Sie sind auf Seite 1von 147

Behaviour of Single Piles under Axial Loading

Analysis of Settlement and Load Distribution

Joana Gonalves Sumares Betencourt Ribeiro

Thesis to obtain the Master of Science Degree in


Civil Engineering

Examination Committee
Chairperson:

Prof. Jos Manuel Matos Noronha da Cmara

Supervisor:

Prof. Jaime Alberto dos Santos

Member of the Committee:

Prof. Peter John Bourne-Webb

May 2013

To my parents
Aos meus pais

May the God of hope fill you with all joy and peace in believing,
so that by the power of the Holy Spirit you may abound in hope.
Romans 15:13

Que o Deus da esperana vos mantenha felizes e cheios da paz que nasce pela f,
para que abundeis na esperana pelo poder do Esprito Santo.
Romanos 15:13

ABSTRACT
The purpose of this thesis is to analyse the behaviour of single piles under axial loading, as far as
settlement and load transfer mechanisms are concerned.
It includes a literary review of two elastic theory-based methods, the Poulos and Davis method and the
Randolph and Wroth method, as well as axisymmetric elastic modelling using the finite element-based
program Plaxis. The results given by each method are organized in dimensionless charts of loadsettlement ratio and proportion of load transferred to the pile base in terms of the pile slenderness
ratio and the soil inhomogeneity and compared amongst each other.
This thesis also includes two comparison studies which involve axisymmetric elastoplastic modelling in
Plaxis, considering the Mohr-Coulomb failure criterion. The first one is a comparison with two previous
finite element simulations subject to very similar conditions: a GEFdyn 3D simulation and a CESARLCPC 2D axisymmetric simulation. It showed that in the simulation performed by Plaxis the value of
load transferred to the pile base was lower than the others, although the total load-settlement curve
was very similar in all three cases. The second one is a case study of the simulation of a static load
test performed on a test pile. It includes the geological description of the site and the justification of the
choice of parameters introduced in the model. Although limited information is available regarding the
geological and geotechnical conditions of the site, the overall results were quite satisfactory.

Key words: axially loaded pile, numerical modelling, soil-pile interaction, Poulos and Davis, Randolph
and Wroth

RESUMO
O objectivo desta tese analisar o comportamento de estacas isoladas sob carga axial, em termos de
assentamento e mecanismos de transferncia de carga.
Inclui uma reviso literria de dois mtodos baseados na teoria da elasticidade, o mtodo de Poulos e
Davis e o mtodo de Randolph e Wroth, bem como a modelao elstica em estado axissimtrico
utilizando o programa de elementos finitos Plaxis. Os resultados obtidos a partir destes mtodos
esto organizados em bacos da relao carga-assentamento e carga da ponta-carga total em
termos do factor de esbelteza da estaca e da variao de rigidez do solo, e so comparados entre si.
Esta tese inclui tambm dois estudos de comparao que incluem modelao elastoplstica em
estado axissimtrico utilizando Plaxis, considerando o critrio de cedncia de Mohr-Coulomb. O
primeiro estudo trata-se de uma comparao com duas simulaes numricas realizadas
anteriormente em condies idnticas: uma simulao 3D utilizando GEDdyn e uma simulao em
estado axissimtrico utilizando CESAR-LCPC. Este estudo mostra que na simulao em Plaxis o
valor da carga transferida para a base era inferior ao das outras simulaes, apesar de a curva de
carga no topo-assentamento no topo ser semelhante nos trs casos. O segundo um caso de estudo
que consiste na simulao de um ensaio de carga esttico executado numa estaca experimental.
Inclui uma descrio geolgica do local e a justificao da escolha de parmetros introduzidos no
modelo. Apesar da pouca informao disponvel relativa s condies geolgicas e geotcnicas do
local, os resultados foram bastante satisfatrios.

Palavras-chave: estacas sob carga axial, modelao numrica, interaco solo-estaca, Poulos e
Davis, Randolph e Wroth

ACKNOWLEDGEMENTS
Firstly, I want to thank Professor Jaime Alberto dos Santos, for proposing this theme and guiding me
through my research. I appreciate the confidence placed in me by permitting that I develop each
chapter according to my will, allowing me to learn for myself. I also acknowledge the transmission of
the value of discipline, organization and patience at work I will not forget the importance of finishing
a task before beginning the next one.
Secondly, I want to thank Joo Cames, for answering my every question so promptly and thoroughly.
His help with working with finite elements was priceless, teaching me how to find solutions
methodically and observing details critically. He explained to me how useful a tool like Plaxis can be,
and also how easily it can become a black hole. I will never forget that no model works well the first
time you run it. I wish him all the luck with his career, and I know he will excel at whatever he
attempts.
The gratitude I feel towards my parents cannot be expressed. I thank my father for encouraging me to
pursue this career and setting up a fine professional example; he taught me that to be an engineer is
not a job, it is a way of life. I thank my mother for listening with infinite patience and being an endless
source of care and support. I thank both for providing me with the best conditions to study and learn
and never pressing or demanding anything in return. There is nothing more a daughter can ask.
I also thank my grandparents, my brothers and the rest of my family, for the good environment I grew
up in, and for allowing me to discharge whenever I go home. I particularly thank my uncle Antnio
Carlos, for clarifying my doubts whenever I needed; I hope to have inherited at least some of his talent
for engineering.
I thank Marta for being the sister I never had, my most trustworthy friend. I thank her for always being
there, and for so often disregarding her own work to help me with mine. We shared almost everything
the last few years, the very best and the very worst moments. There is no one I would rather have
spent this time with.
I thank all my friends and colleagues who have in one way or another helped during university years,
especially Francisco, Caldinhas, Antnio, Miguel Melo, Ana Bento, Tiago Schiappa, Margarida and
Guilherme. Because of them, Lisbon became a home to me; they made studying and working in
projects much easier and more agreeable. I am proud to belong to such a group of civil engineers, as I
am sure they will all be excellent professionals.

CONTENTS
1

Introduction ...................................................................................................................................... 1
1.1

Context and Motivation ............................................................................................................ 1

1.2

Objectives and Methodology ................................................................................................... 2

Elastic Theory-Based Methods for Analysis of Single Axially Loaded Piles ................................... 3
2.1

Introduction .............................................................................................................................. 3

2.2

Poulos and Davis Method ........................................................................................................ 6

2.3

Randolph and Wroth Method................................................................................................. 16

Comparison between Elastic Theory-Based Methods and the Finite Element Method in the

Analysis of Single Axially Loaded Piles ................................................................................................. 33

3.1

Introduction ............................................................................................................................ 33

3.2

The Model .............................................................................................................................. 34

3.2.1

Geometry ....................................................................................................................... 34

3.2.2

Loading .......................................................................................................................... 36

3.2.3

Materials and Interfaces ................................................................................................ 37

3.2.4

The Mesh ....................................................................................................................... 38

3.3

Initial Conditions .................................................................................................................... 38

3.4

Calculations and Results ....................................................................................................... 38

Numerical Validation of Elastoplastic Modelling of a Single Axially Loaded Pile .......................... 47


4.1

Introduction ............................................................................................................................ 47

4.2

The Model .............................................................................................................................. 48

4.2.1

Geometry ....................................................................................................................... 48

4.2.2

Loading .......................................................................................................................... 48

4.2.3

Materials and Interfaces ................................................................................................ 49

4.2.4

The Mesh ....................................................................................................................... 52

4.3

4.3.1

Water pressure generation ............................................................................................ 52

4.3.2

Initial stress field generation .......................................................................................... 53

4.3.3

Introduction of the Pile ................................................................................................... 55

4.4

Initial Conditions .................................................................................................................... 52

Calculation and Results ......................................................................................................... 56

4.4.1

Sensitivity Analysis of the Interface ............................................................................... 56

4.4.2

Plaxis Results Compared to CESAR-LCPC and GEFdyn Results ............................... 57

4.4.3

Analysis Including Soil Dilatancy ................................................................................... 59

Case Study of a Static Load Test Performed on a Test Pile ......................................................... 61


5.1

Introduction ............................................................................................................................ 61

5.2

Geological Characterization .................................................................................................. 62

5.3

The Static Load Test (SLT) ................................................................................................... 63


i

5.3.1

Preparation .................................................................................................................... 63

5.3.2

Procedure ...................................................................................................................... 64

5.3.3

Instrumentation .............................................................................................................. 67

5.3.4

Results ........................................................................................................................... 68

5.4

Numerical Analysis by Plaxis ................................................................................................. 72

5.4.1

The Model ...................................................................................................................... 72

5.4.2

Initial Conditions ............................................................................................................ 79

5.4.3

Calculations and Results ............................................................................................... 83

Concluding Remarks ..................................................................................................................... 91


6.1

Conclusions ........................................................................................................................... 91

6.2

Further Research ................................................................................................................... 93

Bibliography ........................................................................................................................................... 95
References ........................................................................................................................................ 95
Consulted Bibliography...................................................................................................................... 96
Appendixes ............................................................................................................................................ 97
Appendix A Dimension of the Model Used in the Elastic Simulations ........................................... 97
Appendix B1 Load Settlement Ratio in Terms of the Pile Slenderness Ratio ................................ 98
Appendix B2 Proportion of Load Transferred to the Pile Base in Terms of the Pile Slenderness
Ratio ................................................................................................................................................ 107
Appendix C Load Settlement Curves Determined in the Elastoplastic Modelling for Numerical
Validation ......................................................................................................................................... 116
Appendix D1 Load Settlement Curve Determined in the Elastoplastic Modelling for Comparison
with the Static Load Test ................................................................................................................. 120
Appendix D2 Normal Stress along the Pile in the Elastoplastic Modelling for Comparison with the
Static Load Test ............................................................................................................................... 121
Appendix D3 Shaft Load Curves Determined in the Elastoplastic Modelling for Comparison with
the Static Load Test......................................................................................................................... 122
Appendix D4 Load Settlement Curves Determined in the Elastoplastic Modelling for Comparison
with the Static Load Test ................................................................................................................. 123

ii

LIST OF FIGURES
Figure 2.1: (a) Stresses acting in the soil adjacent to the pile; (b) Stresses acting on the pile; (c)
Stresses acting on a division of the pile. Adapted from (Poulos & Davis, 1980), p. 75. ......................... 4
Figure 2.2: Pile under axial loading relevant parameters ..................................................................... 5
Figure 2.3: Settlement-influence factor for a rigid pile in a semi-infinite incompressible soil, I0, in terms
of the pile slenderness ratio, L/d, and of the relation between the base and the shaft diameters, db/d.
(Poulos & Davis, 1980), p.89. .................................................................................................................. 8
Figure 2.4: Correction factor for the pile compressibility, R k: in terms of the relation between the piles
and the soils Youngs modulus, K, and of the pile slenderness ratio, L/d. (Poulos & Davis, 1980), p.89.
................................................................................................................................................................. 9
Figure 2.5: Correction factor for the finite depth of the layer on a rigid base, R h, in terms of the relation
between the total depth of the soil layer and the length of the pile, h/L, and of the pile slenderness
ratio, L/d. (Poulos & Davis, 1980), p.89. .................................................................................................. 9
Figure 2.6: Correction factor for the Poissons ratio of the soil, R, in terms of the soils Poissons
coefficient, , and of the relation between the piles and the soils Youngs modulus, K. (Poulos &
Davis, 1980), p.89. ................................................................................................................................ 10
Figure 2.7: Tip-load proportion for incompressible pile in uniform half-space, 0, in terms of the pile
slenderness ratio, L/d, and of the relation between the base and the shaft diameters, db/d. (Poulos &
Davis, 1980), p.86. ................................................................................................................................ 10
Figure 2.8: Correction factor for pile compressibility, Ck, in terms of relation between the pile and the
soils Youngs modulus, K, and of the pile slenderness ratio, L/d. (Poulos & Davis, 1980), p.86. ........ 11
Figure 2.9: Correction factor for pile compressibility, C, in terms of the soils Poissons coefficient, ,
and of the relation between the piles and the soils Youngs modulus, K. (Poulos & Davis, 1980), p.86.
............................................................................................................................................................... 11
Figure 2.10: Load settlement ratio in terms of the pile slenderness ratio, for different inhomogeneity
factors and =975, according to the Poulos and Davis method. ........................................................... 14
Figure 2.11: Proportion of the load taken by the pile base in terms of the pile slenderness ratio, for
different inhomogeneity factors and =975, according to the Poulos and Davis method. .................... 15
Figure 2.12: (a) Upper and lower soil layers; (b) Separate deformation patters of the upper and lower
soil layers. Adapted from (Randolph & Wroth, 1978), p.1469. .............................................................. 16
Figure 2.13: Hypothetical variation of the radius of influence of the pile, r m. Adapted from (Randolph &
Wroth, 1978), p. 1471. ........................................................................................................................... 18
Figure 2.14: Load settlement ratio in terms of the pile slenderness ratio for rigid piles, for different
inhomogeneity factors, according to the Randolph and Wroth method. ............................................... 25
Figure 2.15: Load settlement ratio in terms of the pile slenderness ratio for compressible piles, for
different inhomogeneity factors and =975, according to the Randolph and Wroth Method. ............... 26

iii

Figure 2.16: Load settlement ratio in terms of the pile slenderness ratio for rigid and compressible
piles (=975), for different inhomogeneity factors, according to the Randolph and Wroth method. ..... 27
Figure 2.17: Load settlement ratio in terms of the pile slenderness ratio for rigid and compressible
piles, for different inhomogeneity factors and soil-pile stiffness ratios, according to the Randolph and
Wroth method. ....................................................................................................................................... 28
Figure 2.18: Proportion of the load taken by the pile base in terms of the pile slenderness ratio for rigid
piles, for different inhomogeneity factors, according to the Randolph and Wroth method. .................. 29
Figure 2.19: Proportion of the load taken by the pile base in terms of the pile slenderness ratio for
compressible piles, for different inhomogeneity factors and =975, according to the Randolph and
Wroth method. ....................................................................................................................................... 30
Figure 2.20: Proportion of the load taken by the pile base in terms of the pile slenderness ratio for rigid
and compressible piles, for different inhomogeneity factors and =975, according to the Randolph and
Wroth method. ....................................................................................................................................... 31
Figure 2.21: Proportion of the load taken by the pile base in terms of the pile slenderness ratio for rigid
piles, for different inhomogeneity factors and soil-pile stiffness ratios, according to the Randolph and
Wroth method. ....................................................................................................................................... 32
Figure 3.1: Geometry of the model. ....................................................................................................... 35
Figure 3.2: Geometry of the model, L=20m. ......................................................................................... 35
Figure 3.3: (a) Point load at the centre; (b) Point load on the side; (c) Distributed load; (d) Prescribed
displacements. ....................................................................................................................................... 36
Figure 3.4: Vertical displacement field, for 1m prescribed displacement at the pile top, L=20m. ......... 38
Figure 3.5: Vertical stress field, for 1m prescribed displacement at the pile top, L=20m...................... 39
Figure 3.6: Vertical stress field near the pile base, for 1m prescribed displacement at the pile top,
L=20m.................................................................................................................................................... 39
Figure 3.7: Normal stress diagram at the pile top, for 1m prescribed displacement at the pile top,
L=20m.................................................................................................................................................... 40
Figure 3.8: Normal stress diagram at the pile base, for 1m prescribed displacement at the pile top,
L=20m both at the pile and at the soil side of the interface. ............................................................... 40
Figure 3.9: Load settlement ratio in terms of the pile slenderness ratio for homogeneous soils (=1)
and =975. ............................................................................................................................................. 41
Figure 3.10: Load settlement ratio in terms of the pile slenderness ratio for inhomogeneity factor
=0,75 and =975.................................................................................................................................. 42
Figure 3.11: Load settlement ratio in terms of the pile slenderness ratio for inhomogeneity factor =0,5
and =975. ............................................................................................................................................. 43
Figure 3.12: Proportion of the load taken by the pile base in terms of the pile slenderness ratio for
homogeneous soils (=1) and =975. ................................................................................................... 44
Figure 3.13: Proportion of the load taken by the pile base in terms of the pile slenderness ratio for
inhomogeneity factor =0,75 and =975. .............................................................................................. 45
Figure 3.14: Proportion of the load taken by the pile base in terms of the pile slenderness ratio for
inhomogeneity factor =0,5 and =975. ................................................................................................ 46

iv

Figure 4.1: Geometry of the model. ....................................................................................................... 48


Figure 4.2: Distribution of nodes in interface elements and respective connection to 15-node triangular
elements. Adapted from Plaxis Manual (Brinkgreve, 2006). ................................................................. 49
Figure 4.3: Mesh. ................................................................................................................................... 52
Figure 4.4: Diagram of pore pressure. .................................................................................................. 53
Figure 4.5: Initial vertical effective stress field, not including the pile. ................................................... 54
Figure 4.6: Initial horizontal effective stress field, not including the pile. .............................................. 55
Figure 4.7: Total load, load transmitted to the pile shaft and load transmitted to the pile base in terms
of the settlement at the pile top, for smooth and rough interfaces. ....................................................... 56
Figure 4.8: Total load, load transferred to the pile shaft and load transferred to the pile base in terms of
the settlement at the pile top, obtained through Plaxis and GEFdyn. ................................................... 57
Figure 4.9: Total load, load transferred to the pile shaft and load transferred to the pile base in terms of
the settlement at the pile top, obtained through Plaxis (for smooth and rough interfaces) and GEFdyn.
............................................................................................................................................................... 58
Figure 4.10: Plastic points due to the Mohr-Coulomb criterion at the pile base for s=8mm. ................ 58
Figure 4.11: Total load, load transferred to the pile shaft and load transferred to the pile base in terms
of the settlement at the pile top, obtained through Plaxis (including soil dilatancy at the base) and
GEFdyn.................................................................................................................................................. 59
Figure 5.1: Geological profile where the static load test was performed. ............................................. 62
Figure 5.2: Pile and soil layers. Adapted from (Santos, 2005). ............................................................. 63
Figure 5.3: (a) Driving of the temporary casing; (b) Welding of the casing; (c) Introduction of the
reinforcing cage. (Viaponte, 2003). ....................................................................................................... 64
Figure 5.4: Reaction system (Viaponte, 2003). ..................................................................................... 64
Figure 5.5: Loading plan. Adapted from (Santos, 2005). ...................................................................... 65
Figure 5.6: Vibrating wire extensometer welded to the reinforcing cage (Viaponte, 2003). ................. 67
Figure 5.7: Depth of each level of extensometers. Adapted from (Santos, 2005). ............................... 68
Figure 5.8: Load at the pile top, measured by pressure gauges, in terms of time in minutes. Adapted
from (Santos, 2005). .............................................................................................................................. 69
Figure 5.9: Measured load at the pile top in terms of the measured settlement of the pile top. Adapted
from (Santos, 2005). .............................................................................................................................. 69
Figure 5.10: Normal stress at different levels along the pile (Santos, 2005). ....................................... 70
Figure 5.11: Normal stress along the pile shaft measured for load steps 4 and 19. Adapted from
(Santos, 2005). ...................................................................................................................................... 70
Figure 5.12: Lateral stress between different levels along the pile (Santos, 2005). ............................. 71
Figure 5.13: Model geometry. ............................................................................................................... 72
Figure 5.14: Mesh. ................................................................................................................................. 79
Figure 5.15: Diagram of pore pressure. ................................................................................................ 80
Figure 5.16: Initial vertical effective stress field, including removed soil and not including the pile. ..... 80
Figure 5.17: Initial horizontal effective stress field, including removed soil but not the pile. ................. 81
Figure 5.18: Total displacements after the removal of soil at the top. .................................................. 82

Figure 5.19: Load at the pile top in terms of the total settlement at the pile top, given by Plaxis. ........ 83
Figure 5.20: Load at the pile top in terms of the total settlement at the pile top, given by the SLT and
Plaxis, for the first loading cycle. ........................................................................................................... 85
Figure 5.21: Load at the pile top in terms of the total settlement at the pile top, given by the SLT and
Plaxis, for the second loading cycle. ..................................................................................................... 86
Figure 5.22: Load at the pile top in terms of the total settlement at the pile top, given by the SLT and
Plaxis, for both loading cycles. .............................................................................................................. 86
Figure 5.23: Normal stress along the pile shaft at the load peaks (steps 4 and 19), given by the SLT
and Plaxis. ............................................................................................................................................. 87
Figure 5.24: Shaft load between different levels along the pile and total applied load, obtained by
Plaxis. .................................................................................................................................................... 88
Figure 5.25 Shaft load by layer of soil and total applied load, obtained by Plaxis. ............................... 88
Figure 5.26: Total load, load transferred to the pile shaft and load transferred to the pile base in terms
of the settlement at the pile top for the second loading cycle, obtained through Plaxis. ...................... 90

vi

LIST OF TABLES
Table 2.1: Limit pile slenderness ratio between rigid and compressible piles, for different values of the
soil-pile stiffness factor, according to (Fleming, 1992). ......................................................................... 28
Table 3.1: Material Properties. .............................................................................................................. 37
Table 3.2: Youngs Modulus of the soil, E. ............................................................................................ 37
Table 4.1: Material properties (Neves, 2001a). ..................................................................................... 49
Table 4.2: Interface properties used in the CSAR-LCPC and in the GEFdyn simulations. ................ 50
Table 4.3: Interface properties used in the Plaxis simulation. ............................................................... 51
Table 5.1: Loading plan. Adapted from (Santos, 2005). ....................................................................... 66
Table 5.2: Depth of each level of extensometers. ................................................................................. 67
Table 5.3: Mobilized shaft load for load steps 4 and 19. Adapted from (Santos, 2005). ...................... 71
Table 5.4: Pile properties....................................................................................................................... 73
Table 5.5: Soil properties....................................................................................................................... 73
Table 5.6: Pile properties (2). ................................................................................................................ 74
Table 5.7: Soil properties (2). ................................................................................................................ 77
Table 5.8: Analytical base resistance, Rb. ............................................................................................. 77
Table 5.9: Analytical shaft resistance, Rs. ............................................................................................. 78
Table 5.10: Shaft load for load step 4. .................................................................................................. 89
Table 5.11: Shaft load for load step 19. ................................................................................................ 89
Table 5.12: Poulos and Davis estimation. ............................................................................................. 84
Table 5.13: Randolph and Wroth (compressible piles) estimation. ....................................................... 84

vii

viii

SYMBOLS
Latin alphabet

A: area of the cross section of the pile


Ab: area of the pile base
As: area of the pile shaft
CK: correction factor for pile compressibility
C: correction factor for Poissons ratio of soil
c: cohesion of the soil
cu: undrained strength of the soil
: average undrained resistance along the pile shaft
d: diameter of the pile shaft
db: diameter of the pile base
E: Youngs modulus of the soil
Einterface: Youngs modulus of the soil-pile interface
EL/2: Youngs modulus of the soil at the middle of the pile
EL: Youngs modulus of the soil at the pile base
Ep: Youngs modulus of the pile
Eoed: Youngs modulus of the soil for oedometer loading conditions
Eoed, interface: Youngs modulus of the soil-pile interface for oedometer loading conditions
G: shear modulus of the soil
Ginterface: shear modulus of the soil-pile interface
GL/2: shear modulus of the soil at the middle of the pile
GL: shear modulus of the soil at the pile base
h: total depth of the soil layer, i. e. distance between the soil surface and the rigid layer
I: coefficient used in the calculation of the total settlement of the pile
I0: settlement-influence factor for a rigid pile in a semi-infinite incompressible soil (=0.5)
Ib: vertical displacement factor for a pile element due to the normal stress at the pile base
Is: vertical displacement factor for a pile element due to the shear stress at the pile shaft
K: relation between the piles and the soils Youngs modulus
Ks: earth pressure coefficient
L: length of the pile
Nc: end-bearing capacity factor
Pb: load transferred to the pile base
Ps: load transferred to the pile shaft
Pt: total (applied) load
r: horizontal distance to the pile axis
ix

r0: radius of the pile shaft


rm: radius of influence of the pile, i.e. maximum distance past which shear stress is negligible
Rb: resistance of the pile base
Rc: total resistance of a pile under compression
Rh: correction factor for finite depth of layer on a rigid base
Rk: correction factor for pile compressibility
Rs: resistance of the pile shaft
R: correction factor for the Poissons ratio of the soil (when <0.5)
Rinter: interface resistance factor
tinterface: thickness of the soil-pile interface
u: radial displacement of the soil
w: vertical displacement of the soil
wb: vertical displacement of the pile base
ws: settlement of the pile shaft
wt: total vertical displacement (settlement) of the pile
x: horizontal coordinate
z: vertical coordinate (depth)

Greek alphabet

: adhesion factor of the pile shaft


0: tip-load proportion for incompressible pile in uniform half-space (=0.5)
: shear strain
: specific weight
: friction angle of the soil-pile interface
: relation between the radius of influence of the pile and the radius of the pile shaft
: depth factor of interaction between layers
: soil-pile stiffness ratio
: Poissons ratio of the soil
: inhomogeneity factor of the soil
: normal stress
0: normal stress at the pile base
z: vertical stress
: average effective vertical stress along the pile shaft
: shear stress
0: shear stress at the pile shaft
: effective friction angle of the soil

ACRONYMS
CPTU: Piezocone Test
FVT: Field Vane Test
GEFdyn : Go Mcanique Elments Finis Dynamique
P&D: Poulos and Davis
R&W: Randolph and Wroth
SCPTU: Seismic Cone Penetration Test
SLT: Static Load Test
SPT: Standard Penetration Test

xi

xii

1 INTRODUCTION

1.1 Context and Motivation


Piles are deep foundations, necessary when the bearing capacity of shallow foundations is not enough
to ensure the support of the superstructure. This superstructure results in vertical forces, due to its
weight as well as additional loads, which are axially transferred to the pile and, through its shaft and
base, to the soil, possibly reaching a stiffer layer.
The analysis of the load transfer mechanism in single piles under axial loading is therefore an
essential basis for deep foundation design. It is very important that the physical interaction between
pile and soil is carefully studied. The settlement analysis is also fundamental, for the maximum
allowable settlement of a foundation is often the most important criterion in its design. Thus, it should
be estimated accurately.
The behaviour of single piles under axial loading, as far as load distribution and settlement along the
pile are concerned, have been analysed through numerous methods. They can be divided into three
main categories, according to (Poulos & Davis, 1980):
1) Load-transfer methods, which involve a comparison between the pile resistance and the pile
movement in several points along its length;
2) Elastic theory-based methods, which employ the equations described in (Mindlin, 1936) for
surface loading within a semi-infinite mass (such as the Poulos and Davis method), or other
analytical formulations that impose compatibility between the displacements of the pile and of
the adjacent soil for each element of the pile (such as the Randolph and Wroth method);
3) Numerical methods, such as the finite element method.
Elastic theory based methods, such as the ones presented in this work, do not explain the behaviour
of the pile near failure. In this thesis, their results are used in comparison with the results of a finite
element method program, Plaxis 2D version 8. Numerical methods are powerful and very useful tools
when used carefully and calibrated with the appropriate tests. They also constitute a valuable way of
performing a sensitivity analysis of the soil parameters.

1.2 Objectives and Methodology


This thesis has three main objectives:
1) To describe two elastic theory-based methods of analysis of single piles under axial loading;
2) To compare solutions given by finite element 2D elastic modelling of piles with the results of
the elastic theory-based methods;
3)

To perform finite element 2D elastoplastic modelling of single piles under axial loading,
validating its results with former numerical simulations and finally analysing a real case study.

This thesis consists of six chapters, the first being the introduction.
Chapter 2 describes two elastic theory-based methods that analyse the behaviour of single piles under
axial loading: the Poulos and Davis method and the Randolph and Wroth method. This review focuses
in settlement and load transfer mechanisms. Results are organizes in dimensionless charts.
Chapter 3 presents the results of 2D axisymmetric elastic modelling of single piles under axial loading,
also organized in dimensionless charts. These are compared to the ones obtained from the analytical
methods.
Chapter 4 compares the results of 2D axisymmetric elastoplastic modelling of a single pile under
vertical loading with the ones obtained by other authors under similar conditions. The objective is to
validate the used finite element method-based program with other similar ones.
Chapter 5 presents a 2D axisymmetric elastoplastic modelling of a single pile in a real case study,
simulating a static load test performed on a test pile. It includes a geological description of the site and
subsequent choice of the attributed soil parameters that calibrate the numerical model.
Chapter 6 list the concluding remarks derived from this thesis, as well as indications on further
research.

2 ELASTIC THEORY-BASED METHODS


ANALYSIS

OF

FOR

SINGLE AXIALLY LOADED

PILES

2.1 Introduction
In this thesis, two elastic theory-based methods are analysed: the Poulos and Davis method,
introduced by (Poulos & Davis, 1968) and the Randolph and Wroth method, firstly described in
(Randolph & Wroth, 1978).
Elastic theory-based methods usually consist of dividing the pile into uniformly-loaded elements, as
shown in Figure 2.1(a). Shear stress, , acts along the shaft, whereas normal stress, , acts on the
base of the pile, as represented in Figure 2.1(b). These are assumed to be uniform in each division
(as shown in Figure 2.1(c)), and the resultant is equal to the total applied load, Pt. Equilibrium and
compatibility between the displacements of the pile and of the soil adjacent to it are imposed for each
element. Usually, there is no imposition of radial compatibility of displacements between pile and soil,
since it is assumed that there is no horizontal movement (du/dz=0).

(a)

(b)

(c)

Figure 2.1: (a) Stresses acting in the soil adjacent to the pile; (b) Stresses acting on the pile; (c) Stresses acting
on a division of the pile. Adapted from (Poulos & Davis, 1980), p. 75.

The methods are distinct because of the different assumptions made on the distribution of shear
stress along the pile. It may be represented as a single point load acting on the axis of each element
or as a uniformly-loaded circular area at the centre of each element, for instance. In the Poulos and
Davis method, shear stress is considered to be uniformly distributed around the circumference of the
pile, which has proved to be an acceptable assumption, especially for shorter piles, according to
(Poulos & Davis, 1980).
For both elastic theory-based methods, a cylindrical pile is considered, of length L and diameter of the
shaft d. Although the possibility of the shaft diameter, d, and the diameter at the pile base, db,
assuming different values is considered in both methods, that case is not analysed in this thesis.
Therefore, in the rest of this document, the diameter is considered to be constant along the entire pile.
The pile radius is represented by r0, and the cross section area by A. As a general rule, the index s
refers to the pile shaft, and the index b to the pile base.
The soil is considered to be an ideal isotropic elastic mass, being the Youngs modulus, E, and the
Poissons ratio, , its linear elastic parameters that are not influenced by the presence of the pile. The
total depth of the soil layer, i.e. the distance between the soil surface and the rigid layer, is
represented by h. The Youngs modulus of the pile, Ep, is assumed to be constant. The Poissons ratio
of the pile is generally not taken into consideration, as it has negligible effect in the overall behaviour.
In fact, even the Poissons ratio of the soil, , has little effect in the end results.
It is assumed that both the pile and the soil are initially stress-free, and there is no residual stress, i.e.
effects of installation are not taken into consideration in any way. It may be argued that this is an
oversimplification; actually, there are quick ways of simulating the installation, such as employing
adjusted shear factor values, as stated in (Poulos & Davis, 1980). However, that is beyond the scope
of this thesis.
There is an applied axial load Pt in the pile head. The settlement and load results refer to the applied
load only: the difference between specific weights, , of the pile and of the soil is not considered.
Since the conditions of this analysis are purely elastic, the interface between the soil and the pile is
considered to be rigid there is no relative movement between them.

The shear modulus of the soil, G, is used instead of the Youngs modulus, E, in (Randolph & Wroth,
1978) because the soil deforms primarily in shear and also because G is assumed to be unaffected by
whether the load is drained or undrained. The shear modulus of the soil may be obtained from the
Youngs Modulus, through eq. (2.1), which results from Hookes law of isotropic linear elasticity:

(2.1)

The relevant parameters that influence the vertical displacement, w, of a floating pile under axial
loading are stated in eq. (2.2):

(2.2)

These parameters are illustrated in Figure 2.2.


It is useful to have dimensionless solutions for the pile behaviour, so as to simplify and quicken their
employment. Results from the methods subsequently presented are therefore arranged in
dimensionless units, such as the pile slenderness ratio, L/r0, the proportion of load transferred to the
pile base, Pb/Pt, and the load settlement ratio, Pt/(wtr0GL).

Figure 2.2: Pile under axial loading relevant parameters

2.2 Poulos and Davis Method


This method, firstly presented in (Poulos & Davis, 1968), allows a quick estimation of both the
proportion of load which reaches the pile base and the total settlement of a pile in the conditions
described in the last section.
It is necessary to determine the values of the stress acting on the pile (see Figure 2.1) that satisfy the
condition of displacement compatibility. As previously stated, only vertical displacement are
considered. In order to obtain the values of shear stress, , normal stress, , and displacement of the
pile top, i.e. total displacement of the pile, wt, expressions that relate vertical displacement with
unknown stresses must be determined, imposing the compatibility conditions and solving the resulting
equations.
The vertical displacement of the soil adjacent to a pile element due to shear stress at the pile shaft is
given by eq. (2.3):

(2.3)

Where:

d: diameter of the pile shaft


Is: vertical displacement factor for the pile element due to the shear stress at the pile shaft
0: shear stress acting at the pile shaft
E: Youngs modulus of the soil

Considering all n pile elements, the resulting vertical displacement of the soil, wt, is provided by eq.
(2.4):

(2.4)

Where:

Ib: vertical displacement factor for the pile element due to the normal stress at the pile base
0: normal stress acting at the pile base
This expression is valid for piles with constant diameter. The mentioned factors are determined using
the integration of the (Mindlin, 1936) equations for the displacement caused by a point load within a
semi-infinite mass. If the presence of a rigid layer at a certain depth is to be accounted for, then the
factor Is is to be altered accordingly.

For calculating the displacement of the pile elements, only axial compression of the pile is considered.
The vertical equilibrium of a cylindrical pile is provided by eq. (2.5):

(2.5)

Where:
: normal stress acting on the pile (average over the cross section)
r0: radius of the pile shaft
Eq. (2.5) may be applied to the pile top, resulting in eq. (2.6):

(2.6)

Where:

Pt: total applied load


A: area of the pile cross section

Eq. (2.5) may also be applied to the pile base, resulting in eq. (2.7):

(2.7)

The displacement compatibility condition is satisfied by imposing the same displacement for the pile
and the soil in each element (rigid interface).
The load settlement ratio is expressed in terms of a coefficient, I, as shown in eq. (2.8):

(2.8)

However, it has been mentioned that the shear modulus, G, is used instead of the Youngs modulus,
E. Thus, eq. (2.8) may be rewritten as eq. (2.9):

(2.9)

The difference between the general shear modulus, G, and the shear modulus at the pile base, G L, is
clarified as the soil inhomogeneity is taken into consideration, further in this chapter. The load
settlement ratio is from now on represented as in eq. (2.9): Pt/(wtr0GL).
This coefficient I is obtained by multiplying other coefficients, as shown in eq. (2.10):

(2.10)

Where:
I0: settlement-influence factor for a rigid pile in a semi-infinite incompressible soil (=0.5)
Rk: correction factor for pile compressibility
Rh: correction factor for finite depth of layer on a rigid base
R: correction factor for the Poissons ratio of the soil (when <0.5)
The values of I0, Rk, Rh and R are plotted in Figure 2.3, Figure 2.4, Figure 2.5 and Figure 2.6,
respectively.

=1

Figure 2.3: Settlement-influence factor for a rigid pile in a semi-infinite incompressible soil, I0, in terms of the pile
slenderness ratio, L/d, and of the relation between the base and the shaft diameters, db/d. (Poulos & Davis,
1980), p.89.

Figure 2.4: Correction factor for the pile compressibility, Rk: in terms of the relation between the piles and the
soils Youngs modulus, K, and of the pile slenderness ratio, L/d. (Poulos & Davis, 1980), p.89.

The correction factor for the pile compressibility, Rk, is function of the relation between the piles
Youngs modulus, Ep, and the soils, E. This relation is represented by K, as shown in eq. (2.11):

(2.11)

The more relatively compressible the pile, the smaller the value of K.

Figure 2.5: Correction factor for the finite depth of the layer on a rigid base, Rh, in terms of the relation between
the total depth of the soil layer and the length of the pile, h/L, and of the pile slenderness ratio, L/d. (Poulos &
Davis, 1980), p.89.

Figure 2.6: Correction factor for the Poissons ratio of the soil, R, in terms of the soils Poissons coefficient, ,
and of the relation between the piles and the soils Youngs modulus, K. (Poulos & Davis, 1980), p.89.

Figure 2.6 confirms the previous statement that the Poissons ratio of the soil, , has no great
influence in the total settlement of the pile, since the correction factor R varies between 0.8 and 1.0,
for normal cases.
The proportion of load transferred to the pile base, Pb/Pt, for a floating pile may be calculated by eq.
(2.12), first presented in (Poulos, 1972):

(2.12)

Where:
0: tip-load proportion for incompressible pile in uniform half-space (=0.5)
CK: correction factor for the pile compressibility
C: correction factor for the Poissons ratio of the soil
The values of 0, CK and C are plotted in Figure 2.7, Figure 2.8 and Figure 2.9, respectively.

Figure 2.7: Tip-load proportion for incompressible pile in uniform half-space, 0, in terms of the pile slenderness
ratio, L/d, and of the relation between the base and the shaft diameters, db/d. (Poulos & Davis, 1980), p.86.

10

Figure 2.8: Correction factor for pile compressibility, Ck, in terms of relation between the pile and the soils
Youngs modulus, K, and of the pile slenderness ratio, L/d. (Poulos & Davis, 1980), p.86.

The piles compressibility has the effect of decreasing the load transferred to the tip. On the other
hand, the load transferred to the tip tends to increase with the relative stiffness of this stratum, and this
is more pronounced for slender piles.

Figure 2.9: Correction factor for pile compressibility, C, in terms of the soils Poissons coefficient, , and of the
relation between the piles and the soils Youngs modulus, K. (Poulos & Davis, 1980), p.86.

The distance to a rigid layer, h, is not present nor has any influence on any term of eq. (2.12). In fact,
the proportion of load which reaches the pile base is not greatly affected by it, when its value is higher
than 2L, according to (Poulos & Davis, 1980). This must be taken into consideration when comparing
results provided by different methods if the depth h is not to be accounted for, then the limit of 2L
must be respected.
There are obviously other factors that may have influence on the proportion of load that reaches the
pile tip, such as the presence of a pile cap resting on the soil surface, or of enlarged bulbs along the
pile, but they are beyond the scope of this study.

11

A layered or a vertically non-homogeneous soil may also be analysed by using the equations in
(Mindlin, 1936) for a uniform mass, if approximate values of Youngs modulus and Poissons ratio at
various points along the pile are employed.
Thus, the stress distribution is assumed to be unaltered, as if the soil was homogeneous, but the soil
displacement at a point adjacent to the pile is function of the soils Youngs modulus at that point. The
result of the soil-displacement equation changes (see eqs. (2.3) and (2.4)), but not the piledisplacements one.
An average Youngs modulus may be calculated using eq. (2.13):

(2.13)

Where:
Ei: Youngs modulus of layer i
hi: thickness of layer i
n: number of layers/divisions of the soil
This may be used when the soil is divided into different layers but the Youngs modulus does not vary
much. In those cases, the solution may be calculated with this new value of the Youngs modulus and
is very close to the one provided by the finite element method (errors inferior to 15%), according to
(Poulos & Davis, 1980). This approach is an approximation, but its solution is considered to be
accurate enough for practical purposes. It must not be forgotten, however, that this is an
approximation and that it does not provide an accurate solution of the load or settlement distribution
along the pile. Only the total values are considered relevant. The variations in the Poissons ratio
along the depth may be ignored, since, as discussed before, this parameter has little influence in the
total settlement of the pile.
A relevant form of soil non-homogeneity is one in which the shear modulus varies linearly with depth.
A measure of this variation is the inhomogeneity factor, ; it is calculated through eq. (2.14):

(2.14)

In the extreme case of =0.5, the shear modulus at the surface must be null this is called a Gibson
soil.
The factor of inhomogeneity enables a comparison between different types of soil, with more or less
vertical inhomogeneity. Since eq. (2.13) is also applicable in this case, it is used in the calculation of ,
through eq. (2.14), using the relation described in eq. (2.1).

12

In Figure 2.10, results given by eq. (2.9) are plotted, for different values of the soil inhomogeneity
factor, . The value of Poissons coefficient of the soil, , is 0.3. The radius of the pile, r0, is equal to
the unity in every case, for simplification reasons. The length of the pile, L, varies between 4m and
100m. The rigid layer is assumed to be at a distance of 2.5L of the surface (h). The Youngs modulus
3

at the pile base, EL, is equal to 8010 kPa in every case its value varying in the rest of the soil
according to .
This has very little influence in the overall results, since the charts are normalized for the soil rigidity;
thus, it only affects the value of K. K assumes the values of 375, 500 and 750 for =1, =0.75 and
=0.5, respectively. The reason why it is not given a constant value is that it is not possible, if the said
values of soil inhomogeneity are to be tested and simultaneously the shear modulus at the pile base,
GL, is to be the same in all cases, for K is a relation between the piles Youngs modulus and the
average shear modulus along the shaft. Nevertheless, K has not great influence either over the load
settlement ratio or over the proportion of load that is transferred to the pile base for normal values of
Ep and GL. Besides, in (Randolph & Wroth, 1978) a similar relation is presented, the soil-pile stiffness
ratio, , calculated through eq. (2.15):

(2.15)

Since both the Youngs modulus of the pile, Ep, and the shear modulus of the soil at the pile base, G L,
are the same for every case, the soil-pile stiffness ratio is constant and =975.
Below the pile, the Youngs modulus is constant and has the same value as at the pile base. The
Youngs modulus of the soil, E, used in its calculation is the one at the middle of the pile. The Youngs
6

modulus of the pile is 3010 kPa.


The chart in Figure 2.10 was built according to non-linear functions created from the few exact points
given by Figure 2.3 to Figure 2.6, since the intermediate values cannot be interpolated linearly.
Therefore, the load settlement ratio was calculated for each natural number of pile slenderness
between 4 and 100. All those values are grouped in tables in Appendix B1.

13

140
120

Pt/(wtr0GL)

100
80

=1
=0,75

60

=0,5
40
20
0
0

10

20

30

40

50

60

70

80

90

100

L/r0
Figure 2.10: Load settlement ratio in terms of the pile slenderness ratio, for different inhomogeneity factors, =0.3,
h=2.5L and =975, according to the Poulos and Davis method.

The chart shows that the load settlement ratio increases with the pile slenderness ratio. This is
expected, since, when subject to the same conditions, longer piles settle less.
The shapes of the three curves are very similar. In fact, the inhomogeneity factor, , influences the pile
compressibility, K, and the expression of the load settlement ratio, eq. (2.9), only; the distance
between the curves increases slightly with the slenderness ratio, L/r0.
The pile settlement ratio increases with the inhomogeneity factor, . Since the same shear modulus at
the pile base is considered for the three cases, the one with the smallest is the one in which the
shear modulus at the middle of the pile (G, in eq. (2.9)) is the smallest, i.e. the average Youngs
modulus Eav is the lowest. Thus, according to eq. (2.9) and Figure 2.4 and Figure 2.6, the load
settlement ratio will also be the lowest. It is expected that, the lower the , the worst the results, since
eq. (2.13) obviously provides a very gross approximation, the grosser the less homogeneous the soil.
Since there were few exact points to be extracted from the original charts, the error associated with
these charts is considerable. The non-linearity of these functions is the cause of irregularity of the
resulting curves.
In Figure 2.11, results given by eq. (2.12) are plotted, for different values of the soil inhomogeneity
factor, . The conditions of the pile and the soil are identical to the ones described for Figure 2.10.
Once again, it is built from non-linear functions created from the few exact points given by Figure 2.7
to Figure 2.9. Therefore, the proportion of load transferred to the pile tip was calculated for each
natural number of pile slenderness between 4 and 100. All those values are grouped in tables in
Appendix B2.

14

0,40
0,35
0,30

Pb/Pt

0,25
=1

0,20

=0,75
0,15

=0,5

0,10
0,05
0,00
0

10

20

30

40

50
L/r0

60

70

80

90

100

Figure 2.11: Proportion of the load taken by the pile base in terms of the pile slenderness ratio, for different
inhomogeneity factors, =0.3, h=2.5L and =975, according to the Poulos and Davis method.

This chart shows that the proportion of load transferred the pile base decreases non-linearly with the
slenderness ratio: its value reduces quickly for low values of L/r 0, but tends to stabilize. In longer piles,
the shaft plays a more important role in the load transfer mechanism, given its dimension. Thus, less
load reaches the pile base. In fact, some studies consider it to be null, for values of the pile
slenderness ratio higher than a certain limit.
Besides, the values of the load settlement ratio increase inversely with , although very slightly. In
fact, the difference in the shear modulus of the soil distribution only affect C k (Figure 2.8) and C
(Figure 2.9), and in this last case the change is negligible. However, it is natural that, in piles with a
higher value of the relation K, more load is transferred to the tip.
Once again, it should not be forgotten that the information used to build this chart has come from
Figure 2.7 to Figure 2.9, and so there is a significant error associated with it.
Although some of these parameters are taken as independent from each other (as the measure of pile
compressibility, K, and the total depth of the soil layer, h, used in the calculation of the load settlement
ratio), and other factors are not considered, this method is very convenient and adequate for practical
purposes.
The Poulos and Davis method has proved to provide relatively good solutions, considering its
simplicity. Its results for the load settlement ratio are usually slightly higher than the ones given by
numerical methods, i.e. settlement values are lower, according to (Poulos & Davis, 1980). In Chapter
3, the pertinence of this statement is tested.

15

2.3 Randolph and Wroth Method


This method, firstly introduced in (Randolph & Wroth, 1978), has been developed in order to explain
the axial load transfer process between pile and soil. It is particularly useful in cases where the soil is
non-homogenous, since the previously developed methods, the Poulos and Davis method amongst
them, had great limitations in that aspect.
Initially, the shaft and base behaviours are studied separately. An imaginary horizontal plane AB at the
depth of the pile base separates base and shaft, as represented in Figure 2.12(a). Thus, it is
considered that above that plane the soil deforms due to the pile shaft only, and that below the plane
the soil deforms due to the pile base only, as shown in Figure 2.12(b). The deformation above and
below the plane is not compatible and that allows for interaction between the upper and lower layers of
soil. This is a simplification which will obviously not provide the exact solution, but that has proved to
be satisfactory.

(a)

(b)

Figure 2.12: (a) Upper and lower soil layers; (b) Separate deformation patters of the upper and lower soil layers.
Adapted from (Randolph & Wroth, 1978), p.1469.

The soil is considered to be linear elastic. Thus, the effects of installation (residual stresses) are
ignored. As explained before, it is also assumed that the parameters of the soil are not affected by the
installation of the pile.
The deformation of the soil surrounding the pile is similar to shearing of concentric cylinders. The
vertical equilibrium of an element of soil is given by eq. (2.16):

(2.16)

Where:

r: horizontal distance to the pile axis


: shear stress
x: horizontal coordinate
z: vertical stress
z: depth

16

Since the shear stress caused by pile loading is much greater than the vertical stress, the second
summand is insignificant. Thus, the integration of the previous equation gives the shear stress in the
soil surrounding the pile, according to eq. (2.17):

(2.17)

Where:

r0: radius of the pile shaft


0: shear stress at the pile shaft
The shear strain of the soil is calculated through eq. (2.18):

(2.18)

As mentioned before, the displacement is considered to be mainly vertical (the radial component is
negligible), so eqs. (2.16) to (2.18) may be rewritten as eq. (2.19):

(2.19)

When including rm in the equation, an upper boundary of the radius of influence of the place, i.e. the
distance past which shear stress becomes negligible, eq. (2.19) may be rewritten as eq. (2.20):

(2.20)

The factor is as a relation between the radius of influence of the pile and the radius of the pile shaft,
as shown in eq. (2.21):

(2.21)

The pile acts on the layer below as a rigid punch; its effect is more significant, i.e. the lower layer
suffers more deformation, nearer the pile. The lower layer restrains the deformation of the upper layer.
The result is the generation of positive vertical stresses ( z>0, compression), which indicates that, by
the equation of soil equilibrium, shear stresses will decrease more rapidly than linearly, contrary to
what eq. (2.17) implies.

17

The consequence is that the magnitude of /r(r) (the variation of shear stress with distance do the
pile) decreases with depth along the pile and, consequently, so does the value of rm, as represented in
Figure 2.13.
.

Figure 2.13: Hypothetical variation of the radius of influence of the pile, r m. Adapted from (Randolph & Wroth,
1978), p. 1471.

Rigid Piles in Homogeneous Soil


For rigid piles, the shaft settlement, ws, is independent of the depth, since there is no shortening (the
pile is considered as non-compressible). Thus, the shear stress at the pile shaft, 0, must also vary,
increasing with depth, so that the shaft displacement given by eq. (2.20) is constant. Thus, the higher
the shear stress, the smaller the distance at which its value is significant.
The variation of rm with depth is generally disregarded, and its value is taken, in the case of
homogeneous soil, as an average, given by eq. (2.22):

(2.22)

Generally, varies between 3 and 5.


The shear stress at the pile shaft may be considered as described in eq. (2.23):

(2.23)

Assuming both this and the radius of influence of the pile, rm, constant with depth, the settlement
relative to the pile shaft is given by eq. (2.24):

(2.24)

18

Since the behaviour of the pile base resembles a rigid punch, the resulting displacement is obtained
by eq. (2.25):

(2.25)

is the factor of interaction between the upper and lower layers of soil.
The factor refers to the stiffening effect of the soil above the loaded area. The value to attribute to
this factor is subject to discussion, but it is generally accepted that to adopt the unity is adequate,
according to (Randolph & Wroth, 1978).
In (Fleming, 1992), this factor is considered as the relation between the base and shaft diameters, as
shown in eq. (2.26):

(2.26)

This factor is included in the following equations, but is considered equal to the unity whenever
calculations are performed.
In a rigid pile, there is by definition no relative displacement within, and so eq. (2.27) is applicable:

(2.27)

Therefore, every point in the pile has the same settlement.


Besides, the total load is given by eq. (2.28):

(2.28)

For rigid piles in a homogeneous soil, the load settlement ratio may be calculated through eq. (2.29):

(2.29)

Also, the fraction of the load that is taken by the base may be calculated by eq. (2.30):

(2.30)

19

Compressible Piles in Homogeneous Soil


Not always may the pile be considered as rigid. For piles with certain properties, relative displacement
within the pile must be taken into consideration. Thus, for compressible piles, the shaft settlement
varies with depth, since shear stress can no longer be assumed as constant along the pile, as in eq.
(2.31):

(2.31)

The pile is considered as elastic, and its compressive strain in depth may be expressed by terms of
the transmitted load, as in eq. (2.32):

(2.32)

The soil-pile stiffness ratio, , is calculated by eq. (2.15).


It is relevant to point out that the Youngs modulus of the pile, Ep, is taken into account in the
calculations relative to compressible piles only, since for rigid piles it is considered to be infinite.
The transmitted load is related to the variable shear stress on the shaft surface, as in eq. (2.33):

(2.33)

Differentiating and combining the last equations will result in the governing differential equation, as in
eq. (2.34):

(2.34)

And its solution is described in eq. (2.35):

(2.35)

Taking the following measure of pile compressibility, described in eq. (2.36):

(2.36)

20

The constants A and B are found by using the boundary conditions at the base of the pile, and
equation (2.35) may be rewritten as eq. (2.37):

([

(2.37)

-1

The term (r0) is very small (inferior to 0.02 in normal cases) and, through eq. (2.25), eq. (2.37)
may be simplified to eq. (2.38):

(2.38)

It is possible to particularize eq. (2.38), expressing the total settlement of the pile in terms of the
settlement of the base, as in eq. (2.39):

(2.39)

The mentioned simplification may also be used, as in eq. (2.40):

(2.40)

The integration of eq. (2.33) and substitution of the latest equations provides the expression of the
total load supported by the pile in terms of depth, as in eq. (2.41):

))

(2.41)

Thus, the expression of the load settlement ratio is provided by eq. (2.42):

][

(2.42)

This allows for a comparison between the pile compressibility and the load deformation behaviour.
Also, the fraction of the load that is taken by the base may be calculated by eq. (2.43):

(2.43)

21

It is also useful to adapt the former equations to the case of a non-homogenous soil. Only vertical nonhomogeneity is considered the concept of radial inhomogeneity not being developed, given its lack
of relevance for the present study.
As in the chapter referent to the Poulos and Davis method, two types of non-homogeneous soils may
be considered: a layered soil (in which every layer has a constant shear modulus) and a soil in which
the shear modulus varies linearly.
For the case of the layered soil, the shear strain distribution is unaltered (see eq. (2.18)), the shear
stress 0 distribution being obtained by multiplying by an appropriate shear modulus.
Once again, the second type of soil inhomogeneity, in which the shear modulus varies linearly, is more
carefully analysed.

Rigid Piles in Non-Homogeneous Soil


The behaviour of a rigid pile in this type of soil is considered. The shear stress, which would be
assumed constant in a homogenous soil, increases approximately linearly with depth in this case. The
distance at which the shear stresses become negligible, rm, will also decrease.
For a pile in an infinite half space, rm may be calculated by eq. (2.44):

(2.44)

And, for a pile in a space where there is a rigid layer at the depth of 2.5L, which is more commonly
used since it allows a comparison with finite element method programs results, rm may be calculated
by eq. (2.45):

(2.45)

The inhomogeneity factor, , is given by eq. (2.14).


Below the pile base, the shear modulus is considered to be constant. However, analyses using the
finite element method have proved that the difference between this and the case where the shear
modulus continues to increase is negligible (inferior to 5% in the total settlement value).
The value of shear stress may be written as in eq. (2.46):

(2.46)

And thus, the shaft settlement, ws, may be calculated through eq. (2.47):

(2.47)

22

The total load taken by the pile shaft may be written as in eq. (2.48):

(2.48)

Eq. (2.48) may be rewritten as eq. (2.49):

(2.49)

And the load settlement ratio may be written as in eq. (2.50):

(2.50)

The fraction of the load that is taken by the base may be calculated by eq. (2.51):

(2.51)

Compressible Piles in Non-Homogeneous Soil


For compressible piles, eq. (2.42), used for homogeneous soils, may be used for non-homogeneous
soils when modified by the introduction of the inhomogeneity factor, as in eq. (2.52):

][

(2.52)

Also, the fraction of the load that is taken by the base may be calculated by eq. (2.53):

(2.53)

23

A relevant statement is that a pile with L/r0=100 could hardly be considered rigid this would be the
case of a pile with L=100m and d=2m, or L=8m and d=16cm. The authors of this method have not
provided any means of distinction between rigid and compressible piles. According to (Fleming, 1992),
eq. (2.54) can be used as a general rule to determine if a pile is may be considered as rigid:

(2.54)

In this case, for the conditions described, eq. (2.54) is calculated as shown in eq. (2.55):

(2.55)

In the charts that display results from the Randolph and Wroth method presented from now on, a black
line represents this limit, and the charts are divided in Rigid and Compressible areas.

In Figure 2.14, results given by eq. (2.50) are plotted, for different values of the soil inhomogeneity
factor, . The factor of interaction between layers, , is given the value of the unity, and so is the pile
radius, r0, in every case. The value of Poissons coefficient, , is 0.3. The relation between the radius
of influence of the pile and the radius of the pile shaft, , is given by equation (2.21), using values of
the radius of influence of the pile, rm, given by equations (2.22) and (2.45), for homogeneous (=1)
and inhomogeneous (1) soils respectively; thus, the rigid layer is assumed to be at the distance of
3

2.5L from the surface. The Youngs modulus at the pile base, EL, is equal to 8010 kPa in every case
its value varying in the rest of the soil according to . Below the pile, the Youngs modulus is
constant and has the same value as at the pile base.
These conditions are the closest possible to the ones used to analyse the Poulos and Davis method
(see Figure 2.10).

24

140
120

Pt/(wtr0GL)

100
=1

80

=0,75
60

=0,5

40
20
0
0

10

20

30

40

50
L/r0

60

70

80

90

100

Figure 2.14: Load settlement ratio in terms of the pile slenderness ratio for rigid piles, for different inhomogeneity
factors, =0.3, h=2.5L and =975, according to the Randolph and Wroth method.

This chart shows that, for rigid piles, the load settlement ratio increases approximately linearly with the
pile slenderness ratio. The equations that originate these curves show very clearly that the
contribution of the pile base is constant, and that the shaft contribution varies linearly with the pile
slenderness.
Besides, the higher the inhomogeneity factor, , the higher the load settlement ratio. As expected, as
decreases, so does the slope of the line that represents the function.
The values that originate these curves are grouped in tables in Appendix B1.

25

In Figure 2.15, results given by eqs. (2.42) and (2.52) are plotted, for different values of the soil
6

inhomogeneity factor. The Youngs modulus of the pile is 3010 kPa. The conditions are similar to the
ones described for Figure 2.14.

140
120

Pt/(wtr0GL)

100
80

=1
=0,75

60

=0,5
40
20
0
0

10

20

30

40

50

60

70

80

90

100

L/r0
Figure 2.15: Load settlement ratio in terms of the pile slenderness ratio for compressible piles, for different
inhomogeneity factors, =0.3, h=2.5L and =975, according to the Randolph and Wroth Method.

This chart shows that, for compressible piles, the load settlement ratio increases non-linearly with the
pile slenderness ratio. This suggests that the total settlement does not reduce infinitely as the pile
slenderness increases, but instead that it stabilizes, indicating the existence of an asymptote. This did
not happen for rigid piles, as in that case the contribution of the shaft for the load settlement ratio
increased infinitely with the pile slenderness.
For compressible piles, the contribution of the base remains the same. However, with the introduction
of the measure of the pile compressibility, L, the proportion of the load taken by the shaft is reduced.
Thus, the load settlement ratio for a given value is lower if the pile is considered compressible than if it
were considered rigid, and the difference increases with the pile slenderness (see fig. Figure 2.16).
This is an overall more realistic approach for high pile slenderness values.
Once again, the load settlement ratio also becomes higher as the inhomogeneity factor approaches
the unity.
The values that originate these curves are grouped in tables in Appendix B1.

26

Figure 2.16 allows a clear comparison between curves relative to rigid and to compressible piles, as
well as the mentioned division.

140
Rigid

Compressible

120

=1 compressible

100
Pt/(wtr0GL)

=0,75 compressible

80

=0,5 compressible
=1 rigid

60

=0,75 rigid

40

=0,5 rigid

20

0
0

10

20

30

40

50

60

70

80

90

100

L/r0
Figure 2.16: Load settlement ratio in terms of the pile slenderness ratio for rigid and compressible piles (=975),
for different inhomogeneity factors, =0.3, h=2.5L and =975, according to the Randolph and Wroth method.

It is now evident how different is the behaviour of each type of pile, rigid or compressible. The domain
of rigid piles is clearly very limited, according to the division criterion used.
This analysis would not be complete, however, if a soil-pile stiffness ratio sensitivity analysis was not
performed, since the presented charts are not normalized by it. Figure 2.17 shows the load settlement
ratio in terms of the pile slenderness ratio for three different values of soil inhomogeneity factor (=1,
=0.75 and =0.5) and for three different values of the soil-pile stiffness ratio (=3,000, =975 and
=300). This sensitivity analysis is only applied to compressible piles, since the soil-pile stiffness factor
is assumed to be infinite in rigid piles.

27

140

120
=1; =3000

Pt/(wtr0GL)

100

=0,75; =3000
=0,5; =3000

80

=1; =975
=0,75; =975
=0,5; =975

60

=1; =300
=0,75; =300

40

=0,5; =300
20

0
0

10

20

30

40

50
L/r0

60

70

80

90

100

Figure 2.17: Load settlement ratio in terms of the pile slenderness ratio for rigid and compressible piles, for
different inhomogeneity factors and soil-pile stiffness ratios, =0.3 and h=2.5L, according to the Randolph and
Wroth method.

This chart shows that the load settlement ratio is quite sensitive to the soil-pile stiffness factor, for any
given value of the soil inhomogeneity factor. However, it should be borne in mind that, for a constant
Youngs modulus of the pile of 30GPa, a soil-pile stiffness factor of 300 requires the shear modulus of
the soil at the pile base to be equal to 100MPa, whereas a soil-pile stiffness factor of 3,000 requires
the shear modulus of the soil at the pile base to be equal to 10MPa, which is quite a wide range. The
curves relative to =3,000 are obviously closer to the rigid pile solution (see Figure 2.14).
Application of eq. (2.54) provides the boundary values between rigid and compressible piles, for the
given values of soil-pile stiffness factor, shown in Table 2.1.

Table 2.1: Limit pile slenderness ratio between rigid and compressible piles, for different values of the soil-pile
stiffness factor, according to (Fleming, 1992).

L/r0

300

8.66

975

15.61

3,000

27.39

This table shows that the boundary between rigid and compressible piles is also quite sensitive to the
soil-pile stiffness factor (as is, in fact, clear in eq. (2.54)). This gives an idea of the utility of the rigid
solution (up to which value of the pile slenderness ratio the rigid solution is applicable).

28

In Figure 2.18, results given by eqs. (2.30) and (2.51) are plotted, for different values of the soil
inhomogeneity factor. Once again, the conditions are the same as the ones described for Figure 2.14.

0,40
0,35
0,30

Pb/Pt

0,25
=1
0,20

=0,75
=0,5

0,15
0,10
0,05
0,00
0

10

20

30

40

50
L/r0

60

70

80

90

100

Figure 2.18: Proportion of the load taken by the pile base in terms of the pile slenderness ratio for rigid piles, for
different inhomogeneity factors, =0.3, h=2.5L and =975, according to the Randolph and Wroth method.

This chart shows that, for rigid piles, the proportion of load that reaches the pile base decreases nonlinearly with the slenderness ratio: its value reduces quickly for low values of L/r0, but tends to
stabilize.
Besides, the values increase inversely with ; this is expectable, since the contribution of the base to
the load settlement ratio is not affected by and the contribution of the shaft is proportional to (see
eq. (2.50)), as it increases with the shear modulus along the pile (see eq. (2.24)).
The values that originate these curves are grouped in tables in Appendix B2.

29

In Figure 2.19, results given by equations (2.43) and (2.53) are plotted, for different values of the soil
inhomogeneity factor.

0,40
0,35
0,30

Pb/Pt

0,25
=1
0,20

=0,75

0,15

=0,5

0,10
0,05
0,00
0

10

20

30

40

50

60

70

80

90

100

L/r0
Figure 2.19: Proportion of the load taken by the pile base in terms of the pile slenderness ratio for compressible
piles, for different inhomogeneity factors, =0.3, h=2.5L and =975, according to the Randolph and Wroth
method.

This chart shows that, in terms of the proportion of load taken by the pile base, piles regarded as
compressible behave very similarly to the ones considered to be rigid. The shapes of the curves are
almost identical, the only difference being that values relative to the compressible piles are slightly
lower. This is due to the fact that, in this case, the calculation of the contribution of the shaft includes
the variation of shear stress with depth, and its results are higher. Besides, the longer the pile, the
more load is transferred to the shaft, resulting in less load reaching the pile base.
It is interesting that the difference between curves =1 and =0.75 is lower than the difference
between curves =0.75 and =0.5.
The values that originate these curves are grouped in tables in Appendix B2.

30

Figure 2.20 allows a clear comparison between curves relative to rigid and to compressible piles, as
well as the mentioned division.

0,40
Rigid

Compressible

0,35
0,30
=1 compressible

Pb/Pt

0,25

=0,75 compressible
=0,5 compressible

0,20

=1 rigid

0,15

=0,75 rigid
=1 rigid

0,10
0,05
0,00
0

10

20

30

40

50

60

70

80

90

100

L/r0
Figure 2.20: Proportion of the load taken by the pile base in terms of the pile slenderness ratio for rigid and
compressible piles, for different inhomogeneity factors, =0.3, h=2.5L and =975, according to the Randolph and
Wroth method.

Once again, a soil-pile stiffness ratio sensitivity analysis is performed. Figure 2.21 shows the
proportion of load transferred to the pile base in terms of the pile slenderness ratio for three different
values of soil inhomogeneity factor (=1, =0.75 and =0.5) and for three different values of the soilpile stiffness ratio (=3,000, =975 and =300).

31

0,40

0,35

0,30

0,25

=1; =975

Pb/Pt

=0,75; =975
=0,5; =975

0,20

=1; =3000
=0,75; =3000

0,15

=0,5; =3000
=1; =300

0,10

=0,75; =300
=0,5; =3000

0,05

0,00
0

10

20

30

40

50
L/r0

60

70

80

90

100

Figure 2.21: Proportion of the load taken by the pile base in terms of the pile slenderness ratio for rigid piles, for
different inhomogeneity factors and soil-pile stiffness ratios, =0.3 and h=2.5L, according to the Randolph and
Wroth method.

When compared to the case of load settlement ratio, the proportion of load transferred to the pile base
is less sensitive to the soil-pile stiffness ratio, as it is to the soil inhomogeneity factor. Besides, the
difference between rigid and compressible solutions of the proportion of load transferred to the pile
base is also not as relevant as the one of the load settlement ratio.
Thus, it is possible to conclude that the soil-pile stiffness ratio has more influence over the load
settlement ratio than over the proportion of load transferred to the pile base.

The authors of this method state that comparisons made between the results given by this method
and by the finite element method show great similarity. They also declare, however, that this method
2

proves to be unstable for very long compressible piles and is considered unsuitable when (L/r0) />20;
it provides unreliable results for values of the pile slenderness ratio higher than 80. In Chapter 3, tests
will be made to confirm these statements.
The list of this methods advantages includes its simplicity, which allows a good estimation of the total
settlement of a pile with no need of a computer. Also, the non-homogeneity of the soil are more
thoroughly taken into account than in the previous elastic theory-based methods.
In the next chapter, charts provided by the Poulos and Davis method and by the Randolph and Wroth
method are compared to each other and to others, given by the finite element method-based program
Plaxis.
32

3 COMPARISON BETWEEN ELASTIC THEORYBASED

METHODS

ELEMENT METHOD

AND
IN THE

THE

FINITE

ANALYSIS

OF

SINGLE AXIALLY LOADED PILES

3.1 Introduction
Nowadays, the use of finite element method-based programs in structural and geotechnical analyses
is generalised. These are particularly useful in study of inelastic behaviour, when other methods are
no longer applicable.
In this chapter, the finite element method program Plaxis 2D version 8 is used to model single piles
behaviour under axial loading. Its results are compared to the results given by the two described
elastic theory-based methods, the Poulos and Davis method and the Randolph and Wroth method.
The purpose is to obtain curves of the load settlement ratio and of the proportion of load taken by the
pile base in terms of the pile slenderness, traced from the results provided by the finite element
program.

33

3.2 The Model


The analysed piles are cylindrical and isolated and the loading is performed axially. These conditions
allow for the use of symmetry; thus, the single piles behaviour is modelled in two dimensions, using
axisymmetry. 15-node triangular elements are used to create the mesh.

3.2.1 Geometry
The model consists of two materials: the pile and the soil. Since axisymmetry is used, only half of the
problem is represented, as shown in Figure 3.2. The length of the pile, L, defines the rest of the
geometry; its value varies between 4m 100m. The pile radius, r0, is given the value of the unity. The
distance between the surface and the rigid layer, h, is 2.5L. This limit will be used in every analysed
case, and has special importance because the proximity of the rigid layer has great influence in the
pile settlement results. In order to compare Plaxis results with Poulos and Davis, it is required that
this distance must be greater than 2L, so it does not influence the proportion of load that reaches the
pile base. However, the Randolph and Wroths expression of the radius of influence of the pile, r m,
which is empirical and subject to discussion, given by equations (2.22) and (2.45), has proved to be
adequate for that particular distance, and there is no proof of its validity for others.
In Randolph & Wroth (1978), it is recommended that the distance between the symmetry axis and the
vertical outer boundary, l, should be equal or higher than 2L, since in pile settlement analyses
performed with closer limits, results have been affected by it.
The choice of these properties is determined by the two mentioned elastic theory-based methods:
since the purpose of this simulation is to enable a comparison with results provided by them, the
conditions must be as similar as possible.
Appendix A shows the dimensions of the model used for each pile slenderness value, L/r0, which
acquires the same value as the pile length, L.
Figure 3.1 shows the general geometry of the model and Figure 3.2 shows the geometry of the model,
in the case when L=20m.

34

Figure 3.1: Geometry of the model.

Free boundary

Horizontal

Horizontal

displacements

displacements

not allowed

not allowed

Vertical and horizontal displacements not allowed


Figure 3.2: Geometry of the model, L=20m.

35

3.2.2 Loading
Plaxis provides several ways of introducing data into the model. It is important, however, to determine
which is the best way to do so, i.e. which way provides results closer to reality.
As far as loading is concerned, there are three possible means of introducing it: using a point load, a
distributed load or a prescribed displacement. The analysis of the diagram of normal stresses in the
top of the pile provides a good evaluation of the quality of the results given by each one of these
cases. The diagram is expected to be uniform with an increase near the soil-pile interface, resulting
from the interaction between the two materials.
Figure 3.3 shows the qualitative diagram of normal stresses for each of the mentioned cases, for a
generic geometry of the pile and load.

Figure 3.3: (a) Point load at the centre; (b) Point load on the side; (c) Distributed load; (d) Prescribed
displacements.

The observation of the diagrams which result from point loads enables an immediate discard of the
possibility of using this type of loading. Although the diagram referent to distributed load is
considerably better, it is clear that results closer to reality are provided by prescribed displacements.
Therefore, this is the method used in the model: a 1m displacement is applied at the pile top, and the
resulting stress is used in the pile-settlement ratio and the proportion of load transferred to the pile
base charts. This may seem an unrealistic approach, but since Plaxis results showed no sign of
numerical error for w=1m and all the materials are linear elastic, any value of the displacement may be
used as an input.
For the boundaries, standard fixities are used: the bottom boundarys vertical and horizontal
displacements are null (representing the rigid layer), the left and right boundaries horizontal
displacements are null and the upper boundary is free, as represented in Figure 3.2.

36

3.2.3 Materials and Interfaces


Two different materials are used in this analysis: reinforced concrete (referred to as pile) and soil.
The simulation is performed under drained conditions. This is an elastic analysis, and therefore it is
considered that the materials do not yield. Since the difference between the specific weight, , of each
material is not accounted for, and that the initial stress state is not taken into consideration being all
the produced stress a consequence of loading only in the elastic theory-based methods, the specific
weight introduced in the program is null for both materials. The soil-pile interface is rigid (parameter
Rinter=1), i.e. there is no slip or gap (relative displacement) between the two materials, when the load is
applied. The value of the Poissons ratio, , is commonly used. The value of the Youngs modulus of
the pile, Ep, is widely used for reinforced concrete, and the one of the soil, E, is variable. Table 3.1
summarizes the described material properties, and Table 3.2 shows the values of the soils Youngs
modulus, E, used for each case (varying according to the inhomogeneity factor, ). Since the charts
are normalized for the soil rigidity, the results would be the same for other Youngs modulus values, if
the soil homogeneity factor remained the same.

Table 3.1: Material Properties.


2

Material Model

Material Type

E [kN/m ]

Pile

Linear Elastic

Drained

0.15

3010

Soil

Linear Elastic

Drained

0.30

Variable

Rinter

Table 3.2: Youngs Modulus of the soil, E.

=1
2

=0.75
3

4010

6010

4010

8010

8010

Ei [kN/m ]

8010

8010

EL/2 [kN/m ]
2

EL [kN/m ]

=0.5

8010

3
3

The soil-pile stiffness ratio, , calculated by eq. (2.15), will consequently have the value of 975 in every
case.

37

3.2.4 The Mesh


Since geometry changes in every case, it is not possible to define a fixed mesh. The dimension of the
problem varies significantly with the simulation (see Appendix A) and the number of elements alters
accordingly. However, care is taken to ensure that the element size is adequate, the mesh being
refined where the highest gradients of stress occur. In most cases, it is proved to be an efficient
procedure to use transition areas: clusters which have the sole purpose of easing the change between
fine and coarse regions of the mesh. All elements are triangular with 15 nodes.

3.3 Initial Conditions


No water table is taken into account. Besides, as previously mentioned, the specific weights of the soil
and of the pile are considered to be equal to zero. Therefore, both the pore pressure and the stress
fields are null.

3.4 Calculations and Results


As previously mentioned, the initial stress field is null. Therefore, the calculation can be performed in
one phase only, which corresponds to the loading.
In order to compare results provided by Plaxis with the ones presented in former chapters, total
displacement and load at the pile base must be computed and analysed.
The maximum displacement is measured at the pile top, and the load at the pile base is calculated
from the diagram of normal stress at the pile base.
As previously mentioned, the load is applied in the form of a prescribed displacement of 1m. The
resulting displacement field within the model boundaries is displayed in Figure 3.4.

Figure 3.4: Vertical displacement field, for 1m prescribed displacement at the pile top, L=20m.

38

This displacement field implies that the model behaves as expected, as far as the settlement is
concerned; the maximum value occurs at the pile top, and is also quite high along the pile, gradually
decreasing with the distance to the pile.
The vertical stress field is displayed in Figure 3.5 and Figure 3.6.

Figure 3.5: Vertical stress field, for 1m prescribed displacement at the pile top, L=20m.

Figure 3.6: Vertical stress field near the pile base, for 1m prescribed displacement at the pile top, L=20m.

39

As expected, high gradients of stress occur at the pile base, while the rest of the soil mass is less
disturbed. The existence of a stress bulb is evident. The maximum and minimum stress points result
from the soil-pile interaction.
Figure 3.7 and Figure 3.8 represent the normal stress diagram at the pile top and at the pile base,
respectively.

Figure 3.7: Normal stress diagram at the pile top, for 1m prescribed displacement at the pile top, L=20m.

Figure 3.8: Normal stress diagram at the pile base, for 1m prescribed displacement at the pile top, L=20m both
at the pile and at the soil side of the interface.

These diagrams attest for the good behaviour of the model, as far as normal stress is concerned; their
shapes are as expected, showing that the load is transmitted to the pile and also how the pile and the
soil interact. Their smoothness shows that the mesh is adequate.
40

The obtained results for every different model geometry are organized in the following charts.
The curves of the load settlement ratio in terms of the slenderness ratio of the pile are presented in
Figure 3.9 (for homogeneous soils, =1), in Figure 3.10 (for non-homogeneous soils, =0.75) and in
Figure 3.11 (for non-homogeneous soils, =0.5).

140
Rigid

Compressible

120

Pt/(wtr0GL)

100
P&D

80
R&W (rigid)
R&W (compressible)

60

Plaxis

40

20

0
0

10

20

30

40

50
L/r0

60

70

80

90

100

Figure 3.9: Load settlement ratio in terms of the pile slenderness ratio for homogeneous soils (=1), =0.3,
h=2.5L and =975.

In the chart above, which refers to homogeneous soils, there is an obvious similarity between the
results given by Plaxis and by the Randolph and Wroth method for compressible piles. Both indicate
the existence of an asymptote. The difference between these results increases with the slenderness
ratio, and shows that, for the same applied load, vertical displacements calculated through Plaxis are
higher than those calculated by the other methods.
Values referent to the Poulos and Davis method are between the Randolph and Wroth methods
referent to rigid and compressible piles. This is expected, since the Poulos and Davis method refers to
rigid piles, allowing for the introduction of a correction factor for pile compressibility, Rk; this factor
should have a lower (thus, more significant) value in the calculation of the load settlement ratio of the
pile, if results provided by Plaxis are assumed to be more correct solutions. The settlement provided
by the Poulos and Davis method is indeed lower than the one provided by the finite element program,
as stated in (Poulos & Davis, 1980).
As stated in (Randolph & Wroth, 1978), when L/r0>80, the difference between the methods and
numerical results is significant.

41

140
Rigid

Compressible

120
100

Pt/(wtr0GL)

P&D

80

R&W (rigid)
R&W (compressible)

60

Plaxis

40
20
0
0

10

20

30

40

50

60

70

80

90

100

L/r0
Figure 3.10: Load settlement ratio in terms of the pile slenderness ratio for inhomogeneity factor =0.75, =0.3,
h=2.5L and =975.

Similar conclusions may be drawn from this chart, as far as the relation between each methods
results is concerned.
As mentioned before, all elastic theory-based methods, as well as the numerical analysis, provide
lower values of the load settlement ratio in this case, comparing to the homogeneous soil.
For =0.75, the existence of an asymptote is even clearer. The shaft contribution to the load
settlement ratio decreases, i.e. the settlement due to the shaft is higher, since the soil does not
develop as much shaft load as in homogeneous soils, given the same shear modulus at the pile base.

42

140
Compressible

Rigid

120

100
Pt/(wtr0GL)

P&D

80

R&W (rigid)
R&W (compressible)

60

Plaxis

40

20

0
0

10

20

30

40

50

60

70

80

90

100

L/r0
Figure 3.11: Load settlement ratio in terms of the pile slenderness ratio for inhomogeneity factor =0.5, =0.3,
h=2.5L and =975.

As previously mentioned, the Poulos and Davis procedure to account for the soil inhomogeneity is to
assume constant stiffness and use the Youngs modulus at the depth corresponding to L/2. This might
be an oversimplification.
The graph of Plaxis shows that, for values of the pile slenderness higher than 70, the load settlement
ratio starts decreasing, although very slightly. This cannot correspond to the real behaviour, because
the total settlement of the pile may not increase with the pile length, if all the other factors remain
constant. Therefore, this decrease may be explained by the insufficient refinement of the mesh and
numerical inaccuracies; however, these differences may be considered negligible.
The curves of the proportion of load taken by the pile base in terms of the slenderness ratio of the pile
are presented in Figure 3.12 (for homogeneous soils, =1), in Figure 3.13 (for non-homogeneous
soils, =0.75) and in Figure 3.14 (for non-homogeneous soils, =0.5).

43

0,40
Rigid

0,35

Compressible

0,30
P&D

Pb/Pt

0,25

R&W (rigid)

0,20
R&W (compressible)

0,15

Plaxis

0,10
0,05
0,00
0

10

20

30

40

50

60

70

80

90

100

L/r0
Figure 3.12: Proportion of the load taken by the pile base in terms of the pile slenderness ratio for homogeneous
soils (=1), =0.3, h=2.5L and =975.

Results delivered by Plaxis are very consistent with the ones provided by the Randolph and Wroth
method for compressible piles.
They are also extremely close to the ones given by the Poulos and Davis. This proximity is expected
to diminish as the inhomogeneity factor reduces, not only because the results of the finite-element
program are subject to more error, but also due to the simplification of the Poulos and Davis method
for non-homogeneous soils.

44

0,40
Rigid

0,35

Compressible

0,30
P&D

0,25
Pb/Pt

R&W (rigid)

0,20

R&W (compressible)

0,15

Plaxis

0,10
0,05
0,00
0

10

20

30

40

50

60

70

80

90

100

L/r0
Figure 3.13: Proportion of the load taken by the pile base in terms of the pile slenderness ratio for inhomogeneity
factor =0.75, =0.3, h=2.5L and =975.

This chart indicates that the finite element program is more sensitive to the change in the
inhomogeneity factor than the Randolph and Wroth method for compressible piles: the vertical shift of
its curve is more evident. In fact, Plaxis results are the ones which vary the most with the increase of
the inhomogeneity of the soil.

45

0,40
Compressible

0,35
0,30
P&D

Pb/Pt

0,25

R&W (rigid)

0,20

R&W (compressible)

0,15

Plaxis

0,10
0,05
0,00
0

10

20

30

40

50

60

70

80

90

100

L/r0
Figure 3.14: Proportion of the load taken by the pile base in terms of the pile slenderness ratio for inhomogeneity
factor =0.5, =0.3, h=2.5L and =975.

This chart shows the same tendency as the last one: once again, results provided by Plaxis are the
ones with the highest relative increase due to change in the soil inhomogeneity.
For L/r00, it is expected that Pb/Pt1. The chart shows that the Randolph and Wroth methods
solution presents a slight inflexion for L/r0=4. This means the method cannot be applied for small
values of the pile slenderness ratio (when it tends to become a shallow foundation).
Nevertheless, the curves obtained by the different methods correspond well. The Randolph and Wroth
solution for compressible piles proves to be adequate for both load distribution and settlement
estimation. Its solutions for vertically non-homogeneous soils are accurate, and are useful for
expeditious calculations.

46

4 NUMERICAL

VALIDATION

ELASTOPLASTIC MODELLING

OF A

OF

SINGLE

AXIALLY LOADED PILE

4.1 Introduction
In this chapter, a calculation performed through the finite element method-based program Plaxis will
be validated with a set of published results provided by other programs, under the same conditions.
This case was first analysed in (Neves, 2001b). In this document, results provided by the software
CSAR-LCPC are displayed a two-dimensional axisymmetric simulation, as the one performed by
Plaxis. The parameters of the different layers of soil were determined in (Neves, 2001a). Later, the
case was analysed again in (D'aguiar, 2008), using the software GEFdyn, in a three-dimensional
calculation. The results from these two different simulations correspond very well.
Results provided by Plaxis will be compared with the ones provided by GEFdyn. This will validate the
model used in the Plaxis simulation, and particular attention is given to the behaviour of the soil-pile
interface.

47

4.2 The Model


The analysed pile is cylindrical and isolated, subject to axial loading. Therefore, it is once again
possible to perform a two-dimensional axisymmetric simulation. 15-node triangular elements are used.

4.2.1 Geometry
The model consists of five different materials: the pile (made of reinforced concrete) and four layers of
soil.
The pile is 10m long, with a 0.4m diameter. Its shaft crosses three different layers of granular soil,
whose thicknesses are 6.3m, 2m and 2.8m, respectively. There is yet another layer of soil underneath
the pile, with thickness of 8.9m, at the bottom of which a rigid layer is assumed to exist.
The water table is located 0.4m above the pile base.
Since axisymmetry is used, only half of the problem is represented. The total dimensions of the model
2

are 1020m , as used in the previous simulations.


Figure 4.1 shows the geometry of the model.

Figure 4.1: Geometry of the model.

4.2.2 Loading
In the last chapter, it was proved that the best way of introducing load into the model is to use
prescribed displacement. Since the general conditions are the same in this chapter (cylindrical pile,
axial loading at the pile top, etc), the same type of loading will be applied.

48

4.2.3 Materials and Interfaces


Five materials were used in this analysis: reinforced concrete (referred to as pile) and four different
layers of granular soil. In one of the layers, there is a distinction between saturated and unsaturated,
since it is crossed by the water table. The simulation is performed under drained conditions.
The pile is modelled as linear elastic, and the soil as elastoplastic, using the Mohr-Coulomb criterion.
In Table 4.1, the material properties introduced in the program are listed. These are exactly the ones
used in the previous simulations.

Table 4.1: Material properties (Neves, 2001a).

L
(m)
Reinforced
Concrete

Pile
Layer
1
Layer

Soil

2
Layer
3
Layer
4

10.0

6.3

2.0

2.8

8.9

Material
Model
Linear
Elastic
MohrCoulomb
MohrCoulomb
MohrCoulomb
MohrCoulomb

Material

c'
3

'

'

k0

E
2

Type

[kN/m ]

[kN/m ]

()

()

[kN/m ]

Drained

24.0

0.3

2910

Drained

16.7

13

26

0.12

0.562

9150

Drained

18.8

12

23

0.12

0.609

13510

Drained

19.8

14

23

0.07

0.609

13570

Drained

20.0

17

23

0.05

0.609

19300

The soil-pile interaction is modelled through an interface. This is, in the case of 15-node triangular
elements, five pairs of nodes that allow for relative displacement between the two materials that are
linked through it. Figure 4.2 represents the geometry of an interface.

Figure 4.2: Distribution of nodes in interface elements and respective connection to 15-node triangular elements.
Adapted from Plaxis Manual (Brinkgreve, 2006).

In Table 4.2, the interface properties of the layers of soil crossed by the pile shaft used in the previous
analyses are listed. The interface model, cohesion and friction angle were used in both the CSARLCPC and the GEFdyn calculations; however, the information in the last three columns of Table 4.2
refers to the latter simulation only, since these parameters were not mentioned in the former.

49

Table 4.2: Interface properties used in the CSAR-LCPC and in the GEFdyn simulations.

' interface

Material

Interface

c'interface

(m)

Model

Model

[kN/m ]

()

Mohr-

Mohr-

Coulomb

Coulomb

23

0.562

410

Mohr-

Mohr-

Coulomb

Coulomb

20

0.609

410

Mohr-

Mohr-

Coulomb

Coulomb

20

0.609

410

Layer 1

6.3

Layer 2

2.0

Layer 3

2.8

k0

Ginterface

Einterface

[kN/m ]

910

910

910

[kN/m ]

17

17

17

As far as the interface is concerned, the two-dimensional analysis performed by Plaxis provides the
user with limited control of the parameters. Actually, there are only two interface properties that may
be altered directly: the interface strength (Rinter) and the virtual thickness factor. The interface is always
elastoplastic, by the Mohr-Coulomb criterion.
The interface strength, Rinter, may vary between 0.01 and 1.00. The upper limit indicates that the
interface is assumed to be weaker and more flexible than the correspondent soil, since the factor Rinter
relates the interface parameters with the soil parameters, according to eq. (4.1) to eq. (4.3):

(4.1)

(4.2)

(4.3)

It is evident that there is no value of R inter that fulfils these three conditions simultaneously, when a set
of material properties as in Table 4.1 and a set of interface properties as in Table 4.2 are used. For
instance, for the first layer of soil, eq. (4.1) results in Rinter=0.462, eq. (4.2) results in Rinter=0.870 and
eq. (4.3) results in Rinter=3.129 (!) this is due to the fact that, in the previous simulations, the interface
was assumed to be less flexible than the adjacent soil, which is simply not possible in this calculation.
Since the friction angle is more important to the interface behaviour than its cohesion, the results from
eq. (4.2) are used directly, as a first iteration. The Poissons ratio of the interface, for Rinter<1, are fixed
and equal to 0.45.

Table 4.3 lists the interface strength used in each layer of soil, as well as the resulting interface
parameters. The different between the interface cohesion, c, shear modulus, G interface, and Youngs
modulus, Einterface, in Table 4.2 and Table 4.3 is evident.

50

Table 4.3: Interface properties used in the Plaxis simulation.

L (m)

Layer 1

6.3

Layer 2

2.0

Layer 3

2.8

Material

Interface

Model

Model

Mohr-

Mohr-

Coulomb

Coulomb

Mohr-

Mohr-

Coulomb

Coulomb

Mohr-

Mohr-

Coulomb

Coulomb

Rinter

c'interface
2

[kN/m ]

' ()

interface

Ginterface
2

Einterface
2

[kN/m ]

[kN/m ]

0.870

11

23

0.45

3092

8966

0.857

10

20

0.45

4430

12846

0.857

12

20

0.45

4657

13506

The other interface-related option than may be changed by the user in Plaxis calculations is the virtual
thickness factor. This factor, when multiplied by the average element size, results in the virtual
thickness of the interface, tinterface. The virtual thickness influences the interface behaviour because it
affects the relative displacements between the pile and the soil, as described in eq. (4.4) and (4.5).

(4.4)

(4.5)

Eoed, interface is the Oedometer modulus of the interface, and may be calculated through eq. (4.6):

(4.6)

Since the values of both the shear modulus and the Youngs modulus of the interface used in the
Plaxis calculation are significantly lower than the ones used in the two previous simulations, the virtual
thickness of the interface is reduced, in an attempt to produce elastic slips closer to the ones obtained
before i.e. lower values of relative displacement between soil and pile. It is important to notice that
these are not calculated in the same way as in the other softwares, and therefore there is no exact
value of the interface virtual thickness to introduce that may be considered correct. The default value
of the interface virtual thickness factor is 0.1. This was changed to 0.01, which is the lowest possible
value resulting in a ten times lower virtual thickness, whereas the shear modulus and the Youngs
modulus used in the former simulations are over ten times higher than the ones used in the present
simulation. Therefore, only a small improvement is to be expected. This alteration will provide a stiff
interface, allowing for less movement.
A few tests were run using this approach, in order to ensure no numerical error derived from it. Indeed,
the results showed no sign of error, and the difference to the ones using the default interface thickness
was almost insignificant.

51

4.2.4 The Mesh


The mesh is constituted by 15-node triangular elements, in a total of 11559 elements and 95613
nodes, with an average element size of 0.131m. The mesh is much more refined around the pile. To
achieve this, two transition areas are drawn, so as to obtain a good connection between highly refined
and coarse parts of the mesh. This mesh is considerably refined, as may be observed in Figure 4.3.

Figure 4.3: Mesh.

4.3 Initial Conditions

4.3.1 Water pressure generation


The initial conditions of the model refer to both the water pressure and the stress field, before any
calculation is performed.
The pore pressure field due to the water table is generated in the first phase, using the specific weight
3

of the water, =10kN/m . The resulting pore pressure diagram, which is constant along the model, is
displayed in Figure 4.4.

52

Figure 4.4: Diagram of pore pressure.

4.3.2 Initial stress field generation


The initial stress field generation is not as immediate as the pore pressure field.
In the two previous analyses, the simulation was carried out with no generation of shaft load due to the
pile installation. Thus, the weight of the pile was completely transmitted to the base, without shear
stress being developed along the shaft.
In order to replicate these conditions, the ideal procedure would be to generate the initial stress field
with the different soil layers as well as the pile, keeping the interface off, so there would be no
relative movement between the pile and the soil. However, the normal stress generation method (the
k0-procedure) carried out by Plaxis provides vertical equilibrium (with the self-weight of the soil) only,
horizontal equilibrium not being ensured, since horizontal stresses are generated from the introduced
value of the coefficient of lateral earth pressure, k 0. Thus, if two different materials are introduced in
the model at the same depth, the k0-procedure does not ensure horizontal equilibrium, according to
(Brinkgreve, 2006), the Plaxis Manual. Given that neither the specific weight nor the k 0 coefficient of
the pile are the same as the soils, and that the geometry of the problem implies the existence of pile
material and soil material at the same depth i.e., horizontal equilibrium cannot be achieved this
apparently ideal way of generating initial stress does not provide adequate results.
Another method of stress generation is available in Plaxis gravity loading. This has the advantage
of granting equilibrium. However, in this type of stress generation, the coefficient of lateral earth
pressure, k0, is calculated through the Poissons ratio, , as in eq. (4.7), and not through the friction
angle, , as in eq. (4.8):

(4.7)

(4.8)

53

In the case of layer 1, for instance, eq. (4.7) results in k0=0.136 (which is not very realistic) and eq.
(4.8) results in k0=0.562. These are indeed very different values. That is the reason why Plaxis
manuals recommend that the Poissons coefficient be adjusted so that eq. (4.7) provides the desired
value of k0. Since the objective of this work is to simulate the conditions of the previous analyses,
attributing a different Poissons coefficient to the soil would not be appropriate, and thus this type of
stress generation should not be used.
For these reasons, and since in this simulation the effects of the installation in the soil cannot be
considered, the initial stress field must be generated from the different layers of soil only, using the k 0procedure. Using the information contained in Table 4.1, the initial stresses were then generated, and
the resulting vertical and horizontal effective stress fields are shown in Figure 4.5 and Figure 4.6,
respectively.

Figure 4.5: Initial vertical effective stress field, not including the pile.

54

Figure 4.6: Initial horizontal effective stress field, not including the pile.

These two figures show satisfactory stress fields, with increasing stress values along the depth, and
constant along the radius. The values correspond to the ones calculated manually, with an error in the
2

order of 0.01kN/m . Thus, this initial stress field is considered to be adequate and is used in the
following calculations.

4.3.3 Introduction of the Pile


The question of how to include the pile in the model is yet to be addressed. The assumption that the
adjacent soil remains undisturbed by the pile installation implies that no shear stress in the soil-pile
interface occurs before the first loading, and consequently that the totality of the pile weight reaches
the base. In order to achieve this, the soil-pile interface should be considered as completely smooth, in
the first step of the calculation, which corresponds to the introduction of the pile with no loading. In the
program, this is achieved by replacing the volume of soil material which occupied the place of the pile
during the initial stress generation by pile material, while maintaining the interface strength R inter set to
the minimum value 0.01, since the introduction of 0.00 is not allowed. This is an artificial step, and
the proper interface strength Rinter value is restored when loading is applied. Immediately after it, the
incremental loading starts, with a null displacement field that is achievable by selecting the option
reset displacements.
Although in the program this step is already part of the calculation, and not of the initial conditions, it is
included in this section because, along with the initial stress generation, it is common to all the
calculations performed subsequently.

55

4.4 Calculation and Results


The following calculations are performed with incremental loading, the size of each step increasing
gradually. This is a way to avoid numerical errors that might result from applying all the load at once.
The result of each increment is displayed as a mark in the correspondent curve. All the points that
constitute the charts are shown in Appendix C.

4.4.1 Sensitivity Analysis of the Interface


In a first observation, it seems clear that the most evident variation of the model used in Plaxis from
the ones used in the two previous calculations (by CESAR-LCPC and GEFdyn) is the soil-pile
interface.
Therefore, a sensitivity analysis of the interface is performed. Two interfaces with extreme values of
interface strength are tested: Rinter=0.01 and Rinter=1.00 (rigid interface). All the other parameters used
are the ones described in the previous section.
In Figure 4.7, the total load Pt (sum of the applied load and the pile total weight), shaft load Ps and
base load Pb in terms of the displacement at the pile top wt are plotted, for the cases of smooth and
rigid interfaces.

700
650
600
550

P (kN)

500
450

Pt Rinter=0,01

400

Ps Rinter=0,01

350

Pb Rinter=0,01
Pt Rinter=1

300

Ps Rinter=1

250

Pb Rinter=1

200
150
100
50
0
0

10

15
wt (mm)

20

25

30

Figure 4.7: Total load, load transmitted to the pile shaft and load transmitted to the pile base in terms of the
settlement at the pile top, for smooth and rough interfaces.

This chart shows that the applied load necessary to cause a certain displacement is much higher
when the interface is rigid. The curves of the load transferred to the pile base are very close, indicating
56

that, as expected, the interface strength has a much greater influence over the shaft load than over
the base load. The shaft load is, as expected, very small in the case of smooth interface.
Since the calculated interface strengths are Rinter=0.87 and Rinter=0.857, the correspondent curves are
expected to be closer to the rigid interface ones.

4.4.2 Plaxis Results Compared to CESAR-LCPC and GEFdyn Results


The calculation with the interface values shown in Table 4.3 is now performed.
In Figure 4.8, the curves of total load, load transmitted to the pile shaft and load transmitted to the pile
base obtained by Plaxis are plotted, along with the ones obtained by CESAR-LCPC and GEFdyn.

500
450
400
350
Pt CESAR-LCPC
Ps CESAR-LCPC
Pb CESAR-LCPC
Pt GEFdyn
Ps GEFdyn
Pb GEFdyn
Pt Plaxis
Ps Plaxis
Pb Plaxis

P (kN)

300
250
200
150
100
50
0
0

10

15
wt (mm)

20

25

30

Figure 4.8: Total load, load transferred to the pile shaft and load transferred to the pile base in terms of the
settlement at the pile top, obtained through Plaxis and GEFdyn.

The curves of total load show correspond very well. However, there is an evident discrepancy in the
other two curves results provided by Plaxis indicate a lower base load, i. e. the base resistance is
not as mobilized as in the other models. As proved in Figure 4.7, changing the value of the interface
strength, Rinter, barely alters these results, since it affects the load that reaches the pile base very little.
In fact, the curve that corresponds to the maximum possible base load (smooth interface, R inter=0.01)
is still below the one given by GEFdyn, as shown in Figure 4.9. Therefore, the calculated value of the
interface strength may be considered acceptable, and the inconsistency of results must be explained
by something else.

57

700
650
600
550
500

P (kN)

450

Pt Plaxis Rinter=0,01

400

Pb Plaxis Rinter=0,01

350

Pt Plaxis Rinter=1

300

Pb Plaxis Rinter=1

250

Pt GEFdyn

200

Pb GEFdyn

150
100
50
0
0

10

15
wt (mm)

20

25

30

Figure 4.9: Total load, load transferred to the pile shaft and load transferred to the pile base in terms of the
settlement at the pile top, obtained through Plaxis (for smooth and rough interfaces) and GEFdyn.

In order to understand the difference of the load transfer behaviour between the GEFdyn model and
the Plaxis model, the failure mechanism is analysed. The plastic points due to the Mohr-Coulomb
failure envelope are displayed in Figure 4.10, for a prescribed displacement of s=8mm (when the
interface along the entire pile is in the plastic state).

Figure 4.10: Plastic points due to the Mohr-Coulomb criterion at the pile base for s=8mm.

This figure shows that the developed failure mechanism includes a wedge. The distribution of the
plastic points does completely resemble the failure mechanism normally attributed to piles under
compression. This indicates that the discrepancies shown in Figure 4.8 might be a consequence of the
difference between the finite element algorithms.
58

4.4.3 Analysis Including Soil Dilatancy


As an attempt to improve these results, a calculation including dilatancy of the soil beneath the pile
base is performed. It is expected that if the soil at the pile base has a positive dilatancy angle, the
base resistance will be more easily mobilized as the load is applied, since the soil will expand. The
introduced value for the angle of dilatancy, , is the same as the friction angle, (==23).
Figure 4.11 shows the results of this calculation, compared to the ones given by GEFdyn.

500
450
400
350
Pt GEFdyn

P (kN)

300

Ps GEFdyn
250

Pb GEFdyn

200

Pt Plaxis

150

Ps Plaxis
Pb Plaxis

100
50
0
0

10

15

20

25

30

wt (mm)
Figure 4.11: Total load, load transferred to the pile shaft and load transferred to the pile base in terms of the
settlement at the pile top, obtained through Plaxis (including soil dilatancy at the base) and GEFdyn.

This chart shows that introducing dilatancy as a soil property does indeed have the effect of mobilizing
the base. In this case, an angle of dilatancy equal to the friction angle is still not enough to obtain
results identical to the ones given by GEFdyn; however, it does approximate the base load curves,
and the difference seems to decrease with the loading.

The simulation performed in this chapter led to the conclusion that there are difficulties in modelling
the behaviour of the pile base in Plaxis, since in this program the base resistance is not as mobilized
as in CESAR-LCPC and GEFdyn.

59

60

5 CASE STUDY

OF A

STATIC LOAD TEST

PERFORMED ON A TEST PILE

5.1 Introduction
In this chapter, a calculation performed by the finite element method-based program Plaxis is
compared to the results of a static load test performed on a test pile.
The static load test is part of a planned set of tests that support the design of the project
Modernizao da Linha do Norte, Sub Troo 14 Azambuja/Vale de Santarm, Viaduto de Santana
do Cartaxo, the construction of a viaduct. It was performed by Mota-Engil, Engenharia e Construo,
S. A. and the results were analysed by Instituto de Engenharia de Estruturas, Territrio e Construo
do Instituto Superior Tcnico (ICIST).
The objective of a static load test is to analyse the load transfer process between the pile and the soil.
The results are therefore carefully analysed and conclusions on the soil behaviour are drawn.

61

5.2 Geological Characterization


The geological prospection included the excavation of 37 boreholes that enabled the execution of
SPT, CPTU, SCPTU and FVT tests, as well as laboratory tests, which included edometric and triaxial
tests. The boreholes were executed until a hard layer was reached.
These tests allowed the distinction of several soil layers, which constitute the lithological profile of the
area. According to the report by (Viaponte, 2003), four different lithostratigraphic units are identified,
following in the order of crescent depth:
- Superficial deposits: dark brown soil, formed by silty sand and sandy organic clay;
- Recent alluvial deposits: constituted by soft grey organic clay (0<NSPT<1), green and grey silty and
clayey sand (2<NSPT<14) or grey silty sand (7<NSPT<17);
- Pleistocene/Miocene: formed by brown and grey very stiff argillaceous marl (10<NSPT);
- Miocene: formed by calcareous marl, constitutes a hard layer in which deep foundations may be
embedded.
The water table in this location is quite high.
In the geological profile drawn from the information given by the borehole which is closest to the test
site (referenced SC22, made by Geocontrole), the referred lithostratigraphic units may be identified,
but the overall profile is still quite different from the one of the place where the pile was tested, made
by Mota-Engil.
For modelling purposes, the chosen geological profile is the one provided by Mota-Engil, which
corresponds to the exact place where the static load test was performed and is shown in Figure 5.1.

Figure 5.1: Geological profile where the static load test was performed.

62

5.3 The Static Load Test (SLT)


5.3.1 Preparation
The static load test (SLT) was performed on a cylindrical concrete pile with a total length of 40.6m and
diameter of 800mm. The pile cap was built 2.1m below the soil surface, and the shaft was extended to
the hard layer. This geometry is displayed in Figure 5.2.

Figure 5.2: Pile and soil layers. Adapted from (Santos, 2005).

A retrievable casing tube was necessary due to the position of the water table and to the properties of
the first layers of soil, which are very soft. This temporary casing extended to the depth of 32.0m (or
29.9m below the pile top) as shown in Figure 5.2, since the soil was considered to be consistent
enough for it to be unnecessary from that point on. As shown in Figure 5.3(a), the casing was
introduced into the soil using static pressure and small rotations, in a gradual way, by welding several
pieces until the necessary length had been achieved (Figure 5.3(b)). The soil inside the casing was
constantly being removed. When the casing reached the depth of 32.0m, the reinforcing cage was
introduced as displayed in Figure 5.3(c).

63

(a)

(b)

(c)

Figure 5.3: (a) Driving of the temporary casing; (b) Welding of the casing; (c) Introduction of the reinforcing cage.
(Viaponte, 2003).

The casing was removed as the pile was being concreted. The concrete had a resistance class of
C35/45 and a Youngs modulus of Ep=33.5GPa. When the concreting process ended, the pile top was
cleaned and the cap was built.

5.3.2 Procedure
th

The execution of the test began on January 20 2005 at 11.10 a.m. Care was taken to ensure that
enough time had passed since the building of the pile to let the vertical displacements and the
properties of the soil to stabilize.
Two hydraulic jacks were used; they exerted pressure against the reaction system, which was
constituted by a metallic truss, since these structures are rigid enough not to deform excessively as
the load is applied, thus being very efficient. The foundations of this reaction system are constituted by
a total of four piles, built by following the same procedure as for the tested pile, described above.
Figure 5.4 shows all these elements.

Figure 5.4: Reaction system (Viaponte, 2003).

64

The usual procedure for static load tests is to apply a maximum load of twice the calculated service
load, Pser; in this case, the correspondent value would be 2Pser=22,800kN=5,600kN. However, since
the maximum load that could be applied to the reaction system was 8,000kN (its maximum
resistance), it was decided to take advantage of that limit.
The test was performed incrementally, the loading and subsequent unloading being executed twice;
the first loading lasted until service load (2,800kN) was reached, and the second until the reaction
system limit (8,000kN) was reached. Before proceeding to the next load step, stabilization criterions
were applied. In the loading phases, it was ensured that either the settlement rate was below 0.3mm/h
or the difference between two consecutive displacement results was under 5% of the total settlement,
before continuing to the next increment. Besides these two criterions, time limits were also
established, the most important being that the step that corresponded to the first peak had to last at
least 2h and the one that corresponded to the second peak had to last at least 12h. During the
unloading, the only criterion was that each step should last 15min, except for the last step of each
loading cycle, which in the first unloading had to last at least 2h and in the second unloading had to
last at least 4h.
Table 5.1 shows the applied load in each load step and its duration. Figure 5.5 traduces this
information in a chart, considering the minimum possible duration for each load step.

8000
7000
6000

Load (kN)

5000
4000
3000
2000
1000
0
0

10

15

20

25

30

Time (h)
Figure 5.5: Loading plan. Adapted from (Santos, 2005).

65

Table 5.1: Loading plan. Adapted from (Santos, 2005).


Duration
Load Step

Applied Load Pt

Proportion of Applied Load Pt/Pser

Minimum

Maximum

700

0.25

30min

4h

1,400

0.50

30min

4h

2,100

0.75

30min

4h

2,800

1.00

2h

4h

2,100

0.75

15min

1,400

0.50

15min

700

0.25

15min

0.00

2h

4h

700

0.25

30min

4h

10

1,400

0.50

30min

4h

11

2,100

0.75

30min

4h

12

2,800

1.00

30min

4h

13

3,500

1.25

30min

4h

14

4,200

1.50

30min

4h

15

4,900

1.75

30min

4h

16

5,600

2.00

30min

4h

17

6,300

2.25

30min

4h

18

7,000

2.50

30min

4h

19

8,000

2.86

12h

20

7,000

2.50

15min

21

5,600

2.00

15min

22

4,200

1.50

15min

23

2,800

1.00

15min

24

1,400

0.50

15min

25

0.00

4h

66

5.3.3 Instrumentation
As previously mentioned, the load was applied through two hydraulic jacks. The capacity of the jack
should exceed the maximum applied load by a considerable margin; for very high values of maximum
load, as in this case, more than one may be necessary.
For the load measurement at the pile top, two different methods were used, so there would be two
sets of results; this redundancy assesses the quality of the results. The first was a pressure gauge,
and the second a load cell. The two sets of results corresponded well and the values provided by the
pressure gauge were the ones used in the interpretation of the test results.
For the settlement measurement at the pile top, two independent methods were used: the first was a
set of dial gauges and transducers and the second was optical levelling. At the pile top, two
mechanical dial gauges and two digital transducers were installed. They were placed on a metallic
structure, and the topographic measurements showed no relative displacement between this and the
reference level during the test. Both the results from the set of dial gauges and transducers and from
the optical levelling corresponded very well.
For the displacement measure along the pile shaft, and consequent determination of the mobilized
resistance, vibrating wire extensometers were used, by welding to the reinforcing cage, as shown in
Figure 5.6. Sets of three extensometers were placed at eight different depths, as shown in Table 5.2
and displayed in Figure 5.7.

Figure 5.6: Vibrating wire extensometer welded to the reinforcing cage (Viaponte, 2003).

Table 5.2: Depth of each level of extensometers.

Level

Depth (m)

20.7

25.7

29.7

31.7

32.7

36.7

37.7

40.7

67

Figure 5.7: Depth of each level of extensometers. Adapted from (Santos, 2005).

Due to the fact that the records at levels 1, 3 and 7 were unusual (unexpectedly high or low, or with
sudden peaks), only the normal stresses provided by the extensometers at the other levels are
considered.

5.3.4 Results
There are three sets of results which must be analysed to understand the behaviour of the soil in this
static load test:
1)

The first is the set of results of the applied load at the pile top, provided by the pressure

gauges. Due to load losses, the measured load may not be exactly the same as the applied load.
Figure 5.8 shows the measured load in terms of the time. When compared to the plan (see Figure
5.5), it becomes evident that the test took longer than the minimum scheduled amount of time, and
that the load had to be adjusted several times to achieve the intended value, as in Pt=7,000kN.

68

Figure 5.8: Load at the pile top, measured by pressure gauges, in terms of time in minutes. Adapted from
(Santos, 2005).

2)

The second set of results provides the vertical displacement at the pile top, measured by the

dial gauges and transducers. These corresponded very well. Figure 5.9 displays the applied load at
the pile top measured by the pressure gauges (shown in Figure 5.8) in terms of the average
settlement measured by the dial gauges and transducers.

8 000
7 000
6 000

Pt (kN)

5 000
4 000
3 000
2 000
1 000
0
0

10

20

30
wt (mm)

40

50

60

70

Figure 5.9: Measured load at the pile top in terms of the measured settlement of the pile top. Adapted from
(Santos, 2005).

69

3)

The results of normal stress along the pile shaft, provided by the vibrating wire extensometers.

As previously mentioned, only 5 out of 8 levels of extensometers provided acceptable results. It should
be borne in mind that this kind of instrument measures the increase of extension with load along the
pile; normal stress is calculated from its results. Figure 5.10 shows the normal stress measured at five
different levels and Figure 5.11 shows the normal stress along the depth for service load (2,800kN,
load step 4) and for maximum load (8,000kN, load step 19).

L2

L4
L5
L6

L8

Figure 5.10: Normal stress at different levels along the pile (Santos, 2005).

2000

N (kN)
4000

6000

8000

0
5
10

z (m)

15
Load Step 4
20
25

Load Step 19

30
35
40
45
Figure 5.11: Normal stress along the pile shaft measured for load steps 4 and 19. Adapted from (Santos, 2005).

70

Figure 5.12 shows the lateral stress measured between two levels, calculated from the normal
stress, and Table 5.3 shows the estimated mobilized shaft resistance measure at load steps 4 and
19.

L45

L24
L56

L68

Figure 5.12: Lateral stress between different levels along the pile (Santos, 2005).

Table 5.3: Mobilized shaft load for load steps 4 and 19. Adapted from (Santos, 2005).

Mobilized Shaft Load (kPa)


Between levels

Load Step 4

Load Step 19

2 and 4

15 to 25

60 to 70

4 and 5

100 to 110

180 to 200

5 and 6

15 to 25

100 to 110

6 and 8

20 to 30

130

The chart in Figure 5.12 appears to indicate that, between levels 6 and 8, the mobilized shaft
resistance has not reached its maximum for the maximum applied load. Given that this corresponds to
the last layer of soil (a very consistent clay), it makes sense that the maximum resistance has not
been reached yet.
Also, it is interesting to notice that the mobilized shaft resistance between levels 4 and 5 is so high
compared to the mobilized shaft resistance between levels 5 and 6, given that the same soil (very stiff
clay) is measured. This is probably due to the effect of the casing (that reached the distance between
levels 4 and 5), which may have disturbed the stress state of the soil at its bottom.

71

5.4 Numerical Analysis by Plaxis


The static load test is now modelled by the finite element-based program Plaxis, in order to compare
its results with the ones provided by the different types of instrumentation described in the previous
section. The geological profile in the model is identical to the one given by the static load test report,
and the different layers of soil are related to the lithostratigraphic units described above.
The analysed pile is cylindrical and isolated, subject to axial loading. Therefore, it is possible to
perform a two-dimensional axisymmetric simulation. 15-node triangular elements are used.

5.4.1 The Model


1) Geometry

The model consists of nine different materials: the pile (made of reinforced concrete) and eight layers
of soil.
The pile is 40.6m long, with a 0.8m diameter. The pile top is not located at the soil surface; its cap was
built 2.1m below. The pile shaft crosses eight different layers of soil. The water table is located 3.4m
below the soil surface, and 1.3m below the pile top; this depth corresponds to 0.4m below the base of
the first layer of soil. For simplification reasons, since 0.4m is a very small length compared to the
overall size of the model and thus difficult to mesh, the water table is introduced so as to coincide
precisely with the base of the first layer of soil.
Since axisymmetry is used, only half of the problem is represented. The chosen dimension of the
2

model is 5055m , since the horizontal dimension should be at least as long as the pile
(50.0m>40.6m) and the distance from the pile base to the lower boundary should be enough not to
affect the behaviour (in this case, the distance of 12.3m is considered enough, for the last layer is very
resistant).
Figure 5.13 shows the geometry of the model.

Figure 5.13: Model geometry.

72

2) Loading

The general conditions in this case are the same as in the previous simulations: cylindrical pile, axial
loading at the pile top, etc. Therefore, once again, the loading is applied through prescribed
displacements.

3) Materials and Interfaces


Nine materials are used in this analysis: reinforced concrete (referred to as pile) and eight different
layers of soil. The soil properties vary considerably from layer to layer, but there is a clear increase in
consistency with depth. The first soil layer is unsaturated, and the seven others are saturated. The pile
is modelled as linear elastic, and every layer of soil as elastoplastic, according to the Mohr-Coulomb
criterion or to the Tresca criterion, depending on if it is drained or undrained, respectively. Table 5.4
and Table 5.5 describe the pile and the soil layers, respectively.

Table 5.4: Pile properties.

Material

L (m)

d (m)

Designation

Material Model

Reinforced concrete

40.6

0.8

Pile

Linear elastic

Table 5.5: Soil properties.

Layer

L (m)

Designation

Material Model

3.0

Superficial deposit

Mohr-Coulomb

12.0

Soft clay

Tresca

9.0

Medium clay

Tresca

4.0

Firm clay

Tresca

2.5

Gravel

Mohr-Coulomb

6.0

Very stiff clay

Tresca

4.0

Dense sand

Mohr-Coulomb

14.5

Hard clay

Tresca

Unlike the case of the last chapter, the information about each material type is very limited. In some
cases, (Viaponte, 2003) suggests a value or a range of values; in others, there is no indication. The
attributed soil properties for each case is described and discussed below.

- Pile
(Santos, 2005) indicates for the piles Youngs modulus the value of Ep=33.5GPa. For the other
properties, values that are generally accepted for reinforced concrete piles are used: specific weight
3

equal to =25kN/m and Poissons ratio of =0.3.


In Table 5.6, the pile properties introduced in the program are listed.

73

Table 5.6: Pile properties (2).


3

Material

Designation

[kN/m ]

E [MN/m ]

Rinter

Reinforced concrete

Pile

25.0

0.3

33500

1.000

- Layer 1: Superficial Deposit


There is little information available about this material. It is assumed that it corresponds to the
superficial deposit described in the geological characterization. In fact, its properties are not expected
to have great influence in the final results, since its thickness is 3m only, and the interface with the pile
only 0.9m long; besides, being the top layer, it is probably the least resistant.
3

It is modelled under drained conditions, =18kN/m , E=25MPa, =0.3 and friction angle =30. For the
interface strength factor, Rinter, it is assumed that the most important parameter is the friction angle
(see eq. (4.2)). Generally, the friction angle of the soil-pile interface is taken as eq. (5.1):

(5.1)

Thus, through eq. (4.2), the correspondent interface strength factor is Rinter=0.63.
- Layer 2: Soft clay
This type of soil is assumed to be a recent alluvial deposit (a1) grey and very soft to soft organic clay
(0<NSPT<1). (Viaponte, 2003) suggests, for this soil, =16kN/m

and E=59MPa. These values

correspond to a coefficient of earth pressure k 0=0.55; this is thus imposed in the program, while the k 0
values of all the other soil layers are the ones generated by default.
Being clay, it is modelled under undrained conditions. For recent alluvial deposits, (Viaponte, 2003)
indicates that the undrained strength, c u, should vary according to eq. (5.2):

(5.2)

In this case, the variation of the undrained strength with depth follows eq. (5.3):

(5.3)

The initial value of undrained strength (at the layer top), is considered to be c u=8kPa. Although the
recommended value of Poissons ratio is =0.49, the maximum value that may be introduced in the
program is =0.35 for undrained conditions. According to (Brinkgreve, 2006), this is due to numerical
issues: for undrained simulations, Plaxis attributes to the pore fluid an implicit Poissons ratio of
=0.495 (0.5) and if the effective Poissons ratio of the soil were higher than 0.35, the stiffness of the

74

water compared to the soils would not be high enough. Therefore, the drained value of the Poissons
ratio is not taken into account into undrained simulations.
As for the interface strength factor, Rinter=0.6 is applied. Generally, this factor is supposed to decrease
with the soil resistance. In this case, since the rule used in drained situations may not be applied, a
value slightly below the used in the previous layer is chosen.

- Layer 3: Medium sandy clay


This type of soil is assumed to be a recent alluvial deposit (a2) green and grey fine sand with high
percentage of silt and clay (2<NSPT<14). Although the geological characterization states that there is a
higher percentage of sand and the static load test report that there is a higher percentage of clay,
there is an evident similarity between the two descriptions. Since the static load test report is
considered to have precedence, this soil is modelled under undrained conditions.
(Viaponte, 2003) suggests, for this soil, =16kN/m

and E=77.7MPa. Eq. (5.2) is once again

applicable, but since the specific weight of the soil is changed, eq. (5.3) becomes eq. (5.4):

(5.4)

The initial value of undrained strength (at the layer top), is considered to be the same as the one in the
bottom of the previous layer: cu=26kPa.
(Viaponte, 2003) suggests a value for the Poissons ratio of =0.44 for this soil.
The interface strength value is the same as in the previous layer: R inter=0.6.
- Layer 4: Firm sandy clay
This layer of soil is compared to the Pleistocene/Miocene (?) soil. It is described by the static load test
report as green and stiff sandy clay. The Pleistocene/Miocene (?) is described as a silty clay with a
high percentage of sand, and NSPT>10.
3

(Viaponte, 2003) suggests for this soil =20kN/m and E=176.1MPa. The soil is modelled under
undrained conditions. For the undrained strength, the value at the bottom of the previous layer,
cu=44kPa, is used, for it is generally considered an appropriate value for stiff clay; this is supposed to
be constant along depth.
(Viaponte, 2003) suggests a value for the Poissons ratio of =0.48 for this soil.
The interface strength value is the same as in the previous layers: R inter=0.6.
- Layer 5: Gravel
This layer is assumed to be part of the Miocene, which, according to (Viaponte, 2003), has the
3

properties of =21.5kN/m and E=745.0MPa. On the other hand, (Viaponte, 2003) also suggests, for
3

gravel in that region, the values of =20.5kN/m and E=100.3MPa.


3

It is decided that the specific weight of the gravel, =20.5kN/m , is more appropriate. As for the
Youngs modulus, the suggested value for the Miocene is clearly too high, and the value for the gravel

75

too low, since the previous layers Youngs modulus is E=176.1MPa, and it is supposed to increase
with depth. Thus, the compromise value of E=180MPa is used.
The simulation is performed under drained conditions, and the attributed value of the Poissons ratio is
=0.35. The friction angle is supposed to be very high, so the value of =45 is chosen. Using eqs.
(4.2) and (5.1), this leads to an interface strength factor of Rinter=0.577. This value is slightly lower than
the one of the previous layer (Rinter=0.6), so it is considered appropriate.
- Layer 6: Very stiff yellow clay
This layer is clearly part of what is assumed to be the Miocene, which, according to (Viaponte, 2003),
3

has the properties of =21.5kN/m and E=745.0MPa and is formed by calcareous marl and clay. The
indication of the specific weight is respected, but not the Youngs modulus, since it is excessively
high, especially given that the value of the previous layer is E=180MPa. The value of E=200MPa
appears to be more appropriate. The indicated Poissons ratio is =0.25.
The simulation is performed under undrained conditions. The undrained strength is assumed to be
constant and equal to cu=200kPa, since that is a generally accepted value for very stiff clay. Since
there is an increase in resistance, it seems adequate to slightly lower the value of the interface
strength factor, and Rinter=0.5 is introduced in the model.
- Layer 7: Dense yellow sand
3

This layer also belongs to the Miocene; thus, a specific weight of =21.5kN/m is used. It is considered
appropriate to use the same Youngs modulus as the previous layer, E=200MPa, and also the same
Poissons ratio, =0.25. This layer is modelled under drained conditions. For dense sand, a friction
angle equal to =40 is generally considered to be appropriate. If eqs. (4.2) and (5.1) were to be
used, a Rinter=0.599 would be introduced in the program; however, since the previous layer is
attributed Rinter=0.5, this value is also used in this one, at least as a first iteration.
- Layer 8: Hard yellow clay
This layer has similar characteristics as layer 6, but, given its depth, is supposed to be more
3

consistent. Most soil properties introduced in the program are the same: =21.5kN/m and E=200MPa.
It also behaves under undrained conditions, but the attributed value of undrained strength is higher:
cu=600kPa (for hard clay: c u>400kPa). The same value of interface strength factor as in the previous
layers is introduced: Rinter=0.5.
In Table 5.7, the soil properties of all layers introduced in the program are listed.

76

Table 5.7: Soil properties (2).

Layer

Material

cu0

cu/z

'

E*

'

Rinter

k0

25.0

0.630

0.500

0.35

59.0

0.600

0.550

0.35

77.7

0.600

1.000

0.35

176.1

0.600

1.000

45

0.35

180.0

0.577

0.293

21.5

200

0.25

200.0

0.500

1.000

Drained

21.5

40

0.25

200.0

0.500

0.357

Undrained

21.5

600

0.25

200.0

0.500

1.000

(m)

Type

[kN/m ]

[kN/m ]

[kN/m /m]

()

3.0

Drained

18.0

30

0.3

12.0

Undrained

16.0

1.5

9.0

Undrained

18.0

26

2.0

4.0

Undrained

20

44

2.5

Drained

20.5

6.0

Undrained

4.0

14.5

[MN/m ]

-4

* For =10 .

In order to verify the consistency of these parameters, the analytical shaft resistance is calculated and
compared to the shaft resistance given by the SLT (see Table 5.3). All the following equations are
found in (Santos, 2008b).
The equation that expresses pile resistance to compression, Rc, is eq. (5.5):

(5.5)

Where:
Rb: base resistance of the pile
Rs: shaft resistance of the pile
In undrained conditions, the base resistance, Rb, is calculated through eq. (5.6):

(5.6)

Where:
Ab: area of the pile base
cu: undrained resistance at the pile base
Nc: load capacity factor (Nc=9, according to (Santos, 2008a))
Table 5.8 shows the calculated base resistance, Rb, for this case.
Table 5.8: Analytical base resistance, Rb.
2

Layer

L (m)

Material Type

Ab (m )

cu (kPa)

Nc

rb (kPa)

Rb (kN)

3.0

Undrained

0.503

600

5400.00

2714.34

77

In undrained conditions, the shaft resistance, Rc, is calculated through eq. (5.7):

(5.7)

Where:
As: area of the pile shaft
: adhesion factor (=Rinter)
: average undrained resistance along the pile shaft

In drained conditions, the shaft resistance, Rs, is calculated through eq. (5.8):

(5.8)

Where:
Ks: earth pressure coefficient
: average effective vertical stress along the pile shaft
: friction angle at the soil-pile interface
Table 5.9 shows the calculated shaft resistance, Rs, for the different soil layers.

Table 5.9: Analytical shaft resistance, Rs.

Material

()

rs (kPa)

Rs (kN)

27.00

20.00

4.91

37.05

0.550

90.00

10.20

307.62

35

1.000

162.00

21.00

475.01

0.600

44

1.000

218.00

26.40

265.40

Drained

0.577

0.293

251.13

30.00

42.47

266.82

6.0

Undrained

0.500

200

1.000

298.75

100.00

1507.96

4.0

Drained

0.500

0.357

356.25

26.67

63.91

642.50

2.2

Undrained

0.500

600

1.000

462.63

300.00

1658.76

Layer

L (m)

Rinter

3.0

Drained

0.630

0.500

12.0

Undrained

0.600

17

9.0

Undrained

0.600

4.0

Undrained

2.5

Type

(kPa)

k0

(kPa)

5161.13

Total

These

tables

indicate

that

this

piles

analytical

resistance

to

compression

is

Rc=2714.34+5161.13=7875.47kN, which seems an acceptable value, given the SLTs results.

78

4) The Mesh

The mesh is created from 15-node triangular elements, in a total of 9970 elements and 84399 nodes,
with an average element size of 0.525m. The mesh is much more refined around the pile. To achieve
this, three transition areas are drawn, so as to obtain a good connection between highly refined and
coarse parts of the mesh. This mesh is considerably refined, as may be observed in Figure 5.14.

Figure 5.14: Mesh.

5.4.2 Initial Conditions


1) Water Pressure Generation

The initial conditions of the model refer to both the water pressure and the stress field, before any
calculation is performed.
The pore pressure field due to the water table is generated in the first phase, using the specific weight
3

of the water, =10kN/m . The resulting pore pressure diagram, which is constant along the model, is
displayed in Figure 5.15.

79

Figure 5.15: Diagram of pore pressure.

2) Initial Stress Field Generation

The initial stress field generation must be executed carefully in Plaxis, as discussed in the previous
chapter. For the same reasons as in the previous chapter (the K0-procedure does not impose
horizontal equilibrium), the pile is not included in this generation. In this case, however, there is
another relevant issue, in the form of the volume of superficial soil removed for the execution of the
pile. Since it is not advisable to use the normal initial stress generation procedure for non-horizontal
surfaces, it is decided to ignore the removal of the soil in the initial stress generation. The values of the
coefficient of earth pressure displayed in Table 5.7 lead to the diagram of effective vertical stresses in
Figure 5.16 and the diagram of effective horizontal stresses in Figure 5.17.

Figure 5.16: Initial vertical effective stress field, including removed soil and not including the pile.

80

Figure 5.17: Initial horizontal effective stress field, including removed soil but not the pile.

These two figures show satisfactory stress fields, with increasing stress values along the depth, and
constant along the radius. The values correspond to the ones calculated manually, with a negligible
error in the area of the pile (since the mesh is much more refined there). Thus, this initial stress field is
considered to be adequate and is used in the following calculations.

3) Removal of Soil at the Top

The next step is to remove the volume of soil at the pile top. According to the profile published by
(Santos, 2005), the ICIST report (see Figure 5.2), the excavation is horizontal at the bottom, where the
pile cap is located, and its slopes have an unclear inclination value. For modelling purposes, it is
important that this excavation does not cause a failure mechanism in the slope; thus, an inclination of
26.57 is used (10m5m). The total displacements in the model after this step are displayed in Figure
5.18.

81

Figure 5.18: Total displacements after the removal of soil at the top.

It is clear that this inclination is sufficient to avoid a failure mechanism. The vertical displacements that
results from this step are also acceptable. The soil removal also causes a slight change in the stress
fields; however, there is no available data to compare with the stress field provided by Plaxis.

4) Introduction of the Pile

Unlike the case of the last chapter, which was an academic analysis that pursued a comparison with
other simulations, this chapter aims to compare results provided by Plaxis with the ones from a static
load test. In the last chapter, the introduction of the pile into the soil was simulated so as not to disturb
the soil, i. e. not generating shear stress. In this case, although there is no direct way to determine the
degree of disturbance caused by the introduction of the pile into the soil, it seems appropriate to
consider it as a non-displacement pile. This is achieved by simply replacing the volume of soil for pile
material, using the interface strength factors listed in Table 5.7. The consequence is that the pile
introduction generated shear stress along the shaft.
As a simplification, the use of the temporary casing is not modelled, although it is probable that its
effect is not entirely negligible.

82

5.4.3 Calculations and Results


The calculation is performed through steps, each one representing one of increments of the static load
test. Since the loading is applied through prescribed displacements, the inversed method is used: the
settlements recorded from the test are introduced in the program, and the consequent load along the
pile is the output.
Since the goal is to compare the results with the ones recorded from the static load test, the reset
displacements option is activated before the first load is applied, for the measurements refer to
variations on settlement and stress that result from loading.
The results provided by Plaxis are presented in comparison to the ones found in the static load test
report.
Figure 5.19 shows the load at the pile top, Pt, in terms of the total settlement at the pile top, wt,
resulting from the Plaxis simulation.

8 000
7 000
6 000

Pt (kN)

5 000
4 000
3 000
2 000
1 000
0
0

10

20

30
wt (mm)

40

50

60

70

Figure 5.19: Load at the pile top in terms of the total settlement at the pile top, given by Plaxis.

This curve indicates that the overall behaviour of the model is quite acceptable, for it is similar to the
results from the static load test, shown in Figure 5.9.

83

As mentioned in chapters 2 and 3, elastic theory-based methods are useful tools for analysing the
settlement and load distribution in early stages of loading, when the soil is still in plastic state. Thus,
the two described methods are applied to this case. Table 5.10 and Table 5.11 show the load
settlement ratio and the proportion of load transferred to the pile base, as well as the relevant
parameters for the calculations. Table 5.7 provides the rest of the necessary information.
Table 5.10: Poulos and Davis estimation.

Eav (kN/m )

I0

Rk

Rh

118892

0.045

1.9

0.8

0.98

53.21

Ck

0.035

0.45

0.85

0.0134

.
Table 5.11: Randolph and Wroth (compressible piles) estimation.

3.09

4.27

50.12

0.0116

The two methods provide similar solutions, which is a good indicator. The Poulos and Davis method
gives slightly higher values, as occurred in the results displayed in chapters 2 and 3.

Figure 5.20 shows the load at the pile top, Pt, in terms of the total settlement at the pile top, wt,
resulting from the static load test and from the calculation performed by Plaxis for the first loading
(load steps 1 to 8). The solution by Randolph and Wroth is also plotted.
The points which constitute this chart are listed in Appendix D1.

84

3 000

Pt (kN)

2 000

SLT
Plaxis

1 000

R&W

0
0

wt (mm)
Figure 5.20: Load at the pile top in terms of the total settlement at the pile top, given by the SLT, Plaxis and the
Randolph and Wroth method, for the first loading cycle.

The curve of the first loading produces satisfactory results, which traduces in a good behaviour of the
model. In this load step, only the first 4 layers of soil are in plastic state, the 4 in the bottom still being
in the elastic state. This is expected, since resistance increases with depth. Thus, the parameters
attributed to the soil are considered to be acceptable.
Results provided by the Randolph and Wroth method show that the total settlement is underestimated,
when compared with the ones from the SLT and Plaxis.
Figure 5.21 shows the load at the pile top, Pt, in terms of the total settlement at the pile top, wt,
resulting from the static load test and from the calculation performed by Plaxis for the reloading (load
steps 8 to 25).
Figure 5.22 shows the load at the pile top, Pt, in terms of the total settlement at the pile top, wt,
resulting from the static load test and from the calculation performed by Plaxis for both loading cycles.

85

8 000
7 000
6 000

Pt (kN)

5 000
SLT

4 000

Plaxis
3 000
2 000
1 000
0
0

10

20

30

40

50

60

70

wt (mm)
Figure 5.21: Load at the pile top in terms of the total settlement at the pile top, given by the SLT and Plaxis, for
the second loading cycle.

8 000
7 000
6 000
5 000
SLT

Pt (kN)

4 000

Plaxis

3 000
2 000
1 000
0
0

10

20

30

40

50

60

70

wt (mm)
Figure 5.22: Load at the pile top in terms of the total settlement at the pile top, given by the SLT and Plaxis, for
both loading cycles.

86

For the last load step, the simulation by Plaxis shows that all the soil layers are in the plastic state.
The similarity of shapes between the curves accounts for the good behaviour of the model.
Figure 5.23 shows the normal stress along the shaft for both loading peaks (step 4 and 19). The points
which originate this chart are listed in Appendix D2.

2000

N (kN)
4000

6000

8000

0
5
10

z (m)

15

SLT - Load Step 4

20

Plaxis - Load Step 4

25

SLT - Load Step 19


Plaxis - Load Step 19

30
35
40
45
Figure 5.23: Normal stress along the pile shaft at the load peaks (steps 4 and 19), given by the SLT and Plaxis.

This chart shows that the shaft load is appropriately distributed, for the slope between levels is close
to the one given by the SLT. It is interesting to notice that the results are better for load step 19. This
accounts for the resistance properties of the soil.
These values refer to the load increment; i.e. normal stress at the initial state (after the pile installation)
was subtracted to the normal stress at the steps 4 and 19, otherwise a comparison to the vibrating
wire extensometers would not make sense. The same procedure was applied to the shaft load.
Nevertheless, the values of initial stress are almost negligible when compared to the ones caused by
loading.

Figure 5.24 shows the mobilized shear stress along the pile for different levels of depth and Figure
5.25 shows the shaft load in every soil layer, as well as the corresponding applied load at the pile top,
resulting from the Plaxis calculations.

87

200

8000

6000
R24

R45

R56

R68

Pt

100

4000

50

2000

Pt (kN)

(kPa)

150

0
0

10
12
Load Step

14

16

18

20

22

24

Figure 5.24: Shaft load between different levels along the pile and total applied load, obtained by Plaxis.

400

8000

350

7000

300

Layer 2

Layer 3

Layer 4

Layer 5

Layer 6

Layer 7

Layer 8

6000
5000

Pt

200

4000

150

3000

100

2000

50

1000

Pt (kN)

(kPa)

250

Layer 1

0
0

10
12
Load Step

14

16

18

20

22

24

Figure 5.25 Shaft load by layer of soil and total applied load, obtained by Plaxis.

88

Table 5.12 and Table 5.13 show a comparison between mobilized shear load between levels,
measured during the SLT, calculated analytically and determined by Plaxis, for load step 4 and 19.
Table 5.12: Shaft load for load step 4.
Load Step 4

Mobilized shaft resistance (kPa)

Between levels

SLT

Plaxis

2 and 4

15 to 25

31

4 and 5

100 to 110

32

5 and 6

15 to 25

28

6 and 8

20 to 30

27

Table 5.13: Shaft load for load step 19.


Load Step 19

Mobilized Shaft Resistance (kPa)

Between levels

SLT

Analytic

Plaxis

2 and 4

60 to 70

48

51

4 and 5

180 to 200

100

105

5 and 6

100 to 110

98

99

6 and 8

130

76

70

For load step 4, the mobilized resistance given by Plaxis is slightly superior to the measurements of
the vibrating wire extensometers, except between levels 4 and 5.
For load step 19, comparison with Figure 5.12 shows that the resistance of the model between levels
2 and 4 and also between levels 5 and 6 is slightly lower than the measured in the SLT. As far as the
resistance between levels 4 and 5 is concerned, it is much lower than the measured. Thus, it is
possible to conclude that the temporary casing affected the area between levels 4 and 5 by increasing
shear resistance, but its influence did not reach the region between levels 5 and 6. Shear resistance
between levels 6 and 8 obtained by Plaxis are lower than the ones obtained in the SLT.
Figure 5.25 shows the distribution of shaft load by the different soil layers; it accounts for the good
behaviour of the model. The layers closer to the surface have lower resistance; the last layer (hard
clay) does not reach the maximum resistance. It is interesting that, in the unloading phase, the top
layers are unloaded before the bottom layer layer 8 loses little load between load steps 19 and 24.
Figure 5.26 shows the total load, base load and shaft load in terms of the settlement at the pile top,
determined by Plaxis, for the second loading cycle. It also shows the measured values for load step 19
of the SLT, since that was the only step of the second loading cycle for which there were results
available. The total load and the base load in terms of the settlement at the pile top provided by the
Randolph and Wroth method for compressible piles is also displayed.
These results concern the displacement caused by the loading alone, disregarding the weight of the
pile.

89

8 000
7 000
Pt Plaxis

6 000

Ps Plaxis
Pb Plaxis

Pt (kN)

5 000

Pt SLT

4 000

Pt SLT load step 19


Ps SLT load step 19

3 000

Ps SLT load step 19


Pt R&W

2 000

Pb R&W

1 000
0
0

10

20

30

40

50

60

wt (mm)
Figure 5.26: Total load, load transferred to the pile shaft and load transferred to the pile base in terms of the
settlement at the pile top for the second loading cycle, obtained through Plaxis.

These curves given by Plaxis are a good indicator, shaft failure occurring for load values between
5,000kN and 6,000kN (the analytic calculation pointed to 5161.13kN).
As far as the analytical methods are concerned, they underestimate the total settlement, as happened
in the calculations performed in chapter 3. The difference in this case, however, is more evident. That
is expected, since the conditions in this case study make it much more complex. One of the most
important factors is the relative movement between soil and pile; the analytical methods consider the
interface strength factor to be equal to 1 and the values introduced in the program were between 0.5
and 0.6. This causes the shear stress along the pile to develop differently: the shaft generates less
load, and therefore the total load curve is formed by lower values. As far as the base load is
concerned, the analytical result is more satisfactory, since the slope of the curve is similar to the slope
of the Plaxis curve their values differing due to the fact that the Randolph and Wroth method does
not take into account the weight of the pile.
All the points that constitute the Plaxis curves in Figure 5.26 are shown in Appendix D4.
In this chapter, the calculations performed by Plaxis were quite satisfactory; the static load test was
accurately simulated. Given the little available information about the site geologic conditions, the
parameters introduced into the model provided good results. It is relevant to state that the goal of this
work was not to manipulate parameters indiscriminately in order to obtain the same results as the
SLTs; it was to introduce in the program acceptable values within the commonly used ranges and
taking into account the geological report recommendations, calibrating the model as possible.
It is possible to conclude that the effect of the casing was not negligible, and the fact that it was not
considered in the model affected the quality of the results.
90

6 CONCLUDING REMARKS

6.1 Conclusions
The purpose of this thesis was to understand the load transfer mechanism in single piles under axial
loading, as well as to determine the load settlement relation.
The first step was to understand the background of elastic analyses, by studying elastic theory-based
methods. This was complemented by an elastic numerical analysis using rigid interfaces.
The second step was to study the soil behaviour when it was no longer linear elastic. In order to
perform inelastic analyses, it was essential to use numerical methods, since it was no longer possible
to resort to the elastic theory-based methods. Both the soil and the soil-pile interface were assumed to
have elastoplastic behaviour, according to the Mohr-Coulomb criterion or to the Tresca criterion.
As far as the first part of this thesis is concerned, in which chapters 2 and 3 are included, the resulting
dimensionless charts of load settlement ratio and proportion of load transferred to the pile base in
terms of the slenderness ratio and of the soil inhomogeneity factor constitutes a quick way to estimate
the total settlement and the base load of a given single pile subject to vertical loading.
Thus, it is possible to conclude that:

The load settlement ratio increases with the pile slenderness ratio whereas the proportion of
load transferred to the pile base decreases with it.

The soil-pile stiffness factor has more influence over the load settlement ratio than over the
proportion of load transferred to the pile base, according to the Randolph and Wroth method.
It also influences the limit of applicability between rigid and compressible pile solutions;

91

The Randolph and Wroth method for compressible piles gave solutions which are very close
to results provided by Plaxis, indicating its adequacy for the studied cases (homogeneous and
vertically non-homogeneous soil with linear stiffness increase with depth). It provides an
expeditious method for estimating settlement and load distribution along the pile. However, it
is not applicable to very short piles;

The Poulos and Davis method does underestimate total settlement, as affirmed by the authors
and confirmed by Plaxis results. This is probably related to the fact that the pile compressibility
is not thoroughly taken into account, especially for high values of pile slenderness;

Results given by Plaxis are more sensitive to the soil inhomogeneity than the elastic theorybased methods, as far as load transferred to the pile base is concerned. This may be due to
oversimplification of the analytical methods (especially in the Poulos and Davis method), but
also to difficulties in the modelling of a high rigidity gradient in the soil;

Chapter 4 concerned the comparison of a simulation by Plaxis with two other finite element programs.
The following conclusions were drawn:

The sensitivity analysis of the interface strength factor, R inter, showed that it is the parameter
that affected the total load and the shaft load the most; however, the influence over the base
load was not very significant.

Although the total load-settlement curves provided by the different programs correspond well,
the discrepancy regarding the proportion of load transferred to the pile base is not negligible.
In the simulation performed by Plaxis, the tip was not as mobilized, even if a smooth interface
was considered;

This indicates that the algorithm used in Plaxis might have significant differences to the other
ones;

When dilatancy in the soil at the pile base was introduced, the base was indeed a lot more
mobilized.

The simulation performed in chapter 5, however, lead to satisfactory results. It is possible to conclude
that:

The goal of this work was not to manipulate parameters indiscriminately in order to obtain the
same results as the SLTs; it was to introduce in the program acceptable values within the
commonly used ranges and taking into account the geological report recommendations,
calibrating the model as possible. Given the little available information of the sites geology,
the number of layers (and consequently of soil parameters) and the considerable
discrepancies between the geological profile of the static load test report and the ones given
by the closest boreholes, as well as the differences between the applied and measured load
and the normal errors associated with in-situ tests, the results provided by Plaxis correspond
very well with the ones from the instrumentation used in the SLT;

Not all the parameters indicated in the geological report were used, since some of them did
not seem appropriate; the ones which were used, however, lead to good results. As for the

92

parameters for which there was no available information, values of the generally used ranges
were successfully applied;

The normal stress distribution along the pile calculated by Plaxis corresponded well with the
SLT results, for the two load steps with available information. As for the shaft load, the
distribution by layer provided by the model was consistent with the expectations regarding the
increase of soil resistance with depth. The last layer of soil, hard clay, did not reach its
maximum resistance; besides, during the unloading phase, the other less resistant layers
unloaded first.

It is possible to conclude that the effect of the casing was not negligible, and the fact that it
was not considered in the model affected the quality of the results. It is likely that the casing
has caused an enlargement of the borehole, increasing stresses in the area;

When applied to this case study, the elastic theory-based methods do not provide satisfactory
load-settlement results. This is due to the number of simplifications assumed by the methods,
such as admitting an average value for the Youngs modulus of the soil. The most relevant
difference is perhaps the interface strength factor, which is equal to 1 in the analytical
methods, whereas the values introduced in the model are between 0.5 and 0.6; therefore,
Plaxis provides lower values of generated shear stress, and thus lower total load.

6.2 Further Research


The described elastic theory-based methods are applicable to piles with non-circular cross sections.
Since 2D axisymmetry was used, it was not possible to validate numerical methods for this case. In
the future, it would be interesting to use 3D modelling to study the behaviour of piles with non-circular
cross sections.
There are non-linear load-transfer methods, which can be applied and compared to the elastic theorybased methods.
Although only floating piles were analysed in this work, the Randolph and Wroth method may also be
applied cases of end-bearing piles, i.e. where the Youngs modulus of the soil increases abruptly at
the pile base. It would be interesting to study the results of numerical analyses in those conditions.
Besides, it would be interesting to draw charts of the load settlement ratio and the proportion of load
transferred to the pile base in terms of the pile slenderness ratio for different values of the soil-pile
stiffness factor obtained by Plaxis, since the simulations made in chapter 3 all refer to the same value.
The calibration of numerical methods through static load tests is an interesting and useful work. In this
thesis, the soil was modelled using the Mohr-Coulomb criterion and the Tresca criterion. It would be
appropriate to apply other material models, such as the hardening soil, to this and other cases, and
analyse the differences. The pile-soil interface could also be modelled using another failure criterion.

93

94

BIBLIOGRAPHY
References
Brinkgreve, R. B. (2006). Plaxis - 2D Version 8 Manual. The Netherlands: Delft University of
Technology and Plaxis b. v.
D'aguiar, S. (2008). Numerical Modelling of Soil-Pile Axial Load Transfer Mechanisms in Granular
Soils. PhD Thesis. cole Centrale des Arts et Manufactures cole Centrale Paris;
Universidade Tcnica de Lisboa Instituto Superior Tcnico.
Fleming, K. W. (1992). Piling Engineering. 3rd Edition. Glasgow: Black Academic and Professional.
Mindlin, R. D. (1936). Force at a Point in the Interior of a Semi-Infinite Soil. In Physics, vol. 7.
Neves, M. M. (2001a). tude du Comportement de Pieux Fors. I. Exprimentations in situ et en
Laboratoire. Bulletin des Laboratoires des Ponts et Chausses, 231, pp. 39-54.
Neves, M. M. (2001b). tude du comportement de pieux fores. II. modelisation par elements finis.
Bulletin des Laboratoires des Ponts et Chausses, 231, pp. 55-67.
Poulos, H. G. (1972). Load-Settlement Prediction for Piles and Piers. JSMFD ASCE vol. 98 SM9, pp.
879-897.
Poulos, H. G., & Davis, E. (1980). Pile Foundation Analysis and Design. New York: Wiley and Sons.
Poulos, H. G., & Davis, E. H. (1968). The Settlement Behaviour of Single Axially-Loaded
Incompressible Piles and Piers. Geotechnique, vol. 18, pp. 351-371.
Randolph, M., & Wroth, C. (1978). Analysis of Deformation of Vertically Loaded Piles. JGED, ASCE
104(GT12), pp. 1465-1488.
Santos, J. A. (2005). Relatrio do Ensaio de Carga numa Estaca Experimental no mbito do Projecto
de Modernizao da Linha do Norte, Viaduto de Santana do Cartaxo. Instituto de Engenharia
de Estruturas, Territrio e Construo (ICIST), Instituto Superior Tcnicno.
Santos, J. A. (2008). Fundaes por Estacas. Aces Verticais. . Elementos Tericos da disciplina de
Obras Geotcnicas. Instituto Superior Tcnico.
Santos, J. A. (2008). Fundaes por Estacas. Formulrio. Elementos Tericos da disciplina de Obras
Geotcnicas. Instituto Superior Tcnico.
Viaponte. (2003). Memria Descritiva e Justificativa Geologia e Geotecnia, Doc n PE 27.5 MDg 2.
Vol. 10 Viaduto de Santana do Cartaxo, Projecto de Modernizao da Linha do Norte
Sub-troo 1.4: Azambuja (excl.) Vale de Santarm (inc.).

95

Consulted Bibliography
Bowles, J. E. (1996). Foundation Analysis and Design, 5th Edition. McGraw-Hill.
Das, B. M. (2006). Principles of Geotechnical Engineering 7th Edition. Cengage Learning.
Fellenius, B. H. (2011). Basics of Foundation Design. Electronic Edition.
Frank, R. (1975). tude Thorique du Comportement des Pieux sous Charge Verticale. Introduction
de la Dilatance. Laboratoire Central des Ponts et Chausses.
Pereira, T. (2008). Estacas Sob Aces Verticais, Considerando os Efeitos da Carga Residual. MsC
Thesis. Universidade Tcnica de Lisboa - Instituto Superior Tcnico.
Tadjbakhsh, S. a. (1985). tude par la Mthode des lments Finis du Comportement lastoplastique
de Sols Dilatants. Application aux Pieux sous Charge Verticale. Laboratoire Central des Ponts
et Chausses.
Timoshenko, S. P. (1936). Theory of Elastic Stability. New York: McGraw-Hill.
Velloso, D. A. (2002). Fundaes. Vol. 2 - Fundaes Profundas. COPPE-UFRK. Oficina de Textos.

96

APPENDIXES
Appendix A Dimension of the Model Used in the Elastic
Simulations

L/r0

h (m)

l (m)

36

90.0

72.0

69

172.5

138.0

10.0

8.0

37

92.5

74.0

70

175.0

140.0

12.5

10.0

38

95.0

76.0

71

177.5

142.0

15.0

12.0

39

97.5

78.0

72

180.0

144.0

17.5

14.0

40

100.0

80.0

73

182.5

146.0

20.0

16.0

41

102.5

82.0

74

185.0

148.0

22.5

18.0

42

105.0

84.0

75

187.5

150.0

10

25.0

20.0

43

107.5

86.0

76

190.0

152.0

11

27.5

22.0

44

110.0

88.0

77

192.5

154.0

12

30.0

24.0

45

112.5

90.0

78

195.0

156.0

13

32.5

26.0

46

115.0

92.0

79

197.5

158.0

14

35.0

28.0

47

117.5

94.0

80

200.0

160.0

15

37.5

30.0

48

120.0

96.0

81

202.5

162.0

16

40.0

32.0

49

122.5

98.0

82

205.0

164.0

17

42.5

34.0

50

125.0

100.0

83

207.5

166.0

18

45.0

36.0

51

127.5

102.0

84

210.0

168.0

19

47.5

38.0

52

130.0

104.0

85

212.5

170.0

20

50.0

40.0

53

132.5

106.0

86

215.0

172.0

21

52.5

42.0

54

135.0

108.0

87

217.5

174.0

22

55.0

44.0

55

137.5

110.0

88

220.0

176.0

23

57.5

46.0

56

140.0

112.0

89

222.5

178.0

24

60.0

48.0

57

142.5

114.0

90

225.0

180.0

25

62.5

50.0

58

145.0

116.0

91

227.5

182.0

26

65.0

52.0

59

147.5

118.0

92

230.0

184.0

27

67.5

54.0

60

150.0

120.0

93

232.5

186.0

28

70.0

56.0

61

152.5

122.0

94

235.0

188.0

29

72.5

58.0

62

155.0

124.0

95

237.5

190.0

30

75.0

60.0

63

157.5

126.0

96

240.0

192.0

31

77.5

62.0

64

160.0

128.0

97

242.5

194.0

32

80.0

64.0

65

162.5

130.0

98

245.0

196.0

33

82.5

66.0

66

165.0

132.0

99

247.5

198.0

34

85.0

68.0

67

167.5

134.0

100

250.0

200.0

35

87.5

70.0

68

170.0

136.0

97

Appendix B1 Load Settlement Ratio in Terms of the Pile


Slenderness Ratio
=1, =975, =0.3, h=2.5L
Load Settlement Ratio, Pt/(wtr0GL)

Pile Slenderness

Randolph and

Ratio,

Poulos and Davis,

Randolph and Wroth,

L/r0

P&D

R&W (rigid piles)

24.96

18.63

18.42

26.08

20.20

19.90

27.45

21.75

21.34

29.03

23.27

22.74

30.71

24.76

24.10

32.42

26.23

25.41

10

34.06

27.67

26.67

11

35.54

29.08

27.89

12

36.86

30.48

29.08

13

37.98

31.86

30.22

14

38.91

33.21

31.33

15

39.63

34.56

32.41

16

40.19

35.88

33.44

17

40.61

37.20

34.45

18

40.94

38.50

35.42

19

41.25

39.78

36.37

20

41.72

41.06

37.28

21

42.15

42.32

38.16

22

42.70

43.58

39.01

23

43.35

44.82

39.84

24

44.09

46.06

40.64

25

44.92

47.29

41.41

26

45.82

48.51

42.15

27

46.78

49.72

42.87

28

47.80

50.92

43.57

29

48.83

52.12

44.24

30

50.14

53.30

44.89

31

51.25

54.49

45.51

32

52.37

55.66

46.12

33

53.47

56.83

46.70

34

54.56

58.00

47.26

35

55.64

59.16

47.80

Wroth, R&W

Plaxis

(compressible piles)
19.30

22.11

24.86

27.35

29.74

32.93

34.87

37.48

39.05

41.83

44.18

46.62

98

36

56.71

60.31

48.32

37

57.75

61.46

48.82

38

58.78

62.60

49.30

39

59.79

63.74

49.77

40

60.53

64.87

50.21

41

61.41

66.00

50.64

42

62.25

67.12

51.06

43

63.04

68.24

51.46

44

63.79

69.36

51.84

45

64.50

70.47

52.20

46

65.17

71.58

52.56

47

65.80

72.68

52.89

48

66.39

73.78

53.22

49

66.95

74.88

53.53

50

67.62

75.97

53.83

51

68.02

77.06

54.11

52

68.40

78.15

54.39

53

68.77

79.23

54.65

54

69.13

80.31

54.90

55

69.51

81.38

55.14

56

69.89

82.46

55.37

57

70.29

83.53

55.59

58

70.72

84.59

55.80

59

71.18

85.66

56.00

60

71.26

86.72

56.19

61

71.68

87.78

56.37

62

72.12

88.83

56.55

63

72.58

89.89

56.71

64

73.04

90.94

56.87

65

73.50

91.99

57.02

66

73.94

93.03

57.16

67

74.37

94.07

57.30

68

74.75

95.11

57.43

69

75.09

96.15

57.55

70

75.63

97.19

57.67

71

75.93

98.22

57.78

72

76.18

99.26

57.89

73

76.39

100.28

57.98

74

76.56

101.31

58.08

75

76.71

102.34

58.17

76

76.84

103.36

58.25

77

76.97

104.38

58.33

78

77.10

105.40

58.41

48.63

51.28

52.90

53.77

99

79

77.25

106.42

58.48

80

77.57

107.43

58.54

81

77.81

108.45

58.60

82

78.08

109.46

58.66

83

78.37

110.47

58.72

84

78.67

111.47

58.77

85

78.96

112.48

58.81

86

79.24

113.48

58.86

87

79.49

114.49

58.90

88

79.70

115.49

58.94

89

79.85

116.49

58.97

90

80.60

117.48

59.00

91

80.80

118.48

59.03

92

80.96

119.47

59.06

93

81.05

120.47

59.08

94

81.05

121.46

59.11

95

81.05

122.45

59.13

96

81.05

123.43

59.14

97

81.05

124.42

59.16

98

81.15

125.40

59.17

99

81.11

126.39

59.18

100

81.19

127.37

59.19

54.17

54.42

54.06

100

=0.75, =975, =0.3, h=2.5L


Load Settlement Ratio, Pt/(wtr0GL)

Pile Slenderness

Randolph and

Randolph and Wroth,

Wroth, R&W

R&W

(rigid piles)

(compressible piles)

18.72

18.85

18.61

19.62

19.92

19.60

20.73

21.08

20.65

21.99

22.25

21.70

23.34

23.43

22.74

24.72

24.60

23.76

10

26.02

25.76

24.74

11

27.17

26.90

25.70

12

28.18

28.03

26.63

13

29.06

29.15

27.53

14

29.77

30.26

28.40

15

30.34

31.35

29.24

16

30.77

32.44

30.05

17

31.09

33.51

30.83

18

31.35

34.57

31.59

19

31.59

35.63

32.32

20

31.95

36.67

33.03

21

32.32

37.71

33.71

22

32.77

38.73

34.36

23

33.31

39.75

35.00

24

33.91

40.76

35.61

25

34.58

41.77

36.19

26

35.30

42.76

36.76

27

36.08

43.76

37.30

28

36.88

44.74

37.83

29

37.71

45.72

38.33

30

38.74

46.69

38.82

31

39.63

47.66

39.28

32

40.51

48.62

39.73

33

41.38

49.58

40.16

34

42.24

50.53

40.57

35

43.10

51.48

40.97

36

43.93

52.42

41.35

37

44.75

53.36

41.71

38

45.56

54.29

42.06

39

46.35

55.22

42.39

40

46.93

56.15

42.71

41

47.61

57.07

43.02

Ratio,

Poulos and Davis,

L/r0

P&D

Plaxis

18.96

20.95

23.01

24.92

26.76

29.22

30.72

32.73

33.94

36.07

37.85

39.68

41.17

101

42

48.26

57.98

43.31

43

48.87

58.90

43.59

44

49.45

59.81

43.86

45

49.99

60.72

44.12

46

50.50

61.62

44.36

47

50.98

62.52

44.60

48

51.43

63.42

44.82

49

51.85

64.31

45.03

50

52.34

65.20

45.24

51

52.74

66.09

45.43

52

53.12

66.97

45.61

53

53.49

67.86

45.79

54

53.86

68.74

45.96

55

54.23

69.61

46.12

56

54.60

70.49

46.27

57

54.99

71.36

46.41

58

55.39

72.23

46.54

59

55.82

73.09

46.67

60

55.95

73.96

46.80

61

56.34

74.82

46.91

62

56.75

75.68

47.02

63

57.16

76.54

47.12

64

57.58

77.39

47.22

65

57.99

78.24

47.31

66

58.39

79.09

47.40

67

58.77

79.94

47.48

68

59.12

80.79

47.56

69

59.43

81.63

47.63

70

59.89

82.48

47.70

71

60.16

83.32

47.76

72

60.39

84.16

47.82

73

60.59

84.99

47.87

74

60.75

85.83

47.93

75

60.89

86.66

47.97

76

61.02

87.49

48.02

77

61.14

88.32

48.06

78

61.26

89.15

48.10

79

61.39

89.97

48.13

80

61.66

90.80

48.16

81

61.86

91.62

48.19

82

62.08

92.44

48.22

83

62.31

93.26

48.24

84

62.55

94.08

48.26

43.04

44.09

44.53

44.61

102

85

62.79

94.90

48.28

86

63.01

95.71

48.30

87

63.20

96.52

48.31

88

63.36

97.34

48.33

89

63.47

98.15

48.34

90

64.05

98.96

48.35

91

64.10

99.76

48.35

92

64.10

100.57

48.36

93

64.10

101.37

48.36

94

64.10

102.18

48.36

95

64.10

102.98

48.36

96

64.23

103.78

48.36

97

64.23

104.58

48.36

98

64.23

105.38

48.36

99

64.23

106.17

48.35

100

64.28

106.97

48.35

44.55

44.03

103

=0.5, =975, =0.3, h=2.5L


Load Settlement Ratio, Pt/(wtr0GL)

Pile Slenderness

Randolph and

Randolph and Wroth,

Wroth, R&W

R&W

(rigid piles)

(compressible piles)

12.48

17.92

17.66

13.08

18.25

17.92

13.82

18.85

18.43

14.66

19.55

19.02

15.56

20.30

19.66

16.48

21.08

20.30

10

17.37

21.86

20.94

11

18.17

22.64

21.56

12

18.88

23.43

22.18

13

19.50

24.21

22.77

14

20.01

24.98

23.35

15

20.42

25.76

23.91

16

20.74

26.52

24.45

17

21.00

27.28

24.97

18

21.20

28.03

25.48

19

21.40

28.78

25.96

20

21.67

29.52

26.43

21

21.95

30.26

26.88

22

22.29

30.99

27.32

23

22.68

31.72

27.73

24

23.13

32.44

28.13

25

23.61

33.15

28.52

26

24.14

33.87

28.89

27

24.70

34.57

29.24

28

25.28

35.28

29.58

29

25.88

35.98

29.90

30

26.62

36.67

30.22

31

27.26

37.36

30.51

32

27.90

38.05

30.80

33

28.53

38.73

31.07

34

29.16

39.41

31.33

35

29.78

40.09

31.58

36

30.39

40.76

31.82

37

31.00

41.43

32.05

38

31.59

42.10

32.27

39

32.17

42.76

32.47

40

32.60

43.43

32.67

41

33.11

44.08

32.86

Ratio,

Poulos and Davis,

L/r0

P&D

Plaxis

17.49

18.58

19.85

21.12

22.32

23.97

24.97

26.32

27.12

28.52

29.67

30.81

31.69

104

42

33.60

44.74

33.04

43

34.06

45.39

33.21

44

34.50

46.04

33.37

45

34.91

46.69

33.52

46

35.30

47.34

33.67

47

35.67

47.98

33.81

48

36.01

48.62

33.94

49

36.34

49.26

34.06

50

36.73

49.90

34.18

51

37.04

50.53

34.29

52

37.34

51.16

34.40

53

37.64

51.79

34.50

54

37.93

52.42

34.59

55

38.23

53.04

34.68

56

38.53

53.67

34.76

57

38.84

54.29

34.84

58

39.16

54.91

34.91

59

39.50

55.53

34.98

60

39.62

56.15

35.05

61

39.94

56.76

35.11

62

40.26

57.37

35.17

63

40.60

57.98

35.22

64

40.93

58.59

35.27

65

41.26

59.20

35.31

66

41.59

59.81

35.36

67

41.90

60.41

35.40

68

42.19

61.02

35.43

69

42.45

61.62

35.47

70

42.82

62.22

35.50

71

43.06

62.82

35.52

72

43.26

63.42

35.55

73

43.45

64.01

35.57

74

43.61

64.61

35.59

75

43.75

65.20

35.61

76

43.89

65.79

35.63

77

44.02

66.39

35.64

78

44.15

66.97

35.66

79

44.29

67.56

35.67

80

44.53

68.15

35.68

81

44.73

68.74

35.68

82

44.93

69.32

35.69

83

45.15

69.90

35.69

84

45.38

70.49

35.70

32.69

33.07

33.01

32.68

105

85

45.60

71.07

35.70

86

45.81

71.65

35.70

87

46.00

72.23

35.70

88

46.17

72.81

35.70

89

46.43

73.38

35.69

90

46.78

73.96

35.69

91

46.95

74.53

35.68

92

47.08

75.11

35.68

93

47.18

75.68

35.67

94

47.25

76.25

35.66

95

47.29

76.82

35.65

96

47.30

77.39

35.64

97

47.38

77.96

35.63

98

47.45

78.53

35.62

99

47.51

79.09

35.61

100

47.54

79.66

35.60

32.24

31.52

106

Appendix B2 Proportion of Load Transferred to the Pile Base


in Terms of the Pile Slenderness Ratio

=1, =0.3, h=2.5L


Proportion of Load taken by the Pile Base, Pb/Pt

Pile Slenderness
Ratio,

Poulos and Davis,

L/r0

P&D

Randolph and

Randolph and Wroth,

Wroth, R&W

R&W

(rigid piles)

(compressible piles)

0.265

0.307

0.305

0.229

0.283

0.281

0.202

0.263

0.261

0.181

0.246

0.243

0.164

0.231

0.228

0.151

0.218

0.215

10

0.139

0.207

0.203

11

0.130

0.196

0.193

12

0.121

0.187

0.183

13

0.114

0.179

0.175

14

0.108

0.172

0.167

15

0.102

0.165

0.160

16

0.097

0.159

0.154

17

0.093

0.154

0.148

18

0.089

0.148

0.142

19

0.085

0.144

0.137

20

0.082

0.139

0.133

21

0.079

0.135

0.128

22

0.076

0.131

0.124

23

0.073

0.127

0.120

24

0.071

0.124

0.116

25

0.068

0.121

0.113

26

0.066

0.118

0.109

27

0.064

0.115

0.106

28

0.062

0.112

0.103

29

0.060

0.110

0.100

30

0.059

0.107

0.098

31

0.057

0.105

0.095

32

0.056

0.103

0.093

33

0.054

0.101

0.090

34

0.053

0.099

0.088

35

0.051

0.097

0.086

36

0.050

0.095

0.084

Plaxis

0.257

0.203

0.167

0.148

0.128

0.110

0.101

0.089

0.083

0.073

0.065

0.056

107

37

0.049

0.093

0.082

38

0.048

0.091

0.080

39

0.047

0.090

0.078

40

0.046

0.088

0.076

41

0.044

0.087

0.074

42

0.043

0.085

0.073

43

0.043

0.084

0.071

44

0.042

0.082

0.069

45

0.041

0.081

0.068

46

0.040

0.080

0.066

47

0.039

0.079

0.065

48

0.038

0.077

0.064

49

0.037

0.076

0.062

50

0.036

0.075

0.061

51

0.036

0.074

0.060

52

0.035

0.073

0.058

53

0.034

0.072

0.057

54

0.034

0.071

0.056

55

0.033

0.070

0.055

56

0.032

0.069

0.054

57

0.032

0.068

0.053

58

0.031

0.068

0.052

59

0.030

0.067

0.051

60

0.030

0.066

0.050

61

0.029

0.065

0.049

62

0.029

0.064

0.048

63

0.028

0.064

0.047

64

0.028

0.063

0.046

65

0.027

0.062

0.045

66

0.027

0.061

0.044

67

0.026

0.061

0.043

68

0.026

0.060

0.043

69

0.025

0.059

0.042

70

0.025

0.059

0.041

71

0.024

0.058

0.040

72

0.024

0.058

0.039

73

0.023

0.057

0.039

74

0.023

0.056

0.038

75

0.022

0.056

0.037

76

0.022

0.055

0.037

77

0.022

0.055

0.036

78

0.021

0.054

0.035

79

0.021

0.054

0.035

0.050

0.040

0.033

0.028

108

80

0.020

0.053

0.034

81

0.020

0.053

0.033

82

0.020

0.052

0.033

83

0.019

0.052

0.032

84

0.019

0.051

0.032

85

0.019

0.051

0.031

86

0.018

0.050

0.030

87

0.018

0.050

0.030

88

0.018

0.049

0.029

89

0.017

0.049

0.029

90

0.017

0.049

0.028

91

0.017

0.048

0.028

92

0.017

0.048

0.027

93

0.016

0.047

0.027

94

0.016

0.047

0.026

95

0.016

0.047

0.026

96

0.015

0.046

0.025

97

0.015

0.046

0.025

98

0.015

0.046

0.025

99

0.015

0.045

0.024

100

0.014

0.045

0.024

0.023

0.020

0.017

109

=0.75, =975, =0.3, h=2.5L


Proportion of Load taken by the Pile Base, Pb/Pt

Pile Slenderness

Randolph and

Randolph and Wroth,

Wroth, R&W

R&W

(rigid piles)

(compressible piles)

0.266

0.303

0.301

0.230

0.287

0.285

0.202

0.271

0.268

0.181

0.257

0.254

0.164

0.244

0.240

0.151

0.232

0.228

10

0.139

0.222

0.217

11

0.130

0.212

0.207

12

0.122

0.204

0.198

13

0.115

0.196

0.190

14

0.108

0.189

0.183

15

0.103

0.182

0.175

16

0.098

0.176

0.169

17

0.094

0.171

0.163

18

0.090

0.165

0.157

19

0.086

0.160

0.152

20

0.083

0.156

0.147

21

0.080

0.152

0.142

22

0.077

0.148

0.138

23

0.074

0.144

0.134

24

0.072

0.140

0.130

25

0.070

0.137

0.126

26

0.068

0.134

0.122

27

0.065

0.131

0.119

28

0.064

0.128

0.116

29

0.062

0.125

0.113

30

0.060

0.122

0.110

31

0.059

0.120

0.107

32

0.057

0.118

0.104

33

0.056

0.115

0.102

34

0.054

0.113

0.099

35

0.053

0.111

0.097

36

0.052

0.109

0.094

37

0.051

0.107

0.092

38

0.049

0.105

0.090

39

0.048

0.103

0.088

40

0.047

0.102

0.086

Ratio,

Poulos and Davis,

L/r0

P&D

Plaxis

0.314

0.251

0.208

0.185

0.161

0.139

0.128

0.114

0.106

0.094

0.084

0.073

0.066

110

41

0.046

0.100

0.084

42

0.045

0.099

0.082

43

0.044

0.097

0.080

44

0.043

0.096

0.078

45

0.043

0.094

0.077

46

0.042

0.093

0.075

47

0.041

0.091

0.073

48

0.040

0.090

0.072

49

0.039

0.089

0.070

50

0.039

0.088

0.069

51

0.038

0.086

0.067

52

0.037

0.085

0.066

53

0.036

0.084

0.065

54

0.036

0.083

0.063

55

0.035

0.082

0.062

56

0.035

0.081

0.061

57

0.034

0.080

0.060

58

0.033

0.079

0.058

59

0.033

0.078

0.057

60

0.032

0.077

0.056

61

0.032

0.076

0.055

62

0.031

0.076

0.054

63

0.030

0.075

0.053

64

0.030

0.074

0.052

65

0.029

0.073

0.051

66

0.029

0.072

0.050

67

0.028

0.071

0.049

68

0.028

0.071

0.048

69

0.027

0.070

0.047

70

0.027

0.069

0.046

71

0.027

0.069

0.045

72

0.026

0.068

0.044

73

0.026

0.067

0.043

74

0.025

0.067

0.043

75

0.025

0.066

0.042

76

0.024

0.065

0.041

77

0.024

0.065

0.040

78

0.024

0.064

0.039

79

0.023

0.064

0.039

80

0.023

0.063

0.038

81

0.023

0.062

0.037

82

0.022

0.062

0.037

83

0.022

0.061

0.036

0.054

0.046

0.039

0.033

111

84

0.021

0.061

0.035

85

0.021

0.060

0.035

86

0.021

0.060

0.034

87

0.020

0.059

0.033

88

0.020

0.059

0.033

89

0.020

0.058

0.032

90

0.020

0.058

0.032

91

0.019

0.057

0.031

92

0.019

0.057

0.030

93

0.019

0.056

0.030

94

0.018

0.056

0.029

95

0.018

0.055

0.029

96

0.018

0.055

0.028

97

0.018

0.055

0.028

98

0.017

0.054

0.027

99

0.017

0.054

0.027

100

0.017

0.053

0.026

0.029

0.025

112

=0.5, =975, =0.3, h=2.5L


Proportion of Load taken by the Pile Base, Pb/Pt

Pile Slenderness

Randolph and

Randolph and Wroth,

Wroth, R&W

R&W

(rigid piles)

(compressible piles)

0.267

0.319

0.316

0.231

0.313

0.310

0.204

0.303

0.299

0.183

0.292

0.287

0.166

0.281

0.276

0.153

0.271

0.265

10

0.141

0.261

0.255

11

0.132

0.252

0.245

12

0.124

0.244

0.236

13

0.117

0.236

0.227

14

0.110

0.229

0.219

15

0.105

0.222

0.212

16

0.100

0.215

0.205

17

0.096

0.209

0.198

18

0.092

0.204

0.192

19

0.088

0.199

0.186

20

0.085

0.194

0.180

21

0.082

0.189

0.175

22

0.079

0.184

0.170

23

0.076

0.180

0.165

24

0.074

0.176

0.161

25

0.072

0.172

0.156

26

0.070

0.169

0.152

27

0.068

0.165

0.148

28

0.066

0.162

0.144

29

0.064

0.159

0.141

30

0.063

0.156

0.137

31

0.061

0.153

0.134

32

0.060

0.150

0.130

33

0.058

0.148

0.127

34

0.057

0.145

0.124

35

0.056

0.143

0.121

36

0.055

0.140

0.118

37

0.054

0.138

0.116

38

0.053

0.136

0.113

39

0.051

0.134

0.110

40

0.050

0.132

0.108

Ratio,

Poulos and Davis,

L/r0

P&D

Plaxis

0.391

0.320

0.270

0.242

0.213

0.186

0.173

0.156

0.146

0.131

0.118

0.104

0.096

113

41

0.050

0.130

0.105

42

0.049

0.128

0.103

43

0.048

0.126

0.101

44

0.047

0.124

0.099

45

0.046

0.122

0.096

46

0.045

0.121

0.094

47

0.044

0.119

0.092

48

0.044

0.118

0.090

49

0.043

0.116

0.089

50

0.042

0.115

0.087

51

0.041

0.113

0.085

52

0.041

0.112

0.083

53

0.040

0.110

0.081

54

0.039

0.109

0.080

55

0.039

0.108

0.078

56

0.038

0.106

0.076

57

0.037

0.105

0.075

58

0.037

0.104

0.073

59

0.036

0.103

0.072

60

0.036

0.102

0.070

61

0.035

0.101

0.069

62

0.034

0.100

0.068

63

0.034

0.099

0.066

64

0.033

0.098

0.065

65

0.033

0.097

0.064

66

0.032

0.096

0.062

67

0.032

0.095

0.061

68

0.031

0.094

0.060

69

0.031

0.093

0.059

70

0.030

0.092

0.058

71

0.030

0.091

0.057

72

0.029

0.090

0.056

73

0.029

0.089

0.054

74

0.028

0.088

0.053

75

0.028

0.088

0.052

76

0.028

0.087

0.051

77

0.027

0.086

0.050

78

0.027

0.085

0.049

79

0.026

0.085

0.048

80

0.026

0.084

0.048

81

0.026

0.083

0.047

82

0.025

0.082

0.046

83

0.025

0.082

0.045

0.080

0.069

0.060

0.052

114

84

0.025

0.081

0.044

85

0.024

0.080

0.043

86

0.024

0.080

0.042

87

0.024

0.079

0.042

88

0.023

0.078

0.041

89

0.023

0.078

0.040

90

0.023

0.077

0.039

91

0.022

0.077

0.039

92

0.022

0.076

0.038

93

0.022

0.076

0.037

94

0.022

0.075

0.036

95

0.021

0.074

0.036

96

0.021

0.074

0.035

97

0.021

0.073

0.034

98

0.021

0.073

0.034

99

0.020

0.072

0.033

100

0.020

0.072

0.033

0.047

0.041

115

Appendix C Load Settlement Curves Determined in the


Elastoplastic Modelling for Numerical Validation

Rinter=0.01

Rinter=1.00

s (mm)

Pb (kN)

Ps (kN)

Pt (kN)

Pb (kN)

Ps (kN)

Pt (kN)

0.0

0.00

0.00

0.00

0.00

0.00

0.00

0.0

29.23

0.42

29.66

29.23

0.42

29.66

0.1

30.07

0.49

30.56

29.54

7.79

37.33

0.2

30.88

0.54

31.42

29.84

15.16

45.00

0.3

31.68

0.60

32.28

30.13

22.53

52.67

0.4

32.44

0.65

33.09

30.42

29.91

60.33

0.5

33.18

0.71

33.89

30.70

37.29

68.00

0.6

33.90

0.76

34.66

30.99

44.67

75.66

0.7

34.61

0.81

35.42

31.27

52.06

83.33

0.8

35.31

0.86

36.17

31.55

59.44

90.99

0.9

35.97

0.92

36.89

31.82

66.83

98.65

1.0

36.62

0.98

37.60

32.09

74.22

106.31

1.5

39.58

1.26

40.84

33.40

111.19

144.59

2.0

42.21

1.55

43.76

34.68

148.14

182.81

2.5

44.58

1.80

46.38

35.90

185.03

220.94

3.0

46.71

2.05

48.76

37.09

221.83

258.92

3.5

48.67

2.30

50.97

38.24

258.49

296.72

4.0

50.45

2.54

52.98

39.36

294.95

334.31

4.5

52.10

2.73

54.83

40.46

331.11

371.57

5.0

53.68

2.86

56.54

41.54

366.79

408.33

5.5

55.45

2.90

58.35

42.58

400.88

443.46

6.0

56.79

2.92

59.71

43.62

432.22

475.84

6.1

57.11

2.92

60.02

43.81

438.13

481.94

6.2

57.43

2.91

60.34

44.02

443.30

487.32

6.3

57.67

2.90

60.58

44.26

447.28

491.54

6.4

57.99

2.89

60.88

44.49

451.27

495.76

6.5

58.27

2.89

61.15

44.71

455.19

499.90

6.6

58.54

2.88

61.42

44.93

458.99

503.92

6.7

58.80

2.87

61.68

45.16

462.63

507.79

6.8

59.07

2.87

61.94

45.39

466.14

511.53

6.9

59.33

2.86

62.19

45.61

469.50

515.11

7.0

59.59

2.86

62.45

45.83

472.61

518.44

7.1

59.91

2.85

62.75

46.05

475.50

521.55

7.2

60.12

2.84

62.97

46.28

478.06

524.34

7.5

60.91

2.82

63.73

47.00

480.14

527.14

8.0

62.20

2.79

64.98

48.08

490.54

538.62

116

9.0

64.44

2.75

67.19

50.25

498.45

548.71

10.0

66.80

2.70

69.50

52.29

505.38

557.67

12.0

71.01

2.63

73.64

56.08

521.24

577.32

14.0

74.97

2.56

77.53

59.50

535.56

595.05

16.0

78.70

2.49

81.19

62.67

549.35

612.02

18.0

82.18

2.45

84.62

65.63

561.87

627.51

20.0

85.46

2.40

87.86

68.36

573.37

641.72

22.0

88.59

2.35

90.94

70.90

584.00

654.90

24.0

91.61

2.30

93.91

73.26

594.27

667.53

26.0

94.48

2.25

96.73

75.54

603.58

679.12

28.0

97.26

2.20

99.47

77.67

612.51

690.18

30.0

99.98

2.16

102.14

79.65

621.22

700.86

117

Rinter=0.870 or Rinter=0.857
=0

=23

s (mm)

Pb (kN)

Ps (kN)

Pt (kN)

Pb (kN)

Ps (kN)

Pt (kN)

0.0

0.00

0.00

0.00

0.00

0.00

0.00

0.0

29.23

0.42

29.66

29.23

0.42

29.66

0.1

29.55

7.76

37.30

29.55

7.75

37.31

0.2

29.85

15.10

44.95

29.87

15.08

44.95

0.3

30.15

22.45

52.60

30.18

22.42

52.60

0.4

30.44

29.80

60.24

30.48

29.76

60.25

0.5

30.73

37.15

67.88

30.79

37.10

67.89

0.6

31.02

44.50

75.52

31.09

44.44

75.54

0.7

31.31

51.86

83.17

31.39

51.79

83.18

0.8

31.59

59.21

90.81

31.69

59.13

90.82

0.9

31.87

66.57

98.44

31.98

66.48

98.46

1.0

32.15

73.93

106.08

32.28

73.83

106.11

1.5

33.48

110.76

144.25

33.72

110.57

144.29

2.0

34.78

147.57

182.35

35.16

147.28

182.44

2.5

36.06

184.27

220.34

36.63

183.86

220.49

3.0

37.41

220.72

258.13

38.22

220.17

258.39

3.5

38.78

256.84

295.62

39.88

256.14

296.03

4.0

40.18

292.45

332.64

41.62

291.63

333.25

4.5

41.63

326.68

368.31

43.45

325.77

369.23

5.0

43.11

357.70

400.82

45.41

356.77

402.18

5.5

44.82

378.61

423.43

47.72

377.99

425.71

6.0

46.60

390.54

437.14

50.22

390.00

440.22

6.1

46.96

391.45

438.41

50.75

391.06

441.81

6.2

47.34

391.14

438.48

51.29

390.99

442.28

6.3

47.71

391.22

438.93

51.83

390.89

442.71

6.4

48.07

391.18

439.26

52.36

390.87

443.23

6.5

48.43

391.23

439.66

52.88

390.87

443.74

6.6

48.78

391.20

439.98

53.39

390.86

444.25

6.7

49.13

391.24

440.37

53.90

390.85

444.75

6.8

49.48

391.20

440.68

54.41

390.84

445.25

6.9

49.82

391.25

441.06

54.91

390.83

445.74

7.0

50.15

391.21

441.37

55.41

390.81

446.23

7.1

50.49

391.26

441.74

55.91

390.80

446.71

7.2

50.82

391.22

442.04

56.40

390.79

447.19

7.5

51.79

391.37

443.15

57.86

390.90

448.76

8.0

53.36

389.68

443.04

60.28

389.01

449.30

9.0

56.24

390.87

447.11

64.79

389.71

454.50

10.0

58.93

391.14

450.07

69.08

389.67

458.76

12.0

63.82

392.08

455.90

77.09

390.75

467.84

118

14.0

68.24

392.07

460.31

84.56

389.96

474.53

16.0

72.30

392.08

464.38

91.60

389.45

481.05

18.0

76.09

392.22

468.31

98.29

389.14

487.43

20.0

79.68

392.29

471.97

104.69

388.94

493.63

22.0

83.07

392.25

475.32

110.85

389.05

499.90

24.0

86.27

392.20

478.47

116.82

388.72

505.53

26.0

89.29

392.28

481.57

122.60

388.15

510.75

28.0

92.18

392.24

484.42

128.23

387.88

516.11

30.0

94.96

392.19

487.15

133.72

387.45

521.17

119

Appendix D1 Load Settlement Curve Determined in the


Elastoplastic Modelling for Comparison with the Static Load Test

s (mm)

P (kN)

0.00

0.15

0.60

426.49

1.55

1049.62

2.65

1668.84

4.25

2474.16

3.80

2153.25

3.25

1762.21

2.35

1131.30

0.80

155.19

1.50

653.12

2.50

1359.82

3.00

1696.23

4.00

2323.66

5.00

2810.61

7.50

3759.55

9.50

4426.71

12.50

5096.28

18.00

6270.01

27.50

6992.60

63.00

7785.68

62.00

7035.90

58.50

4788.97

56.00

3470.83

53.00

2021.50

50.00

812.07

120

Appendix D2 Normal Stress along the Pile in the Elastoplastic


Modelling for Comparison with the Static Load Test

Load Step 4
Level

z (m)

N (kN)

2.1

2474.01

25.7

1305.82

31.7

841.81

32.7

765.20

36.7

489.26

40.7

234.33

Load Step 19
Level

z (m)

N (kN)

2.1

7826.85

25.7

6434.54

31.7

5675.60

32.7

5420.72

36.7

4424.93

40.7

3751.34

121

Appendix D3 Shaft Load Curves Determined in the


Elastoplastic Modelling for Comparison with the Static Load Test

Load

s (mm)

P (kN)

R24

R45

R56

R68

0.00

0.15

1.53

1.44

1.35

1.40

0.60

426.49

6.08

5.36

4.83

4.85

1.55

1049.62

13.34

11.59

10.35

10.32

2.65

1668.84

22.31

19.17

17.05

16.88

4.25

2474.16

32.46

32.94

28.86

28.09

3.80

2153.25

29.00

30.03

26.27

25.52

3.25

1762.21

24.83

26.44

23.08

22.36

2.35

1131.30

18.01

20.57

17.87

17.19

0.80

155.19

7.92

10.35

8.82

8.26

1.50

653.12

11.39

14.92

12.87

12.28

10

2.50

1359.82

18.97

21.45

18.67

18.02

11

3.00

1696.23

22.77

24.71

21.56

20.89

12

4.00

2323.66

30.46

31.30

27.40

26.66

13

5.00

2810.61

34.53

41.26

35.66

34.30

14

7.50

3759.55

43.49

75.78

63.91

49.36

15

9.50

4426.71

48.37

103.76

92.11

51.20

16

12.50

5096.28

48.27

103.23

96.86

56.78

17

18.00

6270.01

48.75

103.23

96.94

66.94

18

27.50

6992.60

49.17

103.36

97.12

68.21

19

63.00

7785.68

51.87

105.75

99.56

70.77

20

62.00

7035.90

46.89

103.08

97.08

67.49

21

58.50

4788.97

25.47

81.10

78.10

50.76

22

56.00

3470.83

24.08

62.22

61.48

35.05

23

53.00

2021.50

18.26

34.69

38.80

19.03

24

50.00

812.07

15.02

3.31

10.74

17.65

Step

122

Appendix D4 Load Settlement Curves Determined in the


Elastoplastic Modelling for Comparison with the Static Load Test

Plaxis
s (mm)

Pb (kN)

Ps (kN)

P (kN)

0.80

233.82

248.35

155.19

1.50

249.75

730.35

653.12

2.50

272.49

1414.30

1359.82

3.00

283.82

1739.39

1696.23

4.00

306.54

2344.10

2323.66

5.00

335.40

2802.19

2810.61

7.50

431.82

3654.71

3759.55

9.50

531.09

4222.60

4426.71

12.50

769.05

4654.21

5096.28

18.00

1199.29

5397.69

6270.01

27.50

1766.71

5552.86

6992.60

63.00

2460.03

5652.62

7785.68

SLT
s (mm)

Pb (kN)

Ps (kN)

P (kN)

Load Step 4

4.25

520.00

2130.00

2650.00

Load Step 19

63.00

2450.00

5400.00

7850.00

123