Xu Xiangtao 2007

INVESTIGATION OF THE END BEARING
PERFORMANCE OF DISPLACEMENT PILES IN SAND
by
Xiangtao XU
BEng, MSc
A dissertation submitted for

the degree of Doctor of Philosophy
THE UNIVERSITY OF WESTERN AUSTRALIA
A Leading University
June 2007
The School of Civil and Resource Engineering
DEDICATION
To
Jianle XU, Jiping LIU
& Ting ZHANG
For
Their Love and Support

DECLARATION
I hereby declare that, except where specific reference is made to the work of
others, the contents of this dissertation are original and have not been submitted in
whole or in part for consideration for any other degree of qualification at this, or
any other, university.
……………………..
Xiangtao XU
June 2007
i
ABSTRACT
The axial performance of piles in sand remains an area of great uncertainty in

geotechnical engineering. Over the years, database studies have shown that the
existing method for offshore piles (e.g. API 2000) is unreliable. There is therefore
a clear need for an improved predictive method, which incorporates the state-of-
the-art understanding of the underlying controlling mechanisms. This Thesis is
dedicated to address the factors influencing the end bearing performance of
displacement piles in siliceous sand with a view to proposing and justifying an
improved design formulation.
Firstly, a database of displacement pile load tests in sand with CPT data was
compiled in collaboration with James Schneider (Schneider 2007). It features the
widest database with also the latest available pile load test data (e.g. Euripides,
Ras Tanajib, Drammen etc) in electronic form. Evaluation of the three new CPT-
based methods (Fugro-05, ICP-05 & NGI-05) against this database has revealed a
broadly similar predictive performance despite their end bearing formulations
being remarkably different. This anomaly promoted the author to extend the
database to include additional tests with base capacity measurements to form new
base capacity databases for driven and jacked piles, which resulted in the UWA-
05 method for end bearing of displacement piles in sand. This method accounts
for the pile effective area ratio, differentiates between driven and jacked piles, and
employs a rational qc averaging technique. The UWA-05, together with other
three methods has been included in the commentary of the 22nd edition of API
RP2A (API 2006) for design of offshore driven piles.
Field tests were performed in Shenton Park, Perth to supplement the database
study and, in particular, to examine the effect of the incremental filling ratio (IFR).
10 open-ended and 2 closed-ended piles were tested in compression followed by
tension. The test results provide strong support for the UWA-05 method for base
capacity evaluation employing the CPT qc values and the effective area ratio.
A series of jacked pile tests was carried out on the UWA beam centrifuge, to
further explore the factors affecting pile base response. In total, four uniform and
four layered centrifuge samples were prepared and tested at various stress levels
and relative densities using three separate pile diameters. The resistance ratio
(qb0.1/qc,avg) is found to be independent of the absolute pile diameter, effective
stress and soil relative density. The tests in layered soil enabled quantification of
the reduction in penetration resistance when a pile/cone approaches a weak layer
and revealed the significant influence on base stiffness of underlying soft clay
layers. The stiffness decay curves (G/GIN vs. w/D, where GIN is initial operational
shear stiffness) measured in static load tests were found to vary with ratios of
GIN/qc, while there was a unique relationship between G/GIN and qb/qc.
A detailed parametric study was carried out (using the FE code PLAXIS) by
idealising pile penetration using a spherical cavity expansion analogue in layered
soil. The numerical predictions compare well with the centrifuge results and their
generalization enabled guidelines to be established for end bearing in layered soil.
ii
ACKNOWLEDGEMENTS
I am very grateful for the interest and help of my supervisor, Prof. Barry Lehane,
who has always been encouraging and generous with his time and has constantly
been on hand to provide invaluable guidance and constructive advice when
needed. He has also consistently provided feedback on my writing, which greatly
improved my academic writing skills. Special thanks are also extended to Prof.
Lehane’s wife, Silke, for her hospitality.
The contributions from a number of people are acknowledged. Firstly, James
Schneider, who introduced me the very efficient tool (Macro) in Microsoft Excel
during our collaboration to develop the UWA database, and also provided
stimulating discussions and help in many other aspects; James Ayers and Greg
Morrig who played a role in setting up and conducting the field test in Shenton
Park; Dr. Christophe Gaudin, centrifuge manager, who always found me a testing
slot in the busy schedule of the UWA centrifuges. I am also grateful for the
invaluable discussions I had with Prof. Tatsunori Matsumoto, Prof. Mark
Randolph, Dr. David White, Dr. Kenneth Gavin and Dr. Fiona Chow. Special
thanks go to Dr. Susan Gourvenec for giving me the opportunity to play a role in
ISFOG and Engineering Camp, which I enjoyed a lot. I am also indebted to Prof.
Hanlong Liu, director of Geotechnical Institute of Hohai University for
introducing me to the research topic on PCC pile. Moreover, I am very grateful
for help from my friend, Jitse Pruiksma in GeoDelft, for his quick response to any
of my questions regarding PLAXIS and MATLAB. Technical support from the
PLAXIS group in the Netherlands is also highly appreciated.
I would like to thank people at the electronic and mechanical workshops for their
respective parts in the design and construction of the apparatus, Gary Davies,
Frank Tan, David Jones, John Breen, Turan Brown, Shane de Catania and Philip
Hortin. Also assistance from Bart Thompson and Don Herley with the centrifuge
testing is much appreciated.
I also would like to thank all the administrative staff in the School of Civil and
Resource Engineering and COFS for their friendship and kind help. In particular,
Dr. Wenge Liu, my hometown fellow, who has always offered kind IT help
whenever it was needed. I would like to acknowledge all my colleagues, past and
present, and academic visitors to the department, some of whom became great
friends. I am especially grateful to Dr. Qin Lu and Dr. Jianguo Zhang, who always
welcomed me in their happy family and made me feel at home.
On a serious note, the financial support (IPRS, UPAIS and ADHOC) I received
throughout my time at the University is gratefully acknowledged.
Finally, my grandparents, parents and sisters, thank you for your love and support
throughout the years. Last but not least, I would like to say ‘thank you’ to my
husband, Ting Zhang, who has always been encouraging and supportive with love
and great passion for life.
iii
CONTENTS
Declaration i
Abstract ii
Acknowledgements iii
Contents iv
Notation viii
Abbreviation xi
CHAPTER 1 INTRODUCTION ........................................................................1-1
1.1 Background...........................................................................................1-1
1.2 Research objectives ..............................................................................1-3
1.3 Organisation of THESIS.......................................................................1-4
CHAPTER 2 LITERATURE REVIEW .............................................................2-1
2.1 Introduction ..........................................................................................2-1

2.2 Pile End Bearing Resistance in Sands ..................................................2-2
2.2.1 Bearing capacity theory....................................................................2-2
2.2.2 Cavity expansion method .................................................................2-4
2.2.3 Correlations with CPT data ..............................................................2-6
2.3 Factors Affecting Pile End Bearing Resistance..................................2-10
2.3.1 Method of installation ....................................................................2-10
2.3.2 Surface scale effect.........................................................................2-12
2.3.3 End condition (open vs. closed) .....................................................2-14
2.3.4 Residual stress ................................................................................2-17
2.3.5 Partial mobilisation.........................................................................2-19
2.3.6 Scale effect in layered soil..............................................................2-20
2.3.7 Assessment of pile base settlement ................................................2-21
2.4 Previous Research for qb in Layered Soil...........................................2-24
2.4.1 Meyerhof approaches (1976-83) ....................................................2-25
2.4.2 Vreugdenhil et al. (1994)...............................................................2-25
2.4.3 van den Berg & Huetink (1996) .....................................................2-27
2.4.4 Ahmadi & Robertson (2005)..........................................................2-28
2.5 Summary.............................................................................................2-29
iv
CHAPTER 3 CPT-BASED METHODS FOR END BEARING OF PILES IN
SILICEOUS SANDS.................................................................... 3-1
3.1 Introduction .......................................................................................... 3-1

3.2 CPT-Based Design Methods for Driven Piles ..................................... 3-2
3.3 Overview of the UWA Driven Pile Database ...................................... 3-7
3.4 Evaluation of Fugro-05, ICP-05 & NGI-05 Methods ........................ 3-12
3.4.1 Using full UWA database .............................................................. 3-12
3.4.2 Using UWA base capacity database .............................................. 3-16
3.5 qc Averaging Technique..................................................................... 3-20
3.6 The UWA-05 Method for Driven Piles.............................................. 3-25
3.7 Jacked Pile Database .......................................................................... 3-30
CHAPTER 4 FIELD TESTS AT SHENTON PARK ........................................ 4-1
4.1 Introduction .......................................................................................... 4-1

4.2 Soil Conditions..................................................................................... 4-1
4.2.1 Shenton Park test site ....................................................................... 4-1
4.2.2 CPT results....................................................................................... 4-2
4.2.3 Seasonal effect ................................................................................. 4-5
4.3 Test Set-up and Procedures.................................................................. 4-6
4.3.1 Test programme ............................................................................... 4-6
4.3.2 Pile installation................................................................................. 4-8
4.3.3 Static load tests ................................................................................ 4-9
4.4 Analysis of Test Results..................................................................... 4-11
4.4.1 Driving records .............................................................................. 4-11
4.4.2 Static load tests results ................................................................... 4-16
4.4.3 Performance of Fugro-05, ICP-05, NGI-05 & UWA-05............... 4-25
4.5 Analysis of Base Stiffness.................................................................. 4-30
CHAPTER 5 CENTRIFUGE TEST APPARATUS AND PROCEDURE........ 5-1
5.1 Introduction .......................................................................................... 5-1

5.2 Centrifuge Modelling ........................................................................... 5-1
5.3 Experimental Apparatus....................................................................... 5-3
5.3.1 Geotechnical centrifuge ................................................................... 5-3
5.3.2 Actuator stiffness test....................................................................... 5-5
5.3.3 Model piles and load cells................................................................ 5-7
5.3.4 Pile cap and guiding plate .............................................................. 5-10
5.4 Testing Programme ............................................................................ 5-12
v
5.5 Material Properties .............................................................................5-15
5.5.1 Superfine silica sand.......................................................................5-15
5.5.2 Kaolin clay .....................................................................................5-15
5.6 Sample Preparation.............................................................................5-16
5.6.1 Sand hopper....................................................................................5-16
5.6.2 Clay mixer and consolidation press................................................5-17
5.6.3 Samples A to G...............................................................................5-19
5.6.4 Sample H ........................................................................................5-19
5.7 Test Details .........................................................................................5-20
5.7.1 Test layout.....................................................................................5-20
5.7.2 Test procedures...............................................................................5-22
CHAPTER 6 CENTRIFUGE TEST RESULTS.................................................6-1
6.1 Introduction ..........................................................................................6-1

6.2 Samples A to G.....................................................................................6-1
6.2.1 Pile Installation.................................................................................6-1
6.2.1.1 Base resistance..........................................................................6-1
6.2.1.2 Correlation with Dr ...................................................................6-5
6.2.1.3 Modelling of models.................................................................6-8
6.2.2 Static Load Tests ..............................................................................6-8
6.2.2.1 Resistance mobilisation curves.................................................6-8
6.2.2.2 Residual base resistance qb,residual ............................................6-13
6.2.2.3 Factors affecting qb0.1/qc .........................................................6-16
6.3 Sample H ............................................................................................6-19
6.3.1 Pile installation ...............................................................................6-19
6.3.2 Static load tests...............................................................................6-22
6.3.3 Post sample excavation ..................................................................6-24
CHAPTER 7 ANALYSIS OF qb IN LAYERED SOIL ....................................7-1
7.1 Introduction ..........................................................................................7-1

7.2 Spherical Cavity Expansion in PLAXIS ..............................................7-2
7.2.1 Use of Plaxis.....................................................................................7-2
7.2.2 Soil models .......................................................................................7-2
7.2.3 Mesh set-up ......................................................................................7-4
7.2.4 Analysis procedure ...........................................................................7-7
7.3 Numerical Model Verification..............................................................7-8
7.3.1 Closed-form solutions to limit pressure plimit ...................................7-8
7.3.2 Comparisons with numerical results ..............................................7-10
vi
7.4 Analysis in Two-Layer Soil Profile ................................................... 7-12
7.4.1 Methodology .................................................................................. 7-12
7.4.2 Analysis program ........................................................................... 7-14
7.4.3 Pressure expansion curves ............................................................. 7-16
7.4.4 Interpretation of results .................................................................. 7-18
7.4.5 Curve fitting ................................................................................... 7-20
7.4.6 Zones of influence.......................................................................... 7-23
7.5 Analysis in Three-Layer Soil Profile ................................................. 7-26
7.5.1 Introduction.................................................................................... 7-26
7.5.2 Weak/Strong/Weak ........................................................................ 7-27
7.5.3 Strong/Weak/Strong....................................................................... 7-29
CHAPTER 8 ANALYSIS AND DISCUSSION ................................................ 8-1
8.1 Introduction .......................................................................................... 8-1

8.2 Assessment of Design Methods for qb0.1 .............................................. 8-1
8.2.1 Driven piles...................................................................................... 8-1
8.2.2 Jacked piles ...................................................................................... 8-5
8.2.3 Implication for design ...................................................................... 8-7
8.3 Evaluation of Pile Base Stiffness, G .................................................... 8-8
8.3.1 Samples A to G ................................................................................ 8-8
8.3.2 Sample H........................................................................................ 8-18
8.4 Base Resistance Response in Layered Soil........................................ 8-19
8.4.1 Analysis procedure in a two-layer soil profile............................... 8-19
8.4.2 Variation of qc/qc,S with H/D ......................................................... 8-20
8.4.3 Variation of GIN with H/D.............................................................. 8-24
8.4.4 Comments on qc averaging technique............................................ 8-27
CHAPTER 9 CONCLUSIONS .......................................................................... 9-1
9.1 Introduction .......................................................................................... 9-1

9.2 Findings of Research............................................................................ 9-1
9.3 Recommendations for Future Work..................................................... 9-5
APPENDIX A………………………………………………………………. A-1
APPENDIX B………………………………………………………………. B-1
APPENDIX C………………………………………………………………. C-1
APPENDIX D………………………………………………………………. D-1
REFERENCE………………………………………………………………. R-1
vii
NOTATIONS
Roman
a Current radius of the spherical cavity
a0 Initial radius of the spherical cavity
D Pile outer diameter
Di Pile inner diameter
D10 Particle size where 10% of the particles are smaller
E Young’s modulus of the soil
Ep Young’s modulus of an equivalent solid pile
E50ref Triaxial loading stiffness at pref
Eoedref Oedometer loading stiffness at pref
Eurref Triaxial unloading stiffness at pref
einitial Initial void ratio
emax Maximum void ratio
emin Minimum void ratio
Gs Particle specific gravity of the soil
G Shear modulus of the soil
G0 Small strain in-situ stiffness
GIN Initial operational shear stiffness
h Distance above pile tip level
H Distance to the interface of the strong/weak soil layers
hcone Height of the sand cone formed beneath pile tip
Ir Rigidity index = G/p'0
k Pile base secant gradient =Δqb/Δ(w/D)

K0NC Earth pressure coefficient for normally consolidated sand
viii
K0OC Earth pressure coefficient for overconsolidated sand
L Embedment length of a pile
Lplug Length of soil column (or plug) within an open-ended pile
m Stress dependent stiffness according to a power law
Ng Inertial acceleration field
Nq Bearing capacity factor
plimit Cavity expansion limiting pressure
pref Reference pressure
Q Normalised base resistance
q' Principal stress difference
qb Pile base resistance
qb0.02 Pile base resistance at 2% of pile tip displacement
qb0.1 Pile base resistance at 10% of pile tip displacement
qb,2LS Pile base resistance evaluated in a two-layer soil
qb,3LS Pile base resistance evaluated in a three-layer soil
qc CPT tip resistance
qc,avg Averaged qc values at pile tip level
qc,tip Cone tip resistance at pile tip level
qc,S or qb,S Unaffected resistance in the strong soil only
qc,W or qb,W Unaffected resistance in the weak soil only
qcNC Value of qc for normally consolidated sand
qcOC Value of qc for overconsolidated sand
QBase Base capacity
QShaft Shaft capacity
QTotal Total capacity
r Radius from the centre of the centrifuge to any point in soil model
Re Effective radius
Rcla Centreline averaged surface roughness
R Pile outer radius (=D/2)
Ri Pile inner radius (=Di/2)
ix
su-PP Undrained shear strength measured using a Pocket Penetrometer
t Pile wall thickness
TS Thickness of the strong soil layer
TW Thickness of the weak soil layer
VS In-situ shear wave velocity
w Pile tip settlement
WH Weight of the pile hammer
Wpile Weight of the pile
z Depth of penetration
Z Normalised depth of penetration, = z/D
ZI Total zone of influence = ZS + ZW
ZS Zone of influence in the strong soil
ZW Zone of influence in the weak soil
Greek
α Stiffness ratio =E/qc
ε1 Axial strain
φ Friction angle
γ Specific unit weight of the soil
ρ Density
σ'1 Major principal stress
σ'3 Minor principal stress
σ'v0 Effective vertical stress
σv0 Total vertical stress
ω Constant angular rotational speed of the centrifuge
x
ABBREVIATION
API American Petroleum Institute

CPT Cone Penetration Test
COV Coefficient Of Variation
EURIPIDES EURopean Initiative on PIles in DEnse Sands
FEM Finite Element Method
FFR Final Filling Ratio
ICP Imperial College Pile
IFR Incremental Filling Ratio
LDT Linear Displacement Transducer
NGI Norwegian Geotechnical Institute
OCR OverConsolidation Ratio
PLR Plug Length Ratio
PP Pocket Penetrometer
SPM Strain Path Method
xi
CHAPTER 1 INTRODUCTION
1.1 BACKGROUND
Piles have been used for foundation purposes since prehistoric times. Their
behaviour, however, is far from completely clear and a substantial volume of
research is being carried out on the subject (Mandolini et al. 2005). The field is also
continually evolving, with development in technologies, methods of analysis and in
design approaches, which make the economics of deep foundations ever more
attractive (Salgado 1995, White 2002). However, significant uncertainty remains in
our ability to estimate the axial capacity of individual piles (Randolph 2003). For
onshore piles, it is possible to adapt and fine-tune our design based on the results of
static pile load test, while this is not an option for offshore piles due to the
prohibitive testing costs.
The American Petroleum Institute (API) recommended practice (RP2A) for fixed
offshore platforms is the most frequently used design code for offshore piles
worldwide. First published in 1969, the code is continually updated to reflect some of
the developments in piling engineering. However, for piles in sand, it has only
undergone some minor adjustments with no significant changes to the original 1969
design equations or input parameters (Schneider 2006). Over the years, its predictive
performance has been examined based on database studies of onshore pile load tests
(Sulaiman & Coyle 1971, Dennis & Olson 1983, Briaud & Tucker 1988, Kraft 1990,
Toolan et al. 1990, Chow 1997, Gavin 1998, NGI 2001, Fugro 2003, Lehane et al.
2005a). It has revealed that in general, the current API RP2A (2000) has significant
bias against both soil consistency and pile geometry. It tends to (i) under-predict the
capacity of short piles in dense sand, (ii) over-predict the capacity of long piles in
loose sand, and (iii) over-estimate the ratio of the tension to compression shaft
capacity.
1-1
Chapter 1 Introduction
The CPT qc value is now commonly used directly in design methods for onshore
piles (e.g. De Cock et al. 2003) and there has been growing general support for
inclusion of such a method in the API recommendations for offshore driven piles in
sand. Therefore, in late 2004, the API sub-committee on piling, chaired by Mr. Harry
Kolk (Fugro Engineers B.V.), requested the University of Western Australia (UWA)
to evaluate three CPT-based methods: Fugro-05 (Kolk et al. 2005a), ICP-05 (Jardine
et al. 2005) and NGI-05 (Clausen et al. 2005), and to assess their predictive
performance against a database including new large scale pipe pile tests at Euripides,
Jamuna Bridge, Ras Tanajib and Drammen. The results of the UWA evaluation
exercise are reported in Lehane et al. (2005a), which considered a significantly wider
database of pile test data than originally envisaged by the API piling sub-committee.
The following observations were made during their study:
• The UWA database of pile load tests with CPT data (which is also described in
Chapter 3 of this Thesis) is larger than that employed for the derivation of the
Fugro-05, ICP-05 and NGI-05 design methods. There is, however, a significant
shortage of test data for piles with dimensions commonly used offshore.
• When tested against the UWA database, the Fugro-05, ICP-05 and NGI-05
methods provide substantially better predictions for the database piles than the
existing API recommendations (i.e. API 2000) and generally do not exhibit
significant bias with respect to pile length, diameter and sand relative density.
The ICP-05 method indicated the lowest coefficient of variation (COV) for
calculated to measured capacities (Qc/Qm) when an equal weighting is given to
each pile test in the database. However, for various categories within the database
the position is less clear. For example, NGI-05 predictions appear best for open-
ended piles in compression while the nominal reliability of Fugro-05 is
approximately the same as that of ICP-05 for open-ended piles in tension.
• Fugro-05 tends to over-predict capacity of piles in compression, particularly for

closed-ended piles. The predicted high radial stresses near the pile tip, and
offsetting high friction fatigue exponent lead to a sensitivity of this method to
cone tip resistance near the pile tip. This sensitivity may not be warranted, and
could be un-conservative for piles driven a short distance into dense layers
underlying soft layers.
1-2
• ICP-05 has a tendency to under-predict end-bearing capacity of open-ended piles

in loose sands or for large diameter thin walled piles. The formulation for end-
bearing of open-ended piles is seen as a significant limitation that may result in
large differences in end-bearing capacity (40 percent) for small changes in pile
diameter (1.1m to 1.2m for the uniform dense sands considered). This behaviour
may lead to arbitrary and unnecessary design decisions related to end-bearing of
large diameter piles in sand.
• NGI-05 appears to over-predict pile capacity in gravely sands, and possibly for
concrete piles. Simplifications in the method’s formulation for assessment of
interface friction angle and in its assumed differences between closed- and open-
ended piles are a limitation, especially when extrapolating outside of the database
used for its calibration.
The three CPT-based methods were therefore concluded to have a broadly similar
predictive performance against the UWA database of load tests. It should be noted
that although refined based on different assumptions, the shaft resistance (τf)
formulations of ICP-05 and Fugro-05 are generally similar (Schneider et al. 2006).
The formulations proposed by NGI-05 differ in format but also assume a near-
proportional relationship between local shaft friction (τf) and qc and allow for the
degradation of τf with distance above the pile tip (h) due to friction fatigue. However,
their formulations relating the pile end-bearing resistance (qb) with CPT end
resistance (qc) are remarkably different (as will be discussed in Chapter 3).
Prompted by such differences and the general need for piling research, this Thesis
presents a study into the end bearing performance of displacement piles in sand. The
research involved careful examination of the existing databases of static load tests in
sand, field and centrifuge testing that targeted uncertainties emerging from the
database review, and supporting numerical analyses.
1.2 RESEARCH OBJECTIVES
The aim of this research is to improve the accuracy and efficiency of the design of
axially loaded displacement piles in siliceous sand, with an emphasis on establishing
and testing a design formulation for the base resistance (qb) of closed- and open-
1-3
ended piles. Factors affecting end bearing, including the installation method, the pile
end condition (open or closed), and soil layering will be quantified to assist design.
Specifically, the key objectives of the research are to:
(1) Compile a database of displacement pile load tests in sand (in collaboration with
James A. Schneider), and extend it to include additional tests with base capacity
measurements to form new base capacity databases for driven and jacked piles.
(2) Assess the predictive performance of the three CPT-based methods (Fugro-05,
ICP-05 & NGI-05) using the new database of pile load tests.
(3) Use the database to develop a new improved design formulation for end-bearing,
which accounts for the effective area ratio of open-ended piles, differentiates
between driven and jacked piles, and employs a rational qc averaging technique.
(4) Validate the new proposals using a new set of field tests to be performed at
Shenton Park.
(5) Develop an improved understanding of the physical mechanisms which govern

the end-bearing of jacked piles and the end-bearing performance in layered soil
through a series of centrifuge investigations.
(6) Provide guidance on the influence of soil layering by numerical analyses (using
PLAXIS) for a two-layer soil stratigraphy.
1.3 ORGANISATION OF THESIS
This Thesis comprises nine Chapters.
Chapter 2 presents a review of the conventional design approaches for base

resistance of axially loaded piles in sand. In addition, the influence of soil layering
on base resistance is highlighted and current correction methods for soil layering are
summarised.
Chapter 3 compares the predictive performance of the three CPT-based methods

(Fugro-05, NGI-05 & ICP-05) and proposes an improved method, UWA-05, based
on the base capacity database of driven pile load tests in siliceous sand. The effects
of pile installation method (driving vs. jacking) and qc averaging techniques are also
discussed.
1-4
Chapter 4 describes the field pile testing programme conducted in Shenton Park,
Perth and compares the results obtained with the four CPT-based methods (including
UWA-05, described in Chapter 3).
Chapter 5 describes the centrifuge modelling laws and provides details of all the
testing apparatus involved in the centrifuge tests. The bulk of this Chapter discusses
the sample preparation and test procedures followed during the experimental
programme.
Chapter 6 presents the centrifuge test results in both uniform and layered soil
samples. Complete installation data and static load test results for the jacked piles are
analyzed and discussed.
Chapter 7 describes the results of PLAXIS numerical analyses that idealise model
pile penetration in layered soil assuming the spherical cavity expansion analogue.
The factors affecting the zone of influence and the reduction of resistance due to the
presence of neighbouring weak layers are identified.
Chapter 8 discusses major trends identified by this research, and compares the results
of the centrifuge modeling with the numerical analyses. The influence of the pile
installation method is further examined by comparing the findings from the database
and centrifuge studies.
Chapter 9 summarises the major conclusions drawn from this research and makes
recommendations for future work.
1-5
CHAPTER 2 LITERATURE REVIEW
2.1 INTRODUCTION
The design of piles has advanced steadily in recent decades largely because of the
substantial volume of research that has been conducted using instrumented model
piles and the extensive efforts made to establish reliable pile load test databases. A
number of Terzaghi and Rankine lectures have been delivered since the 1970s to
address various aspects and developments in the design of piled foundations - both
onshore and offshore (McClelland 1974, Meyerhof 1976, Poulos 1989, O'Neill 2001,
Randolph 2003, Poulos 2005). These lectures have highlighted the increasing
tendency to replace the conventional design methods with cone penetration test (CPT)
based methods for estimating pile capacity.
In fact, many comprehensive research reports (Titi & Abu-Farsakh 1999, NGI 2001,
Salgado et al. 2002, Fugro 2003, Jardine et al. 2005, Lehane et al. 2005a) focused on
evaluation of CPT-based methods. As a result, four new CPT-based methods, namely
Fugro-05 (Kolk et al. 2005a), NGI-05 (Clausen et al. 2005), ICP-05 (Jardine et al.
2005) and UWA-05 (Lehane et al. 2005b) will be included in the commentary of the
22nd edition of American Petroleum Institute’s (API) design guidelines (API 2006).
The topic, is, however, still in continuous evolution with developments in technology
(Bolton 2005, Mandolini et al. 2005).
This Chapter first reviews the current design frameworks for estimating pile end
bearing resistance (qb). These include: (i) conventional bearing capacity theory, (ii)
the cavity expansion method and (iii) correlations with in situ test measurements (e.g.
CPT qc or SPT N values). In fact, the first two approaches are often also employed
for the analysis of cone penetration, while the last one often avoids the need for
assessment of soil properties by using the in-situ test parameter directly in the
formulations. The various factors potentially influencing qb are then discussed.
2-1
Chapter 2 Literature Review
Finally, given its relevance to one component of this Thesis, the state-of-the-art for
end bearing resistance evaluation in layered soils is reviewed. The occasional
conflicting conclusions for end bearing in layered ground are discussed, with
emphasis on their current limitations.
2.2 PILE END BEARING RESISTANCE IN SANDS
2.2.1 Bearing capacity theory
The bearing capacity theory for shallow foundation analysis is based on the plasticity
approach first developed by Prandtl (1920). The extension to deep foundation
analysis requires additional assumptions to be made regarding the failure mechanism
and needs to assume incompressible material with a linear strength envelope and
plane strain conditions. Despite the disagreement between the observed and predicted
pile end bearing performance, the theory is still widely taught and used in practice
primarily only because of its relative simplicity and general acceptance by engineers.
As a result, the main text method in the 22nd edition of America Petroleum Institute
RP2A (API 2006) for estimating pile end bearing capacity in sand takes the
following conventional form:
q b = N q × σ' v 0 ≤ q b,lim it ( 2.1 )
where Nq is dimensionless bearing capacity factor, σ'v0 is the effective overburden

stress at pile tip level and qb,limit is the limiting value depending on soil classification
(e.g. loose, medium dense, dense, and very dense sands).
Over the years, a large variety of formulations for Nq have been published and there
is not yet general agreement on the most accurate formulation. The existing Nq charts
present Nq as a function of friction angle (φ) only. Figure 2.1 shows relationships
between Nq and φ, as proposed by different authors. The scatter of Nq values is
substantial for any given φ value and Nq calculated using any one method is sensitive
to small changes in φ (Poulos 1989). In addition, Gupta (2002) demonstrated that the
rigidity index Ir (=G/p'0), which generally reduces with depth (or stress level), also
has a significant effect on Nq. Research on factors affecting Nq appears to have been
largely ignored in the current design recommendations of API (2006), which
2-2
proposes a constant Nq value that varies only with the soil classification based on
estimated relative density (but with a limiting value qb,limit applied to qb).
Because both φ and Ir reduces with depth (Bolton 1986, Gupta 2002), Nq, which
depends on φ and Ir, also reduces with depth. A decrease in the value of Nq and
increase in the effective overburden stress (σ'v0) are such that qb (=Nq×σ'v0) continues
to increase with depth, while the rate of increase of qb with depth decreases.
However, qb does not reach a limiting value qb,limit as suggested by API (2006) at any
depth (Kulhawy 1984, Gupta 2002). Furthermore, (Klotz & Coop 2001)
demonstrated through a series of centrifuge tests that the introduction of qb,limit in API
RP2A is arbitrary and inconsistent with the mechanisms that govern the soil response
under pile base. In less compressible siliceous sand, this could lead to a conservative
design while in carbonate sands the base capacity can be over predicted.
Figure 2.1 Bearing capacity factor Nq proposed by different authors (Coyle & Castello 1981)
2-3
2.2.2 Cavity expansion method
Bishop et al. (1945) were the first to propose the analogy between cavity expansion
and a pile/cone penetration. The cavity (spherical or cylindrical) expansion method
(CEM) now plays an important role in modern soil mechanics (Salgado & Randolph
2001) and has attracted a significant amount of attention in the last 20 years. Closed-
form and numerical solutions have been presented for soils using progressively more
realistic and experimentally measured soil stress-strain curves (Carter et al. 1986, Yu
& Houlsby 1991, Salgado 1993, Collins & Yu 1996, Salgado et al. 1997, Ladanyi &
Foriero 1998, Cao et al. 2001, Chang et al. 2001, Cudmani & Osinov 2001,
Yasufuku et al. 2001, Schnaid & Mantaras 2003). A full review of CEM and its
various applications has been published by Yu (2000).
D=AC
qb
A C
E β α=45°+φ/2
α β=φ
F
plimit
τ=plimittanφ'
(a) Salgado (1993) (b) Randolph et al. (1994)
Figure 2.2 Assumed relationship between pile/cone tip resistance qb and cavity limit pressures plimit, (a)
cylindrical CEM, and (b) spherical CEM
According to this method, the pile/cone penetration can be simulated by expanding a

cavity of an initial zero radius or finite radius. In either case, as the pressure inside
the cavity increases, the cavity expands radially. At a certain radial displacement, the
expanding pressure will not increase with further radial expansion and reach a steady
state value, referred to as the limit pressure (plimit). The limit pressure plimit from
spherical or cylindrical expansion analysis is then related to the pile/cone resistance
2-4
based on some semi-empirical relationships (Salgado 1993, Randolph et al. 1994), as

shown in Figure 2.2 and summarised below.
Salgado (1993) demonstrated that the pile/cone resistance can be related to the
cylindrical cavity limit pressure plimit by an approximate slip line analysis ( Figure
2.2a). The numerical solution to plimit has been implemented in a program referred as
CONPOINT (Salgado et al. 1997). Stress rotation analysis produces the relationship
given in Equation 2.2 between the pile/cone resistance and plimit ; the equation
parameters (fv, Δλ, φT, Cλ & λ) are described in more detail by Salgado & Prezzi
(2007).
q b = 2f v p lim it exp(2Δλ tan φT )

(1 + C λ ) − C λ (η + 1) − 1
η+1
( 2.2 )
C λ2 η(η + 1)
Vesic (1977), Ladanyi (1961), Randolph et al. (1994) and Yasufuku et al. (2001),
believe that the soil displacements in front of the advancing pile/cone tip may be
considered closer to those undergoing spherical cavity expansion, as schematically
illustrated in Figure 2.3. The value of plimit from spherical cavity expansion can be
determined analytically or numerically depending on the soil models employed.
Based on vertical equilibrium of the soil wedge in Figure 2.2b, the following
relationship between qb and plimit is obtained:
q b = p lim it (1 + tan α ⋅ tan φ) ( 2.3 )
where φ is the soil friction angle and α equals to 60o for a standard cone
penetrometer, but can be taken as (45+φ/2) for the soil wedge under a pile.
CEM offers an alternative framework for interpretation of CPT data and estimation
of pile base resistance, and unlike conventional bearing capacity theory, can take
account both elastic and plastic deformation of the soil during penetration, and also
in an approximate manner, consider the influence of the penetration process on initial
stress states and effect of stress rotations that occur around the cone tip (Yu &
Mitchell 1998). In particular, it can provide greater confidence and more accurate
results if implemented in parametric studies (e.g. Moss et al. 2006). Therefore, as
will be discussed in Chapter 7, CEM is used to study the pile end bearing response in
layered soil. However, in design of piles, CEM has the disadvantage of requiring a
relatively large number of input parameters (compressibility, G, φ, ψ etc), which are
2-5
governed by numerous uncertainties (e.g. relative density, strain level, stress state,
loading history, ageing, disturbance during pile installation and soil anisotropy). An
accurate estimation of these parameters is difficult to attain in practice.
(a) Ladanyi (1961) (b) Yasufuku et al. (2001)

Figure 2.3 Failure mode immediately under the pile tip
2.2.3 Correlations with CPT data
The estimation of pile end bearing resistance utilizing the Cone Penetration Test
(CPT) end resistance (qc) is one of the earliest direct applications of CPT data. In fact,
the CPT was developed in the Netherlands over 70 years ago primarily as a model
pile for the efficient positioning of piles in sand layers below thick compressible
deposits (GeoDelft 1936). Over the years, a number of empirical methods have been
developed to estimate the base and shaft capacity of piles utilizing cone end
resistance qc, e.g. the Dutch method (Van Mierlo & Koppejan 1952, de Ruiter &
Beringen 1979), Schmertmann method (Schmertmann 1978), LCPC method
(Bustamante & Gianeselli 1982), and EF-97 (Eslami & Fellenius 1997). The
formulations for end bearing given by the Dutch and Schmertmann methods are
essentially the same, both being based on the same early research in the Netherlands.
As indicated in Equation 2.4, all of those methods employ a direct relationship

between pile base resistance (qb) and the average cone end resistance in the vicinity
2-6
of the pile base (qc,avg), but very often do not differentiate between end bearing of
closed- and open-ended driven piles.
q b = k c ⋅ q c,avg ( 2.4 )
where kc is the reduction factor relating pile end bearing qb to the averaged cone
resistance qc,avg. As discussed below, values of kc suggested by common design
methods vary from 0.3 to 1.0, depending on the sand relative density, pile installation
method and the qc averaging techniques employed.
(a) (b)
where qc,avg = 0.5 × [ 0.5 × ( qcI + qcII ) + qcIII ]; D: pile diameter
qcI: arithmetic average of qc values below pile tip over a depth which may vary between 0.7D
to 4D as shown in Figures 2.4a & b;
qcII: arithmetic average of the qc values following a minimum path rule recorded below the
pile tip over the same depth of 0.7D to 4D;
qcIII: arithmetic average of the minimum qc values following a minimum path rule recorded
above the pile tip over a height of 8D.
Figure 2.4 Calculation of the average cone tip resistance qc,avg in Dutch method
In the Dutch method, the pile end bearing is governed by the arithmetic averaged
cone resistance over a zone extending from 0.7 to 4 pile diameters below the pile tip
to 8 pile diameters above the pile tip. Figure 2.4 illustrates such a typical CPT
application. In Figure 2.4a, the qc values continue to increase with depth Therefore
qcI equals qcII and is calculated as the averaged qc values over the zone of 0.7D below
the pile tip, and qcIII is calculated as the averaged qc values over the zone of 8D above
pile tip. In Figure 2.4b, qc values do not continue to increase with depth and have
some low minimal. In this case, qcI is calculated as averaged qc values over the zone
of 4D below the pile tip, while qcII is evaluated as the averaged minimum qc values
2-7
(i.e. following minimum path rule: starting from 4D below the pile tip, the qc values
employed for averaging at a certain depth must not be greater than the low minimal
qc values at depth below) over the same zone (i.e. 4D), and finally qcIII is calculated
as the averaged minimum qc values (i.e. following the minimum path rule as
discussed above) over the zone of 8D above pile tip. In addition, as shown in Figure
2.5, recommended values of kc are specified as a function of the overconsolidation
ratio (OCR) and vary from 1.0 for OCR=1 to 0.5 for OCR=6 to 10. Moreover, a
nominal limiting value of 15MPa is applied to qb.
20
fine-coarse sand with OCR=1
very gravelly coarse sand and sand with OCR=2 to 4
fine gravel and sand with OCR=6 to 10
15
kc=1.0
kc=0.7
q b (MPa)
for OCR=1
for OCR=2 to 4
10
5
kc=0.5
for OCR=6-10
0
0 5 10 15 20 25 30 35
qc (MPa)
Figure 2.5 Values of kc for driven piles in sand (Dutch method)
Bustamante and Gianeselli (1982) proposed the LCPC method for the French
Highway Department based on analysis of 197 pile load tests with a variety of pile
types and soil conditions. In this method, kc varies from 0.3 to 0.5 depending on the
soil type and pile installation procedure (Table 2.1), and qc,avg is calculated as the
arithmetic average of qc values in the zone extending 1.5D above and below pile tip
level, as shown in Figure 2.6. Bustamante & Gianeselli (1982) noted that this method
can be applicable to open-ended piles only that are observed to plug during
installation.
2-8
Table 2.1 Values of kc factor for LCPC method
Soil Type Driven Piles Bored Piles

Medium dense to loose
0.5 0.4
sand, silt and gravel
Dense sand and gravel 0.4 0.3
The method by Eslami and Fellenius (1997) was calibrated against 102 pile load tests
with different pile sizes and lengths, installed in different types of soils. The unit end
bearing qb equals to qc,avg with kc equals to unity. The value of qc,avg is calculated as
the geometric mean of qc values at distance 4D below and 8D above the pile tip level
if the pile is installed from weak into strong soil and 2D above and 4D below if pile
is installed through strong soil and into weak soil. Although it is claimed by the
authors that this qc averaging technique (geometric mean) works better in non-
homogeneous natural soil, Xu & Lehane (2005) has shown this method could
significantly underestimate the representative qc,avg value in strongly layered soil.
Step 1: calculate q’c,avg by arithmetic averaging qc values over a zone 1.5D above and below pile
tip level; Step 2: eliminate qc values in the zone that are higher than 1.3 q’c,avg and those are lower
than 0.7q’c,avg in the upper zone of 1.5D; Step 3: calculate qc,avg by averaging the remaining qc
values over the same zone.
Figure 2.6 Calculation of the averaged cone resistance qc,avg (LCPC method)
Kraft (1990), among others, simply considers kc is a correction factor for site
variability and suggested that kc in absence of insufficient site specific data is 0.5 for
very dense, 0.55 for medium-dense and 0.6 for loose sand. It has so far been obvious
that there is no general agreement on the reduction factor kc and the qc averaging
procedure for estimating qb in the literature. Selections of kc and qc,avg are largely
2-9
based on local experience. Where possible, static load tests are required to confirm
the estimated pile capacity.
Nevertheless, Briaud & Tucker (1988), and others, show that CPT based design
methods offer greater reliability than conventional bearing capacity factor Nq based
methods, because of the geometric similarity between piles and CPT instruments.
However, there is a need to further understanding of the underlying mechanism
controlling qb and the difference between various qc averaging techniques. As will be
discussed below, qb depends not only on the pile dimension and the nature of the
sand, but also on the installation method of displacement piles and other factors.
2.3 FACTORS AFFECTING PILE END BEARING RESISTANCE
2.3.1 Method of installation
According to Mandolini et al. (2005), the onshore pile market world is equally
subdivided between displacement (driven, jacked, screwed, etc.) and non-
displacement piles (bored, continuous flight augered, etc). Driven piles dominate the
market for offshore conditions (McClelland 1974).
The effects of installation methods are particularly significant for piles designed to
sustain axial loading. In fact, the bearing capacity and load settlement behaviour of
an axially loaded pile depends primarily on the characteristics of the soil
immediately adjacent to the shaft and below the base of the pile. In those zones the
installation produces considerable variations to the stress state and sand density.
To demonstrate the influence of installation method on pile capacity, Mandolini et al.

(2005) employed a database of 20 load tests on concrete piles installed in the
relatively uniform pyroclastic soils of the eastern Naples area (D=0.3m to 2m,
L=9.5m to 42m, and L/D=16 to 61). In Table 2.2, the results were summarized in
terms of the ratio of pile total capacity Qtotal (defined at 0.1D of pile head
displacement) to the weight of the pile Wpile. It came as no surprise that driven piles
give the largest value (73 times the pile weight) with the lowest coefficient of
variation (COV=8%), while bored piles gives the smallest value and also the larger
scatter.
2-10
Table 2.2 Comparison of total capacities of different pile types (Mandolini et al. 2005)
Pile type Qtotal/Wpile COV(Qtotal/Wpile)

Bored 12.1 0.26
CFA 37.5 0.25
Driven 73.1 0.08
Figure 2.7 Relative stiffness of bored, driven and jacked piles (Deeks et al. 2005)
Mandolini et al. (2005) also studied the effects of pile installation on the axial
stiffness of piles based on 125 load tests in eastern Naples. It was concluded that the
installation method affects the axial stiffness of the piles (within ±20%) much less
than their bearing capacity. The initial stiffness (defined as the initial tangent of a
hyperbola fitted to the first three points on the experimental load-settlement curve)
depends primarily on the small strain shear modulus of the soil. However, it should
be noted that this is in particularly true for long friction piles where the axial pile
2-11
stiffness is primarily controlled by the skin friction along the pile, which is generally
mobilised at very small displacement (e.g. less than 1%D). In fact, the base stiffness
of piles could be significantly affected by the method of installation. Deeks et al.
(2005) have shown that the characteristic secant base stiffness, ks, of the field tested
jacked piles at a settlement of 0.02D is more than 2 and 10 times greater than
recommended design values for driven and bored piles respectively. Figure 2.7
illustrates this important effect for base resistance.
q c (MPa)
0 5 10 15 20 25 30 35
0
1
3
depth (m)
5 D=45mm
6 D=110mm
D=216mm
7
D=320mm
8
9
10
Figure 2.8 Cone resistance in terms of different diameters in dense sand (Kerisel 1961)
2.3.2 Surface scale effect
Tests performed by Kerisel (1961) in very dense dry sand show a marked surface
scale effect, i.e. the penetration resistance at a given depth varies considerably with
the pile dimensions. As shown in Figure 2.8, at relatively shallow depth, penetration
resistance qc reduces considerably with increase in pile diameter (i.e. qc varies from 8
to 27 MPa at depth of 1m). The embedment depth to diameter ratio, L/D, required to
reach a pseudo-constant qc value is approximately 20 for both 45mm and 110mm
diameter piles. The penetration resistance for the pile with a diameter of 320mm is
about two thirds that of a 45mm diameter pile at depth of 3m. This (apparent)
marked influence of the ground surface may reflect a dependence of the failure
mechanism on the L/D ratio, which was exacerbated by the reported ‘silo-effect’ in
2-12
Kerisel’s chamber. As illustrated in Figure 2.9, the failure mode for foundations in
sand were categorized by Vesic (1963) as general shear, local shear and punching
shear failure depending on the sand relative density and L/D ratio.
Relative Density Dr (-)

0 0.5 1
0
General Shear
Embedment Ratio, L/D (-)
Local Shear
circular
foundation
5
Punch Shear long

rectaangular
foundation
10
Figure 2.9 Types of failure at different relative depth of foundations in sand (Vesic 1963)
Figure 2.10 Relationship between qb and qc in sand (Meyerhof 1983)
From theoretical consideration based on rupture formulae, De Beer (1963) tried to

explain Prof. Kerisel’s test results and concluded that for determining the end bearing
capacity of piles in a homogeneous sand layer starting from ground surface it is
2-13
necessary to apply some reduction coefficients to the qc values measured from CPT.
This view is supported by the database of measured base resistance compiled by
Meyerhof (1983) as shown in Figure 2.10, which illustrates that as the embedment
ratio (indicated by Db/B in the Figure) decrease (shallowly embedded into the ground
surface), values of qb are considerably less than qc, while for Db/B greater than 10,
values of qb are approximately equal to that of qc.
2.3.3 End condition (open vs. closed)
Large diameter open-ended steel pipe piles (2-3 meters) are increasingly being used
for bridge and offshore foundations, as they provide superior moment resistance,
high axial capacity and may now be driven using modern pile driving equipment.
Figure 2.11 shows the photos of the large diameter pipe pile (D=2.4m, t=40-70mm
and L=107m) and pile hammer from a recent pile installation demonstration project
for the new San Francisco-Oakland Bay Bridge (SFOBB). This transition to larger
and larger diameter pipe piles raises concerns regarding constructability, but has also
prompted the need for more reliable design methods. For example, design
recommendations of API (2000), which is widely used offshore (but also onshore),
does not consider the effects of the degree of soil displacement on the pile bearing
capacity imposed by closed- and open-ended piles.
Pile Scale
Lifting Hammer
Hammer Scale
Figure 2.11 Photos of large diameter pipe pile and impact hammer (D=2.4m, L=107m)
2-14
A soil plug is known to advance up inside a pipe pile during driving. Many
researchers have shown, both numerically and experimentally, that the behaviour of
this plug could significantly affect the dynamic driving resistance and static bearing
capacity of a pipe pile (Kishida 1977, Smith et al. 1986, Paikowsky et al. 1989,
Paikowsky & Whitman 1990, Brucy et al. 1991, Matsumoto & Takei 1991, O'Neill
& Raines 1991, de Nicola & Randolph 1997, Miller & Lutenegger 1997, Lehane &
Gavin 2001, Lehane & Randolph 2002, Malhotra 2002, Gavin & Lehane 2003, Paik
et al. 2003, White et al. 2003, Paik & Salgado 2004). For instance, if a pipe pile
plugs fully (i.e. no soils enter inside) early during driving, the pile displaces a volume
of soil approximately equal to the volume of an equivalent solid pile, resulting in
increased radial and tip stresses, which can lead to higher driving resistance and
static capacity. On the other hand, large diameter pipe piles, which often tend to
install in fully coring mode, displace a volume equal only to the equivalent pile
volume, resulting in lower radial and tip stresses. This effect is illustrated
schematically by White et al. (2005) in Figure 2.12. It would inevitably influence the
pile performance in both sand and clay. Apart from the main theme of this Thesis,
the performance of pipe pile in clay was also investigated through field and
centrifuge testing programme. Such results are summarised in Appendix D by two
published papers.
Figure 2.12 Schematic streamlines of soil flow and profiles of radial stress; δr=radial displacement of
soil element at pile wall (White et al. 2005)
2-15
The degree of soil plugging (coring, fully plugged or partially plugged as shown in
Figure 2.13) during driving are a complex function of the pile base configuration, in
situ stress, soil properties, and installation method. This installation effect is best
described by the incremental filling ratio (IFR=ΔLp/ΔL, incremental change of plug
length Lp relative to change in pile penetration depth L). The skin friction and end
bearing resistance were shown (Salgado et al. 2002, Gavin & Lehane 2003, White et
al. 2005, Xu et al. 2005) to be a function of the average of IFR during installation
and IFR measured at the final stage of installation (referred to as the final filling ratio,
FFR). However, despite the strong impact of soil plug response during driving on the
shaft and base capacity, there is no current design methods (except UWA-05 method
as will be discussed in Chapter 3) that takes IFR or FFR into account, presumably
because its measurements is still not a standard. There have been recent attempts to
develop a methodology for estimating IFR from the given pile and sand conditions.
Salgado et al (2002) proposed an empirical relationship between IFR, sand relative
density (Dr) and relative penetration depth (L/Di). The procedure is illustrated in
Figure 2.14.
Figure 2.13 Possible failure mechanism for open-ended piles (Smith et al. 1986)
2-16
Figure 2.14 Relationship between normalized IFR [NIFR(%)=IFR(%)/Dn; Dn=(L/Di)] and relative
density Dr (Salgado et al. 2002)
2.3.4 Residual stress
During installation of a displacement pile, the jacking or driving force on top of the
pile head is resisted by upward skin friction and end bearing resistance. At the end of
the installation, the pile tends to rebound with removal of the pile head force. This
will result in a reversal in the direction of the skin friction in the upper part of the
pile (acting downwards now), while the lower part may still remain in compression.
At zero pile head load, the base resistance which is in equilibrium with the
downwards skin friction is referred to as the residual stress or locked-in base
resistance (qb,residual). The development of residual stress is schematically illustrated
in Figure 2.15.
As discussed by Hunter and Davisson (1968), Briaud & Tucker (1984) , Poulos
(1987), Kraft (1991), and others, residual stress can significantly affect the load-
settlement behaviour under static loading in both compression or tension, particularly
in case of high end bearing capacity. The residual stresses do not affect the ultimate
bearing capacity of the pile at very large displacement. However, if they are
neglected (i.e. assumed to be zero), their presence may lead to an overestimation of
the shaft resistance and subsequent underestimation of the base resistance in a
2-17
compression load tests (and vice verse in a tension test). White & Bolton (2005)
highlight their importance when developing correlations between qb0.1 (defined at
pile tip displacement of 0.1D) and qc based on full scale pile load tests.
Figure 2.15 (a) pile in compression during driving, (b) residual stresses with zero pile head load, and
(c) axial load distribution with zero pile head load
The distribution and magnitude of residual stress are suggested (Hunter & Davisson
1968, Briaud & Tucker 1984, Alawneh & Husein Malkawi 2000) to be a function of
ultimate base and total loads, the pile length, relative pile soil stiffness, and
installation method (driving or jacking). Residual stresses, however, cannot still be
reliably evaluated (and their prediction will need to be a subject of future research).
A recent method proposed by Alawneh & Husein Malkawi (2000) for estimation of
residual stresses, which is based on a database study, is briefly illustrated in Figure
2.16 and Equation 2.5. It is shown by these authors that among other factors, pile
flexibility is a key parameter controlling the magnitude of residual stresses for
displacement piles in sand.
L Ap G
q b ,residual = 13158 ⋅ (η) η= ⋅ ⋅
0.724
(kPa); ( 2.5 )
D A Ep
2-18
where L is pile embedment length; D is pile diameter; A is pile cross section area; Ap
is the gross area of the pile; G is the small strain shear modulus of the sand; and Ep is
the Young’s modulus of pile material. The method is evidently highly empirical.
Figure 2.16 Correlation between end residual stress qb,residual and dimensionless pile flexibility factor η
(Alawneh & Husein Malkawi 2000)
2.3.5 Partial mobilisation
Unlike ultimate skin friction, which is often mobilized at relatively small

displacement (<0.01D), steady state end bearing resistance for a pile, qb
(approximately equals to qc from cone penetration test) may only be reached after
very large displacements during static load test (i.e. 4-10D, Randolph 2003).
However, in practice, the end bearing capacity is normally defined at 0.1D of pile tip
displacement, which could therefore be much lower than the steady state penetration
resistance. This apparent scale effect when correlating qb0.1 (mobilized at 0.1D of tip
displacement) to qc due to partial mobilisation is also one of the main reasons for a
reduction factor kc. Lee & Salgado (1999) and Oda & Matsui (2001) have shown
from their numerical analysis that the bearing resistance is not at the ultimate state at
0.1D pile tip settlement (w). This view is also supported by field pile load tests as
shown in Figure 2.17, which is extracted from the UWA pile load test database for
2-19
base capacity (discussed in Chapter 3). The ratio of qb to qc,avg ranges approximately
from 0.4 to 0.75 at normalised displacement w/D of 10%. For all tests, except one at
loose sand site Drammen, there is no tendency for qb to reach qc within w/D of 50%.
0.9
0.8
0.7
q b/q c,avg (-)
0.6
0.5
0.4
Akasaka; 6C
0.3 Drammen; A
Drammen; D/A
0.2 Hoogzand; II-C
Hunter's P; S
0.1 Pigeon R; CEP
0
0 10 20 30 40 50
w/D (%)
Figure 2.17 Normalised end bearing resistance (qb/qc,avg) plotted against normalized displacement
(w/D)
2.3.6 Scale effect in layered soil
The significant influence of soil layering on pile/cone tip resistance has long been
recognized by Begemann (1963) based on early research in Holland where
practically all piles derive their bearing capacity from the pile base embedded in deep
sand layers underlying soft clay. In this case, as the pile/cone approaches the sand
layer, it starts to ‘sense’ the presence of this layer some distance away. Also, after the
pile/cone enters the sand layer, it continues to ‘sense’ the softer mud above it, and
this will result in a lower tip resistance for some distance into the sand layer. The
base resistance will evidently vary with the base diameter in the different layers and
there is an obvious ‘scale effect’ when assessing base resistance of larger diameter
piles using the base resistance profile given by a 36mm diameter cone.
2-20
qc (MPa) qc (MPa) qc (MPa)

0 5 10 0 5 10 15 0 5 10 15 20 25 30 35 40 45
10 0 6
11
weak (a) 1
(b) weak (c)
7
12 2 weak
8
13 3
strong
9
14 4 strong
15 5 10
depth (m)
16 6
strong
11
17 7
12
18 8
weak
13
19 D=36mm 9 D=36mm D=36mm
D=430mm D=114mm D=250mm
20 10 14
Figure 2.18 Scale effect in measured qc of a cone and a pile (a) Plantema (1948), (b) Begemann
(1963), and (c) De Beer (1979)
Nauroy & Le Tyrant (1983) reported that the high stress surrounding the advancing
pile/cone in calcareous sand has an approximate extent of 10 pile diameters. For a
standard cone of 36mm in diameter, this implies that the zone of influence may
extend a distance of 0.4m ahead of the tip, while for a pile of 0.5m in diameter, this
extent equals 5m. Figure 2.18 shows the measured tip resistance during jacking
installation of a standard cone and larger diameter piles. The steady state penetration
resistances differ significantly when travelling from a weak into a strong layer or a
strong into a weak layer.
Although there has been abundant research reported in the literature for pile/cone
penetration in homogenous soil, the penetration resistance in layered soils has not
received significant attention. There is generally no agreement on the zone of
influence ahead and behind a pile/cone in layered soil and on the most appropriate
correction method. Previous research in this area is discussed in Section 2.4.
2.3.7 Assessment of pile base settlement
Poulos (1989) demonstrated that in the realm of pile settlement prediction, the
method of analysis is likely to have less influence than does the geotechnical
characterization of the site. Fleming (1992) also suggested that a simple approach
should be adopted coupled with site experience and mainly used parameters that
2-21
most geotechnical engineers would recognize and understand. Hyperbolic functions

were therefore proposed by Fleming (1992) to describe individual shaft and base
performance under axial loading for bored piles. The base resistance qb was linked to
base displacement w, giving the normalised base resistance qb/qc expressed as
Equation 2.6 below.
qb w D
= ( 2.6 )
q c w D + 0.2 q c G IN
where D is the pile diameter, qc is the CPT tip resistance and GIN is the initial shear
modulus of the soil below pile base.
Although the exact form of pile base response is still debatable, Equation 2.6 does
provide some insights into the factors affecting the degradation of secant stiffness
(G/GIN) with the increased in strain (i.e. w/D) and loading level (qb/qc), which are
summarized in Equation 2.7 and Figure 2.19.
G 0.2 q c G IN
= ; hyperbolic relationship of G/GIN vs. w/D ( 2.7a )
G IN w D + 0.2 q c G IN
G q
=1− b ; linear relationship of G/GIN vs. qb/qc ( 2.7b )
G IN qc
The following can be observed: (i) the stiffness decays with w/D in a hyperbolic
format; the degree of degradation (G/GIN) at certain w/D depends on the initial shear
stiffness ratio GIN/qc; the higher the ratios of GIN/qc, the lower the ratios of G/GIN at
certain w/D; and (ii) there is a unique linear relationship between G/GIN and the
loading level qb/qc.
In fact, by knowing the exact form of stiffness degradation with strain and loading
levels (i.e. G/GIN vs. w/D & qb/qc), the load settlement relationship (i.e. qb/qc vs. w/D)
can be subsequently inferred. Similar proposition to generalise jacked pile base
response has been suggested by White (2006) based on a series of full-scale closed-
ended jacked pile tests (Figure 2.20). However, it is obvious that the form of stiffness
degradation (e.g. G/GIN does not vary linearly with qb/qc for jacked piles) is different
from that proposed by Fleming (1992) for bored piles. This is as expected since the
method of installation can significantly alter the base stiffness response under
loading. The exact form of the stiffness degradation curves for jacked piles will be
investigated in Chapter 8 based on the series of centrifuge conducted in this study.
2-22
1
(a)
0.9
0.8
0.7
0.6
q b/q c (-)
0.5 80
40
0.4 25
18
0.3 15
0.2 8
GIN/qc = 6
0.1 decrease in GIN/qc 4
2
0
0 2 4 6 8 10
w/D (%)
1
(b)
0.9
decrease in GIN/qc
0.8
0.7
G/G IN (-)
0.6
80 0.5
40
25 0.4
18
15 0.3
GIN/qc = 8
6 0.2
4 0.1
2
0
0.01 0.1 1 10 100
w/D (%)
1
(c)
0.9
0.8
0.7
G/G IN (-)
0.6
0.5
0.4
0.3
0.2
0.1
0
0 0.2 0.4
qb/qc (-) 0.6 0.8 1
Figure 2.19 Normalised plots (a) qb/qc vs. w/D, (b) G/GIN vs. w/D, and (c) G/GIN vs. qb/qc (Fleming
1992)
2-23
Figure 2.20 Generalisation of base response for jacked piles (White 2006)
2.4 PREVIOUS RESEARCH FOR qb IN LAYERED SOIL
Previous research (e.g. Meyerhof & Sastry 1978a,b, Vreugdenhil et al. 1994, Joer et
al. 1996, van Den Berg & Huetink 1996, Ahmadi et al. 1999, Berrill et al. 2004)
sought to derive correction factors to the measured qc,D=36mm to obtain the steady state
penetration resistance qb of a much bigger pile in layered soil, and apply a correction
factor to the measured qc,D=36mm value in very thin soil layers to estimate the true
characteristic steady state penetration resistance in the absence of neighbouring
weak/strong layers.
Figure 2.21 Relationship between qb and embedment depth H in layered soil
2-24
2.4.1 Meyerhof approaches (1976-83)
Meyerhof and co-workers (1976, 1977, 1978a, 1978b, 1983) concluded from small
scale laboratory test results that the required depth of penetration into a stronger layer
to develop full resistance of the layer decreased with an decrease in pile diameter and
a decrease in strength difference between the layers. However, the authors
recommended a constant zone of influence in the strong soil (ZS) of 10 pile diameters
(10D). This is shown schematically in Figure 2.21, where the resistance in strong soil
increases linearly with depth from the steady state values in the weak soil of qb,W to
qb,S in the strong soil.
For pile located at distance, H, away from the interface of weak and strong soils, as
shown in Figure 2.21, the steady state resistance qb in the strong soil is given by:
q b = q b ,W +
(q b ,S − q b ,W ) ⋅ H
, or conservatively q b =
q b ,S ⋅ H
( 2.8 )
10 D 10 D
In this proposition, it is assumed implicitly that qb in the weak soil is not influenced
significantly by the nearby strong soil layer and the zone of influence (ZW) is less
than 2D and can be taken as zero.
2.4.2 Vreugdenhil et al. (1994)
In an attempt to quantify the layering effect more precisely, Vreugdenhil et al. (1994)
performed simple linear elastic analyses of pile/cone resistance in multi-layered soil
systems. The objective of the model was to quantify the effect of adjacent layers on
the cone resistance, using the undrained elastic stiffness of an embedded circular
‘foundation’ of diameter, D, as a qualitative analogue of the pile/cone resistance.
Although the penetration problem is far too complicated to be modelled by an elastic
analysis, the author argued that it may prove to be useful since the layering effect at a
far distance from the pile/cone tip must be essentially elastic in nature.
The results from the elastic analysis were presented in terms of normalised resistance
qb/qb,S, as a function of the normalised distance to the interface of two soil layers
(H/D) and of the stiffness ratio of the strong and weak soil, GW/GS. The solution
provided in Equation 2.9 suggests that the qb,W/qb,S ratio for the weak and strong soil
layers equals the stiffness ratio GW/GS. The graphical output for resistance reduction
2-25
curves (qb as function of H/D) are shown in Figure 2.22, where various curves
represent solutions based on different values of GW/GS (or qb,W/qb,S).
2 − λ1
q b q b,S = , in the weak soil, H<0
2k1 ⋅ (1 − λ1 )
( 2.9 )
2 + k1 ⋅ λ1
q b q b,S = , in the strong soil, H≥0
2 + 2k1 ⋅ λ1
1 − 1 k1
where λ 1 = and k 1 = G S G W
1 + (2 ⋅ H D )
2
As shown in Figure 2.22, the zone of influence in the weak soil (ZW) is relatively
small (H<2D) and generally independent of the resistance ratio of qb,W/qb,S; the zone
of influence in the strong soil (ZS) increases as the resistance ratio (qb,W/qb,S)
decreases, ranging from 10D at qb,W/qb,S=0.5 to greater than 40D at qb,W/qb,S=0.05.
This effect points to the inadequacy of the adoption of a constant zone of influence
(e.g. the 10D zone suggested by Meyerhof and his co-workers). In addition, the
approach suggests that, unlike Meyerhof’s proposal, the resistance qb does not vary
linearly with H/D.
0.8
q b/q b,S (-)
0.6 q b,W /q b,S=0.05

GS/GW=20
q b,W /q b,S=0.1
GS/GW=10
0.4 q b,W /q b,S=0.2
GS/GW=5
q b,W /q b,S=0.5
GS/GW=2
Weak Soil Strong Soil
0.2
0
-10 0 10 20 30 40
H/D (-)
Figure 2.22 Resistance reduction curve (qb/qb,S as a function of H/D)
Despite the obvious limitation of the approach (i.e. the assumption of elasticity), Joer
et al. (1996) and Berrill et al. (2004) independently show that the Vreugdenhil (1994)
approach provides a promising basis for a new improved method for evaluating qb in
2-26
layered soil. It was subsequently employed by Moss (2003) for qc correction in very
thin layered soils to obtain the true characteristic qc value for the assessment of
seismic soil liquefaction and is also included in Lunne (1997) for assessment of
layering effect. Such recognition should, however, be treated with caution. For
example, the method predicts that at qb,W/qb,S=0.01 (typical qb values of 0.1MPa and
10MPa in soft clay and dense sand respectively) the value of ZS is as high as 450D.
In other words, it suggested that a standard cone of 36mm in diameter will continue
to feel the presence of the neighboring soft clay at depth of 16m into the sand layer
below, which is practically impossible. It is noted that the method was calibrated
using a limited number of tests for which values of qb,W/qb,S were greater than 0.15.
2.4.3 van den Berg & Huetink (1996)
van den Berg & Huetink (1996) presented an Eulerean large strain finite element
formulation to simulate static soil penetration in both homogeneous and layered soil
profiles (clay on sand; or sand on clay). In the layered soil, the strength of the sand
was characterized by a non-associative Drucker-Prager criterion depending on peak
and critical friction angles and the undrained clay behaviour was modelled using a
Von Mises criterion.
2-27
Figure 2.23 Load displacement curves for homogenous and layered soil profiles: (a) clay over sand &
(b) sand over clay (van Den Berg & Huetink 1996)
The results, as shown in Figure 2.23, suggest that a cone coming out of a sand layer
senses a soft clay layer at a distance of about three times the diameter of the cone
(3D). Simulation of cone penetration from soft clay into sand shows that a
penetrating distance of at least four times the diameter (4D) is needed into the sand
layer to reach the new steady-state cone resistance corresponding to that layer.
However, these results are strongly dependent on the material properties of the layers
and qualitative in nature as only a few cases were computed. To obtain more precise
and definitive results, more refined constitutive models and a parametric study of
these models’ parameters is required.
2.4.4 Ahmadi & Robertson (2005)
Ahmadi & Robertson (2005) presented a numerical modelling procedure using the
FLAC program to evaluate cone tip resistance in sand. The soil was modelled as a
Mohr-Coulomb elastic plastic material with a stress level dependent stiffness. The
procedure was firstly verified against published experimental measurement of qc
from calibration chamber tests. It was then employed to simulate cone penetration in
layered soil: (i) dense over loose sand, (ii) loose over dense sand, and (iii) sand over
clay. Figure 2.24 shows the results of penetration analysis in the layered soils. Value
of ZS (zone of influence in the stronger soil, dense sand) are in the range of 7D to
2-28
12D, while value of ZW (zone of influence in the weaker soil, loose sand and clay)
are in the range of 1D to 4D. In fact, once the cone tips in clay, resistance is
primarily controlled by the strength of the clay and the above sand has little effect.
Considering the case of sand over clay in Figure 2.24c, it is noted that the cone only
start to feel the existence of the clay at distance ~12D above the interface, while
solution by Vreugdenhil et al. (1994) is suggesting a zone of influence of over 100D.
Ahmadi (2000) states that the analyses presented are very time consuming, and as
such predictions for the full range of layered deposits encountered in practice has yet
to be performed.
q c (MPa) q c (MPa) q c (MPa)

0 5 10 15 20 0 5 10 15 20 25 0 5 10 15
0 0 0
(a) (b) (c)
5 5 5
10 10 10
15 Sand, D r=50% 15 15
20 20 20
25 Sand, D r=90% 25 25
30 30 30
z/D (-)
35 35 35
Sand, D r =90%
40 40 40
45 45 45 Sand, D r =70%
Sand, D r=50%
50 50 50 Clay, s u=30kPa
55 55 55
60 60 60
Figure 2.24 Results of cone penetration analysis in layered soil (Ahmadi 2000)
2.5 SUMMARY
The review of design methods for base capacity, factors influencing this capacity,
and previous research in layered soil can be summarized as follows:
• The conventional bearing capacity theory (Nq based) for shallow foundations
in low confining stress regimes does not offer a sound approach for the
analysis of deep foundations, which are better modelled using the contained
failure, cavity expansion method (CEM). However, the parameter required
for CEM, such as shear modulus (G), dilation angle (ψ) and friction angle (φ)
2-29
can be equally difficult to evaluate and vary with stress state, strain level and
sand anisotropy. CEM therefore may not be employed directly as a design
method for pile base resistance, but may be used to improve understanding
and interpretation of pile/cone installation base resistance.
• Cone penetration tests (CPT) based methods should offer the most reliable
way to estimate the pile end bearing resistance. There is however a significant
difference between existing formulations (e.g. Dutch method, LCPC, EF-97
etc). The significant role of soil displacement in estimating pile capacity is
generally ignored.
• For practical design purposes, ultimate end bearing resistance (qb0.1) is

normally defined at tip displacement of 10% of pile diameter (0.1D). Value
of qb0.1 could therefore be a lot lower than the true steady state penetration
resistance, qb or qc due primarily to the base resistance partial mobilisation
during a static load test. A reduction factor kc is generally applied to the CPT
data to account for this effect. In case of presence of non-zero base resistance
(qb,residual, residual stress), qb0.1 could be higher due to prestressing and pre-
stiffening.
• Theoretically, the steady state penetration resistance for a deep embedded pile
and cone (qb and qc) should be the same in homogeneous soil. However,
natural soil comprises layered deposits with different thicknesses and
properties. This layering leads to a pronounced scale effect as the much larger
diameter pile reacts more slowly to soil stratigraphy changes due to the
involvement of a large volume of soil mass.
• The effect of pile/cone penetration in layered soil is not adequately addressed

in literature. Newer approaches such as that reported by Vreugdenhil et al.
(1994) are an improvement, but their limitations need to be addressed before
a general new formulation is adopted in practice.
2-30
CHAPTER 3 CPT-BASED METHODS FOR END
BEARING OF PILES IN SILICEOUS SANDS
3.1 INTRODUCTION
Static design of piles in sand has historically been based on visual classification
methods put within the framework of the bearing capacity theory for base resistance
(qb) and the earth pressure approach for skin friction (τf), but the 22nd Edition of
American Petroleum Institute (API) recommended practice (RP2A) for fixed
offshore structures (2006) has added four ‘new’ CPT based methods in the
commentary. This Chapter examines the relative reliability of the end bearing
formulation of three such methods (Fugro-05, ICP-05 & NGI-05) using a new
database of high quality base capacity measurements for driven piles in siliceous
sand and also develops and examines the end bearing component of the fourth API
CPT method, referred to as UWA-05. The composition of the compiled full UWA
database and assessment of predictive performance of total compression and tension
capacities is discussed by Schneider, Xu and Lehane (2006) and also summarised in
Appendix A.
The four CPT methods were developed using static load test databases of driven piles
in siliceous sand although several jacked piles were included in the development of
ICP-05 and NGI-05. Jacked piles are often employed in urban areas where
conventional dynamic pile driving is not permitted due to noise, vibration and
pollution restrictions. The latter half of this Chapter therefore extends the database
study on driven piles (conducted by the author in collaboration with Prof. Barry
Lehane and Mr. James Schendier) to jacked piles.
3-1
Chapter 3 CPT-Based Methods for End Bearing of Piles in Siliceous Sands
3.2 CPT-BASED DESIGN METHODS FOR DRIVEN PILES
The relatively high cost-effectiveness, reliability and speed of a CPT has contributed
to its ever-increasing popularity for site characterisation, while the similarities
between the geometry and installation of a CPT tool and a displacement pile provide
the basis for direct correlations between the CPT end resistance and pile end bearing.
Difficulties related to the sensitivity of analytical/numerical solutions to the assessed
operational friction angle and to the theoretical treatment of variations in sand
consistency near a pile base have contributed to the move towards CPT-based
correlations.
With the growing application of large diameter open-ended driven piles in both
onshore and offshore environments, a number of recent research efforts have re-
examined the relationship between qc and pile (base and shaft) capacity in siliceous
sands with a recognition of differences between closed- and open-ended piles. These
have resulted in Fugro (Fugro-05), Imperial College (ICP-05), Norwegian
Geotechnical Institute (NGI-05) and University of Western Australia (UWA-05)
methods, as published in Kolk et al. (2005a), Jardine et al. (2005), Clausen et al.
(2005) and Lehane, Schneider & Xu (2005b) respectively.
Part of the motivation behind the development of the UWA-05 method presented in
this Chapter 1 , was the desire to provide a formulation that could account for the
varying levels of soil displacement imposed during pile installation. The behaviour
of open-ended piles is expected to lie between that of ‘full-displacement’ and ‘non-
displacement’ piles, as shown by full-scale field tests (Kishida 1967, Paik et al.
2003), laboratory testing chamber studies (O'Neill & Raines 1991, Foray et al. 1998),
and centrifuge model pile tests (de Nicola & Randolph 1999). Paikowsky &
Whitman (1990), Lehane & Gavin (2001) and others, have shown that this
installation effect may be described by the incremental filling ratio (IFR=ΔLp/ΔL,
incremental change of plug length Lp relative to change in penetration depth L) and
that a more consistent relationship between qb and qc emerges when the final filling
1
Noting that this author focussed on the end bearing formulation for UWA-05 and James
Schneider, who is a PhD student at UWA focussed on the shaft friction formulation.
3-2
ratio (FFR) (i.e. IFR measured over the final stages of pile installation) is included in
a formulation for end bearing. Despite the strong impact of the soil plug response
during driving on base (and shaft) capacity, the FFR is not incorporated in the Fugro-
05, ICP-05 and NGI-05 methods.
Table 3.1 Design methods for base resistance (qb0.1) of driven piles in siliceous sand
Methods End Design equations

Condition
Fugro-05 Closed qb0.1 / qc,avg = 8.5 × ( pa / qc,avg ) 0.5 × Ar 0.25

& Open
ICP-05 Closed qb0.1 / qc,avg = maximum [ 1 – 0.5 × log( D / DCPT ), 0.3 ]
Open if Di ≥ 2.0 × (Dr ─ 0.3) or Di ≥ 0.083 × qc,avg / pa × DCPT, Di in meter, then

the pile is “unplugged”, and qb0.1 / qc,avg = Ar
if not, the pile is “plugged”, and
qb0.1 / qc,avg = maximum [ 0.5 - 0.25 × log( D / DCPT ), 0.15, Ar]
NGI-05 Closed qb0.1 / qc,tip= FDr = 0.8 / ( 1 + Dr2 )
Open qb0.1 = minimum [ plugged qb0.1, unplugged qb0.1 ]

the plugged qb0.1value is calculated as:
qb0.1 / qc,tip = FDr = 0.7 / ( 1 + 3Dr2 )
the unplugged qb0.1 is calculated as:
qb0.1 = qb,ann × Ar + qb,plug × ( 1 - Ar ); qb,ann = qc,tip, and qb,plug= 12τf,avg×L/Di
τf,avg: averaged external skin friction (Clausen et al. 2005);
L: pile embedment depth
UWA-05 Closed qb0.1 / qc,avg = 0.15 + 0.45 × Ar,eff

& Open
Note D: pile outer diameter; Di: pile inner diameter

pa = 100 kPa; DCPT = 0.036 m; Ar (area ratio) = 1 - ( Di / D ) 2 ;
Ar,eff (effective area ratio) = 1 - FFR × ( Di / D ) 2;
FFR: IFR (= ΔLp / Δz) averaged over last 3D of the pile penetration
qc,avg = qc averaged ±1.5D over pile tip level for Fugro-05 & ICP-05
qc,avg = qc averaged using Dutch averaging technique for UWA-05
Dr (nominal relative density) = 0.4 × ln [ qc,tip / 22 / ( pa × σ’v0 ) 0.5 )]
Dr >1 is accepted for NGI method; as decimal
3-3
Table 3.2 Design methods for local shaft friction (τf) of driven piles in siliceous sand
Methods Design equations

0.05 −0.90
⎛ σ' ⎞ ⎛ h ⎞
τ f = 0.08 ⋅ q c ⋅ ⎜⎜ v 0 ⎟⎟ ⎜ ⎟ compression loading for h/R* > 4
⎝ p ref ⎠ ⎝ R *⎠
0.05
⎛ σ' ⎞
Fugro-05 2
τ f = 0.08 ⋅ q c ⋅ ⎜⎜ v 0 ⎟⎟ (4)−0.90 ⎛⎜ h ⎞
⎟ compression loading for h/R* < 4
⎝ p ref ⎠ ⎝ 4R * ⎠
0.15 −0.85
⎛ σ' ⎞ ⎡ ⎛ h ⎞⎤
τ f = 0.045 ⋅ q c ⎜⎜ v 0 ⎟⎟ ⎢ max⎜ R * ,4 ⎟⎥ tension loading
⎝ p ref ⎠ ⎣ ⎝ ⎠⎦
⎡ 0.13
⎛ σ' v 0 ⎞ ⎡ ⎛ h ⎞ ⎤
−0.38
⎤
τ f = a ⎢0.029 ⋅ b ⋅ q c ⎜⎜ ⎟⎟ ⎢ max ⎜ , 8 ⎟⎥ + Δσ' rd ⎥ tan δ f
ICP-05 ⎢⎣ ⎝ p ref ⎠ ⎣ ⎝ R * ⎠⎦ ⎥⎦
a= 0.9 for open ended piles in tension and 1.0 for all other cases
b = 0.8 for piles in tension and 1.0 for piles in compression
τ f = z L ⋅ p ref ⋅ FDr ⋅ Fsig ⋅ Ftip ⋅ Fload ⋅ Fmat ≥ τ min
FDr=2.1(Dr-0.1)1.7
Dr=0.4ln(qc1N/22); nominal relative density which may be greater than 1.0
-0.25
NGI-05 Fsig =(σ’v0/pref)
Ftip = 1.0 for driven open ended and 1.6 for driven closed ended
Fload = 1.0 for tension and 1.3 for compression
Fmat = 1.0 for steel and 1.2 for concrete
τmin = 0.1·σ'v0
ft ⎡ ⎡ ⎛h ⎞ ⎤
−0.5
⎤
τf = ⎢ ⋅ ⋅ ⎟ ⎥ + Δσ' rd ⎥ tan δ f
0.3
0 . 03 q c A r ,eff ⎢ max ⎜ , 2
fc ⎣⎢ ⎢⎣ ⎝ D ⎠ ⎥⎦ ⎦⎥
Ar,eff (effective area ratio) = 1 - IFR × ( Di / D ) 2;
UWA-05 Ar (area ratio) = 1 - ( Di / D ) 2 ;
IFR (= ΔLp / Δz, incremental increase in soil plug length over pile penetration
depth)
ft/fc = ratio of tension to compression capacity (equal to 1 for compression and 0.75
in tension)
τf = local ultimate shaft friction; δf = interface friction angle; pref = reference stress
= 100 kPa
L = pile length; z = element depth; h = height above pile tip
R* = equivalent radius = (R2-Ri2)0.5 where Ri is the internal radius
Notes
For noncircular piles, an equivalent circular is used to assess R* and D
Δσ’rd ≈ (4G /D)×Δy = change in radial stress during pile loading
G≈185qc×qc1N-0.7 = operational level of shear modulus
Δy ≈ 0.02mm = radial displacement during pile loading
2
The formulation for compression shaft friction within 4R* of the pile tip differs slightly from Kolk et
al (2005a) due to a typing error in the original paper.
3-4
The end bearing formulations of the four ‘new’ CPT-based methods to be

incorporated in API (2006), including those for UWA-05 method (which is described
later), are summarised in Table 3.1. The skin friction (τf) formulations are provided
for completeness in Table 3.2 and the reader is referred to Schneider (2007) for a
detailed discussion of driven pile shaft friction. The followings details highlight the
significant features of, and differences between the Fugro-05, ICP-05 and NGI-05
methods for end bearing.
(i) The base resistance is defined by ICP-05 and NGI-05 as the resistance
mobilised at a pile head displacement of 10% of the diameter (D). Such a
definition does not consider a pile’s elastic shortening, which can be substantial
for long piles (but was relatively insignificant for the database of pile tests
employed in the derivation of these methods). Ultimate base resistance
recommended in Fugro-05 and in the 21st edition of API recommendation is
mobilised at a pile tip displacement of 0.1D, and is referred to here as qb0.1.
(ii) The design cone resistance, qc,avg, recommended by Fugro-05 and ICP-05 is the
arithmetic average qc value measured over the interval of 1.5D above to 1.5D
below the pile tip; this averaging technique is employed for both closed- and
open-ended piles, despite the fact that the influence zone in the vicinity of a
typical offshore driven pipe pile (which is usually almost fully coring during
driving) is less than that of a closed-ended pile with the same diameter. NGI-05
relates the pile base resistance with qc measured at pile tip level qc,tip, but
involves a degree of subjectivity in the assessment of a ‘representative’ qc,tip
values in variable strata.
(iii) Fugro-05 expresses base resistance as a function of the square root of qc,avg and
the area ratio Ar (=1-(Di/D)2) raised to the power of 0.25, where Di is the
internal pipe diameter. As Ar is unity for a closed-ended pile, the method
implies that the qb0.1/qc,avg ratio could be greater than unity in relatively loose
sand with qc,avg less than 7.2 MPa. The sand plug is assumed to remains
stationary during static loading for open-ended pile, provided that the pile is
driven to a sufficient depth (≈5D) within the end bearing stratum (Lehane &
Randolph 2002).
3-5
(iv) For closed-ended piles, ICP-05 proposes that the qb0.1/qc,avg ratio varies with the
pile diameter, reducing by a factor of 3 as D increases from 0.1m to 0.9m. The
existence of this ‘scale effect’ is contentious. For example, White & Bolton
(2005) argue that an apparent diameter dependence of this ratio arises because
of the greater zone of influence around a large diameter pile and the consequent
need for a suitable averaging technique for resistance data measured with a
36mm diameter cone. For open-ended piles, ICP-05 assumes implicitly that a
pile plug can fail under static loading and that the occurrence of plug failure can
be assessed using an empirical relationship between pile diameter and the
relative density of the sand (Dr). The ICP-05 formulations are such they can
lead, in certain instances, to the inference of a greater base capacity for a pile
which is adjudged (on the basis of the specified relationship between D and Dr)
to have a failing plug than a pile which is assessed to have a rigid basal plug.
(v) qb0.1 values given by NGI-05 for closed-ended and plugged open-ended piles are
obtained by factoring the cone end resistance at pile tip level (qc,tip) by a relative
density correction coefficient FDr; this adjustment leads to qb0.1/qc,tip ratios that
reduce with increasing relative density of the sand. For open-ended piles, the
ultimate capacity is the minimum of the calculated “unplugged” and “plugged”
capacities. The ‘unplugged’ qb0.1 value is generally significantly greater than the
“plugged” qb0.1 value due to the assumptions made regarding the contribution of
the inner wall shear stress, which is taken to be 3 times the outer shaft friction
along the entire sand plug length. This assumption is not, however, consistent
with measured internal shaft friction distributions on model pipe piles (e.g.
O'Neill & Raines 1991, Lehane & Gavin 2001).
(vi) The three methods predict a qb0.1/qc,avg ratio of less than unity (apart for Fugro-
05 when qc values are low, as discussed above). White & Bolton (2005), and
others, have argued that this may be expected due to partial end bearing
mobilization in a pile test (involving a displacement of only 0.1D compared to
larger ‘steady state’ displacements in a CPT), and to effects arising from partial
embedment and local inhomogeneities (i.e. strong local variations in qc making
the assessment of an appropriate qc,avg difficult).
3-6
3.3 OVERVIEW OF THE UWA DRIVEN PILE DATABASE
A database of load tests on driven piles in sand has been compiled by Prof. Barry
Lehane, Mr. James Schneider and the author at the University of Western Australia
(UWA). As summarised in Table 3.3, the full UWA database contains 49
compression piles and 28 tension piles. More details of the full database are given in
Appendix A. A component of this database comprises (good quality) base capacity
measurements on driven piles with CPT data in siliceous sand. Table 3.4 and Table
3.5 summarise significant details of the extended base capacity databases, which
comprises 20 full-scale static load tests at 10 sites for closed-ended driven piles and
13 full-scale static load tests at 7 sites for open-ended driven piles.
Table 3.3 Summary of the UWA full database characteristics
Closed-ended Open-ended All Piles

Number of piles 44 33 77
Steel 23 33 56
Concrete 21 - 21
Compression 32 17 49
Tension 12 16 28
One of the difficulties in the interpretation of end bearing from static load tests arises
due to the often uncertain residual stresses existing at the pile base prior to testing;
such stresses can be relatively large for closed-ended driven piles (and even larger
for jacked piles; see Chow 1997). Most of the closed-ended piles in the UWA
database were instrumented with strain gauges or a tip load cell, enabling a
reasonable determination of the residual base stresses (qb,residual). Table 3.4 lists the
measured/assumed qb,residual values and includes comments on how these stresses
were estimated. Values of qb,residual are typically between 5% and 25% of the qc value
in the vicinity of the pile tip. Corrections to allow for different definitions of ultimate
capacity (i.e. at a head or tip displacement of 10%D) were not significant because the
average length of the closed-ended piles in the database was only about 13m.
For open-ended piles, measurements of inner and outer skin friction distributions, as
well as theoretical studies reported by Randolph et al. (1991), have indicated that the
internal shear stresses are highest near the tip of the pile and decay exponentially
with distance above the pile tip. Internal skin friction, as a consequence, is very small
beyond a height (h) of two diameters (Di) above the pile tip. This trend coupled with
3-7
the general shortage of reliable strain gauge data very close to the tips of pipe piles
and the associated interpretation difficulties (because of the need to assess the
contribution to the strain/load measurement of internal and external friction),
prompted the author to infer base capacities as the load in the pile wall at a distance
h=2Di above the pile toe (based on the axial load distribution at 10% of pile tip
movement) minus the estimated external shaft friction which was assumed equal to
1% of qc. It was found, in general, that there was good agreement between the
ultimate base capacities inferred in this study (i.e. those provided in Tables 3.4 & 3.5)
and those presented for the same piles by Fugro (2003) and Chow (1997).
Table 3.4 UWA base capacity database of closed-ended driven piles in siliceous sand
Site name; D L qb0.1 qb,residual qc,tip qc,avg (MPa) σ'v0 c
Ref
test No. m) (m) (MPa) (MPa) (MPa) ±1.5D Dutch (kPa)
Akasaka; 6C 0.2 11.0 15.18 1.20 27.5 29.9 25.7 147.4 BCP (1971)
a
Baghdad; P1 0.285 11.0 5.14 2.98 6.5 6.0 4.7 151.9 Altaee (1992a)
a
Baghdad; P2 0.285 15.0 7.10 4.70 6.2 7.1 6.1 189.3 Altaee (1992a)
Drammen; A 0.28 8.0 1.14 0.34 3.0 2.9 2.6 89.5 Gregersen (1973)
Drammen; D/A 0.28 16.0 1.79 1.09 5.0 5.2 4.4 177.2 Gregersen (1973)
b
Hoogzand; II 0.356 6.8 13.96 0.35 29.6 28.6 26.3 97.5 Beringen (1979)
Hsin Ta; TP4 0.609 34.3 3.11 0.89 7.7 9 6.4 310.5 Yen (1989)
Hunter's P; S 0.273 9.2 4.97 1.04 8.6 8.3 6.6 99.6 Briaud (1989)
Kallo; I 0.908 9.7 8.96 0 27.9 27.3 14.7 108.9 De Beer (1979)
Kallo; II 0.539 9.7 10.69 0 28.1 29.3 18.8 109.1 De Beer (1979)
Kallo; III 0.615 9.8 9.73 0 29.2 30.1 16.3 110.2 De Beer (1979)
Kallo; IV 0.815 9.8 9.22 0 28.9 29.0 15.1 110.0 De Beer (1979)
Kallo; V 0.406 9.3 10.74 0 25.9 25.0 16.6 105.3 De Beer (1979)
Kallo; VII 0.609 9.4 8.55 0 26.5 24.9 16.0 105.7 De Beer (1979)
b
Ogeechee; H-12 0.457 6.1 10.66 0.31 10.9 12.2 8.8 74.2 Vesic (1970)
b
Ogeechee; H-13 0.457 8.9 13.16 0.48 15.1 14.8 10.9 103.6 Vesic (1970)
b
Ogeechee; H-14 0.457 12 13.46 0.66 6.7 13.8 10.5 135.6 Vesic (1970)
b
Ogeechee; H-15 0.457 15 16.28 0.85 15.1 15.2 13.1 166.2 Vesic (1970)
Pigeon R; CEP 0.356 6.9 10.96 2.26 18.2 18.1 16.4 89.6 Paik (2003)
b
Sermide; 0.508 35.9 10.17 1.23 16.4 16.3 14.5 315.0 Appendino (1981)
a
: quoted values of qb,residual are those estimated in the original reference; b: quoted values of qb,residual are those
estimated following the procedure proposed by Briaud & Tucker (1984); c: effective vertical stress at pile tip
3-8
Table 3.5 UWA base capacity database of open-ended driven piles in sand
Site name D t L FFR Ar Ar,eff σ'v0d qb0.1a qb,plugb qc,tip qc,avg (MPa)
Ref
test No. m mm m - - - kPa MPa MPa MPa ±1.5D Dutch
Dunkirk Brucy
0.324 19.1 11.3 0.45 0.22 0.65 141 7.3 6.0 26.7 24 21.0
CSa (1991)
Dunkirk Brucy
0.324 12.7 11.3 0.48 0.15 0.59 141 6.6 5.7 26.7 24 21.8
CLa (1991)
Euripides Fugro
0.763 35.6 30.5 0.99 0.18 0.18 320 12.3 7.9 61.5 60.8 55.0
I30c (2003)
Euripides Fugro
0.763 35.6 38.7 0.90 0.18 0.26 403 9.9 5.9 50.8 50.8 47.3
I38c (2003)
Euripides Fugro
0.763 35.6 47.0 0.89 0.18 0.26 488 15.3 11.7 65.9 66.4 53.1
I47c (2003)
Euripides Fugro
0.763 35.6 46.7 0.82 0.18 0.33 477 16.0 12.5 63.3 63.1 53.4
II47c (2003)
Hoogzand c Beringen
0.356 16 7.0 0.66 0.17 0.45 100 12.2 10.2 28.7 28.9 28.1
I (1979)
Hoogzand c Beringen
0.356 20 5.3 0.77 0.21 0.39 82 11.2 7.1 33.8 33.2 29.5
IIIA (1979)
Pigeon R Paik
0.356 32 7.0 0.8 0.33 0.46 94 8.9 5.6 19.5 19.7 18.4
OEP (2003)
Rastanajib Fugro
0.763 38.5 25.0 1.13 0.19 0.19 306 20.7 14.2 92.4 84.9 79.6
25a (2003)
Shanghai c Pump
0.914 20 79.0 0.80 0.09 0.27 637 5.9 5.2 21.5 21.5 21.5
ST1 (1998)
Shanghai c Pump
0.914 20 79.1 0.85 0.09 0.22 637 5.1 4.3 21.5 21.5 21.5
ST2 (1998)
Tokyo c Shioi
2.0 43 30.6 1.08 0.08 0.08 275 2.0 1.6 30.4 23.9 9.7
TP (1992)
a
: at 0.1D pile tip displacement, and inferred from axial pile stress at h=2Di minus external shaft friction from h=0
to h=2Di with τf=0.01qc; b: inferred from qb0.1 by assuming qb,ann=0.6qc,avg (Dutch); c: no measurement of IFR with
FFR assumed equal to PLR; d: effective vertical stress at pile tip
To assist understanding, some background information (especially those related to

degree of soil displacement) on the base capacity database is now summarized:
(i) The closed-ended pile database includes piles with embedded depths ranging
from about 6m to 36m and diameters varying from 0.2m to 0.9m. The lengths
and diameters of the pipe pile database range from 5m to 79m and from 0.324m
to 2m respectively.
(ii) Only two sites (Pigeon River and Shanghai) have not been included in previous
databases studies associated with the Fugro-05, ICP-05 and NGI-05 design
methods, indicating a dearth of high quality data available in the public domain.
Furthermore, while sand sites are naturally variable, only one ‘representative’
CPT profile is reported by the authors for each of the various case histories at 9
of the 17 test sites. The average of the qc profiles in the vicinity of the test pile
was employed for the cases reporting multiple CPTs.
3-9
(iii) The soil plug length was measured continuously during pile driving for eight
pile tests in the database and the corresponding incremental filing ratios (IFR)
are plotted in Figure 3.1. The FFRs were derived for the database piles by
averaging the IFRs (shown on Figure 3.1) measured over the final three
diameters of pile installation. FFR values were taken equivalent to PLR (=final
plug length over pile penetration depth) values in the absence of IFR data.
(iv) At Dunkirk, the first time compression tests (following tension tests) on two
piles, CSa and CLa, are included in the database. Both of these had strain
gauges attached to their outside walls and pile CSa had a 6.5mm thick inner
shoe. Final filling ratios (FFRs) and plug length ratios PLRs (i.e. ratio of plug
length to pile length) were about 0.5 and 0.55 respectively, indicating a
relatively high level of plugging during pile installation. Initial tension tests
performed on the piles are likely to have influenced the assessment of qb0.1/qc,avg.
(v) The EURIPIDES test programme involved a highly instrumented pile (Kolk et
al. 2005c), driven at one location and then load tested at three depths. The pile
was then extracted and driven at an adjacent location to the deepest penetration
achieved in the first installation. The FFRs were in the range of 0.8 to 1.0, while
PLRs were about unity.
(vi) For tests at the Hoogzand site, the strain gauge instrumented pile no. I and un-
instrumented pile no. IIIA (which had a 4mm thick internal shoe) were driven
open-ended, with only plug length at final pile penetration being measured.
Tension tests were performed after compression tests. For pile no. IIIA, the base
resistance was estimated by assuming that the tension shaft capacity was 75% of
the shaft capacity in compression. It should be noted that another pile IIIB,
which was identical to pile IIIA except that the soil plug inside the pile was fixed
by grouting, gave exactly same static capacity as that of IIIA.
(vii) At Pigeon Creek river in Indiana, a comprehensively instrumented, double-
walled, pile system was used for the first time at field scale to separate the
resistance provided by the annulus (qb,ann) and by the plug (qb,plug). The FFR and
PLR were both approximately equal to 0.8.
(viii) The load test performed on a fully instrumented 0.76m diameter pipe pile with
a pile tip depth of 25m at the Ras Tanajib site (Kolk et al. 2005b) was included
3-10
in the database. The PLR was 0.81, which was due to a high level of plugging
during initial driving from 17m to 19m penetration but the FFR was about unity.
(ix) In Shanghai, two 79m long pipe piles were instrumented with strain gauges at 9
levels along the pile shaft and driven through soft clay (~30m) to a dense sand
layer. Plug length ratios (PLR) were around 80% to 85% at final pile
penetration, with no IFR during installation being reported.
(x) In Trans-Tokyo Bay, a 2m diameter pipe pile (fitted with a 9mm thick, 300mm
long external driving shoe) was driven to a depth of 30.6m and load tested to
plunging failure. This ‘plunging failure’ was attributed to the presence of a
relatively soft clay layer close to its tip. The pile was not plugged during driving
– with the final plug level being 2.4m above the sea bed after driving. No strain
gauge instrumentation data along the pile shaft was reported and therefore the
value qb0.1 given in Table 3.5 is that quoted in Shioi (1992).
30
Dunkirk CLs
27 Dunkirk CSa
EURIPIDES I30c
EURIPIDES I38c
24 EURIPIDES I47c
EURIPIDES II47c
Pigeon R OEP
21 Ras Tanajib
18
h/D i (-)
15
12
0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
IFR (-)
Figure 3.1. Examples of measured IFR plotted against normalised distance above pile tip (h/Di) during
pile driving in UWA database (Table 3.5)
3-11
3.4 EVALUATION OF FUGRO-05, ICP-05 & NGI-05 METHODS
3.4.1 Using full UWA database
The full UWA database contains more pile load tests than the base capacity database
(only total capacities were available for many of the tests). The database does,
however, provide a good opportunity to compare the relative predictive performance
of Fugro-05, ICP-05 and NGI-05 for end bearing. It should be noted that for the
purpose of overall evaluation of methods performance (for both base and shaft),
several pile load tests (i.e. in sites Kallo, Salt Lake Valley and Ras Tanajib) in the
base capacity database are not included in the full UWA database due to the presence
of a significant amount of clay or cemented sand along the pile shaft. In the full
UWA database, there are 32 closed-ended piles (CEPs) and 17 open-ended piles
(OEPs) loaded in compression, and 12 CEPs and 16 OEPs loaded in tension. Table
3.6 summarise the predictive performance of the methods in terms of the average and
coefficient of variation (COV) of the calculated to measured total capacities (Qc/Qm).
It can be observed from the Table that although Fugro-05 give reasonable good
predictions for both CEPs and OEPs in tension but the compressive capacities are
overestimated by 15% to 25%, indicating a tendency of this method to over-predict
end bearing. To further investigate such tendencies, the predicted base resistance
ratio (qb0.1/qc,avg or qb0.1/qc,tip) for the compression piles (i.e. 49 piles) in the UWA full
database by each method are compared.
The predicted ratios of qb0.1/qc,avg or qb0.1/qc,tip are plotted on Figure 3.2 against the
key controlling parameters of qc,avg, D, and Dr,tip for Fugro-05, ICP-05 and NGI-05
respectively. The predicted ratios of qb0.1/qc,avg are indicated by the dots, while the
design formulations as outlined in Table 3.1 are represented by curves. It is expected
to see a perfect match between the dots (dark dots for CEPs; light dots for OEPs) and
the curves (dark curve for CEPs; light curve for OEPs), except for some cases where
special design rules of the method apply, i.e. ICP-05 specify qb0.1 of an “unplugged”
pipe pile depends on the area ratio Ar rather than the absolute outer diameter D. In
these Figures, it is intended to show the distribution of the predicted ratios (qb0.1/qc,avg
or qb0.1/qc,tip) with the controlling parameter of each method.
3-12
As shown in Figure 3.2a, the predicted ratios of qb0.1/qc,avg for Fugro-05 exceeded
unity in a number of cases (i.e. ranging from 0.42 to 1.77) for closed-ended piles.
This is consistent with the previous discussions that Fugro-05 could overestimate
base resistance in relatively loose sand (or qc,avg < 7.2 MPa). The predicted ratios (i.e.
ranging from 0.21 to 0.67) for open-ended piles fall into the two light curves which
represent the minimum and maximum area ratios in the UWA full database.
Table 3.6 Assessment of methods performance against UWA full database
Database Method Average COV Number of piles

Fugro-05 1.24 0.39
CEP ICP-05 0.99 0.33
32
Compression NGI-05 1.16 0.40
UWA-05 0.98 0.33
Fugro-05 0.97 0.41
CEP ICP-05 1.02 0.30
12
Tension NGI-05 1.27 0.50
UWA-05 1.00 0.29
Fugro-05 1.14 0.30
OEP ICP-05 0.89 0.28
17
Compression NGI-05 1.01 0.25
UWA-05 0.98 0.19
Fugro-05 0.90 0.32
OEP ICP-05 0.90 0.27
16
Tension NGI-05 1.01 0.35
UWA-05 0.91 0.23
Fugro-05 1.11 0.38
ICP-05 0.95 0.30
Entire database 77
NGI-05 1.11 0.37
UWA-05 0.97 0.27
Figure 3.2b presents the design formulations for ICP-05 with a straight dashed line
for square closed-ended piles, dark curve for circular close-ended piles, and light
curve for plugged open-ended piles. Some light dots not in agreement with the curve
were predictions for unplugged open-ended piles, whose qb0.1 is solely determined by
pile area ratio Ar instead of diameter D. As discussed earlier, in certain instances,
ICP-05 can lead to the inference of a greater base capacity for unplugged pile than
plugged pile. There are two such examples at sites Pigeon River and Euripides
indicated by the two dots above the light curve, where the piles were adjusted to be
plugged under static loading, but the base capacities were calculated based on
recommendations for unplugged piles. The predicted ratios of qb0.1/qc,avg range from
0.37 to 0.7 for closed-ended piles and 0.06 to 0.33 for open-ended piles.
3-13
As noted on Figure 3.2c, the predicted ratios of qb0.1/qc,tip by NGI-05 reduces with an
increase in Dr,tip from 0.8 to 0.34 for closed-ended piles and 0.66 to 0.12 for open-
ended piles. Only one pile test (at the Drammen site - as indicated by one light dot
off the dashed curve) was adjudged by this method to have a failing soil plug under
static loading.
Table 3.7: Averaged predicted ratios (qb0.1/qc,avg or qb0.1/qc,tip) in UWA full compression database
Characteristics CEPs OEPs

Number of piles 32 17
Averaged Fugro-05 0.90 0.33

qb0.1/qc,avg ICP-05 0.54 0.16
or NGI-05 0.62 0.26
qb0.1/qc,tip
UWA-05 0.6 0.26
2.0
Fugro-05
1.8
1.6 0.5 0.25

q b0.1 /q c,avg = 8.5 × (p a /q c,avg ) × Ar
1.4 qc,avg: averaged ±1.5D above and below tip
Ar = 1 for closed-ended pile (CEP)
q b0.1 /q c,avg (-)
1.2 2
Ar = 1 - (Di/D) for open-ended pile (OEP)
1.0
0.8 32 CEPs
17 OEPs
0.6
0.4
0.2
(a)
0.0
0 10 20 30 40 50 60 70
qc,avg (MPa)
3-14
0.9
ICP-05
0.8
0.7
CEP : qb0.1/qc,avg = max { 1.0-0.50×log(D/D CPT), 0.3 }
0.6 P OEP: qb0.1/qc,avg = max { 0.5-0.25×log(D/DCPT), 0.15, Ar }
q b0.1 /q c,avg (-)
U OEP: qb0.1/qc,avg = Ar
0.5 if Di < 0.02×(Dr -30), pile "Plugged"
otherwise "Unplugged"
0.4 DCPT = 36mm; qc,avg: ±1.5D
0.3 Pigeon R
32 CEPs
Euripides
0.2 17 OEPs
0.1
(b)
0.0
0 0.3 0.6 0.9 1.2 1.5 1.8 2.1
D (m)
0.9
NGI-05
0.8
2
CEP : qb0.1/qc,tip = 0.8/(1+Dr )
0.7
2
OEP : qb0.1/qc,tip = 0.7/(1+3Dr )
0.5
0.6 Dr = 0.4×ln { qc,tip/22/(σ’v0pa) }
q b0.1 /q c,tip (-)
0.5 32 CEPs
17 OEPs
0.4
0.3
Drammen
0.2
0.1
(c)
0.0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
Dr,tip (-)
Figure 3.2 Predicted ratios of qb0.1/qc,avg or qb0.1/qc,tip in the full UWA compression database: (a) Fugro-
05, (b) ICP-05, and (c) NGI-05
3-15
The average of the predicted ratios of qb0.1/qc,avg are summarized in Table 3.7 for each
method. The results from UWA-05 (as discussed later) are also included for
comparative purpose. It is obvious from the Table that the ratios predicted by Fugro-
05 are significantly higher than those predicted by other methods. This is also the
main reason why Fugro-05 tends to overpredict the pile total capacities in the UWA
compression database (as indicated by Table 3.6). ICP-05 tends to be quite
conservative for open-ended piles, with the predicted ratio only half of that predicted
by other methods.
3.4.2 Using UWA base capacity database
The predictive performance of the Fugro-05, ICP-05 and NGI-05 methods (outlined
in Table 3.1) was further examined using the UWA base capacity databases (as
summarised in Tables 3.4 & 3.5). The ratios of calculated to measured database qb0.1
values are plotted against diameter (D) on Figure 3.3 for closed-ended piles and
Figure 3.4 for open-ended piles. The average and coefficient of variation (COV) of
the ratio between calculated and measured qb0.1 values are included in these Figures
(also in Table 3.8) and may be taken as measures of each method’s predictive ability.
It would appear from Figure 3.3 that all three methods do not perform well for
closed-ended piles in loose sand at Drammen, Norway. The Fugro-05 and NGI-05
methods also tend to over-predict the base capacity at the Kallo site, where the piles
were driven through soft clay to bear on underlying sand. As will be discussed later,
this overestimation could be attributed to the use of an inappropriate qc averaging
technique for those methods. Although employing the same value of qc,avg as Fugro-
05, the base resistances at Kallo are well predicted by ICP-05 because of this
method’s proposed diameter dependence of end bearing.
The predictions by the three methods for open-ended piles in the UWA database are
summarized in Figure 3.4. Apart from over-predictions for the 2m diameter pile at
Tokyo Bay, both Fugro-05 and NGI-05 provides reasonable predictions for piles in
this database. ICP-05 has the lowest COV but under-predicts the capacity of all but
one pile and under-predicts the base capacity of the pile at Shanghai by a factor of 3.
3-16
In summary, it is evident that despite the wide differences between the formulations
for qb0.1 employed by these three methods, their predictive performances are broadly
similar. This is perhaps not surprising given that the methods were ‘tailored’ for very
similar databases to that employed here. It follows that at least two of the three
respective formulations contains compensating errors and that extrapolation outside
of the database may lead to significant errors. Furthermore, none of the methods
allows for the varying effects of soil displacement (FFR) shown by each of the
database piles (as summarized in Table 3.5) and all use a somewhat arbitrary means
to select qc,avg for layered deposits.
Table 3.8: Summary of statistics for ratios of calculated to measured base resistance (qb0.1, C/qb0.1,M)
Method Recommended Averaging; qc,avg Dutch Averaging; qc,avg [Dutch]

Method Closed-ended Open-ended Closed-ended Open-ended
average cov average cov average cov average Cov
Fugro-05 1.49 0.59 1.25 0.59 1.30 0.64 1.11 0.37
ICP-05 0.92 0.36 0.71 0.30 0.69 0.42 0.60 0.28
NGI-05 1.05 0.47 0.98 0.83 0.88 0.47 0.87 0.66
UWA-05 - 0.89 0.34 1.02 0.19
2.00
Drammen Hsin Ta Fugro-05
qb0.1,C/qb0.1, M=3.99&3.42; qb0.1,C/qb0.1, M= 2.59 average=1.49
1.75
cov=0.59
1.50 Akasaka
q b0.1, C [Fugro-05] /q b0.1, M
Baghdad
1.25 Drammen
Hoogzand
1.00
Hsin Ta
0.75 Hunter's P
Kallo
0.50
Ogeechee
0.25 Pigeon R
(a) Sermide
0.00
0.0 0.2 0.4 0.6 0.8 1.0 1.2
D (m)
3-17
2.00
ICP-05
average=0.92
1.75
cov=0.36
1.50 Akasaka
q b0.1, C [ICP-05] /q b0.1, M
Baghdad
1.25 Drammen
Hoogzand
1.00
Hsin Ta
0.75 Hunter's P
Kallo
0.50
Ogeechee
0.25 Pigeon R
(b) Sermide
0.00
0.0 0.2 0.4 0.6 0.8 1.0 1.2
D (m)
2.00
NGI-05
Drammen average=1.05
1.75 qb0.1,C/qb0.1, M=2.06&2.12; cov=0.47
1.50 Akasaka
q b0.1, C [NGI-05] /q b0.1, M
Baghdad
1.25 Drammen
Hoogzand
1.00
Hsin Ta
0.75 Hunter's P
Kallo
0.50
Ogeechee
0.25 Pigeon R
(c) Sermide
0.00
0.0 0.2 0.4 0.6 0.8 1.0 1.2
D (m)
Figure 3.3. Ratios of calculated to measured base resistance (qb0.1,C/qb0.1,M) for closed-ended piles, (a)
Fugro-05, (b) ICP-05 and (c) NGI-05
3-18
2.00
1.75 Tokyo
qb0.1,C/qb0.1, M=3.6
1.50
q b0.1, C [Fugro-05] /q b0.1, M
1.25
Fugro-05
average=1.25
1.00 cov=0.59
0.75 Dunkirk
Euripides
Hoogzand
0.50
Pigeon R
Ras Tanajib
0.25 Shanghai
(a) Tokyo
0.00
0.0 0.3 0.6 0.9 1.2 1.5 1.8 2.1
D (m)
2.00
1.75
1.50
q b0.1, C [ICP-05] /q b0.1, M
1.25
ICP-05
average=0.71
1.00 cov=0.30
0.75 Dunkirk
Euripides
Hoogzand
0.50
Pigeon R
Ras Tanajib
0.25 Shanghai
(b) Tokyo
0.00
0.0 0.3 0.6 0.9 1.2 1.5 1.8 2.1
D (m)
3-19
2.00
Tokyo
1.75 qb0.1,C/qb0.1, M=3.4
1.50
q b0.1, C [NGI-05] /q b0.1, M
1.25
NGI-05
average=0.98
1.00
cov=0.83
0.75 Dunkirk
Euripides
Hoogzand
0.50
Pigeon R
Ras Tanajib
0.25 Shanghai
(c) Tokyo
0.00
0.0 0.3 0.6 0.9 1.2 1.5 1.8 2.1
D (m)
Figure 3.4. Ratios of calculated to measured base resistance (qb0.1,C/qb0.1,M) for open-ended piles, (a)
Fugro-05, (b) ICP-05 and (c) NGI-05
3.5 qc AVERAGING TECHNIQUE
The importance of the selection of qc,avg in CPT-based methods for pile end bearing
prompted a review, summarised here, of three popular averaging techniques. These
techniques involve the derivation of qc,avg using (i) a simple arithmetic average over a
zone extending from 1.5D above to 1.5D below the pile tip (e.g. as recommended for
the Fugro-05 and ICP-05 design methods), (ii) the geometric mean of qc measured
over a given zone in the pile tip vicinity, as recommended by Eslami & Fellenius
(1997), and (iii) the ‘Dutch’ qc averaging technique commonly employed in North
American and European practice. Eslami & Fellenius (1997) suggest zones of
influence extending from 4D below to 8D above a pile toe when a pile is installed
through a loose layer into a dense layer, but from 4D below to 2D above the toe
when a pile is installed through dense layer into a loose layer. To avoid arbitrary
decisions being made concerning the thickness that constitutes a given layer, the
qc,avg geometric mean value in the following is taken as the minimum of the values
obtained using the two influence zones. In the Dutch method (Schmertmann 1978, de
Ruiter & Beringen 1979), the design cone resistances are averaged over an influence
3-20
zone extending from 8D above, and between 0.7D and 4D below the pile toe
following a ‘minimum path rule’ (as shown in Figure 2.4).
The qc profiles for all case histories in the database were digitised and a computer
code was employed to examine the implications of these and other averaging
techniques. The averaged qc values (qc,avg) are plotted against the cone resistance at
pile tip level, qc,tip for each method on Figure 3.5 for both closed and open-ended
piles.
As shown on Figure 3.5a, arithmetic averaged qc values over ±1.5D are very close to
the cone resistances at the pile tip qc,tip, with averaged ratios of qc,avg/qc,tip equal to
1.03. It is obvious that this method does not reflect the strongly layered stratigraphies
at sites such as Kallo and Tokyo Bay; such layering is likely to be significant for
piles, which have a much larger diameter than a cone penetrometer. In contrast, as
seen on Figure 3.5b and Figure 3.5c, the use of the geometric mean and Dutch
approaches lead to qc,avg/qc,tip ratios significantly different to unity at ‘non-
homogeneous’ sites, with qc,avg/qc,tip being as low as 0.2 at the Kallo site when the
geometric average is employed. In the relatively homogeneous deposits (i.e.
excluding the Kallo and Tokyo Bay sites), the average qc,avg/qc,tip ratio is 1.0 and 0.9
for the geometric mean and Dutch approaches respectively.
The Kallo case history provides an opportunity to investigate the suitability of

averaging techniques as it includes the base resistance measured during continuous
jacked installation of a 250mm diameter cone. The base resistance measured with
this cone and that of a standard 36mm diameter cone are compared on Figure 3.6a
with the base resistance of the 250mm diameter ‘cone’ predicted using the three
averaging techniques applied to the qc data. A second comparison of pile end bearing
with averaged qc values is shown in Figure 3.6b for an instrumented 350mm square
concrete jacked pile in Western Australia, as reported by Lehane et al. (2003). The
base load of this pile was recorded continuously during pile jacking as the pile
moved from a soft clay layer into a 6m thick layer sand and then back into a soft-firm
clay.
3-21
100
(a)
90
Arithmetic average
80
over ±1.5D Akas aka
Baghdad
70
Dram m en
Hoogzand
qc,avg (MPa)
60 Hs in Ta
Hunter's P
50 Kallo
Ogeechee
40 Pigeon R
Serm ide
30 Dunkirk
Euripides
20 Pigeon R
Ras Tanajib
10 Shanghai
Tokyo
0
0 10 20 30 40 50 60 70 80 90 100
qc,tip (MPa)
100
(b)
90
80 Geometric average
Akas aka
70 Baghdad
Dram m en
Hoogzand
qc,avg (MPa)
60 Hs in Ta
Hunter's P
50 Kallo
Ogeechee
40 Pigeon R
Serm ide
30 Dunkirk
Euripides
20 Pigeon R
Ras Tanajib
10 Shanghai
Tokyo
0
0 10 20 30 40 50 60 70 80 90 100
qc,tip (MPa)
3-22
100
(c)
90
80 D utch method
Akas aka
70 Baghdad
Dram m en
qc,avg (MPa)
60 Hoogzand
Hs in Ta
50 Hunter's P
Kallo
Ogeechee
40 Pigeon R
Serm ide
30 Dunkirk
Euripides
20 Pigeon R
Ras Tanajib
10 Shanghai
Tokyo
0
0 10 20 30 40 50 60 70 80 90 100
qc,tip (MPa)
Figure 3.5. Comparisons of qc,avg with qc,tip: (a) Arithmetic average, (b) Geometric average, and (c)
Dutch method
On inspection of the comparisons on Figure 3.6, it appears that, of the three

averaging techniques considered, qc,avg derived using the Dutch method provides a
better match to the general shape and magnitude of the pile end bearing profile. This
match with ‘steady state resistance’ suggests that the static pile end bearing, qb0.1, is
more uniquely related to the Dutch value of qc,avg; this averaging technique is
therefore employed in the development of the UWA-05 method, described below.
As the use of qc,avg values obtained by the arithmetic averaging technique over ±1.5D
may be misrepresentative in case of strong soil stratigraphy change close to the pile
toe, it is considered more appropriate to compare the methods predictive
performance by employing qc,avg values by Dutch technique. Table 3.8 also
summarises the average and COVs of the ratios of calculated to measured end-
bearing resistances when using qc,avg values by Dutch technique. Compared with
statistical values based on qc averaging techniques recommended by each method,
there are slight improvements for Fugro-05 and NGI-05 for open-ended piles,
primarily due to the improved method performance for the pile at Tokyo Bay. The
ratio of calculated to measured capacity reduces significantly for ICP-05 due to the
contentious diameter effect for closed-ended piles and the static plugging criteria for
3-23
open-ended piles. For the 2m diameter pile at Tokyo Bay, the ratio of predicted to
measured capacities using ICP-05 reduces from 1.02 to 0.41 due to differences in
averaging technique.
qc (MPa) qc (MPa)
0 10 20 30 40 0 10 20 30 40
7 7
8 8 (a)
9 9
10 10
11 250mm cone 11 250mm cone
Depth (m)
12 12
13 13
14 14
15 15
16 16
q c,avg [Dutch]
17 17
q c,avg [±1.5D]
18 36mm cone 18
19 19
q c,avg [geometric]
20 20
q c (MPa) q c (MPa)
0 3 6 9 12 15 0 3 6 9 12 15
15 15
Jacked Pile, Jacked Pile, (b)
B=350mm B=350mm
36mm
18 18
q c,avg [±1.5D]
Depth (m)
21 21
24 24
27 27
q c,avg [Dutch]
q c,avg [geometric]
30 30
Figure 3.6. Measured and predicted qc profiles at sites, (a) Kallo, Belgium, and (b) Perth, Australia
3-24
3.6 THE UWA-05 METHOD FOR DRIVEN PILES
The foregoing review of the Fugro-05, ICP-05 and NGI-05 methods and of the qc
averaging techniques prompted the development of a new formulation for end
bearing, which now forms part of the UWA-05 method included in API (2006).
As described, the Dutch averaging technique was considered the more suitable
technique to derive qc,avg. The UWA database of qb0.1 values normalised by the Dutch
qc,avg values are presented on Figure 3.7, which compares qb0.1/qc,avg with pile
diameter and the relative density Dr (which is calculated following NGI’s
formulation in Table 3.1). It is evident that, apart from the Ogeechee River and
Baghdad sites, qb0.1/qc,avg for all test piles is relatively constant at ≈0.6 and
independent of the pile diameter and sand relative density. The qb0.1/qc,avg for the five
457mm diameter piles at Ogeechee River are greater than unity, and remain in excess
of unity even if residual loads are taken to be zero (see Table 3.4). The Baghdad
‘outlier’ appears to arise because of the qb,residual/qc,avg ratio obtained from the residual
load estimated by Altaee (1992b) is over four times the average ratio of residual load
to average cone tip resistance in the database.
The base resistance (qb0.1) of an open-ended pile comprises the annular resistance
(qb,ann) and the sand plug resistance (qb,plug). Experimental data and numerical
analysis indicate that qb,ann varies between about 0.5 and 1.0 times the CPT qc value
(e.g. Finlay et al. 2001, Lehane & Gavin 2001, Paik & Salgado 2003, Jardine et al.
2005). A constant of 0.6 for qb,ann/qc,avg is adopted here (see Equation 3.1a), which is
equivalent to the average qb0.1/qc,avg ratio indicated for closed-ended piles. Lehane &
Randolph (2002) show that the pile will remain fully plugged during static loading
(as long as the plug is longer than ≈5Di) and indicate that the compression of the soil
plug for a pile driven into siliceous sand is a relatively small component of the full
tip displacement. The sand plug base stress at a displacement of 0.1D, i.e. qb,plug
(which is much lower than the true ultimate base stress) is therefore controlled by the
stiffness of the sand below the sand plug. This stiffness is shown to be related to
degree of soil displacement in the vicinity of the tip, as measured by the final filling
ratio, FFR (Lehane & Gavin 2001). The ratio of qb,plug/qc falls between that of a full-
displacement closed-ended pile at FFR=0 (when qb,plug/qc ≈0.6) and that of a non-
3-25
displacement or bored pile when FFR=1. Recommended values of qb0.1/qc,avg at a tip

movement of 0.1D for non-displacement (bored) piles range from 0.15 to 0.23
(Ghionna et al. 1993). It is therefore proposed that qb,plug/qc,avg varies from 0.6 for
FFR=1 to 0.15 when FFR=0, with a (tentative) linear interpolation at intermediate
FFRs (Equation 3.1b).
The normalised plug resistances qb,plug/qc,avg deduced from the measured qb0.1 values
of the database piles (assuming qb,ann/qc,avg =0.6) are plotted against these FFRs on
Figure 3.8. It indicates a trend very similar to that shown by Gavin & Lehane (2005)
with qb,plug/ qc,avg reducing from 0.6 at FFR=0 to 0.15 at FFR=1. A limit of unity is
set on the maximum value of the FFR, as larger values are primarily related to
dilation within the soil plug.
q b,ann / q c,avg = 0.6 ( 3.1a )

q b,plug / q c ,avg = 0.6 − 0.45 × FFR ; FFR = minimum( FFR, 1 ) ( 3.1b )
From Equations 3.1a & 3.1b, the overall base resistance on the total area qb0.1 can be
expressed as:
q b 0.1 / q c,avg = q b,ann × A r + q b,plug × (1 − A r )

( 3.2 )
= 0.15 + 0.45 × [1 − FFR × (D i D ) ]
2
with the effective area ratio defined as Ar,eff=1-FFR×(Di/D)2, Equation 3.2 can be
simplified as:
q b 0.1 / q c,avg = 0.15 + 0.45 × A r ,eff ( 3.3a )
For typical offshore piles with FFR=1 and Ar,eff=Ar:
q b 0.1 / q c,avg = 0.15 + 0.45 × A r ( 3.3b )
Equation 3.3 provides predictions for qb0.1 for fully plugged (or closed-ended piles
with Ar,eff = Ar = 1), partially plugged and fully coring pile installations. The
continuous form of this equation is made possible by employment of the effective
area ratio term, Ar,eff. The equivalent diameter, D*, employed when assessing qc,avg
using the Dutch averaging technique is consequently given as D*=D×(Ar,eff)0.5.
3-26
1.50
(a)
1.25
Akasaka
1.00 Baghdad
q b0.1, M /q c,avg
Drammen
Hoogzand
0.75
Hsin Ta
Hunter's P
0.50
Kallo
Ogeechee
0.25
Pigeon R
Sermide
0.00
0.0 0.2 0.4 0.6 0.8 1.0 1.2
D (m)
1.50
(b)
1.25 Akasaka
Baghdad
1.00 Drammen
q b0.1, M /q c,avg
Hoogzand
Hsin Ta
0.75
Hunter's P
Kallo
0.50
Ogeechee
Pigeon R
0.25
Sermide
0.00
0 0.2 0.4 0.6 0.8 1 1.2
Dr (-)
Figure 3.7. Measured qb0.1 normalised by Dutch qc,avg for closed-ended piles plotted against, (a)
diameter, D, and (b) relative density, Dr=0.4ln(qc,avg/22/(pa×σ’v0)0.5)
The predictions for the database piles employing Equation 3.3 are shown on Figure
3.9 and may be compared directly with those of Fugro-05, ICP-05 and NGI-05 in
Table 3.8. The predictions are seen to be reasonable at most test sites, with the
method having an average of calculated to measured capacities of 0.89 and a COV of
3-27
calculated to measured capacities of 0.34 for closed-ended piles and an average of

1.02 and COV of 0.19 for open-ended piles. Under-predictions at Ogeechee River
and Baghdad may be due to site variability and interpretations of residual stress (as
previously discussed), while over-predictions at Dunkirk may be related to the
influence of load test sequence on load displacement behaviour (these piles were
loaded in tension prior to compression testing). The improvement in predictive
performance provided by this method over the other three methods considered is
marginal for closed-ended piles, but is quite significant for open ended piles,
primarily because of the incorporation of the influence of partial plugging on
performance of open ended piles.
0.7
Equation (3.1b) Dunkirk
0.6 q b,plug /q c,avg =0.6-0.45 × FFR Euripides

Hoogzand
Indiana
0.5
Ras Tanajib
Shanghai
q b,plug /q c,avg
0.4 Tokyo
0.3
0.2
0.1
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2
FFR
Figure 3.8. Ratios of qb,plug/qc,avg plotted against FFR for open-ended piles
Although the FFR may be assumed to be unity for a typical large diameter offshore
pile, its value cannot presently be predicted in advance to assist the design of smaller
diameter piles. In such cases and so that the effect of partial plugging can be roughly
incorporated in a design proposal for qb0.1, an estimate of the likely FFR may be
obtained from the trend line through existing measurements shown on Figure 3.10;
this figure includes FFR and PLR data available for the UWA base capacity database,
as well as PLR measurements from the California (UWA) Department of
Transportation.
3-28
2.00
UWA-05
(a) average=0.89
1.75
cov=0.34
1.50 Akasaka
q b0.1, C [UWA-05] /q b0.1, M
Baghdad
1.25
Drammen
Hoogzand
1.00
Hsin Ta
0.75 Hunter's P
Kallo
0.50
Ogeechee
0.25 Pigeon R
Sermide
0.00
0.0 0.2 0.4 0.6 0.8 1.0 1.2
D (m)
2.00
(b)
1.75
1.50
q b0.1, C [UWA-05] /q b0.1, M
1.25
UWA-05
1.00 average=1.02
cov=0.19
0.75 Dunkirk
Euripides
0.50 Hoogzand
Pigeon R
Ras Tanajib
0.25 Shanghai
Tokyo
0.00
0.0 0.3 0.6 0.9 1.2 1.5 1.8 2.1
D (m)
Figure 3.9. Ratios of calculated to measured base resistance qb0.1 by UWA-05, (a) closed-ended piles,
and (b) open-ended piles
3-29
1.1
1.0
0.9
0.8 FFR measured

0.7 FFR=PLR
FFR (-)
0.6
0.5
0.4 ⎧⎪ ⎡ D (m ) ⎤ 0.2 ⎫⎪
FFR ≈ min ⎨1, ⎢ i ⎥ ⎬
0.3 ⎪⎩ ⎣ 1 .5 (m )⎦ ⎪⎭
0.2
0.1
0.0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
Pile inner diameter, Di (m)
Figure 3.10 Database of FFR (including PLR) measurements (Xu et al. 2005)
3.7 JACKED PILE DATABASE
Jacked piles installed by using static force produce less noise and vibration compared
to conventional dynamic methods of displacement pile installation. They may also be
used for strengthening of existing pile foundations (Medzvieckas & Slizyte 2005).
Recent research (Deeks et al. 2005, Yetginer et al. 2006) has identified that the axial
response of a jacked pile is much stiffer that that of a driven or bored pile due to the
different stress histories applied to the soil. In particular, Yetginer et al. (2006)
attributed this high axial stiffness of a jacked pile to its pile base response: (i) the
high stiffness at the pile base due to the preloading applied during the final
installation stroke, and (ii) the presence of residual base load. Moreover, White &
Lehane (2004) concluded, based on a series of centrifuge tests, that the available skin
friction during static load test depends on the magnitudes and types of cycles
imposed by installation method. Jacked pile installation procedures involve fewer
cycles than driving, and may therefore yield higher shaft friction. However, it is also
pointed out by the authors that any additional friction on a jacked pile may degrade
faster under a cyclic working load than on a driven pile, therefore resulting in
possibly comparable shaft resistance.
3-30
Table 3.9 UWA base capacity database of closed-ended jacked piles in siliceous sand
Site name; D L QTotal,M Qb0.1,M qc,avg σ'v0 qb0.1,M/qc,avg Ref

test No m M kN kN MPa kPa -
Akasaka 0.2 11 1176.5 576 25.62 180.48 0.72 (BCP 1971)
1C
Akasaka 0.2 4 171.3 137 5.79 62.70 0.75 (BCP 1971)
6B
Dunkirk; 0.1016 7.4 308.75 90.3 12.21 114.39 0.91 (Chow 1997)
DK1
Dunkirk; 0.1016 6 246.1 87 10.00 62.29 1.07 (Chow 1997)
DK2
Labenne; 0.1016 6 95.6 36.1 4.39 80.45 1.01 (Lehane 1992)
LB2
Labenne; 0.1016 1.8 52 38 5.49 28.73 0.85 (Lehane 1992)
LB2
Kallo; 0.25 8.9 NA 618.5 14.19 101.10 0.89 (De Beer et al.
CPT250 1979)
Gold Coast 0.4 8 4500 NA 24.74 85.29 NA (Mitchell 2005)
P113
P224
Gold Coast 0.4 7.9 3380 NA 14.92 84.37 NA (Mitchell 2005)
P247
P413
P415
P586
In view of these anticipated/proven differences between a jacked and driven pile (i.e.
a jacked pile may potentially possess increased stiffness and capacity than a driven
pile in similar conditions), the four CPT-based methods (outlined in Table 3.1)
developed from a database of driven piles may not be applicable for jacked piles. In
fact, some degrees of underestimation by the methods are expected if they are
employed to predict the capacities of a database of jacked piles. To further illustrate
this, a database of pile load tests for jacked piles were compiled and analyzed. The
details of each pile load test are summarised in Table 3.9. In total, there are 13
closed-ended jacked piles from 5 sites. Some piles were equipped with tip load cells,
which allow accurate determination of the base resistance. The quoted total and base
capacity are those measured at displacement of 0.1D. Additional information on the
test site is provided below:
(i) Akasaka
The pile tests reported by BCP Committee were carried out at the site of a basement
construction in Akasak, Tokyo. The geological conditions at the site consisted of 4m
3-31
of loose sand fill overlying a thick deposit of diluvial sand (called Tokyo sand
stratum). The testing piles were instrumented with tip load cells, allowing
determination of base resistance during installation and static load tests. Loading
cycles of 10cm penetration were applied to the piles until settlement reached 5 to 9
pile diameters.
(ii) Dunkirk & Labenne
The Imperial College Pile (ICP) was used in the pile testing programmes in the dense
sand site at Dunkirk and the loose to medium dense sand site at Labenne. The ICP is
a heavily instrumented, 101.6mm diameter steel tubular closed-ended pile. Tests
performed were designed to investigate the effect of pile length and depth, direction
of loading and rate of loading. The first time compression tests are included in the
jacked pile database.
(iii) Kallo
In Kallo, standard 36mm CPT tests were conducted with an additional penetration
test involving a 250mm diameter cone. The site consisted of a thick deposit of soft
clay and peat overlying dense sand. The interface between the sand and the soft
deposit was at a depth of ~8.2m. Static load tests were conducted at a shallow
embedment (<1.6m) into the dense sand. The 250mm diameter cone was equipped
with a tip load cell to measure the base resistance, while there was no
instrumentation to measure the total load.
(iv) Gold Coast
Due to environment restrictions, jacked piles were selected as the foundation for the
Gold Coast Convention and Exhibition Centre in Queensland. Installation involved
the use of a group of hydraulic jacks to push the piles into the ground. In total, static
load tests were reported for 7 square closed-ended concrete piles (0.4m by 0.4m)
with most CPTs performed very close-by (within 0.4m to 2m of the piles). Two of
the piles had multiple CPTs (7 CPTs for P413 and 5 CPTs for P415) taken in their
vicinity. As summarised in Figure 3.11, the site is quite variable with coefficient of
variations (COVs) ranging from 30% to 40% for CPTs performed fairly close to each
other. The averaged qc profiles are employed for the analysis of pile load tests for
3-32
P413 and P415. One pile, P214 is excluded from current analysis because no close-
by CPT profile was available (the nearest CPT was 8m away from P214).
The total capacities (QTotal,C) calculated for the jacked piles by the four CPT methods
are summarised in Table 3.10 and the ratios of calculated to measured total capacity
(QTotal,C/QTotal,M) are further illustrated in Figure 3.12, including the averages and
COVs of capacity ratios QTotal,C/QTotal,M. It can be observed that:
(i) The average of QTotal,C/QTotal,M is slightly lower than unity (i.e. 0.99) for Fugro-
05, with the highest COV (=0.35) among other methods. It appears to perform
well for the jacked pile database although the method itself was developed
based on driven piles (Fugro 2003). It is also necessary to be reminded that
Fugro-05 overestimated the total capacities by 24% for the UWA database of
closed-ended driven piles in compression (Table 3.6). Therefore, the
contradictory prediction performance of Fugro-05 for the driven and jacked
database reveal the fact that the method formulation might not capture the real
mechanism controlling pile behaviour, especially for the end bearing component
(since the shaft resistance formulations are broadly similar for the four methods).
(ii) The ICP-05 method appears to provide the best predictions for the database of
jacked piles, with average of 0.98 and COV of 0.17. However it should be noted
that the 7 jacked piles of small diameters (<0.25m) at sites Akaska, Dunkirk,
Labenne and Kallo were included in the database for derivation of ICP-05 end-
bearing formulation (Chow 1997). In particular, those high base resistances (i.e.
high ratios of qb0.1,M/qc,avg) measured from the small diameter jacked piles may
produce an contentious absolute diameter effect, i.e. qb0.1/qc,avg=1-0.5log(D/DCPT)
for closed-ended piles.
(iii) The average and COV of QTotal,C/QTotal,M for NGI-05 are 0.96 and 0.17
respectively. As discussed earlier, a method developed for driven piles is
expected to underestimate the capacity when being applied for jacked piles.
Although NGI-05 seems to provide reasonable predictions for the database of
jacked piles, it overestimates the pile capacities by 16% for the UWA database
of driven closed-ended piles in compression (Table 3.6). It may therefore imply
some compensating errors.
3-33
(iv) The UWA-05 method tends to underestimate the capacities for the database of
jacked piles, with a average and COV of 0.79 and 0.17. This underestimation
(by ~20%) indicates that although the current UWA-05 provides good
predictions for closed-ended driven piles in compression (average=0.98,
COV=0.33, see Table 3.6), it is not applicable for jacked piles. It requires
improvement/modification before potentially being employed for jacked piles.
Table 3.10 Calculated total capacities (QTotal,C) by Fugro-05, ICP-05, NGI-05 and UWA-05
QTotal,C (kN) QTotal,C/QTotal,M (-)

Site name; Fugro- UWA- Fugro- UWA-
test No 05 ICP-05 NGI-05 05 05 ICP-05 NGI-05 05
Akaska; 1C 938 1095 1141 1051 0.80 0.93 0.97 0.89
Akaska; 6B 303 215 205 189 1.77 1.25 1.20 1.10
Dunkirk; DK1 174 251 297 218 0.56 0.81 0.96 0.71
Dunkirk; DK2 154 199 283 174 0.62 0.81 1.15 0.71
Labenne; LB1 71 83 61 78 0.74 0.87 0.64 0.81
Labenne; LB2 74 56 58 47 1.42 1.08 1.11 0.91
Kallo; CPT250 805 752 689 647 NA NA NA NA
GC; P113 3970 4440 3583 3458 0.88 0.99 0.80 0.77
GC; P224 4612 4647 4518 2959 1.12 1.13 1.10 0.72
GC; P247 2986 3373 2808 2103 0.88 1.00 0.83 0.62
GC; P413 2872 2922 2375 2115 1.19 1.21 0.98 0.87
GC; P415 2708 2499 2419 2056 1.02 0.94 0.91 0.78
GC; P586 2917 2551 2931 2194 0.85 0.74 0.85 0.64
Statistics average of QTotal,C/QTotal,M= 0.99 0.98 0.96 0.79
COV of QTotal,C/QTotal,M= 0.35 0.17 0.17 0.17
3-34
q c (MPa) COV (%)

0 5 10 15 20 25 0 20 40 60 80 100
1.0 1.0
1.5 1.5
2.0 averaged 2.0

q c profile
2.5 2.5
3.0 3.0
Depth (m)
3.5 3.5
4.0 4.0
4.5 4.5
5.0 5.0
5.5 5.5
(a)
6.0 6.0
qc (MPa) COV (%)

0 5 10 15 20 25 0 20 40 60 80 100
1.0 1.0
1.5 1.5
2.0 2.0
2.5 2.5
3.0 3.0
Depth (m)
averaged
3.5 3.5
q c profile
4.0 4.0
4.5 4.5
5.0 5.0
5.5 5.5
(b)
6.0 6.0
Figure 3.11 Measured CPT qc profiles adjacent to piles, (a) P413, and (b) P415
3-35
1.8
Akasaka
1.6 Dunkirk
Labenne
1.4
Q Total,C /Q Total,M (-)
Gold Coast
1.2
0.8
(a)
0.6
0.4 Fugro-05
average=0.99
0.2 COV=0.35
0
0 0.2 0.4 0.6 0.8 1
D r (-)
1.4
1.2
1
0.8
0.6
Akasaka
(b)
0.4 Dunkirk
ICP-05 Labenne
0.2 average=0.98 Gold Coast
COV=0.17
0
0 0.2 0.4 0.6 0.8 1
D r (-)
3-36
1.4
1.2
1
0.8
0.6
(c) Akasaka
0.4 Dunkirk
NGI-05 Labenne
0.2 average=0.96 Gold Coast
COV=0.17
0
0 0.2 0.4 0.6 0.8 1
D r (-)
1.2
1
0.8
0.6
0.4 (d)
Akasaka
UWA-05 Dunkirk
0.2 average=0.79 Labenne
COV=0.17 Gold Coast
0
0 0.2 0.4 0.6 0.8 1
D r (-)
Figure 3.12 Ratios of calculated to measured total capacities plotted against the relative density of the
sand, Dr=0.4ln(qc,avg/22/(σ’v0pa)0.5) for UWA database of jacked piles
Based on above discussions, it can be inferred that the driven pile design methods
require further justification and possibly modification before being safely and
efficiently employed for jacked piles as well. Therefore, the measured base resistance
3-37
(qb0.1,M) for some of the instrumented jacked piles, as summarised in Table 3.9, are
further examined. The measured base resistance ratios (qb0.1,M/qc,avg) are plotted
against pile diameter (D) and the relative density of the sand (Dr) on Figure 3.13. The
average and COV of the measured base resistance ratios (qb0.1/qc,avg) are 0.89 and
0.13 respectively. It is obvious that the jacked pile (qb0.1,M/qc,avg=0.89) has
significantly higher base resistance compared to a driven pile (qb0.1,M/qc,avg=0.60, see
Figure 3.7) in similar conditions. In fact, this is also supported by the recent research
findings (as shown in Figure 3.14, a slide from White’s presentation in UWA) from
the field test for 9 closed-ended jacked piles installed in Takasu, Japan (White 2006).
It has been demonstrated that (i) jacked piles have higher base stiffness with base
resistance reaching plunging failure at relative small displacement 2 to 4%D, and (ii)
the ultimate base resistance (reached before 0.1D, also referred as qb0.1) is
approximately equal to the base resistance at the end of installation, which in turn
can be safely inferred from the value of qc,avg by Dutch averaging technique. In other
words, the ratio of qb0.1,M/qc,avg derived from the field tests is approximately 1.0.
1.2
Dunkirk; DK2
1 Labenne; LB2
Kallo; CPT250
Dunkirk; DK1
Labenne; LB2
0.8 Akasak; 6B
q b0.1, M /q c,avg (-)
Akasak; 1C
0.6
0.4
(a)
0.2 average=0.89
COV=0.13
0
0 0.05 0.1 0.15 0.2 0.25 0.3
D (m)
3-38
1.2
Dunkirk; DK2
1 Labenne; LB2
Dunkirk; DK1
Labenne; LB2 Kallo; CPT250

0.8
q b0.1, M /q c,avg (-)
Akasak; 6B
Akasak; 1C
0.6
0.4
(b)
0.2 average=0.89
COV=0.13
0
0 0.2 0.4 0.6 0.8 1
Dr (-)
Figure 3.13 Measured resistance ratios, qb0.1,M/qc,avg plotted against, (a) pile diameter, and (b) relative
density of sand, Dr=0.4ln(qc,avg/22/(σ’v0pa)0.5)
Figure 3.14 Base resistance response for 9 closed-ended jacked piles (D=319mm) installed in Takasu,
Japan (White 2006)
3-39
Although based on a limited number of tests (7 piles), the potential high base
resistance derived (qb0.1/qc,avg=0.89) is also supported by additional recent field tests
reported by White (2006). It is therefore proposed to modify the UWA-05 so that
qb0.1/qc,avg equals 0.9 for jacked piles (see Figure 3.13); this compares to the ratio of
0.6 adopted for driven piles. With this new formulation for end bearing, the method
was applied to the database of closed-ended jacked piles (Table 3.9). The
formulation for shaft resistance of the method remains the same (as outlined in Table
3.2) because the effect of installation cycles (blow counts vs. jacking strokes) on
development of shaft resistance are only qualitative and requires considerable further
research (White & Lehane 2004). The ratios of calculated to measured capacities
(using qb0.1/qc,avg =0.9) are summarised in Figure 3.15 for the database of jacked piles.
As expected, the predictive performance of the modified UWA-05 is greatly
improved with an average and COV of 1.02 and 0.19.
1.6
1.4
1.2
0.8
0.6
Akasaka
0.4 Dunkirk
Modified UWA-05
average=1.03 Labenne
0.2
COV=0.19 Gold Coast
0
0 0.2 0.4 0.6 0.8 1
D r (-)
Figure 3.15 Ratios of calculated to measured capacities using the modified UWA-05 method
Compared to closed-ended jacked piles, there are even less data being reported for
open-ended jacked piles. Gavin & Lehane (2005) reported four open-ended jacked
piles installed in sites Blessington and Kochi. The measured resistance ratios of
qb0.1/qc,avg range from 0.3 to 1.1, depending on the final filling ratios (as listed in
Table 3.11). The two fully plugged piles (FFR=0) behaved like closed-ended piles,
3-40
with ratios of qb0.1/qc,avg around unity, while for the partially plugged piles (FFR~0.6),
the ratios are much lower (<0.6). This indicates a similar trend for the open-ended
driven piles. However, more high quality data need to be gathered before an end
bearing design method for open-ended jacked piles can be established.
Table 3.11 Open-ended jacked piles in sand reported by Gavin & Lehane (2005)
D L qb0.1 qc,avg qb0.1/qc,avg

Site Name FFR Ref
m m MPa MPa -
Blessington 0.11 2 0 15 15 1 (Gavin & Lehane 2003)
Kochi 0.16 7.1 0 10 9 1.1 (White et al. 2000)
Kochi 0.32 8 0.60 1.7 5.8 0.3 (White et al. 2000)
Kochi 0.32 5.7 0.55 5.5 9 0.6 (White et al. 2000)
3-41
CHAPTER 4 FIELD TESTS AT SHENTON PARK
4.1 INTRODUCTION
This Chapter presents the pile testing programme at UWA Shenton Park site. One of
the primary objectives was to provide independent data that could be used in the
assessment of the four CPT-based design methods for end-bearing of driven piles in
sand as discussed in Chapter 3 (Fugro-05, ICP-05, NGI-05 and UWA-05). Other
aspects regarding skin friction will be reported separately by Schneider (2007). In
this Chapter, firstly, the soil conditions at the testing site are discussed based on in
situ CPT tests. Next the experimental set-up and testing procedures are described.
Finally, the results and analysis of the pile installation and static load test results are
presented.
4.2 SOIL CONDITIONS
4.2.1 Shenton Park test site
The location selected for the pile testing program was the UWA research field station
at Shenton Park in Perth. The soil conditions at this site consisted of 5-12m of dry to
moist, yellow sand overlying variably cemented Tamala Limestone. The sand is part
of the calcareous Spearwood dune system. The calcite, however, has been leached
out and the sand is primarily composed of quartz, containing only minor traces of
feldspar and calcium carbonate. The effective particles sizes of the sand, D50, D60 and
D10, are about 0.42±0.02mm, 0.47±0.02mm and 0.21±0.01mm respectively. The
angularity of the sand grains varies from sub-angular to sub-rounded and the
maximum and minimum void ratios (emax and emin) are 0.81±0.01 and 0.45±0.01
respectively. The in situ bulk density is approximately 1670±25kg/m3 and the water
level is below the base of the sand layer.
4-1
Chapter 4 Field Tests at Shenton Park
4.2.2 CPT results
The in situ cone penetration tests (CPT) were performed by Probedrill Pty. Ltd. using
a 12 tonne track mounted CPT rig. In total, twelve (12) and nine (9) CPTs were
conducted close to the pile testing locations on May 3rd and July 10th 2005
respectively. They are designated as C1 to C12 for the first testing period and C13 to
C21 for the second testing period. The results of all the tests are summarised on
Figure 4.1. The averages and coefficient of variations (COVs) for the measured qc
values during the two testing periods are summarised in Figure 4.2 and Figure 4.3. It
can be observed that:
(i) There is a stiff surface crust of ~0.3m in depth, possibly due to compaction from
traffic and animal activity (sheep) over time. Values of qc then increase linearly
with depth to about 5m where they increase significantly, indicating the presence
of the limestone layer.
(ii) Considerable variation existed between the qc values measured during the two
testing periods. For instance, at depth of 3m, qc values are twice as high for the
first testing period than the second. This is most probably attributed to the fact
that the testing area is surrounded by trees, which can have a significant
influence on the measured CPT data; this phenomenon will be further discussed
in the following section.
(iii) Despite the significant variation in qc value between the two testing periods,
there is a good agreement between CPTs performed at the same time, with
similar coefficient of variations (COVs) of ~20% (Figure 4.2 and Figure 4.3),
indicating that even for a relatively uniform sand site, i.e. at Shenton Park, you
could still expect a coefficient of variation in qc of about 20%.
4-2
qc (MPa)
0 2 4 6 8 10 12 14 16 18 20
0.0
0.5
C1 C2
1.0 C3 C4
C5 C6
1.5 First set of CPTs
C7 C8
C1 to C12
C9 C10
2.0
Depth (m)
C11 C12
2.5 C13 C14
C15 C16
3.0
C17 C18
3.5 C19 C20
C21
4.0
Second set of CPTs

4.5
C13 to C21
5.0
Figure 4.1 CPT qc profiles during the two testing periods
qc (MPa) COV (%)

0 2 4 6 8 10 12 14 16 18 20 0 20 40 60 80 100
0.0 0.0
0.5 0.5
1.0 Average of q c values 1.0

from first set of
1.5 1.5
CPTs C1-C12
2.0 2.0
Depth (m)
2.5 2.5
3.0 3.0
3.5 3.5
4.0 4.0
4.5 4.5
(a) (b)
5.0 5.0
Figure 4.2 (a) qc profiles for the first set of CPTs, C1-C12, and (b) COV of measured qc values
4-3
qc (MPa) COV (%)

0 2 4 6 8 10 12 14 16 18 20 0 20 40 60 80 100
0.0 0.0
0.5 0.5
1.0 1.0
1.5 1.5
2.0 2.0
Depth (m)
Depth (m)
Average of q c values
2.5 from second set of 2.5
CPTs C13-C21
3.0 3.0
3.5 3.5
4.0 4.0
4.5 4.5
(a) (b)
5.0 5.0
Figure 4.3 (a) qc profiles for the second set of CPTs, C13-C21, and (b) COV of measured qc values
Figure 4.4 The testing area; note the trees located about 30m away
4-4
4.2.3 Seasonal effect
As discussed in the previous section, the CPTs, conducted by the same company
using the same equipment, show significant variation between the two testing periods,
but are consistent within each testing period. One possible explanation for this is that
seasonal changes in the soil water content could influence the stiffness and strength
of the sand. Figure 4.4 shows schematically that the testing area in Shenton Park is
surrounded by trees at a distance of approximately 30m. The monthly rainfall data
recorded in the year 2004 and 2005 by Water Corporation are plotted on Figure 4.5.
It is apparent that the vast majority of rainfall occurs between May and October and
there is virtually no rain between November and April. The two sets of CPTs at
Shenton Park were indeed conducted in two separate seasons: the end of the dry
season for C1 to C12 and the middle of the wet season for C13 to C21.
300
250
Avg
2004
200 2005
Rainfall (mm)
150
100
50
0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Figure 4.5 Rainfall in Perth for each month in the year 2004 and 2005
The extensive research by Lehane et al (2004) at Shenton Park revealed that there is
a very strong influence of water uptake by trees roots on the stiffness and strength of
Perth’s Spearwood dune sands, and therefore on the measured CPT data. These
effects can be explained by the mechanics associated with relatively small water
content changes at very low levels of saturation leading to the generation of high
4-5
suction pressures in the dry season. With a modest increase in the soil saturation, the
suction pressures decrease significantly during the wet season.
This seasonal effect needs to be accounted for when analyzing CPT data for design
purposes. However, in present study, the analysis of results, (e.g. comparison
between pile and cone tip resistance) is not affected by the seasonal changes as they
were conducted at the same time during the two testing periods. The pile behaviour
would therefore be affected in the same way as a cone. In fact, in the following
discussion, only CPT conducted closest to the pile at the same period was used for
the analysis.
4.3 TEST SET-UP AND PROCEDURES
4.3.1 Test programme
Two sets of experiments were conducted at UWA Shenton Park, the first between
29th April and 4th May, and the second extending from July 6th to July 11th 2005.
Over the two testing periods, 10 open-ended and 2 closed-ended steel pipe piles with
varying outer diameter, wall thickness and length were driven into the ground and
tested in compression followed by tension. Additional tension tests were performed
at various times between July 2005 and September 2006 to study the influence of
time on pile skin friction. The results will be reported separately by Schneider (2007).
The details of the pile configuration are summarized in Table 4.1. The relative
position of the test piles and CPTs are shown in Figure 4.6. It should be noted that
three CPTs (C19 to C20) are not marked in the figure as they were conducted outside
the current testing area for other purposes.
The 12 piles installed at Shenton Park consisted of a steel tube (Grade 250L) with a
25mm thick steel plate welded to the pile head (Figure 4.7a), which was designed to
(i) prevent the pile body from damage during pile driving, (ii) enable the use of a rod
to guide the driving hammer, and (iii) provide attachment points for performing
tension tests. The two closed-ended piles were formed by welding a steel plate to the
toe of the steel tube (Figure 4.7b). Markings at 20mm depth intervals were drawn on
the piles. At the final penetration depth, each pile was generally about 300mm above
4-6
the ground level. This allowed for displacement to occur during the compression and
tension tests.
Table 4.1 Summary of the model pile tests in Shenton Park
Testing Pile End L D t D/t WH Date

Period No condition (m) (mm) (mm) (-) (kg) Installed
P1 Open 4 88.9 2.6 34.2 25 29-Apr
P2 Open 4 42.4 2.6 16.3 25 29-Apr
P3 Closed 4 88.9 5 17.8 25 30-May
First
P4 Open 4 88.9 5 17.8 25 1-May
P5 Open 4 114.3 3.2 35.7 25 1-May
P6 Open 4 88.9 3.2 27.8 25 1-May
P7 Open 2.5 42.4 2.6 16.3 10 6-Jul
P8 Open 2.5 42.4 2.6 16.3 25 6-Jul
P9 Open 2.5 33.7 2.6 13.0 25 6-Jul
Second
P10 Open 3.5 33.7 2.6 13.0 25 6-Jul
P11 Open 2.5 88.9 2.6 34.2 25 6-Jul
P12 Closed 2.5 88.9 2.6 34.2 25 6-Jul
L: pile embedment length; D: pile outer diameter; t: thickness of the pile wall; WH: hammer weight
Figure 4.6 Relative testing locations for the piles and CPTs during the two testing periods
4-7
Figure 4.7 (a) steel plate welded to the pile head, and (b) steel cap welded to the pile toe to form a
closed-ended pile
4.3.2 Pile installation
Measurements of the soil plug height (i.e. the advancement of the soil inside the
open-ended pile) and blow counts were gathered at the end of each 100mm driving
increment. The driving apparatus (as shown in Figure 4.8) consisted of lightweight
aluminium scaffolding of 5m in height, two pulleys, a drop weight (either 25kg or
10kg), a guide rod, pile support, and the soil plug measurement device.
Figure 4.8 (a) schematic set-up of the pile driving system, and (b) photo of pile driving
4-8
The two pulleys were attached to a cross bar at the top of the scaffolding. One
allowed manual control of the guide rod and the drop weight. The other was used to
connect the lighter weight outside and with the heavier weight inside the pile to
measure the soil plug development during pile driving (see Figure 4.8a). Most piles
were driven using the 25kg drop weight, except P7 (Table 4.1). Use of the lighter
drop weight (10kg) for P7 required more blow counts (therefore more loading-
unloading cycles) for the same penetration depth compared to a similar pile, P8. This
use of different drop weights allowed the influence of the number of dynamic
installation loading cycles on the static capacity to be investigated. The guide rod,
going through the centre of the drop weight and down into the pile, was used to
ensure efficient and concentric contact between the drop weight and the pile top plate.
It also served as a reference for maintaining a constant drop height of 0.5m. To
ensure pile verticality during installation, the base of the pile was first positioned
directly under the drop weight using a plumb line. The pile support was then attached
to the upper part of the pile body. A spirit level was used to align the pile vertically
before the pile support was firmly clamped to the scaffolding. The incremental soil
height change inside can be inferred from the height change of the outside light
weight. A closer view of the drop weight, guide rod, pile support and the soil plug
measurement device is shown in Figure 4.9.
4.3.3 Static load tests
Static load tests were performed on all piles to evaluate their compression and
tension capacities. The loading rate was kept constant at ~1mm/min and loading
continues to a pile head displacement in excess of 10% of the pile diameter.
In compression, the pile was loaded by a hydraulic jack placed in between the pile
head and an axial load cell, which was then jacked against the bottom of the 12 tonne
CPT truck (Figure 4.10a). To avoid load concentration, a thick steel plate of larger
diameter was placed on top of the load cell. A LDT (linear displacement transducer)
was used to measure the pile head displacement during loading. It was attached to a
separate beam so when the pile was loaded, any ground movements surrounding the
pile would not interfere with the LDT readings.
4-9
Figure 4.9 (a) photo of the drop weight, guide rod, pile support and the pile, and (b) photo of light
weight and ruler used to measured the soil plug movement
The tension test involved applying a tensile force evenly to the pile head using the
apparatus shown in Figure 4.10b. Firstly, a steel plate was bolted to the existing plate
on the pile head. A long threaded rod was passed through the load cell and the
hydraulic jack and screwed into the steel plate. A nut was then screwed on the other
end of the rod and tightened off at the top of the load cell. In this way, when the
hydraulic jack was extended, a tension force was applied to the pile through the
tensile rod. Meanwhile, the reaction force was transmitted to the ground through the
reaction frame. The footings of the reaction frame were at a spacing of greater than
750mm to avoid any interference of the loaded ground and the pile. The upward
displacement of the pile head was measured using a LDT attached to a separate beam.
The soil column inside the pile was monitored both before and after the static load
tests to check if there was any relative movement between the soil and the pile. This
was done by inserting a long rod down into the pile until it reached the top of the soil
plug. The distance from the top of the pile and that of the soil column was therefore
recorded. If there was any relative movement between the soil column and the pile,
the distances measured before and after the static load tests would differ. However,
4-10
for all piles tested, no relative soil movement was detected. In other words, all piles
were fully plugged during static compression and tension loading.
Figure 4.10 Photo of experimental set-up for, (a) compression, and (b) tension tests
4.4 ANALYSIS OF TEST RESULTS
4.4.1 Driving records
As discussed earlier, data regarding soil plug length (Lp) and blow counts were
gathered at every ~100mm of pile penetration depth (L) for all piles. This allows
calculations of incremental filing ratio (IFR=ΔLp/ΔL, incremental soil plug change
over pile penetration depth), which best quantifies the degree of soil displacement
induced during pile driving. The final filling ratio (FFR) was calculated as the
average of IFR over the last 3D of pile penetration. In addition, plug length ratio
(PLR=Lp/L, final soil plug length over pile penetration depth) can also be calculated.
The driving records, i.e. the accumulated blow counts, IFR (or FFR), and PLR are
summarised in Figures 4.11, 4.12 and 4.13 and Table 4.2 for all piles. It can be
observed that:
(i) The blow count is an indication of the ease of driving. In this experiment, the
drop weight and height were kept constant as 25kg and 0.5m for all piles,
except for P7, where a lighter weight of 10kg was used. As shown in Figure
4.11, the blow counts reduce with a decrease in pile diameter. For the two
closed-ended piles (P3 & P12), the blow counts are significantly higher than
4-11
that for open-ended piles. In addition, the seasonal difference between qc

profiles measured during the two testing period is also reflected in the blow
counts. For instance, when comparing piles installed during the first testing
period with the second (i.e. P1 with P11, P2 with P8, and P3 with P12), it is
found that the blows counts to drive P1, P2 and P3 are much higher than that
for P8, P11 and P12 at penetration depth up to 2.5m. For similar piles (P7 & P8)
with different drop weights, as expected, the number of blows required to
install P7 was about double of that for P8.
(ii) There is a clear reduction in the IFR and PLR for smaller diameter piles and
similar IFR and PLR for common diameter piles during the same testing period.
The wall thickness does not appear to have a significant influence on the soil
plug development presumably because of the small range of wall thickness
investigated (i.e. t=2.6mm, 3.2mm and 5mm). The IFR values were greater
than unity (100%) at the ground surface, which was possibly due to dilation of
the dense soil crust. The IFR for each pile reduced with penetration depth,
indicating an increase in soil plug resistance. It is expected that IFR would
continue to reduce with depth due to a continue increase in soil plug resistance
unless a significant change in soil strength is encountered. However, for the
first series of tests, a slight increase in IFR was observed at depth greater than
2.5m.
(iii) The qc profile, blow counts and IFR for similar piles installed during two
testing period are compared in Figures 4.14 and 4.15. As discussed before, the
higher qc values in the first tests are also reflected in the higher number of
blows required to install similar diameter piles. However, the IFR does not
seem to be significantly affected by the change in soil strength between the two
testing periods or the number of blows; the IFR seems to be primarily a
function of diameter.
(iv) As summarised in Table 4.2, the FFR values range between 15% and 85%. In
general, they reduce with decreasing diameter, but do not follow the simplified
trend well (i.e. FFR = min[1, [Di (m)/1.5(m)]0.2, in Figure 3.10). Therefore, the
applicability of such simplification to small diameter pile requires further
investigation.
4-12
Σ blow counts (-) Σ blow counts (-)

0 500 1000 1500 2000 2500 3000 3500 4000 4500 0 100 200 300 400 500
0 0
P1
0.5 P2 0.5
P3
P4
P5
1 P6 1
P3; Closed-ended
1.5 1.5
D=114.3mm P12; Closed-ended
P7
D=42.4mm D=88.9mm
Depth (m)
2 2
P8 P11
2.5 2.5
D=42.4mm D=88.9mm
3 3
P5
D=114.3mm P9 & P10
3.5 3.5 P7
D=33.7mm
P8
P9
4 4 P10
P2 P1, P4, & P6 P11
D=42.4mm D=88.9mm P12
4.5 4.5
Figure 4.11 Measured accumulated blow counts data for all piles
IFR (%) IFR (%)

0 20 40 60 80 100 120 140 0 20 40 60 80 100 120 140
0 0
P9 & P10
P2
D=33.7mm
0.5 D=42.4mm 0.5 P11
D=88.9mm
1 1
1.5 1.5
Depth (m)
2 2
P5
D=114.3mm
2.5 2.5
3 3 P7 & P8
D=42.4mm
3.5 3.5 P7
P1
P8
P2
4 P4 4 P9
P1, P4, & P6 P5 P10
D=88.9mm P6 P11
4.5 4.5
Figure 4.12 Measured IFR data for the 10 open-ended piles
4-13
PLR (%) PLR (%)

0 20 40 60 80 100 120 140 0 20 40 60 80 100 120 140
0 0
0.5 0.5 P9 & P10

D=33.7mm
1 1 P11
D=88.9mm
1.5 P2 1.5
D=42.4mm
Depth (m)
2 2
2.5 2.5
P5
D=114.3mm
3 3
P7 & P8
D=42.4mm
3.5 3.5 P7
P1
P8
P2
P4 P9
4 P1, P4, & P6 4
P5 P10
D=88.9mm P11
P6
4.5 4.5
Figure 4.13 Measured PLR data for the 10 open-ended piles
q c (MPa) Σ blow counts (-) IFR (%)

0 2 4 6 8 10 0 200 400 600 0 20 40 60 80 100 120 140
0 0 0
P1; First
0.5 0.5 P11; Second 0.5
1 1 1
Depth (m)
1.5 1.5 1.5
2 2 2
2.5 2.5 2.5
3 3 3
Figure 4.14 Comparison between similar diameter piles (P1 & P11) installed during the two testing
periods
4-14
q c (MPa) Σ blow counts (-) IFR (%)

0 2 4 6 8 10 0 50 100 150 200 0 50 100
0 0 0
P2; First; 25kg
P7; Second; 10kg
0.5 0.5 P8; Second; 25kg 0.5
1 1 1
Depth (m)
1.5 1.5 1.5
2 2 2
2.5 2.5 2.5
3 3 3
Figure 4.15 Comparison between similar diameter piles (P2, P7 & P8) installed during the two testing
periods
Table 4.2 Summary of FFR, PLR and Σblow counts for all piles
Testing Pile End L D t WH FFR PLR Σblow

Period No condition (m) (mm) (mm) (kg) (%) (%) counts
P1 Open 4 88.9 2.6 25 69 74 1642
P2 Open 4 42.4 2.6 25 50 41 330
P3 Closed 4 88.9 5 25 0 0 4333
First
P4 Open 4 88.9 5 25 77 78 1840
P5 Open 4 114.3 3.2 25 85 87 1830
P6 Open 4 88.9 3.2 25 77 78 1486
P7 Open 2.5 42.4 2.6 10 26 44 205
P8 Open 2.5 42.4 2.6 25 24 43 90
P9 Open 2.5 33.7 2.6 25 18 32 65
Second
P10 Open 3.5 33.7 2.6 25 15 29 111
P11 Open 2.5 88.9 2.6 25 45 64 305
P12 Closed 2.5 88.9 2.6 25 0 0 451
4-15
4.4.2 Static load tests results
Static load tests were performed for all piles at full penetration to allow for accurate
measurements of the pile tensile and compressive capacities. The results are
summarised in Figures 4.16 and 4.17. The ultimate compressive total (QTotal) and
tensile shaft (Qs-tension) capacity, defined at pile displacement of 10%D, are
summarised in Table 4.3. The base capacity (Qb) is estimated by subtracting the
compressible shaft capacity (i.e. assumed to be Qs-tension/0.75) from the total
compressive capacities Qtotal. This is consistent with UWA-05 design
recommendation (i.e. ft/fc=0.75 in Table 3.2).
Table 4.3 Summary of static load test results
Pile No QTotal (kN) Qs-tension (kN) Qb (kN)

P1 48.4 17.3 25.3
P2 16.5 8.2 5.6
P3 64.5 7.4 54.7
P4 57.4 25.2 23.7
P5 59.5 28.9 21.0
P6 56.0 29.0 17.3
P7 2.9 0.9 1.7
P8 9.7 4.9 3.2
P9 4.7 1.6 2.6
P10 8.9 3.5 4.3
P11 20.6 11.4 5.4
P12 18.7 5.0 12.0
Table 4.4 Summary of the averaged qc values and measured qb0.1 for all piles
qc,avg (MPa)
Pile Ar,eff, qb0.1,M qc,tip
No (-) (MPa) (MPa) ±1.5D a Dutch b
P1 0.39 4.08 10.8 10.8 10.56

P2 0.62 3.96 11.56 11.58 11.3
P3 1 8.81 10.64 10.73 10.33
P4 0.39 3.83 11.06 10.86 10.53
P5 0.24 2.05 11.7 11.72 11.3
P6 0.34 2.79 9.74 9.9 9.58
P7 0.8 1.2 3.32 3.34 3.26
P8 0.82 2.25 3.53 3.53 3.43
P9 0.87 2.91 4.08 4.07 3.97
P10 0.89 4.79 5.29 5.32 5.26
P11 0.6 0.87 3.99 3.98 3.82
P12 1 1.94 2.95 2.96 2.94
a
: the averaged qc value at distance up to ±1.5D at pile tip; b: the averaged qc value using Dutch
method
4-16
60 20
Q Total Q Total
18
50
16
14
40
12
Load (kN)
30 10 Q s -tens ion
8
20 Q s -tens ion P1, Open-ended 6 P2, Open-ended
D o =88.9mm
D o =42.4mm
t=2.6mm 4
10 t=2.6mm
L=4m
2 L=4m
25kg Hammer
25kg Hammer
0 0
80 0 10 20 30 40 80 0 10 20 30 40 50
w/D (%) w/D (%)
Q Total Q Total
70 70
60 60
50 50
Load (kN)
40 40
30 30 Q s -tens ion
P3, Closed-ended P4, Open-ended
20 D o =88.9mm 20 D o =88.9mm
t=5mm t=5mm
Q s -tens ion L=4m L=4m
10 10
25kg Hammer 25kg Hammer
0 0
90 0 10 20 30 40 90 0 10 20 30 40 50
w/D (%) w/D (%)
Q Total
80 80
Q Total
70 70
60 60
Load (kN)
50 50
40 40
Q s -tens ion Q s -tens ion
30 30
P5, Open-ended P6, Open-ended
D o =114.3mm, D o =88.9mm,
20 20
t=3.2mm t=3.2mm
L=4m L=4m
10 10
0 0
0 10 20 30 40 0 10 20 30 40 50
w/D (%) w/D (%)
Figure 4.16 Static load test results for piles P1 to P6
4-17
4 12
Q Total
Q Total
10
3
8
Q s -tens ion
Load (kN)
2 6
Q s -tens ion 4
P7, Open-ended P8, Open-ended
1 D o =42.4mm, D o =42.4mm,
t=2.6mm t=2.6mm
2
L=2.5m L=2.5m
0 0
6 0 10 20 30 40 10 0 10 20 30 40
w/D (%) w/D (%) Q Total
Q Total 9
5
8
7 Q s -tens ion
4
6
Load (kN)
Q s -tens ion
3 5
4
2
P9, Closed-ended 3 P10, Open-ended
D o =33.7mm D o =33.7mm
t=2.6mm 2 t=2.6mm
1
L=2.5m L=3.5m
1
0 0
25 0 10 20 30 40 25 0 10 20 30 40
w/D (%) Q Total w/D (%)
Q Total
20 20
15 15
Load (kN)
Q s -tens ion
10 10
P11, Open-ended Q s -tens ion P12, Closed-ended

D o =88.9mm D o =88.9mm
5 t=2.6mm 5 t=2.6mm
L=2.5m L=2.5m
0 0
0 10 20 30 40 0 10 20 30 40
w/D (%) w/D (%)
Figure 4.17 Static load test results for piles P7 to P12
4-18
qc (MPa) qb0.1 (MPa)

0 2 4 6 8 10 12 0 2 4 6 8 10 12
0 0.0
Raw qc Fugro-05
0.5 0.5 ICP-05
qc,±1.5Do
qc,Dutch NGI-05
1 1.0
UWA-05
Measured
1.5 1.5
Depth (m)
2 2.0
2.5 2.5
3 3.0
3.5 P1 3.5 P1
Tip depth Tip depth
4 4.0
4.5 4.5

0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14
0 0.0
Raw qc Fugro-05
0.5 0.5 ICP-05
qc,±1.5Do
qc,Dutch NGI-05
1 1.0
UWA-05
Measured
1.5 1.5
Depth (m)
2 2.0
2.5 2.5
3 3.0
3.5 P2 3.5 P2
Tip depth Tip depth
4 4.0
4.5 4.5
Figure 4.18 qc and predicted qb0.1 profiles for piles P1 and P2
4-19

0 2 4 6 8 10 12 0 2 4 6 8 10 12
0 0
Raw qc Fugro-05
0.5 0.5 ICP-05
qc,±1.5Do
qc,Dutch NGI-05
1 1
UWA-05
Measured
1.5 1.5
Depth (m)
2 2
2.5 2.5
3 3
3.5 P3 3.5 P3
Tip depth Tip depth
4 4
4.5 4.5

0 2 4 6 8 10 12 0 2 4 6 8 10 12
0 0
Raw qc Fugro-05
0.5 0.5 ICP-05
qc,±1.5Do
qc,Dutch NGI-05
1 1 UWA-05
Measured
1.5 1.5
Depth (m)
2 2
2.5 2.5
3 3
3.5 P4 3.5 P4
Tip depth Tip depth
4 4
4.5 4.5
4-20

0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14
0 0
Raw qc Fugro-05
0.5 0.5 ICP-05
qc,±1.5Do
qc,Dutch NGI-05
1 1 UWA-05
Measured
1.5 1.5
Depth (m)
2 2
2.5 2.5
3 3
3.5 P5 3.5 P5
Tip depth Tip depth
4 4
4.5 4.5

0 2 4 6 8 10 12 0 2 4 6 8 10 12
0 0
Raw qc Fugro-05
0.5 0.5 ICP-05
qc,±1.5Do
qc,Dutch NGI-05
1 1
UWA-05
Measured
1.5 1.5
Depth (m)
2 2
2.5 2.5
3 3
P6 P6
3.5 3.5
Tip depth Tip depth
4 4
4.5 4.5
4-21
qc (MPa) q b0.1 (MPa)

0 2 4 6 8 0 2 4 6 8
0 0
Raw qc Fugro-05
0.5 qc,±1.5Do0.5 ICP-05
qc,Dutch NGI-05
1 1 UWA-05
Measured
1.5 1.5
Depth (m)
2 2
P7 P7
Tip depth Tip depth
2.5 2.5
3 3
3.5 3.5
4 4
q c (MPa) q b0.1 (MPa)

0 2 4 6 8 0 2 4 6 8
0 0
Raw qc Fugro-05
0.5 qc,±1.5Do0.5 ICP-05
qc,Dutch NGI-05
1 1 UWA-05
Measured
1.5 1.5
Depth (m)
2 P8 2 P8
Tip depth Tip depth
2.5 2.5
3 3
3.5 3.5
4 4
4-22

0 2 4 6 8 0 2 4 6 8
0 0
Raw qc Fugro-05
0.5 qc,±1.5Do0.5 ICP-05
qc,Dutch NGI-05
1 1 UWA-05
Measured
1.5 1.5
Depth (m)
2 P9 2 P9
Tip depth Tip depth
2.5 2.5
3 3
3.5 3.5
4 4
qc (MPa) q b0.1 (MPa)

0 2 4 6 8 0 2 4 6 8
0 0
Raw qc Fugro-05
0.5 qc,±1.5Do0.5 ICP-05
qc,Dutch NGI-05
1 1 UWA-05
Measured
1.5 1.5
Depth (m)
2 2
2.5 2.5
3 P10 3 P10
Tip depth Tip depth
3.5 3.5
4 4
4-23

0 2 4 6 8 0 2 4 6 8
0 0
Raw qc Fugro-05
0.5 qc,±1.5Do0.5 ICP-05
qc,Dutch NGI-05
1 1 UWA-05
Measured
1.5 1.5
Depth (m)
2 P11 2 P11
Tip depth Tip depth
2.5 2.5
3 3
3.5 3.5
4 4

0 2 4 6 8 0 2 4 6 8
0 0
Raw qc Fugro-05
0.5 qc,±1.5Do0.5 ICP-05
qc,Dutch NGI-05
1 1 UWA-05
Measured
1.5 1.5
Depth (m)
2 P12 2 P12
Tip depth Tip depth
2.5 2.5
3 3
3.5 3.5
4 4
4-24
4.4.3 Performance of Fugro-05, ICP-05, NGI-05 & UWA-05
The predictive performance of the four CPT-based methods, Fugro-05, ICP-05, NGI-
05 & UWA-05 (as discussed in Chapter 2), was further evaluated against the
measured base capacity of the 12 piles tested at Shenton Park. For the first series of
piles installed (P1 to P6), the averaged qc profile from two CPTs adjacent to each pile
was used as the input data for prediction of the pile base capacity. As discussed in
earlier section, due to the seasonal influence on the stiffness of strength parameters
of the sand, the qc values measured in the second series of tests were significantly
lower than those in the first. As a result, the qc profile measured from the CPT
adjacent to each pile in the second series was used to evaluate the pile base capacity.
The qc profiles and qb0.1 profiles calculated by the four methods for each pile are
summarized in Figure 4.18 to Figure 4.23. The qc values (qc,tip) at the pile tip level,
averaged qc values (qc,avg) and measured/estimated base resistance (qb0.1,M) are
summarized in Table 4.4. The effective area ratios, calculated as a function of the
measured FFR and pile diameters, Ar,eff=1-FFR×(Di/D)2, are also listed in Table 4.4.
As shown, there is little difference between values of qc,tip and qc,avg obtained by
simplified arithmetic average or Dutch averaging method because the testing site is
relatively uniform. Therefore the choice of the qc averaging techniques will not
influence the predictive performance of each method. Instead, the formulation of the
method plays the key role in assessing its predictive performance.
Table 4.5 summarizes the calculated base resistance qb0.1,C and the ratios of
calculated to measured resistance (qb0.1,C/qb0.1,M) for the four methods. Values of
qb0.1,C/qb0.1,M are further plotted against the cone tip resistance (qc,tip) on Figure 4.24.
The average and coefficient of variation (COV) of the ratios between calculated and
measured qb0.1 values are included in these Figures and may be taken as measures of
each method’s predictive ability. It appears that the UWA-05 method provides the
most accurate predictions for the base resistance, with an average of calculated to
measured capacities of 1.03 and a COV of calculated to measured capacities of 0.36.
Fugro-05 is the least conservative method with an average of 1.88 and a COV of 0.43.
ICP-05 and NGI-05 tend to over-predict the base resistance by ~20%, with COVs of
0.38 and 0.48 respectively. Together with the discussions in Chapter 2, it is obvious
that the improved predictive performance of UWA-05 may be attributed to its
4-25
incorporation of the varying effects of soil displacement (FFR) in the formulation.

None of the other three methods explicitly take this into consideration.
Table 4.5 Calculated base resistance, qb0.1,C, by Fugro-05, ICP-05, NGI-05 and UWA-05 methods
qb0.1,C (MPa) qb0.1,C/qb0.1,M (-)

Pile qb0.1,M
No (MPa) Fugro- UWA- Fugro UWA-
ICP-05 NGI-05 ICP-05 NGI-05
05 05 -05 05
P1 4.08 5.13 4.34 3.54 3.43 1.26 1.06 0.87 0.84
P2 3.96 6.34 5.58 3.62 4.82 1.60 1.41 0.91 1.22
P3 8.81 8.81 8.62 6.21 6.20 1.00 0.98 0.70 0.70
P4 3.83 6.01 4.36 3.57 3.44 1.57 1.14 0.93 0.90
P5 2.05 5.29 4.39 3.63 2.93 2.58 2.14 1.77 1.43
P6 2.79 5.16 3.98 3.43 2.89 1.85 1.42 1.23 1.03
P7 1.20 3.40 1.61 1.99 1.66 2.83 1.34 1.65 1.38
P8 2.25 3.50 1.70 2.05 1.78 1.55 0.76 0.91 0.79
P9 2.91 3.96 2.06 2.19 2.15 1.36 0.71 0.75 0.74
P10 4.79 4.53 2.70 2.68 2.90 0.94 0.56 0.56 0.61
P11 0.87 3.11 1.60 2.17 1.61 3.58 1.84 2.49 1.85
P12 1.94 4.63 2.38 2.28 1.76 2.39 1.23 1.18 0.91
average of qb0.1,C/qb0.1,M 1.88 1.22 1.16 1.03
COV of qb0.1,C/qb0.1,M 0.43 0.38 0.48 0.36
4
(a)
3.5 P11; 88.9 Fugro-05
average=1.88
3 COV=0.43
P7; 42.4
P5; 114.3
2.5
q b0.1,C /q b0.1,M
P12; 88.9
2
P6; 88.9
P8; 42.4 P4; 88.9 P2; 42.4
1.5
P9; 33.7 P1; 88.9
1 P10; 33.7 P3; 88.9
0.5
0
0 2 4 6 8 10 12 14
qc,tip (MPa)
4-26
2.5
(b)
ICP-05 P5; 114.3
2 average=1.22
COV=0.38 P11; 88.9
q b0.1,C /q b0.1,M
1.5
P7; 42.4 P6; 88.9 P2; 42.4
P12; 88.9
P4; 88.9
1 P1; 88.9
P3; 88.9
P8; 42.4 P9; 33.7
P10; 33.7
0.5
0
0 2 4 6 8 10 12 14
qc,tip (MPa)
3
(c)
NGI-05
2.5 P11; 88.9 average=1.16
COV=0.48
2
q b0.1,C /q b0.1,M
P5; 114.3
P7; 42.4
1.5
P12; 88.9 P6; 88.9

1 P4; 88.9 P2; 42.4
P8; 42.4
P9; 33.7 P1; 88.9
P10; 33.7 P3; 88.9
0.5
0
0 2 4 6 8 10 12 14
qc,tip (MPa)
4-27
2
P11; 88.9
(d)
1.8 UWA-05
1.6 average=1.03
COV=0.36
P7; 42.4 P5; 114.3
1.4
q b0.1,C /q b0.1,M
1.2 P2; 42.4
1 P6; 88.9
P12; 88.9 P4; 88.9
0.8 P1; 88.9
P8; 42.4 P9; 33.7
0.6 P10; 33.7 P3; 88.9
0.4
0.2
0
0 2 4 6 8 10 12 14
qc,tip (MPa)
Figure 4.24 Ratios of calculated to measured base resistance (qb0.1,C/qb0.1,M) for piles P1 to P12, (a)
Fugro-05, (b) ICP-05, (c) NGI-05, and (d) UWA-05
The dependence of pile base resistance on the varying effects of soil displacement is
further illustrated by plotting the ratios of qb0.1,M/qc,avg against the pile effective area
ratios, Ar,eff in Figure 4.25. The normalized soil plug resistance qb,plug, deduced from
the measured qb0.1 values (assuming qb,ann/qc,avg =0.6) are plotted against the final
filling ratios (FFR) on Figure 4.26. Also included in these figures are the design
recommendations by UWA-05 method. It can be observed, in general, that the base
and plug resistance ratios, qb0.1/qc,avg & qb,plug/qc,avg, depend on the degree of soil
displacement, as best represented by FFRs. The design recommendation by UWA-05
shows a reasonable agreement to the testing results. Some deviations from the UWA-
05 design lines may be explained by the following:
(i) The first series of field tests (involving the deeper piles) shows a good fit to the
UWA-05 design line, with the exception of the closed-ended pile (P3). The
high ratio of qb0.1./qc,avg (=0.85 > 0.6) of this pile could be attributed to an
underestimation of shaft capacity and hence over prediction of the base
capacity. The number of blows required to install P3 is over 4300, which is
unusual for field scale piles. Therefore the assumption of the ratio of tensile to
4-28
compressive shaft capacity of 0.75 may not be valid for this large number of
blows.
(ii) The shorter piles (P7 to P12) during the second testing period appear to deviate
from the UWA-05 design line. These piles had quite low tensile capacities (in
particular for the small diameter piles P7 to P10). It is noteworthy that after
installation, a gap between the surrounding soil and the pile could be observed
at the pile head, indicating the effect of pile whip were likely to be larger for
the shorter pile. The existence of the gap may influence the pile behaviour
differently when being loaded in compression and tension. As a result, the
compressive shaft capacity might be underestimated for those shorter piles.
0.9 P10; 33.7

P3; 88.9
0.8
P9; 33.7
0.7 P12; 88.9
P8; 42.4
0.6
q b0.1 /q c,avg
UWA-05
0.5 q b0.1 /q c,avg =0.15+0.45Ar,eff
0.4 P1; 88.9

P4; 88.9 P2; 42.4 P7; 42.4
0.3 P6; 88.9
P11; 88.9
0.2 P5; 114.3
0.1
0
0 0.2 0.4 0.6 0.8 1
Ar,eff (-)
Figure 4.25 Ratios of qb0.1/qc,avg [Dutch] plotted against pile effective area ratios Ar,eff
4-29
1.2
P10; 33.7
1
P9; 33.7 UWA-05

0.8 q b,plug /qc,avg =0.6-0.45×FFR
q b,plug /q c,avg
P8; 42.4
0.6
P2; 42.4 P1; 88.9

0.4 P4; 88.9
P7; 42.4
P6; 88.9
P11; 88.9
0.2
P5; 114.3
0
0 0.2 0.4 0.6 0.8 1
FFR (-)
Figure 4.26 Ratios of qb,plug/qc,avg [Dutch] plotted against final filling ratios FFR
4.5 ANALYSIS OF BASE STIFFNESS
The load transfer program, RATZ (Randolph 2003) was used to estimate the pile
base load displacement response curves. A hyperbolic base spring was specified and
the soil’s equivalent linear stiffness, E, (i.e. assuming a rigid punch, Equation 4.1)
was determined for the best fit spring at various displacement ratios. The back-
figured E values at varying displacement ratios (w/D=0.01, 0.02 & 0.1) are plotted
against FFR on Figure 4.27. This figure indicates a strong and clear dependence of
the base stiffness on degree of soil plugging. As the FFR reduces (more plugging
occurs), the base stiffness increases and hence so does the base resistance qb0.1
defined at displacement of 0.1D. During pile installation, the soil beneath the pile tip
undergoes a degree of prestressing, which is well known to lead to a stiffer soil
response. The greater the degree of plugging (lower the FFR), the larger the stress is
induced on the ground beneath the plug. Figure 4.27 also shows the expected
reduction in base stiffness as the displacement increases. This arises because the soil
is not a linear elastic material and its stiffness reduces with increasing strain level.
4-30
π(1 − ν 2 ) × D × q b
E= ( 4.1 )
w
where ν is Poissons ratio, D is pile outer diameter, qb is the pile base resistance and
w is the corresponding pile tip displacement.
There are many correlations between base stiffness, E, and CPT cone resistance, qc.
Typically researchers propose that E is proportional to qc i.e. E=αqc where α is a
constant depending on the type of soil. Poulos (1989) recommends α of 5 for
normally consolidated sand and 7.5 for over consolidated sand. These correlations
are usually used for settlement calculations and typically are limited to displacement
of 1-2% of pile diameter. As shown in Figure 4.27, the α ratio is around 7 for a
closed-ended pile with FFR=1 at displacement of 0.1D. This is fairly close to the
recommendations (5 to 7.5) of Poulos (1989), but the level of displacement (0.1D) is
far beyond the normal working displacement limits. At a typical working level
displacement of 0.02D, the best fit α value of 20 for the closed-ended Shenton Park
piles is about 4 times stiffer than that suggested by Poulos (1989). However, for a
large diameter pipe pile, which is usually installed in fully coring mode with FFR=1,
the data on Figure 4.27 suggests lower α values of between 2 and 8.
35
0.01D
30
displacement= 0.02D
25 0.10D
E/q c,avg (-)
20
15
10
0
0 0.2 0.4 0.6 0.8 1
FFR (-)
Figure 4.27 Pile base stiffness as a function of qc,avg [Dutch], FFR and pile tip displacement
4-31
CHAPTER 5 CENTRIFUGE TEST APPARATUS AND
PROCEDURE
5.1 INTRODUCTION
The potential for using the centrifuge modelling technique in civil engineering was
first recognized by the French Engineer Edouard Philips (1869) for modelling the
steel structures. It was not until the 1930s that the first geotechnical centrifuge tests
were reported by Professor Bucky (1931, 1934, 1935). In the past 30 years, the field
has benefited greatly from advances in electronics, miniature instrumentation and
computers. The centrifuge modelling technique has now become increasingly
widespread all over the world with over 100 centrifuges in operation for geotechnical
basic research, preliminary design and educational purposes.
The series of model pile tests discussed in this Thesis was carried out on the fixed
beam geotechnical centrifuge facility at the University of Western Australia (UWA)
to study pile end-bearing performance in uniform and layered soil profiles. In this
Chapter, the modelling principle and experimental apparatus will be introduced first.
The rest of the Chapter then discusses the test programme, sample preparation and
test details.
5.2 CENTRIFUGE MODELLING
The constitutive behaviour of soil is highly nonlinear and depends on current stress
level and stress history. Therefore, to simulate accurately a prototype using a small
scale model, the self-weight stresses must be reproduced in the model. The
centrifuge is considered ideal for producing identical self-weight stresses in the
model to that in the prototype by applying an increased “gravitational” (N times the
Earth’s gravity g) acceleration to physical models. For instance, if a container is
5-1
Chapter 5 Apparatus and Experimental Procedure
filled with soil up to a depth of 0.3m, placed on the centrifuge and subject to a
centrifugal acceleration of 100g, the stresses are increased by a factor of 100 and the
0.3m deep model represents 30m of prototype soil. The centrifuge enables small
scale models in the soil of 0.3m in depth to feel the same stresses as a full scale
prototype in the soil of 30m in depth.
The basic design of all geotechnical centrifuges is very similar and involves a
rotating beam or channel with soil samples at one end. As shown in Figure 5.1, the
applied inertial acceleration field (Ng) is given by ω2r (Equation 5.1) where ω is the
constant angular rotational speed of the centrifuge and r is the radius from the centre
‘o’ to any point in the soil model.
N ⋅ g = ω2 ⋅ r ( 5.1 )
Figure 5.1 Inertial stresses in a centrifuge model induced by rotation about a fixed axis correspond to
gravitational stresses in the corresponding prototype (Taylor 1995)
While the vertical effective and total stresses (σ'v0 and σv0) in the prototype increase
linearly with depth (i.e. constant 1g), the stress field in the model is parabolic due to
a slight variation in acceleration (i.e. variation of Ng with radius r). To minimise the
error in stress distribution, an effective radius, Re, for calculating the rotational speed
ω for a given N value, should be measured from the central axis to one-third the
depth of the model (Schofield 1980, Taylor 1995), so that for most geotechnical
centrifuges (i.e. with a typical radius), the maximum error in stress profile is minor
and generally less than 3% of the prototype stress. In fact, for model tests discussed
5-2
in this Thesis, where the sample height is less than 300mm, this scaling error in stress
distribution difference is quite small and can be safely neglected.
The principles and scaling laws of geotechnical centrifuge modelling have been
explained in detail by Taylor (1995). The scaling laws relevant to the pile tests
discussed in this Thesis are summarised in Table 5.1 below.
Table 5.1 Centrifuge model scaling relationships (Taylor 1995)
Parameter Prototype Model scale Model/Prototype

Gravity level 1g Ng N
Stress σ σ 1
Strain ε ε 1
Density ρ ρ 1
Length L L/N 1/N
Force F F/N2 1/N2
Mass M M/N3 1/N3
5.3 EXPERIMENTAL APPARATUS
5.3.1 Geotechnical centrifuge
The UWA fixed beam geotechnical centrifuge (Figure 5.2) was installed in April
1989 and started spinning in August 1990. The model 661 centrifuge was made by a
French centrifuge manufacturer, Acutronic. It has a maximum payload of 400kg at
up to 100g, decreasing to 200kg at the maximum acceleration level of 200g. The
machine features a swinging platform at a radius of 1.8m from the axis of the
centrifuge, upon which a strongbox containing soil samples can be mounted. As the
centrifuge rotates, the cradle freely swings up such that the increasing acceleration
always acts perpendicularly to the base of the test package. The centrifuge and
associated equipments are described in detail by Randolph et al. (1991).
The strongbox (i.e. the container for the soil sample) is a rectangular box with
internal dimensions of 390mm by 650mm by 300mm high. The strongbox weighs
about 70kg and permit models with up to 250mm of saturated soil (about 200kg) to
be tested at maximum acceleration level of 200g. It is made of separate base and side
plates with watertight seals in between, so that one side may be removed to assist any
particular requirements. Drainage holes were drilled in the base and sides and water
may be directed to a standpipe. There are screw holes at 25mm spacing on the four
5-3
top edges of the box to host the testing equipment (e.g. actuator, camera, light, etc).
The testing area is generally limited as shown in Figure 5.3.
Figure 5.2 Fixed beam geotechnical centrifuge at the University of Western Australia
The electrical powered actuator used for pile installation and static load tests has two
degrees of freedom. Its maximum vertical and horizontal traveling distances are
250mm and 180mm respectively, with a maximum displacement rate of 3mm/s. The
main communication with the centrifuge is via power and instrumentation slip-rings
which are housed in the main axis of the centrifuge. A flight computer is mounted on
the low-g central platform of the centrifuge to handle fast A/D conversions and
provide controls for the driving actuators, as well as the multiplexing. The control
room for the centrifuge houses the main control system for operating the centrifuge,
including two computers for controlling the centrifuge rotational speed and logging
and displaying data respectively, video monitors, and patch board and amplifier racks
for instrumentation. The logging computer is also linked to the flight computer, and
hence allows data transmission and actuator control through data acquisition
software.
5-4
50 25 Holes for the screw 25 50
20 20
25 650 25
105
180 Test Area 390
100 100
105
25 25
20 20
50 25 25 50
Figure 5.3 Plan view of the UWA strongbox (650mm×390mm×300mm) and test area
5.3.2 Actuator stiffness test
The displacement of the model piles was measured using the linear displacement
transducer (LDT) fixed to the top of the driving actuator. As shown in Figure 5.4a,
the actuator was bolted to the strongbox by three screws on both its ends. The
stiffness of this loading system was measured and found to be far more flexible than
expected (ideally, it should be rigid). Consequently the displacement measured by
the LDT includes a displacement component due to movement of the actuator. This
displacement component needs to be determined to correctly interpret the load
displacement behavior of the test piles. A comprehensive series of actuator stiffness
tests was therefore designed and conducted.
The set-up employed for the actuator stiffness determinations is shown in Figure 5.4b.
The actuator bolted to the strongbox was connected to a steel rod with near infinite
stiffness (~1MN/mm), which was then threaded to an 8kN load cell (LC1). A steel
ball was placed in between LC1 and another 8kN load cell (LC2) supported by a
near-rigid steel block placed in the middle of the strongbox. The use of the steel ball
and two load cells reduced the potential for application of any bending moments
during axial loading and also enabled verification of the induced load by comparing
values from both load cells when the actuator was driven downwards. With this set-
5-5
up, the actuator stiffness can be assessed by logging the induced load and the
displacement of the potentiometer by means of the data acquisition system.
Figure 5.4 Set-up of the actuator stiffness test: (a) overview, (b) details of the loading arrangement
6000
5000 Loading stiffness

= 4583N/mm
4000
Load (N)
3000
Loading
Unloading
2000
1000
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3
Displacement (mm)
Figure 5.5 Results of an actuator stiffness test
5-6
Although tests were carried out at various g levels and distances from the side of the
box, results were similar in terms of actuator stiffness. This may be attributed to the
limited testing g level (i.e. 22.5g to 80g) and testing area (Figure 5.3). A typical plot
of the measured compression load versus potentiometer displacement for the actuator
is shown in Figure 5.5. The test consisted of a few cycles of loading and unloading.
As shown, the actuator unloading stiffness is non-linear. On the basis of the linear fit
to the loading cycle, the compression loading stiffness of the actuator for the case
shown on Figure 5.5 is 4583N/mm. The average of a number of tests was found to be
4.5kN/mm, which is in general agreement with the results (i.e. 4kN/mm) reported by
de Nicola (1996).
5.3.3 Model piles and load cells
In total, three model piles with diameters of 6mm, 9.5mm and 16mm (Figure 5.6)
were used throughout the centrifuge test series. The use of piles of different
diameters enabled (i) investigation of any specific scale dependence on pile end
bearing during installation and load testing (ii) modeling of models, i.e. when being
tested at varying centrifugal accelerations, model piles of different diameter could
represent prototype pile of the same diameter, and (iii) examination of the influence
of absolute pile diameter in layered soil. The three model piles were manufactured
from stainless steel with the surface sand blasted to an average centerline roughness
of about 5μm. The details of the pile body design, in particular the pile head and tip,
are shown in Figure 5.7 and 5.8. The model piles are closed-ended with hollow shaft.
The pile head was the same for all three model piles, therefore allowing the use of
the same cap (as discussed in the next section) during pile installation. Each model
pile was fitted with one tip load cell, which measured the pile tip resistance during
installation and static load tests. As shown in Figure 5.8, the three load cells were
instrumented with strain gauges glued on the outer surface. The thickness of the wall
for the load cells was designed to ensure these walls would not undergo strains
greater than 1000μm under the maximum applied load. In addition, silicone rubber
was used to seal gaps between the load cells and pile end to prevent water from
leaking inside the pile and intrusion of sand particles.
5-7
Figure 5.6 Model piles fitted with tip load cells
18 18 18
18 18 18
1 1 1
9
2 23 23
14
Pile Length Pile Body Pile Body Pile Body

215mm
6 9.5 16
(a) (b) (c)
Figure 5.7 Details of pile head design (unit in mm)
5-8
1.5 1 2
3.0 7.5 12
2.5
Wire Outlet M2 Wire Outlet M3
Wire Outlet
2
4.5 2
3.7 4 3 2
Strain Gauges
10 1 1.5 1.5
Strain Gauges
0.75 9 Strain Gauges
9
3 0.5 6 0.5
4.8
0.2 0.4 5 5 M3.5 0.5 M5.5 0.5
4
1.5 1.5
0.5 6 0.5 11 0.5
7.5 14
(a) (b) (c)
6.0 9.5 16
Figure 5.8 Details of pile tip load cells (unit in mm)
Figure 5.9 Set-up of calibration for the tip load cells
5-9
2000
1800
1600
loading
1400
unloading
1200
Force (N)
1000 y = 342.54x + 211.77

2
R = 0.9999
800
600
Calibration Factor (CF) = 342.54 N/Volts
400
200
0
-1 0 1 2 3 4 5
Volts
Figure 5.10 An example of the calibration factor for the 6mm model pile
The axial tip load cells were calibrated by placing a hanger supporting dead weights.
The set-up is shown schematically in Figure 5.9. The model pile was supported by a
fixed beam. The hanger with a conical tip was placed right above pile tip. The dead
weight was loaded onto the hanger and then unloaded incrementally. The calibration
factors were then determined by using a first order linear regression method as
shown in Figure 5.10. The raw values (volts) from each strain gauge were correlated
to the applied vertical load. This calibration procedure was repeated a number of
times for each load cell and little variation in the calibration factors was observed.
The averaged calibration factors derived for the three load cells are summarized in
Table 5.2. Each load cell was calibrated before and after each series of centrifuge
tests, as a check on these calibration factors.
Table 5.2 Summary of the calibration factors for the three tip load cells
Model piles 6mm 9.5mm 16mm

Calibration factor (N/Volt) 343.72 449.00 1627.01
5.3.4 Pile cap and guiding plate
As shown in Figure 5.11, the model piles were jacked into soil sample using a
specially designed pile cap threaded to an 8kN capacity load cell, which was
connected to the driving actuator. The cap was designed so that the model piles could
5-10
be fully unloaded before performing any static load test, i.e. there was no fixed
connection between the model pile and the pile cap and this allowed the development
of residual base load after fully unloading to be recorded by the tip load cell; Teflon
was used inside the pile cap to reduce any possible friction. In addition, as the pile
head was not fixed to the actuator, a guiding plate (Figure 5.12) threaded to the
bottom of the actuator was fabricated to provide additional initial horizontal support
to the pile at the tip level. Teflon plugs manufactured to accommodate piles of
different diameters were used to reduce friction between the pile shaft and the
guiding plate during jacked installation and static load tests.
threaded
to load cell
M10
24
18
Pile Head 20
0.5
Telfon Plug
Figure 5.11 Design of pile cap for jacking installation (unit in mm)
7.5 Telfon
10.5 10.5 26.5 Plug 3.5
Screw 19.5 Screw
Hole Hole 8 16
10.5 27.5 25 3.5

R13.5
Telfon
26.5 Plug 2.5
R12.5
66 26.5 3.5 8 9.5
27.5 25 3.5
Telfon
7.5 26.5 Plug 2.5
Screw 8 6 1.0
Hole 19.5
10.5 27.5 25 3.5

(a) (b)
57
Figure 5.12 Guiding plate, (a) plan view, and (b) cross-section view (unit in mm)
5-11
5.4 TESTING PROGRAMME
In total, 8 centrifuge samples (Table 5.3) denoted as A to H were prepared and tested
during the course of this research. The testing program may be divided into four
groups comprising four different soil profiles as further illustrated on Figure 5.13.
There were four samples (A, B, D and E) of uniform dry sand, two samples (C & G)
of two-layer dry sand, one sample (F) of three-layer dry sand and one sample (H) of
saturated sand over clay. The uniform samples serve as references for the layered
samples.
Figure 5.13 Testing samples, (a) uniform dry sand: A, B, D & E, (b) two-layer dry sand: C & G, (c)
three-layer dry sand: F, and (d) saturated sand over clay: H
A series of static load tests using the three model piles (6mm, 9.5mm & 16mm) were
performed in samples of different relative densities at various depths and acceleration
(Ng) levels. These enabled investigation of the influence of relative density (Dr),
absolute pile diameter (D) and stress level (σ'v0) on the pile end bearing performance,
such as the development of residual stress and the load displacement response. In
addition, two different jacking methods were used: (i) monotonic jacking (MJ)
installation, comprising a monotonic push at 1mm/s to the planned testing depth, and
(ii) jacking (J) installation, comprising cycles of fixed downward displacement
(10mm at 1mm/s) followed by unloading to zero head load at 0.05mm/s. For some
5-12
tests performed in sample H, the rate of penetration was varied from 0.002mm/s to
1mm/s to study penetration rate effects.
The test details are summarized in Table 5.4. In total, 54 good quality model pile
tests were conducted. The maximum number of the tests in each sample is limited by
the boundary consideration. For samples A to G, there are two tests performed using
each pile diameter. Each pile test is represented by 3 letters and 1 number, signifying
the test sample (A to H), model pile diameter (6 for 6mm, 9 for 9.5mm & 16 for
16mm), g level (g1=22.5, g2=37.9 & g3=60) and installation method (J for jacking,
MJ for monotonic jacking). For static load test, the depth No (i.e. -1, -2 & -3) is
attached after the pile test name (e.g. A6g2MJ-1, indicating the static load test for
pile A6g2MJ at depth 1). For sample H, only model piles of 6mm and 16mm in
diameter were used as they represented the two extreme diameters. In total, there are
12 pile tests (8 for 6mm and 4 for 16mm model piles). All piles were installed by
monotonic jacking and tested at acceleration of 80g. Also since there is only one
static load test for each test, the pile test (and the static load test) were represented
simply by one letter and two numbers, i.e. H for the test sample, 6 and 16 for the
diameters of the tested pile, and a number (e.g. -1 to -8) as indication of the number
of tests.
Table 5.3 Summary of the tested samples A to H
Sample Relative density, Dr (%)

No Note
Layer 1 Layer 2 Layer 3
A 54 - - Medium dense sand
B 36 - - Loose sand
C 45 92 - Medium dense / dense sand
D 93 - - Dense sand Dry
E 88 - - Dense sand OCR=1
Dense / Medium dense / Dense

F 92 50 83
sand
G 92 54 - Dense / Medium dense sand
Saturated
H 96 Clay - Dense Sand / Clay
OCR=2.4
5-13
Table 5.4 Test details for samples A to H (unit in model scale)
No Test Name g level Diameter Depth 1 Depth 2 Depth 3

- - - mm mm mm mm
1 A6g2MJ 37.9 6 180.1 - -
2 A6g3MJ 60 6 117.5 184.4 -
3 A9g2MJ 37.9 9.5 180.0 - -
4 A9g3MJ 60 9.5 113.4 179.7
5 A16g1MJ 22.5 16 179.6 - -
6 A16g3MJ 60 16 116.9 187.5 -
7 B6g3MJ 60 6 120.1 180.3 -
8 B6g3J 60 6 180.3 - -
9 B9g3MJ 60 9.5 119.8 179.9 -
10 B9g3J 60 9.5 119.8 179.9 -
11 B16g3MJ 60 16 119.6 179.6 -
12 B16g3J 60 16 119.6 179.5 -
13 C6g3MJ 60 6 119.7 159.8 179.6
14 C6g3J 60 6 119.5 159.3 179.2
15 C9g3MJ 60 9.5 119.3 - -
16 C9g3J 60 9.5 119.1 129.1 -
17 C16g3MJ 60 16 118.4 157.5 -
18 C16g3J 60 16 118.2 157.4 -
19 D6g3J 60 6 119.4 149.5 179.2
20 D6g3MJ 60 6 119.3 149.4 179.2
21 D9g3J 60 9.5 119.0 - -
22 D9g3MJ 60 9.5 118.9 - -
23 D16g3J 60 16 117.9 147.7 -
24 D16g3MJ 60 16 118.0 147.7 -
25 E6g2MJ 37.9 6 119.6 159.5 -
26 E6g3MJ 60 6 119.5 159.4 -
27 E9g2MJ 37.9 9.5 119.3 159.0 -
28 E9g3MJ 60 9.5 118.8 - -
29 E16g2MJ 37.9 16 118.7 157.9 -
30 E16g3MJ 60 16 118.0 137.5 -
31 F6g2MJ 37.9 6 123.8 163.6 183.5
32 F6g3MJ 60 6 123.8 163.6 183.5
33 F9g2MJ 37.9 9.5 123.5 163.3 183.1
34 F9g3MJ 60 9.5 123.4 163.2 -
35 F16g2MJ 37.9 16 123.2 162.6 182.2
36 F16g3MJ 60 16 122.8 162.0 181.5
37 G6g2MJ 37.9 6 123.8 163.7 183.7
38 G6g3MJ 60 6 123.8 163.7 183.6
39 G9g2MJ 37.9 9.5 123.6 163.5 183.4
40 G9g3MJ 60 9.5 123.4 163.3 183.2
41 G16g2MJ 37.9 16 123.1 162.9 182.7
42 G16g3MJ 60 16 122.8 162.4 182.1
43 H6-1 80 6 154.0 - -
44 H6-2 80 6 152.0 - -
45 H6-3 80 6 146.0 - -
46 H6-4 80 6 139.0 - -
47 H6-5 80 6 193.5 - -
a
48 H6-6 80 6 no SLT; fast penetration test
b
49 H6-7 80 6 no SLT; slow penetration test
50 H6-8 80 6 89.0 - -
51 H16-1 80 16 184.0 - -
52 H16-2 80 16 150.0 - -
53 H16-3 80 16 125.0 - -
c
54 H16-4 80 16 no SLT; slow penetration test
a
: constant rate penetration at 1mm/s; b: first 70mm penetration at 0.02mm/s, then the rest at
0.002mm/s; c: first 100mm penetration at 0.02mm/s, then the rest at 0.002mm/s.
5-14
5.5 MATERIAL PROPERTIES
5.5.1 Superfine silica sand
Commercially available super fine (SF) silica sand was used as the sandy soil in all
tests conducted. The soil specific gravity (Gs) is 2.65. The minimum and maximum
void ratios (emin & emax) are 0.49 and 0.78 and the maximum and minimum dry
densities of this sand (γmax & γmin), are 17.4 and 14.6 kN/m3. The particle size
distribution curve is presented in Figure 5.14. The measured D50 of the super fine
silica sand is 0.19mm, where D50 is the average effective particle size.
120
100
Kaolin Clay
Percentage passing (%)
80
60
40
Superfine
20 silica sand
0
0.001 0.01 0.1 1
Particle diameter (mm)
Figure 5.14 Grading curves for materials used in centrifuge tests
5.5.2 Kaolin clay
The commercially available kaolin clay used for the soft clay stratum in the
centrifuge models had already been extensively characterised in previous projects
undertaken in UWA (Stewart 1992, Watson 1999, House 2002) and the properties
are summarized in Table 5.5 below. A grading curve for the kaolin is show on Figure
5.14.
5-15
Table 5.5 Engineering properties of UWA Kaolin clay - after Stewart (1992)
Property Symbol Value

Liquid limit LL 61%
Plastic limit PL 27%
Plasticity index Ip 34%
Soil particle density Gs 2.6
Angle of internal friction φ’ 23o
Critical state frictional constant M 0.92
Voids ratio at p’=1kPa on critical state line ecs 2.14
Slope of normal consolidation line λ 0.205
Slope of swelling line κ 0.044
Parameter Λ=( λ- κ)/ κ 0.785
Spacing ratio r 2.14
5.6 SAMPLE PREPARATION
5.6.1 Sand hopper
The sand samples were prepared by air pluviation with the sand hopper at UWA
(Figure 5.15), whose speed of travel is under automatic control, while the opening
width and the relative height can be adjusted manually. The relative density of the
sample depends on the pouring height and sand flow rate. To achieve a relatively
uniform sand sample, the following steps are necessary: (i) adjust the opening width
of the hopper, which in turn determines the flow rate, (ii) fill the hopper with sand
with the bottom line opening shut, (iii) position the strongbox under the frame and in
line with the traveling path of the hopper, (iv) make reference markings inside the
strongbox and beside the hopper at height intervals of 20mm, (v) adjust the height of
the sand hopper to the specified value, (vi) set the traveling speed and distance of the
hopper, (vii) start the pouring and raise the hopper at every 20mm increase in sample
height and (viii) level the top surface of the sample by a vacuum device.
After completion of above procedures, the total weight of the strongbox soil was then
measured. The weight of the soil was derived by deducting the weight of the
strongbox from the measured total weight. Since the volume of the soil is known, the
dry density (ρd) of the soil can be calculated. Based on the relationship between the
5-16
void ratio (e) and the dry density (ρd), the relative density of the sand sample (Dr) can
be calculated as follows:
e = G sρ w ρ d − 1 ( 5.2 )
D r = (e max − e ) (e max −e min ) ( 5.3 )
Figure 5.15 Sand hopper and example of sand sample preparation
5.6.2 Clay mixer and consolidation press
To prepare the clay sample, the kaolin clay was firstly mixed in slurry form with a
nominal water content of 120% (twice the liquid limit) using a mechanical drum
mixer as shown in Figure 5.16a. Mixing was performed under a 100kPa vacuum for
at least 2 hours before the slurry was carefully scooped from the mixer into the
strongbox. Prior to placing the slurry, a sheet of filter paper was placed over a lower
coarse sand drainage layer with a thickness of about 10mm to prevent mixing of the
sand and clay. Some free water (~5mm) was allowed on the surface of the filter
paper. To ensure that no air is entrapped, the placement of slurry started from the
middle of the strongbox.
When the required quantity of slurry had been placed in the strongbox, a filter cloth
was placed over the surface. The sample was then preconsolidated in the laboratory
5-17
by placing the strongbox in a consolidation press (Figure 5.16b) and loaded by

means of an air pressurized rigid piston. The drainage outlets from the piston and the
base of the strongbox were connected together into an overflow tank so that the top
and bottom of the sample were at the same piezometric level. The consolidation
pressure was applied in two steps, firstly to a low pressure of 20kPa, after the rate of
consolidation fell below 1mm/hr, then to a maximum sample pressure of 85kPa.
Once at the maximum pressure, as shown in Figure 5.17, it normally took at about 70
hours for the completion of primary lab consolidation (rate of settlement dropped
below 0.1mm/hr).
Figure 5.16 Equipments for clay sample preparation, (a) clay mixer, and (b) strongbox in
consolidation frame
100
80
Settlement (mm)
60
40
20
0
0 20 40 60 80 100 120 140 160
Time (hours)
Figure 5.17 Consolidation data for the kaolin clay
5-18
5.6.3 Samples A to G
As discussed earlier, the relative density of the sand sample depends on dropping
height and sand flowing rate (controlled by the traveling velocity and opening width
of the sand hopper). The three parameters for preparation of the dry sand samples (A
to G) are summarised in Table 5.6. For the layered soil samples (C, F & G), before
placing another layer of different relative density, the surface of the current layer was
leveled and the total weight was measured to allow calculation of its relative density.
Table 5.6 Controlling parameters for preparation of samples A to G
Sample Dr Height Opening Dropping Travelling

(%) (mm) width (mm) height (m) velocity (mm/s)
A Layer 1 54 280 4 1.0 65
B Layer 1 36 280 4.5 0.6 65
C Layer 1 45 104 4.2 0.8 65
Layer 2 92 176 1.2 1.1 120
D Layer 1 93 280 1.2 1.1 120
E Layer 1 88 280 2 1.1 120
F Layer 1 92 94 1.2 1.1 120
Layer 2 50 70 4 0.8 65
Layer 3 83 116 2 1.0 100
G Layer 1 92 104 1.2 1.1 120
Layer 2 50 176 4 0.8 65
5.6.4 Sample H
The layered sample H (sand over clay), was prepared to study the influence of a
marked soil stratigraphy change on the pile end-bearing behaviour. In this sample,
the layer of clay was prepared first. When the primary consolidation of the clay was
completed, the strongbox was removed from the consolidation frame and allowed to
swell overnight. The water level was kept at the sample surface, which was covered
with filter cloth. To provide the clay sample with both horizontal and vertical
drainage, at the four corner of the strongbox, four holes were made using the clay
sampler of 23mm in inner diameter into which coarse sand was carefully placed. The
unit weight of the clay was founded to be 17.5kN/m3. At one side of the strongbox, a
5-19
light straw was placed on top of the clay layer to estimate the settlement of the clay
sample under the testing g field. Pore pressure transducers (referred to as STOMPI at
UWA) were installed at one corner of the strongbox to measure the pore pressure
during sample consolidation.
The surface filter cloth cover was removed just before pouring the upper sand layer
and the total weight of the strongbox was measured at this stage. The opening of the
sand hopper was fixed at ~2mm, the drop height at ~1.2m and the traveling speed at
120mm/s to produce a very dense upper sand layer (Dr=96%). The surface of the
sand sample was then leveled using the vacuum cleaner. The measurement of the
total weight of the strongbox enabled determination of the relative density of this
sand layer.
The whole package was then mounted in the swinging platform of the centrifuge.
The clay sample was further consolidated under a g level of 190g for four days
before the package was ramped down to its testing g level of 80 for one day before
any tests started. When the package was first subjected to a constant acceleration of
190g, the pore pressure transducers in the clay showed values generally in excess of
the expected hydrostatic values, although the transducers near the surface usually
indicated values slightly below expected hydrostatic values. Pore pressures were
observed to change slowly and were monitored over the next 48 hours as they fell
(and in some cases rose) to static levels, as the excess pore pressures and suctions
dissipated. For the duration of the test, water was continually fed through a hydraulic
slip ring into the external standpipe at a rate of about 1 litre per hour to counteract
evaporation. Any excess water was allowed to overflow from the standpipe into the
centrifuge enclosure.
5.7 TEST DETAILS
5.7.1 Test Layout
In order to avoid the influence of the strongbox boundary on the penetration

resistance, the testing locations were planned according to the recommendations by
Bolton et al. (1999) for siliceous sands. The distance between the model pile testing
location to the nearest boundary or another testing site was fixed to no less than 10
5-20
times the pile diameter (10D). The pile tip was fixed to in excess of 5 times the pile
diameter (5D) from the rigid bottom boundary. The test layouts for samples A to G
and H are detailed in Figures 5.17 and 5.18 below.
650
100
Potential 6mm Potential
95
9.5mm 16mm 16mm 9.5mm

390
125 125 75 75 125 95 125
Potential 6mm Potential
100
Figure 5.18 Test layout for dry sand samples A to G
650
Water Inlet
Sand Drain
100
6mm 6mm 6mm 6mm
95
16mm 16mm 16mm 16mm

390
150 125 100 125 150

95
6mm 6mm 6mm 6mm
100
Figure 5.19 Test layout for saturated sample H (sand over clay)
5-21
5.7.2 Test procedures
As shown in Table 5.7, the test procedures for the dry samples A to G were generally
similar, but different steps were required for the saturated sample H. The major
differences relate to the waiting time allowed before a load test and the displacement
rate adopted during a load test. The relatively long waiting time (one hour) and slow
loading rate (0.002mm/s) employed for tests in sample H ensured full dissipation of
excess pore water pressure before a load test and fully drained condition during the
load test.
Table 5.7 Summary of the test procedures
1) Mount the testing package (i.e. strongbox containing soil sample) to the swinging
platform of the beam centrifuge;
2) Fix the testing equipments, including actuator, model pile, light and camera, to the
strongbox;
3) Start the data logging system;
4) Accelerate the testing package to the planned g level;
For dry samples A to G For saturated sample H
5) Wait a few minutes; 5) Wait about one hour for pore water
pressure dissipation;
6) Install the model pile to the planned testing depth;

7) Unload the model pile by pulling back the pile cap for a few millimetres;
8) Wait five minutes; 8) Wait one hour for dissipation of

excess pore water pressure;
9) Perform the static load test at loading
rate of 0.05mm/s; 9) Perform the static load test at loading
rate of 0.002mm/s;
10) Unload the model pile;

11) Jack the model pile to another testing depth if applicable;
12) Perform static load test at specified loading rate;
13) Stop the centrifuge and pull out the model pile manually;
14) Move the actuator to another testing location;
15) Repeat procedures 2) to 14) to perform another test.
5-22
CHAPTER 6 CENTRIFUGE TEST RESULTS
6.1 INTRODUCTION
This Chapter presents results and analysis of the centrifuge tests in uniform and
layered soil samples (Table 5.3). A complete list of the pile tests, along with static
load data, can be found in Appendix B, with only typical test results discussed in this
Chapter. Unless stated otherwise, all results are presented in model scale units.
The Chapter concentrates on the end bearing response of model piles during jacking
installation and subsequent static load testing. The analysis of test results is divided
into two parts. Tests performed in samples A to G, which involved dry uniform and
layered sand, are discussed first; the tests conducted in sample H comprising
saturated sand over clay are then evaluated. A detailed synthesis of the results in
layered soil, in the context of a numerical analysis (as discussed in Chapter 7), is
provided in Chapter 8.
6.2 SAMPLES A TO G
6.2.1 Pile Installation
6.2.1.1 Base resistance
As described in Section 5.4, all model pile tests in Samples A to G were installed at a
constant jacking rate of 1mm/s. Therefore, the uniformity of the centrifuge samples
could be assessed by the measured base resistance of the model piles during jacking.
The installation base resistance is termed as qc in this Thesis to be distinct from the
base resistance (qb) mobilised during a static load test. No discussion of shaft friction
(which represented a very small component of the total resistance) is presented here
6-1
Chapter 6 Centrifuge Test Results
as the measured total loads did not have sufficient resolution to enable frictions to be
inferred to the required degree of accuracy.
Figure 6.1 shows the resulting base resistance (qc) during pile installation in the
uniform medium dense sand sample A (Dr=0.54). For the purpose of comparison, the
qc values presented were extracted from the raw data at depth increments of 0.1mm.
This increment is considered sufficiently small for an accurate representation of the
virgin penetration data (but does not show the complete unloading-reloading cycles,
in particular, the residual base resistance qb,residual after unloading). The effects of
qb,residual are discussed separately in the next section.
q c (MPa) COV (%)

0 2 4 6 8 10 12 14 16 18 20 0 2 4 6 8 10 12 14 16 18 20
0 0
20 20
40 Unloading 40
reloading cycles
60 60
80 80
z (mm)
100 100
120 120
140 140
160 A6g3MJ 160

A9g3MJ
180 180
A16g3MJ
200 200
Figure 6.1 An example of measured base resistance during pile installation (Monotonic Jacking)
It can be observed that, for the first ~10mm to ~20mm, values of qc are close to zero
and do not increase with penetration depth (z). This is caused by the ability of the
model pile to fall freely in vertical direction under its own weight at both 1g and at
the target g level at the start of the test. For example, the 6mm model pile could
penetrate into the soil by ~10mm at 1g under its own weight, while there could be
another ~10mm fall when the centrifuge was ramped up to the target g level.
Therefore the recorded qc values over the initial ~20mm were not representative of
6-2
the true soil resistance and should be neglected. The qc profile of the 16mm model
pile is generally much smoother than those of the 6mm and 9.5mm piles, indicating
that the smaller diameter pile is more sensitive to small variations in local soil
conditions.
Also shown in Figure 6.1 is the coefficient of variation (COV) of qc values measured
from the three model piles of various diameters. At a relatively deep penetration, the
COV consistently ranged from 5% to 10% in all samples, reflecting a low degree of
variation in sample relative density using the adopted sample preparation technique.
The degree of non-uniformity could be further improved (COV < 5%) by employing
an automatic sand pouring technique such as recently reported by Madabhushi et al.
(2006).
150
120
90
Q (-)
60
30 A6g3MJ
A9g3MJ
A16g3MJ
0
0 5 10 15 20 25 30 35
Z (-)
Figure 6.2 Normalised base resistance Q plotted against normalised penetration depth Z
The values of qc generally increased linearly with depth z (or stress level σ'v0),
indicating a linear relationship between qb and σ'v0 for a certain relative density; this
trend differs from the nonlinear trend obtained in calibration chamber tests but is
consistent with other centrifuge testing results (Bolton & Gui 1993, Bolton et al.
1999, Deeks & White 2006). For dimensional analysis, the base resistance is
normalised with respect to the overburden stress, and the penetration depth is
normalised with respect to the pile diameter. The normalised base resistance Q and
normalised penetration depth Z as suggested by Bolton et al. (1999) are defined in
6-3
Equations 6.1 and 6.2. The data on Figure 6.1 are re-plotted using these parameters
in Figure 6.2, which clearly shows a two-phase behaviour, depending on the
normalised depth ratio Z (Gui & Bolton 1998). At relatively shallow depths, it is
evident that the normalised resistance ratio Q increases with depth ratio Z, while at Z
>5, a steady state value is attained. The existence of a constant Q value is used in the
next section to confirm a relationship between qc and Dr in centrifuge testing
proposed by Tatsuoka et al. (1990).
Q = (q c − σ v 0 ) σ' v 0 ( 6.1 )
Z=z D ( 6.2 )
Figure 6.3 explores the influence of the installation method on base resistance qc.
The qc profiles measured during jacking and monotonic jacking installation (J or MJ,
described in Section 5.4) in the uniform dense sample D (Dr=0.93) are compared. For
piles of a given diameter, values of qc are very close for the two types of installation.
Slight differences between them may be attributed to the degree of density variation
in the sample as discussed above. In the cycles of the jacked (J) installation, the qc
values are seen to increase quickly during reloading and rejoin the virgin qc profile of
the MJ installation.
qc (MPa) qc (MPa) qc (MPa)

0 10 20 30 40 50 60 0 10 20 30 40 50 60 0 10 20 30 40 50 60
0 0 0
20 D6g3J 20 D9g3J 20 D16g3J

D6g3MJ D9g3MJ D16g3MJ
40 40 40
60 60 60
80 MJ 80 80
z (mm)
z (mm)
z (mm)
MJ
100 100 100
120 120 120 J

J
140 140 140
MJ
160 J 160 160
180 180 180

(a) (b) (c)
200 200 200
Figure 6.3 Effects of installation method (J vs. MJ)
6-4
The effects of soil layering on the pile penetration base resistance are illustrated in
Figure 6.4. In both cases (medium dense sand underlying or overlying dense sand),
the change in the shape of the qc profile clearly indicated the existence of another
layer. Model piles of 6mm, 9.5mm & 16mm in diameter responded differently in
terms of the absolute influence distance over which the pile base senses another layer.
However, this difference is expected to diminish if distances are normalised by the
respective pile diameters. This effect is further explored in Chapter 7.
qc (MPa) qc (MPa)
0 5 10 15 20 25 30 35 40 45 0 5 10 15 20 25 30 35
0 0
(a) (b)
20 20
40 40
Dr=0.45 Dr=0.92
60 60
80 80
z (mm)
z (mm)
100
Interface Interface
100
120 120
140 Dr=0.92 140 Dr=0.50
160 160 G6g3MJ

C6g3MJ
180 C9g3MJ 180 G9g3MJ
C16g3MJ G16g3MJ
200 200
Figure 6.4 Base resistance measured in layered soil, (a) sample C, and (b) sample F
6.2.1.2 Correlation with Dr
As discussed above, the relationship between base resistance and overburden stress
in the centrifuge differs from that obtained in 1g calibration chamber tests. Such an
example is also illustrated in Figure 6.5, with the measured cone tip resistances from
five centrifuge laboratory in Europe showing good consistency (Bolton et al. 1999).
Typical relationships between qc, σ'v0 and Dr proposed based on chamber and
centrifuge CPT results for normally consolidated sand are given in Equation 6.3
(Baldi et al. 1986) and Equation 6.4 (Tatsuoka et al. 1990) respectively. Equation 6.3
predicts a significantly different stress level dependence on qc than Equation 6.4.
Figure 6.5 illustrates this difference for the case of sands with Dr value of 0.84.
6-5
1 qc
Dr = ln
2.41 157(σ' v 0 )0.55
( 6.3 )
⎛ q − σv0 ⎞
D r (%) = −85 + 76 log10 ⎜⎜ c ⎟⎟ = −85 + 76 log10 Q ( 6.4 )
⎝ σ' v 0 ⎠
Figure 6.5 Comparison of measured (Bolton et al. 1999) and predicted CPT results
The values of Dr in this series of centrifuge tests ranged from 0.36 to 0.93 for
samples A to G. The applicability of Equation 6.4 to this series of tests is assessed on
Figure 6.6 by plotting the normalised resistance ratio Q against the soil relative
density; some results from Bolton & Gui (1993) are also included on this Figure. As
shown, Equation 6.4 provides a reasonable fit to the experimental results and would
therefore be recommended for tests in the centrifuge. However, the outstanding
difference between chamber and centrifuge qc values is certainly worthy of further
research.
The base resistance for overconsolidated sand can be inferred from that in normally
consolidated sand based on the recommendation (Equation 6.5) by Schmertmann
(1978), where the resistance ratio (qcOC/qcNC) is a function of the overconsolidation
ratio (OCR).
6-6
q OC
c q cNC ≅ 1 + 0.75 × (K OC
0 K 0NC − 1) ( 6.5a )
K OC
0 K NC
0 = OCR 0.42
( 6.5b )
q OC
c q NC
c ≅ 1 + 0.75 × (OCR 0.42
− 1) ( 6.5c )
120
100 Equation 6.4

Dr(%) = -85 + 76 × log10Q
80
D r (%)
60
40
Samples A-G
Bolton & Gui 1993
20
Tatsuoka et al. 1990
0
0 50 100 150 200 250 300
Q (-)
Figure 6.6 Applicability of Tatsuoka (1990) formulation for this series of centrifuge tests
140
120
100
80
Q (-)
60
40
A16g1MJ
20
A9g2MJ
(a) A6g3MJ
0
0 5 10 15 20 25 30 35
Z (-)
6-7
300
250
200
Q (-)
150
100
E9g2MJ
50
(b) E6g3MJ
0
0 5 10 15 20 25 30
Z (-)
Figure 6.7 Modelling of models (a) in Sample A, Dr=0.54, and (b) in Sample E, Dr=0.88
6.2.1.3 Modelling of models
It was considered of interest to test model piles of different diameters at g levels that
would lead to the same prototype diameter. In some centrifuge tests, model piles of
6mm, 9.5mm & 16mm in diameter were tested at 60g, 39.7g & 22.5g representing a
pile of 0.36m in prototype scale. Comparisons between the measured resistance
ratios, Q, are shown Figure 6.7. There is a favourable agreement between Q values
from model piles tests at these different g levels.
6.2.2 Static Load Tests
6.2.2.1 Resistance mobilisation curves
Figure 6.8 and 6.9 show typical examples of the base resistance mobilised during
jacked installation and static load test. The model pile of 6mm in diameter was
installed at 10mm jacking strokes into the uniform very dense silica sand sample D
(with Dr=0.93) and load tested at 60g at a constant rate of penetration (CPR of
0.05mm/s) at three different depths. The CPR test was employed for all the load tests
in the centrifuge due to simplicity in control and implementation. Deeks & White
6-8
(2006) show that the results from a CRP and a maintained load test (MLT) are
comparable for centrifuge pile tests in dry siliceous sand.
q c (MPa)
0 10 20 30 40 50
20
Pile Test: D6g3J
40 Diameter=6mm
g level = 60
60 Jacked installation
Rate of penetration
80 during installation: 1mm/s
during SLT : 0.05mm/s
100 zoom in
z (mm)
see following figure

SLT at Depth1
120
140
SLT at Depth2
160
180
increase in residual base resistance SLT at Depth3

200
qb,residual with depth (or stress level)
Figure 6.8 Resistance mobilised for pile test (D6g3J) during jacked installation (fully unloading at
depth of every 10mm) and static load tests (SLT)
It is obvious from Figure 6.8 that base resistance is not influenced by the rate of
penetration (from 0.05mm/s to 1mm/s) in the dry sand since there is not a distinct
difference between installation and static loading resistance. Huy et al. (2004) also
demonstrated based in a series of calibration chamber tests that there was only a
slight increase in pile base resistance (~4%) for rates of penetration ranging from
1mm/s to 250mm/s. In absence of the rate effect, the resistance mobilisation curves
in each jacking stroke can also provide additional information. For instance, the
residual base resistance (qb,residual) at the start of each jacking stroke is seen to
increase with depth (or with the effective stress level). However, the resistance
mobilisation curves during a jacking stroke are not treated and analysed in the same
6-9
way as a static load test due to much fewer data points logged at the fast rate. In
particular, the loss of data points at the start of loading (say at typical displacement
of 0.1D) makes detailed and accurate analysis of stiffness and resistance mobilisation
impossible.
qc (MPa)
0 5 10 15 20 25 30 35
115
Static load test

117 D6g3J-1 the q c value measured at
the end of the final jacking
stroke (before a SLT) Depth1
119
121 residual base resistance

(q b,res idual) measured before
z (mm)
static load test (SLT)

123
base resistance (q b )
125 mobilisation curve
127
129 displacement loop
Figure 6.9 Resistance mobilisation curve for D6g3J-1 (zoom in of Figure 6.8)
Figure 6.9 shows the details of the static load test for pile D6g3J at depth 1 (one of
the ‘zoom-in’ sections indicated in Figure 6.8). The base resistance measured at the
end of the final jacking stroke is referred to here as qc, for consistency with
discussions in previous chapters. It represents the steady state resistance, which
theoretically should be independent of the diameter of the pile (or cone) in relatively
uniform soil. In view of the density variation in the centrifuge samples (COV of 5 to
10%), the qc value at the end of the final jacking stroke prior to each SLT is
considered to be most representative and is used later for normalisation purposes.
6-10
At the end of the final jacking stroke, to mimic real pile installation, the pile cap was
normally pulled back a distance of ~1mm, enabling a complete zero head load; this
also allowed residual base loads to develop. Figure 6.9 illustrates that this unloading
process actually occurred very fast once the pile cap was pulled back (at
displacement < ~0.2mm). The residual base resistance (qb,residual) shows a slight
decrease with time. It should be mentioned that the displacement loop shown in
Figure 6.9 reflected only the movement of the pile cap during this unloading-
reloading loop, and not that of the pile. The true displacement of the pile tip was
inferred from that of the pile top allowing for corrections of the actuator stiffness
(~4.5kN/mm, as discussed in Chapter 5) and pile compression.
Figure 6.10 shows the mobilisation of base resistance qb for a series of three static
load tests conducted at different depths. The pile tip displacement (w) has been
normalised by the pile diameter, and the base resistance (qb) normalised by the
corresponding steady state penetration resistance qc. As shown in Figure 6.10a, there
is an increase in qb with increasing pile embedment (or stress level σ'v0), which is
coupled with an increase in the initial slope (or stiffness) of the load-displacements
curves. In addition, qb continues to increase with further displacement, but at a
decreasing rate. This trend is similar to that observed in other centrifuge studies
conducted in UWA (de Nicola 1996, Bruno 1999) and consistent with field
observations (as shown in Figure 2.17). The residual base resistances (qb,residual) also
increase slightly with depth but these are relatively small compared to the values of
qc (i.e. qb,residual=3.5, 4.6 & 5.7MPa, while qc=29, 37 & 45MPa for the three depths).
As shown in Figure 6.10b, when being normalised by qc, there is an excellent
agreement between the normalised resistance mobilisation curves at the three depths,
and qb/qc does show any dependence on depth (or σ'v0). In addition, the initial slopes
of the normalised curves (i.e. qb/qc vs. w/D) are similar in magnitude.
Figure 6.11 shows the mobilisation of base resistance qb for piles of same diameter at
a similar vertical stress level but in samples of different relative densities. As
expected, values of qb (and also qb,residual) increase with an increase in relative density.
It is interesting to note from Figure 6.11b that the initial slopes (stiffness) of the
normalised resistance curves decrease with increase in sand relative density (Dr). At
a relatively small displacement ratio (say w/D=2%), the ratios of qb/qc increase
6-11
significantly (=0.42, 0.64, & 0.78) with a decrease in Dr (=0.93, 0.54 & 0.36).
However, at large displacement ratio (=10%), the ratios of qb/qc range from 0.9 to
0.95 and are less influenced by Dr.
50
D6g3J-3; at z = 179mm
45
qc = 45MPa
40
D6g3J-2; at z = 150mm
35 qc = 37MPa
D6g3J-1; at z = 119mm
30
q b (MPa)
qc = 29MPa
25
20
15
residual base resistance
10 qb,residual = 3.5 MPa for D6g3J-1
= 4.6 MPa for D6g3J-2
5 = 5.7 MPa for D6g3J-3 (a)
0
0 20 40 60 80 100 120 140
w/D (%)
1.2
0.8
D6g3J-1
q b/q c (-)
D6g3J-2
0.6
D6g3J-3
0.4
0.2
(b)
0
0 20 40 60 80 100 120 140
w/D (%)
Figure 6.10 Static load test results for pile D6g3J at three different depths, (a) qb mobilisation curve,
and (b) normalised qb mobilisation curve
The observed trends from Figure 6.10 and Figure 6.11 are consistent with all the
other static load tests conducted in this series of centrifuge testing. Further detailed
analyses of the stiffness and stiffness decay curves from the normalised resistance
curves will be presented in Chapter 8.
6-12
50
(a)
45 D6g3J-3; D r=0.93
qc = 45MPa
40
35
residual base resistance
30 qb,residual = 5.7 MPa for D6g3J-3
q b (MPa)
= 2.2 MPa for A6g3MJ-2

25 = 0.9 MPa for B6g3J-1
20
A6g3MJ-2; D r=0.54
15 qc = 16MPa
10 B6g3J-1; D r=0.36
5 qc = 7MPa
0
0 10 20 30 40 50
w/D (%)
1.2
(b)
B6g3J-1 ; D r = 0.36
0.8
A6g3MJ-2 ; D r = 0.54
q b/q c (-)
D6g3J-3 ; D r = 0.93
0.6
Increase in sand
0.4 relative density D r
0.2
0
0 10 20 30 40 50
w/D (%)
Figure 6.11 Resistance mobilisation curves for pile tests at uniform sand samples of different relative
densities (Dr) but similar stress levels (σ'v0)
6.2.2.2 Residual base resistance qb,residual
Table 6.1 summarises the results (e.g. qc, qb,residual & qb0.1) of the 91 static load tests
performed in samples A to G. As discussed in Chapter 2, it is necessary to assess the
magnitude of the residual base resistance (qb,residual) to accurately predict the load
displacement response of a pile. For all 91 static load tests, values of qb,residual were
6-13
relatively low compared to the steady state resistance qc, with averaged qb,residual/qc
ratios of 0.08 (ranging from 0.02 to 0.18 as shown in Figure 6.12); these ratios were
observed to increase slightly with an increase in σ'v0 and Dr. However, the relatively
small magnitude of qb,residual and the variation in soil density in each sample make
interpretation difficult; and no consistent relationship was found between qb,residual
and the pile and soil parameters.
0.3
Sample A q b,res idual/q c
Sample B
0.25 Sample C average = 0.08
Sample D COV = 0.38
Sample E
Sample F
0.2 Sample G
q b,residual /q c (-)
0.15
0.1
0.05
0
0 10 20 30 40 50
qc (MPa)
Figure 6.12 Normalised residual base resistance (qb,residual/qc) plotted against qc
5
Calculated q b,residual (MPa)
4 q b,res idual /q c=0.08
Alawneh & Husein Malkawi (2000)

0
0 1 2 3 4 5 6
Measured q b,residual (MPa)
Figure 6.13 Predictions of residual base resistance (qb,residual)
6-14
Table 6.1 Summary of the 91 static load test results in samples A to G
No Test Name g level D L σ'v0 qc qb,residual qb0.1

- - - mm mm kPa MPa MPa MPa
1 A6g2MJ-1 37.9 6 180.1 110.2 12.1 1.1 11.1
2 A6g3MJ-1 60.0 6 117.5 113.8 11.1 1.3 10.4
3 A6g3MJ-2 60.0 6 184.4 178.5 16.2 2.2 14.9
4 A9g2MJ-1 37.9 9.5 180.0 110.1 12.2 1.2 10.1
5 A9g3MJ-1 60.0 9.5 113.4 109.8 12.5 0.7 10.4
6 A9g3MJ-2 60.0 9.5 179.7 174.0 19.2 1.9 15.7
7 A16g1MJ-1 22.5 16 179.6 65.2 7.1 0.4 6.6
8 A16g3MJ-1 60.0 16 116.9 113.2 10.0 0.8 9.4
9 A16g3MJ-2 60.0 16 187.5 181.6 17.0 1.2 16.4
10 B6g3MJ-1 60.0 6 120.1 112.7 2.9 0.1 2.9
11 B6g3MJ-2 60.0 6 180.3 169.2 4.7 0.3 4.5
12 B6g3J-1 60.0 6 180.3 169.2 7.0 0.9 6.6
13 B9g3MJ-1 60.0 9.5 119.8 112.5 6.5 0.4 5.6
14 B9g3MJ-2 60.0 9.5 179.9 168.9 10.3 0.8 8.6
15 B9g3J-1 60.0 9.5 119.8 112.5 6.5 0.4 5.6
16 B9g3J-2 60.0 9.5 179.9 168.9 10.3 0.8 8.6
17 B16g3MJ-1 60.0 16 119.6 112.3 5.3 0.2 5.1
18 B16g3MJ-2 60.0 16 179.6 168.6 8.3 0.5 8.0
19 B16g3J-1 60.0 16 119.6 112.3 5.9 0.3 5.8
20 B16g3J-2 60.0 16 179.5 168.5 9.7 0.6 9.3
21 C6g3MJ-1 60 6 119.7 115.3 23.7 1.1 21.1
22 C6g3MJ-2 60 6 159.8 157.1 36.9 3.5 33.2
23 C6g3MJ-3 60 6 179.6 177.7 43.2 5.1 38.4
24 C6g3J-1 60 6 119.5 115.1 23.8 1.6 23.0
25 C6g3J-2 60 6 159.3 156.6 42.0 5.4 37.8
26 C6g3J-3 60 6 179.2 177.2 47.7 7.5 41.0
27 C9g3MJ-1 60 9.5 119.3 114.9 28.4 1.4 22.4
28 C9g3J-1 60 9.5 119.1 114.7 26.7 1.4 22.4
29 C9g3J-2 60 9.5 129.1 125.2 31.8 2.4 27.6
30 C16g3MJ-1 60 16 118.4 114.0 20.4 0.7 18.4
31 C16g3MJ-2 60 16 157.5 154.6 31.3 1.8 28.8
32 C16g3J-1 60 16 118.2 113.8 21.3 0.9 19.5
33 C16g3J-2 60 16 157.4 154.5 36.8 2.2 32.5
34 D6g3J-1 60 6 119.4 124.1 29.2 3.5 26.1
35 D6g3J-2 60 6 149.5 155.5 36.6 4.6 32.8
36 D6g3J-3 60 6 179.2 186.3 45.4 5.7 40.5
37 D6g3MJ-1 60 6 119.3 124.1 31.6 2.7 27.8
38 D6g3MJ-2 60 6 149.4 155.4 40.5 4.0 35.8
39 D6g3MJ-3 60 6 179.2 186.3 48.1 5.5 42.6
40 D9g3J-1 60 9.5 119.0 123.8 34.0 2.4 27.4
41 D9g3MJ-1 60 9.5 118.9 123.7 32.6 2.1 26.2
42 D16g3J-1 60 16 117.9 122.6 23.7 1.3 21.5
43 D16g3J-2 60 16 147.7 153.6 30.6 1.9 27.9
44 D16g3MJ-1 60 16 118.0 122.7 25.1 1.4 22.3
45 D16g3MJ-2 60 16 147.7 153.6 34.0 2.2 30.0
46 E6g2MJ-1 37.9 6 119.6 77.2 19.4 2.5 17.3
47 E6g2MJ-2 37.9 6 159.5 101.4 29.2 3.7 25.9
48 E6g3MJ-1 60 6 119.5 122.1 30.7 3.9 27.4
49 E6g3MJ-2 60 6 159.4 160.5 40.2 5.1 35.8
50 E9g2MJ-1 37.9 9.5 119.3 77.1 18.9 1.0 15.1
51 E9g2MJ-2 37.9 9.5 159.0 101.2 27.7 1.9 22.1
52 E9g3MJ-1 60 9.5 118.8 121.5 30.3 1.8 23.6
53 E16g2MJ-1 37.9 16 118.7 76.7 15.4 0.7 14.6
54 E16g2MJ-2 37.9 16 157.9 100.5 24.9 1.6 22.2
55 E16g3MJ-1 60 16 118.0 120.7 27.4 1.4 23.9
56 E16g3MJ-2 60 16 137.5 139.5 34.0 2.2 29.5
57 F6g2MJ-1 37.9 6 123.8 79.8 9.5 1.1 9.0
58 F6g2MJ-2 37.9 6 163.6 104.0 18.0 1.7 16.1
59 F6g2MJ-3 37.9 6 183.5 116.8 22.5 2.4 20.3
60 F6g3MJ-1 60 6 123.8 126.3 13.0 2.3 12.3
61 F6g3MJ-2 60 6 163.6 164.6 22.3 2.9 20.6
62 F6g3MJ-3 60 6 183.5 184.9 29.3 3.1 26.3
63 F9g2MJ-1 37.9 9.5 123.5 79.6 10.7 0.9 8.9
64 F9g2MJ-2 37.9 9.5 163.3 103.8 21.6 1.7 17.5
65 F9g2MJ-3 37.9 9.5 183.1 116.5 26.7 1.7 21.0
66 F9g3MJ-1 60 9.5 123.4 126.0 15.8 1.3 13.1
67 F9g3MJ-2 60 9.5 163.2 164.2 26.6 1.8 21.6
68 F16g2MJ-1 37.9 16 123.2 79.4 10.2 0.5 9.5
69 F16g2MJ-2 37.9 16 162.6 103.3 18.4 1.1 16.5
70 F16g2MJ-3 37.9 16 182.2 115.9 23.1 1.4 20.8
71 F16g3MJ-1 60 16 122.8 125.3 15.5 0.8 14.2
72 F16g3MJ-2 60 16 162.0 163.0 27.2 1.4 24.0
73 F16g3MJ-3 60 16 181.5 182.9 33.8 2.2 30.4
74 G6g2MJ-1 37.9 6 123.8 80.3 9.9 1.4 9.3
75 G6g2MJ-2 37.9 6 163.7 104.7 12.2 1.3 11.5
76 G6g2MJ-3 37.9 6 183.7 116.9 12.6 1.7 12.0
77 G6g3MJ-1 60 6 123.8 127.1 15.9 1.9 13.2
78 G6g3MJ-2 60 6 163.7 165.7 15.9 2.3 15.1
79 G6g3MJ-3 60 6 183.6 185.0 17.2 2.2 16.3
80 G9g2MJ-1 37.9 9.5 123.6 80.2 10.2 1.2 8.5
81 G9g2MJ-2 37.9 9.5 163.5 104.6 12.0 1.1 10.1
82 G9g2MJ-3 37.9 9.5 183.4 116.7 13.0 1.2 10.8
83 G9g3MJ-1 60 9.5 123.4 126.7 17.1 1.2 14.2
84 G9g3MJ-2 60 9.5 163.3 165.3 21.4 2.0 17.7
85 G9g3MJ-3 60 9.5 183.2 184.6 24.0 2.2 20.0
86 G16g2MJ-1 37.9 16 123.1 79.9 10.3 0.7 9.8
87 G16g2MJ-2 37.9 16 162.9 104.2 13.2 0.9 12.2
88 G16g2MJ-3 37.9 16 182.7 116.3 14.9 1.2 13.9
89 G16g3MJ-1 60 16 122.8 126.2 15.2 0.9 14.1
90 G16g3MJ-2 60 16 162.4 164.4 20.1 1.6 18.5
91 G16g3MJ-3 60 16 182.1 183.5 23.6 1.8 21.2
6-15
The prediction method for qb,residual proposed by Alawneh & Malkawi (2000) as
discussed in Chapter 2 is examined here. Figure 6.13 shows the comparison between
the measured and predicted qb,residual values. Also shown are predictions assuming a
constant qb,residual/qc ratio of 0.08. It is not a surprise to see a significant under-
estimation by the method (Alawneh & Husein Malkawi 2000), which may be
attributed to the relatively small database used for its development.
6.2.2.3 Factors affecting qb0.1/qc
Although the pile base resistance continues to increase with displacement, the
ultimate base resistance is defined at tip displacement of 10% of pile diameter and
referred to as qb0.1. This is consistent with previous concepts and definitions.
In total, 91 good quality static load tests (Table 6.1) were conducted using three
model piles (6mm, 9.5mm & 16mm in diameter) at various stress levels in samples
of different relative densities. These were designed to examine the possible factors
(i.e. qc, D & Dr) affecting the qb0.1/qc ratio of closed-ended piles - as suggested by the
three new design methods discussed in Chapter 3: Fugro-05 (qc), ICP-05 (D) & NGI-
05 (Dr). In the UWA-05 method, a constant ratio of qb0.1/qc is suggested for closed-
ended piles (=0.6 for driven & 0.9 for jacked piles). The proposal of de Nicola (1996)
that qb0.1/qc reduces with effective stress level (σ'v0) is also examined.
Figure 6.14 summarises the variation of resistance ratios (qb0.1/qc) with steady state
resistance (qc), pile diameter (D), sand relative density (Dr) and effective vertical
stress (σ'v0). The resistance ratios (qb0.1/qc) range from 0.78 to 0.98, with an average
of 0.89 and coefficient of variation (COV) of 0.06. There is no systematic variation
of qb0.1/qc values with qc, D, Dr and σ'v0. In addition, the average ratio of 0.89 from
the centrifuge pile tests is in agreement with base capacity database of jacked piles
(i.e. qb0.1/qc=0.89, Figure 3.13).
Although there is evidently no systematic dependence of qb0.1/qc on Dr, it will be

shown from the analyses of base stiffness (discussed in Chapter 8) that there is a
potential for qb/qc to increase with a decrease in Dr (which is a function of qc and σ'v0)
at relative small displacements (say w/D of 2%).
6-16
1.2
0.8
q b0.1 /q c (-)
0.6 q b0.1 /q c
Sample A average = 0.89
Sample B COV = 0.06
0.4 Sample C
Sample D
Sample E
0.2 Sample F
Sample G
(a)
0
0 10 20 30 40 50
qc (MPa)
1.2
0.8
q b0.1 /q c (-)
0.6 q b0.1 /q c
Sample B COV = 0.06
0.4 Sample C
Sample D
Sample E
0.2 Sample F
Sample G
(b)
0
4 6 8 10 12 14 16 18
Diameter (mm)
6-17
1.2
0.8
q b0.1 /q c (-)
0.6 q b0.1 /q c
Sample B COV = 0.06
0.4 Sample C
Sample D
Sample E
0.2 Sample F
Sample G
(c)
0
0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Dr (-)
1.2
0.8
q b0.1 /q c (-)
q b0.1 /q c
0.6 average = 0.89
COV = 0.06
Sample A
Sample B
0.4 Sample C
Sample D
Sample E
0.2 Sample F
Sample G
(d)
0
50 75 100 125 150 175 200
σ 'v0 (kPa)
Figure 6.14 Ratios of qb0.1/qc plotted against (a) steady state penetration resistance, qc, (b) pile
diameter, D, (c) the relative density of the sand, Dr, and (d) vertical effective stress, σ'v0
6-18
6.3 SAMPLE H
6.3.1 Pile installation
In the saturated sample H (OCR ~2.4), a dense sand layer was underlain by a clay
layer to enable study of the pile end bearing response in strongly layered soil with
sharp transition in stiffness and strength. In total, 12 model piles of 6mm and 16mm
in diameter were installed and some were load tested (as summarized in Table 5.4).
Most piles were installed at a relative “fast” penetration rate of 1mm/s, while a few
piles (tests H6-7, H6-8 & H16-4) were installed at a “slow” rate of either 0.02mm/s
or 0.002mm/s.
Figure 6.15 summarises the base resistance measured during the jacked installation
of all piles. The interface between the sand and clay layer is at depth of about 153mm.
In general, the base resistance (referred to as qc here since it is the steady state
penetration resistance) increases with penetration depth, z. Upon reaching its peak
value at z≈100mm, qc reduces significantly while the pile tip is still well above the
interface. When the pile tip penetrates through the sand/clay interface, qc reduces
rapidly to a relative constant value. Moreover, there is a significant difference
between qc values measured by the 6mm and 16mm diameter model piles in the
upper sand layer. For instance, the peak qc values measured at z≈100mm are 20MPa
for the 6mm model pile, but only 14MPa for the 16mm model pile. This difference
indicates that great care must be taken when comparing the resistance of a steadily
penetrating cone (36mm in diameter) and a (much larger diameter) pile in strongly
layered soil. Meyerhof (1983) has suggested correction factors for such a scale effect
based on a constant zone of influence (=10 times pile diameter). However, as will be
discussed in Chapter 7, the zone of influence (distance ahead and behind the pile tip)
is not actually a constant multiple of penetrometer diameter but depends on the
resistance ratio between the soil layers.
Traditionally, pile/cone penetration in sands has been considered to be drained. The

tip resistance qc is therefore assumed to be independent of the rate of penetration.
However, as revealed in Figure 6.15, qc values (in the saturated sand) actually
increase with an increase in the rate of penetration from 0.002mm/s to 1mm/s. For
6-19
example, H6-7, was installed to ~70mm at 0.02mm/s and then to its full penetration
depth at 0.002mm/s. The peak resistance, qc, measured in this test is ~16.5MPa,
which is about 18% lower than that (~20MPa) measured in the “fast” penetration
tests (at 1mm/s). There appears to be no significant difference between qc values at
rates of 0.02mm/s and 1mm/s. As the pile penetrates into the clay, as expected, the qc
values are higher at 0.002mm/s (when penetration is potentially close to being fully
drained) than that at 1mm/s (when conditions are expected to be undrained).
The drop in resistance in sand in the slow tests has an important implication for the
interpretation of the static load tests, which were conducted at the slow rate of
0.002mm/s. As will be shown in the following section, the averaged resistance ratio
(qb0.1/qc) in sand is only 79%, lower than that (89%) observed from field database
and centrifuge tests in dry sand, therefore indicating a strong influence of the rate. In
fact, a sharp drop can be observed when the rate was switched to 0.002mm/s (as
shown in Figure 6.15b for H16-4).
Leung et al. (1996) argue that qc dependency on the rate of penetration may be
attributed to the creep and particle crushing behaviour of the sand at pile tip. They
inferred from centrifuge pile tests and one dimensional compression tests that stress
relaxation (creep) takes place at and around a pile tip area due to progressive
breakdown of sand particles under sustained load (or at very slow penetration) under
both fully saturated and dry conditions. Eiksund & Nordal (1996) reported a series of
model pile tests in dry dense silica sand under loading rates of 0.8 to 1100mm/s,
which gave almost the same resistance, indicating no loading rate effect (i.e. as
indicated in the tests on Samples A-G). However, Al-Mhaidib (1999) demonstrated
using 45 model pile tests in loose, medium dense and dense dry sand samples, that
the bearing capacity of piles increases with an increase in the rate of loading (from
0.0002mm/s to 0.0167mm/s) and the relationship between the bearing capacity and
loading rate can be represented by a straight line on a log-log plot. It is obvious the
loading rate in these tests is considerably lower than in the tests of Eiksund & Nordal
(1996) and it would appear that, at these slow rates, the creep of the sand under the
pile tip is significant. The rate of penetration in the slow test in sample H is
0.002mm/s which falls into this slow range and explains the lower resistance being
measured at this rate.
6-20
qc (MPa)
0 2 4 6 8 10 12 14 16 18 20 22
0
H6-1
20 H6-2
H6-3
H6-4
40 H6-5
H6-6
H6-7
60 H6-8
H6-7; penetration rate:
80 0.02mm/s for first 70mm &
z (mm)
0.002mm/s for the rest

100
120 1mm/s for the first ~100mm
& 0.002mm/s for the rest
140
Sand
160 Clay
H6-1, H6-2, H6-3, H6-4, H6-5 & H6-6
180 penetration rate: 1mm/s
(a)
200
qc (MPa)
0 3 6 9 12 15
0
H16-1
20 H16-2
H16-3
40 H16-4
60 the depth at which

rate switched from
80 0.02 to 0.002mm/s
z (mm)

100 0.02mm/s for first ~100mm
& 0.002mm/s for the rest
120
H16-1, H16-2 & H16-3
140 penetration rate: 1mm/s
Sand
160 Clay
180
(b)
200
Figure 6.15 The penetration resistance for all the model pile tests, (a) 6mm, and (b) 16mm
6-21
6.3.2 Static load tests
As discussed above, the end resistance mobilised at a penetration rate of 0.002mm/s

is lower than that at 1mm/s. Figure 6.16 further illustrates this effect for a typical
static load test (H6-8) conducted in the sand layer. The qc value at the final jacking
stroke (at 1mm/s) is ~20MPa. For the very first displacement of 0.25D (1.5mm), the
pile was load tested at 0.002mm/s and there is an 8% drop in base resistance
(qb=18.5MPa). Then the rate of penetration was increased to 0.05mm/s, and finally to
1mm/s. It is obvious from the Figure that there is an increase in qb with higher rates.
At depth (z) of 97.5mm, the rate was switched back to 0.002mm/s for the rest of the
pile installation. Once again, there is an abrupt drop in the base resistance after
lowering the rate (20MPa compared to 17.5MPa).
q c (MPa)
0 5 10 15 20 25
80
Static load test: H6-8
82
rate of penetration
84 during jacking: 1mm/s
86 qb,residual
88
SLT; rate: 0.002mm/s
qc = 20MPa
z (mm)
qb = 18.5 MPa
90
rate: 0.05mm/s
92
94
rate: 1mm/s
96
98
rate: 0.002mm/s
100
Figure 6.16 Resistance mobilised for pile test H6-8 at various rates
Table 6.2 summarizes the results (e.g. qc, qb,residual & qb0.1) of the 9 static load tests
conducted in sample H. Figure 6.17 plots the residual base resistance ratio
(qb,residual/qc) against pile embedment length. Although the magnitude of qb,residual (0.8
to 2.8MPa) is still as small as those observed in samples A to G, the ratios of
6-22
qb,residual/qc are considerably higher, especially for pile tests close to the sand clay
interface. In fact, for one load test in clay (H6-5), the qb,residual/qc ratio reaches 0.88
(with qc=1.13MPa). The reason is because, although the pile tipped in clay or close
to the clay interface, the pile shaft was mostly embedded in sand. Upon unloading,
the tendency for pile rebound mobilized negative skin friction along the pile shaft,
which was balanced by the residual base resistance (qb,residual). Therefore, in this
sample, qb,residual/qc increased with the pile embedment length, as qb,residual remained
relatively constant while qc reduced.
Table 6.2 Summary of static load test results for sample H
Test
No g level D L σ'v0 qc qb,residual qb0.1
Name
- - - mm mm kPa MPa MPa MPa
1 H6-1 80 6 154 130 2.1 1.1 2.0
2 H6-2 80 6 152 127 2.6 1.2 2.1
3 H6-3 80 6 146 123 4.3 1.8 3.1
4 H6-4 80 6 139 118 8.1 2.8 5.8
5 H6-5 80 6 194 155 1.1 1.0 1.4
6 H6-8 80 6 89 74 20.1 2.7 18.3
7 H16-1 80 16 184 147 1.3 0.8 1.5
8 H16-2 80 16 150 125 2.2 0.9 1.8
9 H16-3 80 16 125 105 6.0 1.4 4.3
1
q c=1.13 MPa
0.9
Sand/Clay
0.8 Interface
0.7
q b,residual /q c (-)
0.6 1.32
0.5 2.13
4.25 2.62
0.4 2.16
8.11
0.3
5.99
0.2
20.11
0.1 Sand Clay
0
80 90 100 110 120 130 140 150 160 170 180 190 200
Pile embedment length (mm)
Figure 6.17 Residual base resistance ratio (qb,residual/qc) plotted against pile embedment length
6-23
The ultimate base resistance ratio, qb0.1/qc, defined at 10% of pile tip displacement, is
summarized in Figure 6.18. It ranges from 0.7 to 0.9 for static load test in sand and
0.96 to 1.27 for those in clay. The low ratios of qb0.1/qc in sand can be attributed to
the creep behaviour of sand beneath the pile base at slow loading rate (i.e.
0.002mm/s). In addition, the presence of an adjacent clay layer has the potential to
cause reduced base stiffness (as discussed in Chapter 8) and therefore lower
resistance at limited displacement (say 10% of pile diameter).
1.4
H6-5
Sand/Clay
1.2 Interface
H16-1
1 H6-8 H6-1
H16-2
q b0.1 /q c (-)
0.8 H6-2
H16-3 H6-3
H6-4
0.6
0.4
0.2
Sand Clay
0
80 90 100 110 120 130 140 150 160 170 180 190 200
Pile embedment length (mm)
Figure 6.18 Summary of resistance ratios (qb0.1/qc) for pile tests in sample H
6.3.3 Post sample excavation
Immediately after completion of all the tests in this sample, the water was drained
out the soil sample. One side of the strongbox was taken off to assist close
examination of the testing area around the tested pile. As shown in Figure 6.19, a
steel blade was used to cut off the sticky hard clay, which was constantly lubricated
with water. As the sample excavation proceeded, the testing areas could be exposed,
such as those shown in Figure 6.20 for three tested with the 6mm diameter piles. It
was noticed that a sand cone formed immediately under the pile tip. The size of sand
cone evaluated for each test is summarized in Figure 6.21. The ratios of the height of
6-24
the sand cone to the pile diameter (hcone/D) range from 0.5 to 1.0, with an average
value of 0.7.
Figure 6.19 Photo of the sample cutting
Figure 6.20 Photo of the exposed testing areas around three 6mm model pile tests
6-25
1.2
0.8
hcone /D
0.6
0.4
0.2 hcone
0
0 20 40 60 80 100
Distance into the clay layer (mm)
Figure 6.21 Height of the sand cone formed below the pile tip
During the sample excavation, a soil pocket penetrometer (PP, as shown in Figure
6.22) was used to examine the shear strength variations of the clay layer. Tests were
performed in a number of locations so that the strength data over the full depth of the
clay and along the strongbox width (650mm) could be obtained. Figure 6.23
summarises the results of such a simple investigation. It is obvious that values of PP
undrained strengths su-PP are highest in the middle of the strongbox with lower values
close to the sides of the strongbox. Apart from locations close to the bottom drainage
sand layer, su-PP values, as expected, generally increase with depth with minimum
values recorded close to the sand clay interface. At locations nearest the bottom
coarse sand drainage layer, values of su-PP are relative low as well.
6-26
Figure 6.22 Photo of post sample examination using the soil pocket penetrometer
80
70
60
50
s u-PP (kPa)
40
30
15
20 40
distance (mm) from the
bottom of the strongbox = 65
10 90
115
0
0 65 130 195 260 325 390 455 520 585 650
Distance from the side of the strongbox (mm)
Figure 6.23 Distribution of PP undrained shear strength during sample post-excavation
6-27
CHAPTER 7 ANALYSIS OF qb IN LAYERED SOIL
7.1 INTRODUCTION
The literature review presented in Chapter 2 demonstrated that there is a lack of

understanding of the pile base resistance behaviour in layered soil. Centrifuge results
reported in Chapter 6 illustrates that pile base resistance in the strong soil is
influenced by the approaching weak soil at a greater distance from the layer
boundary, whereas base resistance in the weak soil is less influenced by the
neighbouring strong soil. This is in general agreement with previous chamber studies
in layered soil in the literature. However, there remains a need to quantify the
influence of stiffness and strength difference between layers on the base resistance
performance.
This Chapter presents the results of PLAXIS numerical analyses that idealise model
pile penetration in layered soil assuming the spherical cavity expansion analogue.
The primary goal is to investigate factors affecting the zone of influence and the
reduction of steady state penetration resistance due to the presence of neighbouring
weak layers. This approach is considered an improvement on a similar elastic
analysis in layered soil presented in Chapter 2, primarily because the zone of
influence is likely to be highly dependent on the stiffness non-linearity of the soil.
Firstly the accuracy of the proposed numerical model for spherical cavity expansion
was tested by comparing computed pressure-expansion curves with those given by
the closed-form solutions of Yu & Houlsby (1991) for expansion in an elastic
perfectly plastic Mohr-Column soil. Thereafter, the more advanced non-linear elastic
Hardening Soil model as described by Schanz et al. (1999), which is coded in
PLAXIS, is employed for the analysis in layered soil profiles. Detailed parametric
studies are then performed for a two-layer soil profile (e.g. weak underlain or
overlain by strong soil). As a result, procedures are suggested to evaluate the zone of
7-1
Chapter 7 Analysis of qb in Layered Soil
influence and reduction of steady state resistance in a two-layer soil system. Finally,
base resistances computed in two different three-layer soil profiles are presented and
discussed.
7.2 SPHERICAL CAVITY EXPANSION IN PLAXIS
7.2.1 Use of Plaxis
As discussed in Chapter 2, the steady state base resistance qb (or qc) is related to the
cavity expansion limit pressure, plimit. It has been only possible to solve plimit
analytically in some special cases and numerical treatment is necessary especially
when adopting more realistic soil constitutive models. The analytical solution to
cavity expansion in layered soil has only be derived in a linear elastic soil system
(e.g. Sayed & Hamed 1987). The commercial finite element code PLAXIS 8.2 is
therefore used here for the cavity expansion analysis in layered soil; this program
includes the Mohr-Coulomb (MC) and Hardening Soil (HS) constitutive models. The
updated mesh option available in PLAXIS (based on updated Lagrangian approach)
also allows more realistic modelling of the cavity expansion involving large strain
deformation.
7.2.2 Soil models
There are several built-in constitutive models in PLAXIS. For validation of the
proposed numerical procedure (discussed in later sections), the elastic perfectly
plastic Mohr-Coulomb (MC) is used. However, as pointed out by Salgado et al.
(1997), the use of a linear elasticity, perfect plastic model is of limited relevance in
practice. Therefore, the Hardening Soil (HS) model is employed for cavity expansion
analysis in layered soils. In this section, the constitutive models (MC and HS) are
briefly introduced and their predictions for a drained triaxial test is then compared.
The MC model requires the following input parameters, i.e. E or G (Young’s or shear
modulus), ν (Poisson’s ratio), c (cohesion), φ (friction angle) and ψ (dilation angle).
The HS model resembles a MC model, but has additional features such as a limited
dilation of interfaces, stress and strain dependent Young’s modulus and a yield
7-2
function that includes a cap to model plasticity effects due to increase in isotropic
stress. Some basic characteristics of the HS model are listed in Table 7.1.
400
(a)
350
300
250
q' (kPa)
200
150
100 HS Model
50 MC Model
0
0 0.01 0.02 0.03 0.04 0.05 0.06
ε 1 (-)
100000
90000 (b) Drained Triaxial Test
σ'3=100kPa
80000
70000
60000
E (kPa)
50000
40000
30000
20000
10000
0
0.00001 0.0001 0.001 0.01 0.1
ε 1 (-)
Figure 7.1 Results of a drained triaxial test using HS and MC models, (a) principal stress difference
versus axial strain, and (b) modulus versus axial strain
To further illustrate the difference between HS and MC models, they are employed
to simulate a drained triaxial test. A triaxial test can simply be modeled by means of
an axisymmetric geometry of unit dimensions (1x1), which represents a quadrant of
the soil specimen. The analysis procedure can be found in the Material Models
Manual of PLAXIS 8.2. A set of model parameters as listed in Table 7.2,
representing dense sand, is considered.
7-3
Table 7.1 Basic characteristics of the Hardening Soil (HS) model
Description Parameters
Stress dependent stiffness according to a power law m
Plastic straining due to primary deviatoric loading E50ref
Plastic straining due to primary compression Eoedref
Elastic unloading/reloading Eurref, νur
Reference pressure pref
Failure according to the Mohr-Coulomb criterion c, φ and ψ
The results of the drained triaxial are presented in the Figure 7.1 in terms of stress-
strain and modulus degradation curves. As shown, in contrast to MC model, the
deviator stress-strain relationship of HS model is much more typical of real soil. For
the MC model, the stiffness remains at a constant value (50000kPa as specified
initially), whereas the stiffness decreases with increase in strain for HS model.
Table 7.2 Soil parameters for HS and MC models
Parameter HS MC Unit
E50ref 50000 50000 b kPa
Eurref 50000 NA kPa
Eoedref 150000 NA kPa
pref 100 NA kPa
Cohesion, c 0.2 0.2 kPa
Friction angle, φ
o
40 40
Dilation angle, ψ 10 a o
10
Poisson’s ratio, νur 0.2 0.2 -
Power, m 0.5 NA -
K0nc (=1-sinφ) 0.357 0.357 -
a
, limited dilation applies (emin=0.49, emax=0.78 & einitial=0.50); b, Young’s Modulus for MC model
7.2.3 Mesh set-up
All calculations were performed with an axi-symmetric mesh comprising triangular

elements each with 15 nodes and 12 gauss stress points, as shown in Figure 7.2. The
radius of the mesh domain is 12m and the height is 24m. The initial radius of the
spherical cavity (referred to as a0) represented by a linear elastic soil cluster is 0.1m
unless otherwise stated. A typical mesh for analysis of spherical cavity expansion in
uniform soil is shown in Figure 7.3. The expansion of the cavity can be achieved by
applying positive volumetric strain to the spherical soil cluster. The radial
displacement and pressure during expansion can be obtained by selecting appropriate
nodes and gauss points for output. In all the analyses, to be consistent in data output,
the spherical soil cluster was always divided into four triangular elements as shown
7-4
in Figure 7.4. Nine nodes (denoted as A to I) and ten stress points (denoted as J to S)
were selected for output of radial displacements and expanding pressure respectively
based on the symmetry of the problem. The maximum number of nodes or stress
points allowed in PLAXIS is ten. The average of the results from those nodes and
stress points enable the determination of the pressure expansion curve.
Figure 7.2 Nodes and stress points for the triangular element
Figure 7.3 Typical mesh for spherical cavity expansion in PLAXIS
7-5
A
B
C J
K
D L
M N
E
O P
Q
R
F
S
G
H
I
Nine Nodes: A to I Ten Stress Points: J to S
Figure 7.4 Selected nodes and stress points from the spherical soil cluster for output
To study the dependence of results on mesh coarseness, three calculations following

the analysis procedure detailed in the next section were performed with a coarse
mesh (~200 elements), medium coarse mesh (~700 elements) and fine mesh (~2000
elements). As illustrated in Figure 7.5, the difference in the pressure expansion
curves is small. This indicates the results are not sensitive to the mesh coarseness
provided sufficient elements are around zones close to the cavity. To be optimum in
both calculation accuracy and time, the medium coarse mesh with about 700
elements was employed for all analyses.
700
600
Expanding pressure, p (kPa)
500
ref
E 50 =5000kPa
ref
400 p =120kPa
o
φ=20
300 o
ψ=0
200
Coarse; 200 elements
100 Medium; 700 elements
Fine; 2000 elements
0
0 0.1 0.2 0.3 0.4 0.5
Radial displacement, u (m)
Figure 7.5 Pressure expansion curves by coarse, medium coarse and fine meshes
7-6
7.2.4 Analysis procedure
The procedure for the numerical analysis of spherical cavity can be divided into four
sequential steps: (i) “input”, including mesh set-up and initial stress generation, (ii)
“calculate”, including defining calculation phases and selecting nodes and stress
points for output, (iii) “output”, including extracting data by the PLAXIS curve
program, post-processing data to generate pressure expansion curves and evaluation
of the limit pressure plimit, and (iv) calculation of steady state resistance qb from plimit.
Step by step instructions are detailed below:
(i) Set-up and generate the axi-symmetric mesh. The size of the mesh is 12m in
radius and 24m in height with a cavity of 0.1m initial radius located at depth of
12m. At zones close to the cavity, local mesh refinement is necessary. The soil
cluster for the cavity is always divided into four triangular elements. Standard
fixity is applied to the mesh (total fixity at the bottom, free at the top and free
in vertical direction only in left and right hand boundaries). MC or HS models
are assigned to the surrounding soil depending on the type of analysis, while
the soil cluster inside the cavity is linear elastic with E=2000kPa and ν=0.2.
(ii) Generate water and initial stress condition. The water level is at the surface and
the initial effective stresses are generated following the K0 procedure (σ'v0=γ'z;
σ'h0=K0×σ'v0). Value of K0 is set to be unity in all cases.
(iii) Select the nodes and stress points for output. According to the symmetry of the
problem, nine nodal points (A to I) and ten stress points (J to S) are selected.
As will be discussed later, the average of the radial displacement of the nine
nodes and average of the radial stress of the ten stress points are used to
generate the pressure expansion curves.
(iv) Define the calculation phases. The first calculation phase generates initial
stresses and ensures equilibrium. Positive volumetric strain is then applied
incrementally (i.e. +10% in each phase) to the cavity soil cluster in subsequent
calculation phases. This will result in incremental expansion of the cavity. For
this large strain problem, it is necessary to use the Updated Mesh option in
PLAXIS. Moreover, in the iterative procedure, the “arch-length control option”
needs to be de-selected as it will sometimes result in unstable displacement.
7-7
(v) Open the curve program in PLAXIS, plotting, for example, the radial
displacements from the nine nodes against the maximum effective principal
stresses (i.e. σ'1, the expanding pressure) from the ten stress points.
(vi) Extract the data into a spreadsheet (excel) for post processing. The results from
the ten stress points and nine nodes are averaged to give the pressure (p) and
radial displacement (u). The maximum pressure (pmax) from the p-u curve is
determined as the limit pressure plimit.
(vii) As discussed in Chapter 2, this limit pressure (plimit) is then related to the
steady state end resistance, qb (or qc). The relationship between qb and plimit
from a spherical cavity expansion analysis is reintroduced in Equation 7.1 for
completeness.
q b = p lim it (1 + tan α ⋅ tan φ) ( 7.1 )
where φ is the internal angle of friction and α equals to 60o for a standard cone
penetrometer, but can be taken as (45+φ/2) for a pile.
7.3 NUMERICAL MODEL VERIFICATION
7.3.1 Closed-form solutions to limit pressure plimit
The pressure expansion curve and limit pressure (plimit) for spherical cavity
expansion can be evaluated using the closed-form expression by Yu & Houlsby
(1991). The solutions are based on an elastic-perfectly plastic soil with a Mohr-
Coulomb failure criterion and a constant rate of dilation. The steps for constructing
pressure expansion curve and calculating limit pressure are given in the following:
(a) Choose input parameters for the soil, i.e. E (Young’s modulus), ν (Poisson’s
ratio), c (cohesion), φ (the friction angle), ψ (the dilation angle), and p0 (in situ
mean effective stress). Parameter ‘m’ is used to indicate cylindrical analysis
(m=1) or spherical analysis (m=2). The initial radius of the cavity is referred as
‘a0’, while the current radius during expansion is referred as ‘a’.
(b) Calculate the following terms from the input parameters:
7-8
E ( 7.2 )
G=
2(1 + ν )
E ( 7.3 )
M=
1 − ν (2 − m )
2
2c cosφ ( 7.4 )
Y=
1 − ν 2 (2 − m )
1 + sin φ ( 7.5 )
α=
1 − sin φ
1 + sin ψ ( 7.6 )
β=
1 − sin ψ
α ⋅ (β + m ) ( 7.7 )
γ=
m ⋅ (α − 1) ⋅ β
Y + (α − 1) ⋅ p 0 ( 7.8 )
δ=
2 ⋅ (m + α ) ⋅ G
⎧ (β + m ) ⋅ (1 − 2 ν ) ⋅ [Y + (α − 1) ⋅ p 0 ]⋅ [1 + (2 − m ) ⋅ ν]⎫ ( 7.9 )
η = exp ⎨ ⎬
⎩ E ⋅ (α − 1) ⋅ β ⎭
ξ=
[1 − ν ⋅ (2 − m )]⋅ (1 + m ) ⋅ δ
2
( 7.10 )
(1 + ν ) ⋅ (α − 1) ⋅ β
⎡ m ⋅ ν ⋅ (α + β ) ⎤
× ⎢α ⋅ β + m ⋅ (1 − 2 ν ) + 2 ν −
⎣ 1 − ν ⋅ (2 − m )⎥⎦
(c) for pressure ‘p’ less than the pressure ‘p1’ required to initiate plasticity,
p1=2mGδ+p0, calculate the current cavity radius ‘a’ from the small-strain
elastic expression (a-a0)/a0=(p-p0)/(2mG).
(d) for a given value of ‘p’ (greater than p1 and less than the limit pressure
plimit) ,calculate the cavity pressure ratio ‘R’ from the following Equation:
R=
(m + α ) ⋅ [Y + (α − 1) ⋅ p]
α ⋅ (1 + m ) ⋅ [Y + (α − 1) ⋅ p]
( 7.11 )
(e) Evaluate a/a0 from the following Equation and the radial displacement ‘u’
(u=a-a0):
β (β + m )
a ⎧ R −γ ⎫
=⎨ ⎬ ( 7.12 )
a 0 ⎩ (1 − δ )(β+ m ) β − (γ η)Λ 1 (R , ξ ) ⎭
Evaluate sufficient terms in the infinite series to obtain an accurate value of

Λ1 – a few terms will usually be sufficient:
7-9
∞
Λ 1 (x , y ) = ∑ A1n ( 7.13 )
n =0
A 1n = y n ln x n! if n = γ
A 1n =
yn
n!(n − γ )
[ ]
x n−γ − 1 otherwise
( 7.14 )
The procedure from (d) to (e) can be repeated to construct the complete cavity
pressure expansion relationship. By setting (a/a0) to approach +∞ in Equation 7.12,
the limit pressure plimit can then be obtained. It was found that plimit depends strongly
on both the angle of friction and the angle of dilation, as well as the stiffness
properties of the soil.
7.3.2 Comparisons with numerical results
To validate the accuracy of the PLAXIS approach, the computed pressure-expansion

relationship are compared with those of the closed-form solutions deduced, as
described in the previous section, employing the Equations of Yu & Houlsby (1991).
In total, as listed in Table 7.3, five cases (MC1 to MC5) were considered. Cases
MC1, MC2 and MC3 cover a wide range in values of E, φ and ψ, which leads to
strong differences between the respective limit pressure values. In addition, MC4 and
MC5 are cases whose initial radius of the cavity for PLAXIS mesh input is 0.3m and
0.5m instead of 0.1m, while keeping the size the mesh unchanged as 12m×24m. The
two analyses were intended to confirm the adequacy of the mesh size and the
presumed normalisation (discussed later).
As shown Figure 7.6, the numerical results, obtained using the finite element
technique (PLAXIS updated mesh analysis), generally show very good agreement
with closed-form solutions. The difference between the limit pressure calculated by
PLAXIS and the closed-form solution is within 10%, except for case MC5, where the
initial cavity radius is 0.5m. This can be explained by the fact that as the initial cavity
radius increases, the ratio of mesh size to cavity radius is greatly reduced from 120
(=12/0.1) to 24 (=12/0.5). The small difference in calculated limit pressures between
MC3 and MC4 also confirms that the selected mesh boundary (12m × 24m) is
sufficient to represent an infinite large soil mass for a spherical cavity with radius ≤
0.3m. In addition, for this particular problem, it is better to have a small cavity to
7-10
ensure minimal variation in initial mean effective stress (p0) inside the cavity. For
these reasons, an initial radius of 0.1m is employed for all the following analyses,
unless otherwise stated.
Table 7.3 Material parameters and initial stresses adopted in the verification calculations
Name p0 a0 E ν φ ψ p1 kPa* p2 kPa* (p2-p1)/ p2

kPa m MPa − o o
Analytical PLAXIS %
MC1 120 0.1 5 0.2 20 0 8916 8433 -5.4
MC2 120 0.1 100 0.2 42 12 550 571 3.8
MC3 120 0.1 50 0.2 40 0 3180 3075 -3.3
MC4 120 0.3 50 0.2 40 0 3180 2956 -7.0
MC5 120 0.5 50 0.2 40 0 3180 2763 -13.1
*
limit pressure, plimit
600
(a)
500
400
p (kPa)
300 MC1 Closed-form

MC1 PLAXIS
200
MC1: a 0=0.1m
100
E=5MPa, φ=20 , ψ =0
o o
0
1 2 3 4 5 6
a/a0 (-)
10000
(b)
9000
8000
7000
6000
p (kPa)
5000
4000 MC2 Closed-form
3000 MC2 PLAXIS
2000 MC2: a 0=0.1m

E=100MPa, φ=42 , ψ =12
o o
1000
0
1 2 3 4 5 6
a/a0 (-)
7-11
3500
(c)
3000
2500 MC3 Closed-form

a 0 =0.1m
MC3 PLAXIS; a0=0.1m
2000
p (kPa)
a 0 =0.3m
MC4 PLAXIS; a0=0.3m
a 0 =0.5m
MC5 PLAXIS; a0=0.5m
1500
1000 MC3, 4 &5: a 0=0.1m,0.3m & 0.5m

E=50MPa, φ=40 , ψ =0
o o
500
0
1 2 3 4 5 6
a/a0 (-)
Figure 7.6 Comparison between closed-form and numerical (PLAXIS) results: (a) MC1 (b) MC2 and
(c) MC3, MC4 & MC5
7.4 ANALYSIS IN TWO-LAYER SOIL PROFILE
7.4.1 Methodology
The steady state penetration resistance qb can be related to plimit using the relationship
given in Equation 7.1. To study layering effects on plimit (thus qb), numerical
spherical cavity expansion analyses can be performed in layered soil and results can
then be compared with equivalent expansions in a one layer soil. The idea for cavity
expansion in a two-layer soil profile is schematically shown in Figure 7.7. The cavity
with an initial radius of 0.1m is located at depth of 12m. The distance from the centre
of the cavity to the interface of two soil layers is defined as Hcavity, which is ‘positive’
if the cavity is embedded in the strong soil and otherwise is negative. Two basic
layered profiles have been examined: (a) strong underlain by weak soil, and (b) weak
underlain by strong soil.
In this study, the boundary of the two soil layers is altered with reference to the
location of the centre of the cavity at a depth of 12m i.e. changing Hcavity from
positive to negative values examines the motion of the cavity moving from the strong
soil into the weak soil. This approach avoids the need to evaluate the values of plimit
(or qb) in a one layer soil at separate depths. For instance, in Figure 7.7a, for a given
Hcavity value less than 12m (say 2m), the calculated value of qb can be compared with
7-12
the value of qb,S in one layer of strong soil (Hcavity=12m). The value of qb,S needs to
be evaluated only once as the effective stress and stiffness are those predefined at
depth of 12m. The qb/qb,S ratio is essentially a resistance reduction factor arising due
to the presence of another weak layer.
In summary, five steps were followed for the analyses in layered soil: (i) calculate
plimit in a one soil layer (strong Hcavity=12m, or weak Hcavity=-12m), (ii) calculate plimit
in a layered soil where Hcavity can be varied from positive to negative values by
moving the elevation of the boundary between the two soil layers, (iii) evaluate qb
from the limit pressure plimit; the resistances in a one layer, strong or weak soil, are
represented by qb,S and qb,W respectively, (iv) compare qb in layered soil with qb,S,
and (v) evaluate the influence of Hcavity on ratios of qb/qb,S. An example of such an
approach will be discussed in the following sections.
Figure 7.7 Schematic drawing of analysis in a two-layer soil profile: (a) strong over weak soil, and (b)
weak over strong soil
7-13
7.4.2 Analysis program
A relatively extensive parametric study was carried out with soils of various stiffness
and strength. An effective stress analysis with the hardening soil model was
employed for the study (except for one case where a total stress analysis was used for
the clay with undrained shear strength su specified). The basic hardening soil model
parameters for the soils are listed in Table 7.4.
Table 7.4 Basic hardening soil model parameters
Basic parameters Value Unit

Eurref/E50ref 3 -
Eoedref/E50ref 1 -
pref 120 kPa
K0 1.0 -
γsat 20 kN/m3
c (cohesion) 0.2 kPa
ν (Poisson’s ratio) 0.2 -
In total, 19 cases were analyzed for cases involving a wide range of resistance ratios
between the strong and weak soil layers. Thee cases can be generally divided into
five groups, as presented in Table 7.5 and summarized as follows:
Group1: The limit pressure plimit is known to depend strongly on Young’s modulus
(E), friction angle (φ) and dilation angle (ψ). The dilation angle ψ was set to zero in
this set of analyses. The relative weak soil is always underlain by a stronger soil.
Different values of E50ref (5, 10, 25 and 50MPa) and φ (20o, 30o and 40o) are
employed to study the individual and combined effects of stiffness and friction angle
on the cavity pressure expansion curves.
Group2: The effect of the dilation angle is studied in this group of analyses. Keeping
E50ref and φ, constant, values of ψ were specified as 0o (from Group 1), 5o and 10o.
The weaker soil is underlain by stronger soil.
Group3: More realistic soil parameters are selected to represent dense sand, loose
sand and clay in this group of analyses. The sand is always assumed fully drained,
while the clay is either drained or undrained. The exponent ‘m’ (see Table 7.1) is 0.5
for the sand and 1.0 for the clay. The initial void ratio is 0.5 (Dr=97%) for the dense
sand and 0.68 (Dr=34%) for the loose sand with emin=0.49 and emax=0.78. Dilation
ceases when the void ratio reaches emax. The five cases in Group3 involve undrained
7-14
clay underlain or overlain by dense sand (No10 & 11), loose sand underlain or
overlain by dense sand (No12 & 13) and drained clay underlain by loose sand (No
14).
Group4: To further increase the resistance difference between the strong and weak
soil (i.e. a lower qb,W/qb,S ratio), the stiffness (E50ref ) and strength parameters (φ and
ψ) of the dense sand were changed from 50MPa, 40o and 10o in Group3 to 100MPa,
42o and 12o in the Group 4 analysis. The initial void ratio was again set equal to 0.5.
The four cases in this group are drained or undrained clay underlain or overlain by
dense sand.
Group5: In this last analysis, the effective stress analysis with hardening soil model
for the clay is replaced by a total stress analysis. The undrained clay has a constant
undrained shear strength su of only 6kPa. The parameters for the strong soil/dense
sand are the same as those employed for the Group 4 analyses.
Table 7.5 Analysis program in a two-layer soil system
Group No Upper Layer Lower layer

E50ref φ ψ Type E50ref φ ψ Type
Group1 1 5 40 0 Drained 50 40 0 Drained
2 10 40 0 Drained 50 40 0 Drained
Group3 10 5 20 0 Undrained 50 40 10 Drained
11 50 40 10 Drained 5 20 0 Undrained
16 5 20 0 Undrained 100 42 12 Drained
Group5 19* 5 - - Undrained 100 42 12 Drained
* the undrained shear strength, su=6kPa.
7-15
7.4.3 Pressure expansion curves
For each of the 19 cases listed in Table 7.5, analyses were performed with typical
Hcavity values of +12m, +6m, +3m, +2m, +1m, +0.5m, +0.2m, 0m, -0.5m, -1.5m, -3m,
and -12m. By definition, for cavity expansion in a one layer soil, either strong or
weak, the value of Hcavity equals to +12m and -12m respectively. The number of
analyses for Hcavity<0m is less than that for Hcavity>0m. The reason is because once
the cavity is embedded in the weak soil, the limit pressure plimit is largely controlled
by the soil properties of the weak soil, while the neighbouring strong soil has less
influence. Values of plimit for Hcavity<0m are therefore broadly similar. However, if
embedded in strong soil, the value of plimit could be significantly influenced by the
neighbouring weak soil and varies over a significant range of Hcavity values.
Figure 7.8 presents a typical example of the pressure expansion curves for case No18
in Group 4, in which a drained dense sand (Dr=97%) is underlain by undrained soft
clay. It can be observed that as the cavity approaches the clay layer (i.e. as Hcavity
reduces), the pressure at a given radial displacement is considerably reduced, except
at the very start of the pressure expansion curve. Additionally, as enlarged in Figure
7.8b, once the cavity is located in the clay layer, the pressure is largely controlled by
the clay properties and 20 times (~500kPa compared to ~10,000kPa) lower than that
in the sand.
12000
(a)
case: No18 Hcavity = +6m & 12m
10000
Expanding pressure, p (kPa)
Hcavity = +3m
Hcavity = +2m
8000
Hcavity = +1m
6000
4000 Hcavity = +0.5m
Hcavity = +0.2m
2000
Hcavity = 0m
0
0 0.05 0.1 0.15 0.2 0.25
7-16
500
(b)
Expanding pressure, p (kPa) 450 case: No18 Hcavity = -0.5m
400
Hcavity = -1.5m, -3.0m & -12m
350
300
250
200
150
100
50
0
0 0.05 0.1 0.15 0.2 0.25
Figure 7.8 Pressure expansion curve for case No18, (a) Hcavity≥0, and (b) Hcavity<0
The maximum expanding pressure (i.e. plimit=pmax) at the corresponding radial

displacement, u, are tabulated in Table 7.6. It shows that values of ‘u’ reduce as the
Hcavity reduces. In cases for which Hcavity is negative, there is no peak pressure
obvious from the pressure expansion curve, and therefore the pressure at u=0.2m (or
a/a0=3) is taken, with little loss in accuracy, as plimit. The results for all 19 cases are
summarised in Appendix C in similar format as shown in Figure 7.6 and Table 7.6.
Table 7.6 Summary of plimit and corresponding radial displacement ‘u’ for case No18
Hcavity (m) plimit u (m) a/a0

Sand Only; +12 10140 0.2018 3.0
+6 10084 0.2062 3.1
+3 9362 0.1240 2.2
+2 8534 0.0669 1.7
+1 6328 0.0539 1.5
+.5 3889 0.0261 1.3
+0.2 2117 0.0272 1.3
0 842 0.0500 1.5
-0.5 565 0.2000 3.0
-1.5 451 0.2000 3.0
-3 418 0.2000 3.0
Clay Only; -12 403 0.2000 3.0
7-17
7.4.4 Interpretation of results
Results described in the previous section illustrate that the cavity limit pressure (or
maximum achievable pressure) plimit in the strong soil can be reduced significantly
due to the presence of a neighbouring weak layer. Before generalizing all the results,
it is necessary to return to the relationship between a pile of diameter ‘D’ and a
cavity of current radius ‘a’. This is shown schematically in Figure 7.9. The distances
from the pile tip and cavity centre to the interface between two layers are defined as
H and Hcavity respectively. Based on the assumption that the angle of the soil wedge,
α, equals 45+φ/2, the distance H can be derived as a function of distance Hcavity and
current cavity radius ‘a’, i.e. H=Hcavity±a×sinφ. In addition, the pile diameter ‘D’ can
be related to current cavity radius ‘a’ as, D=2a×cosφ. Therefore, the normalized
distance from the pile tip to the interface of the two soil layers can be derived as:
H H cavity ± a ⋅ sin β H cavity

= = ± 0.5 tan φ ( 7.15 )
D 2a ⋅ cos β 2a ⋅ cos φ
where ‘a’ is the current cavity radius (a=a0+u), and φ is the soil friction angle. The
sign ‘+’ applies for strong over weak soil, and ‘-’ applies for weak over strong soil.
Figure 7.9 Schematic diagram of relationship between geometry of the cavity and the pile
7-18
The resistance ratio of qb1/qb2 between two soil layers can be evaluated using
Equation 7.16, and, to reflect the resistance reduction due to the presence of another
weak layer, the resistance qb in the two-layer soil system is best normalized by the
resistance qb,S in the strong soil only.
q b1 p lim it1 1 + tan φ1 ⋅ tan (45 + φ1 2 )

= ×
q b 2 p lim it 2 1 + tan φ2 ⋅ tan (45 + φ2 2 )
( 7.16 )
1.1
case No18
1
Normalised resistance, q b/q b,S (-)
0.9
0.8
0.7
0.6
0.5
Weak Soil: Strong Soil:
0.4
Undrained Clay Dense Sand
0.3
q b,W /q b,S=0.02 0.2

0.1
0
-20 -10 0 10 20 30
Normalised distance, H/D (-)
Figure 7.10 Resistance reduction as a function of normalised distance (case No18)
With Equations 7.15 and 7.16, the resistance reduction curve can be constructed,
where the resistance ratios qb/qb,S are plotted against the normalized interface
distance H/D. One such example is shown in Figure 7.10 for case No18, as discussed
in previous section. The qb/qb,S ratio decreases rapidly from unity as pile tip
approaches the interface (as H/D reduces), and finally to qb,W/qb,S=0.02 when the pile
enters the weak layer with H/D<0. The zone of influence over which the pile does
not sense the presence of the weak layer is shown to be 14D for case No18.
The normalization scheme as discussed above can be further validated by performing

the same analyses but with a different initial cavity radius (i.e. not 0.1m). For
instance the analysis for case No6 is repeated by keeping the same soil properties for
the strong and weak soil, but using a cavity of initial radius a0 of 0.3m. The ratios of
qc normalised by the value of qcS with no presence of the weak layer are plotted
against H and H/D respectively in Figure 7.11. As shown, when plotted against H,
7-19
the resistance reduction curves predicted using the two initial radii differ from each
other. However when plotted against H/D, they fall into one curve.
1.1
(a)
1
0.9
0.8
0.7
q b/q b,S (-)
0.6
0.5 a0=0.1m
a 0=0.1m
0.4 a0=0.3m
a 0=0.3m
0.3
0.2
0.1
0
-6 -4 -2 0 2 4 6 8 10 12
Distance, H (m)
1.1
(b)
1
0.9
0.8
0.7
q b/q b,S (-)
0.6
0.5
aa0=0.1m
0=0.1m
0.4
aa0=0.3m
0=0.3m
0.3
0.2
0.1
0
-6 -4 -2 0 2 4 6 8 10 12 14 16
Normalised distance, H/D (-)
Figure 7.11 Comparison of results with different initial radii (a0=0.1m & 0.3m) for case No6
7.4.5 Curve fitting
The resistance reduction curves for all the 19 cases are summarised in Figure 7.12.
This Figure highlights the importance of the soil properties in each of the two soil
layers; full details of each resistance reduction curve are made available in Appendix
C. The resistance ratios (qb,W/qb,S) cover a wide range from 0.006 (case No19) to 0.8
(case No8) for the 19 cases examined. The resistance reduction curve depends on the
7-20
stiffness and strength properties of both the strong and weak soil layers, and is
therefore found to be a function of the resistance ratio (qb,W/qb,S) of the two layers.
1.1
No8: q b,W/q b,S=0.80 1

0.9
0.8
0.7
q b/q b,S (-)
0.6
Weak Soil Strong Soil
0.5
0.4
0.3
0.2
0.1 No19: q b,W/q b,S=0.006

0
-20 -15 -10 -5 0 5 10 15 20 25 30
H/D (-)
Figure 7.12 Summary of resistance reduction curves for all 19 cases
To quantify the dependence of the resistance reduction curve on the ratio of qb,W/qb,S,
it is necessary to present the resistance reduction curves (i.e. qb/qb,S plotted against
H/D) in mathematical format. After trial and error, the algebraic equation in Equation
7.17 was found to represent the resistance reduction curves calculated by PLAXIS
reasonably well.
y = A 0 + (1 − A 0 ) ⋅ exp[− exp(A1 + A 2 ⋅ x )] ( 7.17 )

where variables ‘y’ and ‘x’ represent resistance ratio (qb/qb,S) and normalized
distance (H/D) respectively, the constant A0 equals to resistance ratio between the
weak and strong soil (qb,W/qb,S) and the constants A1 and A2 are fitting coefficients
which provide best agreement between the fitting curve and resistance reduction
curve computed by PLAXIS. Coefficient A1 controls the relative position of the
fitted curve in the horizontal axis, while A2 controls the shape of the fitted curve (i.e.
the rate of resistance reduction with H/D).
One such fitting exercise is shown in Figure 7.13 for cases No8 and No19 discussed
above. The dotted curves are the calculated results, while the smooth curves are fitted
curves using Equation 7.17, but with different best fit A1 and A2 values. For case No8,
7-21
where A0 is 0.80, with A1=0.15 and A2=-0.80, the fitted curve fits the PLAXIS
predictiosn reasonably well. Similarly for case No18, where A0 is only 0.006, with
A1=1.1 and A2=-0.32, the fitted curve represents the PLAXIS results quite well.
1.1
Fitted curve with
1
A0= qb,W/qb,S= 0.80
A1= 0.15 and A2= -0.80 0.9
0.8 Fitted curve with

No8: q b,W/q b,S=0.80
0.7 A0= qb,W/qb,S= 0.006
q b/q b,S (-)
A1= 1.1 and A2= -0.32

0.6
W eak Soil 0.5 Strong Soil
0.4
Curve fitting:
y=A0 +(1-A0 )exp[-exp(A1 +A2 x)] 0.3
where
0.2
y=q b/q b,S; x=H/D
A0 =q b,W /q b,S 0.1
No19: q b,W/q b,S=0.006
0
-20 -15 -10 -5 0 5 10 15 20 25 30
H/D (-)
Figure 7.13 Comparison of deduced (Figure 7.12) and calculated (Equation 7.17) resistance reduction
curves
1.5
(a)
1.2
1
Fitting coefficient, A
0.9
0.6 A1 = -0.22ln(A0 ) + 0.11

2
R = 0.96
0.3
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
A0 (-)
7-22
-0.2
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
-0.3
2
Fitting coefficient, A (b)
-0.4
-0.5
A2 = -0.11ln(A0 ) - 0.79
2
-0.6 R = 0.92
-0.7
-0.8
-0.9
A0 (-)
Figure 7.14 Fitting coefficients A1 and A2 plotted against resistance ratio A0 (=qb,W/qb,S)
Similar exercises were performed for all the 19 cases and are summarized in
Appendix C. The fitting coefficients A1 and A2 were found to be a function of ratios
of qb,W/qb,S (defined as A0). Current data presented in Figure 7.14 suggest a
logarithmic relationship between A1 (or A2) and A0. However, it should be noted that
the lowest resistance ratio A0 used to derive such a relationship was 0.006. Therefore,
for lower ratios (<0.006), as summarized in Equations 7.18 and 7.19, upper limits
(1.5 and -0.2) are suggested to be applied for fitting coefficients A1 and A2.
A1 = −0.22 ln (A 0 ) + 0.11 ≤ 1.5 ( 7.18 )

A 2 = −0.11 ln(A 0 ) − 0.79 ≤ −0.2 ( 7.19 )
7.4.6 Zones of influence
The extent or zone of influence is important to pile design in layered soil. For
example, a decision must be made regarding the relative level at which the pile end
bearing is not significantly influenced by the neighbouring weak soil.
The definition of the zone of influence is illustrated in Figure 7.15 by the resistance
reduction curve of case No18 (as discussed above). Theoretically, the zone of
influence in the strong soil (ZS) is defined as distance up to where the resistance qb
equals to the unaffected resistance in the strong soil qb,S. In other words, the
resistance ratio qb/qb,S equals unity when the pile tip is outside of the zone of
influence (ZS). Similarly, the zone of influence in the weak soil (ZW) is defined as
7-23
distance down to where the resistance qb equals to the unaffected resistance in the
weak soil qb,W. The total zone of influence ZI is the sum of ZS and ZW and defines a
distance over which the pile base resistance is a function of its relative tip level and
the relative stiffness and strength of the strong and weak soil layers. Outside this
zone of influence, the pile base resistance in the layer is not affected by presence of
another layer. For instance, in Figure 7.15, if positioned at 4D (H/D=4) above the
interface of a strong over weak soil system, there will be around 55% reduction in
resistance with qb/qb,S=0.45. However, at a distance greater than 14D, the resistance
ratio is unity and not influenced by the underlain weak layer. The zone of influence
in the strong soil (ZS) is therefore 14D. In contrast, the zone of influence in the weak
soil (ZW) is much less, i.e. ZW < 2D. This is consistent with previous discussions that
once the pile tip is in the weak soil, the base resistance is largely controlled by the
soil properties of the weak layer.
40
q b/qb,S =1.0
H: Distance from pile tip to the interface
D: pile diameter
30 H/D: Normalised distance
ZS: zone of influence in the strong soil
ZW: zone of influence in the weak soil
20
H/D
10
H/D=4; ZS /D
qb/qb,S=0.45
Interface of two soil layers

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1
ZW /D <2 q b/q b,S
-10
q b,W /qb,S =0.02

-20
Figure 7.15 Definition of the zone of influence
7-24
16
(a)
14
12
10
Z S /D
8 ZS/D = -2.52ln(A0) + 1.41

2
6 R = 0.92
4
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
A0 (-)
1.4
1.2
0.8 ZW /D = 0.26ln(A0) + 1.38

Z W /D
2
0.6 R = 0.90
0.4
0.2
(b)
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
A0 (-)
Figure 7.16 Normalized zone of influence plotted against resistance ratio A0 (=qb,W/qb,S)
Therefore, values of zone of influence (ZS, ZW & ZI) for all the 19 cases analysed can
be determined from the derived resistance reduction curves (as summarised in
Appendix C). Here, values of ZS are taken at a resistance ratio (qb/qb,S) of 0.95,
instead of unity. This is because, in practice, if the resistance ratio is greater than
0.95, the influence of the weak layer can be considered negligible. On the other hand,
if the resistance in the weak soil qb,W is considerably lower than that in the strong soil
qb,S (say qb,W/qb,S < 5%), values of ZW are defined at a resistance ratio of
0.05+0.95qb,W/qb,S. This definition avoids subjective interpretations of ZW in case of
small qb,W/qb,S ratios.
7-25
It is found that the zone of influence depends strongly on the resistance ratio between
the weak and strong soil qb,W/qb,S (referred as A0). Results are summarised in Figure
7.16, where the normalised zones of influence (ZS/D & ZW/D) are plotted against A0.
The best fitted curves to the data on Figure 7.16 show a logarithm relationship
between the zone of influences and A0 (=qb,W/qb,S), which is presented by Equations
7.20 to 7.22 below.
ZS D = −2.52 ln (A 0 ) + 1.41 ( 7.20 )

Z W D = 0.26 ln (A 0 ) + 1.38 ( 7.21 )
Z I D = −2.26 ln (A 0 ) + 2.79 ( 7.22 )
7.5 ANALYSIS IN THREE-LAYER SOIL PROFILE
7.5.1 Introduction
Spherical expansion numerical analyses in the two-layer soil profile provide a basis
for estimating pile steady state penetration resistance in layered soil profile. For
example, for the case shown in Figure 7.17a, tip resistance measured from a standard
cone (DCPT=36mm) can be regarded as the true steady state resistance. The resistance
ratio between the two soil layers can be taken as the measured resistance ratio of the
cone, qc,W/qc,S, and the resistance reduction curve can then be constructed as
described in Section 7.4. For a pile of diameter (D) much larger than that of a
standard cone (DCPT), a resistance reduction factor (qb/qb,S) can be calculated for any
H/D value using Equation 7.17. The steady state resistance for the pile at a distance
relatively close to the interface can then be obtained by applying this resistance
reduction factor to the measured resistance of the cone at the same level.
In this section, the discussion is extended to pile end bearing in three-layer soil
profiles, such as shown by Figure 7.17b & 7.17c. Compared to the two-layer soil
system, an additional parameters, TS (thickness of the strong soil layer) or TW
(thickness of the weak soil layer), in addition to H (defined as distance to the
interface in a two-layer soil profile, and the shorter distance to the two interfaces in a
three-layer soil profile) also play an important role in determination of the adjusted
resistance. The same numerical analysis procedure was employed as for a two-layer
soil profile, but with the spherical cavity embedded in a three-layer soil system.
7-26
Attention is focussed on the base resistance in a strong soil layer due to presence of
the neighbouring weak soil layers (i.e. H≥0). This is because the base resistance in
the weak soil layer is not significantly influenced by a neighbouring strong layer (e.g.
ZW is less than 2D). Also, in practice, end bearing piles are rarely tipped into a weak
soil layer (e.g. clay) if there is a more competent adjacent strong layer (e.g. dense
sand).
Figure 7.17 Schematic drawing of piles in layered soil profile (a) two-layer, (b) weak over strong over
weak, and (c) strong over weak over strong
7.5.2 Weak/Strong/Weak
As shown in Figure 7.17b, the three-layer system comprises a middle strong soil
layer underlain and overlain by weak soil layers (i.e. weak-strong-weak). Compared
to the two-layer soil profile in Figure 7.17a, there is an additional underlain weak
layer. Therefore, the resistance in the strong soil will further reduce due to the
introduction of another weak layer below. As the thickness of the strong soil layer
increases, the resistance reduction factors in the strong soil will be the same as those
in a two-layer soil profile. To investigate the degree of further resistance reduction
due to another weak layer, one series of PLAXIS analyses were performed by
varying both H (actually varying Hcavity in the program) and TS. Due to the limited
7-27
scope of this study, only one set of soil parameters, which were the same as those for
case No18 (see Table 7.5), was investigated.
Table 7.7 Analysis program in the three-layer soil profile (see Figure 7.17b)
Hcavity TS-Hcavity H/D (TS-2H)/D plimit qb,3LS/qb,2LS

(m) (m) (-) (-) (kPa) (-)
0.2 0.2 0.4 0.9 1836 0.87
0.2 0.5 0.5 2.3 2012 0.95
0.2 1.0 0.6 5.2 2117 1.00
0.2 2.0 0.6 10.3 2117 1.00
0.2 12-0.2 * 0.6 63.3 2117 1.00
0.5 0.5 2.2 0.9 3306 0.85
0.5 1 2.3 3.6 3772 0.97
0.5 1.5 2.1 5.9 3889 1.00
0.5 2 2.1 8.5 3889 1.00
0.5 12-0.5 * 2.2 62.3 3889 1.00
1 1 4.7 0.9 5442 0.86
1 1.5 4.2 3.2 5988 0.95
1 2 4.0 5.3 6284 0.99
1 3 4.3 10.3 6321 1.00
1 12-1 * 3.9 49.0 6328 1.00
2 2 8.0 0.9 7789 0.91
2 2.5 7.6 2.9 8247 0.97
2 4 7.1 8.4 8461 0.99
2 6 7.2 16.3 8525 1.00
2 12-2 * 7.6 41.2 8534 1.00
3 3 10.3 0.9 9036 0.97
3 4 9.6 4.2 9270 0.99
3 6 8.8 10.1 9362 1.00
3 12-3 * 8.6 27.9 9362 1.00
6 6 14.5 0.9 9934 0.99
6 7 14.9 3.5 10033 1.00
6 9 14.3 8.3 10079 1.00
6 12-6 * 12.7 14.1 10084 1.00
*
at TS =12m, this is the same as a two-layer soil profile (i.e. weak over strong soil layer)
The analyses performed and corresponding results are summarised in Table 7.7. The
spherical cavity was again fixed in position at depth of 12m (see Figure 7.7), with
Hcavity varied from 0.2m, 0.5m, 1m, 2m, 3m and 6m. Then the underlain weak layer
was introduced by reducing the thickness of the strong soil from 12m to TS=2Hcavity.
At any given value of Hcavity, by comparing the calculated resistance (denoted as
qb,3LS) in the three-layer soil profile (2Hcavity<TS<12m) with that (denoted as qb,2LS) in
a two-layer soil profile (TS=12m), a further resistance reduction factor (i.e.
qb,3LS/qb,2LS) due to another weak layer is quantified.
As shown in Figure 7.18, the ratios of qb,3LS/qb,2LS are plotted against normalised
distance, (TS-2H)/D, and grouped as one curve for similar H/D values. Ratios of
7-28
qb,3LS/qb,2LS at certain H/D values increase rapidly with (TS-2H)/D and exceed 0.95 at
(TS-2H)/D values in excess of about 3.0. The lowest resistance reduction factor
(qb,3LS/qb,2LS=0.85) occurs when the pile tip is embedded near the middle of a
relatively thin strong soil layer.
1.00
0.98
0.96
0.94 qb,3LS/qb,2LS=0.95
q b,3LS /q b,2LS
0.92
average
0.90
H/D=0.6
0.88
H/D=2.2
0.86 H/D=4.2
H/D=7.5
0.84
(TS-2H)/D=3.0 H/D=9.3
0.82 H/D=14.1
0.80
0 2 4 6 8 10 12 14 16
(TS -2H)/D
Figure 7.18 Resistance ratio between qb in a two-layer and a three-layer soil profile
In summary, to evaluate the resistance reduction in a three-layer soil profile (i.e.

Figure 7.17b), there are two steps to follow: (i) choose the distance closest to the
interfaces as H, based on which, the resistance reduction factor (qb/qb,S) in a two-
layer soil profile can be first evaluated from Equation 7.17, and (ii) evaluate the
additional resistance reduction factor (qb,3LS/qb,2LS) due to another weak layer from
Figure 7.18. The final resistance reduction factor is a product of the two factors
discussed above. It is noted that in a three-layer soil profile, the most critical
evaluation of the resistance reduction depends on ratios of qb/qb,S in a two-layer soil
profile, which could range from as low as 0.006 to 1.0, while the further resistance
reduction factors of qb,2LS/qb,3LS are in general greater than 0.85 for a ratio of qb,W/qb,S
as low as 0.02.
7.5.3 Strong/Weak/Strong
The existence of a clay seam below the pile tip can significantly reduce pile base
capacity (and stiffness). As shown in Figure 7.17c, the three-layer soil system
7-29
consists of a weak soil layer embedded in the middle of two strong soil layers. The
thickness of the weak layer (TW) is as critical as the distance to the interface (H) in
determination of the pile base resistance. The resistance (qb) at certain distance will
increase to the unaffected resistance (qb,S) in strong soil as the thickness of the weak
layer reduces to zero. To quantify this effect, one series of PLAXIS analysis was
performed by varying both H and TW. As for the analysis described in Section 7.5.1,
only one set of soil parameters (the same as case No18), was investigated.
Table 7.8 Analysis program in the three-layer soil profile (see Figure 7.17c)
Hcavity TW H/D TW/D plimit qb/qb,S

(m) (m) (-) (-) (kPa) (-)
0.2 0.0 a 0.9 0.0 10140 1.00
0.2 0.1 1.2 0.4 4371 0.43
0.2 0.2 1.5 1.0 3324 0.33
0.2 1.0 1.4 5.0 2472 0.24
0.2 3.0 1.4 14.0 2123 0.21
0.2 12.0 b 1.5 63.5 2117 0.21
0.5 0.0 a 1.6 0.0 10140 1.00
0.5 0.2 2.5 0.8 6385 0.63
0.5 0.5 2.8 2.3 5043 0.50
0.5 1.0 2.8 4.7 4513 0.45
0.5 2.0 3.0 10.1 4171 0.41
0.5 12.0 b 3.1 64.0 3889 0.38
1.0 0.0 a 2.7 0.0 10140 1.00
1.0 0.2 3.5 0.6 8242 0.81
1.0 0.5 4.2 1.9 7362 0.73
1.0 1.0 4.3 3.8 6740 0.66
1.0 2.0 4.9 8.9 6411 0.63
1.0 12.0 b 4.8 52.5 6328 0.62
2.0 0.0 a 4.9 0.0 10140 1.00
2.0 0.5 6.2 1.4 9299 0.92
2.0 1.0 6.9 3.2 8882 0.88
2.0 2.0 6.8 6.3 8583 0.85
2.0 3.0 8.6 12.2 8556 0.84
2.0 12.0 b 8.5 48.4 8534 0.84
a
, one layer of strong soil; b, two-layer soil profile (i.e. strong over weak)
The analyses performed and corresponding results are summarised in Table 7.8. The
spherical cavity was again fixed in position at depth of 12m (see Figure 7.7), with
Hcavity varied from 0.2m, 0.5m, 1m, and 2m. The thickness of the weak soil layer (TW)
was then increased from zero (i.e. one layer of strong soil) to 12m (i.e. two-layer,
strong over weak soil). The calculated resistance qb,3LS in this three-layer soil profile
7-30
can be compared with the unaffected resistance qb,S in one layer strong soil only. As
shown in Table 7.8, for any value of Hcavity, the resistance ratio qb,3LS/qb,S reduces as
the thickness of the weak soil layer increases. Ratios of qb,3LS/qb,S are grouped
together for similar H/D values and plotted against the normalised thickness ratio
TW/D in Figure 7.19. It is clear that resistance reduction factors qb,3LS/qb,S depends on
the thickness of the weak layer (TW) and also on the distance from pile tip to the
interface (H). In general, for thickness TW > 3D, the ratios of qb,3LS/qb,S are
approaching the ones (qb,2LS/qb,S) in a two-layer soil profile. Again, the reader is
reminded that the results presented in Figure 7.19 are only for one set of soil
parameters and may not be generally applicable for all cases.
1.1
0.9
0.8 qb,2LS/qb,S=0.84
0.7
q b,3LS /q b,S
0.6 qb,2LS/qb,S=0.62
0.5
0.4
average qb,2LS/qb,S=0.38
0.3
H/D=1.3
0.2 H/D=2.6
qb,2LS/qb,S=0.21
H/D=4.1
0.1
H/D=7.0
0
0 3 6 9 12 15
T W /D
Figure 7.19 Effect of reduction in resistance due to the underlain clay seam
7-31
CHAPTER 8 ANALYSIS AND DISCUSSION
8.1 INTRODUCTION
This Chapter provides a discussion of three principal aspects of this research, and
illustrates how these inter-relate to the overall Thesis theme of base performance of
displacement piles in sand. Firstly, the four primary CPT-based design methods
(Fugro-05, ICP-05, NGI-05 & UWA-05) are assessed based on the database study
(Chapter 3), Shenton Park field tests (Chapter 4) and centrifuge tests (Chapter 6).
The base resistance mobilisation curves from the 100 high quality centrifuge SLTs
conducted in soil samples A to H (Chapter 6) are then interpreted to provide a
framework for settlement prediction. Finally, a method for predicting pile base
response in a two-layer soil profile is summarized. Comparisons are drawn between
the results shown by the centrifuge tests (Chapter 6) and numerical analyses (Chapter
7) to provide further insights into qc averaging techniques.
8.2 ASSESSMENT OF DESIGN METHODS FOR qb0.1
8.2.1 Driven piles
As discussed in Chapter 3, four new ‘offshore’ CPT-based design methods, including

UWA-05, are now included in the commentary of the 22nd Edition of American
Petroleum Institute RP2A (API 2006). UWA-05 evolved following the evaluation of
the predictive performance of the other three CPT based methods (Fugro-05, NGI-05
& ICP-05) using a new and comprehensive pile load test database. UWA-05 is
shown to be to be the best performing method, as measured in terms of the average
and coefficient of variation of the ratios of calculated to measured capacities
(average=0.97 & COV=0.27, Table 3.6) in the UWA database containing 77 pile
load tests. The improved predictive performance of UWA-05 is attributed to its
incorporation of recent developments in our understating of driven piles in sand, such
8-1
Chapter 8 Analysis and Discussion
as effects of soil displacement during installation, friction fatigue, influence of

loading direction, sand-pile interface friction angle and change in radial stress during
loading.
In fact, as discussed in Chapter 3, the incorporation of final filling ratio (FFR) into
the design formulation of UWA-05 for pile ultimate end bearing resistance qb0.1
(defined at 0.1D of pile tip displacement) has greatly enhanced its predictive
performance. This effect was further explored through a field pile testing programme
in Shenton Park, Perth (as discussed in Chapter 4) with careful measurements of
IFRs, and thus FFRs.
The UWA base capacity database was extended to include results from Shenton Park
pile testing programme 1 . As summarised in Table 8.1, ICP-05 and UWA-05 appear
to have comparable predictive performance for closed-ended piles while for open-
ended piles (Table 8.2), the average ratio of qb0.1,C/qb0.1,M for UWA-05 is close to
unity and has, by far, the lowest COV of all the methods. Despite the small size of
the current base capacity database (22 closed-ended & 23 open-ended piles), this
improvement is rather promising and indicates the importance of incorporation of
degree of soil displacement (i.e. FFR) in the design of open-ended piles.
Figure 8.1 illustrates the variation of the soil plug capacity (qb,plug) with FFR. Also
shown is the design formulation adopted by UWA-05, which evidently provides a
reasonable fit to the Shenton Park data, with the exception of the short 33.7mm
diameter piles. As explained in Chapter 4, this discrepancy may be attributed to a
low estimation of compressive shaft capacity (from the tension capacity) due to a
visible gap that formed around the 33.7mm diameter pile during installation.
Results from Shenton Park pile tests are consistent with evaluation of the base
capacity database presented for driven piles in Chapter 3 that no dependence of the
qb0.1/qc,avg ratio on qc,avg, D or Dr is evident for closed-ended piles - as suggested by
Fugro-05, NGI-05 & ICP-05. For open-ended piles, the degree of soil displacement
induced during pile installation plays a key role in controlling the pile base capacity
1
As discussed in Chapter 4, base capacities were inferred from total compressive and tensile shaft
capacities.
8-2
that can be mobilised. This effect is best quantified by final filling ratio (FFR) and is
incorporated into design as illustrated by Equation 3.3 in UWA-05.
Table 8.1 Summary of method performance for the database of closed-ended driven piles
qc,avg
Site name; D L σ'v0 qb0.1,M qb0.1,C/qb0.1,M
Dutch
test No. m m kPa MPa MPa Fugro-05 ICP-05 NGI-05 UWA-05
Akasaka; AK 0.2 11 147 25.7 15.2 0.97 1.24 0.78 1.02
Baghdad; P1 0.285 11 152 4.7 5.1 1.28 0.64 0.90 0.55
Baghdad; P2 0.285 15 189 6.1 7.1 1.01 0.55 0.64 0.52
Drammen; A 0.28 8 90 2.6 1.1 3.99 1.39 2.06 1.36
Drammen; D/A 0.28 16 177 4.4 1.8 3.42 1.61 2.12 1.48
Hoogzand; II 0.356 6.75 98 26.3 14.0 1.03 1.03 0.81 1.13
Hsin Ta; TP4 0.609 34.25 311 6.4 3.1 2.59 1.11 1.84 1.23
Hunter's P; S 0.273 9.15 100 6.6 5.0 1.56 0.94 1.07 0.80
Kallo; I 0.908 9.69 109 14.7 9.0 1.57 0.91 1.25 0.99
Kallo; II 0.539 9.71 109 18.8 10.7 1.36 1.13 1.05 1.06
Kallo; III 0.615 9.82 110 16.3 9.7 1.52 1.19 1.18 1.01
Kallo; IV 0.815 9.8 110 15.1 9.2 1.57 1.01 1.24 0.98
Kallo; V 0.406 9.33 105 16.6 10.7 1.25 1.10 0.99 0.93
Kallo; VII 0.609 9.37 106 16.0 8.5 1.57 1.12 1.26 1.12
Ogeechee; H-12 0.457 6.1 74 8.8 10.7 0.88 0.51 0.55 0.49
Ogeechee; H-13 0.457 8.9 104 10.9 13.2 0.79 0.51 0.58 0.50
Ogeechee; H-14 0.457 12 136 10.5 13.5 0.74 0.46 0.34 0.47
Ogeechee; H-15 0.457 15 166 13.1 16.3 0.64 0.42 0.51 0.48
Pigeon R; CEP 0.356 6.87 90 16.4 11.0 1.04 0.83 0.76 0.90
Sermide; 0.508 35.85 315 14.5 10.2 1.07 0.68 0.97 0.86
ShentonP; P3 0.0889 4 72 10.3 8.8 1.00 0.98 0.70 0.70
ShentonP; P12 0.0889 2.5 45 2.94 1.9 2.39 1.23 1.18 0.91
average of ratios of qb0.1,C/qb0.1,M 1.51 0.94 1.04 0.89
Statistics COV of ratios of qb0.1,C/qb0.1,M 0.57 0.35 0.46 0.33
8-3
Table 8.2 Summary of method performance for the database of open-ended driven piles
qc,avg
Site name; D t L FFR σ'v0 qb0.1,M qb,plug qb0.1,C/qb0.1,M
Dutch
test No. m mm m - kPa MPa MPa MPa Fugro-05 ICP-05 NGI-05 UWA-05
Dunkirk; CSa 0.324 19.1 11.3 0.45 141 21.0 7.3 6.0 1.24 0.86 0.72 1.28
Dunkirk; CLa 0.324 12.7 11.3 0.48 141 21.8 6.6 5.7 1.25 0.95 0.79 1.38
Euripides; I30c 0.7631 35.55 30.45 0.99 320 55.0 12.3 7.9 1.10 0.83 0.75 1.03
Euripides; I38c 0.7631 35.55 38.7 0.90 403 47.3 9.9 5.9 1.25 0.86 0.93 1.28
Euripides; I47c 0.7631 35.55 46.95 0.89 488 53.1 15.3 11.7 0.93 0.73 0.71 0.93
Euripides; II47c 0.7631 35.55 46.65 0.82 477 53.4 16.0 12.5 0.87 0.67 0.66 0.99
Hoogzand; I 0.356 16 7 0.66 100 28.1 12.2 11.3 0.76 0.59 0.39 0.81
A
Hoogzand; III 0.356 20 5.25 0.77 82 29.5 11.2 9.5 0.93 0.74 0.43 0.86
Pigeon R; OEP 0.356 32 7.04 0.80 94 18.4 8.9 5.6 1.01 0.56 0.46 0.74
Ras Tanajib; 25a 0.7633 38.5 25 1.13 306 79.6 20.7 14.2 0.79 0.69 0.54 0.91
Shanghai; ST1 0.914 20 79 0.80 637 21.5 5.9 5.2 1.14 0.31 1.36 0.99
Shanghai; ST2 0.914 20 79.1 0.85 637 21.5 5.1 4.3 1.33 0.36 1.58 1.06
Tokyo; TP 2.018 43 30.6 1.08 275 9.7 2.0 1.6 3.61 1.02 3.44 0.93
ShentonP; P1 0.0889 2.6 4 0.69 72 10.6 4.1 4.3 1.26 1.06 0.87 0.84
ShentonP; P2 0.0424 2.6 4 0.5 72 11.3 4.0 4.4 1.60 1.41 0.91 1.22
ShentonP; P4 0.0889 5 4 0.77 72 10.5 3.8 4.2 1.57 1.14 0.93 0.90
ShentonP; P5 0.1143 3.2 4 0.85 72 11.3 2.0 2.1 2.58 2.14 1.77 1.43
ShentonP; P6 0.0889 3.2 2.5 0.77 45 9.6 2.8 3.0 1.85 1.42 1.23 1.03
ShentonP; P7 0.0424 2.6 2.5 0.26 45 3.3 1.2 1.3 2.83 1.34 1.65 1.38
ShentonP; P8 0.0424 2.6 2.5 0.24 45 3.4 2.3 2.5 1.55 0.76 0.91 0.79
ShentonP; P9 0.0337 2.6 3.5 0.18 63 4.0 2.9 3.4 1.36 0.71 0.75 0.74
ShentonP; P10 0.0337 2.6 2.5 0.15 45 5.3 4.8 5.6 0.94 0.56 0.56 0.61
ShentonP; P11 0.0889 2.6 2.5 0.45 45 3.8 0.9 0.9 3.58 1.84 2.49 1.85
average of ratios of qb0.1,C/qb0.1,M 1.55 0.94 1.10 1.03
Statistics COV of ratios of qb0.1,C/qb0.1,M 0.54 0.49 0.67 0.28
1.2
Dunkirk
Euripides
1 Hoogzand
D=33.7mm Pigeon R
Ras Tanajib
Shanghai
0.8
Equation (3.1b) Tokyo
q b,plug /q c (-)
ShentonP; I
q b,plug /q c,avg =0.6-0.45 × FFR
ShentonP; II
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2
FFR (-)
Figure 8.1 Ratios of qb,plug/qc vs. FFR for open-ended piles in the extended database
8-4
8.2.2 Jacked piles
Jacked piles may potentially possess increased base stiffness and capacity as well as
higher skin friction than a driven pile in similar conditions (White & Lehane 2004,
Deeks et al. 2005, Yetginer et al. 2006) and therefore deserve a separate evaluation.
However, a few jacked piles were included in the databases for derivation of ICP-05
and NGI-05 (i.e. design methods for driven piles). For instance, in ICP-05, the
dependency of qb0.1/qc,avg on pile absolute diameter (D) is contentious (White &
Bolton 2005, Xu & Lehane 2005) and may be caused by inclusion of the jacked piles
and inappropriate interpretation of qc,avg value in strongly layered soil. The base
resistance of jacked piles indicated by the database summarised in Chapter 3 and the
series of centrifuge model pile tests presented in Chapter 6 revealed the following:
• As summarised in Figure 8.2, the average of the qb0.1/qc,avg ratios from 7 field and
91 centrifuge closed-ended jacked pile tests equals ≈0.9, which is about 50
percent higher than that (≈0.6) for driven closed-ended piles. There is no
systematic variation of qb0.1/qc,avg with qc,avg, D & Dr for the closed-ended jacked
piles. However, as will be discussed in Section 8.3, the ratio of qb/qc,avg at relative
low (working level) displacements (such as qb0.02/qc,avg at a displacement of 2%D)
tends to increase with reducing Dr.
• Fugro-05, NGI-05 & ICP-05 appear to provide reasonably good predictions for
the database containing 13 closed-ended jacked piles (Table 3.10). The average
of calculated to measured total capacities (QTotal,C/QTotal,M) is close to unity, i.e.
0.99 for Fugro-05, 0.98 for ICP-05 and 0.96 for NGI-05. However, this degree of
agreement is fortuitous since some degree of underestimation is expected when
applying a method for driven piles (with qb0.1/qc,avg≈0.6) to jacked piles (with
qb0.1/qc,avg≈0.9). In fact, UWA-05 for driven piles exhibited a 20%
underestimation for the jacked pile database.
• The UWA-05 method was extended to jacked piles based on the evidence from
the field and centrifuge piles, and necessitated increasing the qb0.1/qc,avg ratio for
closed-ended jacked piles to 0.9. This increase led to an average of
QTotal,C/QTotal,M ratios of 1.03 (Figure 3.15)
8-5
• Significant residual stresses (qb,residual) could be locked in at the pile base during
jacking or driving installation. It can therefore lead to a stiffer pile base response
and higher base resistance at small displacement. In the current database of
closed-ended piles (summarised in Table 3.4 & Table 3.9), the residual resistance
ratios, qb,residual/qc,avg are generally less than 15% for driven piles, while those for
jacked piles could reach up to 50% (e.g. pile tests at Labenne). In the centrifuge,
qb,residual/qc,avg for the jacked model piles ranges from 2% to 18%, with an average
of 8%, which is in general higher than the 5% value reported by de Nicola (1999)
for the centrifuge driven pile tests.
• Some field tests (Gavin & Lehane 2005, White 2006) have shown that fully
plugged open-ended jacked piles (FFR=0) behave like a closed-ended jacked pile
with ratio of qb0.1/qc,avg close to unity. Paik & Salgado (2004) also demonstrated
through a series of calibration chamber tests that jacked piles have higher bearing
capacities than driven piles, mostly due to the more effective development of the
soil plug (i.e. lower IFRs) in the absence of dynamic effects. However, further
research and more field data are required before a design method can be put
forward for jacked open-ended piles.
1.2
0.8
q b0.1,M /q c,avg (-)
0.6
average = 0.89
COV = 0.07
0.4
7 field tests
0.2
91 centrifuge tests
0
0 10 20 30 40 50
qc,avg (MPa)
Figure 8.2 Ratios of qb0.1,M/qc,avg for closed-ended jacked piles based on field and centrifuge tests
8-6
8.2.3 Implication for design
The API design guidelines evolved from onshore practice but are widely used for the
design of offshore piles, which are predominately large diameter open-ended driven
piles. It is therefore worth comparing the design performance of the four new CPT-
based methods for such a pile.
qc (MPa) Q Total (MN)

0 10 20 30 40 50 60 0 20 40 60 80 100 120
0 0
5 5 Fugro-05
ICP-05
10 10 NGI-04
15 15 UW A-05
20 20
25
Depth (m)
25
Depth (m)
30 30
35 35
40 40
45 45
50 50
55 55
60 (a) 60 (b)
65 65
Q Base (MN) Q Shaft (MN)
0 10 20 30 40 50 0 20 40 60 80
0 0
5 5
10 10
15 15
20 20
25 25
Depth (m)
Depth (m)
30 30
35 35
40 40
45 45
50 50
55 55
60 (c) 60 (d)
65 65
Figure 8.3 An example of method performance at a site in the Gulf of Mexico
Figure 8.3a shows a typical design qc profile for a site currently being studied in the
Gulf of Mexico. The pile under consideration for development is 2.44m (96inch) in
diameter with a wall thickness of 44.5mm (1.75inch). Figure 8.3 summarises the
8-7
predicted compression capacity (total, shaft & base) with depth for Fugro-05, NGI-
05, ICP-05 & NGI-05 (design formulations as outlined in Tables 3.1 & 3.2). It can be
observed that the total capacities predicted by Fugro-05 and NGI-05 are generally
comparable and greater than those predicted by ICP-05 and UWA-05 methods. The
shaft capacity given by ICP-05 are generally greater than those predicted by other
methods for piles greater than 30m in length, but its predictions for total capacity fall
below the other three methods. This arises, as for the predicted site example, because
of the relatively low end bearing resistance predicted by ICP-05 for large diameter
pipe piles, and is consistent with its performance in the full UWA database (i.e. the
qb0.1/qc,avg ratios are only half of those given by three other methods for open-ended
driven piles; see Table 3.7).
8.3 EVALUATION OF PILE BASE STIFFNESS, G
8.3.1 Samples A to G
The base performance of a pile is influenced by a number of factors, such as pile

characteristics, soil properties and installation method (see Chapter 2). To take these
and other factors into account and to evaluate the overall pile base response, it is
useful to back-analyse the operational base stiffness values from the static load test
results. As illustrated in Figure 8.4, the pile base stiffness is calculated as the secant
gradient (k) from the resistance mobilisation curves. To correctly evaluate ‘k’, the
residual base resistance (qb,residual) needs to be subtracted from the base resistance (qb).
By making use of the elastic rigid punch solution (Equation 8.1a), the operational
shear stiffness of the soil (G) can be evaluated from the pile base secant gradient ‘k’
as shown in Equation 8.1b. The value for Poisson’s ratio (ν) has been assumed to be
0.2, which is typical for sandy soils at intermediate strain levels.
w π(1 − ν )
= × qb ( 8.1a )
D 8G
π(1 − ν ) q b π(1 − ν )
G= × = ×k ( 8.1b )
8 w D 8
8-8
30
qc =29.2MPa
25
20 k
Pile test
q b (MPa)
1 D6g3J-1
15
Pile base secant gradient

10 k = Δ (q b - q b,res idual) / Δ (w / D)
k
5
1
0
0 2 4 6 8 10 12 14 16 18 20
w/D (%)
Figure 8.4 Example illustrating the calculation of pile base secant stiffness ‘k’
180
160
140 Pile test

D6g3J-1
120
G (MPa)
100
80
60
40
20
(a)
0
0 2 4 6 8 10 12 14 16 18 20
w/D (%)
8-9
180
160
140 Pile test

D6g3J-1
120
G (MPa)
100
80
60
40
20
(b)
0
0 0.2 0.4 0.6 0.8 1 1.2
qb/qc (-)
Figure 8.5 Variation of the operational shear stiffness G with (a) w/D, and (b) qb/qc
35
Sample A
30 Sample B
Sample C
Rix & Stokoe (1992)
-0.75
Sample D
25 (G 0 /q c)Average = 291 × q c1N Sample E
Range≈Average×(1±0.5) Sample F
Sample G
G IN/q c (-)
20
15
10
0
0 100 200 300 400 500
q c1N (-)
Figure 8.6 Initial shear stiffness ratios deduced from the 91 static load tests in samples A to G,
qc1N=(qc/pa)/(σ’v0/pa)0.5, with pa=100kPa
8-10
Figure 8.5 shows the variation of operational shear stiffness of the soil (G) with pile
tip displacement ratio (w/D) and resistance ratio (qb/qc) inferred from the resistance
mobilisation curve (in Figure 6.10a) based on Equation 8.1. As expected, G reduces
with an increase in pile tip displacement ratio (or increase in strain level) and
resistance ratio. The stiffness at relatively small displacements (w/D<2%) is referred
to here as the initial operational shear stiffness of the soil, GIN, which equals 164MPa
for the pile test D6g3J-1 as shown in Figure 8.5. For this particular pile test D6g3J-1,
the stiffness G remains unchanged (with G/GIN=1) up to a displacement ratio of ~2%
and a resistance ratio of ~0.5. The term GIN is employed here to distinguish it from
the very small strain in-situ stiffness, G0, which is expected to be operational at w/D
values less than 0.01%. The linear displacement transducer (LDT) used in this
experiment does not allow accurate evaluation of G value at such small displacement
level.
The values of GIN were evaluated following the same procedure as for the example
given in Figure 8.5 for all 91 static load tests and these were subsequently normalised
by the penetration resistance qc – which is actually the steady state pile end bearing
resistance 2 . Figure 8.6 summarises the inferred normalised initial shear stiffness
(GIN/qc). The back-analysed GIN/qc values, including tests in layered samples, are
seen to generally follow the average trendline proposed by Rix & Stokoe (1992) for
the very small strain in-situ stiffness values (G0), although GIN exceeds G0 in the
dense sand samples (i.e. for qc1N>250, in Figure 8.6).
The similarity between GIN and G0 presumably arises because of the high degree of
over-consolidation induced in the sand below the pile base by the final jacking stroke
prior to the static load test. The equivalent modulus is approximately linear elastic to
normalised tip displacement w/D value in excess of 0.5%. This linearity over such an
extensive strain range suggests that decreases in shear stiffness with strain from the
small strain value (expected to operate at w/D<0.01%) is compensated for by the
increase in stress level.
2
It should be mentioned that these “qc” values were taken as the pile base resistance mobilised during
the jacking stage preceding the static load test. For tests in non-homogeneous profiles, these qc values
are affected by the presence of weaker/stronger layers and should be considered as the corrected
steady state pile end bearing value - which may be derived using the procedure described in Section
8.4.
8-11
In summary, it would appear from Figure 8.6 that the base stiffness of jacked piles at
working levels of displacement may be approximated employing the in-situ G0 value
as the operational shear stiffness (which is now measured routinely); this is discussed
further in Section 8.4.3.
1.2
(a) Test Name ; G IN /qc
1 B6g3J-1; 15.24
A6g3MJ-2; 8.56
D6g3J-1; 5.62
0.8 D6g3J-2; 5.01
D6g3J-3; 4.71
G/G IN (-)
G IN/qc increases
0.6
0.4
Sample Dr
B 0.36
0.2 A 0.54
D 0.93
0
0.1 1 10 100
w/D (%)
1.2
(b)
0.8
G/G IN (-)
0.6
Test Name; G IN /q c
B6g3J-1; 15.24
0.4
A6g3MJ-2; 8.56
D6g3J-1; 5.62
0.2 D6g3J-2; 5.01
D6g3J-3; 4.71
0
0 0.2 0.4 0.6 0.8 1 1.2
qb/qc (-)
Figure 8.7 Examples of stiffness decay curves for pile tests at various σ'v0 and Dr, (a) G/GIN vs. w/D,
and (b) G/GIN vs. qb/qc
8-12
The degradation of stiffness seen on Figure 8.5 controls our ability to predict pile
settlement under static loading. This degradation is examined further here for the
series of centrifuge tests on jacked piles in sand samples A to G.
Figure 8.7 summarises the stiffness decay curves (G/GIN vs. w/D & qb/qc) for the
typical pile tests, previously discussed in Section 6.2.2 (i.e. Figures 6.10 & 6.11).
Pile tests, B6g3J-1, A6g3MJ-1 & B6g3J-3 were performed at similar stress levels in
samples of different relative densities, while D6g3J-1, D6g3J-2 & D6g3J-3, formed a
series of tests involving tests at three depths on the same pile. Also shown are the
initial stiffness ratios (GIN/qc) determined for each static load test, which increase
with an increase in the relative density of the soil sample, ranging from 4.7 for
D6g3J-3 to 15.2 for B6g3J-1; this is of course consistent with the findings from
Figure 8.6. The values of GIN/qc decrease only slightly with an increase in σ'v0, from
5.62 for D6g3J-1 to 4.71 for D6g3J-3, showing a minor effect of stress level on the
normalised stiffness ratio.
It is obvious from Figure 8.7a that the GIN/qc ratio is a key factor in determining the
decay of stiffness with displacement/strain level (GIN/G vs. w/D). At a certain w/D
value, the reduction in stiffness is greater for a higher ratio of GIN/qc. For example, at
w/D=10%, values of G/GIN range from 0.16 to 0.50 as ratios of GIN/qc reduce from
15.2 to 4.7. However, such a strong dependency on GIN/qc is not apparent for the
relative stiffness degradation with loading level (GIN/G vs. qb/qc, in Figure 8.7b). The
relative stiffness reduction with qb/qc curves are seen to be relatively independent of
the initial stiffness ratio GIN/qc, suggesting that there may be a single relationship
between G/GIN and qb/qc for all the jacked pile tests.
Stiffness decay curves such as those shown in Figure 8.7 are evaluated from the
resistance mobilisation curves (qb/qc vs. w/D) for all the 91 centrifuge static load
tests and results are summarized in Appendix B.2. Based on the observations from
Figure 8.7, it was decided to categorize the results according to the values of initial
stiffness ratio (GIN/qc). As shown in Figure 8.6, ratios of GIN/qc range from 3.8 to
29.5, with most values in the range of 4 to 10. The static load tests were sorted into
11 groups, each with similar ratios of GIN/qc. Within each group, the data were
averaged to give one representative stiffness decay curve. Such an example (avg
G/GIN vs. w/D & qb/qc) is shown in Figure 8.8 for ratios of GIN/qc of ~4.6. As
8-13
summarized, there are a total of 12 static load tests in this group of tests conducted in
different samples. The averaged stiffness decay curves for the other ten groups can
be found in Appendix B.3.
1.2
average GIN/qc= 4.56
1
0.8
G/GIN (-)
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2
qb/qc (-)
1.2
1
0.8
G/GIN (-)
0.6
0.4
0.2
0
0.1 1 10 100
w/D (%)
No Name GIN/qc GIN qc σ'v0

- - - MPa MPa kPa
1 C6g3MJ-1 4.74 112 23.7 115
2 C6g3MJ-3 4.64 200 43.2 178
3 C6g3J-2 4.81 202 42.0 157
4 C6g3J-3 3.80 182 47.7 177
5 C16g3MJ-2 4.39 137 31.3 155
6 C16g3J-2 4.91 180 36.8 155
7 D6g3J-3 4.71 214 45.4 186
8 D6g3MJ-3 4.44 214 48.1 186
9 D16g3J-2 4.88 150 30.6 154
10 D16g3MJ-2 4.68 159 34.0 154
11 E9g3MJ-1 4.32 131 30.3 122
12 F9g2MJ-3 4.43 118 26.7 116
Figure 8.8 Averaged stiffness reduction curves (G/GIN vs. qb/qc & w/D) for ratios of GIN/qc ~4.6
8-14
Figure 8.9 presents the results of stiffness decay curves for each categorized group
and tabulates the averaged GIN/qc ratios for static load tests within that group. A
similar trend was found to that shown on Figure 8.7. The degree of stiffness
reduction with the displacement/strain level (i.e. w/D) depends on the ratio of GIN/qc,
while the relationship between G/GIN and qb/qc is relatively unique and independent
of GIN/qc. Stiffness decay curves and hence the full pile base load-settlement
response may therefore be derived from a measurement of G0 (~GIN) and qc. In fact,
as shown in Figure 8.9a, the relatively unique relationship between G/GIN and qb/qc
can be described by the modified hyperbolic relationship given in Equation 8.2. By
substituting Equation 8.2 into 8.1b, the base resistance mobilization curve (qb/qc vs.
w/D) can be produced using Equation 8.3. The reduction of G/GIN with w/D can also
be calculated based on Equations 8.2 and 8.3. Such calculated results for two
extreme averaged ratios of GIN/qc (=4.6 & 29.5) are shown in Figure 8.10. There is a
good agreement between measured and calculated stiffness decay curves (G/GIN vs.
w/D). However, it should be noted that Equations 8.2 and 8.3 may be only applicable
for jacked piles, since the installation method, and other factors may significantly
alter this relationship.
G G IN = 1 − (q b q c )
4
( 8.2 )
w π(1 − υ) (q b − q b ,residual ) q c
= ⋅ ( 8.3 )
1 − (q b q c )
4
D 8 G IN q c
As indicated by Equation 8.3, at w/D=10%, where the ultimate base resistance (qb0.1)
is defined, for GIN/qc ratios of 30 to 4 (the typical range for this series of centrifuge
tests), the resistance ratio qb0.1/qc varies from 0.79 to 0.97, which is consistent with
the centrifuge results (from 0.78 to 0.98, with an average of 0.89). It also follows that
the variation of qb0.1/qc (0.79 to 0.97) depends on the relative density of the sand
since GIN/qc is known to vary (reduce) with Dr. The ratios of qb0.1/qc should therefore
increase with a decrease in the relative density of the sand. This trend is, however,
not very pronounced for this series of tests (average of ratios of qb0.1/qc=0.89,
COV=0.06) and may be associated the displacement level (w/D=10%) under
consideration. At a typical working settlement ratio of w/D=2%, Equation 8.3
indicates that the resistance ratio qb0.02/qc range from 0.25 to 0.87 depending on the
relative density of the sand. This is verified on Figure 8.11 which plots qb0.02/qc for
8-15
all static load tests. Evidently, qb0.02/qc reduces with an increase in the ratio of GIN/qc
and therefore decreases as Dr increases.
1.2
(a) 3.8 < GIN/q c < 29.5 Equation 8.3
g
1 G/GIN = 1 - f × (q b/q c )
f =1 & g =4
0.8
4.56
G/G IN (-)
0.6 5.44
6.38
7.51
average 8.37
0.4 GIN/qc= 9.68
12.72
14.94
0.2 16.43
21.03
29.52
0
0 0.2 0.4 0.6 0.8 1 1.2
qb /qc (-)
1.2
(b) 3.8 < GIN/qc < 29.5
1 4.56
5.44
6.38
average 7.51
0.8
GIN/q c = 8.37
9.68
G/G IN (-)
12.72
0.6 14.94
16.43
21.03
GIN/qc increases 29.52
0.4
0.2
0
0.1 1 10 100
w/D (%)
Figure 8.9 Categorized stiffness reduction curves based on similar ratios of GIN/qc, (a) G/GIN vs. qb/qc,
and (b) G/GIN vs. w/D
8-16
1.2
Measured; G IN/qc =4.6
1
0.8
Calculated using Calculated using
G/G IN (-)
Equations 8.2 & 8.3; Equations 8.2 & 8.3;

0.6 G IN/qc =29.5 G IN/qc =4.6
0.4
0.2
Measured; G IN/qc =29.5

0
0.1 1 10 100
w/D (%)
Figure 8.10 Comparisons of measured and calculated (using Equations 8.2 & 8.3) stiffness decay
curves (G/GIN vs. w/D)
0.9
0.8
0.7
0.6
q b0.02 /q c (-)
0.5
0.4 Sample A
Sample B
0.3 Sample C
Sample D
0.2 Sample E
Sample F
0.1 Sample G
0
0 5 10 15 20 25 30 35
G IN/q c (-)
Figure 8.11 The variation of measured resistance ratio qb0.02/qc at w/D=2% with GIN/qc
8-17
8.3.2 Sample H
The stiffness decay curves are summarized in Figure 8.12 for pile tests in sample H.
A similar trend to that seen in samples A to G is observed, where the stiffness decay
curve (G/GIN vs. w/D) depends on the initial stiffness ratio GIN/qc. The higher the
initial stiffness ratio (GIN/qc), the lower the stiffness reduction ratio, G/GIN, at the
same displacement ratio w/D. Also shown in the figure are the stiffness decay curves
calculated using Equations 8.2 and 8.3, which provides a reasonable fits for the
results in samples A to G. For static load test H6-8, there is a good agreement
between the measured and calculated results. This is because this test was conducted
at an embedment far from the sand and clay interface (i.e. about 10D above the
interface) and the effect of clay on base resistance mobilisation was small. However,
for other tests performed close to the interface or in the clay, the calculated stiffness
decay curves are well above those measured, indicating that the presence of adjacent
weak clay causes a significant reduction in pile base stiffness. Pile settlement
prediction methods clearly need to take this effect into account.
1.2
calculated with
G IN /q c=8.5
1
Test Name; GIN/qc
H6-8; 8.5
0.8 H6-2 & H6-4; 10.0
H6-1 & H6-3; 14.8
G/G IN (-)
H16-2 & H-3; 25.1

0.6 calculated; 25
calculated; 8.5
0.4
increase in
G IN /q c
0.2
calculated with
(a) G IN /q c=25
0
0.1 1 10 100
w/D (%)
8-18
1.2
Test Name; GIN/qc

1
H6-5; 9.6
H16-1; 8.8
0.8 calculated; 10
G/G IN (-)
calculated with
0.6
G IN /q c=10
increase in
0.4 G IN /q c
0.2
(b)
0
0.1 1 10 100
w/D (%)
Figure 8.12 Stiffness decay curves (G/GIN vs. w/D) for pile static load tests in sample H: (a) in the
upper sand layer, and (b) in the lower clay layer
8.4 BASE RESISTANCE RESPONSE IN LAYERED SOIL
8.4.1 Analysis procedure in a two-layer soil profile
The analysis procedure evolved from results of numerical analyses simulating pile
installation in a two-layer soil profile. The procedure is summarized as follows:
(i) Evaluate the resistance ratio between the two soil layers, qc,W/qc,S (referred as
A0), i.e. from in situ CPT test.
(ii) Calculate fitting coefficients A1 and A2 using the following Equation:
A1 = −0.22 ln (A 0 ) + 0.11 ≤ 1.5

( 8.4 )
A 2 = −0.11 ln(A 0 ) − 0.79 ≤ −0.2
(iii) Assess the resistance reduction ratio (qc/qc,S) as a function of the normalised
distance to the interface of the two soil layers, H/D, by Equation 8.5.
q c q c,S = A 0 + (1 − A 0 ) ⋅ exp[− exp(A1 + A 2 ⋅ H D )] ( 8.5 )
Alternatively, the variation of qc/qc,S with H/D could also assessed using curves
presented in Figure 8.13 for typical A0 values.
8-19
1.1
1.0
0.9
0.8
0.7 0.005
q c /q c,S (-)
0.01
0.6
0.03
0.5
0.1
0.4 0.3
decrease in A 0
0.3 A 0= 0.5
0.2 0.7
0.1 0.9
0.0
-5 0 5 10 15 20 25
H/D (-)
Figure 8.13 Plots of qc/qc,S vs. H/D for various A0 (=qc,W/qc,S) ratios
8.4.2 Variation of qc/qc,S with H/D
As detailed in Chapters 5 and 6, model piles of three different diameters (6mm,

9.5mm & 16mm) were tested in four uniform (A, B, D & E) and four layered (C, F,
G & H) soil samples. Samples A to G contain dry normally consolidated (OCR=1)
sand deposits, while sample H is saturated with OCR≈2.4. In fact, the tests in the
uniform samples A (Dr of 0.54) and D (Dr of 0.93) provide a good reference as the
unaffected steady state penetration resistance, i.e. qc,W, & qc,S for tests in layered
samples C, G and F, where a medium dense sand layer (Dr of 0.45 to 0.50) was
underlain or overlain by a very dense sand layer (Dr of 0.92). In sample H, a very
dense sand layer (Dr of 0.96) was underlain by a soft clay layer. The unaffected
steady state penetration resistance for the dense sand, qc,S, could be estimated from
the correlations between base resistance and Dr (Equation 6.4) with proper
consideration for the overconsolidation (Equation 6.5).
Figure 8.14 summarizes the measured base resistance (qc) from the three model piles
(6mm 9.5mm & 16mm in diameter) during jacked installation in the layered soil
samples. Also shown are the averaged base resistance profiles measured in the
uniform samples A & D (i.e. the unaffected resistance, qc,W & qc,S) or estimated (i.e.
qc,S) from Equations 6.4 & 6.5. In sample C, qc starts to deviate from the measured
8-20
qc,W profile at some distance above the interface of the two sand layers. It then
gradually increases with further penetration into the underlying dense sand until it
reaches the qc,S profile. As shown, there is an absolute diameter effect (i.e. scale
effect) on the measured qc. For instance, at depth z of 140mm, the qc values from the
16mm are lower than those from the 6mm & 9.5mm piles which appears to agree
with qc,S. This is as expected since the large diameter pile reacts more slowly to
changes in soil stratigraphy and therefore requires greater responding distance.
Similar observations are made in the layered samples F & G. The zone of influence
is about 5 times the pile diameter in the layered sand samples C, F & G. This scale
effect gets more pronounced in sample H, which features a sharp change in soil
strength between the dense sand and soft clay. In the upper sand layer, the measured
qc values of the 16mm pile are much lower than that of the 6mm pile. There is
basically no difference in qc values once the piles begin penetrating into the lower
clay layer. The zone of influence extends up to ~12D in sample H, which is
significantly higher than that in the layered sand samples.
qc (MPa) qc (MPa)
0 10 20 30 40 50 0 10 20 30 40 50
0 0
C6g3 F6g3
20 20 F9g3
C9g3
C16g3 40 F16g3
40
60 avg qc,S in 60 avg qc,S in

sample D sample D
80 80
interface
z (mm)
z (mm)
100 interface 100
120 120
140 140
160 avg qc,W in 160 avg qc,W in

sample A sample A
180 180
(a) (b)
200 200
8-21
Figure 8.14 The measured pile base resistance in layered soil samples (a) C, (b) F, (c) G, and (d) H
Based on the information (qc, qc,S and depth of the interface) on Figure 8.14, the
resistance reduction curves (qc/qc,S vs. H/D) can then be produced. Such results are
summarized in Figure 8.15. Also shown are the resistance reduction curves
calculated following the procedure (detailed in Section 8.4.1) devised based on the
numerical analyses in Chapter 7. There is generally good agreement between the
measured and calculated resistance reduction curves. The variation of qc/qc,S with
H/D is not linear and the zone of influence is not a constant of 10D as suggested by
Meyerhof (1983), and others. Instead, they depend strongly on A0=qc,W/qc,S, and can
be estimated reasonably well following the recommendations emanating from the
numerical analyses for a two-layer soil. In the case of a multi-layered soil profile (e.g.
three-layer), an additional reduction in resistance can be caused by an increase in the
thickness of the weak soil layer or the presence of another weak layer. Although two
such cases were analyzed in Chapter 7 for a three-layer soil profile, further research
is required before any recommendation can be readily put forward for more
complicated multi-layered ground. Ideally as a result, a more sound qc averaging
technique could then be proposed to estimate the steady state penetration for a (large
diameter) pile from a standard CPT (DCPT=36mm). The applicability of currently
8-22
available qc averaging techniques (discussed in Chapter 3) are compared in the

following with those put forward here for a two-layer soil.
1.2
C6g3
C9g3
C16g3 1
Calculated
0.8
q c /q c,S (-)
0.6
0.4
Dr=0.45 0.2 Dr=0.92 (a)

A0=0.40
0
-8 -6 -4 -2 0 2 4 6 8 10 12
H/D (-)
1.2
F6g3
F9g3
1
F16g3
Calculated
0.8
q c /q c,S (-)
0.6
0.4
Dr=0.5 0.2 Dr=0.92 (b)

A0=0.45
0
-8 -6 -4 -2 0 2 4 6 8 10 12
H/D (-)
8-23
1.2
G6g3
G9g3
1
G16g3
Calculated
0.8
q c /q c,S (-)
0.6
0.4
Dr=0.5 0.2 Dr=0.92 (c)

A0=0.45
0
-8 -6 -4 -2 0 2 4 6 8 10 12
H/D (-)
1.2
H6
H16
1
Calculated
0.8
q c /q c,S (-)
0.6
0.4
Clay 0.2 Dr=0.96 (d)

A0=0.025
0
-8 -6 -4 -2 0 2 4 6 8 10 12 14 16 18
H/D (-)
Figure 8.15 Comparison of measured and calculated resistance reduction curves (qc/qc,S vs. H/D) in
layered soil samples, (a) C, (b) F, (c) G, and (d) H
8.4.3 Variation of GIN with H/D
As illustrated in Figure 8.6, the inferred normalised initial pile base stiffness GIN/qc is
shown to be a function of normalised penetration resistance qc1N (i.e.
=(qc/pa)/(σ’v0/pa)0.5, pa=100kPa) in both uniform and layered samples. The qc value
8-24
employed for normalisation in this figure is the steady state pile base penetration
resistance (and not actually the CPT qc value), as it reflects the existence of
neighboring weak or strong soil layers. In fact, as illustrated in previous section, the
‘qc value’ could reduce significantly when the pile approaches a weaker layer. The
pile base GIN value is therefore also expected to reduce due to presence of a weaker
layer.
To examine the influence of soil layering on pile base stiffness, the inferred GIN
values in sample ‘H’ normalised for stress level (i.e. GIN/(σ’v0/pa)0.5) are plotted on
Figure 8.16 against the normalised distance (H/D) from the layer interface. As
expected, the normalised GIN values reduce with H/D, and are lowest for static load
tests performed in the clay layer (i.e. H/D<0). At greater distance away from the
interface, for instance H/D~10, the underlying clay will have the minimum influence
on GIN. In fact, as illustrated in Figure 8.12a, the stiffness decay curve (G/GIN vs.
w/D) for the test (H6-8) at H/D~10 is in keeping with the observations in the sand
samples (i.e. A to G), while other tests conducted closer to the layer interface display
a much faster stiffness decay than those in sand samples.
200 1
180 0.9
160 0.8
(MPa)
/200 (-)
140 0.7
0.5
0.5
120 0.6
G IN /(σ 'v0 /p a )
G IN /(σ 'v0 /p a )
100 0.5
80 0.4
60 0.3
40 0.2
Clay Sand
20 0.1
0 0
-10 -5 0 5 10 15 20
H/D (-)
Figure 8.16 Values of normalised GIN plotted against normalised distance H/D in sample H
As there were no data to allow values of GIN to be evaluated in a single layer of

dense sand (=GIN,S), the variation of GIN/GIN,S with H/D could not be determined
8-25
with accuracy. Therefore, as discussed earlier, the GIN value measured in the test in
dense sand at H/D~10 (for which GIN/(σ’v0/pa)0.5≈200) is taken as a reasonable
estimate of GIN,S and is used to present the trend for GIN/GIN,S in Figure 8.16 using a
secondary ‘y’ axis. This assumption leads to a GIN,W/GIN,S of about 0.044, which is
larger than the qc resistance ratio (i.e. qc,W/qc,S=0.025). This discrepancy is likely to
be due to the use of the undrained qc,W clay value as the drained qc,W value leads to a
qc,W/qc,S ratio comparable to GIN,W/GIN,S =0.044.
Figure 8.16 highlights the significant influence that an underlying weak clay layer
can have on both end bearing capacity and initial base stiffness GIN, e.g. for a pile
base at a distance of ~3D from a weak layer, the pile base stiffness and capacity are
only about 50% of the values in a single dense sand layer.
200
1.1
180
1
160
0.9
/170 (-)
140
(MPa)
0.8
120 0.7
0.5
0.5
100 0.6 G IN /(σ 'v0 /p a )

G IN /(σ 'v0 /p a )
0.5
80
0.4
60
0.3
Sample C 40
0.2
Sample G 20 0.1
0 0
-15 -10 -5 0 5 10 15
H/D (-)
Figure 8.17 Values of normalised GIN plotted against normalised distance H/D in samples C & G
The GIN values (normalised for stress level) inferred from static load tests in the two-
layer sand samples C & G are summarised in Figure 8.17. As described in Section
5.4, a loose sand layer (Dr=45%) was underlain by dense sand (Dr=92%) in sample C,
while a dense sand layer (Dr=92%) was underlain by loose sand (Dr=50%) in sample
G. As all static load tests were performed in the lower layer, data points for H/D>0 in
Figure 8.17 were obtained from tests in sample C, while those with H/D<0 were
derived from tests in sample G. Again due to the unavailability of GIN,S values, the
8-26
secondary y-axis on Fig. 8.17 normalises the data by the (mean) maximum GIN value
inferred in the dense sand i.e. GIN/(σ’v0/pa)0.5 ~170. This normalization leads to an
estimated GIN,W/GIN,S ratio of about 0.6, which is higher than the resistance ratio
qc,W/qc,S=0.45, but lower than the G0,W/G0,S ratio of 0.8 implied by the Rix and
Stokoe (1992) for G0 (which indicates that G0 varies with qc0.25). The ratio of 0.6
suggests that GIN varies with qc0.66. It is clear, however, that further data in addition
to those indicated on Figures 8.6, 8.16 and 8.17 are required to assist assessment of
the GIN value.
8.4.4 Comments on qc averaging technique
In practice, the qc profile obtained from the in situ CPT often requires to be averaged
to properly reflect the steady state resistance for a pile of much large diameter for
design purposes because of the non-uniform nature ground. Three popular averaging
techniques (i.e. Dutch, arithmetic & geometric) were discussed in Chapter 3 and are
further examined here for an idealised two-layer soil profile. The pile considered is
1m in diameter. A strong soil is underlain by a weak soil with the interface being
located at depth z of 15m. The resistance in the upper weak soil layer (qc,W) is kept
constant as 1MPa, and values of qc,S can be varied accordingly. For this illustration,
ratios of qc,W/qc,S are set to be 0.1 and 0.5 (thus qc,S equals 10MPa & 2MPa
respectively). The results of averaged qc profiles from the three averaging techniques
are summarised in Figure 8.18. The qc profile assessed following the
recommendation in Section 8.4.1 (i.e. Equation 8.5) to be the ‘true’ base response, is
shown for comparison purpose. It can be seen that, for both cases (qc,W/qc,S of 0.1 &
0.5), the averaged qc profiles by Dutch averaging technique are in better agreement
with the “true” qc profile than other two methods. The Dutch averaging technique is
also more conservative than the other two methods, because of the imposed
minimum path rule.
8-27
q c (MPa) q c (MPa)
0 1 2 3 4 5 6 7 8 9 10 11 0.8 1 1.2 1.4 1.6 1.8 2 2.2
10 10
Geometric mean
Geometric mean
12 12 Arithematic
Arithematic avg
avg over ±1.5D
over ±1.5D
14 14
interface interface
16 16
z (m)
z (m)
Equation 8.5 Equation 8.5
18 18
20 Dutch qc avg 20 Dutch qc avg

technique technique
22 22
q c,W /q c,S =1/10=0.1 qc,W /qc,S =1/2=0.5
(a) (b)
24 24
Figure 8.18 Comparison of qc averaging techniques in an idealized two-layer soil profile with a 1m
diameter closed-ended pile
8-28
CHAPTER 9 CONCLUSIONS
9.1 INTRODUCTION
This Thesis provides an investigation of the end bearing performance of

displacement piles in siliceous sand and employs (i) a full database review, (ii) field
pile testing, (iii) centrifuge model pile testing, and (iv) numerical analysis to advance
understanding. This final Chapter summarises the principal findings arising from the
research, and then provides recommendations for future work
9.2 FINDINGS OF RESEARCH
The author together with James Schneider (Schneider 2007) developed the UWA
database (UWA-05) containing 90 high quality pile load tests with CPT data (77
driven & 13 jacked piles). A sub-database containing base capacity data was also
compiled. These results together with additional field tests performed in the moist
medium dense siliceous sand at Shenton Park and an extensive series of centrifuge
tests have shown:
(i) Three newly published CPT-based methods (Fugro-05, ICP-05 & NGI-05)
evaluated using the UWA-05 database of driven piles have a similar predictive
performance (Table 3.6). This is despite the fact that their end bearing
formulations for qb0.1 (i.e. at pile tip displacement of 10%D) are remarkably
different (Table 3.1). In particular, the difference in base capacity between
closed- and open-ended piles is attributed to various factors, such that the ratio
of qb0.1;Open/qb0.1;Closed depends on the area ratio Ar in Fugro-05, on both Ar and D
in ICP-05, but on sand relative density Dr in NGI-05. These dependencies
contrast with evidence provided in this Thesis and elsewhere, indicating that the
ratio is primarily a function of the degree of soil displacement induced during
installation (i.e. as measured by IFR or FFR, which could be subsequently
influenced by Dr, D, soil layering and other factors).
9-1
Chapter 9 Conclusions
(ii) These three new methods for driven piles (Fugro-05, ICP-05 & NGI-05) appear
to provide reasonably good predictions for the database of jacked piles, with the
average of calculated to measured total capacities (QTotal,C/QTotal,M) close to unity
(Table 3.10). However, this degree of agreement is fortuitous since this Thesis
has shown that qb0.1/qc,avg is approximately 0.6 for driven piles but about 0.9 for
jacked piles.
(iii) In practice, the in situ CPT qc needs to be averaged to properly reflect the steady
state penetration resistance of a pile (because the latter’s much larger diameter).
The examination of three popular averaging techniques (Section 3.5) has
demonstrated that the Dutch method performs better than other two methods
especially when strong soil stratigraphic changes occur near the pile tip.
Incorrect evaluation of values of qc,avg may result in erroneous conclusions on
the key factors controlling the ratio of qb0.1/qc,avg. Moreover, the Dutch method is
also shown to perform better in an idealised two-layered soil profile, in
comparison with the results from numerical analysis (Section 8.4.4).
(iv) Based on a thorough review of the base capacity databases (driven & jacked),
Equations 9.1 and 9.2 have been proposed for evaluation of pile end bearing
(qb0.1). No systematic variation of qb0.1/qc,avg with qc,avg, D & Dr (as suggested by
Fugro-05, NGI-05 & ICP-05) for closed-ended piles has been revealed and
constant qb0.1/qc,avg ratios of 0.6 & 0.9 are recommended for driven and jacked
closed-ended piles respectively. The effects of soil displacement (i.e. fully
coring, partially plugged, or fully plugged) on qb0.1 have been quantified by
incorporating the effective area ratio term (Ar,eff). As illustrated in Table 3.6,
these new formulations (for both qb0.1 & τf) lead to a comparable performance as
ICP-05 for closed-ended driven piles, but represent a very significant
improvement for open-ended driven piles.
For driven closed- & open-ended piles:
q b0.1 / q c,avg = 0.15 + 0.45 × A r ,eff ( 9.1 )
9-2
where effective area ratio A r ,eff = 1 − FFR × (D i D )

2
For jacked closed-ended piles:
q b 0.1 q c = 0.9 ( 9.2 )
(v) A series of driven pile tests (2 closed- & 10 open-ended) was conducted at the
relative uniform sand site in Shenton Park, Perth. The independent data obtained
were used to further examine the predictive performance of the four CPT-based
methods. It is shown that UWA-05 method best predicts the base capacity of
driven piles in sand (Table 4.5). This appears to be because this is the only
design method which explicitly allows for the effects of partial plugging on qb0.1.
(vi) In the slightly moist sand at Shenton Park, the measured CPT data and end
bearing capacities were sensitive to the seasonal effects (i.e. high qc values in
dry season, but lower qc values in wet season) due to presence of closely
surrounded trees (~30m from test site). This effect needs to taken into account in
the design of piles in any similar conditions.
(vii) The base stiffness of the driven piles at Shenton Park depended strongly on the
degree of plugging during installation (i.e. FFR), suggesting that the level of
pre-stress applied to the sand during installation can provide a guide to the
subsequent load-displacement response.
(viii) Results from 91 centrifuge static load tests on closed-ended jacked piles (in sand
deposits at various stress levels and with a range of relative densities) revealed
that ratio of qb0.1/qc does not vary systematically with qc, D, & Dr, and has a
constant value of about 0.9 (which agrees with the database study of field piles,
in Figure 8.2).
(ix) However, the resistance ratio in the centrifuge tests at relatively small
displacements, such as qb0.02/qc at a working displacement level w/D of 2%, does
increase with a decrease in Dr. Moreover, the back-analysed initial shear
stiffness ratios of the pile base (GIN/qc at w/D of ~2%) are seen to follow the
average trendline proposed by Rix & Stokoe (1992) for the very small strain in
situ stiffness (G0). This response is compatible with the high degree of over-
9-3
consolidation induced in the sand below the pile base by the preceding jacking
stokes. The ratio of GIN/qc is a key factor in determining the decay of stiffness
with displacement/strain level (i.e. G/GIN vs. w/D). At a certain w/D value, the
reduction in stiffness is greater for a higher ratio of GIN/qc. On the other hand
there appears to be a single relationship between G/GIN and qb/qc, which could
be represented by modified hyperbolic model, i.e. G/GIN = 1 - (qb/qc)4.
(x) The pile base resistance in sand was found to be rate dependent over the range
of 0.002mm/s to 1mm/s in saturated sample H due to the creep effect.
(xi) The base response of jacked centrifuge piles in sand is significantly influenced
by the presence of a weak clay layer within about 8D of the pile tip (i.e. in
sample H). These piles are characterised by relatively low initial stiffness ratios
(GIN/qc) and a faster decay of G/GIN with w/D.
(xii) Centrifuge tests in layered soil samples indicate that pile base installation
resistance and the base stiffness (e.g. GIN) in the strong soil are significantly
influenced by the approaching weak layer at a greater distance from the layer
interface, whereas the ones in the weak soil are less influenced by the
neighbouring strong soil. The normalised zone of influence (ZS/D) in the strong
soil ranges from about 5 for resistance ratios qc,W/qc,S of ≈ 0.4 to about 12 for
qc,W/qc,S ≈ 0.025 for the steady state penetration resistance.
(xiii) The trends indicated by penetrometer installation in layered soils in the

centrifuge could be replicated in numerical analyses, where the pile penetration
in a layered soil was idealised assuming the spherical cavity expansion analogue.
Based on analysis of 19 cases involving a wide range of resistance ratios
(qc,W/qc,S), an analysis procedure to evaluate variation of qc/qc,S with H/D has
been proposed (Section 8.4.1). The normalised zone of influence (ZS/D & ZW/D)
is found to be a function of the resistance ratio qc,W/qc,S.
9.3 RECOMMENDATIONS FOR FUTURE WORK
(i) The UWA database of pile load tests is an extension of databases used to derive
Fugro-05, ICP-05 & NGI-05. It is however still small in size. To further verify
the predictive performance of the CPT-based design methods, it is necessary to
9-4
maintain and update the UWA database by adding in any new available data. In
particular, efforts are required to collect more data for jacked piles before any
method could be put forward for general design use.
(ii) Although a simple tentative method for estimation of the degree of plugging (i.e.
FFR, in Figure 3.10) has been proposed based on limited available field data,
there is still a lack of data for a general evaluation of controlling factors for
determination of FFR. Further research is therefore needed in order to better
quantify the degree of plugging in a prediction method. Measurement of IFR is
advocated and should become routine in the same way that instrumentation (to
monitor concrete pressure vs. extraction rate variations) is now an integral part
of continuous flight augered piling.
(iii) As shown in Chapter 8, the effect of soil layering on pile base stiffness (e.g. GIN)
seems to be less pronounced than that on the steady state penetration resistance
(qc). It requires further investigation.
(iv) Spherical cavity expansion numerical analyses in the two-layer soil profile have
provided the basis for estimating the reduction in resistance (i.e. qc/qc,S). This
can be extended to study the base resistance response in a multi-layer soil profile,
which could potentially further improve the qc averaging technique.
9-5
APPENDIX A DATABASE ASSESSMENT OF CPT
BASED DESIGN METHODS
A.1 INTRODUCTION
In Chapter 3, the UWA database of high quality base capacity measurements is

discussed and employed to examine the relative reliability of the end bearing
formulation of four CPT based design methods (Fugro-05, ICP-05, NGI-05 & UWA-
05). This Appendix is intended to provide a complete reference to the full UWA
database of pile load tests in siliceous sands. It includes two papers which examine
the predictive performance of 7 pile design methods against the UWA database of
static load tests (i.e. 49 compression & 28 tension tests) on driven piles in siliceous
sands with adjacent CPT profiles. Tables A1 to A5 provide details of the pile load
tests considered in the full UWA database.
Reference:
Schneider, J. A., Xu, X., and Lehane, B. M. 2007a. Database assessment of CPT
based design methods for axial capacity of driven piles in siliceous sands: Part I –
overview of methods, database and predictive performance. Journal of Geotechnical
and Geoenvironmental Engineering, Submitted.
Schneider, J. A., Lehane, B. M. and Xu, X., 2007b. Database assessment of CPT
based design methods for axial capacity of driven piles in siliceous sands: Part II –
application to reliability analyses for offshore platforms. Journal of Geotechnical and
Geoenvironmental Engineering, Submitted.
A-1
Database assessment of CPT based design methods for axial capacity of driven piles
in siliceous sands: Part I – overview of methods, database and predictive performance
by
James A. Schneider1, Xiangtao Xu2, and Barry M. Lehane3
Prepared for ASCE Journal of Geotechnical and Geoenvironmental Engineering
1
Corresponding author
PhD student, School of Civil and Resource Engineering, The University of
Western Australia, Australia
2
3
Professor, School of Civil and Resource Engineering, The University of
1
ABSTRACT
Numerous CPT based methods exist for calculation of the axial pile capacity in sands,
but no clear guidance is presently available to assist designers in the selection of the
most appropriate method. To assist in this regard, this paper examines the predictive
performance of a range of pile design methods against a newly compiled database of
static load tests on driven piles in siliceous sands with adjacent cone penetration test
(CPT) profiles. Seven driven pile design methods are considered, including the
conventional American Petroleum Institute (API) approach, simplified CPT alpha
methods and four new CPT based methods, which are now presented in the
commentary of the 22nd edition of the API recommendations. Mean and standard
deviation database statistics for the design methods are presented for the entire 77 pile
database, as well as for smaller subset databases separated by pile material (steel and
concrete), end condition (open vs. closed), and direction of loading (tension vs.
compression). Certain methods are seen to exhibit bias towards length, relative
density, cone tip resistance and pile end condition. Other methods do not exhibit any
apparent bias (even though their formulations differ significantly) due to the limited
size of the database subsets and the large number of factors known to influence pile
capacity in sand. The database statistics for the best performing methods are
substantially better than those for the API approach and the simplified alpha methods.
Improved predictive reliability will emerge with an extension of the database and the
inclusion of additional important controlling factors affecting capacity.
2
INTRODUCTION
The static axial bearing capacity (Qt) of a deep foundation is given as the sum of the
shaft capacity (Qs) and base capacity (Qb):
z tip
Q t = Qs + Qb = P ∫ τ dz + q
f
z tip − L emb
b Ab (1)
where P is the pile perimeter, τf is the local ultimate shaft friction, qb is the ultimate
unit base resistance, Ab is the pile base area, ztip is the tip depth and Lemb is the
embedded pile length. The value of qb (which is zero for a pile loaded in tension) is
normally limited to that mobilized at a pile tip settlement of 10% of the pile diameter.
A wide range of empirical approaches are currently used to calculate τf and qb for
driven piles in coarse grained soils. This range has arisen because of the inadequacy
of existing theoretical methods for prediction of the soil response during and after pile
installation coupled with the large number of variables known to affect pile axial
capacities in sand (Randolph 2003). Furthermore, Dennis & Olson (1983) concluded
that while soil density influences shaft friction and base capacity of open ended piles,
interpretation of the influence of density is highly uncertain due to poor definition of
soil properties along the length of a pile. Use of the cone penetration test (CPT) for
site characterization provides a large amount of repeatable information on the vertical
variability of soil strength and compressibility and therefore leads to increased design
reliability (i.e., Briaud & Tucker 1988). This paper and its companion paper
(Schneider et al. 2007) seek to assist designers in making informed decisions
regarding axial pile capacity based on an understanding of the basis of a given
empirical approach and of the associated relative level of uncertainty and bias.
While site specific static (and dynamic) load tests reduce the level of uncertainty for
onshore practice, load tests are prohibitively expensive in the offshore environment
and reliance is placed on extrapolation of empirical correlations derived for onshore
applications (McClelland et al. 1969). Although the derivation of τf and qb in equation
3
(1) for offshore piles in sand has historically been based on ‘visual classification
methods’ placed within frameworks of bearing capacity theory and an earth pressure
approach (McClelland et al. 1969), four Cone Penetration Test (CPT) based methods
are now included in the commentary of the 22nd Edition of American Petroleum
Institute (API) recommended practice (RP2A) for fixed offshore structures (2006).
The CPT is often thought of as a model pile and has a long history of use for
estimations of static axial pile capacity (Delft 1936). There are, however, a wide
variation of CPT based design methods in use worldwide (e.g. DeCock et al. 2003)
because the factors controlling axial pile capacity are influenced by more soil
parameters than those affecting the CPT tip resistance, qc, (e.g. Lehane et al. 1993).
Empirical design methods rely on calibration with databases of static load tests. The
number of published high quality case histories with adjacent CPT profiles remains
surprisingly small and hence previous calibration exercises make arbitrary and often
subjective assumptions to allow for issues such as (i) the effect of the pile end
condition (closed, open, or partially plugged), (ii) the interface friction angle between
the in-situ sand and concrete, steel or timber pile shafts, (iii) design CPT qc profile,
(iv) the absence of reliable axial load data near pile tips to assess the relative
contributions of shaft friction and base resistance, and (v) the ratio of the shaft
friction in tension to that in compression. Adequate allowance for these factors
requires division of a database into relatively small database subsets.
This paper evaluates the predictive performance of the API RP2A static axial pile
design method and six CPT qc methods against a new database of load tests on driven
piles in siliceous sands at sites with CPT data1. Twenty design methods were assessed
during the initial phases of this study and two of the best performing simplified CPT
‘alpha’ methods are considered here. A general overview of the seven selected
methods is provided in the following after an initial description of the terms
uncertainty and bias, which aims to place the subsequent assessment of design
1
All CPT data were collected with electric cone penetrometers except adjacent to the concrete piles at
Drammen and pipe piles at Padre Island. Mechanical CPTs at Drammen were shown to be equivalent to electric
CPTs performed at a later date (Lunne et al. 2003) while the mechanical and electric CPT qc values in the medium
dense sands at Padre Island were considered equal.
4
method predictive performance in context. Method reliability is discussed in a
companion paper (Schneider et al. 2007) and a detailed discussion of end bearing
formulations is presented in Xu et al. (2007). Driven pile capacities in more
compressible calcareous/micaceous sands or in sands with high CPT friction ratios (Fr
> 3%) are not discussed in this paper.
UNCERTAINTY AND BIAS
The level of uncertainty associated with empirical correlations (such as used for
driven piles) can be reduced through increased site investigation and testing (Zhang et
al. 2004) and improved incorporation in the model of the mechanics governing the
behavior. The database of load tests on driven piles with adjacent CPT data reported
here is influenced to some degree by uncertainty in input parameters as well as model
uncertainty. This study minimizes errors associated with poor and discontinuous site
investigation data provided by the standard penetration test (i.e. Dennis & Olson
1983) as well as those induced by the pile capacity dependence on the installation
technique so that an assessment of the level of uncertainty related to any given
method’s capacity prediction may be obtained. Uncertainty in pile capacity
predictions due to site variability is not quantified since one representative CPT
profile was generally available for each site.
Statistical bias is defined as the systematic distortion of an expected result due to

neglect of controlling variables (Simpson & Weiner 1989). The influence of bias on
the quantification of (relative) reliability is discussed throughout this and the
companion paper. Developing a better understanding of where that bias lays and how
it influences calculations is necessary when using new design methods in practice.
The models/methods discussed in this paper make various assumptions related to
variables that influence the axial capacity of driven piles in siliceous sands. Linear
trend lines are fitted to plots of the ratio of the natural log of calculated to measured
capacity [ln(Qc/Qm)] against a design parameter (i.e., as defined in the Appendix, qc,
qc1N, D, L, Leff, IFR, time) to examine bias in design methods, or bias in subset
database characteristics. The influence of extrapolation bias due to design method
5
formulation on the calculation of relative reliability is discussed in the companion
paper (Schneider et al. 2007).
API METHOD
The API (2000, 2006) design method for assumes that both τf and qb varies in direct
proportion with the free field vertical effective stress (σ'v0), but its imposed limiting
values on both shaft friction (τf,lim) and end bearing (qb,lim) generally control the
capacity of long piles, such as those used offshore. τf is given as:
τ f = K f ⋅ σ' v 0 tan δ f = β ⋅ σ' v 0 ≤ τ f ,lim (2)
where Kf is a coefficient of lateral earth pressure and δf the interface friction angle
between the soil and pile wall. For open-ended pipe piles driven unplugged, it is usual
to adopt a Kf value of 0.8 for both tension and compression loading. Values of Kf for
full displacement piles (plugged or closed-end) are assumed to be 1.0, with
recommended values of δf contained in Table 1a. API (2006) combines Kf and δf into
the parameter β (=Kf·tanδf) in Table 1b, implying that it is not appropriate to modify
the empirical method with a measured δf.
The unit end bearing at a tip displacement of 10% of the pile diameter, qb0.1, is
calculated using the bearing capacity factor (Nq) and the effective overburden stress
(σ'v0), which must be less than a limiting value (qb,lim); see Table 1a/b:
q b 0.1 = N q σ' v 0 ≤ q b,lim (3)
Statistics from the database study of API RP2A presented in this paper are based on
the parameters in Table 1a for API-00. API (2006) removes some soils types from
Table 1a (which were considered to result in over-estimations of pile capacity), and
therefore API (2006) could not be applied to a large number of sites in the database.
CPT ‘ALPHA’ METHODS
Although sleeve friction, fs, measured during a CPT has been related to pile shaft
friction (e.g. Begemann 1965), due to greater variability in fs measurements (among
6
other issues), qc based methods for shaft friction generally dominate practice. These
methods can simply be referred to as “alpha” methods, since qc is related to qb and τf
by a factor, α, i.e.
q b = α b ⋅ q c,avg ≤ q b ,lim (4)
qc
τf = ≤ τ f ,lim (5)
αs
The calculation of pile end bearing from qc was the focus of early research in the
Netherlands, as piles were typically driven through soft clays to bear on a dense sand
layer. Meyerhof (1956) extended the research to include pile shaft friction, proposing
approximate values of αs of 200 in sandy soils with αb =1 (when the tip depth to
diameter ratio exceeded 10). In a summary of current European practice, De Cock et
al. (2003) report design αb values ranging from 0.4 to 1.0 and αs values ranging from
50 to 400 for sandy soils. Some alpha methods impose limits on the maximum values
of qb and τf.
Two popular versions of the alpha method are discussed in this paper, namely LCPC-
82 (Bustamane & Gianeselli 1982) and EF-97 (Eslami & Fellenius 1997). Parameters
for each of these methods are presented in Table 2. LCPC-82 is not considered
applicable to open ended piles (due to the database employed for its calibration),
while the single EF-97 formulation is presented as being equally applicable to pipe
piles, concrete piles and H-piles.
CPT BASED METHODS FOR OFFSHORE PILES
The four CPT based methods now included in the commentary of the new 22nd
edition of the API RP 2A recommendations (2006) are referred to as:
• Fugro-05 (Kolk et al. 2005a)
• ICP-05 (Jardine et al. 2005)
• NGI-05 (Clausen et al. 2005)
• UWA-05 (Lehane et al. 2005a).
7
These ‘offshore’ CPT methods are intended to better reflect the mechanisms that
influence capacity of displacement piles in sand, and their formulations are provided
in Tables 3 and 4. The more complicated form of these expressions, compared to the
‘alpha method’, has arisen due to attempts by the respective authors to allow for
factors, other than qc, affecting pile capacity. This higher level of detail is clearly
warranted given that the methods are based on calibrations with static tests on
relatively small onshore piles while their intended use is for much larger offshore
piles.
Formulations for shaft capacity
The local ultimate friction (τf) that can develop on the shaft of a displacement pile in
sand is a function of the radial stress after installation and equalization (σ'rc), the
change in radial stress during loading (Δσ'rd), and the interface friction angle (δf) as
(Lehane et al. 1993):
τ f = (σ' rc + Δσ' rd ) tan δ f = σ' rf tan δ f (6)
Two notable features included in the shaft capacity calculations for these four
methods are: (i) the recognition that τf in a given soil horizon reduces as a pile is
driven deeper, i.e., “friction fatigue” (Toolan et al. 1990, Lehane et al. 1993,
Randolph et al. 1994, White & Lehane 2004, and others) and (ii) open ended piles
tend to have lower τf values than closed ended piles. These issues are important for
long, large diameter offshore piles, which are primarily driven open ended in a coring
manner.
Although refined based on different assumptions, the shaft friction formulations of

ICP-05, Fugro-05, and UWA-05 are generally similar. The equation formats are
based on studies with the instrumented jacked closed ended Imperial College Model
Pile (Lehane et al. 1993, Chow 1997), which showed that radial stress after
installation and equalization at a given depth is related to cone tip resistance and the
distance above the pile tip (h) normalised by the pile radius (R). ICP-05 was extended
8
from the closed ended jacked model pile studies to open end driven piles using
studies performed at the Dunkirk test site in France and a thorough evaluation of a
database of pile load tests (Chow 1997).
Fugro-05 was modified from the design equations of Lehane & Jardine (1994) and
Chow (1997) and was primarily calibrated using τf values measured in load tests on
heavily instrumented 0.76m diameter driven pipe piles for EURIPIDES (Kolk et al.
2005b), Ras Tanajib II (Kolk et al. 2005c), and the micaceous sands at Jamuna Bridge
(Fugro 1995). A (relatively small) database of additional loads tests was used to assist
the calibration exercise (Kolk et al. 2005a). The method assumes that the interface
friction angle is relatively constant between sand types after installation of the pile
and that Δσ'rd is minimal for all database piles.
The separate effects of the pile end condition and friction fatigue are incorporated in
the Fugro-05 and ICP-05 methods using the h/R* term, where R* = (R2 – Ri2)0.5 and
Ri is the internal radius of a pipe pile. UWA-05 avoids the use of this lumped
approach and proposes a (h/D)-0.5 term to allow for friction fatigue and the following
effective area ratio term, Ars,eff, to allow for varying levels of soil displacement
induced in any given soil horizon during pile installation (White et al. 2005):
2
Di
A rs ,eff = 1 − IFR (7)
D2
where IFR is the incremental filling ratio. The value of Ars,eff is a measure of the soil
displacement induced during installation, which has been shown experimentally by
Gavin & Lehane (2003), and others, to influence the radial stresses developed on the
pile shaft. As Ars,eff equals (R*/R)2 when IFR is equal to unity, the h/R* term in
Fugro-05 and ICP-05 can be expressed as a product of (h/D)-c times the area ratio, Ar,
raised to the power of ‘c/2’ i.e. there is inter-dependence of the two terms capturing
the effects of friction fatigue and end condition, each of which is controlled by
different factors. Friction fatigue depends on the number of installation cycles (White
and Lehane 2004), while the degree of partial plugging tends to reduce as the pile
diameter increases.
9
The formulations proposed by NGI-05 differ in format to the other three CPT
methods. Cone tip resistance is included through a specified correlation between
nominal relative density and normalized cone tip resistance. Friction fatigue is based
on the floating triangle distribution of τf using z/L degradation of local shaft friction
(Toolan et al. 1990), as opposed to the pile diameter dependent terms, h/D or h/R,
recommended by other methods. As parameters for differences in end condition and
material type are fixed (unlike the three other offshore CPT based methods), they are
tied to the characteristics of the database and additional bias may arise when
extrapolating to other soil types and pile geometries.
All methods assume that the ultimate shaft friction mobilized in tension is lower than
that in compression. Fugro-05 assumes different τf distributions under tension and
compression loading while the three other methods apply equivalent ft/fc multiples on
the total shaft friction of between about 0.7 and 0.8.
Formulations for end bearing

End bearing formulations for the four ‘offshore’ CPT methods are summarized in
Table 4. Each method is quite different, with the ratio of qb to an average qc value in
the vicinity of the pile tip being assumed to vary with the pile diameter in ICP-05,
with relative density in NGI-05, with cone tip resistance in Fugro-05 and with the
effective area ratio in UWA-05. A critical review of these end bearing formulations is
presented in Xu et al. (2007).
DATABASE CHARACTERISTICS
A database of static load tests was compiled by the authors and is referred to here as
the UWA database. It was derived primarily from the geotechnical literature, with
information on additional case histories provided by Fugro Engineers, B.V., the
California Department of Transportation (Caltrans), the Florida Department of
Transportation, the Louisiana Transportation Research Center (LTRC), Imperial
College London, and the Norwegian Geotechnical Institute (NGI). Primary
differences between this database study and previous studies include:
10
y All CPT data were digitized to a depth interval of 0.1m or smaller. The
digitization of profiles minimizes ambiguity and errors associated with selection
of a “design profile”. The study showed that this level of detail was required to
obtain an adequate assessment of offshore CPT methods (each of which predicts
distributions of τf /qc along the pile shaft).
y Although the database contains static pile test data at sites which only contain
SPT data, these data were not included in this assessment of CPT qc-based
correlations because of the significant uncertainty associated with deriving an
equivalent qc value from an SPT N blowcount and the reduced level of detail in
vertical soil variability provided by a typical SPT profile.
y Piles included in the database were limited to impact driven piles. No jacked
piles, vibratory driven piles or bored piles were included due to the influence of
installation method on pile capacity.
y The database was limited to sites where the pile tip was bearing in a siliceous
sand layer and where siliceous sand contributed to more than 50% of the shaft
capacity.
In total, the data from over 200 pile load tests were processed. This database was
filtered and subdivided into five categories comprising a total of 77 load tests; the
statistics presented in this paper relate to these load tests. Typical reasons for
exclusion of pile load tests from database analyses are summarized above. In
addition, (i) re-tests on piles at the same pile embedment; (ii) piles with a diameter
less than 0.2m and (iii) piles with an embedded length less than 5m were all omitted
from consideration. Characteristics of the pile load tests considered within this study
are contained in Tables A1 to A5 of the appendix to this paper.
The database subsets examined are:

• Closed Ended Concrete Piles in Compression (CECC) – 17 piles
• Closed Ended Steel Piles in Compression (CECS) – 15 piles
• Closed Ended Piles in Tension (CET) – 12 piles
• Open Ended Steel Pipe Piles in Compression (OEC) – 17 piles
• Open Ended Steel Pipe Piles in Tension (OET) – 16 piles
11
The database of closed ended piles in tension was not separated further into steel and
concrete piles due to the small number of load tests from different sites for this
category of test. The use of this database subset was justified later when insignificant
bias was observed when comparing closed ended steel and concrete piles in
compression.
Pile capacity in tension was defined as the maximum uplift load less the pile weight.
Pile capacity in compression was defined at a tip displacement of 10% of the pile
diameter (0.1D). In 6 of the 77 cases, neither a tip displacement of 0.1D or a plunging
failure was reached during a load test. For those cases, hyperbolic extrapolation using
the Chin (1978) method was used to estimate capacity at a displacement of 0.1D.
Extrapolation typically led to a 5% increase in the maximum measured capacity and
did not exceed 15%. Maximum measured loads and associated tip displacements are
also indicated in Tables A1 to A5, when available.
Histograms of soil relative density, pile geometry, and load test characteristics are
presented in Figure 1 for the UWA database. Nominal relative density for an assumed
normally consolidated sand (Dr) has been estimated as a function of the normalized
cone tip resistance (qc1N) using the following empirical relationship (Jamiolkowski et
al. 2003):
D r = 0.35 ln(q c1N 20 ) where (8a)
qc1N=(qc/pref)/(σ'v0/pref)0.5 and pref=100 kPa (8b)

To account for soil layering, the plotted relative densities have been weighted based
on the shaft friction distribution estimated using the UWA-05 method.
For 64 of the 77 piles, CPT friction ratios were also available at the test sites.
Weighted average normalized tip resistance and friction ratio (Fr) for each site are
plotted on a soil behavior type chart in Figure 2, which was developed based on
recommendations of Robertson & Wride (1998).
12
Based on Figures 1 and 2, it is apparent that:
y Piles in the database typically have a capacity less than 5 MN.
y The database of 77 piles consists primarily of piles with diameters less than 0.8m.
y While pile lengths vary from 5m to 80m, most piles are between 10m and 20m.
y The 77 pile database contains a relatively even and wide range of nominal sand
relative density, although individual subset databases may be biased towards
specific sand densities.
y The time between installation and load testing is typically less than 4 weeks.
y Most soils in the database classify as clean sand to silty sand in Zone 6 with some
points in Zone 5 (silty sand to sandy silt). Assuming a near zero value of the pore
pressure parameter Bq, these soil behavior types correspond to Zones 5 and 4 of
the Eslami & Fellenius (1997) classification chart mentioned in Table 2.
Future analyses of the omitted tests will provide more information to assist CPT
based design for piles in non-siliceous sands, residual soils and aged deposits. It is
noted that all of the prediction methods considered tended to over-predict pile
capacity in calcareous or micaceous sands and under-predict pile capacity in residual
soils. In addition, measured capacities at a given site tended to increase with time.
CALCULATION PROCEDURES
A spreadsheet was developed to calculate axial pile capacity in sandy soils based on
CPT parameters. For shaft friction calculations, discretization for numerical
integration was set at 0.1m intervals. Method input and the accuracy of numerical
integration using the spreadsheet solution was validated using integral solutions of
pile capacity in uniform sand profiles. Soils with constant relative density of
approximately 0.4 and 0.8 were analysed for open and closed ended piles of 10m and
70m length, with diameters of 0.5m and 1m. The verification studies showed that
spreadsheet calculations were within 0.5% of the integral solutions for all methods
and cases using a discretization interval of 0.1m (Lehane et al. 2005b).
The following assumptions were also made in the absence of site specific data:
y The in-situ bulk unit weight was 19 kN/m3.
13
y For the ICP-05 and UWA-05 methods, if δf was not measured in a laboratory
interface tests, δf was estimated as a function of D50 or sand description (i.e., fine,
medium, coarse) as discussed in Lehane et al. (2005a). If no grading data were
available, δf was assumed to be 29o.
y For UWA-05, when profiles of IFR were not available, IFR was assumed equal to
the final filling ratio (FFR) and taken as either the plug length ratio (PLR) or
estimated as a function of pile inner diameter (Lehane et al. 2005a).
y Based on a number of studies relating CPT data to shaft friction of displacement
piles in clay (e.g. Lehane et al. 2000) and to avoid introducing unnecessary
complexity, the analyses assumed a fully equalised ultimate shaft friction value of
qt/35 to calculate the contribution of the clay layers within the sand profiles at the
test sites. The time between installation and load testing of the test piles was such
that, for those involving clay layers, full equalisation of pore pressures was
expected to be effectively completed.
Additionally, the equivalent radius (Req) or diameter (Deq) for square piles (for use in
‘offshore’ pile design method ‘friction fatigue’ formulations) is derived assuming
equivalence in area i.e. Deq = 2B/π0.5, where B is the side width of the square pile
(Jardine et al. 2005).
RELATIVE PERFORMANCE OF METHODS
The ratio of the calculated to the measured capacities (Qc/Qm) for each method was
evaluated and the predictive performance was expressed in terms of the geometric
mean (μgR) and the standard deviation of the natural log of the Qc/Qm ratios (σlnR).
These statistical parameters, as well as sample median, are summarized in Tables 5
through 8 for each of the database subsets discussed above, as well for the entire
database. The LCPC-82 method may not have been initially intended for application
to open ended piles, but performance is presented for discussion purposes. To prevent
bias towards sites with multiple tests of similar pile geometry, the average Qc/Qm
ratio of similar piles at the same site was used in statistical analysis. A pile with a
similar geometry was defined as one constructed of the same material with the same
end condition in a sand deposit where the qc values at the respective pile locations are
within 10% of their average and where diameters and pile lengths differ by less than
14
15%. Multiple piles of similar geometry at the same site are indicated in the Tables
A1 to A5.
Figure 3 illustrates the well known (Toolan et al. 1990, Chow 1997) bias towards
density (weighted average normalized cone tip resistance along the pile shaft) and tip
depth for API-00. Figure 4 shows bias plots towards weighted average cone tip
resistance for the six CPT based methods discussed. To account for the influence of
layered profiles on average parameters, cone tip resistance is weighted by the inferred
local shaft friction at a specific depth. The shaft friction distribution of the UWA-05
method is used for weighted averaging, although, similar values are obtained when
using shaft friction distributions of ICP-05 or NGI-05.
Inspection of Figures 3 and 4 along with statistics provided in Tables 5 through 8

indicates inconsistent performance between methods for each database subset. Some
of the differences are a function of the soil consistency and pile geometry within each
database subset, although differences also result from method formulation. General
observations of method performance are summarized as follows:
y API-00 has the highest coefficient of variation (COV 2 ) for each of the subset
databases, which is two to four times higher than the COV of CPT based
methods.
y Bias in length and density result in apparently conservative statistics towards the
database for API-00, although the same method gives Qc/Qm =3.4 for the tension
test on the 23.5m long pile in loose sand at Drammen, Norway. This tendency
towards un-conservatism for longer piles in looser sand has been identified
previously (e.g., Toolan et al. 1990) and is discussed in more detail in Schneider
et al. (2007).
y As indicated in Figure 4, all CPT methods significantly overpredicted the
compression capacity of closed ended steel piles at Lock & Dam 26 by
approximately a factor of two. It is uncertain whether the poor predictions
resulted from scale effects on CPT qc measurements in the gravelly sand, site
2
COV=[exp(σln2)-1]0.5, and is approximately equal to σln for σln < 0.5.
15
variability or other factors. In the original reference, Briaud et al. (1989b) also
note that the measured capacities are unusually low as compared to predictions
using CPT based methods.
y No consistent trends were observed regarding the ability of a given method to
predict the difference in capacity between a steel and concrete pile. Each method
accounts for steel and concrete piles in different ways, with EF-97 and Fugro-05
not distinguishing between the two material types. Evidently, due to the small
database of sites and pile sizes, variability in method performance due to pile
geometry and soil consistency exceeds variability in pile capacity due to material
type.
y LCPC-82 tends to have a bias towards cone tip resistance. On average, the
method over-predicts the capacity of database piles by about 40%. Predictions
can, however, exceed measured capacities by a factor of up to 4.
y The use of a constant αs for sand in the EF-97 method tends to eliminate the bias
towards cone tip resistance observed for LCPC-82.
y The simplified α CPT methods over-predict the capacity of open-ended piles in
tension and compression. As these methods do not recommend differences in
capacity calculation for open-ended piles, and may or may not have been intended
for analysis of these types of piles, differences in method performance highlight
the importance of distinguishing open and closed ended piles. Predictions for
piles in compression are much worse than for piles in tension, indicating
significant errors in the respective end bearing formulations; see Xu et al. (2007).
y Fugro-05 tends to over-predict the capacity of piles in compression and slightly
under-predict the capacity of piles in tension. This tendency is believed to be
partly related to an end bearing formulation, which allows qb/qc to be greater than
unity (Xu et al. 2007), and to the comparatively high friction fatigue (h/R*)
exponent of 0.85 to 0.9.
y NGI-05 tends to be slightly less conservative (μg ≈ 0.9 to 1.1) and less precise
(σlnR = 0.25 to 0.4) than UWA-05 and ICP-05 (μg ≈ 0.85 to 1.0, σlnR ≈ 0.2 to 0.3).
y UWA-05 generally has improved performance for the database of open ended
piles in compression (μgR=0.96; σlnR=0.19) as compared to other methods. This is
primarily considered due to treatment of the influence of partial plugging on end
16
bearing (as well as shaft friction) within the framework for UWA-05 (Xu et al.
2007).
SUMMARY AND CONCLUSIONS
Static axial pile design in sands is still an area of great uncertainty. This paper has
discussed a number of static design methods, which roughly cover the evolution from
an earth pressure based approach to simplified CPT alpha methods, to more detailed
CPT based methods developed for extrapolation to offshore piles. A relatively large
database of pile test sites in siliceous sands with CPT data is presented, which covers
a representative range of sand relative density, although pile geometry is more typical
of onshore conditions than those offshore. Database studies show that the COV has
reduced with each of those improvements in design theory, with API-00 having a σlnR
of 0.4 to 0.9 which is biased to database subset characteristics, CPT alpha methods
having a σlnR on the order of 0.3 to 0.6, and offshore CPT methods having a σlnR of
0.2 to 0.4.
While this trend of improving statistics is encouraging, the following limitations

should be noted:
y As pile geometry (length, diameter, end condition, direction of loading) and soil
profile (density, in situ stress state, soil layering) appear to be the most significant
factors influencing axial pile capacity for the range of pile sizes within the
database, the relative size of the subset databases is too small to assess the
influence of minor features. These minor features may be of significant
importance when extrapolating to pile geometries and soil conditions outside of
the database.
y Many more load tests would be necessary to attempt an unbiased regression
analysis, and presented statistics are significantly biased towards the soil and pile
conditions within this database. Direct application of statistics from database
studies to calibration of resistance and safety factors and reliability analyses
should be used with caution, as the potential for extrapolation bias may be more
significant than σlnR. Applications of these results to reliability analyses for
offshore piles are addressed in more detail by Schneider et al. (2007).
17
Withstanding these limitations, the evaluation presented in this paper has indicated
that:
y The API-00 method performs poorly against database piles. An investigation into
bias of the method (Schneider et al. 2007) is required to explain why the method
has shown acceptable performance in practice.
y CPT alpha methods need to distinguish between open and closed ended piles.
Assuming qb/qc of unity within EF-97 leads to much larger errors for open ended
piles than qb/qc of 0.4 to 0.5 recommended by LCPC-82. Extrapolation of alpha
methods to piles outside of database characteristics may be expected to lead to
greater uncertainty than the four ‘offshore’ CPT-based methods.
y The more detailed method formulation of UWA-05 (which is based on recent
research into the controlling mechanisms influence pile capacity in sands) is
thought to be the primary reason for its slightly better predictive performance
against the database than that of ICP-05, NGI-05 and Fugro-05.
y None of the published methods are equipped to deal with the complex time
dependant behavior of driven piles in sands.
ACKNOWLEDGEMENTS
The authors gratefully acknowledge the support provided by the Australian Research
Council. The first and second authors were also supported through International
Postgraduate Research Scholarships and University Postgraduate Awards from the
University of Western Australia. The authors wish to acknowledge the assistance and
willingness to share data and opinions of various personnel involved in the
development of the Fugro-05, ICP-05, and NGI-05 methods.
REFERENCES
Altaee, A., Fellenius, B. H. & Evgin, E. (1992). "Axial load transfer for piles in sand.
I. Tests on an instrumented precast pile." Canadian Geotech. J., 29(1), 11-20.
API (2000). Recommended Practice for Planning, Designing, and Constructing Fixed
Offshore Platforms – Working Stress Design, API RP2A, 21st Edition,
American Petroleum Institute, Washington, D.C.
18
API (2006). DRAFT Recommended Practice for Planning, Designing, and
Constructing Fixed Offshore Platforms – Working Stress Design, API RP2A,
22nd Edition, American Petroleum Institute, Washington, D.C.
Appendino, M. (1981). "Interpretation of axial load tests on long piles." Proc., 10th
Int. Conf. Soil Mech. Foundation Eng., Stockholm, Vol. 2, 593-598.
Axelsson, G. (2000). "Long term setup of driven piles in sand." PhD Thesis, Royal
Institute of Technology, Stockholm.
Baldi, G., Bellotti, R., Ghionna, V., Jamiolkowski, M., and Pasqualini, E. (1986).
"Interpretation of CPTs and CPTUs; 2nd part: drained penetration of sands."
Proc., 4th Int. Geotech. Seminar, Singapore, 143-156.
BCP-Committee (1971). "Field tests on piles in sand." Soils and Foundations, 11(2),
29-49.
Begemann, H.K.S. Ph. (1965). "The maximum pulling force of a single tension pile
calculated on the basis of results of the adhesion jacket cone." Proc. 6th Int.
Conf. Soil Mech. Foundation Eng., 2, Montreal, 229-233.
Beringen, F. L., Windle, D. & Van Hooydonk, W. R. (1979). "Results of loading tests
on driven piles in sand." Recent development in the design and construction of
piles, ICE, London, 213-225.
Briaud, J.-L., and Tucker, L.M. (1988). "Measured and predicted axial response of 98
piles," J. Geotech. Eng., 114(9), 984-1001.
Briaud, J.-L., Tucker, L. M., and Ng, E. (1989a). "Axially loaded 5 pile group and
single pile in sand." Proc., 12th Int. Conf. Soil Mech. Foundaton Eng., Rio de
Janeiro, 1121-1124.
Briaud, J.-L., Moore, B. H., and Mitchell, G. B. (1989b). "Analysis of pile load tests
at Lock and Dam 26." Foundation Engineering: Current Principles and
Practices, Evanston, 925-942.
Bustamante, M., and Gianeselli, L. (1982). "Pile bearing capacity prediction by
means of static penetrometer CPT." Proc., 2nd European Symp. on
Penetration Test., Amsterdam, 493-500.
Chin, F.K. (1978). "Diagnosis of pile condition." Geotech. Eng., 85-104.
Chow, F. C. (1997). "Investigations into the behaviour of displacement piles for
offshore foundations." PhD Thesis, Imperial College, London.
19
Clausen, C. J. F., Aas, P. M., and Karlsrud, K. (2005). "Bearing capacity of driven
piles in sand, the NGI approach." Proc., Int. Symp. Frontiers Offshore
Geomech. ISFOG, Perth, 677-681.
De Cock, F., Legrand, C., and Huybrechts, N. (2003). "Overview of design methods
of axially loaded piles in Europe - Report of ERTC3-Piles, ISSMGE
Subcommittee." Proc., 8th European Conf. on Soil Mech. and Geotech. Eng.,
Prague, 663-715.
Delft Laboratory of Soil Mechanics (Delft) (1936). "The predetermination of the
required length and the prediction of the toe resistance of piles." Proc. 1st Int.
Conf. on Soil Mech. and Foundation Eng., Cambridge, 181-184.
Dennis, N.D., Jr., and Olson, R.E. (1983). “Axial capacity of steel pipe piles in sand.”
Proc. of the Conf. on Geotech. Practice in Offshore Eng., Austin, Texas,
USA, 389-402.
Eslami, A., and Fellenius, B. (1997). "Pile capacity by direct CPT and CPTu methods
applied to 102 case histories." Canadian Geotech. J., 34, 886-904.
Fugro (1995). "Reduced scale pile load tests, Jamuna Bridge, Bangladesh." Report
No. K-2380/206.
Gavin, K. G., and Lehane, B. M. (2003). "The shaft capacity of pipe piles in sand."
Canadian Geotech. J., 40(1), 36-45.
Gregersen, O. S., Aas, G., and DiBiagio, E. (1973). "Load tests on friction piles in
loose sand." Proc., 8th Int. Conf. Soil Mech. and Foundation Eng., 2.1,
Moscow, 109-117.
Jamiolkowski, M. B., Lo Presti, D. F. C., and Manassero, M. (2003). "Evaluation of
relative density and shear strength of sands from cone penetration test." Soil
behaviour and soft ground construction, Geotechnical Special Publication No.
119, Reston, VA, 201-238.
Jardine, R. J., Chow, F. C., Overy, R. F., and Standing, J. R. (2005). ICP design
methods for driven piles in sands and clays. Thomas Telford, London, 97.
Jardine, R.J., and Standing, J. R. (2000). "Pile load testing performed for HSE cyclic
loading study at Dunkirk, France." Report OTO 2000 007, Health & Safety
Executive, London.
20
Jendeby, L., Noren, C., and Rankka, K. (1994). "Friction piles in sand - prediction of
bearing capacity and load/displacement curve." Proc., Int. Conf. and
Exhibition on Piling and Deep Foundations, 5, DFI 94, Bruges, 3.6.1-3.6.5.
Kolk, H. J., Baaijens, A. E., and Senders, M. (2005a). "Design criteria for pipe piles
in silica sands." Proc., Int. Symp. Frontiers Offshore Geomech. ISFOG, Perth,
711-716.
Kolk, H. J., Baaijens, A. E., and Vergobi, P. (2005b). "Results of axial load tests on
pipe piles in very dense sands: the EURIPIDES JIP." Proc., Int. Symp.
Frontiers Offshore Geomech. ISFOG, Perth, 661-667.
Kolk, H. J., Baaijens, A. E., Al-Shafei, K., and Dakhil, O.A. (2005c). "Axial load
tests on pipe piles in very dense sands at Ras Tanajib." Proc., Int. Symp.
Lehane, B.M., Jardine, R.J., Bond, A.J., and Frank, R. (1993). "Mechanisms of shaft
friction in sand from instrumented pile tests." J. Geotech. Eng., 119 (1), 19-
35.
Lehane, B.M., and Jardine, R.J. (1994). "Shaft capacity of driven piles in sand: a new
design approach." Proc., 7th Int. Conf. on Behaviour of Offshore Struct., MIT,
Cambridge, MA, 23-36.
Lehane, B. M., Chow, F. C., McCabe, B. A., and Jardine, R. J. (2000). "Relationships
between shaft capacity of driven piles and CPT end resistance." Proc. of the
Institution of Civil Eng., Geotech. Eng., 143(2), 93-101.
Lehane, B.M., Schneider, J.A., and Xu, X. (2005a). “The UWA-05 method for
prediction of axial capacity of driven piles in sand.” Proc., Int. Symp.
Lehane, B.M., Schneider, J.A., and Xu, X. (2005b). “A review of design methods for
offshore driven piles in siliceous sand.” UWA Report GEO 05358, The
University of Western Australia, Perth, 105.
Lunne, T., Long, M., and Forsberg, C.F. (2003). "Characterisation and engineering
properties of Holman, Drammen sand." Characterisation and Eng. Prop. of
Nat. Soils, 1121-1148.
21
McClelland, B., Focht, J.A., and Emrich, W.J. (1969). “Problems in design and
installation of offshore piles.” J. Soil Mech. Found. Div., 95(SM6), 1491-
1514.
McClelland, B. (1974). "Design of deep penetration piles for ocean structures." J.
Geotech. Eng. Div., 100(GT7), 709-747.
Meyerhof, G.G. (1956). "Penetration tests and bearing capacity of cohesionless soils."
J. Soil Mech. and Foundation Div., 82(SM1), Paper 886, 19.
Nevels, J. B. J., and Snethen, D. R. (1994). "Comparison of settlement predictions for
single pile in sand based on penetration test results." Proc., Conference on
Vertical and Horizontal Deformations of Foundations and Embankments,
College Station, TX, 1028-1038.
Olson, R. E., and Shantz, T. J. (2004). "Axial load capacity of piles in California in
cohesionless soils." Deep Foundations 2002, Geotechnical Special
Publication No. 116, ASCE, Reston, VA, 1-15.
Paik, K., Salgado, R., Lee, J., and Kim, B. (2003). "Behavior of open- and closed-
ended piles driven into sands." J. Geotech. Geoenviron. Eng., 129(4), 296-
306.
Pump, W., Korista, S., and Scott, J. (1998). "Installation & load tests of deep piles in
Shanghai alluvium." Proc., VII Int. Conf. on Piling & Deep Foundations, 1,
Vienna, DFI, 31-36.
Randolph, M.F., Dolwin. J., and Beck, R. (1994). "Design of driven piles in sand."
Géotechnique, 44(3), 427-448.
Randolph, M. F. (2003). "Science and empiricism in pile foundation design."
Robertson, P.K., and Wride, C.E. (1998). “Evaluating cyclic liquefaction potential
using cone penetration test.” Canadian Geotech. J., 35(3), 442-459.
Schneider, J.A., Lehane, B.M., and Xu, X. 2006. "Database assessment of CPT based
design methods for axial capacity of driven piles in siliceous sands: Part II –
application to reliability analyses for offshore platforms." J. Geotech.
Geoenviron. Eng., submitted.
Shioi, Y., Yoshida, O., Meta, T., and Homma, M. (1992). "Estimation of bearing
capacity of steel pipe pile by static loading test and stress-wave theory (Trans-
22
Tokyo Bay Highway)." Application of stress-wave theory to piles, Balkema,
Rotterdam, 325-330.
Simpson, J.A., and Weiner, E.S.C. (1989). The Oxford English Dictionary, 2nd
Edition, Vol. II (B.B.C. - Chalypsography), Clarendon Press, Oxford.
Titi, H., and Abu-Farsakh, M. (1999). "Evaluation of bearing capacity of piles from
cone penetration tests data." Louisiana Transportation Research Centre.
Toolan, F. E., Lings, M. L., and Mirza, U. A. (1990). "An appraisal of API RP2A
recommendations for determining skin friction of piles in sand." Proc., 22nd
Offshore Technol. Conf., OTC 6422, Houston, Tex., 33-42.
Tveldt, G., and Fredriksen, F. (2003). "N18 Ny motorvegbru I Drammen
Prøvebelasting av peler." Fjellsprengningsteknikk, Bergmekanikk/Geoteknikk,
37.1 - 37.32.
Vesic, A. S. (1970). "Tests on instrumented piles, Ogeechee River site." J. of the Soil
Mech. and Foundations Div., 96(SM2), 561-584.
White, D.J., and Lehane, B.M. (2004). "Friction fatigue on displacement piles in
sand." Géotechnique, 54(10), 645-658.
White, D. J., Schneider, J. A., and Lehane, B. M. (2005). "The influence of effective
area ratio on shaft friction of displacement piles in sand." Proc., Int. Symp.
Williams, R. E., Chow, F. C., and Jardine, R. J. (1997). "Unexpected behaviour of
large diameter tubular steel piles." Int. Conf. on Foundation Failures,
Singapore.
Xu, X., Schneider, J.A., and Lehane, B.M. (2007). "End bearing of open and closed
ended driven piles in siliceous sand using the CPT." Canadian Geotech. J.,
submitted.
Yen, T.-L., Chin, C.-T., and Wang, R. F. (1989). "Interpretation of instrumented
driven steel pipe piles." Foundation Engineering: Current Principles and
Practices, Evanston, IL, 1293-1308.
Zhang, L., Tang, W.H., Zhang, L., and Zheng, J. (2004). “Reducing uncertainty of
prediction from empirical correlations.” J. Geotech. Geoenviron. Eng., 130(5),
526-534.
23
Table 1a. API RP2A (2000) design guidelines for non-cohesive soils*
Density*** Soil Soil-Pile Limiting Shaft Nq Limiting Unit End
Description Friction Friction Bearing Values
Angle, δf Values kips/ft2 kips/ft2 (MPa)
Degrees (kPa)
Very loose Sand 15 1.0 (47.8) 8 40 (1.9)
Loose Sand-Silt **
Medium Silt
Loose Sand 20 1.4 (67.0) 12 60 (2.9)

Medium Sand-Silt**
Dense Silt
Medium Sand 25 1.7 (81.3) 20 100 (4.8)
Dense Sand-Silt**
Dense Sand 30 2.0 (95.7) 40 200 (9.6)
Very Dense Sand-Silt**
Dense Gravel 35 2.4 (114.8) 50 250 (12.0)
Very Dense Sand
*
The parameters listed in this table are intended as guidelines only. Where detailed information such as in situ
cone tests, strength tests on high quality samples, model tests, or pile driving performance is available, other
values may be justified.
**
Sand-silt includes those soils with significant fractions of both sand and silt. Strength values generally increase
with increasing sand fractions and decrease with increasing silt fractions.
***
The following expression of Baldi et al. (1986) is recommended by API (2006) for evaluation of relative
density (Dr) from CPT data in clean siliceous sands: Dr=1/2.93 ln[(qc/pref)/(21.5(p’0/pref)0.51)]; where pa is a
reference stress taken as 100 kPa and p'0 is the initial in situ mean effective stress [(σ'v0 + 2σ'h0)/3].
24
Table 1b. API RP2A (2006) design guidelines for non-cohesive soils1
Relative Soil Shaft Limiting Shaft Nq Limiting Unit End
Density2 Description Friction Friction Bearing Values
Factor, Values kips/ft2 kips/ft2 (MPa)
β3 (kPa)
Very Loose Sand Not Not Not Not
Loose Sand Applicable5 Applicable5 Applicable5 Applicable5
Loose Sand-Silt4
Medium Dense Silt
Dense Silt
Dense Gravel
Medium Dense Sand-Silt4 0.29 1.4 (67) 12 60 (3)
Medium Sand 0.37 1.7 (81) 20 100 (5)
Dense Sand-Silt4
Dense Sand 0.46 2.0 (96) 40 200 (10)
Very Dense Sand-Silt4
Very Dense Sand 0.56 2.4 (115) 50 250 (12)
1
The parameters listed in this table are intended as guidelines only. Where detailed information such as CPT
records, strength tests on high quality samples, model tests, or pile driving performance is available, other values
may be justified.
2
The following definitions for relative density description are applicable:
Description Relative Density [%]
Very Loose 0 - 15
Loose 15 - 35
Medium Dense 35 - 65
Dense 65 - 85
Very Dense 85 – 100
The expression of Baldi et al. (1986) (presented in Table 1a) is recommended for evaluation of Dr from CPT data
in clean siliceous sands
3
Shaft friction parameter β (equivalent to the “K tanδf” term used in previous API RP2A WSD publications) was
introduced in this publication to avoid confusion with the δf parameter used in the Commentary.
4
Sand-Silt includes those soils with significant fractions of both sand and silt. Strength values generally increase
with increasing sand fractions and decrease with increasing silt fractions.
5
Design parameters given in previous API RP2A publications for these soil/relative density combinations may be
un-conservative. Hence it is recommended to either use CPT based methods from the Commentary or apply
engineering judgment for these soils.
25
Table 2. Input parameters for selected CPT alpha methods in sandy soils
Method Applicable Applicable qc αb End Bearing αs τf,lim Reference
Pile Type Material Range Averaging (kPa)
(MPa)
LCPC- Closed Steel <5 0.5 Arithmetic 120 35 Bustamane &
82 5 to 12 0.5 mean +/-2D 200 80 Gianeselli 1982
> 12 0.4 200 120
Closed Concrete <5 0.5 Arithmetic 60 35 Bustamane &
5 to 12 0.5 mean +/-2D 100 80 Gianeselli 1982
> 12 0.4 150 120
EF-97 pipe, Steel & Class 51 1 Effective 250 None Eslami &
square, Concrete Class 41 1 geometric 100 None Fellenius 1997
round, average2
octagonal,
H section,
round, and
triangular
1
Specific soil classification based on qc, fs and u2 as described in Eslami & Fellenius (1997); Class 5 -
sand and gravel; Class 4 – silty sand and sandy silt; other categories not discussed.
2
Zone of averaging varies from 4b below pile tip to 8b above pile tip if pile is installed from weak soil
into dense soil, and from 4b below pile toe to 2b above pile toe when pile installed from dense soil into
a weak soil.
26
Table 3. Offshore CPT based design methods for local shaft friction of driven piles in
siliceous sand
0.05 −0.90
⎛ σ' ⎞ ⎛ h ⎞
τ f = 0.08 ⋅ q c ⋅ ⎜⎜ v 0 ⎟⎟ ⎜ ⎟ compression loading for h/R* > 4
⎝ p ref ⎠ ⎝ R *⎠
0.05
⎛ σ' ⎞
Fugro-05 3
τ f = 0.08 ⋅ q c ⋅ ⎜⎜ v 0 ⎟⎟ (4)−0.90 ⎛⎜ h ⎞
⎟ compression loading for h/R* < 4
⎝ p ref ⎠ ⎝ 4R * ⎠
0.15 −0.85
⎛ σ' ⎞ ⎡ ⎛ h ⎞⎤
τ f = 0.045 ⋅ q c ⎜⎜ v 0 ⎟⎟ ⎢max⎜ R * ,4 ⎟⎥ tension loading
⎝ p ref ⎠ ⎣ ⎝ ⎠⎦
⎡ ⎛ σ' ⎞
0.13
⎡ ⎛ h ⎞⎤
−0.38
⎤
τ f = a ⎢0.029 ⋅ b ⋅ q c ⎜⎜ v 0 ⎟⎟ ⎢max⎜ R * ,8 ⎟⎥ + Δσ' rd ⎥ tan δ f
ICP-05 ⎢⎣ ⎝ p ref ⎠ ⎣ ⎝ ⎠⎦ ⎥⎦
a= 0.9 for open ended piles in tension and 1.0 for all other cases
b = 0.8 for piles in tension and 1.0 for piles in compression
τ f = z L ⋅ p ref ⋅ FDr ⋅ Fsig ⋅ Ftip ⋅ Fload ⋅ Fmat ≥ τ min
FDr = 2.1(D r − 0.1)
1 .7
D r = 0.4 ln (q c1N 22 ) nominal relative density (which may be

greater than 1.0)
( )
NGI-05
Fsig = σ'v 0 p p a
0.25
Ftip = 1.0 for driven open ended and 1.6 for driven closed ended
Fload = 1.0 for tension and 1.3 for compression
Fmat = 1.0 for steel and 1.2 for concrete
τmin = 0.1·σ'v0
ft ⎡ ⎡ ⎛h ⎞ ⎤
−0.5
⎤
τf = ⎢0.03 ⋅ q c ⋅ A rs,eff ⎢max⎜ ,2 ⎟ ⎥ + Δσ' rd ⎥ tan δ f
0 .3
fc ⎢⎣ ⎣⎢ ⎝ D ⎠ ⎦⎥ ⎥⎦
UWA-05 Ars,eff (effective area ratio) = 1 - IFR × ( Di / D ) 2;
Ar (area ratio) = 1 - ( Di / D ) 2 ;
IFR (= ΔLp / Δz), when IFR is not measured, average IFR = min[1, (Di(m)/1.5)0.2 ]
ft/fc = ratio of tension to compression capacity (equal to 1 for compression and 0.75 in
tension)
τf = local ultimate shaft friction; δf = interface friction angle; pref = 100 kPa = reference
stress
L = pile length; z = element depth; h = height above pile tip
R* = equivalent pile radius =(R2-Ri2)0.5 where Ri is the internal pile radius =Di/2
Notes For noncircular piles, an equivalent circular is used to assess R* and D
Δσ’rd ≈ 4GΔy/D = change in radial stress during pile loading
G ≈ 185qc × qc1N-0.7 = operational level of shear modulus; see equation 8b for calculation
of qc1N
Δy ≈ 0.02mm = radial displacement during pile loading
3
The format of the equation for compression shaft friction within 4R* of the pile tip differs slightly from Kolk
et al (2005a) due to a typing error in that paper.
27
Table 4. Offshore CPT based design methods for base resistance of driven piles in
siliceous sand
End
Condition
Closed- &
Fugro-05 qb0.1 / qc,avg = 8.5 × ( pref / qc,avg ) 0.5 × Ar 0.25
Open-Ended
Closed-Ended qb0.1 / qc,avg = maximum [ 1 - 0.5 × log( D / DCPT ), 0.3 ]
if Di ≥ 2.0 × (Dr ─ 0.3) or Di ≥ 0.083 × qc,avg / pref × DCPT, Di in meter, then
ICP-05 the pile is “unplugged”, and qb0.1 / qc,avg = Ar
Open-Ended
if not, the pile is “plugged”, and
qb0.1 / qc,avg = maximum [ 0.5 - 0.25 × log( D / DCPT ), 0.15, Ar]
Closed-Ended qb0.1 / qc,tip= FDr = 0.8 / ( 1 + Dr2 )
qb0.1 = minimum [ plugged qb0.1, unplugged qb0.1 ]
the plugged qb0.1value is calculated as:
qb0.1 / qc,tip = FDr = 0.7 / ( 1 + 3Dr2 )
NGI-05 the unplugged qb0.1 is calculated as:
Open-Ended
qb0.1 = qb,ann × Ar + qb,plug × ( 1 - Ar ); qb,ann = qc,tip, and qb,plug = 12τf,avg × L /
Di
τf,avg: averaged external skin friction (Clausen et al. 2005); L: pile
embedment depth
Closed- &
UWA-05 qb0.1 / qc,avg = 0.15 + 0.45 × Arb,eff
Open-Ended
D: pile outer diameter; Di: pile inner diameter
pref = 100 kPa; DCPT = 0.036 m; Ar (area ratio) = 1 - ( Di / D ) 2 ;
Arb,eff (effective area ratio) = 1 - FFR × ( Di / D ) 2;
FFR: IFR (= ΔLp / Δz) averaged over last 3D of the pile penetration
qc,avg = qc averaged ±1.5D over pile tip level for Fugro-05 & ICP-05
Note
methods
qc,avg = qc averaged using Dutch averaging technique for UWA-05
method
Dr = 0.4 × ln ( qc1N / 22 ); as decimal, nominal relative density which may
be greater than 1.0 for NGI method
28
Table 5. Performance of design methods for closed ended concrete (CECC) and steel
piles (CECS) in compression
CECC CECS
μgR median σlnR μgR median σlnR
API-00 0.99 0.90 0.36 0.55 0.63 0.46
LCPC-82 1.20 1.23 0.27 0.95 0.84 0.33
EF-97 1.25 1.38 0.17 1.22 0.96 0.33
Fugro-05 1.16 1.24 0.33 1.14 1.02 0.33
ICP-05 0.94 0.93 0.24 0.88 0.85 0.35
NGI-05 0.96 0.99 0.30 1.10 0.89 0.33
UWA-05 0.88 0.89 0.24 0.93 0.84 0.33
Table 6. Performance of design methods for closed (CEC) and open (OEC) ended
piles in compression
CEC OEC
API-00 0.76 0.81 0.50 0.72 0.84 0.59
LCPC-82 1.08 1.12 0.31 1.44 1.41 0.21
EF-97 1.24 1.23 0.25 2.34 2.38 0.28
Fugro-05 1.16 1.16 0.32 1.13 1.18 0.31
ICP-05 0.92 0.89 0.29 0.87 0.98 0.29
NGI-05 1.02 0.94 0.32 1.00 1.02 0.25
UWA-05 0.90 0.85 0.28 0.96 0.95 0.19
Table 7. Performance of design methods for closed (CET) and open (OET) ended
piles in tension
CET OET
API-00 0.84 0.79 0.90 0.73 0.66 0.64
LCPC-82 1.33 1.44 0.65 1.26 1.23 0.46
EF-97 1.02 1.09 0.47 1.25 1.21 0.25
Fugro-05 0.78 0.89 0.38 0.92 0.85 0.32
ICP-05 0.88 0.84 0.31 0.96 0.92 0.16
NGI-05 0.92 0.90 0.42 1.04 0.97 0.29
UWA-05 0.85 0.89 0.29 0.97 0.98 0.19
Table 8. Performance of design methods for entire 77 pile database

Entire Database
μgR median σlnR
API-00 0.76 0.78 0.60
LCPC-82 1.23 1.27 0.40
EF-97 1.40 1.36 0.42
Fugro-05 1.04 1.11 0.35
ICP-05 0.91 0.92 0.27
NGI-05 1.01 0.96 0.31
UWA-05 0.92 0.90 0.25
29
50 50
45 45
40 40
35 35
Frequency (%)
Frequency (%)
30 30
25 25
20 20
15 15
10
10
5
5
0
0
<1
2
4
6
8
10
12
14
16
18
20
> 21
0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
Capacity at w tip = 0.1D (MN) Nominal Pile Diameter or Width (m)
50 50
45 45
40 40
35 35
Frequency (%)
Frequency (%)
30 30
25 25
20 20
15 15
10 10
5 5
0 0
5 15 25 35 45 55 65 75 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1
Pile Length (m) Weighted Average Dr along Shaft
50 50
45 45
40 40
35
Frequency (%)
35
Frequency (%)
30
30
25
20 25
15 20
10 15
5 10
0
5
5
15
25
35
45
55
65
75
85
95
> 100
0
Time between installation and load testing 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1
(days) Average Dr at Pile Tip
Figure 1. Frequency of occurrence of soil and pile parameters within analysed

database
30
1000
7
8&9
6
qc1N
100
CEC CET
4
OEC OET
10
0.1 1 10
Fr (%)
Zone Description Zone Description

3 Clays; silty clay to clay 6 Sands; clean sand to silty sand
4 Silt mixtures; clayey silt to silty clay 7 Gravelly sand to dense sand
5 Sand mixtures; silty sand to sandy 8&9 Very stiff overconsolidated /
silt cemented sands to fine grained soils
Figure 2. Soil behavior type classification (after Robertson & Wride 1998) for
average CPT parameters corresponding to 64 load tests within the database
31
10 10
API-00 CEC CET
OEC OET
Qc [API-00]/Qm
Qc [API-00]/Qm
1 1
CEC CET
API-00
OEC OET
0.1 0.1
10 100 1000 1 10 100
qc1N,avg along pile shaft tip depth, z tip (m)
Figure 3. Bias towards density (weighted qc1N) along pile shaft and effective pile
length for API-00
32
10 10
Drammen Padre EURIPIDES Ia
Island
Hound Point
Qc [LCPC-82]/Qm
Qc [EF-97]/Qm
1 1
CEC CET CEC CET
OEC OET LCPC-82 EF-97 OEC OET

0.1 0.1
1 10 100 1 10 100
qc,avg along pile shaft (MPa) qc,avg along pile shaft (MPa)
10 10
Fugro-05 ICP-05
Lock & Dam 26 Lock & Dam 26
Qc [Fugro-05]/Qm
Qc [ICP-05]/Qm
1 1
CEC CET CEC CET
OEC OET OEC OET

0.1 0.1
1 10 100 1 10 100
10 10
NGI-05 Lock & Dam 26 UWA-05
Lock & Dam 26
Qc [UWA-05]/Qm
Qc [NGI-05]/Qm
1 1
CEC CET CEC CET
OEC OET OEC OET

0.1 0.1
1 10 100 1 10 100
Figure 4. Potential bias towards weighted average qc along pile shaft for CPT
methods
33
Appendix
Tables A1 through A5 presents summary information for load tests in the database discussed
in this paper. A short description of each column is listed below.
Column Heading Description
1 ID ID number for UWA database
2 Site Site name based on descriptions in reference
3 Pile No. Pile identification
4 Pile Material from which pile was constructed
Material
5 Pile Shape Exterior shape of pile
6 B or D (m) Outer width (B) of square or octagonal piles or diameter (D) of
circular piles
7 t (mm) Wall thickness of open ended piles
8a ztip Tip depth of pile. Only one pile (ID 116 at Hunters Point) had an
[Leff] embedded length less that tip depth, with that value contained in a
(m) footnote below the table.
8b [Leff] Of interest is the “effective length” of the pile, which is the length
(m) behind the tip in which 80% of the pile shaft friction is estimated
to be generated using the UWA-05 method. The effective length,
Leff, is contained in brackets [] below the tip depth. Effective
length using the distribution of ICP-05 or NGI-05 is similar.
9 Water Depth to water table at time of driving
Table
(m)
10 Time Time between installation and load testing
(days)
11 Avg qc1N Weighted average normalized cone tip resistance, based on UWA-
shaft 05 shaft friction formulation
12 Avg qc1N Average normalized cone tip resistance at pile tip, using Dutch
tip averaging technique
13 Avg Fr (%) Weighted average CPT friction ratio, based on UWA-05 shaft
friction formulation
14 Dbase Qt Pile capacity at a tip displacement of 0.1D, used in database
(MN) assessment of Qc/Qm
15a Max Qt Maximum load measured in pile load test
(MN)
15b [w (mm)] Maximum displacement measured in pile load test. This value is
contained in [brackets] below the maximum load value.
16 Clay Qs Estimated value of shaft friction occurring in clay layers
(MN)
17 Avg IFR Weighted average incremental filling ratio (IFR) of open ended
pile, based on UWA-05 shaft friction formulation. The IFR was
estimated from (a) continuous field measurements; (b) plug length
ratio measured at the end of driving; (c) as a function of pile inner
diameter after Lehane et al (2005a).
18 No. Piles Number of “similar” piles at same site. Qc/Qm for piles at a site
with this value greater than one were averaged prior to assessment
of database statistics to minimize the potential for bias towards a
given site.
19 Reference Reference which contains information describing pile load test.
34
Table A1. Characteristics of load tests on concrete closed ended piles in compression (CECC)
ID Site Pile Pile Pile B or t ztip Water Time Avg Avg Avg Dbase Max Clay Avg No. Reference
No. Material Shape D (mm) [Leff] Table (days) qc1N qc1N Fr Qt Qt (MN) Qs IFR Piles
(m) (m) (m) shaft tip (%) (MN) [w (mm)] (MN)
101 Baghdad p1 Concrete Square 0.253 - 11 6.2 88 60 38 2.63 0.98 1.1 - Closed 1 Altaee et al.
[6.8] [110] 1992
102 Baghdad p2 Concrete Square 0.253 - 15 6.0 42 54 49 2.60 1.61 1.61 - Closed 1 Altaee et al.
[9.4] [29] 1992
104 Cimarron River p2 Concrete Octagonal 0.610 - 19.5 1.0 ? 111 37 0.42 3.56 3.56 - Closed 1 Nevels &
[13.1] [65] Snethen 1994
105 Drammen A Concrete Circular 0.280 - 8 1.7 ? 48 25 0.39 0.28 0.29 - Closed 1 Gregersen et
[4.9] [39] al. 1973
106 Drammen D/A Concrete Circular 0.280 - 16 1.7 ? 39 32 0.40 0.49 0.50 - Closed 1 Gregersen et
[11] [41] al. 1973
107 Drammen E Concrete Circular 0.280 - 7.5 1.7 ? 50 21 0.40 0.21 0.21 - Closed 1 Gregersen et
[4.5] [32] al. 1973
[7.6] [30] al. 1973
[10.7] [35] al. 1973
[12.2] [32] al. 1973
[12.9] [40] al. 1973
112 Fittja Straits D Concrete Square 0.235 - 12.8 2.0 5 42 39 0.44 0.34 0.36 0.03 Closed 1 Axelsson
[8.2] [35] 2000
117 Jonkoping p23 Concrete Square 0.235 - 16.8 1.3 >1 93 112 0.44 1.72 1.72 0.06 Closed 3 Jendeby et al.
[10.9] [50] 1994
118 Jonkoping p25 Concrete Square 0.235 - 17.8 1.3 <1 114 144 0.46 1.65 1.50 0.06 Closed 3 Jendeby et al.
[10.9] [14] 1994
119 Jonkoping p26 Concrete Square 0.275 - 16.2 1.3 >1 91 78 0.44 1.40 1.40 0.07 Closed 3 Jendeby et al.
[10.5] [50] 1994
127 Ogeechee River H-2 Concrete Square 0.406 - 15.2 1.5 0.5 130 112 0.25 2.75 3.16 - Closed 1 Vesic
[8.2] [133] 1970
130 Tickfaw River TP2 Concrete Square 0.610 - 32 1.0 ? 41 24 0.77 2.96 2.60 1.04 Closed 1 Titi & Abu-
[26.1] [19] Farsakh 1999
131 Tickfaw River TP1 Concrete Square 0.610 - 25.9 1.0 ? 43 11 0.77 2.47 2.35 0.53 Closed 1 Titi & Abu-
[20.3] [25] Farsakh 1999
35
Table A2. Characteristics of load tests on steel closed ended piles in compression (CECS)
ID Site Pile Pile Pile D t ztip Water Time Avg Avg Avg Dbase Max Clay Avg No. Reference
No. Material Shape (m) (mm) [Leff] Table (days) qc1N qc1N Fr Qt Qt (MN) Qs IFR Piles
(m) (m) shaft tip (%) (MN) [w (mm)] (MN)
100 Akasaka 6C Steel Circular 0.200 - 11 9.0 ? 168 189 - 1.21 ~1.5 - Closed 1 BCP Com.
[4.9] [1000] 1971
103 Cimarron River p1 Steel Circular 0.660 - 19 1.0 ? 112 39 0.43 3.57 3.58 - Closed 1 Nevels &
[12.9] [80] Snethen 1994
113 Hoogzand II Steel Circular 0.356 - 6.8 3.2 ? 430 388 0.86 2.85 3.10 - Closed 1 Beringen et al.
[2.6] [64] 1979
114 Hsin Ta TP4 Steel Circular 0.609 - 34.3 2.0 33 62 36 - 4.26 4.26 0.72 Closed 2 Yen et al.
[20.2] [78] 1989
115 Hsin Ta TP6 Steel Circular 0.609 - 34.3 2.0 30 74 28 - 4.91 4.40 0.61 Closed 2 Yen et al.
[23.9] [21] 1989
116 Hunter's Point S Steel Circular 0.273 - 9.2* 2.4 24 88 63 0.27 0.44 0.50 - Closed 1 Briaud et al.
[6.1] [83] 1989a
120 Ogeechee River H-12 Steel Circular 0.457 - 6.1 1.5 0.5 155 111 0.38 2.08 2.14 - Closed 1 Vesic
[2.6] [130] 1970
121 Ogeechee River H-13 Steel Circular 0.457 - 8.9 1.5 0.5 144 118 0.33 2.64 2.81 - Closed 1 Vesic
[4.5] [132] 1970
122 Ogeechee River H-14 Steel Circular 0.457 - 12 1.5 0.5 134 83 0.28 3.21 3.56 - Closed 1 Vesic
[6.7] [131] 1970
123 Ogeechee River H-15 Steel Circular 0.457 - 15 1.5 0.5 131 102 0.25 3.95 3.83 - Closed 1 Vesic
[8.1] [61] 1970
124 Pigeon Creek 1 Steel Circular 0.356 - 6.9 3.0 4 203 190 0.55 1.50 1.77 0.01 Closed 1 Paik et al.
[3.0] [150] 2003
125 Sermide S Steel Circular 0.508 - 35.9 0.0 ? 91 81 - 5.49 5.62 0.29 Closed 1 Appendino
[19.6] [84] 1981
126 Lock&Dam 26 3-1 Steel Circular 0.305 - 14.2 0.0 35 235 167 0.38 1.17 1.32 - Closed 3 Briaud et al.
[7.8] [76] 1989b
[7.9] [33] 1989b
[8.0] [76] 1989b
* embedded length for Hunter’s Point in 7.8m
36
Table A3. Characteristics of load tests on closed ended piles in tension (CET)
200 Baghdad p1 Concrete Square 0.253 - 11 6.2 200 62 - 2.63 0.58 0.58 - Closed 1 Altaee et al.
[6.8] [65] 1992
201 Drammen A Concrete Circular 0.280 - 8 1.7 ? 49 - 0.39 0.09 0.09 - Closed 1 Gregersen et
[4.9] [18] al. 1973
202 Drammen D/A Concrete Circular 0.280 - 16 1.7 ? 38 - 0.40 0.25 0.25 - Closed 1 Gregersen et
[11] [37] al. 1973
203 Drammen E Concrete Circular 0.280 - 23.5 1.7 ? 49 - 0.41 0.29 0.29 - Closed 1 Gregersen et
[12.9] [37] al. 1973
204 Hoogzand II Steel Circular 0.356 - 6.8 3.2 ? 430 - 0.86 1.21 1.21 - Closed 1 Beringen et al.
[2.6] [57] 1979
205 Hsin Ta TP5 Steel Circular 0.609 - 34.3 2.0 28 54 - - 2.63 2.45 0.18 Closed 1 Yen et al.
[17.5] [21] 1989
206 Ogeechee River H-16 Steel Circular 0.457 - 15 1.5 1.5 131 - 0.25 1.54 1.54 - Closed 1 Briaud et al.
[8.1] [10] 1989b
207 Lock&Dam 26 3-2 Steel Circular 0.305 - 11 0.0 35 221 - 0.39 0.54 0.54 - Closed 3 Briaud et al.
[6.0] [62] 1989b
208 Lock&Dam 26 3-5 Steel Circular 0.356 - 11.1 0.0 27 222 - 0.40 0.61 0.61 - Closed 3 Briaud et al.
[6.1] [43] 1989b
209 Lock&Dam 26 3-8 Steel Circular 0.406 - 11.1 0.0 28 222 - 0.40 0.90 0.90 - Closed 3 Vesic
[6.1] [60] 1970
210 I-880 2-T Steel Circular 0.610 - 10.7 0.0 16 455 - 1.99 2.00 2.00 0.21 Closed 2 Olson &
[4.8] [32] Shantz 2004
211 I-880 2-W Steel Circular 0.610 - 12.3 0.0 20 459 - 2.15 3.20 3.20 0.29 Closed 2 Olson &
[5.8] [38] Shantz 2004
37
Table A4. Characteristics of load tests on steel open ended pipe piles in compression (OEC)
300 SFOBB Bent Steel Circular 0.610 12.5 13.3 0.0 25 318 96 1.53 2.87 2.71 0.66 0.83c 1 Olson &
E31R [5.2] [28] Shantz 2004
301 Drammen 16 Steel Circular 0.813 12.5 11 3.0 2 58 29 0.58 1.21 1.60 - 0.88c 1 Tveldt &
[6.3] [204] Fredriksen
2003
[7.0] [NA] Fredriksen
2003
[12.0] [NA] Fredriksen
2003
304 Dunkirk zdh C1 Steel Circular 0.457 13.5 10 4.0 68 299 244 0.85 2.94 2.82 - 0.78c 1 Jardine &
[6.6] [34] Standing 2000
305 EURIPIDES Ia Steel Circular 0.763 35.6 30.5 1.0 7 246 307 1.22 8.25 11.60 - 0.99a 1 Kolk et al.
[7.3] [260] 2005b
306 EURIPIDES Ib Steel Circular 0.763 35.6 38.7 1.0 2 293 231 1.43 13.00 16.26 - 0.97a 1 Kolk et al.
[8.8] [249] 2005b
307 EURIPIDES Ic Steel Circular 0.763 35.6 47 1.0 11 284 230 1.32 19.50 23.41 - 0.96a 2 Kolk et al.
[15.2] [260] 2005b
308 EURIPIDES II Steel Circular 0.763 35.6 46.7 1.0 6 296 230 1.54 19.50 21.53 - 0.95a 2 Kolk et al.
[14.6] [190] 2005b
309 Hoogzand I Steel Circular 0.356 16.0 7 3.2 37 425 385 0.86 2.27 2.50 - 0.66b 1 Beringen et al.
[2.8] [64] 1979
310 Hoogzand III Steel Circular 0.356 20.0 5.3 3.2 19 420 428 0.90 1.85 2.00 - 0.77b 1 Beringen et al.
[1.7] [64] 1979
311 Hound Point p Steel Circular 1.220 24.2 26 0.0 21 144 138 - 7.00 7.50 0.44 0.95c 1 Williams et al.
[7.5] [215] 1997
313 Pigeon Creek 2 Steel Circular 0.356 32.0 7 3.0 4 202 192 0.56 1.03 1.28 0.02 0.83a 1 Paik et al.
[3.2] [135] 2003
314 Shanghai ST-1 Steel Circular 0.914 20.0 79 0.5 23 95 81 - 15.56 16.36 1.72 0.80b 2 Pump et al.
[45.5] [121] 1998
315 Shanghai ST-2 Steel Circular 0.914 20.0 79.1 0.5 35 95 82 - 17.08 17.82 1.74 0.85b 2 Pump et al.
[46.9] [130] 1998
316 Trans-Tokyo TP Steel Circular 2.000 34.0 30.6 0.0 52 296 58 - 34.68 34.68 1.32 1.00b 1 Shioi et al.
Bay [17.4] [203] 1992
317 SEUS MT Steel Circular 0.324 38.1 42.4 3.7 21 123 86 1.29 3.58 3.72 0.32 0.50b 1 -
[25.2] [91]
38
Table A5. Characteristics of load tests on steel open ended pipe piles in tension (OET)
400 Los Coyotes 5 Steel Circular 0.356 11.2 14.9 5.0 2 146 - 2.15 1.51 1.51 0.64 0.74c 1 Olson &
[8.9] [26] Shantz 2004
401 SFOBB Bent Steel Circular 0.610 12.5 13.3 0.0 25 320 - 1.55 1.34 1.34 0.66 0.83c 1 Olson &
E31R [5.3] [34] Shantz 2004
402 Dunkirk zdh C1 Steel Circular 0.457 13.5 10 4.0 69 299 - 0.85 0.82 0.82 - 0.78c 1 Jardine &
[6.6] [46] Standing 2000
403 Dunkirk zdh R1 Steel Circular 0.457 13.5 19.3 4.0 9 234 - 0.99 1.45 1.45 - 0.78c 1 Jardine &
[12.6] [31] Standing 2000
404 EURIPIDES Ia Steel Circular 0.763 35.6 30.5 1.0 7 246 - 1.22 3.00 3.00 - 0.99a 1 Kolk et al.
[7.3] [76] 2005b
405 EURIPIDES Ib Steel Circular 0.763 35.6 38.7 1.0 2 293 - 1.43 9.75 9.75 - 0.97a 1 Kolk et al.
[8.8] [36] 2005b
406 EURIPIDES Ic Steel Circular 0.763 35.6 47 1.0 11 284 - 1.32 13.75 13.75 - 0.96a 2 Kolk et al.
[15.2] [72] 2005b
407 EURIPIDES II Steel Circular 0.763 35.6 46.7 1.0 7 296 - 1.54 11.00 11.00 - 0.95a 2 Kolk et al.
[14.6] [76] 2005b
408 Hoogzand I Steel Circular 0.356 16.0 7 3.2 37 425 - 0.86 0.82 0.82 - 0.66b 1 Beringen et al.
[2.8] [20] 1979
409 Hoogzand III Steel Circular 0.356 20.0 5.3 3.2 19 420 - 0.90 0.53 0.53 - 0.77b 1 Beringen et al.
[1.7] [10] 1979
410 Hound Point p Steel Circular 1.220 24.2 34 0.0 11 172 - - 3.86 3.86 0.41 0.95c 1 Williams et al.
[13] [25] 1997
411 Hound Point p Steel Circular 1.220 24.2 41 0.0 4 140 - - 3.74 3.74 0.34 0.95c 1 Williams et al.
[18.7] [NA] 1997
412 I-880 2-P Steel Circular 0.610 19.1 12.3 0.0 28 461 - 2.17 2.00 2.00 0.40 0.82c 1 Olson &
[5.8] [43] Shantz 2004
413 SEUS MT Steel Circular 0.324 38.1 42.4 3.7 28 130 - 1.37 2.44 2.44 0.32 0.50b 1 -
[25.2] [62]
414 Padre Island A Steel Circular 0.508 12.7 14.6 1.0 2 127 - - 0.48 0.48 - 0.87b 1 McClelland
[9.9] [37] 1974
415 Padre Island A Steel Circular 0.508 12.7 17.1 1.0 2 115 - - 0.65 0.65 - 0.80a,b 1 McClelland
[11.9] [NA] 1974
39
Database assessment of CPT based design methods for axial capacity of driven piles
in siliceous sands: Part II – application to reliability analyses for offshore platforms
by
James A. Schneider1 Barry M. Lehane2 and X. Xu3
Prepared for ASCE Journal of Geotechnical and Geoenvironmental Engineering
1
Corresponding author
2
Professor, School of Civil and Resource Engineering, The University of
3
ABSTRACT
This paper presents the results of analyses that quantify the relative reliability of a
range of methods, including new recommendations in the American Petroleum
Institute RP2A (2006), for evaluation of the axial capacity of a driven pile in siliceous
sand; the study employs the load test database described in a companion paper
(Schneider et al. 2007). It is shown that a new method included in the commentary of
API (2006), referred to as UWA-05, has the highest relative reliability index of the
methods considered and that it provides the safest extrapolation of capacity from the
database (comprising piles with a mean diameter of about 450mm) to large diameter
pipe piles commonly used offshore. The paper also shows that the low incidence of
failure of offshore foundation systems designed using previous API recommendations
is not associated with low levels of uncertainty or conservatism in recommendations
for single pile axial capacity and can only be explained because of the exclusion of
relevant factors from reliability assessments. It is concluded that the reliable and
efficient design of offshore piled foundations requires improvements in our
understanding of mechanisms influencing foundation system capacity.
2
INTRODUCTION
Reliability analyses are now being performed more routinely in geotechnical
engineering (e.g. Christian 2004, Hicks 2005), and have been used to aid the
assessment of safety factors for piled foundations and to assist the re-certification of
fixed offshore platforms (Tang et al. 1990, Lacasse & Goulois 1989, Bea et al. 1999,
Paikowsky et al. 2004, Jardine et al. 2005). Input parameters for these analyses
comprise the mean and standard deviation of the pile resistance for particular loading
conditions. Values for the mean and standard deviation of the foundation resistance
may be derived from database studies that summarise the performance of a given
predictive method against a database of measurements using the ratio of calculated
(or predicted) capacity to the measured capacity (Qc/Qm); a companion paper
(Schneider et al. 2007) presents such an assessment for the axial capacity of driven
piles in siliceous sands with adjacent cone penetration test (CPT) profiles. The results
of that study allow designers to assess the relative merits of four new CPT based
design methods, which have been introduced to the commentary of the 22nd edition of
the American Petroleum Institute (API) Recommended Practice for Fixed Offshore
Structures (RP2A) (API 2006). The four ‘offshore’ CPT driven pile design methods
for siliceous sands are referred to as Fugro-05 (Kolk et al. 2005a), ICP-05 (Jardine et
al. 2005), NGI-05 (Claussen et al. 2005), and UWA-05 (Lehane et al. 2005a).
Schneider et al. (2007) present a discussion of the formulations for each of those
methods and compare these with the API (2000, 2006) main text methods (API-00,
API-06), and two CPT ‘alpha’ methods, LCPC-82 (Bustamante & Gianeselli 1982)
and EF-97 (Eslami & Fellenius 1997).
A reasonable design method should result in consistently acceptable levels of

reliability for an expected range of conditions, or a low occurrence of failure. Low
occurrence of failure may result from two primary sources (Lacasse & Nadim 1994):
(i) low uncertainty in a method through an accurate assessment of all
physical mechanisms influencing a design problem; or
(ii) high levels of conservatism in the design method or in application of the
design method.
3
Uncertainty in the estimation of resistance (σln,R) can be further broken down into
uncertainty in the model (σln,Rm), uncertainty in the input parameters to that model
(σln,Ri) as well as uncertainty due to small sample size (σln,Rs), which Wu et al. (1989)
express as:
σ ln,R = σ ln, Rm + σ ln,Ri + σ ln,Rs

2 2 2
(1)
The effects of sample size are not considered further here as the primary focus is to
quantify relative levels of uncertainty/reliability for different design methods.
Whether uncertainty exists within a predictive model or within the input parameters
for that model affects how an engineer approaches an investigation for a specific site.
For example, improving the degree of knowledge related to the input parameters of a
given design method is of little use if the predictions associated that method have a
high level of uncertainty. Additionally, design method development cannot be
established reliably if the vertical distribution of soil type and strength is poorly
quantified (Dennis & Olson 1983). This paper explores whether the observed high
level of reliability for offshore piled foundations is related to (i) low levels of
uncertainty, (ii) conservative bias in design method formulation, or (iii) time
dependent foundation system reliability. The potential for extrapolation bias due to
differences in characteristics of database piles and long, large diameter open ended
offshore piles is also examined.
Statistics relating the ratio of calculated to measured capacity (Qc/Qm) are provided in
the companion paper (Schneider et al. 2007) for the seven design methods
considered. However, these only provide an indication of actual reliability if they are
adjusted to account for uncertainty due to database size and characteristics (Wu et al.
1989, Zhang et al. 2004). Offshore pile design loads in compression are typically an
order of magnitude larger than capacities indicated by the existing database of load
tests on onshore piles and the capacity extrapolated from these smaller piles depends
significantly on the particular design method employed. Figure 1 provides an
illustration of this sensitivity by presenting predictions for the capacity of a 2.4m
diameter open ended pile in a Gulf of Mexico sand profile (Lehane et al. 2005b).
Changes in pile capacity based on CPT qc methods tend to mirror the CPT qc profile
4
in Figure 1, although, the relative magnitude of variations in capacity induced by soil
layering is method specific. For a compression load of 70 MN and tension load of 30
MN, it is evident that required pile tip depths vary by up to 15m (≡25% of the total
pile length), depending on the designer’s preferred prediction method. The choice of
method can be guided by assessment of each method’s predictive performance for a
database of load tests on single piles. However, the reliability of a given method
requires consideration of the ability of a given method to encapsulate the most
important factors affecting axial pile capacity.
RELATIVE RELIABILTY OF EXISTING DESIGN METHODS

The probability of a foundation collapse and its associated costs are major concerns
for owners and operators. To quantify foundation reliability for offshore structures, it
is necessary to assess the uncertainty and bias in (i) loads, (ii) transmission of loads
via the structure to the foundation elements and (iii) foundation capacity. The
database assessment of CPT based axial pile design methods described in Schneider
et al. (2007) only provides information related to foundation capacity and therefore
the reliability values derived in the following using this database should be regarded
as measures of relative (or nominal) reliability related to the axial capacity of piles -
and not actual reliability. Employing a database of static load test data to determine
relative levels of reliability is justified as the axial capacity of offshore piles is
determined using static design methods (McClelland 1974, Randolph et al. 2005)
even though the loads on offshore foundations are predominantly cyclic (McClelland
1974). Static capacity and effects of cyclic loading are addressed separately for
offshore piles (API 2000, 2006).
The recommendations in the first edition of API RP2A (1969) considered that for pile
penetrations greater than 30m (the approximate depth at which limiting shaft friction
values were applied in sand), a factor of safety of as low as 2 could be used for the
combination of operating loads with frequent storm events, and a factor of safety of
as low as 1.5 could be employed for maximum design loads including the effects of
extreme storm events (McClelland 1969). If “below average” site investigation
information was available, minimum factors of safety were raised to 3 and 2 for
5
operating and extreme environmental conditions respectively (Pelletier et al. 1993).
API (2006) currently maintains minimum recommended factors of safety between 1.5
and 2, and suggests that higher factors of safety be considered when designing piles
in siliceous sands using CPT based methods.
Factors of safety can be related to the nominal probability of failure through the
reliability index, β (Christian 2004).
p f = 1 − Φ (β ) (2)
where Φ is the cumulative distribution function (CDF) of the standard normal

distribution. For this study, load and resistance are assumed to be uncorrelated. The
correlation between load and resistance (ρQR) implies designer experience, i.e., if
designers are faced with higher loads for a similar design situation, they will select a
solution with a higher resistance (Harr 1987). Designer experience is highly variable
and ρQR and should be taken as zero without additional documentation. For
uncorrelated variables, the reliability index (β) can be calculated for a lognormal
distribution as (Christian 2004):
⎛R⎞
ln⎜ ⎟
β= ⎝Q⎠ (3)
σ 2 ln R + σ 2 ln Q
where σlnR is the standard deviation of the natural log of the resistance (R), and σlnQ is
the standard deviation of the natural log of the load (Q). The standard deviation of the
natural log of a variable is essentially equal to the coefficient of variation (COV) for
COV less than 0.5, and those two parameters are related as:
[
COV = exp(σ 2 ln ) − 1 ]
0.5
(4)
6
For quantification of uncertainty in pile design methods, σlnR is often characterized
using a database of pile load test results, with uncertainty in design loading conditions
typically taken as a nominal value inferred from separate studies. When using the
lognormal distribution, the mean (λ) and standard deviation (σlnR) are defined using
the natural logarithm of the variables (Christian 2004), i.e., the ratio of calculated to
measured capacity (Qc/Qm) for load and resistance. The geometric mean of the load
and resistance are equal to exp(λ), and identified as μgQ and μgR in this paper and the
companion paper (Schneider et al. 2007). The reliability index can be expressed as a
function of the applied factor of safety (FSA), the bias (μg), and the standard deviation
(σln):
⎛ μ gQ ⎞
ln⎜ FS A ⎟
⎜ μ ⎟
β= ⎝
gR ⎠
(5)
σ 2 ln R + σ 2 ln Q
The discussion presented here refers to a nominal reliability index since many factors
influencing uncertainty (σlnR and σlnQ) within the design process are ignored. For
nominal design loading conditions, σlnQ is taken here as 0.15 (Lacasse & Nadim
1994) and μgQ is assumed to be unity1. Considering unbiased extrapolation, a nominal
probability of failure is a lower bound of the actual system performance (Christian
2004) and factors of safety based on nominal probability of failure (or nominal β)
values may be unconservative. However, Bea et al. (1999) suggest that other factors
influencing foundation system design may result in higher system reliabilities than
implied by the nominal β value. Therefore while this index is useful for assessment of
relative performance of design methods, it cannot be used for statistical quantification
of actual factors of safety or reliability without inclusion of all significant features
influencing uncertainty and bias.
1
Noting that σlnQ varies with the proportion of live and dead loads.
7
Method performance is quantified using the geometric mean (μgR) and standard
deviation of the natural log (σlnR) of the ratio of calculated to measured pile capacity
(Qc/Qm). Figure 2 summarizes these parameters for the database and methods
discussed in Schneider et al. (2007), and also plots the nominal reliability indices (β)
inferred from these parameters for a factor of safety of 2 (API 2006). Statistics are
presented for the 77 piles in the full database and for subset databases comprising
closed-ended piles loaded in compression and tension (CEC & CET) and open-ended
piles loaded in compression and tension (OEC & OET). The same conclusions are
drawn when using the sample geometric mean or median for calculations in Figure 2.
It is apparent that the values of μgR of the LCPC-82, EF-97 and Fugro-05 methods
vary significantly between subset databases i.e. these methods do not provide the
same level of under-prediction or over-prediction for all pile types. Such bias points
to inadequacies in these design approaches (e.g. see Schneider et al. 2007 and Xu et
al. 2007) and to the high level of uncertainty associated with extrapolation to
conditions outside of those relevant to the (relatively small) subset databases.
Inconsistent predictive performance is also evident on examination of the range of
standard deviations (σlnR) for a given method e.g. API-00 and LCPC-82 indicate very
wide differences between the respective σlnR values for various pile types.
The nominal reliability indices plotted on Figure 2 show that ‘offshore’ CPT based
pile design methods perform better than API-00, LCPC-82 and EF-97, with β in the
range of 2 to 3.5 for statistics from the full 77 pile database. While these indices are
generally consistent with target reliability values of 2.6 to 3.2 for serviceability and
ultimate limit state conditions respectively (Phoon & Kulhawy 2002), the actual
prediction uncertainty is higher (and hence β is lower) due, for example, to the small
database sizes and the extent of the extrapolation required. The relatively low levels
of reliability for API-00 against a database of onshore load tests implied by Figure 2,
despite the apparent low probability of failure observed for offshore foundations, is
addressed through parametric studies of method formulation discussed later.
8
Of greater concern when extrapolating to predict the capacity of large offshore driven
piles is the potential for bias in method formulation, which may not be apparent in a
database comprising much smaller piles. The strong effect of bias on the nominal
reliability index is illustrated in Figure 3 by plotting the β variation with the ratio of
μgQ to μgR for the ‘offshore’ CPT based methods using the σlnR value for the entire 77
pile database, σlnQ of 0.15 (Lacasse & Nadim 1994) and FSA of 2. Evidently, un-
conservative extrapolation will significantly reduce the nominal β values when using
a relatively low FSA of 2, and some conservative bias or larger factors of safety are
needed to reach acceptable target levels of reliability.
QUANTIFYING EXTRAPOLATION UNCERTAINTY

Experimental observations made over the past 20 years have highlighted many of the
important factors influencing axial pile capacity in siliceous sands and may be used to
reduce the level of uncertainty associated with extrapolation using current design
methods. A summary of these factors, and if or how they are incorporated in each
design method, is presented in Table 1; it is acknowledged that this table does not
include all variables influencing pile capacity in sand2. The full formulations for each
method are provided in Schneider et al. (2007).
Offshore CPT qc-based methods

As seen in Table 1, the UWA-05 method accounts for a greater number of factors
known to affect the base and shaft capacity of driven piles in sand. Based on a series
of recent peer-reviewed publications in this area, this method assumes a diameter
dependence of both the degree of partial plugging during pile installation (as
described by the incremental filling ratio, IFR) and the increase in lateral stress
(Δσ'rd) during pile loading. The IFR and the value of ∆σ'rd are significant when
predicting the capacity of smaller diameter piles in the database, but typical large
diameter offshore piles core during installation (i.e. IFR ≈ 1) and ∆σ'rd may be
expected to be negligible (Lehane & Jardine 1994). The effects of assuming IFR=1
2
There are many aspects in need of further study including (i) differences in friction mobilization between tension
and compression piles, (ii) ageing and cyclic degradation for large diameter piles in a variety of conditions and
(iii) pile performance in materials such as calcareous sands, micaceous sands, residual silty sands, aged sand
deposits, and sandy silts.
9
and Δσ'rd=0 are examined for the UWA-05 method in Table 2, which shows that the
‘offshore simplification’ (i.e. IFR=1 and ∆σ'rd=0, Lehane et al. 2005a) leads to
predicted capacities that are 10% to 25% more conservative than the measured
capacities. This table does not indicate that this simplification of UWA-05 is
conservative for all pile geometries, but it does imply that the other methods which
do not account for Δσ'rd or IFR are potentially un-conservative when extrapolated to
offshore conditions.
The variations with pile diameter of calculated to measured capacity ratios (Qc/Qm)
for database piles using the CPT-based ‘offshore’ methods are plotted on Figure 4.
There is no clear tendency for any of the methods to over/under-predict capacities for
the range of diameters in the database. For UWA-05, the capacity ratios (Qc/Qm)
derived using estimates of IFR and Δσ'rd are compared with those calculated
assuming IFR=1 and Δσ'rd=0. The evaluation assuming IFR=1 and Δσ'rd =0 clearly
has a lower mean, but no obvious bias with pile diameter is evident despite the fact
that two parameters which vary with pile diameter were removed for the method’s
formulation. This tendency highlights the difficulty in calibrating a design method
involving a relatively large number of parameters using a small number of pile test
results, each of which has an associated level of uncertainty e.g. only four of the 77
piles in the database have a diameter exceeding 1m.
The diameter dependence of capacities predicted by ‘offshore’ CPT methods is

compared on Figure 5 by plotting the ratio of the predicted capacity to that given by
UWA-05 for open-ended piles with the CPT qc profile shown in Figure 1 and a pile
tip depth of 40m, 50m, and 60m. Despite each method having a broadly similar
nominal reliability index when assessed against the existing database (see Figure 2),
it is evident that there are very considerable differences between the methods. The
choice of an appropriate pile diameter for a particular application requires a judgment
to be made on the ability of a given design method to capture the physical
mechanisms governing the diameter dependence of pile capacity.
10
There is an obvious and large diameter effect when comparing UWA-05 and Fugro-
05 at a fixed pile length. The difference is more significant in tension than
compression because of the higher friction fatigue exponent (h/D-0.85 and h/D-0.9)
proposed Fugro-05 relative to that presumed by UWA-05 (h/D-0.5); see Schneider et
al. (2007).
Although assumptions relating to radial stress change during pile loading (Δσ'rd) are
similar for both ICP-05 and UWA-05, the UWA-05 method accounts explicitly for
the influence of partial plugging during installation of open ended piles. It is seen that
both methods provide broadly similar capacities within the pile diameter range typical
of the database. However, as the diameter increases beyond this range (i.e. D >0.8m),
the UWA-05 predictions for tension capacity fall below those of ICP-05 and are
about 70% of the ICP-05 predictions for piles with D>1.5m. Based on Figure 3, this
bias implies that the nominal relability index, β, of UWA-05 for large diameter
offshore piles in tension is higher than that of ICP-05 by a value of about 1. The two
methods have similar capacities in compression due to the conservative assumption
of ICP-05 for end bearing (Xu et al. 2007). The ‘jump’ in the ratio of QICP-05 to QUWA-
05 at D=1.5m arises because of the ICP-05 presumed sharp transition between a pile
failing statically in a coring as opposed to plugged mode. This discontinuity does not
arise for UWA-05, for which failure in a coring mode is not considered possible if the
plug length is greater than 5D.
Considering the influence of IFR and Δσ'rd quantified within UWA-05, it would be
expected that the ratio of the NGI-05 to UWA-05 capacity would increase with
diameter for a given pile length. However, Figure 5 shows an opposite trend because
the friction fatigue formulation for NGI-05 is based on z/L rather than (L-z)/D =h/D,
and consequently the average shaft friction predicted by NGI-05 for a fixed pile
length does not vary with the pile diameter. While a reasonable fit to database piles
can be achieved within the NGI-05 framework, assumptions related to method
formulation are often inconsistent with observations from experimental studies and
may lead to poor performance for piles outside of the NGI database (i.e., Hound Point
site in Figure 4).
11
The differences between the various methods’ pile capacity predictions evidently
result primarily from bias in method formulation. This bias is of critical importance to
the reliability of predictions for the capacity of large offshore piles and is more
significant than that inferred from statistical analysis of the existing database of
(smaller) piles. The exclusion or inappropriate treatment of factors known to
contribute to pile diameter effects contribute to bias and are likely to lead to incorrect
(and potentially un-conservative) engineering decisions.
API main text recommendations

The API-00/API-06 recommendations do not incorporate many of the factors listed in
Table 1 and hence it is not surprising that the method’s predictive performance for the
database piles is poor e.g. the nominal reliability index plotted on Figure 2 is on
average only 50% of the average index derived for UWA-05. However, given the
virtual absence of pile failure offshore, it may be inferred that the actual reliability is
relatively high (and that the method is potentially over-conservative in many
instances).
It appears that the higher than anticipated level of reliability of the API-00
recommendations results from compensating errors; these errors lead to trends of
average shaft friction (τavg) for piles with a length greater than 30m in dense sands
that are broadly similar to the ‘offshore’ CPT methods, each of which contain a
friction fatigue term. The variation with pile length of τavg values predicted by API-06
and UWA-05 are compared in Figures 6 and 7 for 1.2m diameter pipe piles in
uniform medium dense sand and dense silica sand (with respective normalised CPT
end resistance values, qc1N, of 60 and 180); the parameters βs=τavg/σ'v0,avg and
αs=qc,avg/τavg are used for this comparison. It is noted that local shaft friction at a
given depth varies significantly for these two methods, which will result in
inconsistent relative performance in layered soil deposits.
Figures 6 and 7 indicate that:
12
(i) For pile lengths greater than 30m, both API-00 and UWA-05 predict similar
rates of change of αs and βs with pile length. However, UWA-05 predicts
greater βs values in dense sand than in medium dense sand but its predicted αs
values are relatively independent of sand density. In contrast, API-00 predicts
that βs is essentially constant but that αs is strongly density dependant.
Comparison of the LCPC-82 and EF-97 methods by Schneider et al. (2007)
support the use of a constant αs (EF-97) over αs varying with density (LCPC-
82) for piles of similar geometry but differing soil density.
(ii) For shorter piles (typical of the database), the differences between API-00 and
UWA-05 predictions for βs and αs are more significant, with API-00 showing
constant βs values and αs reducing with pile length (to L=25m for the
example illustrated on these figures).
These differences in uniform sand profiles illustrate a density and length bias for API-
00/API-06 that is consistent with the method’s predictive performance for database
piles (Schneider et al. 2007). Consequently, the relatively low value of μgR for API-00
calculated for the database piles (typically ≈0.7; see Figure 2) cannot be equated to
prediction conservatism for long and large diameter offshore piles, and it is inferred
that actual reliability of this method arises because of other factors, such as:
y designer conservatism in the selection of soil parameters and the estimation of

environmental loads;
y designer experience with similar pile geometries, soil conditions and loading
types;
y an increase of pile shaft capacity with time (although effects of cyclic degradation
may also be expected to increase with time);
y a reserve of capacity in compression, which is nominally defined at a tip
displacement of 0.1D; and
y conservatism in the estimate of the resistance of the total foundation system e.g.
redundancy of foundation elements, exclusion of the resistance in compression
provided by mudmats on jacket structures, among other issues.
13
Many of these factors are also noted by Bea et al. (1999) in their review of the
reliability of offshore foundation performance in clay soils. The influence of time and
pile head displacement of a 0.76m diameter open ended pile in dense sand
(EURIPIDES, Kolk et al. 2005b) is illustrated in Figure 8. It may be seen that an
extra 20% (or more) of short term capacity may be available due to the mobilization
of end bearing, although significant deformations will occur. Of greater significance
is the increase, by at least 50%, of the peak compression capacity over the 1.5 year
period between initial loading and re-loading. This strong ageing characteristic is
associated primarily with increases in shaft friction (e.g. Chow et al. 1998); the rate of
increase with time varies with the sand properties and the pile’s loading history and
geometry, but is presently difficult to quantify reliably.
These additional factors are potential sources of conservative bias and appear to be a
major reason why offshore piled foundations have been designed with high levels of
reliability at low factors of safety (but, evidently, not with low values of uncertainty
in the evaluation of resistance). These sources of bias need to be considered in the
assessment of foundation system reliability when designing for different type of
loading conditions, soil types and pile geometries. Load and resistance factor design
(LRFD) provides a framework which is more suited for quantification of these site
specific differences than use of global factors of safety. It is noteworthy, for example,
that a majority of offshore design experience is related to the performance of ‘jacket
structures’, which resist wave induced moments in ‘push-pull’ by generating
compression and tension pile loads. However, newer offshore foundation solutions
involving anchor piles (which are permanently in tension) cannot rely on a
conservative assessment of capacity in compression and, unlike jacket structures,
their reliability relates strongly to the estimated pile shaft friction.
CONCLUSIONS
The companion paper (Schneider et al. 2007) presents a comparison of the predictive
performance of methods included in API (2006) for evaluation of the axial capacity
14
of driven pile foundations in siliceous sands. This paper uses the database of pile tests
assembled to investigate the relative reliability of these methods. It is shown that:
• The nominal (or relative) reliability index of the UWA-05 method is generally
greater than that of the other methods considered in API (2006). This is likely to
be because UWA-05 specifically caters for the following factors which research
over the past 20 years has shown to have an important influence on driven pile
capacity:
- dependence of shaft friction and base capacity on the CPT qc value;
- reduction in local friction (τf) with continued pile penetration (i.e. friction
fatigue);
- influence of soil displacement during installation on shaft friction and base
capacity;
- variability in the coefficient of friction between the soil and pile;
- changes in radial stress (and hence available friction) at the soil pile interface
during pile loading;
- effect of variability in qc near the pile tip (soft layers) on end bearing capacity.
• Extrapolation bias due to method formulation may be more significant than
extrapolation uncertainty, and will significantly influence reliability of offshore
foundations designed at low factors of safety. The UWA-05 method provides a
safer extrapolation of capacity for large offshore piles from the existing database
of pile test results and other methods in API (2006) are potentially un-
conservative.
• The assessment of actual reliability and true factors of safety for use in efficient
design requires quantification of additional factors such as the time dependence
of shaft friction and foundation system response in addition to assessment of
uncertainties related to site variability, method formulation and environmental
load estimation.
• In the absence of site specific pile load tests, uncertainty in extrapolation bias can
only be accounted for by a better understanding of mechanisms which influence
pile capacity.
15
ACKNOWLEDGEMENTS
The authors gratefully acknowledge the support provided by the Australian Research
Council. The first and third authors were also supported through International
Postgraduate Research Scholarships and University Postgraduate Awards from the
University of Western Australia. Fugro Engineers B.V. is thanked for providing
access to data for the EURIPIDES pile load tests.
REFERENCES
Offshore Platforms, API RP2A, 1st Edition, American Petroleum Institute,
Washington, D.C.
Offshore Platforms – Working Stress Design, API RP2A, 21st Edition,
American Petroleum Institute, Washington, D.C.
API (2006). DRAFT Recommended Practice for Planning, Designing, and
Constructing Fixed Offshore Platforms – Working Stress Design, API RP2A,
22nd Edition, American Petroleum Institute, Washington, D.C.
Bea, R.G., Jin, Z., Valle, C., and Ramos, R. (1999). “Evaluation of reliability of
platform pile foundations.” J. Geotech. Geoenviron. Eng., 125(8), 696-704.
Bustamante, M., and Gianeselli, L. (1982). “Pile bearing capacity prediction by
means of static penetrometer CPT.” Proc., 2nd Europ. Symp. on Penetration
Test., Amsterdam, 493-500.
Chow, F.C., Jardine, R.J., Brucy, F., and Nauroy, J.F. (1998). “Effects of time on
capacity of pipe piles in dense North Sea sand.” J. Geotech. Geoenviron.
Eng., 124(3), 254-264.
Christian, J.T. (2004). “Geotechnical engineering reliability: How well do we know
what we are doing?” J. Geotech. Geoenviron. Eng., 130(10), 985-1003.
Claussen, C.F.J., Aas, P.M., and Karlsrud, K. (2005). “Bearing capacity of driven
piles in sand, the NGI approach.” Proc., Int. Symp. Frontiers Offshore
Geomech. ISFOG, Perth, 677-681.
16
Dennis, N.D., Jr., and Olson, R.E. (1983). “Axial capacity of steel pipe piles in sand.”
Proc. of the Conf. on Geotech. Practice in Offshore Eng., Austin, Texas,
USA, 389-402.
Eslami, A., and Fellenius, B. (1997). “Pile capacity by direct CPT and CPTu methods
applied to 102 case histories.” Canadian Geotechnical Journal, 34, 886-904.
Harr, M.E. (1987). Reliability based Design in Civil Engineering, Dover Publications,
Inc. New York, 291.
Hicks, M.A. (2005). “Risk and variability in geotechnical engineering.”
Jardine, R.J., Chow, F.C., Overy, R.F., and Standing, J.R. (2005). ICP design
methods for driven piles in sands and clays, Thomas Telford, London, 98.
Kolk, H.J., Baaijens, A.E., and Senders, M. (2005a). “Design criteria for pipe piles in
silica sands.” Proc., Int. Symp. Frontiers Offshore Geomech. ISFOG, Perth,
711-716.
Kolk, H. J., Baaijens, A. E., and Vergobi, P. (2005b). "Results of axial load tests on
pipe piles in very dense sands: the EURIPIDES JIP." Proc., Int. Symp.
Lacasse, S., and Goulois, A. (1989). “Reliability analysis of axial pile capacity.”
Proc., 12th Int. Conf. Soil Mech. and Found. Eng., Rio de Janeiro, 845-848.
Lacasse, S., and Nadim, F. (1994). “Reliability issues and future challenges in
geotechnical engineering for offshore structures.” Proc., 7th Int. Conf. on
Behaviour of Offshore Struct., Invited Papers, MIT, Cambridge, MA, 9-38.
Lehane, B.M., and Jardine, R.J. (1994). "Shaft capacity of driven piles in sand: a new
design approach." Proc., 7th Int. Conf. on Behaviour of Offshore Struct., MIT,
Cambridge, MA, 23-36.
Lehane, B.M., Schneider, J.A., and Xu, X. (2005a). “The UWA-05 method for
prediction of axial capacity of driven piles in sand.” Proc., Int. Symp.
Lehane, B.M., Schneider, J.A., and Xu, X. (2005b). “A review of design methods for
offshore driven piles in siliceous sand.” UWA Report GEO 05358, The
University of Western Australia, Perth, 105.
17
McClelland, B. (1969). “Fixed Structures: Foundations.” in Handbook of Ocean and
Underwater Engineering, McGraw-Hill, New York, 8:98-8:125.
McClelland, B. (1974). "Design of deep penetration piles for ocean structures." J.
Geotech. Eng. Div., 100(GT7), 709-747.
Paikowsky, S., Birgisson, B., McVay, M., Nguyen, T., Kuo, C., Baecher, G., Ayyub,
B., Stenersen, K., O’Malley, K., Chernauskas, L., and O’Neill, M. (2004).
“Load and resistance factor design (LRFD) for deep foundations.” NCHRP
Report 507, Transportation Research Board, Washington, D.C., 76.
Pelletier, J.H., Murff, J.D., and Young, A.C. (1993). “Historical development and
assessment of the current API design methods for axially loaded piles.” Proc.,
25th Offshore Technol. Conf., OTC 7157, Houston, Tex., 253-282.
Phoon, K.K., and Kulhawy, F.H. (2002). “Drilled shaft design for transmission
structures using LRFD and MRFD.” Deep Foundations 2002, Geotechnical
Special Publication No. 116, ASCE, Reston, VA, 1006-1017.
Randolph, M.F., Cassidy, M., Gourvenec, S., and Erbrich, C. (2005). Challenges of
offshore geotechnical engineering, Proc., 16th Int. Conf. Soil Mech. and
Geotech. Eng., Osaka, 123-176.
Schneider, J.A., Xu., X., and Lehane, B.M. (2007). Database assessment of CPT
based design methods for axial capacity of driven piles in siliceous sands: Part
I – overview of methods, database and predictive performance, J. Geotech.
Geoenviron. Eng., submitted.
Tang, W.H., Woodford, D.L., and Pelletier, J.H. (1990). “Performance reliability of
offshore piles.” Proc., 22nd Offshore Technol. Conf., OTC 6379, Houston,
Tex., 299-308.
Wu, T.H., Tang, W.H., Sangrey, D.A., and Baecher, G.B. (1989). “Reliability of
offshore foundations – State of the art.” J. Geotech. Eng., 115(2), 157-178.
Xu, X., Schneider, J.A., and Lehane, B.M. (2007). "End bearing of open and closed
ended driven piles in siliceous sand using the CPT." Canadian Geotech. J.,
submitted.
Zhang, L., Tang, W.H., Zhang, L., and Zheng, J. (2004). “Reducing uncertainty of
prediction from empirical correlations.” J. Geotech. Geoenviron. Eng., 130(5),
526-534.
18
Table 1. Inclusion of primary mechanisms influencing capacity of piles in design
methods
API-00 LCPC-82 EF-97 Fugro-05 ICP-05 NGI-05 UWA-05
σ'rc = f(qc) No Variable Constant Constant Constant Variable Constant
Tension
friction
capacity No No No Yes Yes Yes Yes
lower than
compression
Limiting Limiting
Friction
shaft shaft No h/D h/D z/L h/D
Fatigue
friction friction
Increase in
lateral stress
No No No No Yes No Yes
during pile
loading
Interface
friction angle
See Note
influenced by No No No No Yes Yes
4
relative
roughness
Plugged end
bearing / See Note
Yes Yes Yes Yes Yes Yes
internal shaft 3
friction
Pile
See Note See Note
displacement No No Yes Yes Yes
1 1
ratio on σ'rc
Incremental
Filling Ratio No No No No No No Yes
on σ'rc
Incremental
Filling Ratio No No No No No No Yes
on qb/qc
Time Effects See Note 2

1
API-00 and NGI-05 allow for differences in capacity for open as compared to closed ended piles, but do not account for
differences between thick walled open ended piles and thin walled open ended piles.
2
No method explicitly accounts for time effects. Approximate design times have been recommended based on performance
relative to the database for certain methods, but implications of these design times and influence of time effects after that design
time still warrants additional study, particularly for large diameter piles.
3
For large diameter piles or piles in loose sands, internal shaft friction or pile plugging behaviour is not explicitly included in
ICP-05.
4
NGI-05 recommends higher shaft friction for concrete piles as opposed to steel piles, which may be related to interface friction
angle. This simplification which does not include particle size effects does not appear to work well for all cases in the database.
19
Table 2. Comparison of UWA-05 database statistics to case assuming IFR=1 and
Δσ'rd = 0
Database IFR Δσ'rd Geometric Standard
Mean, Deviation
μgR of log,
σlnR
Closed Ended Piles Measured or f(Di) f(D, qc, σ'v0) 0.90 0.28
in Compression 1 0 0.80 0.30
(CEC)
Closed Ended Piles Measured or f(Di) f(D, qc, σ'v0) 0.85 0.29
in Tension (CET) 1 0 0.69 0.31
Open Ended Piles in Measured or f(Di) f(D, qc, σ'v0) 0.96 0.19
Compression (OEC) 1 0 0.74 0.27
Open Ended Piles in Measured or f(Di) f(D, qc, σ'v0) 0.97 0.19
Tension (OET) 1 0 0.76 0.18
Total Database Measured or f(Di) f(D, qc, σ'v0) 0.92 0.25
1 0 0.76 0.27
20
qc (MPa)
0 10 20 30 40 50 60
0
10
20
30
Depth (m)
40
50
60
70
80
Qshaft,T (MN)
0 10 20 30 40 50 60 70
0
API-00
10
Fugro-05
20 ICP-05
Tip Depth (m)
NGI-05
30
UWA-05
40
50
60
70 Tension
80
Qtotal,C (MN)
0 20 40 60 80 100 120 140
0
API-00
10 Fugro-05
20 ICP-05
NGI-05
Tip Depth (m)
30
UWA-05
40
50
60
70 Compression
80
Figure 1. Comparison of design total capacities for API (2006) main text and
commentary methods at a Gulf of Mexico location (Lehane et al. 2005b).
The dashed lines indicate pile tip depths using API-00 for tension (30
MN) and compression (70 MN)
21
2.5
CEC
2 CET
OEC
μ gR 1.5 OET
All
1
0.5
0
API-00
NGI-05
ICP-05
EF-97
LCPC-82
Fugro-05
UWA-05
1
0.9 CEC
0.8 CET
0.7 OEC
0.6 OET
σlnR
0.5 All
0.4
0.3
0.2
0.1
0
API-00
NGI-05
ICP-05
EF-97
LCPC-82
Fugro-05
UWA-05
4
3.5 FSA=2
nominal reliabilty index, β
3
2.5
2
1.5 CEC
1 CET
0.5 OEC
0 OET
-0.5 All
-1
NGI-05
ICP-05
API-00
LCPC-82
EF-97
Fugro-05
UWA-05
Figure 2. Statistical values (μgR, σlnR & β) for design methods based on database of
driven piles in siliceous sands presented by Schneider et al. (2007).
22
5
Fugro-05
Nominal Reliability Index, β

ICP-05
4
NGI-05
3 UW A-05
1 FSA =2
Full Database
0
0.4 0.6 0.8 1 1.2 1.4
System Bias, μ gQ/μ gR
Figure 3. Nominal reliability index as a function of system bias for design methods
and database of driven piles in siliceous sands presented by Schneider et al. (2007)
23
2.5 CEC CET
OEC OET
2
Qc [Fugro-05]/Qm
1.5 Lock & Dam 26
0.5
Fugro-05
0
0 0.5 1 1.5 2
Diameter, D (m)
2.5 2.5
Lock &
CEC CET NGI-05
Dam 26
2 2
OEC OET
Hound
Qc [NGI-05]/Qm
Qc [ICP-05]/Qm
1.5 Lock & Dam 26 1.5 Point
1 1
0.5 0.5 CEC CET

ICP-05
OEC OET
0 0
0 0.5 1 1.5 2 0 0.5 1 1.5 2
Diameter, D (m) Diameter, D (m)
2.5 2.5
CEC CET CEC CET
2 2
OEC OET OEC OET
Qc [UWA-05]/Qm
Qc [UWA-05]/Qm
1.5 Lock & Dam 26 1.5 Lock & Dam 26
1 1
UWA-05
0.5 0.5
IFR=1
UWA-05
Δσ'rd=0
0 0
0 0.5 1 1.5 2 0 0.5 1 1.5 2
Figure 4. Potential bias towards diameter for database piles
24
1.5 1.5
1.4 40m Fugro-05/UWA-05 1.4 Fugro-05/UWA-05
1.3 50m 1.3
QFugro-05/QUWA-05
QFugro-05/QUWA-05
1.2 60m 1.2 40m
1.1 1.1 50m
1 1 60m
0.9 0.9
0.8 0.8
0.7 Compression 0.7 Tension
0.6 Profile: Figure 1 0.6 Profile: Figure 1
0.5 0.5
0 0.5 1 1.5 2 2.5 0 0.5 1 1.5 2 2.5
1.5 1.5
1.4 40m ICP-05/UWA-05 1.4
1.3 50m 1.3
1.2 60m 1.2
QICP-05/QUWA-05
QICP-05/QUWA-05
1.1 1.1
1 1
0.9 0.9
ICP-05/UWA-05 40m
0.8 0.8
0.7 0.7 50m
Compression Tension
0.6 Profile: Figure 1 0.6 Profile: Figure 1 60m
0.5 0.5
0 0.5 1 1.5 2 2.5 0 0.5 1 1.5 2 2.5
1.5 1.5
1.4 1.4
1.3 1.3
1.2 1.2
QNGI-05/QUWA-05
QNGI-05/QUWA-05
1.1 1.1
1 1
0.9 0.9
NGI-05/UWA-05 40m NGI-05/UWA-05 40m
0.8 0.8
0.7 50m 0.7 50m
Compression Tension
0.6 Profile: Figure 1 60m 0.6 Profile: Figure 1 60m
0.5 0.5
0 0.5 1 1.5 2 2.5 0 0.5 1 1.5 2 2.5
Figure 5. Diameter bias for ‘offshore’ CPT pile design methods as compared to
UWA-05 for CPT qc profile in Figure 1 and D/t = 50
25
β s = τavg /σ'vo,avg
0 0.5 1 1.5 2
0
10
20 typical database piles
30
API-06; qc1N=60
40
depth (m)
API-06; qc1N=180
50 UWA-05; qc1N=60
60 UWA-05; qc1N=180
shaft friction in
70 compression
D = 1.2m
80 t = 32mm
IFR=1
90 o
δ = 29
100
Figure 6. Comparison of average shaft friction in compression to average vertical
effective stress in uniform (constant relative density) silica sand profiles for UWA-05
and API-00/API-06. Note: qc1N=(qc/pref)/(σ'v0/pref)0.5 and pref = 100 kPa
α s = qc,avg / τavg
0 100 200 300 400 500 600 700 800
0
10
20 typical database piles
30
API-06; qc1N=60
40
depth (m)
API-06; qc1N=180
50
UWA-05; qc1N=60
60 UWA-05; qc1N=180
shaft friction in
70
compression
D = 1.2m
80
t = 32mm
IFR=1
90 o
δ = 29
100
Figure 7. Comparison of average shaft friction in compression to average cone tip
resistance in uniform (constant relative density) silica sand profiles for UWA-05 and
API-00/API-06
26
35
I-47m; 11 days
30
II-46.7m; 6 days
25 II-46.7m; 541 days

Force (MN)
20
15 Pile tip
displacement
of 0.1D
10
0
50 0 -50 -100 -150 -200 -250 -300
pile head displacement (mm)
Figure 8. Influence of displacement and time on compression capacity of 0.76m

diameter pipe piles for EURIPIDES (data from Kolk et al. 2005b)
27
APPENDIX B STATIC LOAD TEST RESULTS
B.1 INTRODUCTION
In total, 100 static load tests (91 in samples A to G and 9 in sample H) were
conducted involving piles of different diameters (D=6, 9.5 & 16mm) at various stress
levels (σ'v0 from to 65kPa to 190kPa) in uniform and layered soil samples. The
relative density of the sand (Dr) ranges from 0.36 to 0.96. Table B.1 and Table B.2
summarise the static load test results (i.e. qc, qb,residual, qb0.1 & GIN) in sample H and
samples A to G. Pile dimensions (D & L) are presented in model scale.
This Appendix presents the normalised resistance mobilisation curve (qb/qc vs. w/D)
and stiffness decay curves (G/GIN vs. w/D & qb/qc) for each static load tests and also
provides an accompanying summary table.
The static load tests in samples A to G were then sorted into 11 groups, each with a
similar ratio of GIN/qc. Within each group, the data were averaged to give one
representative stiffness decay curves, which are summarised in the final section of
this Appendix.
Table B.1 Summary of the results for the 9 static load tests conducted in sample H
No Test Name g level D L H/D σ'v0 qc qb,residual qb0.1 GIN qb,residual/qc qb0.1/qc GIN/qc
- - - mm mm - kPa MPa MPa MPa MPa - - -
1 H6-1 80 6 154 0.00 130 2.1 1.1 2.0 34 0.50 0.96 16.2
2 H6-2 80 6 152 0.00 127 2.6 1.2 2.1 25 0.44 0.80 9.7
3 H6-3 80 6 146 1.00 123 4.3 1.8 3.1 57 0.43 0.74 13.5
4 H6-4 80 6 139 1.83 118 8.1 2.8 5.8 83 0.34 0.71 10.2
5 H6-5 80 6 194 -6.75 155 1.1 1.0 1.4 11 0.88 1.27 9.6
6 H6-8 80 6 89 10.17 74 20.1 2.7 18.3 171 0.13 0.91 8.5
7 H16-1 80 16 184 -1.94 147 1.3 0.8 1.5 12 0.58 1.13 8.8
8 H16-2 80 16 150 0.19 125 2.2 0.9 1.8 50 0.42 0.85 23.2
9 H16-3 80 16 125 1.75 105 6.0 1.4 4.3 126 0.24 0.71 21.0
B-1
Appendix B Static Load Test Results
Table B.2 Summary of the results for the 91 static load tests conducted in samples A to G
No Test Name g level D L σ'v0 qc qb,residual qb0.1 GIN qb,residual/qc qb0.1/qc GIN/qc
- - - mm mm kPa MPa MPa MPa MPa - - -
1 A6g2MJ-1 37.9 6 180.1 110.2 12.1 1.1 11.1 118.2 0.09 0.92 9.74
2 A6g3MJ-1 60 6 117.5 113.8 11.1 1.3 10.4 105.1 0.12 0.93 9.43
3 A6g3MJ-2 60 6 184.4 178.5 16.2 2.2 14.9 138.4 0.13 0.92 8.56
4 A9g2MJ-1 37.9 9.5 180.0 110.1 12.2 1.2 10.1 94.4 0.09 0.83 7.77
5 A9g3MJ-1 60 9.5 113.4 109.8 12.5 0.7 10.4 95.9 0.05 0.84 7.69
6 A9g3MJ-2 60 9.5 179.7 174.0 19.2 1.9 15.7 118.4 0.10 0.82 6.18
7 A16g1MJ-1 22.5 16 179.6 65.2 7.1 0.4 6.6 70.1 0.05 0.92 9.84
8 A16g3MJ-1 60 16 116.9 113.2 10.0 0.8 9.4 99.4 0.08 0.95 9.99
9 A16g3MJ-2 60 16 187.5 181.6 17.0 1.2 16.4 159.4 0.07 0.96 9.38
10 B6g3MJ-1 60 6 120.1 112.7 2.9 0.1 2.9 86.7 0.02 0.97 29.52
11 B6g3MJ-2 60 6 180.3 169.2 4.7 0.3 4.5 99.1 0.07 0.95 21.03
12 B6g3J-1 60 6 180.3 169.2 7.0 0.9 6.6 106.2 0.14 0.95 15.24
13 B9g3MJ-1 60 9.5 119.8 112.5 6.5 0.4 5.6 96.8 0.06 0.87 15.00
14 B9g3MJ-2 60 9.5 179.9 168.9 10.3 0.8 8.6 125.1 0.07 0.84 12.19
15 B9g3J-1 60 9.5 119.8 112.5 6.5 0.4 5.6 95.0 0.06 0.86 14.62
16 B9g3J-2 60 9.5 179.9 168.9 10.3 0.8 8.6 125.8 0.08 0.84 12.20
17 B16g3MJ-1 60 16 119.6 112.3 5.3 0.2 5.1 87.3 0.05 0.96 16.43
18 B16g3MJ-2 60 16 179.6 168.6 8.3 0.5 8.0 108.0 0.06 0.95 12.94
19 B16g3J-1 60 16 119.6 112.3 5.9 0.3 5.8 88.3 0.06 0.98 14.90
20 B16g3J-2 60 16 179.5 168.5 9.7 0.6 9.3 132.0 0.07 0.95 13.58
21 C6g3MJ-1 60 6 119.7 115.3 23.7 1.1 21.1 112 0.05 0.89 4.74
22 C6g3MJ-2 60 6 159.8 157.1 36.9 3.5 33.2 232 0.09 0.90 6.30
23 C6g3MJ-3 60 6 179.6 177.7 43.2 5.1 38.4 200 0.12 0.89 4.64
24 C6g3J-1 60 6 119.5 115.1 23.8 1.6 23.0 146 0.07 0.96 6.11
25 C6g3J-2 60 6 159.3 156.6 42.0 5.4 37.8 202 0.13 0.90 4.81
26 C6g3J-3 60 6 179.2 177.2 47.7 7.5 41.0 182 0.16 0.86 3.80
27 C9g3MJ-1 60 9.5 119.3 114.9 28.4 1.4 22.4 156 0.05 0.79 5.48
28 C9g3J-1 60 9.5 119.1 114.7 26.7 1.4 22.4 167 0.05 0.84 6.28
29 C9g3J-2 60 9.5 129.1 125.2 31.8 2.4 27.6 192 0.08 0.87 6.02
30 C16g3MJ-1 60 16 118.4 114.0 20.4 0.7 18.4 106 0.03 0.90 5.19
31 C16g3MJ-2 60 16 157.5 154.6 31.3 1.8 28.8 137 0.06 0.92 4.39
32 C16g3J-1 60 16 118.2 113.8 21.3 0.9 19.5 114 0.04 0.92 5.35
33 C16g3J-2 60 16 157.4 154.5 36.8 2.2 32.5 180 0.06 0.89 4.91
34 D6g3J-1 60 6 119.4 124.1 29.2 3.5 26.1 164 0.12 0.89 5.62
35 D6g3J-2 60 6 149.5 155.5 36.6 4.6 32.8 183 0.13 0.90 5.01
36 D6g3J-3 60 6 179.2 186.3 45.4 5.7 40.5 214 0.12 0.89 4.71
37 D6g3MJ-1 60 6 119.3 124.1 31.6 2.7 27.8 175 0.09 0.88 5.54
38 D6g3MJ-2 60 6 149.4 155.4 40.5 4.0 35.8 215 0.10 0.88 5.31
39 D6g3MJ-3 60 6 179.2 186.3 48.1 5.5 42.6 214 0.11 0.89 4.44
40 D9g3J-1 60 9.5 119.0 123.8 34.0 2.4 27.4 206 0.07 0.80 6.07
41 D9g3MJ-1 60 9.5 118.9 123.7 32.6 2.1 26.2 213 0.07 0.80 6.53
42 D16g3J-1 60 16 117.9 122.6 23.7 1.3 21.5 120 0.05 0.91 5.07
43 D16g3J-2 60 16 147.7 153.6 30.6 1.9 27.9 150 0.06 0.91 4.88
44 D16g3MJ-1 60 16 118.0 122.7 25.1 1.4 22.3 129 0.06 0.89 5.16
45 D16g3MJ-2 60 16 147.7 153.6 34.0 2.2 30.0 159 0.06 0.88 4.68
46 E6g2MJ-1 37.9 6 119.6 77.2 19.4 2.5 17.3 115 0.13 0.89 5.93
47 E6g2MJ-2 37.9 6 159.5 101.4 29.2 3.7 25.9 158 0.13 0.89 5.42
48 E6g3MJ-1 60 6 119.5 122.1 30.7 3.9 27.4 211 0.13 0.89 6.86
49 E6g3MJ-2 60 6 159.4 160.5 40.2 5.1 35.8 221 0.13 0.89 5.51
50 E9g2MJ-1 37.9 9.5 119.3 77.1 18.9 1.0 15.1 113 0.05 0.80 5.96
51 E9g2MJ-2 37.9 9.5 159.0 101.2 27.7 1.9 22.1 171 0.07 0.80 6.18
52 E9g3MJ-1 60 9.5 118.8 121.5 30.3 1.8 23.6 131 0.06 0.78 4.32
53 E16g2MJ-1 37.9 16 118.7 76.7 15.4 0.7 14.6 113 0.05 0.95 7.34
54 E16g2MJ-2 37.9 16 157.9 100.5 24.9 1.6 22.2 131 0.06 0.89 5.25
55 E16g3MJ-1 60 16 118.0 120.7 27.4 1.4 23.9 157 0.05 0.87 5.73
56 E16g3MJ-2 60 16 137.5 139.5 34.0 2.2 29.5 194 0.06 0.87 5.69
57 F6g2MJ-1 37.9 6 123.8 79.8 9.5 1.1 9.0 82 0.12 0.95 8.60
58 F6g2MJ-2 37.9 6 163.6 104.0 18.0 1.7 16.1 118 0.10 0.89 6.55
59 F6g2MJ-3 37.9 6 183.5 116.8 22.5 2.4 20.3 129 0.11 0.90 5.74
60 F6g3MJ-1 60 6 123.8 126.3 13.0 2.3 12.3 110 0.18 0.95 8.52
61 F6g3MJ-2 60 6 163.6 164.6 22.3 2.9 20.6 168 0.13 0.92 7.54
62 F6g3MJ-3 60 6 183.5 184.9 29.3 3.1 26.3 183 0.11 0.90 6.24
63 F9g2MJ-1 37.9 9.5 123.5 79.6 10.7 0.9 8.9 71 0.08 0.83 6.66
64 F9g2MJ-2 37.9 9.5 163.3 103.8 21.6 1.7 17.5 113 0.08 0.81 5.23
65 F9g2MJ-3 37.9 9.5 183.1 116.5 26.7 1.7 21.0 118 0.06 0.78 4.43
66 F9g3MJ-1 60 9.5 123.4 126.0 15.8 1.3 13.1 114 0.08 0.83 7.19
67 F9g3MJ-2 60 9.5 163.2 164.2 26.6 1.8 21.6 140 0.07 0.81 5.26
68 F16g2MJ-1 37.9 16 123.2 79.4 10.2 0.5 9.5 78 0.05 0.94 7.70
69 F16g2MJ-2 37.9 16 162.6 103.3 18.4 1.1 16.5 118 0.06 0.90 6.40
70 F16g2MJ-3 37.9 16 182.2 115.9 23.1 1.4 20.8 145 0.06 0.90 6.29
71 F16g3MJ-1 60 16 122.8 125.3 15.5 0.8 14.2 127 0.05 0.92 8.20
72 F16g3MJ-2 60 16 162.0 163.0 27.2 1.4 24.0 170 0.05 0.88 6.24
73 F16g3MJ-3 60 16 181.5 182.9 33.8 2.2 30.4 221 0.06 0.90 6.54
74 G6g2MJ-1 37.9 6 123.8 80.3 9.9 1.4 9.3 86 0.14 0.94 8.71
75 G6g2MJ-2 37.9 6 163.7 104.7 12.2 1.3 11.5 101 0.10 0.94 8.30
76 G6g2MJ-3 37.9 6 183.7 116.9 12.6 1.7 12.0 103 0.13 0.95 8.12
77 G6g3MJ-1 60 6 123.8 127.1 15.9 1.9 13.2 122 0.12 0.83 7.68
78 G6g3MJ-2 60 6 163.7 165.7 15.9 2.3 15.1 134 0.15 0.94 8.37
79 G6g3MJ-3 60 6 183.6 185.0 17.2 2.2 16.3 143 0.13 0.95 8.29
80 G9g2MJ-1 37.9 9.5 123.6 80.2 10.2 1.2 8.5 71 0.12 0.83 6.95
81 G9g2MJ-2 37.9 9.5 163.5 104.6 12.0 1.1 10.1 77 0.09 0.84 6.39
82 G9g2MJ-3 37.9 9.5 183.4 116.7 13.0 1.2 10.8 85 0.09 0.84 6.55
83 G9g3MJ-1 60 9.5 123.4 126.7 17.1 1.2 14.2 124 0.07 0.83 7.25
84 G9g3MJ-2 60 9.5 163.3 165.3 21.4 2.0 17.7 132 0.09 0.83 6.19
85 G9g3MJ-3 60 9.5 183.2 184.6 24.0 2.2 20.0 152 0.09 0.83 6.33
86 G16g2MJ-1 37.9 16 123.1 79.9 10.3 0.7 9.8 83 0.07 0.95 8.07
87 G16g2MJ-2 37.9 16 162.9 104.2 13.2 0.9 12.2 98 0.07 0.93 7.42
88 G16g2MJ-3 37.9 16 182.7 116.3 14.9 1.2 13.9 107 0.08 0.93 7.17
89 G16g3MJ-1 60 16 122.8 126.2 15.2 0.9 14.1 120 0.06 0.92 7.91
90 G16g3MJ-2 60 16 162.4 164.4 20.1 1.6 18.5 133 0.08 0.92 6.63
91 G16g3MJ-3 60 16 182.1 183.5 23.6 1.8 21.2 147 0.08 0.90 6.22
B-2
B.2 RESULTS OF INDIVIDUAL STATIC LOAD TEST
B.2.1 Sample A
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
A6g2MJ-1 A6g2MJ-1
0.2 0.2
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
Summary Table for Test A6g2MJ
1 g level= D L σ'v0 qc qb,residual
37.9 mm mm kPa MPa MPa
0.8 A6g2MJ-1 6.0 180.1 110.2 12.1 1.1
G/GIN (-)
0.6
Dr= qb0.1 GIN qb,residual/qc qb0.1/qc GIN/qc
0.4
0.54 MPa MPa - - -
A6g2MJ-1 A6g2MJ-1 11.15 118.19 0.09 0.92 9.74
0.2
0
0.1 1 10 100
w/D (%)
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
A6g3MJ-1 A6g3MJ-1
0.2 A6g3MJ-2 0.2 A6g3MJ-2
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
60 mm mm kPa MPa MPa
0.8 A6g3MJ-1 6.0 117.5 113.8 11.1 1.3
G/GIN (-)
A6g3MJ-2 6.0 184.4 178.5 16.2 2.2

0.6
0.4
0.54 MPa MPa - - -
A6g3MJ-1 A6g3MJ-1 10.37 105.11 0.12 0.93 9.43
0.2 A6g3MJ-2 A6g3MJ-2 14.87 138.39 0.13 0.92 8.56
0
0.1 1 10 100
w/D (%)
B-3
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
A9g2MJ-1 A9g2MJ-1
0.2 0.2
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 A9g2MJ-1 9.0 180.0 110.1 12.2 1.2
G/GIN (-)
0.6
0.4
0.54 MPa MPa - - -
A9g2MJ-1 A9g2MJ-1 10.07 94.45 0.09 0.83 7.77
0.2
0
0.1 1 10 100
w/D (%)
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
A9g3MJ-1 A9g3MJ-1
0.2 A9g3MJ-2 0.2 A9g3MJ-2
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 A9g3MJ-1 9.0 113.4 109.8 12.5 0.7
G/GIN (-)
A9g3MJ-2 9.0 179.7 174.0 19.2 1.9

0.6
0.4
0.54 MPa MPa - - -
A9g3MJ-1 A9g3MJ-1 10.43 95.92 0.05 0.84 7.69
0.2 A9g3MJ-2 A9g3MJ-2 15.73 118.38 0.10 0.82 6.18
0
0.1 1 10 100
w/D (%)
B-4
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
A16g1MJ-1 A16g1MJ-1
0.2 0.2
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 A16g1MJ-1 16.0 179.6 65.2 7.1 0.4
G/GIN (-)
0.6
0.4
0.54 MPa MPa - - -
A16g1MJ-1 A16g1MJ-1 6.57 70.05 0.05 0.92 9.84
0.2
0
0.1 1 10 100
w/D (%)
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
A16g3MJ-1 A16g3MJ-1
0.2 A16g3MJ-2 0.2 A16g3MJ-2
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 A16g3MJ-1 16.0 116.9 113.2 10.0 0.8
G/GIN (-)
A16g3MJ-2 16.0 187.5 181.6 17.0 1.2

0.6
0.4
0.54 MPa MPa - - -
A16g3MJ-1 A16g3MJ-1 9.44 99.35 0.08 0.95 9.99
0.2 A16g3MJ-2 A16g3MJ-2 16.37 159.44 0.07 0.96 9.38
0
0.1 1 10 100
w/D (%)
B-5
B.2.2 Sample B
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
B6g3MJ-1 B6g3MJ-1
0.2 B6g3MJ-2 0.2 B6g3MJ-2
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
Summary Table for Test B6g3MJ
0.8 B6g3MJ-1 6.0 120.1 112.7 2.9 0.1
G/GIN (-)
B6g3MJ-2 6.0 180.3 169.2 4.7 0.3

0.6
0.4
0.36 MPa MPa - - -
B6g3MJ-1 B6g3MJ-1 2.86 86.71 0.02 0.97 29.52
0.2 B6g3MJ-2 B6g3MJ-2 4.50 99.11 0.07 0.95 21.03
0
0.1 1 10 100
w/D (%)
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
B6g3J-1 B6g3J-1
0.2 0.2
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
Summary Table for Test B6g3J
0.8 B6g3J-1 6.0 180.3 169.2 7.0 0.9
G/GIN (-)
0.6
0.4
0.36 MPa MPa - - -
B6g3J-1 B6g3J-1 6.63 106.24 0.14 0.95 15.24
0.2
0
0.1 1 10 100
w/D (%)
B-6
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
B9g3MJ-1 B9g3MJ-1
0.2 B9g3MJ-2 0.2 B9g3MJ-2
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 B9g3MJ-1 9.0 119.8 112.5 6.5 0.4
G/GIN (-)
B9g3MJ-2 9.0 179.9 168.9 10.3 0.8

0.6
0.4
0.36 MPa MPa - - -
B9g3MJ-1
B9g3MJ-1 5.62 96.81 0.06 0.87 15.00
0.2 B9g3MJ-2 B9g3MJ-2 8.64 125.10 0.07 0.84 12.19
0
0.1 1 10 100
w/D (%)
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
B9g3J-1 B9g3J-1
0.2 B9g3J-2 0.2 B9g3J-2
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 B9g3J-1 9.0 119.8 112.5 6.5 0.4
G/GIN (-)
B9g3J-2 9.0 179.9 168.9 10.3 0.8

0.6
0.4
0.36 MPa MPa - - -
B9g3J-1 B9g3J-1 5.61 94.96 0.06 0.86 14.62
0.2 B9g3J-2 B9g3J-2 8.65 125.80 0.08 0.84 12.20
0
0.1 1 10 100
w/D (%)
B-7
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
B16g3MJ-1 B16g3MJ-1
0.2 B16g3MJ-2 0.2 B16g3MJ-2
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 B16g3MJ-1 16.0 119.6 112.3 5.3 0.2
G/GIN (-)
B16g3MJ-2 16.0 179.6 168.6 8.3 0.5

0.6
0.4
0.36 MPa MPa - - -
B16g3MJ-1 B16g3MJ-1 5.09 87.33 0.05 0.96 16.43
0.2 B16g3MJ-2 B16g3MJ-2 7.96 108.01 0.06 0.95 12.94
0
0.1 1 10 100
w/D (%)
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
B16g3J-1 B16g3J-1
0.2 B16g3J-2 0.2 B16g3J-2
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 B16g3J-1 16.0 119.6 112.3 5.9 0.3
G/GIN (-)
B16g3J-2 16.0 179.5 168.5 9.7 0.6

0.6
0.4
0.36 MPa MPa - - -
B16g3J-1 B16g3J-1 5.81 88.27 0.06 0.98 14.90
0.2 B16g3J-2 B16g3J-2 9.28 132.01 0.07 0.95 13.58
0
0.1 1 10 100
w/D (%)
B-8
B.2.3 Sample C
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
C6g3MJ-1 C6g3MJ-1
0.2 C6g3MJ-2 0.2 C6g3MJ-2
C6g3MJ-3 C6g3MJ-3
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
Summary Table for Test C6g3MJ
0.8 C6g3MJ-1 6.0 119.7 115.3 23.7 1.1
G/GIN (-)
C6g3MJ-2 6.0 159.8 157.1 36.9 3.5

0.6 C6g3MJ-3 6.0 179.6 177.7 43.2 5.1

0.4
0.45/0.92 MPa MPa - - -
C6g3MJ-1 C6g3MJ-1 21.13 112.26 0.05 0.89 4.74
0.2 C6g3MJ-2 C6g3MJ-2 33.18 232.05 0.09 0.90 6.30
C6g3MJ-3 C6g3MJ-3 38.35 200.05 0.12 0.89 4.64
0
0.1 1 10 100
w/D (%)
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
C6g3J-1 C6g3J-1
0.2 C6g3J-2 0.2 C6g3J-2
C6g3J-3 C6g3J-3
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
Summary Table for Test C6g3J
0.8 C6g3J-1 6.0 119.5 115.1 23.8 1.6
G/GIN (-)
C6g3J-2 6.0 159.3 156.6 42.0 5.4

0.6 C6g3J-3 6.0 179.2 177.2 47.7 7.5

0.4
0.45/0.92 MPa MPa - - -
C6g3J-1 C6g3J-1 22.99 145.80 0.07 0.96 6.11
0.2 C6g3J-2 C6g3J-2 37.80 202.00 0.13 0.90 4.81
C6g3J-3 C6g3J-3 40.96 181.55 0.16 0.86 3.80
0
0.1 1 10 100
w/D (%)
B-9
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
C9g3MJ-1 C9g3MJ-1
0.2 0.2
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 C9g3MJ-1 9.0 119.3 114.9 28.4 1.4
G/GIN (-)
0.6
0.4
0.45/0.92 MPa MPa - - -
C9g3MJ-1 C9g3MJ-1 22.37 155.57 0.05 0.79 5.48
0.2
0
0.1 1 10 100
w/D (%)
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
C9g3J-1 C9g3J-1
0.2 C9g3J-2 0.2 C9g3J-2
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 C9g3J-1 9.0 119.1 114.7 26.7 1.4
G/GIN (-)
C9g3J-2 9.0 129.1 125.2 31.8 2.4

0.6
0.4
0.45/0.92 MPa MPa - - -
C9g3J-1 C9g3J-1 22.40 167.44 0.05 0.84 6.28
0.2 C9g3J-2 C9g3J-2 27.63 191.55 0.08 0.87 6.02
0
0.1 1 10 100
w/D (%)
B-10
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
C16g3MJ-1 C16g3MJ-1
0.2 C16g3MJ-2 0.2 C16g3MJ-2
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 C16g3MJ-1 16.0 118.4 114.0 20.4 0.7
G/GIN (-)
C16g3MJ-2 16.0 157.5 154.6 31.3 1.8

0.6
0.4
0.45/0.92 MPa MPa - - -
C16g3MJ-1 C16g3MJ-1 18.40 106.13 0.03 0.90 5.19
0.2 C16g3MJ-2 C16g3MJ-2 28.84 137.24 0.06 0.92 4.39
0
0.1 1 10 100
w/D (%)
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
C16g3J-1 C16g3J-1
0.2 C16g3J-2 0.2 C16g3J-2
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 C16g3J-1 16.0 118.2 113.8 21.3 0.9
G/GIN (-)
C16g3J-2 16.0 157.4 154.5 36.8 2.2

0.6
0.4
0.45/0.92 MPa MPa - - -
C16g3J-1 C16g3J-1 19.48 113.73 0.04 0.92 5.35
0.2 C16g3J-2 C16g3J-2 32.54 180.31 0.06 0.89 4.91
0
0.1 1 10 100
w/D (%)
B-11
B.2.4 Sample D
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
D6g3J-1 D6g3J-1
0.2 D6g3J-2 0.2 D6g3J-2
D6g3J-3 D6g3J-3
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
Summary Table for Test D6g3J
0.8 D6g3J-1 6.0 119.4 124.1 29.2 3.5
G/GIN (-)
D6g3J-2 6.0 149.5 155.5 36.6 4.6

0.6 D6g3J-3 6.0 179.2 186.3 45.4 5.7

0.4
0.93 MPa MPa - - -
D6g3J-1 D6g3J-1 26.13 164.36 0.12 0.89 5.62
0.2 D6g3J-2 D6g3J-2 32.83 183.45 0.13 0.90 5.01
D6g3J-3 D6g3J-3 40.47 213.68 0.12 0.89 4.71
0
0.1 1 10 100
w/D (%)
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
D6g3MJ-1 D6g3MJ-1
0.2 D6g3MJ-2 0.2 D6g3MJ-2
D6g3MJ-3 D6g3MJ-3
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
Summary Table for Test D6g3MJ
0.8 D6g3MJ-1 6.0 119.3 124.1 31.6 2.7
G/GIN (-)
D6g3MJ-2 6.0 149.4 155.4 40.5 4.0

0.6 D6g3MJ-3 6.0 179.2 186.3 48.1 5.5

0.4
0.93 MPa MPa - - -
D6g3MJ-1 D6g3MJ-1 27.77 174.87 0.09 0.88 5.54
0.2 D6g3MJ-2 D6g3MJ-2 35.84 215.20 0.10 0.88 5.31
D6g3MJ-3 D6g3MJ-3 42.58 213.54 0.11 0.89 4.44
0
0.1 1 10 100
w/D (%)
B-12
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
D9g3J-1 D9g3J-1
0.2 0.2
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 D9g3J-1 9.0 119.0 123.8 34.0 2.4
G/GIN (-)
0.6
0.4
0.93 MPa MPa - - -
D9g3J-1 D9g3J-1 27.36 206.36 0.07 0.80 6.07
0.2
0
0.1 1 10 100
w/D (%)
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
D9g3MJ-1 D9g3MJ-1
0.2 0.2
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 D9g3MJ-1 9.0 118.9 123.7 32.6 2.1
G/GIN (-)
0.6
0.4
0.93 MPa MPa - - -
D9g3MJ-1 D9g3MJ-1 26.16 213.14 0.07 0.80 6.53
0.2
0
0.1 1 10 100
w/D (%)
B-13
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
D16g3J-1 D16g3J-1
0.2 D16g3J-2 0.2 D16g3J-2
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 D16g3J-1 16.0 117.9 122.6 23.7 1.3
G/GIN (-)
D16g3J-2 16.0 147.7 153.6 30.6 1.9

0.6
0.4
0.93 MPa MPa - - -
D16g3J-1 D16g3J-1 21.48 120.20 0.05 0.91 5.07
0.2 D16g3J-2 D16g3J-2 27.92 149.57 0.06 0.91 4.88
0
0.1 1 10 100
w/D (%)
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
D16g3MJ-1 D16g3MJ-1
0.2 D16g3MJ-2 0.2 D16g3MJ-2
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 D16g3MJ-1 16.0 118.0 122.7 25.1 1.4
G/GIN (-)
D16g3MJ-2 16.0 147.7 153.6 34.0 2.2

0.6
0.4
0.93 MPa MPa - - -
D16g3MJ-1 D16g3MJ-1 22.26 129.40 0.06 0.89 5.16
0.2 D16g3MJ-2 D16g3MJ-2 29.97 159.12 0.06 0.88 4.68
0
0.1 1 10 100
w/D (%)
B-14
B.2.5 Sample E
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
E6g2MJ-1 E6g2MJ-1
0.2 E6g2MJ-2 0.2 E6g2MJ-2
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
Summary Table for Test E6g2MJ
0.8 E6g2MJ-1 6.0 119.6 77.2 19.4 2.5
G/GIN (-)
E6g2MJ-2 6.0 159.5 101.4 29.2 3.7

0.6
0.4
0.92/0.5/0.83 MPa MPa - - -
E6g2MJ-1
E6g2MJ-1 17.34 115.28 0.13 0.89 5.93
0.2 E6g2MJ-2 E6g2MJ-2 25.94 158.09 0.13 0.89 5.42
0
0.1 1 10 100
w/D (%)
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
E6g3MJ-1 E6g3MJ-1
0.2 E6g3MJ-2 0.2 E6g3MJ-2
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 E6g3MJ-1 6.0 119.5 122.1 30.7 3.9
G/GIN (-)
E6g3MJ-2 6.0 159.4 160.5 40.2 5.1

0.6
0.4
0.92/0.5/0.83 MPa MPa - - -
E6g3MJ-1
E6g3MJ-1 27.40 210.60 0.13 0.89 6.86
0.2 E6g3MJ-2 E6g3MJ-2 35.81 221.23 0.13 0.89 5.51
0
0.1 1 10 100
w/D (%)
B-15
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
E9g2MJ-1 E9g2MJ-1
0.2 E9g2MJ-2 0.2 E9g2MJ-2
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 E9g2MJ-1 9.0 119.3 77.1 18.9 1.0
G/GIN (-)
E9g2MJ-2 9.0 159.0 101.2 27.7 1.9

0.6
0.4
0.92/0.5/0.83 MPa MPa - - -
E9g2MJ-1
E9g2MJ-1 15.06 112.64 0.05 0.80 5.96
0.2 E9g2MJ-2 E9g2MJ-2 22.08 171.10 0.07 0.80 6.18
0
0.1 1 10 100
w/D (%)
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
E9g3MJ-1 E9g3MJ-1
0.2 0.2
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 E9g3MJ-1 9.0 118.8 121.5 30.3 1.8
G/GIN (-)
0.6
0.4
0.92/0.5/0.83 MPa MPa - - -
E9g3MJ-1
E9g3MJ-1 23.56 130.75 0.06 0.78 4.32
0.2
0
0.1 1 10 100
w/D (%)
B-16
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
E16g2MJ-1 E16g2MJ-1
0.2 E16g2MJ-2 0.2 E16g2MJ-2
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 E16g2MJ-1 16.0 118.7 76.7 15.4 0.7
G/GIN (-)
E16g2MJ-2 16.0 157.9 100.5 24.9 1.6

0.6
0.4
0.92/0.5/0.83 MPa MPa - - -
E16g2MJ-1
E16g2MJ-1 14.64 112.95 0.05 0.95 7.34
0.2 E16g2MJ-2 E16g2MJ-2 22.21 130.51 0.06 0.89 5.25
0
0.1 1 10 100
w/D (%)
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
E16g3MJ-1 E16g3MJ-1
0.2 E16g3MJ-2 0.2 E16g3MJ-2
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 E16g3MJ-1 16.0 118.0 120.7 27.4 1.4
G/GIN (-)
E16g3MJ-2 16.0 137.5 139.5 34.0 2.2

0.6
0.4
0.92/0.5/0.83 MPa MPa - - -
E16g3MJ-1
E16g3MJ-1 23.85 156.73 0.05 0.87 5.73
0.2 E16g3MJ-2 E16g3MJ-2 29.46 193.67 0.06 0.87 5.69
0
0.1 1 10 100
w/D (%)
B-17
B.2.6 Sample F
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
F6g2MJ-1 F6g2MJ-1
0.2 F6g2MJ-2 0.2 F6g2MJ-2
F6g2MJ-3 F6g2MJ-3
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
Summary Table for Test F6g2MJ
0.8 F6g2MJ-1 6.0 123.8 79.8 9.5 1.1
G/GIN (-)
F6g2MJ-2 6.0 163.6 104.0 18.0 1.7

0.6 F6g2MJ-3 6.0 183.5 116.8 22.5 2.4

0.4
0.92/0.5/0.83 MPa MPa - - -
F6g2MJ-1
F6g2MJ-1 9.02 81.76 0.12 0.95 8.60
0.2 F6g2MJ-2 F6g2MJ-2 16.12 118.09 0.10 0.89 6.55
F6g2MJ-3 F6g2MJ-3 20.26 129.23 0.11 0.90 5.74
0
0.1 1 10 100
w/D (%)
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
F6g3MJ-1 F6g3MJ-1
0.2 F6g3MJ-2 0.2 F6g3MJ-2
F6g3MJ-3 F6g3MJ-3
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 F6g3MJ-1 6.0 123.8 126.3 13.0 2.3
G/GIN (-)
F6g3MJ-2 6.0 163.6 164.6 22.3 2.9

0.6 F6g3MJ-3 6.0 183.5 184.9 29.3 3.1

0.4
0.92/0.5/0.83 MPa MPa - - -
F6g3MJ-1
F6g3MJ-1 12.35 110.34 0.18 0.95 8.52
0.2 F6g3MJ-2 F6g3MJ-2 20.59 167.82 0.13 0.92 7.54
F6g3MJ-3 F6g3MJ-3 26.28 182.77 0.11 0.90 6.24
0
0.1 1 10 100
w/D (%)
B-18
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
F9g2MJ-1 F9g2MJ-1
0.2 F9g2MJ-2 0.2 F9g2MJ-2
F9g2MJ-3 F9g2MJ-3
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 F9g2MJ-1 9.0 123.5 79.6 10.7 0.9
G/GIN (-)
F9g2MJ-2 9.0 163.3 103.8 21.6 1.7

0.6 F9g2MJ-3 9.0 183.1 116.5 26.7 1.7

0.4
0.92/0.5/0.83 MPa MPa - - -
F9g2MJ-1
F9g2MJ-1 8.86 71.16 0.08 0.83 6.66
0.2 F9g2MJ-2 F9g2MJ-2 17.45 113.01 0.08 0.81 5.23
F9g2MJ-3 F9g2MJ-3 20.95 118.22 0.06 0.78 4.43
0
0.1 1 10 100
w/D (%)
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
F9g3MJ-1 F9g3MJ-1
0.2 F9g3MJ-2 0.2 F9g3MJ-2
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 F9g3MJ-1 9.0 123.4 126.0 15.8 1.3
G/GIN (-)
F9g3MJ-2 9.0 163.2 164.2 26.6 1.8

0.6
0.4
0.92/0.5/0.83 MPa MPa - - -
F9g3MJ-1
F9g3MJ-1 13.10 113.82 0.08 0.83 7.19
0.2 F9g3MJ-2 F9g3MJ-2 21.57 139.72 0.07 0.81 5.26
0
0.1 1 10 100
w/D (%)
B-19
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
F16g2MJ-1 F16g2MJ-1
0.2 F16g2MJ-2 0.2 F16g2MJ-2
F16g2MJ-3 F16g2MJ-3
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 F16g2MJ-1 16.0 123.2 79.4 10.2 0.5
G/GIN (-)
F16g2MJ-2 16.0 162.6 103.3 18.4 1.1

0.6 F16g2MJ-3 16.0 182.2 115.9 23.1 1.4

0.4
0.92/0.5/0.83 MPa MPa - - -
F16g2MJ-1
F16g2MJ-1 9.51 78.23 0.05 0.94 7.70
0.2 F16g2MJ-2 F16g2MJ-2 16.46 117.54 0.06 0.90 6.40
F16g2MJ-3 F16g2MJ-3 20.82 145.39 0.06 0.90 6.29
0
0.1 1 10 100
w/D (%)
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
F16g3MJ-1 F16g3MJ-1
0.2 F16g3MJ-2 0.2 F16g3MJ-2
F16g3MJ-3 F16g3MJ-3
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 F16g3MJ-1 16.0 122.8 125.3 15.5 0.8
G/GIN (-)
F16g3MJ-2 16.0 162.0 163.0 27.2 1.4

0.6 F16g3MJ-3 16.0 181.5 182.9 33.8 2.2

0.4
0.92/0.5/0.83 MPa MPa - - -
F16g3MJ-1
F16g3MJ-1 14.21 126.75 0.05 0.92 8.20
0.2 F16g3MJ-2 F16g3MJ-2 23.98 169.75 0.05 0.88 6.24
F16g3MJ-3 F16g3MJ-3 30.39 221.41 0.06 0.90 6.54
0
0.1 1 10 100
w/D (%)
B-20
B.2.7 Sample G
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
G6g2MJ-1 G6g2MJ-1
0.2 G6g2MJ-2 0.2 G6g2MJ-2
G6g2MJ-3 G6g2MJ-3
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
Summary Table for Test G6g2MJ
0.8 G6g2MJ-1 6.0 123.8 80.3 9.9 1.4
G/GIN (-)
G6g2MJ-2 6.0 163.7 104.7 12.2 1.3

0.6 G6g2MJ-3 6.0 183.7 116.9 12.6 1.7

0.4
0.92/0.5 MPa MPa - - -
G6g2MJ-1
G6g2MJ-1 9.30 86.34 0.14 0.94 8.71
0.2 G6g2MJ-2 G6g2MJ-2 11.53 101.33 0.10 0.94 8.30
G6g2MJ-3 G6g2MJ-3 11.99 102.61 0.13 0.95 8.12
0
0.1 1 10 100
w/D (%)
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
G6g3MJ-1 G6g3MJ-1
0.2 G6g3MJ-2 0.2 G6g3MJ-2
G6g3MJ-3 G6g3MJ-3
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 G6g3MJ-1 6.0 123.8 127.1 15.9 1.9
G/GIN (-)
G6g3MJ-2 6.0 163.7 165.7 15.9 2.3

0.6 G6g3MJ-3 6.0 183.6 185.0 17.2 2.2

0.4
0.92/0.5 MPa MPa - - -
G6g3MJ-1
G6g3MJ-1 13.20 122.20 0.12 0.83 7.68
0.2 G6g3MJ-2 G6g3MJ-2 15.05 133.53 0.15 0.94 8.37
G6g3MJ-3 G6g3MJ-3 16.34 142.83 0.13 0.95 8.29
0
0.1 1 10 100
w/D (%)
B-21
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
G9g2MJ-1 G9g2MJ-1
0.2 G9g2MJ-2 0.2 G9g2MJ-2
G9g2MJ-3 G9g2MJ-3
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 G9g2MJ-1 9.0 123.6 80.2 10.2 1.2
G/GIN (-)
G9g2MJ-2 9.0 163.5 104.6 12.0 1.1

0.6 G9g2MJ-3 9.0 183.4 116.7 13.0 1.2

0.4
0.92/0.5 MPa MPa - - -
G9g2MJ-1
G9g2MJ-1 8.47 70.74 0.12 0.83 6.95
0.2 G9g2MJ-2 G9g2MJ-2 10.09 76.53 0.09 0.84 6.39
G9g2MJ-3 G9g2MJ-3 10.84 84.88 0.09 0.84 6.55
0
0.1 1 10 100
w/D (%)
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
G9g3MJ-1 G9g3MJ-1
0.2 G9g3MJ-2 0.2 G9g3MJ-2
G9g3MJ-3 G9g3MJ-3
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 G9g3MJ-1 9.0 123.4 126.7 17.1 1.2
G/GIN (-)
G9g3MJ-2 9.0 163.3 165.3 21.4 2.0

0.6 G9g3MJ-3 9.0 183.2 184.6 24.0 2.2

0.4
0.92/0.5 MPa MPa - - -
G9g3MJ-1
G9g3MJ-1 14.16 123.76 0.07 0.83 7.25
0.2 G9g3MJ-2 G9g3MJ-2 17.72 132.24 0.09 0.83 6.19
G9g3MJ-3 G9g3MJ-3 19.97 152.30 0.09 0.83 6.33
0
0.1 1 10 100
w/D (%)
B-22
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
G16g2MJ-1 G16g2MJ-1
0.2 G16g2MJ-2 0.2 G16g2MJ-2
G16g2MJ-3 G16g2MJ-3
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 G16g2MJ-1 16.0 123.1 79.9 10.3 0.7
G/GIN (-)
G16g2MJ-2 16.0 162.9 104.2 13.2 0.9

0.6 G16g2MJ-3 16.0 182.7 116.3 14.9 1.2

0.4
0.92/0.5 MPa MPa - - -
G16g2MJ-1
G16g2MJ-1 9.75 83.26 0.07 0.95 8.07
0.2 G16g2MJ-2 G16g2MJ-2 12.19 97.65 0.07 0.93 7.42
G16g2MJ-3 G16g2MJ-3 13.93 107.20 0.08 0.93 7.17
0
0.1 1 10 100
w/D (%)
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
G16g3MJ-1 G16g3MJ-1
0.2 G16g3MJ-2 0.2 G16g3MJ-2
G16g3MJ-3 G16g3MJ-3
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 G16g3MJ-1 16.0 122.8 126.2 15.2 0.9
G/GIN (-)
G16g3MJ-2 16.0 162.4 164.4 20.1 1.6

0.6 G16g3MJ-3 16.0 182.1 183.5 23.6 1.8

0.4
0.92/0.5 MPa MPa - - -
G16g3MJ-1
G16g3MJ-1 14.07 120.29 0.06 0.92 7.91
0.2 G16g3MJ-2 G16g3MJ-2 18.47 133.10 0.08 0.92 6.63
G16g3MJ-3 G16g3MJ-3 21.17 146.88 0.08 0.90 6.22
0
0.1 1 10 100
w/D (%)
B-23
B.2.8 Sample H
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
0.2 H6-1 0.2 H6-1
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
Summary Table for Test H6-1
1 g level D L σ'v0 qc qb,re
0.8 H6-1 6 154.1 129.7 2.1 1.1
G/GIN (-)
0.6
Dr qb0.1D GIN qb,re/qc qb0.1/qc GIN/qc

0.4
0.96 MPa MPa - - -
H6-1 2.0 34.4 0.50 0.96 16.16
0.2 H6-1
0
0.1 1 10 100
w/D (%)
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
0.2 H6-2 0.2 H6-2
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 H6-2 6 150.9 127.3 2.6 1.2
G/GIN (-)
0.6

0.4
0.96 MPa MPa - - -
H6-2 2.1 25.5 0.44 0.80 9.72
0.2 H6-2
0
0.1 1 10 100
w/D (%)
B-24
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
0.2 H6-3 0.2 H6-3
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 H6-3 6 146.0 123.2 4.3 1.8
G/GIN (-)
0.6

0.4
0.96 MPa MPa - - -
H6-3 3.1 57.3 0.43 0.74 13.48
0.2 H6-3
0
0.1 1 10 100
w/D (%)
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
0.2 H6-4 0.2 H6-4
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 H6-4 6 139.9 118.0 8.1 2.8
G/GIN (-)
0.6

0.4
0.96 MPa MPa - - -
H6-4 5.8 83.0 0.34 0.71 10.24
0.2 H6-4
0
0.1 1 10 100
w/D (%)
B-25
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
0.2 H6-5 0.2 H6-5
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 H6-5 6 194.5 154.6 1.1 1.0
G/GIN (-)
0.6

0.4
0.96 MPa MPa - - -
H6-5 1.4 10.8 0.88 1.27 9.58
0.2 H6-5
0
0.1 1 10 100
w/D (%)
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
0.2 H6-6 0.2 H6-6
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 H6-6 6 87.5 73.8 20.1 2.7
G/GIN (-)
0.6

0.4
0.96 MPa MPa - - -
H6-6 18.3 171.3 0.13 0.91 8.52
0.2 H6-6
0
0.1 1 10 100
w/D (%)
B-26
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
0.2 H16-1 0.2 H16-1
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 H16-1 6 181.5 146.6 1.3 0.8
G/GIN (-)
0.6

0.4
0.96 MPa MPa - - -
H16-1 1.5 11.6 0.58 1.13 8.77
0.2 H16-1
0
0.1 1 10 100
w/D (%)
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
0.2 H16-2 0.2 H16-2
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 H16-2 6 148.7 125.4 2.2 0.9
G/GIN (-)
0.6

0.4
0.96 MPa MPa - - -
H16-2 1.8 50.1 0.42 0.85 23.21
0.2 H16-2
0
0.1 1 10 100
w/D (%)
B-27
1.2 1.2
1 1
0.8 0.8
G/GIN (-)
qb/qc (-)
0.6 0.6
0.4 0.4
0.2 H16-3 0.2 H16-3
0 0
0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2
w/D (%) qb/qc (-)
1.2
0.8 H16-3 6 124.9 105.3 6.0 1.4
G/GIN (-)
0.6

0.4
0.96 MPa MPa - - -
H16-3 4.3 125.8 0.24 0.71 21.00
0.2 H16-3
0
0.1 1 10 100
w/D (%)
B-28
B.3 GROUPED STIFFNESS DECAY CURVES - SAMPLES A TO G
B.3.1 G/GIN plotted against qb/qc
B-29
1.2
1
0.8
G/GIN (-)
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2
qb/qc (-)

- - - MPa MPa kPa
1 C6g3MJ-1 4.74 112 23.7 115
2 C6g3MJ-3 4.64 200 43.2 178
3 C6g3J-2 4.81 202 42.0 157
4 C6g3J-3 3.80 182 47.7 177
5 C16g3MJ-2 4.39 137 31.3 155
6 C16g3J-2 4.91 180 36.8 155
7 D6g3J-3 4.71 214 45.4 186
8 D6g3MJ-3 4.44 214 48.1 186
9 D16g3J-2 4.88 150 30.6 154
10 D16g3MJ-2 4.68 159 34.0 154
11 E9g3MJ-1 4.32 131 30.3 122
12 F9g2MJ-3 4.43 118 26.7 116
B-30
1.2
1
0.8
G/GIN (-)
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2
qb/qc (-)

- - - MPa MPa kPa
1 C9g3MJ-1 5.48 156 28.4 115
2 C16g3MJ-1 5.19 106 20.4 114
3 C16g3J-1 5.35 114 21.3 114
4 D6g3J-1 5.62 164 29.2 124
5 D6g3J-2 5.01 183 36.6 155
6 D6g3MJ-1 5.54 175 31.6 124
7 D6g3MJ-2 5.31 215 40.5 155
8 D16g3J-1 5.07 120 23.7 123
9 D16g3MJ-1 5.16 129 25.1 123
10 E6g2MJ-1 5.93 115 19.4 77
11 E6g2MJ-2 5.42 158 29.2 101
12 E6g3MJ-2 5.51 221 40.2 161
13 E9g2MJ-1 5.96 113 18.9 77
14 E16g2MJ-2 5.25 131 24.9 101
15 E16g3MJ-1 5.73 157 27.4 121
16 E16g3MJ-2 5.69 194 34.0 139
17 F6g2MJ-3 5.74 129 22.5 117
18 F9g2MJ-2 5.23 113 21.6 104
19 F9g3MJ-2 5.26 140 26.6 164
B-31
1.2
1
0.8
G/GIN (-)
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2
qb/qc (-)

- - - MPa MPa kPa
1 A9g3MJ-2 6.18 118 19.2 174
2 C6g3MJ-2 6.30 232 36.9 157
3 C6g3J-1 6.11 146 23.8 115
4 C9g3J-1 6.28 167 26.7 115
5 C9g3J-2 6.02 192 31.8 125
6 D9g3J-1 6.07 206 34.0 124
7 D9g3MJ-1 6.53 213 32.6 124
8 E6g3MJ-1 6.86 211 30.7 122
9 E9g2MJ-2 6.18 171 27.7 101
10 F6g2MJ-2 6.55 118 18.0 104
11 F6g3MJ-3 6.24 183 29.3 185
12 F9g2MJ-1 6.66 71 10.7 80
13 F16g2MJ-2 6.40 118 18.4 103
14 F16g2MJ-3 6.29 145 23.1 116
15 F16g3MJ-2 6.24 170 27.2 163
16 F16g3MJ-3 6.54 221 33.8 183
17 G9g2MJ-1 6.95 71 10.2 80
18 G9g2MJ-2 6.39 77 12.0 105
19 G9g2MJ-3 6.55 85 13.0 117
20 G9g3MJ-2 6.19 132 21.4 165
21 G9g3MJ-3 6.33 152 24.0 185
22 G16g3MJ-2 6.63 133 20.1 164
23 G16g3MJ-3 6.22 147 23.6 184
B-32
1.2
1
0.8
G/GIN (-)
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2
qb/qc (-)

- - - MPa MPa kPa
1 A9g2MJ-1 7.77 94 12.2 110
2 A9g3MJ-1 7.69 96 12.5 110
3 E16g2MJ-1 7.34 113 15.4 77
4 F6g3MJ-2 7.54 168 22.3 165
5 F9g3MJ-1 7.19 114 15.8 126
6 F16g2MJ-1 7.70 78 10.2 79
7 G6g3MJ-1 7.68 122 15.9 127
8 G9g3MJ-1 7.25 124 17.1 127
9 G16g2MJ-2 7.42 98 13.2 104
10 G16g2MJ-3 7.17 107 14.9 116
11 G16g3MJ-1 7.91 120 15.2 126
B-33
1.2
1
0.8
G/GIN (-)
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2
qb/qc (-)

- - - MPa MPa kPa
1 A6g3MJ-2 8.56 138 16.2 179
2 F6g2MJ-1 8.60 82 9.5 80
3 F6g3MJ-1 8.52 110 13.0 126
4 F16g3MJ-1 8.20 127 15.5 125
5 G6g2MJ-1 8.71 86 9.9 80
6 G6g2MJ-2 8.30 101 12.2 105
7 G6g2MJ-3 8.12 103 12.6 117
8 G6g3MJ-2 8.37 134 15.9 166
9 G6g3MJ-3 8.29 143 17.2 185
10 G16g2MJ-1 8.07 83 10.3 80
B-34
1.2
1
0.8
G/GIN (-)
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2
qb/qc (-)

- - - MPa MPa kPa
1 A6g2MJ-1 9.74 118 12.1 110
2 A6g3MJ-1 9.43 105 11.1 114
3 A16g1MJ-1 9.84 70 7.1 65
4 A16g3MJ-1 9.99 99 10.0 113
5 A16g3MJ-2 9.38 159 17.0 182
B-35
1.2
1
Measured
0.8
G/GIN (-)
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2
qb/qc (-)

- - - MPa MPa kPa
1 B9g3MJ-2 12.19 125 10.3 169
2 B9g3J-2 12.20 126 10.3 169
3 B16g3MJ-2 12.94 108 8.3 169
4 B16g3J-2 13.58 132 9.7 169
B-36
1.2
1
0.8
G/GIN (-)
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2
qb/qc (-)

- - - MPa MPa kPa
1 B6g3J-1 15.24 106 7.0 169
2 B9g3MJ-1 15.00 97 6.5 112
3 B9g3J-1 14.62 95 6.5 113
4 B16g3J-1 14.90 88 5.9 112
B-37
1.2
1
0.8
G/GIN (-)
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2
qb/qc (-)

- - - MPa MPa kPa
1 B16g3MJ-1 16.43 87 5.3 112
B-38
1.2
1
0.8
G/GIN (-)
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2
qb/qc (-)

- - - MPa MPa kPa
1 B6g3MJ-2 21.03 99 4.7 169
B-39
1.2
1
0.8
G/GIN (-)
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2
qb/qc (-)

- - - MPa MPa kPa
1 B6g3MJ-1 29.52 87 2.9 113
B-40
B.3.2 G/GIN plotted against w/D
B-41
1.2
1
0.8
G/GIN (-)
0.6
0.4
0.2
0
0.1 1 10 100
w/D (%)

- - - MPa MPa kPa
1 C6g3MJ-1 4.74 112 23.7 115
2 C6g3MJ-3 4.64 200 43.2 178
3 C6g3J-2 4.81 202 42.0 157
4 C6g3J-3 3.80 182 47.7 177
5 C16g3MJ-2 4.39 137 31.3 155
6 C16g3J-2 4.91 180 36.8 155
7 D6g3J-3 4.71 214 45.4 186
8 D6g3MJ-3 4.44 214 48.1 186
9 D16g3J-2 4.88 150 30.6 154
10 D16g3MJ-2 4.68 159 34.0 154
11 E9g3MJ-1 4.32 131 30.3 122
12 F9g2MJ-3 4.43 118 26.7 116
B-42
1.2
1
0.8
G/GIN (-)
0.6
0.4
0.2
0
0.1 1 10 100
w/D (%)

- - - MPa MPa kPa
1 C9g3MJ-1 5.48 156 28.4 115
2 C16g3MJ-1 5.19 106 20.4 114
3 C16g3J-1 5.35 114 21.3 114
4 D6g3J-1 5.62 164 29.2 124
5 D6g3J-2 5.01 183 36.6 155
6 D6g3MJ-1 5.54 175 31.6 124
7 D6g3MJ-2 5.31 215 40.5 155
8 D16g3J-1 5.07 120 23.7 123
9 D16g3MJ-1 5.16 129 25.1 123
10 E6g2MJ-1 5.93 115 19.4 77
11 E6g2MJ-2 5.42 158 29.2 101
12 E6g3MJ-2 5.51 221 40.2 161
13 E9g2MJ-1 5.96 113 18.9 77
14 E16g2MJ-2 5.25 131 24.9 101
15 E16g3MJ-1 5.73 157 27.4 121
16 E16g3MJ-2 5.69 194 34.0 139
17 F6g2MJ-3 5.74 129 22.5 117
18 F9g2MJ-2 5.23 113 21.6 104
19 F9g3MJ-2 5.26 140 26.6 164
B-43
1.2
1
0.8
G/GIN (-)
0.6
0.4
0.2
0
0.1 1 10 100
w/D (%)

- - - MPa MPa kPa
1 A9g3MJ-2 6.18 118 19.2 174
2 C6g3MJ-2 6.30 232 36.9 157
3 C6g3J-1 6.11 146 23.8 115
4 C9g3J-1 6.28 167 26.7 115
5 C9g3J-2 6.02 192 31.8 125
6 D9g3J-1 6.07 206 34.0 124
7 D9g3MJ-1 6.53 213 32.6 124
8 E6g3MJ-1 6.86 211 30.7 122
9 E9g2MJ-2 6.18 171 27.7 101
10 F6g2MJ-2 6.55 118 18.0 104
11 F6g3MJ-3 6.24 183 29.3 185
12 F9g2MJ-1 6.66 71 10.7 80
13 F16g2MJ-2 6.40 118 18.4 103
14 F16g2MJ-3 6.29 145 23.1 116
15 F16g3MJ-2 6.24 170 27.2 163
16 F16g3MJ-3 6.54 221 33.8 183
17 G9g2MJ-1 6.95 71 10.2 80
18 G9g2MJ-2 6.39 77 12.0 105
19 G9g2MJ-3 6.55 85 13.0 117
20 G9g3MJ-2 6.19 132 21.4 165
21 G9g3MJ-3 6.33 152 24.0 185
22 G16g3MJ-2 6.63 133 20.1 164
23 G16g3MJ-3 6.22 147 23.6 184
B-44
1.2
1
0.8
G/GIN (-)
0.6
0.4
0.2
0
0.1 1 10 100
w/D (%)

- - - MPa MPa kPa
1 A9g2MJ-1 7.77 94 12.2 110
2 A9g3MJ-1 7.69 96 12.5 110
3 E16g2MJ-1 7.34 113 15.4 77
4 F6g3MJ-2 7.54 168 22.3 165
5 F9g3MJ-1 7.19 114 15.8 126
6 F16g2MJ-1 7.70 78 10.2 79
7 G6g3MJ-1 7.68 122 15.9 127
8 G9g3MJ-1 7.25 124 17.1 127
9 G16g2MJ-2 7.42 98 13.2 104
10 G16g2MJ-3 7.17 107 14.9 116
11 G16g3MJ-1 7.91 120 15.2 126
B-45
1.2
1
0.8
G/GIN (-)
0.6
0.4
0.2
0
0.1 1 10 100
w/D (%)

- - - MPa MPa kPa
1 A6g3MJ-2 8.56 138 16.2 179
2 F6g2MJ-1 8.60 82 9.5 80
3 F6g3MJ-1 8.52 110 13.0 126
4 F16g3MJ-1 8.20 127 15.5 125
5 G6g2MJ-1 8.71 86 9.9 80
6 G6g2MJ-2 8.30 101 12.2 105
7 G6g2MJ-3 8.12 103 12.6 117
8 G6g3MJ-2 8.37 134 15.9 166
9 G6g3MJ-3 8.29 143 17.2 185
10 G16g2MJ-1 8.07 83 10.3 80
B-46
1.2
1
0.8
G/GIN (-)
0.6
0.4
0.2
0
0.1 1 10 100
w/D (%)

- - - MPa MPa kPa
1 A6g2MJ-1 9.74 118 12.1 110
2 A6g3MJ-1 9.43 105 11.1 114
3 A16g1MJ-1 9.84 70 7.1 65
4 A16g3MJ-1 9.99 99 10.0 113
5 A16g3MJ-2 9.38 159 17.0 182
B-47
1.2
1
0.8
G/GIN (-)
0.6
0.4
0.2
0
0.1 1 10 100
w/D (%)

- - - MPa MPa kPa
1 B9g3MJ-2 12.19 125 10.3 169
2 B9g3J-2 12.20 126 10.3 169
3 B16g3MJ-2 12.94 108 8.3 169
4 B16g3J-2 13.58 132 9.7 169
B-48
1.2
1
0.8
G/GIN (-)
0.6
0.4
0.2
0
0.1 1 10 100
w/D (%)

- - - MPa MPa kPa
1 B6g3J-1 15.24 106 7.0 169
2 B9g3MJ-1 15.00 97 6.5 112
3 B9g3J-1 14.62 95 6.5 113
4 B16g3J-1 14.90 88 5.9 112
B-49
1.2
1
0.8
G/GIN (-)
0.6
0.4
0.2
0
0.1 1 10 100
w/D (%)

- - - MPa MPa kPa
1 B16g3MJ-1 16.43 87 5.3 112
B-50
1.2
1
0.8
G/GIN (-)
0.6
0.4
0.2
0
0.1 1 10 100
w/D (%)

- - - MPa MPa kPa
1 B6g3MJ-2 21.03 99 4.7 169
B-51
1.2
1
0.8
G/GIN (-)
0.6
0.4
0.2
0
0.1 1 10 100
w/D (%)

- - - MPa MPa kPa
1 B6g3MJ-1 29.52 87 2.9 113
B-52
APPENDIX C NUMERICAL ANALYSIS RESULTS
C.1 INTRODUCTION
As discussed in Chapter 7, the penetration of a pile in layered soil was analyzed by

assuming the spherical cavity expansion analogue using the FEM code PLAXIS. In
total, 19 cases (as summarized in Table 7.5 or Table C.1 below) were analyzed
involving a wide range of resistance ratios (A0=qb,W/qb,S) between the strong and
weak soil layers. This Appendix presents the results of pressure expansion (p vs. u)
and resistance reduction (qb/qb,s vs. H/D) curves for each case on Figures C.1 to C.19.
Also summarized are the fitted curves to the inferred resistance reduction curves.
Table C.1 Analysis program in a two-layer soil system
Group No Upper Layer Lower layer

E50ref φ ψ Type E50ref φ ψ Type
Group3 10 5 20 0 Undrained 50 40 10 Drained
16 5 20 0 Undrained 100 42 12 Drained
Group5 19* 5 - - Undrained 100 42 12 Drained
* the undrained shear strength, su=6kPa.
C.2 PRESSURE EXPANSION AND RESISTANCE REDUCTION

CURVES
C-1
Appendix C Numerical Analysis Results
4000
12
3500 3
2
3000 1
Hcavity (m)= 0.5
2500 0.2
0
p (kPa)
-0.2
2000 -0.5
-3
1500 -12
1000
500 decrease in Hcavity

0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
u (m)
1.2
1
Inferred
0.8 Fitted
qb/qb,S (-)
0.6
0.4
0.2
0
-10 -5 0 5 10 15 20
H/D (-)
φ=
o
Upper Layer: Weak 40
φ=
o
Lower Layer: Strong 40
Hcavity u plimit H/D qb/qb,S
m m kPa - -
12 0.2124 3623 24.6 1.00
3 0.1555 3526 7.2 0.97
2 0.1100 3447 5.8 0.95
1 0.0570 2992 3.7 0.83
0.5 0.0539 2448 1.7 0.68
0.2 0.0784 1948 0.3 0.54
0 0.1306 1653 -0.4 0.46
-0.2 0.2000 1518 -0.9 0.42
-0.5 0.2000 1317 -1.5 0.36
-3 0.2000 1235 -6.9 0.34
-12 0.2000 1224 -26.5 0.34
A0= 0.34
qb/qb,S=A0+(1-A0)exp[-exp(A1+A2×H/D)] A1= 0.30
A2= -0.65
Figure C.1 Pressure expansion and resistance reduction curves for case No1
C-2
4000
12
3500 3
2
3000 1
Hcavity (m)= 0.5
2500 0.2
p (kPa) 0
-0.2
2000 -0.5
-3
1500 -12
1000

0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
u (m)
1.2
1
Inferred
0.8 Fitted
qb/qb,S (-)
0.6
0.4
0.2
0
-10 -5 0 5 10 15 20
H/D (-)
φ=
o
φ=
o
m m kPa - -
12 0.2124 3623 24.6 1.00
3 0.1582 3561 7.2 0.98
2 0.1306 3484 5.2 0.96
1 0.0931 3177 3.0 0.88
0.5 0.0835 2809 1.4 0.78
0.2 0.0797 2435 0.3 0.67
0 0.1352 2170 -0.4 0.60
-0.2 0.1767 2013 -0.9 0.56
-0.5 0.2395 1872 -1.4 0.52
-3 0.1591 1731 -8.0 0.48
-12 0.1693 1680 -29.5 0.46
A0= 0.46
A2= -0.70
C-3
4000
12
3500 3
1
3000 0.5
Hcavity (m)= 0.2
0
2500 -0.2
p (kPa)
-0.5
2000 -3
-12
1500
1000

0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
u (m)
1.2
1
Inferred
0.8 Fitted
qb/qb,S (-)
0.6
0.4
0.2
0
-10 -5 0 5 10 15 20
H/D (-)
φ=
o
φ=
o
m m kPa - -
12 0.2124 3623 24.6 1.00
3 0.1333 3572 8.0 0.99
1 0.1199 3443 2.5 0.95
0.5 0.1310 3322 1.0 0.92
0.2 0.0983 3174 0.2 0.88
0 0.1627 3056 -0.4 0.84
-0.2 0.1383 2979 -1.0 0.82
-0.5 0.1468 2809 -1.7 0.78
-3 0.1437 2703 -8.5 0.75
-12 0.2198 2680 -24.9 0.74
A0= 0.74
A2= -0.73
C-4
4000
12
3500 3
2
3000 1
Hcavity (m)= 0.5
2500 0.2
p (kPa) 0
-0.2
2000 -0.5
-3
1500 -12
1000

0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
u (m)
1.2
1
Inferred
0.8 Fitted
qb/qb,S (-)
0.6
0.4
0.2
0
-10 -5 0 5 10 15 20
H/D (-)
φ=
o
φ=
o
m m kPa - -
12 0.2124 3623 20.3 1.00
3 0.1335 3602 6.7 0.99
2 0.1337 3592 4.4 0.99
1 0.1096 3411 2.4 0.94
0.5 0.0750 3001 1.3 0.83
0.2 0.0602 2434 0.5 0.67
0 0.1145 2019 -0.3 0.40
-0.2 0.1327 1513 -1.0 0.23
-0.5 0.1073 1324 -2.0 0.20
-3 0.1341 1274 -8.8 0.19
-12 0.2354 1224 -23.8 0.18
A0= 0.18
A2= -0.65
C-5
4000
12
3500 3
2
3000 1
Hcavity (m)= 0.5
2500 0.2
p (kPa)
0
-0.2
2000 -0.5
-3
1500 -12
1000

0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
u (m)
1.2
1
Inferred
0.8 Fitted
qb/qb,S (-)
0.6
0.4
0.2
0
-10 -5 0 5 10 15 20
H/D (-)
φ=
o
φ=
o
m m kPa - -
12 0.2124 3623 21.9 1.00
3 0.1325 3593 7.2 0.99
2 0.1346 3534 4.6 0.98
1 0.1278 3531 2.2 0.97
0.5 0.0875 3377 1.3 0.93
0.2 0.1002 3126 0.3 0.86
0 0.1110 2918 -0.4 0.67
-0.2 0.1409 2641 -1.0 0.52
-0.5 0.1360 2445 -1.8 0.48
-3 0.1774 2412 -7.5 0.48
-12 0.3110 2287 -19.5 0.45
A0= 0.45
A2= -0.75
C-6
4000
12
3500 3
2
3000 1
Hcavity (m)= 0.5
p (kPa)2500 0.2
0
-0.2
2000 -0.5
-3
1500 -12
1000

0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
u (m)
1.2
1
Inferred
0.8 Fitted
qb/qb,S (-)
0.6
0.4
0.2
0
-10 -5 0 5 10 15 20
H/D (-)
φ=
o
φ=
o
m m kPa - -
12 0.2124 3623 20.3 1.00
3 0.1353 3548 6.6 0.98
2 0.1062 3486 5.0 0.96
1 0.0604 2945 3.1 0.81
0.5 0.0449 2236 1.7 0.62
0.2 0.0360 1611 0.6 0.44
0 0.1141 1024 -0.3 0.20
-0.2 0.2000 767 -0.9 0.11
-0.5 0.2000 709 -1.5 0.11
-3 0.2000 599 -6.9 0.09
-12 0.2000 580 -26.5 0.09
A0= 0.09
A2= -0.50
C-7
4000
12
3500 3
2
3000 1
Hcavity (m)= 0.5
2500 0.2
p (kPa)
0
-0.2
2000 -0.5
-3
1500 -12
1000

0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
u (m)
1.2
1
Inferred
0.8 Fitted
qb/qb,S (-)
0.6
0.4
0.2
0
-10 -5 0 5 10 15 20
H/D (-)
φ=
o
φ=
o
m m kPa - -
12 0.2124 3623 21.9 1.00
3 0.1278 3531 7.3 0.97
2 0.1085 3467 5.3 0.96
1 0.0597 2983 3.3 0.82
0.5 0.0404 2359 1.8 0.65
0.2 0.0474 1737 0.5 0.48
0 0.1195 1338 -0.4 0.31
-0.2 0.2000 1134 -0.9 0.22
-0.5 0.2000 1092 -1.5 0.22
-3 0.2000 895 -6.9 0.18
-12 0.2000 875 -26.5 0.17
A0= 0.17
A2= -0.63
C-8
6000
12
3
5000 2
1.5
4000 Hcavity (m)= 1
p (kPa) 0.5
0.2
3000 0
-0.5
-1.5
2000 -3
-12
1000
decrease in Hcavity
0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
u (m)
1.2
1
Inferred
0.8 Fitted
qb/qb,S (-)
0.6
0.4
0.2
0
-10 -5 0 5 10 15 20
H/D (-)
φ=
o
φ=
o
Hcavity u plimit
H/D qb/qb,S
m m kPa- -
12 0.2389 5550
22.7 1.00
3 0.2531 5499
5.1 0.99
2 0.2221 5469
3.6 0.99
1.5 0.2228 5459
2.6 0.98
1 0.1613 5406
2.1 0.97
0.5 0.2136 5370
0.6 0.97
0.2 0.1347 5037
0.1 0.91
0 0.2081 4766
-0.4 0.86
-0.5 0.2079 4507
-1.5 0.81
-1.5 0.3285 4495
-2.7 0.81
-3 0.2381 4468
-6.2 0.81
-12 0.3106 4417
-19.5 0.80 0.80
A0= 0.80
A2= -0.80
C-9
6000
12
3
5000 2
Hcavity (m)= 1.5
4000 1
0.5
p (kPa)
0.2
3000 0
-0.5
-1.5
2000 -3
-12
1000
decrease in Hcavity
0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
u (m)
1.2
1
Inferred
0.8 Fitted
qb/qb,S (-)
0.6
0.4
0.2
0
-10 -5 0 5 10 15 20
H/D (-)
φ=
o
φ=
o
Hcavity u plimit
H/D qb/qb,S
m m kPa- -
12 0.2389 5550
22.7 1.00
3 0.2389 5550
5.4 1.00
2 0.2600 5490
3.2 0.99
1.5 0.2569 5446
2.3 0.98
1 0.2380 5407
1.5 0.97
0.5 0.1630 5194
0.8 0.94
0.2 0.1316 4704
0.1 0.85
0 0.1748 4257
-0.4 0.77
-0.5 0.2311 3788
-1.4 0.68
-1.5 0.2263 3642
-3.4 0.66
-3 0.1894 3605
-7.2 0.65
-12 0.1894 3605
-27.5 0.65 0.65
A0= 0.65
A2= -0.75
C-10
6000
12
3
5000 2
Hcavity (m)= 1.5
4000 1
p (kPa) 0.5
0.2
3000 0
-0.2
-0.5
2000 -2
-12
1000
decrease in Hcavity
0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
u (m)
1.2
1
Inferred
0.8 Fitted
qb/qb,S (-)
0.6
0.4
0.2
0
-10 -5 0 5 10 15 20
H/D (-)
φ=
o
φ=
o
Hcavity u plimit
H/D qb/qb,S
m m kPa- -
12 0.2063 5412
20.7 1.00
3 0.1554 5169
6.1 0.96
2 0.0891 4916
5.4 0.91
1.5 0.0771 4488
4.3 0.83
1 0.0694 3936
3.0 0.73
0.5 0.0398 2880
1.7 0.53
0.2 0.0339 1793
0.6 0.33
0 0.0559 801-0.3 0.11
-0.2 0.2000 467-0.9 0.05
-0.5 0.2000 417-1.5 0.04
-2 0.2000 408-4.8 0.04
-12 0.2000 398-26.5 0.04
A0= 0.04
A2= -0.48
C-11
6000
12
5000 2.5
1.5
Hcavity (m)= 1
4000
0.5
0.2
0
p (kPa)
-0.2
3000 -0.5
-2
-12
2000
1000
decrease in Hcavity
0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
u (m)
1.2
1
Inferred
0.8 Fitted
qb/qb,S (-)
0.6
0.4
0.2
0
-10 -5 0 5 10 15 20
H/D (-)
φ=
o
Upper Layer: Strong 40
φ=
o
Lower Layer: Weak 20
m m kPa - -
12 0.2063 5412 26.0 1.00
2.5 0.1490 5191 7.0 0.96
1.5 0.0835 4750 5.8 0.88
1 0.0814 4209 4.0 0.78
0.5 0.0448 3007 2.7 0.56
0.2 0.0424 1760 1.3 0.33
0 0.0528 778 0.3 0.11
-0.2 0.2000 539 -0.2 0.05
-0.5 0.2000 542 -0.7 0.05
-2 0.2000 439 -3.4 0.04
-12 0.2000 398 -21.1 0.04
A0= 0.04
A2= -0.45
C-12
6000
12
3
5000 1
0.5
Hcavity (m)= 0.2
4000 0
p (kPa) -0.2
-0.5
3000 -2
-12
2000
1000
decrease in Hcavity
0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
u (m)
1.2
1
Inferred
0.8 Fitted
qb/qb,S (-)
0.6
0.4
0.2
0
-10 -5 0 5 10 15 20
H/D (-)
φ=
o
φ=
o
m m kPa - -
12 0.2063 5412 23.0 1.00
3 0.1290 5320 7.5 0.98
1 0.1319 4987 2.2 0.92
0.5 0.0800 4500 1.3 0.83
0.2 0.0828 3728 0.3 0.69
0 0.1460 3154 -0.4 0.51
-0.2 0.1491 2663 -0.9 0.39
-0.5 0.1424 2392 -1.8 0.35
-2 0.1166 2307 -6.4 0.33
-12 0.1518 2300 -31.5 0.33
A0= 0.33
A2= -0.60
C-13
6000
12
2
5000 1
0.5
Hcavity (m)= 0.2
4000 0
-0.2
p (kPa)
-0.5
3000 -3
-12
2000
1000
decrease in Hcavity
0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
u (m)
1.2
1
Inferred
0.8 Fitted
qb/qb,S (-)
0.6
0.4
0.2
0
-10 -5 0 5 10 15 20
H/D (-)
φ=
o
φ=
o
m m kPa - -
12 0.2063 5412 26.0 1.00
2 0.1095 5264 6.7 0.97
1 0.1048 4932 3.6 0.91
0.5 0.0957 4573 2.1 0.85
0.2 0.0825 3826 1.1 0.71
0 0.1427 3212 0.4 0.53
-0.2 0.2050 2854 -0.1 0.41
-0.5 0.1608 2467 -0.8 0.36
-3 0.1568 2297 -6.6 0.33
-12 0.1518 2300 -28.1 0.33
A0= 0.33
A2= -0.60
C-14
2500
12
3
2000 2
Hcavity (m)= 1.5
1
p (kPa)1500 0.5
0.2
0
-0.5
1000 -1.5
-3
-12
500
decrease in Hcavity
0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
u (m)
1.2
1
Inferred
0.8 Fitted
qb/qb,S (-)
0.6
0.4
0.2
0
-10 -5 0 5 10 15 20
H/D (-)
φ=
o
φ=
o
Hcavity u plimit
H/D qb/qb,S
m m kPa- -
12 0.3181 2351
15.1 1.00
3 0.1934 2281
5.3 0.97
2 0.1159 2238
4.7 0.95
1.5 0.0911 2166
4.0 0.92
1 0.0722 2012
2.9 0.86
0.5 0.0448 1663
1.7 0.71
0.2 0.0654 1247
0.5 0.53
0 0.1646 963-0.2 0.34
-0.5 0.2000 657-1.3 0.19
-1.5 0.2000 618-3.3 0.18
-3 0.2000 611-6.3 0.18
-12 0.2000 607-24.2 0.18
A0= 0.18
A2= -0.60
C-15
12000
12
6
10000 3
Hcavity (m)= 2
8000 1
0.5
p (kPa)
0.2
6000 0
-0.5
-1.5
4000 -3
-12
2000
decrease in Hcavity
0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
u (m)
1.2
0.8 Inferred
qb/qb,S (-)
Fitted
0.6
0.4
0.2
0
-10 -5 0 5 10 15 20 25
H/D (-)
φ=
o
φ=
o
Hcavity u plimit
H/D qb/qb,S
m m kPa- -
12 0.2018 10140
21.0 1.00
6 0.1667 10010
11.8 0.99
3 0.0884 9035
8.3 0.89
2 0.0588 7998
6.5 0.79
1 0.0312 5568
3.9 0.55
0.5 0.0302 3564
1.9 0.35
0.2 0.0404 2036
0.6 0.20
0 0.0628 1156
-0.3 0.08
-0.5 0.2000 741-1.6 0.04
-1.5 0.2000 630-3.8 0.03
-3 0.2000 610-7.2 0.03
-12 0.2000 607-27.4 0.03
A0= 0.03
A2= -0.32
C-16
12000
12
10000 6
3
Hcavity (m)= 2
8000 1
p (kPa) 0.5
0.2
6000 0
-0.5
-1.5
4000 -3
-12
2000
decrease in Hcavity
0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
u (m)
1.2
1
Inferred
0.8 Fitted
qb/qb,S (-)
0.6
0.4
0.2
0
-10 -5 0 5 10 15 20 25
H/D (-)
φ=
o
φ=
o
Hcavity u plimit
H/D qb/qb,S
m m kPa- -
12 0.2018 10140
21.0 1.00
6 0.1712 9997
11.6 0.99
3 0.0905 9027
8.2 0.89
2 0.0583 8060
6.5 0.79
1 0.0457 5871
3.5 0.58
0.5 0.0280 3716
1.9 0.37
0.2 0.0286 2034
0.6 0.20
0 0.0516 881-0.3 0.06
-0.5 0.2000 430-1.6 0.02
-1.5 0.2000 410-3.8 0.02
-3 0.2000 407-7.2 0.02
-12 0.2000 403-27.4 0.02
A0= 0.02
A2= -0.32
C-17
12000
12
6
10000 3
Hcavity (m)= 2
8000 1
0.5
p (kPa)
0.2
6000 0
-0.5
-1.5
4000 -3
-12
2000
decrease in Hcavity
0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
u (m)
1.2
Inferred
0.8
qb/qb,S (-)
Fitted
0.6
0.4
0.2
0
-10 0 10 20 30
H/D (-)
φ=
o
φ=
o
Hcavity u plimit
H/D qb/qb,S
m m kPa
- -
12 0.2018 10140
27.2 1.00
6 0.1812 10229
14.8 1.01
3 0.0986 9383
10.6 0.93
2 0.0900 8436
7.5 0.83
1 0.0439 6023
5.1 0.59
0.5 0.0379 3732
2.9 0.37
0.2 0.0272 2117
1.5 0.21
0 0.1000 1150
0.3 0.08
-0.5 0.2000 830
-0.7 0.04
-1.5 0.2000 658
-2.5 0.03
-3 0.2000 611
-5.1 0.03
-12 0.2000 607
-21.1 0.03
A0= 0.03
A2= -0.32
C-18
12000
12
6
10000 3
Hcavity (m)= 2
8000 1
p (kPa) 0.5
0.2
6000 0
-0.5
-1.5
4000 -3
-12
2000
decrease in Hcavity
0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
u (m)
1.2
1
Inferred
0.8 Fitted
qb/qb,S (-)
0.6
0.4
0.2
0
-10 0 10 20 30
H/D (-)
φ=
o
φ=
o
Hcavity u plimit
H/D qb/qb,S
m m kPa
- -
12 0.2018 10140
27.2 1.00
6 0.2062 10084
13.6 0.99
3 0.1240 9362
9.5 0.92
2 0.0669 8534
8.5 0.84
1 0.0539 6328
4.8 0.62
0.5 0.0261 3889
3.1 0.38
0.2 0.0272 2117
1.5 0.21
0 0.0500 842
0.3 0.06
-0.5 0.2000 565
-0.7 0.03
-1.5 0.2000 451
-2.5 0.02
-3 0.2000 418
-5.1 0.02
-12 0.2000 403
-21.1 0.02
A0= 0.02
A2= -0.32
C-19
12000
12
6
10000 3
Hcavity (m)= 2
8000 1
0.5
p (kPa)
0.2
6000 0
-0.5
-1.5
4000 -3
-12
2000
decrease in Hcavity
0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
u (m)
1.2
0.8 Inferred
qb/qb,S (-)
Fitted
0.6
0.4
0.2
0
-10 0 10 20 30
H/D (-)
φ=
o
Upper Layer: Weak 0
φ=
o
Hcavity u plimit
H/D qb/qb,S
m m kPa
- -
12 0.2018 10140
19.9 1.00
6 0.1751 9931
10.9 0.98
3 0.0863 9286
8.1 0.92
2 0.0576 8113
6.3 0.80
1 0.0290 5269
3.9 0.52
0.5 0.0179 2721
2.1 0.27
0.2 0.0392 1191
0.7 0.12
0 0.0362 461
-0.2 0.02
-0.5 0.2000 186
-1.6 0.01
-1.5 0.2000 177
-3.8 0.01
-3 0.2000 176
-7.2 0.01
-12 0.2000 173
-27.4 0.01
A0= 0.01
A2= -0.32
C-20
APPENDIX D PILE PERFORMANCE IN CLAY
D.1 INTRODUCTION
As discussed in Chapter 2, the capacity of a displacement pile is strongly influenced

by the degree of soil displacement induced during installation. Apart from the main
theme of this Thesis, this effect was also explored for piles in clay using a field and
centrifuge testing programme. In the field test near Shanghai, the displacements and
stresses were measured in the vicinity of two pipe piles during their installation in
soft lightly overconsolidated clay. The centrifuge experiments were conducted in the
drum centrifuge at UWA and involved tests on a series of model piles to study the
axial performance of single and group displacement piles in clay. This work was
conducted with some collaboration from the Geotechnical Institute of Hohai
University (GeoHohai) in Nanjing, China. Two papers have been published and are
provided in this Appendix.
D.2 PIPE PILE INSTALLATION EFFECTS IN SOFT CLAY
Reference: Xu, X., Liu, H. L., and Lehane, B. M. 2006. Pipe pile installation effects
in soft clay. Proceedings of the Institution of Civil Engineers, Geotechnical
Engineering, 159(GE4): 285-296.
D.3 CENTRIFUGE STUDIES OF SINGLE AND GROUP

DISPLACEMENT PILES IN CLAY
Reference: Xu, X., Lehane, B. M., Gaudin, C., Zhang, T., and Liu, H. L. 2006.
Centrifuge studies of single and group displacement piles in clay. In Proceedings of
the 6th International Conference on Physical Modelling in Geotechnics, Hongkong.
pp. 895-900.
D-1
Physical Modelling in Geotechnics – 6th ICPMG ’06 – Ng, Zhang & Wang (eds)
© 2006 Taylor & Francis Group, London, ISBN 0-415-41586-1
Centrifuge studies of single and group displacement piles in clay
X. Xu, B.M. Lehane & C. Gaudin

University of Western Australia, Perth, Australia
T. Zhang & H.L. Liu

Geotechnical Research Institute of Hohai University, NanJing, China P.R.
ABSTRACT: A series of model pile tests has been performed using the University of Western Australia drum
centrifuge to study the axial performance of single and group displacement piles in clay. Open and closed-ended
piles were jacked into kaolin clay with an overconsolidation ratio of 2. Each group comprised five piles at spacing
ratios, s/D, of 2, 3, & 4 and was subjected to compressive loading followed by tensile loading. Single piles were
also installed and tested in the same way as the group piles to provide a reference for comparison of the pile
groups’ performance. The results provide useful insights into the significant influence of pile tip area ratio (Ar )
and spacing ratio (s/D) on the stiffness and capacity of single and group piles in clay.
1 INTRODUCTION 1 to 2 m, with wall thickness (t) of 140 to 200 mm.

The diameter to wall thickness ratio of a PCC pile
Pile groups supporting buildings, bridge piers and (D/t∼7) is much lower than that for a typical offshore
high-level highway embankments usually have a large pipe pile (D/t∼40) and this difference prompted inves-
ultimate resistance and their design is governed by tigation in the centrifuge testing programme of three
the need to limit displacements under working loads different tip area ratios, varying from unity (for closed-
(e.g. Zhang & Ng 2005). Routine pile group settlement ended piles) to 0.46 (typical of a PCC pile) to 0.13
predictions often assume interaction factors between (typical of a steel pipe pile used in marine and off-
the piles in a group based on linear elastic solutions shore construction). The PCC pile is used as a form
and consequently tend to over-predict group settle- of ground improvement and therefore the centrifuge
ment considerably (Jardine et al. 1986, McCabe & experiments were conducted in soft-firm clay with an
Lehane 2006). There are, however, few well docu- overconsolidation ratio (OCR) of 2. All pile groups
mented case histories to ‘calibrate’ non-linear 3D tested contained five piles arranged in the same con-
finite element computations and to allow refinement figuration as the full scale field tests in soft clay (also
of new approximate pile group settlement prediction with OCR ≈ 2) reported by McCabe & Lehane (2006).
methods such as that proposed by Castelli & Maugeri The closed-ended piles installed in the centrifuge at a
(2002). Moreover, there has been no previous system- spacing to pile diameter ratio (s/D) of 3 could be com-
atic study into the influence of the pile end condition pared directly with these field tests. The centrifuge
on the behaviour of a displacement pile group in clay. tests examined two further typical spacing ratios to
This paper presents the results of such a study using allow assessment of the most efficient combination of
centrifuge model piles in lightly overconsolidated clay. spacing and pile area ratio.
This experimental research has a number of appli-
cations, but was motivated primarily by the increasing
popularity of a new form of large diameter open-ended 2 TEST APPARATUS AND PROCEDURE
cast-in-place concrete pipe pile (PCC pile) for soft
ground improvement applications in China (Liu et al. 2.1 Sample preparation
2003a, Liu et al. 2003b). The PCC pile has a high shaft The experiments were undertaken in the 0.6m radius
friction to concrete volume ratio (employing only 50% geotechnical drum centrifuge at the University of
of the concrete of an equivalent solid concrete pile) Western Australia (UWA). The centrifuge has twin
and avoids the need for reinforcement to deal with concentric shafts that are coupled with a precision
handling and installation stresses; it therefore offers servo motor, allowing the central tool table to be
a cost-effective solution to resist compression loads. rotated differentially to the outer sample contain-
The outer diameter (D) of a PCC pile ranges from ment channel, which is 300 mm in width and 200 mm
895
Total Load Cell
Load Cells
10.2mm
Pile3 Pile4
Model
Piles
125 Hook
105 S/D=2, 3, &4
Pile1 125
Pile2 Pile5
sample depth=150mm;
Ø10.2
Model pile
(a) Pile length=105mm (b) 125
Figure 2. Photo of pile installation.
(ii) open thin-walled (Ar = 0.13) piles. Closed-ended

piles were formed by fitting the open-ended piles with
removable flat tip plugs. The group configuration of
five piles is shown on Figure 1b. The pile cap, shown
on Figure 1c, was fabricated from aluminium and
designed to be light, but relatively rigid. The spacing
ratio (s/D) is defined as the distance between the centre
(c) of the centre pile to the centre of a corner pile normal-
ized by the pile outer diameter. The spacing could be
Figure 1. Model piles and pile group layout (a) sectional easily adjusted by moving the pile head connectors
view; (b) plan view, and (c) photo. along the tracks built into the pile cap; see Figures 1b
& c. The group tests employed a total axial load cell
in depth. A major advantage of the drum centrifuge (8 kN capacity) on top of the pile cap and three load
is the large plan area of the test sample, which could cells (1 kN capacity) on top of pile Nos. 1, 2 and 4; see
accommodate a number of single and group piles tests Figure 1b. Details regarding the design of the pile cap
in one sample. A complete description of the drum and load cells are given by Gaudin et al. (2005).
centrifuge facility at UWA is given by Stewart et al.
(1998). 2.3 Test details
Sample preparation involved placement of a
10 mm-thick sand drainage layer at the base of the All model piles were jacked into the kaolin sample
drum channel prior to the addition of kaolin clay slurry (under 125 g) at a constant rate of 1 mm/s using a spe-
at a water content of 120%; this slurry was poured into cially designed hook (as shown on Figure 2), which
the drum channel at 20 g via a hose fitted in a spe- allows detachment from the pile at final penetration
cialized sample placement actuator. Once the channel depth. The centre pile (No. 1) was installed first and
was full, the sample was left to consolidate at 250 g. was followed by installation of pile Nos. 2, 3, 4 and
Three miniature pore pressure transducers, pre-placed 5 (see Figure 1b). The final embedment length of all
in the channel on purpose-built stands, facilitated piles was 105 mm.
monitoring of progress of soil consolidation. As the After the five piles had been installed, the pile cap
clay settled, further clay slurry was added until a was mounted on the actuator and driven towards the
final sample height of 150 mm was achieved after full pile group in order to achieve the connection. Static
consolidation. The centrifuge acceleration was then load tests (compression followed by tension) were per-
reduced to 125g and the clay allowed swell to an over- formed at a constant rate of 0.0025 mm/s after at least
consolidation ratio (OCR) of 2. All foundation tests 100 minutes for single piles and 120 minutes for group
were performed at this OCR. The experimental results piles1 . Although accurate measurements of the clay
will be presented in model scale in later sections. plug height in the pipe piles after extraction could not
be made, the clay plug length was estimated to be close
2.2 Instrumented model piles to the full embedded length for all tests.
Figure 1a shows a schematic drawing of the model
piles used in the tests described below. The model piles 1
The equalization period of 100 minutes was assessed using
were made from aluminum and had an outer diameter the database of Lehane (1992) to be sufficient to allow at least
of 10.2 mm. The surface of the piles was sandblasted. 90% of excess pore pressures to dissipate. The loading rate
Three pile configurations were investigated: (i) corresponds to a normalised velocity (vD/cv ) of 0.1 and hence
closed-ended; (iii) open thick-walled (Ar = 0.46); and may be considered effectively drained (Randolph 2004).
896
Table 1. Testing program. As shown on Figure 3, the measured net T-bar tip
resistances (qnet ) were relatively consistent across the
Test t sample and over the testing duration. The undrained
name Pile End s mm D/t Ar shear strengths suT-bar , which are also shown on Fig-
ure 3, were estimated using the ‘standard’ bearing
GC4D Group Closed 4D – – 1
factor of 10.5. At pile tip level (L = 105 mm), the aver-
GC3D Group Closed 3D – – 1
GC2D Group Closed 2D – – 1 age suT-bar value was 20.7 kPa and the average suT-bar
SC Single Closed – – – 1 value along the pile shaft was 9.6 kPa. The corres-

GOB4D Group Open thick 4D 1.35 7.6 0.46 ponding mean undrained strength ratio, suT-bar /σv0 , of
GOB3D Group Open thick 3D 1.35 7.6 0.46 0.23 is considerably lower than the expected value of
GOB2D Group Open thick 2D 1.35 7.6 0.46 0.2 × OCR0.8 = 0.2 × 20.8 = 0.35; such low strength
SOB Single Open thick – 1.35 7.6 0.46 ratios are a characteristic feature of T-bar tests per-
GOA3D Group Open thin 3D 0.35 29 0.13 formed in the UWA drum centrifuge and are currently
GOA2D Group Open thin 2D 0.35 29 0.13 the subject of further research.
SOA Single Open thin – 0.35 29 0.13
Test name convention: G-Group, S-Single; C-Closed,

OB-Open Thick-Walled, OA-Open Thin-Walled; 4, 3 &
2D-Pile centre to centre spacing (D = 10.2 mm). 3 EXPERIMENTAL RESULTS
3.1 Single pile

qnet (kPa)
The load displacement response for the single ref-
0 50 100 150 200 250 300
0 0 erence piles (SC, SOB & SOA) is summarized on
15
10 Figure 4. The ultimate compression total capacities
20
30
(Qult ) are defined as those mobilised at 10% of pile
30
40 head displacement. It should be noted that pile weights
50 of ∼10 N (SC & SOA) and ∼20 N (SOB) have been
depth (mm)
(kPa)
45
60 accounted for – as indicated by the non-zero start point
σ'v0
60 70
80
on Figure 4.Table 2 provides a summary of the ultimate
75 Pile Tip Level 90 (Qult ), base (Qb ) and shaft (Qs ) capacities for these sin-
90
100 gle piles; these results assume that the shaft capacity
110 (Qs ) in compression is equal to the capacity measured
120
105
130
in the tension test that followed the compression test.
0 3 6 9 12 15 18 21 24 27 It is clear from these results that the pile area ratio
suT-bar (kPa) (Ar ) had a significant effect on the ultimate capacities
of these piles. Shaft frictions increase in a systematic

Figure 3. Variation of T-bar resistance with σv0 and depth. fashion as Ar increases – and hence as the degree of
displacement induced to the soil increases. End bear-
The experimental programme is summarized in ing capacities for the thick-walled pipe pile and the
Table 1, which also provides details of the notation closed-ended pile are similar but are 50% higher than
used for the tests. Due to space limitations, this the end bearing capacity of the thin-walled pipe pile
paper focuses on presentation and discussion of the (with Ar = 0.13).
compression test results. The strong correlation between Qs and Ar is in
keeping with trends reported by Miller & Lutenegger
(1997), but more pronounced than that inferred from
2.4 Soil properties
field tests by Chow (1997), and others, who show that
The kaolin soil used in the experiments had a liquid the shaft capacity of a typical pipe pile (with Ar < 0.2)
limit (LL) of 61%, plastic limit (PL) of 27% and angle is only about 10% less than that of a closed-ended pile.
of internal friction φ = 23◦ (Stewart 1992). Tests on The same field data, however, show erratic trends and
three core samples taken after completion of the cen- it is clear that further research is required. For example,
trifuge tests indicated an average bulk unit weight Bogard & Matlock (1990) show that equalized radial
of 16.6 kN/m3 and average water content of 50%. effective stresses on the pipe piles were up to 40% less
T-bar penetrometer tests, first introduced by Stewart & than the corresponding stresses on closed-ended piles
Randolph (1991), were carried out in-flight to deter- in Empire clay, but tension capacities for the two pile
mine the soil strength profile throughout the two weeks types were apparently very similar.
testing period. In all, 17 T-bar tests were performed Pile capacity predictions were performed using
around the sample at various stages during the testing the API design method (API 2000) and the Imperial
period (using the 5 mm diameter, 25 mm long T-bar). College ICP method (Jardine et al. 2005). The results,
897
120 1.3
1.2
100
1.1
α = τf/SuT-bar(-)
80
1
QTotal (N)
60 0.9
40 0.8
SC; Ar=1
measured
SOB; Ar=0.46 0.7 ICP
20 SOA; Ar=0.13 API
0.6
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0
0 1 2 3 4 5 6 7 8 9 10 Ar (-)
w/D (%)
1.6 40
measured
Figure 4. Measured load displacement response for single 1.4 ICP 35
API
piles (SC, SOB & SOA) under compression loading. 1.2 30
1 25
qb/suT -bar (-)

qb/qc (-)
Table 2. Ultimate capacities for single reference piles. 0.8 20
0.6 15
SC SOB SOA
Single piles Ar = 1 Ar = 0.46 Ar = 0.13 0.4 10
0.2 5
At Qs (N) 40a 30a 25a 0 0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
w/D = 10 Qb (N) 64 60 41
Ar (-)
% Qult (N) 104 90 66
a
measured tensile shaft capacities. Figure 5. Comparison between measured and predicted
resistance by API and ICP methods: (a) shaft; (b) base.
which are compared with measured capacities on
Figure 5, indicate that: qb /qc ratio for the centrifuge open-ended piles and
does not capture the observed relationship with Ar .
– The measured average α ratio2 (=τf /suT-bar )
increases with an increase in Ar . API (2000) does
not incorporate a dependence of α on Ar and hence 3.2 Pile group
its prediction of α (of unity) applies to the three The load displacement relationships measured for the
pile types considered. The α value predicted using pile groups under compression loading are presented
the ICP method for the closed-ended pile is close on Figure 6. As detailed in Table 1, a total of eight
to the measured value (assuming a residual friction groups were tested in compression.
angle of 17◦ (Chen 2006)), but this method over- Pile group capacity efficiency, ηcapacity , is defined
estimates the skin friction significantly for the two as the ratio of the pile group capacity to the number of
for open-ended piles. piles (5) times the single pile capacity. In this paper,
– The measured base resistance qb also increases with the ultimate capacity is defined at a pile group/head
the value of Ar , but as stated previously, the ultimate displacement of 10%D.
end bearing for the piles with Ar = 0.46 and 1.0 are ηcapacity values were derived from Figure 6 and are
closely comparable. It would appear that the lower plotted as a function of area ratio and spacing ratio
end bearing offered by the pile with Ar = 0.13 is (s/D) on Figure 7a. It is apparent from this figure that
due to plug failure during its load test. the ηcapacity value for the group with closed-ended piles
– The inferred qb /suT-bar ratios are considerably larger increases with Ar and is ≈0.8 at s/D = 3; this ηcapacity
than those given by API (2000). Comparable ratios value is about the same as that measured in comparable
to those measured in the centrifuge were also tests in the field (McCabe & Lehane 2006). Interest-
reported for field piles by Chow (1997). ingly, the ηcapacity values for the groups with the pipe
– The ICP method relates qb to the measured cone piles do not vary with s/D (for s/D between 2 and 4), but
resistance qc , which in this case is derived from as might be expected, ηcapacity increases as Ar reduces.
empirical relationship between qc and suT-bar (qc = While ηcapacity values of less than unity need to
15suT-bar + σv0 ). Although the interpreted ratios of be acknowledged for ultimate limit state calculations,
qb /qc for the closed-ended pile compare well with most pile group designs are dominated by the service-
the ICP method, this method under-predicts the ability limit state, for which the stiffness efficiency
ηg as defined below is of more interest. Values of
2
Note that these α values may be expected to be higher than ηg depend upon pile interaction factors, which are
conventional values, which are usually based on undrained known to vary with factors such as the soil stiffness
strengths in triaxial compression. non-linearity, s/D, L/D, Ar , number of piles in the
898
500 1.2
Ar=0.13 at w/D=10%
450
1
400 Ar=0.46
350
0.8
300
ηcapacity(-)
QTotal (N)
Ar=1
250 0.6
200
SC; Ar=1 0.4
150 GC4D
100 GC3D
GC2D 0.2
50
0 0
0 1 2 3 4 5 6 7 8 9 10 1.5 2 2.5 3 3.5 4 4.5
w/D (%)
Spacing ratio, s/D (-)
400
350 0.75
Ar=0.13 at w/D=1%
300 0.7
0.65
250
QTotal (N)
0.6 Ar=0.46
200
ηg(-)
0.55
150 SOB; Ar=0.46 Ar=1
GOB4D 0.5
100
GOB3D 0.45
GOB2D PIGLET;Linear
50 Elastic Solution
0.4
0
0 1 2 3 4 5 6 7 8 9 10 0.35
w/D(%) w(mm) 0.3
1.5 2 2.5 3 3.5 4 4.5
350
Spacing ratio, s/D (-)
300
250 Figure 7. Reduction factors plotted against s/D and Ar

(a) capacity reduction factor ηcapacity ; and (b) stiffness reduc-
200
tion factor ηg .
QTotal (N)
150
SOA; Ar=0.13 The trends indicated by ηg on Figure 7b are com-
100
GOA3D pared with stiffness efficiency factors computed for
50 GOA2D
closed-ended piles using the PIGLET computer pro-
0 gram (Randolph 2003); this program is typical of the
0 1 2 3 4 5 6 7 8 9 10
w/D(%) w(mm) type of pile group programs used worldwide in stand-
ard design and derives interaction factor between piles
Figure 6. Measured load displacement response for pile using the expressions for a linear elastic continuum
groups (a) GC; (b) GOB and (c) GOA. developed in Randolph & Wroth (1979). It is seen on
Figure 7b that PIGLET under-predicts the observed
group, position of the pile within the group, rigidity of ηg value for the closed-ended pile group by between
the pile cap. 15% and 20%, but predicts the observed trend with
s/D very well.
where kg is the axial stiffness (load per unit displace-
3.3 Optimum pile configuration
ment) of the pile group, ks is axial stiffness of a single
pile and n is the number of piles in the group. Based on above discussion, although single closed-
As shown on Figure 7b, ηg values range from 0.4 ended piles have a higher stiffness and capacity, their
to 0.75 at a pile head displacement of 1%D. ηg for use may not be economical when employed in a group
the thin-walled pile group with Ar = 0.13 is signifi- because of a relatively low stiffness efficiency factor
cantly greater than that of the piles with Ar = 0.46 and compared to a group containing thin-walled pipe piles.
1, indicating reduced levels of interaction between the For instance, the load displacement responses for pile
thin-walled pipe piles. The ηg value of 0.48 observed groups at a typical spacing ratio (s/D) of 3 are plotted
for the group with closed-ended piles with s/D = 3 on Figure 8. At typical working levels of displacement
is in good agreement with the mean value of 0.45 (e.g. w/D = 1%), for piles with Ar values of 1.0, 0.46
reported by McCabe & Lehane (2006) for a pile and 0.13, kg is 120 N/mm, 135 N/mm and 144 N/mm
group in the field with the same plan group configur- respectively. It is therefore apparent that, when the ser-
ation. Evidence from the single pile tests suggests that viceability limit state dictates the design, thin-walled
the similarity between ηg values for the groups with piles can provide a solution with comparable per-
Ar = 0.46 and 1.0 arises because of their comparable formance to closed-ended piles but with savings on
(and relatively large) end bearing resistances. material costs (if end plates are not a feasible option).
899
350 Castelli, F., & Maugeri, M. 2002. Simplified nonlinear analy-
sis for settlement prediction of pile groups. Journal of
300
Geotechnical and Geoenvironmental Engineering 128(1):
250 76–84.
Chen, W. 2006. Performance of suction caissons during
QTotal (N)
200
GC3D; Ar=1 installation and pullout in the marine clay. PhD thesis,
150 GOB3D; Ar=0.46 University of Western Australia, Perth.
GOA3D; Ar=0.13 Chow, F.C. 1997. Investigations into the behaviour of dis-
100
placement piles for offshore foundations. PhD, Imperial
50 College, London.
Gaudin, C., Lehane, B.M., & Wallis, P.F. 2005. A centrifuge
0
0 1 2 3 4 5 6 7 8 9 10 study of the monotonic and cyclic resistance of piles and
w/D (%) pile groups in sand. International Symposium on Frontiers
in Offshore Geotechnics, Perth: 749–755.
Figure 8. Comparison of load displacement response for Jardine, F.M., Chow, F.C., Overy, R.F., & Standing, J.R. 2005.
GC3D, GOB3D and GOA3D. ICP design methods for driven piles in sands and clays.
London: Thomas Telford.
Jardine, R.J., Potts, D.M., Fourie, A.B., & Burland, J.B. 1986.
4 CONCLUSIONS Studies of the Influence of Nonlinear Stress-Strain Char-
acteristics in Soil Structure Interaction. Geotechnique
The observations described in this paper reveal several 36(3): 377–396.
important points concerning the performance of single Lehane, B.M. 1992. Experimental investigations of pile
and group piles’ performance: behaviour using instrumented field piles. PhD thesis,
Imperial College.
– The pile area ratio (Ar ) has a significant influence Liu, H.L., Fei, K., Ma, X.H., & Gao, Y.F. 2003a. Cast-in-situ
on the ultimate capacities of single and group piles. concrete thin-wall pipe pile with Vibrated and steel tube
Shaft friction tends to increase in proportion to Ar mould technology and its application. (I): Development
(from the ‘wished-in place’ value at Ar close to and design. Rock and Soil Mechanics 24(2): 164 –168 (in
zero) whereas base capacity increases significantly Chinese).
at the area ratio corresponding to the transition from Liu, H.L., Hao, X.Y., Fei, K., & Chen, Y.H. 2003b. Field pour
coring to plugged behaviour during load testing. concrete thin wall cased pile technology and its appli-
cation. (II) Application and in-situ test. Rock and Soil
– Pile group capacity efficiency (ηcapacity ) for the Mechanics 24(3): 372–375 (in Chinese).
pile group configuration investigated was ≈0.8 for McCabe, B.A., & Lehane, B.M. 2006. The behavior of axi-
closed-ended piles with s/D = 3 and increased to ally loaded pile groups driven in clayey silt. Journal of
values closer to unity for thin-walled pipe piles. Geotechnical and Geoenvironmental Engineering,
– Stiffness efficiency for thin-walled pipe piles is sig- March.
nificantly higher for piles with low area ratios and Miller, G.A., & Lutenegger, A.J. 1997. Influence of pile plug-
is such that the overall group stiffness is relatively ging on skin friction in overconsolidated clay. Journal of
independent of the pile end condition. It follows that Geotechnical and Geoenvironmental Engineering 123(6):
thin-walled pipe piles may be employed without any 525–533.
Randolph, M.F. 2003. Analysis and design of pile groups.
loss in performance (compared to full displacement Users’ Manual, Version 4-2, University of Western
piles) under working load. Australia, Perth.
– The experiments have validated the effectiveness of Randolph, M.F. 2004. Characterization of soft sediments for
the ground improvement schemes employing PCC offshore applications. Proceedings of the 2nd Geotech-
piles. nical and Geophysical Site Characterization, Porto,
Portugal: 209–232.
Randolph, M.F., & Wroth, C.P. 1979. Analysis of the ver-
ACKNOWLEDGEMENTS tical deformation of pile groups. Geotechnique 29(4):
423–439.
This study was funded by a joint research project Stewart, D.P. 1992. Lateral loading of piled bridge abutments
between UWA and the Geotechnical Institute of Hohai due to embankment construction. PhD thesis, University
University. This support is gratefully acknowledged. of Western Australia, Perth.
Stewart, D.P., Boyle, R.S., & Randolph, M.F. 1998. Experi-
ence with a new drum centrifuge. Proceedings of the
REFERENCES International Conference Centrifuge ’98, Tokyo: 35–40.
Stewart, D.P., & Randolph, M.F. 1991. A new site investi-
API. 2000. RP2A-WSD Recommended Practice for Planning, gation tool for the centrifuge. Proceedings of the Inter-
Designing and Constructing Fixed Offshore Platforms- national Conference on Centrifuge 91, Boulder, USA: 531.
working Stress Design (21 ed.), Washington. Zhang, L.M., & Ng, A.M.Y. 2005. Probabilistic limiting tol-
Bogard, J.D., & Matlock, H. 1990. In-situ pile segment model erable displacements for serviceability limit state design
experiments at Empire, Louisiana. 22nd OTC, Houston, of foundations. Geotechnique 55(2): 151–161.
Texas: 459–467.
900
REFERENCE
Ahmadi, M. M., Byrne, P. M., and Campanella, R. G. 1999. Numerical simulation of

CPT tip resistance in layered soil. In Asian Institute of Technology 40th Year
Conference, New Frontiers & Challenges. November. pp. 1-8.
Ahmadi, M. M. 2000. Analysis of cone tip resistance in sand. PhD Thesis,
Department of Civil Engineering, the University of British Columbia,
Vancouver.
Ahmadi, M. M., and Robertson, P. K. 2005. Thin-layer effects on the CPT qc
measurement. Canadian Geotechnical Journal, 42(5): 1302-1317.
Alawneh, A. S., and Husein Malkawi, A. I. 2000. Estimation of post-driving residual
stresses along driven piles in sand. Geotechnical Testing Journal, 23(3): 313-
326.
Al-Mhaidib, A. I. 1999. Bearing capacity of model pile in sand under different
loading rates. In Proceeding of the 9th International Offshore and Polar
Engineering Conference, Brest, France. May 30-June 4. pp. 724-730.
Altaee, A., Fellenius, B. H., and Evgin, E. 1992a. Axial load transfer for piles in sand.
I. Tests on an instrumented precast pile. Canadian Geotechnical Journal,
29(1): 11-20.
Altaee, A., Evgin, E., and Fellenius, B. H. 1992b. Axial load transfer for piles in sand.
II. Numerical analysis. Canadian Geotechnical Journal, 29(1): 21-30.
API. 2000. Recommended Practice for Planning, Designing and Constructing Fixed
Offshore Platforms - Working Stress Design (21 ed.). Washington.
API. 2006. Recommended Practice for Planning, Designing and Constructing Fixed
Offshore Platforms - Working Stress Design (22 ed.), Washington.
Appendino, M. 1981. Interpretation of axial load tests on long piles. In Proceedings
of the 10th International Conference on Soil Mechanics and Foundation
Engineering, Stockholm. 15-19 June. pp. 593-598.
Baldi, G., Bellotti, R., Ghionna, V. N., Jamiolkowski, M., and Pasqualini, E. 1986.
Interpretation of CPTs and CPTUs; 2nd part: drained penetration of sand. In
Proceedings of the 4th International Geotechnical Seminar, Singapore. pp.
143-156.
BCP. 1971. Field tests on piles in sand. Soils and Foundations, 11(2): 29-49.
Begemann, H. K. S. P. 1963. Use of static soil penetrometer in Holland. New
Zealand Engineering, 18(2): 41-49.
R-1
Reference
Beringen, F. L., Windle, D., and Van Hooydonk, W. R. 1979. Results of loading tests
on driven piles in sand. In Proceedings of the Conference on Recent
Development in the Design and Construction of Piles, ICE, London. 21-22
March. pp. 213-225.
Berrill, J. b., Canou, J., Le Kouby, A., and Foray, P. 2004. The effect of layering on
cone resistance: calibration chamber tests. In Proceedings of 9th Australia
New Zealand Conference on Geomechanics, Auckland. pp. 648-653.
Bishop, R. F., Hill, R., and Mott, N. F. 1945. Theory of indentation and hardness
tests. Proceedings of the physical society, 57: 147-159.
Bolton, M. D. 1986. Strength and dilatancy of sands. Geotechnique, 36(1): 65-78.
Bolton, M. D., and Gui, M. W. 1993. The study of relative density and boundary
effects for cone penetration tests in centrifuge. London: Cambridge
University.
Bolton, M. D., Gui, M. W., Garnier, J., Corte, J. F., Bagge, G., Laue, J., et al. 1999.
Centrifuge cone penetration tests in sand. Geotechnique, 49(4): 543-552.
Bolton, M. D. 2005. Presentation, Piles in sand: science, sociology or engineering?
In Proceedings of the First International Symposium on Frontiers in Offshore
Geotechnics, Perth. 19-21 September.
Briaud, J. L., and Tucker, L. M. 1984. Piles in sand: a method including residual
stresses. Journal of Geotechnical Engineering, 110(11): 1666-1680.
Briaud, J. L., and Tucker, L. M. 1988. Measured and predicted axial response of 98
piles. Journal of Geotechnical Engineering, 114(9): 984-1001.
Briaud, J. L., and Tucker, L. M. 1989. Axially loaded 5 pile group and single pile in
sand. In Proceedings of the 12th International Conference on Soil Mechanics
and Foundation Engineering, Rio de Janeiro, Br. pp. 1121-1124.
Brucy, F., Meunier, J., and Nauroy, J. F. 1991. Behavior of pile plug in sandy soils
during and after driving. In Proceedings of the 23rd Annual Offshore
Technology Conference, Houston, Texas. 6-9 May. pp. 145-154.
Bruno, D. 1999. Dynamic and static load testing of driven piles in sand. PhD Thesis,
School of Civil and Resource Engineering, University of Western Australia,
Perth.
Bucky, P. B. 1931. Use of models for the study of mining problems. American
Institute of Mining and Metallurgical Engineers, Technical Publication, 425:
3-28.
Bucky, P. B., and Fentress, A. L. 1934. Application of principles of similitude to
design of mine workings. American Institute of Mining and Metallurgical
Engineers, Technical Publication, 529: 3-20.
Bucky, P. B., Solakian, A. G., and Baldin, L. S. 1935. Centrifugal method of testing
models. Civil Engineering, 5(5): 287-290.
Bustamante, M., and Gianeselli, L. 1982. Pile bearing capacity prediction by means
of static penetrometer CPT. In Proceedings of the 2nd European Symposium
on Penetration Testing, Amsterdam. 24-27 May. pp. 493-500.
R-2
Reference
Cao, L. F., Teh, C. I., and Chang, M. F. 2001. Undrained cavity expansion in
modified Cam Clay I: Theoretical analysis. Geotechnique, 51(4): 323-334.
Carter, J. P., Booker, J. R., and Yeung, S. K. 1986. Cavity expansion in cohesive
frictional soils. Geotechnique, 36(3): 349-353.
Chang, M. F., Teh, C. I., and Cao, L. F. 2001. Undrained cavity expansion in
modified Cam Clay II: Application to the interpretation of the piezocone test.
Geotechnique, 51(4): 335-350.
Chow, F. C. 1997. Investigations into the behaviour of displacement piles for
offshore foundations. PhD Thesis, Department of Civil and Environmental
Engineering, Imperial College, London.
Clausen, C. J. F., Aas, P. M., and Karlsrud, K. 2005. Bearing capacity of driven piles
in sand, the NGI approach. In Proceedings of the First International
Symposium on Frontiers in Offshore Geotechnics, Perth. 19-21 September.
pp. 677-681.
Collins, I. F., and Yu, H. S. 1996. Undrained cavity expansions in critical state soils.
International Journal for Numerical and Analytical Methods in Geomechanics,
20(7): 489-516.
Coyle, H. M., and Castello, R. R. 1981. New design correlations for piles in sand.
Journal of the Geotechnical Engineering Division, 107(7): 965-986.
Cudmani, R., and Osinov, V. A. 2001. The cavity expansion problem for the
interpretation of cone penetration and pressuremeter tests. Canadian
Geotechnical Journal, 38(3): 622-638.
De Beer, E., Lousberg, D., De Jonghe, A., Carpentier, R., and Wallays, M. 1979.
Analysis of the results of loading tests performed on displacement piles of
different types and sizes penetrating at a relatively small depth into a very
dense layer. In Proceedings of the Conference on Recent Development in the
Design and Construction of Piles, ICE. 21-22 March. pp. 199-211.
De Beer, E. E. 1963. Scale effect in transposition of results of deep-sounding tests on
ultimate bearing capacity of piles and caisson foundations. Geotechnique,
13(1): 39-75.
De Cock, F., Legrand, C., and Huybrechts, N. 2003. Overview of design methods of
axially loaded piles in Europe - Report of ERTC3-Piles, ISSMGE
Subcommittee. In Proceeding of the 8th European Conference on Soil
Mechanics and Geotechnical Engineering, Prague. pp. 663-715.
de Nicola, A. 1996. The performance of pipe piles in sand. PhD Thesis, School of
civil and resource engineering, the University of Western Australia, Perth.
de Nicola, A., and Randolph, M. F. 1997. Plugging behaviour of driven and jacked
piles in sand. Geotechnique, 47(4): 841-856.
de Nicola, A., and Randolph, M. F. 1999. Centrifuge modelling of pipe piles in sand
under axial loads. Geotechnique, 49(3): 295-318.
de Ruiter, J., and Beringen, F. L. 1979. Pile foundations for large North Sea
structures. Marine Geotechnology, 3(3): 267-314.
R-3
Reference
Deeks, A. D., White, D. J., and Bolton, M. D. 2005. A comparison of jacked, driven
and bored piles in sand. In Proceedings of the 16th International Conference
on Soil Mechanics and Geotechnical Engineering. 12-16 September. pp.
2103-2106.
Deeks, A. D., and White, D. J. 2006. Centrifuge modelling of jacked piles. In
Proceedings of the 6th International Conference on Physical Modelling in
Geotechnics, Hongkong. pp. 821-826.
Dennis, N. D., and Olson, R. E. 1983. Axial capacity of steel pipe piles in sand. In
Proceedings of the Conference on Geotechnical Practice in Offshore
Engineering, Austin, Texas. pp. 389 - 402.
Eiksund, G., and Nordal, S. 1996. Dynamic model pile testing with pore pressure
measurements. In Proceedings of the 5th International Conference of
Application of Stress-wave Theory to Piles, Orlando. pp. 1-11.
Eslami, A., and Fellenius, B. H. 1997. Pile capacity by direct CPT and CPTu
methods applied to 102 case histories. Canadian Geotechnical Journal, 34(6):
886-904.
Finlay, T. C., White, D. J., Bolton, M. D., and Nagayama, T. 2001. Press-in piling:
the installation of instrumented steel tubular piles with and without driving
shoes. In Proceedings of 5th International Conference on Deep Foundation
Practice, Singapore. 4-6 April. pp. 199-208.
Fleming, W. G. K. 1992. A new method for single pile settlement prediction and
analysis. Geotechnique, 42(3): 411-425.
Foray, P., Balachowski, L., and Colliat, J. L. 1998. Bearing capacity of model piles
driven into dense overconsolidated sands. Canadian Geotechnical Journal,
35(2): 374-385.
Fugro. 2003. Axial pile capacity design method for offshore driven piles in sand.
Report to API, P-1003, Issue 3.
Gavin, K. 1998. Experimental investigation of open and closed ended piles in sand.
PhD Thesis, Department of Civil, Structural & Environmental Engineering,
University of Dublin, Trinity College, Dublin.
Gavin, K., and Lehane, B. 2003. End bearing of small pipe piles in dense sand. In
BGA International Conference on Foundations, Innovations, Observations,
Design and Practice, Dundee, United Kingdom. pp. 321-330.
Gavin, K., and Lehane, B. M. 2005. Estimating the end bearing resistance of pipe
piles in sand using the final filling ratio. In Proceedings of the First
International Symposium on Frontiers in Offshore Geotechnics, Perth. 19-21
September. pp. 717-723.
Gavin, K. G., and Lehane, B. M. 2003. The shaft capacity of pipe piles in sand.
Canadian Geotechnical Journal, 40(1): 36-45.
GeoDelft (Delft Geotechnics). 1936. The pre-determination of the required length
and the prediction of the resistance of piles. In Proceeding of the First
International Conference on Soil Mechanics and Foundation Engineering,
Cambridge. 22-26 June. pp. 181-194.
R-4
Reference
Ghionna, V. N., Jamiolkowski, M., Lancellotta, R., and Pedroni, S. 1993. Base
capacity of bored piles in sands from in situ tests. In Proceedings of the 2nd
International Geotechnical Seminar on Deep Foundations on Bored and
Auger Piles, Ghent. pp. 67-75.
Gregersen, O. S., Aas, G., and DiBiagio, E. 1973. Load tests on friction piles in loose
sand. In Proceeding of the 8th International Conference on Soil Mechanics
and Foundation Engineering, Moscow. pp. 109-117.
Gui, M. W., and Bolton, M. D. 1998. Geometry and scale effects in CPT and pile
design. In Geotechnical Site Characterization, Balkema, Rotterdam. pp. 1063-
1068.
Gupta, R. C. 2002. Estimating bearing capacity factors and cone tip resistance. Soils
and Foundations, 42(6): 117-127.
House, A. R. 2002. Suction caisson foundations for buoyant offshore foundations.
PhD Thesis, School of Civil and Resource Engineering, University of
Western Australia, Perth.
Hunter, A. H., and Davisson, M. T. 1968. Measurements of pile load transfer. In
Performance of Deep Foundations: a Symposium Presented at the Seventy-
first Annual Meeting American Society for Testing and Materials, San
Francisco. pp. 106-117.
Huy, N. Q., Dijkstra, J., van Tol, A. F., and Holscher, P. 2004. Influence of loading
rate on the bearing capacity of piles in sand. In 16th International Conference
on Soil Mechanics and Geotecnical Engineering. pp. 2125-2128.
Jardine, F. M., Chow, F. C., Overy, R. F., and Standing, J. R. 2005. ICP design
methods for driven piles in sands and clays. Thomas Telford, London.
Joer, H. A., Randolph, M. F., and Liew, Y. H. 1996. Interpretation of cone resistance
in layered soils. In Proceedings of 7th Australia New Zealand Conference on
Geomechanics, Adelaide. pp. 92-97.
Kerisel, J. 1961. Deep foundations in sands: variation of ultimate bearing capacity
with soil density, depth, diameter and speed. In Proceedings of the 5th
International Conference on Soil Mechanics and Foundation Engineering. pp.
73-83.
Kishida, H. 1967. Ultimate bearing capacity of pipe piles in sand. In Proceedings of
the 3rd Asian Regional Conference on Soil Mechanics and Foundation
Engineering. Sep 25-28. pp. 196-199.
Kishida, H. 1977. Behaviour of sand plugs in open-ended steel pipe piles. In
Proceedings of the 9th International Conference on Soil Mechanics and
Foundation Engineering, Tokyo. pp. 601-604.
Klotz, E. U., and Coop, M. R. 2001. An investigation of the effect of soil state on the
capacity of driven piles in sands. Geotechnique, 51(9): 733-751.
Kolk, H. J., Baaijens, A. E., and Senders, M. 2005a. Design criteria for pipe piles in
silica sands. In Proceedings of the First International Symposium on Frontiers
in Offshore Geotechnics, Perth. 19-21 September. pp. 711-716.
R-5
Reference
Kolk, H. J., Shaffei, K. A., and Baaijens, A. E. 2005b. Axial load tests on pipe piles
in very dense sands at Ras Tanajib. In Proceedings of the First International
Symposium on Frontiers in Offshore Geotechnics, Perth. 19-21 September.
pp. 765-771.
Kolk, H. J., Vergobbi, P., and Baaijens, A. E. 2005c. Results from axial load tests on
pipe piles in very dense sands: The EURIPIDES JIP. In Proceedings of the
First International Symposium on Frontiers in Offshore Geotechnics, Perth.
19-21 September. pp. 661-667.
Kraft, L. M. 1990. Computing axial capacity in sands for offshore conditions. Marine
Geotechnology, 9(1): 61-92.
Kraft, L. M. 1991. Performance of axially loaded pipe piles in sand. Journal of
Geotechnical Engineering, 117(2): 272-296.
Kulhawy, F. H. 1984. Limiting tip and side resistance: fact or fallacy? In Proceedings
of a Symposium, ASCE Geotechnical Division, National Convention, San
Francisco. pp. 80-98.
Ladanyi, B., and Foriero, A. 1998. Numerical solution of cavity expansion problem
in sand based directly on experimental stress-strain curves. Canadian
Geotechnical Journal, 35(4): 541-557.
Ladanyi, M. B. 1961. Discussion. In Proceedings of the 5th International Conference
on Soils Mechanics and Foundation Engineering. pp. 270-271.
Lee, J. H., and Salgado, R. 1999. Determination of pile base resistance in sands.
Journal of Geotechnical and Geoenvironmental Engineering, 125(8): 673-683.
Lehane, B. M. 1992. Experimental investigations of pile behaviour using
instrumented field piles. PhD Thesis, Department of Civil and Environmental
Engineering, Imperial College, London.
Lehane, B. M., and Gavin, K. G. 2001. Base resistance of jacked pipe piles in sand.
Lehane, B. M., and Randolph, M. F. 2002. Evaluation of a minimum base resistance
for driven pipe piles in siliceous sand. Journal of Geotechnical and
Geoenvironmental Engineering, 128(3): 198-205.
Lehane, B. M., Pennington, D., and Clark, S. 2003. Jacked end-bearing piles in the
soft alluvial sediments of Perth. Australian Geomechanics, 38(3): 123-133.
Lehane, B. M., Ismail, M. A., and Fahey, M. 2004. Seasonal dependence of in situ
test parameters in sand above the water table. Geotechnique, 54(3): 215-218.
Lehane, B. M., Schneider, J. A., and Xu, X. 2005a. A review of new API design
methods for displacement piles in siliceous sand (No. GEO05358). Perth:
University of Western Australia.
Lehane, B. M., Schneider, J. A., and Xu, X. 2005b. The UWA-05 method for
prediction of axial capacity of driven piles in sand. In International
Symposium on Frontiers in Offshore Geotechnics, Perth. pp. 683-689.
Leung, C. F., Lee, F. H., and Yet, N. S. 1996. Role of particle breakage in pile creep
in sand. Canadian Geotechnical Journal, 33(6): 888-898.
R-6
Reference
Lunne, T., Robertson, P. K., and Powell, J. J. M. 1997. Cone penetration testing in
geotechnical practice. Blackie Academic & Professional, London.
Madabhushi, S. P. G., Houghton, N. E., and Haigh, S. K. 2006. A new automatic
sand pourer for model preparation at the University of Cambride. In
Proceedings of the 6th International Conference on Physical Modelling in
Geotechnics, Hongkong.
Malhotra, S. 2002. Axial load capacity of pipe piles in sand: Revisited. In Deep
Foundations Congress 2002 - International Perspective on Theory, Design,
Construction, and Performance, Orlando, FL. Feb 14-16. pp. 1230-1246.
Mandolini, A., Russo, G., and Viggiani, C. 2005. Pile foundations: experimental
investigations, analysis and design. In Proceedings of the 16th International
Conference on Soil Mechanics and Geotechnical Engineering, Osaka, Japan.
pp. 177-213.
Matsumoto, T., and Takei, M. 1991. Effects of soil plug on behaviour of driven pipe
piles. Soils and Foundations, 31(2): 14-34.
McClelland, B. 1974. Design of deep penetration piles for ocean structures. Journal
of Geotechnical Engineering, 100(GT7): 709-747.
Medzvieckas, J., and Slizyte, D. 2005. Installation of jacked piles in sandy soils. In
Proceedings of the 16th International Conference on Soil Mechanics and
Geotechnical Engineering, Osaka. pp. 2149-2152.
Meyerhof, G. G. 1976. Bearing capacity and settlement of pile foundations. Journal
of the Geotechnical Engineering Division, 102(3): 195-228.
Meyerhof, G. G., and Valsangkar, A. J. 1977. Bearing capacity of piles in layered
soils. In Proceedings of the 8th International Conference on Soil Mechanics
and Foundation Engineering, Moscow. pp. 645-650.
Meyerhof, G. G., and Sastry, V. V. R. N. 1978a. Bearing capacity of piles in layered
soils. Part 1. Clay overlying sand. Canadian Geotechnical Journal, 15(2):
171-182.
Meyerhof, G. G., and Sastry, V. V. R. N. 1978b. Bearing capacity of piles in layered
soils. Part 2. Sand overlying clay. Canadian Geotechnical Journal, 15(2): 183-
189.
Meyerhof, G. G. 1983. Scale effects of ultimate pile capacity. Journal of
Geotechnical Engineering Division, 109(6): 797-806.
Miller, G. A., and Lutenegger, A. J. 1997. Influence of pile plugging on skin friction
in overconsolidated clay. Journal of Geotechnical and Geoenvironmental
Engineering, 123(6): 525-533.
Mitchell, P. 2005. Personal Communication.
Moss, R. E. S. 2003. CPT-based probabilistic assessment of seismic soil liquefaction
initiation. PhD Thesis, Civil and Environmental Engineering, University of
California, Berkeley.
R-7
Reference
Moss, R. E. S., Seed, R. B., and Olsen, R. S. 2006. Normalizing the CPT for
Overburden Stress. Journal of Geotechnical and Geoenvironmental
Engineering, 132(3): 378-387.
Nauroy, J. F., and Le Tirant, P. 1983. Model tests of piles in calcareous sands. In
Proceedings of the Conference on Geotechnical Practice in Offshore
Engineering, Austin, Tex, USA. pp. 356-369.
NGI. 2001. Bearing Capacity of Driven Piles in Sand (No. 525211-2): Norwegian
Geotechnical Institute.
Oda, K., and Matsui, T. 2001. Effect of pile diameter on bearing mechanism of cast-
in-place bored piles. In Proceedings of the 11th International Offshore and
Polar Engineering Conference, Stavanger, Norway. pp. 639-646.
O'Neill, M. W., and Raines, R. D. 1991. Load transfer for pipe piles in highly
pressured dense sand. Journal of Geotechnical Engineering, 117(8): 1208-
1226.
O'Neill, M. W. 2001. Side resistance in piles and drilled shafts. Journal of
Geotechnical and Geoenvironmental Engineering, 127(1): 3-16.
Paik, K., and Salgado, R. 2003. Determination of bearing capacity of open-ended
piles in sand. Journal of Geotechnical and Geoenvironmental Engineering,
129(1): 46-57.
Paik, K., Salgado, R., Lee, J., and Kim, B. 2003. Behavior of open- and closed-ended
piles driven into sands. Journal of Geotechnical and Geoenvironmental
Engineering, 129(4): 296-306.
Paik, K., and Salgado, R. 2004. Effect of pile installation method on pipe pile
behavior in sands. Geotechnical Testing Journal, 27(1): 78-88.
Paikowsky, S. G., Whitman, R. V., and Baligh, M. M. 1989. New look at the
phenomenon of offshore pile plugging. Marine Geotechnology, 8(3): 213-230.
Paikowsky, S. G., and Whitman, R. V. 1990. Effects of plugging on pile performance
and design. Canadian Geotechnical Journal, 27(4): 429-440.
Philips, E. 1869. De l'equilibre des solides elastiques semblables. Ecole
Centrale/Ecole Polytechnique, France.
Plantema, G. 1948. Results of a special loading-test on a reinforced concrete pile, a
so-called pile sounding; interpretation of the results of deep-soundings,
permissible pile loads and extended settlemetn observations. In Proceedings
of the 2nd International Conference on Soil Mechanics and Foundation
Engineering. pp. 112-117.
Poulos, H. G. 1987. Analysis of residual stress effects in piles. Journal of
Geotechnical Engineering, 113(3): 216-229.
Poulos, H. G. 1989. Pile behaviour - theory and application. Geotechnique, 39(3):
365-415.
Poulos, H. G. 2005. Pile behavior - Consequences of geological and construction
imperfections. Journal of Geotechnical and Geoenvironmental Engineering,
131(5): 538-563.
R-8
Reference
Prandtl, L. 1920. Harte plashercher Korper. Nachrichten Ges. Wissenschafter

Gottingen.
Pump, W., Korista, S., and Scott, J. 1998. Installation and load tests of deep piles in
Shanghai alluvium. In Proceedings of the 7th International Conference on
Piling and Deep Foundations, Vienna. pp. 31-36.
Randolph, M. F., Jewell, R. J., Stone, K. J. L., and Brown, T. A. 1991. Establishing a
new centrifuge facility. In Proceedings of the International Conference on
Centrifuge 91, Boulder/Colorado, USA. Jun 13-14. pp. 3-9.
Randolph, M. F., Leong, E. C., and Houlsby, G. T. 1991. One-dimensional analysis
of soil plug in pipe piles. Geotechnique, 41(4): 587-598.
Randolph, M. F., Dolwin, J., and Beck, R. 1994. Design of driven piles in sand.
Randolph, M. F. 2003. RATZ, Version 4.2, Load transfer analysis of axially loaded
piles.
Randolph, M. F. 2003. Science and empiricism in pile foundation design.
Rix, G. J., and Stokoe, K. H. 1992. Correlation of initial tangent modulus and cone
resistance. In Proceedings of the International Symposium on Calibration
Chamber Testing, Potsdam, New York. pp. 651-362.
Salgado, R. 1993. Analysis of penetration resistance in sands. PhD Thesis, School of
Civil Engineering, University of California, Berkeley.
Salgado, R. 1995. Design of piles in sands based on CPT results. In Proceedings of
10th Pan American Conference on Soil Mechanics and Foundation
Engineering, Guadalajara, Mexico. pp. 1261-1274.
Salgado, R., Mitchell, J. K., and Jamiolkowski, M. 1997. Cavity expansion and
penetration resistance in sand. Journal of Geotechnical and Geoenvironmental
Engineering, 123(4): 344-354.
Salgado, R., and Randolph, M. F. 2001. Analysis of cavity expansion in sand.
International Journal of Geomechanics, 1(2): 175-192.
Salgado, R., Lee, J. H., and Kim, K. 2002. Load tests on pipe piles for development
of CPT-bassed design method (No. FHWA/IN/JTRP-2002/4). West Lafayette:
Purdue University.
Salgado, R., and Prezzi, M. 2007. Computation of cavity expansion pressure and
penetration resistance in sands. International Journal of Geomechanics,
ASCE, ASCE, scheduled for publication in the August 2007 issue.
Sayed, S. M., and Hamed, M. A. 1987. Expansion of cavities in layered elastic
system. International Journal for Numerical and Analytical Methods in
Geomechanics, 11(2): 203-313.
Schanz, T., Vermeer, P. A., and Bonnier, P. G. 1999. The hardening soil model:
formulation and verification. In Beyond 2000 in Computational Geotechnics -
10 Years of Plaxis, Rotterdam.
R-9
Reference
Schmertmann, J. H. 1978. Guidelines for cone test, performance, and design (No.
FHWATS-78209): U.S. Federal Highway Administration.
Schnaid, F., and Mantaras, F. M. 2003. Cavity expansion in cemented materials:
Structure degradation effects. Geotechnique, 53(9): 797-807.
Schneider, J. A., Xu, X., and Lehane, B. M. 2007a. Database assessment of CPT
based design methods for axial capacity of driven piles in siliceous sands:
Part I – overview of methods, database and predictive performance. Journal
of Geotechnical and Geoenvironmental Engineering, Submitted.
Schneider, J. A., Lehane, B. M. and Xu, X. 2007b. Database assessment of CPT
based design methods for axial capacity of driven piles in siliceous sands:
Part II – application to reliability analyses for offshore platforms. Journal of
Geotechnical and Geoenvironmental Engineering, Submitted.
Schneider, J. A. 2006. Impact of offshore site investigation practice on reliability of
axial pile design in siliceous sand. In Proceedings of the 31st Annual
Conference on Deep Foundations, Washington D.C. 4-6 Oct. pp. 623-638.
Schneider, J. A. 2007. Methods to increase the reliability of axial capacity analyses
for offshore piles in granular soils. PhD Thesis, School of Civil and Resource
Engineering, The University of Western Australia, Perth.
Schofield, A. N. 1980. Cambridge geotechnical centrifuge operations. Geotechnique,
30(3): 227-268.
Shioi, Y., Yoshida, O., Meta, T., and Homma, M. 1992. Estimation of bearing
capacity of steel pipe pile by static loading test and stress-wave theory
(Trans-Tokyo Bay Highway). In Proceedings of Application of Stress-Wave
Theory to Piles. pp. 325-330.
Smith, I. M., To, P., and Willson, S. M. 1986. Plugging of pipe piles. In Proceedings
of the 3rd international conference on numerical methods in offshore piling,
Nantes. pp. 53-73.
Stewart, D. P. 1992. Lateral loading of piled bridge abutments due to embankment
construction. PhD thesis, School of Civil and Resource Engineering,
University of Western Australia, Perth.
Sulaiman, I. H., and Coyle, H. M. 1971. Predicted behaviour of axially loaded piles
in sand. In Proceedings of the 3rd Annual Offshore Technolgy Conference,
Houston. pp. 643-653.
Tatsuoka, F., Zhou, S., Sato, T., and Shibuya, S. 1990. Evaluation method of
liquefaction potential and its application, in Report on Seismic Hazards on
the Ground in Urban Areas. Tokyo, Japan: Ministry of Education of Japan.
Taylor, R. N. 1995. Geotechnical centrifuge technology. Blackie Academic.
Titi, H., and Abu-Farsakh, M. 1999. Evaluation of bearing capacity of piles from
cone pnetration tests data: Louisiana Transportation Research Center.
Toolan, F. E., Lings, M. L., and Mirza, U. A. 1990. An appraisal of API RP2A
recommendations for determining skin friction of piles in sand. In
Proceedings of the 22nd Annual Offshore Technology Conference, Houston,
Texas. May 7-10. pp. 33-42.
R-10
Reference
van Den Berg, P., and Huetink, H. 1996. Eulerean finite element model for
penetration in layered soil. International Journal for Numerical and Analytical
Methods in Geomechanics, 20(12): 865-886.
Van Mierlo, W. C., and Koppejan, A. W. 1952. Lengte en draagvermogen van
heipalen. Bouw.
Vesic, A. B. 1963. Bearing capacity of deep foundations in sand. National Research
Council, Highway Research Board, Research Record(39): 112-153.
Vesic, A. S. 1970. Tests on instrumented piles, Ogeechee River site. Journal of the
Soil Mechanics and Foundations Division, Proceeding of the American
Society of Civil Engineering, 96(SM2): 561-584.
Vesic, A. S. 1977. Design of pile foundations: Synthesis of Highway Practice No. 42,
Nat. Coop. Highway Research Prog., Transportation Research Board,
Washingtong, D.C.
Vreugdenhil, R., Davis, R., and Berrill, J. 1994. Interpretation of cone penetration
results in multilayered soils. International Journal for Numerical and
Analytical Methods in Geomechanics, 18(9): 585-599.
Watson, P. G. 1999. Performance of skirted foundations for offshore structures. PhD
Thesis, School of Civil and Resource Engineering, University of Western
Australia, Perth.
White, D. J. 2002. An investigation into the behaviour of pressed-in piles. PhD
Thesis, Churchill College, University of Cambridge, London.
White, D. J., Sidhu, H. K., Finlay, T. C., Bolton, M. D., and Nagayama, T. 2000.
Press-in piling: the influence of plugging on driveability. In Proceedings of
the 8th International Conference of the Deep Foundations Institute, New York,
pp. 299-310.
White, D. J., and Lehane, B. M. 2004. Friction fatigue on displacement piles in sand.
White, D. J., and Bolton, M. D. 2005. Comparing CPT and pile base resistance in
sand. ICE, Geotechnical Enginnering, 158: 3-14.
White, D. J., Schneider, J. A., and Lehane, B. M. 2005. The influence of effective
area ratio on shaft friction of displacement piles in sand. In Proceedings of
the First International Symposium on Frontiers in Offshore Geotechnics,
Perth. 19-21 September. pp. 741-747.
White, D. J. 2006. Presentation in the University of Western Australia.
Xu, X., and Lehane, B. M. 2005. Evaluation of end-bearing capacity of closed-ended
pile in sand from cone penetration data. In Proceedings of the First
International Symposium on Frontiers in Offshore Geotechnics, Perth. 19-21
September. pp. 733-739.
Xu, X., Lehane, B. M., and Schneider, J. A. 2005. Evaluation of end-bearing
capacity of open-ended piles driven in sand from CPT data. In Proceedings of
the First International Symposium on Frontiers in Offshore Geotechnics,
Perth. 19-21 September. pp. 725-731.
R-11
Reference
Xu, X., Lehane, B. M., Gaudin, C., Zhang, T., and Liu, H. L. 2006. Centrifuge
studies of single and group displacement piles in clay. In Proceedings of the
6th International Conference on Physical Modelling in Geotechnics,
Hongkong. pp. 895-900.
Xu, X., Liu, H. L., and Lehane, B. M. 2006. Pipe pile installation effects in soft clay.
Proceedings of the Institution of Civil Engineers, Geotechnical Engineering,
159(GE4): 285-296.
Yasufuku, N., Ochiai, H., and Ohno, S. 2001. Pile end-bearing capacity of sand
related to soil compressibility. Soils and Foundations, 41(4): 59-71.
Yen, T. L., Chin, C. T., and Wang, R. F. 1989. Interpretation of instrumented driven
steel pipe piles. In Proceedings of Congress on Foundation Engineering,
Current Principles and Practices, Evanston, IL, USA. pp. 1293-1308.
Yetginer, A. G., White, D. J., and Bolton, M. D. 2006. Field measurements of the
stifness of jacked piles and pile groups. Geotechnique, 56(5): 349-354.
Yu, H. S., and Houlsby, G. T. 1991. Finite cavity expansion in dilatant soils: loading
analysis. Geotechnique, 41(2): 173-183.
Yu, H. S., and Mitchell, J. K. 1998. Analysis of cone resistance: review of methods.
Yu, H. S. 2000. Cavity expansion methods in Geomechanics.
R-12

Xu Xiangtao 2007

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Xu Xiangtao 2007

Uploaded by

Copyright:

Available Formats

INVESTIGATION OF THE END BEARING

PERFORMANCE OF DISPLACEMENT PILES IN SAND

A dissertation submitted for

THE UNIVERSITY OF WESTERN AUSTRALIA

Jianle XU, Jiping LIU

& Ting ZHANG

Their Love and Support

The axial performance of piles in sand remains an area of great uncertainty in

CHAPTER 1 INTRODUCTION ........................................................................1-1

CHAPTER 2 LITERATURE REVIEW .............................................................2-1

2.1 Introduction ..........................................................................................2-1

3.1 Introduction .......................................................................................... 3-1

CHAPTER 4 FIELD TESTS AT SHENTON PARK ........................................ 4-1

4.1 Introduction .......................................................................................... 4-1

CHAPTER 5 CENTRIFUGE TEST APPARATUS AND PROCEDURE........ 5-1

5.1 Introduction .......................................................................................... 5-1

CHAPTER 6 CENTRIFUGE TEST RESULTS.................................................6-1

6.1 Introduction ..........................................................................................6-1

CHAPTER 7 ANALYSIS OF qb IN LAYERED SOIL ....................................7-1

7.1 Introduction ..........................................................................................7-1

CHAPTER 8 ANALYSIS AND DISCUSSION ................................................ 8-1

8.1 Introduction .......................................................................................... 8-1

CHAPTER 9 CONCLUSIONS .......................................................................... 9-1

9.1 Introduction .......................................................................................... 9-1

k Pile base secant gradient =Δqb/Δ(w/D)

α Stiffness ratio =E/qc

γ Specific unit weight of the soil

σ'1 Major principal stress

σ'3 Minor principal stress

σ'v0 Effective vertical stress

σv0 Total vertical stress

ω Constant angular rotational speed of the centrifuge

API American Petroleum Institute

• Fugro-05 tends to over-predict capacity of piles in compression, particularly for

• ICP-05 has a tendency to under-predict end-bearing capacity of open-ended piles

1.2 RESEARCH OBJECTIVES

Specifically, the key objectives of the research are to:

(5) Develop an improved understanding of the physical mechanisms which govern

1.3 ORGANISATION OF THESIS

This Thesis comprises nine Chapters.

Chapter 2 presents a review of the conventional design approaches for base

Chapter 3 compares the predictive performance of the three CPT-based methods

2.2 PILE END BEARING RESISTANCE IN SANDS

2.2.1 Bearing capacity theory

q b = N q × σ' v 0 ≤ q b,lim it ( 2.1 )

where Nq is dimensionless bearing capacity factor, σ'v0 is the effective overburden

2.2.2 Cavity expansion method

(a) Salgado (1993) (b) Randolph et al. (1994)

According to this method, the pile/cone penetration can be simulated by expanding a

based on some semi-empirical relationships (Salgado 1993, Randolph et al. 1994), as

q b = 2f v p lim it exp(2Δλ tan φT )

q b = p lim it (1 + tan α ⋅ tan φ) ( 2.3 )

(a) Ladanyi (1961) (b) Yasufuku et al. (2001)

2.2.3 Correlations with CPT data

As indicated in Equation 2.4, all of those methods employ a direct relationship

Figure 2.5 Values of kc for driven piles in sand (Dutch method)

Table 2.1 Values of kc factor for LCPC method

Soil Type Driven Piles Bored Piles

2.3 FACTORS AFFECTING PILE END BEARING RESISTANCE

2.3.1 Method of installation

To demonstrate the influence of installation method on pile capacity, Mandolini et al.

Pile type Qtotal/Wpile COV(Qtotal/Wpile)

2.3.2 Surface scale effect