typically, the pile driving hammer selected 65 blows per 0.25 m at 70-rn penetration, and
for a particular installation may be large enough to fiom 45 to 85 blows per 0.25 m at 80-m
drive the piles to design penetration, but not large penetration.
enough to overcome the long-term static capacity. With the exception of Pile C-3, all piles were
In clay, the skin fiiction during driving is driven to design penetration. Pile C-3 was driven
generally much smaller than that mobilized under to refusal at 56.8-m penetration, 1.2 m above
static loading because large excess pore pressures design penetration. Pile C-3 was accepted because
are generated during continuous driving. the required compressive capacity was obtained
CAPWAP analyses can be used to estimate the fiom the upper bound static pile capacity curve
distribution of the soil resistance to driving along computed using Ras Tanajib pile design
the length and at the toe of a pile during parameters, and the required compressive capacity
continuous driving and after a set-up period. By was confirmed by the CAPWAP analyses, the
combining these results, it is possible to proof test Case-Goble bearing capacity, and the hindcast
a pile without the expense of mobilizing a larger resistance to driving obtained fiom wave equation
hammer to the site. analyses. The required compressive design load
High strain dynamic testing can also be used for Pile C-3 was 26.1 MN. At fmal penetration,
to evaluate the capacity of cast in situ piles (also soil resistance to driving determined by Case-
known as drilled shafts or bored piles). Hussein, Goble (Fig. I), CAPWAP (Fig. 2), and wave
Likins, and Rausch (1996) made recommendations equation analyses were 28.7, 31.8, and 33.4 MN,
concerning the hammer weight, drop height, and respectively. As shown in Fig. 3, the Case-Goble
cushion thickness. The hammer weight should be bearing capacity was greater than the lower bound
at least equal to 1.5 percent of the expected compressive pile capacity.
capacity. The hammer drop height should be 6.2 In our second case history, piles were
approximately 8.5 percent of the pile length, with driven to rehsal in a hard silty clay stratum
a minimum value of 2 m. The thickness of the between penetrations of 118.9 and 134.3 ft with a
plywood cushion is determined from: Vulcan 560 hammer. Piles F-5 and G-6 were
driven without interruption to penetrations of 120
t = L2/2D
and 134 ft, respectively. Pile G-5 had delays of 22
where: t = cushion thickness, mm; and 40 minutes that had little effect on the driving.
L = pile length, m; and Pile F-6 had a delay of 3.1 hours that almost
D = pile diameter, m. resulted in pile refusal, as shown in Fig. 4.
The minimum cushion thickness is 100 mm, and The restart blow count for Pile F-6 was 730
the cushion thickness should be increased 150 mm bpf at 99-ft penetration. At this depth, the
when the pile length exceeds 30 m. required compressive capacity of 4523 kips was
not obtained fkom the upper bound static pile
6. CASE HISTORIES 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
Case histories are presented to illustrate the use of
our pile acceptance criteria for very dense sand, accepted because the minimum blow count was
hard clay, and rock. A fourth case history obtained. Although not a requirement for pile
illustrates the use of combined CAPWAP acceptance, the required compressive capacity was
confrmed by the CAPWAP capacity of 5878 kips
analyses.
6.1 In our fwst case history, piles were driven (Fig. 5), the Case-Goble bearing capacity of 5200
to design penetrations of 35 to 84 m using Menck kips, and the soil resistance to driving back-
MRBS 2500-S, 3000, 4600, and 5000 air/steam calculated from wave equation analyses of 6050
kips.
hammers, and a Menck MHU 1000 hydraulic
hammer. Cone penetrometer tests indicated a 6.3 In our third case history, piles were
driven to practical refusal in a strong gypsum
predominantly dense to very dense silty sand
353
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
Figure 1. Log of Driving System Performance (Pile C-3, Menck MRBS 5000 Hammer)
354
Blow No. 2787
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
stratum between penetrations of 58.3 and 62.2 A exceeded, the required tensile capacities were
with a Vulcan 560 hammer. Driving was obtained from the lower bound static pile capacity
terminated because stresses exceeded 85 percent curve, the required compressive capacities were
of yield in the Grade B steel above the driving obtained from the upper bound static pile capacity
shoe. curve computed using rock end bearing, the
The piles were accepted because the required required tensile and compressive capacities were
penetration for adequate lateral capacity was 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 6.4 Our fourth case history is for a series of
by the Case-Goble bearing capacities and soil redrive tests performed with a hammer that is
resistances to driving obtained fiom wave clearly not large enough to mobilize the full soil
equation analyses performed using measured resistance. The soil resistance mobilized during a
driving system performance data and the field series of redrive tests performed on a 1.6-m-
blow count. diameter steel pipe pile driven to a penetration of
The driving system performance data for Pile 26 m in a very silty clay is presented in Fig. 8.
B-2, shown on Fig. 6, indicates a Case-Goble The four lower bound soil resistance profiles are
bearing capacity of less than 1460 kips to 28-ft for continuous driving. The pile was redriven by
penetration, and less than 2140 kips to 60-ft applying only two consecutive hammer blows
penetration. The required compressive capacity of with a PMJ-400 hydraulic hammer after delays of
3390 kips was obtained at 62-fi penetration, and 6 minutes, 15 minutes, 33 minutes, 2 hours, and
was confirmed by the CAPWAP capacity of 4823 66 hours. The resistance generally increases with
kips (Fig. 7), the Case-Goble bearing capacity of time, but the resistance after a 33-minute delay
5072 kips, and the hindcast capacity of 6050 kips. appears to be slightly smaller than the resistance
357
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
Figure 4. Log of Driving System Performance (Pile F-6, Vulcan 560 Hammer)
358
Blow No. 170
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
mobilized after a 15-minute delay, and the analysis is about 4400 kips (19.6 MN), as shown
resistance after a 66-hour delay is less than the in Fig. 9. In the standard CAPWAP analysis, the
resistance mobilized after a 2-hour delay. The resistance mobilized was only 3305 kips (14.7
maximum resistance overcome by the pile driving MN).
hammer is about 3300 kips (14.7 MN). The To help put this in perspective, Fig. 10 shows
hammer is too small to mobilize the hll soil that the resistance mobilized by a Vulcan 060
resistance during the redrive tests. This is shown hammer is about 3300 kips (14.7 MN), and the
very clearly for the redrive test performed after the resistance mobilized by a Vulcan 560 hammer is
66-hour delay. Almost no soil resistance was about 4100 kips (18.2 MN), assuming the same
mobilized over the bottom quarter of the pile. soil and pile parameters. Comparing the three
In our combined CAP WAP anaIyses, we have hammers shown in Fig. 10, we see that by
assumed that the soil resistance mobilized during mobilizing a hammer having 50 percent more
continuous driving is a lower bound, i.e., the soil rated energy results in only a 10 percent increase
resistance mobilized on a particular pile segment in the maximum soil resistance overcome, and
is assumed to be the larger of the actual resistance mobilizing a hammer having 150 percent more
mobilized or the resistance mobilized during rated energy results in a 37 percent increase in the
continuous driving. The soil resistance mobilized maximum resistance overcome.
after a 66-hour delay in the combined CAPWAP
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
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
Figure 6b. Log of Driving System Performance (Pile B-2, Vulcan 560 Hammer)
362
Blow No. 739
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
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
366
Figure 10. Driving Resistance-Blow Count Curves
367
Courage, W.M.G. and Bielefeld, M. W. (1992), van Foeken, R.J., Daniels, B., and Middendorp, P.
"TNO automatic signal matching," Proceedings of (1 996), "An improved method for the real time
the Fourth International Conference on the calculation of soil resistance during driving," Fifth
Application of Stress-Wave Theory to Piles, The International Conference on the Application of
Hague, The Netherlands, September 2 1-24, 1992. Stress-Wave Theory to Piles, Orlando, Florida,
September 1 1- 13, 1996.
Helfrich, S.C., Wiltsie, E.A., Cox, W.R., and
AI-Shafei, K.A. (1985), "Pile Load Tests in Dense Wiltsie, E.A., Stevens, R.F., and Vines, W.R.
Sand: Planning, Instrumentation, and Results," (1984), "Pile Installation Acceptance in Strong
Proceedings, 7th Offshore Technology Soil," Proceedings, 2nd International Conference
Conference, Vol. 1, pp. 55-64. on the Application of Stress Wave Theory on
Piles, pp. 72-78.
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.
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
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.
370
The statistic data of static testing of 20 large into full play even when the shaft frictional
diameter diving casing cast-in-situ piles among 6 resistance has been exerted to an extreme extent.
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 5 PROBAl3LE EXISTING PROBLEMS
resistance in miscellaneous soil, clay and residual
sandy clay are obtained. The mean values of them After driving the former of large diameter diving
are respectively 57kPa, 79kPa and IOSkPa. casing cast-in-situ piles has been finished, the pile
bottom is required to be examined with lamp.
Meanwhile, some effective technology measures
4 ANALYSIS OF BEARING CAPACITY have been taken. There, some common faults of
MECHANISM quality in small diameter diving casing cast-in-situ
piles, such as pile tip mixing mud, etc., can be
The CAPWAP analyses of testing piles above have avoided, and the qualified ratio of testing of large
shown that the shaft resistance and toe resistance of diameter pile piles exceeds far that of the sinalf
this type of pile are all high. It follows from this diametcr piles.
result that the mechanism of bearing capacity of this However, when carrying out testing of a pile in SF
type of pile is as follows: in the respect of toe garden in Xiamen of Fujian province, the author also
resistance, because of high penetration ability of this discussed a quality problem. The diameter of this
pile equipment which can impact former to hardpan pile with a bearing stratum of residual sandy-clay is
even to weathered rock, the bearing stratum absorbs equal to Q> 600mm, and penetration length is 3 1Am.
great energy passing by steel pile shoe and is By means of static load experiment, its ultimate
compacted, and its modulus of deformation bearing capacity is only 2 150kN, which is only one
increases greatly, which is similar to the driven pre- half of most entrusted load. After sending the pile
formed concrete pile, but its bottom area is far 970mn1, then high strain dynamic testing is carried
greater than driven pre-formed concrete pile, What out. Its ultimate bearing capacity was calculated
is the main reason why the toe resistance of this pile only 2620kN either. It follows from the curve that
can reach so high value. the pile was damaged seriously on the location of
In the respect of frictional resistance, also similar 1 I m beneath the pile top. (The geological coiuinti
to driven pre-formed pile, the pile former penetrate around pile, pile profile and pile model and
into the soii strata by blowing of rani, soil around CAPWAP final results are shown respectivcly i n
former is extruded. Soil around pile rebounds and fig.4, fig.5 and fig.6) This location happened to be
combines close with concrete of pile after casting the boundary between muck and sandy clay, So w e
concrete. But because the interface between pile and can infer the reason of causing this problem. Firstly,
soil around driven pre-formed pile is smooth, while when the pile penetrate densely into the saturated
the interface between pile and soil around large soft soil during the process of construction, soil
diameter diving casing cast-in-situ pile is rough, strata around the pile is crushed to cause excess pore
meanwhile large diameter diving casing cast-in-situ water pressure. The pressure makes the ground
pile has not the mudcake problem of bored pile, hunch and the pile bear a horizontal pushing force. If
therefore, the shaft frictional resistance of this pile is the concrete intensity of adjacent piles is still very
larger than that of pre-formed pile under the normal low, the pile perhaps damaged by the horizontal
stress. thrust force. Secondly, at the boundary between
Analyzing result of CAPWAP program of high hardpan and soft soil, the speed of pulling former is
strain testing indicates that the end bearing of pile is so rapid that it causes the pile mixing muck.
about 35-45% in all load of pile top, and its mean In fact, on the other field with a deep muck such
value is 38.9%, which is larger than that of cominoii as DH garden introduced above, many piles are
bored cast-in-situ pile under the same geological found some defects existing on the location of mitck
condition. The result also shows that this pile is by testing. The reason may be the same as that of SF
buoyant pile while its lengt~i-diameterratio (LiD) is garden.
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 4 CONCLUSIONS
simultaneously to resist the load of pile top, thus the
pile obtain higher bearing capacity. For the bored 1. The testing result of large diameter diving
cast-in-situ pile, because of the influence of casing cast-in-situ pile indicates: Its shaft frictional
sediment, the effect of toe resistance may not bring resistance and toe resistance can resist together the
371
Fig.3 CAPWAP final result of testing pile in DH 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
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
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 drop heights of0.2, 0.4, 0.6, 0.8, 1.0, 1.2, and 1.5 m.
The elastic (rebound) and permanent (set)
Dynamic load tests with increasing energy has displacements were recorded through paper and
been used in Brazil for 10 years and allows the pencil, placed at the top of the pile.
bearing capacity determination of the soil- Based on the measured rebounds, the mobilized
foundation system from a load- displacement curve load was calculated according to UTO et al. (1985)
defined by a certain number of points. and VELLOSO (1987) propositions.
The rebound technique has also been spread in
Brazil as a driving control tool of all piles in a
construction site.
- Displacenieri f parcels C2 arid C3 deierniirinfrori.
In practice, the measured rebound signal is the C2 and C3 displacement parcels are the first and
wave equation solution. The rebound recording at major variables in the load test assessment using the
the top of the pile with pencil and paper can be used rebound technique. Inadequate assumptions of these
to estimate the mobilized load for the corresponding variables can lead to misrepresentative mobilized
energy level. Assessment of the mobilized load - load values
displacement curve for several energy levels allows Scatter results were obtained for the formerly
the ultimate load estimation. assumed values C3 = 2.5 mm and C = set. Thus, the
displacement parcels C2 and C3 were determined by
AOKI (1986) proposition
2 DYNANAMIC TESTS PERFORMED DURING
PILE DRIVING
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.
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
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
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.
379
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.
Figwe 7 - Comparisons between static and dynamic load tests.
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 dynamic load test with rebound measurement de estacas pre-fabricadas via repique elastic0 da cravaGdo
is a very straightforward and, because it does not PublicaCSo da ABMS/NRSP. ABEF E IE/SP, p 48
require any type of equipment, almost costless AOKI. N (1989) A new dynamic load test concept In XI1
INTEFWACIONAL CONFERENCE ON SOIL
technique. MECHANICS AND FOUNDATION ENGINEERING. TC
380
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
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
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.
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
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.
389
Figure 1. Test setups for single steel rail (A = accelerometer;
T = strain transducer)
Figure 2. Force and velocity traces for different test setups steel
rail type TR 32. resistance is 29% higher than test (a) where great
deviation occured.
In this example, it is clear the necessity of a good
The results of the CAF'WAPC@analyses base on control of the test procedure and instrumentation set
wave up matching are presented in Figure 3 for the up to carry out dynamic test in such piles otherwise,
force (F) signal. The quality match for both cases is their results should be use with caution.
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. 3 INSTRUMENTATION OF COMPOSITE PILE
For test (b), where force (F) and velocity
multiplied by pile impedance (ZV) signals are The difficulties and the experience with dynamic
congruent in the beginning, the mobilized soil 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.
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.
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
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
393
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-
Figure 1 - Installation of screw pile vide this resistance verification.
created around the pile that reduces lateral support. 2 DYNAMIC PILE TESTING
However, in very soft soils it may be an advantage
in reducing negative skin friction. It is understood The Author uses the PAK Pile Driving Analyzer
that at least one company is developing a new from Pile Dynamics Inc. together with the associated
pressed fitting, which does not extend beyond the CAPWAP signal matching software. The method
outside diameter of the steel tube. The Author has and current “state-of-the-art” has been described in
not dynamically tested a pile with the new joint. Goble et a1 ( I 996).
Concerns have been raised by some clients that, A special cable drop hammer and frame was
based on design calculations, the connection be- made up by one of the Contractors specifically for
tween the screw flight and the pile tube is a weak dynamic testing of his screw piles. See Figure 2.
point in the system as almost all pile resistance to The drop weights comprise 2tonne and 4tOnr1e solid
vertical load is supposed to be provided by the screw circular billets of steel of about 350mrn diameter and
flight. Concerns include the stiffness of the auger 2.6m and 5m long respectively. The guide frame
flight and the strength of the welded connection to allows for a maximum hammer stroke of about 2m.
the steel tube. Although these concerns are well The frame is also supported laterally by 4 guy wires
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 Figure 2 - Test Rig
torque can provide a quantitative measure of vertical
load resistance, which is mostly provided by end
394
that are generally tied to adjacent screw piles, being Independent Gaoscience 04-Apr-97
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
Figure 4 -Summarv table for SP9
3.1 Sydney International Airport
This was the first project where the Author had en- side of the pile the Author does not accept that it
countered screw piles and the design appeared to be truly is involved in the travel of the stress wave and
under early development. Several configurations we almost never model the flights as anything other
were tested, being 89x5.5mm tubes with single than supplying a point resistance with low stiffness.
350mm screw flights near the toe (working load Owing to the small shaft diameter conventional
1OOkN) and 168x7mm tubes with either two 700mm (2Fi2V) gauges were placed individually at 90” in-
screw flights (working load 500kN) or with 5 flights tervals around the piles. Gauges were bolted to the
(working load 600kN). All piles penetrated the pile using drilled and tapped holes as with conven-
ground about 8m. Ground conditions were moder- tional testing of larger steel tube piles. Testing was
ately dense to dense medium sands. 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-
Figure 3 -Plots for SP9
sistance. However, the testing was generally
stopped owing to high dynamic stresses in the shaft
rather than reaching the ultimate geotechnical resis-
CAPWAF’ modelling of the piles should not in- tance of the piles with very high “sets.”
clude the screw flight as taking part in the travel of The piles with 5 flights could not be “set” at all
stress waves along the pile ie a significant increase and they were probably significantly over-designed.
in shaft impedance over a very short distance, of The Author has not seen this design used again.
only a few millimetres, at the screw flight is not re- We CAPWAPed data from all piles. During
quired. As the flights are only a plate attached to the 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 ~
396
correlation is shown in Figure 7. The static load test Independent Geoscience 06-Oct-99
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
4 CONCLUSIONS
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
398
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
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.
399
The over-consumption on this pile was 105%, ie 4 DYNAMIC PILE TESTING
more than 2 times the required volume of concrete
was used during construction. The contractor’s The test piles were cast above ground level inside a
equipment provided for measuring pressure and vol- steel sleeve of about 4mm wall thickness and about
ume throughout construction of the pile so the pile the same diameter as the pile for about 2 5 pile di-
profile could be estimated. The contractors record ameters above the surrounding ground This was
for pile 68 is shown in Figure 1 . Most of the extra done at the same time as casting the pile or as soon
concrete consumption is shown as a cone between 5 as possible after casting the pile in order to keep the
and IOm depth. 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
3 STATIC PILE TESTING removed This leaves a substantial steel collar at the
top to reinforce the pile top during impacts of a drop
A static load test had been conducted on pile 68 weight and allows the test equipment to be attached
prior to the dynamic pile testing. The contractor was at a level where there is a regular smooth surface
careful to avoid the dynamic testing consultant be- with no additional impedance that might interfere
coming aware of the static test results before the dy- with strain measurements The location of the test
namic test results were reported. The static test was gauges had a diameter close to that of the pile shaft
conducted with several cycles in accordance with the and had similar reinforcing
Australian Piling Code. Applied load was measured The Author uses the PAK model Pile Driving
using the jack pressure only. There are shortcom- Analyzer from Pile Dynamics Inc together with the
ings to this system that are described below. Dis- associated CAPWAP signal matching software The
placement was measured using 3 dial gauges. A method and current “state-of-the-art” has been de-
check of pile displacement was also taken using a scribed in Goble et al (1996) The option to test
level survey. Load was applied by jacking against a with 4 strain gauges was not adopted for piles of this
reaction beam and displacement of this beam was size
also measured by level survey.
400
The contractor supplied a “Hydroquip” HQ5 hy- sonnel that conducted the test claim this was not the
draulic piling hammer to strike the piles. Some re- case and consider that the jack did not reach the end
building of the hammer’s hydraulic valving had been of it’s travel. However the measurements of the re-
conducted to increase energy transfer efficiency. action beam, which was also deflecting elastically,
Highest energy transfer efficiency during this testing show the same behaviour, with almost no deflection
was 76%’ which we consider to be high for a hy- during the last load application cycle. This is shown
draulic hammer with a Stonne rain striking a solid in figure 3. The Author considers that the maximum
concrete pile of this diameter. load applied to the pile did not exceed 1700kN.
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
Figure 3 - Hastings St Static Test Reaction Beam
The static results are summarized in Figure 2,
below together with the CAPWAP load vs deflection As the static test was conducted On this pile prior
prediction plotted on the same axes.
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-
Figure 2 - Hastings St Load vs Deflcctioii sary to “preload” the piles by “driving” them after
construction
It can be seen that in the static test when the pile If the inaxiinuin applied load during the static
reaches a displacement of slightly less than 151nni load test was 1700kN this correlates well with the
the inferred load increases but there is no corre- “break point’ at about 1800kN in the CAPWAP load
sponding deflection of the pile. The Author consid- vs deflection prediction for this pile The un-
ers that this is impossible and that there must have load/reload stiffness shown i n each of the load cy-
been an error in the test measurements The most cles of the static test is also of interest as this corre-
likely error was that the jack reached the end of its lates quite well with the initial loading stiffness
travel or jammed such that although there was an in- shown in the dynamic test results
crease in jack pressure and hence inferred load on The stiffness of the static test on initial loading
the pile, in reality load did not increase and the pile during each of the cycles also correlates reasonably
consequently did not deflect. The contractor’s per- well with the dynamic test prediction after the
40 1
“break point.” The Author considers the dynamic Accuracy of the results for bored piles appears
test was the first time the pile experienced sufficient similar to driven pre-formed piles provided the pile
deflection to generate a resistance of more than shaft can be realistically modeled. This requires
about 1700kN and so after this point the lower initial some knowledge of the shape of the pile shaft.
loading stiffness is valid. The load-displacement behaviour of “cfa” piles in
The load vs deflection behaviour shown in other sand appears to be related to previous load history
dynamic tests at the site were similar, with similar with initial loading to any level of load being less
initial loading stiffness and stiffness after the “break stiff than reloading. Consequently dynamic testing
point” but they showed a much lower “break point” should be conducted with as few blows as possible if
and a typical example is shown in Figure 4. The it is necessary to predict initial load stiffness.
Author considers this is because these other piles 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
6 CONCLUSIONS
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
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.
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
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 .
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
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.
407
3.1 Test Pile I
The PDA test results for TP1 are summarised in Total Resistance
Table 1 and the CAPWAP@'analyses results are
summarised in Table 2.
408
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).
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.
7 REFERENCES
41 0
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
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.
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 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 )
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.
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
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.
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.
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)
419
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.
Figure 6. Comparison between Equation (1) and Case-
Goble capacity, hydraulic hammers 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.
Equation (1) is actually the same as “true” Hiley Viljakainen, J. 1998. Steel pipe pile’s dynamic load testing and
formula or Paikowsky’s (1 994) uncorrected Energy driving simulation. Master Thesis. Tampere University of
Approach prediction. The correlation in Figure 6 may Technology. Tampere. (in Finnish).
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.
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
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
2 PROJECT SITE CHARACTERISTICS The Jamuna river is as wide as 15km during flood
season and 5km during dry season. The main stream
2.1 /+(?jt?cfdefciih of Jamuna river shifts each year so river training
The Jamuna Multi Purpose Bridge(Figure l), Bang- work is needed.
ladesh, consisted of a 4.8km long combined road, rail
42 1
2 2 Pile details proximately 3m. The piles had the “Shear keys”
The piles for the 4 8km Jamuna Bridge are large located neat the pile top and also near the pile bot-
diameter open ended steel pipe piles There is a total tom(Figure 3 ) .
of 121 piles and all are driven on a 1H 6V inclina- These keys were approximately 20mm thickness,
tion The Piles are either 3 15m or 2 50 diameter semicircular in shape, located in alternate planes per-
steel pipes of variable wall thickness in the range of pendicular to the pile axis, and extended from about
40 to 60mm, and driven in two or three sections The 3m below the pile top to about 13m below the pile
splice is accomplished by use of common offshore top. The shear keys at the bottom began one diame-
“stabbing guides” which are internal pipes of slightly ter above the pile toe and extended about 13m from
smaller diameter with pointed end to assist in align- the pile toe. Wall thickness for the 3.15m diameter
ment of the two sections, the thickness of the stab- piles was 60mm at the top, reducing 55mm, then
bing guides is 25mm and the length of those is ap- 5Omm, then40mm, before terminating at the toe with
422
Figure 3. Stabbing Guide and Shear Key
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 The CPT qc profiles are erratic, probably due to
pressure remained essentially hydrostatic during CPT the dense state of the sand, with high negative pore
penetration. Inclusions (lenses) of fine sand and silt at water pressure readings (down to atmospheric) ob-
random levels were also indicated. CPT results con- tained during the CPT pushes. These responses are
sistently show a linear increasing tip bearing pressure probably caused by a combination of dilatancy and a
as a fhction of depth. 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
3.2 Gr.nvel Lnyer
Holocene “Upper Sand”. From both the borehole and
The gravel layer consists of a mixture of both sand CPT data, the “Lower Sand” contains inclusions
and gravel. The gravel sizes range from fine to coarse (both layers and lenses) of very dense sandy silt and
(up to 6Omm diameter) and occasional cobbles very stiff to hard clay.
(maximum diameter 80mm) The gravel at approxi-
mately PWD-70111 was found in all boreholes and the The “Lower Sand” is clearly a different layer than
majority of CPTs The layer is of variable grading, the “Upper Sand’, both in terms of its geology (age)
both vertically and laterally The layer often exhibits and geotechnical (density) characteristics.
a “sandwich” layering, consisting of a thick layer of
gravel, underlain by sand, and thereafter gravel Oc- The detail soil characteristics is summarized in
casionally the complete layer is gravel. The gravel Table 1
layer is approximately 7m thick and dense
The excess pore pressure again remained essen-
4 DYNAMIC BEHAVIOR
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 4 I P / I ~Lmgrh r < c m
are founded) with gravel sizes up to 60mm diameter Soil plug measurements were made to investigate soil
All piles were driven to a depth of 74 to 76m (toe plug behavior during driving by using a so-called
elevation PWD-72m) plug follower, a ballast weight placed on top of the
soil plug This weight was connected to a wire
3.3 L O M W Snmi Lnyer guided to the outside of the pile shaft to enable the
tneasurements of soil plug movements relative to the
Micaceous sands encountered at elevations varying
pile head
between approximately PWD-7 1 m and PWD-80m,
The plugging can be indicated by the specific re-
and exended to depths beyond those investigated
covery ratio and plug length ratio A better indication
(GL-I 20m) Their in-situ density is probably dense
424
Table 1. Soil Characteristics of Jamuna Site
425
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.
426
Below this level (which roughly corresponds to the one of the most important case study about the one
top of the shear key location) the unit friction is of the largest piles driven in normally consolidated
about 250 to 350 kPa, and probably represents the sands. Dynamic pile testing was carried out to inves-
sum of both inner and outer friction (Figure 9) tigate dynamic pile behavior and soil plug length was
In addition, the inner soil plug will be to be re- also checked.
moved by water jetting and air lifting A concrete
plug will then be installed followed by base grouting Piles soon after driving showed a skin-friction
Thus final inner friction should be discounted en- dominant pile behavior, that is, 92% of total capacity
tirely, setup on the outside has not been filly being developed by skin resistance. Quakes values
achieved at the end of driving, and the soil failure and Smith damping factors were almost constant re-
mechanism at ultimate load for the pile toe will be gardless of pile diameters. Typically the shaft and toe
changed due to plug installation and base grouting quakes (QS and QT) were 0.20 to 0.27cm and 0.27
Thus the capacity of the soil and its dynamic re- to 0.4cm, respectively. The Smith damping factors
sponse are only vaguely related to driveability or in- for shaft and toe (SS and ST) were about 0.45 to
stallation of the piles themselves It is expected that 0.61s/m and 1.24 to 1.69s/m, respectively.
the long term service load these piles will very sig-
nificantly greater than that presented here, provided
that the base grouting is properly performed 6 REFERENCE
427
PDI (1995), PDA - Pile Driving Analyzer : PAK
I'
ynamic load test of cast in place pile using a free fall hammer
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.
429
Figure 1. Soil profile and schematic of the hammer device
sandy silt, as well as a clayey silt matrix depending The upper 3.0 m of the pile was specially cast in
on their rock of origin. place to allow the support of the hydraulic hammer.
The dynamic testing was done on a production This pile was dug using bentonite slurry up to the
pile. The area was investigated by SPT borings to bedrock depth, then the steel reinforcement cages
refusal, one located exactly where the test pile was were introduced and finally the fresh concrete was
executed. placed, prepared to have a characteristic compressive
Figure 1 presents the soil profile from the SPT strength of 20 MPa. The time interval between
boring, which presents the sequence of layers as casting and testing the pile was of 26 days, so
described below. complete cure of the concrete could be assumed.
Presence of 1.9m superficial landfill layer, above The combination of the vertical and horizontal
2.8 m of soft organic silty clay layer. Below the clay forces including bending moments, acting on the pile
layer there is a fine sand layer with gravels (0,95 m) top, resulted in the predicted working load of
and, after this, there is a sandy silt residual soil layer. 2,260 kN for the particular tested pile. The tested
At 11.26 rn depth appears, the top of the bedrock pile is one of a group of 8 piles, supporting a single
constituted by granite and gnaisse. The water level in pillar of the elevated way of the metro line.
this place is located 3.4 m below the land surface. 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
3 DESCRIPTION OF THE TESTING AND pile, a value of 26,100 MPa was adopted as the
IMPACT DEVICE elastic modulus of the pile.
3.I Characteristics of the testedpile
3.2 Characteristics of the hammer
The main characteristics of the tested pile are:
1.20 m diameter, total length of 10.50 m, and length The device designed to apply high impact forces on
below the ground level of only 6.40 m. the head of the bored pile consists of a free fall self-
430
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 5 blows. The test was interrupted because the quality
pile, a 20 mm thick by 1.0 m diameter steel wired of the signals was considered satisfactory to make
rubber pad was installed, in order to evenly CAPWAP analysis later.
distribute compressive stresses on the pile head, For each blow, set measurements were collected
absorbing the dynamic impact, and preserving the from the shaft of the pile, in order to evaluate the
structural integrity of the pile top, above the pile-soil interaction behavior. For the first blow,
electronic sensors which will measure data to be 15 mm of permanent set was observed and
associated with the test. practically no permanent set was measured in the
following blows.
3.3 Test procedure
Instrumentation for the dynamic test followed the 4 RESULTS
recommendations of ASTM D-4945/89, AASHTO
Designation T 298-93, and ABNT - NBR 13208 Figure 4 presents the force and velocity x EA/c
(Brazilian Standard). curves as function of time, for the fifth and last
In the present case, considering the 1.2 m diameter blow, showing a good proportionality. Table 1
of the pile, four strain transducers and four presents the results of the maximum Case Resistance
accelerometers were used, located by pairs on each (RMX) for each blow applied, for three values of Jc
90" along the pile perimeter. Sensors were installed (Case damping factor) and considering the wave
at 3.46 m from the top and for better adherence of velocity of 3,200 m/s.
the sensors onto the shaft, its' surface was polished. Table 2 presents the main parameter obtained by
Field testing was done using the PDA (Pile CAPWAP analysis for the five applied blows.
Driving Analyzer), PAK-95 model, manufactured by During the analysis it was observed that the best
Pile Dynamics Inc. The Case Method was used to match of the curves was obtained when it was
interpret the data in the field. 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
Figure 3. Detail for the fixation of the hammer to the energy (blow 4).
pile top The skin friction resistance reached the maximum
value for the third blow; after this, despite of
The test was done by a sequence of blows with increasing energy, the skin friction value was
variable drop heights as a dynamic load test with smaller, suggesting that the saturation of the
increasing energy. This type of test has been available resistance was reached.
discussed by Aoki, 1989; Niyama & Aoki, 1991; The toe resistance increased during the sequence
Hussein et al, 1992; Aoki & Mello, 1992, Beim & of blows, showing that its total mobilization only
Aoki, 1996, Paraiso & Mello, 1998, among others. occurs after the saturation of skin friction resistance.
The test was carried out with drop heights of Considering the maximum value of resistance
0.4 m, 0.4 m, 0.8 m, 1.2 m and 1.0 m, a total of obtained by CAPWAP analysis it was determined
432
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 AASHTO T 298-93 - American Association of State
(2,260 kN), the safety factor would be greather than Highway Officials Standard method of test for high-
3. However, considering that no permanent set was strain dynamic testing of piles.
verified during the application of the blows, it is ABNT NBR 13.208/94 - AssociaqBo Brasileira de
possible to say that, probably, the total available T6cnicas - Estacas de Carregamento
resistance was not mobilized by the blows. 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,
Figure 6. Distribution of total resistance for the five Brazil (in Portuguese).
_ .
blows. Paraiso, S. and Mello, L.G. - 1998 - Variable energy
dynamic load test on a 1.0 m diameter CFA pile.
BAP 111, November.
5 CONCLUSIONS
ACKNOWLEDGEMENTS
REFERENCES
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.
435
Figure 1. Summary of the results of the driveability study for triangular friction resistance distribution.
436
Figure 2. Schematic of the hammer device
28 days was of approximately 42 MPa. The concrete Figure 3. Device for impact application
of the piles was placed more than 30 days before the
dynamic tests. 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
5 IMPACT FORCE APPLICATION DEVICE an ordinary crane available at the job site. In this
picture the piles located in the proximities can also
The impact force applicatjon device was developed be observed, with part of the reiforcement cages
and built by GEOBRAS, responsible for the appearing from the pile heads.
execution of the piles in this job site, under general
orientation of IPT - Institute for Technological
Research of Silo Paulo, considering the results 6 INSTRUMENTATION
obtained in the previous drivability study.
The device was composed of a steel tube having The instrumentation consisted of a pair of force
1.0m diameter and 8.0 m total length? closed in the transdutores and accelerometers, used in high strain
bottom by two steel plates of 25mm thickness that dynamic test. They were fixed on the concrete
also served to support the system upon the pile top surface of the pile, through two windows done in
and also as the pile helmet. diametrically opposite positions on the steel casing
The impact was generated by dropping a surface, located 2,5m from the pile top. It was
cylindrical steel mass of 8,600 kg weight, hoisted by decided to install the sensors on the concrete surface
on hydraulic motor fastened in the top of this steel and not on the steel casing, because the perfect
tube, where an automatic de-hook system was adhesion of both materials could not be verified.
designed. The fall height was controlled visually
through spaced holes of 10 cm diameter drilled
along the steel tube. 7 TEST DESCRIPTION
The hammer cushion consisted of a wood disk
800 mm in diameter and with 100 mm thickness. The high strain dynamic test was conducted
The pile cushion three plywood disks consisted of adopting the procedure widely used in Brazil, which
800 mm in diameter and with 25 mm of thickness consists in the application of a series of blows with
each one.
437
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
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
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
ACKNOWLEDGEMENTS
REFERENCES
440
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
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.
44 1
STRUCTURAL ALIGNMENTS - linear, sub- 3 BEARING CAPACITY PREDICTIONS
vertical geological features, where differentiated
movements occur along the length; and Non-confined simple compression tests were
DESCONTINUITY J - structures present in the bed performed on core samples of basalt rock with a
of the river at an elevation located between 240 and 22% 5 R.Q.D. 5100% whose results, presented in
250 meters, a sub-horizontal zone which is Chart 1, were submitted to a probability analysis in
weathered and fractured with a high rate of the concept of Gauss, allowing a 95% margin of
permeability (spill 9. reliability (a characteristic value):
The area is also characterized geologically by an
excessive, thick mantle of soil, saprolito, and Chart 1 - Results of the compression tests.
weathered rock, with a depth varying from 10,O to
20,O meters in the hillsides, and that could surpass
I$ CS I Rupture Tension -OC (MPa)
30,O meters in other localized areas (Antunes 1 1 89.0
Sobrinho, 1999 and Infanti, 1999). 2 1 52.5
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
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.
in that:
4,, =o,.k,d
cr, =
average unconfined compressive strength
of rock core;
K ~ , , = empirical factor; and
d' = depth factor.
: 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
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 - The hammer weight should, at least, be equal to
Dynamic Load Test on cast in place piles with high 1.5% of the anticipated static test mobilized load;
load capacity have had routine application in the - The hammer drop height should be approximately
USA, Europe, Asia etc. and in Brazil. The technical 8.5% of the pile length, with a minimum value of
literature available gives an account of several cases 2.0m; and
of dynamic tests on cast in place piles, as discussed - A plywood cushion thickness of t = L2 / 2 0 ,
by Rausche and Seidel (1984), Balthaus et al. where thickness ( t ) is expressed in and pile
(1985)Townsend et al. (1991), Seidel et al. (1996), length ( L ) and pile diameter ( D ) are in meters, a
Sawai et al. (1996) etc. minimum value of lOOmm is suggested, with an
The option of using the high strain dynamic load additional 15Omm being added when the pile
test (DLT) instead of the static load test, which is
length exceeds 30m.
routine in many places, comes from the
attractiveness of the economics and the scheduling In this particular case, it was postulated that a ram
benefits. The cost of the high strain dynamic test in of 120kN, with a maximum free fall height of 2 . 5 ~
Brazil is 25% less than the cost of a static load test and a plywood cushion thickness of 150mm by 1.l m
(SML-slow maintained load), and represents a diameter would be used to comply with the
reduction of 85% in time consumption. recommendations.
The device designed to apply high impact forces The diameter of the plywood cushion should be
on the head of the bored pile consists of the free fall about 80 to, at most, 90 percent of the shaft diameter
self-propulsion hammer with special features to test in order to be able to centralize the blow. The
piles with high load capacity and variable diameters, thickness of the steel striker plate to be placed above
varying from 70cm to 250cm. the top of the pile should range from 50mm to
The selection of the proper hammer size is 100- its minimum diameter being equal to the
essential for successful high-strain dynamic testing diameter of the cushion the ram, ideally, is a simple
hammer weight, and drop height. The cushion drop weight which can be raised to a variable height.
details must be appropriately chosen so that hammer In order to facilitate, rams can be made in segments
impact causes sufficient pile movement in order to
which can be assembled at the job site. The ram to
mobilize the required soil resistance, and to assure
that dynamic stresses in the shaft will not impair its be used needs to be “guided” to assure that the pile
structural integrity. top and ram bottom are perfectly parallel during
The hammer apparatus must be constructed in a impact, in order to assure uniform contact stresses
way to facilitate mobility around job sites and assure during impact.
a uniform impact to the head of the pile (Hussein et Between the steel striker plate and the top of the
al., 1996). pile a 20mm thick steel wired rubber pad was
The free fall hammer conceived and in use in installed with a 1.0m diameter in order to evenly
Brazil consists of the following main components, distribute compressive stresses on the pile head,
shown in Figure 3 : the steel striker plate, the absorbing the dynamic impact, and preserving the
cylindrical - 2 pieces - steel cage (both of these structural integrity of the pile top above the
components being referred to as the hammer), the electronic sensors, which measures data to be
set of round 20kN steel rams allowing modulation associated with the test.
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.
446
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
It is observed that the measured value of zero set Azevedo contractor responsible for the construction
in all the blows leads to the consideration that the of the bridge on the Uruguay River and, especially,
pile still presented an additional point load to be to engineer Claudio Menin, project planner and
mobilized, however, the value of the load mobilized responsible for the structural calculations of the
in the fifth blow indicated factor of safety larger bridge.
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
7 REFERENCES
the dynamic test can be interpreted in the following
way: AASHTO T 298/93 - American Association of State
- Ultimate Bearing Capacity (5th blow) = 13,856kN Highway Officials Standard - Method of test for
- Ultimate Bearing Capacity, considering the friction high-strain dynamic testing of piles.
resistance (4th blow) added to the point resistance ABNT NBR 13208/94 - Associaggo Brasileira de
th blow) = 14,556kN Normas TCcnicas - Estacas Ensaio de
The test showed the saturation of the skin friction Carregamento Dinilmico, in Portuguese.
in the 4th blow and maximum point mobilized in the ASTM D4945/89 - American Society for Testing
5th blow, whose obtained values present a close and Materials Standard - Test method for high-
approximation with those considered in the semi- strain dynamic testing of piles.
rnpiric analyses in the project phase. Antunes Sobrinho, J. et al. - 1999 - Criterios de
The dynamic test in piles of high load capacity projeto na construqgo da barragem da UHE Ita.
has become a reliable tool and has reduced costs and XXIII Seminario Nacional de Grandes Barragens.
shortened evaluation deadlines for the quality of this In Portuguese, Belo Horizonte, Brazil.
type of deep foundation, as long as the hammer to be Aoki, N. - 1989 - A new dynamic load test concept.
used presents the available potential energy to the Proc. Discussion Session on Pile Drivability, XI1
load to be mobilized. ICSMFE, Rio de Janeiro.
Aoki, N. & Mello, V. F. B. - 1992 - Dynamic
6 ACKNOWLEDGEMENTS loading test curves. Proc. Fourth International
Conference on the Application of Stress-Wave
The authors wish to thank the support given by Theory to Piles. Hauge, Holland.
the Company CBPO-Grupo Odebrecht,and Mello de 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
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
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 of the penetration which occurs between the first and
movement of the rod, the frequency characteristics second impacts due to the rebound of the rod and/or
of the stress wave and the relation between the rod ground before the second impact. In this case, during
penetration and the frequency of the stress wave will the completion of the SPT there are at least two
be introduced, while the study is still proceeding. impacts of the hammer, and this evidence should be
considered in the wave equation analysis.
Figure 3 is a diagram summarizing the measured
2 PENETRATION VS. TIME DIAGRAMS penetrations of the rod as a finction of time during
the part of the SPTs at a depth of l l m . The
Figure 2 is an example o f a diagram showing records maximum penetration and the final penetration for
of the penetration of the rod with time and the stress each blow are around 12 to 13mm and 4 to 7mm,
in the rod with time (stress wave) measured near the respectively, and the rebounds of the rod and/or
top of drill rod during one blow of the SPT. The ground are between 6mm and 8mm. While the final
figure shows that the hammer strikes the rod again penetrations (set) per blow which contribute to the
(second impact) about 14ms after the hammer blow N-value are scattered, the maximum penetration per
(first impact) and the penetration increases blow is almost identical in all cases.
accordingly. The penetration at the time of the Attention should be paid to this from the stress
second impact is smaller than the maximum values 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)
3 FREQUENCY CHARACTERISTICS OF
STRESS WAVES
453
speed c- = ,/A J>= . i i 111
~ for the material, based on 0 1 ISkHz, respectively, assuming the wave
one-dimensional wave dynamics However, the propagation speed is = Jty. while the
speed depends on the shape of the members, and predominant frequencies shown in Figure 5 (a), (b)
decreases when two-dimensional wave dynamics are are calculated to be 0 19kHz and 0 074kHz,
applied For example, the speed in the hammer or respectively, which are about 65% of the natural
anvil would be about 70% of the one-dimensional frequencies
wave propagation speed for the material In Figure 5 , more than four prominent
The waves are partially reflected at the ends of frequencies can be seen including the predominant
the rod system and at the junctions between the frequency, these peaks are referenced to as A, B, C,
different cross sections so that various waves are D ..., from low to high frequency The natural
propagated in the rod following the hammer blow In frequency is located between the frequencies of the
order to examine the frequency characteristics of the peaks B and C According to the time-power
stress waves, the method of Wavelet transformation diagram (not included in this paper), the peak B
was chosen, with which it is possible to obtain a occurs about the time of the second impact between
thi-ee-dimensioiial diagram giving the frequency, the hammer and the rod
time and power spectra As shown in Figure 6, the reciprocal of the time
Figure 5 shows two frequency vs power between the first and the second impacts
diagrams obtained by Wavelet transformation of corresponds well with the fiequency at the peak B
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 4 RELATION BETWEEN POWER AND
PENET RATION
I Pztietration 71nm
001 01 1 10
Frecluel1cy (hHz)
(b) at a depth of 171n
Figure 6 Reciprocal of the tiinc bctnceii 1 “ and 2’ld 1111-
figiirc 5 Pciictration 1 s poucr spcclrum (Fqita 2000) pacts plottcd agaiiist fiequcnc! at peak B (Fujita 2000)
454
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
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
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.
457
where T, = energy content of the first compression where To = maximum energy applied to the rod-
wave, F(0,t) and v(0,t) = force and velocity at the sampler-soil system (ENTHRU), F(0,t) and v(0,t) =
section 0, and t, = cutoff time at the arrival of the force and velocity at the section 0.
first tension pulse. The maximum energy transferred to the top of
Figure 2 presents the corresponding force-time sampler-soil system is expressed by
diagram in the section x = 0 at the top of the rod.
(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.
Figure 2. Force-time diagram at the top of the rod.
kinetic energy
The efficiency factor q (Schmertmann & Palacios
1979, ASTM 1986 & Belincanta 1998) is
1'
-r
458
kinetic energy is dissipated at the end of the hammer where pmax = maximum downwards displacement, pe
impact. = recovered elastic displacement, and pp= permanent
Figure 5 presents the displacement - resistance sampler penetration.
curve in the section x = 0 at the top of rod-sampler- The recovered elastic displacement is correspon-
soil system. 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
Figure 5. Resistance - displacement curve at the top of the time when the compressive wave arrives the top
rod-sample-soil system. of the sampler-soil system, and t2 is the time when all
the kinetic energy is dissipated at the end of the
According to figure 5 the maximum kinetic energy hammer impact.
reaching the rod-sampler-soil system at the time tl is 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-
where T = total kinetic energy, and To = maximum tion energy is transformed into work done in the
energy applied to the rod-sampler-soil system permanent penetration of the sampler into the soil
(ENTHRU) . and in recovered elastic energy accumulated in the
The maximum downwards displacement (pmax) sampler-soil system.
corresponding to the maximum mobilized resistance The difference (To - TA) in figure 4 can be consid-
(R), occurs at the point B of the curve, when the ered roughly equal to the elastic energy accumulated
velocity becomes equal to zero and the kinetic energy in the rod length t.
T is transformed in potential deformation energy ac- When the rod length tends towards zero the ki-
cumulated in the rod-sampler-soil system, not taking netic energy TA tends towards To and the efficiency
into account others energies losses, 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-
where V = potential deformation energy accumulated mum and not a minimum as indicated in figure 3.
in the rod-sampler-soil system, equal to the area Figure 6 presents :he probable relationship be-
OBCO in figure 5. tween the efficiency q measured at the top of sam-
At the end of hammer impact, correspondent to pler-soil system and the rod length.
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
0 20 40 60 80 100
(9) 0
efficiency (Yo)
5
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
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)
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
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.
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 model parameters were also obtained via CAPWAP
indicated that the shaft resistances increase with match procedure. The database can firther be broken
increasing applied loads, with no signs of leveling off down into two sets. One set of data contains 208 pile
in the mobilization curves. As for the pile toe, he cases (120 individual piles) that have both dynamic
indicated that the trend sharply increases in a more testing and subsequent static load test results. The
like curvilinear manner with settlement measured at second set of data contains a total of 403 pile cases
pile toe. A summary of soil quake values as deduced with only dynamic pile testing results.
from the SPT-N, and pile tests is presented in Table To investigate the significance of soil type and
1, below. time of dynamic testing (i.e., end of driving- EOD,
and Beginning Of Restrike- BOR), a statistical
Table 1 Summary of Measured Quake Values as analysis of the Smith model parameters was
Obtained from the SPT-N. and Pile Load Tests * conducted on the Paikowsky’s database, particularly
Author the quake and the damping factor. Table 2 lists the
means and standard deviations of the Smith model
parameters for different soil types.
(1967)
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.
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
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
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 I 160.123 0.27 0.213 0.052 0.384 0.134
18 1 0.37 I 0.293 I 0.077 I 0.36 I 0.151 I 170.239 0.179 0.141 0.133 0.201 0.070
19 0.524 0.123 0.027 0.25 0.105 18 0.37 0.364 0.288 0.077 0.442 0.155
20 0.116 0.189 0.018 0.212 0.089 190.524 0.16 0.126 0.027 0.313 0.110
21 0.078 0.224 0.251 0.182 0.076 1201 0.116 (0.2471 0.195 1 0.018 10.275 1 0.096 1
22 0.06 0.141 0.03 0.232 0.097 21 0.078 0.274 0.216 0.251 0.233 0.082
23 I 0.305 I 0.151 I 0.036 10.318 I 0.134 22 0.06 0.18 0.142 0.03 0.293 0.103
A ~ ~0.177I I 0.181 I 0.111 l0.288 I 0.121 23 0.305 0.188 0.149 0.036 0.396 0.139
* Not included in the average value. AV 0.177 0.227 0.180 0.111 0.358 0.120
* Not included in the average value.
values for shaft damping factor J along the pile shaft
can be obtained. The predicted damping values in ( 5 ) . In this way, the predicted average value (0.121)
Table 8 and in the subsequent Tables are all length- is close to the C average value (0.1 11). The
weighted average values along the pile shaft. It can comparison between C N W M and predicted values
be seen that there is a difference in the shaft damping are shown in Figures 3 and 4 for shaft and toe
J between the C N W N results and the predicted damping, respectively. Figure 5 shows the sirmlar
values. Nevertheless, their average values are very comparison for the modified toe damping.
close. As for project 13, the shaft damping J from Figure 6 presents the correlation between the
WAd) is too big, so the shaft damping factor J of predicted damping factor J and §PT N values, where
s project is not included in the average v the damping values are shown for all SPT N values
Columns (3) and (4) in Table 8 are the Clap from all projects. The damping values were too high,
and the predicted values of the toe damping fa so author recommends that the damping J be
ly. Similar to the shaft damping, the modified with a modification factor of 0.75. Then the
values for toe damping shown in column correlations between the predicted damping J and the
(3) are the average values. As for the toe damping in SPT-W values are those shown in, in which, two
column (4), if there is more than one pile with approximate fitting curves for clay and sand are
different pile toe elevations, then average value is plotted. The equations for these two approximate
used. Similar to the shaft damping, the values for toe fitting curves are as follows.
damping factor J from CAPWAP for projects 1 and
11 are too, therefore, they were not included in the 0 .637 N
Clay
calculation of the average values. It can be seen that 62 . 5 + N
the difference between the CBBWAP results and the J = 3
predicted values is rather significant. The author 0 ,3375 N
Sand
recornmends that the calculated toe damping factor J 62 . 5 + N
being modified by a reduction factor of 0.42. The
modified toe-damping factor J is shown in column
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
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
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)
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 Common Soil and Rock Tests”, Proc. Of the
result, if the hammer efficiency is high, the increased Congress on Foundation Engineering: Current
energy is resisted by the dynamic soil resistance Principles and Practices, Evanston illinois,
indicated by an increased damping J. Table 7 along June, 1989, Vol. 2, 1026-10389.
with Figures 7 and 8 provide the results and 7. Terzaghi, K. and Peck, R. B. (1967). “Soil
correlation between Mechanics in Engineering Practice”, John
predicted damping for h Wiley and Sons, Inc., New York.
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.
COWCLUSIONS
467
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
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..
The SPT Working Party of the ISSMFE Technical For applying wave equation analysis to the SPT, LJto
Committee on Penetration Testing presented the et al.( 1974) proposed a model using the generalized
“International Reference Test Procedures on SPT”, solution of St. Venant assuming the reflection coef-
at ISOPT-1, Qrland in 1988. In Document 4- ficient to be the boundary condition at the bottom of
Calibration Methods, a method using stress wave the rod and that the rod is supported on rigid ground
equation analysis was introduced for the adjustment and struck by a hammer which is considered to be a
of the N-values in the SPT to account for different rigid body. A series of experiments were carried out
magnitude of driving energy, as a means of quality using an oscilloscope and strain gauges etc. for
control for the SPT. Many papers related to the ap- monitoring the stress waves and a high-speed movie
plication of stress wave theory to the SPT have been camera for recording the movement of the top of the
contributed and the measurements of the energy rod. The calculated and measured maximum com-
transfer ratio have increasingly been made in actual pressive stresses in the rod were 154 M N m 2 and
practice in the field in USA and other countries. about 84 MN/m’, respectively.
However, few high frequency measurements of the
movement or penetration per blow of the rod system
have been published or reported in connection with 3 STUDIES BY MTSUMOTO ET AL.
the stress wave equation analysis of the SPT.
The application of the stress wave equation ana- Matsumoto et al. (1992) have applied the method of
lysis to the SPT in Japan started after the study of two-point strain measurements proposed by Lund-
Uto et al. in the 1970‘s. Most of the studies carried berg & Henchoz (1977) to a series of instrumented
out include high quality measurements of the pene- SPTs that they carried out. The upper gauge was lo-
tration during the SPT in order to establish the cated on the rod at a point of 225 rnrn below the an-
mechanics or dynamics of the SPT which are not vil which had a height of 80 mm, and the distance
only for the purpose of quality assurance of N-values between the upper and lower gauges was 600 mm.
but also for attempting to find the shearing strength The total length between the top of anvil and the
ox characteristics of the soil. bottom of the sampler was 3.39mYthe hammer was
This paper will introduce a state-of-the art report operated by means of the rope and cathead method
on studies concerned with the application of stress and the energy transfer ratio was found to be be-
wave equation analysis to the SPT or related matters tween 0.6 and 0.9. The SPT was carried out based on
which have been done in Japan. These are mostly the Japanese Industry Standard (JIS) from a depth of
written in Japanese and include unpublished data. 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.
- 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
47 1
rnitted perfectly to the connecting rod system or
sampler.
wh = R,.S
472
inversely proportional to the energy transfer ratio ratios given by Kovacs et al. and others, by the
(4. method of the International Reference Test Proce-
Figure 8 given by Nagasaki et al. (1998) shows dures assuming the N-value is inversely proportional
three sets of measured data with different energy to the energy transfer ratio and the method proposed
transfer ratio being plotted on a Fujita’s diagram. by Fujita assuming the N-value is inversely propor-
Each was obtained by a blow of the SPT hammer at tional to the square root of the energy transfer ratio.
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- 7 RELATION BETWEEN THE ANGLE OF THE
duced intentionally by changing the free fall height HAMMER BLOW AND ENERGY TRANSFER
of the hammer. RATIO
The three plots in Figure 8 are located nearly on a
straight line which is orthogonal to the lines of the The energy transfer ratio of the SPT represents the
energy transfer ratio. This means that the magnitude efficiency with which energy is transferred from the
of the penetration is proportional to the square root kinetic energy of the hammer to the energy trans-
of the energy transfer ratio e, not proportional to e. mitted to the rod system, and it is generally under-
As the N-value is defined as the number of blows stood that its magnitude depends upon the method of
necessary to achieve a penetration of 300mm, and hammer release, e.g. larger for the trip method 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
473
smaller for the rope and cathead method. tween hammer blows should be determined prop-
Fujita et al. (1994) have frequently observed that erly.
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 9 INFLUENCE OF PORE WATER AND
of the hammer and the centerline of the rod are usu- OVERBURDEN PRESSURE ON THE N-
ally not in a straight line and usually are not com- VALUE
pletely vertical. The amount of the deviation is likely
to depend on the method of hammer release and Studies of the influence of water content and over-
workmanship of the operator, and these many affect burden pressure on the N-value are interesting issues
the magnitude of the energy transfer ratio. to be examined. Nagasaki et al. (1997) have contrib-
To examine the effects of angled contact during uted the results of a series of laboratory SPTs using
the hammer blow of the SPT on the energy transfer dry and saturated Toyoura sands with relative densi-
ratio by applying the wave equation analysis, a seri- ties of about 40%, 60% and 80% prepared uniformly
es of laboratory tests simulating the SPT were car- in a heavy steel tank, employing two types of pres-
ried out using a l m long instrumented steel bar with surized air bags for simulating overburden pressures
a diameter of 25mm, an anvil with a surface inclined of 100kPa, 300kPa and 5OOkPa.
at 0,1,2 or 3 degree and a steel donut type hammer Figure 10 gives the results of the test, and shows
weighing 49N provided with an inside diameter that the N-value increases with overburden pressure.
suitable for making contact at an angle. The results The N-value of the dry sand at relative density of
show that when the slope of the surface of the an\d 60% is approximately equal to that of saturated sand
is increased, the energy transfer ratio becomes and the N-values of dry sand at the densities of 40%
smaller as shown in Figure 9(a), and the penetration and 80% are smaller and larger than that of the satu-
of the rod per blow also becomes smaller as shown rated sand, respectively. On the other hand the N-
in Figure 9(b). Figure 9 indicated that the SPT appa- values are larger and smaller respectively, when
ratus should be set up accurately in a straight vertical another type of pressurized air bag is employed for
line in order to ensure the free fall of the hammer. the saturated sand as shown in Figure 11.
474
Table 3. Comparison of I'\;-valuc between dry and saturated REFERENCES
sands at relative dcnsity of 40,60 and 80%.(after Nagasaki-
et al. 19'97). Fujita, K., 0.Kusakabe. 1988. On the evaluation of static
DJ 40%
~ Dr- 60% D r - 80% bearing capacity. B.H.Fellenius(ed), Proc. 3"' Int.
Gibbs-l-loltz (1957) Nrr>N,. Nd> N , N(J>N, Con$ on Application of Stress-wave Theory to Piles.
Y anase (1963) N(s>NLV N ( J > N , Nct>N,br 525-534. Ottawa:
Yaniacla et al. (1004) Ni>N,, N(s=N,. Nis<N,. Fujita, K., R.Tanaka, M.Umemura, & T.Une. 1994. An-
Fu,jita et al. (1996) N(i>N,, Ncs'N, Ncr<N,. gled blow of hammer against head of piles. Proc. 49'"
Nagasaki et a1.(1'997) Nci<Nw N(s-=Nw Nrl>N,. Annual Con$ of JSCE. 3: 954-955. (in Japanese).
A',; : .Y-\~alucof d r y sand. N , : N-value of saturated sand Fujita, K. 1997. Referable N-values of unstandardized
standard penetration tests. J. of JSCE. 82(12): 24-27.
(in Japanese).
These results are summarized in Table 3, which Fujita, K. 1997. Mechanism of SPT, interpretation and
also includes other contributions by Gibbs & Holtz evaluation of N-values. Kisoko. 25( 12): 2-13. Tokyo:
(1957), Yanase (1963) and Yamada et al. (1994). Sogodoboku-kenkyusho. (in Japanese).
According to Table 3, there are two cases in which Matsumoto, T., H.Sekiguchi, H.Yoshida, & K.Kita. 1992.
the N-value of dry sand is larger than that of saturat- Significance of two-point strain measurement in SPT.
ed sand, and three other cases in which the N-value Soils and Foundations, (32)2: 67-82.
of dry sand is approximately equal to that of saturat- Nagasaki, T., K.Fujita, T.Une. 1996. Dynamic resistance
ed sand at relative density of 60%, while it is either of SPT in dry sand subjected to overburden pressure.
larger or smaller at the relative density of 40% and Proc. 51" Annual Conf of JSCE. 3A: 758-759. (in
smaller or larger at the density of 80%. Japanese).
Table 3 suggests that either the N-value is not al- Nagasaki, H., K.Fujita, Y .Imamura, K.Nohara &
ways proportional to the effective stress in the model K.Yamamoto. 1997. N-value, energy transfer ratio and
ground or the N-value of the saturated sand is sig- dynamic resistance in saturated Toyoura sands. Proc.
nificantly affected by the type of air pressure bags 32"d Annual Con$ on Geotechnical Engineerng. JGS.
employed for the laboratory tests. 191-192. (in Japanese).
Nagasaki, H., K.Fujita, T.Yoshinaga & H.Yamaguchi.
1998. A rational adjustment of SPT N-value by square
10 CONCLUSIONS root of energy transfer ratio. Proc. 53rdAnnual Con$
of JSCE. 3A: 762-763. (in Japanese).
The application of stress wave equation analysis to Uto, K., M.Fuyuki, H.Kondo, M.Morihara & H.Matsu-
the SPT is very useful for the quality control of the mura. 1974. Consideration of N-Value and strength of
SPT as well as for the quality assurance of the N- ground from a view-point of wave theory. Proc. Fac-
value. The laboratory experiments on the SPT and ulty of Engineering, Tokai University. (in Japanese).
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
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
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.
479
Figure 2. Installation and design process of vibratory driven
piles
480
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.
48 I
the sinusoidal force described by eq. (1) whereas the O’Neill and Vipulanandan (1 989) provide an expres-
mass Mib is subjected to gravity and the suspension sion of the theoretical power required to maintain the
force T. The net quasi-stationary action on the pile vibrating regime of a dynamic mass in the absence of
resulting from the carrier operation and vibrator is soil reaction but accounting for the presence of the
the weight of the vibrator mass and its clamp Mcl de- isolating spring and the bias mass Mib. That formula
ducted by the suspension force: 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
2.3 VibratorMovemerit driving analyses conducted to date.
The movement of the vibrated body will depend on
its so-called dynamic mass and the soil resistance. 2.4 Types of vibrators
Specifications of vibrators often list a “maximum
amplitude” S,. That number [generally expressed in Two main types of vibrators are commercially avail-
mm] corresponds to the total (i.e. double) amplitude able: hydraulic and electrical. In both cases, the
of movement for a free hanging vibrator, thus as- motor is housed in the vibrator and powered through
suming a dynamic mass consisting of the exciter a transmission line connected to a separate or carrier-
block M,b and the clamping device ml: mounted diesel-hydraulic or diesel-electric power
pack (see Fig. 3a). Hydraulic vibrators are lighter
Ssp = 2 so = 2 nze/(Meh + Mc,) than their electrical counterparts, because of the
smaller size of the motor. The adjustment of the op-
It should be noted that the double amplitude does erating frequency is more readily available on the hy-
not depend on the operating frequency, as the center draulic vibrators, which also explain why they are
of mass of the free mechanical system remains sta- more commonly used.
tionary, irrespective of the frequency. The amplitude Five types of vibrators can be distinguished based
of the free hanging pile to be vibrated will always be on operating frequency and eccentric moments, as
smaller than the specified amplitude, as can be de- summarized in Table I .
rived from eq. (4a), where the dynamic mass is in- It can be noted that initial improvements of the vi-
creased by that of the pile (PAL). 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
with s = actual (single) amplitude of the dynamic
can adjust on-the-fly its effective eccentricity by
mass. shifting the phase angle between a multiple of 4
A power for the vibrator is often listed in the masses. The claimed advantage of such an adjust-
specifications. It generally corresponds to the nomi- ment is to avoid “soil resonance”, a term coined after
nal power of the motor actuating the eccentric the observation that vibration levels pass through a
masses. It does not correspond to standardized op- peak upon vibration start-up and shut-down. This
erational conditions of the vibrator in action. Power phenomenon will however be shown later not to be
consumption is indeed dependent upon testing con- necessarily related to a particular fkequency.
ditions. Barkan suggests that under pile vibratory Vibrator choice amongst practicionners is gener-
conditions, the power follows a squared velocity law: ally based on experience and field verification.
W[kW] = M,, . n . ( s w ) ~=Pt.K Rodger and Littlejohn (1980) have summarized that
body of experience into a table recommending fi-e-
Experimental verification of that law shows the n quency and amplitude parameters for different piles
value to depend on soil type and pile type; a range of and soil types. Those recommendations are repro-
15 to 50 Hz is observed. duced herein as Table 2.
Type Frequency range [rpm] Eccentric moment Maximum centrifugal Free hanging double
[kg.ml force FNJ amplitude [mm]
“Standard frequency” 1300-1800 up to 230 up to 4,600 up to 30
High fTequency 2000-2500 6 to 45 400 to 2,700 13 to 22
Variable eccentricity 2300 10 to 54 600 to 3300 14 to 17
Excavator accessory 1800 to 3000 1 to 13 70 to 500 6 to 20
Resonant dnver 6000 50 20,000 (in theory) Self destructing
482
Cohesive soils Dense cohesionless soils 1 Loosc cohesionless soils
All cases Low point resistance Ihgh pomt resistance ] Heavy piles Light piles
wlgh acceleration &gh acceleration Low frequency. Large displacement amplitude IIigh acceleration
Low displacement amplitude
Predominant side resistancc Predominant side resistance. Predominant end resistance Predominant side resistance.
Requires high acceleration for Iiequlres high acceleration Require? lugh displacement aniplitude and low Requires high acceleration
either sheanng or thixotropic for fluidization frequency for maulmum impact to permit elasto- for fluid17ation
transformation plastic penetration
483
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):
484
'kin = Laboratory results conducted at constant cyclic
strain show that in many soils, the degradation index
with Tkin = kinetic ultimate
shear strength [ P a l after N cycles can be approximated by the following
xsta = "static" ultimate
shear strength [Wa] relationship as suggested by Idriss et a1 (1 978):
= shear strain rate [s-]]
The advantage of that mathematical form is that re- A = A'-' (10)
sistance does not vanish as the strain rate goes to- The exponent t, called degradation parameter, de-
wards zero. The power law also requires the strain pends mainly on the amplitude of the cyclic strain and
rate to vary by orders of magnitude to provide tangi- the nature of the material (PI), as suggested by Do-
ble increases in both the modulus and the ultimate bry and Vucetic (1988) and as indicated in Fig. 6
strength. The J coefficient and n exponent depend on Vucetic, 1993). It is noteworthy that the degradation
the nature of the soil. Based on pile driving data, parameter assumes a zero value at strains smaller
than a cyclic "threshold" shear strain, y,, The thresh-
n=0.2 and J=0.3 have been suggested for plas-
old strain increases with the plasticity of the soil, as
tic soils. J should therefore essentially depend on the suggested in Fig. 6.
plasticity of the soil and become quite limited for
granular materials.
3.7 Soil liquefaction
Vibration induced compaction of saturated sands has
3.6 Degradation Law received attention not only from the earthquake en-
When subjected to undrained cyclic loading in- gineering community, but also from vibro-
volving a number N of large strain cycles, the soil compaction specialists.
structure continuously deteriorates, the pore pressure Recent advances tend to indicate that build up of
increases, and the secant shear modulus decreases pore pressures (eventually leading to liquefaction)
with N. This process known as cyclic stiffness deg- and volume reduction of cyclically loaded materials
radation can be best characterized on the basis of are the expression of the same phenomenon, i.e. the
strain controlled tests for the type of loading in- irreversible tendency for a particulate arrangement to
volved with the vibratory penetration of piles. Typi- achieve a denser packing when sheared back and
cal results of strain-controlled tests are sketched in forth.
Fig. 5 , where the degradation is clearly expressed by Under drained conditions, the volume reduction is
the decrease of the amplitude of the peak stress mo- immediate. Under undrained conditions, the tendency
bilized at successive cycles. for volume reduction is expressed by an increase in
The quantification of the degradation process calls the pore water pressure (see Fig. 7), such that the
for the introduction of the degradation index A, de- effective stress is reduced to a value that may be
fined by: close to zero. It is then necessary to wait for the soil
to consolidate in order to see the volume reduction
(9) 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 6. Effect of Plasticity Index (PI) on soil degradation Figure 7. Build up of residual pore pressure in different sands
(Vucetic, 1993) in undrained cyclic strain-controlled tests (Dobry et al., 1982)
485
to the problem of the vibrations induced by a verti- 4.1 Conceptual model testing
cally vibrating pile. It also allows one to evaluate Tests have been conducted by several Russian re-
potential changes of the void ratio based on a cyclic searchers to investigate the “vibro-viscous” resis-
strain rather than stress history, as supported by labo- tance of soils. In particular, Barkan (1963) reports
ratory drained tests conducted on sands by Youd on the sphere test, shown in Fig. Sa, where a steel
(1972). That framework of analysis entranced by the ball is sunk into a vibrated soil vessel with the assis-
threshold cyclic strain concept embodies in a single tance of a bias force. Penetration speed is shown to
model the intrinsic relationship between degradation obey Stokes sedimentation law (see Fig. 8b), allow-
and pore pressure build-up, with the advantage that it ing one to determine an equivalent viscosity p . The
can be applied to general categories of soils (sands to inverse of that equivalent kinematic viscosity
clays) [cm.s/kg] was shown to vary linearly with the rela-
The excess pore pressure generated during cyclic tive level of acceleration (a/g), passed a threshold
loading has been shown (see Fig. 7) to increase with value of approximately 1.4 for a dry sand (see Fig.
the shear strain and the number of cycles for a given 8c). The influence of the water content on the “vibro-
soil type. The damage parameter K approach (Finn, viscosity factor l/p” of a sand vibrated at constant
1981) can be adopted to evaluate the excess pore a/g is also shown in Fig. 8d, highlighting the near
pressure 6u resulting from a particular strain history, total loss of vibro-penetrability at optimal water
as characterized by the following equations : content.
d~ / d o = A / 4 In (1 + rd2 )
* (11)
with Relative Energy Loss given by Eq. 5, and 4.2 Pile-soil interface testing
K =<ex’ (16) with C = 5 and Soil shear strength resisting the pulling out of a vi-
5 = length of strain path (12) brating steel plate against a normal stress controlled
= 4 N yc, for constant amplitude cycles medium sand (vibratory direct shear box) has been
investigated in the early days by Levchinsky and
Savtchencko (Barkan, 1963). The friction coefficient
4 PILE VIBRATORY DRIVING TESTING (tan @ = do) was shown to decrease with cyclic am-
plitude and frequency. The ultimate relative reduc-
The above discussion of soil behavior under cyclic tion of the friction was also shown to increase with
loading does not encompass the particular geometry the grain size within the investigated range shown in
of the profile-soil interface, nor does it consider the Fig. 9. Fig 9 shows that the sand vibratory friction
continuous penetration of the profile that leads to angle can easily drop to !h to 1/5 of its static value.
successive exploration stages into “virgin” soil be-
havior. That is why a number of experiences have
been conducted to reveal soil-structure interaction 4.3 Reduced scale tests
w i t h a vibratory framework. Based on the ambition Testing of model profiles in soil tanks were initially
and complexity of the tested interface, one can cate- attempted by Bernhard (1968), Schmid and Hill
gorize various experiences reported in the literature (1966), continued by Rodger and Littlejohn (1980),
as conceptual, interface, and both reduced and full- Billet and SBert (1 985) and O’Neill et a1 (1990),
scale testing.
and more recently by Viking (1 998) and Holeyman et 4.4 Full scale tests
a1 (1999). The tests were generally conducted using a Because of inconsistencies in the conclusions derived
lightweight vibrator acting on a heavily instrumented from reduced scale tests, research has been con-
profile. Monitoring included strain gauges, acceler- ducted in several countries based on full-scale tests.
ometers and displacement transducers. The soil used Early hll-scale programmes have been conducted by
was exclusively sand (dry, moist, or saturated), Barkan (1963) and Davisson (1970). Other pro-
placed at a controlled density, and in some cases, grammes have been conducted by manufacturers on
confined at a controlled radial stress. Monitoring of specific equipment, but lead to a limited diffusion of
the soil response involved accelerometers, total stress their conclusions. More recently, collective European
and pore pressure cells during installation as well as programs have provided actual penetration speed,
compactlon and-in situ testing ager instalYation. but within soil conditions that cannot be controlled,
Insighthl observations relative to the vibratory toe only characterized. Monitoring nowadays involvies
resistance have been reported by Schmid (1966), acceleration, strain, pore pressure, penetration speed,
who identified three regimes, depending on the mag- making the tested profile a hlly instrumented probe.
nitude of the driving force : Such programs have produced results that have not
a sinusoidal resistance domain, for a driving force been hlly analyzed (BBRI, 1994, Sipdis, 1997); oth-
lower than the “resistance threshold” ers are being presently conducted (KTH, 1999) or
an impact domain, when the upward force ex- planned (IREX, 2000). Publication of such research
ceeds the soil uplift resistance; the toe of the pile results would be appreciated by the profession.
alternately separates from the soil and tamps it
a phase instability domain, when the downward
force exceeds the soil compressive resistance. 5 PILE-SOIL-VIBRATOR INTERACTION
Rodger and Littlejohn (1980) call upon accelera- MODELS
tion amplitude to distinguish:
- the elastic state (a<0.6g) 5.1 Types of models
- the trans-threshold state (0.6g<a<l.5g), wherein
Models that have been suggested by various authors
most of the shear strength reduction takes place, differ in the way they account for mechanical engi-
- the fluidized response state (a>l.5g).
neering principles. We will review models purely
Although their views are contradicted by some of based on (1) force equilibrium, (2) momentum con-
Barkan’s observations, these three different states are servation, (3) energy conservation, and (4) integra-
stated to be confirmed by dynamic direct shear tests tion of the laws of motion.
performed by others.
Results of‘tarik experiments’have’ been reporte‘d’in
terms of friction reduction coefficients, potential op- 5.2 Force equilibrium models
timal operation, and have shed some light on hnda- The force models aim at predicting whether a vibra-
mental soil behavior. Correlations have been estab- tor can or cannot overcome an estimated soil resis-
lished between penetration speed and parameters in- tance. They will not provide an estimate of the driv-
duced by the vibrator (amplitude, frequency) and by ing speed. Jonker (1987) and Warrington (1989)
the soil (grain size, relative density, and lateral have suggested, respectively:
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 force generated by the vibrator, per eq. (1)
for the toe resistance inertia forces of dynamic mass, = Mdp . a
optimum operation of the hammer required at surcharge force, per eq. ( 3 )
times that the frequency or eccentric moment be empirical factor of shaft resistance outside
reduced, while energy transfer was of the order pipe pile,
of 40% of the hll theoretical power produced by soil resistance outside pile shaft,
the vibrator empirical factor of shaft resistance inside
a number of observations cannot be explained. pipe pile,
Although reduced scale models are of use, they soil resistance inside pile shaft,
suffer from improper boundary conditions (at the soil resistance at Dile toe.
tank limits) that significantly prevent the vibration For sheet-piles Tunke; Company recommends to
energy from propagating away from its source. replace x. z [m2] with 2.81 times the sheet-pile
width.
487
5.3 Energy based models 5.5.1 Single degree of freedom (SDOF)
Simplest models of the vibrator suggest that the dy-
Energy based models assume the following general
namic mass be the focus of attention, thereby as-
form:
suming that the pile behaves as a rigid body. New-
R.v, = p, .W, + ( F , +F').v, (1 4 4 ton's second law can therefore be applied to the
dynamic mass :
leading to a direct estimate of the penetration speed :
mem2sin(wl)
vP=p, W,/ ( R -F,
* -F,) (14b) a= (174
hrfIn,
With:
R = soil resistance, where
v,, = average rate of penetration in m/s, U
,, =Meb +Md+ M , ( 17b)
W, = theoretical power delivered to the system,
Fi = inertia forces of dynamic masses. Holeyman (1993) has suggested a method that
Davisson's formula (1 970) to estimate the bearing integrates the inertial effects of the excess force. That
capacity for the Bodine Resonant Driver suggests : excess force is defined as the difference between the
sinusoidal driving force and the opposing soil resis-
p, =I--v*s,. R/1000 w, (15) tance. A distinction is made between the skin ftiction,
where s,: is an empirically determined set [mm/cycle] which is reversible (Uplift resistance = Downward
representing all energy losses. resistance) and the toe resistance, which cannot pro-
Warrington (1989) has coined eq. (14b) as the duce uplift resistance. Attention is also paid to the
'Vibdrive' formula provided a value of 0. I is used for clutch resistance, which is combined with the skin
p, and the power Wt is calculated according to his friction.
procedure, 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
5.4 Momentum conservation models the severity of degradation. The method involves an
Schmid ( I 968) has suggested a formula implying iterative procedure to identifL the coexisting accel-
that, for steady-state penetration, the momentum of eration and soil resistance (17b). The driving speed is
the total mass of the vibrator. additional bias mass obtained by intuitively integrating the net downward
(MJ, and pile accrued by gravity over a vibration cy- and upward accelerations over a complete cycle. The
cie be balanced by the soil resistance impulse: method have been verified and liquefaction parame-
ters hrther refined through calibration with full-scale
(Ads+hd,,,b+M,)g T= 1 6;Rdt=&T, (164 tests (BBIPI, 1994)
Gonin (1998) has followed a similar approach that
with T, = contact time between pile toe and soil analytically integrates the effects of an excess force,
as shown in Fig. 10. The integration is however per-
within a cycle and a = coefficient between 0.5 and 1,
formed soleby on the toe resistance, while the skin
generally assumed to be 213. friction influence is accounted for in terms of damp-
Conversely, the penetration speed follows a linear ing of the driving force. In addition, the wave equa-
trend passed the threshold acceleration amln,which tion theory is used to estimate the displacement ac-
becomes a key parameter to successfdly apply the crued at the toe over the period of net force
niethod and estimate T, : exceedance.
489
At the shear modulus of the soil in contact with Moulai-Khatir et al. (1994) have developed to-
the profile is nearly zero, and very little energy can gether with the University of Houston, the so-called
pass through the fluidized surrounding zone. The soil VPDA computer program (for Vibratory Pile Driving
in the vicinity of the profile cannot anymore follow Analysis) wherein the action of the hammer is re-
the profile movement, from which it uncouples itsel, placed by a static surcharge load and a sinuso‘idal
resulting in a lower level of vibration. That example load. The soil model was modified from Smith‘s
demonstrates that apparent resonance of soil vibra- original in that hyperbolic mobilization curves were
tion may be no more than the transient combination adapted for the shafi and toe resistance, as shown in
of increased rotation speed and soil degradation. The Fig. 15. A simple viscous damper was used to model
model can also shed light on “damping” as it clearly damping along the shaft, while no damping was
separates geometric damping from the energy losses deemed necessary at the pile toe.
attributable to viscous and hysteretic behavior. It should also be noted that the GRLWEAP pro-
gram has included in its latest versions (GRL,, 1998)
5.5.3 Longitudinal I-D models the capacibilty to model vibratory hammers.
Few authors have adapted Smith‘s (1960) classic
lumped parameters model to represent the longitudi-
nal behavior of a pile subjected to vibratory driving. 6 VIBRQ-DRIVABILITY ANALYSIS
Gardner (1981) and Chua et al. (1981) have devel-
oped a wave-equation computer code where the vi- Most of the models discussed in the previous sec-
brator is represented by a two-mass system, sepa- tions should be able to provide a reasonnable match
rated by a sofi spring, while the exitor black is of calculations with relevant field observations pro-
subjected to a sinuso‘idal force (cfi-. eq. (I)); as vided the models parameters are properly calibrated.
shown in Fig. 14. The soil behavior is represented by The use of energy balance methods is discouraged
spring-slider-dasplot systems, according to Smith’s by the author, while force equilibrium methods are of
early suggestion. limited use because they do not provide vibratory
Middendorp and J o k e r (1988), as well as Ligterink penetration speed. Momentum based methods m y
et al. (1990) used the TNOWAVE computer pro- produce a penetration speed very similar to that ob-
gram to analyze the driveability of offshore vibratory tained through integration of the laws of notion of a
driven pipe piles, based on the methods of character- rigid body. Finally, wave equations methods should
istics. The authors identify the need for a soil model not produce penetration speeds significantly different
able to describe the degradation of the soil resistance form those obtained form a rigid body analysis, pro-
as a function of the oscillation history, and warn that vided the vibrator speed is lower than the resonant
soil parameters may depend on opera- ting fiequency frequency of the pile, which is generally the case.
and pile movement amplitude.
Figure 14. Longitudinal I-D model Figure 15. Resistance mobilization for (a) skin friction (b) toe
compression (After Moulai - Khatir et al., 1994)
490
Exceptions to that general case include the Bodine At each depth z the vibratory pile driving resis-
Resonant driver and very long piles for offshore ap- tance is calculated :
plications (LMO m).
In the author’s opinion, the most critical parameter Rbase =qd
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
where $2is the pile section, x the pile perimeter and
and recent consistent experimental data is cruelly
missing. The reliability of the predicted vibro- D the pile penetration.
penetration log will strongly depend on the degrada-
tion parameters adopted to assess the vibratory
penetration resistance form the soil investigation re- 7 BEARING CAPACITY FROM
sults. The author’s experience leads him to use the INSTALLATION MQNITOR1ING
following crude ultimate degradations coefficients :
0.15 in sand, 0.4 in silt, and 0.65 in clay for skin fric- Because soil resistance degradation is significant
tion; as well as 0.55 in sand, 0.7 in silt and 0.85 for during vibratory driving, one should expect it a chal-
end bearikg. lenge to estimate the static bearing capacity from the
A more involved assessment of the degradation end of penetration vibratory performance of a driven
coefficient has been suggested (Holeyman, 19%) profile.
based on CPT test results. In that method, the soil In the impact driving practice, it is recognized that
driving resistance is obtained by interpolation be- end of driving (EQD) data generally provides a safe
tween a static value and an ultimately degraded estimate of the pile capacity; that is why beginning of
value. The static base (qs) and shaft (zs) resistance restrike (BQR) or “retap” data is strongly advised to
profiles derived from Cone Penetration (CPT) tests the owner who wishes to tap the value of letting the
results, i.e. from the cone resistance qc and local unit soil set up. If the end of Vibratory driving (EOV)
skin friction f, (El cone). data is used, methods to estimate the static capacity
The ultimately liquefied base (ql) and shaft (z1) should allow for recovery of soil degradation, as
unit soil resistances are derived based on an expo- highlighted in Fig. 2. However significant uncer-
nential law as expressed below : 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),
rl=r,[(l-1/ A).e-l”;R+1/A] (1 8b) formulae, even more so than already much detracted
where (impact) pile driving formulae. A limited number of
= liquefied soil base resistance [Wa] such PVD formulae have been published; however
41
21 = liquefied soil shaft resistance [@a] only one has been, to the author’s knowledge been
FR = friction ratio as measured in a CPT test with extensively field tested. The “Snipe” formula is
E l cone (percentage of the mantle friction to therefore the only one that will be discussed in this
the cone resistance, i e FR = 100 f, /qc) paper.
A = empirical liquefaction factor expressing the The formula is a field-based method was devel-
loss of resistance attribuable to liquefaction oped in the former Soviet Union according to
(A will be higher for saturated and loose Steffanof and Boshinov (1 977). The following em-
sands and is chosen in the range of 4 to 10) pirical formula is used to predict the static bearing
The driving base (qd) and shaft (zd) unit resis- capacity Q,, :
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
QU = load capacity, in [kN];
w = power used by the vibrator to drive the pile,
in [kW]
where Ft = total weight (force) of vibro-hammer and pile,
qd = driving base unit resistance in [kld]; = (M,ib + MJg
zd = driving shaA unit resistance 1/p = empirical loss coefficient (in Soviet practice
a = acceleration ratio (= dg)of the pile, as ob- 143 is safety taken to be 5 in cohesionless
tained from Eq. (1 7a)
49 1
soils) reflecting the influence of driving on drivability. The observed frequency was 38 fi.
soil properties. From the measurement results, one can observe
The bearing capacity in of vibro-driven pipe piles that :
has been verified by A monitoring the behavior of - the vibration amplitude at the pile top (0.65 mm
the finished product the beginning of impa zero to peak) is considerably less than the nominal
strike. This of course requires that a specific vibration amplitude which is,
procedure be enforced a sufficient time aft
EOV installation, in order to allow pore pressures to me
-- -
me -
26000kg.mm
= 2.3mm
dissipate and accrue soil setup. M +M (6500 + 4820)kg
The monitoring of the installation of vibratory - the amplitude at the pile base (0.45 mm zero to
driven piles is not at all as widely spread as for im-
pact driven piles. Recent improvements in the field peak) is smaller than the amplitude at the pile top
monitoring devices FaSA., “NO-System, etc) now (0.65 mm)
allows the geotec’nnical engi It would appear that the pile base amplitude (0.45
celeration, stress and energy mm) is not sufficient to allow the pile to penetrate as
equivalent interpretation of the stress-strain behaviour for clayey soils is primarily
formula or CAPWAP method, available for nearly 15 elastic for small amplitudes. ossible explanations
years in the impact driven products. for that observation were considered :
- An important soil (i.e. clay) mass was sticking to
Field monitoring provides a tremendous advan-
tage in controlling the effective performance of the the vibrating pile, leading to a more important vi-
vibratory hammer, as illustrated by the following case brating mass, leading to a smaller vibration am-
history reported by Holeyman et al. (1996). plitude.
- The vibratory hammer was unable to deliver the
That case involved the installation of a 20.6 m
long tubular steel pile with a thickness of 9.5 mm and required energy, and thus maintain its nominal
a diameter of 1 m on a site in Kortrijk ( plitude or frequency. A characteristic of the
shows the subsoil profile as depicted by a CPT test C variable eccentric hammers is that a lack of
(cone M4) performed at the site. The water table power results in a reduction of vibration amplitude
was encountered at a depth of -1.8 m. The upper (rather than a reduction of frequency (Houze,
twelve meters consist of very soft river deposits; be- 1994)).
- A smaller amplitude at the pile base was obtained
low the underlying sandy layer was found a very stiff
tertiary clay layer in which the tube had to be driven. due to the elasticity of the pile.
A preliminary calculation using eqs. (1 7) through By applying the observed vibration amplitude to
(20) pointed out that the necessary time to install the the calculation model (Figure 181, a much better cor-
pile to a depth o f 2 0 m with a PTC 3 0 W V vibratory relation between the calculated and the observed
hammer was 13% minutes. However, driving met penetration time was obtained. The pile was placed
rehsal at a depth of 11 m. at the bottom of an excavation at -2.5 m and pene-
The reason for the difficult driving and the differ- trated 4.5m under its own weight. As a result, ob-
ence between the predicted and the observed pene- served and calculated penetration rates are reported
tration speed was explained by measurements taken starting at level - 7 m. Figure 18 evidences that the
during the actual driving of the pile. The pile vibra- difference for the predicted and observed penetration
tion amplitude was measured by means of a velocity times for the site in Kortrijk was not due to an incor-
transducer placed at the pile head and a velocity rect estimation of the dynamic soil resistance but due
transducer (protected by a cover) at the pile toe. to an incorrect estimation of the vibration amplitude,
Figure 17 shows the monitored amplitude of vibra- which happened to be limited by the nominal power
tion at the pile top and at the pile base upon loss of of the power pack. A more powefil power pack
Figure 16. Subsoil profile site at Kortrijk Figure 17. Record from the vibration amplitude upon refusal
492
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
Figure 18. Predicted and Observed penetration log at Kortrijk
site compared with prehcted log using actual vibratory am- zur Beschreibung des Vibrationsrammens in komngen
plitude Bodcn" Doctoral Thesis, University of Karlsnihe. Ger-
many
Dobry. R , Ladd. R S . Yokel. F V . Chrang. R M and Fowl!.
was brought on site and the piles could be vibrated to
D (E982) "Prechction of Pore Water Pressure BuiPdup and
design depth using the same vibrator.
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
After reviewing the present state-of-the-art of vibra-
52 1-525,
t o driving,
~ ~ the following suggestions for krther
Dobry. R and Vircetrc M (1987) "State-o€-the-Art Report
consideration are offered '
Dvnamic Properties and Response of Soft Clay Deposits"
soil mechanics research is needed in the area of
Proceeding\ of the In11 Svmponurn on Geolechnicnl &,a
large cyclic deformation to better understand and
of.S'ofi Soils. Mexico City. Vol 2. pp 51-87
assess the effects of degradation and liquefaction
Drnevrch I J P Hall. J R . Jr , and kchart. F E Jr (1967)
under those extreme conditions,
"Effects o f Amplit~dcof Vibration on the Shear Modulus
full scale vibratory driving tests, with extensive
of Sand Proceediiigs of the Iiifernatroi?al Svmposium on
field monitoring, will be required preferably to
Fave Propagatioii nnd Dvnat~iicPropertier of Earth hfu-
reduced scale laboratory tests, which suffer from
ferialr, Albtrquerque. N M . pp 189-199
improper energy dissipation boundary conditions,
potential and transferred power of vibrators need Finn. Vd D L (1981) "Liquefaction Potential Developments
to be better defined, as well as modeled for better Since 1976". Proceedings, Intl Conf oii Recent Ad-
description of the mechanical behavior of vibra- vaiices in Geofechnical Earthquake Eiigineering and wrl
l A ~ i i a ~ ? ~Sifc sLouis,
, Missouri. Vol I€, pp 655-681
tois,
peak vibration of the soil surrounding a profile Garclner SherriIl. (1987) "Analysis of vibratory driven pile"
upon vibrator start up does not necessarily imply ~ r o cof 2'ld Inl ~ o n Of M ~ e e ~orni~atron,
p Luuenibourg,
soil resonance; it can also result from the combi- 5-7 May, pp 29-56
nation of increasing fi-equency and degrading soil Gonin. J (1998) QueIques reflevions sur Ie vibrofonqage. Re-
vue Fraiqai re de Geoiechnique, No 83. 2'"Ie trimestre. pp
resistance,
monitoring of vibrated profiles is recommended 35-39
with the view to emulate the benefits accrued by Harchn. B 0 and Black, W L (1968) "Vibration Modulus of
a similar practice for driven profiles, and Normally Consolidated Clay mrnal of the Soil Ale-
chmnc I aim' Fowm'atio~rsllil n, ASCE. Vol 94. No
procedures for vibratory loading tests should be
SM2. Proc Fapeer 5833. pp 353-369
developped
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
NCES
August 1985. Vol 1. pp 173-176
Barkan. D D . (1963). Methodes de vibration clans la con- Holeyman. A (1988) "Modeling of Pile Dvnainic Behavior at
struction. Dunod. Fans, 302 p (Plreiich traidatioi? of the Pile Base during Driving," Proceedings of the 3rd Iii-
Originnl 117 1?uwa~i" Vibsometod v StroitelJSlVe. 1960) feriiafionnl Confereiice on the Applicnfion of Stress-Ffni e
BBRI (2994) High performance vlbralopv pile h v e r s base on Thenrv fo Piles. Otlataa. May 1988. pp 173-285
novel electromagnetic actuation systems and Iinproved UiI- Holeyman. A (19932) "WUPERVIBI An analytical model-
derstandng of soil dynamics, Progress reports of the based computer prograni to evaluate the penetration speed
BR17I9ElJRAhl research contrucl CTYI-0.561, 19% of vibratory dnven sheet Piles". Research report prepared
Ben&arcL R K . (1967), Fluichzation phenoniena ~n soils dur- for BBH. June. 23p
ing vibro-cornpaction and vibro-pile-hvmg and -prrllrng Holcymamr, A (1993b) "HYPERVIP3IIa. An detailed numenca!
Hanover. NH 1967. 58 pp CIS Alrnzv Cold Xeg1on.s Re-
~
model proposed for Future Computer Implementation to
search and Engineering evaluate the penetration speed of vibratory chrrven sheet
493
Piles", Research report prepared for BBRI, September, 316>National Research Council, Washington, DC. Vol. 1
54p. pp. 1-51. ISBN 0-309-04613-0
Holeyman, A. & Legrand C. (1994). Soil Modelling for pile ONeilI Michael W., Vipulanandan, C., Wong, D., (1990).
vibratory dnving, International Conference on Design and "Laboratory modelling of vibro-driven piles". Journal of
Construction of Deep Foundations, Vol. 2, pp 1165-1178, Geotechnical Engineering, ASCE, Vol. 116, No 8: pp.
Orlando, U.S.A.. 1994. 1190-1209
Holeyman, A., Legrand, C., and Van Rompaey, D., (1996). A Roberlson, P.K. and Wride, C.E. (1998). "Evaluating cyclic
Method to preQct the driveability of vibratory dnven piles, liquefaction potential using the cone penetration test", Ca-
Proceedings of the 3rd International Conference on the nadian Geotechnical Journal, No. 35, pp. 442-459.
Application of Stress-Wave Theory to Piles, pp 1101-1112, Rodger, A.A. and Littlejohn, G.S. (1980). "A study of vibra-
Orlando, U.S.A.. 1996. tory h v i n g in granular soils". Geotechnique, Vol. 30, no
Howe, C. (1994). HFV Amplitude control vibratory ham- 3, pp. 269-293.
mers : piling efficiency without the vibration inconven- Seed, H.B. and Idnss, I.M. (1970). "Soil Moduli and Damping
ience, in DFI 94, pp. 2.4.1 to 2.4.10, Proceedings of the Factors for Dynamic Response Analyses." Earthquake En-
FiBh International Conference and Exhibition on Piling gineering Research Center, College of Engineering, Uni-
aiid Deep Foundations, Bruges, Belgium, 1994. versity of California, Berkeley, Report No. EERC 70-10.
Idnss, I.M., Dobry, R, and Singh. R.D. (1978). "Nonlinear Seed, H.B. and De Alba, P. (1986) YJse of SPT and CPT
Behavior of Soft Clays during Cyclic Loadmg." J. Geo- Tests for Evaluating the Liquefaction Resistance of Sands"
technical Engineering Div. ASCE, 104(GT12), pp. 1427- Proc. INSITU 86, V A 22 p.
1447. Schmid, W.E., (1969). "Driving resistance and bearing ca-
IREX (1998), VibrofonGage des pieux et palplanches - Etude pacity of vibrodriven model piles". American Society of
exploratoire, (by Le Tirant, P., Borel, S., Gonin, H., Guil- Testing and A4aterials Special Techn., Publ. 444, pp. 362-
laume. D. and Longueval, A.) 375
Jonker, G., (1987). "Vibratory Pile Driving Hammers for Oil Schmid, W.E. & Hill, H.T. (1967). "A rational dynamic equa-
Installation and Soil Improvement Projects". Proc. of tion for vibro driven piles in sand". Symp. Dynamic Prop-
Alineteenth Annual Ofshore Technology Conf:, Dallas. erties of Earth Materials. New Mexico Universityy:p. 349.
Texas, OTC 5422, pp. 549-560. Smith E.A.L., (1960). "Pile-dnving analysis by the wave
Kondner, R. L., (1963). "Hyperbolic Stress-Strain Response: equation". Jourrzal of the Soil A4echanics and Foundatioris
Cohesive Soils." .Journal of the Soil Mechanics and Foun- Divisions, ASCE, Vol. 86, August 1960.
dations Division, ASCE, Vol. 89; No. SMI, pp. 115-143. Vilung, K., (1997). "Vibratory dnven pile and sheet piles - a
Jan. literature survey", Report 3035, Div. of Soil and Rock Me-
Ligterink, A., van Zandwijk, C., Wddentorp, P., (1990). "Ac- chanics, Rqyal Institute of Technology, Sweden, 75 p.
curate vertical pile installation by using a hydraulic vibra- Vucetic, M. and Dobry, R. (1988). "Degradation of Marine
tory hammer on the Arboath project". Proc. 2Jnd annual Clays Under Cyclic Loading. ASCE Journal of Geotechni-
Technology Conf: in Houston, Texas, May 7-10, pp. 315- cal Engineering,Vol. 114, No.2,pp. 133-149.
326. Vucetic, M. and Dobry, R. (1991). "Effect of Soil Plasticity on
Masing, G. ( 1926), "Eigenspannungen und Verfeistigung Cyclic Response." ASCE Journal of Geotechnical Eiigi-
beim Messing", Proceedings of Second International iieering, Vol. 117, No. 1, pp. 89- 107.
Congress ofApplied Mechanics~pp. 332-33 5. Vucetic. M. (1993). "Cyclic Threshold Shear Strains of Sands
Mdendorp; P. and Jonker, G. (1988). Prediction of Vibratory and Clays", Research Report, UCLA Dept. of Civil Engi-
Hammer Performance by Stress wave Analysis, Preprint to neering, May 1993.
the 3'd Int. Cone on the Application of Stresswavre Theory Vucetic, M. (1994). "Cyclic Threshold Shear Strains of Sands
to Piles, Ottawa and Clays", Paper in print, ASCE Journal of Geotechnical
Moulai-Khatir. Re&. ONeill, Michael W., Vipulanandan, C.: Engineering
(1994). "Program VPDA Wave Equation Analysis for Vi- Wamngton, Don. C.: (1989). "Driveability of Piles by vibra-
bratory Driving of Piles", Report to The U.S.A. Army tion". Deep Foundations Institute 141h Annual members
Corps of Engineers Waterways Experiments Station. Dept. Con$, Baltimore, Maryland, USA pp. 139-154.
of Civil and Environmental Engineering, UHCE 94-1. Westerberg. E.. Massarch, K. Rainer. Eriksson, K., (1995).
Univ. of Houston, Texas, August 1994, 187 pp. "Soil resistance during vibratory pile dnving". Proc. to hit.
Novak, M., Nogami, T., and Aboul-Ella, F. (1978). "Dynamic Symposium on Cone Penetration Testing, Linkoping, Swe-
Soil Reactions for Plane Strain Case", J. Engrg. Mech. den, Vol. 3, Report 3.41, pp. 241-250.
Div., ASCE, 104(4), 953-959. You4 L.T. (1972). Compaction of sands by repeated shear
NRC (1985). "Liquefaction of Soils During Earthquakes." Na- straining. Journal of the Soil Mechanics and Foundaiioiw
tional Research Council Committee on Earthquake Eng- Division, 1972, Proc. ASCE, Vol. 98, SM7, pp. 709-725.
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
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) (c> 2000 Balkema, Rotterdam, ISBN 90 5809 150 3
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.
495
Calibration against real site data is difficult because conceptual modei includcs lumped rani an3 anvil
of the many physical parameters involved. mcwes separated by a damped cushion which is
However, good agreement of both vertical and represented by a spring and dashpot in parallel.
radial surface vibrations, defined by velocity-time These rest on the pile which is modelkd by a
traces, has been achieved for a number of site dashpot, see Figure 1.
records. These include bearing piles installed by
vibrodriver at the second Severn crossing and
impact driving on the M66 near Manchester.
496
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.
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.
1.0
[ A 0 --3 1
0.3
0.0
9 -0.5
d
i -1.0
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.
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
498
for the second stage of the computational procedure Taken overall, and recognising the ground
in order to determine the displacement-time variability, the correlations between the equivalent
functions on the pile shaft. velocity-time traces are generally encouraging. The
main discrepancy appears to be due to a P-wave
In the third stage, the generated displacement-time from the pile shaft which may be caused by an
functions were imposed onto the shaft nodes of the eccentric strike. This is not simulated in the model.
“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 4 SIMULATION OF VIBRO-DRIVING OF PILES
5.5m, 10m and 16.5m are compared with the
measured vibrations in Figures 6 and 7 respectively. The process of pile installation by vibratory driver
is fundamentally different from impact driving, and
Figure 6. Radial ppv’s at 5Sm, 10m and 16.5m Figure 7. Vertical ppv’s at 5Sm, 10m and 16.5rn
respectively. (Measured = dashed line) respectively. (Measured = dashed line).
499
so modelling of the process must also be different The spring and dashpot constants are given by
from the above procedure for estimating vibrations
k = -4.GR (1 1)
from impact hammers. In practice, the method is I-v
best suited to saturated fine sands, and liquefaction kR
is induced by the vibration of the pile, which then ~=0.85-
sinks into the ground under the combined weight of
driver plus pile.
The time-stepping computation generally stabilises
Many site records have shown that ground into steady state after 1.0-1.5s with time increments
vibrations resulting from vibro-driving are of 0.001s.
continuous and sinusoidal in character, in marked
contrast to the transient nature of vibrations caused In the second stage, a €%/E mesh is established,
by impact piling. similar to that in Figure 3, and the sinusoidal
displacements of the shaft nodes and the base,
The technique for modelling pile response to cyclic computed in stage one, are imposed in the form of
excitation, and outgoing ground waves, is now Fourier series. It is essential in an on-going analysis
proposed as a two-stage procedure. that there is no spurious reflection from an artificial
boundary, and this is achieved by the infinite
In stage one, the objective is to establish a model elements around the periphery of the model.
for rigid body vertical oscillation of the pile in Material damping is applied in the form of a
response to the cyclic excitation of the vibro-driver. Rayleigh damping ratio given by
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
where U,, =m = undamped frequency of
comprises a two-surface contact, with a Coulomb vibration. A typical damping ratio of 5% has been
friction model. A static computation is made for the suggested by Massarsch (1992) for the elastic range
soil/shaft normal stresses due to geostatic stress. A of soil deformations.
dynamic analysis is then conducted to set up steady
state response of the rigid pile to the cyclic The output from the computation gives ground
excitation from the hammer of surface vibrations as sinusoidal vertical and
horizontal particle velocities. These can then be
calibrated against site data.
F(t) = m.e.w2. (10)
500
Figure 9. Radial ppv’s at 5.7m, 14.5m and 32.9m Figure 10. Vertical ppv’s at 5.7m, 13.5m and 32.9m
respectively. (Measured = dashed line) respectively. (Measured = dashed line)
the measured site data. The influence of the various Increasing the material damping from 5% to 8%
parameters were found to be as follows: only causes a small reduction in the ppv’s
The distribution of shaft friction with depth
affects the proportions of the ppv’s with relation The computed ppv’s at the ground surface for a
to each other at various stand-off distances. This casing installed to a depth of 1l m at a frequency of
may be due to the interaction of the ground 15.8 Hertz and a penetration speed of about 0.18mds
waves propagated from the shaft and the toe as are shown in Figures 9 and 10. These computations
observed by Attewell et al, 1991. assumed a small strain stiffness of 50 x 106 Pa for
The magnitude of the dashpot constant affects the soft/firm clay with v = 0.4 and p = 1750 kg/m3.
the magnitude of the ppv’s
Doubling the spring stiffness only causes a The computed ppv’s show good agreement with
small reduction in the magnitude of the ppv’s both the radial and vertical ppv’s measured on site
over a considerable stand-off distance.
501
6. CONCLUSIONS loading. Int. J. for Num. arzd Analytical
Methods in Geomechanics. 19:307-329
Simplified procedures have been developed for the Eurocode 3 1996 Design of steel structures, Pt 5
computation of ground surface vibrations caused by Piling CENflC 250/SC3/PT5
impact driving and vibratory driving of pre-formed Head, J.M. & F.M. Jardine 1992. Grouizd-borne
piles. vibrations arising from piling. TN 142
London: CIRIA
The procedures have been shown to be viable by Lysmer, J. & F.E.Richart 1966. Dynamic response
comparisons with site measurements at two offootings to vertical loading. J. of the S.M &F
locations, provided that appropriate values for Div. Proc. A.S.C.E 92; SM1 65-91
certain parameters can be identified. Further pile Mabsout, M.E. & J.L. Tassoulas 1994. A finite
driving simulations are currently being undertaken element model for the simulation of pile
for a large number of sites in the U.K. (Uromeihy, driving. Int. J. for Nunz. Methods in Eizg.
1990) in order to refine the various parameters, and 37:257-278
ultimately, to develop general guidance on the most Massarsch, K.R. 1992. Static and dynamic soil
appropriate parameters for each set of displacements caused by pile driving.
circumstances. Application of Stress- Wave Theory to Piles, ed
Barends, F.B.J. Balkema Rotterdam : 15-24
The procedures do not require detailed knowledge Matthews, M.C., V.S. Hope & C.R.I. Clayton 1996.
of site conditions and are therefore particularly The use of surface waves in the determination
useful as a preliminary design tool. of ground stiffness profiles. Proc. Instrz. Civ.
Erzgrs Geotech. Eizgng, 1 19:84-95.
The methods are now established for a further phase Ramshaw, C.L., A.R. Selby & P. Bettess 1998.
of the work, which will be to add typical building Computation of the transmission of waves from
forms to the finitdinfinite element meshes, for pile driving. Ground dynamics and ninn made
estimation of building response with full soil- processes. ed Skipp, B. 0. T.Telford Publ.
structure interaction. London : 1 15- 128.
Selby A.R. 1991 Ground vibrations caused by pile
installation Proc of 4th Int Conf on Piling arzd
ACKNOWLEDGEMENTS. Deep Foundations. Stresa DFUTespa :497-502
Stain, R.T. 1992. SIMBAT - a dynamic load test
The funding of this work by EPSRC through Grant for bored piles. Piling: European practice and
No. GNLl8679 is gratefully acknowledged. 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
REFERENCES. Uromeihy, A. 1990. Ground vibration
measurements with special reference to pile
Attewell, P.B. & I.W. Farmer 1973. Attenuation of driving. PhD thesis. Uni. of Durham. UK
ground vibrations from pile driving. Ground Wolf, J.P. 1988. Soil-Structure-Interaction Analysis
Engineering. 6.4:26-29. in Time Domain. Prentice-Hall, New Jersey.
Attewell, P.B., A.R. Selby & A. Uromeihy. 1991 Wiss J.F. 1967. Damage effects of pile driving
Non-monotonical decay of ground surface vibrations. Highways Res B. 155 : 14-20
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
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.
(1 996), 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
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
~~
5.1 Case 1
506
Figure 4. Maximum PPV of ground vibrations at distances Figure 5 . 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 of 7.0, 28.3 and 150.2 m from TP-2, TP-4, TP-6 and TP-8
driven piles driven piles
150.2 m from group 2 of TP-2, TP-4, TP-6, TP-8 In pile group 2, the effects of TP-4 pipe pile shell
piles. Maximum PPV of radial and vertical ground and TP-8 fluted pile shell were major for radial and
vibrations are depicted in Figures 4 and 5 . Radial vertical components at all distances from driven
ground vibrations had higher intensity in comparison piles.
with other components of ground vibrations.
Transverse horizontal ground vibrations were not @ Driving non-displacement piles induced lower
considered in the study because their level was 3-5 ground vibrations, but the effects of TP-1 and TP-3
times smaller than level of radial horizontal ground H-piles were dominant for radial and vertical
vibrations. components at distance of 9.1 and 30.5 m from
Four different pile types were tested. H-piles driven piles at the pile penetration length between
were non-displacement piles and the rest of piles 6.4-7.3 m with the pile toe in dense clayey sand.
were displacement ones.
The analysis of soil vibration records, measured at @ Effects of displacement piles were more
the same distances from each group of piles with substantial on amplification of ground vibrations
different impedances and driven by the same than those from piles with impedance reduction.
hammer, revealed some interesting observations.
Ground vibrations increased due to pile
e In general, piles being driven as displacement penetration depth of 6.4-7.3 m and then started to
piles generated the greatest ground vibrations, but decrease. This effect is related to soil conditions
there effects were varied. In pile group 1, the effect rather than to pile length.
of concrete TP-7 pile was dominant for radial
component at distances of 30.5 and 152.4 m and for @ Restrikes were made for TP-2 - TP-6 piles.
vertical component at all distances from the source. Restrikes did not actually increase the level of
ground vibrations.
Figure 6 . Maximum PPV of ground vibrations at various Figure 7. Maximum PPV of ground vibrations at disranxs
distances from driven precast pile and H-pile of 7.0 and 9.1 rn versus soil resistance to pile driving
508
and suimxmding ground, but on the other hand. point source of vibrations. This assumption might
according t ~equation
! (31, increasing pile impedance be correct at dismce. D, from driven pile (Svinkin
decreases PPV of pile and ground vibrations. An 1996)
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 Where c, = velocity of surface wave propagation in
intensity of ground vibrations is not o b l h s . the ground; remaining parameters are the same as
Presented case histories revealed that pile above.
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 8 CONCLUSIONS
generated similar ground T ibrduons though the latter
had pile impedance 2.9 rimes less. In Case 2, che File impedance affects force and velocity at the pile
dri\.ing of precdst concrete piles induced PP\’s of head in opposite ways. Therefore it can be expected
ground \ lbrarions IZCX rhe prics three times more that pile impedance does not significantly affect
tlian the driving oc the H-piles eke11chough the latter energy transferred to the pile. Analysis of published
had pile impedance 1.6 times smaller. experimental data and wave equation simulation
cunfiims that the impedance of steel piles has a
negligible effect on ETR for air and hydraulic
7 . 2 Displacernmr piles hammers. Additional research IS needed to clarify
the impedance effect of concrete piies on ETR for
Piles being driven as displacement piles transfer all hammer modek.
p a t e r force to the surrounding ground. The more An increase of hammer energy magnifies ground
pronounced effects of displacement piles on ground vibrations until the pile impedance allows the force
\ ibrations occurs predominantly at distances less transmitted to the pile and the surrounding sod to
t h m 10 m from pile driving. 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
7 . 3 Pile-soil load transfer vibrations is not obvious.
The more pronounced effects of displacement piles
The dynamic force transferred to the pile at each on ground vibrations occurs predominantly at
hammer blow is then transferred to surrounding distances less than 10 m from pile driving.
ground through the pile toe as a concentrated load In many situations, a driven pile. can be
and through the pile shaft as a non-uniformly considered as a point source of vibrations.
distributed load.
During driving. the various soil resistances to pile
penetration are developing as the pile penetration ACKNOWLEDGEMENT
depth is increasing. Generally speaking, dynamic
loads transferred from the pile shaft and toe to the The writers are thankful to Mr. Anatol Longinow,
ground should be increasing with increasing pile Project Manager at Wiss, Janney, Elstner
penetration depth. However, this does not always Associates, Inc. for valuable comments.
occur. In Case 1. the intensity of ground vibrations
was independent of the pile penetration depth and
depended on soil properties. Maximum PPV of REFERENCES
ground vibrations were recorded during driving all
piles through dense clayey sand at depth 6.4 to Abe, S. & G. Thendean 1996. Hammer performance
7.6 ni. evaluation. In F. Townsend, M.Hussein &
Pile-soil load transfer is realized by means of both M. McVay (eds), Proceedings of the Fifth
concentrated loads from the pile toe and distributed International Conference on the Application of
loads generated along the pile shaft. However, for Stress Wave Theory to Piles, Orlando, Florida:
many practical purposes, a pile can be deemed as a 9 12-927.
509
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
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.
d‘w
1 INTRODUCTION EI 7 + KW 10 K=k.d (1.3)
dx
The pile foundation inay be under a horizontal load /F Coefficient of soil reaction, d= Diameter of the
as well as under an axial load. The pile problems, pile, the equation number (1.3) is obtained. The “k”
which are under the effect of the horizontal force, value is given in the literature approximately for the
appears frequently at quay structures, petrol produc- soil classes (Bowles 1988). The aim of the study is;
tion plants and marine pilings. The horizontally first, to determine in testing manner of the “I? value
loaded pile is closely related with the beam prob- through mounting a vibrating apparatus on the head
lems, based on the elastic foundation, and is an ap- of the pile, then, through adding the technical terms,
plication to head of pile in the pile-soil system of the taking into account tlie dynamic effects establish-
external forces and moments. According to the ment of a numerical model and its evaluation. The
beam theory based on the elastic foundation: new equation obtained is:
a‘w a2w
El- + ,U---- + k-&v + Kw = 0 (1.4)
ax‘ at2 at
In this equation Theoretically, the pile head flexibility, i.e. tlie
$4) : The horizontally displacement of the pile horizontal displacement curves, which are the coun-
x : The depth in soil terpart for a unit force, are obtained for differcnt val-
E l : Pile stiffness ues of coefficient of soil reaction. The “k” value of
12 : Soil reaction. the curve is the coefficient of soil reaction, with
The soil reaction “p” is a function of the pile which curves experiinentally obtained flexibility
characteristics and (1.1) the solubility of the equa- values fits.
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 2 EXPERIMENTAL STUDY
of the pile “ w ” , the soil reaction which effect to the
unit length of the pile will be found by the following The experimental study has been carried out one of
equation: the composite pile of the quay at Clierbourg
(France). The composite piles have been formed
~=-K.w ( 1.2) through driving of steel tube piles of l n i of diameter
In this equation, K represents the ratio of soil re- and then filling them with concrete without rein-
action on a point of pile to the displacement at the forcement. The experiment has been carried out by
same point. If the equations (1.1) and (1.2) are casting after 2 months of the vibrating apparatus on a
united: 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.’)
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)
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)
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
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
514
FLEXIBILITY
LEGEND
v Experitnental values
Unstable measunngs
- Theoretical vzlues
5 CONCLUSION
515
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
516
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Salkema, Rotterdam, ISBN 905809 150 3
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.
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:
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 11. Curve of the maximal shear changed with
foundation modulus
521
Figure 13. Curve of the maximal shear changed with
coefficient of viscosity of soil Figure 15. Cur-vc ol displaceriient of pile at different
time
522
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.
523
This Page Intentionally Left Blank
Application of Stress-Wave Theory fo Piles, Niyama SC Beim (eds) C) 2000 Balkema, Rotterdam, ISBN 90 5809 1503
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.
525
@ Inspect present conditions of surrounding 3 VIBRATION ENVIRONMENT
buildings.
Perform damage susceptibility study to establish Pile driving operations generate different kinds of
vibration control limits. waves in the ground, but Rayleigh waves have the
Measure vibration background at the area under largest practical interest. Rayleigh waves become
investigation. predominant over other wave types at distances
Assess problems such as cracking of building or smaller than a pile length. Impact hammers induce
foundation failure. transient ground vibrations with the dominant
First of all a preliminary desk study of a layout of frequency between 7-50 Hz (Svinkin 1999), and
the area for a preconstruction inspection and a site vibratory hammers induce steady-state ground
walk-over observation should be made. Existing vibrations in the frequency range of 5-40 Hz
cracks found in buildings have to be marked. It is (Warrington 1992).
necessary to distinguish cosmetic and structural The intensity of ground vibrations at the surface
cracks. Most attention should be paid to cracks in and the certain depth are comparable. Linehan et al.
the structures themselves. The width of cracks (1992) studied the effect of pile driving in the
should be measured with a proper ruler. proximity of the gas line with a pipe diameter of
Determining the cause of cracking is important to 1 m. The sheet and H piles were driven in sandy
predict lengthening and dilatation of old cracks soils with vibratory and diesel hammers,
under the vibration effect of pile driving. respectively. For both types of driving, particle
For assessment of the dynamic effect on velocities on the pipe were less than those measured
surrounding structures it is necessary to take into at the surface at a distance of 1.5 m but became
account the thresholds of damage, cracking and similar at distances about 10-12 m from driving.
perception. Buildings inspected under the contract Barkan (1962) analyzed the variation with depth of
requirements are commonly classified depending on vertical displacements in loessial clay induced by a
structure susceptibility to cracking and proximity of pile driver and concluded that at small depths about
structures to pile driving operations. of 0.2 to 0.5 wavelength changes in vibration
Structure susceptibility is usually related to the displacements are relatively small.
threshold of cosmetic cracking (Dowding 1996) and
depends on a degree of degradation of the building
structural and nonstructural systems. Apparently, it 3.1 Deterministic approach
makes sense to broader consider this terminology as
susceptibility of the building-soil system depending Statistical analysis is used for evaluation of the test
on degradation of building systems, utilization of data from blasting (Dowding 1996). Such an
buildings and soil conditions. It is important for approach for assessment of other construction
certain cases. For example, a building, located in vibrations should be carefully used because ground
the proximity of the driven piles, identified as vibrations generated by the dynamic forces applied
having low susceptibility and built on liquefiable to the soil directly or transferred through driven
soils might have substantially larger deformations piles have deterministic nature.
than a building identified as having high The effect of large plastic soil deformations at the
susceptibility but erected on non-liquefiable soils at contact area under falling mass on ground vibrations
long distance from the driven piles. Certainly, for was studied with falling mass of 15.0 tonnes at a site
some sites only the threshold of cosmetic cracking where soil deposits were mostly fine sands with
could be sufficient. natural moisture content (Svinkin 1996a). The drop
Measurement of a vibration background at soil and height was 10 m. Many impacts were performed at
buildings with sensitive equipment and/or the same spot; consequently, large plastic soil
computerized technology should be a part of the deformations occurred at the point of impact.
preconstruction survey. The vibration measurement Records of ground vibration displacements at various
might reveal microvibrations and vibrations induced distances from the place of impact on the ground are
by industrial machinery located nearby. A waveform depicted in Fig. 1.
recorder has to be used for these measurements. Comparison was made for records obtained for
Assessment of the dynamic effect of pile two equal impacts with different degrees of plastic
installation on surrounding structures can be made soil deformations at the contact area. In particular,
on the basis of prediction of soil and structures vibrations were measured at distance of 43 m for the
vibrations from pile driving. first and ninth impacts, and at a distance of 57 m for
526
amplitude reduction of Rayleigh waves, generated by
an earthquake, between two points at distance rl and
r2 from the source as (Golitsin 1912)
r = Z rn
r=33 rn
527
y described above. So, equation (2) has distances from the sources. Perhaps the cause of
disadvantages similar to the equation (1). this phenomenon was a higher attenuation of surface
Wiss (1981) found the ground velocity-distance- waves in the loose soil deposits where Franki piles
energy relationship, so-called scaled-distance were driven. This observation underlines the
approach, to calculated the peak ground velocity at significance of soil contribution to the formation of
surface distance, D, from a source normalized with ground vibrations.
energy as
528
written as follows
where I, = impulse force transmitted from machine Figure 2. Layout of machine foundation, place of impact
to foundation, kNs; fI1z = circular natural frequency loads on ground and geophones
of vertical vibrations of foundation, rad/s; f n d =
circular natural frequency of vertical damped
vibrations of foundation, rad/s; CY = effective 5 CONTROL OF VIBRATIONS
damping constant, rad/s; \(t-7) = impulse response
function at the output point under consideration; 7 = There are three ways of how ground vibrations from
variable of integration. construction sources can effect adjacent and distant
Examples of predicted results are shown in Fig. 2, structures. First, excitation frequencies from ground
and 3. Measurement and prediction of vertical and vibrations do not match structural natural
horizontal ground surface displacements were made frequencies. Second, ground vibrations may
at diverse distances from the foundation under a generate the conditions of resonance in the building.
sizeable drop hammer with a falling weight of Third, foundation settlements result from relatively
147.2 kN and a maximum drop height of 30.0 m. small ground vibrations. For all situations, sensitive
The soil at the site consisted of about 1.6 m of loose equipment in buildings will be exposed to vibrations.
sand followed by about 6.8 m of medium density The tolerable vibration level can be verified before
sand and 1 m of sandy clay underlain by about 10 m beginning of construction activities at a site.
of slightly moist sand. The water table was about For structures which natural frequencies do not
6 m below the ground surface. The Rayleigh wave match excitation frequencies of ground vibrations,
velocity was 270 m/sec. A layout of the machine structural responses would be lower than the ground
foundation, the place of impact on the ground and vibrations. Prediction of the maximum peak particle
geophones is displayed in Fig. 2. velocity independent of frequency could be used.
Predicted and measured vertical and horizontal However, it is necessary to point out that low
components of ground surface vibrations at eight structure vibrations cannot prevent high vibrations of
locations are shown in Fig. 3. It can be seen that equipment installed in the building. Dowding (1996)
good agreement is matched in time domain vibration demonstrated a case where PPV on the ground and
records, except for horizontal vibrations at two the second floor were 2.03 and 1.52 mm,
locations close to the foundation. This can be respectively, but the table on the floor had PPV of
explained by the different wave paths from the 10.5 mm.
foundation under the operating machine and the Levin (1969) found that 3-4 cycles of ground
place of impact on the ground. The distance vibrations could generate conditions of resonance in
between these two sources was 18.7 m. Lack of the building. Such conditions can be evaluated on
coincidence of the two dynamic sources slightly the basis of the prediction of complete time domain
affected ground vibrations at a distance from the records.
machine foundation. Agreement of predicted and Significant foundation settlements caused by pile
measured vibration displacements is quite driving in vulnerable sands can result from PPV of
satisfactory. The differences between the peak ground vibrations as low as 2.5 to 5.1 mm/s (Lacy
predicted and measured vibration amplitudes are less & Gould 1985). Such velocity values are much less
than 30 % at distances larger than 43.0 m from the than 51 mm/s which is considered as the tolerable
foundation. For some individual points amplitudes safety limit for buildings. Therefore at sites with a
actually coincide. predominant granular soil deposits, the maximum
Spectrum analysis of predicted and measured time peak particle velocity that varies with frequency
histories revealed that both records have similar should be predicted at the area where intensity of
frequency domain curves with the same dominant vibration is expected higher than 2.5 m / s .
frequency. Moreover, predicted records are slightly
dependent on the dynamic parameters of the source
(Svinkin, 1999).
529
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 Academy of Science News, Vol. 6, No. 2,
Heckman, W.S. & D.J. Hagerty 1978. Vibrations
Pile driving is a wide-spread source of construction associated with pile driving. Journal of the
vibrations which may harmfully affect surrounding Construction Division, A X E , 104(C04): 385-394.
buildings. It is important to accurately predict Lacy, H.S. & J.P. Gould 1985. Settlement from
vibrations of ground, structures, and sensitive pile driving in sands. Proceedings of ASCE
devices prior to the beginning of construction Symposium on Vibration Problems in Geotechnical
activities to avoid the undesirable effect of generated Engineering, Detroit, Michigan, G. Gazetas and
vibrations. E.T. Selig, Editors, pp. 152-173.
The predriving survey should be made to ensure Levin, G.E. 1969. Dynamic influence of drop
safety and serviceability of adjacent structures and/or hammer foundations on surrounding structures (in
distant structures with sensitive equipment like Russian). Proceedings of the Second Conference
hospital facilities and offices with precision on Dynamics of Bases and Foundations: 147-152,
instruments. Moscow: NIIOSP
Empirical equations provide only calculation of a Linehan, P. W., A. Longinow & Dowding, C .H.
vertical amplitude of ground vibrations and not 1992. Pipeline response to pipe driving and
always with sufficient accuracy. For pile driving, adjacent excavation. Journal of Geotechnical
the scaled distance approach with calculated peak Division, ASCE, 118(2): 300-316.
particle velocity of the source is probably the most Massarsch, K.R. 1992. Keynote lecture: Static and
appropriate method for predicting upper limits of the dynamic soil displacements caused by pile driving.
peak particle velocity of ground vibrations. In F. Barends (ed.), Proceedings of the Fourth
The impulse response function prediction method International Conference on the Application of
(IRFP) has obvious advantages in predicting time Stress- Wave Theory to Piles: 15-24. Rotterdam:
domain ground and structure vibrations prior to the Balkema.
beginning of construction activities. Smith, G.M. & G.L. Downey 1968. Advanced
The IRFP method provides opportunity to control engineering dynamics. Scranton: International
the tolerable vibration level of soil, structure and Textbook Company.
sensitive equipment in advance. This is an Svinkin, M.R., 1992. Pile driving induced
important step in preventing intolerable vibrations vibrations as a source of industrial seismology. In
from pile installation. F. Barends (ed.), Proceedings of the Fourth
International Conference on the Application of
Stress-Wave Theory to Piles: 167-174. Rotterdam:
REFERENCES Balkema.
Svinkin, M.R. 1996a. Overcoming soil uncertainty
Barkan, D.D. 1962. Dynamics of bases and in prediction of construction and industrial
foundations. New York: McGraw Hill Co. vibrations. In C.D. Shackelford, P. Nelson, and
Brettmann,T. & B. Cotton 1999. Case history: M.J.S. Roth (eds.), A X E , Proceedings of
estimating ground vibrations caused by pile Uncertainty in the Geologic Environment: From
driving. DFI 24th Annual Members ’ Conference, theory to Practice, Geotechnical Special
Decades of Technology - Advancing into the Publications No. 58. 2: 1178-1194.
Future: 161-171. Svinkin M. R. 1996b. Velocity-impedance-energy
Broers, F. & H.A. Dieterman 1992. Environmental relationships for driven piles. In F. Townsend,
impact of pile-driving. In F. Barends (ed.), M. Hussein and M. McVay (eds.), Proceedings of
Proceedings of the Fourth International the Fifh International Conference on the
Conference on the Application of Stress-Wave Application of Stress-Wave Theory to Piles: 870-
Theory to Piles: 61-68, Rotterdam: Balkema. 890.
Dowding, C.H., 1996. Construction Vibrations. Svinkin, M.R. 1999. Prediction and calculation of
Upper Saddle River: Prentice Hall. construction vibrations. DFI 24th Annual
Gerasch, W.J. 1998. A way to predict vibrations. Members ’ Conference, Decades of Technology -
Seventh International Conference and Exhibition Advancing into the Future: 53-69.
on Piling and Deep Foundations: 3.3.1-3.3.13. van Staalduinen & P.H.Waarts 1992. Prediction of
Rickmansworth: Westrade Group Ltd. vibrations due to pile driving. In F. Barends (ed.),
Golitsin B.B., 1912. On dispersion and attenuation Proceedings of the Fourth International
of seismic surface waves (in German). Russian Conference on the Application of Stress-Wave
53 1
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
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.
....,...... ..... ,/
The eccentric moment M , determines the ampli-
a) b) tude of both the displacement and the unbalanced
force, but it has not yet been established how the
igure la-b. Two schematic plots of the driveability curve. best choice of static moment should be determined
concerning the driveability. Several authors, (e.g.
The slope of the t-z curve, described in Figure la, Rodger et al. 1980) have stated that the displacement
represents the global rate of penetration vp. The en- amplitude should be chosen (high or small) to drive
largement of the t-z curve (Figure la) illustrates the the pile efficiently at the chosen driving frequency
principle of the penetrative motion of the sheet pile, but none gives the underlying assumptions or why it
i.e. the different up- and downward parts of the dis- should be so.
placement amplitude. The conducted field tests reviled that the effi-
The pattern of the recorded v,,-z curve can be used <
ciency of the vibro-equipment is another essential
as a tool to study the change in dynamic soil resis- parameter. The peak compressive force measured at
tance. Rao (1993), involved in the extensive research the sheet pile head during the field test should be
work at University of Texas at Houston, stated that theoretically equal to the sum of the applied static
the penetration speed can be subdivided into three surcharge force, maximum unbalanced force and de-
states, ranging from slow to high penetration speeds, celeration of the vibrator, expressed according to
(Table 1). Equation (2).
534
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).
535
penetration. The rate of penetration has been ob- 2.4 Pei i elra 1ioii 171CCIINIZi st i t s of lli c system
served to be higher in saturated sand than in dry
The difficulty in applying a driveability formula
sand, but the difference scems to be more pro- comes from the fact that a great deal of trade-off
nounced in the case of low displacement pilcs. Re- must be made between the accuracy of the candidate
corded pore pressure readings near the pile shaft (in formula and the practicability of applying it to real
the sand) have been observed to be higher near the problems. This means that the key issue is to access
pile shaft than those away from the oscillating pile the magnitudes of both the dynamic soil resistance
shaft. The pore pressure development during driving and the friction force R, in the sheet pile clutch.
is closely related to the driveability. However, stud- However, the magnitude of the theoretical forcing
ies by O’Neill et al. (1989) could not relate all of the function of the vibro-driver Fd, i.e. sum of static sur-
shear strength loss to the pore pressure build-up charge force F, and unbalanced force F,, is easier to
during driving. This could be explained either by the estimate. It varies sinusoidally in time with driving
fact that liquefaction occurs near the oscillating sheet frequency and eccentric moment according to Fig-
pile, or by shear strength reduction due to the in- ure 3a.
duced cyclic motion of the soil particles during The motion of the sheet pile u(t) is described by a
driving, or a combination of both mechanisms. downward directed sinusoidal displacement, corre-
The relative density is the soil parameter that has lated in time with the driving force Fd, see Figure 3b.
been reported to have the most significant influence The dynamic shaft resistance R, (Figure 3c) varies
on the global rate of penetration. The rate of pene- between positive and negative, correlated with the
tration has been found to decrease with increasing up- and downward motion of the sheet pile.
relative density. The induced volume changes, (den- The dynamic toe resistance R, varies between zero
sification or dilation) in the soil are believed to give and maximum, also correlated with the motion u(t)
rise to a characteristic shape of the driveability vp-z of the sheet pile, and reaches it’s peak value at the
curve, see Figure lb. lower end of the up- and downward penetrative mo-
The content of fines (greater than 12%) is a ne- tion u(t) of the sheet pile, see Figure 3d.
glected soil parameter that usually dictates if the use No theoretical explanation has been presented for
of vibro-drivers will be successful or not, a fact that how the generated clutch friction R, varies during
the penetrative motion. However, it is well known
has been experienced by several authors, but none
that the dynamic clutch friction R, is of great signifi-
gives the underlying exp!anation. cance.
Figure 3. Schematic descriptions of the relationship between driving force, penetrative motion, shaft- and toe resistance.
536
3 FIELDTEST
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.
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
Corps of Engineers Waterways Experimental Station.,
-35
UHCE 94-1, Univ. of Houston, Texas, 187 pp.
O'Neill Michael W., Vipulanandan, C., (1989), Laboratory
0,11 0.15 0.15 0,16 0,16 0.17
1' Iml evaluation of piles installed with vibratory drivers. National
Cooperative Highway Research Program, Report No. 316,
igure 10. Load transfer curves of sheet pile B1 near the toe. National Research Council, Washington, DC. Vol. 1. pp. 1-
51. ISBN 0-309-04613-0.
The negative toe resistance noticed in Figures 9b- Rao, Pramod M., (1993), Effect of pile geometry and soil satu-
ration in the behavior of non displacement piles installed by
10, is assumed to be related to inertia effects in vibration., M.Sc. Thesis, Dept. of Civil and Environmental
combination with a small portion of shaft resistance Engineering, University of Houston, Texas.
on the 600 mm long part of the sheet pile below the Rodger, A.A., Littlejohn, G.S., (1980), A study of uibrarory
position of the toe sensors. driving in granular soils. Geotechnique, Vol. 30, No. 3, pp.
269-293.
5 CONCLUSIONS 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.
Based on this study and the range of variables in- Viking, K., (1998), Driveability studies of vibro-driven model
vestigated, the following primary conclusions can be piles in non-cohesive soils -laboratory simulations., Licen-
advanced: tiate Thesis 2029, Dept. of Civil and Environmental Engi-
1.)The presented instrumentation system has proven neering, Royal Institute of Technology.
to successfully register the dynamic soil resistance. Viking, K., Green, J., Nilsson, C-O., (2000), Uppmatta
2.) The position of applied sensors made it possible markvibrationer av vibroinstallerad spont., 13th Nordic
Geotechnical Conf. on Geotechnics and Sustainable Devel-
to document field related properties of vibro- opment., Helsinki, June 5 - 7, 2000 (in press).
capacity and penetrative resistance.
540
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 0 2000 Balkema, Rotterdam, ISBN 90 5809 1503
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
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
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
Typical measurements of the pile forces and accel-
erations and the corresponding velocities and dis-
Figure 3 TJ pica1 ground accelerometer (GA) arid ~olalpres- placements with time for a single blow are shown in
w e cell (TPC) iiistalla~ioncross-~ectioii 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).
3 . MEAS U REMENTS Figure 5 shows the pile tip experiencing significant
acceleration with corresponding high velocity and
Analysis of the pile and soil gage data focused on displacement during driving near the soil instr-umen-
rneasurements taken during passage of the pile tip tation. The measurements of high acceleration and
approximately 3-5m above and below the soil in- displacement at the pile tip, coupled with the low
strumentation locations This corresponded to a driving resistance of the pile through the siltp sand
range of soil starting at the interbedded layer (depth layer as shown in Figwe 4, make this pile ideal for
16 3 lm) and ending within the fine to medium sand studying the efTects of soil inertia.
layer (depth 25 58111) Examination of the pile measurements during
penetration at and near the silty sand layer showed a
consistent drop between the maximum pile top, pile
4. PILE RESPONSE middle, and pile tip accelerations. Figure 6 s h o w
the maximuin pile middle and tip accelerations nor-
Problems experienced during Test Pile # 1 installation malized to maximum pile top acceleration with
limited pile acceleration measurements at both the depth. The ratio between pile middle and pile top
pile iniddle (gage 1 -8-APCB-A) and pile tip (gage 1 - *
accelerations was 0.85 0.07, while the ratio be-
16-APCB-A). Gage 1 - 16-APCB-A experienced tween pile tip and pile top was 0.58 7t 0.05 (mean Sr 1
equipment failure at pile tip depth of 18.90m, directly s.d.). This consistent range suggests that a reason-
above the silty sand layer Data acquisition system able estimation of pile tip acceleration can be made
problems with the PDA resulted in a gap in pile mid- fi-on1the measured values at the pile top. The ratios
dle acceleration measurements between depths 20.42 presented in Figure 6 are not necessarily representa-
and 23.48m. Measurernetits of the dynamic gages at tive of standard piles.
the pile top and the interior electrical resistance strain
gages were continuous throughout the driving of
Test Pile,# 1 near the soil instnmentation. 5 . SOIL RESPONSE
Permanent set for each blow was determined
through a detailed analysis of a video taken during 5.1 Soil Veiocip Meastiremenfs
the installation of Test Pile # 1 , Figure 4 shows these
Acceleration measurements taken by the ground ac-
measurements converted into driving resistance for.
celerometers (GA) were processed in the following
pile penetration around the silty sand layer.
manner:
544
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.
'i
--
I .-_- I
I Vertical
I -I Horizontal
Axis
1 -
PILE
Figure 7 Transfomied ground accelerometer axes
545
with the upward velocities being larger, indicating Figure 8. The maximum and minimum vertical, hori-
the soil is moving primarily downward (positive). zontal, and tangential velocities are presented as
Figure 9 was prepared based on the data shown in 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 .
Figure 9. Maximum and minimum measured velocities with Figure 10. Maximum and minimum measued velocities with
depth for GA1. depth for GA2.
546
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- Figure 12. Pore and total pressure measurements \\4th depth
forms. The results of the Fourier transforms are pre- for Gnd PZ-8 and TPC I.
sented in Figure 1 1.
The results presented in Figure 11 show that at the
pile top, the acceleration waveform largely consists in frequency content would occur within the accel-
of a range of frequencies between 1 to 3000Hz. At eration wave in the soil mass. If these frequency re-
the pile tip the acceleration wave is primarily com- ductions did occur, then the ground acceleration
prised of frequencies between 1 to 600Hz. This sub- measurements should have accurately recorded the
stantial reduction suggests that the soil attenuates the behavior of the soil mass due to pile penetration.
acceleration wave as it travels down the pile. Al-
though these observations do not reveal if high fie- 5.3 Total Sod and Pore Water Pressure Measure-
quency signals are present within the ground accel- ments.
eration measurements, they suggest that a reduction
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-
Figure 1 1. Typical test pile acceleration measurements and quisition system, did not allow for detailed pore pres-
corresponding Fourier transform analysis. sure measurements per blow
547
CONCLUSIONS REFERENCES
Based on the measurements presented in this paper, Holscher, P. 1995. Dynamical response of satiwated and drv
the following conclusions are made: soils. Delft: Delft University Press.
Measurements of pile and soil forces, accelera- Holscher, P. and Barends, F.B.J. 1996. In-situ Measurement
tions, pore and total pressures during pile driving of Soil-motion near the Toe of a dynamically Loaded Pile.
are possible and provide insight into the behavior In F.C. Townsend, M. Hussein. & M.C. McVay (4).Pro-
of the soil during pile penetration. ceedings3 ShInternationaI Coq/erence 011 tlie Applicatioii
Over the range of pile penetration examined, the of Stress-Wave Theory to Piles, 11-13 September, 1996:
majority of soil disturbance occurred at or near 26-36. Orlando, Florida.
the pile tip. Paikowsky. S.G. and Chen. Y.L. 1998. Field and Laboratory
Changes in total and pore water pressure around Testing of the Physical Characteristics of the Subsurface at
a driven pile are caused primarily by movement the Newbury Bridge Site. Research Report subirtitted to
of the soil around the pile tip. the Massachusetts Highwav Department, January 1998.
The accuracy of the ground acceleration meas- Boston, Massachusetts.
urements and corresponding velocity and dis- Paikowsky, S.G. and Chernauskas. L. 1996. Soil Inertia and
placement integrations at the Newbury Site may Use of Pseudo Viscous Damping Parameters. In F.C.
have been affected by the filtering of high fre- Townsend, M. Hussein, & M.C. McVay (ed), Proceedings,
quencies. To avoid this problem, ground accel- 5IhInternational Colf i r m c e on the Application qf Stress-
erometers in fbture research into soil accelera- Wave Theorv to Piles, 11-13 September, 1996: 203-216.
tions around driven piles should be capable of Orlando. Florida.
recording data at the same range of frequencies Paikowsky, S.G. and Hajduk. E.L. 1999. Design and Con-
generated within the pile. struction of an Instrumented Test Pile Cluster. Research
Examination of the pile measurements showed Report submitted to the A fassachusetis Highwqv Depart-
that the frequency content of the acceleration ment, September 1999. Boston, Massachusetts.
wave decreased as the wave traveled through the Paikowsky, S.G. and Stenersen. K. 2000. The performance of
pile. This attenuation suggests that the frequen- the dynamic methods, their controlling parameters and
cies within the soil mass are also attenuated in AASHTO 2001 Deep Foundation Specifications. Pro-
relation to those generated within the pile. ceeditigs, dhInternational Cotfirence on the .Ipplication
Additional, detailed examination of the Newbury of Stress-Wave Theory to Piles, 11-13 September, 2000.
Site measurements is suggested in order to assist in S I i Pado City. Brazil.
planning fbture research to develop a reliable, physi- Smith. E. 1960. Pile Driving Analysis by the Wave Equation.
cally based soil inertia model to incorporate into pile Journal of Soil Mechanics and Foundations, Aniericaii So-
driving formulations. ciety of Civil Engineers: 3 5-6 1. A irgust 1960.
ACKNOWLEDGEMENTS
548
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
€? 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.
551
world. Many engineers are not aware that their crea- idea of Fellenius at that time was that dynamic pile
tive solutions can be considered art and that they act 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.
552
Bermingham made a first prototype (Figure 3) and ist-engineer” can be observed, especially the gravel
started experiments by shooting masses upwards in catching mechanism based on the reliable gravity of
Hamilton, Ontario in April 1988. He determined the the earth (Fig.4). With the gravel mechanism a
feasibility of accelerating a mass upwards from the gravel container is placed around the reaction mass
top of the foundation rather than dropping a mass and the space between gravel container and reaction
onto the foundation 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.
3 FIRST DEVELOPMENTS
553
this point onwards the direction of Statnamic was also proved the simplicity and practicality of the sys-
upwards. A second model was built and sent to tem in the field and the first load test comparisons
TNO in Holland where the instrumentation would be proved an unexpectedly close agreement with con-
developed. In the laboratory at TNO, Statnamic ventional static load tests. The success of this first
tests were performed using a calibrated load cell and prototype enabled Berminghammer to manufacture
a new laser measuring system developed specifically of a full-scale 5MN tester.
for the Statnamic test. Both measuring systems
worked very well the first day and they have re-
mained virtually unchanged to this date.
554
than a static load prior to plunging the pile. In stiff The first approach to analyzing failure was to
non-cohesive soils and rock sockets it was observed look to the displacement curve and to analyze the
that there was very close agreement between Stat- rate of change of displacement, or velocity of the
namic load test results and static load tests per- pile. Normalizing the load and plotting load vs. ve-
formed along side. It was also observed that the se- locity was examined in an effort to pinpoint the load
quence of loading a foundation had a great effect on at which the velocity begins to increase. This only
the perceived similarity of test results and this had to worked well when the foundation experienced a
be taken into account. It was also observed that dur- plunging failure, and it did not work well when the
ing a typical Statnamic test the pile would reach pile was in a cohesive material.
maximum displacement at some time after peak
force had been applied. In other words the pile Statnamic test results were also evaluated with a
would continue to move downwards while the ap- simple 2.5mm offset method, which was analogous
plied load at the pile top was decreasing. At the to the Davisson failure criterion but much more con-
point where the pile reached maximum displacement servative. All three of these methods of determining
the velocity of the pile was zero and then the pile the point of failure were far too subjective to be of
would begin to rebound as the load was further de- any great value.
creased. This observation which was present in
nearly all test results except very stiff piles on rock, In January of 1993, while reviewing the results of
lead to the development of the Unloading Point pile 7 at Texas A&M the author noticed that during
Method (UPM) by the author (P. Middendorp, the unloading of the test the velocity of the shaft
1992). 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.
4 METHODS OF ANALYSIS PDA users had observed the significance of the
point of zero velocity in the 1970's and some at-
From the very start of development there has tempts were made to make use of it. However, dur-
been a determined effort to make Statnamic a means ing pile driving the point of zero velocity at the pile
of measuring rather than a prediction method. This head does not correspond to zero velocity anywhere
has meant putting a very strong emphasis on using else in the pile unless the pile is very short and rigid.
accurate measuring equipment and recording equip-
ment. The measured data will then be more reliable The author's observation provided both a practi-
and may then be examined more closely. From the cal means of determining a significant point on the
beginning we have been observing the behavior of static load displacement curve and also a means of
foundations subjected to very rapid loading with a estimating the damping coefficient directly from the
view to being able to better understand the mecha- test results rather than from a soil boring. This
nism of failure during a Statnamic test. In the end it Unloading Point Method (UPM) assumed that the
is hoped that Statnamic testing will stand alone as a damping was a constant, which was zero when the
rapid test with a distinct method of analysis, which velocity was zero, and that the pile was behaving as
will measure the load deflection behavior and de- an elastic body, which could be treated as a lumped
termine the factor of safety of the foundation.
555
mass and a spring. Subsequent research has concen- measure of toe displacement is desirable. All results
trated on testing these assumptions and determining are based on measured quantities.
the limit of their validity. The unloading point
method has provided a very simple universal method F =E,E,A,
of analysis and was first published at the 4ti1Interna-
tional Conference on Stress Waves in 1992. The ba- where F; is the measured force at level i, E; is the
sic principles of the method have been presented in measured strain at level i (typically an average of all
the Appendix. strain gages at level i), E, is the calculated (or as-
sumed) elastic modulus at level i, and A; is the calcu-
The UPM method is based on the assumption that lated (or assumed) area at level i
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 where U ; is the calculated displacement at level i,
duration of the load, c the stress wave velocity and L L; is the length of the pile segment between levels i
the pile length. In this way it was possible to indi- and i+l
cate when stress wave phenomena could be ne-
glected and when they should be taken into account.
556
The derived static load displacement curve can be Hydraulic catching systems eliminate the need for
drawn at each strain gage elevation using FSTAT,~and gravel and gravel structure since the upward moving
U,. reaction masses are caught at the top of their flight
by four hydraulic actuators (or rams). These 3.2m
Each of the above variables represents an entire stroke rams are activated by four low pressure (100
data set measured over time. bar) nitrogen accumulators, which store compressed
nitrogen gas over hydraulic oil. As the weight on the
The S-UPM was first used for the Taipei Finan- rams is released during a test, the compressed nitro-
cial Center for 80m piles in 1999 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
5 THE HYDRAULIC CATCH MECHANISM (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.
557
can be filled with sand, gravel, water or any combi-
nation to attain the required mass
558
Figure 15. Artist’s impression of a water re-
action pile-driving tool.
Figure 13. Water reaction mass containers Statnamic testing using water reaction mass was first
done by Berminghammer in 1998 in Hamilton Har-
the test load. The necessary 5% orlO% reaction mass bour. These tests went to 600kN. The first field-
would be provided by water confined within a vessel testing was performed for the Port of Lake Charles
and submerged below the surface of the water. This in Louisiana in 1999 by Applied Foundation Testing
weightless reaction mass makes it possible to per- assisted by Benninghammer. The loads were up to
form very large tests of longer duration than are 5 MN.
practical today. The use of water reaction will also
make it possible to drive piles underwater with a
tool, which will be virtually weightless.
9 PILE DRIVING WITH WATER REACTION
10 EVENTS
Figure 14. Water reaction mass testing In March of 2000, the Japanese Geotechnical So-
ciety (1998) published a standard for “Method for
Rapid Load Test of Single Piles”
559
An ASTM standard for Statnamic type pile test- The velocity dependent soil behavior can be split
ing is in progress. up in soil damping phenomena and strain rate de-
pendency. Soil damping phenomena can be derived
straightforward from a Statnamic test. Strain rate
1 1 RESEARCH AND DEVELOPMENTS dependency for fine-grained soils is still subject to
study, for example by the University of Sheffield
Companies and universities have accumulated up UK (A.F.L. Hyde. Et all, 1998). Well-documented
to date more than 700 case histories throughout the data from pile load test projects is becoming avail-
world. The total number of contract Statnamic tests able to support the insight in strain rate effects
worldwide has exceeded 1000, with testing now oc- (Holyman, et al., 2000).
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- 12 CONCLUSIONS
ing 300. Similar numbers of contract tests have now
been performed in the United States. The UK and The success of Statnamic stems for a significant
Japan are close behind in their numbers of tests. part from the concepts and ideas generated by “art-
ist-engineers” and “engineer-artists”.
An important contribution to the development of
Statnamic was supplied by the Japanese Geotechni- According to Brand1 (2000): an excellent engi-
cal Society, which established the research group of neer requires not only a firm theoretical knowledge
Rapid Pile Load Test Methods in 1993. Professor but also comprehensive experience as well as engi-
Osamu Kusakabe of Tokyo Institute of Technology neering feeling and intuition in equal parts. The au-
chaired the group. A strong promoter and initiator of thor would like to add: the ability to be creative and
Statnamic in Japan was Makoto Tsuzuki of Fugro think in unconventional ways.
Japan. The research group included 30 private insti-
tutes and companies as members. The activities of The success of Statnamic can be further explained
the Research Group aimed at cataloging the existing by the high degree of international cooperation and
knowledge about rapid load tests, examining the ba- research, which has brought the technology to the
sic characteristics and the applicability of the test, forefront .
and producing scientific interpretations of the Stat-
namic test results. The Statnamic community originated from the
stress wave community and the author is sure that
The University of South Florida (1998) con- they will remain in close contact. Both have a com-
ducted over 150 Statnamic tests in conjunction with mon interest in the research of dynamic phenomena
privately and federally funded test programs. The of soils and the development of tools for the load
tests programs have included: (1) axial load tests on testing of piles. The incorporation of the ideas of
piles and shafts in sands, clays, or rock-sockets, (2) “artist-engineers” and “engineer artists” will guaran-
lateral load tests on pile groups and shafts, and (3) tee more marvels in the development of pile testing
plate load tests on sands and full-scale spread foot- applications and other fields of engineering.
ings on sands and vibro-compacted soils (stone col-
umns).
13 REFERENCE:
The application of Statnamic produces excellent
results in stiff andor granular soils, although loading Bermingham, P., Janes, M., ”An innovative ap-
rate effects have to be taken into account. The influ- proach to load testing of high capacity piles”, Pro-
ence of soil viscosity alongside buildup of pore wa- ceedings of the International Conference on Piling
ter pressure in fine grained soils requires further de- and Deep Foundations, London, 1989
velopment of analysis tools and experience.
(E.L.Hajuk et. al. 1998). Middendorp, P. Bermingham, B Kuiper, Statnamic
load testing of foundation piles. 4th International
The soil viscosity shows up in two different Conference on Stress Waves, The Hague, Balkema,
ways: 1992
- Creeping, this means continuing settlements
under constant pile load Galluzi, P., Mechanical Marvels, Invention in the
- Velocity dependent soil behavior age of Leonardo, ISBN 88-09-20959-1, Instituto e
Museo di Storia della Scienza, Florence, 1996.
Creeping phenomena cannot be determined with
Statnamic, dynamic load testing and in many cases Baudson, M., Panamarenko, Paris 1996
not even with static load testing.
560
Fellenius, B., Welcome from the Chairman, First In-
ternational Statnamic Seminar, Vancouver, 1995
561
14 APPENDIX
Assumption:
The long duration Statnamic load F,,, allows modelling of the pile as a
zoncentrated mass (M) and springs
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
562
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 1503
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
sis, it cannot be applied to a complex system such as
At present, the primary approach to investigating a pile-raft foundation or batter piles Moreover, it is
pile behavior under statnamic loading is to make a obvious for a horizontal loading case that the analy-
comparison between the results of full-scale stat- sis should be done in three-dimensional space
namic and static load tests. Unfortunately, this cali- In order to develop a unified tool that can be ap-
bration method has an inherent problem in its testing plied for various types of problems, a three-
consistency, as was addressed by Amir & Amir dimensional finite element analysis program called
(1995). Under the uncertainty and complexity of the DYSC (Dynamic and Static systems analysis Code)
target ground, the situation whereby any model can is originally developed No-tension criteria and a
be validated by calibrating statnamic load tests simple elasto-plastic model based on the Drucker-
against corresponding static load tests can be dem- Prager theory are applied as the yield functions for
onstrated as follows: the model ground An interface layer element is in-
serted at the interface between the pile and the soil
1 If a comparison is made of the results of tests on For the pile body, a linear elastic relation is used
the same pile, the tests have to be done in se- since the applied force is lower than its yielding
quence. Therefore, the quality of the later tests point
invariably deteriorates due to the induced residual It is thought that before applying the DYSC to the
stresses from the former tests. simulation of piles, its fidelity should be checked by
2 If a comparison is made between the results of a physical model For calibration purposes, a com-
tests on different piles, significant samples should parison of the analytical results to the laboratory test
be tested in order to achieve some degree of reli- results is preferred to the results from the field tests
ability over the ground uncertainty. The reasons are due to the complexity of the ground
and the qualities of the geotechnical parameters re-
lnstead of studying the data from the field tests, the quired for the analysis
mechanism of the piles during statnamic loading can In our previous study, a small-scale statnamic
be thoroughly investigated by the numerical analysis loading device (3SLD-Mkl) was used to conduct
approaches. Applying the fact that the system of in- laboratory tests of piles under vertical loading (Ki-
terest is axis-symmetry, Matsumoto (1998) applied a mura et a1 1998) This device is used to conduct ex-
2D-FEM program to analyze the behavior of a single periments under a controllable environment Conse-
pile under vertical loading. Although a two- quently, guaranteeing that the series of tests are done
dimensional analysis is simpler and the solution can under the same conditions is possible Accompanied
be found faster than with a three dimensional analy- 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
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
0.0 0.2 0.4 0.6 0.8 1.0 1.2
Di sp1acein eiit (m in )
(4) Figure 3. Load-displacement relations from statiiaiiuc simula-
tion and experiment.
565
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
5 CONCLUSIOK
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
567
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 1503
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.
569
eleven elevations are used to measure the bending
moments along the shaft
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
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.
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
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
00 02 04 06 08 10 I 2
Displacement (mm )
Flgurc A 1 Load-displaccincnt rclallons of a inodcl pilc undcr
1 crlical loading
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
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.
575
2 THEORETICAL BACKGROUND 5. The unloading force of the static soil resistance
overcomes the opposing forces and pushes the pile
2.1 Unloading Point Method (UPM)
upwards. At the end, final settlement can be ob-
The long duration of loading (N, > 12) allows the served.
pile to be modelled as a rigid mass on which the fol- The static load behaviour can be derived as fol-
lowing forces act: lows:
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; or
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; At each point, the inertia force and the damping
Static resistance force, FU-SlatlC; force can be calculated, resulting in the following
Water pore pressure, Fp; static resistance:
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- All terms of equation (2) can be determined from
ity, the pore pressure resistance is included in the the Statnamic loading data, except the damping fac-
damping. Figure 1 shows the Statnamic load- tor Cmean. In the unloading point method, at maxi-
displacement with dynamic components. mum displacement the velocity and therefore damp-
ing force is zero. The damping factor is calculated
fi-om the average of
576
'automatic' matching routine to help the engineer in Usually, time signals are obtained froin measure-
the matching process Early automatic matching rou- ments. revealing information of lower pile parts and
tines in 1984 calculated pile shapes from measured soil layers up to the reflection of tlie pile toe with in-
signals (Ypey 1984) and determined the pile shape of creasing time It is therefore useful that the method
several model piles measured in research conditions considers new data points iteratively, updating the
without soil As soil influences were not considered, parameter values and their reliabilities
however, the accuracy of practical results was v e q To satisfi the above properties, tlie Kalman fil-
low I n 1 989, more advanced research programs be- tering method was chosen The I<alman filter was
gan (Staalduinen 1989) I n 1992, a routine for auto- originally developed to identie the behaviour of dy-
matically calculating measured signals was developed namical systems and has proben to be a reliable tool
(Courage 1994) 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
The routine used in automatic signal matching of soil 1986) For parameter estimation, more information
or pile parameters has a dif-f-'erentbasis now instead can be found in Hendrilts et al (1990) and Courage
of directlg. calculating the pile shapes and soil models, et a1 (1994)
TNOWAVE matches signals as an engineer would, The method is schematically depicted in Figure 2
with a "search strategy" to find the best fit between With the impact signal and initial pal-ameter Lalues
calculated and measured signals Superbision by an given by tlie user, TNOWAVE calculates the re-
engineer, with experience in signal matching, geo- sponse of the system by the forward model This re-
technical Anowledge, and practical experience in sponse together v. ith the measured response is then
foundation engineering, is still 1-equired Automatic supplied to tlie Kalman algorithm as implemented
signal matching is not replacing the engineer, but of- nithin TNOW 4VE automatic signal matching It is
fers support to perform matches more easily and based on the difyerence betmeen both responses, the
faster 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
The p remi ses for deve1opi ng autoin at i c signa 1 Proper gain is calculated to update the actual pa-
matching in TNOWAVE were rameter \ alues Besides iie\b parameter values,
A mathematical model \n/hich accurately describes po,/erior covariance for the parameters are also cal-
the behaviour of foundation piles in soil layers, culated
with an appropriate paraineterization of the pile The procedui-e is repeated until a convergence
properties and pile-soil interaction In criterion is met, for instance small residuals in calcu-
TNOWAVE, the mathematical model is based on lated and measured response, small updates for the
the method of characteristics, developed by Voi- parameters, or a maximum number of iterations As
tus van Haniine ( 1 974, 1977), Middendorp described above, the data can be considered itera-
(1 966), and others tively, updating values from new data points All
* An accurate and efficient computational algo- features are incorporated in TNOWAVE by means of
rithm to solve the boundary value problem de- a user-friendly menu-structure for input data and
scribed by the tnathematical model Such an algo- control parameters as well as the output and presen-
rithm is available in TNOWAVE and is referred tation of results calculated
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
577
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
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
28
20
12 I
4 1
4
0 100 200
-Calc. and --measured Time [ms]
579
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.
Figurc 9: Results of Statnamic prediction for displacement (a) Bielefeld. M.W.. 1991. Prediction of Installation of Sheet Piles
and force at top of pile (b). 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
following parameters have been used to model the Driving PreQction for Impact Hammers and Vibratory
Statnamic device: combustion chamber area 0.2082 Hanuners. Fourth Inrerna~ionalConference on the .4ppli-
m2; initial chamber volume 0.12426 m’; vent diame- cation of Stress Wave Theorv t o Piles. The Hague. 21 -24
ter 0.15875 m; vent length 0.1016 m; vent distance September.
Courage. W.M.G. and Bielefeld. M.W.. 1992. TNOWAVE
0.1524; plenum volume 1.24542 m3; plenum vent automatic signal matching. Application of S/re.s.s-J b l w
area 0.02 194 m. Theory to Piles, F.B.J. Barends (ed.) Balkema. Rotterdam.
The Statnamic simulation starts with the pre- ISBN 9054100826.
loading of the pile by the reaction mass. After 20 ms, Courage, W.M.G. and Foeken, R.J. van, 1992. TNOWAVE
he1 combustion in the Statnamic chamber begins. Automatic Signal Matching for Dy-namic Load Testing.
The pressure in the chamber increases rapidly and as Application of 5’tre.w-Wave Theoiy to Piles. F.B.J. Barends
(ed.) Balkema, Rotterdam, ISBN 9054100826.
a result the mass is accelerated and launched. The re- Foeken, van R.J., Daniels, B. and Middendorp, P., 1996. An
sulting reaction on the pile top introduces forces improved method for the real time calculation of soil re-
causing stress waves to travel in the pile. Shaft fric- sistance during h v i n g . Application of Stress- Wave Theory
tion and toe resistance will be mobilised. The dis- to Piles, F.B.J.Barends (ed.) Balkema, Rotterdam.
placement of the pile will have its influence on the Hendriks, M.A.N., Oomens, C.W.J., Jans, H.W.J, Janssen,
J.D. and Kok, J.J., 1990, A Numerical Experimental Ap-
burning fuel. The calculated displacement and force proach for the Mechanical Characterization of Composites.
as hnction of time is shown in Figure 9. In V. Askegaard (ed), Proceedings of the ninth Interna-
The predicted jump height of the reaction mass is tional Conference on Experimental Mechanics, Aaby Tryk.
4.46 m, a conservative result since friction forces Copenhagen.
from gravel are not considered. The calculated force Horvath R.G., 1990, Statnamic, an accurate and innovative
and displacement are about 10% to high. A good load test method for high capacity deep foundations. Foun-
dation Drilling, Volume XXVIII, No. 11,
comparison can be obtained by using 10% less hel, Jazwinski, A.H. 1970, Stochastic Processes and Filtering
but can not be currently confirmed by other back Theoty. Academic Press, New York and London.
analysis. Kalman, R.E., 1961. A new Approach to Linear Filtering and
PreQction. Trans. ASME.
Koten, H. van, Middendorp, P. & Brederode, P. van.,1980. An
4 CONCLUSIONS analysis of dissipative wave propa-gation in a pile. Intl.
Seminar on the Application of Stress-Wave Theory on
Piles, Stockholm, 4-5 June.
For wave numbers greater than 12, the UPM (un- Middendorp, P., Foeken, R.J. van, 1998. When to apply Dy-
loading point method) can be straightforwardly ap- namic Load Testing or Statnamic Load Testing, Second
580
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
583
Figwe 1. Soil profile and results of site investigations at the test site.
584
pile driving. The dynamic signals (time variations of the shaft resistance along the top 3 5m of the pile
of accelerations and strains at the pile head) were at 6 days after the pile driving may be attributed to
monitored by a PDA (Pile Driving Analyzer) and an the fact that a gap between the pile shaft and the
FPDS (Foundation Pile Diagnostic System), which surrounding soil was generated by lateral
are widely used in the world on a commercial basis. movements of the pile near the ground surface
Two different computer programs, CAPWAPC and during the pile driving. The shaft resistance along
TNOWAVE, were employed for the wave matching the top 3.5m of the pile recovered after a rest period
analyses of the dynamic load test signals to estimate of 30 days, although it is difficult to comment on
the static load-displacement curve for the pile. this mechanism.
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 4 DYNAMIC LOAD TEST R E S a T S
conducted 30 days after the initial pile driving. The
4.1 The resirifsdirring the irzitiai driwig
measured load-disp~acementcurves were compared
with those estimated from the dynamic load tests on Figure 5 shows the change in the total resistance, Rt,
pile D to examine the applicability of the dynamic with penetration depth of the test piles during the
load test to the estimation of the static load- initial driving. The total resistance, Rt, was
displacement curve for a pile. estimated by the Case method (Goble et al. 1975).
The Statnamic test was also carried out on pile S The total resistance of each test pile increases almost
after the completion of the series of static load tests,
52 days after the initial pile driving (Nishimura et al. Load on pile head (MN)
1995).
585
linearly with the increasing penetration depth and It is recommended that the relations (EHvs R,and
attains about 2 MN at the end of initial driving. ~ 5') as shown in Figure 6 be obtained prior to
E I vs
Therefore, the bearing characteristics of all the test the actual pile driving, in order to assure that the pile
piles are comparable with each other. 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
4.2 The ii$'iierrce of hammer driving energy
actual pile capacity or it is a limitation due to the
The static pile capacity, R,,estimated from the insufficient hammer driving energy. The relations
dynamic load test may vary with the change in the shown in Figure 6 are useful for the selection of an
driving energy actually transmitted to the pile To appropriate hammer for each site.
investigate this aspect in detail, the dynamic load
tests were carried out on pile D changing the
4.3 Set-irp phenomerza
hammer driving energy after different time intervals
from the initial pile driving Dynamic load tests were performed after 8 different
Figure 6 shows the change in the static resistance, time intervals measured from the end of initial
R,,estimated from the wave matching analysis of the driving, in order to investigate the set-up phenomena
dynamic load test signals using CAPWAPC, as a of a steel pipe pile (pile D) driven in sandy ground.
function of the nominal hammer driving energy, EH. Figure 7 shows the change in the total resistance, Rt,
The measured values of set per blow, S, are also which was estimated using the Case method, with
shown in Figure 6 The static resistance and the set elapsed time after the pile driving. Two dynamic
per blow increase gradually with increasing hammer monitoring systems, PDA and FPDS, were used to
driving energy to 75 kNm. The static resistance record the dynamic signals. The total resistance, Rt,
attains it's peak value of 2 MN when the hammer reaches it's peak 60 min after the initial pile driving,
energy is greater than 80 kNm On the contrary, the and remains almost constant after that time. The
set per blow increases steeply for hammer driving peak total resistance is 1.2 times R, measured at the
energies greater than 75 kNm, indicating that the end of initial driving, that is to say, the "set-up ratio"
pile capacity is fully mobilized by hammer driving for the total resistance is 1.2 in this case.
energies greater than approximately 75 kNm.
Figure 7 Increase 111 total resistance. R,. nit11 elapsed tiine after
inihal driving (pile D)
586
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
dunng each blou at different depths of the pile tip (pile D) -: 32:
'5
1-
587
Shaft resistance, z (kN/niL) Shaft resistance, z (kN/m’) calculated distribution of the shaft resistance along
oO 100 200 300 0 100 200 300 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
(a) 6 days after EOID (b) 30 days after EOID using the wave matching analyses are compared
Figure 11. Distributions of shaft resistance estimated from re-
with the load-displacement curves measured in the
driving tests using TNOWAVE. together with static load test static load tests in Figure 12. The estimated load-
results (piles D and Sj. 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).
588
Pile h e a d load (MN)
-
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
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
590
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
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.
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
592
3 . 2 Test objectives
Figure 2 presents a breakdown by objective of 3.4 Pile length
Statnamic tests at the test site. The Statnamic tests Figure 4 presents a breakdown of the number of
have been conducted to estimate the bearing sites with respect to pile length and pile type. Large
capacity and load-displacement curves of newly shares of precast concrete piles having a length less
constructed piles in 57% of the test sites, and of than 10m and cast-in-situ concrete piles having a
existing piles in 12% of the test sites. The Statnamic length less than 20 m are seen, whereas most steel
test was employed in 14% of the test sites for the pipe piles range in length from 10 to 40 m, 70 m in
development of new pile construction methods. The maximum. The longer lengths of the steel pipe piles
remaining were research projects, concerning can be attributed to the fact that most steel pipe piles
aspects of the Statnamic test itself such as loading are offshore piles, while all the concrete piles are
characteristics, dynamic effects, applicability to the constructed on land.
estimation of static load-displacement curves and so
on.
593
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.
z
z-
-U
CCI
-
0
-0
-a,
.-
Q
Q
Q
0 30 60 90 120 150
Time (ms)
Figure 8. Definition of loading duration.
594
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
Figure 9. Breakdown of the test number with respect to loading forces
relative loading duration and pile type. The relationships between the pile head force,
Fstn, and the pile head displacement, w, are shown in
The Unloading Point Method (Midendorp et a1 Figures 11, 12 and 13 for the cast-in-situ concrete
1992, Kusakabe & Matsumoto 1995) is usually piles (CISC), the steel pipe piles (SPP) and the
employed to estimate a static load-displacement precast concrete piles (PHC), respectively In these
curve for the pile in the case of Tr > 5 . In cases of figures, Fsttl is normalized by the maximum pile
Tr < 5 , the wave-propagation analysis (wave head force, I ~ ~ t ~ land
( zw( m ~~),
is normalized by the pile
matching analysis) is usually conducted to estimate diameter, d The F&v curves may be classified
the pile capacity and the corresponding static load- into two groups In one group, the F\t,,-w curves
displacement curve (for example, Ochiai et a1 1996). have a characteristic 'tear drop' shape in which the
pile displacement, w,tends to increase and decrease
3.8 Pcrr.allel test nielhods with increasing and decreasing the pile head force,
FqtI1, regardless of the kind of pile In the other
The types of load tests other than the Statnamic test group, the Fstn-~i' curves have shapes far from the
conducted at the test site are given in Figure 10 in 'tear drop' shape In the Statnamic tests classified
the form of a breakdown by the number of sites into the latter group, the pile displacement continues
The Statnamic test (STN) was carried out alone at to increase after the start of unloading of I i m ,
56% of the test sites, whereas dynamic load test although the recovery of the displacement occurs at
(DLT) and/or static load test (SLT) were conducted a later unloading stage of F,t,, Such behavior of the
in parallel at the remaining sites. At most test sites J;stn-wcurve is notable for cast-in-situ concrete piles
where the Statnamic test alone was conducted, it which have relatively large masses compared to
was confirmed that the pile had bearing capacity steel pipe piles and precast concrete piles
greater than the design load that was derived from According to single mass modeling of a pile
pile design codes. during Statnamic loading (Middendorp et a1 1992),
On the other hand, Statnamic test results were used the total soil resistance, Fsoli,is given as
for calibration of a dynamic load test at the test sites I;'. -1-
where both load tests were carried out for quality soil - 'stn - M a
assessment and driving control of the other piles at 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.
Figire 10 Breakdonn of site number uith respect to load test The Fsoil - IQ curves are again classifiable into two
methods groups: the group with elastic responses in which the
595
I”soli- it’ curves have the characteristic ‘tear drop’ elastic or the plastic response type is possible from
shape, and the group with plastic responses in which the measured pile head displacement immediately
plunging loads are clearly detected in the F s o -~ w~ aRer the completion o f the ~tatnamictest, this may
curves. It may be judged that the pile exceeded the be useful An overview of the measured pile head
‘static’ yield load in the Statnamic tests classified displace~entsis provided below for this purpose
into the group with plastic responses, although the The maximum and residual pile head displacements
dynamic soil resistance, F,,, which depends on the are shown in Figures 17, 18 and 19 for cast-in-situ
pile penetration rate, is not corrected for. 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
If the classification o f Statnamic test results into the can be seen that the Statnamic tests classified into
596
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
597
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
pile obtained in the Statnamic test For the Rot t crdaiii. Ba lkcnia
Statnamic tests classified into the elastic response Nisluinura. S . A Shibata, K Yaiiwshita, N K ~ t aN Ogita &
group, F,o,l(md,) almost equals FI I P for most cases M lsluda. 1998 One &ineiisional stress wile simulahon
On the contrary, for the Statnamic tests classified malj sis of Stahiamc load tests, Statnanuc loading test,
Pmc 2 ~ dhit Srntmmrc Senrincw T o h 0, Japan
into the plastic response group, t(\ol~(l,,d\)exceeds
Balkeina (to be pnnted)
FITp by 5 to 30% of F1.1 p, indicating that sufficient Rcscarch Connnittcc on Rapid Pilc Load Tcst Mcthods. 1998
load was applied to obtain the static pile capacity Rcscarcli Actii itics tovard Standardmtion of Rapid Pilc
The difference between F\ol~(,nrl\) and FIIIP may be Load Tcsl Mctliods i n Japan. Statriunic loading tcst
regarded as the dynamic soil resistance which l’roc 2nd lnr Sfntnniiiic Seiirrnar. Tolq 0. Japan Balkeiiia
depends on the pile penetration rate (to be published)
Figure 22 presents the relationships between JGS (Japanese Geotecluucal Socieg ) Conuiiittee on
Standardization of Pile Loaduig Tests. I998 Draft of
and k i ~ pThis figure indicates that the
k~~l,~llla,) “Method for Rapid Load Test of Single Piles (JGS 1815 -
planned maximum force could be determined to be 2000)”. Statnrunic loadmg test. l’roc 2nd Int S’tntnrnrr~c
I I to 1 3 times the expected static pile capacity 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 -
5 CONCLUDING REMARKS Kjushu Airport, Proc 4th 1171 Conf on the. fpplimnfion of
Sfress-Ifirve Iheorv to Piles (B J Barends Ed ), The
An overview of the case studies of the Statnamic Hape. The Netherlands 997- 1014 Rotteddain Balkeina
tests conducted in Japan has been presented in this Tada. K . M Ohnishi. Y Cliono. S Nishiniura & S Sakaiiioto.
paper, showing their chronology and statistics. The I997 An cffcct of loading Lclocity r a m at latcral
loading test for a steel pipe pde. l’roc 321id -lni?iinl
results of the Statnamic tests also have been Meerrng, Japanese Geotecluucal Societj . Kuinanioto
reviewed from the view point of whether the 1601-1602 (in Japanese)
ultimate capacity of the piles was obtained from the Terada, M . R Nakatsuka, K Olmiwa, N Akisato, K
Statnamic test or not. It was shown that the ultimate Honkoshi & S Nishiinura, 1998 Applicafion of the
pile capacity was obtained in a large number of the Statnamc loadng test for large-scale pier construcfion
Statnamic tests, encouraging the use of the Statnarmc loadmg test. Proc 2nd Int Statnainrc Seminar.
Tokyo, Japan Balkema (to be published)
Statnamic test to estimate the design pile capacity.
598
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.
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.
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
604
Figure 8. Comparison of distributions of shaft friction.
5 DISCUSSION
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 Figure 10. Comparison of load-displacement curves.
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 6 CONCLUSIONS
in 1100 hours (46 days) and dynamic tests and
Statnamic test had generated large pore water Safety of the bearing capacities for steel pipe piles
pressures around the pile. seated on the thin bearing stratum were confirmed
Figure 9 shows obvious correlation between the by the result of Statnamic and Dynamic load tests. In
recovery of the pore pressure and the set up addition, the site-specific driving formula to control
phenomena of the shaft resistances in elapsed time driving of the production piles was developed based
after the end of initial driving. on the test results.
The enforcement of the load tests at the wharf in
Nagoya port materialized economical and reliable
5.2 Toe resistance pile installation in such tricky soil conditions.
Toe resistances obtained fi-om the load tests were Following interesting result was obtained from the
compared with the design values estimated. Those series of Dynamic and Statnamc tests.
values are shown in Table 16.
Shaft resistance of the alluvial clayey stratum
shows large recovery in its set-up behavior.
Table 16. Comparison of toe resistances. Pore water pressures around the pile in the alluvial
r - - - Statnamic (KN) Dynamic(KN) Designed value
Lilbf: clayey stratum had increased at the time of
(Matching) (Matching) (KN) Dynamic and Statnamic tests and settled down
v -1 5069 3378 4760
natural pressure within 1100 hours (46 days).
v-2 7076 442 1 7140
v-3 5069 2813 6170
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
The toe resistance was not fully mobilized in that of without the cross-rib plates in the load-
every case; therefore the effect of the thin bearing displacement curves.
stratum (gravel with sand; 4.5m to 6.5m thickness)
could not identify from the test results.
Figure 10 shows the relationship between force REFERENCES
and displacement at the head of test piles during
Statnamic test. Stiffness of load-displacement curve Japanese Geotechnical Society, 1998. Draft of “Method for
of the V-2 pile with the cross-rib plates at the pile Rapid Load Test of Single Piles (JGS 1815 - 2000)”,
toe is larger than those of the V-1 and V-3 piles Proc. 2nd Int. Statnamic Seminar, Tokyo, Japan.
Balkema (to be published)
without the cross-rib plates. This certainly suggests Karkee, M., T. Horiguchi and H. Kishida, 1997. Static and
that increase in load capacity of pile due to the dynamic tests for evaluation of the vertical load bearing
presence of the cross-rib plates at pile toe. However, capacity of piles, Proc. 22nd Annual Member’s
the piles were not loaded up to yield values, Conference, Deep Foundations Institute : 199-214.
therefore it is difficult to verify an effect of the Kusakabe, 0 & T. Matsumoto, 1995. Statnamic tests of
Shonan test program with review of signal interpretation,
cross-rib plates on allowable resistance of the pile Middendorp, Proc. 1st Int. Statnamic Seminar,
toe. 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
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.
609
provide the mass needed to perform a Statnamic test, was able to act on the mass of water as well as the
thus eliminating the mobilization cost of the con- mass of the steel container. As expected. the jump-
crete or steel masses. ing height of the mass was greatly reduced. Unfor-
Middendorp and Courage of ” N O Building and tunately, this configuration would place greater
Construction Research performed one of the first physical demands on the catching mechanism.
theoretical studies of water as Statnamic reaction The third model presented the most attractive so-
mass in 1995. In this analysis three configurations lution. In this model the top of the steel container
of the underwater container were examined. was left open, and valves or ‘trap-doors‘ were intro-
The first mathematical model described a com- duced to the bottom of the container. This cdnfigu-
pletely closed steel container, filled with water and ration allowed for the same mass of water to be
submerged. The steel container was assumed to re- ‘contained’ as in the first two models. thus providing
main entirely below the water surface throughout the equal inertial reaction for the Statnamic loading
test. After the initial upward acceleration of the event. During the upward acceleration of the Stat-
Statnamic test, it was calculated that the continued namic device the valves in the container bottom re-
upward movement of the container would be very mained closed. During the deceleration phase of the
large due to the upward momentum of such a large event the contained moving mass of water was al-
volume of water. The deceleration force of gravity lowed to flow through the container rather than ‘lift’
could only act on the mass of the steel container and the container along with it.
not on the mass of water, which was weightless Modest jumping heights were observed in this
while submerged, and moving upward very quickly. container configuration, even when the container
The main advantage of this first model was that a remaining completely submerged throughout the
relatively small amount of mass would need to be event. This was the most favorable model for the
’caught‘ by whatever type of catching mechanism design of an over-water catching mechanism.
was to be devised. The disadvantage was that an un- A schematic of this third model is shown in Fig-
realistically large jumping height of the mass would ure 1 . Figure I also contains the definitions of the
need to be accommodated. 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’
610
Most of the remaining questions centered on the pile. The top of the container was constructed with a
physical concerns of assembling such a test. The flat top with six 250 mm holes. The container was
only remaining task was to undertake a physical ex- fabricated with a flat bottom with six 250 mm holes.
periment. The purpose of the bottom holes was to allow the
container to fill with water as it was lowered over
2.2 Experiment the test pile. Each 250 mm bottom hole was also
equipped with a hinged door that opened in the up-
In the spring of 1998, a series of 12 Statnamic load ward direction. These doors were designed to re-
tests were performed in Hamilton, Canada. These main closed during the upward acceleration of the
tests were conducted using a 0.6 MN Statnamic de- container to prevent the water from simply flowing
vice and a prototype submerged reaction mass con- through in the downward direction. The doors did
tainer. allow water to flow through the container in the up-
ward direction after the completion of a test, to pre-
2.2.1 Apparatus vent the creation of a low-pressure area at the bot-
Testing was conducted using a 0.6 MN Statnamic tom of the container due to the upward momentum
device. A hydraulic catching mechanism was also of the contained water and to act as the valves pro-
used to facilitate repetitive testing. The testing appa- posed by the Middendorp and Courage model. A
ratus, along with the fabricated reaction mass as- photo of the submerged reaction mass assembly is
sembly is shown in Figure 2. shown in Figure 4.
61 1
and bottom of the reaction mass assembly. Two sets For discussion purposes, the results of three typi-
of three rods were used. The first set was 8 m long cal loading cycles will be discussed.
while the second set were 4 m long. The two sets
were coupled together to provide the required length 2.2.3.1 Series 1, Test 4 - Results
to perform one series of tests with the top of the re- The first test to be discussed was performed dur-
action mass assembly 3.3 m below the water surface ing the first series of tests prior to the addition of the
(masses stayed below the water surface during test- ‘drag flange’. The peak Statnamic load was 480 kN,
ing), and another set of tests with only 0.3 m of with a measured upward acceleration of the reaction
submergence (masses jumped out of the water dur- mass assembly of 10 g. Typical upward accelera-
ing testing). tions for the same test, performed with a typical
The test pile was a 20 m long closed end steel amount of concrete reaction mass would have been
pipe pile 323 mm in diameter with a wall thickness approximately 15 g. The drag of the water provided
of 8.0 mm. The pile was installed using an MIST a reduction in the upward acceleration of approxi-
V16 vibratory hammer with a peak driving force of mately 50%.
1.1 MN. On land, with conventional reaction masses, the
The soil at the test location consisted of 7.8 m of jumping height of the masses would have been ap-
water, overlying soft marine silt sediments to a depth proximately 2.3 m, while the measured jumping
of 33 m. The purpose of the foundation was to pro- height of the water reaction masses was only 1.7 m -
vide a consistent elastic load deflection behavior for a reduction of 30%.
all tests with the loading capacity of the 0.6 MN The pressure transducers mounted on the reaction
Statnam i c device. mass assembly suggested peak drag forces on the
After the first seven load tests using the above re- underside of the reaction mass assembly of 35 kN.
action mass configuration, a 250 mm wide flange The peak drag force on the top of the assembly was
was welded to the outer edge of the water container. 58 kN. These forces occurred at the same time as
This flange (added to the bottom outer edge), essen- the peak Statnamic load. Together the drag forces
tially doubled the plan area of the reaction mass as- accounted for 93 kN, or about 19% of the applied
sembly. This was done to investigate the effects of load. Significant changes in the lateral pressure on
drag on the behavior of the underwater reaction the container were not observed suggesting minimal
mass assembly. drag along the sides of the reaction mass assembly.
The total reaction mass for this testing was 3350
kg, with 1490 kg (44%) provided by the Statnamic 2.2.3.2 Series 2, Test 10 - Results
cylinder, silencer and reaction mass assembly. The This series of tests was performed with the addi-
remaining 1860 kg (56%) of the reaction mass was tional drag flange welded to the bottom of the reac-
provided by the contained water. tion mass assembly.
The peak Statnamic load for this test was 496 kN.
2.2.2 lnstrume ntation The load was very near to the previously discussed
Accelerometers were placed on the reaction mass test, however the 3% increase in the peak force was
flange of the Statnamic device. These measure- achieved without increasing the amount of Stat-
ments were used to determine the acceleration of the namic fuel. This was likely achieved by a combina-
under water reaction mass assembly. The jumping tion of the increased mass of the drag flange ( 1 20
height of the Statnamic device was measured after kg), as well as the increased drag at the peak load.
each load cycle. The peak upward acceleration of the reaction
Pressure transducers were mounted in various lo- mass assembly was approximately 9 g rather than 10
cations, both inside and outside the water reaction g. And the jumping height of the masses was only
mass assembly. These measurements were made to 1.4 m, down from the previous 1.7 m, a reduction of
help investigate the behavior of the water both inside 18%.
and outside the steel container. The pressure transducers indicated similar pres-
In addition to the above special instrumentation, sures to the previous testing, both on the top and bot-
the typical Statnamic load cell, laser displacement tom of the reaction mass assembly. The increased
sensor, and pile accelerometer was also used. area on the bottom of the container translated into an
estimated peak drag force of 85 kN.
2.2.3 Typical Results The addition of the drag flange increased the total
For all of the loading cycles performed in this ex- drag resistance to approximately 140 kN or 28% of
periment, the test pile was not displaced beyond its the peak Statnamic load.
elastic range, providing an essentially rigid test
foundation. Peak displacements of 7-8 mm were 2.2.3.3 Series 2, Test I ? - Results
typical for the test pile, with net displacements of Test 12 provided data on one of the highest peak
zero. loads of all the tests performed. This test was also
612
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
613
Figure 7 Statnamic equipment being placed on a test pile
614
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 Figure 9 Load vs Time using concrete and steel as reaction
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 when the catching cylinders were deactivated. As a
the reaction masses. Unfortunately, data was not consequence, the descending silencer and reaction
collected on the upward acceleration of the reaction masses produced a sharp impact load on the catching
mass assembl>i. From the test data the force-time frame. Several of the steel teeth as well as the me-
curves give the most information about the behavior chanical latches showed noticeable wear. The catch-
of the testing apparatus. Jumping heights were re- ing device functioned well nonetheless, and proved
corded, but are not presented here. the idea of a mechanical catching mechanism to be
Figure 8, shows the force-time curve for one of sound.
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. 4 SUMMARY AND CONCLUSIONS
The similarity between the two graphs shows the ef-
fectiveness of the water reaction mass assembly. The main conclusion drawn from the work described
Closer examination of Figure 8 shows a loading in this paper was that water could be used success-
event of just over 100 ms duration. Figure 9 is at fully as reaction mass for Statnamic testing.
least 10 ms shorter in duration. This was the ex- The theoretical work by Middendorp and Cour-
pected result since the land test was performed with age was useful in the design of the water containers.
a total reaction mass of 22,000 kg, whereas the water In particular, the use of a valve at the base of the wa-
test used closer to 25,000 kg. Although the results ter containers was highlighted in their work as an
were not surprising they do offer direct evidence of important element. In the 1998 experimental testing
the effectiveness of the contained water as reaction the mass of the steel containers was high in relation
mass for the Statnamic test. to the mass of contained water, hence the Midden-
Also of ncte was the successful operation of the dorp and Courage valves had only a moderate effect.
mechanical catching mechanism. As the Statnamic The containers used in the Lake Charles contract
silencer reached the high-point of its jump it was testing omitted the valves entirely. As the percent-
noted that the subsequent downward movement was age of contained water increases in relation to the
abrupt - similar to the 1998 experimental testing 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
Figure 8 Load vs Time usin$ water as reaction mass used to catch the water reaction masses. It also
615
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
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.
617
tion compared to dynamic load testing. tests indeed showed the same or similar results as
static tests.
As an example the force time diagrams of dy-
namic load test, a STN load test and a static load test During the first years the desire to use STN for
have been presented in Fig. I . The long duration of higher loads quickly rose. Statnamic devices to
the STN loading keeps the pile under constant pres- measure up to 30 MN were developed and are still
sure and tension stresses cannot develop. The central used in larger projects. From the beginning the eco-
location of the STN loading device on the pile top nomic gains in these larger projects were evident.
guaranties an axial introduction of the load onto the
pile. 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
The long duration of the STN loading causes for steadily growing, having performed up to present
all pile levels a similar displacement behaviour as some 300 STN tests in the USA alone. The way
can be observed with static load testing. This justi- STN was introduced in the USA was identical to its
fies a simple modelling of pile and soil, in which acceptance in the Far East. First comparative studies
stress wave phenomena do not have to be taken into were carried out between static and STN load tests
account. The pile is considered to be a mass on to confirm the reliability and accuracy of STN be-
which the STN force, the inertia force and soil resis- fore the transportation authorities adopted the
tance are acting (Fig. 2). method.
618
be dominant over quality assurance and as a result model for the strain rate sensitivity of clayley soils
often a lot of construction works go without regular under STN loading conditions.
testing of foundation piles. The reason is that static
load tests are considered too expensive, time con- Demonstrations have been trend-setting in adopt-
suming and therefore cumbersome. Dynamic load ing this new technology into the market. The method
testing is not always recommended because of its used by TNO Profound to introduce this new con-
limitations. cept is to organise seminars and training courses for
users and potential users to acquaint themselves with
For STN to be attractive applications are to be this new technology. This introduction involves a
found in the lower pile loads testing range up to 2.5- theoretical part in which the geo-technical back-
3.5 MN and competition with established low cost ground of STN is being explained and a practical
static and dynamic testing methods is strong. part in which the actual testing is demonstrated.
In addition, regulations in most European countries
are not yet adjusted in considering STN as an ac- Collaboration exists with the leading Universities
ceptable means for pile Ioad testing replacing or in and Research Institutes in the countries involved se-
conjunction with static load tests. curing adequate technical support, acceptance and
being able to open steps towards introduction of
STN into the national regulations. This means train-
3 INTRODUCING STATNAMIC IN THE ing, educating and continuously introducing techni-
NETHERLANDS AND EUROPE cians into the world of STN.
In response to the increased interest in statnamic New countries covered are Spain and Turkey,
in Europe TNO has decided to set up TNO Profound where they are very interested to get STN intro-
(Professional Foundation Diagnostics) responsible to duced. Preparations are made for a first STN demon-
introduce STN further in Europe. In combination stration project in Barcelona. In Poland and Hungary
with the knowledge base in foundation technology progress is made to get STN into the regulations and
present at TNO Building and Construction Research perform static-STN demonstrations in 2000. Infra-
a strong basis has made to make STN a success fol- structural EU supported investments for the pre-
lowing the examples in the Far East and the USA. accession countries are the driving force for the in-
troduction of STN.
3.1 Stntnarnic load testing with a 4MN device
3.2 Major considerations for the introduction of
Statnanzic on the European continent
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 Regulations
continent. This has opened up the market for STN in The tradition in most European countries is to
Europe. Steps are taken to position a second 16 MN conduct static load tests on specially made test piles.
device in the Central European market with the aim Loads to be applied range between 1.5 - 2.0 times
to have a STN testing facility available on local the design load. These static load tests require a lot
markets to conduct testing services at commercially of preparation work and piles are to be carefully se-
attractive prices. lected and prepared for testing. This is a costly and
cumbersome activity. For this reason static load tests
In 1999 TNO Profound has been engaged in several are virtually abandoned in The Netherlands for proof
research projects in Europe in which the results of testing and are becoming less popular in other coun-
Static and Dynamic Load Testing have been com- tries counties like Belgium as well. Time and con-
pared with STN test results on the same and/or simi- struction costs are increasingly under pressure
lar piles. Projects have been carried out with Jacbo squeezing cumbersome testing practices to the bare
minimum.
(The Netherlands), WTCB (Belgium), the Technical
In Europe most countries the use of alternative
University of Budapest (Hungary). Preparations are load testing methods is not foreseen in the national
being made to initiate similar projects in Germany, regulations. This hampers the wider introduction and
although Germany has a strong tradition in static acceptance of other innovative and more convenient
load testing the willingness to apply STN in Ger- methods. Mainly in Germany dynamic load testing
many as well is rising now STN tests become eco- has gained interest because of its convenience to use.
nomically more attractive. TNO Building and Con- A pile driving hammer or drop hammer can be used.
struction Research is participating in a research Piles to be tested can be selected randomly. The
project with the University of Sheffield to develop a same counts for STN where the reaction mass re-
619
quired is only 5% of maximum load to be applied. The price of STN and dynamic load testing is largely
Construction engineers of local organisations may influenced by the costs of mobilizing the testing
decide to accept DLT or STN when judged suffi- equipment. When the testing equipment is used for
ciently suitable for its purpose. Their application de- testing of one pile only the price remains relatively
pends on the availability of a testing capacity and is high. When more piles are tested on one project site
mostly used on a case-by-case basis, where conven- the price per pile drops considerably. When
ience, time pressure and economics play a major comparing STN with DLT, pile preparation for STN
role. For cast-in-situ piles STN is the preferred is easier, but more time is required for the
method (P.Middendorp, 2000). TNO Profound is installation of the testing device. Four to 6 piles can
promoting to consider also alternative testing meth- be tested a day with a STN device with catching
ods like STN into the national regulations of EU mechanism.
countries. This is believed to be an essential prereq-
uisite to get STN widely accepted in Europe. STN
Transportation costs in Europe are considerably
demonstrations with the involvement of the acade-
higher than in the USA and allowable transportation
mia and regulation authorities are required to get de-
weights per truck are maximized. Therefore in
cision makers acquainted with the method.
Europe it is preferred to move empty containers to
be filled on the site with load material, whereas in
Statrzanzic as the only feasible alternative for test- the USA it’s easier to transport concrete/steel reac-
iiig
tion masses.
The major applications of STN in The Nether-
lands have been cases where the alternative testing
methods where not seriously considered because of 4 CASE STUDIES OF STATNAMIC IN THE
technical limitations (DLT on cast-in-situ piles NETHERLANDS
would not generate satisfactory results), time pres-
sure or undesired excessive costs. STN was not re- 4.1 Statizamic pile testing in Rotterdam
quested to replace routine static load testing, but was
considered to identify the load bearing capacity of Project Data
existing piles for re-use or in situations of doubt Location : Rotterdam Harbour
about the performance of disturbed soil profiles : Wilton-Feijenoord
where conventional calculation and testing methods Time : June 1999
could not be applied. Groundwater seepage has been Client : Van Hattum & Blankevoort
one of the major causes. STN would also be ideal in Testing : Prefabricated piles 420x420
situations where a lot of piles are to be tested in a 4 piles, 23 m long (new)
short time frame. For such larger projects STN is 2 piles, 24 m long (existing)
still lacking sufficient EU based reference projects. Load applied : 2 -4 MN
620
Fig. 4. STN Testing at 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
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.
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 Middendorp, P., Foeken, R.J. van, 1998, When to
Statnamic proved to be a simple and effective Apply Dynamic Load Testing and Statnamic Test-
way to test the performance of cast-in-situ founda- ing, 2nd Statnamic Seminar, Tokyo
tion piles in cases where CPT tests would give
doubtful results. STN testing confirmed that in some Middendorp, P, Ginneken, G.J.J. van, Foeken,
cases the interpretation of CPT tests alone could lead R.J. van, 2000, The advantages and disadvantages of
to questionable conclusions. With STN an effective
tool was provided to verify independently the actual x
D namic Load Testing and Statnamic Load Testing.
6' International Conference on the Application of
performance of the pile tested. Stress Wave Theory to Piles, Sao Paulo, Brazil
5 CONCLUDING REMARKS
6 REFERENCES
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
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
625
- economy
- mobilization of capacity
- chance on pile damage
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.
626
of elasticity (E) of the concrete and the pile cross The stress wave velocity is calculated from the
section (A). time (T) it takes for a stress wave to travel over the
pile length (L) from the pile head to the pile toe and
F=E. A.& (1) back to the pile head (Fig. 5). For this method it is
required that the reflection coming from the pile toe
The accurate determination of the properties E is clearly visible in the signals. In Fig. 6 the force
and A for bored piles is difficult in many cases. and velocity times impedance signals of two dy-
namic load test are presented. The first case shows a
To calculate the force from the measured strain in clear toe reflection and the stress wave velocity can
a pile during DLT we need to know the cross section be calculated accurately. If the toe reflection is not
and the modulus of elasticity of the concrete at the visible one has to estimate the toe reflection time.
measuring level. For piles with homogeneous mate- However an error in the estimated toe reflection time
rial the stress wave velocity (c) is used to calculate (T) and stress wave velocity (c) will result in a con-
the E-modulus with siderable error in the calculation of the E-modulus.
For example a 5% error in the stress wave velocity
E = c2. p will result in a 10% error for the E-modulus and cor-
responding load in the pile. Another option in this
c = 2L/T (3) case is to rely on an estimate for the E-modulus from
the pile material properties.
Knowing or estimating the stress wave velocity c
we can calculate the pile load at the measuring level
with the formula Making an estimate on the E-modulus is difficult
because it is not a constant value but depends on the
F= c’.p.A. (4) age and the quality of the concrete (Franklin,
1971)(Fig. 7), the loading rate (Sparks et all, 1973 ),
So the derived stress wave velocity has a strong and even the temperature of the concrete (Abbasi
influence on the value of the load measured in the 1990). For example, for static load testing the
pile. An error in the measured load will result in an modulus of elasticity for concrete is in the range of
error for the pile capacity prediction. 28 GPa to 32 GPa while for dynamic load testing it
is in the range of 32 GPa to 52 GPa
627
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 during construction. The quality of the concrete near
stress wave velocity c tor cast in situ piles is the fact the toe will in general be better than the quality of
that the concrete is not homogeneous. The concrete the concrete near the pile head. This also means that
quality will vary over the cross section and over the the stress wave velocity will vary with the pile
pile axis. The concrete in contact with the soil will length. So the stress wave velocity calculated with
be of lesser quality than the concrete in the center of c=2L/T is a mean value for the whole pile. The
the pile and the shaft area that has been in contact modulus of elasticity calculated from the toe reflec-
with the soil might be the location where the strain tion represents a mean value for the pile and there
transducers are mounted. The concrete quality dif- can be a considerable difference with the modulus of
ference over the pile length is caused by the pouring elasticity at measuring level.
procedure and the difference in concrete pressure
628
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
629
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
ate tension waves. Cast in situ piles are not designed The capacity of driven piles is mobilized at rela-
to withstand high tension stresses. As soon as allow- tive small displacements.
able tension stress levels are reached the impact en-
ergy has to be reduced to prevent pile damage. As a Both DLT and STN are performed after a set up
result, DLT has to be stopped at a stage where full period. For DLT the pile load displacement behavior
capacity has not yet been mobilized. is calculated by a signal matching technique (CAP-
WAPTM,TNOWAVE) in most cases.
With STN the duration of the loading is long
enough to keep the pile is under constant compres- For STN the pile load displacement behavior is
sion and tension stresses will not occur. To prevent determined by a direct method, the Unloading Point
bending stresses the piston of the statnamic device is Method (UPM) and in some cases by signal match-
installed exactly on or near the center of the pile ing.
head cross section. The launching of the reaction
mass, and the resulting push load starts from the 3.3 Economy
center of the pile.
DLT has the advantage that the pile driving ham-
mer used for pile installation can also be used for re-
3 THE APPLICATION OF DLT AND STN ON driving the piles after a set-up period. However,
PRECAST DRIVEN PILES when the pile driving hammer has to be used for
constant production, an additional pile driving ham-
For pre-cast driven piles both DLT and STN are mer or drop hammer has to be mobilized. When the
performed after a setup period after pile installation. mobilization of the full pile capacity is requested,
This allows the soil to recover from driving induced the production hammer might not be sufficient to
disturbances like pore water pressure. In most cases mobilize pile capacity after the set up period and an
the soil will regain strength during the setup period. additional heavier hammer has to be mobilized. For
STN the same economical conditions are applicable
For DLT strain and acceleration transducers are as mentioned in paragraph 2.5 (Ginneken, van G.J.J.,
mounted on the pile shaft near the pile head. The 2000)
load displacement behavior is calculated by signal
matching.
630
Table 1. Preferences for DLT or STN with respect to 4 CONCLUSIONS
economy for driven piles
Driven piles For bored concrete piles, auger piles and caissons
Soil set up I DLT I STN Preferred 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
dium for the preference of statnamic load testing in the
medium to ** ***** STN case of cast in situ piles are:
high
1. Accuracy in load measurement
STN is not dependent on pile material and cross
3.4 Mobilization of capacity section properties
2. No influence of cross sectional variations
Set up phenomena can increase the soil resistance STN results are not influenced by cross sectional
considerably. The pile driving hammer used for pile variations over the pile length
installation might not be able to mobilize the full 3. No tension during compressive testing
pile capacity in such a case. STN long duration loading will keep pile under
Another reason that capacity can not be mobilized constant compressive pressure
with DLT is that the load cannot be increased be- 4. Concentric loading
cause compression or tension stresses becoming too Easy placement of STN loading device in center
high. of the pile
5. Pile and soil response closer to static
To mobilize the pile capacity a STN device will With STN the pile moves as one unit, similar to
be sent to the building site with at least a corre- static load tests. Stress wave phenomena can be ne-
sponding loading capacity. Only when the piles are glected resulting in a simple method of analysis
over-designed will the full bearing capacity not be
mobilized. 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
3.5 Chance of pile damage can be the use of the production rig for testing. A big
advantage for STN is the fact that maximum avail-
For DLT there are some cases with a chance of able energy can be used to mobilize capacity and
pile damage. In the case of low friction and a soft that that testing does not have to be stopped when
toe response tension waves will be generated during tensional stresses become too high like with DLT.
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 5 REFERENCES
the pile will experience damage. In the case of a pile
with a hard toe response, for example pile toe on Abbasi, A.F., Al-Tayyib, 1990. Effect of hot
rock, the compressive stresses at the pile toe can weather on pulse velocity and modulus of elasticity
theoretically be two times higher than the maximum of concrete. Materials and Structures, 1990, 23,
compression stress at the pile head. This is caused ~~334-340
by the superposition of compression stress waves at
the pile toe. So, if during DLT the compression Ginneken, van G.J.J., 2000, Introduction Stat-
stresses at the pile head are higher than half the namic Load Testing in Europe, Case Studies in the
compressive strength of the pile material, collapse of Netherlands. Proceedings of the Sixth International
the pile material at the pile toe will occur. In this Conference on the Application of Stress-wave The-
case piles can only be tested up to half the compres- ory to Piles, Sao Paulo.
sive strength of the pile material, which may not cor-
respond with the capacity of the pile. Holeyman, A.E., 1992. Keynote Lecture: Tech-
nology of Pile Dynamic Testing. Proceedings of the
For STN the pile is kept under constant compres- Fourth International Conference on the Application
sion and tension waves are suppressed. Superposi- of Stress-wave Theory to Piles, The Hague, F.B.J.
tion of compression waves at the pile will not occur. Barends, Editor, A.A. Balkema Publishers, ppl95-
As with SLT piles can be tested near to the compres- 215.
sive strength of the shaft.
Franklin, R.E., King, T.M.J. 1971, Relations be-
tween compressive and indirect-tensile strength of
631
concrete, Road Research Laboratory, RRL Report
LR 412
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
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.
635
bearing capacity with the dynamic increasing energy The upper bound and lower bound theorems are
test (Aoki 1997). used to determine the static failure load from energy
The DIET has been largely used in Brazil as an transformations and equilibrium of forces consid-
alternative to the traditional static load test and the erations in the theoretical soil mechanics limit analy-
constant energy dynamic test (Aoki & Mello 1992, sis (Chen & Liu 1990). By the upper bound theorem,
Beim & Aoki 1996). The static and dynamic results the load is not less than that of the actual failure
compare reasonably well (Aoki 1989c, Niyama & load, determined by equating the external rate of
Aoki 1991). The DIET can also be used to analyze work to the internal rate of energy dissipation in an
time effects on pile behavior “set up” after short time assumed velocity field which satisfies: (a) the veloc-
intervals. 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
2 RESISTANCE - DISPLACEMENT CURVE equilibrium. Under this limit load the system loses
AND THE ULTIMATE PILE RESISTANCE its capability to store more elastic recoverable energy
of deformation and starts to displace as a rigid body.
The goal of a loading test is the determination of the By the lower bound theorem, the stress field must
ultimate resistance or bearing capacity of an isolated satisfy the stress boundary conditions and must no-
foundation pile. The isolated foundation pile is a where violate the materials yield condition.
system constituted by the structural pile element and The most popular dynamic loading test is the im-
the soil layers along the shaft and under the pile toe. pact of a hammer on the cushion - pile - soil system.
This system’s response to the action of an axial In this case the pile bearing capacity can be predicted
load applied to the pile head is the displacement of by the old pile dynamic formulae or by the applica-
this point, which is a complex function of many in- tion of the stress wave theory to the hammer impact
dependent variables. Among them, the most impor- (Whitaker & Bullen 1981, Goble et al. 1975). The
tant are: (a) time history of the applied force; (b) ge- statnamic test has been developed as an alternative
ometry of the pile and the soil layers; (c) physical dynamic test (Middendorp 1993).
properties of the pile and the soils; (d) peculiarities During the hammer impact the time histories of
of pile installation that change material’s properties; deformation and acceleration are recorded at the in-
and (e) elapsed time after the pile installation. strumented section and the velocities and displace-
During the static loading the applied force and the ments are calculated by integrating the measured ac-
vertical movement are measured at each loading celerations. The force history is obtained from the
stage. In this case the applied force is equal to the deformation and the applied kinetic energy is evalu-
mobilized resistance and the response of the pile - ated by integrating the force and the displacement.
soil system to the static loading is the well-known The dynamic resistance or total resistance to the
load - settlement curve. The slow maintained loading pile penetration is evaluated from the analysis of
(SML), the quick maintained loading (QML), the forces and velocities. The correspondent static re-
constant rate of penetration (CRP) and the cyclic sistance is evaluated by using the classical Smith‘s
loading are examples of static test procedures (Fel- model (Smith 1960) or the derived model of CASE
lenius 1980, Whitaker & Cooke 1961). and CAPWAP methods (Goble et al. 1981, Rausche
Nowadays, the definition of the static pile bearing et al. 1985, Goble & Likins 1996). The ultimate re-
capacity or static ultimate resistance, which is the sistance from one single hammer impact is defined
maximum reaction capacity of the pile - soil system, by the Davisson‘s limit load criterion (Davisson
is a matter of discussion in the pile foundation engi- 1972). By this criterion the ultimate resistance is de-
neering practice. Nevertheless, it is recognized that fined by the elastic toe displacement equal to (0.15 +
the shape of the load - settlement curve should be in- B/120) inches, where B is the toe dimension.
cluded in this definition (Reese 1972, Fellenius Finally it can not be forgotten that the soil resis-
1980) and this can be done by the consideration of tance changes with the time and neither the static nor
the deformation energy of the pile - soil system. the dynamic loading can predict the future behavior
For a given curve there are many interpretations, of the pile - soil system, from the measured values at
each one leading to a particular ultimate resistance. the time of the test.
The absolute or relative displacement values and the It can be concluded that the definition of the ulti-
relationship between the elastic and permanent com- mate resistance is not clear in both static and dy-
ponents of the displacement after the unloading or namic loading tests.
the time dependent behavior are some of the criteria
for the establishment of the static bearing capacity
2.1 Static force - displacement curve response.
(Vesic’ 1975, de Beer 1988). The absolute displace-
ment equal to 10% of the pile dimension is a typical The loading is called static when the force (Q) is
definition of the ultimate resistance (Terzaghi 1942). reached in stages of infinitesimal increments (dQ)
and each stage is of infinite time duration. Thus, the
636
frequency of static loading is equal to zero. How- if the zero load condition is reached in the inverse
ever, in practice, there is no static loading once the path described for the static loading. In the actual
increment of force (AQ) and the time duration (At) example of a building construction, the unloading
have finite values. path does not exist while the load-unloading cycle is
Figure 1 presents the static force - displacement observed only for the live load.
curve, resulting from the force increment (AQ) and The potential energy of deformation (V) stored in
displacement increment (Ap) and the consequent in- the system after the static loading can be elastic re-
crement of the deformation energy (AV). 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 -
Figure 1. Static force - displacement curve. displacement curves is discussed thereafter.
The rate of internal energy of deformation corre-
sponding to the force increment (AQ) is the power 2.2 Shape of the static load - settlement curve.
Figure 2 presents the typical static load -settlement
g=AV I At (1) curve in the case where the ultimate resistance is
where 9=
power or rate of the internal energy of de- well defined by the final shape of the curve. This
formation increment, AV = increment of energy de- curve correlates the settlement (p) to the load (Q)
formation, and At = increment of time loading. where it is implicit that the static reaction (R) is al-
The average power of the actual static or dynamic ways equal to the static load (Q).
loading is variable from some watts to various kW. The failure is characterized when the pile head
Under the action of the static load (Q), the dis- displaces from point D to point A under the action of
placement of the pile head (point D) is expressed by the constant limit load (Q,), equal to the ultimate
system resistance (Ru).
P = Pe + Pp (2)
where p = pile head displacement, pe = elastic dis- 0 QU
637
(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
638
When the curve is expressed by Veen's expres- the structural pile material strength. This is ex-
sion (8), the limit complementary energy becomes pressed by the following condition:
constant and its value is expressed by
QSQu<R, (14)
v,, = -Q "
o! where Q = applied force, Q, = ultimate soil resis-
Therefore the physical ultimate resistance can be tance, and R, = structural pile strength.
defined by the limit complementary energy of ex- In some cases, where the structural pile strength is
pression (13). This criteria has the advantage of in- reached before the ultimate soil resistance has been
volving both the curve shape and the ultimate resis- mobilized, the system fails by the weaker structural
tance in the definition of the failure behavior of the link and the following condition prevails:
pile - soil system (Aoki 1997).
Figure 5 presents the case where the OD path of Q u > Re (15)
the curve is similar to that presented in figures 1 to
4. The special feature in this case is that the settle- where Q, = ultimate soil resistance, and Re = struc-
ment increases linearly with increasing loading up to tural pile strength.
the maximum load test value Qmax,without the ulti- This is the case where the pile toe is placed either
mate resistance characterization. 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)
639
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-
where Peu,s = limit elastic pile displacement, and Qu tem. Under the action of this constant load, the pile
= pile toe ultimate soil resistance. head displaces from point D to any point A on the
Considering expression (1 5 ) for materials of vertical line. Observe that, in this loading path, all
perfectly elastic - plastic behavior, the limit elastic the deformations are permanent.
pile head displacement just before the structural fail- The potential energy of deformation (V) stored in
ure of the pile will be: the system at point A is expressed by
/ \
where Peu,p = limit elastic pile displacement, just be- where V = potential energy of deformation, Q, =
fore the structural failure of the pile. ideal ultimate resistance, p = total displacement, plv
The limit elastic energy of deformation, when the = permanent displacement, and Peu,s = limit elastic
ultimate soil resistance is smaller than the structural pile displacement.
pile strength, can be evaluated by The unloading path is line AC. The limit recov-
ered elastic energy of deformation is expressed by
640
penetration under a single blow of the hammer where R,-, = dynamic shaft resistance, n = number of
changes with increasing penetration. However the sections, Rf ,m = maximum local shaft static resis-
expression ’dynamic resistance against penetration’ tance, J, = local Smith’s damping factor, and v(m,t)
has no definite meaning, unless this term is applied = velocity in the shaft section (m) at time (t).
to the final resistance represented by the abscissa The total dynamic resistance (static + dynamic) or
(QJ of the vertical asymptote to the penetration simply dynamic resistance is not mobilized at the
curve (Terzaghi 1943). This observation shows the same time at all pile sections and is computed by the
importance of the shape of the resistance - displace- following expression:
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- where Rt = dynamic resistance, R,, = dynamic toe re-
sion (6). According to Terzaghi, the question is: is sistance, and Rf, = dynamic shaft resistance.
the kinetic energy of the hammer blow enough to From expressions (26), (27) and (28) it is possi-
mobilize the ultimate resistance of pile- soil system? ble to conclude that the corresponding static resis-
Figure 7 presents the idealized dynamic toe re- tance is
sistance - pile toe displacement curve corresponding
to the idealized load - settlement curve for elastic n
perfectly plastic materials of figure (6). R = R, + E R f m
1
E
cd
Besides the simplified rheological behavior, a
fundamental feature of Smith’s model is that the dis-
_1
64 1
At this point the resulting kinetic energy (T) was Despite such difficulties, the reliability of the pile
stored as potential energy of deformation (V) in the bearing capacity prediction can be improved by the
pile - soil system. The potential energy of deforma- dynamic increasing energy testing (DIET), based on
tion (V) and the maximum kinetic energy (T) are Hamilton's principle of energy conservation which
represented by the area OABO. will be discussed thereafter.
The measurements of the maxiqum total (dy-
namic + static) resistance (RJ, the maximum dy-
namic displacement (D) and the maximum kinetic 3 HAMILTON'S PRINCIPLE OF ENERGY
energy (T) can be done, for example, by monitoring CONSERVATION.
the hammer impact with PDATMsystem.
The dynamic loading cycle is unique and it is not The ultimate pile resistance has been predicted by
possible to separate the loading and unloading stages the dynamic pile driving formulas and by the appli-
as in the static loading case. The applied forces and cation of the stress wave theory to model the ham-
the corresponding mobilized resistances vary during mer impact. In both cases the repeatability of the
the impact time and only the maximum values at displacements under impacts of same energy shows
each pile section are of importance. that the dynamic loading is a reliable procedure.
The dynamic unloading path starts soon after the The fundamental problem to evaluate the energy
maximum displacement (D) is reached and ends at lost in the pile driving system was clarified when the
the time (t2) when the final pile penetration is equal ENTHRU concept was introduced in practice (Goble
to the permanent displacement or set (S). This un- et al.1975). From this time, the actual kinetic energy
loading path is represented by curve AC in figure 8. available in the pile - soil system has been currently
The area ABCA is the potential recoverable elastic measured during the pile driving, allowing the
energy of deformation (V,) and the area OACO is measurement of the hammer impact efficiency.
the work done (W,,,) by the non-conservative forces. Apart from this use, not much has been made to pre-
The maximum pile head displacement under dy- dict the static resistance from the direct energy
namic loading measured by either the old cardboard measurement (Broms & Choo 1988).
and pencil system or the double integration of meas- This keynote lecture brings back the old Hamil-
ured acceleration is equal to ton's principle of energy conservation that has not
been yet mentioned in the traditional pile dynamic
D=K+S (30) literature (Aoki 1997). It is a valuable tool to better
understand the energy transformations in the pile -
where D = maximum pile head displacement under soil system, subjected to the action of a dynamic
dynamic loading, K = rebound or elastic displace- loading, in the case of the ultimate resistance analy-
ment, and S = set or permanent displacement. sis.
The elastic displacement component is equal to 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-
where C2 = elastic shortening of the pile, and C3 = tem, including the work done by damping and any
elastic displacement due to deformations of the soil arbitrary external forces at any time interval tl to t2,
layers under the pile toe (pile toe quake). is equal to zero. In variational form, Hamilton's
In figure 8 the maximum displacement (D) is principle of energy conservation is expressed by
equal to segment OB, the elastic displacement (K) is
equal to segment BC and the set (S) is equal to seg- ft': S (T - V ) dt + S,:S ( Wnc ) dt = 0 (32)
ment OC. The set (S) is usually measured by the
pencil and cardboard technique. where T = system total kinetic energy, V = system
Although fundamental, there is not much interest potential energy, including the energy of deforma-
in measuring the actual dynamic resistance - pile tion and the potential energy of any external conser-
head displacement curve from one hammer impact. vative forces, W,, = work done by the non-
The ideal shape indicated in figure 7 for the toe dis- conservative forces acting in the time interval, in-
placement is based on the ideal static response of cluding the damping forces and any arbitrary exter-
figure 6. The actual static curves of figures 2 to 5 are nal loading, and 6 = variation during the time inter-
being simulated by the ideal curve of figure 6. val (t2 - tl).
The CAPWAPB analysis simulates the shape of According to Hamilton's principle, during the dy-
the static load - settlement curve based on the stress namic loading corresponding to one impact of the
wave theory applied to one single hammer impact hammer, it follows that: (a) at the time (t = tl), the
measurements, by zeroing the velocity in figure 8. In kinetic energy applied to the pile - soil system is
this curve the toe ultimate resistance is defined by equal to (T); (b) at the time when the displacement
Davisson's limit load criteria. reaches the maximum value (D), all kinetic energy is
642
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
643
0 D (displacement) S (permanent set) A K (elastic rebound) 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
Figure 9. Dynamic resistance - displacement curve change in the system response involves energy and
work done variations under increasing loading,
plot is in good agreement with the direct measure- which will be analyzed under the light of Hamilton’s
ment with the PDA system, within the measurements principle of energy conservation.
accuracy. If the initial condition of the pile-soil sys- The subsequent search for the ultimate resistance
tem has changed after each increasing loading stage, in the DIET procedure is a similar problem to the
the measured potential energy of deformation would prediction of the failure load in the static loading.
not fit the calculated value from the dynamic resis- Considering that the energy and the work done under
tance - displacement curve. increasing energy rates can be measured, it is natural
The maximum displacement at point A corre- to analyze their variations in order to check the ex-
sponding to the last applied blow of 3.0 m of height istence of a dynamic upper bound of the failure load
was 20.3 mm and the mobilized dynamic resistance similar to the case of the static loading.
was 2 180 kN. At this point the stored potential en- For static loading, the upper bound theorem de-
ergy deformation was equal to 29.9 kJ. termines the limit load by equating the rate of inter-
After unloading, the permanent displacement was nal energy dissipation to the rate of the work done by
9.0 mm. The area ABCA is equal to the recovered the external forces in the system. This theorem states
elastic energy of deformation and the area OACO is that the variation of the internal potential energy is
equal to the work done by the non - conservative equal to the variation of the work done by the exter-
forces including the damping forces. According to nal forces, if the rates of variation are referred to the
the Hamilton’s principle of energy conservation, the same time interval.
following equation can be applied at the end of any The load and the analysis are called “limit” be-
impact of increasing energy: cause it is not possible to apply a load greater than
this limit value. This limit is equal to the ultimate re-
T = V = V, -t W,, (35) sistance of pile-soil system.
where V, = recoverable elastic energy of energy of In the dynamic increasing energy test it is possi-
deformation, and W,, = work done by the non - con- ble to measure the variation of the internal potential
servative forces. energy and the work done at any dynamic loading
Assuming that the unloading path AC is linear, level. Similarly, the limit dynamic resistance can be
the elastic energy of deformation will be determined by the upper bound theorem, as follows:
(a) choose an energy of deformation (VU) and
V,- R,(W2) (36) suppose that it fulfills the limit upper bound theo-
where Rt = dynamic resistance, and K = rebound or rem;
elastic displacement. (b) calculate the variation of the energy of defor-
In that case the work done is mation for the impacts of energy greater than the
value under analysis by the expression
W,CZR~(W~)-T (37)
AV=V-Vu ( 3 8)
where Wn, = work done by non - conservative forces.
According to the measurements the wave takes 16 where AV = energy of deformation variation,
millisecond after the impact to reach point A, re- V = current energy of deformation, and VU = limit
sulting an average test power equal to 1868 kW. energy of deformation under analysis.
644
(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.
645
AW (kJ) 17 AV (kJ) 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 -
Figure 1 1. Dynamic loading: post failure behavior. dynamic displacement curve corresponding to the
dynamic resistance - displacement curve of figure 9.
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 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 kN weight hammer. The sedimentary soil layers are
static resistance - displacement curve from DIET constituted by 12 m of soft clayey silt soft clay layer,
performed one day after the driving and the load - overlying 4.0 m of hard micaceous sandy silt and
settlement curve obtained in the static SML test pe- 12.0m of very hard micaceous clayey silt, in the
formed 45 days later. In this case the concrete pile of Tertiary Itambe Formation, Parana State. The
diameter of 50 cm, 10 cm of wall thickness and 16.2 maximum static load was 2600 and the maximum
m of length, was driven with a 50 kN hammer 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.
648
mum mobilized static resistance in the DIET was 62 EOD + t = 10’ t = 20’
YOof the structural pile strength corresponding to the = 40’ t = 80’ o t = 13days
maximum displacement of 3 YOof the pile diameter.
Figure 18. Correlation (4) DIET‘S and (2) SML test Figure 19. Initial setup in very soft clay layers.
curves on large diameter piles with the same length
The soil was constituted by 5 m thickness soft
Piles I and I1 were subjected only to the dynamic clayey sand layer (N=2) and 9 m of very soft marine
increasing energy test. The set of the points shows sandy silty clay (N=l), overlaying 12 m thickness
that the curve trends are in good agreement with the soft to medium fine clayey sand (1 <N<7).
static results of pile A. The shape of the curves are These sedimentary layers rest on 2 m thickness
typical of piles working by the point. The general very dense, fine to medium silty sand young residual
comparison between the DIET and SML test curves gnaissic soil (25 <N<37).
(figures 14 to 18), shows that the load - settlement In this example the measurements are done just at
curve of the conventional SML static loading and the the end of the driving (EOD) and after 10 minutes,
Smith‘s static resistance - displacement curve of the 20 minutes, 40 minutes, 80 minutes and 13 days af-
dynamic increasing energy test show a reasonable ter the EOD. For comparison 16 points correspond-
agreement for engineering purpose. ing to the last 35 cm of driving were registered.
It can be concluded that the dynamic increasing By taking the referential line corresponding to 15
energy test (DIET), as currently used in Brazilian mm of displacement, it can be seen that the setup ef-
foundation practice, improves the reliability of the fect was 70% in 13 days, but half of this value was
pile bearing capacity prediction. developed in the first 80 minutes.
This measurement confirms the local practice that
it is enough to stop the driving for some minutes to
5.2 Setup measured in increasing energy test get smaller sets due to the soil freezing. In the con-
In coastal sedimentary soft clay formations it is ventional static load test it was not possible to per-
well established that the soil setup takes place very form this kind of research.
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 6 CONCLUSIONS
the pile test assemblage time. An advantage of the
dynamic increasing energy test is that the loading The dynamic increasing energy test (DIET) is a
can be performed at time intervals as short as de- promising procedure to evaluate the behavior of the
sired. pile - soil system. Its potentialities have not been
Figure 19 presents the results of a typical investi- fully explored and more experience is necessary be-
gation of the soil setup in the case of a concrete pile, fore its reliability on the prediction of the pile bear-
of 26.6 m of length, external diameter of 42 cm and ing capacity from measured dynamic pile behavior
8 cm of wall thickness, driven with a free fall ham- can be confirmed.
mer of 42 kN of weight, in the sedimentary marine The authors express their gratitude to SCAC Fun-
coastal of the Baixada Santista soil Formation in Siio daqiies e Estmturas Ltda, by the support given in the
Paulo State. last 18 years of research on the behavior of driven
649
reinforced concrete piles under dynamic loading Proc. Int. Seminar on Application of Stress-Wave Theory
conditions. 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 &
REFERENCES McVay, Proc. Int. Conf on Application of Stress-Wave
Theory to Piles (5): 263-273. Orlando.
Aoki, N. 1989a. Prediction of the Behavior of Vertical Driven Koten, H.; Sandbergen, R. & Starink, T.W. 1988. Model test
Pile under Static and Dynamic Conditions. Drivability of on static and dynamic bearing capacity of piles. Proc. Int.
Piles (2): 55-61; Proc.Disciission Session 14, Int. Con$ Con$ on Application of Stress-Wave Theory to Piles (3):
Soil Mech. and Fozind. Eng. (12), Rio de Janeiro, August 368-376. Ottawa: Canada.
I9S9. Japan : Japanese Soc. Soil Mech. and Found. Eng. Langhaar, H.L. 1962. Energy methods in applied mechai7ics.
Aoki, N. 1989b. A new dynamic load test concept. DrivabiliOJ New York: John Wiley.
of Piles (I): 1-4; Proc.Discussion Session I4, Int. Con$ Middendorp, P. 1993. First experiences with statnamic load
Soil Mech. and Foui7d. Eng. (IZ), Rio de Janeiro, August testing of foundation piles in Europe. In W.F.van Impe
1989. Japan : Japanese Soc. Soil Mech. and Found. Eng. (ed.), Proc. of the 2'"' Int. Geotechnical Seminar on Deep
Aoki, N. 1 9 8 9 ~Discussion
. on Driveability of Piles. Proc. Int. Foundations on Bored and Auger Piles: 265- 272. Rotter-
Con$ Soil Mech. Found. Eng. (I2), Rio de Janeiro, Discus- dam: Balkema.
sion Session I j : ( 5 ) : 2977-2978. Rotterdam: Balkema. Niyama, S. & Aoki, N. 1991.Correlation between dynamic and
Aoki, N. & de Mello, V.F.B. 1992. Dynamic loading test static loading tests in Experimental Research of EPUSP-
curves. In Barends (ed.), Proc. Iift. Coif on Application of Campus. (in Portuguese). Proc. Seininario Engenharia de
Stress- Wave Theory to Piles (4); 525-530. Rotterdam: FundaGGes Especiais - SEFE 11, SE0 Paiilo (1): 285-293.
Balkema. Poskit, T.J. 1992. Keynote lecture: Problems of reconciling
Aoki, N. 1997. Evaluation of the ultimate bearing capacity of stress wave measurements with theory. In B.J. Barends
piles driven with increasing energy (in Portuguese). Doctor (ed.), Proc. lilt. Conf on Application of Stress- Wave The-
of Engineering thesis, School of Engineering of Sao Car- oty to Piles (4): 495-507. Rotterdam: Balkema
l o ~ University
, of Sao Paiilo: 1 10. Randolph M.F. & Deeks A.J. 1992. Keynote lecture: Dynamic
Aoki, N. & Cintra, J.C.A. 1997. New interpretations of the dy- and static soil models for axial response. In B.J. Barends
namic loading curves for driven piles based on the energy (ed.), Proc. Int. Conf on Application of Stress- Wave The-
approach. In M. Almeida (ed.), Recent Developments in oty to Piles (4): 3- 14. Rotterdam: Balkema.
Soil and Pavement Mechanics: 467-472. Rotterdam: Rausche, F.; Goble, G.G. & Likins Jr., G.E. 1985. Dynamic
Balkema. determination of pile capacity. Journal of Geotechnical
Beer, E.E. 1988. Different behavior of bored and driven piles. Engineering ASCE: v.111 n.3: 367-383.
In W.F.van Impe (ed.), Proc. Int. Geotechnical Seminar: Reese, L.C. 1972. Axial capacity. Proc. Specialty Conference
Deep Foundations on Bored and Auger Piles: 47-82. Rot- on Petformance on Earth and Earth-supported Structures.
terdam: Balkema. Purdue Ut7iversity - ASCE, June.v.111: 20 1207.
Beim, J.W. & Aoki, N. 1996. Dynamic load test method with Smith, E.A.L. 1960. Pile driving analysis by the wave equation.
variable energy. In F.C. Townsend & M.Hussein & McVay, Journal of Soil Mech. and Found Division, ASCE, v.86,
Proc. Int. Con$ on Application of Stress-Wave Theory to SM4: 35-61.
Piles (5): 274-28 1. Orlando. Terzaghi, K. 1942. Discussions. Pile-driving formula. Progress
Bernardes, G.P. 1989. Dynamic and static testing of large report of the Committee on the bearing value of pile foun-
model piles in sand. Doctor of Engineering thesis, Depart- dations. Discussion. Proc. American Society of Civil Engi-
ment of Civil Engineering, Norwegian Imtitute of Technol- neers, v.68 n.2: 31 1-323.
ogy, Trondheim. Terzaghi, K. 1943. Theoretical soil mechanics: 137- 143. New
Broms, B.B. & Choo, L.P. 1988. A simple pile driving formula York: John Wiley.
based on stress-wave measurements. Proc. Int. Con$ on Vesic', A.S. 1975. Principles of pile foundation design. Soil
Application of Stress-Wave Theoiy to Piles (3): 59 1-600. Mechanics Series No. 38. Duke University, School of Engi-
Ottawa: Canada. neering.
ChelIis, R.D. 1951. Pile foundations. New York: McGraw-Hill. Vrouwenvelder, A. & Dieterman, H.A. 1992. General report 4:
Chen, W.F. & Liu, X.L. 1990. Limit analysis in soil mechanics. Reliability of predictions. In B.J. Barends (ed.), Proc. Int.
Developments in Geotechnical Engineering (52). Amster- Conf: on Application of Stress-Wave Theory to Piles (4):
dam: Elsevier. 509- 5 17. Rotterdam: Balkema.
Clough, R.W. & J. Penzien 1975. Dynamics ofstrzxtures. New Veen, C. 1953. The bearing capacity of a pile. Proc. Int. Conf
York: McGraw-Hill. on Soil Mech. and Found. Eng. (3). ( 2 ) :84-90. Zurich.
Davisson, M.T. 1972. High Capacity Piles. Proc. of Lectitre Weele, A.F. 1993. Quality assesment foundation piles after in-
Series. Innovations it7 Foundation Construction. Soil Mech. stallation. In W.J.van Impe (ed.), Proc. Int. Geotechnical
and Found. Div. Illinois Section. ASCE and Dept. of Civil Seminar: Deep Foundations on Bored and Auger Piles (2)
Eng. Illinois Inst. of Technology. Chicago. BAP 11: 459-467. Rotterdam: Balkema.
Fellenius, B.T. 1980. The analysis of results from routine pile Whitaker, T. & Bullen, F.R. 198 1. Pile testing and pile-driving
load tests. Ground Engineering v. I3 n. 6, September: 19-3 1. formulae. In F.E. Young (ed.), Piles and Foundations
Fujita, K. & Kusakabe, 0. 1988. On the evaluation of static 12 1-134. London: Thomas Telford.
bearing capacity. Proc. Int. Conf on Application of Stress- Whitaker, T. & Cooke, R.W. 1961. A new approach to pile
Wave Theory to Piles (3): 525-534. Ottawa: Canada. testing. Proc. Int. Con$ on Soil Mech. and Found,. Eng.
Goble, G.G.; Likins Jr., G.E. & Rausche, F. 1975. Bearing ca- (4). (2): 17 1-1 76. Paris: Dunod.
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
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.
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
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
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 methods are highly close with the second knee
matching results are given in Figure 3 and Figyre 4. points almost identical.
Comparing both the capacities and the load- 3. For pile #27 of Bigui Garden, two piles of
movement curves, we can see that: Meidi Shore Garden and two of Huarun Dope Plant,
1. The dynamic result of each test pile is a little which were driven with a diesel hammer, the main
less than its static value, except pile #85 of Gaozan resistance is at or near the pile toe and the static
Market. loading test curves are belong to "slight slope" with-
2. For piles mainly depending on the friction re- out steeply downward sections. The total settlement
sistance, such as: three piles of Gaozan Market and couldn't achieve the failure criterion according to
pile #113 of Yujing Garden, which were installed by Code, therefore the ultimate bearing capacity
silent piling, failure loads were obtained in static couldn't be determined. Sd is a little more than Ss
tests and soil resistance sufficiently mobilized in dy- for this type, the ratios are from 1.1 to 1.3. As to the
namic tests, the capacity errors between the static load-movement curves of two methods, three of five
and dynamic are not more than 11%, settlement ra- piles are analogous. While pile #429 and pile #709
tios(Sd/Ss) =l, the load-movement curves of two are much the same as others at the
653
beginning and separate obviously from the middle which is different from the static loading test. The
part, and that the load-movement curves from dy- pile capacity is gained by slowly loading in static
namic tests are below the static ones. tests. When the dynamic test is being performed,
We have given Some preliminary analyses to impact is applied on pile top for many times with a
above correlation results: heavy hammer. In general, the upper resistance
Transient impact load is applied to mobilize the around the pile shaft is mainly mobilized by former
soil resistance around pile in the dynamic testing, blows. While the lower shaft resistance and end-
654
Figure 4. CAF'WAF' results of Pile #429 of Huarun
bearing capacity act gradually under later blows, at and soil around the pile. Commonly mobilization of
that time, pile has possibly lost small part of top shaft resistance requires 5-1 Omm displacement and
friction resistance. This is the reason why the dy- 10-20mm is needed for sandy soil while for mobili-
namic results are sometimes conservative. zation of toe resistance, even more displacement
Besides, according to transfer mechanism of ver- may be required.
tical load for pile foundations, adequate mobilization The displacement may come to 20-30mm in the
of shaft and toe resistance needs a certain displace- dynamic testing. Such a displacement is enough for
ment of the pile, which is related to the pile diameter friction piles to be mobilized adequately. That is
655
why three piles of Gaozan Market and pile #113 of However, for the short (less than 1Om) or extra
Yujing Garden have good correlation Nevertheless long (more than 40m) piles where the resistance ac-
if lower soil layers have good conditions or the pile cording to the safety requirement of design is hard to
toe is built-in firmly, it will be difficult to have all be mobilized, the static test is recommended. For
resistance mobilized For example, in Huarun site, pipe pile foundations of the first grade building, in
the middle soil layer around the pile is gravel sand order to ensure safety, it is suggested that testing
and the pile toe is built in intensely weathered rock piles should be driven in the design stage and dy-
The penetration of the end driving was only namic testing should be performed during the driv-
1 2-1 5cmllOblows (normally 2 0-5 OcdlOblows) ing and restriking. And then, static loading test
The dynamic testing was performed one week after should be carried out. After the comparison of two
driving Thus, it should be expected that soil resis- testing results, dynamic method may serve as an ac-
tance of piles is hard to be hlly mobilized in this ceptance test for working piles.
site Today in Shunde district, a great number of pipe
It should be explained that although the dynamic pile foundations and square pile foundations are
results are more conservative than the static ones for tested and accepted by the dynamic testing (More
the two piles of Huarun See Table 2, #429 is a than seven thousand prestressed pipe piles have been
400mm dia pile with a design capacity of 1400kN, tested in recent seven years). Considerable testing
and the soil resistance mobilized in dynamic testing cost has been saved from this. As for the advantage
is up to 4734kN, which is more than three times of in enlarging the testing quantity, ensuring safety and
the design value, #709, 300mm dia , design capacity quickening the tempo of construction, it is much
is lOOOkN, the dynamic capacity is 2943kN, near more than economical.
three times more than the design value Both of re- 2. In Shunde, the design capacity of the
sults are more than twice of the design vertical load prestressed pipe piles is always on the low side. For
capacity of pipe pile defined in Technical Specifica- 300mm dia. piles, design capacity is less than
tion for Prestressed Concrete Pipe Pile Foundations lOOOkN, 400mm dia. less than 1500kN and 500mm
(DBJ/T 15-22-98 1998), which is depend on the ma- dia. less than 2000kN. The measured capacities are
terial strength of pile shaft It can be calculated by close to or even more than three times of the design
following formu1a values
(such as the five of nine piles in the cases chosen in
R1, = 0.3 (fee - o p c ) A (1) this paper) The annual amount of the working pipe
piles is up to tens of thousands, it will cause enor-
Where R, = the design vertical load capacity of mous waste and losses. We suggest that the design
the pile shaft; f,, = the compression strength of the vertical load capacity defined in Technical Specifi-
centrihgal concrete pipe pile; oPc= the effective cation for Prestressed Concrete Pipe Pile Founda-
prestressing force of the concrete pile; A = the sec- tions, which is depend on the material strength of
tion area of the pipe pile. pile shaft, should be applied for designing, if the pile
This capacity for 400mm dia. pile is about length and soil conditions are similar to the Huarun
4200kN and for 3OOmm dia. pile is about 2400kN and Meidi site. In this way, both safety and economy
Hence, it is insignificant for design engineers to in- would be achieved.
crease impact force in order to gain more resistance We are gratehl to Prof. De-qing Li for his sug-
in dynamic testing, In addition, the ratios between gestion on this paper.
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 REFERENCES
dynamic testing and static loading test are not much
at all. DBJ/T 15-22-98 1998. Teclinical Specjfication for Prestressed
,
656
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkerna, Rotterdam, ISBN 90 5809 150 3
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.
657
H - hammer drop The driving of steel piles on land is usually
L - pile length performed with the hammer striking directly the top
A - pile section area of the pile. Since no cap was used, and the Weap
E - modulus of elasticity of pile program can not be run without a cap, a solution was
s - pile penetration found by modelling the pile 0,lO m shorter and
SO - elastic pile displacement considering the pile top as the cap. The same
procedure was introduced in the Dinexp program.
The DGI (1978) bulletin presents some The data preparation and analysis of each pile
recommendations concerning mean value of pile followed the sequential procedure:
length to be considered for piles shorter than 20 i) Establishment of soil profile (layering and
times the width of the pile. All the piles of the data estimation of soil parameters).
base are longer than that limit, so the actual pile ii) Estimation of soil resistance distribution during
length was considered in the analysis. It is also driving for the embedded pile length.
mentioned in the DGI (1978) bulletin that in case of iii) Establishment of driving data (hammer, pile).
no direct measurements, q should be assumed as 0.7. iv) Performance of driving analysis.
In the present analysis a q value of 0.7 was adopted, v) Drawing the driving curve.
as it also corresponds to the usual procedure in the vi) Introduction of the blow counts from the pile
application of the Danish formula in Brazil. driving bulletin in the driving curve.
The driving resistance will be evaluated in this vii) Determination of the driving resistance.
paper for the two records usually considered in The estimation of soil resistance distribution
Brazil: (1) the blow counts for the last 50 cm during driving (step ii above) has been performed
penetration and (2) the pile set at final penetration, according to Semple and Gemeinhardt (1981) and
related to the 10 last successive blows. For the Stevens et all (1982).
whole data base, both records (1) and (2) were taken The soil damping at the pile point (Jp) and at the
with a drop hammer falling from one metre height. pile shaft (Js) introduced in the wave equation
Table 1 presents the data base containing the analyses were (Aas, 1993):
complete set of information related to the 19 driving Sand layers: J, = 0.1 s/m
records, including SPT borings extending below pile J, = 0.5 s/m
toe. The columns include, from left to right, the job Clay layers: J, = 0.65 s/m
number, a description of a simplified soil profile, the J, = 0.15 s/m
pile designation and type, the pile section area, the The quake values were assumed as 2.54 mm
pile length, the hammer weight, the blow counts for (Smith 1960).
the last 50 cm penetration (record number (1)) and The driving resistance was obtained from the
the pile penetration for 10 blows (record number driving curve, both for the blow count related to the
(2)). last 50 cm penetration of continuous driving and
also for the 10 successive blows.
The soil resistance during driving calculated by
3 THE WAVE EQUATION ANALYSES the Danish formula and the Wave equation (Weap
and Dinexp program) are indicated for the two
Two wave equation programs were used to estimate records in Table 2.
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 4 STATISTICAL ANALYSES
1988). Details concerning the use of the programs
are discussed elsewhere, GRLWEAP (1988), Costa Two statistical analyses were carried out. A one
et a1 (1 988) and Danziger (1 99 1). population analysis which included the whole data
The data needed to run the program were base and a two population analysis, in which the
available in the pile driving bulletin, except soil data were separated according to soil type mostly
nature and the soil resistance distribution along pile present along the pile length. Unfortunately, the data
shaft and toe. The soil nature was obtained by means base included predominantly clayey soils. Piles
of the SPT boring profile close to the pile. The embedded mainly in sandy soils included only five
estimation of soil resistance during driving is cases, which were not sufficient for a statistical
necessary to make a proper choice of toe resistance analysis.
percentage and side resistance distribution along pile
shaft.
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
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
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 ~~
662
lasting just some minutes. Redriving the pile days REFERENCES
later would probably show much stronger effect.
The analyses are only related to piles in clay. Aas, P M 1993 Personal communication, from
Piles in sand included very few cases for which Norwegian Geotechnical Institute Experience
time effect is not expected to occur. The results are Bussab, W 1988 Analise de F'nridncia e de
indicated in Table 5. RegressLio. SLio Paid0 Atual Editora LTDA
Chellis, R D 1961 Pile Foundatioris. McGraw-Hill
Table 5 . Soil Resistance during Driving. Time effect Book Company New York Second Edition
from end of continuous driving and additional 10 Costa, A M , L F R Moreira, N F F Ebecken,
sucessive blows. A L G A Coutinho, L Landau & J L D Alves
1988 Recent Application of Computer Methods
X Variable Y Variable B Correlation for Drivability Analysis of Offshore Piles in
Coefficient Coefficient Brazil Proceedings of the Iriteriiatronal
Danish (1) Danish (2) 1.20 0.9925 Conference on Computer Modellitig r n Oceari
Weap (1) Weap (2) 1.03 0.9897 Str1~tici.e.665-672 Veneza, Italy
Dinexp (1) Dinexp (2) 0.92 0.9964 Danziger, B R 1991 Dynaniic arm'ysis of pile
drising D Sc Thesis COPPE, Federal
According to what was expected, time effect was not University of Rio de Janeiro Brazil (in
significant within the time lag separating both Portuguese)
records. DGI 1978 Code of Practice for Fouridation
E~gi~eeririgBulletin I T' 32 The Danish
Geotecnical Institute
6 CONCLUSIONS Goble, G G , F Rausche & G E Linkins 1980 The
analysis of pile driving - A state of the art
The paper analyses the driving data from steel piles report Proceedings of the 2nd Iriteniatiorial
using both the Danish driving formula and the wave Corference on the Applicatrori of Stress Wave
equation. Theory or?Piles, pp 131-161 Stockholm
Good correlations were found between soil GRLWEAP 1988 Mamal for Wave Epntiori
resistance during driving determined by the Danish Aiialysrs of Pile Driving Goble Rausche Likins
formula and two different wave equation programs. and Associates, Inc
High correlation coefficients were obtained when Semple, R M & P Gemeinhardt 1981 Stress
the data base was grouped as one population and History Approach to Analysis of Soil Resistance
also when the data was separated according to soil to Pile Driving Proceedrrigs of the 131h Offshore
nature mostly present along pile shaft. Technology Coiiference. Vol 1, 465-48 1
Similar results were found when comparing both Houston, Texas, USA
the Danish formula and the wave equation solution Smith, E A L 1960 Pile Driving Analysis by the
and two distinct wave equation programs. Such Wave equation. ,Joimal of the Sorl Mecharircs
results assure the confidence of the Danish driving arid Foinidatroris Divisioti ASCE Vol 127, part
formula in evaluating the soil resistance during I, 1145-1193
driving for steel piles with characteristics similar to Stevens, R , A Wiltsie & T H Turton 1982
those of the data base. Evaluating pile drivability for hard clay, very
The conclusion above is particularly usehl when dense sand and rock Proceedirigs of the I 4 rh
the pile penetration is deeper than the SPT boring, Ofshore Techriology Coifererice Paper No
which represents a very common situation in the 4205, Vol 1, pp 465-481 Houston, Texas,
engineering practice in Brazil. For those cases, the USA
driving formula may be the only tool available to
verify piling performance.
ACKNOWLEDGEhENTS
663
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
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.
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
666
resistance. The groundwater table lies approximately 3.3 DyFianiic f e s f i q
2.0 m below ground level, and the soil being rela-
Dynamic testing was performed on the three piles
tively well-graded varies between a silty sand and a and the 21 rods at different points in time from the
gravelly sand. The relative density (Dr) is estimated end of driving, up to a maximum of 72, 143 and 2 16
from the CPT to lie between 35-50 % using a rela- days respectively for the piles, and up to a maximum
tionship by Jamiolkowski et al. (1985). The sand is of 69 days for some of the rods. The tests were per-
considered to be normally consolidated with respect formed using a Pile Driving Analyzer (PDA), to-
to its geological history. A mineralogical inspection gether with two strain gauges and two accelerome-
of the sand showed that it mainly consists of hard ters attached close to the top of the pile (or rod).
minerals, such as quartz and feldspars. To determine the static capacity and its approxi-
Table 2. Evaluation of CPT results
mate distribution along the shaft and on the toe,
CAPWAP analyses were performed on the measured
Depth Soil type 9c signals for seven of the rods and all the piles. In ad-
[ml [MPa) dition, the static capacity was also calculated for all
0-2 Clay the tests using the Case Method (RMAX) with
2-1 Silty sand 2-3 damping factors of Jc=0.7 for the piles and Jc= 0.3
1-7 Sand 3-6 for the rods. These values were chosen as they pro-
7-13 Silty sand 3-5 duced capacities that best agreed with the CAPWAP
13-19 Sand 5 -9 results
19-21 Silt? sand 3 -8 The drop height was held constant at 0.5 m for all
21-23 Sand 7-9 the dynamic testing on the rods, while the drop
23- Sand 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
3.2 Irisfallatiori of piles and rods
was measured by simply marking the pilehod be-
Figure 1 shows the location of the piles and rods. tween blows with a sharp object. The accuracy was
The rods (Hl-H21) were installed with the Swedish estimated to be approximately a half-millimetre.
standard equipment (type “Borros”) used for per-
forming dynamic penetration tests (dynamic prob-
ing). The equipment is fairly widespread outside 4 RESULTS AND ANALYSES
Scandinavia and is classified between a DPH and a
DPSH. The weight of the hammer is 63.5 kg and the 4.1 Bearing capacity versiis time
drop height is 50 cm (E,,, = 63.5.g.0.5 = 311 Nm),
with an efficiency of approximately 80%. The rod is Figure 2 shows the relative increase in bearing ca-
a 32 mm diameter solid steel rod, normally equipped pacity (Q.’Ql) versus the logarithm of time, where
with an enlarged 45 mm diameter cone shaped toe to (21 is the capacity one day after end of driving
minimise shaft friction. However, to resemble the (EOD). As can be seen, the capacity increased
behaviour of normal pile driving, the dynamic roughly linear with the logarithm of time. In par-
probing in this study was performed without an en- ticular, the increase in bearing capacity was ob-
larged toe. The rods were driven to a depth of 19.1 served to be on average approximately 40% per log
m below ground. cycle of time for both the piles and the rods. Moreo-
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.
Figure 2. Increase in bearing capacity of piles and rods.
667
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).
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
669
4.4 Comparison between the methods The results from the sensitivity study are pre-
sented in Table 6 in terms of the correlation factor,
In the previous sections the empirical method and
the WEAP-based method were compared against
RZ,and the constants, a and b, that produced the best
curve-fit using Equation (1). The range for the set
measured field data. Since both methods showed
per blow that was used in the regression analyses
very good correlation with measured data, it could
was between roughly 5 and 25 millimetres.
be instructive to compare the methods with one an-
The most interesting result fi-om these analyses
other. In Figure 5 and 6 it is shown that Equation (1)
was that the variation in the constant b was relatively
can be best-fit to the WEN analyses with very good
small, ranging fi-om 0.73 to 0.87, where low values
Correlation.
were linked to small damping factors, and corre-
A sensitivity analysis was performed to see if this
good correlation could be also achieved for a wide spondingly, high values were linked to high damp-
spectrum of soil models. In this case 14 WEAP ing factors. On the other hand, the toe quake was
shown to have very little influence on b. Some of the
analyses were performed for the same pile and
results are also illustrated in Figure 7.
hammer as above. The drop height was chosen to 0.5
m, the pile length was 20 m, and the shaft distribu- The influence of pile length was also examined. It
tion chosen as constant with depth. 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.
670
Nevertheless, it would be of great benefit to be able 5 CONCLUSIONS
to obtain a rough estimate of the relative set-up at a
site by just striking the piles at different points in Both methods, the empirical and the WEAP based,
time. According to Equation (2), a halving of the set, showed that it is possible to evaluate the relative in-
for instance, would indicate a relative increase in ca- crease in bearing capacity with very good accuracy
pacity of 63 % for b = 0.7. If b was instead 0.6, the from pile penetration data.
increase in capacity would be 52%, the difference The particular advantage with the empirical
only being 11 % in the estimated set-up. This ex- method is its simplicity, making it ideal for produc-
ample demonstrates that the variability in the rela- tion control purposes. Furthermore, it was noted that
tive set-up as a function of set is small, if b lies the constant, b, can be expected to range between a
within the mentioned range. relatively narrow span, approximately 0.6-0.8, which
The authors would, however, like to emphasise increases the value of the method even more.
that the two methods presented in this paper are by The WEAP based method, however, is suggested
no means intended to replace pile testing for set-up for use in more detailed evaluations of the relative
prediction. Their main purpose is to provide an esti- set-up, as it is possible to account for time-
mate of the set-up in, for instance, an early testing dependent changes in the soil parameters.
stage, or during production control. The results also showed that there is very good
If we consider the following dynamic formula correlation between the empirical equation and the
(the Energy Approach) derived from the energy WEAP analyses for a variety of possible soil mod-
equation by Paikowsky & Chernauskas (1992), and els. An interesting observation was that the choice of
and D,,,, are measured quantities:
where E,:,,,a, dynamic soil parameters had little influence on the
b-value.
It can be seen that if D,,,,, N" S, i.e. the total quake ACKNOWLEDGEMENTS
(pile and soil) is very small in relation to the set, the
following expression is derived: The research was funded by the Development Fund
of the Swedish Construction Industry, the Swedish
s (5) Council for Building Research, Skanska AB. Fur-
thermore, important contributions were made by the
In fact the expression is a conventional dynamic Commission on Pile Research, Stabilator AB, the
formula (initially presented by Sanders in 1851) in Swedish National Railway Administration and the
its simplest form, thus disregarding the energy Swedish National Highway Administration. This is
losses It also very similar to Equation (1) Never- acknowledged with gratitude.
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 REFERENCES
sets, and further, for short piles with high stiffness
(impedance) This was also shown in the WEAP A\elsson G (1998a). "Long-term set-up of dn\ en piles in non-
sensitivity study However, a value less than 0 9, but cohesn e soils e\ aluated froin dpnaiiuc tests on penetration
probably in most cases less than 0 8, is more realis- rods" Proc I '' hit Co~f on Site Chnracterizntioii, .Itlantn
tic, as near to perfect conditions (a ideal plastic soil Axelsson G (1998b). "Long-term set-up of dn\ en piles in non-
and an infinitely rigid pile) are not very likely to be cohesive soils". Liceritiafe thesis, Kovnl Itistirute of Tech-
nology, Stochliolni
encountered in the field Changes in soil damping Ayelsson G (1998c)."Long-tenn increase in shaft capacih of
over time will also have an influence on h, in as dmen piles in sand" Proc JthInr Coif on Case Histories
much as an increase in soil damping over time will 117 Geotech Eiigng , St Louis, l\fissouri, paper I 25
produce a smaller value, than if the damping was Alelsson G (2000). "Set-up of dmen piles in sand - The ef-
constant fect of constrained dilatancj on the sliafi belia~iourdunng
loading". Proc 1171 Coiif oii Geotechnical R- Geological
It is also possible to reason that for a normal Eiig , GeoEng2000, .\feIboume, -4ustrnlin, (in press)
hammer to pile impedance ratio and for non-extreme Alelsson G & Westin A (2000). "Torque tests on dmen rods
soil conditions, a value of h greater than about 0 6 is for predichon of pile set-up ' , Proc Iiit Colif, Geo-Dewer
assumed reasonable Again under the circumstances ,7000, Denver, Colorado, (IM press)
that the ultimate capacity is mobilised Castclli R J & Hussein M (1 998). "Pile foundation construc-
At the moment, however, site specific determina- tion for the Buckinan Bridge. Jacksoni ille. Florida". Proc
tion of the constant h, by dynamic load testing and 4th Itit Cotif 0 1 1 Cave Ifirtorier 117 GeotecJi Eiig17g Si
I,OUIJ,.I fi s souri',paper 11 01
set measurements, is highly recommended. at least .
Chov F C Jardine R J . Naro! J F and Brucj F (1998) "Ef-
until further calibration of the constant b has been fects of time on the capacih of pipe piles in dense inaniie
performed for a variety of field cases sand". -J GeotecJ?Engug , ASCE 1'01 121(3),p p 251-261
671
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
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.
673
For reasons such as lack of sufficient reaction When settlements are same, sd~~]=&r-]and Sdr=SSr,
load, limit of hydraulic jack, limiting strength of re- the absolute error of maximum loading between both
action beams, the ultimate bearing capacity of the P-S curves is:
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- E of matching P-S curve relating to P-S curve of
tions of the People's Republic of China( 1987). static is:
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.
4 CONCLUSIONS
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.
677
steel has fyk=1500 MPa. In instrumentation CA-50 consecutive stages, with load reductions equal to
steel bars (L=0.60m and +=12.5mm) were used, in 25% of the total load achieved in the test. The soil
which strain-gages were installed, bound in was excavated from the head of the pile up to the
complete bridge. These bars were installed at depths depth of 0.60m7 keeping this area as reference
of 0.60m, 5.Om, 10.0m and 14.0m, inside an iron section to determine its Elasticity Modulus.
sheath left inside the pile while it was being made.
Cement slurry was injected, so elements could be
connected (Albuquerque 1996). 6. DYNAMIC LOAD TEST AND
Driving system used is a free-falling hammer, with a REBOUNDMETER SYSTEM
ram of 1,650 kg of mass, falling from a height of
approximately 0.5m. A basic set of instruments and other commonly used
equipment (force transducers, accelerometer etc)
were used (PDA) for dynamic load tests. These tests
4. IPT-CASE SYSTEM allowed checking integrity and bearing capacity of
the piles (Albuquerque 1996). While the piles were
Developed by IPT ( S ~ OPaul0 State Institute of being driven, measurements of elastic rebound took
Technological Research) this is a computer-based place were performed using the Reboundmeter
data acquisition and analysis system to be used to System developed by IPT, which was in testing
monitor dynamic pile driving. It digitizes, analyses period (Machado 1995)
and records strength and speed signals produced by The Reboundmeter System is a technique to assess
the instruments (acceleration meters, specific soil mobilized resistance as a pile is being driven,
deformation transducers) placed in a section near through the elastic rebound value observed. This
the pile head while it is being driven (Figure 1). technique is an alternative to other methods such as
By analyzing these signals, several important values conventional load test and dynamic driving
are obtained in sifu, such as an estimation of pile monitoring.
bearing by the Case Method, maximum strength and The value of elastic rebound (K) is obtained by
speed, maximum transferred energy, maximum and measuring the displacement caused by the hammer
final displacement, and evaluation of pile structure blow. The displacement curve along time is
integrity. measured for a section near the pile head and, using
The purpose of developing this system is to supply measured values of maximum (M x F) and final (S)
the technical area with a domestic and potentially displacements of the section observed, elastic
less expensive alternative to PDA 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
5. STATIC LOAD TEST (W.
The Reboundmeter developed by IPT records and
The directions established by NBR 12131/91 were processes signals coming from a sensor. These
followed, adopting the slow load (SML). The loads signals describe the displacement of an
were made in consecutive stages, with load increases instrumented section of the pile along time, as a
of 40kN, until the load in which the settlements result of the driving hammer blows.
indicated that the rupture of the connection pile-soil The system has a software that controls data
was reached. The unloading was made in 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
678
Table 5. Results obtained from the static load tcst
Pilc QiZi (kN) Q,,li(W Ql, (W
2 219 43 262
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)
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, Albuquerque, P J R 1996 A~idhse do
were analyzed individually for each pile. The results Contportanwito de Estaca Pre Moldadu de
obtained from the three analysis for pile 1 were very Peqrreiio Dicintetro, Insfruvneizfnda em Solo
close. For piles 2 and 3 the results were a bit farther Residual de Diabasio da Regifio de Canipmas
than for pile 1. DissertagBo de Mestrado 154p Campinas
It was verified that the values of QI, obtained by the Feagri-Unicamp
Rebound Meter are very different for the two piles Aoki, N & Alonso, U R 1989 Correlation
analyzed (1 and 3) among the analyzed formulas Between Different Evaluation Procedures of
(Table 3). Static and Dynamic Load Tests and Rebound
Comparing these results with the ones obtained from In XI1 I C S M F E, Rio de Janeiro
the static load test, it is possible to observe, Proceedrngs...,v 2, p 1 1 15- 1 1 16
therefore, that the Uto et al. (1985) formula offered Aoki, N & Velloso, D A 1975 Urn Metodo
superior values, around 330kN. Regarding the Aproximado para a Estimativa da Capacidade
Chellis (195 1) method, modified by Velloso (1985), de Carga de Estacas In Cotigresso
this has presented values of a lower level, around Parianterrcaiio de Mec&ica dos Solos e
180kN. Nevertheless, in this case the Chellis (195 1)- Eqqenharm de Fimda@s, V, Buenos Aires,
Velloso (1985) formula must be used, because the Arms. .., p 367-376
Uto et al. (1985) formula is valid for piles with point Aoki, N 1986 Co’ontroleIrA’itu da Capacidode de
in compact sand. C’arga de Estacas Pre-Fabricadas L’la Repiqire
Based on the data shown in Table 2, it is possible to Elbstico da Crava@o SZo Paulo ABMS,
observe that, related to point load, skin friction and ABEF e IE-SP
ultimate load, the results obtained from static and Aoki, N 1989 A New Dynamic Load Test Concept
dynamic load tests were very close. However the In XI1 I C S M F E, Discussion Session 14,
same cannot be said about the methods obtained Technical Committee on Pile Driving,
through the rebound, because they present large Drivability of Piles, 1989, Rio de Janeiro
variation. Proceedings..., v i, p 1-4
Aoki, N 1991 Carga Admissive1 de Estacas Atraves
de Ensaios Dinimicos In Seminario de
9. CONCLUSIONS Engenharia de Fundagaes Especiais I1 - SEFE
11, SBo Paulo, Anais ..., v 2, p 269-292, SP
The dynamic load test through the CAPWAPC, Case Aoki, N , Alonso, U R & Trindade, 0 A 1990
and IPT-Case analyses, has shown reliability as a Aplicaqiio de Registrador de Deslocamento
means to determine pile side and point loads. The Diniimico (RDD) na Avaliagiio da Carga
value of ultimate load obtained from the dynamic Mobilizada em Estacas Cravadas In Simposio
load test was placed 10% below the one obtained sobre InstrumentagBo Geotecnica de Campo
from the static load test. SINGE0 90, Rio de Janeiro Anais..., p 45-51
Although the Rebound Meter System did not Chellis, R D 1951 Pile FourzdatronA. Theory-
indicate results comparable to the ones from analysis Desrgti-Pratrce New York McGraw-Hill Book
CAPWAPC, Case and IPT-Case, shows great Company lnc
potential for utilization, mostly due to its simplicity IPT - Instituto de Pesquisa Tecnologica do Estado de
and low cost. However, an improvement in its Siio Paulo SJA 1994 Ensnios de Carregamento
calculation formulas is required. It is important to Dinbniico iia Faculdade de ErzgeiTharra
emphasize that the Uto et al. (1985) formula is only Agricola da Uriicantp, em Estacas Pre
valid for piles in sand. Moldadas de Concreto Relatorio no 32 075
Siio Paulo IPTDCC - AMSFEICC
Machado, J R A 1995 A Avaha@o da Capacidode
REFERENCES de Carga de Estacas, com Base 110 Repique
Elastico Medrdo no Final da Crnvap?o
AssociagBo Brasileira de Normas Tecnicas 1991 DissertagBo de Mestrado 265p SBo Paulo
MB - 3472191 Esfacas - Provas de Carga EPUSP
Estbtica. Rio de Janeiro, AssociagBo Brasileira Monacci, M D 1995 Estudo da Colapsibilrdade de
de Normas Tecnicas, 4p. Rio de Janeiro: unz Solo do ramp0 Expertnwztal da Facirldade
A.B.N.T. de Eigenharia Agricola - Urircamp.
680
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
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.
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
-
~
684
CAPWAP FINAL RESULTS
Total CAPWAP Capacity: 11838.7; along Shaft 8804.4; at Toe 3034.3 I ~ N
. ._.- -. -. -.- ._ ._.- ._.- -. _. _
_ . _
. ._._._. _. = = _ _
- -_- _
- -_= = = _ _ _ _ _ _ _ _ -===___-__________________:===_
1x838 7
-
~
685
Fig.7 The PITWAP final results of pile L74-4 The comparative curves of piles before and after
indicated the break at the location of 5.5m from pile pressure grouting can be seen as follows.
top (shown in first curve). The second curve showed
the improvement of pressure grouting.
686
thickness of 8cm and with thickness of 32cm 4. CONCLUSIONS
strongly weathered granite to the medium weathered
granite. Pressure grouting (shown as the second The pressure grouting to pile foundation is a new
curve) obviously strengthened the bearing stratum. 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
687
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 1503
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.
689
of frequency contents in rapid and dynamic loading Other details of the different indicators are discussed
tests on bored precast piles. The concept of the subsequently in the respective sections that follow,
normalized wave length is introduced and the where attempt is also made to evaluate their
applicable range of the dimensionless frequency a0 suitability in representing the interface between
is also investigated. The extent of the dynamic effect dynamic and rapid loading tests.
in load transfer through shaft resistance is compared
based on the data from static, dynamic and
Statnamic tests. The results may be usefbl in 3 TEST SITES AND STATIC LOAD TRANSFER
defining the basis for disregarding the stress wave
phenomenon in rapid loading tests. 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
2 INDICATORS OF THE INERFACE shown in Figure 2, where the ground condition, steel
pipe pile profile and its instrumentation details are
It is evident that the interface between dynamic and depicted. Circled numbers indicate sections along
rapid loading tests has to be evaluated depending on the pile where strain is measured. The shear wave
how the pile body and the pile soil interface respond velocity VS profile in Figure 2 is estimated from the
to the respective loading methods. Basically, the correlation with standard penetration test N-values
nature of loading method employed is reflected in (Japan Road Association, 1990) given by:
the dynamic characteristics of the resulting load and
velocity time histories recorded during the test. 80N1I3 (Sand)
Prospective indicators considered for investigation vs ={
1 0 0 ~ (clay)
~ 1 ~
are also derived from the time histories, and include:
(a) relative duration of loading T,, (b) average rate of The relation of the static load POapplied at the
loading re, (c) normalized wave length ;1’2L~,(d) pile head and the strain measurement at level 0 is
dimensionless frequency an, and (e) time lag given in Figure 3. Approximately a straight line
between load and movement peaks t,,,.Here, 3, is the relationship may be noted except at very high strain.
wave length and LP is the pile length. Discussions The Young’s modulus of the steel pipe material ES is
are based on actual observations from loading tests. 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.
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.
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.
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:
4.3 ~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 from actual loading test This is because the soil
Statnamic tests. The values of the time lag t,,. shear modulus G degrades at higher strain level
between load and movement are given in Table 1 Conversely, for k,=I. 7kiVm' nini in Equation 7, the
along with other relevant parameters. The difference effective yalue of G works out ,to be about
in between dynamic and Statnamic tests is not I IO5kN ~ i i -compared to 2 756JkN ni- at low strain
evident. That is, t,,. is not found to be usefbl for level (a difference of about 25 times)
distinction between rapid dynamic loading Similarly, for the section 6-1of the test pile in
conditions based on the loading test data considered Figure 2, where L's is again about 1 3 0 d s on the
in this study. Further study would be needed for average, the effective value of the soil ,shear
more conclusive outcome. 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
k,s = -
2'9G and c ,
2nro
= JG'P =Vsp (7)
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
where G is the soil shear modulus, ro is the pile part of the pile, the unit shaft resistance may be
radius and p is mass density of soil. In what follows, obtained from Equation 6. For this, the values of k,y
some investigations are made concerning the based on the value of G estimated from the
and c , ~
difference that may be expected in the behavior of static test can be utilized. The shaft resistance thus
load transfer by shaft resistance. obtained is compared with that of the static loading
case in Figure 7. Large difference in the shaft
5.1 Load transfer in dynamic and static tests resistance behavior between dynamic and static
loading, in terms of the initial slope as well as in
As mentioned above, static as well as dynamic terms of the magnitude, may be noted.
loading tests were conducted on the steel pipe pile In Figure 4 the static shaft resistance tends to
shown in Figure 2. The unit shaft resistance obtained increase up to a movement of more than 20mm. In
from the static test is given in Figure 4, where the contrast, the maximum movement affected in the
bilinear load transfer relation is found to represent dynamic test is only about 8mm. However, there
well the observed behavior. It is also seen that k, is appears to have been a clear permanent movement
about l.7kN/mZ/ntmat the upper part of the ground of the shaft at the end of the dynamic loading cycle
(section E-Cl) where VSis about I30mis in Figure 2. in Figure 13, in spite of the relatively small
Considering G = ~ ( V Sat) ~low strain level? the value maximum movement. The result has very important
of k, works out to be about 42.5kN/m-/mmfrom implication, because the maximum shaft movement
Equation 7, much larger than the value obtained at which a permanent movement may be expected in
694
a dynamic test may not be coincident with the Similarly, Figure 15 compares the trend of load
movement at which the shaft resistance is fully transfer relation in Statnamic and static tests for the
mobilized in static test. This aspect needs to be test on steel pipe pile in Figure 8 (test no. 4 in Table
adequately reflected in defining the soil model in the 1). The two load transfer relations are noted to be
analysis for dynamic test. even more similar to one other when compared to
those in Figure 14.
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 Figure 15. Sl& resistance in Statnamic test on Steel pipe pile
initial slope of the static load transfer function as
well as the Statnamic load transfer fbnction may be It would be interesting to see if the indicators in
defined by Equation 7. Assuming the fraction, noted Table 1 reflect the better resemblance of Statnamic
above for the static loading test depicted in Figure 2, test with the static load transfer trend in Figure 15.
to be applicable here as well, the load transfer Table 1 shows that test no. 1 (Figure 14) has smaller
relations for the Statnamic tests (no. 1 and 4 in Table values of A,,, (25.5) and T,. (12.96) and higher rate of
1) may be obtained as shown in Figures 14 and 15. 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
Figure 11. Shaft resistance in Statnarnic test on PHC pile Attempt has been made in this paper to propose
suitable indicators representing the interface
Figure 14 gives the relative comparison of the between rapid (Statnamic) and dynamic loading
initial slope of load transfer function in static test tests. The range of the values of such indicators are
with that in the Statnamic test on the pile shown in evaluated based on the actual loading test
Figure 7 (test no. 1 in Table 1). It can be noted that observations. To provide certain consistency in the
the difference in magnitude as well as in the trend of comparison between several Statnamic and dynamic
the initial slope between the Statnamic and static tests, low pass filter is used to remove components
tests is much smaller when compared to that in the frequency range higher than 1OOHz. The
between dynamic and static test shown in Figure 13. concept of normalized wave length is introduced and
This is a clear indication that the load transfer in utilized in the investigation. The load transfer
Statnamic test tends to be much closer to the static characteristics in static, rapid and dynamic tests have
trend when compared to the conventional dynamic been compared and investigated. Following
test. conclusions are drawn from the study:
695
i. The relative duration of loading Tr tends to have REFERENCES
a proportional relation with the normalized
wave length A,,, at the peak spectral velocity ASTM 1999. Standard test method for piles wider rapid axial
component. Alrj is referred to as the relative cotnpressive load (Designation: D 0000-99)
Bruno. D. & M. F. Randolf 1999. Dynamic and static load
wave length in this paper. testing of model piles dnven into dense sand. J. Geotecli. 61.
11. Three of the indicators investigated, namely the Geoenv. Eizgrg., ASCE. 125(11): 988-998.
relative duration of loading T., the average rate Danziger. B. R.. A. M. Costa, F. R. Lopes & M. P. Pacheco
of loading re, and the relative wave length A,,j, 1999. Back analysis of offshore pile dnving with an
are found to be well suited to depict the improved soil model. Geotechnique, 49(6): 777-799.
Hirayama, H. 1990. Load-settlement analysis for bored piles
distinction between dynamic and rapid loading using liperbolic transfer functions; Soils & Foundations.
tests. Certain range of values of these VO1.30. No.1: 55-64.
parameters may be defined to serve as the Japan Road Association 1990. Specifications for highway
interface between dynamic and rapid loading bridges, Part IV, Earthquake resistant design (in Japanese).
tests. Karkee, M. B. & H. Kishida 1999. Dynamic, intermediate and
iii. Based on the loading test data analyzed in this static loading tests and application of test results in design
practice. Keynote paper: Proc. 3Ih Int. Col$ Deep
study, the applicable range of the dimensionless Foundation Practice incorporating PILET.4LK '99.
frequency a0 is noted to be less than about 0.2 Singapore, 29-30 July 1999: 13-26.
for rapid loading tests. In comparison, dynamic Karkee, M. B. 1999. Developments in low noise and low
tests seem to involve a0 of up to 3, and may be vibration methods of pile installation in Japan. Proc. 1Ith
more. Asian Regional Col$ Soil Mechanics & Geotechnical
11'. The pile shaft load transfer relation in static
Engineering, Seoul, Korea, Vol. I1
Karkee, M. B., H. Horiguclii & H. Kislida 1999. Limit state
loading test can be represented by bilinear formulation for the vertical resistance of bored PHC
approximation quite satisfactorily, as evidenced nodular piles based on field load test results. Proc. I I "
by the comparison between measured and .4siaii Regional CO@ Soil Mechanics & Geotechnical
computed load movement behavior at different Engineering, Seoul, Korea, 1999: 237-240.
points along the pile length. Karkee. M. B., H. Horiguchi & H. Kishida 1997. Static and
dynamic tests for evaluation of vertical bearing capacity of
1). The load transfer by shaft resistance in case of piles. Deep Foundation Institute, 2?ld Aiinual Metriber's
rapid loading test tends to be relatively closer to Conference, Toronto, Canada. 1997: 199-214.
the static behavior when compared to the case Karkee. M. B., 1. Kojima & Y. Slunoda 1995. Influence of
of a dynamic test. different test conditions on the results of Statnaniic load test
VI. Different extent of resemblance in load transfer at Shonan test site. Proc. I"' Int. Statnaiiiic Seminar,
P'aiicouver, Canada. 1995:137-147.
behavior between rapid and static tests can be
Matsunioto. T. & S. Nisliiniura 1995. Wave propagation
expected in practice, indicating the need to look plienomena in Statnamic test of a steel pipe pile. Proc. 5'"
at the individual cases in detail before deciding Int. Conf. Application of Stress-wave Theor?: to Piles.
to adopt a certain method of analysis for Orlando. Florida. 1996: 1015-1030
estimating the static load bearing behavior. Middendorp. P.. P. Bermingham and B. Kuiper 1993. Statanic
1'11. The extent of resemblance between rapid and testing of foundation pile. Proc. -!Ih Int. Con$ Application of
Stress-wave Theoty to Piles, the Hague, 1993: 585-588.
static loading tests in relation to the load Randolph. M. F. 1990. Analysis of the dynamic pile driving.
transfer knction is adequately reflected in the Developments in soil mechanics IV: .-lnlxmcet/
values of the indicators proposed in this paper. geotechnical analvsis. Elsevier Applied Science Publishers.
Closer resemblance with static loading may be Poulos. H. G. 1998. Pile testing - from the designer's point of
expected in case of rapid loading test with larger view. Preliminary Proc. 2"" Int. Statnamic Seminar. Tokyo.
values of relative duration of loading Tr, and the Schellingerhout, A. & E. Revoorte 1996. Pseudo static pile
load tester. Proc. 5'" Int. Conf. Application of Stress-wave
relative wave length A,,,, and a smaller value of Theon- to Piles, Orlando. Florida. 1996: 103 1-1037.
the average rate of loading re. Smith. E. A. L. 1960. Pile driving analysis by the wave
equation. J. SoilMecli. Found. Div., ..ISCE, 86(1): 35-61.
ACKNOWLEDGEMENT
696
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.
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.
699
C, = &G= pV, =G/VS (3) respectively, and n is the porosity (n = 0.25 for the
mudstone).
Although the shaft model shown in Figure 6 was The soil parameters except for the maximum shaft
used also for the internal shaft resistance, the resistance, f, and the maximum base resistance, q b ,
radiation damping was set to be zero since the were determined by means of Equations (2) through
inward radiation does not occur in the soil plug. (9) with v = ves.The values off and q b alone were
In the base model (Figure 7), the slider represents established by the wave matching. The static load-
the maximum base (toe) resistance, q b , the spring displacement curve was calculated using the
represents the elastic deformation of the soil below identified f and q b , and the drained value of the base
the pile base, kb, the dashpot expresses the damping, spring, kb, and the shaft spring, under static
cb, and the mass is the lumped soil mass, Mb, loading. The ks(stat,c) was estimated according to the
respectively. The soil parameters, kb, c b and Mb are following equations (Randolph 1991):
given as follows for soil plug base and for the annular
pile base (Deeks 1992):
for soil plug base;
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 @
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
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
-8
construction for drivability assessment. A total of 70
Figure 10. Distributions of shaft resistance estimated from piles were driven at abutment A2, piers P2, P3 and Pq
initial driving and re-driving tests, together with static load in April, 1996.
test results. 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
indicates that a large set-up of the shaft resistance distance of 1 m from the pile head for the dynamic
occurred during the rest period of 66 hours. The load tests.
results from the re-driving test are comparable with In the pile construction work in 1995, 28 initial
the static load test results. driving tests and 33 re-driving tests were conducted
The load-displacement curve derived from the within 4 days without a delay of the usual pace of the
wave matching analysis of the re-driving test is fairly pile construction work. The initial driving tests were
comparable with the static load test result (Figure performed on all the piles except the four piles which
11). The result of the wave matching analysis 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
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
Figure 12. An example of wave matching for the re-driving required ultimate capacity. However, for the pile at
test of a pile at abutment A,. 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.
Several piles underwent multiple re-driving tests. In The frequency distributions of the estimated
the pile construction work in 1996, 70 initial driving ultimate static resistance, Qu, of the piles obtained
tests (one test for each pile) and 116 re-driving tests after the elapsed time of 100 minutes in abutment AI,
(several piles underwent multiple re-driving tests) 200 minutes in pier PI, and 1000 minutes in abutment
were conducted within 13 days. 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
4.2 Wave matching results distributions seem to take the form of normal
The procedure of the wave matching analysis distribution.
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
5 SUFFICIENT NUMBER OF DYNAMIC LOAD failure to be 5% for the characteristic value and 0.1%
TESTS for the design value, the value of k is calculated as
1.64 for the characteristic value and 3.0 for the
Let us discuss the sufficient number of load tests to design value.
be conducted at a site to determine the distribution of Using the results for pier P:! and abutment A2
pile capacity. The design value, Qd, is assumed to be shown in Figure 15, the relationships between the
given by 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
in which Qk is the characteristic value and YQ is the number of piles (16 for pier PI and 18 for abutment
partial safety factor. A2). For each test number, n, n data were selected
If the bearing capacity of the piles takes the form from the measured bearing capacities in random
of normal distribution as shown in Figure 16, Q d and fashion. Then, the values of ,UQ and VQ were
Q k are expressed as (Murakami et a1 1988): 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
in which ,qis the average of Q, VQis the coefficient average and the standard deviation of Q d and Q k
of variance and k is a factor that depends on the were calculated for each test number, n.
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
Figure 18. Variations of average and standard deviation of Qd
and Qkwith of number of load tests at Abutment A?.
6 CONCLUSIONS
REFERENCES
The results of the wave propagation analyses of the
dynamic load tests of a total of 102 steel pipe piles Matsumoto, T., Y . Michi & M. Hayashi 1997. Reliability of
for the foundations of Noetsu Bridge No.3 have been dynamic load testing on steel pipe piles in soft rock. PTOC.
14th ICSMFE, Hamburg, v01.2 : 1185-1 188.
presented in this paper. Rational soil models Randolph, M. F. & H.A. Simons 1986. An improved soil
(Randolph & Simons 1986; Deeks 1992) were used model for one-dimensional pile driving analysis. Proc. 3rd
in the wave matching analyses. A wave matching Int. Con$ on Num. Methods in Offshore Piling 1- 17.
704
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
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.
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
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-
Figure 2. Static load tests results. creased with the succession of the blows. An effi-
ciency ranging from 12.2 to 38.6% was observed.
5 . The performing of additional SPT borings just The dynamic load test comprised seven blows with
beside the piles. drop heights ranging from 60 to 170 cm. The maxi-
These investigations and the results of all the mum tension stresses below sensors (TSX) were
available data are presented and discussed in the small for all blows. The maximum compressive
following items. stress (CSX) of 28.5 MPa generated by the seventh
blow caused the pile top material to fail just below
the steel reinforcement.
4 LOW STRAIN TESTS The CFA-1 pile received six blows with drop
heights comprised between 40 and 140 cm. The
Since the penetration of the piles was well con- driving system efficiency increased from 1 1.7 to
trolled, the low strain tests allowed to calculate the 32.1% during the test. The tension stresses were
wave speed, resulting 3550 m/s and 3800 m / s for the slightly higher than i n the CFA-2 pile, but as ex-
CFA-1 and the CFA-2 piles respectively. These pected the successive increase in compressive
wave velocities are somewhat reduced. stresses caused the failure of the pile top immedi-
Although the concrete used in both piles was es- ately below the steel ring. The compressive stresses
sentially the same, the higher wave speed of the that controlled the end of the test (15.0 MPa at the
CFA-2 pile might be associated to a material having gages) were lower than the CFA-2 ones. It can be
a slightly better performance (Kormann et al., 1999). related to bending effects due to eccentric blows or
concrete properties.
Figure 3 shows the records of force and velocity
5 HIGH STRAIN DYNAMIC TESTING multiplied by the impedance for the two piles. For
the CFA-1 pile, a wave speed of 3000 m/s is suit-
In order to perform the dynamic testing the pile tops able. The CFA-2 pile shows a wave speed of
needed to be reinforced. The reinforcement was car- 3250 d s . These findings are in full agreement with
ried out using longitudinal steel bars and a cylindri- the wave speeds measured in the PIT tests, which
cal steel ring mounted at the top of the piles due to the level of deformation usually present
(Fig. la). It is important to note that the reinforce- higher values. The ratio between the wave speed
ment reduced the pile embedded length by approxi- from both tests for the CFA-2 pile to the CFA-1 one
mately one meter which resulted in some loss of was the same regardless the type of test (1.07- 1.OS).
shaft friction near the top of the piles. The dynamic Due to the reduced length of the piles, the width of
load tests were conducted in Dez/99. the pulse impact is considerably higher than 2L/c.
For a proper selection of the hammer to be used Despite that, it will be demonstrated that good re-
for dynamic testing of cast-in-place piles, Hussein et sults can be achieved.
al. (1996) suggest a ram weight about 1.5% of the The traces of the force and the velocity of the
pile static resistance to be verified. A 30 kN free fall CFA-1 pile suggest a high skin friction (Fig. 3a).
hammer was selected for the high strain dynamic The succession of the increasing energy blows
tests. One cushion made by wood plate was used to causes the force signal to be close to the velocity
reduce the impact maximum stress. signal after 4L/c. It can be related to increased pile
Two piezoeletric accelerometers and two strain point displacements. A considerable pile set (9,4
transducers were attached to the piles. A Pile Driv- mm) was measured at the last blow (Table 2). It is
ing Analyzer, PAL model, was employed expected a high mobilization of the soil resistance.
(PDI, 1999). In order to mobilize as much as possi- The first record of the CFA-2 pile (Fig. 3b) shows
ble the soil resistance, the tests employed increasing an increase in the velocity just after 2L/c. It is re-
hammer energies (e.g. Aoki 1989, Bernardes 1989, lated to a hammer interference. The subsequent rec-
Beim & Aoki 1996). The drop height was increased ords do not present such feature.
in 20 cm increments until the compressive stresses
709
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- 7 CONCRETE TESTING
locity record. It can be due to compressive upward
waves that are related to a high skin friction. In order to evaluate the concrete density, a total of
The comparison between the measured sets and 17 samples of the material of the two piles were
rebounds of the two piles also reflects the distinct collected just below the steel reinforcement and at
behaviors observed both in the static load tests and the bottom of the excavation.
in the dynamic records. Despite the higher energies Table 3 presents the mean results of the tests. It is
applied to the CFA-2 pile, its sets are lower than a important to note that the concrete specific wei ht of
half of the CFA-1 ones. The measured rebounds of the piles - in the range of 19.13-19.52 kN/m - is 5
the CFA-2 pile were as high as 5.5 mm, while in the low. It is due to the high porosity of the material.
CFA- 1 pile the maximum rebounds were 2.3 mm. These findings are in agreement with the low wave
All blows presented in Figure 3 were submitted to speeds measured. It can be noted that despite the low
CAPWAP analysis. The results will be presented in concrete density and probably the low resistance the
the item 9. piles supported well the loads both of the static and
the dynamic tests.
711
Table 4. Complementary SPT boring data. were found, in special for the CFA-1 pile. For the
CFA- 1 CFA-2 two piles the ratio between the toe quake and the
Depth Blows130 cni Blows/30 cm Blows130 cm Blows130 cm
below (first and (second and (first and (second and
measured set is approximately constant - a repre-
ground second 15 third 15 cm) second 15 third 15 cm) sentative range is 1.O- 1.7.
level cm) NSI’T cm) NSPT
The Case Method damping JC values of Table 5
(m) does not exhibit a trend and should be disregarded.
4.0 16 27 16 27 The shaft and toe Smith damping values of the two
5.0 21 30128cm 22 30 1 2 5 cni piles showed dependent on the energy level. At the
6.0 28 30123cm 19 30
7.0 30129cni * 17 19
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
changes the conditions of the soil. In effect, even the damping and quake of the Smith model are not con-
pile installation changes the natural state of stresses. stant soil parameters, but are dependent on the en-
However, since the conditions of the ground around ergy level.
the two piles were similar at the time of the com- The use of the radiation damping - that may ac-
plementary SPT boring, a qualitative comparison count for the soil motion associated with the pile
can be made among them. Table 4 presents the re- movcment in drilled shafts - did not lead to im-
sults. proved results. It seems in agreement with the
The following remarks can be made: smooth surface of the piles revealed by the visual in-
1. The complementary SPT borings data are in spection and the stiff characteristics of the soil.
agreement with the SPTT-1 information. Rausche et al. (1996) discusses the applicability
2. The new borings revealed a lens of silty sand of the Multiple Blow Analysis (MBA). They rec-
just above the point of the CFA-2 pile and 60 cm ommend such analysis for variable energy blows and
below the CFA-1 pile. These thin layer is approxi- provide an example for a drilled shaft. However, for
mately 40 cm thick and it probably appears in the the piles here studied such analysis did not seem
soil profile as presented in Figure 1 a. suitable. The MBA analysis keeps constant the end
3. The N s p ~values along the shaft of the CFA-2 bearing. That is not in agreement with a expected in-
pile are not higher than the CFA-1 ones. At the point creased mobilization of the toe resistance due in-
of the CFA-2 pile the N s p ~values are lower than the creasing hammer energies. In this study, the Resid-
CFA-1 ones. ual Stress Analysis (RSA) option also did not
4. On the basis of the known limitations of the improve the results.
SPT testing, the blow count suggests that the higher The mobilized capacities of the two piles are
resistance of the CFA-2 pile could not be explained compatible with the results of the static load tests. At
only by a stronger soil around the shaft or at the pile the end of the dynamic tests, for the CFA-1 and the
point. The hypothesis of the influence of the grout CFA-2 piles respectively, the maximum mobilized
pressure in the performance of the pile cannot be capacities were 10.7% and 48.1% higher than the
disregarded. Also, since the sand lens intercepts the static capacities reported by Korrnann et al. (1999).
CFA-2 pile just above the tip, a distinct response of The high CAPWAP resistance of the CFA-2 pile
this soil to the installation procedure - such as an suggests that the limited settlements of the CFA-2
slight enlargement - might affect the pile end bear- pile in the static tests may have conducted to an un-
ing. 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-
9 CAPWAP ANALISYS: COMPARISONS namic test. These preloading leads to a set of resid-
WITH STATIC LOAD TESTS ual stresses both in the pile and in the soil (e.g Mas-
sad 1992, Maiorano et al. 1996). The comparison
The records of both the CFA-1 and CFA-2 piles between static and dynamic load tests performed on
were submitted to conventional CAPWAP analysis. the same pile should account for the residual
In order to perform a suitable analysis eight pile and stresses. The set of residual stresses does not change
soil segments were employed. A concrete specific the ultimate load. However, the form of static load-
weight of 20.0 kN/m3 was adopted. The main results ing curves can be markedly affected. The residual
are presented in Table 5. loads generated by a cycle of static or dynamic
The skin quake values of the two piles lies in a loading increase the pile head stiffness.
narrow range (0.97- 1.622 mm) and they do not de- In addition, as discussed by Dkcourt (1998), the
pend on the energy level. On the other hand, the toe meaning of the physical failure cannot be applied to
quakes exhibit a trend to increase with the energy bored piles. Cast-in-place piles may develop in-
transferred to the pile. It can be related to a growing creasing point resistance over to large displace-
mobilization of the toe resistance. High toe quakes ments.
712
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, around the piles. The actual diameter of the two
it is expected that the succession of static or dynamic piles, ranging between 33 and 34 cm, is lower than
cycles of loading will lead to increased pile capaci- the nominal diameter (35 a n ) . It shows that there is
ties. For instance, Liu et al. (1996) points out that no reason for the high concrete overconsumption re-
when the dynamic test is performed after the static ported by the installation monitoring system. The
loading, in average a 10% increased pile capacity CFA-I pile shows some small soil intrusions over to
can be found. 4 cm inside the concrete mass. The wave speed of
On the basis of the discussion above, it seems the CFA-1 pile (3000 m/s) was lower than the wave
suitable to compare static and dynamic tests per- speed of the CFA-2 pile (3250 m/s).
formed at the same pile as a succession of cycles of 2. The concrete testing of the CFA-1 and the
loading. In order to perform such analysis, the per- CFA-2 piles showed a reduced specific weight
manent displacement at the end of a cycle of loading (19.52-19.13 kN/m3) and a high porosity. Despite
that, the structural behavior of the piles was satis-
can be added to the next one (e.g. Seitz 1984, Ni- factory both in the static and dynamic load tests.
yama & Aoki 1991). However, it should be noted that the concrete spe-
Figure 5 shows the static load tests and the cific weight is currently employed in wave equation
CAPWAP static simulations superimposed as cyclic analysis to calculate the dynamic elastic modulus of
loading. In order to plot the calculated curves, the the pile. The use of an overestimated concrete spe-
measured sets were employed. The range of 0.928- cific weight would produce an overestimated com-
I . 1 15 is representative for the ratio among the meas- putation of the cast-in-place pile capacity. In addi-
ured sets and the CAPWAP ones. The unloading of tion, the low specific weight suggests that the effects
the blows number 4 to 7 of the CFA-2 pile was ad- in the wave equation results of a possible nonlinear
justed as a straight line between the CAPWAP stress-strain relationship of the pile material should
maximum displacement (DMX) and the measured be investigated.
set. 3. The distinct behavior observed between the
The simulations plotted in Figure 5 shows a re- two piles in the static load tests was confirmed by
markable agreement with the static load tests. The the dynamic load tests. Despite their similarity, the
CAPWAP inferred static behavior follows within pile-soil interaction of the two piles is markedly dis-
reasonable limits the trend of the static load tests. tinct. The CAPWAP analysis suggest that the skin
For the CFA-I pile simulations, it can be ob- friction of the CFA-2 pile is higher than the CFA-1
served that the succession of blows increases the end one. On the basis of the current soil investigation
bearing mobilization and reduces the toe stiffness. performed, the higher capacity of the CFA-2 pile
The gradient of the last straight portion of the load- docs not scem to be explained only by a stronger soil
ing cycles increases with the succession of the around the shaft or at the pile point. The hypothesis
blows. In effect, the ratio RT / Toe Quake decreases of the influence of the grout pressure on the pile-soil
when the drop height increases. It can be related to a interaction cannot be disregarded. Also, since a sand
nonlinear stress-strain behavior of the soil. Despite lens intercepts the CFA-2 pilc just above the tip, a
distinct response of this soil to the installation pro-
the lower displacements, the same trend can be ob-
cedure might affect the pile end bearing.
served in the CFA-2 pile data. 4. For the two piles, the maximum mobilized ca-
pacities in the dynamic tests ( 1 1 14 and 21 82 kN) are
10 CONCLUSIONS in reasonable agreement with the results of the static
load tests (1006 and 1473 kN). The higher
At the current stage of the research here presented, CAPWAP capacities are due to the higher displace-
the following conclusions can be outlined: ments achieved in the dynamic tests. A comparison
1. The visual shaft inspection of the two CFA between static and dynamic load tests performed on
piles revealed a very smooth and regular surface. It the same pile should account for the preloading ef-
is consistent with the stiff and hard clayey soil fects. Cast-in-place piles may develop increasing
713
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
Figure 5. Static load tests and CAPWAP static simula- em argila sobreadensada: comportamento em provas de
tions for the CFA-1 (a) and the CFA-2 (b) piles. carga est6ticas e dinbmicas. Proc. SEFE l V , S3o Paulo,
Brazil.
Likins, G.E. & M.H. Hussein 1995. High strain dynamic test-
point resistance over to large displacements. If the ing of drilled shafts and cast-in-place piles. Deep Foulid.
soil resistance degradation is reduced, it is expected liist., 20th Aiiriuul Menibera Cot$ Meet., Cliurlestoii.
that the succession of static or dynamic cycles of Liu, C., Q. Lin & F. Shi 1996. Determining the bearing capac-
loading will lead to increased pile capacities. ity of large-diameter bored cast-in-situ piles by high strain
dynamic pile-testing. Proc. 5th lrit. Coi$ Applic. of Stress-
5 . In order to compare the results of the static and Wuve Tlieoiy to Piles: 797-804.
dynamic load tests, the permanent displacement at Maiorano, R.M.S., C. Viggiani & M. Randolph 1996. Residual
the end of a cycle of loading should be added to the stress system arising from different methods of pile instal-
next one. As shown in Figure 5 , the CAPWAP static lation. Proc. 5th lilt. Coilf: Applic, of'Stress-Wuve Theory to
simulations shows a remarkable agreement with the Piles: 5 18-528.
trend of the static load tests. Massad, F. 1992. Sobre a interpretaq5,o de provas de carga em
6. The results here presented demonstrate the po- estacas, considerando as cargas residuais na ponta e a re-
tential of the high strain dynamic testing as a valu- vers3o do atrito lateral. Park I: solos relativamente ho-
mogEneos. Solos e Roclzas 15 (2): 103-115.
able tool for the assessment of the load-settlement Niyania, S. & N. Aoki 1991. Correla@o entre provas dc carga
behavior of drilled shafts. dinbmica e esthtica no campo experimental da
EPUSP/ABEF. 2'. Seni. de Erig. de FuridugGes EspeciLiis,
SCO Ptiiilo, B t - ~ i ~ i285-293.
l:
ACKNOWLEDGMENTS PDI 1999. Pile Dt-iviiig Aiidyzer, PAL Model, Users Mariual.
Cleveland: Pile Dynamics.
The authors would like to express their gratitude Rausche, F., B. Richardson & G. Likins 1996. Multiple blow
CAPWAP analysis of pile dynamic records. PI-oc. 5th bit.
to Eng. Alexandre Chwist (Estacas Premold), Eng. Cor!f:Applic. of Stress-Wcive Theor? to Piles: 435-446.
Janies Barossi (Sondar) and Norberto E. Calliari for Salamuni, E. 1998. Tectdriicci dci bucia Sediiiientar de Curitibu
their support in performing the tests here reported. (PI?). P1i.D. Thesis, Instituto de GeociEncias e CiEncias
Exatas - UNESP. Rio Claro. Brazil.
Seidel, J. & F. Rausche 1984. Correlation of static and dynamic
REFERENCES 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-
Aoki, N. 1989. A new dynamic load test concept. Proc. Dis-
cohesive soils. Proc. 211d lrit. Cot$ Applic. of Stress- Wuve
cussion Session 14, TC Pile DiYvirig, X I 1 lilt. Coi$ Soil
Tlieorj to Piles: 201 -209.
Mech. Fouiidcition Erig., Rio de Janeiro: 1-4.
714
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
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.
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
716
Figure 4. ‘Revised’ CAPWAP@ and ISL
compression test results.
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
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.
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.
720
It is interesting to discuss what happened with Conclusions: Dynamic Load Tests are affected by
this correlation. Due to the unreal short period of local geotechnical characteristics as well as by soil
time allowed to finish all the driving works, the En- conditions at the moment of carrying out the meas-
gineer decided to perform the Dynamic Load Test urements. This is the reason why the frst DLT did
immediately after driving- although GEOTECNICA not provide reliable information, because the be-
CIENTEC intended to persuade him not to do so be- haviour of the pile-soil interaction during testing was
cause a less total resistance was going to be obtained different than that at the design stage.
due to the influence of pore pressure. It is the authors’ opinion that in order to perform
The geotechnical information, described on Fig- a better geotechnical design of piles, more efficient
ure 3, shows the stratigraphic profile throughout and foundation works less expensive, it must be to
SPT and CPT tests. Because discrepancies were de- carry out CPTs tests during the soil investigation
tected during the original borings in different areas programme. In this way better Signal Matches will
of the site, Dutch Cone was used. It was also very be obtained between calculated and measured Upper
useful the Pile Driving Prediction (PDP), performed Travelling Waves, because more exact soil models
before the first pile was driven in order to select the can be evaluated.
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 Figure 4. DLT and SLT load-settlement curves.
slow dissipating pore pressure cause a reduction of
the bearing capacity.
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.
721
Figure 6. DLT and SLT load-settlement curves.
4 REFERENCES.
723
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
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.
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 BBRI and CEBTP and results provided to predic-
expressed an interest to make a prediction on that tors.
basis. Also characteristics needed to interpret the
measurements were provided in part in Table 1 of
the companion paper (in particular dimensions and 3 RECIEVED PREDICTIONS
properties of the pile heads extension) and in part in 3.1 Reporting format
the files containing the measurements. Additional
characteristics (wave propagation speed, density, The predictions are reported herein under an anony-
pile impedance, etc.) needed to further analyze the mous format in order not to stigmatize those with
measurements were distributed together with the less accurate predictions. Each prediction is however
measurements. labeled with a code corresponding to the prediction
Interested parties obtained the digital files of the type. Each predictor is thus enabled to position his
events by e-mail, which required the structuring of a own prediction within the cluster of results and en-
vast amount of information, totaling more than 8 couraged publishing his prediction procedure, using
Mbytes of digital records. For each pile type a di- the present paper as a reference.
rectory was established (for example the Prefab di-
rectory), containing subdirectories according for
3.2 Predictions types
each pile of that specific pile type. The ‘pile number’
subdirectory (for example subdirectory pile A7 in Results from 10 predictors had been received on
the directory Prefab) contained a word file: (e.g. November S’’ 1999, the ultimate submittal date.
‘A7info.doc’) and was further subdivided into the Contractors had also predicted the ultimate bearing
following subdirectories: TNO files, ASCII files, capacity of their own piles (they were not asked to
PDA files, and Displacement files. The word supply the load-settlement curves.)
file ‘A7info.doc’ gave supplementary information According to the reference data used to cast those
about the dynamic load test on the pile A7 (the blow predictions, the following labels have been used:
numbers, the drop height, and field notes). It was - “CPT” for predictors using the CPT results.
thus possible for the predictors to reprocess the raw - “PMT” for predictors using the PMT results.
signals using adjusted pile parameters, number of - “LAB” for predictors using the laboratory results.
samples, etc. - “DLT” for predictors using the Dynamic Load Test
results.
- “STN” for predictors using the Statnamic Test re-
2.3 Meas urements nomin a I interpretat ion sults.
The choice of the relevant moduli and sections is The CPT predictors used different methods, in-
often considered as part of the predictors’ art and was cluding ultimate state design as well as load transfer
purposely left open to some degree, as is usually the curves. All the contractors’ predictions were made
case for cast-in place piles. using CPT results and De Beer’s 1974 method. The
It was emphasized that all files had been uni- PMT predictors used the pressiometric approach that
formly acquired using a nominal modulus of ap- provides stress-displacement relationships for the
proximately 40,000 MPa and a nominal wave speed shaft and the base. The LAB predictor used a load-
of 4,000 m/s at the measurement section (i.e. in the transfer functions method based on plasticity indi-
concrete of the cast head). The pile heads were cyl- ces.
inders with a diameter of 0.4 m, except for the Pre- The DLT predictors’ methods included either
fab pile where the current 0.35x0.35m section pre- CAPWAP or SIMBAT: the soil parameters in a
vailed. It was the predictor’s responsibility to assess model are adjusted to get the best match between the
the measurement section modulus adequate for his measured and the predicted signals of a Dynamic
prediction. It was also emphasized that the concrete Load Test. SIMBAT is an empirical method con-
of the tested pile below the added head had a differ- verting the dynamic reaction to a static reaction.
ent modulus and a different section. It was also the The STN predictor used the Unloading Point
predictor’s responsibility to assess the appropriate Method (UPM) to predict the static load test. It was
section and modulus for the shaft. mentioned by the predictor that, due to strain rate
Peripheral information allowing the predictor to sensitivity of clayey soils, a 30% reduction coeff-
perform that important assessment included Table 1 cient had to be applied on the usual UPM method. A
of the companion paper, the digital signal them- hyperbolic approximation of that reduced function
selves (e.g. impedance match or 2L/c check), integ- was then calculated. This is the reason why those
rity tests of the piles. Strength and ultrasonic wave predictions are labeled as “0.7 STN”.
speed measurements on concrete samples cast at the It should be noted that the STN predictor was not
time of installation of the piles and concreting of the provided with the results of the dynamic load tests,
pile heads were made available finally. Low strain and that no TNO-WAVE prediction was submitted.
testing had been performed on all piles both the
728
3.3 Predictions classes The ultimate capacity was considered reached
Predictions had to be made before static pile load when the pile head settlement was equal to 10%
tests were performed in order to qualify as Class A Obase.It should be noted that a maximum Constant
type predictions. If predictions were made after the Rate of Penetration ( C W ) of 0.6 m d m i n was en-
static pile load tests, they qualified as C type predic- forced towards the end of the loading procedure for
tions, according to accepted definitions of predic- all piles (except for piles A1 and A4).
tions classes (Lambe, 1973).
Each prediction for each pile can be classified ac- 4.2 Results
cording to its submittal date relative to the date of Figures 1 and 2 show the various load (Q) - settle-
static loading. Table 1 shows that most of the pre- ment (s) curves for the long and the short piles, re-
dictions are Class C. The only Class A predictors spectively.
were CPT 1 (except for piles A1 & A4) and Con- “SLT” refers to the Static Loading Test. The pre-
tractors Atlas and Fundex. Other Class A predictions dictors’ curves are also identified using the labels
were those of predictor DLT2 for A3 pile and of discussed in Section 2.2. “Contractor” refers to the
predictor PMTl for C2 to C4 piles. 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
4 STATIC LOAD TESTS estimated by the BBRI. It is a “box” corresponding
4.1 Procedurefor static load tests to: - 8AQ<Q<lOAQ
- 9.75 % a b a s e s < 10.25 % OZlbase
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 5 DISCUSSION
to cause bearing failure:
- A pre-load stage of maximum 5% of Q,,, was The load-settlement curves resulting from the static
applied in order to check the measurement load tests show a good proportionality between load
equipment and the centricity of the applied force, and settlement up to 5 mm settlement. Beyond that
- 10 maintained load steps with equal AQ until Q point, the curves deviate from their initial linear
reaches Q,,, trend. After evidencing a peak resistance, the pile
- No intermediate unloading cycles
settles under a slowly decreasing load beyond set-
- Duration of maintained load step of 60 minutes tlements exceeding 10 to 20 mm.
- Load test was performed until a pile head settle- Pile C3 exhibits an unusually high peak, with the
ment 2 15% Obase was reached load increasing very rapidly up to about 2300 kN
- When the pile head settlement has reached a within a 6 to 20 mm settlement range. This was due
value of 25 mm, subsequent load steps can be to a problem in the pressure regulation of the hy-
applied using a smaller increment (AQ/2), in or- draulic jack. After reaching a settlement of 6 mni,
der to refine the pile load-settlement curve as it the pile started to settle at a rate of penetration of ap-
approaches failure, proximately 48 m d m i n . This explains that the peak
- Unloading in 5 steps of 10 min. each, except for capacity of the SLT (about 2300 kN) is an overesti-
final unloading (30 min at least of monitoring). mate of the pile ultimate static bearing capacity. In
The system used to apply the maintained loads on the absence of that regulation problem, the peak
the piles called upon a sophisticated hydraulic regu- would have been reached under a load of approxi-
lation that guaranteed a tolerance of 5 kN. That sys- mately 1500 kN. This test evidences the influence of
tem had just been developed by the BBRI. The 3 the rate of penetration on the assessment of piles ul-
MN reaction was provided by a kentledge consisting timate bearing capacities.
of concrete blocks. Besides load and settlement PMT and LAB predictions are consistently on the
monitoring, extensometers provided longitudinal safe side of the SLT curve. PMTl and PMT2 curves
strains along 5 to 7 shaft segments along the pile are very similar, which tends to show that the pressi-
length. The results provided by those more detailed onietric approach is consistently applied.
measurements are to be reported elsewhere. BBRI predictions (targets “windows”) were good
Such a procedure requires a value for the ultimate except for Piles B3, B4, C1 and C4 where the ulti-
capacity R, of each pile. Those capacities were esti- mate capacity had been overestimated.
mated by the BBRI and the national experts using CPTl predictions were extremely accurate for,
De Beer’s method based on the CPT tests results (De A1 to A4 but overestimated the results for B and C
Beer, 1974). The load increments A Q were actually: piles, probably because of an overestimated installa-
- R,,/8 for Atlas, Fundex, Prefab and Olivier piles. tion coefficient.
- RJ0 for De Waal and Omega piles. Concerning the DLT and STN predictions, it can
be observed that maximum transient pile displace-
ments rarely exceed 15 mm That might explain why
729
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- while TNO provided an FPDS5 system free of
dicting pile behavior under large displacements. charge. Bachir Dali, Paris, using an in-house laser
The predictions (DLT1 to 4 and 0.7 STN) are quite system he developed, acquired displacements on a
good within the service load range of the load- voluntary basis. The authors would like to thank
settlement curves. DLTl and DLT 3 are very close Mr. J. -M. Couvreur, Research Assistant at UCL,
to the SLT ultimate capacities even though they for his management of communication with the
did not predict the pile behavior for settlements predictors and of the information, including prepa-
greater than 20 mm. DLT2 (Capwap method), ration of comparative figures. Finally this exercise
DLT4 (Simbat) and 0.7 STN overestimates of the would not have been possible without countless
ultimate capacities would warrant the following hours accrued amongst the several prediction team
approximate reductions: that bravely took up the challenge. Many thanks to
- DLT2by25%, those heroes; some of who will hopefully identify
- DLT4by50%. themselves through follow-up publications.
- 0.7 STN by 25 % (which means the 30 % re-
duction coefficient initially taken by that pre-
dictor should have been 50 %). 7 REFERENCES
If such reductions were applied to these predic-
tions, they would however not fit as well the initial Lambe T.W. (1973), Predictions in soil engineering,
part of the SLT curves. The reduction of dynamic Giotechnique XXIII, n02, p 149-202.
soil resistance to its static value still needs to be Holeyman, A. et al., Design of axially loaded piles - Bel-
clarifi ed . gian practice in Design of axially loaded piles - European
Practice, edited by De Cock and Legrand, Balkema, Rotter-
dam, 1997, pp.57-82
6 ACKNOWLEDGMENTS Holeyman et a1 (1999a), Prediction Invitation, Belgian
Research Project “Ground Displacement Screwed Piles at
The authors would like to acknowledge the Bel- Sint-Katelijne Waver”, August 20‘” 1999.
gian Ministry of Economic Affairs for its financial Holeyman et a1 (1999b), Additional Reference Document
support of the testing program (Convention CC for Predictors, Belgian Research Project “Ground Displace-
CIF - 562). The ‘De Nayer’ Institute at Sint- ment Screwed Piles at Sint-Katelijne Waver”, August 29“’
Katelijne-Waver made the test site available. Con- 1999.
tractors installed their own piles free of charge. Holeyman et a1 (2000), Preparation of an International
Profound performed the Statnamic tests and Pile Dynamic Testing Prediction Event, Proceedings of the
CEBTP the integrity tests also free of charge. Dr. VIth SWC, Sao Paulo.
Klingmuller, of GSP, Mannheim loaned a PDA
732
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
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 Six different types of ground displacement piles
Building Research Institute (BBRI) has been con- were installed and tested: one prefab and five cast-
ducted in order to establish the performance of dif- in-place screwed types:
ferent types of cast-in-place ground displacement - Atlas pile, installed by Franki Co.
screwed piles. A national advisory committee under - De Waal pile, installed by De Waal CO
chairmanship of the first two authors directed the - Fundex pile, installed by Fundex Co.
program, which included installation and testing of - Olivier pile, installed by Olivier Co.
30 test piles. The program has also been followed by - Omega pile, installed by Socofonda Co.
an international panel of experts selected amongst Figures 1 through 5 illustrate the installation pro-
members of ITC 18 (International Committee on cess of the five ground displacement screwed piles
Pile Foundations, of the International Society of Soil and define the pile toe level relative to the geometry
Mechanics and Geotechnical Engineering). of the augedscrew tip.
A total of 30 loading tests could be performed at a Five piles of each type have been installed on the
site located in Sint-Katelijne-Waver, some 20 km test site to accommodate the following conditions
north of Brussels, according to the following sched- for each pile type:
ule: - One short pile for static load testing
- 6 Statnamic tests and 12 dynamic tests took place - One long pile for static load testing
within the first and second week of August 1999; - One short pile for dynamic load testing
- 12 Static pile tests were performed between Sep- - One long pile for dynamic load testing
tember Znd and October 12* 1999. - One long pile for Statnamic testing
That timing allowed the organization of a Class-A The short and long piles had an approximate
type prediction event (Lambe, 1973), with the view depth of 7.5 m and 1 1.7 m, respectively.
to document the profession’s ability to estimate The different pile types, their testing destination,
those new piles behavior based on standard investi- their nominal shaft and base diameters for geotech-
gation means as well on dynamic testing. nical bearing capacity calculations, and their meas-
An extensive soil investigation was performed as ured pile base depths are listed in Table 1.
part of the research project, including in situ tests A total number of 30 piles were thus installed ac-
(CPT, PMT, SPT, DMT, SASW, ...) and laboratory cording to the pile layout shown on figure 6. The
tests on undisturbed samples. Whereas the present following load tests were to be performed on the
paper provides a description of the pile types and following piles referenced according to their grid
subsurface conditions, its companion (Holeyman et line locations :
al, 2000) focuses on load tests results and their com-
parison.
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
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 )
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.
4 PREDICTION ORGANIZATION
730
measurements warranted the prior permission of the 7 REFERENCES
BBRI. Once this was done, they were sent the com-
plete information, available as laboratory and in situ De Beer, E., Lousberg, E., Wallays, M., Carpentier, R., De
investigation and dynamic load test results (Holey- Jaeger, J., & Paquay, J. 1977. Bearing capacity of displace-
man et al, 1999b). 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
4.2 Prediction Format Lambe T.W. ( 1973), Predictions in soil engineering,
It was requested that the prediction submittal in- Geotechnique XXIII, n02, p 149-202.
clude: Holeyman et a1 (1997) Belgian practice in Design of axially
- A description of the used model(s), with a list of loaded piles - European Practice, edited by De Cock and Le-
governing parameters, grand, Balkema, Rotterdam, 1997, pp.57-82
- The type of soil investigation method on which Holeyman et a1 (1999a), Prediction Invitation, Belgian Re-
the calculations were based, search Project “Ground Displacement Screwed Piles at Sint-
- A detailed calculation methodology, with spe- Katelijne Waver”, August 20‘h 1999.
cific references (data provided, standards, publi- Holeyman et a1 (1 999b), Additional Reference Document
cations, ...) and derivation of governing pa- for Predictors, Belgian Research Project “Ground Displace-
rameters, ment Screwed Piles at Sint-Katelijne Waver”, August 29Ih
- A separation between pile base resistance and 1999.
shaft resistance; Holyeman et a1 (2000), Results of an International Pile Dy-
- A criterion for the ultimate pile bearing capacity namic Testing Prediction Event, Proceedings of the VIth SWC,
The predicted static load-settlement behavior of the Sao Paulo
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.
5 CONCLUSION
6 ACKNOWLEDGMENTS
739
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 1503
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.
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
743
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
744
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 Therefore, the small values for the ultimate capacity
of the dynamic load test signals In the cases of the derived from the re-driving tests for cases 3 and 5
static load tests, Q, and Cl, were separately estimated can be attributed to the fact that the set-up
when axial forces of the pile were measured, phenomena have not been completed at the time
otherwise 0 alone was measured Note that Q is instants of the re-driving tests
defined as the load corresponding to a pile toe In case 8, the ultimate capacity was not obtained in
displacement of 10% of the pile diameter the static load test due to insufficient load capacity of
The load-displacement curve estimated from the the loading device Therefore. case 8 was excluded
wave matching analysis of the re-driving test signals from the comparison of the load-displacement curves
is compared with the load-displacement curve obtained from the dynamic and static load tests
obtained from the static load test in Figures 13 to 19 Figure 12 shows the change with time of the set-
for 7 tests out of the 13 comparative tests For the 7 up ratio of each pile after the end of initial pile
selected cases, set per blow, S, was greater than driving The values of set-up ratio are in a wide
1 2mm In cases 1 and 2, S was almost equal to 0 range from 1 1 to 10, although the set-up ratio is
indicating insufficient driving energy to estimate the lower than 4 if the re-driving test is conducted within
ultimate pile capacity In fact, the ultimate capacity 3 or 4 days after the end of initial pile driving It
derived from the re-driving tests for cases 1 and 2 may be very difficult to predict the set-up ratio
are notably smaller than the ultimate capacity accurately at the current stage of the pile technology
obtained from the static load tests Figures 13 to 19 show the load-displacement
In cases 3, 5 and 9, the time intervals between the curves obtained from the static load test and derived
end of initial pile driving and the re-driving tests from the re-driving test, together with the profiles of
were short (14 hours in cases 3 and 5 , 15 min in the soil layers and the SPT N-values, for piles 4, 6, 7
case 9), not allowing the completion of the set-up and 10 to 13 where re-driving tests were conducted
phenomena of the piles In these cases, the ultimate at time instants sufficiently after the end of initial
capacity derived from the re-driving tests is clearly driving to permit a full recovery of the pile capacity
smaller than those obtained from the static load tests Various commercial computer programs such as
It is interesting to note that the same hammer
CAPWAPC and TNOWAVE, and KWAVE
(DSL45) was used in cases 3, 4 and 5 and an
ultimate capacity of 4 9MN was derived in case 4 (developed by Matsumoto & Takei (1991)) were
745
Figure 14. Load- displacement cunes of case No.6 capacity and the measured bearing capacity is seen in
some cases, the initial portion of the derived load-
used for the wave matching analyses to derive the displacement curve is fairly coincident with the
static load-displacement curves. However, the measured curve, suggesting negligible influence of
empirical soil model developed by Smith (1 960) was the computer program and the operator
used in all of the computer programs. Even with such Table 3 shows a comparison of the calculated yield
conditions, the derived load-displacement curve is in load and the yield load obtained in the static load test
good agreement with the static load test results for of each pile. The yield load is defined as the load
each case. While some discrepancy between the corresponding to the first rapid increase in the pile
derived ultimate bearing displacement on the load-displacement curve. The
746
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%
747
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.
Figure 22. Wave matching results with the use of the rational
soil models.
4 CONCLUSIONS
Shaft resisitance, z (kN/m2) 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
-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.
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
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.
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
752
and 1.2m. All dynamic testing was conducted using
PDA equipment designed and manufactured by PDI
-USA.
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
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.
757
Another result of these International Conferences Geotechnical Engineer, particularly in South
are the Alliance of Pile Testing Laboratory Engi- America, during the Preliminary Project, Project
neers (APTLY), an association recently set up dur- Preparation and Construction stages. In most cases,
ing 1995 by users and manufacturers of these tech- Geotechnical Engineers participated only in the
nology, as well as the periodic Seminars (among “soil investigation” and failed to be included in the
them, Control Quality on Piles: Uses & Abuses) on preparation of a Geotechnical Research Plan neces-
the occasion of the annual meetings of the Deep sary for proper subsoil identification based on the
Foundation Institute (DFI). available data, subsoil characteristics and site spe-
That is how nowadays, the Quality Control cific requirements.
through Non Destructive Tests (QC-NDT), has be- Upon completing of the “soil investigation”-
gun to be included in the Technical Specifications with the scope and vision of a merely formal re-
for the construction of piled foundations in most quirement in most cases- specialists are kept away
Civil Engineering works, by private companies or from the tasks within their specific field of exper-
governments from those countries that have tise, such as the analysis of foundation alternatives,
adopted this technology. The availability of this determination of load capacity of foundations and
technology has resulted in an early detection of prediction of foundation behaviour, geotechnical
defects likely to jeopardize the stability of deep and structural design (which is beyond a straight-
foundations and the superstructure resting on them forward determination of the “allowable stress” as
and reduced the occurrence of accidents likely to is mistakenly believed), selection of the construc-
lead to the loss of both lives and assets. tion method (or alternatives) better suited to local
conditions, design of containment and anchor
structures, excavation procedures, groundwater
3 QUALITY CONTROL ON PILES abatement project, drainage network and relocation
of digging material, soil compaction from mechani-
In the late ’70s and early ’ ~ O S the
, situation of most cal, chemical or thermal causes, soil filling project
industrialised countries and some “developing” (including sanitary fillings and embankments).
countries can be summarised as follows: Last but not least, the necessary Quality Control on
soil movements and/or upgrading, contention and
anchor, underground and open pit excavations, and
3.1 Growth of Construction.
foundations, among other things.
This growth was characterised by a strong flow of Worse still, some of these tasks are sometimes
capital to the construction and/or upgrading of in- carried out by the Client without the support of
frastructure projects, such as roads and bridges, properly equipped laboratories or personnel lacking
thermal, nuclear and hydro power stations and as- the necessary experience or post-graduate training
sociated transformer stations and transmission and, in some cases, personnel alien to the disci-
lines, industrial and petrochemical plants, airports, pline or lacking the necessary skills. It can then be
port facilities and housing developments. In many concluded that Foundation and Contention Project
cases, these projects required the construction of Design and Construction and consequent Safety and
piled foundations with high design loads. This re- Economics are, at least, “unstable”. This is unac-
quirement led to the introduction of new technology ceptable for any structure designed under the Static
for large diameter cast-in place concrete piles under Principles.
bentonite mud. Because the usual design criteria have failed to
catch up with the current building practices in de-
veloped countries and the quality precepts associ-
3.2 Inadequate and/or insufficient Quality Control
ated with safety and economics are slow in being
A long-standing general concern in the construction adopted, deep foundations and contention structures
community has been the safety of foundation continue to be, theoretically, over- (or hyper ?) di-
works, in particular those associated with excava- mensioned with the resulting loss or poor use of,
tions and piles. This concern arose from the very among other things, intellectual resources, materi-
nature of these structures, the imperfect knowledge als, time and labour ... in other words, money.
of the soil profile, wide range of building methods The above circumstances and the many cases of
and customs, general lack of effective quality con- structural and foundation pathologies we were in-
trol methods, lack of tests and measurements on thc volved in our region, paved the way for a study of
behaviour of the soil-structure interaction and a the way in which these problems were approached
certain blend of pseudo-confidence and fatalistic and solved in developed countries. Our study al-
resignation (or hope ?) on “...pile contractors surely lowed us to be in contact with the European and
know what they are doing ...” U.S. experience, scope and limitations of Non-
In addition to the above conditions, there has Destructive Testing technology derived from the
bcen a marked absence of involvement from the
758
Stress Wave Theory and to decide on its introduc- struction” [9] organised by the Institution of Civil
tion to the services provided by our Company. Engineers- ICE- in London on November 4, 1998)
Thus, in 1986, the first tests were performed by clearly show the incidence of the possibility of cost
us using equipment developed and manufactured by optimisation consistent with the Project Develop-
TNO - Building and Construction Research (DeIft, ment Stage and Costs of the sometimes unavoidable
The Netherlands) [7]. and sometimes unforeseeable -Scope Changes de-
pending on the same parameter. In these cases,
3.3 Market Demand: Quality Assurance (QA). while the immediate increased costs are paid by the
Owner or Concession Operator, the final actual
The current trend in most industries is to attain the costs are paid by the Community at large.
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 Figure 1. Changcs in “Reduction Cost Potential” and
cases, on rapid Non-Destructive Tests (NDTs) suf- “Change Order Costs” relative to the time schedule.
ficiently known all over the industrialised world,
numeric analysis methods and simulations, and 3.4 Needs of Industrialized Countries
control procedures developed for rapid accurate di-
agnosis of the detected anomaly. World concern for the energy crisis generated by
The fitness of these tools for their intended pur- geo-political instability in the Middle East attrib-
pose is dependent, prior to the implementation of a uted great priority to the search, exploration and
successful Quality Control System, on a strict per- operation of new oil and gas fields, in particular
formance of quality standards associated with the those located in the North Sea and the Mexican
scope of the Architectural and Engineering Prelimi- Gulf as a result of their geographical proximity to
nary Project, study of the available data, project the natural high demand markets. As an immediate
feasibility study, essential construction service result of the above, there was the urgent need to
availability study, regulatory and environmental build “off-shore” and “near-shore’’ platforms
provisions, basic geotechnical studies (including which, as usual in most cases, are supported by
feasibility of using local materials for fillings and piles.
construction, and final disposal of digging materials This drove to rapid reliable solutions to predict
and waste generated by the works), Executive Proj- and check the bearing capacity of piles under ex-
cct, analysis of alternatives, as well as geotechnical treme performance requirements and environmental
and structural foundation, excavation and conten- conditions calling for the traditional Static Load
tion and anchor design and, if needed, abatement, Tests (SLT) to check the design assumptions
drainage and/or digging materials relocation. adopted. For this reason, a number of International
Successful performance of these tasks will en- Conferences and Seminars have been held since the
sure an efficient construction in terms of quality, ‘60s to present new developments in Dynamic
safety and cost-efficiency. This requires specialists Tests and Numerical Analysis Methods on pile
in each trade to meet the Quality Assurance (QA) driving and load capacity associated with off-shore
standards for the project. The increasingly accepted structures.
principle of having a geotechnical specialist during When evidence was obtained of the actual suit-
each stage of project preparation and implcmenta- ability of certain tests and analysis of results, Sonic
tion has led the most developed countries to stage Integrity Tests (SIT) emerged, almost naturally, as
Technical Meetings, Seminars, and Congresses to the “younger brother” of other prior tests with
analyse this issue in greater detail [8]. similar performance assumptions relative to the
Figure 1 (quoted by R.P.Thompson during the Stress Wave Theory.
Seminar on “The Value of Geotechnics in Con-
759
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. P Equipment FPDS-1.
In the mid 80s, the FPDS-I was developed, based
The application of the above equations to the on a UNIX workstation and a MC68000 (1 6 bit mi-
different situations arising daily in the pile industry- cro-processor chip), with built-in keyboard, disk
coupled to the powerful new "high tech" develop- drive, graphical display, and printedplotter. The
ments- have paved the way for a rational analysis of signal was digitised (12 bit) at an early stage after
the signals obtained from SITS, PDAs and DLT retrieving the signal to ensure that all subsequent
tests, as a result of which it has been able to differ- processing was of the highest possible standard.
entiate the following cases: The system had certain automatic functions, but the
76@
skilled operator could also override, magnify, or processor) and a subsystem with an analogue-to-
clarify a result to facilitate interpretation. The digital (A/D) converter board and a specially de-
FPDS-1 system is still used today by clients for signed board for integrity testing. As with FPDS-2,
SIT. 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
Photograph 2. Equipment FPDS-I . cap, accelerometer sensitivity, and hammer shaft
stiffness produces a load pulse signal perfectly rep-
resentative of the impact amplitude and phase.
9 Equipment FPDS-2. Since force is proportional to velocity at the pile
With FPDS-2, the computer and signal conditioner head, the shape of the force and velocity traces are
were separated, a philosophy continued with subse- identical until a reflection is recorded. By compar-
quent systems. The FPDS-2 system, introduced in ing the force signal to the velocity signal, defects in
1986, was based on a portable IBM-PC/AT com- the first 2 meters of the pile are revealed. The force
patible computer running under MS-DOS, equipped is equivalent to the pile head impedance ’ velocity.
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.
761
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.
762
Software: Built in time and frequency domain 5.1.2 Problems associated to drived piles
software. Impedance profile. PITWAP (calculates
pile shape given a soil model, based on wave equa- 5.1.3. Problems associated to external causes in
tion). the manufacturing process.
763
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).
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 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 quality of the soil in the building site (see Figure
of the disbelief showed by the Engineer, we pro- 12) the planners decided that the foundation should
posed the excavation of the pile and it could be be piled. Franki type piles were adopted.
seen a series of anomalies (deep craks, loss of
coating, etc.) which are indicated in Photograph 10.
+--@--
+ I
Photograph 10. Pile #55. View of the observed damages.
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
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
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