The University of Sydney

School of Civil Engineering

Analysis of Piled-Raft Foundations
with
Piles of Different Lengths and Diameters



by




Helen Sze Wai Chow, B.E, M.E., M.B.A.





A thesis submitted for the Degree of Doctor of Philosophy
in the University of Sydney

August 2007
i
SYNOPSIS

In recent years, there have been an increasing number of structures using piled rafts as the
foundation to reduce the overall and differential settlements. For cases where a piled raft
is subjected to a non-uniform loading, the use of piles with different sizes can improve
the performance of the foundation. Extensive research work has been performed in the
past to examine the behaviour of piled rafts. However, most of the research was focused
on piled rafts supported by identical piles, and the use of non-identical piles has not
received much attention.

In this thesis, the behaviour of piled rafts supported by non-identical piles is examined by
the use of a computer program APRILS based on the finite layer and finite element
methods. The finite layer method is used for the analysis of the layered soil system. The
application of this method to different shapes of loadings is presented and has been
shown to be in good agreement with the theoretical solutions. The finite element method
is used for the analysis of the raft and piles. Full interaction between raft, piles and soil
which is of major importance in the behaviour of piled rafts is considered in the analysis.
Among the four different types of interaction present in the piled raft foundation, the
interaction between piles plays an important role.

Interaction between non-identical piles subjected to both horizontal and vertical loadings
is examined. It is found that for a pair of piles under consideration, if the diameter and
length of one of the piles is a multiple of the diameter or length of the other pile, the
reciprocal theorem holds.

Program APRILS can be used for the analysis of piled rafts subjected to horizontal and
vertical loadings in which the base of the raft is treated as being rough and slip along the
pile-soil interface is allowed in the analysis. By having a rough raft base, there is a
significant effect on the vertical movement of the raft due to the lateral resistance of the
soil and the piles. For piled rafts embedded in a non-homogeneous layered soil, the
ii
modulus of each layer of soil is used in the computation and accurate solutions are
obtained without the use of an averaging technique.

The performance of piled rafts is affected by several factors such as the soil conditions,
pile dimensions and arrangements. Detailed examinations of these factors have shown
that the displacement of the raft and the proportion of load carried by piles are
significantly affected by them. For piled rafts subjected to non-uniform vertical or
horizontal loads, the use of non-identical piles can improve the performance of the piled
rafts. For vertically loaded piled rafts, longer piles are prefered to be used underneath the
heavily loaded region while for horizontally loaded piled rafts, larger diameter piles are
more preferable.

Several field cases are studied in this thesis, and they show that the solutions obtained
from program APRILS are in good agreement with field measurements. The use of the
insitu test results and back-analysis is used to obtain the correlation between the modulus
of the soil and field test results for laterally loaded piles.
iii
PREFACE

This thesis is submitted to the University of Sydney, Australia, for the degree of Doctor
of Philosophy. The work described in this thesis was carried out by the candidate during
the years 2002-2007 in the School of Civil Engineering at the University of Sydney,
under the supervision of Professor John Small.

The By-laws of the University of Sydney require a candidature for the degree of Doctor
of Philosophy to indicate the originality of the work. Unless otherwise indicated in the
text, the candidature submits that the work presented in this thesis is original and it
includes the following:

The solution of problems for different types of interaction due to the application
of horizontal or vertical loadings.

The numerical method (Method II) for the analysis of piled rafts with piles of
different sizes and embedded in layered soils subjected to uniform or non-uniform
horizontal and vertical loadings.

The solutions for problems involving rafts, pile groups and piled rafts subjected to
uniform or non-uniform horizontal and vertical loadings in homogeneous or non-
homogeneous soils.

The technique for predicting the modulus of soil based on the field test results.

iv
The following supporting papers have been written based on the work described in this
thesis:
1. Small, J. C., Zhang, H. H. and Chow, H. (2004). Behaviour of Piled Rafts with
Piles of Different Lengths and Diameters., Proc. 9
th
Australia New Zealand
Conference on Geomechanics, 8-11 February, Auckland, New Zealand, Vol. 1, pp.
123-129.
2. Chow, H. and Small, J. C. (2005a). Behaviour of Piled Rafts with Piles of
Different Lengths and Diameters under Vertical Loading. Geotechnical Special
Publication No. 130-142, Geo-Frontiers, 2005, Austin, USA, pp. 841-855.
3. Chow, H. and Small, J. C. (2005b). Finite Layer Analysis of Combined Pile-Raft
Foundations with Piles of Different Lengths., Proc. 11
th
International Conference
on Computer Methods and Advances in Geomechanics (IACMAG), 19-24 June,
Torino, Italy, Edited by G. Barla and M. Barla, Vol. 2, pp. 429-436.
4. Chow, H. S. W. and Small, J. C. (2006a). Settlement of a Piled-Raft Foundation
considering Lateral Pile Resistance., Proc. 10th International Conference on
Piling and Deep Foundations, 31 May 2 June, Amsterdam, The Netherlands, pp.
232-241.
5. Chow, H. S. W. and Small, J. C. (2006b). Analysis of piled raft foundations with
piles of different lengths subjected to horizontal and vertical loads., Proc. 6th
European Conference on Numerical Methods in Geotechnical Engineering
(NUMGE06), 6-8 September, Graz, Austria, Edited by H. Schweiger, pp. 583-588.
6. Chow, H. and Small, J. C. (2007). Effect of a thin soft layer on the settlement of
piled rafts (accepted for publication in 10
th
Australia New Zealand Conference
on Geomechanics).
v
ACKNOWLEDGEMENTS

This thesis was made possible by the great contribution of my supervisor Professor John
Small. I am deeply indebted to him for his outstanding supervision, encouragement and
guidance throughout the period of my candidature.

I would like to express my appreciation to Dr. Tim Hull and Nigel Balaam for their
valuable advice. Thanks are extended to the staff at the Centre for Geotechnical Research,
Mr. Ross Barker and Antonio Reyno for their assistance and friendship.

I would like to thank the staff and postgraduates in the School of Civil Engineering.
Special thanks are due to my fellow research students: Dr. Ezzat William, Dr. Bosco
Poon, Ms. Nooshin Jabiri, Mr. Jurgen Becque, Mr. Ryan Chen, Mr. Benoit Gilbert, Mr.
David Cao, Mr. Derrick Yap, Mr. Niphan Yaiaroon, Mr. Frank Zhu, Mr. Ali Hanandeh,
Mr. Thanh Binh Nguyen, Mr. Cao Hung Pham and Mr. Tayakorn Chandranqsu for their
friendship and understanding.

Finally, I would like to thank my family, especially to my mother for her understanding,
encouragement, support and patience, making it possible for me to pursue the challenging
work of my interest.
vi
NOTATION


The following are some of the more frequently used symbols in this thesis. Symbols used
infrequently are not included in the list.

English Letters
A cross-sectional area of pile
A
i
area of element i in the raft
A
x
axial force along the pile
a radius of ring or circular load
B or B
r
width of raft
C circumference of pile
c
a
pile-soil adhesion
D or d diameter of pile
E

Youngs modulus
E
p
Youngs modulus of pile
E
r
Youngs modulus of raft
E
s
Youngs modulus of soil
F vector of nodal forces
G shear modulus of soil
h depth of soil
I influence factor
[I
p
] influence matrix of pile group
[I
r
] influence matrix of the pinned raft
[I
s
] influence matrix of the soil
[I
sp
] influence matrix of the pile enhanced soil continuum
[I
pr
] influence matrix of the pinned piled raft
[k]
i
stiffness of pile element i
[K
p
] stiffness matrix of pile group
vii
K
rs
raft-soil stiffness matrix
[K
s
] stiffness matrix of the soil
L or l length of the pile
L
r
length of the raft
M
b
bending moment along the pile
M
x
, M
y
total applied bending moments in the x- and y-directions respectively
M
z
bending moment about the z-axis
[P
p
] vector of nodal forces acting at the nodes along the pile
P
pr
, P
r
contact stresses on the raft elements
[P
s
] vector of nodal forces at the soil nodes
[P
sp
] vector of interface loads between the raft and the pile enhanced soil
P
u
ultimate load capacity of pile
P
x
, P
y
, P
z
total applied loadings in the x-, y- and z-directions respectively
Q total applied load
Q
x
, Q
y
, Q
z
total external loads in the x-, y- and z-directions applied to the pile group
q uniform distributed load
r distance from the centre of the loaded area
S shear forces along the pile
s spacing between piles
s
u
shear strength of soil
S
z
Hankel transform for vertical loads
T
x
, T
y
Hankel transform for horizontal circular and ring loads in the x- and y-
directions respectively
t
r
thickness of the raft
u horizontal displacement
U
x
, U
y
, U
z
transform of displacements in the x-, y-, and z-directions
u
x
, u
y
, u
z
displacements in the x-, y-, and z-directions
v vertical displacement
W width of raft
x, y, z Cartesian co-ordinates

viii
Greek Letters

pp
pile-pile interaction factor

ps
pile-soil interaction factor

sp
soil-pile interaction factor

ss
soil-soil interaction factor
displacement vector

l
vector of soil displacements due to a unit surface load in the x-, y- and z-
directions

p
vector of pile displacements at the nodes in the x-, y- and z-directions

pr0
vector of displacements at the centre of each raft element or at the nodes
of each pile in the x-, y- and z-directions for the pinned raft

r
vector of displacements at the centres of the raft elements in the x-, y- and
z-directions

r0
vector of displacements at the centres of each raft element in the x-, y- and
z-directions for the pinned raft

s
vector of soil displacements at the soil nodes or at the centres of soil
elements

x
,
y
,
z
translation of the pinned raft or piled raft in the x-, y- and z-directions
vector of strains
pile-soil relative stiffness

r
Poissons ratio of the raft

s
Poissons ratio of the soil

x
,
y
(1) rotations of the pile head about the x- and y- axes
(2) rotations of the pinned raft about the x- and y- axes
(3) rotations of the pinned piled raft about the x- and y- axes

z
rotations of the pinned raft or pinned piled raft about the z-axis
vector of stresses
ix
CONTENTS
Synopsis i
Preface iii
Acknowledgements v
Notation vi
Contents ix
CHAPTER 1 INTRODUCTION 1
1.1 BACKGROUND 1
1.2 OBJECTIVES 2
1.3 THESIS OUTLINE 3

CHAPTER 2 LITERATURE REVIEW 6
2.1 INTRODUCTION 6
2.2 RAFT FOUNDATIONS 6
2.2.1 Analytical Methods 7
2.2.2 Boundary Element Methods 7
2.2.3 Finite Element Methods 9
2.2.4 Hybrid Approach 11
2.3 PILE GROUP FOUNDATIONS 12
2.3.1 Simplified Analytical Method 12
2.3.2 Hybrid Method 13
2.3.3 Boundary Element Method 14
2.3.4 Finite Element Method 19
2.3.5 Infinite Layer Method 20
2.3.6 Finite Layer Method 21
x
2.4 PILED RAFT FOUNDATIONS 23
2.4.1 Approximation Method 23
2.4.2 Boundary Element Method 24
2.4.3 Finite Element Method 26
2.4.4 Combined Boundary Element and Finite Element Method 29
2.4.5 Combined Finite Layer and Finite Element Method 32
2.4.6 Variational Approach 32
2.5 CONCLUSIONS 34

CHAPTER 3 FINITE LAYER ANALYSIS OF LAYERED SOIL 38
3.1 INTRODUCTION 38
3.2 ANALYSIS OF LAYERED SOIL 39
3.2.1 Circular and Ring Loads 39
3.2.1.1 Basic Equations for Elastic Materials 39
3.2.1.2 Hankel Transforms 42
3.2.2 Rectangular Loads 46
3.2.2.1 Basic Equations for Elastic Materials 47
3.2.2.2 Fourier Transformations 48
3.2.2.3 Transformed Equations in the , Co-ordinate System 48
3.2.3 Stiffness Relationship 52
3.2.4 Transform of Loads 53
3.2.4.1 Hankel Transform of Vertical and Horizontal Ring Load 53
3.2.4.2 Hankel Transform of Vertical and Horizontal Circular Load 54
3.2.4.3 Hankel Transform of Vertical and Horizontal Rectangular Load 55
3.2.5 Solutions for Displacements 55
3.2.5.1 Soil Displacements for Vertical Ring or Circular Loads 55
3.2.5.2 Soil Displacements for Horizontal Ring or Circular Loads 56
3.2.5.3 Soil Displacements due to Vertical Rectangular Loads 57
3.2.5.4 Soil Displacements due to Horizontal Rectangular Loads 58
3.3 NUMERICAL INTEGRATION 59
3.4 NUMERICAL EXAMPLES 62
xi
3.4.1 Soil subjected to Vertical Loadings 62
3.4.1.1 Ring Loads 62
3.4.1.2 Circular Loads 63
3.4.1.3 Rectangular Loads 64
3.4.2 Soil subjected to Horizontal Loadings 65
3.4.2.1 Ring Loads 65
3.4.2.2 Circular Loads 66
3.4.2.3 Rectangular Loads 67
3.5 CONCLUSIONS 68

CHAPTER 4 Analysis of Rafts and Pile Groups in Layered Soil 91
4.1 INTRODUCTION 91
4.2 ANALYSIS OF RAFTS 92
4.2.1 Raft Subjected to Membrane Action 92
4.2.2 Raft Subjected to Bending Action 94
4.2.3 Combination of Membrane and Bending Actions 96
4.2.4 Rafts on Layered Soils 96
4.2.4.1 Influence Matrix for the Raft 96
4.2.4.2 Influence Matrix for the Soil 98
4.2.4.3 Analysis of Foundations 99
4.3 ANALYSIS OF PILE GROUPS 101
4.3.1 Stiffness Matrix of Pile Group 101
4.3.2 Stiffness Matrix for Layered Soil 103
4.3.3 Pile Groups in Layered Soil 104
4.4 NUMERICAL EXAMPLES 105
4.4.1 Examples for Rafts 105
4.4.2 Examples for Pile Groups 110
4.5 CONCLUSIONS 115


xii
CHAPTER 5 SOIL-STRUCTURE INTERACTION 140
5.1 INTRODUCTION 140
5.2 EFFECT OF SUPERSTRUCTURE ON FOUNDATIONS 140
5.3 INTERACTION MECHANISM OF PILED RAFT 142
5.3.1 Pile-Pile Interaction 142
5.3.2 Pile-Soil Interaction 144
5.3.3 Soil-Pile Interaction 144
5.3.4 Soil-Soil Interaction 146
5.4 NUMERICAL EXAMPLES 146
5.4.1 Pile-Pile Interaction 146
5.4.2 Pile-Soil Interaction 152
5.4.3 Soil-Pile Interaction 153
5.4.4 Soil-Soil Interaction 153
5.5 CONCLUSIONS 154

CHAPTER 6 ANALYSIS OF PILED RAFTS 171
6.1 INTRODUCTION 171
6.2 METHODS OF ANALYSIS 172
6.2.1 Method I Isolated Raft and Pile Group Embedded in Layered Soil 173
6.2.2 Method II Piled Raft and the Layered Soil 174
6.2.2.1 Analysis of the Isolated Piled Raft 174
6.2.2.2 Analysis of the Layered Soil 178
6.2.2.3 Analysis of a Piled Raft in a Layered Soil 180
6.2.3 Comparison of the Two Methods 181
6.2.4 Non-Linear Analysis of Piled Raft 184
6.2.5 Verification of Results with Other Numerical Solutions 185
6.3 CONCLUSIONS 194


xiii
CHAPTER 7 PERFORMANCE OF PILED RAFTS 233
7.1 INTRODUCTION 233
7.2 PARAMETRIC STUDIES 234
7.2.1 Types of Foundation 235
7.2.2 Soil Conditions 237
7.2.3 Dimension of Piles 241
7.3 CONCLUSIONS 249

CHAPTER 8 Application of Program to Piled Raft Foundations 285
8.1 INTRODUCTION 285
8.2 CASE STUDIES 285
8.2.1 Case 1: Centrifuge Model Tests 285
8.2.2 Case 2: Messe-Torhaus Building, Frankfurt, Germany 288
8.2.3 Case 3: Liquid Gas Terminal, Gdansk, Poland 289
8.2.4 Case 4: Roosevelt Bridge, Florida 291
8.2.5 Case 5: Bridge Foundation, Portugal 294
8.3 CONCLUSIONS 298

CHAPTER 9 Conclusions and Recommendations for Future Research
319
9.1 CONCLUSIONS 319
9.2 RECOMMENDATIONS FOR FUTURE RESEARCH 325


APPENDIX I Exact Solution for Auxiliary Function 327
APPENDIX II Flexibility Matrix for a Single Layer of Material
328
xiv
APPENDIX III Boundary Conditions for Circular and
Rectangular Loadings Applied to a Layered
System 330
APPENDIX IV Shape Function for Raft Elements (8 Noded
Isoparametric Element) 331
REFERENCES 332
Chapter 1 - Introduction
1
CHAPTER 1
INTRODUCTION

1.1 BACKGROUND
For most piled raft foundations, the primary purpose of the piles is to act as settlement
reducers. The proportion of load carried by the piles is considered as a secondary issue in
the design. Over the past decades, extensive research work has been presented, aimed at
improving the accuracy in the prediction of the behaviour of piled rafts.

In the design of piled rafts, design engineers have to understand the mechanism of load
transfer from the raft to the piles and to the soil to predict (i) the behaviour of the raft
which includes the settlements, bending moments and the proportion of load carried by
the raft, and (ii) the behaviour of the piles which includes the displacements and load
distributions along the piles. Interactions between piles, raft and soil are of major concern
in the analysis. The concept of interaction between piles introduced by Poulos (1968) was
used in the analysis of pile groups and can be extended to the analysis of piled rafts.

Methods that have been used for analysis range from simplified calculations to numerical
methods such as the boundary element method (Butterfield and Banerjee, 1971b; Brown
and Wiesner, 1975; Kuwabara, 1989; Mendona and de Paiva, 2000) and the finite
element method (Hooper, 1973; Ottaviani, 1975; Chow 1987a; Liu and Novak, 1991;
Katzenbach and Reul, 1997; Prakoso and Kulhawy, 2001; Reul and Randolph, 2003). In
early years, because of the limited availability of computer memory and processing speed,
the use of numerical methods was confined to simple problems. In the last two decades
due to the rapid development in computer technologies, numerical methods such as full
three-dimensional finite element methods are often used to solve complex problems.

A finite layer technique based on the flexibility approach was developed by Small and
Booker (1984 and 1986) to determine the soil deflections of horizontally layered soil
Chapter 1 - Introduction
2
subjected to different shaped loadings by the use of Fourier transforms. This technique
was used in combination with the finite element technique for the analysis of rafts, pile
groups and piled rafts embedded in layered soils (Zhang and Small, 1991, Lee and Small,
1991a and b; Ta and Small, 1995; Southcott and Small, 1996) subjected to vertical
loadings and then extended to horizontal loadings (Zhang and Small, 1999 and 2000). In
the above analyses, the raft was modelled as a thin elastic plate and analysed by a two-
dimensional finite element method, the piles were modelled as rod or beam elements and
analysed by a one-dimensional finite element method and the horizontally layered soil
was analysed by the finite layer method. Full interaction between the raft, piles and soil
was taken into account in the analyses. However, there are some limitations with these
methods:
(1) an identical size for each of the piles and a uniform thickness for the raft was used
(2) the piles are contained within the raft element which may not be a good simulation of
the loads acting at the pile heads for large piled rafts.
(3) analysis is separated into an isolated raft and a pile group embedded in the soil and
so there is no moment transfer from the pile head to the raft.
(4) applied moments have to be transformed into equivalent forces along the edges of the
raft elements

1.2 OBJECTIVES
For cases where the superstructures on the piled rafts apply non-uniform loadings, the use
of piles of different dimensions underneath the raft might be preferable to minimize the
overall and differential settlements and the tilting of the superstructures. Tilting can be a
significant problem for a tall building as a small tilt can mean a large horizontal
movement at the top of the building. Most of the recent research has been performed for
piled rafts with piles of identical size.

The objective of this thesis is to develop a numerical method for the analysis of piled
rafts with piles of different lengths and diameters by the use of the finite layer method for
the analysis of the layered soil and the finite element method for the analysis of the piles
Chapter 1 - Introduction
3
and the raft. This new method overcomes the limitations mentioned in (1 4) above and
has the following features:
(1) both horizontal and vertical loadings can be applied to the piled raft by treating the
base of the raft as being rough.
(2) the piled raft is analysed as a whole structure by attaching the piles to the nodes on
the raft
(3) piles can have different lengths and diameters and the raft can have different
thicknesses and any shape
(4) applied moments are directly transferred from the raft to the pile heads
(5) slip along the pile-soil interface is allowed in the analysis

Effects of the dimensions of piles on different kinds of interaction and on the overall
behaviour of the piled raft subjected to horizontal and vertical loadings will be examined
by the use of the new method developed in this thesis. Results obtained from this new
method are compared with those from existing methods and also with field measurements.

1.3 THESIS OUTLINE
The thesis is presented in the following eight chapters.

Chapter 2 presents a brief review of previous research work on rafts, pile groups and
piled raft foundations. The review outlines the development of different numerical
methods for the analysis of the different types of foundations.

Chapter 3 presents the general concept of the finite layer theory and the application of the
theory to solve for the displacements in a layered soil subjected to vertical or horizontal
circular, ring or rectangular loads at the soil surface or soil interface. The displacements
can be obtained at any point in a layered soil system. The accuracy of the solutions
depends on the numerical integration schemes.

Chapter 1 - Introduction
4
Chapter 4 presents the use of the finite layer method in the analysis of rafts resting on the
soil surface and pile groups embedded in layered soils. The foundation system is analysed
by a combination of the finite layer and finite element methods. The analysis of rafts is
performed by the use of a two-dimensional finite element method which models the raft
as a thin plate. The raft can be of any shape and is divided into a number of rectangular
elements. The raft elements can have different sizes and thicknesses. The soil surface
underneath the raft is divided into a number of rectangular elements corresponding to the
rectangular elements of the raft and is analysed by the finite layer method as described in
Chapter 3 for rectangular loads. The analysis of piles is performed by the use of a one-
dimensional finite element method in which the piles are modelled as beam elements.
The loads acting along the pile shaft are modelled as a series of ring loads and the base
load is modelled by a uniform circular load at the pile base. The layered soil is analysed
by the finite layer method for circular and ring loads. Results obtained from the present
method for the raft and pile groups subjected to horizontal and vertical loadings are
presented in this chapter. Comparisons with the results from the other numerical methods
are also presented which show that the present method is capable of producing accurate
results.

Chapter 5 presents the interaction mechanism of piled rafts which is comprised of four
different types of interaction. Methods used for the computation of each type of
interaction are described. Interaction factors between two identical piles embedded in
different soil conditions are presented and compared with other existing methods.
Interaction factors between two non-identical piles and the effect of the size (i.e. diameter)
and length of piles on different kinds of interactions are also presented in this chapter.

Chapter 6 presents a new method (Method II) for the analysis of piled rafts in a layered
soil system based on the theories described in Chapters 4 and 5. This new method is
compared with Method I which was developed in previous research work based on the
same theories. In the new method (Method II), the piled raft is analysed as a whole
structure and the piles are attached to the nodes on the raft. The interface between the raft
and soil is treated as being rough. Slip along the pile-soil interface is also allowed in the
Chapter 1 - Introduction
5
new method. Comparison of the results between a smooth and a rough rigid base shows
that the raft interface has significant effect on the behaviour of the raft. The examples
presented in this chapter show that the number of piles, pile length and raft thickness also
have effects on the behaviour of piled rafts with identical piles.

Chapter 7 presents comparisons of the performance of different types of foundations. The
comparisons show that piled raft is effective in reducing the overall and differential
settlement and also the skin friction at the pile heads for vertical loadings. Factors such as
the soil conditions, pile spacing and pile dimensions on the performance of piled rafts
with identical or non-identical piles with respect to the displacements, bending moments
and load distributions for the rafts and piles are carefully examined. Non-uniform or
uniform loadings are applied to the piled rafts.

Chapter 8 presents several case studies that include centrifuge model tests, commercial
buildings in Europe and full scale load tests for bridges. Load tests and back analysis of a
load test on a single pile are used to obtain the soil modulus that gives a good fit to
measured performance in the analyses and then this modulus is used to determine
correlation factors that can be used with insitu test results.

Chapter 9 summarizes the main conclusions drawn from the research work and
suggestions for future research work in this field.
Chapter 2 Literature Review
6
CHAPTER 2
LITERATURE REVIEW

2.1 INTRODUCTION
In foundation design, rafts, pile groups and piled rafts are commonly used to support
structures. Extensive research work has been carried out and published in the past
decades, and different analysis methods have been developed that can be classified into
several categories: empirical, analytical and numerical methods. In this chapter, a brief
review of the techniques developed for the analysis of raft, pile group and piled raft
foundations are presented.

2.2 RAFT FOUNDATIONS
In the design of raft foundations, the soil can be treated (I) as a series of individual
springs known as a Winkler model (1867) or (II) as a continuum. The Winkler model
treats the soil as a series of springs and assumes that the pressure at any point on the
surface of the soil is related to the modulus of subgrade reaction (or Winkler spring
stiffness) and the deflection of the soil. The spring stiffness depends on the settlement
characteristic of the soil and the geometry of the foundation. This model can be easily
handled by mathematical equations and can produce reasonably accurate results.
However, it neglects the interaction between each individual spring and the supporting
soil is therefore not modelled as a continuum. An alternative approach that treats the
supporting soil as an elastic continuum can better represent the physical behaviour of the
supporting soil. The soil parameters used in this approach depend on the field stress state
and have to be carefully evaluated (Hain and Lee, 1974). Different methods ranging from
one-dimensional to full three-dimensional models have been developed for the analysis
of raft foundations.

Chapter 2 Literature Review
7
2.2.1 Analytical Methods
The use of analytical methods for the analysis of rafts on elastic foundations has been
investigated by numerous researchers. However, this approach is limited to simple
geometrical shapes of the raft and homogeneous soils. Zhemochkin and Sinitsyn (1962)
obtained the analytical solution by assuming that the contact pressures between the raft
and soil were uniform blocks of pressure. The deflections of the raft and the soil due to
the pressure could be determined by considering the compatibility of the displacements of
the raft and the soil at a number of points beneath the raft.

Brown (1969a) employed a similar method for the analysis of a circular raft on an elastic
foundation of finite depth. The raft was divided into a number of equal width annular
elements and the contact stress was assumed to be uniformly distributed over the annular
elements. Solutions from the analysis were based on the solution presented by Burmister
(1956) for a two layer system subjected to a surface point load. Based on the integral
transform techniques presented by Sneddon (1951), Brown (1969b) later presented an
improved method which provided greater accuracy and less computation. In this
improved method the contact stress was represented by a series of mathematical functions
instead of uniform annular pressures. The same method was used by Booker and Small
(1983) for analysis of liquid storage tanks resting on homogeneous soils.

Kay and Cavagnaro (1983) presented a method for the prediction of settlement for raft
foundations by the use of field parameters in which the soil can have numerous sublayers
having different properties. The raft was replaced by an equivalent uniformly loaded
circular area such that the influence of the raft stiffness was considered in the assessment
of differential settlement.
2.2.2 Boundary Element Methods
The boundary element method is a powerful tool that can be applied in engineering
applications as only the boundary has to be discritized which reduces the amount of
computer memory and the time to solve the problem. Katsikadelis and Armenkas (1984a
and 1984b) and Costa and Brebbia (1985 and 1986) used the boundary integral equation
Chapter 2 Literature Review
8
method for the analysis of plates resting on a Winkler type elastic foundation. In this
method the boundary of the plate was divided into a finite number of elements with a
node defined at the midpoint of each element. Each boundary element was approximated
by a curve so that the boundary of the plate can be approximated by straight line or
curved line segments. The domain was assumed to be bounded by a continuous curve. In
order to reduce the domain integrals, Costa and Brebbia (1985 and 1986) suggested that
the domain integrals have to be transformed into boundary integrals.

Bezine (1988) developed a new boundary element method for bending of plates on a
Winkler foundation which used an original boundary integral equation method. The
fundamental solution is for plate flexure problems based on the Kirchhoffs theory. This
method required the discretization of the boundary into a number of straight elements and
the domain into a number of rectangular panels. The distribution of contact pressures at
the interface was replaced by the equivalent loads applied at each node of the discretised
domain. The solution was obtained by Gauss-Legendre integration over the elements on
the boundary and the panels on the domain.

Analysis of unilaterally supported plates on elastic foundations have been presented by
Sapountzakis and Katsikadelis (1992) and Hu and Hartley (1994). Sapountzakis and
Katsikadelis (1992) presented a boundary element solution for unilaterally supported
plates resting on a homogeneous or non-homogeneous elastic foundation. The soil was
modelled by independent springs with the subgrade reaction dependent linearly or non-
linearly on the deflection. The contact between the plate and soil was unbonded and
separation contact between the plate and soil had been considered. Discretization along
the boundary and within the domain was required and the solution was obtained by an
iterative process. Hu and Hartley (1994) carried forward the same work by incorporating
an elastic half-space model into the analysis. A direct boundary element method was
developed which took into account the interaction between an elastic plate and an elastic
half-space. The plate and the elastic foundation were separated into two mathematical
models. The boundary of the plate was divided into boundary elements and plate bending
was modelled by the thin plate boundary integral formulation. The contact area was
Chapter 2 Literature Review
9
divided into a set of contact elements or quadratic elements. The contact pressure within
the contact elements was assumed to be constant, while for the contact pressure within a
quadratic element was expressed by a second order polynomial.

de Paiva and Butterfield (1997) presented a formulation for the analysis of plate-soil
interaction where the interface was divided into triangular elements in which the
displacements and the subgrade reaction varied linearly. Linear functions were used for
the approximation of the displacements and bending moment of the boundary elements
and the tractions on the triangular elements on the plate-soil interface. The shear force on
the boundary was approximated by reaction forces applied to the nodes of the elements.
2.2.3 Finite Element Methods
The first solution which employed the finite element method for the analysis of
foundation structures on an elastic halfspace was obtained by Cheung and Zienkiewicz
(1965). The behaviour of the raft was obtained by the finite element technique in which
the raft was divided into a number of rectangular elements joined at a discrete number of
nodal points. The soil was modelled either by the Winkler model in which interactions
between springs were not considered or by the elastic continuum model in which
separation between the raft and the soil was not allowed when negative reactions existed.
The stiffness matrix for the whole system was formed by combining the stiffness of the
soil (which was derived by using the Boussinesq equation) with the stiffness matrix of the
plate bending elements. Contact stresses were represented by equivalent forces applied at
nodal points of the finite element mesh. Cheung and Nag (1968) extended this method to
incorporate the shear stresses beneath the raft and examined the effects of uplift between
the raft and the soil. Svec and Gladwell (1973) developed an improved method for the
analysis of a thin plastic plate on an elastic half space. The plate and the surface of the
elastic half-space which was in contact with the plate were divided into a number of 10
noded triangular elements. The continuous contact pressure distribution beneath the plate
was represented by a cubic polynomial on each of the triangular regions. The
displacements at the surface of the elastic half-space due to the contract pressure were
determined from the Boussinesq equation.
Chapter 2 Literature Review
10

Wardle and Fraser (1974) and Fraser and Wardle (1976) extended the approach of
Cheung and Zienkiewicz (1965) to a multi-layered soil system with isotropic or cross-
anistropic properties. The layered soil system was divided into a number of horizontal
layers of uniform thickness with infinite lateral extent. The loaded surface of the soil
mass was discretized into surface elements corresponding to the raft elements, and the
raft was modelled by conventional finite elements. The contact between the raft and the
soil was assumed to be smooth. The surface element stiffness matrix of the soil was
derived from the surface settlements due to uniformly loaded rectangular areas by using
the integral transform techniques of Gerrard and Harrison (1971). With the use of surface
elements and integral transform techniques, the required computer storage and
computational time were less than those required for the three-dimensional finite element
method.

The approach of Cheung and Zienkiewicz (1965) was also extended by Wood and
Larnach (1974 and 1975) to include layered soils and time-dependent consolidation
effects in the analysis. Wood (1977) then extended the method further to include applied
moments. The raft can be of irregular shape subjected to non-uniform loadings and
resting on a non-homogeneous soil mass.

Hain and Lee (1974) suggested that in the analysis of raft foundations, the structure,
foundation and supporting soil have to be analysed as a system. The stiffness of the
structure can have an influence on the distribution of loads and moments transferred to
the raft. The structure-raft-supporting soil system was analysed by the substructure
method developed by Przemieniecki (1968). The supporting soil was modelled by both
the Winkler model and the linear elastic model. Results have revealed that there were
significant differences in the behaviour of the raft predicted by the use of different soil
models for the supporting soil. Flexibility of the raft has significant effects on the
distribution of column loads and moments. Results have shown that the linear elastic
continuum model provided a more realistic solution to the behaviour of the raft and is
more preferable to use in modelling the supporting soil.
Chapter 2 Literature Review
11

Sharma et al. (1984) used the finite element method for the analysis of rafts of any shape
resting on an elastic half space. The raft was modelled by using eight noded
isoparametric quadrilateral plate bending elements in which shear deformations were
taken into consideration. The global plate bending stiffness matrix was formed by
assembling the stiffness matrices for each element. The distribution of contact pressure in
an element was represented in terms of shape functions and the vertical deflection at the
node due to the contact pressure on an element was obtained by the Boussinesq solution.
2.2.4 Hybrid Approach
Zhang and Small (1991) presented a method for the analysis of soil-raft interaction. This
method employed the finite layer technique to determine the behaviour of the soil and the
finite element technique for the analysis of the raft. The contact pressure between the raft
and the soil was represented by uniform blocks of pressure. The response of the soil due
to the contact pressure was obtained by the Fourier transform technique. This method can
be used for the analysis of rafts of any shape in plan and subjected to uniform,
concentrated or eccentric loads. The elastic soil can be anisotropic or non-homogeneous.

Mandal and Ghosh (1999) presented a coupled finite element and boundary element
approach for the prediction of elastic settlement of a raft on a semi-infinite elastic
continuum. The raft was modelled by isoparametric plate bending finite elements and the
raft-soil interface was modelled by boundary elements. The domain of the boundary was
divided into a number of isoparametric quadrilateral quadratic elements. The raft was
divided into elements corresponding to the boundary elements of the soil and the
response of the soil due to the load was obtained from the Mindlin solution for a point
load.

Rashed (2005) developed a new boundary/domain element method for the analysis of a
raft on elastic foundations. Shear deformable plate bending theory was used to model the
raft, the soil was modelled by continuous springs following the Winkler model and the
Chapter 2 Literature Review
12
raft domain was divided into trapezoidal or general shaped cells. The associate domain
integral was replaced by using an equivalent boundary integral along each cell contour.
2.3 PILE GROUP FOUNDATIONS
2.3.1 Simplified Analytical Method
Randolph and Wroth (1978) developed an approximate closed form solution for the
analysis of single vertically loaded piles. In this approach, the soil was divided into an
upper and a lower layer in which the base of the upper layer corresponded to the level of
the base of the pile. The deformation of the upper layer was due to the load acting along
the pile shaft, whereas the deformation of the lower layer was due to the load acting on
the pile base. For the upper layer, the deformation of the soil around the pile shaft was
modelled as shearing of concentric cylinders (Cooke, 1974). For the lower layer, the base
of the pile was assumed to act as a rigid punch on the surface of the layer, and this layer
was acting as a restraint on the deformation of the upper layer. This approach was then
extended to the analysis of pile groups by the same authors (Randolph and Wroth, 1979)
by incorporating the interaction between loaded piles. The interaction factors for the pile
shaft and base were considered separately. For rigid pile groups, the interaction factors
were computed using an approximate closed-form expression, while for compressible
pile groups, the interaction factors were obtained by an iterative procedure to establish a
relationship which expressed the shaft displacement in terms of the pile head and pile
base displacement. The overall displacement of a pile with the presence of adjacent
loaded piles was obtained by the principle of superposition. This approach was limited to
piles of the same embedded length.

Lee (1993) presented an approach which was modified from expressions employed by
Randolph and Wroth (1978). The expression for approximating the load-settlement
behaviour of a pile was modified for a compressible pile embedded in a soil with the
stiffness increasing linearly with depth. The interaction factors were evaluated directly
from the approximate closed-form analytical model for compressible piles. The
settlement of pile groups could then be obtained by the principle of superposition.
Chapter 2 Literature Review
13

Guo and Randolph (1996 and 1999) developed a program GASGROUP which employed
an exact closed form solution for predicting the settlement of pile groups in non-
homogeneous soil by using a load transfer approach. The stiffness of the soil was
assumed to increase with some power of depth (Booker et al., 1985). The pile-soil
interaction was represented by a series of independent springs along the pile shaft and at
the pile base. The load transfer factors take into account the pile slenderness ratio, the soil
non-homogeneity factor and the Poissons ratio. The interaction factor between the two
identical piles was computed by modifying the load transfer factor in the closed form
solution for a single pile to account for the presence of a neighbouring pile. The
settlement of an individual pile in the group was then computed by the superposition of
the interaction factors.

Lee and Xiao (2001) presented an approach for non-linear analysis of pile groups in
multi-layered soils. The approach employed the solution presented by Randolph and
Wroth (1979) to simulate the interaction between two piles and a hyperbolic function as
the load transfer function to model the non-linear behaviour between the shear stress and
displacement of the pile shaft along the pile-soil interface. The non-linear displacement
was approximated as displacement discontinuity to model the non-linear local shear
displacement developed at the disturbed soil around the pile shaft. The interaction effect
between two identical piles was assumed to be linearly elastic and the shaft and base
interaction for individual piles in the group was considered separately. By the principle of
superposition, the load transfer function for the pile group was obtained which accounts
for the non-linear local shear displacement surrounding the pile shaft and the interactive
effect of adjacent piles.

2.3.2 Hybrid Method
A hybrid model was proposed by ONeill et al. (1977) for the analysis of pile groups. The
response of the pile was modelled by the load-transfer method and the pile-soil-pile
interaction was determined based on Mindlins solution. In this approach, by using the
Chapter 2 Literature Review
14
solution for a single pile (i.e. interaction effects were ignored), the response of individual
piles in the group was determined. The computed soil reactions were then used to
determine the additional soil displacements at the nodes of other piles in the group using
Mindlins solution. The additional soil displacements were used to adjust the load
transfer curve to account for the group effects. The solution was obtained following an
iterative procedure. Chow (1986) refined the approach by considering the pile-soil-pile
interaction directly in the hybrid model. The response of the individual pile was
computed by the load transfer curve presented by Randolph and Wroth (1978) and the
non-linear behaviour of the soil was approximated by a hyperbolic function. The
interaction between piles based on Mindlins solution was computed by replacing the
continuous distributed loads at the pile shaft and pile base by point loads at the nodes on
the piles. The non-linear response of an individual pile was approximated by introducing
a discontinuity in the soil. Since the non-linearity is only confined to a narrow zone
adjacent to the pile it was assumed that it would not have any effect on other piles, and
therefore, the interaction between piles would remain elastic.

Leung and Chow (1987) extended the approach by Chow (1986) to analyse laterally
loaded pile groups. For lateral loading, the soil response was modelled by the modulus of
subgrade reaction approach. The pile was divided into a number of discrete beam
elements and the soil was modelled by non-linear springs at the nodal points on the pile
elements. The pile and soil stiffness were determined by the conventional finite element
method. The individual pile response was obtained from a load-transfer curve. The
interaction between piles was obtained by Mindlins solution for a unit horizontal point
load in a homogeneous, isotropic elastic half-space. An averaging procedure was used for
the analysis of pile groups in non-homogeneous soil.
2.3.3 Boundary Element Method
Poulos and Davis (1968) presented the solution for a single pile by using Mindlins
equations. Poulos (1968) further extended the method to the analysis of pile groups by
introducing an interaction factor, , to examine the interaction effect between two
identical piles. The interaction factor was defined as
Chapter 2 Literature Review
15



In the analysis, each pile was divided into a number of cylindrical elements. Each
element was subjected to a uniform load around the periphery of the element and a
uniform circular load at the circular base of the pile as shown in Figure 2.1. The shaft of
the pile was assumed to be perfectly rough while the base was assumed to be perfectly
smooth such that shear stresses were not developed on the base. The vertical
displacement of the soil adjacent to the pile was expressed as
[ ] [ ] [ ] ( )[ ] [ ] [ ] ( )
b b b
p I I p I I
2 1 2 1
+ + + = (2.2)
where [] = displacement of the soil adjacent to the pile
[
1
I] = vertical displacement influence factors for elements due to a uniform load
on each element on pile 1
[
2
I] = vertical displacement influence factors for elements due to a uniform load
on each element on pile 2
[
1
I
b
] = vertical displacement influence factors for the pile base due to a uniform
load on the base of pile 1
[
2
I
b
] = vertical displacement influence factors for the pile base due to a uniform
load on the base of pile 2
[p] = uniform shear load on pile shaft
[p
b
] = uniform vertical stress on pile base

The displacement factors were obtained by integration of the Mindlin equation for
vertical displacement due to a point load within a semi-inifinite soil mass. By considering
the compatibility of the vertical displacement (i.e. unit displacement, = 1), equation (2.2)
can be solved to obtain the distribution of the shear stress along the pile shaft and the
vertical stress on the pile base and subsequently the displacement of the pile can be
determined. For a group of m piles, the displacement of an individual pile in the group
was obtained by superposition

=
=
=
m j
j
kj j k
P
1
1
(2.3)
additional settlement due to an adjacent pile
=
settlement of a pile under its own load
_____________________________________
(2.1)
Chapter 2 Literature Review
16
where
kj
= interaction factors for piles k and j
P
j
= load on pile j

1
= displacement of a single pile under unit load

Poulos (1971) further extended the method to laterally loaded pile groups by introducing
the interaction factor for rotation,

, which is defined as



In the analysis, each pile was assumed to be a vertical strip with a length and breadth of L
and d respectively. The pile was divided into (n + 1) elements and each of the elements
was subjected to a uniform horizontal stress. The length of the elements at the pile top
and base were L/2n, whereas the length of elements along the pile shaft was L/n as shown
in Figure 2.2. The lateral displacements at the soil surface can be obtained from the
method presented by Poulos (1968) by replacing the vertical loads with horizontal loads.
The displacement of the pile can be expressed as

+ =

=
k
m j
k j
j
Hkj j H k
H H
1

(2.5)
where H
j
= load on pile j

H
= unit reference displacement of a single free-head pile under a unit
horizontal load

Hk
= interaction factor for pile k and j

Butterfield and Banerjee (1971a) presented the analysis of compressible piles and pile
groups with a rigid floating pile cap based on the integration of the Mindlin solution. The
pile shaft was divided into a number of equal segments and the base into a number of
rings. For the analysis of a compressible pile, the pile was assumed to be perfectly
bonded to the soil medium. The same method was also used for the analysis of
compressible pile groups with the pile cap in contact with the ground (Butterfield and
Banerjee, 1971b). The cap-soil and pile-soil interfaces were divided into discrete
additional rotation due to adjacent pile

=
rotation of pile under its own load
_____________________________________
(2.4)
Chapter 2 Literature Review
17
elements. The displacements of the soil due to the direct stresses acting on the interfaces
were obtained from Mindlins equations. Results from this method have shown that the
load-displacement behaviour of similar pile groups with floating or contacting caps were
of little difference while the load distribution for groups with contacting caps differed
substantially from those groups with floating caps.

Chin et al. (1990) presented a method based on the analytical layered solutions of Chan et
al. (1974) for the analysis of an axially loaded vertical pile group embedded in a layered
soil with the pile cap not in contact with the ground. The foundation system was
decomposed into (i) a pile group subjected to the pile-soil interaction forces along the pile
shaft and base and the applied loads (ii) the soil continuum subjected to interface forces.
The piles were divided into a number of axial bar elements. The flexibility coefficient
which defined as the displacement at a node due to unit interaction force at another node
was obtained from the analytical two-layered solution of Chan et al. (1974) while for the
nodes on the same pile, the displacement was obtained from the Gaussian quadrature
method by applying a uniformly distributed force over the element. Slip at the pile-soil
interface was allowed by limiting the pile-soil interaction forces as described by Chow
(1986). Pile groups with piles of different radii, lengths and moduli can be analysed by
this approach.

Lee and Poulos (1990) developed an approach for the analysis of pile groups in non-
homogeneous soil. This method involved the development of soil models which account
for the soil modulus of all soil layers and the horizontal non-homogeneity of the soil due
to soil disturbance caused by pile installation. The approach was a modification of the
methods by Poulos (1979) and Yamashita et al. (1987). Poulos (1979) used some
equivalent soil moduli computed from moduli of the influencing and influenced elements.
Yamashita et al. (1987) modified the method by computing an equivalent soil modulus
from the weighted averages of the soil modulus at every layer.

Xu and Poulos (2000) developed a fully coupled load-deformation computer program
GEPAN for the analysis of single piles and pile groups subjected to three-dimensional
Chapter 2 Literature Review
18
loadings and ground movements. The analysis was based on the principles of the three-
dimensional boundary element method and incorporated the effects of defective piles,
soil movements and on/off pile loadings. The global matrix for the governing equation
was derived using the concept of hierarchical structures and a basic influence factor
matrix. The piles were assumed to be circular in cross-section and each pile was divided
into a series of cylindrical elements on the shafts and ring elements on the bases and
discontinuities. The cylindrical elements were then divided into several sub-elements.
The soil-pile interface was modelled by soil elements and pile elements which were
meshed in partly cylindrical or annular surfaces. In the analysis, the cylindrical or annular
boundary elements were transformed into rectangular elements by mathematical
transforms. The rectangular elements were then divided into a number of smaller
rectangles and interaction between the elements was obtained from the integration of
Mindlins equation. The program GEPAN can be applied to a variety of pile problems
such as (i) the off-line effects of piles which is horizontal pile head movement due to
vertical load on the pile group, (ii) interaction factors between piles and pile groups, (iii)
pile groups containing defective piles, (iv) non-linear and elastoplastic analysis. Results
for the off-line effect of piles have shown that for a highly compressible and closely
spaced pile group, loading the piles axially would cause significant horizontal pile
movements.

Basile (2003) extended the approach proposed by Butterfield and Banerjee (1971a) and
developed a computer program PGROUPN which was based on a complete non-linear
boundary element formulation. The program PGROUP accounts for the interactions
between piles, group stiffening effects, load-deformation coupling and non-linearity of
the soil. The pile-soil interface was discretized into a number of cylindrical elements and
the base was represented by a circular element. A substructuring technique was employed
such that the piles and the surrounding soil were considered separately. For the soil
domain, the flexibility matrix was obtained from the integration of the Mindlin functions.
The piles were modelled by simple beam-columns and Bernoulli-Euler beam theory was
applied to obtain the flexibility matrix. Slip was allowed in the computation by limiting
Chapter 2 Literature Review
19
the stresses at the pile-soil interfaces and the pile base. The non-linear soil behaviour was
approximated by a hyperbolic stress-strain relationship.
2.3.4 Finite Element Method
Ottaviani (1975) used the three-dimensional finite element method for the analysis of
vertically loaded pile groups with or without pile caps. Due to the complexity of the
single element stiffness computation and large number of elements, the piles and the soil
were assumed as weightless linearly elastic homogeneous media for the examination of
the load transfer mechanism. It was found that the presence of a cap would cause a non-
uniform distribution of load among the piles of the group. If the cap is in contact with the
soil surface, reduction of the shear stress in the soil around the upper portion of the pile
was found.

Chow (1987a) presented a method based on elasticity theory for the analysis of axially
and laterally loaded pile groups embedded in isotropic non-homogeneous soils. The axial
and lateral group response was assumed to be uncoupled. The foundation system was
decomposed into two systems (i) a pile group subjected to external loads and pile-soil
interaction forces, and (ii) a layered soil continuum subjected to pile-soil interaction
forces. The piles were discretized into discrete elements and the soil was discretized into
constant elements at the pile-soil interface with uniformly distributed vertical or lateral
forces acting over each element. The pile group and the soil continuum were analysed
separately. The load deformation relationship of the soil was determined using the
flexibility approach in which the soil flexibility coefficients were evaluated using the
finite element method with a Fourier series. By applying equilibrium of the pile-soil
interaction forces and the compatibility of the pile and soil displacements, the load
deformation relationship of the pile was determined and expressed as
([K
p
] + [K
s
]){w
p
} = {Q} (2.6)
where [K
p
] = stiffness matrix of piles
[K
s
] = stiffness matrix of the soil obtained by inverting the soil flexibility matrix,
[F
s
], i.e [K
s
] = [F
s
]
-1

{w
p
} = vector of deformations at the pile nodes
Chapter 2 Literature Review
20
{Q} = vector of external applied loads

The reinforcing effect of all piles in the group was considered in the formulation. Non-
homogeneity of the soil was taken into account by incorporating a continuously varying
soil stiffness into the numerical integration process during the formulation of the element
stiffness matrices. The soil non-linearity was assumed to be elasto-plastic and was
analysed by limiting the shear forces or lateral soil pressures to some values and the
excess soil forces at the nodes were then redistributed to other nodes. Chow (1989)
extended the approach to analyse pile groups in cross-anisotropic soils. This was done by
replacing the constitutive isotropic soil model with a cross-anisotropic soil model. Results
have shown that the effect of soil anisotropy on small pile groups embedded in
homogeneous soils was small, however, the effect on large pile groups in non-
homogeneous soils was significant.
2.3.5 Infinite Layer Method
Guo et al. (1987) presented an inifinite layer method for the analysis of piles. This
method was based on the finite layer concept presented by Cheung (1976). The analysis
was carried out in the cylindrical co-ordinate system as shown in Figure 2.3. The layered
soil mass was divided into a number of horizontal layers, each layer was represented by
an infinite layer element. The pile was divided into a number of elements along its length
and was treated as a solid bar. Interaction between the soil and pile was defined by the
compatibility conditions along the pile-soil interface. Cheung et al. (1988) extended the
method to the analysis of pile groups with caps by incorporating an iteration procedure to
determine the interaction between two identical piles. The cap was discretized into a
number of eight noded isoparametric bending elements and analysed by Mindlin thick
plate bending theory. The connection between the cap and piles was assumed to be a
sliding ball joint such that only the vertical forces were transmitted from the cap to the
pile heads.
Chapter 2 Literature Review
21
2.3.6 Finite Layer Method
The finite layer method developed by Small and Booker (1984 and 1986) was first
introduced into the analysis of axially loaded piles embedded in isotropic and cross-
anisotropic layered soils by Lee and Small (1991a). This method is similar in principle to
that of the infinite layer method of Guo et al. (1987). A single pile embedded in a layered
soil was treated as two components: (i) the single isolated pile, and (ii) the layered soil.
The pile was divided into a series of one-dimensional two noded elastic cylindrical solid
elements. The forces acting on the nodes of the elements were treated as point loads and
were evaluated from the uniformly distributed interaction forces acting on the
circumferential area of each pile shaft element and for the base node from the uniform
pressure acting over the area of the pile base. The layered soil component was subjected
to the interaction forces along the pile-soil interface which were represented by a series of
ring loads on the soil nodes and a uniform load on the base. The displacements at all soil
nodes, which form the soil flexibility matrix, were determined by applying a unit load at
each soil node in turn. The soil stiffness matrix was obtained by inverting the flexibility
matrix. Combining the pile and soil stiffness formed the stiffness matrix for the system
and displacements of the pile group were then obtained by solving a set of stiffness
equations. Lee and Small (1991b) modified this approach for the analysis of laterally
loaded piles by replacing the vertical loads acting along the pile shaft and base with
horizontal loads.

Southcott and Small (1996) extended the approach of Lee and Small (1991a) to the
analysis of vertically loaded pile groups. Instead of using the flexibility approach to
generate the flexibility matrix for the soil, the stiffness method was used so that loads
could be applied at the layer interfaces.

Zhang and Small (2000) proposed two methods based on the finite layer theory to analyse
axially and laterally loaded pile groups embedded in homogeneous and non-
homogeneous soils. The principle of the methods is similar to that used by Lee and Small
(1991a and b). The finite layer theory was employed for the layered soil and simple beam
theory for the piles. The piles were divided into a series of finite elements and the soil
Chapter 2 Literature Review
22
was divided into corresponding layers. Interaction and stiffness methods were developed
to generate the influence matrices for the soil and the pile group. In the interaction
method, each pair of piles in the pile group was considered in turn to compute the soil
influence and pile influence matrices. The soil influence matrix was formed by applying
unit ring loads (to each node along the shaft in turn) or a circular load (to the base node)
to compute the displacement at each node of the layered soil. The pile influence matrix
was formed by pinning the top of the pile to stop rigid body rotations and translations of
the pile and a similar method was used to form the soil influence matrix. For the stiffness
method, the soil stiffness matrix was formed by first generating the soil influence matrix.
The influence matrix was formed by applying a unit ring load or circular load (base node)
to each node on all piles in turn to obtain the displacement of the soil. The soil stiffness
matrix was obtained by inverting the soil influence matrix. The pile stiffness matrix was
generated by assembling the stiffness matrix of all the elements of the piles. Comparison
of results between the two methods has shown that the stiffness method is the most
accurate method, however, it is not practicable for large pile groups as it requires a large
amount of computer memory. The shortcoming of the interaction method is that the
formation of the pile influence matrix only considers the interaction between a pair of
piles and ignores the effects of other piles in the pile group. Therefore, the stiffness
method is more suitable for small pile groups with any pile spacing while the interaction
method is suitable for large pile groups with large pile spacing. Zhang and Small (2000)
further extended the method to include the pile cap in the analysis. The analysis was
separated into three parts: the cap, the piles and the layered soil. The cap was assumed to
be a thin elastic plate and analysed by the finite element method. The element division of
the cap was such that the pile head fitted within one element of the cap. In the analysis of
the cap, the cap has to be restrained from rigid body rotations and translations by pinning
two corner nodes. The influence matrix of the pinned cap was generated by applying a
uniform horizontal or vertical load to each of the cap elements in turn to obtain the
central displacements of all cap elements. The influence matrices of the piles and the
layered soil were obtained from the iteration method presented by Zhang and Small
(1999). By considering the equilibrium of interaction forces and compatibility of
Chapter 2 Literature Review
23
displacements between the pile heads and the cap, the contact stresses can then be
obtained and displacement of the raft can then be computed by solving the equations.

2.4 PILED RAFT FOUNDATIONS
2.4.1 Approximation Method
One approach that treated the raft as a thin plate, the piles as springs and the soil as an
elastic continuum, was used by Hongladaromp, Chen and Lee (1973) in which the
interaction effects between the piles were ignored. Poulos (1994) developed a program
GARP (Geotechnical Analysis of Raft with Piles) which employed a finite difference
method for the raft with the consideration of the interaction effects between the piles and
raft. Allowances were made for the piles to reach their ultimate capacities and local
bearing failure of the raft.

Russo (1998) employed a similar method where the piles and soil were modelled by
linear or non-linear interacting springs. The soil displacements were calculated using
Boussinesqs solution thus yielding a closed form solution. The non-linear behaviour of
the piles was modelled by the assumption of a hyperbolic load-settlement curve for a
single pile. This method has the limitation of only allowing for pure vertical interaction
between the raft, piles and soil.

Randolph (1983) presented a method to compute the interaction between a single pile and
a circular raft. A flexibility matrix method was then used to calculate the overall stiffness
of the piled raft foundation by combining the individual stiffness of a single pile-raft unit.

Clancy and Randolph (1993) employed a hybrid method which combined finite elements
and analytical solutions. The raft was modelled by two-dimensional thin plate finite
elements, the piles were modelled by one-dimensional rod finite elements and the soil
response was calculated by using an analytical solution. The pile was attached to a raft
element at a common node, such that the vertical freedoms are common at the connected
Chapter 2 Literature Review
24
nodes. Mindlins solution was used to compute the interaction between the components.
Effects of the pile and raft stiffnesses on displacements and bending moments of the
foundation were examined and it was demonstrated that the differential displacements
and bending moments were dependent on the raft-soil stiffness ratio which was
introduced by Hain and Lee (1978). The load sharing and the average displacement of the
raft were dependent on the pile-soil stiffness ratio. This method took into account the
non-linearity of pile behaviour and slip was allowed to occur at the pile-soil interface.
However, this method is limited to homogeneous soil conditions.

Kitiyodom and Matsumoto (2003) presented a similar approach to Hain and Lee (1978),
but the piles were modelled by elastic beams and the interactions between structural
members were approximated by Mindlins solutions. The foundations can be subjected to
both axial and lateral loads and embedded in non-homogeneous soil. This approach
incorporated both the vertical and lateral resistance of the piles and the base of the raft in
the analysis.
2.4.2 Boundary Element Method
In this method, discretization is only required on the boundary of the system under
consideration. This technique requires the transformation of the governing partial
differential equation into an integral equation. As only the boundaries have to be
discretized, the number of sets of equations to be solved is generally smaller than the
finite element or finite difference methods. Solutions such as stresses and displacements
can be obtained directly by solving the system of equations. Since only the boundaries
are discretized, interpolation errors are confined to the boundaries. As this method
provides a direct and accurate solution for the analysis, is fast, and requires a moderate
amount of computer storage space, it can be used for the analysis of large pile groups.

Butterfield and Banerjee (1971b) employed the boundary element method to study the
behaviour of a pile group embedded in an ideal elastic half space with a perfectly rigid
cap not in contact with the ground. Soil-structure interaction was taken into account in
Chapter 2 Literature Review
25
the analysis. Mindlins solution was used to describe the soil response and the interaction
effects.

Brown and Wiesner (1975) used the boundary element method to analyse a strip footing
supported by equally spaced identical piles embedded in an isotropic homogenous elastic
half space. In this method, the raft and piles were divided into a number of zones in
which interface forces or pressures acted on the corresponding zones. Application of
Mindlins solution was used to determine the interaction relationships due to the interface
forces.

Kuwabara (1989) described a boundary element analysis based on elastic theory to
examine the behaviour of a piled raft foundation in a homogeneous elastic soil mass. In
the analysis, the raft was assumed to be rigid but compressibility of the piles was
considered. The raft was discretized into a series of rectangular elements and the pile was
discretized into a series of shaft and base elements. Poulos (1993) extended the method to
incorporate the effect of free-field soil movement, load cutoffs for the pile-soil and raft-
soil interfaces to examine the interaction mechanism between the piled raft and a soil
subjected to externally imposed vertical movement. The analysis is implemented via a
computer program PRAWN (Piled Raft With Negative Friction)

Mendona and de Paiva (2000) presented a boundary element method for the analysis of
piled rafts in which full interaction between the raft, piles and the soil is considered.
Unlike the other approaches, discretization of the foundation system was not required in
this approach. The soil was represented by a Mindlin elastic linear homogeneous half
space. The raft was assumed to be a thin plate and was represented by integral equations.
The pile was represented by a single element and the shear stresses along it were
approximated by a second-degree polynomial. The interaction between the raft and soil
was analysed by dividing the interface into triangular elements and the subgrade reaction
was assumed to vary linearly across each element.
Chapter 2 Literature Review
26
2.4.3 Finite Element Method
The finite element method is one of the most powerful tools for the analysis of piled rafts.
It requires the discretation of both the structural foundation system and the soil. In order
to reduce the computational effort, problems are sometimes simplified to an
axisymmetric problem or a plane-strain problem.

An early example of the analysis of a piled raft (the Hyde Park Barracks) was given by
Hooper (1973), in which an axisymmetric model with eight noded isoparametric elements
was used. In the analysis, approximation of the equivalent stiffness of the pile group was
made such that each concentric row of piles was modelled by a continuous annulus with
an overall stiffness that was equivalent to the sum of the stiffnesses of the individual piles.
The soil was assumed to be a linear elastic isotropic material with the modulus increasing
linearly with depth. In order to incorporate the additional stiffening effect of the
superstructure into the analysis, an equivalent raft thickness which had the same bending
stiffness as the combined raft and the superstructure was introduced. However, Hoopers
results have shown that the contribution of the stiffening effect of the superstructure on
the behaviour of the piled raft was relatively small in the case of the Hyde Park Barracks,
although this may not be true in all cases. Ottaviani (1975) applied this method to the
analysis of a rigid raft resting on a pile group embedded in a homogeneous medium. The
piles and the soil were assumed to be a weightless linearly elastic homogeneous media.

Chow and Teh (1991) presented a numerical method to examine the behaviour of a rigid
piled raft embedded in a non-homogeneous soil. The raft was discretized into square sub-
elements. The base of the raft was assumed to be perfectly smooth and the interface of
the raft and the soil medium was approximated by square subdivisions (Chow 1987a).
The soil was assumed to be a linearly elastic, isotropic material and the Youngs modulus
assumed to increase linearly with depth. The piles were assumed to have a circular cross-
section and were discretized into two noded elements at the pile-soil interface (Chow,
1987b). Interactions between the piles, the raft and the soil were taken into account and
the vertical deformation of the soil was determined by the principle of superposition in
which equilibrium of the raft-pile-soil system was considered.
Chapter 2 Literature Review
27

Liu and Novak (1991) employed the finite element method to examine the behaviour of a
raft supported by a single pile at the centre. In the analysis, the cap was assumed to be
circular and to contact the soil perfectly. Nine noded isoparametric elements were used to
model the cap, the pile and the near-field soil medium. The cap, the pile and the
surrounding soil medium were modelled by finite elements. Mapped infinite elements
proposed by Danjanic and Owen (1984) are used to model the far-field soil medium to
stimulate an unbounded domain. The pile was assumed to be linearly elastic and the soil
was assumed to be either elastic or elastic perfectly plastic. A weak zone around the pile
with lower strength and modulus was introduced to account for the slip at the pile-soil
interface. This method allows for the analysis of a piled raft where the raft is embedded
as well as in contact with the ground. The use of inifinite elements to model the far-field
soil medium reduced the number of elements required significantly which made the
analysis more efficient than the conventional finite element method.

Wiesner (1991) presented a method for the analysis of a circular piled raft that was
constructed in Cairns. The raft was treated as a thin elastic plate and modelled by
rectangular plate bending finite elements. The reaction forces acting on the raft-soil
interfaces were assumed to be rectangular blocks of uniform vertical stresses. The piles
were represented by elastic cylinders and the soil was assumed to be linearly elastic. The
reaction forces on the pile-soil interfaces were treated as uniform vertical shear stresses
along the pile shaft and as a uniform vertical stress at the pile base. To take interaction
into account, the reciprocal theorem was applied to the pile, and influence factors were
calculated based on elastic theory. Non-linear behaviour of piles and soil were considered
by comparing the reaction forces with the limiting values during the iterative process.

Katzenbach and Reul (1997) described a structural model which employed the finite
element method for the geometrical modelling of the continuum, an elastoplastic
constitutive model to describe the soil behaviour and a step-by-step analysis for
numerical simulation. The piles were modelled by 3-dimensional isoparametric finite
elements and the raft was modelled by shell elements. A realistic stress-strain behaviour
Chapter 2 Literature Review
28
of the soil was simulated by a constitutive model for the soil which consisted of two main
yield-surface segments: a pressure dependent perfectly plastic shear failure surface and a
compression cap yield surface. Katzenbach et al. (2000) used the same structural model
to carry out parametric studies on the behavioiur of piled rafts in Frankfurt Clay. Reul
(1998) refined the structural model by the use of infinite elements at the borders to model
the soil as an elastic halfspace.

Smith and Wang (1998) proposed the use of iterative techniques with the finite element
method to examine the behaviour of a non-uniformly loaded piled raft. Unlike the
traditional techniques, this approach can reduce the size of the stiffness matrices, thus
reducing the number of equations to be solved in the system.
Prakoso and Kulhawy (2001) analysed piled raft foundations by the use of linear elastic
and non-linear plane strain finite element models which involved the analysis of a three-
dimensional piled raft as a two-dimensional strip piled raft. The analysis was performed
by a geotechnical finite element code PLAXIS (Vermeer and Brinkgreve, 1995) and six
noded triangular elements were used to model the piled raft and the soil. As the rows of
piles were simplified into strips, the in-plane row of piles has to be simplified into a plane
strain pile with an equivalent pile Youngs modulus E
eq
in terms of the number of piles in
the row considered, the dimensions of the pile and the dimensions of the raft:
p r
p p rowi p
eq
D L
E A n
E

=
where
n
p-row i
= number of piles in row i
A
p
= area of pile cross section
E
p
= Youngs modulus of the pile
L
r
= length of the raft
D
p
= pile diameter

Further simplification is necessary for the elastic-plastic modelling such that the
compression of the equivalent plane strain piles and the in-plane row of piles should be
the same. Interface elements were used to model the resistance along the pile shaft.
Chapter 2 Literature Review
29

Reul and Randolph (2003) presented a three-dimensional elasto-plastic finite element
method for the analyses of piled raft foundations in overconsolidated clay Frankfurt
clay. The analysis was implemented by the program ABAQUS. In the finite element
method, the soil was modelled by hexahedron elements and the piles were modelled by
triangular prism elements. The circular piles were modelled by square piles with the same
shaft circumference. The interfaces between the raft and soil, and between the pile and
soil were modelled by thin solid continuum elements and were assumed to be perfectly
rough. The soil was modelled by a cap model to simulate the non-linear behaviour. The
cap model consisted of three yield surface segments: the pressure-dependent, perfectly
plastic shear failure surface, the compression cap yield surface, and the transition yield
surface. The same model was used by these authors to perform a parametric study of a
piled raft with different pile configurations and subjected to non-uniform vertical loading
(Reul and Randolph, 2004).

Maharaj and Gandhi (2004) proposed a non-linear finite element method for the analysis
of a piled raft subjected to a uniformly distributed load. This method combined an
incremental iterative procedure with a Newton-Raphson method to solve the non-linear
equations involved in a plasticity analysis. The raft, pile and soil were discretized into
eight noded brick elements. The raft and piles were assumed to be linearly elastic and the
non-linear behaviour of the soil was modelled by the Drucker-Prager yield criterion
(1952).
2.4.4 Combined Boundary Element and Finite Element Method
Hain and Lee (1978) developed an analysis to examine the behaviour of a flexible raft
supported by a group of compressible friction piles embedded in an elastic homogeneous
or non-homogeneous material with the consideration of the ultimate capacity of the piles.
The analysis combined the finite element method for the analysis of the raft and the
boundary element method for the analysis of the piles and soil. The raft was treated as a
thin elastic plate and the pile group - supporting soil system was modelled by the use of
the Mindlin equation. However, the connection between the raft and the pile was
Chapter 2 Literature Review
30
assumed to be a sliding ball joint which implied that no moments or lateral forces were
transferred between the raft and pile heads. In the analysis, they suggested that the
behaviour of the piled raft would depend on the relative flexibility of the raft and the
relative stiffness of the pile to the soil. Four different interactions between the piles, raft
and soil were introduced and thoroughly considered in the analysis. In addition, a load
cut-off procedure was introduced to account for the development of the ultimate load
capacities of the piles.

Kakurai et al. (1987) used a similar method with several modifications to examine the
settlement behaviour of a piled raft foundation on soft ground. The raft was modelled by
beam and bending elements. The piles and soil were modelled as vertical springs
supporting the raft at selected nodal points. The reaction forces on the raft were applied at
the nodes of the finite element mesh used for analysing the raft. The area surrounding the
supporting soil node was converted into a circle with an equivalent area as shown in
Figure 2.4 and a circular uniform load was applied to the surface of the soil for the
analysis. The theory of elasticity was used to analyse the interactions among the surface
load, soil and piles.

Mandolini and Viggiani (1997) presented an analysis to predict the settlement of piled
raft foundations. The method is capable of taking into account the soil-structure
interaction and non-linear behaviour at the pile-soil interface. The piles were analysed by
the boundary element method and the behaviour of a pile group embedded in an elastic
continuum was then analysed based on the use of interaction factors. The raft was
analysed by the use of the finite element method and the interaction between the piles,
raft and soil was represented by a linear elastic model. To stimulate the non-linear
behaviour, a stepwise linear incremental procedure was used and a hyperbolic load-
settlement relationship for a single pile was assumed.

A mixed technique based on the finite element method and boundary element method
was developed by Franke et al. (2000) to model the three-dimensional nature of a piled
raft. Plate-bending finite elements were used to model the stiffness of the superstructure,
Chapter 2 Literature Review
31
whereas the piles and soil were modelled by non-linear elastic springs and were attached
to each node of the finite element mesh of the raft. Boundary elements at the raft-soil and
pile-soil interfaces were used to model the contact pressures between the raft and soil and
between the piles and soil respectively. The non-linear pile response was described by (i)
a hyperbolic shear stress-shear strain relationship for the soil adjacent to the pile shaft; (ii)
a boundary element solution to obtain the skin friction distribution; (iii) a hyperbolic
load-deformation relationship for the soil near the pile base. In the analysis, the effects of
construction sequence and pile installation on the non-linear pile response were also
taken into account.

A coupled boundary element and finite element formulation was described by Mendona
and de Paiva (2003) in which full interaction of the structure has been incorporated into
the analysis. The pile was represented by a single element and the stresses along the shaft
were approximated by a quadratic function and the pressure at the pile tip was assumed to
be constant over the cross-section of the base. The soil was represented by a Mindlin
linear-elastic homogeneous half space. The raft was analysed by the finite element
method through the use of a combination of flat triangular elements based on discrete
Kirchhoff Theory and the hybrid stress model. The plate-soil interface was modelled by
triangular elements to account for the plate-soil interaction. The analysis dealt with a raft
subjected to vertical load only.

Sinha (1997) described an analysis for piled raft foundations in expansive soil using the
finite element method to model the raft and the boundary element method to model the
piles. The raft was analysed as a plate resting on an elastic soil medium and was
discretised into four noded rectangular elements. The pile was discretised into cylindrical
elements and analysed by the boundary element method, and the soil was assumed to be a
homogeneous elastic soil mass. Non-linear behaviour which included lift off of the raft
from the soil and local soil yield underneath the raft, slip at the soil-pile interface and
yielding of the soil under the pile base were incorporated into the analysis. The effects of
free field soil movement have been considered in the analysis in which the ground
movements due to the process of swelling and shrinking of the soil were considered.
Chapter 2 Literature Review
32
2.4.5 Combined Finite Layer and Finite Element Method
Ta and Small (1996) presented an approach based on the finite layer technique developed
by Small and Booker (1984, 1986) to compute the behaviour of piled rafts subjected to
vertical loads in layered soils. The soil was divided into a series of horizontal layers; the
raft was treated as a thin elastic plate and the piles were divided into rod elements
corresponding to the soil layers. The soil was analysed by the finite layer method and the
raft and piles were analysed by the finite element method. Ta (1996) proposed two
approximation methods (Type I and II) which may be used to compute interactions
between the piles or piles and raft more efficiently. Displacement at any point on the soil
surface can be approximated by a closed form polynomial equation. A type I
approximation is limited to piled rafts with square raft elements of equal size and
identical piles. A circular uniform load can then be used to represent the square block of
contact pressure under the raft element. For the Type II approximation, the elements of
the raft can be of different sizes, however, the interaction method which was used to
determine the interaction does not account for the group effect.

Zhang (2000) presented a method for the analysis of a piled raft subjected to both loads
and moments in any direction. The analysis is implemented by the program APRAF. The
raft was assumed to be a thin elastic plate and analysed by a finite element method
developed by Bogner et al. (1965) that is based on a four noded element. The pile was
modelled by beam elements and analysed by a simple finite element method. The layered
soil was assumed to be elastic and modelled by the finite layer technique. The interaction
between the pile and soil which forms the pile-soil flexibility matrix was computed by the
finite layer method developed by Small and Booker (1986).
2.4.6 Variational Approach
This approach makes use of the principle of minimum potential energy to simulate the
response of the foundation system. Discretisations are only required at the interface
between the raft and soil. Shen et al. (1999) developed a variational approach for the
analysis of pile groups with a rigid pile cap not in contact with the ground. An extension
of the method was developed by the same authors for the analysis of the cap in contact
Chapter 2 Literature Review
33
with the ground (Shen et al., 2000). Interaction between the cap, piles and soil was
simulated by the use of Mindlins solution and finite series were used to represent the
deformation and shear stresses along the piles. The soil was modelled as an isotropic
elastic half space and the piles were assumed to be linearly elastic. The analysis involved
the isolation of the pile group-soil system from the pile-cap foundation and the interface
between the pile cap and soil was discretized into a number of elements as shown in
Figure 2.5. The pile group-pile cap system was analysed by the minimum potential
energy principle. The same approach was used by Shen and Teh (2002) which has
incorporated lateral loading into the analysis. The piles were modelled as flexible beams
and the non-linear behaviour of the pile head was approximated in the elastic analysis by
a reduction in the soil modulus near the ground surface.

An extension of the above method for vertically loaded pile groups was made by Chow et
al. (2001) for the analysis of piled raft foundations. The raft was assumed to be a thin
elastic plate. The deformations of the raft and pile group were each represented by a finite
series. The piled raft and pile group were analysed by the use of the principle of
minimum potential energy. The load-settlement relationship at the interface between the
pile heads and soil were incorporated into the raft analysis, and the behaviour of the pile
and soil were assumed to be elastic-perfectly plastic. In order to account for the ultimate
bearing capacity of the pile and soil, an initial stress technique was used to limit the pile
loads and soil reaction pressures to the ultimate values.

Liang and Chen (2004) presented a modified variational approach based on the above
method. In their modified approach, the flexible raft was isolated from the piled raft
foundation and the deformation of the raft was represented by an analytical function. The
base of the raft was assumed to be smooth and the connection between the raft was
assumed to be a sliding ball joint such that only the vertical forces could be transmitted
from the raft to the pile heads. Interactions between the raft, piles and soil were described
by an approximate closed-form analytical method.

Chapter 2 Literature Review
34
2.5 CONCLUSIONS
This chapter has outlined the existing methods used for the analysis of rafts, pile groups
and piled raft foundations. The following conclusions may be drawn:

(1) An elastic continuum model is preferable to a Winkler spring model in modeling the
supporting soil for the foundation as it provides a realistic and better representation of
the physical behaviour of the soil.
(2) Except for the finite element and finite layer methods, the analysis of foundations in
non-homogeneous soil has been carried out either by averaging the soil modulus, by
assuming the modulus increases linearly with depth (Gibson soil) or some power of
depth.
(3) Among the rigorous methods, the boundary element method is more practicable for
the analysis of large pile groups as it is easy to compute the interaction factors for the
soil from the analytic expressions.
(4) Mindlins solution has been incorporated into different kinds of approaches for the
determination of soil displacements and interaction factors.
(5) For most of the methods, the non-linear analysis was carried out by an iterative
procedure such that the forces acting on the pile-soil interface were limited to an
assumed value and forces exceeding the limited value were then redistributed.
(6) Finite layer and finite element methods can be used for the analysis of foundations
embedded in non-homogeneous soil without the need for approximation.
(7) For pile groups or piled raft foundations subjected to both axial and lateral loads, the
axial and lateral responses of the foundation are often assumed to be uncoupled.
Chapter 2 Literature Review
35
Figure 2.1 Stresses on piles and soil for a 2 pile group subjected to axial load

s
P
P
p
j
i

p
b
Pile 1

Pile 2

p
b
(a) Stresses on piles

p
j
p
b p
b
(b) Stresses on soil

H, M
d
S
d
H, M

M
H
M
H
L/n

L/n

1
2
i
j
n
n+1
p
1
1
2
i
j
n
n+1
p
j
p
n+1
3

p
1
p
j
p
n+1
Pile 1 Pile 2
L
(a) Stresses acting on piles
p
1
p
j
p
n+1
1
2
i
j
n
n+1
3

1
2
i
j
n
n+1
3

p
1
p
j
p
n+1
(b) Stresses acting on soil
Figure 2.2 Stresses acting on piles and soil for a 2 pile group subject to lateral loads
Chapter 2 Literature Review
36
Nodal points of supporting soil
Nodal points of pile
Equivalent loaded area
Beam elements for raft
Figure 2.4 Model of the soil with circles of equivalent area
(a) Model of the piles and raft (b) Model of the soil
Figure 2.3 Infinite layer in cylindrical co-ordinate system and bar element for pile
(b) Bar element for pile
(a) Infinite layer in cylindrical system
t
i
r
0
u

u
r

0

r

z, u
z
= 0

u
rp
u
zp
z

r
0
u

t
i
Chapter 2 Literature Review
37



















q
s
P
j P
i
p
p
P
pj
P
pi

i
j

j
x
z
(a) Model of the pile cap, pile and soil system
x
y
Soil element
Pile
(b) Discretisation of the interface between the soil and pile cap
Figure 2.5 Model of the piled raft foundation
Chapter 3 Finite Layer Analysis of Layered Soil
38
CHAPTER 3
FINITE LAYER ANALYSIS OF LAYERED
SOIL

3.1 INTRODUCTION
In the past few decades, numerical techniques have been developed for the analysis of
rafts, piles and piled raft foundations in layered soils. The most commonly used
approaches to model the soil mass are finite element methods (Ottaviani, 1975; Chow,
1986; Katzenbach and Reul, 1997; Reul and Randolph, 2003 and 2004)) and boundary
element methods based on Mindlins equations (Butterfield and Banerjee, 1971a; Brown
and Wiesner, 1975; Hain and Lee, 1978; Poulos and Davis, 1980). However, there are
some drawbacks to these approaches. The finite element method requires the creation of a
mesh with a large number of elements and large computer storage, to model the infinite
soil mass while the application of the boundary element method to layered soil is only an
approximation.

Due to the fact that natural deposits are often horizontally layered and earthworks such as
landfills consist of horizontal layers of materials with different properties, the analysis of
such horizontally layered soil deposits can be solved in a similar way to that of a one-
dimensional problem. For horizontally layered soil, Fourier series can be used to express
the field quantities such as displacements and stresses. The use of a series can then be
used to simplify a two- or three-dimensional problem into a problem involving a single
spatial dimension. Cheung (1976) introduced the finite strip and finite layer methods,
which are based on Fourier series and a stiffness approach, into the analysis of
foundations and rigid pavements (Cheung and Fan, 1979). Small and Booker (1984)
developed a finite layer method, which was based on a flexibility approach, to determine
soil deflections of horizontally layered soil subjected to strip loadings by the use of
Fourier transforms (Small and Booker, 1984). For circular loadings Hankel transforms
Chapter 3 Finite Layer Analysis of Layered Soil
39
were used and for rectangular loadings double Fourier transforms (Small and Booker,
1986) were applied. The use of the flexibility approach to assemble the matrix can
overcome the disadvantage of the stiffness approach in dealing with incompressible
materials that occur in the conventional finite layer method.

3.2 ANALYSIS OF LAYERED SOIL
3.2.1 Circular and Ring Loads
Circular and ring loads can be used to simulate the force acting along pile shafts or at pile
bases in the analysis of piles, pile groups and piled rafts. This type of load is treated as an
axially-symmetric loading and it is convenient to use cylindrical co-ordinates to represent
their axis system, where r represents the radial, represents the tangential and z
represents the vertical directions as shown in Fig. 3.1.

3.2.1.1 Basic Equations for Elastic Materials
Stress-strain Relationship
For cross-anisotropic materials, the stress-strain relationship may be expressed as
= D (3.1)
where = (
r
,

,
z
,
rz
,
z
,
r
)
T
is the vector of stress components
= (
r
,

,
z
,
rz
,
z
,
r
)
T
is the vector of strain components
D is the matrix of elastic constants

=
2 / ) ( 0 0 0 0 0
0 2 / 0 0 0 0
0 0 2 / 0 0 0
0 0 0
0 0 0
0 0 0
b a
f
f
d c c
c a b
c b a
D
The elastic constants, a, b, c and d can be defined in terms of the Youngs modulus and
Poissons ratios as follows (Gerrard, 1967)
Chapter 3 Finite Layer Analysis of Layered Soil
40
) 2 1 )( 1 (
) 1 (
vh hv h h
vh hv h
E
a


+

= (3.2a)

) 2 1 )( 1 (
) (
vh hv h h
vh hv h h
E
b


+
+
= (3.2b)

vh hv h
vh h
E
c

2 1
= (3.2c)

vh hv h
h v
E
d

2 1
) 1 (

= (3.2d)

where E
h
= modulus of elasticity in the horizontal direction
E
v
= modulus of elasticity in the vertical direction

h
= Poissons ratio for the effect of horizontal stress on complementary
horizontal strain

hv
= Poissons ratio for the effect of horizontal stress on complementary
vertical strain

vh
= Poissons ratio for the effect of vertical stress on complementary
horizontal strain

The following relationship between elastic constants also holds for cross-anisotropic
materials

vh
hv
v
h
E
E

= (3.3)
The elastic constant f is a shear modulus and can be denoted as 2G
v
. The constant f is
limited by the condition (Koning, 1957)

) 2 1 (
1
1 )( 1 (
2 / 1
hv vh h h
h
h hv vh hv vh
hv vh
h
E
f

+

+
(3.4)

Chapter 3 Finite Layer Analysis of Layered Soil
41
and f > 0

In addition, the limit conditions for the Poissons ratios are

h
+
hv
1

vh
< 0.5
1 -
h
2
hv

vh
> 0
(1 +
h
)(1 -
h
- 2
hv

vh
) > 0

For an isotropic material
=
h
=
vh
=
hv
E = E
v
= E
h

and so the equations for the elastic constants can be rewritten as
) 2 1 )( 1 (
) 1 (

= = =
E
a d a
iso

) 2 1 )( 1 (

+
= = =
E
b c b
iso

and
f = 2G
v

Equilibrium Equations
By assuming that the loading is independent of for a circular or ring load, the
equilibrium equations may be written as
0 =

r z r
r rz r

(3.5a)
0 = +

r z r
rz z rz

(3.5b)
0
2
= +

r z r
r z r

(3.5c)

Chapter 3 Finite Layer Analysis of Layered Soil
42
Strain-displacement Relationship
The strain-displacement relationship may be expressed as (Poulos and Davis, 1974)
= u (3.6)
where u = (u
r
, u

, u
z
)
T
is the vector of displacement components
= (
r
,

,
z
,
rz
,
z
,
r
)

=
0 ) / 1 / ( 0
0 / 0
/ 0 /
/ 0 0
0 0 / 1
0 0 /
r r
z
r z
z
r
r

3.2.1.2 Hankel Transforms
Hankel transformations are used to simplify the preceding equations (Small and Booker,
1986)

=
0
0
) ( dr r J ru U
z z
(3.7a)

=
0
1
) ( ) , ( ) , ( dr r J u u r U U
r r

(3.7b)
The equivalent inverse transformations are

=
0
0
) ( d r J U u
z z
(3.7a)

=
0
1
) ( ) , ( ) , (

d r J U U u u
r r
(3.7b)
where J
0
(r) and J
1
(r) are Bessel functions of the first kind, and is the Hankel
transform parameter.

Applying a Hankel transform to the strain-displacement equation (3.6) results in the
following expression
{ } Ud r J z r N r J z r N

+ =
0
1 2 0 1
) ( ) , , ( ) ( ) , , ( (3.8)
where U = (U
r
, U

, U
z
)
T

Chapter 3 Finite Layer Analysis of Layered Soil
43

=
0
2
0
0 0
0
0 0 0
0 0
1
0 0
1
) , , ( ,
0 0
0 0 0
0 0 0
0 0
0 0 0
0 0
) , , (
2 1
r
z
z
r
r
z r N z z r N

Substituting equation (3.8) into the stress-strain equation (3.1) to obtain the stresses in
terms of the displacement coefficients U gives
= D f(U
r
, U
z
, U

) (3.9)
Then, applying a transform to the equilibrium equations (3.5a) to (3.5c) we obtain

0
2
=

+
f
z
U
U
z z
U
c aU
r
z
z
r
(3.10a)
0
2
=

z
U
d cU
z
f
z
U
U
z
r
r
z
(3.10b)
0
2 2
) (
2
2
2
=

z
U f
U
b a

(3.10c)
It may be noticed that in the last equation (3.10c), the term U

is uncoupled and the

equation may be solved directly. The first two equations (3.10a) and (3.10b) containing
the terms U
r
and U
z
are coupled equations and must be solved together.

Solution of Coupled Equations
The coupled equations (3.10a) and (3.10b) may now be solved simultaneously. The
following substitution may be made to simplify the equations
z
Uz
c aU M
r

+ = (3.11a)
2
f
z
U
U T
r
z

+ = (3.11b)
Chapter 3 Finite Layer Analysis of Layered Soil
44
z
U
d cU N
z
r

+ = (3.11c)
Equations (3.10a) and (3.10b) become
M(,z)S = 0 (3.12)
where S = (M,N,T)
T

z
z
z M
0
0
) , (
To satisfy the equations of equilibrium (3.12), an auxiliary function may be introduced
+

=
2
2
2
B
z
M (3.13a)
z
T

= (3.13b)
=
2
N (3.13c)
Combining equations (3.11) and (3.13) to obtain the following expressions in terms of the
auxiliary function gives
+

=
2
2
2
B
z
A U
r
(3.14a)

2
2
2
C
z
B
z
U
z
(3.14b)
z
F U
z
U
z
r

(3.14c)
where
A = d/(ad-c
2
), B = c/(ad-c
2
)
C = a/(ad-c
2
), F = 2/f
The exact solution of the auxiliary function is given in Appendix I.

Flexibility Relationship
Consider a layer of material that occupies the region h z h as shown in Figure 3.2 in
which the notation N and T are defined as the normal and shear stress coefficients
Chapter 3 Finite Layer Analysis of Layered Soil
45
respectively. The subscripts p and m indicate the stress coefficients of the lower and
upper surface of the layer.

To obtain the flexibility relationship for the layer, the stress boundary conditions are
firstly applied at the lower and upper surface of the layer.
S
zz
= N
p
,

T
rz
= T
p
for z = + h
S
zz
= N
m
,

T
rz
= T
m
for z = - h
Then, values of displacements on the surfaces of a layer may be found from equation
(3.14a) and (3.14b) in terms of the stress coefficient. The flexibility relationship between
the transformed stresses and displacements for any layer may then be expressed as

i
= F
i
P
i
(3.15)
where

i
= (U
zp
, U
rp
, -U
zm
, -U
rm
)
T

P
i
= (N
p
, T
p
, N
m
, T
m
)
T

F
i
is the flexibility matrix given in Appendix II.

Solution of Uncoupled Equations
To solve the uncoupled equation (3.10c), a Hankel transform is applied to the shear stress
and shear strain in the stress-strain law of equation (3.1). The transformation of shear
stress and shear strain may be expressed as

=
0
1
) (

d r J T
z z
(3.16a)

=
0
1
) (

d
z
U
r J
z
u
z
(3.16b)
The stress-strain relationship
z z
f


2
= may be rewritten as
z
U f
S
z

2
(3.17)
The solution for U

may be obtained from equation (3.10c) as follows

U

= C
1
sinh(kz) + C
2
cosh(kz) (3.18)
where C
1
and C
2
are constants and
Chapter 3 Finite Layer Analysis of Layered Soil
46
f
b a
k

=
By substituting equation (3.18) into equation (3.17), the solution for the coefficient S
z

may be obtained
[ ] ) sinh( ) cosh(
2
2 1
kz C kz C k
f
S
z
+ =

(3.19)
By combining equations (3.18) and (3.19) and applying the boundary conditions at the
upper and lower surface of the layer, the flexibility relationship for the uncoupled terms
may be obtained

m z
p z
m
p
S
S
kh anh kh ech
kh ech kh anh
k
F
U
U

) 2 ( cot ) 2 ( cos
) 2 ( cos ) 2 ( cot
(3.20)
where F = 2/f
The overall flexibility relationship for the layered system (soil having n layers) can be
obtained by summing the flexibility matrix for each layer. We may write the flexibility
relationship of the n layered system as
= F P (3.21)
where

=
=
n
i
i
1
,

=
=
n
i
i
P P
1
,

=
=
n
i
i
F F
1

and apply appropriate boundary conditions to the equations to solve for the interface
forces. The boundary conditions for a circular loading applied to a layered system are
given in Appendix III.
3.2.2 Rectangular Loads
Rectangular loads are often used to simulate the contact stresses under a raft or piled raft.
The following section summarizes the method developed by Small and Booker (1986) for
the analysis of this type of load. Figure 3.3 shows a three-dimensional Cartesian co-
ordinate system to represent the axis system of the load where x, y represent the lateral
directions and z represents the vertical direction.

Chapter 3 Finite Layer Analysis of Layered Soil
47
3.2.2.1 Basic Equations for Elastic Materials
Stress-strain Relationship
The stress-strain relationship may be expressed as
= D (3.22)
where = (
xx
,
yy
,
zz
,
xy
,
yz
,
zx
)
T
is the vector of stress components
= (
xx
,
yy
,
zz
,
xy
,
yz
,
zx
)
T
is the vector of strain components
D is the matrix of elastic constants defined in equation (3.1)

Equilibrium Equations
The equations of equilibrium for stresses at any point in the x, y and z directions may be
expressed as
0 =

y z x
xy
xz xx

(3.23a)
0 =

x z y
xy zy yy

(3.23b)
0 =

y x z
zy
xz zz

(3.23c)

Strain-displacement Relationship
The strain-displacement relationship may be expressed as
= u (3.24)
where u = (u
x
, u
y
, u
z
)
T
is the vector of displacement components, and,

=
x z
y z
x y
z
y
x
0
0
0
0 0
0 0
0 0

Chapter 3 Finite Layer Analysis of Layered Soil
48
3.2.2.2 Fourier Transformations
A double Fourier transformation may be applied to the field quantities to simplify the
above equations. The stress and displacement transforms may be expressed as:

+
= dxdy e iu iu iu U U U
y x i
z y x z y x
) (
2
) , , (
4
1
) , , (

(3.25a)

+
= dxdy e i i i T T T S S S
y x i
zx yz xy zz yy xx zx yz xy zz yy xx
) (
2
) , , , , , (
4
1
) , , , , , (

(3.25b)
where the transformed quantities are all functions of , , z.
Inverse Fourier transforms can be applied to equation (3.25a) to get the actual
displacements


d d e U iU iU u u u
y x i
z y x z y x
) (
) , , ( ) , , (
+

=

(3.26)
and a similar expression may be used for the stress components.
To solve the problem, an auxiliary co-ordinate system , may be introduced in the x - y
plane and this co-ordinate system is considered to make an angle with the x - y co-
ordinate system as shown in Figure 3.3. The transformations of any point from one
Cartesian co-ordinate system into another Cartesian co-ordinate system can be written as
= x cos + y sin (3.27a)
= -x sin + y cos (3.27b)
Multiplying equation (3.27a) by and equation (3.27b) by , and then adding them
together gives the following equation
x + y = ( cos + sin ) + ( cos - sin ) (3.27c)
Assume = cos , = sin
then x + y = which is independent of .

3.2.2.3 Transformed Equations in the , Co-ordinate System
Stress-strain Relationship
The stress-strain relationship in the x y co-ordinate system can be rewritten in the
, and z co-ordinate system as
= D (3.28)
where = (

,
zz
,

,
z
,
z
)
T
Chapter 3 Finite Layer Analysis of Layered Soil
49
= (

,
zz
,

,
z
,
z
)
T

D is defined in equation (3.1)
As the field quantities are independent of the direction, the derivative of the
displacement, stress or strain with respect to will become zero, therefore 0 =

u

and the stress-strain relationship can be written as

z
z
z
z
F
C B
B A

0 0
0
0
(3.29)
where
2
c ad
d
A

= ,
2
c ad
c
B

=
2
c ad
a
C

= ,
f
F
2
=

The equilibrium equations can be written in terms of the new co-ordinate system as
follows
0 =

z
z

(3.30a)
0 =

z
zz
z

(3.30b)
0 =

z
z

(3.30c)
Equations (3.30a) and (3.30b) may be solved in the same manner as in Section 3.2.1.
Equation (3.30c) is uncoupled from the other two equations and may be solved separately.
As the stress, strain and displacement are independent of the direction, the expressions
for the stresses, strain and displacement with reference to the - z plane may be defined
in the following forms

2
0
) , , , , , ( ) , , , , , ( d d e iT iT iT S S S
i T
z z zz
T
z z zz
(3.31)

2
0
) , , , , , ( ) , , , , , ( d d e iG iG iG E E E
i T
z z zz
T
z z zz
(3.32)
Chapter 3 Finite Layer Analysis of Layered Soil
50

2
0
) , , ( ) , , ( d d e U iU iU u u u
i T
z
T
z
(3.33)
Substituting equation (3.31) and (3.32) into equation (3.29) leads to
E = RS (3.34)
where E = (E

, E
zz
, G
z
)
T
, S = (S

, S
zz
, T
z
)
T
R =

F
C B
B A
0 0
0
0

and E

, G
z
, S

, T
z
are obtained from the transformation of axes.
The equilibrium equations become
M(, z)S = 0 (3.35)
where M(, z) =

z
z
0
0

The strain-displacement relationship becomes
E = - N((, z)U (3.36)
where U = (U

, U
z
)
T

Solution of Coupled Equations
To satisfy equations (3.35), an auxiliary function is introduced in the same manner as
was done for the case of circular loadings, i.e.
z
S T
z

= =

(3.37a)
= =
2

zz
S N (3.37b)
+

= =
2
2
2

B
z
S M (3.37c)
Substituting equations (3.37a-c) into equation (3.34) lead to
+

=
2
2
2

B
z
A U (3.38a)

2
2
2
C
z
B
z
U
z
(3.38b)
Chapter 3 Finite Layer Analysis of Layered Soil
51
z
F U
z
U
z

(3.38c)
Equations (3.38a-c) have the same form as equations (3.14a-c) for circular loadings and
can be solved by the same technique. Hence the flexibility relationship for the ith layer in
a layered system may be expressed as

i
= F
i
P
i
(3.39)
where
i
= (U
zp
, U
p
, -U
zm
, -U
m
)
T

P
i
= (N
p
, T
p
, N
m
, T
m
)
T

F
i
is the flexibility matrix given in Appendix II.


Solution of Uncoupled Equations
Equation (3.30c) is an uncoupled equation which is independent of equations (3.30a) and
(3.30b). Substituting each term in the uncoupled equation into equation (3.28) gives
z
U
f
S
z

2
(3.40a)
The uncoupled equation may be transformed and expressed as
0 =

+
z
S
S
z

(3.40b)
Substituting the terms of equation (3.30c) into the stress-strain relationship (3.28) gives
the following equations
z
S
f z
U

2
=

(3.41a)

S
b a
U
) (
2

= (3.41b)
Substituting equations (3.41a) and (3.41b) to equation (3.40b) gives
z
z
S
f
b a
z
S

2
2
2
(3.42)
The general solution may then be written
S
z
= C
1
cosh (kz) + C
2
sinh (kz) (3.43)
where C
1
, C
2
are constants and
Chapter 3 Finite Layer Analysis of Layered Soil
52

=
f
b a
k
Substituting equation (3.43) into equation (3.41a) gives
[ ] ) sinh( ) cosh(
1 2
2 1
kz C kz C
k f
U + =

(3.44)
which leads to the flexibility relationship

) 2 ( cot ) 2 ( cos
) 2 ( cos ) 2 ( cot
kh anh kh ech
kh ech kh anh
k
F
U
U
m
p

(3.45)

3.2.3 Stiffness Relationship
The stiffness relationship for a layer of material is given as
P
i
= K
i

i
(3.46)
where K
i
is the stiffness matrix of layer i.
The stiffness matrix K
i
for individual layer may be obtained by inverting the flexibility
matrix [F
i
]
-1
.

The total stiffness matrix [K] can be obtained by summing the stiffness matrix for n
layers. The global stiffness relationship may be expressed as
P

= K (3.47)

where

=
=
n
i
i
1

=
=
n
i
i
P P
1

=
=
n
i
i
K K
1

Consider loading the soil surface of an n layered system by a horizontal or vertical
uniform load as shown in Figure 3.2, the continuity of the internal stresses at the interface
between two consecutive layers requires
(N
m
)
i+1
= -(N
p
)
i
(3.48a)
(T
m
)
i+1
= -(T
p
)
i
(3.48b)
Chapter 3 Finite Layer Analysis of Layered Soil
53
and the continuity of the displacements requires
(U
xm
)
i+1
= (U
xp
)
i
(3.49a)
(U
zm
)
i+1
= (U
zp
)
I
(3.49b)
Therefore, the terms in the force vector will be zero, except at the soil surface (where the
load is applied) and at the soil base. The force and displacement vector may be written as
P = (T
1
, N
1
, 0, 0, ,T
n
, N
n
) (3.50a)
= (U
x1
, U
z1
, U
x2
, U
z2
, , U
xn
, U
zn
) (3.50b)
By applying the boundary conditions to the soil base, the stress and displacement
transform can be eliminated. The boundary conditions for a rectangular loading applied
to a layered system are given in Appendix III.
3.2.4 Transform of Loads
As stated in the previous sections, loads along the pile shafts and at the base are treated as
a series of unit ring loads applied to the soil interface and a uniform circular load applied
to the pile base. Loads acting on the raft are treated as uniform rectangular loads applied
to the soil surface. To simulate these loadings applied to the soil, it is necessary to obtain
the transforms of the loads.

3.2.4.1 Hankel Transform of Vertical and Horizontal Ring Load
To obtain the transform of a vertical ring load, we have to first consider a uniform
annular load q with an inner radius r
in
, and outer radius r
out
. Hence, the total load P
an
may
be defined as
q r r P
in out an
) (
2 2
=

The Hankel transform for the load may be expressed as

= =
2 2
1 1
0
) ( ) (
) (
in out
in in out out an
r
r
z
r r
r J r r J r P
dr r rqJ S
out
in

(3.51)

Chapter 3 Finite Layer Analysis of Layered Soil
54
For a ring load, the outer radius r
out
approaches r
in
, the transform of the load may then be
expressed as
) (
2
0
r J
P
S
z

= (3.52)
where J
0
(r) and J
1
(r) are the Bessel functions
= Hankel transform parameter

Consider a horizontal ring load h
x
with a radius r, the total load H
x
can be obtained by the
same method as for the vertical ring loads. The transform of the horizontal load in the x-
direction may be expressed as
) (
4
0
2
r J
iH
T
x
x

= (3.53a)
and therefore the load in the y-direction may be expressed as
) (
4
0
2
r J
iH
T
y
y

= (3.53b)

where T
x
= transform of uniform horizontal load in the x-direction
T
y
= transform of uniform horizontal load in the y-direction
H
x
= total horizontal load in the x-direction
H
y
= total horizontal load in the y-direction
r = radius of the ring load
i = 1
3.2.4.2 Hankel Transform of Vertical and Horizontal Circular Load
For a uniform vertical circular load, the inner radius r
in
of the ring load approaches zero.
From equation (3.51), the transform of the circular load may be written as
) (
1
r J
qr
S
z

= (3.54)
As for the horizontal ring load, the transform of the horizontal circular load in the x-
direction can be expressed in terms of the uniform shear load h
x

2
) (
1
r rJ ih
T
x
x
= (3.55a)
Chapter 3 Finite Layer Analysis of Layered Soil
55
and the transform in the y-direction in terms of h
y

2
) (
1
r rJ ih
T
y
y
= (3.55b)

3.2.4.3 Hankel Transform of Vertical and Horizontal Rectangular Load
For a uniformly distributed load q of width 2a and length 2b applied vertically on a
rectangular area, the transform of the vertical load may be expressed in the form


=
a
a
b
b
z
dxdy y x q S ) cos( ) cos(
4
1
2

(3.56a)
which leads to

) sin( ) sin(
2
b a q
S
z
= (3.56b)
For a uniform distributed load h
x
and h
y
applied horizontally in the x- and y-directions,
the transformed shear load is given as


2
) sin( ) sin( b a ih
T
x
x
= (3.57a)


2
) sin( ) sin( b a ih
T
y
y
= (3.57b)
3.2.5 Solutions for Displacements
3.2.5.1 Soil Displacements for Vertical Ring or Circular Loads
When a layered soil is subjected to a vertical uniform circular or ring load, the
displacements in the x-, y- and z-directions at the layer interfaces may be expressed as

=
0
1
) ( cos d r J U u
r x
(3.58a)

=
0
1
) ( sin d r J U u
r y
(3.58b)

=
0
0
) ( d r J U u
z z
(3.58c)
Chapter 3 Finite Layer Analysis of Layered Soil
56
3.2.5.2 Soil Displacements for Horizontal Ring or Circular Loads
To obtain the solution for the soil displacement due to a horizontal ring or circular load,
we have to resolve the shear loads T
x
and T
y
into the plane. From equation (3.53a),
the , components of the shear load in the x-direction may be written as

cos ) (
4
0
2
r J
iT
T
x
= (3.59a)

sin ) (
4
0
2
r J
iT
T
x
= (3.59b)
and the , components of the shear load in the y-direction may be written as

sin ) (
4
0
2
r J
iT
T
y
= (3.60a)

cos ) (
4
0
2
r J
iT
T
y
= (3.60b)
Resolving the transform of the displacement U
x
in the plane gives


2 2
sin cos U U U
x
= (3.61)
where

cos U U = and

sin U U =
Substituting equation (3.61) to equation (3.26), the solution for the displacement may be
written as



d d e U U i u
y x i
x
) ( 2 2
) sin cos (
+

=

(3.62)
where x = r cos , y = r sin
= cos , = sin
Equation (3.61) may be rewritten in terms of the cylindrical co-ordinates

d d e U U i u
r i
x
) cos( 2 2
2
0
) sin cos (

=

(3.63)
Put = , d = d

d d e U U i u
r i
x
cos 2 2
2
0
) sin cos ( =

Knowing that

d n e
i
z J
iz
n
n
) cos(
1
) (
0
cos

= , the solution for the displacement in the x-

direction may be expressed as
Chapter 3 Finite Layer Analysis of Layered Soil
57


d r J
U U
r J
U U
u
x

+
+
= ) (
2
) ( 2 cos
2
2
0 2
(3.64)
To obtain the displacement in the y-direction, U
y
is resolved in the plane giving


cos sin cos sin U U U
y
=
The solution may be obtained by the same method as was used for the x-direction and
expressed as

+ =

d r J U U u
y
) ( cos sin ) ( 2
2
(3.65)
The solution of the displacement in the z-direction may be obtained by putting
cos
z z
U U =
which gives

= d r J U u
z z
) ( cos 2
1
(3.66)

3.2.5.3 Soil Displacements due to Vertical Rectangular Loads
The solution for the soil displacement due to a uniform vertical rectangular load can be
obtained from equation (3.26) and expressed in terms of the co-ordinates

+
=
0
2
0
) (

d d e iU u
y x i
x x
(3.67a)

+
=
0
2
0
) (

d d e iU u
y x i
y y
(3.67b)

+
=
0
2
0
) (

d d e U u
y x i
z z
(3.67c)
Resolving the transformed displacements in the plane gives
U
x
= U

cos , = cos
U
y
= U

sin , = sin
Equation (3.67a-c) can then be rewritten in the form

=
0
2
0
) cos( ) sin( 4

d d y x U u
x
(3.68a)

=
0
2
0
) sin( ) cos( 4

d d y x U u
y
(3.68b)

=
0
2
0
) cos( ) cos( 4

d d y x U u
z
(3.68c)
Chapter 3 Finite Layer Analysis of Layered Soil
58
3.2.5.4 Soil Displacements due to Horizontal Rectangular Loads
To obtain the solution of the soil displacement due to a horizontal rectangular load, the
shear load T
x
and T
y
have to be resolved in the plane. From equation (3.57a) the
components of the shear load in the x-direction may be written as

cos
) sin( ) sin(
2
b a ih
T
x
= (3.69a)

sin
) sin( ) sin(
2
b a ih
T
x
= (3.69b)
and the components of the shear load in the y-direction (equation 3.57b) may be written
as

sin
) sin( ) sin(
2
b a ih
T
y
= (3.70a)

cos
) sin( ) sin(
2
b a ih
T
y
= (3.70b)
Resolving the transform of displacement in the plane as was done for the circular or
ring load and substituting into equation (3.26) gives

d d y x U U N u
x
) cos( ) cos( ) sin cos (
2
0
2
0
2

= (3.71a)

d d y x U U N u
y
) sin( ) sin( ) sin sin cos (
2
0
2
0

+ = (3.71b)

=
0
2
0
(cos

d d U N u
z z
(3.71c)
where


2
sin sin 4 b a
N =
Equation (3.71a-c) may then be rewritten in terms of and

d d y x U U N u
x
) cos( ) cos( ) (
2
0
2
0
2

= (3.72a)

d d y x U U N u
y
) sin( ) sin( ) (
0
2
0

+ =

(3.72b)

=
0
2
0
) cos( ) sin(

d d y x U N u
z
(3.72c)
Chapter 3 Finite Layer Analysis of Layered Soil
59

3.3 NUMERICAL INTEGRATION
To obtain the actual values of field quantities, inverse integrals with infinite limits of
integration have to be evaluated. This can be done by the use of Gaussian integration and
the infinite integral can be truncated at a value of A provided that A is large enough to
provide a good approximation to the integral. Gaussian integration of a function of may
be carried out as follows

= =
G
i
i i
A
A
f W d f d f
1
) ( ) ( ) ( (3.73)
where
i
= value of at a selected Gauss point
W
i
= Gaussian weights at the Gauss points
G = number of Gauss points
The finite range of integration (from A to +A) is then divided into 2n blocks of length t.
The finite integral will then have the integration limits between nt and +nt as shown in
Figure 3.4. The accuracy of the integration depends on the integration scheme which is a
combination of the number and length of the blocks.

Consider a uniform circular load of magnitude q with a radius of a applied vertically or
horizontally on a homogeneous elastic half space. The modulus and Poissons ratio of the
soil are assumed to be E (= 10MPa) and (= 0.3) respectively. Four different integration
schemes listed in Table 3.1 were considered. The effects of the integration schemes on
the normalised vertical (v) and horizontal (u) surface displacements of the soil subjected
to vertical and horizontal loads are shown in Figures 3.5 and 3.6 respectively. It may be
seen that for both vertical and horizontal loadings, integration schemes 3 and 4 can
produce accurate solutions at distances up to x = 40a, however, schemes 1 and 2 can only
produce accurate solutions at distances up to x =14a and x = 28a respectively. With the
use of the same number of integration blocks, smaller block sizes (t) will give more
accurate solutions for the points far away from the centre of the load. Although the
integration range must be large enough to provide an accurate solution, making the range
Chapter 3 Finite Layer Analysis of Layered Soil
60
too large does not improve the degree of accuracy of the solution. Therefore, appropriate
combinations of the number and size of blocks are required to provide correct solutions.

Integration Scheme

1 2 3 4
No. of Blocks (n) 20 20 20 20
Size of Blocks (t) 10/a 5/a 1/a 0.5/a
Table 3.1 Integration Schemes for different integration ranges

Furthermore, the maximum integration range is found to be inversely proportional to the
size of the loaded area which determines the size of the integration blocks. To
demonstrate the effect of the size of the loaded area on the accuracy of solutions, consider
a uniform circular load of q with a radius of a applied horizontally to a homogeneous
elastic half space. The number of integration blocks (n) was chosen as 20 and the size of
the integration blocks (t) is then determined by the inverse of the size of the loaded area
(i.e. t = 1/a). Different sizes of the loaded area were considered as listed in Table 3.2.

Integration Scheme
1 2 3 4
Radius of
Loaded Area (a)
0.5 1 2 4
Size of Blocks (t) 4a a a/4 a/16
Table 3.2 Integration schemes for different radius of loaded area

Figure 3.7 shows the normalised displacement in the vertical (v) and horizontal (u)
directions when a uniform circular load was applied horizontally in the x-direction for
different radius of loadings. It can be observed that the accuracy of solution can be
guaranteed if the size of the integration blocks is determined by the method mentioned
above. Figures 3.8a and b show the normalised vertical and horizontal displacements for
different radius of loading area when the number and size of blocks remain unchanged.
The number of integration blocks was again chosen as 20 and the size of the blocks was
Chapter 3 Finite Layer Analysis of Layered Soil
61
chosen as 10. The radius of the loading area (a) varied between 0.5 and 4.0. For a = 0.5,
the solutions become inaccurate at about 10 radii from the centre of the loading. As the
radius of the loading area increases, the solutions become inaccurate at distances closer to
the centre of loadings. For a = 4, the solution is accurate to about 4 radii from the load
centre. It can be observed that for small radius of loaded area, a large size of integration
blocks is required to provide an accurate solution.

For the three dimensional rectangular loading, a double integration is required to invert
the double Fourier transform and the integral is in the form of

2
0 0
) , ( ) , ( d d H y x h (3.74)
The first numerical integration ranges from 0 to . The integration range of this infinite
integral has to be large enough to obtain accurate solutions. The infinite integral can be
evaluated in the same way as for a ring or circular load as described previously.
Therefore, the integration range required is again inversely proportional to the size of the
loaded area. If the loadings are of different sizes, the size of the integration blocks can be
determined by inverting the average of the width and length of the loaded areas. The
integration scheme also depends on the distance from the centre of the load, and for
accuracy, the block size has to be made smaller for greater distances from the load.

The second numerical integration ranges from 0 to 2. The effect of the integration
schemes on the accuracy of the solution is examined by considering a uniform
rectangular load of q with a width of 2a and a length of 2b applied vertically or
horizontally on the soil surface. The modulus and Poissons ratio of the soil are assumed
to be E (= 10MPa) and (= 0.3) respectively. The number and size of the integration
blocks were chosen as 20 and 1/a respectively. The integration schemes for integration
are listed in Table 3.3. Figures 3.9a and b show the normalised vertical and horizontal
displacements for uniform vertical rectangular loads. Figures 3.10a and b show the
displacements for a uniform horizontal rectangular load in the vertical and horizontal
directions respectively. It can be seen that for the loadings in both directions, the effect of
the number of blocks used for the integration on the solutions is remarkably small. Use
Chapter 3 Finite Layer Analysis of Layered Soil
62
of more integration blocks will improve the accuracy slightly but will also increase the
computation time. Therefore, one block of width /2 is recommended.

Integration Scheme
1 2 3 4 5 6
Width (a) /
Length (b) of
Loaded Area
0.5/0.5 0.5/0.5 0.5/1.0 0.5/1.0 1.0/0.5 1.0/0.5
No. of Blocks 1 5 1 5 1 5
Table 3.3 Integration schemes for integration

3.4 NUMERICAL EXAMPLES
3.4.1 Soil subjected to Vertical Loadings
3.4.1.1 Ring Loads
A ring load with a radius a and a load q (load per unit length) is applied vertically to the
surface of a semi-infinite mass as shown in Figure 3.11 in this example. From Poulos and
Davis (1974) the analytical solution for the vertical displacements along the axis of
loading which was expressed as

+
+
+
+
=
) (
) 1 ( 2
) (
) 1 (
2 2
2
2
1
2 2
z a
z
z a E
a q
u
z

(3.75)
where q = applied load (per unit length)
a = radius of the ring load
z = depth from the soil surface
E = Youngs modulus of the soil
= Poissons ratio of the soil
The present method was used to obtain the vertical displacement and these were
compared with the analytical solution. The Youngs modulus and Poissons ratio of the
soil were chosen as E and = 0.499 respectively. The soil was assumed to be infinitely
Chapter 3 Finite Layer Analysis of Layered Soil
63
deep and the number and size of the integration blocks were chosen to be 20 and 1/a
respectively. The displacements obtained from the present method are in good agreement
with the analytical solutions as shown in Figure 3.12.

3.4.1.2 Circular Loads
A uniform circular load with radius a and a uniform load q (load per unit area) applied
vertically to the surface of a semi-infinite mass as shown in Figure 3.13 is next
considered. From Poulos and Davis (1974) the solution for a uniform circular load
applied vertically and the vertical displacements along the axis of loading was expressed
as

+
+

=
2
2
2
1 ) 1 ( 2
1 1
) 1 ( 2
a
z
a
z
a
z
a
z
E
qa
u
z

(3.76)
The parameters of the soil and the integration scheme used in the present method are the
same as in the previous example. Figure 3.14 shows that the vertical displacements along
the axis of loading obtained from the present method are in good agreement with the
analytical solution.

Alhvin and Ulery (1962) presented an analytical solution for the displacement of the soil
at different points beneath the loaded area. Vertical displacement at three different depths
(z/a = 0, 2, 4) obtained from both methods are presented in Figure 3.15. The results
obtained from the present method are in good agreement with the analytical solutions.
The displacement at different depth ratios approaches the same value when the distance
ratio r/a is greater than 4.

The behaviour of a soil subjected to a uniform circular load applied at a depth c below the
soil surface as shown in Figure 3.16 has been examined by Nishida (1966). The analytical
solution for the vertical displacement at the centre of the load can be expressed as
E
qaI
u
o
z
= (3.77)
Chapter 3 Finite Layer Analysis of Layered Soil
64
where I
o
= the influence factor (0.862 for = 0.5 as derived by Nishida (1966)). For the
vertical displacement along the axis of loading, the expression can be written as
{ }
{ }
]
) (
2 ) )( 4 3 (
) (
2
) ) ( (
) ( 2
2 ) )( 4 3 (
) (
) (
) (
) ( ) ( ) 8 12 5 (
) ( ) ( ) 4 3 [(
) 1 ( 4
) 1 (
2 2
2
3 2 2
2
2
2 2
2
2 2 2
2 2
c z a
cz c z
c z
cz
c z a
c z cz
c z
cz c z
c z a
c z
c z
c z c z a
c z c z a
E
q
+ +
+

+
+
+ +
+

+
+
+
+

+
+ + + + +
+

+
=

(3.78)

Figure 3.17 shows comparisons of the vertical displacement obtained from the present
method and the analytical solution where the load is applied at a depth c/a = 5. The
results from the present method are in good agreement with the analytical solution. As
expected, the displacement is a maximum at the depth c/a = 5 where the load is applied.

3.4.1.3 Rectangular Loads
In this example, a uniform rectangular load q (load per unit area) of length a and width b
in plan is applied vertically to the surface of a semi-infinite mass as shown in Figure 3.18.
The ratio of a/b was taken as 1.0 and the Youngs modulus and Poissons ratio were
chosen to be E and = 0.499 respectively.

Giroud (1968) presented theoretical solutions for the vertical surface displacements at
any point in a soil subjected to uniform rectangular loading and these can be expressed as
qbI
E
u
z
2
1
= (3.79)
where I = influence factor for vertical surface displacement.

Harr (1966) also presented solutions for vertical displacement beneath the corner of the
rectangular load and these may be expressed as
)
1
2 1
)( 1 (
2
B A
E
qb
u
z

= (3.80)
u
z

Chapter 3 Finite Layer Analysis of Layered Soil
65
where
1 1
1 1
ln
1
1
(ln
2
1
2 2
2 2
2 2
2 2
+ +
+ + +
+
+ +
+ + +
=
n m
n m
m
m n m
m n m
A

2 2
1
1
tan
2
n m n
m n
B
+ +
=

Figure 3.19 shows the comparison of the vertical displacement beneath the corner of the
rectangular load between the present method and Harrs solutions (1966). The results
from the present method are in excellent agreement with the theoretical solutions.

3.4.2 Soil subjected to Horizontal Loadings
3.4.2.1 Ring Loads
If the radius of the ring load is small enough or the distance between the point of interest
and the loading centre is big enough, the displacements due to the ring load should be
fairly close to the displacements due to a point load provided that the total loads are the
same. From Poulos and Davis (1974), some expressions for the displacements due to a
horizontal point load acting along the surface of a semi-infinite mass as shown in Figure
3.20 were presented. The horizontal displacement was expressed as

+
+ +
+
= )
) (
)( 2 1 ( 1
2
) 1 (
2
2
2
2
z R
x
z R
R
R
x
ER
Q
u
x

(3.81)
and the vertical displacement was expressed as

+
+
=
z R
x
R
xz
ER
Q
u
z
) 2 1 (
2
) 1 (
2

(3.82)
where Q = concentrated point load = 2aq
q = ring load (per unit length)
R = distance between the point load and point of interest =
2 2 2
z y x + +
Figure 3.21 shows a comparison of the normalised displacements from the ring load and
concentrated point load at 3 different distance to radius ratios (i.e. x/a = 4, 12, 20). The
ring load with radius a and load q (per unit length) and an equivalent point load are
considered to act on the surface. The soil has a modulus, E and Poissons ratio, = 0.499.

Chapter 3 Finite Layer Analysis of Layered Soil
66
Figures 3.21a and b show the normalised horizontal and vertical displacements beneath
different points respectively. At distances close to the loading centre (x/a = 4), there
exists some differences between the ring load and point load. However, as the distances
increase, the differences become small and approach the same values.

3.4.2.2 Circular Loads
As for the ring load, displacements due to a uniform circular load on the surface of a
semi-infinite mass as shown in Figure 3.22 will be fairly close to the displacements due
to a concentrated load if the radius of the circular load is small enough or the point of
interest is far enough from the loading centre.

Figures 3.23a and b show the normalised horizontal and vertical displacements beneath
different points (r/a = 4, 12, 20) for a uniform circular load with radius a and load q (per
unit area) applied horizontally to the surface of a semi-infinite mass with the same soil
parameters as in the previous example. The displacements in both horizontal and vertical
directions for the circular load and point load are slightly different at close distances (r/a
= 4), but as the distance increases, the displacements approach the same values.

After comparing the displacements between a circular load and a point load, we now
examine the behaviour of a circular load on a multi-layer system. A uniform circular load
with radius a and a total load Q is applied horizontally to the surface of a two layered soil
system with a thickness h for the upper layer and infinitely deep for the lower layer. The
Youngs moduli are chosen to be E
1
and E
2
and Poissons ratio are chosen as
1
and
2
as
shown in Figure 3.24.

The theoretical solutions for a uniform horizontal circular loading on a two layered soil
system have been presented by Westmann (1963). The vertical and horizontal
displacement can be expressed as
cos
2
a G
Q
I u
z
u z
= (3.83)
Chapter 3 Finite Layer Analysis of Layered Soil
67

sin
2
a G
Q
I u
u
= (3.84)
where

u u
I I
z
, are the influence factors for vertical and horizontal surface displacement.
Figure 3.25 shows that the results from the present method are in good agreement with
the theoretical solutions in which E
1
= 2E
2
= 40,
1
=
2
= 0.5, h = 2a and the bottom layer
is infinitely deep. It can be seen that the horizontal displacement (Figure 3.25a) is a
maximum at the centre of the loaded area while the vertical displacement (Figure 3.25b)
is at a maximum where the ratio of distance to radius r/a is about 2.

When the distance from the loading is far enough, the displacement of the soil due to a
ring load should be fairly close to the displacement due to a circular load. To demonstrate
that this is so, a ring load and a uniform circular load with radius a and a total load Q are
applied horizontally to the surface of a semi-infinite mass. The Youngs modulus and
Poissons ratio are chosen as E (= 20MPa) and (= 0.5) respectively for the uniform soil
analysed.

The variations of the normalised horizontal and vertical displacements beneath 3 different
positions (r/a = 2, 4, 10) from the centre of the loaded area are presented in Figures 3.26a
and b. It can be seen that at r/a = 2, there exists some differences between the
displacements in both directions at points near the surface (i.e. z/a < 4). As the distance
from the loaded centre increases, the displacements of the soil due to the ring and circular
loads tend to be the same. Figure 3.26a shows that the horizontal displacements at
different positions beneath the loaded area approach the same value at z/a > 8. Figure
3.26b shows that the maximum vertical displacement occurs near the surface (i.e. z/a 1)
where r/a = 2. As the distance from the centre of loading increases, the locations at where
the maximum displacements occur will be further away from the soil surface.

3.4.2.3 Rectangular Loads
Giroud (1969) presented solutions for the vertical and horizontal surface displacements
for the soil subjected to a horizontal uniform rectangular load q with a length a and a
width b and on the surface of a semi-infinite mass as shown in Figure 3.27. The Youngs
Chapter 3 Finite Layer Analysis of Layered Soil
68
modulus and Poissons ratio of the soil were chosen to be E (= 20MPa) and (= 0.5)
respectively for this example. The expression for the horizontal displacement at the
corners of the load is

+ +
+
+ +

+
=
b
b a a
a
b
a
b a b
qa
E
u
x
2 2 2 2
ln ln ) 1 (
) 1 (

(3.84)
) (
) 1 (
2 2
b a b a q
E
u
y
+ +
+
=

(3.85)
in which the displacements in the y-direction at corners A and C are positive and at
corners B and D are negative (Figure 3.27).

Figure 3.28 shows the comparison of the normalised corner displacements (in the x- and
z-directions) between the present method and the theoretical solution from Giroud (1969).
The normalised displacements are plotted against the width-length ratio (b/a). It can be
observed that the results from the present method are in excellent agreement with the
analytical solutions.

3.5 CONCLUSIONS
The Finite Layer method has been discribed in this chapter. This method can simplify a
two- or three-dimensional problem into a one-dimensional problem through the
application of a Fourier or Hankel transformation of the field quantities. Problems with
ring, circular and rectangular loads can be analysed by this method. The soil material can
have different material properties or can be multi-layered.

The application of the finite layer method to the analysis of problems with ring, circular
and rectangular loading in both horizontal and vertical directions has been demonstrated
through the above examples. The following conclusions can be reached:

(1) The accuracy of the solution obtained from the finite layer method depends on the
numerical integration schemes used to invert the transformation of the field
quantities. For ring and circular loads, the integration schemes depend on the radius
Chapter 3 Finite Layer Analysis of Layered Soil
69
of the loaded area and the distance from the centre of loading. For rectangular load,
double integration is required and the schemes for the first integration (with respect
to parameter ) depend on the half width of the rectangular loaded area. For different
types of loading, if the size of the loaded area is small, the size of the integration
blocks needs to be large to achieve comparable accuracy. Thus, the integration range
is inversely proportional to the size of the loaded area but also depends on the
distance at which the solution is evaluated.
(2) Comparisons of the solutions obtained from the finite layer method and the analytical
solutions for different shaped loadings in the horizontal or vertical directions have
shown that the present method provides accurate and reliable solutions.
(3) If the size of the loaded area is small enough or the distance from the loading centre
is far enough, the displacements due to uniform circular, ring or rectangular loadings
will be fairly close to the displacements due to a point load.



Chapter 3 Finite Layer Analysis of Layered Soil
70

rz

rz

rz

rz
x

r

z

r
d

x

Figure 3.1 Stresses on an element in a cylindrical (r, , z) co-ordinate system
Chapter 3 Finite Layer Analysis of Layered Soil
71
i
T
p
N
p
T
m
N
m
z

h
h
x
Layer i
Layer i + 1
T
p
N
p
N
m
T
m
U
rp
U
zp
U
zm
U
rm
z

Figure 3.2 Transform of normal and shear stresses and displacements for layer i
y

x

Rectangular
Loaded Area
Figure 3.3 Three dimensional Cartesian co-ordinate system for rectangular load
Chapter 3 Finite Layer Analysis of Layered Soil
72
f()

t t t t t t
Block 1 Block 3 Block 4 Block 5 Block 6 Block 2
Figure 3.4 Numerical integration scheme
Chapter 3 Finite Layer Analysis of Layered Soil
73
Figure 3.5 The effect of integration schemes on the normalized
vertical and horizontal displacements of a soil
subjected to vertical circular loads.
(b) Normalised horizontal displacement of soil
(a) Normalised vertical displacement of soil
-0.05
0.00
0.05
0.10
0.15
0.20
0 5 10 15 20 25 30 35 40
Normal i sed Di stance , x /a
N
o
m
r
a
l
i
s
e
d

D
i
s
p
l
a
c
e
m
e
n
t
,

v
/
a
n=20 t=10
n=20 t=5
n=20 t=1
n=20 t=0.5
-0.014
-0.012
-0.010
-0.008
-0.006
-0.004
-0.002
0.000
0.002
0.004
0 5 10 15 20 25 30 35 40
Normal i sed Di stance, x /a
N
o
r
m
a
l
i
s
e
d

D
i
s
p
l
a
c
e
m
e
n
t
,

u
/
a
n=20 t=10
n=20 t=5
n=20 t=1
n=20 t=0.5
Chapter 3 Finite Layer Analysis of Layered Soil
74
(a) Normalised vertical displacement of soil
(b) Normalised horizontal displacement of soil
Figure 3.6 The effect of integration schemes on the normalized
vertical and horizontal displacements of a soil
subjected to horizontal circular loads.
-0.05
0.00
0.05
0.10
0.15
0.20
0.25
0 5 10 15 20 25 30 35 40
Normal i sed Di stance, x/a
N
o
r
m
a
l
i
s
e
d

D
i
s
p
l
a
c
e
m
e
n
t
,

u
/
a
n=20 t=10
n=20 t=5
n=20 t=1
n=20 t=0.5
-0.004
-0.002
0.000
0.002
0.004
0.006
0.008
0.010
0.012
0.014
0 5 10 15 20 25 30 35 40
Normal i sed Di stance, x /a
N
o
r
m
a
l
i
s
e
d

D
i
s
p
l
a
c
e
m
e
n
t
,

v
/
a
n=20 t=10
n=20 t=5
n=20 t=1
n=20 t=0.5
Chapter 3 Finite Layer Analysis of Layered Soil
75
0.00
0.05
0.10
0.15
0.20
0.25
0 10 20 30 40 50
Normal i sed Di stance, x /a
N
o
r
m
a
l
i
s
e
d

D
i
s
p
l
a
c
e
m
e
n
t
,

u
/
a
a = 0.5, t=2
a = 1, t=1
a = 2, t=0.5
a = 4, t=0.25
-0.002
0.000
0.002
0.004
0.006
0.008
0.010
0.012
0.014
0 10 20 30 40 50
Normal i sed Di stance, x /a
N
o
r
m
a
l
i
s
e
d

D
i
s
p
l
a
c
e
m
e
n
t
,

v
/
a
a = 0.5, t=2
a = 1, t=1
a = 2, t=0.5
a = 4, t=0.25
(a) Normalised vertical displacement of soil
(b) Normalised horizontal displacement of soil
Figure 3.7 The effect of the radius of load on the normalized
vertical and horizontal displacements of the soil
subjected to a uniform circular horizontal load.
No. of Blocks (n) = 20
No. of Blocks (n) = 20
Chapter 3 Finite Layer Analysis of Layered Soil
76
-0.05
0.00
0.05
0.10
0.15
0.20
0.25
0 10 20 30 40 50
Normal i sed Di stance, x/a
N
o
r
m
a
l
i
s
e
d

D
i
s
p
l
a
c
e
m
e
n
t
,

u
/
a
a = 0.5
a = 1
a = 2
a = 4
-0.010
-0.005
0.000
0.005
0.010
0.015
0 10 20 30 40 50
Normal i sed Di stance, x /a
N
o
r
m
a
l
i
s
e
d

D
i
s
p
l
a
c
e
m
e
n
t
,

v
/
a
a = 0.5
a = 1
a = 2
a = 4
(a) Normalised vertical displacement of soil
(b) Normalised horizontal displacement of soil
Figure 3.8 Normalized vertical and horizontal displacements of
the soil subjected to a uniform circular horizontal
load when the number and size of blocks remain
unchanged.
No. of Blocks (n) = 20
Size of Blocks (t) = 10
No. of Blocks (n) = 20
Size of Blocks (t) = 10
Chapter 3 Finite Layer Analysis of Layered Soil
77
(a) Normalised vertical displacement of soil
(b) Normalised horizontal displacement of soil
Figure 3.9 The effect of the number of integration blocks on
the normalized displacement of the soil subjected to
a vertical rectangular load.
-0.05
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0 5 10 15 20 25 30 35 40
Normal i sed Di stance, x/a
N
o
r
m
a
l
i
s
e
d

D
i
s
p
l
a
c
e
m
e
n
t
,

v
/
a
a =b =0.5, 1 block
a =b =0.5, 5 blocks
a=0.5, b=1.0, 1 block
a=0.5, b=1.0, 5 blocks
a=1.0, b=0.5, 1 block
a=1.0, b=0.5, 5 blocks
-0.030
-0.025
-0.020
-0.015
-0.010
-0.005
0.000
0 5 10 15 20 25 30 35 40
Normal i sed Di stance, x/a
N
o
r
m
a
l
i
s
e
d

D
i
s
p
l
a
c
e
m
e
n
t
,

u
/
a
a =b =0.5, 1 block
a =b =0.5, 5 blocks
a =0.5, b =1.0, 1 block
a =0.5, b =1.0, 5 blocks
a=1.0, b=0.5, 1 block
a=1.0, b=0.5, 5 blocks
Chapter 3 Finite Layer Analysis of Layered Soil
78
(a) Normalised vertical displacement of soil
(b) Normalised horizontal displacement of soil
Figure 3.10 The effect of the number of integration blocks on
the normalized displacement of a soil subjected to a
horizontal rectangular load.
0.000
0.005
0.010
0.015
0.020
0.025
0.030
0 5 10 15 20 25 30 35 40
Normal i sed Di stance, x/a
N
o
r
m
a
l
i
s
e
d

D
i
s
p
l
a
c
e
m
e
n
t
,

v
/
a
a =b =0.5, 1 block
a =b =0.5, 5 blocks
a=0.5, b=1.0, 1 block
a=0.5, b=1.0, 5 blocks
a=1.0, b=0.5, 1 block
a=1.0, b=0.5, 5 blocks
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0 5 10 15 20 25 30 35 40
Normal i sed Di stance, x/a
N
o
r
m
a
l
i
s
e
d

D
i
s
p
l
a
c
e
m
e
n
t
,

u
/
a
a =b =0.5, 1 block
a =b =0.5, 5 blocks
a =0.5, b =1.0, 1 block
a =0.5, b =1.0, 5 blocks
a=1.0, b=0.5, 1 block
a=1.0, b=0.5, 5 blocks
Chapter 3 Finite Layer Analysis of Layered Soil
79
Figure 3.12 Comparison of vertical displacement along the axis
of loading between the finite layer method and the
analytical solution.
Figure 3.11 Uniform vertical ring loading on the surface of a semi-infinite
mass
a
r
z
r
Load
q/unit length
z
O
0.000
0.100
0.200
0.300
0.400
0.500
0.600
0.700
0.800
0.900
0 5 10 15 20 25
z/a
u
z
E
/
2
a
q
Poulos & Davis (1974)
Finite Layer Method
0 =
a
r

Chapter 3 Finite Layer Analysis of Layered Soil
80
Figure 3.13 Uniform circular loading on the surface of a semi-infinite
mass
a
r
z
r
Load
q/unit area
z
O
Figure 3.14 Comparison of vertical displacement along the axis
of loading between the finite layer method and the
analytical solution
0
5
10
15
20
25
0 0.1 0.2 0.3 0.4 0.5 0.6
z
/
a
u
z
E/a
2
q
Poulos & Davis (1974)
Finite Layer Method
0 =
a
r

Chapter 3 Finite Layer Analysis of Layered Soil
81
Figure 3.15 Comparison of vertical displacement at various
distances from the loading centre between the finite
layer method and the analytical solution
0
0.1
0.2
0.3
0.4
0.5
0.6
0 5 10 15
r/a
E
u
z
/

a
2
q
Ahlvin & Ulery (1962)
Finite Layer Method
z/a = 0
2
4
Chapter 3 Finite Layer Analysis of Layered Soil
82
c
r
Load
q/unit area
z
a
O
Figure 3.16 Uniform circular loading applied at a depth c below
the surface
Figure 3.17 Comparison between the finite layer method and
the analytical solution for vertical displacement
along the axis of loading
0
2
4
6
8
10
12
14
16
18
20
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
u
z
E/a
2
q
z
/
a
Nishida (1966)
Finite Layer Method
0 =
a
r

Chapter 3 Finite Layer Analysis of Layered Soil
83
Figure 3.18 Uniform vertical rectangular loading on the surface of
a semi-infinite mass
Figure 3.19 Comparison between the finite layer method and
the analytical solution for vertical displacement
beneath the corner of a rectangular load
b
a
x
Load
q/unit area
O
z
y
0
2
4
6
8
10
12
0 0.1 0.2 0.3 0.4 0.5
u
z
E/qa
z
/
a
Harr (1966)
Finite Layer Method
1 =
b
a

Chapter 3 Finite Layer Analysis of Layered Soil
84
Q
z
x
2 2 2
z y x R + + =
Point (x,y,z)
Figure 3.20 Horizontal point load acting on the surface of a semi-infinite mass
(a) Horizontal displacement
(b) Vertical displacement
Figure 3.21 Comparison of displacement between a ring load
and a concentrated point load
0
5
10
15
20
25
30
35
40
45
0 0.05 0.1 0.15 0.2 0.25 0.3
u
x
E/2aq
z
/
a
Point Load x/a =4
Ring Load, x/a =4
Point Load x/a =12
Ring Load, x/a =12
Point Load x/a =20
Ring Load, x/a =20
0
5
10
15
20
25
30
35
40
45
0 0.01 0.02 0.03 0.04 0.05
u
z
E/2aq
z
/
a
Point Load x/a =4
Ring Load, x/a =4
Point Load x/a =12
Ring Load, x/a =12
Point Load x/a =20
Ring Load, x/a =20
Chapter 3 Finite Layer Analysis of Layered Soil
85
Figure 3.22 Uniform horizontal circular loading on the surface of
a semi-infinite mass
a
Total load Q
r
z
z
r
r
Chapter 3 Finite Layer Analysis of Layered Soil
86
(a) Horizontal displacement
(b) Vertical displacement
Figure 3.23 Comparison of displacement between a uniform
circular load and a concentrated point load
0
5
10
15
20
25
30
35
40
45
0 0.05 0.1 0.15 0.2 0.25 0.3
u
x
E/a
2
q
z
/
a
Point Load r/a = 4
Circular Load r/a = 4
Point Load r/a = 12
Circular Load r/a = 12
Point Load r/a = 20
Circular Load r/a =20
0
5
10
15
20
25
30
35
40
45
0 0.01 0.02 0.03 0.04 0.05
u
z
E/a
2
q
z
/
a
Point Load r/a = 4
Circular Load r/a = 4
Point Load r/a = 12
Circular Load r/a = 12
Point Load r/a = 20
Circular Load r/a =20
Chapter 3 Finite Layer Analysis of Layered Soil
87
Figure 3.24 Uniform circular loading on the surface of a 2 layered soil
system
a
Total load Q
r
z
h Layer 1
Layer 2
E
1
,
1
,
G
1
= E
1
/2(1+
1
)
E
2
,
2
,
G
2
= E
2
/2(1+
2
)
Perfectly
adhesive
interface
r
Chapter 3 Finite Layer Analysis of Layered Soil
88
0.0
0.2
0.4
0.6
0.8
1.0
0 2 4 6 8 10 12
r/a
u
x
E
1
/
Q
Westmann
Finite Layer Method
0.000
0.005
0.010
0.015
0.020
0.025
0 2 4 6 8 10 12
r/a
u
z
E
1
/
Q
Westmann
Finite Layer Method
Figure 3.25 Comparison between the finite layer method and the
theoretical solution for horizontal and vertical displacements
on the surface along the x-axis.
(a) Horizontal surface displacement due to a uniform horizontal
circular load
(b) Vertical surface displacement due to a uniform horizontal
circular load
2
0
=
=
a
h
a
z

2
0
=
=
a
h
a
z

Chapter 3 Finite Layer Analysis of Layered Soil
89
Figure 3.26 Comparison between a horizontal ring load and a
horizontal circular load of the displacements at
different positions from the centre of the load.
(a) Horizontal displacement
(b) Vertical displacement
0
2
4
6
8
10
12
14
16
18
20
0 0.05 0.1 0.15 0.2 0.25 0.3
u
x
E/Q
z
/
a
Ring Load, r/a = 2
Circular Load, r/a = 2
Ring Load, r/a = 4
Circular Load, r/a = 4
Ring Load, r/a = 10
Circular Load, r/a = 10
0
2
4
6
8
10
12
14
16
18
20
0 0.01 0.02 0.03 0.04 0.05
u
z
E/Q
z
/
a
Ring Load, r/a = 2
Circular Load, r/a = 2
Ring Load, r/a = 4
Circular Load, r/a = 4
Ring Load, r/a = 10
Circular Load, r/a = 10
Chapter 3 Finite Layer Analysis of Layered Soil
90
b
a
x
Load
q/unit area
A
z
y
B
C
D
Figure 3.27 Uniform horizontal rectangular
loading on the surface of a semi-
infinite mass
0
0.02
0.04
0.06
0.08
0.1
0.12
0 1 2 3 4 5 6 7
b/a
u
z
E
/
q
Giroud (1969)
Finite Layer Method
0
0.1
0.2
0.3
0.4
0.5
0.6
0 1 2 3 4 5 6 7
b/a
u
x
E
/
q
Giroud (1969)
Finite Layer Method
(a) Horizontal displacement due to a uniform horizontal rectangular load
(b) Vertical displacement due to a uniform horizontal rectangular load
Figure 3.28 Comparison between the finite layer method and the
theoretical solution for displacements beneath the corner of a
horizontal rectangular load.
1
, 0
= =
=
b
y
a
x
a
z

1
, 0
= =
=
b
y
a
x
a
z

Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
91
CHAPTER 4
ANALYSIS OF RAFTS AND PILE GROUPS
IN LAYERED SOIL

4.1 INTRODUCTION
The analysis of a raft foundation involves the separation of the foundation system into an
isolated raft and the supporting soil and takes account of the interactions between the soil
and raft. The distribution of contact pressures acting on the interface between the raft and
the soil depends on the rigidity of the raft. For pile groups, the analysis may be performed
by separating the system into a pile group and the supporting soil. In a similar way to the
raft analysis, the full soil-structure interactions have to be taken into account.

Finite element techniques have often been used for the analysis of rafts by different
researchers such as Cheung and Zienkiewicz (1965), Cheung and Nag (1968), and Svec
and Gladwell (1973). The raft can be treated as a plate and the soil can be treated as (I) a
series of isolated springs known as a Winkler model, in which the contact pressure at any
point on the base of the raft is proportional to the deformation of the soil at that point or
(II) as an elastic half-space (or elastic continuum) in which the behaviour of the soil can
be obtained from a number of closed-form solutions. For a multilayered soil system,
Wardle and Fraser (1974), Fraser and Wardle (1976) used the finite element technique to
analyse a rectangular raft subjected to uniformly distributed loads. The finite layer
method was used by Zhang and Small (1991) for the analysis of rafts on an elastic half
space subjected to different types of loadings.

The behaviour of pile groups has been examined extensively by the use of different
techniques. The most widely used technique is the interaction method which was
introduced by Poulos (1968) to calculate the settlement between two piles and hence the
pile group. Other techniques including the boundary element method (Butterfield and
Banerjee, 1971a and b; Banerjee and Davies, 1977) and the finite element method
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
92
(Ottaviani, 1975; Valliappan, et al. 1974; Pressley and Poulos, 1986; Chow, 1987a) have
been used for settlement analysis. The boundary element method is generally more
efficient and economical to use compared with the finite element method for a large
number of piles. The finite layer method (Lee and Small, 1991a and b; Ta and Small,
1995; Southcott and Small, 1996; Zhang and Small, 1999 and 2000) was used for the
analysis of pile groups subjected to vertical or horizontal loadings.

In this chapter, the analysis of raft and pile groups through the use of the finite layer and
finite element methods will be presented. Loadings and moments in different directions
can be applied to the foundation. The soil behaviour is modelled by the use of finite layer
techniques and the raft and piles are modelled by finite element techniques.

4.2 ANALYSIS OF RAFTS
The analysis of rafts is carried out by the finite element technique in which the raft is
divided into 8 noded isoparametric rectangular elements. Each element has 40 degrees of
freedom. The raft element can be subjected to the membrane and bending actions that are
shown in Figure 4.1. The displacement due to the membrane action is considered
independent of the displacement due to the bending action, therefore it can be considered
separately. For the bending action, the displacement field for an individual element can
be described in terms of the vertical nodal displacement and the rotations about the x and
y axes. For the membrane action, the displacement field can be described in terms of the
nodal displacements in the x and y directions. The nodal numbering of the element is
shown in Figure 4.2.
4.2.1 Raft Subjected to Membrane Action
When the raft is subjected to membrane action, the displacement of an element can be
expressed in terms of the shape functions and nodal displacements as follows
u = N
e
{a
m
}
e
(4.1)
where N
e
= (N
1
, N
2
, , N
8
), are the shape functions of an element as given in Appendix
IV
a
m
= (u
1
,v
1
,u
2
, v
2
, , u
8
, v
8
)
T
is the nodal displacement vector for an element
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
93
The strain at any point in the element can be expressed in terms of the nodal
displacements in the form of

m
= [B
m
]
e
{a
m
}
e
(4.2)
where
( )
( )
( )

=
i
i
i
i
i
xy
i
y
i x
i
x
v
y
u
y
v
x
u

(4.3)
and

=
x
N
y
N
y
N
x
N
B
i i
i
i
m
0
0
, for i = 1, 2, , 8 (4.4)
The relationship between the nodal forces, F, and nodal displacements, a, for an element
can be expressed as
{F
m
}
e
= [K
m
]
e
{a
m
}
e
(4.5)
where F
m
= (P
x1
, P
y1
, P
x2
, P
y2
, , P
x8
, P
y8
) is the nodal force vector
K
m
= element stiffness matrix in the form of
dA B D B t K
e m e m
A
T
e m m
] [ ] [ ] [

= (4.6)
where t = thickness of the element
A = area of the element
D
m
= matrix related to the material properties
For an isotropic material, the matrix D can be written as

=
2
) 1 (
0 0
0 1
0 1
1
2

Et
D
m
(4.7)
where E and are Youngs modulus and Poissons ratio of the element respectively.
The stress-strain relationship can be written as
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
94
= D
m

m
(4.8)
Substituting equation (4.2) into equation (4.3) gives
= D
m
B
m
a
e
(4.9)
where = (N
x1
, N
y1
, N
xy1
, N
x2
, N
y2
, N
xy2
, , N
x8
, N
y8
, N
xy8
) is the membrane force vector.
4.2.2 Raft Subjected to Bending Action
When the raft is subjected to bending action, the vertical displacement of an element
can be expressed in terms of the shape functions and nodal displacements as follows
= N
e
{a
b
}
e
(4.10)
where a
b
= (w
1
,
x1
,
y1
, w
2
,
x2
,
y2
, , w
8
,
x8
,
y8
)
T
is the nodal displacement vector
N
e
is the shape function of the element as defined in Section 4.2.1.
The terms
x
and
y
are defined as the rotations of the nodes about the x and y axes
respectively in the form of
( )
( )

i
i
i
y
i x
y
w
x
w

, for i = 1, 2, , 8 (4.11)
The curvatures at a point on the element can be expressed in terms of the nodal
displacement in the form of

b
= [B
b
]
e
{a
b
}
e
(4.12)
where

=
i
i
i
i
y x
w
y
w
x
w
2
2
2
2
2
2
(4.13a)
and
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
95

=
i
i
i
i
i i
i
i
b
N
y
N
N
x
N
x
N
y
N
y
N
x
N
B
0
0
0
0 0
0 0
, for i = 1, 2, , 8 (4.13b)

The expression relating the nodal forces, F, and the nodal displacements, a, can be
written as
{F
b
}
e
= [K
b
]
e
{a
b
}
e
(4.14)
where F
b
= (P
z1
, M
x1
, M
y1
, P
z2
, M
x2
, M
y2
, , P
z8
, M
x8
, M
y8
)
T
is the nodal force vector
K
b
= element stiffness matrix in the form of
dA B D B K
e b e b
A
T
e b b
] [ ] [ ] [

= (4.15)
For an isotropic material, the matrix D
b
can be written as

=
2
2
2
3
) 1 ( 5
0 0 0 0
0
) 1 ( 5
0 0 0
0 0
2
1
0 0
0 0 0 1
0 0 0 1
) 1 ( 12
t
t
Et
D
b

(4.16)
where E, and t are defined as in Section 4.2.1.

For equation (4.11), the stress-strain relationship for bending action can be written as
= D
b

b

= D
b
B
b
a
e
(4.17)

where = (M
x1
, M
y1
, M
xy1
, S
x1
, S
y1
, M
x2
, M
y2
, M
xy2
, S
x2
, S
y2
, M
x8
, M
y8
, M
xy8
, S
x8
, S
y8
)
T
is
the vector of moments and shears per unit length for bending action.
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
96
4.2.3 Combination of Membrane and Bending Actions
The nodal forces for an element subjected to combined actions can be obtained from the
expression
{F}
e
= [K]
e
{a}
e
(4.18)
where {F}
e
= (P
x1
, P
y1
, P
z1
, M
x1
, M
y1
, P
x2
, P
y2
, P
z2
, M
x2
, M
y2
, , P
x8
, P
y8
, P
z8
, M
x8
, M
y8
)
T

is the nodal force vector
{a}
e
= (u
1
, v
1
, w
1
,
x1
,
y1
, u
2
, v
2
, w
2
,
x2
,
y2
, , u
8
, v
8
, w
8
,
x8
,
y8
)
T
is the nodal
displacement vector
[K]
e
= [K
m
]
e
+ [K
b
]
e
is the element stiffness matrix

For a raft with n elements, the total stiffness matrix [K] can be obtained by the sum of n
element stiffness matrices [K]
e
, and the relationship between the nodal forces and nodal
displacements for the whole raft can be written as
{F} = [K] {} (4.19)
where { } { } ( )

=
=
n
i
i
e
F F
1

[ ] [ ] ( )

=
=
n
i
i
e
K K
1

{ } { } ( )

=
=
n
i
i
e
a
1


The nodal displacement of the raft can be obtained simply by applying the appropriate
boundary conditions and solving equation (4.19).

4.2.4 Rafts on Layered Soils
4.2.4.1 Influence Matrix for the Raft
The analysis of a raft founded on a layered soil has to be separated into an isolated raft
and a layered soil. The raft is assumed to be a plate and is divided into a number of
rectangular elements as shown in Figure 4.3. Each element has 8 nodes and forty degrees
of freedom. The soil can be divided into horizontal layers if it is layered and the surface
of the soil is divided into rectangular regions corresponding to the raft elements. The
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
97
contact and shear stresses between the raft and the soil are assumed to be uniform
rectangular blocks of pressure acting horizontally or vertically over each raft element and
each soil region. The actual displacement at the centre of each raft element can be
expressed as
{
r
}

= [I
r
] P
r
+ a
x
+ b
y
+ c
z
+ d
x
+ e
y
+ f
z
+
r0
(4.20)
where
[I
r
] = influence matrix of the pinned raft
P
r
= contact stresses on the raft elements

r0
= vector of displacements at the centre of each raft element for the pinned
raft under the applied load

x
= horizontal translation of the raft in the x-direction

y
= horizontal translation of the raft in the y-direction

z
= vertical translation of the raft in the z-direction

x
=

x-rotation of the raft about the first pinned point

y
= y-rotation of the raft about the first pinned point

z
= z-rotation of the raft about the first pinned point
a = (1, 0, 0, 1, 0, 0, )
T

b = (0, 1, 0, 0, 1, 0, )
T

c = (0, 0, 1, 0, 0, 1, )
T

d = (0, 0, x
1
x
p
, 0, 0, x
2
x
p
, , 0, 0, x
n
x
p
)
T
e = (0, 0, y
1
y
p
, 0, 0, y
2
y
p
, , 0, 0, y
n
y
p
)
T
f = (y
1
y
p
, x
1
x
p
, 0, y
2
y
p
, x
2
x
p
, 0, , y
n
y
p
, x
n
x
p
, 0)
T
x
p
, y
p
= co-ordinates of the pin
x
i
, y
i
= co-ordinates of the centres of each raft element

The influence matrix of the raft [I
r
] and the vector of the displacements at the centres of
the raft elements for the pinned raft under applied load {
r0
} are obtained by restraining
the raft from rigid body rotations and translations. As shown in Figure 4.4, the raft is
pinned at two points in which the first point (node 1) is completely fixed against the rigid
body rotations and translations and the second point (last node on the raft) is fixed in the
x-direction against the rotation of the raft about the z-axis. The entries of the raft
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
98
influence matrix [I
r
] are formed by applying a unit uniform pressure in the x-, y- and z-
directions on each raft element of the pinned raft in turn to compute the displacements at
the centre of all elements (x-, y- and z-directions). The vector of the displacements at the
centres of the raft elements for the pinned raft under the applied load {
r0
} is obtained by
using the plate theory discussed in Section 4.2.1-4.2.3.

4.2.4.2 Influence Matrix for the Soil
When the load is applied to the raft, the load is transferred from the raft to the soil
through the contact stresses. The soil may consist of several different material layers and
can be analysed by the finite layer method described in Chapter 3. The entries of the soil
influence matrix [I
s
] are formed by applying a unit uniform block of pressure in the x-, y-,
and z-directions to each rectangular element on the soil in turn and calculating the soil
displacements at the centres of all elements at the interface of the raft and soil.

The soil influence matrix can be generated as follows:
(1) Apply a uniform pressure in the x-direction to the first element and calculate the soil
displacements of each element in the x-, y- and z-directions. These displacements
will form the first column of [I
s
].
(2) As for the step 1, apply a uniform pressure in the y-direction to the first element to
obtain the second column of [I
s
].
(3) Apply a uniform pressure in the z-direction to the first element to obtain the third
column of [I
s
].
(4) Repeat the above steps by applying uniform pressure to each element in turn to
construct the influence matrix [I
s
].

Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
99
The displacement of the soil in all directions due to the contact stresses at all rectangular
elements can be obtained by superposition of solutions resulting in

zn
yn
xn
z
y
x
z
y
x
zznn zynn zxnn zzn zyn zxn zzn zyn zxn
yznn yynn yxnn yzn yyn yxn yzn yyn yxn
xznn xynn xxnn xzn xyn xxn xzn xyn xxn
n zz n zy n zx zz zy zx zz zy zx
n yz n yy n yx yz yy yx yz yy yx
n xz n xy n xx xz xy xx xz xy xx
n zz n zy n zx zz zy zx zz zy zx
n yz n yy n yx yz yy yx yz yy yx
n xz n xy n xx xz xy xx xz xy xx
zn
yn
xn
z
y
x
z
y
x
p
p
p
p
p
p
p
p
p
I I I I I I I I I
I I I I I I I I I
I I I I I I I I I
I I I I I I I I I
I I I I I I I I I
I I I I I I I I I
I I I I I I I I I
I I I I I I I I I
I I I I I I I I I
M
L
M O M
L
M
2
2
2
1
1
1
2 2 2 1 1 1
2 2 2 1 1 1
2 2 2 1 1 1
2 2 2 22 22 22 21 21 21
2 2 2 22 22 22 21 21 21
2 2 2 22 22 22 21 21 21
1 1 1 12 12 12 11 11 11
1 1 1 12 12 12 11 11 11
1 1 1 12 12 12 11 11 11
2
2
2
1
1
1

(4.21)

The above equation can be written as

s
= [I
s
] P
s
(4.22)
where
s
= (
x1
,
y1
,
z1
,
x2
,
y2
,
z2
, ,
xn
,
yn
,
zn
) is the vector of displacements
at the centres of the elements in the x-, y- and z-directions respectively.
P
s
= (p
x1
, p
y1
, p
z1
, p
x2
, p
y2
, p
z2
, , p
xn
, p
yn
, p
zn
) is the force vector of pressure on
the elements in the x-, y- and z-directions respectively.

4.2.4.3 Analysis of Foundations
Assuming lift-off does not occur between the raft and soil, the compatibility of the
displacements requires that the displacements of the raft and soil are equal, i.e.

s
=
r
(4.23)
The equilibrium of forces at the raft and soil interface are equal but opposite in sign, i.e.
P = P
s
= -P
r
(4.24)
Combining equations (4.20) and (4.22) will give
([I
r
]+[I
s
])P a
x
b
y
c
z
d
x
e
y
f
z
=
r0
(4.25)



Considering equilibrium of the raft, the total applied loads P
x
, P
y
and P
z
must be equal to
the total reaction forces due to the contact pressures, i.e.
a
T
P = P
x
(4.26a)
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
100
b
T
P = P
y
(4.26b)
c
T
P = P
z
(4.26c)
where a
T
= (A
1
, 0, 0, A
2
, 0, 0, , A
n
, 0, 0)
b
T
= (0, A
1
, 0, 0, A
2
, 0, , 0, A
n
, 0)
c
T
= (0, 0, A
1
, 0, 0, A
2
, , 0, 0, A
n
)

The applied moments M
x
, M
y
and M
z
about the x, y and z axes must be equal to the
moments due to the contact pressures, i.e.
d
T
P = M
x
(4.26d)
e
T
P = M
y
(4.26e)
f
T
P = M
z
(4.26f)
where d
T
= (0, 0, A
1
(x
1
-x
p
), 0, 0, A
2
(x
2
-x
p
), , 0, 0, A
n
(x
n
-x
p
))
e
T
= (0, 0, A
1
(y
1
-y
p
), 0, 0, A
2
(y
2
-y
p
), , 0, 0, A
n
(y
n
-y
p
))
f
T
= (A
1
(y
1
-y
p)
, -A
1
(x
1
-x
p
), 0, A
2
(y
2
-y
p
), -A
2
(x
2
-x
p
), 0, , A
n
(y
n
-y
p
), -A
n
(x
n
-x
p
), 0)
A
i
is the area of element i in the raft
x
i
, y
i
, x
p
and y
p
are defined in equation (4.20).

Combining equations (4.25) and (4.26a to f) gives
[ ] [ ]

+
z
y
x
z
y
x
r
z
y
x
z
y
x
T
T
T
T
T
T
r s
M
M
M
P
P
P
P
f
e
d
c
b
a
f e d c b a I I
0
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0

(4.27)
where a, b, c, d, e and f are defined in equation (4.20).

Equation (4.27) can then be solved for the contact pressures P between the raft and soil,
and the rigid body translations and rotations relative to the first pinned node.

Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
101
4.3 ANALYSIS OF PILE GROUPS
Like the raft foundation, the analysis of pile groups (i.e. piles with no raft) is carried out
by separating the foundation into a pile group and the layered soil system as shown in
Figure 4.5. The finite element technique is used for the analysis of the individual piles in
which the piles are assumed to have a solid circular cross-section with diameter d and
length l. The piles are modelled as beam elements and each pile is divided into n
elements. The soil between the pile head and pile base can be divided into layers
corresponding to the number of elements in the pile. The soil beneath the pile base can be
divided into one or more layers depending on the soil profile.

For piles subjected to general loadings, the vertical component of the load is resisted by
shear forces developed along the pile shaft and normal contact stresses at the pile bases;
the horizontal component of the load is resisted by lateral earth pressures along the pile
shafts and shear stresses at the pile bases.

4.3.1 Stiffness Matrix of Pile Group
Consider an individual pile element in the pile group. The element is treated as a beam
with 2 nodes and five degrees of freedom (u
x
, u
y
, u
z
,
x
,
y
) at each node. The relationship
between the nodal forces and nodal displacements can be expressed in the form
F
i
= [k]
i

i
(4.28)
where F
i
= (p
xi
, p
yi
, p
zi
, M
xi
, M
yi
, p
x(i+1)
, p
y(i+1)
, p
z(i+1)
, M
x(i+1)
, M
y(i+1)
)
T
is the vector of
forces and moments at the ends of an element

i
= (
xi
,
yi
,
zi
,
xi
,
yi
,
x(i+1)
,
y(i+1)
,
z(i+1)
,
x(i+1)
,
y(i+1)
)
T
is the vector of
displacements and rotations at the ends of an element
and [k]
i
is the stiffness of the pile element i in the form of
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
102
[ ]

=
i
x
i
x
i
x
i
x
i
y
i
y
i
y
i
y
i i
i
x
i
x
i
x
i
x
i
y
i
y
i
y
i
y
i
x
i
x
i
x
i
x
i
y
i
y
i
y
i
y
i i
i
x
i
x
i
x
i
x
i
y
i
y
i
y
i
y
i
L
EI
L
EI
L
EI
L
EI
L
EI
L
EI
L
EI
L
EI
L
EA
L
EA
L
EI
L
EI
L
EI
L
EI
L
EI
L
EI
L
EI
L
EI
L
EI
L
EI
L
EI
L
EI
L
EI
L
EI
L
EI
L
EI
L
EA
L
EA
L
EI
L
EI
L
EI
L
EI
L
EI
L
EI
L
EI
L
EI
k
4
0 0
6
0
2
0 0
6
0
0
4
0 0
6
0
2
0 0
6
0 0 0 0 0 0 0 0
6
0 0
12
0
6
0 0
12
0
0
6
0 0
12
0
6
0 0
12
2
0 0
6
0
4
0 0
6
0
0
2
0 0
6
0
4
0 0
6
0 0 0 0 0 0 0 0
6
0 0
12
0
6
0 0
12
0
0
6
0 0
12
0
6
0 0
12
2 2
2 2
2 3 2 3
2 3 2 3
2 2
2 2
2 3 2 3
2 3 2 3

where E = Youngs modulus of the pile
I
x
, I
y
= second moment of area of the pile about the x- and y-axes
A = cross-sectional area of the pile
L
i
= length of element i

Suppose the pile group has m piles and the pile is divided into n elements, the stiffness
matrix of a pile group can be obtained by first forming the stiffness matrix for a single
pile then adding the single pile stiffness matrices together to form the total stiffness
matrix [K
p
] for the pile group. The stiffness matrix for a single pile can be formed by the
sum of n individual element stiffness matrices and the total stiffness matrix can be
formed by the sum of m single pile stiffness matrices. The stiffness matrix for the pile
group can be expressed as
[ ]

=
=
n m
i
i p
k K
1
(4.30)
(4.29)
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
103
If the pile group is subjected to external loads Q
x
, Q
y
and Q
z
and each of the piles will
carry the loads q
xi
, q
yi
and q
zi
as shown in Figure 4.5, the load-displacement relationship
of the pile group can be expressed as
K
p

p
= Q + P
p
(4.31)
where
p
= (u
x1
, u
y1
, u
z1
,
x1
,
y1
, u
x2
, u
y2
, u
z2
,
x2
,
y2
,

,

u
x(mxn)
, u
y(mxn)
, u
z(mxn)
,
x(mxn)
,

y(mxn)
)
T
is the vector of nodal displacements
Q = (q
x1
, q
y1
, q
z1
, 0, 0, 0, 0, 0, 0, 0, , q
x(n+1)
, q
y(n+1)
, q
z(n+1)
, 0, 0, , 0, 0, 0, 0, 0,
q
x(2n+1)
, q
y(2n+1)
, q
z(2n+1)
, 0, 0, , 0, 0, 0, 0, 0 )
T
is the vector of external loads
and moments applied to the pile group
P
p
= (p
px1
, p
py1
, p
pz1
, M
x1
, M
y1
, p
px2
, p
py2
, p
pz2
, M
x2
, M
y2
, , p
px(mxn)
, p
py(mxn)
, p
pz(mxn)
,
M
x(mxn)
, M
y(mxn)
) is the vector of nodal forces and moments acting at the nodes
along the pile.

4.3.2 Stiffness Matrix for Layered Soil
In the layered soil system, the soil is subjected to the forces transferred from the pile.
These forces are treated as a series of ring loads at the soil nodes and a uniform circular
load on the soil beneath the pile base. The displacements of the soil due to these ring
loads and circular loads will form the columns of the influence matrix of the soil [I
s
] and
can be obtained by the finite layer method. The influence matrix is formed by applying a
unit ring load to the soil nodes or a uniform circular load at the pile base in turn.

Consider a single pile divided into n elements so that the pile will have (n+1) nodes. The
node at the pile base is represented by b. The soil influence matrix will be of size 3(n+1)
by 3(n+1) and can be constructed as in Section 4.2.4.2 by replacing the uniform pressure
with a unit ring load or a uniform circular load and applying these ring loads at the soil
nodes.

Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
104
The displacement of the soil at each soil node can be obtained and expressed as

szb
syb
sxb
sz
sy
sx
s
zzbb
s
zybb
s
zxbb
s
zzb
s
zyb
s
zxb
s
yzbb
s
yybb
s
yxbb
s
yzb
s
yyb
s
yxb
s
xzbb
s
xybb
s
xxbb
s
xzb
s
xyb
s
xxb
s
b zz
s
b zy
s
b zx
s
zz
s
zy
s
zx
s
b yz
s
b yy
s
b yx
s
yz
s
yy
s
yx
s
b xz
s
b xy
s
b xx
s
xz
s
xy
s
xx
s
zn
s
yn
s
xn
s
z
s
y
s
x
p
p
p
p
p
p
I I I I I I
I I I I I I
I I I I I I
I I I I I I
I I I I I I
I I I I I I
M
M
L L
M O M
M O M
L L
M
1
1
1
1 1 1
1 1 1
1 1 1
1 1 1 11 11 11
1 1 1 11 11 11
1 1 1 11 11 11
1
1
1

(4.32)

The influence matrix of the soil for the pile group can be obtained since the columns of
the matrix are the deflections at all soil nodes due to the application of the unit ring loads
in the x-, y-, z-directions at each node in turn.

The soil-displacement relationship can be written as

s
= [I
s
] P
s
(4.33)
where
s
= (
x1
,
y1
,
z1
,
x2
,
y2
,
z2
, ,
xn
,
yn
,
zn
) is the vector of displacements
at the nodes
P
s
= (p
sx1
, p
sy1
, p
sz1
, p
sx2
, p
sy2
, p
sz2
, , p
sxn
, p
syn
, p
szn
) is the force vector at the
nodes
The soil stiffness matrix [K
s
] of a layered soil can then be calculated by inverting the
influence matrix of the soil which can be expressed as
[K
s
] = [I
s
]
-1
(4.34)
Equation (4.33) can be re-written in the stiffness matrix form as
K
s

s
= P
s
(4.35)
4.3.3 Pile Groups in Layered Soil
Combining equations (4.31) and (4.35) gives the expression for the load-displacement
relationship of a pile group embedded in a layered soil,
K
p

p
+ K
s

s
= Q + P
p
+ P
s
(4.36)
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
105
Assuming that slip does not occur along the pile-soil interfaces, the compatibility of the
displacements requires that the displacements of the pile and soil are equal, i.e.

=
p
=
s
(4.37)
The equilibrium of forces at the pile and soil interface are equal but opposite in sign, i.e.
P
s
= -P
p
(4.38)
Combining equations (4.36) to (4.38) gives
(K
p
+ K
s
) = Q (4.39)
If the applied load on the pile group Q is known, the displacements at the nodes along the
piles can be solved by equation (4.39).

4.4 NUMERICAL EXAMPLES
4.4.1 Examples for Rafts
Example 1 Raft on isotropic homogeneous half space
This example involves a square raft of size 6m x 6m and a thickness of 0.8m resting on a
homogeneous soil having a modulus (E
s
) of 1.0MPa and a Poissons ratio of 0.15. The
raft is subjected to a concentrated load of 1000kN at the centre and its modulus and
Poissons ratio are 40MPa and 0.15 respectively.

Figure 4.6 shows the comparisons of settlements at the centreline of the raft obtained by
the present finite layer (FL) method with those presented by Cheung and Nag (1968) and
Mandal and Ghosh (1999). Except for the settlement where the concentrated load was
applied, the settlements elsewhere along the centreline of the raft obtained by the
different methods may be seen to be in excellent agreement. The higher value of the
central settlement obtained by the present method could be due to the refinement of the
mesh for the raft in the vicinity of the point load. The mesh used for the results computed
here is shown in Figure 4.7.

Example 2 Raft on layered materials Savings Bank Building, Adelaide
The Savings Bank Building is located in the city of Adelaide in South Australia. The
building was founded on Pleistocene Hindmarsh Clay underlain by Hallet Cove
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
106
Sandstone. A rectangular raft of size 33.5m x 39.5m with a thickness of 0.9m was used as
the foundation of the building. The raft was subjected to a uniformly distributed load of
134kPa and the modulus and Poissons ratio of the raft were taken as 25,000MPa and
0.15 respectively.

Kay and Cavangaro (1983) carried out settlement computation using different methods
by assuming the clay layer consisted of (i) a 2m layer with a modulus of 44MPa, (ii) an
8m thick layer with a modulus of 60MPa underlain by a layer of Hallet Cove sandstone.
For all soil layers the Poissons ratios are taken as 0.2. Analyses have been carried out by
considering the modulus of the sandstone as 500MPa or as a very large value.

In the present analysis, the raft was divided into 1360 elements and the depth of soil in
consideration was about 80m from the bottom of the raft. The modulus of the sandstone
was taken as 500MPa and 10000MPa in the analyses. The measured and computed
settlements are summarised in Table 4.1 and presented in Figure 4.8. When the modulus
was taken as 500MPa, except for the conventional 1D approach, the computed
settlements at the centre of the raft from the different approaches were in good agreement
and ranged from 26mm to 28mm. However, they all over-predicted the actual settlement.
If the modulus was taken as a large value (~ 10000MPa), the settlements at the centre
reduced significantly to around 20mm for all approaches. The computed differential
settlements ranged from 6 12 mm. The discrepancies between computed and measured
settlements could be due to the soil parameters used in the analyses which may be
different from the actual soil conditions or could be due to the stiffness of the structure
which was not included in the analysis.
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
107

Methods
Settlements at
centre (mm)
Differential
settlements (mm)
Conventional: 1D approach 22 16
Hookes Law approach
Hallet Cove Sandstone E = large value 20 -
Hallet Cove Sandstone E = 500MPa 26 -
Kay and Cavangaro (1983)
Hallet Cove Sandstone E = large value 20 10
Hallet Cove Sandstone E = 500MPa 26 12
Mandal and Ghosh (1999)
FE & BE approach (E = 500MPa)
28 8
Present FL
Hallet Cove Sandstone E = 10000MPa 20 6
Hallet Cove Sandstone E = 500MPa 27 9
Measurement 16 -18 7-11
Table 4.1 Comparison of computed and measured settlements

Example 3 Raft on layered materials with a rigid base
In this example, the raft is resting on layered material underlain by a rigid base. The raft
is square in plan with a size of 10 x 10m and a thickness of 0.5m. A uniformly distributed
load of 100kPa is applied to the raft and the modulus and Poissons ratio of the raft are
taken as 15,000MPa and 0.2 respectively. The soil consists of four layers underlain by a
rigid base. Each layer has a thickness of 10m with the moduli and Poissons ratios as
given in Table 4.2.

For settlement computation, the depth of soil being considered was taken as five times
the width of the raft as there is little influence on settlements from the soil below that
level (Mandal and Ghosh, 1999). Thus, the depth of soil for this example was taken as
50m from the bottom of the raft.

Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
108
Layer Modulus, E
s
(MPa) Poissons ratio,
I 100 0.3
II 80 0.3
III 60 0.3
IV 100 0.3
V (Rigid base) 10000 0.3
Table 4.2 Parameters for different soil layers

Analyses have been carried out by the present and finite element methods and compared
with the published results. In the finite element analysis, one quarter of the raft was
analysed and the finite element mesh is shown in Figure 4.9. Infinite elements were used
along the edges of the mesh so as to simulate the infinite lateral boundaries that are
implied by the finite layer analysis. In the present finite layer analysis, the raft was
divided into 400 elements and the soil was divided into 10 layers. Two analyses have
been carried out by considering the multi-layered soil medium (I) with different moduli
as specified in Table 4.2, (II) as a homogeneous layer with the equivalent modulus E
s
=
81.9MPa (Fraser and Wardle, 1976).

Contours of settlements computed by the present method and the finite element method
for analysis I are shown in Figures 4.10a and b respectively. The maximum settlements at
the centre of the raft were 9.5mm and 9.2mm from the present and finite element
analyses respectively. Figure 4.11 presents the results obtained from the different
methods clearly showing that by assuming the multi-layered soil medium to be a
homogeneous layer, the settlement of the raft would be overpredicted by about 10 %.
Table 4.3 summarises the comparisons of the calculated settlements with published
results.

Example 4 Raft subjected to a uniform horizontal load
A square raft of size 10 x 10m with a thickness of 0.5m is embedded in a homogeneous
soil having a depth of 50m below the bottom of the raft. The modulus and Poissons ratio
of the soil are 10MPa and 0.35 respectively. The raft, with a modulus and Poissons ratio
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
109
of 15000MPa and 0.2 respectively, is subjected to a uniform horizontal pressure of
100kPa.

Methods
Settlements at centre
(mm)
Differential settlements
(mm)
Fraser and Wardle
(approximate method)
10.7 2.9
Fraser and Wardle (FE
method equivalent modulus)
11.4 2.7
Mandal and Ghosh (FE-BE
method)
10.6 2.8
Present FL
(I) Multi-layered 9.5 2.5
(II) Homogeneous 10.9 2.9
Finite element 9.2 2.2
Table 4.3 Comparisons of settlements at the centre and differential settlements

This example was analysed by the present and finite element methods for comparison.
For the present method, the whole raft was analysed, while one quarter of the raft was
analysed by the finite element method. In the finite element analysis, appropriate
boundary conditions have to be applied to produce anti-symmetrical results about the y-
axis. Figure 4.12 illustrates the typical deformed shape of the raft in the x-y plane. Along
the x-axis the raft is fixed against movement and rotation in the y-direction while along
the y-axis, the raft is fixed against movements in the y- and z-directions and rotation in
the y-direction.

Deformed shapes and contours of settlement for the raft obtained from the present and
finite element methods are shown in Figures 4.13 and 4.14. Comparisons of displacement
in the horizontal x- and vertical z-directions along the centre of the raft between the two
methods are shown in Figure 4.15. When the raft is subjected to a horizontal load, the
displacement in the direction of the load (x-direction) is about ten times larger than the
displacements in the other two directions (y- and z-directions). The horizontal x-
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
110
displacements computed from the present method were slightly larger than from the finite
element method while the vertical displacements are in reasonably close agreement.
Contours of bending moments, M
xx
, computed from the present method are shown in
Figure 4.16 and comparisons between the two methods show good agreement as
indicated in Figure 4.17.

Example 5 Raft subjected to a uniform vertical load
The raft in the above example is subjected to a uniform vertical pressure of 100kPa.
When a uniform vertical pressure is applied to the raft, the raft deforms vertically
downwards and the displacement of the raft in the horizontal directions (x- and y-
directions) are relatively small compared with the vertical displacement. Comparisons
between the present and finite element methods of the vertical displacement and bending
moments are shown in Figures 4.18a and b and the results are in good agreement.

4.4.2 Examples for Pile Groups
Example 1 3x3 pile group subjected to a vertical load
In this example, the behaviour of a 3x3 pile group embedded in a homogeneous soil was
examined by the use of DEFPIG (Poulos, 1990) and the present method. The piles have a
length of L and diameter of d. The piles have a spacing of 8d and were embedded in a soil
layer with a depth of 50m. Figure 4.19 shows the configuration of the pile group and soil
profile. A total load of P
v
was applied vertically to the pile group with each of the piles
carrying a load of P
v
/9. The modulus of the piles and soil are assumed to be E
p
and E
s

respectively and the pile to soil stiffness ratio E
p
/ E
s
was taken to be 200.

Comparisons of the results are shown in terms of dimensionless parameters: (i) vertical
displacement influence factor I
v
which is defined as
v
P
d E
I
s
v

= (4.40)
and (ii) axial forces influence factor C
a
defined as
v
a
P
A
C = (4.41)
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
111
where = vertical displacement of the pile
E
s
= modulus of the soil
d = diameter of the pile
P
v
= total load on the pile group
A = axial force in the pile at any level
Figures 4.20a and b show the displacements and axial forces respectively along the piles
at different locations respectively. The results obtained from the present method are in
good agreement with the DEFPIG results.

Example 2 3x3 pile group subjected to a horizontal load
In the above example, the load P
H
was applied horizontally to each of the pile heads in
the group as shown in Figure 4.19. As for the above example, comparisons of the results
are shown in terms of the (i) horizontal displacement influence factor, I
H
,

defined as
H
s
H
P
d uE
I = (4.42)
(ii) shear force influence factor C
s
defined as
H
s
P
S
C = (4.43)
(iii) bending moment influence factor C
b
defined as

d P
B
C
H
b
= (4.44)
where u = horizontal displacement of the pile
S = shear force along the pile
B = bending moment along the pile

The present solution was compared with the results obtained from DEFPIG and the finite
element analysis. In the finite element analysis, the piles were modelled by square
elements with the same cross-sectional area as the circular piles. The horizontal
displacements and shear forces along the pile at different locations are shown in Figures
4.21 and 4.22 respectively. In the present method, the shear forces were calculated by the
sum of the horizontal ring loads acting at the nodes, while in the finite element analysis,
the shear forces were calculated from the shear stresses acting on each element, and in
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
112
DEFPIG, the shear forces were calculated from the pressure acting on the face of each
element (i.e. shear force = pressure x projected area of the pile element). The present
solutions were in good agreement with the DEFPIG and finite element solutions except at
the pile head where the displacements and shear forces given by DEFPIG were slightly
larger. Figure 4.23 shows that the bending moments along the pile obtained by the
present method were in also in good agreement with the other solutions.

Example 3 Square raft supported by a 2x2 pile group
Kitiyodom and Matsumoto (2003) conducted analyses of a square rigid raft supported by
a 2x2 pile group as shown in Figure 4.24a. A vertical (P
v
) or horizontal (P
H
) load was
applied to the foundation. Five cases with different soil profiles as shown in Figure 4.24b
were analysed by a computer program PRAB and the finite element method. For case 1,
the foundation is embedded in a homogeneous soil, for cases 2 to 5, the soil is non-
homogeneous. In the computer program PRAB, a hybrid model was used in which the
raft was modelled by finite elements, the soil was modelled as springs and the
interactions between structural elements were determined by Mindlins solutions. In the
finite element analysis, the foundation and soil were discretised into eight noded
hexahedron solid elements and modelled as linear elastic materials. The circular pile was
modelled by an octagon inscribed in the circle. In the analysis, a gap of 10mm was
maintained between the raft base and the soil surface to ensure the applied load was
resisted by the pile group only.

The modulus of the raft was calculated by the raft-soil stiffness ratio, K
rs
(Brown, 1975),
which is defined as
4
2 3
3
) 1 ( 4
r s
s r r r
rs
L E
t B E
K

= (4.45)
where E
r
= modulus of the raft
B
r
= width of the raft
L
r
= length of the raft
t
r
= thickness of the raft

s
= Poissons ratio of the soil
E
s
= modulus of the soil
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
113
For a rigid raft, K
rs
= 10 and if it is assumed that E
s
= 25MPa, then the modulus of the raft,
E
r
was calculated to be 23970MPa.

For vertical loading, vertical displacement and axial force along the pile were calculated
and for lateral loading, horizontal displacement, shear force and bending moment along
the pile were calculated. The results are presented in terms of the dimensionless
parameters as given in equations (4.40) to (4.44).

Figures 4.25 to 4.29 show the comparison of results computed by different methods for
the different cases. For all five cases, the present method tends to overestimate the
vertical displacement of the pile group. Except for the computation of vertical
displacements, there is reasonably good agreement between the results from the present
method and those from the program PRAB and the finite element method.

Under vertical loading, the vertical displacement influence factors I
v
along the pile for
case 1 are larger than in the other cases. As the bottom layers become stiffer, the
influence factors I
v
will become smaller. The piles are displacing fairly uniformly down
to the pile base. The axial force factors C
a
along the pile decrease linearly with depth for
case 1, when the piles are embedded in non-homogeneous soil, the axial forces will
decrease in a non-linear manner depending on the stiffness of the soil as shown in Figures
4.25(b) to 4.29(b).

Under horizontal loading, the horizontal displacement influence factors I
H
for case 1 are
larger than the other cases as for the vertical loading. As the top layer becomes stiffer, the
horizontal displacement reduced significantly. Unlike the vertical displacement along the
pile, large horizontal displacements are observed at the pile top and then decrease along
the pile. For all cases presented, the first 5m of the pile from the pile head is subjected to
large horizontal displacement and below that depth the pile displaces almost uniformly.

The bending moments along the pile can be worked out from the second derivatives of
the pile displacements, as for the horizontal displacements, the bending moments for case
1 are higher than those for the other cases. The bending moment at the pile head reduces
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
114
as the top layer gets stiffer. The moments approach zero at a depth of 5m below the pile
head which implies that there is no curvature in the piles beyond this level. The shear
forces along the piles can be worked out from the first derivatives of the bending
moments, therefore, the shear forces are also dependent on the displacements of the pile.
For piles subjected to horizontal loadings, the stiffness of the soil at the top layer has
significant effect on the behaviour of the pile group.

Example 4 Small pile group in an infinite layer
In this example, pile groups embedded in a homogeneous soil of infinite depth were
examined. Configurations of the pile groups are shown in Figure 4.30. The spacings
between the piles were taken as 5 x the radius of a pile (r) and the Poissons ratio of the
soil was taken as 0.3. The pile groups were assumed to be attached to flexible pile caps
such that the load carried by each of the piles in the pile group will be identical.

Results obtained by the present method were compared with those by (i) Guo and
Randolph (1999) using the program GASGROUP which was based on the superposition
principle together with a closed form expression for the interaction factors, (ii)
Butterfield and Banerjee (1971a) using the boundary element approach. The results are
shown in Figure 4.31. The responses of the pile groups were described in terms of
dimensionless parameters pile head stiffness defined as P/Gr in which P is the load
acting on the pile head, G is the shear modulus of the soil, r is the radius of the pile and
is the settlement at the pile head.

Figures 4.31a and b show the variation of pile head stiffness P/Gr of the pile groups
with different pile length to radius ratios, L/r, for the value of of 6000 and
respectively where is defined as pile-soil relative stiffness (= E
p
/G where E
p
is the
Youngs modulus of a pile and G is the shear modulus of the soil). The results obtained
by the present method show good agreement with the published results. The pile head
stiffness increases with the L/r ratio but decreases with the number of piles in the group.
The pile soil relative stiffness has a considerable effect on the pile head stiffness at L/r
40.
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
115
4.5 CONCLUSIONS
In this chapter, the analyses of rafts and pile groups subjected to horizontal or vertical
loadings have been discussed thoroughly. The analyses are carried out by the
combination of finite element and finite layer methods. The application of the present
approach to the analyses of foundations has been demonstrated through the examples
presented above. The following conclusions can be reached:

(1) The present solutions for settlements, shear forces and bending moments are in
good agreement with the solutions of other existing methods such as the finite
element method and the boundary element method.
(2) For a raft subjected to a concentrated load, the finite element mesh where the load
is applied has to be refined to obtain more accurate solutions.
(3) For a raft under horizontal loading, the displacement of the raft in the direction of
the applied load is relatively large compared with the displacements in the other
directions.
(4) For a vertically loaded square pile group with loads of equal magnitude applied to
the pile heads, the pile at the centre will have the maximum displacements, the
pile at the corner will have the minimum displacement, and the edge pile will
have an intermediate displacement.
(5) For a horizontally loaded square pile group, the pile at the centre will have the
maximum displacements as for the vertically loaded pile group however the
maximum bending moments are carried by the piles at the corners.
(6) For vertically loaded pile groups embedded in non-homogeneous soil media, the
load carried by the pile will decrease in a non-linear manner, while for
horizontally loaded pile groups, the stiffness of the soil media from the surface
down to half of the embedded length of the pile will have a significant effect on
the behaviour of the pile group.
(7) For a small pile group embedded in homogeneous soil of infinite depth, the pile
head stiffness increases with increasing pile length to radius (L/r) ratio but
decreases with increasing number of piles in the group.
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
116






y
x
Element i
1
8
7
2 3
6 5
4
Figure 4.2 Nodal numbering of the 8-noded isoparametric rectangular element
z

x

y

u
i
, P
xi
v
i
, P
yi

(b) Bending action (a) Membrane action
Figure 4.1 Displacements and forces for a rectangular element subjected to
membrane and bending actions
Element i

yi

xi
z

x

y

Element i w
i
, P
zi
M
yi
M
xi
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
117
z

x

y

Raft Rectangular
Elements
Soil
Figure 4.3 Rectangular elements for raft

z

x

y

Raft
Point 1
Point 2
fixed against
x-movement
fixed
completely
Figure 4.4 Nodal fixity for analysis of the raft
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
118
q
z1
q
x1
q
z(n+1)
q
z(2n+1)
q
x(n+1)
q
x(2n+1)
p
pz1
p
pz2
p
pzn
p
px1
p
px2
p
pxn
p
pz(n+1)
p
pz(n+2)
p
px(n+1)
p
px(n+2)
p
px2n
p
pz2n
p
pz(2n+1)
p
pz(2n+2)
p
px(2n+1)
p
px(2n+2)
p
px3n
p
pz3n
p
sxn
p
sx2n
1
2
3
4
n
n+1
n+2
2n
2n+1
p
sz1
3n
2n+2
p
sz2
p
szn
p
sx1
p
sx2
p
sz(n+1)
p
sz(n+2)
p
sx(n+1)
p
sx(n+2)
p
sz(2n+1)
p
sx(2n+1)
p
sz(2n+2)
p
sx(2n+2)
p
sz2n
p
sz3n
p
sx3n
Figure 4.5 Interface forces of pile group in layered soil
(b) Forces acting on the pile group
(a) Forces acting on the soil interface
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
119
100
150
200
250
300
350
-3 -2 -1 0 1 2 3
Di stance from the centre poi nts (m)
S
e
t
t
l
e
m
e
n
t

(
m
m
)
Present FL
Mandal & Ghosh
(1999)
Cheung & Nag
(1968)
Figure 4.6 Comparisons of settlements along the centerline of the
centrally loaded square raft
POINT & DISTRIBUTED LOADS
Raft on isotropic homogeneous half space
R a ft s i z e : 6 x 6 m
T h i c k n e s s : 0.8 m
C o nc e ntr a te d l o a d a t c e n tr e
U ni fo r m l o a ds
Sc a l e
1.000E +03
Figure 4.7 Finite element mesh of the raft subjected to concentrated
load for the present analysis
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
120
Figure 4.8 Settlements along the centerline of the raft for Saving Bank Building
Figure 4.9 Mesh of the raft on layered materials for the finite element analysis
MATERIAL TYPES
Raft on layered materials with rigid base
Mate r i a l s
1
2
3
4
5
x y
z
0
5
10
15
20
25
30
0 5 10 15 20 25 30 35
x-distance (m)
S
e
t
t
l
e
m
e
n
t

(
m
m
)
Present FL - E = 500 MPa Present FL - E = 10000 MPa
Hooke's Law approach - E = 500 Hooke's Law approach - E = large value
Kay & Cavangaro - E = 500 Kay & Cavangaro E = large value
Mandal & Ghosh - E = 500 Conventional: 1 D approach
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
121
(b) Contours of settlement of the raft computed by finite element method
(a) Contours of settlement of the raft computed by the present method
Figure 4.10 Contours of settlement of the raft computed by the present
and infinite element method.
CONTOURS OF VERTICAL DISPLACEMENT
Raft on layered materials with rigid base
R a ft: 10x 10 m
T hi c k nes s : 0. 5 m
R i gi d Ba s e
C o nto ur L e gend
4. 500E+00
5. 000E+00
5. 500E+00
6. 000E+00
6. 500E+00
7. 000E+00
7. 500E+00
8. 000E+00
8. 500E+00
9. 000E+00
9. 500E+00
1. 000E+01
CONTOURS OF Z-DISPLACEMENT
Raft on layered materials with rigid base
C o nto ur L e ge nd
- 9.000E+00
- 8.000E+00
- 7.000E+00
- 6.000E+00
- 5.000E+00
- 4.000E+00
- 3.000E+00
- 2.000E+00
- 1.000E+00
0. 000E+00
x y
z
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
122
Figure 4.11 Comparisons of settlement by different methods
Figure 4.12 Typical deformed shape of a raft subjected to uniform horizontal load
x
y
u
y
= 0
u
z
= 0

y
= 0
u
y
= 0

y
= 0
u
y
= 0
u
z
= 0

y
= 0

y
0
0.002
0.004
0.006
0.008
0.01
0.012
0 2 4 6 8 10
x -di stance (m)
S
e
t
t
l
e
m
e
n
t

(
m
)
Present FL - layered
Finite element - layered
Present FL - equivalent modulus
Fraser & Wardle (approximate method)
Fraser & Wardle (FE method)
Mandal & Ghosh (FE-BE method)
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
123
DEFORMED RAFT
Raft subjected to uniform horizontal load
Raft size: 10 x 10 m
Thickness: 0.5 m
y x
CONTOURS OF X-DISPLACEMENT
Raft subjected to uniform horizontal load
Raft size: 10 x 10 m
Thickness: 0.5 m
Contour Legend
9.478E+01
9.480E+01
9.482E+01
9.484E+01
9.486E+01
9.488E+01
9.490E+01
9.492E+01
9.494E+01
9.496E+01
Figure 4.13 Deformed shape and contours of settlement of the raft by the present method
(b) Contours of x-displacement of the raft subjected to a uniform horizontal load
(a) Deformed shape of the raft subjected to a uniform horizontal load
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
124
(a) Deformed shape of the raft subjected to a uniform horizontal load
(b) Contours of settlement of the raft subjected to a uniform horizontal load
Figure 4.14 Deformed shape and contours of settlement of the raft for the finite
element method
100100 100 100 100 100100 100100
DEFORMED MESH
PILED RAFT (CHECK FOR PILE LOAD)
x y
z
CONTOURS OF X-DISPLACEMENT
PILED RAFT (CHECK FOR PILE LOAD)
C o nto ur L e ge nd
- 8.000E+01
- 7.000E+01
- 6.000E+01
- 5.000E+01
- 4.000E+01
- 3.000E+01
- 2.000E+01
- 1.000E+01
0.000E+00
1.000E+01
x y
z
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
125
(a) Horizontal x-displacement of the raft
(b) Vertical z-displacement of the raft
Figure 4.15 Comparisons of horizontal and vertical displacement of
the raft between the present and finite element methods
0.00
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.10
0 2 4 6 8 10
x-di stance from edge (m)
H
o
r
i
z
o
n
t
a
l

x
-
d
i
s
p
l
a
c
e
m
e
n
t

(
m
)
Present FL
Infinite element
-0.010
-0.008
-0.006
-0.004
-0.002
0.000
0.002
0.004
0.006
0.008
0.010
0 2 4 6 8 10 12
x-di stance from edge (m)
V
e
r
t
i
c
a
l

z
-
d
i
s
p
l
a
c
e
m
e
n
t

(
m
)
Present FL
Infinite element
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
126
Figure 4.17 Comparisons of bending moments in the raft
between the present and finite element methods
.0
.0
.0
.0
.0
0.0
0.0 2.0 4.0 6.0 8.0 10.0 12.0
0.0
2.0
4.0
6.0
8.0
10.0
CONTOURS OF MOMENT IN XX-DIRECTION
Raft subjected to uniform horizontal load
Raft size: 10 x 10 m
Thickness: 0.5 m
Contour Legend
-1.000E+01
-5.000E+00
0.000E+00
5.000E+00
1.000E+01
1.500E+01
2.000E+01
2.500E+01
3.000E+01
3.500E+01
Figure 4.16 Contours of bending moments in the raft computed by the present method
-15
-10
-5
0
5
10
15
0 2 4 6 8 10
x-di stance from edge (m)
M
o
m
e
n
t
,

M
x
x

(
k
N
m
/
m
)
Present FL
Infinite element
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
127
(b) Comparisons of bending moments in the raft on centre line
(a) Comparisons of vertical displacement of the raft on centre line
Figure 4.18 Comparisons of vertical displacement and bending moments
of the raft between the present and finite element methods
50
60
70
80
90
100
0 2 4 6 8 10 12
x-di stance from edge (m)
V
e
r
t
i
c
a
l

d
i
s
p
l
a
c
e
m
e
n
t

(
m
m
)
Present FL
Infinite element
-300
-250
-200
-150
-100
-50
0
0 2 4 6 8 10
x-di stance from edge (m)
M
o
m
e
n
t
,

M
x
x

(
k
N
m
/
m
)
Present FL
Infinite element
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
128
Figure 4.19 Configuration of the pile group and soil profile
100 , 30 , 200 = = =
d
h
d
L
E
E
s
p

d
8d
1 2 3
4 5 6
7 8 9
Example 1:
Vertical load
L
h

s
= 0.3
P
v
/9
P
v
/9 P
v
/9
P
H
/9
P
H
/9 P
H
/9
Example 2:
Horizontal load
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
129
(b) Axial forces along the piles
(a) Displacements along the piles
Figure 4.20 Comparisons of displacements and axial forces along the
piles for vertical loading
0
3
6
9
12
15
0 0.01 0.02 0.03 0.04
I
v
D
e
p
t
h

(
m
)
PILE 1 - Present FL
PILE 1 - DEFPIG
PILE 2 - Present FL
PILE 2 - DEFPIG
PILE 5 - Present FL
PILE 5 - DEFPIG
0
3
6
9
12
15
0 0.02 0.04 0.06 0.08 0.1 0.12
C
a
D
e
p
t
h

(
m
)
PILE 1 - Present FL
PILE 1 - DEFPIG
PILE 2 - Present FL
PILE 2 - DEFPIG
PILE 5 - Present FL
PILE 5 - DEFIG
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
130

(a) Displacements along pile 1
(b) Displacements along pile 2
(c) Displacements along pile 5
Figure 4.21 Comparisons of horizontal displacements along the piles
0
3
6
9
12
15
0.00 0.05 0.10 0.15
I
u
D
e
p
t
h

(
m
)
PILE 1 - Present FL
PILE 1 - DEFPIG
PILE 1 - Finite element
0
3
6
9
12
15
0.00 0.05 0.10 0.15
I
u
D
e
p
t
h

(
m
)
PILE 2 - Present FL
PILE 2 - DEFPIG
PILE 2 - Finite element
0
3
6
9
12
15
0.00 0.05 0.10 0.15
I
u
D
e
p
t
h

(
m
)
PILE 5 - Present FL
PILE 5 - DEFPIG
PILE 5 - Finite element
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
131

(a) Shear forces along pile 1
(b) Shear forces along pile 2
(c) Shear forces along pile 5
Figure 4.22 Comparisons of shear forces along the piles
0
3
6
9
12
15
-0.05 0 0.05 0.1 0.15
C
s
D
e
p
t
h

(
m
)
PILE 1 - Present FL
PILE 1 - DEFPIG
PILE 1 - Finite element
0
3
6
9
12
15
-0.05 0 0.05 0.1 0.15
C
s
D
e
p
t
h

(
m
)
PILE 2 - Present FL
PILE 2 - DEFPIG
PILE 2 - Finite element
0
3
6
9
12
15
-0.05 0 0.05 0.1 0.15
C
s
D
e
p
t
h

(
m
)
PILE 5 - Present FL
PILE 5 - DEFPIG
PILE 5 - Finite element
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
132
(b) Bending moments along pile 2
Figure 4.23 Comparisons of bending moments along the piles
(a) Bending moments along pile 1
(c) Bending moments along pile 5
0
3
6
9
12
15
0 0.02 0.04 0.06
C
b
D
e
p
t
h

(
m
)
PILE 2 - Present FL
PILE 2 - DEFPIG
PILE 2 - Finite element
0
3
6
9
12
15
0 0.02 0.04 0.06
C
b
D
e
p
t
h

(
m
)
PILE 1 - Present FL
PILE 1 - DEFPIG
PILE 1 - Finite element
0
3
6
9
12
15
0 0.02 0.04 0.06
C
b
D
e
p
t
h

(
m
)
PILE 5 - Present FL
PILE 5 - DEFPIG
PILE 5 - Finite element
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
133
Figure 4.24 Configuration of foundation and soil profile
(Kitiyodom and Matsumoto, 2003)
1.5 m 0.75 m 0.75 m
0
.
7
5

m

0
.
7
5

m

1
.
5

m

0.9 m
10 m
0.4 m
D
S
0.3 L
0.3 L
0.4 L
h
V
H
Case 1 Case 2 Case 3 Case 4 Case 5
E
s
E
s
2E
s
4E
s
4E
s
E
s
2E
s
E
s
E
s
2E
s
E
s
4E
s
4E
s
2E
s
E
s
h = 2L, L/d = 25,
s
= 0.3, E
p
/E
s
= 1000
(b) Soil profile for different cases
(a) Configuration of foundation
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
134
(a) Vertical displacements along the piles
Figure 4.25 Comparisons of calculated solutions for pile group. Case 1
(Kitiyodom and Matsumoto, 2003)
(b) Axial forces along the piles
(c) Horizontal displacements along the piles
(d) Shear forces along the piles
(e) Bending moments along the piles
0
2
4
6
8
10
0 0.01 0.02 0.03 0.04 0.05
I
v
D
e
p
t
h

(
m
)
Present FL
FEM
PRAB
0
2
4
6
8
10
0 0.03 0.06 0.09 0.12 0.15
I
H
D
e
p
t
h

(
m
)
Present FL
FEM
PRAB
0
2
4
6
8
10
0 0.1 0.2 0.3 0.4 0.5
C
a
D
e
p
t
h

(
m
)
Present FL
FEM
PRAB
0
2
4
6
8
10
-0.1 0 0.1 0.2 0.3
C
s
D
e
p
t
h

(
m
)
Present FL
FEM
PRAB
0
2
4
6
8
10
-0.6 -0.4 -0.2 0 0.2
C
b
D
e
p
t
h

(
m
)
Present FL
FEM
PRAB
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
135
0
2
4
6
8
10
0 0.01 0.02 0.03 0.04 0.05
I
v
D
e
p
t
h

(
m
)
Present FL
FEM
PRAB
(a) Vertical displacements along the piles
Figure 4.26 Comparisons of calculated solutions for pile group. Case 2
(Kitiyodom and Matsumoto, 2003)
(b) Axial forces along the piles
(c) Horizontal displacements along the piles (d) Shear forces along the piles
(e) Bending moments along the piles
0
2
4
6
8
10
0 0.1 0.2 0.3 0.4 0.5
C
a
D
e
p
t
h

(
m
)
Present FL
FEM
PRAB
0
2
4
6
8
10
0 0.03 0.06 0.09 0.12 0.15
I
H
D
e
p
t
h

(
m
)
Present FL
FEM
PRAB
0
2
4
6
8
10
-0.1 0 0.1 0.2 0.3
C
s
D
e
p
t
h

(
m
)
Present FL
FEM
PRAB
0
2
4
6
8
10
-0.6 -0.4 -0.2 0 0.2
C
b
D
e
p
t
h

(
m
)
Present FL
FEM
PRAB
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
136
(a) Vertical displacements along the piles
Figure 4.27 Comparisons of calculated solutions for pile group. Case 3.
(Kitiyodom and Matsumoto, 2003)
(b) Axial forces along the piles
(c) Horizontal displacements along the piles (d) Shear forces along the piles
(e) Bending moments along the piles
0
2
4
6
8
10
-0.4 -0.3 -0.2 -0.1 0 0.1 0.2
C
b
D
e
p
t
h

(
m
)
Present FL
FEM
PRAB
0
2
4
6
8
10
-0.1 0 0.1 0.2 0.3
C
s
D
e
p
t
h

(
m
)
Present FL
FEM
PRAB
0
2
4
6
8
10
0 0.1 0.2 0.3 0.4 0.5
C
a
D
e
p
t
h

(
m
)
Present FL
FEM
PRAB
0
2
4
6
8
10
0 0.03 0.06 0.09
I
H
D
e
p
t
h

(
m
)
Present FL
FEM
PRAB
0
2
4
6
8
10
0 0.01 0.02 0.03 0.04 0.05
I
v
D
e
p
t
h

(
m
)
Present FL
FEM
PRAB
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
137
(a) Vertical displacements along the piles
Figure 4.28 Comparisons of calculated solutions for pile group. Case 4.
(Kitiyodom and Matsumoto, 2003)
(b) Axial forces along the piles
(c) Horizontal displacements along the piles (d) Shear forces along the piles
(e) Bending moments along the piles
0
2
4
6
8
10
-0.3 -0.2 -0.1 0 0.1
C
b
D
e
p
t
h

(
m
)
Present FL
FEM
PRAB
0
2
4
6
8
10
-0.1 0 0.1 0.2 0.3
C
s
D
e
p
t
h

(
m
)
Present FL
FEM
PRAB
0
2
4
6
8
10
0 0.1 0.2 0.3 0.4 0.5
C
a
D
e
p
t
h

(
m
)
Present FL
FEM
PRAB
0
2
4
6
8
10
0 0.03 0.06 0.09
I
H
D
e
p
t
h

(
m
)
Present FL
FEM
PRAB
0
2
4
6
8
10
0 0.01 0.02 0.03 0.04 0.05
I
v
D
e
p
t
h

(
m
)
Present FL
FEM
PRAB
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
138
(a) Vertical displacements along the piles
Figure 4.29 Comparisons of calculated solutions for pile group. Case 5.
(Kitiyodom and Matsumoto, 2003)
(b) Axial forces along the piles
(c) Horizontal displacements along the piles
(d) Shear forces along the piles
(e) Bending moments along the piles
0
2
4
6
8
10
0 0.01 0.02 0.03 0.04 0.05
I
v
D
e
p
t
h

(
m
)
Present FL
FEM
PRAB
0
2
4
6
8
10
0 0.1 0.2 0.3 0.4 0.5
C
a
D
e
p
t
h

(
m
)
Present FL
FEM
PRAB
0
2
4
6
8
10
0 0.02 0.04 0.06 0.08 0.1
I
H
D
e
p
t
h

(
m
)
Present FL
FEM
PRAB
0
2
4
6
8
10
-0.1 0 0.1 0.2 0.3
C
s
D
e
p
t
h

(
m
)
Present FL
FEM
PRAB
0
2
4
6
8
10
-0.3 -0.2 -0.1 0 0.1
C
b
D
e
p
t
h

(
m
)
Present FL
FEM
PRAB
Chapter 4 Analysis of Rafts and Pile Groups in Layered Soil
139


0
20
40
60
80
100
0 20 40 60 80 100
L/r
P
/
(
G
r
w
)
Present FL
Guo & Randolph (1999)
Butterfield & Banerjee (1971)
0
20
40
60
80
100
0 20 40 60 80 100
L/r
P
/
(
G
r
w
)
Present FL
Guo & Randolph (1999)
Butterfield & Banerjee (1971)
Figure 4.30 Configurations of pile groups
s
2r
s
s
s
s

(a) Pile head stiffness for = 6000
(b) Pile head stiffness for =
Figure 4.31 Comparisons of pile head stiffness for different pile
groups in infinite homogeneous soil layer (Guo and
Randolph, 1999)
2-pile group
3-pile group
4-pile group
2-pile group
3-pile group
4-pile group
Chapter 5 Soil-Structure Interaction
140
CHAPTER 5
SOIL-STRUCTURE INTERACTION

5.1 INTRODUCTION
Soil-structure interaction refers to the interaction between a structure, its foundation and
the subsoil. As the term soil-structure interaction covers a broad field, the work presented
in this chapter will be restricted to the analysis of raft or piled raft foundations under
static load.

Traditionally, the design work for structures has been separated from that of the
foundations by the structural and geotechnical engineers, thus the stiffness of the
superstructure is neglected. However, the structural stiffness can have effects on the
distribution of column loads and bending moments transmitted from the structure to the
foundation, therefore interaction analysis which accounts for the structural stiffness has to
be considered.

Piled rafts are composite structures which are comprised of three elements: the piles, raft
and the supporting soil. Loads applied to the raft are transferred to the soil through the
piles, therefore, it is necessary to take into account the interaction among the three
elements. Four different types of interaction have to be considered in the analysis. In this
chapter, the interaction mechanism for piled raft foundations is discussed and the use of
the finite layer technique to compute the interaction factors is presented.

5.2 EFFECT OF SUPERSTRUCTURE ON FOUNDATIONS
Methods for the analysis of soil-structure interaction have been developed over many
years. Lee and Brown (1972) developed an analysis by treating the structure, foundation
and soil system as an integral unit. The soil was treated as a Winkler or linear elastic
model. The method was applied to the analysis of a multi-bay frame. Results have shown
Chapter 5 Soil-Structure Interaction
141
that the maximum moment in the foundation decreases with increasing flexibility of the
foundation.

Fraser and Wardle (1975) used the finite element method to analyse a two bay portal
frame on a layered cross-anisotropic elastic continuum in which the frame was modelled
by beam elements, the raft by plate elements and the soil surface by surface elements.
The elastic continuum was assumed to consist of a number of horizontal layers of
uniform thickness with infinite lateral extent. It was found that the differential
displacements depend on the stiffness of the structural frame and cross anisotropy of the
soil has significant effects on the interaction. Brown (1975) showed that for structures on
strip footings, the differential displacements in the raft were dependent on the relative
stiffness of the structure. Zhao and Cao (1985) used the substructure method to analyse
a twelve storey frame structure with two basements founded on soft clay. The structure,
raft and the soil were analysed as a whole system. Results have shown that a variation in
the stiffness of the structure, raft or soil would cause a redistribution of the force in the
system. As the superstructure contributes additional stiffness to the raft, bending moment
in the raft was reduced which led to forces being transferred to the superstructure and
resulted in increasing the bending moment in the structure.

Yao and Zhang (1985) have shown that relative section rigidity of the frame has an effect
on the differential settlements and the forces on the raft, however, it has no influence on
the moments in the raft. The thickness of the raft has significant effect on the moments in
the frame and also the forces in the raft. Brown and Yu (1986) showed that the stiffness
of the structure increases progressively during construction which could have significant
effect on the differential settlements in the raft. Noorzaei et al. (1991) studied the effect
of the slab and raft thickness on the behaviour of the overall frame-raft-soil system. The
slab was taken into account as part of the superstructure. Results have shown that
increasing the raft stiffness would lead to an increase of bending moments and a
reduction of differential settlement and contact pressure in the raft. The moments in the
superstructure members were redistributed. The slab thickness has significant effect on
the bending moments in the raft and the superstructure (Viladkar et al. 1992).
Chapter 5 Soil-Structure Interaction
142

Zhang and Small (1994) employed the finite layer technique to analyse three-dimensional
framed buildings on raft foundations. It was demonstrated that an increase in the relative
stiffness of the frame will result in a decrease in the differential settlements in the raft.
Small (2001 and 2002) used very stiff elements on the raft to represent the extremely stiff
structural elements such as shear walls or solid cores. For very flexible structures, the
effect of the stiffness of the superstructure on the behaviour of the foundation is not
significant, therefore, it can be neglected in the analysis of the raft. However, if the
superstructure is stiff, it needs to be included into the analysis in some way.

5.3 INTERACTION MECHANISM OF PILED RAFT
The concept of interaction factors has been widely adopted for the analysis of pile groups
and piled rafts (Hain and Lee, 1978, Kakurai et al., 1987, Kitiyodom and Matsumoto,
2003) since its introduction by Poulos (1968). Davis and Poulos (1972) suggested that the
analysis of a piled raft involves the interaction between the piles and the cap. In their
analysis, the interaction was computed based on the integration of the Mindlin solution.
Finite layer techniques developed by Small and Booker (1986) were used to compute the
interaction factors by Ta and Small (1995) and Zhang and Small (1999) for the analysis
of pile groups in layered soils and then extended to piled rafts (Ta, 1996) and Zhang
(2000). In their analyses, piles used to support the raft have to be identical. In this section,
four different interactions (i) pile-pile, (ii) pile-soil, (iii) soil-pile and (iv) soil-soil as
shown in Figure 5.1 for a layered system are calculated by the finite layer method. The
piles considered in the calculation can be of different diameters and lengths.

5.3.1 Pile-Pile Interaction
Interaction between piles is an important consideration in the analysis of pile groups and
piled rafts. This interaction is defined as the additional deflection of an unloaded pile
caused by an adjacent loaded pile. The interaction factor,
pp
, may be expressed as
additional deflection of an unloaded pile
deflection of a loaded pile in isolation

pp
=
__________________________________
(5.1)
Chapter 5 Soil-Structure Interaction
143
This interaction can be calculated by considering a pair of piles with a distance s between
them as shown in Figure 5.2. The piles can have different lengths and diameters and are
divided into a number of beam elements. The soil is divided into several layers
corresponding to the pile elements.

The computation starts with the formation of the influence matrix for the soil. The
influence matrix is computed by applying a horizontal or vertical unit ring load at each
soil layer along the pile shaft and a horizontal or vertical unit uniform circular load at the
pile base in turn. Deflections computed at the locations of these ring and base loads form
the columns of the soil influence matrix [I
s
]. The soil influence matrix is in the form of
equation (4.32) in Chapter 4. Deflections of the soil at the locations of the ring or uniform
loads can be expressed as

s
= I
s
P
s
(5.2)
where
s
= vector of soil displacement at the nodes
I
s
= soil influence matrix
P
s
= vector of forces acting at the soil interfaces
The soil stiffness matrix [K
soil
] can be obtained by inverting the soil influence matrix and
the load-displacement relationship for the soil can be rewritten as
K
s

s
= P
s
(5.3)
where [K
s
] = [I
s
]
1

After the soil stiffness matrix has been constructed, the pile stiffness matrix [K
p
] for the
two piles under consideration can be assembled. The pile stiffness matrix is simply the
sum of individual element stiffness matrices. The stiffness matrix for individual pile
elements is given in equation (4.29) in Chapter 4. The load-displacement relation for the
pile can be expressed as

[K
p
]
p
= P
p
+ Q
p
(5.4)

where
p
= vector of deflections at the nodes of the pile
P
p
= vector of forces applied at the nodes of the pile
Q
p
= vector of the applied load at the pile head
Chapter 5 Soil-Structure Interaction
144
By assuming that no slip occurs along the pile shaft, the deflection of the soil and
deflection at the pile shaft must be the same. Thus, the compatibility of displacement is
=
s
=
p
(5.5)
The equilibrium of the interaction forces at the pile-soil interface and the pile bases are
P
s
= -P
p
(5.6)
Combining equations (5.2) to (5.5) gives
[K
c
] {} = {P} (5.7)
where [K
c
] = [K
s
] + [K
p
]
By solving equation (5.7), the horizontal and vertical deflections along the pile shaft and
the pile base of the pair of piles in consideration can then be determined.

5.3.2 Pile-Soil Interaction
This is the interaction between a loaded pile and the soil surface at a distance s from the
loaded pile as shown in Figure 5.3. The pile-soil interaction,
ps
, may be expressed as
It can be obtained by using the same approach as for the pile-pile interaction in Section
5.3.1. However, the stiffness matrix is set up for a single pile only. By applying a unit
load on the pile head of the single pile, the magnitudes of the ring loads along the pile
shaft and circular load at the pile base can be computed. These loads are then applied
back to the soil to solve for the horizontal and vertical surface deflections at the desired
positions by the finite layer method as described in Chapter 3.

5.3.3 Soil-Pile Interaction
This is the interaction between a uniform rectangular load applied to the soil surface and
an unloaded pile at a distance s from the centre of the loaded soil surface as shown in
Figure 5.4. The soil-pile interaction,
sp
, may be expressed as
additional deflection of an unloaded soil surface
deflection of a loaded pile in isolation

ps
=
_________________________________________
(5.8)
additional deflection of an unloaded pile
deflection of a loaded soil surface

sp
=
__________________________________
(5.9)
Chapter 5 Soil-Structure Interaction
145
This interaction is computed by considering the soil first. The deflection of the soil
consists of two components (i) due to the ring loads along the pile shaft and (ii) the
surface load applied to the soil. The deflection due to the ring load can be computed by
using equation (5.2), whilst the deflection due to the surface load can be computed simply
by the finite layer method. The load-deflection relationship can be written as
{
s
}= [I
s
]{P
s
} + {
l
} (5.10)
where
s
= vector of deflections of the soil along the pile shaft and at the pile base
computed at the nodes of each pile
P
s
= vector of loads acting on the soil along the pile shaft and at the base

1
= vector of deflections of the soil due to a unit surface load at a distance s from
the pile computed at the nodes of the pile
For a single pile, the load-deflection relationship can be written as
{
p
} = [I
p
]{P
p
} + a
x
+ b
y
+ c
z
+ d
x
+ e
y
(5.11)
where
p
= vector of deflections at the nodes of each pile element
I
p
= pile influence matrix

x
,
y
,
z
= horizontal (x, y) and vertical (z) translations for the pinned pile

x
,
y
= x- and y-rotations for the pinned pile
a = (1, 0, 0, 0, 0, 1, 0, 0, 0, 0,, 1, 0, 0, 0, 0)
T
b = (0, 1, 0, 0, 0, 0, 1, 0, 0, 0,, 0, 1, 0, 0, 0)
T

c = (0, 0, 1, 0, 0, 0, 0, 1, 0, 0,, 0, 0, 1, 0, 0)
T

d = (0, 0, x
1
, 1, 0, 0, 0, x
2
, 1, 0,, 0, 0, x
n
, 1, 0)
T

e = (0, 0, y
1
, 0, 1, 0, 0, y
2
, 0, 1,, 0, 0, y
n
, 0, 1)
T

The pile influence matrix can be obtained by (i) assembling the stiffness matrix for a
single pile, (ii) applying the boundary conditions to pin the pile at the pile head, (iii)
applying unit loads at each node to the pinned pile.

The deflection of the pile can be computed by combining equations (5.10) and (5.11)
with the assumption that no slip occurs along the pile shaft (i.e.
p
=
s
) and the
interaction forces at the pile-soil interface are the same in magnitude but of opposite sign
(i.e. P = P
p
= - P
s
). The combined equation is written as
{
p
} = [I
sp
]{P} + a
x
+ b
y
+ c
z
+ d
x
+ e
y
(5.12)
Chapter 5 Soil-Structure Interaction
146
where [I
sp
] = [I
s
] + [I
p
].

5.3.4 Soil-Soil Interaction
This is the interaction between a uniformly loaded surface and the unloaded soil surface
at a distance s from the centre of the loaded surface as shown in Figure 5.5. The soil-soil
interaction,
ss
, may be expressed as
The interaction can be computed by using the finite layer method described in Chapter 4.
A uniform rectangular load in the horizontal or vertical direction is applied on the soil
surface to each area in turn to calculate the deflection of the soil at the desired positions.

5.4 NUMERICAL EXAMPLES
5.4.1 Pile-Pile Interaction
Example 1 Vertically loaded rigid piles embedded in a homogeneous elastic soil
This example shows the interaction between two rigid piles embedded in a homogeneous
elastic soil with a modulus of 25MPa and a Poissons ratio of 0.499. The piles were
identical with a spacing s between them. As shown in Figure 5.6a, each pile has a
diameter (D) of 1m, the pile length (L) to diameter ratio, L/D, was taken as 25 and the
overall soil depth (h) to pile length ratio, h/L, was taken as 2.5. A vertical load was
applied to the pile head of Pile 1 and the deflections at the pile head of Pile 2 at various
spacings were computed.

The interaction factors determined from the present finite layer method were compared
with those by Chow (1987a) based on the finite element method, and by Poulos (1968)
based on the integral equation method as shown in Figure 5.6b. There is a good
additional deflection of an unloaded soil surface
deflection of a loaded soil surface

ss
=
_________________________________________
(5.9)
Chapter 5 Soil-Structure Interaction
147
agreement between the solutions. The interaction factor decreases as the spacing to
diameter (s/D) ratio increases and approaches zero at s/D 25.

Example 2 Vertically loaded piles embedded in non-homogeneous layered soil
This example demonstrates the calculation of interaction factors for pile groups
embedded in layered soils. Two identical piles were involved and the piles have a
diameter of 1m and a length of 25m. A vertical load was applied to one of the piles and
four cases with different soil profiles as shown in Figure 5.7a were analysed. The
Poissons ratio for all layers of the soil was taken as 0.3.

The present solutions are in good agreement with the solutions from the boundary
element method (Chow, 1986), the finite element method (Chow, 1987a) and the discrete
layer method (Lee, 1991) for cases 1, 2, and 4 as shown in Figure 5.7b. For case 3, where
a soft layer exists in between two much stiffer layers, the present solution is in excellent
agreement with the finite element approach. The interaction factors were overestimated
by the boundary element approach and were underestimated by the discrete layer
approach. The interaction factors increase as the top layer of the soil gets stiffer. This
implies that as the soil gets stiffer, the head of the unloaded pile deflects more when the
other pile is loaded.

Example 3 Vertically loaded piles embedded in soil with underlying stiff layer
This example shows the influence of a soft layer underlying a stiff layer beneath the pile
base. Two identical piles with a diameter of 1m and a length of 25m were involved. A
stiff layer with a thickness of t is lying beneath the pile base as shown in Figure 5.8a. The
modulus of the pile E
p
was assumed to be 25,000MPa and the overall depth of the soil
media was 2 x length of the pile.

The interaction factors for different stiff layer thickness to diameter (t/D) ratios are
shown in Figure 5.8b. The factors obtained from the present method are in good
agreement with those obtained from the finite element method by Chow (1987a). As the
t/D ratio increases, the interaction factor decreases which implies that when the stiff layer
Chapter 5 Soil-Structure Interaction
148
beneath the pile base gets thicker, it restricts the unloaded pile from deflecting and the
piles are acting more as end bearing piles.

Example 4 Vertically loaded piles embedded in a semi-infinite soil mass
In this example, two identical piles were embedded in a semi-infinite soil mass. The piles
have a diameter of 1m and a modulus of 25,000MPa. The Poissons ratio of the soil was
taken as 0.499. Analyses have been carried out for pile lengths of L/D = 25 and 50, and
for pile to soil stiffness ratios E
p
/E
s
= 100 and 1000 as shown in Figure 5.9a. The present
solutions were compared with the solutions from Poulos (1968) and El Sharnouby and
Novak (1990) in which the deflections of the piles were computed based on the Mindlin
solution.

Figures 5.9a and b show the interaction factors for L/D = 25 and 50 respectively. There
is a fair agreement among the solutions obtained from the present finite layer method,
and the solutions of El Sharnouby and Novak (1990) and Poulos (1968) for L/D = 25. For
L/D = 50, the present solutions are in good agreement with El Sharnouby and Novak
(1990) and Poulos (1968) for E
p
/E
s
= 1000. For E
p
/E
s
= 100, the present solutions are in
close agreement with El Sharnouby and Novak (1990), however, the solutions from
Poulos (1968) are slightly higher. For the cases that have been analysed, the interaction
factors predicted by the present method are smaller, this could be due to the number of
elements into which the piles were discretised and the method of solution. Poulos (1968)
used 10 elements in the pile and the integral equations to approximate the deflection of
the pile while El Sharnouby and Novak (1990) used 50 elements in the pile and a point
load in an elastic half space. In the present method, each of the piles was discretised into
25 elements to obtain the solutions.

Example 5 Horizontally loaded piles embedded in homogeneous soil
In this example, two identical piles embedded in a homogeneous soil and directly
standing on a very stiff base layer were considered. A horizontal load was applied to one
of the pile heads. The angle between the line joining the pile centres and the direction of
Chapter 5 Soil-Structure Interaction
149
loading (termed as the departure angle) is as shown in Figure 5.10a. Analyses have
been carried out for pile to soil stiffness ratios E
p
/E
s
= 80 and 8000.

The interaction factors for lateral displacements of two free-head piles are shown in
Figures 5.10b (E
p
/E
s
= 80) and 5.10c (E
p
/E
s
= 8000). The interaction factors for stiffer
piles (E
p
/E
s
= 8000) are higher than those for compressible piles (E
p
/E
s
= 80). When the
relative location of the piles is in the same direction as the applied load (i.e. = 0), the
interaction factors are higher than those in the perpendicular direction (i.e. = 90). This
implies that the applied load would have significant influence on the piles located in the
same direction as the load but this influence becomes less significant when the piles are
located perpendicular to the load. The present solutions are compared with the finite
element solutions by Chow (1987a), the approximate solutions based on finite element
method by Randolph (1981) and the integral equation by Poulos and Davis (1980). It can
be seen that for low stiffness piles (E
p
/E
s
= 80), the present solutions are in good
agreement with Chows and Randolphs solutions. For stiffer piles (E
p
/E
s
= 8000), the
present solutions are in good agreement with Chows solution and slightly different from
the Randolph solution. For both cases, the predictions from Poulos and Davis (1980) are
slightly higher than for the other solutions, and this could be due to the number of
elements used in the analysis.

Example 6 Laterally loaded piles embedded in non-homogeneous layered soil
This example involves two identical piles embedded in a layered soil of finite depth. One
of the piles was subjected to a lateral load and four cases were analysed. The pile
configuration and soil properties as shown in Figure 5.11a are the same as in Example 2.

Figure 5.11b shows the interaction factor at various spacing to diameter (s/D) ratios for
different cases. It can be seen that the interaction factors are almost the same for all cases
which implies that for piles embedded in layered soil of finite depth considered here, the
change in modulus of the soil would have little influence on the interaction factor. The
present solutions are compared with the solutions obtained from a computer program
PRAB developed by Kitiyodom and Matsumoto (2003) and the finite element method.
Chapter 5 Soil-Structure Interaction
150
The program PRAB uses an averaging technique to approximate the interactions between
the piles. In the finite element analysis, the pile was modelled by an octagon inscribed in
the circle and the soil was modelled by a linear elastic material. The present solutions are
in good agreement with the solutions from the computer program PRAB and the finite
element method for Case 1. For the other cases, there are minor differences between the
present method and the finite element method. This could be due to the number of
elements used to model the piles and the boundaries of the soil used in the finite element
analysis.

Example 7 Piles of different sizes subjected to axial or lateral loads
In the previous examples, the interaction between two identical piles subjected to
horizontal or vertical loads was examined. When non-uniform loadings are applied to the
foundation, the use of piles of different sizes may be required, therefore, it is necessary to
examine interaction between two non-identical piles. In this example, three different
cases are presented to show the interaction between piles of different lengths and
diameters. The piles with a modulus of 30,000MPa, are embedded in an infinitely deep
homogeneous soil with a modulus of 4MPa and a Poissons ratio of 0.3. The lengths and
diameters of piles considered in this example are listed in Table 5.1 and the piles have a
spacing s between them. The analyses were carried out by applying a vertical or
horizontal load to either one of the piles and computing the interaction factor between the
two piles.

Pile Length (m) Pile Diameter (m)
Case L1 L2 D1 D2
1 25.0 12.5 1.0 1.0
2 25.0 25.0 1.0 0.5
3 25.0 12.5 1.0 0.5
Table 5.1 Dimensions of piles for different cases

Chapter 5 Soil-Structure Interaction
151
Case 1 - Same diameters and different lengths
This case involves two piles (Pile 1 and Pile 2) with the same diameter (D1 = D2) but
where the length of Pile 1 is twice the length of Pile 2 (L1 = 2 x L2) as shown in Figure
5.12a. Figure 5.12b shows the interaction factors for piles subjected to a vertical load,
where it can be seen that the interaction factors for loading the longer pile (Pile 1) are
higher than for loading the shorter pile (Pile 2). The interaction factors for piles subjected
to a horizontal load are shown in Figure 5.12c. For piles having the same diameters, the
interaction factors for loading the longer piles are almost the same as for loading the
shorter piles. This shows that the length of the piles has little effect on the interaction
between the piles when a horizontal load is applied as long as the piles are longer than the
critical depth (the depth at which lateral deflection is small).

Case 2 Different diameters and same lengths
This case involves two piles of the same length (L1 = L2) but the diameter of Pile 1 is
twice the diameter of Pile 2 (D1 = 2 x D2) as shown in Figure 5.13a. The interaction
factors for vertically loaded piles are shown in Figure 5.13b. The interaction factors when
the larger pile (Pile 1) is loaded are twice the magnitude of the factors when the smaller
pile (Pile 2) is loaded. This is because the reciprocal theorem can be applied. The same
trend for the interaction factors for horizontally loaded piles is observed as shown in
Figure 5.13c. The vertical interaction factors are slightly smaller than those for case 1,
this shows that as the diameter of the pile increases, the stiffness of the pile also increases.
When a load is applied to a stiff pile, the deflection of the loaded pile will be small and
the corresponding interaction forces induced on an adjacent unloaded pile will be smaller
which results in a reduction in the deflection of the unloaded pile. The horizontal
interaction factors when loading Pile 1 are about the same compared to those in Case 1
but when Pile 2 is loaded, the interaction factors are reduced significantly. This shows
that the larger pile is more resistant to moving horizontally than when a smaller pile is
loaded. From this case, it can be observed that the diameter of the piles has significant
effect on the interaction between piles subjected to vertical or horizontal load.


Chapter 5 Soil-Structure Interaction
152
Case 3 Different diameters and different lengths
In this case, the piles are of different diameters and lengths in which the length and
diameter of pile 1 is twice of the length and diameter of pile 2 (D1 = 2 x D2 and L1 = 2 x
L2) as shown in Figure 5.14a. The interaction factors for vertical and horizontal loads are
shown in Figures 5.14b and c respectively. It is observed that the interaction factors when
loading the larger and longer pile (Pile 1) are twice as big as those computed when
loading the smaller and shorter pile (Pile 2). The vertical interaction factors are slightly
higher than in Case 2 as the stiffness of the shorter pile has reduced due to the reduction
of the pile length. The horizontal interaction factors are the same as in Case 2 which
implies that the length of the pile may not have significant effect on the horizontal
interaction between piles past a critical length.

The interaction factors for different pile arrangements presented in this example have
shown that the diameter of piles has significant effect on the interaction between piles
especially when the piles are subjected to horizontal loads. From Cases 2 and 3, the
interaction factors for loading the larger pile of twice the diameter as the smaller pile are
twice as big as those for loading the smaller pile which implies that the reciprocal
theorem holds.

5.4.2 Pile-Soil Interaction
Example 1 Different pile lengths and diameters
This example demonstrates the interaction between a pile and the soil surface in which a
load is applied to the pile. Two different sizes of pile as listed in Table 5.2 were
considered and the pile was subjected to a vertical or horizontal load as shown in Figure
5.15a. The properties of soil and piles are the same as in Example 7 in Section 5.4.1.
Pile 1 Pile 2
Length (m) 25 12.5
Diameter (m) 1 0.5
Table 5.2 Lengths and diameters of Pile 1 and Pile 2

Chapter 5 Soil-Structure Interaction
153
The interaction factors were plotted against the spacing to diameter ratio (s/D) where the
diameter is for the loaded pile. The vertical and horizontal interaction factors for a pile
and the soil surface are shown in Figures 5.15b and c respectively. The interaction factors
when Pile 1 is loaded are about the same as when Pile 2 is loaded for both a vertically
and horizontally loaded pile. This shows that diameter and length of the loaded pile does
not have significant effect on the interaction between a pile and the soil surface for cases
such as the one chosen here, but for other geometries this may not be so.

5.4.3 Soil-Pile Interaction
Example 1 Different pile lengths and diameters
This example involves a rectangular soil surface load with a length and breadth of S1
loaded vertically or horizontally and a pile as shown in Figure 5.16a. The lengths and
diameters of the piles considered are listed in Table 5.2. The properties of soil and piles
are the same as in Example 7 in Section 5.4.1.

The interaction factors are plotted against the spacing to the length of the loaded surface
rectangle (s/S1) ratios. The vertical and horizontal interaction factors for two different
sizes of piles are shown in Figures 5.16b and c respectively. The vertical and horizontal
interaction factors between the soil surface load and Pile 2 (smaller and shorter pile) are
higher than those for Pile 1 which shows that Pile 2 is deflecting more than Pile 1 with
the same magnitude of load applied on the soil surface. However, the interaction factors
decrease to the same value as s/S1 exceeds 10 and then approach zero.

5.4.4 Soil-Soil Interaction
Example 1 Soil surface subjected to vertical and horizontal loads
This example involves a square uniform load applied to the soil surface with a length and
breadth of S1 in plan and a point on the soil surface where the square region of the soil
surface is loaded vertically or horizontally as shown in Figure 5.17a. The soil properties
are the same as in Example 1 in Section 5.4.3. The vertical and horizontal interaction
Chapter 5 Soil-Structure Interaction
154
factors are plotted against the ratios of s/S1 as shown in Figures 5.17b and c respectively.
The interaction factor decreases rapidly as s/S1 increases from 1 to 2. At s/S1values
exceeding 6, the interaction factor decreases gradually and approaches zero.

Note that the soil surface load may be rectangular and need not be square as in these
examples.

5.5 CONCLUSIONS
In this chapter, soil-structure interaction which governs the behaviour of rafts and piled
rafts has been discussed. In the analysis of piled rafts, four different types of interaction:
(i) pile-pile, (ii) pile-soil, (iii) soil-pile, (iv) soil-soil have to be considered. The
application of the finite layer method to computing the interaction factors was
demonstrated through the above examples and the following conclusions can be reached:

(1) By taking into account the stiffness of the superstructure in analysing rafts or
piled rafts, the differential deflections of the raft can be reduced. However, for
very flexible structures, the raft can be analysed alone without great error.
Therefore, neglecting the superstructure in the analysis can be conservative.
(2) The present solutions are in good agreement with the solutions of other existing
methods such as the finite element method, discrete layer method and boundary
element method.
(3) The present method can give reasonably accurate solution for foundations
embedded in non-homogeneous soil without the use of averaging techniques.
(4) If the piles are embedded in soil with the existence of a stiff layer beneath the pile
bases, the stiff layer can resist the deflection of a pile such that the interaction
factor decreases as the thickness of the stiff layer increases. If the stiff layer is
thick enough, the piles will act as end-bearing piles.
(5) The accuracy of the solution depends on the number of elements into which the
pile is discretised.
Chapter 5 Soil-Structure Interaction
155
(6) For horizontally loaded piles, the interaction factors for piles located in the same
direction as the applied load are higher than those where the piles are located
normal to the applied load.
(7) The interaction between two non-identical piles is significantly affected by the
diameter of the pile especially for horizontally loaded piles. If the diameter and
length of one of the piles is twice that of another pile in consideration, the
reciprocal theorem holds.
(8) The interaction factors between the soil surface and a short and small diameter
pile are higher than those for a long and large diameter pile. However, as the
distance between the soil surface and the pile increases, the interaction factors
approach zero.

Chapter 5 Soil-Structure Interaction
156
(b) Pile to soil
(a) Pile to Pile
(c) Soil to pile
(d) Soil to soil
Figure 5.1 Types of interaction in a piled raft foundation
Chapter 5 Soil-Structure Interaction
157

Figure 5.2 Interaction between a loaded pile and an unloaded pile Pile-pile interaction

Figure 5.3 Interaction between a loaded pile and the soil surface Pile-soil interaction

s
P
z

x
P
x
Pile 1

Pile 2

s
P
z
Unloaded soil
surface

z
P
x

x
s
P
z

x
P
x
Pile 1

Pile 2

Chapter 5 Soil-Structure Interaction
158
Figure 5.4 Interaction between a loaded surface and an unloaded pile Soil-pile interaction

s
Loaded soil
surface

x
Figure 5.5 Interaction between a loaded surface and an unloaded surface Soil-soil interaction

s
Loaded soil
surface

x
Chapter 5 Soil-Structure Interaction
159
Figure 5.6 Pile arrangement and vertical deflection interaction factors for 2 identical piles

(b) Vertical deflection interaction factors for 2 rigid piles

(a) Two rigid piles embedded in a homogeneous elastic soil

P
L
s
h
D
1000 , 5 . 2 , 25 = = =
s
E
p
E
L
h
D
L

Pile 1 Pile 2
0.0
0.2
0.4
0.6
0.8
1.0
0 5 10 15 20 25
s/D
I
n
t
e
r
a
c
t
i
o
n

f
a
c
t
o
r
,

p
p
Present FL
CHOW (1987a)
POULOS (1968)
Chapter 5 Soil-Structure Interaction
160

D
s
0.3 L
0.3 L
0.4 L
h
P
Case 1 Case 2 Case 3 Case 4
E
s
2E
s
4E
s
4E
s
2E
s
E
s
E
s
2E
s
4E
s
4E
s
2E
s
E
s
1000 , 2 , 25 = = =
s
p
E
E
L
h
D
L

(a) Soil profiles for different cases
Case 1 Case 2
Case 3 Case 4
(b) Interaction factors for different cases
Figure 5.7 Soil properties and vertical interaction factors for piles embedded in layered soils
0.0
0.2
0.4
0.6
0.8
1.0
0 2 4 6 8 10 12
s /D
I
n
t
e
r
a
c
t
i
o
n

f
a
c
t
o
r
,

p
p
Present FL
CHOW (1987a)
CHOW (1986)
LEE (1991)
0.0
0.2
0.4
0.6
0.8
1.0
0 2 4 6 8 10 12
s /D
I
n
t
e
r
a
c
t
i
o
n

f
a
c
t
o
r
,

p
p
CHOW (1987a)
CHOW (1986)
Present FL
LEE (1991)
0.0
0.2
0.4
0.6
0.8
1.0
0 2 4 6 8 10 12
s/D
I
n
t
e
r
a
c
t
i
o
n

f
a
c
t
o
r
,

p
p
Present FL
CHOW (1987a)
CHOW (1986)
LEE (1991)
0.0
0.2
0.4
0.6
0.8
1.0
0 2 4 6 8 10 12
s /D
I
n
t
e
r
a
c
t
i
o
n

f
a
c
t
o
r
,

p
p
CHOW (1987a)
CHOW (1986)
Present FL
LEE (1991)
Chapter 5 Soil-Structure Interaction
161
D
s
L
h
P
E
s
5E
s
E
s
1000 , 2 , 25 = = =
s
p
E
E
L
h
D
L

t
(b) Interaction factors for different thickness of the stiff layer
Figure 5.8 Interaction factors for 2 pile group embedded in soil with underlying stiff layer
(Chow, 1987a)
(a) Soil layer profile
0
0.1
0.2
0.3
0.4
0.5
0.6
0 5 10 15 20 25
s/D
I
n
t
e
r
a
c
t
i
o
n

f
a
c
t
o
r
,

p
p
Present FL
CHOW (1987a)
t/D
0
5
10
25
Chapter 5 Soil-Structure Interaction
162
0
0.2
0.4
0.6
0.8
1 2 3 4 5 6
s/D
I
n
t
e
r
a
c
t
i
o
n

f
a
c
t
o
r
,

p
p
Present FL
El Sharnouby and Novak (1990)
Poulos (1968)
0
0.2
0.4
0.6
0.8
1
1 2 3 4 5 6
s/D
I
n
t
e
r
a
c
t
i
o
n

f
a
c
t
o
r
,

p
p
Present FL
El Sharnouby and Novak (1990)
Poulos (1968)
P
L
s
D
Pile 1 Pile 2
=
=
s
p
E
E
D
L

25 and 50
100 and 1000
Ep/Es = 1000
Ep/Es = 100
L/D = 25
Ep/Es = 1000
Ep/Es = 100
L/D =50
(c) Interaction factors for L/D = 50
(b) Interaction factors for L/D = 25
(a) Piles embedded in semi-infinite soil mass
Figure 5.9 Comparison of interaction factors for piles embedded in semi-finite soil mass
Chapter 5 Soil-Structure Interaction
163
0
0.1
0.2
0.3
0.4
0.5
0.6
0 2 4 6 8 10 12 14
s/D
I
n
t
e
r
a
c
t
i
o
n

f
a
c
t
o
r
,

p
p
h
Present FL
POULOS AND DAVIS (1980)
RANDOLPH (1981)
CHOW (1987a)
0
0.1
0.2
0.3
0.4
0.5
0.6
0 2 4 6 8 10 12 14
s/D
I
n
t
e
r
a
c
t
i
o
n

f
a
c
t
o
r
,

p
p
h
Present FL
POULOS AND DAVIS (1980)
RANDOLPH (1981)
CHOW (1987a)
= = =
s
p
E
E
D
L
, 499 . 0 , 25 80 and 8000
H
L
s
D
Pile 1 Pile 2

(c) Interaction factors for E
p
/E
s
= 8000
(b) Interaction factors for E
p
/E
s
= 80
(a) Piles embedded in homogeneous soil and stiff base layer
= 0
= 90
= 0
= 90
Figure 5.10 Horizontal interaction factors for pile groups embedded in homogeneous soil
Chapter 5 Soil-Structure Interaction
164
D
s
0.3 L
0.3 L
0.4 L
h
H
Case 1 Case 2 Case 3 Case 4
E
s
2E
s
4E
s
4E
s
2E
s
E
s
E
s
2E
s
4E
s
4E
s
2E
s
E
s
1000 , 2 , 25 = = =
s
p
E
E
L
h
D
L

0.0
0.2
0.4
0.6
0.8
1.0
0 2 4 6 8 10 12
s/D
I
n
t
e
r
a
c
t
i
o
n

f
a
c
t
o
r
,

p
p
h
Present FL
FEM
PRAB
0.0
0.2
0.4
0.6
0.8
1.0
0 2 4 6 8 10 12
s/D
I
n
t
e
r
a
c
t
i
o
n

f
a
c
t
o
r
,

p
p
h
Present FL
FEM
PRAB
0.0
0.2
0.4
0.6
0.8
1.0
0 2 4 6 8 10 12
s/D
I
n
t
e
r
a
c
t
i
o
n

f
a
c
t
o
r
,

p
p
h
Present FL
FEM
PRAB
0.0
0.2
0.4
0.6
0.8
1.0
0 2 4 6 8 10 12
s/D
I
n
t
e
r
a
c
t
i
o
n

f
a
c
t
o
r
,

p
p
h
Present FL
FEM
PRAB
(a) Soil profiles for different cases
Case 1 Case 2
Case 3 Case 4
(b) Interaction factors for different cases
Figure 5.11 Soil properties and lateral interaction factors for piles embedded in layered soils
(Kitiyodom and Matsumoto, 2003)
Chapter 5 Soil-Structure Interaction
165
0
0.1
0.2
0.3
0.4
0.5
0.6
0 2 4 6 8 10 12
s/D1
I
n
t
e
r
a
c
t
i
o
n

f
a
c
t
o
r
,

p
p
h
Pile 1 to Pile 2 - Load Pile 1 only
Pile 2 to Pile 1 - Load Pile 2 only
Figure 5.12 Interaction factors for piles of different lengths and same diameters Case 1
(a) Pile arrangement
(b) Vertical interaction factors for loaded piles
(c) Horizontal interaction factors for loaded piles
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0 2 4 6 8 10 12
s/D1
I
n
t
e
r
a
c
t
i
o
n

f
a
c
t
o
r
,

p
p
v
Pile 1 to Pile 2 - Load Pile 1 only
Pile 2 to Pile 1 - Load Pile 2 only
=
=
2
1
2
1
D
D
L
L

2
1
P
L1
s
D2
Pile 1 Pile 2
D1
L2
P
H
H
7500 =
s
p
E
E

s
= 0.3
Chapter 5 Soil-Structure Interaction
166
0
0.1
0.2
0.3
0.4
0.5
0.6
0 2 4 6 8 10 12
s/D1
I
n
t
e
r
a
c
t
i
o
n

f
a
c
t
o
r
,

p
p
h
Pile 1 to Pile 2 - Load Pile 1 only
Pile 2 to Pile 1 - Load Pile 2 only
Figure 5.13 Interaction factors for piles of same lengths and different diameters Case 2
(a) Pile arrangement
(b) Vertical interaction factors for loaded piles
(c) Horizontal interaction factors for loaded piles
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0 2 4 6 8 10 12
s/D1
I
n
t
e
r
a
c
t
i
o
n

f
a
c
t
o
r
,

p
p
v
Pile 1 to Pile 2 - Load Pile 1 only
Pile 2 to Pile 1 - Load Pile 2 only
=
=
2
1
2
1
D
D
L
L

1
2
P
L1
s
D2
Pile 1 Pile 2
D1
L2
P
H
H
7500 =
s
p
E
E

s
= 0.3
Chapter 5 Soil-Structure Interaction
167
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0 2 4 6 8 10 12
s/D1
I
n
t
e
r
a
c
t
i
o
n

f
a
c
t
o
r
,

p
p
v
Pile 1 to Pile 2 - Load Pile 1 only
Pile 2 to Pile 1 - Load Pile 2 only
0
0.1
0.2
0.3
0.4
0.5
0.6
0 2 4 6 8 10 12
s/D1
I
n
t
e
r
a
c
t
i
o
n

f
a
c
t
o
r
,

p
p
h
Pile 1 to Pile 2 - Load Pile 1 only
Pile 2 to Pile 1 - Load Pile 2 only
Figure 5.14 Interaction factors for piles of same lengths and different diameters Case 3
(a) Pile arrangement
(b) Vertical interaction factors for loaded piles
(c) Horizontal interaction factors for loaded piles
=
=
2
1
2
1
D
D
L
L

2
2
L2
P
L1
s
D2
Pile 1 Pile 2
D1
P
H
H
7500 =
s
p
E
E

s
= 0.3
Chapter 5 Soil-Structure Interaction
168
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0 2 4 6 8 10 12
s/D1 or s/D2
I
n
t
e
r
a
c
t
i
o
n

f
a
c
t
o
r
,

p
s
v
Pile 1 to Soil - Load Pile 1
Pile 2 to Soil - Load Pile 2
0
0.1
0.2
0.3
0.4
0.5
0.6
0 2 4 6 8 10 12
s/D1 or s/D2
I
n
t
e
r
a
c
t
i
o
n

f
a
c
t
o
r
,

p
s
v
Pile 1 to Soil - Load Pile 1
Pile 2 to Soil - Load Pile 2
Figure 5.15 Interaction factors for a loaded pile and soil surface
(a) Pile arrangement
(b) Vertical interaction factors for pile and soil
(c) Horizontal interaction factors for pile and soil
=
=
2
1
2
1
D
D
L
L

2
2
P
L1
s
Pile 1
D1
H
P
s
D2
Pile 2
L2
H
7500 =
s
p
E
E

s
= 0.3
Chapter 5 Soil-Structure Interaction
169
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0 2 4 6 8 10
s/S1
I
n
t
e
r
a
c
t
i
o
n

f
a
c
t
o
r
,

s
p
v
Soil to Pile 1
Soil to Pile 2
0.00
0.05
0.10
0.15
0.20
0.25
0 2 4 6 8 10
s/S1
I
n
t
e
r
a
c
t
i
o
n

f
a
c
t
o
r
,

s
p
v
Soil to Pile 1
Soil to Pile 2
Figure 5.16 Interaction factors for a loaded soil surface and a pile
(a) Different sizes of piles considered
(b) Vertical interaction factors for soil and pile
(c) Horizontal interaction factors for soil and pile
=
=
2
1
2
1
D
D
L
L

2
2
L1
s
Pile 1
D1
P
H
S1
s
D2
Pile 2
L2
H
P
S1
7500 =
s
p
E
E

s
= 0.3
Chapter 5 Soil-Structure Interaction
170

Figure 5.17 Interaction factors for a soil surface load and a point on the soil surface
(a) Rectangular load and a point on the soil surface
0
0.1
0.2
0.3
0.4
0 2 4 6 8 10 12
s/S1
I
n
t
e
r
a
c
t
i
o
n

f
a
c
t
o
r
,

s
s
v

0
0.1
0.2
0.3
0.4
0 2 4 6 8 10 12
s/S1
I
n
t
e
r
a
c
t
i
o
n

f
a
c
t
o
r
,

s
s
h
(b) Vertical interaction factors for a soil surface load and a point on the soil surface
(c) Horizontal interaction factors for a soil surface load and a point on the soil surface
s
P
H
S1
E
s
= 4 MPa

s
= 0.3
Chapter 6 Analysis of Piled Rafts
171
CHAPTER 6
ANALYSIS OF PILED RAFTS

6.1 INTRODUCTION
As the use of piled raft foundations as an alternative to conventional piled foundation for
tall buildings has been increasing, different techniques have been developed for
performing analyses over the last decade. Piled rafts are composite structures comprised
of the piles, raft and soil. Such foundations will be subjected to the vertical loadings
transferred directly from the structure and horizontal loadings mostly due to wind loads.
These loads are transferred to the soil through the raft and the piles. Unlike the
conventional piled foundation design in which the piles are designed to carry the majority
of the load, the design of a piled-raft foundation allows the load to be shared between the
raft and piles and it is necessary to take the complex soil-structure interaction effects into
account.

Davis and Poulos (1972) developed an approach to analyse the piled raft system by
considering interaction between pile-cap units. Hain and Lee (1978) presented a method
which considered the foundation system as a flexible elastic plate supported by a group of
compressible friction piles. This method took into account the interactions between the
piles, raft and soil. However, the connection between the raft and the pile was assumed to
be a sliding ball joint which implied that no moments or lateral forces were transferred
between the raft and pile heads. Poulos (1994) developed an approximate method where
the raft was modeled as a thin elastic plate and the piles were modeled as a series of
springs. Three-dimensional finite element techniques have been used by Katzenbach and
Reul (1997) and Reul and Randolph (2003 and 2004). All of the above methods with the
exception of the three-dimensional finite element methods may only deal with piled raft
foundations subjected to vertical loads but not the horizontal loads.

Chapter 6 Analysis of Piled Rafts
172
Small and Booker (1986) developed a finite layer technique to determine soil deflections
of horizontally layered soil. This technique was used for the analysis of piled raft
foundations under vertical loads by Ta (1996) and extended to horizontal loads by Zhang
(2000). Program PIRAF (Ta, 1996) was developed for the analysis of vertically loaded
piled rafts and Program APRAF was developed for the analysis of horizontally and
vertically loaded piled rafts.

In this chapter, the method used by Ta (1996) and Zhang (2000) is briefly discussed. A
program APRILS which is based on an improved approach is then presented. Two
methods of analysis are also examined, called Method I and Method II. In Method I, the
analysis is separated into an isolated raft and a pile group embedded in a horizontally
layered soil. In Method II, the analysis is separated into a piled raft and the horizontally
layered soil.

6.2 METHODS OF ANALYSIS
For a piled raft foundation, the raft is subjected to vertical and horizontal external
loadings and moments as shown in Figure 6.1a. These loadings are then transferred as
interface forces from the raft to the pile and the soil through (i) the contact and shear
stresses at the raft-soil interface, (ii) the shear and lateral earth pressure along the pile
shaft, and (iii) the contact and shear stresses at the pile base as shown in Figure 6.1b. The
contact stresses at the raft-soil interface are represented by rectangular blocks of uniform
pressure and the shear stresses are represented by a series of uniform rectangular shear
stresses. The shear and lateral earth pressure along the pile shaft are approximated as a
series of ring loads applied at the soil interface in both vertical and horizontal directions.
The stresses at the pile base are approximated as a uniform circular load in both
horizontal and vertical directions with the same diameter as the pile base. The raft and
piles are analysed by the finite element method and the layered soil is analysed by the
finite layer method. The analysis takes into account the soil-structure interaction as
discussed in Chapter 5.
Chapter 6 Analysis of Piled Rafts
173
6.2.1 Method I Isolated Raft and Pile Group Embedded in Layered
Soil
Program PIRAF was developed by Ta (1996) based on this method to analyse piled raft
systems subjected to vertical loads and moments in the x- and y-directions. Program
APRAF (Analysis of Piled Raft Foundations) was developed by Zhang (2000) based on
the same method to analyse piled rafts subjected to concentrated or uniform loads and
moments in the x-, y-, and z-directions. The raft can be in contact with the ground surface
or free of the ground surface (i.e. not in contact). The method for the analysis of a piled
raft subjected to both horizontal and vertical loadings is discussed in this section in detail.

The analysis of the foundation system was carried out by separating the foundation
system into two components (i) an isolated raft and (ii) a pile group embedded in a
horizontally layered soil as shown in Figure 6.2a. The raft was assumed to be a thin
elastic plate and analysed by the finite element method. The raft was divided into
rectangular elements with each pile head fitted within one of the raft elements as shown
in Figure 6.2b. The displacements at the centre of the raft elements due to the applied unit
loads can be determined from the method described in Section 4.2 in Chapter 4. For
horizontally and vertically loaded piled rafts (Program APRAF), two nodes of the raft
have to be pinned as shown in Figure 6.2b. Node 1 is completely fixed in all directions
and node 2 is fixed in the y-direction only to resist the raft from rotating about the z-axis.
However, for vertically loaded piled rafts (Program PIRAF) only one node (node 1) of
the raft has to be completely fixed in all directions. The influence matrix of the pinned
raft was generated by applying a unit uniform load in the x-, y- and z-directions to all raft
elements in turn and for those elements containing a pile, a unit moment in the x- and y-
directions in addition to the unit uniform load has to be applied in turn. The unit moment
applied at the centre of the raft element was transformed into equivalent forces applied to
the edges of the element.

The pile group was analysed by the finite element method and each pile was divided into
a number of beam elements corresponding to the soil layers. Four interactions among the
piles and the soil described in Chapter 5 were considered. These interactions are the
Chapter 6 Analysis of Piled Rafts
174
deflections at the pile head or at the centre of the soil surface element due to a unit load
or moment at the pile head of another pile or due to a uniform pressure on another soil
surface element. These deflections then form the columns of the pile-soil influence
matrix. The forces and moments acting on the raft were transferred to the pile heads as
interface forces as shown in Figure 6.3. These interface forces consisted of the loads and
moments applied to the pile heads and loads applied onto the soil surface. The deflections
and rotations at the pile heads or at the centre of the soil surface elements were obtained
by superposition. Combining the piles and raft influence matrices and taking into account
the compatibility of displacements and equilibrium of interaction forces, the load-
deformation behaviour of the piled raft can be expressed as
([I
r
] + [I
sp
]){P
sp
} aD
x
bD
y
cD
z
d
x
e
y
f
z
=
r0
(6.1)
where
[I
r
] = influence matrix of the pinned raft
[I
sp
] = influence matrix of the pile enhanced soil continuum
[P
sp
] = vector of interface loads between the raft and the pile-enhanced soil
{
r0
} = vector of the displacements at the centres of the pinned raft elements
containing a pile due to the loads applied to the pinned raft
D
x
, D
y
, D
z
,
x
,
y
,
z
are the translations and rotations of the pinned raft
{a} to {f} are auxiliary vectors dependent on the condition of each raft element.

6.2.2 Method II Piled Raft and the Layered Soil
This method involves the separation of the foundation into an isolated piled raft and the
layered soil. The piles are rigidly attached to the raft at the nodes on the raft as shown in
Figure 6.4a. The raft and piles are analysed by the finite element method. The layered soil
is analysed by the finite layer method.

6.2.2.1 Analysis of the Isolated Piled Raft
The raft is divided into a number of isoparametric rectangular elements and the piles
attached to the nodes on the raft are divided into a series of beam elements. Each of the
raft elements has eight nodes and forty degrees of freedom while each of the pile
Chapter 6 Analysis of Piled Rafts
175
elements has 2 nodes with ten degrees of freedom as shown in Figure 6.4b. The node
numbering for the piled raft system is shown in Figure 6.4c.

Figure 6.5 shows the free body diagram for the piled raft and the layered soil. The raft is
subjected to external vertical and horizontal forces and contact and shear stresses that
exist between the raft and soil. The contact stresses between the raft and soil are
represented by rectangular blocks of uniform pressure and the shear stresses are
represented by a series of uniform rectangular shear stresses. The pile group is subjected
to interface forces between the pile and soil. The actual displacement at the centres of the
raft elements and at the nodes of the piles can be expressed as
{
pr
}

= [I
pr
] {P
pr
}

+ a
x
+ b
y
+ c
z
+ d
x
+ e
y
+ f
z
+
pr0
(6.1)
where
[I
pr
] = influence matrix of the pinned piled raft
{P
pr
} = contact stresses on the raft elements
{
pr0
} = vector of displacements at the centre of each raft element and at the
nodes of each pile for the pinned raft under the applied load

x
= horizontal translation of the pinned raft in the x-direction

y
= horizontal translation of the pinned raft in the y-direction

z
= vertical translation of the pinned raft in the z-direction

x
= rigid body rotation of the pinned raft in the x-direction

y
= rigid body rotation of the pinned raft in the y-direction

z
= rigid body rotation of the pinned raft in the z-direction
a = (1, 0, 0, 1, 0, 0, )
T

b = (0, 1, 0, 0, 1, 0, )
T

c = (0, 0, 1, 0, 0, 1, )
T

d = (0, 0, x
r1
x
p
, 0, 0, x
r2
x
p
, , 0, 0, x
rn
x
p
, -(z
i
z
1
), 0,
x
i
x
p
, , -(z
m
z
1
), 0, x
m
x
p
)
T

e = (0, 0, y
r1
y
p
, 0, 0, y
r2
y
p
, , 0, 0, y
rn
y
p
)
T
f = (y
r1
y
p
, x
r1
x
p
, 0, y
r2
y
p
, x
r2
x
p
, 0, , y
n
y
p
, x
n
x
p
, 0)
T
x
p
, y
p
= co-ordinates of the pin
x
ri
, y
ri
= co-ordinates of the centres of each raft element
Chapter 6 Analysis of Piled Rafts
176
x
i
, y
i
= co-ordinates of the nodes of each pile
z
1
= co-ordinates of the first node of the pile

As for the method used in Chapter 4 for the analysis of a raft, the separate piled raft is
analysed by restraining rigid body translations and rotations. There is done by choosing
two corner nodes of the piled raft as the points of restraint as shown in Figure 6.6. Point 1
(the first node of the piled raft) is completely fixed in all directions and Point 2 (the last
node of the piled raft) is fixed in the x-direction only to resist the piled raft from rotating
about the z-axis.

Assume the piled raft consists of n raft elements and m nodes for the piles. The influence
matrix of the pinned piled raft can be generated as follows:
(1) Form the total stiffness matrix of the piled raft by the addition of the element
stiffness matrices of the raft elements and pile elements
(2) Apply a unit uniform load in the x-direction to the first element of the piled raft and
calculate the x-, y- and z-displacements at the centre of each element on the raft and
at each node of the piles, this forms the first column of the influence matrix
(3) As for step 2, apply a unit uniform load in the y-direction to the first element of the
piled raft to obtain the second column of the matrix
(4) Apply a unit uniform load in the z-direction to the first element of the piled raft to
obtain the third column of the matrix
(5) Repeat steps 2-4 by applying uniform load to each raft element in turn to obtain the
columns of the matrix due to loading the raft elements (i.e up to 3 x n columns)
(6) Apply a unit load in the x-direction to the first node of the first pile and calculate
the x-, y- and z-displacements at the centre of each element on the raft and at each
node of the piles, this forms the column (3n + 1) of the influence matrix
(7) Apply a unit load in the y-direction to the first node of the first pile to obtain the
column (3n + 2) of the influence matrix
(8) Apply a unit load in the z-direction to the first node of the first pile to obtain the
column (3n + 3) of the influence matrix
Chapter 6 Analysis of Piled Rafts
177
(9) Repeat steps 6-8 by applying a unit load to each node on the piles in turn to
construct the columns of the influence of the matrix due to loading the pile nodes
(10) The size of the influence matrix of the pinned piled raft [I
pr
] would be 3(n + m) x
3(n + m) and the matrix is in the form of:
where D
ss
, D
sp
, D
ps
and D
pp
are the sub-matrices defined as
D
ss
= deflections at the centres of the raft elements under unit uniform loads in the x-, y-
and z-directions on each of the raft elements
D
sp
= deflections at the nodes of the piles under unit uniform loads in the x-, y-, and z-
directions on each of the raft elements
D
ps
= deflections at the centres of the raft elements under unit loads in the x-, y-, and z-
directions at each node of the piles
D
pp
= deflections at the nodes of the piles under unit loads in the x-, y- and z-directions
at each node of the piles
The vector of the displacements at the centres of the raft elements for the pinned raft {
pr0
}
under applied load can be obtained simply by applying the load to the raft and calculating
the displacements at the centres of the raft elements and at the nodes of the piles, the size
of the vector is 3 (n + m) and in the form of:

pr0
= (
rx1
,
ry1
,
rz1
,
rx2
,
ry2
,
rz2
,,
rxn
,
ryn
,
rzn
,
px(3n+1)
,
py(3n+1)
,
pz(3n+1)
,,

px3(n+m)
,
py3(n+m)
,
pz3(n+m)
)
T

where
ri
= displacements at the centres of the raft elements

pi
= displacements at the nodes of the piles



D
ss D
ps
D
pp
D
sp

=
pr
I
(6.2)
Chapter 6 Analysis of Piled Rafts
178
6.2.2.2 Analysis of the Layered Soil
The analysis of the layered soil is carried out by the finite layer method as discussed in
Chapter 3. The soil is divided into layers corresponding to the elements of the piles and
the soil surface consists of rectangular elements corresponding to the elements on the raft.
Figure 6.5 shows that the loads acting on the soil are transferred from the piled raft to the
soil which consist of (i) the contact and shear stresses at the raft-soil interface which are
represented by rectangular blocks of uniform pressure, (ii) the shear and lateral earth
pressures along the pile shafts which are approximated as a series of ring loads in both
horizontal and vertical directions applied at the soil interface between each layer and (iii)
the normal contact and shear stresses at the pile bases which are approximated as uniform
circular loads with the same diameter as the pile bases. If the number of the ring loads is
large enough, the approximation to the continuous contact and shear stress along the pile
shaft is sufficiently close to give accurate results for the soil behaviour.

The load-deflection relationship may be expressed as

s
= [I
s
] P
s
(6.3)
where

s
= vector of soil deflections in the x-, y- and z-directions due to the forces
acting on the raft-soil and pile-soil interfaces
I
s
= influence matrix for the layered soil
P
s
= vector of forces acting on the raft-soil and pile-soil interfaces

Assume the soil surface consists of n surface elements and the soil media consists of m
interface nodes which correspond to the number of nodes on the piles. The influence
matrix for the soil can be obtained by the following steps:
(1) Apply a unit uniform load in the x-direction to the first soil surface element and
calculate the deflections of the soil at the centres of all elements on the soil surface
and at the nodes of the soil layer interfaces. These are the entries for the first
column of the matrix
(2) Apply a unit uniform load in the y-direction to the first soil surface element to
obtain the second column of the matrix
Chapter 6 Analysis of Piled Rafts
179
(3) Apply a unit uniform load in the z-direction to the first soil surface element to
obtain the third column of the matrix
(4) Repeat steps 1 to 3 by applying uniform load to each of the soil surface elements in
turn to obtain the columns of the matrix due to loading the surface elements (i.e up
to 3 x n columns)
(5) Apply a unit ring load in the x-, y- and z-directions in turn to the first node on the
layer interface and calculate the deflections in the three directions at the centres of
all soil surface elements and at all nodes at the layer interfaces. This forms the first
three columns of the influence matrix (i.e columns (3n + 1) for loading in the x-
direction, (3n + 2) for the y-direction and (3n + 3) for the z-direction). Repeat the
step by applying the ring load to each node in turn. For the last node at the pile base
interface, a unit uniform circular load in the x-, y- and z-directions is applied in turn
to obtain the corresponding columns for the matrix.
(6) Repeat step 5 for each interface node to construct the influence matrix.

The size of the soil influence matrix would be the same as the piled raft influence matrix
and is of the following form:
where S
ss
. S
sp
, S
ps
and S
pp
are the sub-matrices defined as
S
ss
= deflections at the centres of the surface elements under unit uniform loads in the
x-, y- and z-directions on each of the surface elements
S
sp
= deflections at the interface nodes of the soil under unit uniform loads in the x-, y-,
and z-directions on each of the surface elements
S
ps
= deflections at the centres of the surface elements under unit ring loads in the x-, y-,
and z-directions at each interface node of the soil
S
ss S
ps
S
pp
S
sp

=
s
I
(6.4)
Chapter 6 Analysis of Piled Rafts
180
S
pp
= deflections at the interface nodes of the soil under unit ring loads in the x-, y- and
z-directions at each interface node of the soil

6.2.2.3 Analysis of a Piled Raft in a Layered Soil
Assume no slip occurs along the pile-soil interface, the compatibility of the
displacements requires equal displacements of the soil and the piled raft, i.e
pr
=
s
and
the equilibrium of forces at the piled raft and the soil is such that the forces are equal in
magnitude but of opposite sign, i.e P = P
pr
= -P
s
. Combining equations (6.1) and (6.3)
will give:
([I
pr
] + [I
s
]){P} a
x
b
y
c
z
d
x
e
y
f
z
=
pr0
(6.5)
a
T
{P} = P
x
(6.6a)
b
T
{P} = P
y
(6.6b)
c
T
{P} = P
z
(6.6c)
d
T
{P} = M
x
(6.6d)
e
T
{P} = M
y
(6.6e)
f
T
{P} = M
z
(6.6f)
where
a
T
= (A
1
, 0, 0, A
2
, 0, 0, , A
n
, 0, 0, 1, 0,0, 1, 0, 0, )
b
T
= (0, A
1
, 0, 0, A
2
, 0, , 0, A
n
, 0, 0, 1, 0, 0, 1, 0, )
c
T
= (0, 0, A
1
, 0, 0, A
2
, , 0, 0, A
n
, 0, 0, 1, 0, 0, 1, )
d
T
= (0, 0, A
1
(x
r1
- x
p
), 0, 0, A
2
(x
r2
- x
p
), , 0, 0, A
n
(x
rn
- x
p
), -(z
i
- z
1
),
0, (x
i
-x
p
), , -(z
m
z
1
), 0, (x
m
x
p
))
e
T
= (0, 0, A
1
(y
r1
y
p
), 0, 0, A
2
(y
r2
y
p
), , A
n
(y
rn
y
p
), 0, -(z
j
z
1
),
(y
i
y
p
), , 0, -(z
m
z
1
), (y
m
y
p
))
f
T
= (A
1
(y
r1
y
p
), -A
1
(x
r1
x
p
), 0, A
2
(y
r2
y
p
), -A
2
(x
r2
x
p
), 0, ,
A
n
(y
rn
y
p
), -A
n
(x
rn
x
p
), 0, (y
j
y
p
), -(x
j
x
p
), 0, , (y
m
y
p
),
-(x
m
x
p
))
x
p
, y
p
= co-ordinates of the pin
x
ri
, y
ri
= co-ordinates of the centres of each raft element
x
i
, y
i
= co-ordinates of the nodes of each pile
z
1
= co-ordinates of the first node of the pile
Chapter 6 Analysis of Piled Rafts
181
A
i
= area of element i in the raft
P
x
, P
y
, P
z
are the total external loads applied to the piled raft in the x-, y- and z-
directions, respectively.
M
x
, M
y
, M
z
are the total external moments applied to the piled raft about the x, y
and z axes, respectively.

The interface forces on the soil surface elements and along the pile shaft and the rigid
body translations and rotations can be solved by using equations (6.5) to (6.6f). The
results are then substituted into equation (6.1) to calculate the displacements at the nodes
of the piled raft.

Program APRILS (Analysis of Piled Rafts In Layered Soils) was developed based on this
method for the analysis of rafts and piled rafts subjected to concentrated loads, uniform
or non-uniform loads, and moments in the x, y and z directions. The raft can be of any
shape and have different thicknesses and the piles can have different lengths and
diameters.

6.2.3 Comparison of the Two Methods
The two methods discussed in the previous section are conceptually the same. The
differences between the two methods are summarised in Table 6.1.
Method I Method II (APRILS)
Separation of
Foundation
(1) Isolated raft, (2) Pile
group and layered soil
(1) Isolated piled raft, (2)
layered soil
Pile Connection Attached to the raft element Attached to the raft at the node
Load applied at the Pile
Head
Uniform pressures Concentrated loads
Table 6.1 Differences between Methods I and II

Chapter 6 Analysis of Piled Rafts
182
Method II has several advantages over method I:
(1) The piles are attached to the raft at the nodes which reduces the size of the mesh in
terms of the total number of nodes and elements which then reduces the number of
equations that need to be solved.
(2) The raft and piles are analysed as a whole structure such that the applied loads and
moments are directly transferred from the raft to the pile. Therefore, the
transformation of the applied moments into equivalent forces acting along the edges
of the raft elements is not necessary.
(3) The influence matrix of the piled raft is generated by applying unit uniform load to
the raft elements or unit load to the nodes of the piles in turn without the need to
apply unit moment which reduces the computation time for matrix formation.
(4) Better simulation of the loads at the pile heads can be obtained for the analysis of
large piled rafts as the loads behave as concentrated loads instead of uniform
pressures at the pile head.

In the following example, a piled raft was analysed by both methods and results were
compared. A square piled raft is supported by 16 piles of identical size and length. The
raft has a size of 8 x 8m and a thickness of 1m. The piles are 10m long with a diameter of
0.5m. The overall depth of the soil is assumed to be 2 x pile length (i.e. 20m). The
configuration and pile arrangement of the piled raft is shown in Figure 6.7. The modulus
and Poissons ratio of the soil are taken as 10MPa and 0.3 respectively. The raft and piles
have the same modulus of 25,000MPa and Poissons ratio of 0.3. For both analysis, each
pile was divided into 10 elements and the soil was divided into 20 layers. Figures 6.8a
and b show the finite element meshes for the piled raft used in the Method I and II
analyses. In the Method I analysis, the raft was divided into 144 elements and the element
which contains the pile has to have the same area as the pile cross-sectional area. In the
Method II analysis, the raft was divided into 64 elements which was approximately half
of the number of elements used in the Method I analysis. The use of fewer elements in
the analysis means that the number of equations that need to be solved is reduced thus
requiring less time for computation.

Chapter 6 Analysis of Piled Rafts
183
Figures 6.9a and b show the contours of settlement obtained from Method I and Method
II analyses. Figure 6.10a shows the comparison between the settlement along the centre
line of the raft for both methods. The settlement obtained by the Method II analysis was
about 6% less than that obtained from the Method I analysis. The axial load distributions
along the centre (P1) and corner (P2) piles are shown in Figure 6.10b. The load
distribution for the corner piles from both analyses are in good agreement except at the
pile head. For the pile at the centre, the load distribution obtained from Method II was
smaller than that obtained from Method I. However, the differences in the load
diminished with depth along the pile. For both corner and centre piles, the differences of
load at the top of the piles are larger than at the base of the piles, this implies that the
influence of the type of the load applied at the pile head would become less significant as
the embedment depth of the pile increases. As the depth between the load at the pile head
and the pile node of interest increases, the effect of the uniform load or concentrated load
at the pile head on the load in the pile will approach a similar value. The contours of
bending moment in the x-direction for Methods I and II are shown in Figures 6.11a and b
respectively. The contours are slightly different especially at the pile heads where
different pile-raft connection exists. Figure 6.11c shows the comparison between the two
methods of the moments in the x-direction at the Gauss points close to the centre line of
the raft and it may be seen that the two results are in good agreement.

The difference in the results obtained by both analyses could be due to (i) different
integration schemes used in the analyses as the size of the integration blocks is
determined from the average size of the raft elements, (ii) the load applied at the pile
head is represented by a concentrated load in the Method II analysis, (iii) the piled raft is
analysed as a single structure in Method II so that forces and moments acting on the raft
are directly transmitted to the piles while in the Method I analysis, the raft is separated
from the pile group and the moments acting on the raft have to be transformed into
equivalent forces.

Chapter 6 Analysis of Piled Rafts
184
6.2.4 Non-Linear Analysis of Piled Raft
The typical load-settlement behaviour of a piled raft is shown in Figure 6.12. When the
piled raft is loaded below P
e
, both the piles and raft are behaving elastically. When the
piled raft is loaded between P
e
and P
u
, the interface between the pile and soil starts to slip.
When the pile capacity is fully mobilised, the load in excess of P
e
is carried by the raft
only. As the load reaches P
u
, both the piles and raft have reached their ultimate capacities
and fail to carry additional loads. Such non-linear behaviour of the foundation can be
simulated by various approaches: (i) assumption of a hyperbolic load-settlement curve
(Russo, 1998; Mandolini and Viggiani, 1997) (ii) limiting the load carried by the raft and
piles (Horikoshi and Randolph, 1998) (iii) a soil model consisting of yield surface
segments, eg. Drucker-Prager yield criterion (Reul and Randolph, 2003; Maharaj and
Gandhi, 2004).

Program APRILS has the capability to take into account the non-linear response of the
pile by limiting the loads acting on the pile-soil interface to simulate the slip that occurs
at the interface. The analysis is implemented through an incremental-iterative process.
The load is applied in increments and the forces acting on the raft-soil and pile-soil
interfaces are computed for each increment. These interface forces are then compared
with the limiting contact pressures acting on the raft or the limiting ring loads along the
pile shaft and base load at the pile base. These limiting loads are computed from the shear
strength of the soil, s
u
. The limiting ring loads along the pile shaft for pile element i (for a
pile in clay), LR
i
, is
LR
i
= c
a
C z (6.7)
and the limiting base load at the pile base, LB, is
LB = N
c
s
u
A (6.8)
where c
a
= pile-soil adhesion = s
u
x
C = circumference of pile
z = length of pile element i
= adhesion factor and is a function of s
u

N
c
= bearing capacity factor (~ 9 for piles in clay)
A = area of pile base
Chapter 6 Analysis of Piled Rafts
185
6.2.5 Verification of Results with Other Numerical Solutions
Program APRILS based on Method II will be used in the following examples except for
example 1 in which Method I was used and checked against an analysis using the
program GARP.
Example 1 Raft supported by a 5x5 pile group in homogeneous soil
In this example, a square raft on a 5x5 pile group as shown in Figure 6.13 was analysed
by the program APRILS based on both method I and II as stated above. The method I
results were checked against the program GARP analysis while the results from the
program APRILS were checked against a three-dimensional finite element analysis.
Properties of the piled raft and the soil are listed in Table 6.2. The piled raft was
subjected to a vertical uniform load of 50kPa and the overall depth of the soil was taken
as 2 times the length of the pile.

Method I and GARP analyses (smooth raft base)
Figures 6.14a and b show the meshes of the raft for the GARP and (Method I) analyses.
In the GARP analysis, the piles are attached to the raft at the nodes while in the Method I
analysis, each pile is contained within a raft element. For the GARP analysis, the
interaction factors were computed up to a distance of s/d = 30. The influence between a
raft element and a pile was calculated at a depth of one-third of the pile length and
therefore the depth of influence factor, , was taken as 0.333. For both analyses, the raft
is isolated from the pile group and the base of the raft is treated as being smooth.

Results of the analysis are shown in Figure 6.15a for the settlement along the centre line
of the raft and in Figure 6.15b for the pile load at different locations. The settlement and
pile loads estimated by Method I are higher than those estimated by GARP in which the
settlement is overestimated by about 7%. As shown in Figure 6.15b, the piles at the
corners carried the maximum load followed by the piles at the edges and the pile at the
centre carried the minimum load. The contours of moments in the x-direction for the
GARP and Method I analyses are shown in Figures 6.16a and b respectively. Moments
computed at Gauss points close to the piles along row 1 and the centre row are shown in
Figures 6.17a and b. The moments obtained from Method I are in excellent agreement
Chapter 6 Analysis of Piled Rafts
186
with those from the GARP analysis. The differences in the settlement and pile load could
be due to the fact that the load at the pile head is treated as a concentrated load in the
GARP analysis and as a uniform pressure in the Method I analysis.

Parameters Value
Modulus of soil 10MPa
Poissons ratio of soil 0.3
Modulus of raft and piles 20,000MPa
Poissons ratio of raft and piles 0.2
Length of piles 20m
Diameter of piles 1.128m
Spacing of piles 4m
Overlap of cap at edges of raft 1m
Thickness of raft 1m
Table 6.2: Parameters used for the analysis

APRILS and finite element analyses (rough raft base)
The mesh used for the finite element analysis is shown in Figure 6.18a. One quarter of
the piled raft was analysed because of the symmetry of the example. Infinite elements
were used at the edges so that the solutions would not be affected by the location of the
side boundaries. The piles are shown as material 2 in Figure 6.18a which are modelled by
square elements with the same cross-sectional area as used for the APRILS and GARP
analyses such that the amount of vertical compression in the piles for different analyses
will be the same. The cross-sectional area of the pile was taken as 1m
2
which meant that
the square pile in the finite element analysis has a side width and breadth of 1m. Figure
6.18b shows the mesh of the piled raft for the APRILS analysis in which the piles are
attached to the raft at the node and the piled raft is considered as a whole system in the
analysis. The diameter of the pile was taken as 1.128m which has the same area as a pile
in the finite element analysis. For both the finite element and the APRILS analyses, the
base of the raft is treated as being rough.

Chapter 6 Analysis of Piled Rafts
187
Comparisons of the horizontal displacement and vertical settlement along the centre line
of the raft are shown in Figures 6.19a and b respectively. From the APRILS analysis, the
vertical settlement is about 8% more than the FE analysis, while the horizontal
displacement is in good agreement with the finite element analysis. The bending
moments at different positions of the raft are shown in Figure 6.20. The moments
obtained from APRILS are plotted close to the centre of the piles, and so high bending
moments are observed at the pile positions. These high bending moments are due to the
use of a concentrated load applied at the pile head. When the moments are plotted further
from the centre of the piles, the high bending moments reduced significantly. As uniform
pressures are applied at the pile head in the finite element analysis, the moments plotted
at different distances from the centre of the piles do not change significantly.

The responses of the piles in terms of compression and load distribution along the piles at
different locations are shown in Figures 6.21a and b respectively. The compression of the
piles obtained from the APRILS analysis are in good agreement with those from the finite
element results whereas the load distribution along the piles is slightly different from the
finite element results. Such differences could be due to the fact that the connection of the
pile to the raft at the node in APRILS is different from the finite element analysis.

Comparison between smooth and rough raft base
The base of the raft interface can have significant effect on the behaviour of the
foundation system. In the analysis of a piled raft with a smooth raft base, only the vertical
resistance is considered, whereas both the lateral and vertical resistance are considered in
the analysis of a piled raft with a rough raft base. By considering the lateral resistance in
the analysis of a vertically loaded piled raft, the lateral movement of the piles can be
reduced and this results in the reduction in the overall settlement of the raft as shown in
Figure 6.22a as well as leading to a reduction in the bending moments as shown in Figure
6.22b.

Chapter 6 Analysis of Piled Rafts
188
Example 2 Piled raft in homogeneous soil subjected to horizontal uniform load
In this example, the behaviour of a piled raft subjected to horizontal loadings is examined.
The piled raft in the previous example was subjected to a horizontal uniform load of
50kPa in the x-direction. The results obtained from the APRILS analysis were compared
with those from a finite element analysis.

Figure 6.23a shows that APRILS overestimated the horizontal displacement by about
11%, however, the vertical displacement is in excellent agreement with the finite element
analysis as shown in Figure 6.23b. It is clearly shown that when the raft is subjected to
horizontal loadings, the raft would undergo large displacement in the direction of the load
and the displacements in the vertical directions are relatively small.

The behaviour of the piles at different locations due to a horizontal load is shown in
Figures 6.24 and 6.25. The horizontal displacement in the x- and y-directions and the
compression (relative to the pile base) along the pile at three different locations (centre
P1, edge P2 and corner P3) are shown in Figures 6.24a, b and c respectively. When a
horizontal load in the x-direction is applied to the raft, the pile at the centre only displaces
in the direction of the applied load while the piles at the edges displace horizontally in the
direction of the applied load as well as vertically and the piles at the corners displace in
all three directions. The horizontal x-displacement (Figure 6.24a) and compression
(Figure 6.24c) of the piles obtained from the APRILS analysis are slightly different from
the finite element analysis while the horizontal y-displacements (Figure 6.24b) along the
pile are in excellent agreement. The comparisons of bending moments and shear forces
along the piles between the APRILS and finite element analyses are shown in Figures
6.25a and b respectively. The piles at the corners (P3) carry the maximum bending
moments and shear forces followed by the piles at the edges (P2) and the pile at the
centre carries the minimum. The bending moments and shear forces obtained from the
APRILS analysis are in good agreement with the finite element analysis as shown in
Figure 6.25.

Chapter 6 Analysis of Piled Rafts
189
Example 3 Square raft supported by a 2x2 pile group
In this example, a square raft supported by a 2x2 pile group as shown in Figure 6.26a is
analysed by APRILS and results are compared with those obtained from a computer
program PRAB as described in Chapter 4 and the finite element method as presented by
Kitiyodom and Matsumoto (2003). The foundation was subjected to either a vertical (P
v
)
or horizontal (P
H
) load and five cases with different soil profiles as shown in Figure 6.26b
were analysed. Except for case 1 where the embedded soil consisted of homogeneous
layers, the foundation was embedded in non-homogeneous soil layers for all cases.
Program PRAB and the finite element method used by Kitiyodom and Matsumoto (2003)
have been discussed in Section 4.4.2. The modulus of the raft was taken as 23970MPa
which was calculated from the raft-soil stiffness ratio, K
rs
, as defined in Equations (4.45)
by assuming that E
s
= 25MPa and K
rs
= 10. The results are presented in terms of the
dimensionless parameters (I
v
, C
a
, I
H
, C
s
and C
b
) as defined in Equations (4.40) to (4.44)
in Chapter 4.

The responses of the piles due to the vertical or horizontal loads for different cases are
shown in Figures 6.27 to 6.31. All four piles underneath the raft have the same
displacements and were carrying equal loads. The vertical displacement and axial force
along the pile are presented for piled rafts subjected to vertical loading while the
horizontal displacement, shear force and bending moments along the piles are presented
for piled rafts subjected to horizontal loading. For all cases subjected to the horizontal
load, the results obtained from APRILS are in good agreement with those from PRAB
and the finite element method. However, for cases subjected to the vertical load, APRILS
tends to overestimate the vertical displacement along the pile but shows good agreement
with the axial force along the pile from PRAB and the finite element method.

Under vertical load, the vertical displacement factors I
v
decrease as the modulus of the
soil at the pile base increases. For cases 4 and 5 with a stiff soil as the top layer underlaid
by a soft soil at the bottom, the displacement factors I
v
are larger than for cases 2 and 3
which indicates that the stiffness of the soil at the pile base contributes more to the
resistance of the vertical movement of the pile. The axial force factor C
a
is dependent on
Chapter 6 Analysis of Piled Rafts
190
the stiffness of the soil layers as shown in Figures 6.27(b) to 6.31(b) as the soil at the pile
base gets softer, the base load of the pile is smaller. Under horizontal load, the stiffness of
the soil layer close to the pile head has significant effect on the lateral resistance of the
horizontal movement of the pile which in turn has influences on the bending moments
and the shear forces. With a stiff layer at the pile head (cases 4 and 5), the horizontal
displacement factor I
H
decreases significantly compared with the other cases.

Example 4 Rectangular piled raft subjected to concentrated loads
In this example, the effects of the number of piles, pile length and raft thickness on the
behaviour of the rectangular piled raft are examined (Van Impe, 2001). Results from both
linear and non-linear analyses of the piled raft are presented. The rectangular raft with a
size of 10m x 6m and a thickness of 0.5m is supported by piles with a diameter of 0.5m.
The raft has a bearing capacity of 0.3MPa and the pile has a load capacity of 0.873MN in
compression. The piled raft is embedded in a homogeneous soil with a modulus E
s
of
20MPa and the Poissons ratio
s
of 0.3. Figure 6.32 shows the pile arrangement for the
piled rafts and the soil profile for different cases that are considered in this example. A
load of 12MN was applied vertically as concentrated loads to the raft as shown in Figure
6.33. Program APRILS was used for the analysis and results were compared with results
from program GARP and the Poulos-Davis-Randolph (PDR) method (Van Impe, 2001).

Effect of number of piles
To examine the performance of a piled raft with different numbers of piles, three cases
were analysed with the number of piles varying from 3 to 15 as shown in Figure 6.32.
Program APRILS-V which is a simplified version of APRILS was used for the analysis.
Program APRILS-V treats the base of the raft as being smooth and only vertical loads
and moments in the x- and y-directions can be applied to the piled raft. Results from
APRILS-V are compared with those from program GARP and the Poulos-Davis-
Randolph (PDR) method (Van Impe, 2001). The PDR method is a simplified analysis
method which is based on elastic theory to estimate the stiffness of the raft. The pile
group stiffness is approximated by multiplying the group settlement ratio with the
stiffness of a single pile which is obtained by a closed form approximation.
Chapter 6 Analysis of Piled Rafts
191
The load-settlement curves for three different cases are shown in Figure 6.34. There is a
good agreement among the three analyses. By increasing the number of piles from 3 to 9,
the load capacity of the piled raft increases by 21% but for an increase from 9 to 15 piles
only an increase of 18% in the load capacity is obtained. The load capacity of the piled
raft increases as the number of piles increases. However, as the number of piles reaches
the optimum state, further increase in the number of piles will not have a significant
effect on the settlement. This is shown in Figure 6.35 for the load of 12MN where the
optimum state is reached when 15 piles are used, and a further increase in the number of
piles beyond this optimum number of piles will have little contribution in reducing the
settlement.

Effect of pile length Linear and non-linear analysis
The piled raft in case 2 (in Figure 6.33) was used to demonstrate the effect of varying the
pile length on the behaviour of the piled raft. The length to diameter ratio (L/d) varies
from 0 to 50. The results are presented in terms of the maximum settlement, the
differential settlement between the centre and the corner of the raft, the maximum
moment in the raft and the proportion of load carried by the piles.

1. Linear analysis
Figure 6.36 shows a good agreement among the results from the linear analysis which
were carried out for a smooth raft base by APRILS-V, GARP and a rough raft base by
APRILS and the finite element method except for the differential settlement. From the
APRILS-V and GARP analyses, it has been shown that the differential settlement is
sensitive to the size of the raft elements for the ratio L/d < 20. In order to improve the
accuracy of the results, finer meshes with smaller elements along the edges of the raft as
shown in Figure 6.37a were used in both analyses for L/d < 20. For L/d > 20, the results
from a finer mesh as shown in Figure 6.37a will be the same as those from a coarse mesh
as shown in Figure 6.33b. For the APRILS and finite element analyses, accurate results
can be obtained by the use of reasonably sized elements as shown in Figures 6.33b and
6.37b respectively.

Chapter 6 Analysis of Piled Rafts
192
By treating the raft base as being smooth, APRILS-V and GARP tend to overestimate the
differential settlement as shown in Figure 6.36b. This could be due to the fact that there is
no resistance in the lateral movement of the raft especially for short piles. However, when
the raft base is treated as being rough, the lateral resistance of the piles contributes to the
reduction of the vertical movement of the raft. It is clearly shown that as L/d increases,
the differences in the differential settlement between the results from a smooth raft base
and a rough raft base analysis becomes less.

By increasing the pile length, the settlement, differential settlement and maximum
moment decrease while the percentage of overall load carried by the piles increases as
shown in Figure 6.36.

2. Non-linear analysis
The non-linear analysis of the piled raft was carried out by the programs APRILS-V and
GARP. The bearing capacity of the raft and ultimate load capacity of the pile are 0.3MPa
and 0.873MN respectively. In program APRILS-V, the analysis is carried out in
increments by limiting the ring load applied at the soil interface and the contact pressures
on the soil surface as described in Section 6.2.4. In program GARP, the analysis is
carried out by changing the stiffness of the pile in each load increment according to the
following hyperbolic load-settlement curve
2
0
1

=
u
f pile
P
P
R k k (6.9)
where k
pile
= modified pile stiffness
k
0
= original pile stiffness
P = computed load on pile
P
u
= ultimate load capacity of pile
R
f
= hyperbolic factor
and the hyperbolic factor was taken as 0.9 in this example. Figure 6.38 shows that there is
a good agreement between the APRILS-V and GARP analyses except for the proportion
of load carried by the piles. The load distribution is obtained from the computed pile
Chapter 6 Analysis of Piled Rafts
193
displacement and the stiffness of piles which is constant throughout the APRILS-V
analysis, while the stiffness of the pile varies in the GARP analysis.

Figure 6.38 shows that as the pile length increases to a ratio of L/d = 30, the settlement,
differential settlement and maximum moment in the raft decrease while the load carried
by the piles increases. However, for L/d > 30, the effect of increasing the pile length
diminishes.

Comparison of results for the linear and non-linear analyses as shown in Figures 6.36 and
6.38 has shown that in the non-linear analysis, due to the slip along some areas of the pile
shafts, the settlement, differential settlement and maximum moment in the raft are higher
than for the linear analysis while the load carried by the piles are lower.

Effect of raft thickness
The effect of raft thickness on the behaviour of a piled raft can be demonstrated by
analysing the piled raft in case 2 (of Figure 6.33) by varying the raft thickness to pile
diameter ratio (t/d) from 0 to 5. The analyses were carried out by programs ARPILS-V
and GARP. Results from the linear and non-linear analyses are shown in Figures 6.39 and
6.40 respectively. The results from APRILS-V are in good agreement with GARP except
for the load carried by the piles from the non-linear analysis. By comparing Figures 6.39
and 6.40, the settlement from the linear analysis is smaller than that from the non-linear
analysis, whereas the differential settlement and moment for thick rafts are smaller in the
non-linear analysis. The load carried by the piles from the linear analysis is higher than
that in the non-linear analysis. From both analyses, the maximum settlement decreases
rapidly for thin rafts (ie. 0.2 < t/d < 1). However, increasing the raft thickness has no
significant effect for raft thicknesses of t/d > 1. Increasing the raft thickness leads to a
significant decrease in the differential settlement and an increase in the maximum
moment for t/d values < 2, however, it has little effect on the load carried by the piles.

By comparing the linear analysis results in Figures 6.35, 6.36 and 6.39, it can be shown
that (i) increasing the raft thickness is an effective way to reduce differential settlement,
Chapter 6 Analysis of Piled Rafts
194
(ii) increasing the pile length is a more effective way to improve the performance of the
piled raft than increasing the number of piles or the raft thickness.

6.3 CONCLUSIONS
This chapter presents two methods (Method I & II) for the analysis of piled raft
foundations in layered soil. The approach used in Method II is an improvement from the
Method I approach and comparison between the two methods was presented. The
analysis of piled raft foundations based on Method II is implemented by the computer
programs APRILS-V for vertical loading and APRILS for both vertical and horizontal
loadings. Program APRILS-V is capable of carrying out non-linear analysis through the
incremental-iterative process. The examples presented above have demonstrated the use
of the programs for the analyses of several different problems and the following
conclusions can be reached:

(1) By connecting the piles at the nodes of the raft and considering the piles and raft as
a whole structure in Method II, the number of elements used for the analysis could
be reduced which leads to a reduction in the number of equations that needs to be
solved for the problem compared with Method I. As the load acting on the pile
heads is treated as a concentrated load, it can better simulate the loads acting on the
pile heads for large piled rafts.
(2) Solutions obtained from the two approaches were found to have differences in the
settlement and pile load close to the pile heads as different types of load were
applied.
(3) With the use of Method II, high bending moments were observed in the raft at the
locations of the piles, this is due to the fact that a concentrated load is applied.
However, the bending moments decrease significantly with increasing distance
from the pile head.
(4) Solutions obtained from APRILS-V and APRILS are in good agreement with the
solutions of other methods.
Chapter 6 Analysis of Piled Rafts
195
(5) By treating the base of the raft as being rough, the overall settlement and bending
moments of the raft reduce significantly from the smooth base result as the lateral
resistance of the piles reduces the lateral movement of the piles and therefore the
vertical movement of the raft.
(6) For a horizontally loaded piled raft, large displacements of the raft and piles are
observed in the direction of the applied load. The piles at the corners will carry the
maximum shear forces and bending moments.
(7) For the cases examined in example 3 for vertically loaded piled rafts embedded in
non-homogeneous soil, least vertical settlement occurs when the pile base is in the
stiffest layer while for horizontally loaded piled rafts, least horizontal movement
occurs when the pile head is in the stiffest layer.
(8) When the pile is allowed to slip along the pile-soil interface, a pile may be loaded
to its ultimate capacity which results in an increase in the overall settlement,
differential settlement and bending moment of the foundation whereas the load
carried by the piles would be limited to its capacity.
(9) By increasing the number of piles supporting the raft up to the optimum number,
the settlement of the raft can be reduced significantly. In order to improve the
overall performance of a piled raft, increasing the pile length is one of the most
effective and economical ways as the overall settlement, differential settlement, and
bending moments of the raft can be reduced while the load carried by the piles can
be increased.
(10) For the 9 pile group examined here, the thickness of the raft has no effect on the
percentage of load carried by the piles. For thin rafts, increasing the raft thickness
reduces the overall settlement of the raft, however, this is accompanied by a
significant increase in the bending moment. As the raft thickness increases the
differential settlement approaches zero, and for very thick rafts there is no
significant effect on the overall settlement.

Chapter 6 Analysis of Piled Rafts
196



Ring loads
acting on
soil nodes
t
r
P
z
P
x
M
x
External forces
Pile
Interface forces
between the raft
and soil
Interface forces
between the piles
and soil
Circular loads acting
on pile base
(a) External forces and moments acting on a piled raft
(b) Interface forces acting on a piled raft
Figure 6.1 External and interface forces acting on a piled raft
M
x
M
y
P
y
P
z
P
x
M
z
Chapter 6 Analysis of Piled Rafts
197
(a) Isolated raft and pile group embedded in a layered soil
Pile
Elements
Soil Elements
Pin 1 fixed in
all directions
Pin 2 fixed in
y-direction only
y
x
(b) Piles fitted within raft elements of a pinned raft
Figure 6.2 Separation of the raft and pile group and a pinned raft
M
x
P
y
M
y
P
z
P
x
M
z
Chapter 6 Analysis of Piled Rafts
198
Interface forces transferred from raft to pile heads and soil surface
Pile
Soil
{P
sp
}
Ring loads
acting on
soil nodes
Circular loads
acting on pile
base
x
z
t
r
P
z
P
x
M
x
External forces
{P
r
}
Interface forces between
the raft and piles
Interface forces between
the raft and the soil
Figure 6.3 Interface forces acting on the isolated raft and pile group
Chapter 6 Analysis of Piled Rafts
199
M
y
M
x
P
y
P
z
P
x
M
z
Pile
element j
v
j1
u
j1
w
j1

yj1

xj1
v
j2
u
j2
w
j2

yj2

xj2

zi1
v
i1
u
i1
w
i1

yi1

xi1

zi8
v
i8
u
i8
w
i8

yi8

xi8
Raft Element i
Figure 6.4 Analysis of the piled raft system
(a) Piles attached to the raft at the nodes
(b) Degrees of freedom for nodes on the raft and pile elements
17
18
19
20
21
39
40
41
42
43
22
23
33
34
44
45
1
2
9
10
11
14
15
17
18
19
20
21
31 36
(c) Node numbering for the piled raft system
Chapter 6 Analysis of Piled Rafts
200
z
t
r
P
z
P
x
M
x
External forces
{P
r
}
Pile
Soil
Interface forces transferred from piled raft to the soil
Interface forces
between the piles
and soil
Ring loads
acting on soil
nodes
Circular loads
acting on pile base
x
Interface forces
between the raft
and soil
Figure 6.5 Free body diagram for the isolated piled raft and the layered soil
Point 1
fixed in all
directions
Point 2
fixed in x
direction
only
Figure 6.6 Nodal fixity for the analysis of an isolated piled raft
Chapter 6 Analysis of Piled Rafts
201
Figure 6.7 Configuration and pile arrangement of piled raft foundation
Raft: Pile: Soil:
E
r
= 25,000 MPa E
p
= 25,000 MPa E
s
= 10 MPa

r
= 0.3 d = 0.5 m
s
= 0.3
t
r
= 1 m
Rigid Rough
Base
10 kPa
t
r
20 m

10 m

0.5 m

x
2 m
1 m
y
8

m

8 m
P2
P1
Chapter 6 Analysis of Piled Rafts
202
Figure 6.8 Finite element meshes of the piled raft for the Method I & II analyses
(a) Finite element mesh of the piled raft for the Method I analysis
(b) Finite element mesh of the piled raft for the Method II analysis
PLOT OF FINITE ELEMENT MESH
Mesh for Method I analysis
R a ft o n 4x 4 pi l e gr o up
P i l e di a me te r 0.5m
Ep/Es = 2500
PLOT OF FINITE ELEMENT MESH
Mesh for Method II analysis
R a ft o n 4x 4 pi l e gr o up
P i l e di a me te r 0.5m
Ep/Es = 2500
Chapter 6 Analysis of Piled Rafts
203
0. 0 2. 0 4. 0 6. 0 8. 0 10. 0
0. 0
2. 0
4. 0
6. 0
8. 0
0. 0 2. 0 4. 0 6. 0 8. 0 10. 0
0. 0
2. 0
4. 0
6. 0
8. 0
0. 0 2. 0 4. 0 6. 0 8. 0 10. 0
0. 0
2. 0
4. 0
6. 0
8. 0
CONTOURS OF VERTICAL DISPLACEMENT
Settlement for Method I
R a ft o n 4x 4 pi l e gr o up
P i l e di a me te r 0.5m
Ep/Es = 2500
C o nto ur L e ge nd
2. 670E +00
2. 680E +00
2. 690E +00
2. 700E +00
2. 710E +00
2. 720E +00
2. 730E +00
2. 740E +00
2. 750E +00
2. 760E +00
0. 0 2. 0 4. 0 6. 0 8. 0 10. 0
0. 0
2. 0
4. 0
6. 0
8. 0
0. 0 2. 0 4. 0 6. 0 8. 0 10. 0
0. 0
2. 0
4. 0
6. 0
8. 0
0. 0 2. 0 4. 0 6. 0 8. 0 10. 0
0. 0
2. 0
4. 0
6. 0
8. 0
CONTOURS OF VERTICAL DISPLACEMENT
Settlement for Method II
R a ft o n 4x 4 pi l e gr o up
P i l e di a me te r 0.5m
Ep/Es = 2500
C o nto ur L e ge nd
2. 510E +00
2. 520E +00
2. 530E +00
2. 540E +00
2. 550E +00
2. 560E +00
2. 570E +00
2. 580E +00
2. 590E +00
2. 600E +00
(a) Contours of settlement for Method I analysis
(b) Contours of settlement for Method II analysis
Figure 6.9 Contours of settlement of piled raft obtained from Method I & II analyses
Chapter 6 Analysis of Piled Rafts
204
0.0
0.5
1.0
1.5
2.0
2.5
3.0
0 2 4 6 8 10
Di stance from edge (m)
S
e
t
t
l
e
m
e
n
t

(
m
m
)
Method I
Method II
0
2
4
6
8
10
12
0 10 20 30 40 50 60 70
Pi l e l oad (kN)
D
e
p
t
h

(
m
)
Method I
Method II
(a) Comparison of settlement between Method I & II analyses
(b) Comparison of load distribution along the piles between Method I & II analyses
Figure 6.10 Comparison of settlement of the piled raft and load distribution along the
piles between Method I & II analysis
Centre
Pile P1
Corner
Pile P2
Chapter 6 Analysis of Piled Rafts
205
0. 0 2. 0 4. 0 6. 0 8. 0 10. 0
0. 0
2. 0
4. 0
6. 0
8. 0
0. 0 2. 0 4. 0 6. 0 8. 0 10. 0
0. 0
2. 0
4. 0
6. 0
8. 0
0. 0 2. 0 4. 0 6. 0 8. 0 10. 0
0. 0
2. 0
4. 0
6. 0
8. 0
CONTOURS OF MOMENT IN XX-DIRECTION
Moments for Method I
R a ft o n 4x 4 pi l e gr o up
P i l e di a me te r 0. 5m
Ep /E s = 2500
C o nto u r L e ge nd
- 5 .00 0E+00
0. 000E +00
5. 000E +00
1. 000E +01
1. 500E +01
2. 000E +01
2. 500E +01
3. 000E +01
3. 500E +01
4. 000E +01

(b) Contours of bending moment
for Method II analysis
-20.0
-16.0
-12.0
-8.0
-4.0
0.0
4.0
0 2 4 6 8
Di stance from edge (m)
M
o
m
e
n
t

(
k
N
m
/
m
)
Method I
Method II
(c) Comparison of bending
moments between Method I
and II analyses (moment in
x-direction at y = 4.211m)
Figure 6.11 Bending moments of piled raft obtained from Method I and II analyses
0. 0 2. 0 4. 0 6. 0 8. 0 10. 0
0. 0
2. 0
4. 0
6. 0
8. 0
0. 0 2. 0 4. 0 6. 0 8. 0 10. 0
0. 0
2. 0
4. 0
6. 0
8. 0
0. 0 2. 0 4. 0 6. 0 8. 0 10. 0
0. 0
2. 0
4. 0
6. 0
8. 0
CONTOURS OF MOMENT IN XX-DIRECTION
Moments for Method II
R a ft o n 4x 4 p i l e g r o u p
P i l e di a me te r 0 .5m
Ep /E s = 2500
C o nto u r L e ge n d
- 4.00 0E+0 0
- 2.00 0E+0 0
0. 000 E+00
2. 000 E+00
4. 000 E+00
6. 000 E+00
8. 000 E+00
1. 000 E+01
1. 200 E+01
1. 400 E+01
1. 600 E+01
(a) Contours of bending moment
for Method I analysis
Chapter 6 Analysis of Piled Rafts
206
S
e S
u
P
u
P
e
Piles & raft
are elastic
Pile capacity reached
&
raft is elastic
Pile and raft ultimate
capacity reached
Figure 6.12 Typical load-settlement behaviour of a piled raft
Figure 6.13 Configuration and pile arrangement of the square piled raft

4.0 m
1.5m
y
1
9

m

19 m
P1
P3
P2
Row 1
Row 5
Rigid Rough
Base
20 m
50 kPa
40 m
1m
1.128 m
Chapter 6 Analysis of Piled Rafts
207
Figure 6.14 Meshes of the raft for the GARP and the Method I analyses
PLOT OF FINITE ELEMENT MESH
Comparison with GARP
AP R I L
R a ft o n 5 x 5 pi l e gr o up
P i l e l e ngth 20m
(a) Mesh of the piled raft for the GARP analysis
(b) Mesh of the piled raft for the Method I analysis
PLOT OF FINITE ELEMENT MESH
Comparison with APRILS
GARPX
Raft on 5 x 5 pile group
Pile length 20m
Chapter 6 Analysis of Piled Rafts
208
0
5
10
15
20
25
30
35
40
0 5 10 15 20
Di stance from edge (m)
S
e
t
t
l
e
m
e
n
t

(
m
m
)
Method I
GARP
(a) Comparison of settlement along the centre line of the raft
(b) Comparison of pile load at different pile locations
Figure 6.15 Comparison of settlement and pile load for the GARP and Method I analyses
0
200
400
600
800
1000
1200
centre edge corner
Pi l e l ocati on
P
i
l
e

l
o
a
d

(
k
N
)
GARP
Method I
Chapter 6 Analysis of Piled Rafts
209
Figure 6.16 Contours of moments in x-direction for the GARP and Method I analyses
(a) Contours of moment in x-direction for the GARP analysis
(b) Contours of moment in x-direction for the Method I analysis
0.0 5.0 10.0 15.0 20.0
0.0
5.0
10.0
15.0
20.0
CONTOURS OF MOMENT IN XX-DIRECTION
Moments for GARP analysis
GARPX
Raft on 5 x 5 pile group
Pile length 20m
Contour Legend
-1.500E+02
-1.000E+02
-5.000E+01
0.000E+00
5.000E+01
1.000E+02
1.500E+02
2.000E+02
2.500E+02
3.000E+02
0.0 5.0 10.0 15.0 20.0
0.0
5.0
10.0
15.0
20.0
0.0 5.0 10.0 15.0 20.0
0.0
5.0
10.0
15.0
20.0
0.0 5.0 10.0 15.0 20.0
0.0
5.0
10.0
15.0
20.0
CONTOURS OF MOMENT IN XX-DIRECTION
Moments for Method I
Method I
Raft on 5 x 5 pile group
Pile length 20m
Contour Legend
-5.000E+01
0.000E+00
5.000E+01
1.000E+02
1.500E+02
2.000E+02
2.500E+02
3.000E+02
3.500E+02
4.000E+02
x
y
Chapter 6 Analysis of Piled Rafts
210
-350
-300
-250
-200
-150
-100
-50
0
50
100
0 5 10 15 20
Di stance from edge (m)
M
o
m
e
n
t

(
k
N
m
/
m
)
Method I (at y = 9.789m)
GARP (at y = 9.894m)
-300
-250
-200
-150
-100
-50
0
50
100
0 5 10 15 20
Di stance from edge (m)
M
o
m
e
n
t

(
k
N
m
/
m
)
Method I (at y = 1.789m)
GARP (at y = 1.894m)
(a) Comparison of x-moment in the raft close to the first row of piles
(b) Comparison of x-moment of the raft close to the centre row of piles
Figure 6.17 Comparison of moment in the x-direction at different positions between the
GARP and Method I analyses.
Chapter 6 Analysis of Piled Rafts
211
MATERIAL TYPES
PILED RAFT (5X5 PILE GROUP)
Ma te r i a l s
1
2
x y
z
PLOT OF FINITE ELEMENT MESH
Mesh for APRILS analysis
R a ft o n 5 x 5 pi l e gr o up
P i l e l e n gth 20m
Ep/Es = 2500
(a) Mesh of the piled raft for the finite element analysis
(b) Mesh of the piled raft for the APRILS analysis
Figure 6.18 Meshes of the piled raft for the finite element and APRILS analyses
Chapter 6 Analysis of Piled Rafts
212
(a) Comparison of vertical settlement along the centre line of the raft
(b) Comparison of horizontal displacement along the centre line of the raft
Figure 6.19 Comparison of vertical settlement and horizontal displacement along the
centre line of the raft between APRILS and finite element analyses.
0
5
10
15
20
25
30
35
0 5 10 15 20
Di stance from edge (m)
V
e
r
t
i
c
a
l

s
e
t
t
l
e
m
e
n
t

(
m
m
)
APRILS
FE
-0.03
-0.02
-0.01
0
0.01
0.02
0.03
0 5 10 15 20
Di stance from edge (m)
H
o
r
i
z
o
n
t
a
l

d
i
s
p
l
a
c
e
m
e
n
t

(
m
m
)
APRILS
FE
Chapter 6 Analysis of Piled Rafts
213
-300
-200
-100
0
100
200
300
400
0 5 10 15 20
Di stance from edge (m)
M
o
m
e
n
t

(
k
N
m
/
m
)
APRILS (at y = 9.659m)
FE (at y = 9.606m)
APRILS (at y = 9.923m)
FE (at y = 9.894m)
Figure 6.20 Comparison of bending moment at different locations in the raft
Chapter 6 Analysis of Piled Rafts
214
(a) Comparison of compression of piles at different locations
(b) Comparison of pile load for piles at different locations
Figure 6.21 Comparison of compression and pile load for piles at different
locations between the APRILS and finite element analyses.
0
5
10
15
20
25
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Compressi on (mm)
P
i
l
e

d
e
p
t
h

(
m
)
APRILS
FE
Corner P3
Edge P2
Centre P1
0
5
10
15
20
25
0.0 0.2 0.4 0.6 0.8 1.0
Pi l e l oad (MN)
P
i
l
e

d
e
p
t
h

(
m
)
APRILS
FE
Centre P1
Edge P2
Corner P3
Chapter 6 Analysis of Piled Rafts
215
Figure 6.22 Comparison of settlement and bending moments in the raft
between smooth and rough raft bases by different analyses.
0
5
10
15
20
25
30
35
40
0 5 10 15 20
Di stance from edge (m)
S
e
t
t
l
e
m
e
n
t

(
m
m
)
Method I - smooth raft base GARP - smooth raft base
APRILS - rough raft base FE - rough raft base
-400
-300
-200
-100
0
100
200
0 5 10 15 20
Di stance from edge (m)
M
o
m
e
n
t

(
k
N
m
/
m
)
Method I - smooth raft base (at y = 9.789m)
GARP - smooth raft base (at y = 9.894m)
APRILS - rough raft base (at y = 9.923m)
FE - rough raft base (at y = 9.894m)
(a) Comparison of settlement along the centre line of the raft
(b) Comparison of bending moments in the raft
Chapter 6 Analysis of Piled Rafts
216
-15
-10
-5
0
5
10
15
0 5 10 15 20
Di stance from edge (m)
V
e
r
t
i
c
a
l

d
i
s
p
l
a
c
e
m
e
n
t

(
m
m
)
APRILS
FE
(a) Comparison of horizontal displacement along the centre line of the raft
(b) Comparison of vertical displacement along the centre line of the raft
Figure 6.23 Comparison of horizontal and vertical displacements along the centre
line of the raft between APRILS and finite element analyses.
0
10
20
30
40
50
60
70
0 5 10 15 20
Di stance from edge (m)
H
o
r
i
z
o
n
t
a
l

d
i
s
p
l
a
c
e
m
e
n
t

(
m
m
)
APRILS
FE
Chapter 6 Analysis of Piled Rafts
217
(a) Comparison of horizontal x-displacement
along piles P1, P2 and P3
0
4
8
12
16
20
0 20 40 60 80
x -di spl acement (mm)
P
i
l
e

d
e
p
t
h

(
m
)
APRILS - P1
FE - P1
APRILS - P2
FE - P2
APRILS - P3
FE - P3
0
4
8
12
16
20
0 0.5 1 1.5 2 2.5
y-di spl acement (mm)
P
i
l
e

d
e
p
t
h

(
m
)
APRILS - P3
FE - P3
(b) Comparison of horizontal y-displacement
along pile P3
(c) Comparison of compression along
piles P2 and P3
Figure 6.24 Comparison of horizontal displacements and compression along the pile
at different pile locations between APRILS and finite element analyses.
0
4
8
12
16
20
0 0.1 0.2 0.3 0.4
Compressi on (mm)
P
i
l
e

d
e
p
t
h

(
m
)
APRILS - P2
FE - P2
APRILS - P3
FE - P3
Chapter 6 Analysis of Piled Rafts
218
(b) Comparison of shear forces along the pile at different pile locations
(a) Comparison of bending moments along the pile at different pile locations
Figure 6.25 Comparison of bending moments and shear forces along the pile at
different pile locations between APRILS and finite element analyses.
0
4
8
12
16
20
-2.5 -2 -1.5 -1 -0.5 0 0.5 1
Moment Mx (MNm/m)
P
i
l
e

d
e
p
t
h

(
m
)
APRILS - P1
FE - P1
APRILS - P2
FE - P2
APRILS - P3
FE - P3
0
4
8
12
16
20
-0.2 0.0 0.2 0.4 0.6 0.8 1.0
Shear force (MN)
P
i
l
e

d
e
p
t
h

(
m
)
ARPILS - P1
FE - P1
ARPILS - P2
FE - P2
ARPILS - P3
FE - P3
Chapter 6 Analysis of Piled Rafts
219
Figure 6.26 Configuration of foundation and soil profile.
1.5 m 0.75 m 0.75 m
0
.
7
5

m

0
.
7
5

m

1
.
5

m

0.9 m
10 m
0.4 m
D
S
0.3 L
0.3 L
0.4 L
h
V
H
Case 1 Case 2 Case 3 Case 4 Case 5
E
s
E
s
2E
s
4E
s
4E
s
E
s
2E
s
E
s
E
s
2E
s
E
s
4E
s
4E
s
2E
s
E
s
h = 2L, L/d = 25,
s
= 0.3, E
p
/E
s
= 1000
(b) Soil profile for different cases
(a) Configuration of foundation
Chapter 6 Analysis of Piled Rafts
220
0
2
4
6
8
10
0 0.01 0.02 0.03 0.04 0.05
I
v
D
e
p
t
h

(
m
)
APRILS
FEM
PRAB
0
2
4
6
8
10
0 0.05 0.1 0.15
I
H
D
e
p
t
h

(
m
)
APRILS
FEM
PRAB
0
2
4
6
8
10
0 0.05 0.1 0.15 0.2 0.25
C
a
D
e
p
t
h

(
m
)
APRILS
FEM
PRAB
0
2
4
6
8
10
-0.3 -0.2 -0.1 0 0.1
C
b
D
e
p
t
h

(
m
)
APRILS
FEM
PRAB
(a) Vertical displacements along the piles
Figure 6.27 Comparisons of calculated solutions for piled raft subjected to vertical
loads (a) and (b), and horizontal loads (c), (d) and (e). Case 1.
(Kitiyodom and Matsumoto, 2003)
(b) Axial forces along the piles
(c) Horizontal displacements along the piles (d) Shear forces along the piles
(e) Bending moments along the piles
0
2
4
6
8
10
-0.05 0 0.05 0.1 0.15
C
s
D
e
p
t
h

(
m
)
APRILS
FEM
PRAB
Chapter 6 Analysis of Piled Rafts
221
0
2
4
6
8
10
0 0.01 0.02 0.03 0.04 0.05
I
v
D
e
p
t
h

(
m
)
APRILS
FEM
PRAB
0
2
4
6
8
10
0 0.1 0.2 0.3 0.4 0.5
C
a
D
e
p
t
h

(
m
)
APRILS
FEM
PRAB
0
2
4
6
8
10
0 0.05 0.1 0.15
I
H
D
e
p
t
h

(
m
)
APRILS
FEM
PRAB
0
2
4
6
8
10
-0.3 -0.2 -0.1 0 0.1
C
b
D
e
p
t
h

(
m
)
APRILS
FEM
PRAB
(a) Vertical displacements along the piles (b) Axial forces along the piles
(c) Horizontal displacements along the piles (d) Shear forces along the piles
(e) Bending moments along the piles
0
2
4
6
8
10
-0.05 0 0.05 0.1 0.15
C
s
D
e
p
t
h

(
m
)
APRILS
FEM
PRAB
Figure 6.28 Comparisons of calculated solutions for piled raft subjected to vertical
loads (a) and (b), and horizontal loads (c), (d) and (e). Case 2.
(Kitiyodom and Matsumoto, 2003)
Chapter 6 Analysis of Piled Rafts
222
0
2
4
6
8
10
0 0.01 0.02 0.03 0.04 0.05
I
v
D
e
p
t
h

(
m
)
APRILS
FEM
PRAB
0
2
4
6
8
10
0 0.1 0.2 0.3 0.4 0.5
C
a
D
e
p
t
h

(
m
)
APRILS
FEM
PRAB
0
2
4
6
8
10
0 0.02 0.04 0.06
I
H
D
e
p
t
h

(
m
)
APRILS
FEM
PRAB
0
2
4
6
8
10
-0.3 -0.2 -0.1 0 0.1
C
b
D
e
p
t
h

(
m
)
APRILS
FEM
PRAB
(a) Vertical displacements along the piles
Figure 6.29 Comparisons of calculated solutions for piled raft subjected to vertical
loads (a) and (b), and horizontal loads (c), (d) and (e). Case 3.
(Kitiyodom and Matsumoto, 2003)
(b) Axial forces along the piles
(c) Horizontal displacements along the piles (d) Shear forces along the piles
(e) Bending moments along the piles
0
2
4
6
8
10
-0.05 0 0.05 0.1 0.15
C
s
D
e
p
t
h

(
m
)
APRILS
FEM
PRAB
Chapter 6 Analysis of Piled Rafts
223
0
2
4
6
8
10
0 0.01 0.02 0.03 0.04 0.05
I
v
D
e
p
t
h

(
m
)
APRILS
FEM
PRAB
0
2
4
6
8
10
0 0.01 0.02 0.03 0.04
I
H
D
e
p
t
h

(
m
)
APRILS
FEM
PRAB
0
2
4
6
8
10
-0.15 -0.1 -0.05 0 0.05
C
b
D
e
p
t
h

(
m
)
APRILS
FEM
PRAB
(a) Vertical displacements along the piles
Figure 6.30 Comparisons of calculated solutions for piled raft subjected to vertical
loads (a) and (b), and horizontal loads (c), (d) and (e). Case 4.
(Kitiyodom and Matsumoto, 2003)
(b) Axial forces along the piles
(c) Horizontal displacements along the piles (d) Shear forces along the piles
(e) Bending moments along the piles
0
2
4
6
8
10
-0.05 0 0.05 0.1 0.15
C
s
D
e
p
t
h

(
m
)
APRILS
FEM
PRAB
0
2
4
6
8
10
0 0.1 0.2 0.3 0.4 0.5
C
a
D
e
p
t
h

(
m
)
APRILS
FEM
PRAB
Chapter 6 Analysis of Piled Rafts
224
0
2
4
6
8
10
0 0.01 0.02 0.03 0.04 0.05
I
v
D
e
p
t
h

(
m
)
APRILS
FEM
PRAB
0
2
4
6
8
10
-0.01 0 0.01 0.02 0.03 0.04
I
H
D
e
p
t
h

(
m
)
APRILS
FEM
PRAB
0
2
4
6
8
10
-0.15 -0.1 -0.05 0 0.05
C
b
D
e
p
t
h

(
m
)
APRILS
FEM
PRAB
(a) Vertical displacements along the piles
Figure 6.31 Comparisons of calculated solutions for piled raft subjected to vertical
loads (a) and (b), and horizontal loads (c), (d) and (e). Case 5.
(Kitiyodom and Matsumoto, 2003)
(b) Axial forces along the piles
(c) Horizontal displacements along the piles (d) Shear forces along the piles
(e) Bending moments along the piles
0
2
4
6
8
10
-0.05 0 0.05 0.1 0.15
C
s
D
e
p
t
h

(
m
)
APRILS
FEM
PRAB
0
2
4
6
8
10
0 0.1 0.2 0.3 0.4 0.5
C
a
D
e
p
t
h

(
m
)
APRILS
FEM
PRAB
Chapter 6 Analysis of Piled Rafts
225
(a) Case 1 Raft on 3 piles
(b) Case 2 Raft on 9 piles
(c) Case 3 Raft on 15 piles
(d) Properties of the piled raft and the supporting soil
Figure 6.32 Configuration and pile arrangement of the piled rafts
1m
2m
2m
1m
1m 4m 4m 1m
P2
P2
P2
P1
P1
P1
P1
P1
P1
1m
2m
2m
1m
1m 2m 2m 2m 2m 1m
P2
P2
P2
P1
P1
P1
P1
P1
P1
P1
P1
P2
d = 0.5m
L = 10m
t
r
= 0.5m
h = 20m
Applied loads:
P2 = 2P1
E
p
= E
r
= 30000 MPa

p
=
r
= 0.2
E
s
= 20 MPa

s
= 0.3
Bearing capacity
of raft = 0.3 MPa

Pile capacity =
0.873 MN
(Compression)
1m
2m
2m
1m
1m 4m 4m 1m
P2
P2
P2
P1
P1
P1
P1
P1
P1
Chapter 6 Analysis of Piled Rafts
226
(a) Mesh of piled raft for Case 1
(b) Mesh of piled raft for Case 2
(c) Mesh of piled raft for Case 3
Figure 6.33 Meshes of piled rafts for different cases used in the APRILS analyses
POINT &DISTRIBUTED LOADS
Non-linear analysis of piled raft
R a ft o n 3 pi l e s
Be a r i ng c apa c i ty = 0. 3MP a
P i l e c a pac i ty = 0. 873MN
U ni fo r m l o ads
Sc a l e
5. 00E+00
POINT &DISTRIBUTED LOADS
Non-linear analysis of piled raft
R aft on 3 pi l e s
Be a r i ng c a pa c i ty = 0. 3MP a
P i l e c a pa c i ty = 0.873MN
U ni fo r m l o a ds
Sc a l e
5. 00E+00
POINT &DISTRIBUTED LOADS
Non-linear analysis of piled raft
R a ft o n 15 pi l es
Be ar i ng c a pa c i ty = 0.3 MP a
P i l e c a pa c i ty = 0.873 MN
U ni fo r m l o a ds
Sc al e
5.00E+00
Chapter 6 Analysis of Piled Rafts
227
(a) Load-settlement curve for Case 1
Raft on 3 piles
(b) Load-settlement curve for Case 2
Raft on 9 piles
(c) Load-settlement curve for Case 3 Raft on 15 piles
Figure 6.34 Comparison of load-settlement curves
0
5
10
15
20
25
30
0 20 40 60 80 100 120 140
Central settl ement (mm)
G
r
o
u
p

L
o
a
d

(
M
N
)
PDR
APRILS-V
GARP
0
5
10
15
20
25
30
35
0 20 40 60 80 100 120 140
Central settl ement (mm)
G
r
o
u
p

L
o
a
d

(
M
N
)
PDR
APRILS-V
GARP
0
5
10
15
20
25
0 20 40 60 80 100 120 140
Central settl ement (mm)
G
r
o
u
p

L
o
a
d

(
M
N
)
PDR
APRILS-V
GARP
Figure 6.35 Effect of number of piles on settlement for the load of 12MN
0
10
20
30
40
50
60
0 5 10 15 20 25 30
No. of pi l es
C
e
n
t
r
a
l

s
e
t
t
l
e
m
e
n
t

(
m
m
)
APRILS-V
GARP
Chapter 6 Analysis of Piled Rafts
228
(a) Effect on max. settlement
(b) Effect on differential settlement
(c) Effect on max. moment
(d) Effect on load carried by piles
Figure 6.36 Effect of pile length on the behaviour of a piled raft (linear analysis).
0
20
40
60
80
0 20 40 60
L/d
M
a
x
.

S
e
t
t
l
e
m
e
n
t

(
m
m
)
GARP
APRILS-V
APRILS
FE
0
4
8
12
16
0 20 40 60
L/d
D
i
f
f
e
r
e
n
t
i
a
l

s
e
t
t
l
e
m
e
n
t

(
m
m
)
GARP
APRILS-V
APRILS
FE
0.0
0.2
0.4
0.6
0.8
1.0
0 20 40 60
L/d
M
o
m
e
n
t

(
M
N
m
/
m
)
GARP
APRILS-V
APRILS
FE
0
20
40
60
80
100
0 20 40 60
L/d
%

L
o
a
d

o
n

p
i
l
e
s
GARP
APRILS-V
APRILS
FE
Chapter 6 Analysis of Piled Rafts
229
PLOT OF FINITE ELEMENT MESH
APRILS-V (smooth raft base)
R a ft o n 9 pi l e s
L /d = 4
F i ne r e l e me n ts
MATERIAL TYPES
PILED RAFT (9 PILE GROUP)
Ma te r i a l s
1
2
x y
z
(a) APRILS-V and GARP analysis - Refined mesh of the piled raft in Case 2
(b) Finite element analyses - Mesh for the piled raft in Case 2
Figure 6.37 Meshes for the piled raft in Case 2 for the APRILS-V, GARP and
finite element analyses
Chapter 6 Analysis of Piled Rafts
230
(a) Effect on max. settlement
(b) Effect on differential settlement
(c) Effect on max. moment
(d) Effect on load carried by piles
Figure 6.38 Effect of pile length on the behaviour of a piled raft (non-linear analysis).
0
20
40
60
80
0 20 40 60
L/d
M
a
x
.

s
e
t
t
l
e
m
e
n
t

(
m
m
)
GARP
APRILS-V
0
4
8
12
16
0 20 40 60
L/d
D
i
f
f
e
r
e
n
t
i
a
l

s
e
t
t
l
e
m
e
n
t

(
m
m
)
GARP
APRILS-V
0.0
0.2
0.4
0.6
0.8
1.0
0 20 40 60
L/d
M
a
x
.

m
o
m
e
n
t

(
M
N
m
/
m
)
GARP
APRILS-V
0
20
40
60
80
100
0 20 40 60
L/d
%

L
o
a
d

o
n

p
i
l
e
s
GARP
APRILS-V
Chapter 6 Analysis of Piled Rafts
231
(a) Effect on max. settlement
(b) Effect on differential settlement
(c) Effect on max. moment
(d) Effect on load carried by piles
Figure 6.39 Effect of raft thickness on the behaviour of a piled raft (linear analysis).
0
10
20
30
40
0 1 2 3 4 5
t/d
M
a
x
.

s
e
t
t
l
e
m
e
n
t

(
m
m
)
GARP
APRILS-V
0
5
10
15
20
0 1 2 3 4 5
t/d
D
i
f
f
e
r
e
n
t
i
a
l

s
e
t
t
l
e
m
e
n
t

(
m
m
)
GARP
APRILS-V
0.0
0.2
0.4
0.6
0.8
1.0
1.2
0 1 2 3 4 5
t/d
M
a
x
.

m
o
m
e
n
t

(
M
N
m
/
m
)
GARP
APRILS-V
0
20
40
60
80
100
0 1 2 3 4 5
t/d
%

L
o
a
d

o
n

p
i
l
e
s
GARP
APRILS-V
Chapter 6 Analysis of Piled Rafts
232

(a) Effect on max. settlement
(b) Effect on differential settlement
(c) Effect on max. moment
(d) Effect on load carried by piles
Figure 6.40 Effect of raft thickness on the behaviour of a piled raft (non-linear analysis).
0
10
20
30
40
50
60
0 1 2 3 4 5
t/d
M
a
x
.

s
e
t
t
l
e
m
e
n
t

(
m
m
)
GARP
APRILS-V
0
10
20
30
40
0 1 2 3 4 5
t/d
D
i
f
f
e
r
e
n
t
i
a
l

s
e
t
t
l
e
m
e
n
t

(
m
m
)
GARP
APRILS-V
0.0
0.2
0.4
0.6
0.8
1.0
0 1 2 3 4 5
t/d
M
a
x
.

m
o
m
e
n
t

(
M
N
m
/
m
)
GARP
APRILS-V
0
20
40
60
80
100
0 1 2 3 4 5
t/d
%

L
o
a
d

o
n

p
i
l
e
s
GARP
APRILS-V
Chapter 7 Performance of Piled Rafts
233
CHAPTER 7
PERFORMANCE OF PILED RAFTS

7.1 INTRODUCTION
When using a raft alone as a foundation results in excessive settlements and the use of
pile groups is too costly, a piled raft is a feasible solution. The use of a piled raft as the
foundation for buildings has proven to be an effective and economic way to control total
and differential settlements as well as improving bearing capacities.

The performance of a piled raft can be influenced by several factors such as the
conditions of the supporting soil, relative stiffness between pile and soil, loading
conditions, size and length of the piles, and pile arrangement. In the design of piled raft
foundations, it is necessary to take account of these factors to achieve the objective of
economic construction with satisfactory performance.

Optimised design of piled rafts has been addressed by several researchers (Tandjiria et al.,
1999, Valliappan et al., 1999, Kim et al., 2001, Bezerra et al., 2005, El-Mossallamy et al.,
2006). Based on the distribution of contact pressure at the raft-soil interface, Randolph
(1994) stated that by placing piles under the centre region of a raft, the differential
settlement could be reduced substantially. When a raft is subjected to non-uniform
loadings, the overall and differential settlements and the tilting of the building are of
special importance (Reul and Randolph 2004).

In this chapter, the effect of each of the factors mentioned above on the performance of
piled rafts will be examined. The results will be presented in terms of the (i) overall and
differential settlements of the raft, (ii) bending moment in the raft, (iv) displacement
along the piles, (v) load distribution and skin friction along the piles and (vi) bending
moment in the piles for horizontal loads.

Chapter 7 Performance of Piled Rafts
234
7.2 PARAMETRIC STUDIES
In the following examples (except for example 3), the results are presented in terms of
dimensionless parameters:
(i) normalised vertical displacement I
rz
, and vertical differential displacement, I
dz
in the
raft
BW q
D E u
I
z
s rz
rz
= (7.1),
BW q
D E u
I
z
s rz
dz

= (7.2)
(ii) normalised horizontal displacement I
rx
in the raft
BW q
D E u
I
x
s rx
rx
= (7.3)
(iii) normalised moment per unit length in the raft, I
M
BW q
M
I
z
x
M
= (7.4)
(iv) normalised moment per unit length along the pile, I
pM

DBW q
M
I
x
x
pM
= (7.5)
(v) normalised vertical displacement , I
pz
, and compression, I
comp
, along the pile
BW q
D E u
I
z
s pz
pz
= (7.6),
BW q
D E u
I
z
s pz
comp

= (7.7)
(vi) normalised horizontal displacement, I
px
, along the pile
BW q
D E u
I
x
s px
px
= (7.8)
(vii) normalised axial force, C
a
, and shear force, C
s
along the pile
BW q
Ax
C
z
a
= (7.9)
BW q
S
C
x
s
= (7.10)
(viii) normalised skin friction, F, along the pile

P
DL f
F

= (7.11)
where u
rx
and u
rz

are the horizontal and vertical displacements in the raft
u
rz
is the differential vertical displacement in the raft
u
px
and u
pz

are the horizontal and vertical displacements in the pile
Chapter 7 Performance of Piled Rafts
235
u
pz
is the differential vertical displacement in the pile

f is the shear stress along the pile shaft
E
s
is the modulus of the soil
D is the diameter of the pile
B and W are the length and width of the raft
L is the length of the pile
q
x
and q
z
are the uniform horizontal and vertical applied loads on the raft
M
x
is the moment in the raft or along the pile in the x-axis direction
A
x
and S are the axial and shear forces along the pile
P is the total vertical load acting at the pile head

7.2.1 Types of Foundation
(a) Example 1 - A single pile and single piled raft
In this example, the distribution of skin friction along a single pile (Case 1) and a single
piled raft (Case 2) embedded in a uniform soil is examined. For both cases, the pile has a
diameter of D and a length of L. Figure 7.1 shows the details of the single pile and single
piled raft subjected to a vertical load of P. The L/D ratio was taken as 25. The modulus of
the pile and the soil are E
p
and E
s
respectively. The Poissons ratio,
s
, of the soil was
taken to be 0.499.

Figure 7.2 shows the comparison of the distribution of skin friction along the pile for
different pile compressibilities

=
s
p
E
E
K = 50 and 5000 with the theoretical solutions for
a single pile from Poulos and Davis (1980). From the APRILS analysis, the skin friction
along the pile shaft is computed by dividing the ring load acting at each node of the pile
by the surface area of each of the pile elements. The skin friction of the first pile element
at the pile head is computed from the average of the ring loads at the first and second
nodes of the pile. The results obtained from APRILS for a single pile were in excellent
agreement with the theoretical solutions. For a compressible pile (K = 50), high skin
Chapter 7 Performance of Piled Rafts
236
friction was developed near the head of the pile while for an incompressible pile (K =
5000), the development of skin friction is relatively uniform along the pile. When a raft is
attached to the pile, the skin friction near the pile head reduces significantly and even
becomes negative (i.e downdrag) at the head of the stiff pile as shown in Figure 7.2. The
reduction in the skin friction at the pile head is due to the fact that the presence of the raft
causes the soil to move downward relative to the pile at the surface while for a pile alone,
the pile moves downward relative to the soil along its whole length.

(b) Example 2 - A raft, piled raft and pile group
In this example, a uniform vertical or horizontal load of 20kPa is applied to a raft, a piled
raft and a capped pile group as shown in Figure 7.3. For the analysis of the pile group, the
pile cap is off the ground such that the applied load is carried by the piles only. The raft
and the pile cap are square in plan with a size of 10m (B) x 10m (W) and a thickness of
0.5m. The piled raft is supported by 9 piles at the centre and the pile group consists of 25
piles each of diameter 0.5m and length 10m. The foundations are constructed in a 30m
deep layer of homogeneous soil with a modulus of 10MPa and a Poissons ratio of 0.3.
The piles and raft have a modulus of 20,000MPa and a Poissons ratio of 0.3.

Figure 7.4 shows the normalised settlements, differential settlements and bending
moments for different types of foundation under a uniform vertical load. The differential
displacements were calculated as the difference between the displacement at the centre
(point A) and at the edge (Point B) and also at the corner (Point C) of the raft or cap as
shown in Figure 7.3. Figures 7.4a and b show a comparison of the normalised vertical
displacements and differential displacements among the three different types of
foundations. With the use of a piled raft with 9 piles underneath the central region of the
raft, the maximum displacement reduced by over 30% and the differential displacements
reduced by over 80% compared with the raft foundation. The use of a capped pile group
with 25 piles reduced the maximum displacement by a further 12% compared with the
piled raft, however, the differential displacement was about 2 times larger than the piled
raft. Figure 7.4c shows the bending moments along section a-a of the foundations. The
use of a piled raft or a pile group has reduced the sagging moments by about 30% and
Chapter 7 Performance of Piled Rafts
237
60% compared to the raft alone and hogging moments were observed at the pile locations
that are closest to the edge.

The behaviour of the three different foundation types subjected to a horizontal uniform
load is presented in Figure 7.5. The use of 9 piles to support the raft has reduced the
horizontal displacement of the raft by about 8% and the use of 25 piles as the foundation
has reduced the displacement by about 10% compared with the horizontal displacement
of the raft alone as shown in Figure 7.5a. The bending moments along section a-a of the
three different types of foundation are shown in Figure 7.5b. For the raft alone, the
bending moment is extremely small, however, when piles are used underneath, high
moments are induced at the pile locations.

This example demonstrated that the piled raft with 9 piles which is about one-third of the
number of piles in the 25 pile groups gives a similar reduction in the maximum
displacement of the raft. For a vertically loaded foundation, the differential settlement for
the piled raft is the least among the three types of foundation. If the design of the
foundation is aimed at reducing differential settlement, a piled raft is an economical
design which can reduce the overall and differential settlements as well as the bending
moment in the raft. For a horizontally loaded foundation, the piles resist the lateral
movement, however, the bending moment increases significantly in the raft due to the
presence of the piles.

7.2.2 Soil Conditions
(a) Example 3 - Piled rafts embedded in homogeneous and non-homogeneous soil
When the foundation is embedded in a non-homogeneous soil, a common practice used in
analysis is to calculate the weighted average modulus of the soil (Poulos and Davis, 1980;
Chin et al., 1990; Hong et al., 1999). The weighted average E
av
for a soil having n layers
is defined as
n
n n
av
t t t
E t E t E t
E
+ + +
+ + +
=
L
L
2 1
2 2 1 1
(7.12)
where t
i
= thickness of layer i
Chapter 7 Performance of Piled Rafts
238
E
i
= elastic modulus of layer i

In this example, a piled raft embedded in a non-homogeneous soil (NHomo) is analysed
and the results are compared with a piled raft embedded in a homogenous soil (Homo)
with weighted average modulus. The analyses were carried out by APRILS and the finite
element method (FE). A square raft is supported by 25 piles with a diameter of 1.128m
and a length of 20m. The piled raft is subjected to a uniform vertical or horizontal load of
50kPa. Figure 7.6 shows the layout of the piled raft and the soil profile. The non-
homogeneous soil consisted of four equal thickness layers with different moduli and
Poissons ratios. The equivalent modulus of soil in the homogeneous soil layer was
computed by the weighted average modulus for each layer. The properties of the piled
raft and the supporting soil are listed in Table 7.1.

Soil
Parameters Pile Raft
Non-
homogeneous
(NHomo)
Homogeneous
(Homo)
30 (0-10m)
50 (10-20m)
70 (20-30m)
Modulus, E (MPa) 20,000 20,000
100 (30-40m)
62.5 (0-40m)
0.15 (0-30m) 0.15 (0-30m)
Poissons ratio, - 0.2.
0.25 (30-40m) 0.25 (30-40m)
Table 7.1 Properties of the piled raft and the supporting soil

For a vertically loaded piled raft, comparisons of the results between the non-
homogeneous and the equivalent homogeneous soils are presented in Figures 7.7, 7.8 and
7.9. The results obtained from APRILS were in good agreement with those from the
finite element method. From both methods of analysis, the settlement and bending
moment of the raft in the non-homogeneous soil model were larger than those in the
equivalent homogeneous soil model as shown in Figures 7.7a and b respectively. For
Chapter 7 Performance of Piled Rafts
239
both settlement and bending moments in the raft, maximum differences between the two
soil models were observed at the centre of the raft and the differences diminish towards
the edge of the raft. Figure 7.8 shows the compression along the pile relative to the pile
base at different locations. The piles embedded in non-homogeneous soil model
compressed slightly more than those embedded in the homogeneous soil model. This
could be due to the existence of the softer layers near the pile head in the non-
homogeneous soil model.

The pile load distribution and skin friction along the pile are shown in Figure 7.9. The
load carried by the piles in the non-homogeneous soil model is higher than that in the
homogeneous soil model for each of the piles. The skin friction developed along the pile
head to the half length of the pile (i.e. z < L/2) in the non-homogeneous soil model was
slightly lower than that developed in the homogeneous soil model. This could be due to
the fact that in the non-homogeneous soil where the stiffer layers were underlying the soft
layers near the pile head, the load acting on the pile was transferred to the stiffer layers
near the bottom and therefore less skin friction was developed in the softer layers.

Comparisons of the behaviour of a horizontally loaded piled raft embedded in non-
homogeneous and homogeneous soil are shown in Figures 7.10 and 7.11. The horizontal
displacements of the raft obtained from APRILS were about 10% more than those from
the finite element analysis for both cases as shown in Figure 7.10a and the bending
moments in the raft obtained from APRILS and the finite element analysis were in good
agreement except at the pile locations as shwon in Figure 7.10b. However, the horizontal
displacements and bending moments along the pile were in good agreement with the
finite element analysis as shown in Figure 7.11. By using an equivalent homogeneous
soil model, the displacements and bending moments in the raft and at the pile heads were
underestimated by about 40%. This could be due to the fact that the moduli of the top
layers in the non-homogeneous soil model were lower than in the homogeneous soil
model which resulted in larger displacements and bending moments. The differences in
the pile displacements diminish down the pile and approach the same magnitude at the
pile base.
Chapter 7 Performance of Piled Rafts
240

This example shows that for vertically loaded piled rafts embedded in non-homogeneous
soil, the behaviour in some cases can be approximated by the use of a homogeneous soil
with a weighted average modulus. If the moduli of the layers are very different, then care
would need to be taken in using a weighted average method. However, for horizontally
loaded piled rafts, use of the weighted average technique can lead to significant error.
When the raft is subjected to horizontal loads, the stiffness of the soil near the pile head
will have a significant effect on the horizontal movement of the foundation. The weighted
average technique may not be able to represent the actual behaviour of the soil under
horizontal loads.

(b) Example 4 - Effect of pile-soil stiffness ratio
This example shows the effect of pile-soil stiffness on the behaviour of a square piled raft.
A square raft is supported by a 3x3 pile group embedded in a homogeneous soil as shown
in Figure 7.12. The Poissons ratios for the soil and raft were chosen to be 0.3 and the
ratio of the raft modulus to pile modulus, E
p
/E
r
was taken as 1.0. The ratio of the
thickness of the raft to pile diameter (t
r
/D) was taken as 1.0.

Figures 7.13 and 7.14 show the variation of the normalised displacement and proportion
of the applied load carried by the piles with the pile-soil stiffness ratio K for different
ratios of spacing to diameter (i.e. s/D = 6 and 9). The displacement presented in the
figures is that of at the centre of the raft. For the piled raft subjected to vertical loading,
the displacement of the raft decreases with increasing pile-soil stiffness and spacing to
diameter ratio as shown in Figure 7.13a. The proportion of load carried by the piles
increases rapidly when 100 < K < 2000, and where K > 2000 the proportion of load
remains unchanged. Increasing the spacing between the piles will lead to a significant
reduction in the proportion of load carried by the piles as shown in Figure 7.13b.

For the piled raft subjected to horizontal loading, the horizontal displacement of the raft
decreases with increasing pile-soil stiffness and spacing to diameter ratio as shown in
Figure 7.14a while the proportion of load carried by the piles continues to increase with
Chapter 7 Performance of Piled Rafts
241
increasing pile-soil stiffness but decreases with increasing pile spacing as shown in
Figure 7.14b. It should be noted that the raft becomes larger with larger pile spacing as
shown in Figure 7.13a.

At the same value of pile-stiffness K ratio, the proportion of load carried by the piles
subjected to vertical load is higher than that of subjected to horizontal load as shown in
Figures 7.13b and 7.14b. This shows that under horizontal loading, the soil in contact
with the raft provides more resistance to movement than when the piled raft is subjected
to vertical loading. As the pile spacing increases, the load carried by the piles decreases
thus the contribution of the soil to resisting movement is becoming greater. For
horizontally loaded piled rafts, the resistance of the soil at the surface is more important
to the behaviour of the piled raft than those subjected to the vertical loading.

7.2.3 Dimension of Piles
(a) Example 5 - Effect of the pile length
When a piled raft is subjected to non-uniform loadings, the use of piles with different
lengths may help to reduce the differential and overall settlement. This example shows
the effect of pile length on the behaviour of a piled raft subjected to non-uniform loadings.
A square piled raft supported by 25 piles was subjected to uniform pressures of q and 2q
at the edge and central regions of the raft respectively as shown in Figure 7.15a. These
uniform pressures were applied vertically or horizontally to the raft. The piles underneath
the central and edge regions of the raft have lengths of L
1
and L
2
respectively as shown in
Figure 7.15a. The Poissons ratios of the raft and the soil were taken as 0.3 and the ratio
of the pile modulus to soil modulus E
p
/E
s
was taken as 2500. Analyses were carried out
for pile length ratios (L
1
/L
2
) that varied from 1 to 3.5 for vertical load and from 1 to 1.5
for horizontal load.

For a vertically loaded piled raft, results are presented in Figures 7.16 for the behaviour
of the raft and in Figures 7.17 and 7.18 for the behaviour of the piles. The differential
Chapter 7 Performance of Piled Rafts
242
displacement between the centre and the corner of the raft decreases with increasing
L
1
/L
2
ratio as shown in Figure 7.16a. Figure 7.16b shows that the overall vertical
displacement along the centre line of the piled raft reduced by 40% as L
1
/L
2
increased
from 1 to 3.5. As the vertical displacement of the raft reduced by increasing L
1
, the
magnitude of the bending moment in sagging has reduced while the hogging moment has
increased at the pile locations as shown in Figure 7.16c. The vertical displacements along
the piles located at the centre (P1), edge (P2) and corner (P3) are shown in Figure 7.17.
The centre pile (P1) has the maximum displacement which decreases significantly as
L
1
/L
2
increases. However, the effect becomes less significant for the piles located further
from the centre (i.e. Pile P3). Figure 7.18a shows the axial force along the centre pile
versus the depth (z) relative to the longer pile (L
1
). By increasing the L
1
/L
2
ratio from 1 to
1.5, there is a significant increase in the axial load carried by the pile. Figures 7.18b and c
show the axial force along the edge and corner piles relative to the shorter pile (L
2
)
respectively. For the edge pile (P2), the axial load along the pile decreases as L
1
/L
2

increases and significant decrease in the axial load is observed for 1 < L
1
/L
2
< 1.5. For
the corner pile (P3), increasing the ratio of L
1
/L
2
has no significant effect on the axial
load along the pile as shown in Figure 7.18c.

For a horizontally loaded piled raft, the results for different L
1
/L
2
ratios are shown in
Figure 7.19 for the horizontal displacement and bending moment of the raft and in
Figures 7.20 and 7.21 for the horizontal displacement, bending moment and shear forces
versus the length of the shorter pile (L
2
) at different pile locations. From the results, it can
be seen that an increase in the length (L
1
) of the piles underneath the central region of the
raft has no significant effect on the behaviour of the piled raft. Despite the increase in the
length of the centre pile, the total shear forces or the bending moment carried by the
centre pile remain the same as shown in Figures 7.21. This could be explained by critical
length of the pile. The lateral capacity of the pile would increase with the pile length up
to the critical length, further increase in the length beyond the critical length would have
no improvement in the performance of the piled raft subjected to lateral loads.

Chapter 7 Performance of Piled Rafts
243
(b) Example 6 - Effect of the pile diameter
The piled raft as shown in Figure 7.22 is used to demonstrate the effect of using larger
piles underneath the heavily loaded (central) region of the raft on the behaviour of the
piled raft. The properties of the soil and piles and the applied loads are the same as in the
above example. The piles underneath the central and edge region have diameters of D
1

and D
2
respectively with an identical length of L. Analyses were carried out for pile
diameter ratios D
1
/D
2
that vary from 1 to 3. The ratios of the raft thickness to D
2
(t
r
/D
2
)
and the pile length to D
2
(L/D
2
) were taken as 1 and 20 respectively.

Figures 7.23 to 7.26 show the behaviour of a vertically loaded piled raft supported by
piles of different diameters. The variation of the differential displacement with different
D
1
/D
2
ratios is shown in Figure 7.23. The differential displacement was computed from
the displacement at the centre with reference to the displacement at the corner of the raft.
The differential displacement decreases with increasing D
1
/D
2
ratios, however, the
contribution of increasing D
1
to the differential displacement becomes less at D
1
/D
2
>2.
The overall vertical displacement along the centre line of the raft and the bending
moment in the raft are shown in Figures 7.24a and b respectively. By increasing the
diameter of the piles underneath the heavily loaded region, the vertical displacement of
the raft especially at the heavily loaded region has been reduced. As a result of the
reduction in the displacement, the sagging moment of the raft under the heavily loaded
region is reduced while the moment at the edge of the raft has little change for different
D
1
/D
2
ratios. For the behaviour of the piles, as the D
1
/D
2
ratio increases, the displacement
of the centre pile decrease. However, there is an increase in the displacement of the piles
at the edge and corner as shown in Figures 7.25a, c and e. The effect on the centre pile is
more significant than on the edge and corner piles. It is observed that at 1 < D
1
/D
2
< 1.5
the displacement of the centre pile decreases slightly. However, as D
1
/D
2
further
increases to 2, there is a significant reduction in the pile displacement and any further
increases in the ratio would result in a small reduction in the displacement as shown in
Figure 7.25a. For the piles at the edge and corner piles, the displacement along the pile
increases in the range 1 < D
1
/D
2
< 2, and for D
1
/D
2
>2 there is no significant effect on the
pile. Figures 7.25b, d and f have shown that the use of larger piles underneath the heavily
Chapter 7 Performance of Piled Rafts
244
loaded region has significant effect on the compression of the centre pile but has no effect
on the edge and corner piles. As the diameter of the pile becomes larger, the pile is
becoming more stiff. The axial force along the piles at different locations is shown in
Figure 7.26 where it may be seen that the force on the centre pile is increasing with
increasing D
1
/D
2
ratio, while the force on the edge and centre pile remains unchanged. It
is observed that at 1 < D
1
/D
2
< 1.5, there is a significant increase in the axial force along
the centre pile, at D
1
/D
2
> 1.5 there is only a small increase in the force carried by the pile.
It is clearly shown that by increasing the diameter of the piles underneath the heavily
loaded region up to 2 times the diameter of the piles in the other region could reduce
displacement of the raft and piles by increasing the load carried by the larger diameter
pile.

Figures 7.27 to 7.29 show the variation of the behaviour of a horizontally piled raft with
different D
1
/D
2
ratios. By increasing the diameter of piles underneath the heavily loaded
region of the raft, the horizontal displacement of the raft reduces slightly. However, the
bending moment of the raft increases significantly when the diameter of the larger pile is
twice the diameter of the smaller pile (i.e. D
1
/D
2
= 2) as shown in Figures 7.27a and b.
Figures 7.28 and 7.29 show the horizontal displacement, bending moment and shear force
along the piles. For the horizontal displacement of the piles as shown in Figure 7.28,
there is a small change as the ratio of D
1
/D
2
increases from 1 to 2 but virtually no change
for larger D
1
/D
2
values. This is more clearly shown in Figure 7.29 in which the bending
moment and shear force at the pile head has significantly increased by two times as the
ratio of D
1
/D
2
increases from 1 to 2. However, the effect becomes less significant for
further increase in the diameter of the centre piles. For the edge and corner piles, the
displacement and shear forces along the piles remain unchanged for different D
1
/D
2
ratios.

It can be seen in this example, the optimum ratio of D
1
/D
2
is where the diameter of the
centre piles is twice of the diameter of the edge and corner piles (i. e. D
1
/D
2
= 2) for both
horizontal and vertical loading of a piled raft.


Chapter 7 Performance of Piled Rafts
245

(c) Example 7 Rectangular piled raft subjected to non-uniform vertical loadings
In this example, a rectangular piled raft with 18 piles is subjected to a uniform load of 2q
over the region A and another uniform load of q over the region B as shown in Figure
7.30. The raft has a length of B which is 2 times longer than the width, W (i.e. B/W =
2.0). The piled raft is embedded in a homogenous soil with a modulus of E
s
. The piles
located underneath the region A have a diameter of D
1
and length of L
1
, while the piles
underneath the region B have a diameter and length of D
2
and L
2
respectively. The
modulus of the piles is E
p
. The raft has a thickness of t
r
and a modulus of E
r
. The ratio of
the pile modulus to soil modulus E
p
/E
s
was taken as 2500. Analyses were carried out for
three cases with different L
1
/L
2
and D
1
/D
2
ratios as shown in Figure 7.30. In case 1, the
piles underneath region A have a length 2 times longer and a diameter 1.5 times larger
than those underneath region B. In case 2, the piles have the same diameter in both
regions but the pile length underneath region A is twice as long as for piles underneath
region B. Case 3, where piles underneath both regions have the same length and diameter,
was used as a benchmark.

Results from cases 1 and 2 are compared with those from case 3. Figure 7.31a shows that
the maximum vertical displacement that occurred at the edge of region A in cases 1 and 2
were about 16% and 14% less than those in case 3 respectively. The differential
displacements which were computed from the displacement at points M and N in Figure
7.30 have been reduced by 55% in case 1 and 43% in case 2 as shown in Figure 7.31b.
The bending moments for the different cases are shown in Figure 7.31c which indicates
that the sagging moment of the raft has been reduced significantly in cases 1 and 2.

The displacement along the pile at different locations (Pile P1 to P6 as shown in Figure
7.30) for different cases is shown in Figure 7.32. For the piles underneath region A, the
displacement along piles P1, P2 and P3 is reduced significantly by the use of longer piles
in case 2, however, the use of larger and longer piles in case 1 has resulted in the
displacement being almost the same as in case 2. For piles underneath region B, the
displacement along the piles P4 and P5 was reduced by about 11% in case 2. However,
Chapter 7 Performance of Piled Rafts
246
the displacement along those two piles was increased in case 1. The displacement along
pile P6 was reduced by 17% and 21% in cases 1 and 2 respectively

The use of piles with different lengths and diameters underneath the piled raft leads to
redistribution of axial load along the pile. Figure 7.33 shows the distribution of axial
force along the piles at different locations. For all pile locations except pile P6 which is
located underneath region B as shown in Figure 7.30, the load carried by the piles was
increased in cases 1 and 2. The increase in the length and diameter of the piles has little
effect on piles P2 and P5 as shown in Figures 7.33c and d. For piles P3 and P6, the effect
was significant in which the load carried by pile P3 was increased and led to a decrease in
the load carried by pile P6 as shown in Figures 7.33e and f.

In this example, it can be observed that by increasing the length of the pile under region
A in case 2 would result in significant reduction in the overall and differential
displacements, moments and pile displacements associated with an increase in the load
carried by the piles. Further increasing the diameter of the piles underneath region A in
case 1 would result in a further small reduction in displacements and moments compared
with case 2.

Examples 5 to 7 have demonstrated that the use of longer piles underneath the heavily
loaded region is more effective in reducing the overall and differential displacement than
the use of larger diameter piles under vertical loadings. For horizontal loadings, the use of
larger diameter piles is an effective way to reduce the horizontal displacement of the
piled raft but will cause an increase in the bending moments and shear forces along the
pile.

(d) Example 8 Piled raft with different raft thickness
In this example, a square piled raft with 101 piles as shown in Figure 7.34a was analysed
by APRILS-V. On the inner region of the piled raft, a uniform load of 500kPa and nine
concentrated loads of 1000kN were applied while on the outer region a uniform load of
300kPa was applied as shown in Figure 7.34b. The raft has dimensions of 60m x 60m
Chapter 7 Performance of Piled Rafts
247
with the thicknesses of the raft for the inner and outer regions being 3.5m and 2m
respectively. Two types of piles were used to support the raft (i) Type I 40 piles and
(ii) Type II 61 piles. The spacings between piles vary between 4m for Type I and 8m
for Type II piles. The overall depth of the soil used for the analysis was taken as 2 times
the length of Type II piles. Properties of the soil, pile and raft used for the analysis are
listed in Table 7.2.

Parameters Soil Pile Raft
Modulus (MPa) 20 25000 25000
Poissons ratio 0.2 - 0.3
Table 7.2 Properties of soil, pile and raft used for the analysis.

Three cases were analysed (i) Case 1 piles have different lengths and diameters, (ii)
Case 2 piles have the same lengths and diameters, the length of the pile is the average
length of the piles in Case 1 and the diameter of the piles was taken as the larger diameter
of the piles in Case 1, (iii) Case 3 piles have different lengths but the same diameters.
The lengths and diameters of the piles in each case are shown in Figure 7.34a.
Comparisons of the results from different cases are presented in Figures 7.35, 7.36 and
7.37. Figure 7.35a shows that the settlement obtained from Case 1 was smaller than that
from Case 2 and 3. It is observed that from Case 1, the increase in the diameter of the
piles underneath the inner region could decrease the overall settlement. In Case 2, the use
of piles with the average lengths and larger diameters as in Case1 led to a reduction in the
settlement of the outer region as the piles underneath this region were larger and longer
than those in Case 1. However, the settlement of the inner region was increased because
the piles underneath this region were shorter than those in Case 1. In Case 3 where the
piles have the same diameters but different lengths, the settlement was slightly larger than
in Case 1 as the piles underneath the inner region have a smaller diameter. Figure 7.35b
shows that the differential settlement for Case 1 was about 5% less than Case 3 and about
38% less than Case 2. Figure 7.35c shows that the bending moments in the piled raft for
Case 2 are the largest while Cases 1 and 3 have about the same magnitude.

Chapter 7 Performance of Piled Rafts
248
Figure 7.36 shows the displacement along the piles at four different locations (P1 to P4)
relative to the length and diameter of the Type I piles in Case 3 as shown in Figure 7.34a.
The pile at the centre (P1) of the raft has the maximum displacement while the
displacement of the pile at the corner (P4) was a minimum as expected. The
displacements of piles for Cases 1 and 3 were about the same. For piles underneath the
inner region (Piles P1 and P2), the pile displacements for Case 2 were larger than that of
Cases 1 and 3 as shown in Figures 7.36a and b. This is due to the fact that the length of
the piles underneath the inner region in Case 2 was shorter than those in Cases 1 and 3.
For piles underneath the outer region (Piles P3 and P4) where the piles in Case 2 were
longer than that of in Cases 1 and 3, the pile displacements in Case 2 were less than those
of the other two cases as shown in Figures 7.36c and d.

The axial forces along the pile at different pile locations relative to the length of the
longer pile (L
1
) as shown in Figures 7.37a and b and relative to the length of the shorter
pile (L
2
) for Figures 7.37c and d. Among the piles at different locations, the centre pile
P1 was carrying the least load. For Cases 1 and 3, increasing the diameter of the piles
underneath the inner region resulted in increasing the pile load for piles P1 and P2 as
shown in Figures 7.37a and b. This indicated that the increase in the diameter of the pile
would increase the cross-sectional area and this resulted in an increased axial load carried
by the piles. For piles P3 and P4, the axial loads were about the same for both cases as
shown in Figures 7.37c and d which showed that the effect of increasing the diameter of
the piles underneath the inner region on the axial load distribution is more significant at
the centre pile than at the corner pile. For Case 2 where identical piles were used, the
load carried by pile P3 which was located at the edge of the raft was larger than that of
the other two cases. This is due to the fact that the length and diameter of the pile P3 were
longer and larger than those in Cases 1 and 3 which increases both the cross-sectional and
surface area of the pile and leads to an increase in the axial load carried by the pile.

This example has further demonstrated that the use of longer piles underneath a heavily
loaded region of a vertically loaded piled raft (Case 3) is effective in reducing the overall
and differential settlement and bending moment of the raft as well as the displacement
Chapter 7 Performance of Piled Rafts
249
along the pile as mentioned in the previous examples. Further use of larger diameter piles
(Case 1) led to a further reduction in the overall and differential settlement, however, the
bending moments in the raft would remain about the same.

7.3 CONCLUSIONS
In this chapter, the use of piled raft foundations and the effect of soil conditions and pile
dimensions on the performance of a piled raft have been discussed through the examples
presented above. The following conclusions can be reached:

(1) Comparison of the skin friction between a single pile foundation and a single piled
raft has shown that there is a significant reduction in the skin friction near the pile
head of a single piled raft as compared to the single pile alone.
(2) For vertical loads, piled rafts can be used to minimize the overall and differential
displacement as well as the bending moment of the raft. For horizontal loads, the
piles underneath the raft can help to resist the lateral movement of the raft, however,
these piles will induce bending moments into the raft.
(3) For vertically loaded piled rafts embedded in non-homogeneous soil, the modulus
computed from the weighted average technique can provide a good approximation
of the behaviour of piled rafts as shown in this example, but if the moduli of the
layers are very different, this may not be so. For horizontally loaded piled rafts, this
technique would lead to significant error in the prediction of the behaviour of the
foundation as the stiffness of the soil near the pile head controls the lateral
movement of the foundation.
(4) Increasing pile-soil stiffness leads to a decrease in the displacement of the piled raft
and an increase in the proportion of load carried by the piles for a piled raft
subjected to both vertical and horizontal loads.
(5) Increasing pile spacing (where the raft size increases also) leads to a decrease in the
displacement of the piled raft and a decrease in the proportion of load carried by the
piles for both a vertically and horizontally loaded piled raft.
Chapter 7 Performance of Piled Rafts
250
(6) For horizontally loaded piled raft, the proportion of the load carried by the piles is
lower than that of subjected to vertical loading. This showed that the contribution
of the soil to resist movement for horizontally loaded piled raft is more than that of
for vertically loaded piled raft.
(7) For piled rafts subjected to vertical non-uniform loadings, the use of longer piles
underneath the heavily loaded region leads to a significant reduction in the overall
and differential displacement as well as the bending moment of the piled raft. The
effect of longer centre pile is more significant than the corner pile.
(8) For piled rafts subjected to non-uniform horizontal loadings, the lateral resistance
of the piles is governed by the critical length of the piles. The use of longer piles
may not have significant effect on the behaviour of the piled raft if the length of
piles exceeds the critical length.
(9) The effect of using larger diameter piles on a vertically loaded piled raft is less
significant compared with the effect of using longer piles. However, the effect is
more significant for a horizontally loaded piled raft.
(10) For large piled rafts subjected to non-uniform vertical loadings, the use of longer
piles underneath the heavily loaded region would lead to a significant reduction in
the overall and differetical settlements and bending moments in the raft compared
with a raft supported by piles of the same length. Further increasing the diameter of
piles would lead to a further small reduction in the overall and differential
settlement but would have little effect on the bending moments.
Chapter 7 Performance of Piled Rafts
251
L
z
D
P
L
z
D
t
4D 4D
P
s
p
E
E
K
D
L
=
= 25

0
0.2
0.4
0.6
0.8
1
-1 0 1 2 3 4 5
F
z
/
L
APRILS - piled raft
APRILS - pile
Poulos & Davis
(1980) - pile
K = 50
K = 5000
Case 1 single pile
Case 2 - single piled raft
Figure 7.2 Normalised skin friction distribution along the pile for different pile
compressibilities (Poulos and Davis, 1980)
Figure 7.1 Details of a single pile (Case 1) and a single piled raft (Case 2)

s
= 0.5
Chapter 7 Performance of Piled Rafts
252
x
2m
1m
y
(a) Raft
(b) Piled raft with 9 piles
(c) Pile group with 25 piles
(d) Soil profile
Figure 7.3 Different types of foundation and soil profile
1
0

m

10 m
x
y
a a
A
B
C
y
3 m
2 m
1
0

m

10 m
x
a a
A
B
C
1
0
m

10 m
a a
A
B
C
Rigid Rough
Base
10 m
20 kPa
30 m
0.5 m
0.5 m
E
s
= 10 MPa

s
= 0.3
E
p
= 20,000 MPa
E
r
= 20,000 MPa

r
= 0.3
Chapter 7 Performance of Piled Rafts
253
0.000
0.005
0.010
0.015
0.020
0.025
0.030
0.035
0.040
0 0.2 0.4 0.6 0.8 1
x /B
I
z
Raft
Piled Raft (9 piles)
Pile Group (25 piles)
-0.020
-0.015
-0.010
-0.005
0.000
0.005
0.010
0 0.2 0.4 0.6 0.8 1
x/B
I
M
(a) Normalised vertical
displacement along the
centre line of foundations
(c) Normalised bending
moment along section a-a
of foundations
(b) Normalised differential
displacement of foundations
0
0.001
0.002
0.003
0.004
0.005
0.006
Between A & B (Edge) Between A & C (Corner)
I
d
z

Raft
Piled Raft (9 piles)
Pile Group (25 piles)
Figure 7.4 Normalised settlements, differential settlements and bending
moments for different types of foundation under vertical loadings
Chapter 7 Performance of Piled Rafts
254
-0.04
-0.03
-0.02
-0.01
0
0.01
0.02
0.03
0.04
0 0.2 0.4 0.6 0.8 1
x /B
I
M
(a) Normalised horizontal displacement along the centre line of foundations
(b) Normalised bending moment along section a-a of foundations
Figure 7.5 Normalised horizontal displacements and bending moments for
different types of foundation under horizontal loadings
0.000
0.010
0.020
0.030
0.040
0.050
0 0.2 0.4 0.6 0.8 1
x /B
I
x
Raft
Piled Raft (9 piles)
Pile Group (25 piles)
Chapter 7 Performance of Piled Rafts
255
Figure 7.6 Layout of the piled raft and soil profile
x
4 m
y
1.5 m
B

=

1
9

m

W = 19 m
P1
P3
P2
L = 20 m
Rigid Rough
Base
50 kPa
H
1
= 10 m
t = 1m
D = 1.128 m
Layer I
Layer II
Layer III
Layer IV
H
2
= 10 m
H
3
= 10 m
H
4
= 10 m
a a
Chapter 7 Performance of Piled Rafts
256
(a) Settlement along the centre line of the raft
(b) Bending moment along section a-a of the raft (y = 9.894m)
Figure 7.7 Comparison of settlement and bending moment of the piled
raft embedded in non-homogeneous and homogeneous soils
-100
-50
0
50
100
0 5 10 15 20
Di stance from edge (m)
M
o
m
e
n
t

(
k
N
m
/
m
)
APRILS - NHomo
APRILS - Homo
FE - NHomo
FE - Homo
3.0
3.5
4.0
4.5
5.0
5.5
6.0
6.5
0 5 10 15 20
Di stance from edge (m)
S
e
t
t
l
e
m
e
n
t

(
m
m
)
APRILS - NHomo
APRILS - Homo
FE - NHomo
FE - Homo
Chapter 7 Performance of Piled Rafts
257
(a) Compression along the
pile at the centre P1
(b) Compression along the
pile at the edge P2
(c) Compression along the
pile at the corner P3
Figure 7.8 Comparison of compression along the pile for the piled raft embedded in
non-homogeneous and homogeneous soils
0
5
10
15
20
0 0.1 0.2 0.3 0.4 0.5 0.6
Compressi on (mm)
D
e
p
t
h

(
m
)
APRILS - NHomo
APRILS - Homo
FE - NHomo
FE - Homo
0
5
10
15
20
0 0.1 0.2 0.3 0.4 0.5 0.6
Compressi on (mm)
D
e
p
t
h

(
m
)
APRILS - NHomo
APRILS - Homo
FE - NHomo
FE - Homo
0
5
10
15
20
0 0.1 0.2 0.3 0.4 0.5 0.6
Compressi on (mm)
D
e
p
t
h

(
m
)
APRILS - NHomo
APRILS - Homo
FE - NHomo
FE - Homo
Chapter 7 Performance of Piled Rafts
258
(a) Pile load distribution
along the centre pile P1
Figure 7.9 Comparison of pile load distribution and skin friction along the
pile embedded in non-homogeneous and homogeneous soils
(b) Skin friction along
the centre pile P1
0
5
10
15
20
0 200 400 600 800
Pi l e Load (kN)
D
e
p
t
h

(
m
)
APRILS - NHomo
APRILS - Homo
FE - NHomo
FE - Homo
0
5
10
15
20
-10 0 10 20 30
Ski n Fri cti on (kPa)
D
e
p
t
h

(
m
)
APRILS - NHomo
APRILS - Homo
FE - NHomo
FE - Homo
0
5
10
15
20
0 200 400 600 800
Pi l e Load (kN)
D
e
p
t
h

(
m
)
APRILS - NHomo
APRILS - Homo
FE - NHomo
FE - Homo
0
5
10
15
20
0 10 20 30
Ski n Fri cti on (kPa)
D
e
p
t
h

(
m
)
APRILS - NHomo
APRILS - Homo
FE - NHomo
FE - Homo
(c) Pile load distribution
along the edge pile P2
(d) Skin friction along
the edge pile P2
0
5
10
15
20
0 200 400 600 800
Pi l e Load (kN)
D
e
p
t
h

(
m
)
APRILS - NHomo
APRILS - Homo
FE - NHomo
FE - Homo
(e) Pile load distribution
along the corner pile P3
0
5
10
15
20
-10 0 10 20 30
Ski n Fri cti on (kPa)
D
e
p
t
h

(
m
)
APRILS - NHomo
APRILS - Homo
FE - NHomo
FE - Homo
(f) Skin friction along
the corner pile P3
Chapter 7 Performance of Piled Rafts
259
-400
-300
-200
-100
0
100
200
300
400
0 5 10 15 20
Di stance from edge (m)
M
o
m
e
n
t
,

M
x

(
k
N
m
/
m
)
APRILS - NHomo
APRILS - Homo
FE - NHomo
FE - Homo
(a) Horizontal displacement along the centre line of the raft
(b) Bending moment along section a-a of the raft (y = 9.894m
for FE and y = 9.923 for APRILS)
Figure 7.10 Comparison of horizontal displacement and bending moment for
a piled raft embedded in non-homogeneous and homogeneous
soils
0
4
8
12
16
20
0 5 10 15 20
Di stance from edge (m)
H
o
r
i
z
o
n
t
a
l

d
i
s
p
l
a
c
e
m
e
n
t

(
m
m
)
APRILS - Nhomo
APRILS - Homo
FE - NHomo
FE - Homo
Chapter 7 Performance of Piled Rafts
260
(a) Horizontal displacement
along the centre pile P1
(b) Bending moment along
the centre pile P1
(c) Horizontal displacement
along the edge pile P2
(d) Bending moment along
the edge pile P2
(e) Horizontal displacement
along the corner pile P3
(f) Bending moment along
the corner pile P3
Figure 7.11 Comparison of horizontal displacement and bending moment along
a pile embedded in non-homogeneous and homogeneous soils
0
4
8
12
16
20
0 5 10 15 20
Hori zontal di spl acement (mm)
D
e
p
t
h

(
m
)
APRILS - NHomo
APRILS - Homo
FE - NHomo
FE - Homo
0
4
8
12
16
20
-500 0 500 1000
Moment, Mx (kNm/m)
D
e
p
t
h

(
m
)
APRILS - NHomo
APRILS - Homo
FE -NHomo
FE - Homo
0
4
8
12
16
20
0 5 10 15 20
Hori zontal di spl acement (mm)
D
e
p
t
h

(
m
)
APRILS - NHomo
APRILS - Homo
FE - NHomo
FE - Homo
0
4
8
12
16
20
-500 0 500 1000
Moment, Mx (kNm/m)
D
e
p
t
h

(
m
)
APRILS - NHomo
APRILS - Homo
FE -NHomo
FE - Homo
0
4
8
12
16
20
0 5 10 15 20
Hori zontal di spl acement (mm)
D
e
p
t
h

(
m
)
APRILS - NHomo
APRILS - Homo
FE - NHomo
FE - Homo
0
4
8
12
16
20
-500 0 500 1000 1500
Moment, Mx (kNm/m)
D
e
p
t
h

(
m
)
APRILS - NHomo
APRILS - Homo
FE -NHomo
FE - Homo
Chapter 7 Performance of Piled Rafts
261
y
s
0.5s
W

B
x
D
t
r
s
q
x
or q
z
L
h

s
= 0.3,
r
= 0.3
Figure 7.13 Variation of normalized vertical displacement and proportion
of load carried by piles with pile-soil stiffness for vertical load
Figure 7.12 A 3x3 piled raft configuration and soil profile
(a) Variation of normalized
vertical displacement with
pile-soil stiffness
(b) Variation of proportion of
load carried by piles with
pile-soil stiffness
, 1 =
r
p
E
E
, 1 =
D
t
r
, 2 =
L
h
1 =
W
B
20 =
D
L

0.000
0.005
0.010
0.015
0.020
0.025
0.030
0.035
100 1000 10000 100000
K = E
p
/E
s
I
z
s/D = 6
s/D = 9
0
20
40
60
80
100 1000 10000 100000
l og K
%

o
f

l
o
a
d

c
a
r
r
i
e
d

b
y

p
i
l
e
s
s/D = 6
s/D = 9
Chapter 7 Performance of Piled Rafts
262
(a) Variation of normalized horizontal displacement with pile-soil stiffness
(b) Variation of proportion of load carried by piles with pile-soil stiffness
Figure 7.14 Variation of normalized horizontal displacement and proportion
of load carried by piles with pile-soil stiffness for horizontal load
0.00
0.01
0.02
0.03
0.04
0.05
0.06
100 1000 10000 100000
K = E
p
/E
s
I
x
s/D = 6
s/D = 9
0
10
20
30
40
50
60
100 1000 10000 100000
l og K
%

o
f

l
o
a
d

c
a
r
r
i
e
d

b
y

p
i
l
e
s
s/D = 6
s/D = 9
Chapter 7 Performance of Piled Rafts
263
, 1 =
r
p
E
E
, 1 =
D
t
r
, 5
2
=
L
h
1 =
W
B

20
2
=
D
L
2500 =
s
p
E
E

(a) Configuration of the piled raft
.0
.0
0.0
5.0
0.0
5.0
.0
.0
0.0
5.0
0.0
5.0
0.0 5.0 10.0 15.0 20.0 25.0 30.0
0.0
5.0
10.0
15.0
20.0
25.0
PILE LOCATIONS & DISTRIBUTED LOADS
NON-UNIFORMLY LOADED PILED RAFT
Uniform loads
Load 1 - 2q
Load 2 - q
Load = 2q
Load = q
Figure 7.15 Configuration and mesh of the piled raft
(b) Mesh of the piled raft for the APRILS analysis
x
0.2 W
0.1 W
y
W
B
P1
P3
P2
0.6 B
0.6 W
2q
q

s
= 0.3

r
= 0.3
Rigid Rough
Base
h
L
2

2q
0.6 B
q
L
1
D
t
r
Chapter 7 Performance of Piled Rafts
264
Figure 7.16 Normalised vertical and differential displacement and bending
moment of the piled raft for different L
1
/L
2
ratios
(b) Normalised vertical
displacement of the piled raft
for different L
1
/L
2
ratios
(a) Variation of normalised
differential displacement of the
piled raft with L
1
/L
2
ratio
0
0.002
0.004
0.006
0.008
0.01
0.0 1.0 2.0 3.0 4.0
L
1
/L
2
I
d
z
0.000
0.002
0.004
0.006
0.008
0.011
0.013
0.015
0 0.2 0.4 0.6 0.8 1
x /B
I
z
L1/L2 = 1.0 L1/L2 = 1.5 L1/L2 = 2.0
L1/L2 = 2.5 L1/L2 = 3.0 L1/L2 = 3.5
-0.01
-0.008
-0.006
-0.004
-0.002
0
0.002
0.004
0 0.2 0.4 0.6 0.8 1
x /B
I
M
(c) Normalised bending moment of
the piled raft for different L
1
/L
2

ratios
Chapter 7 Performance of Piled Rafts
265
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
0 0.005 0.01 0.015
I
pz
z
/
L
2
L1/L2 = 1.0
L1/L2 = 1.5
L1/L2 = 2.0
L1/L2 = 2.5
L1/L2 = 3.0
L1/L2 = 3.5
0.0
0.2
0.4
0.6
0.8
1.0
1.2
0 0.002 0.004 0.006 0.008 0.01
I
pz
z
/
L
2
0.0
0.2
0.4
0.6
0.8
1.0
1.2
0 0.002 0.004 0.006 0.008
I
pz
z
/
L
2
(a) Normalised vertical
displacement along the centre
pile P1 for different L
1
/L
2
ratios
(b) Normalised vertical
displacement along the edge pile
P2 for different L
1
/L
2
ratios
(c) Normalised vertical
displacement along the corner
pile P3 for different L
1
/L
2
ratios
Figure 7.17 Normalised vertical displacement along the pile at
different pile locations for different L
1
/L
2
ratios
Chapter 7 Performance of Piled Rafts
266

0.0
0.2
0.4
0.6
0.8
1.0
0 0.005 0.01 0.015 0.02 0.025
C
a
z
/
L
2
0.0
0.2
0.4
0.6
0.8
1.0
0 0.005 0.01 0.015 0.02
C
a
z
/
L
2
(a) Normalised axial force along
the centre pile P1 versus z/L
1

(b) Normalised axial force along
the edge pile P2 versus z/L
2

(c) Normalised axial force along
the corner pile P3 versus z/L
2

Figure 7.18 Normalised axial force along the pile at different pile
locations for different L
1
/L
2
ratios
0.0
0.2
0.4
0.6
0.8
1.0
0 0.01 0.02 0.03 0.04 0.05
C
a
z
/
L
1
L1/L2 = 1.0
L1/L2 = 1.5
L1/L2 = 2.0
L1/L2 = 2.5
L1/L2 = 3.0
L1/L2 = 3.5
Chapter 7 Performance of Piled Rafts
267
`
(a) Normalised horizontal displacement of
the piled raft for different L
1
/L
2
ratios
(b) Normalised bending moment of the
piled raft for different L
1
/L
2
ratios
Figure 7.19 Normalised horizontal displacement and bending
moment of the piled raft for different L
1
/L
2
ratios
0.0150
0.0152
0.0154
0.0156
0.0158
0.0160
0 0.2 0.4 0.6 0.8 1
x/B
I
x
L1/L2 = 1.0
L1/L2 = 1.5
-0.004
-0.003
-0.002
-0.001
0
0.001
0.002
0.003
0.004
0 0.2 0.4 0.6 0.8 1
x /B
I
M
Chapter 7 Performance of Piled Rafts
268
(a) Normalised horizontal
displacement along the
centre pile P1
(b) Normalised horizontal
displacement along the
edge pile P2
(c) Normalised horizontal
displacement along the
corner pile P3
Figure 7.20 Normalised horizontal displacement along the pile
at different pile locations for different L
1
/L
2
ratios
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
0 0.005 0.01 0.015 0.02
I
px
z
/
L
2
L1/L2 = 1.0
L1/L2 = 1.5
0.0
0.2
0.4
0.6
0.8
1.0
0 0.005 0.01 0.015 0.02
I
px
z
/
L
2
0.0
0.2
0.4
0.6
0.8
1.0
0 0.005 0.01 0.015 0.02
I
px
z
/
L
2
Chapter 7 Performance of Piled Rafts
269

(a) Bending moment along
the centre pile P1
(c) Bending moment along
the edge pile P2
(e) Bending moment along
the corner pile P3
(b) Shear force along the
centre pile P1
(d) Shear force along the
edge pile P2
(f) Shear force along the
corner pile P3
Figure 7.21 Normalised bending moments and shear forces along the
pile at different pile locations for different L
1
/L
2
ratios
0
0.3
0.6
0.9
1.2
1.5
-0.01 0 0.01 0.02 0.03
I
PM
z
/
L
2
L1/L2 = 1.0
L1/L2 = 1.5
0
0.3
0.6
0.9
1.2
1.5
-0.002 0 0.002 0.004 0.006 0.008
C
s
z
/
L
2
0
0.2
0.4
0.6
0.8
1
-0.01 0 0.01 0.02 0.03 0.04
I
PM
z
/
L
2
0
0.2
0.4
0.6
0.8
1
-0.005 0 0.005 0.01
C
s
z
/
L
2
0
0.2
0.4
0.6
0.8
1
-0.02 0 0.02 0.04 0.06
I
PM
z
/
L
2
0
0.2
0.4
0.6
0.8
1
-0.005 0 0.005 0.01 0.015
C
s
z
/
L
2
Chapter 7 Performance of Piled Rafts
270
, 1 =
r
p
E
E
, 1
2
=
D
t
r
, 5 =
L
h
1 =
W
B
20
2
=
D
L
2500 =
s
p
E
E

x
0.2 W
0.1 W
y
W
B
P1
P3
P2
0.6 B
0.6 W
2q
q

s
= 0.3

r
= 0.3
D
1
D
2
Rigid Rough
Base
h
L
2q
0.6 B
q
t
r
Figure 7.22 Configuration of the piled raft
0
0.002
0.004
0.006
0.008
0.01
0 0.5 1 1.5 2 2.5 3 3.5
D
1
/D
2
I
d
v
Figure 7.23 Variation of normalised differential displacement of the piled raft with D
1
/D
2
ratios
Chapter 7 Performance of Piled Rafts
271
0.006
0.007
0.008
0.009
0.01
0.011
0.012
0.013
0.014
0 0.2 0.4 0.6 0.8 1
x /B
I
Z
D1/D2=1.0
D1/D2=1.5
D1/D2=2.0
D1/D2=2.5
D1/D2=3.0
-0.01
-0.008
-0.006
-0.004
-0.002
0
0.002
0.004
0.006
0 0.2 0.4 0.6 0.8 1
x /B
I
M
D1/D2 = 1.0 D1/D2 = 1.5 D1/D2 = 2.0
D1/D2 = 2.5 D1/D2 = 3.0
Figure 7.24 Normalised vertical and bending moment of the piled raft for different D
1
/D
2
ratios
(a) Normalised vertical displacement of the piled raft for different D
1
/D
2
ratios
(b) Normalised bending moment of the piled raft for different D
1
/D
2
ratios
Chapter 7 Performance of Piled Rafts
272
(a) Normalised vertical displacement
along the centre pile P1
Figure 7.25 Normalised vertical displacement and compression along
the pile at different pile locations for different D
1
/D
2
ratios
0.0
0.2
0.4
0.6
0.8
1.0
0.0E+00 1.0E-04 2.0E-04 3.0E-04
I
comp
z
/
L
D1/D2 =1.0
D1/D2 =1.5
D1/D2 =2.0
D1/D2 =2.5
D1/D2 =3.0
0.0
0.2
0.4
0.6
0.8
1.0
0.011 0.012 0.013 0.014
I
pz
z
/
L
(b) Normalised compression
along the centre pile P1
0.0
0.2
0.4
0.6
0.8
1.0
0.0085 0.009 0.0095 0.01
I
pz
z
/
L
0.0
0.2
0.4
0.6
0.8
1.0
0.0E+00 5.0E-05 1.0E-04 1.5E-04 2.0E-04
I
comp
z
/
L
(d) Normalised compression
along the edge pile P2
(c) Normalised vertical displacement
along the centre pile P2
0.0
0.2
0.4
0.6
0.8
1.0
0 0.002 0.004 0.006 0.008 0.01
I
pz
z
/
L
0.0
0.2
0.4
0.6
0.8
1.0
0.0E+00 5.0E-05 1.0E-04 1.5E-04
I
comp
z
/
L
(f) Normalised compression
along the corner pile P3
(e) Normalised vertical displacement
along the corner pile P3
Chapter 7 Performance of Piled Rafts
273
0.0
0.2
0.4
0.6
0.8
1.0
0 0.01 0.02 0.03 0.04 0.05
C
a
z
/
L
D1/D2 = 1.0
D1/D2 = 1.5
D1/D2 = 2.0
D1/D2 = 2.5
D1/D2 = 3.0
0.0
0.2
0.4
0.6
0.8
1.0
0 0.005 0.01 0.015 0.02 0.025
C
a
z
/
L
D1/D2 = 1.0
D1/D2 = 1.5
D1/D2 = 2.0
D1/D2 = 2.5
D1/D2 = 3.0
0.0
0.2
0.4
0.6
0.8
1.0
0 0.005 0.01 0.015 0.02
C
a
z
/
L
D1/D2 = 1.0
D1/D2 = 1.5
D1/D2 = 2.0
D1/D2 = 2.5
D1/D2 = 3.0
(a) Normalised axial force along
the centre pile P1 for different
D
1
/D
2
ratios
Figure 7.26 Normalised axial force along the pile at different pile locations for
different D
1
/D
2
ratios
(b) Normalised axial force along
the edge pile P2 for different
D
1
/D
2
ratios
(c) Normalised axial force along
the corner pile P3 for different
D
1
/D
2
ratios
Chapter 7 Performance of Piled Rafts
274
0.015
0.0152
0.0154
0.0156
0.0158
0.016
0 0.2 0.4 0.6 0.8 1
x/B
I
x
D1/D2=1.0
D1/D2=2.0
D1/D2=2.5
D1/D2=3.0
-0.015
-0.01
-0.005
0
0.005
0.01
0.015
0 0.2 0.4 0.6 0.8 1
IM
x
/
B
Figure 7.27 Normalised horizontal displacement and bending moment of a
piled raft for different D
1
/D
2
ratios
(a) Normalised horizontal displacement of a piled raft for different D
1
/D
2
ratios
(b) Normalised bending moment of a piled raft for different D
1
/D
2
ratios
Chapter 7 Performance of Piled Rafts
275
0.0
0.2
0.4
0.6
0.8
1.0
0 0.005 0.01 0.015 0.02
I
px
z
/
L
D1/D2 = 1.0
D1/D2 = 2.0
D1/D2 = 2.5
D1/D2 = 3.0
0.0
0.2
0.4
0.6
0.8
1.0
0 0.005 0.01 0.015 0.02
I
px
z
/
L
0.0
0.2
0.4
0.6
0.8
1.0
0 0.005 0.01 0.015 0.02
I
px
z
/
L
(a) Normalised horizontal
displacement along the
centre pile P1
(b) Normalised horizontal
displacement along the
edge pile P2
(c) Normalised horizontal
displacement along the
conrner pile P3
Figure 7.28 Normalised horizontal displacement along the pile at
different pile locations for different D
1
/D
2
ratios
Chapter 7 Performance of Piled Rafts
276
0
0.2
0.4
0.6
0.8
1
-0.02 0 0.02 0.04 0.06
I
pM
z
/
L
D1/D2 = 1.0
D1/D2 = 2.0
D1/D2 = 2.5
D1/D2 = 3.0
0.0
0.2
0.4
0.6
0.8
1.0
-0.005 0 0.005 0.01 0.015 0.02
C
S
z
/
L
0
0.2
0.4
0.6
0.8
1
-0.01 0 0.01 0.02 0.03 0.04
I
pM
z
/
L
0.0
0.2
0.4
0.6
0.8
1.0
-0.005 0 0.005 0.01
C
S
z
/
L
0
0.2
0.4
0.6
0.8
1
-0.02 0 0.02 0.04 0.06
I
pM
z
/
L
0.0
0.2
0.4
0.6
0.8
1.0
-0.005 0 0.005 0.01 0.015
C
S
z
/
L
(a) Normalised bending moment
along the centre pile P1
(b) Normalised shear force
along the centre pile P1
(c) Normalised bending moment
along the edge pile P2
(d) Normalised shear force
along the edge pile P2
(e) Normalised bending moment
along the corner pile P3
(f) Normalised shear force
along the corner pile P3
Figure 7.29 Normalised bending moment and shear force along the
piles at different pile locations for different D
1
/D
2
ratios
Chapter 7 Performance of Piled Rafts
277

Ratio
Case No. Pile Length - L
1
/L
2
Pile Diameter - D
1
/D
2
1 2.0 1.5
2 2.0 1.0
3 1.0 1.0
Figure 7.30 Configuration of rectangular piled raft
, 1 =
r
p
E
E
, 2
2
=
D
t
r
, 4
2
=
L
h
2 =
W
B

25
2
2
=
D
L
2500 =
s
p
E
E


q
2q
L
1

L
2
h
D
1

D
2
Rough rigid base

s
= 0.35,
r
= 0.15

t
r
x
y
W
B
0.3W 0.25W
0.2W
0.3W
P1
P2
P3 P6
P4
P5
Region A 2q Region B - q
M
N
Chapter 7 Performance of Piled Rafts
278
0.000
0.004
0.008
0.012
0.016
0.020
0 5 10 15 20
x /B
I
z
Case 1
Case 2
Case 3
0.000
0.002
0.004
0.006
0.008
0.010
0.012
0.014
Case 1 Case 2 Case 3
I
d
z
(b) Normalised differential
displacement of the piled raft
(a) Normalised overall vertical
displacement of the piled raft
(c) Normalised bending moment
of the piled raft
Figure 7.31 Normalised overall and differential vertical displacement
and bending moment of the piled raft for different cases
-0.04
-0.03
-0.02
-0.01
0.00
0.01
0.02
0.0 0.2 0.4 0.6 0.8 1.0
x /B
I
M
Case 1
Case 2
Case 3
Chapter 7 Performance of Piled Rafts
279
0.0
0.5
1.0
1.5
2.0
0 0.005 0.01 0.015 0.02
I
pv
z
/
L
2
Case 1
Case 2
Case 3
0.0
0.5
1.0
1.5
2.0
0 0.005 0.01 0.015 0.02
I
pv
z
/
L
2
0.0
0.5
1.0
1.5
2.0
0 0.005 0.01 0.015 0.02
I
pv
z
/
L
2
0.0
0.2
0.4
0.6
0.8
1.0
0 0.002 0.004 0.006 0.008
I
pv
z
/
L
2
0.0
0.2
0.4
0.6
0.8
1.0
0 0.002 0.004 0.006 0.008
I
pv
z
/
L
2
0.0
0.2
0.4
0.6
0.8
1.0
0 0.005 0.01 0.015
I
pv
z
/
L
2
(a) Normalised vertical displacement
along Pile P1
(b) Normalised vertical displacement
along Pile P4
(c) Normalised vertical displacement
along Pile P2
(e) Normalised vertical displacement
along Pile P3
(d) Normalised vertical displacement
along Pile P5
(f) Normalised vertical displacement
along Pile P6
Figure 7.32 Normalised vertical displacement along the pile at different locations
Chapter 7 Performance of Piled Rafts
280
0.0
0.2
0.4
0.6
0.8
1.0
0 0.02 0.04 0.06 0.08
C
a
z
/
L
1
Case 1
Case 2
Case 3
0.0
0.2
0.4
0.6
0.8
1.0
0 0.02 0.04 0.06 0.08 0.1
C
a
z
/
L
1
0.0
0.2
0.4
0.6
0.8
1.0
0 0.01 0.02 0.03 0.04 0.05
C
a
z
/
L
1
0.0
0.2
0.4
0.6
0.8
1.0
0 0.005 0.01 0.015 0.02
C
a
z
/
L
2
0.0
0.2
0.4
0.6
0.8
1.0
0 0.005 0.01 0.015 0.02 0.025
C
a
z
/
L
2
0.0
0.2
0.4
0.6
0.8
1.0
0 0.005 0.01 0.015 0.02
C
a
z
/
L
2
(a) Normalised axial force along Pile P1 (b) Normalised axial force along Pile P4
(c) Normalised axial force along Pile P2
(e) Normalised axial force along Pile P3
(d) Normalised axial force along Pile P5
(f) Normalised axial force along Pile P6
Figure 7.33 Normalised axial force along the pile at different locations
Chapter 7 Performance of Piled Rafts
281

Pile Type I Pile Type II
Length, L
1
(m) Diameter, D
1
(m) Length, L
2
(m) Diameter, D
2
(m)
Case 1 35 1.5 20 1.0
Case 2 28 1.5 28 1.5
Case 3 35 1.0 20 1.0
0. 0 10. 0 20. 0 30. 0 40. 0 50. 0 60. 0 70. 0 - 10. 0
0. 0
10. 0
20. 0
30. 0
40. 0
50. 0
60. 0
0. 0 10. 0 20. 0 30. 0 40. 0 50. 0 60. 0 70. 0 - 10. 0
0. 0
10. 0
20. 0
30. 0
40. 0
50. 0
60. 0
0. 0 10. 0 20. 0 30. 0 40. 0 50. 0 60. 0 70. 0 - 10. 0
0. 0
10. 0
20. 0
30. 0
40. 0
50. 0
60. 0
POINT &DISTRIBUTED LOADS
Piled raft with 101 piles
U ni fo r m l o ads
5.00E+02
3.00E+02
Sc al e
1.000E+03
Inner
Outer
(a) Configuration of the piled raft for different cases
(b) Applied loads on different regions of the piled raft
Figure 7.34 Configuration and load application of piled raft for different cases
0. 0 10. 0 20. 0 30. 0 40. 0 50. 0 60. 0 70. 0 - 10. 0
0. 0
10. 0
20. 0
30. 0
40. 0
50. 0
60. 0
0. 0 10. 0 20. 0 30. 0 40. 0 50. 0 60. 0 70. 0 - 10. 0
0. 0
10. 0
20. 0
30. 0
40. 0
50. 0
60. 0
PLOT OF FINITE ELEMENT MESH
Piled raft with 101 piles
Type I
Type II
Pile
P1
P2
P4
P3
A
B
Chapter 7 Performance of Piled Rafts
282
(c) Normalised bending moment in
the piled raft at y = 30.422m
(a) Normalised overall settlement
along the centre line of the piled
raft
Figure 7.35 Normalised overall and differential settlement and
bending moment for the piled raft in different cases
0.003
0.004
0.005
0.006
0.007
0.0 0.2 0.4 0.6 0.8 1.0
x /B
I
z
Case 1 Case 2 Case 3
0.0000
0.0005
0.0010
0.0015
0.0020
0.0025
Case 1 Case 2 Case 3
I
d
z
(b) Normalised differential
settlement of the piled raft
-0.030
-0.025
-0.020
-0.015
-0.010
-0.005
0.000
0.0 0.2 0.4 0.6 0.8 1.0
x /B
I
M
Case 1
Case 2
Case 3
Chapter 7 Performance of Piled Rafts
283
(a) Normalised vertical displacement
along Pile P1
(b) Normalised vertical displacement
along Pile P2
(c) Normalised vertical displacement
along Pile P3
(d) Normalised vertical displacement
along Pile P4
Figure 7.36 Normalised vertical displacement along the piles at different locations
0
0.3
0.6
0.9
1.2
1.5
1.8
0.000 0.002 0.004 0.006 0.008
I
pz
z
/
L
2
Case 1
Case 2
Case 3
0
0.3
0.6
0.9
1.2
1.5
1.8
0.000 0.002 0.004 0.006
I
pz
z
/
L
2
0
0.3
0.6
0.9
1.2
1.5
1.8
0.000 0.001 0.002 0.003 0.004
I
pz
z
/
L
2
0
0.3
0.6
0.9
1.2
1.5
1.8
0.000 0.002 0.004 0.006
I
pz
z
/
L
2
Chapter 7 Performance of Piled Rafts
284

(a) Normalised axial force along Pile P1
(b) Normalised axial force along Pile P2
(c) Normalised axial force along Pile P3
(d) Normalised axial force along Pile P4
Figure 7.37 Normalised axial force along the piles at different locations
0
0.2
0.4
0.6
0.8
1
0.000 0.005 0.010 0.015 0.020
C
a
z
/
L
2
0
0.2
0.4
0.6
0.8
1
0.000 0.005 0.010 0.015
C
a
z
/
L
2
0
0.2
0.4
0.6
0.8
1
0.000 0.002 0.004 0.006 0.008
C
a
z
/
L
1
Case 1
Case 2
Case 3
0
0.2
0.4
0.6
0.8
1
0.000 0.005 0.010 0.015 0.020
C
a
z
/
L
1
Chapter 8 Application of Program to Piled Raft Founddations
285
CHAPTER 8
APPLICATION OF PROGRAM TO PILED
RAFT FOUNDATIONS

8.1 INTRODUCTION
The application of the present program APRILS to the analysis of piled raft foundations
with different pile lengths and diameters and subjected to uniform and non-uniform
loadings has been demonstrated in the previous chapter. In this chapter, numerical
analyses are performed on several case studies. The case studies presented include
centrifuge model tests, residential and commercial buildings in Europe and full scale tests
on bridge foundations.

8.2 CASE STUDIES
8.2.1 Case 1: Centrifuge Model Tests
Thaher and Jessberger (1991a and b) investigated the behaviour of piled raft foundations
through a series of centrifuge model tests. The raft foundation was square in plan with
dimensions of 180mm by 180mm wide and was 15mm thick. The supporting piles were
symmetrically arranged and the model soil has a thickness of 360mm. Examination of the
effects of the number of piles, pile diameters and pile lengths on the behaviour of the
foundation systems have been performed. Pile groups consisting of 8, 12 and 16 piles
were used in the tests as shown in Figure 8.1, the pile diameters varied between 6.7 and
16.7mm and the lengths varied between 135 and 315mm. The raft and piles in the tests
were modelled as relatively high quality concrete with the Youngs modulus and
Poissons ratio assumed to be 20,000MPa and 0.16 respectively. The model soil was an
overconsolidated stiff clay similar to London clay or Frankfurt/Main clay. The Youngs
modulus of the soil was taken as 9.8MPa as determined by Poulos (1994) with the use of
Chapter 8 Application of Program to Piled Raft Founddations
286
the theoretical solutions obtained from the computer program GARP, and the Poissons
ratio was taken as 0.45. A total load of 19.44kN was applied uniformly to the raft. The
model tests were conducted in three test series: Series I number of piles varied; Series
II length of piles varied and Series III - diameter of piles varied. Table 1 lists the pile
layout and dimensions adopted for the different test series.

Series No. Test No.
Number of
Piles
Pile Length
(mm)
Pile Diameter
(mm)
1 0 - -
2 8 135 16.7
3 12 135 16.7
I
4 16 135 16.7
1 8 135 16.7
2 8 225 16.7 II
3 8 315 16.7
1 8 225 6.7
2 8 225 11.7 III
3 8 225 16.7
Table 8.1: Pile Layout and Dimensions

APRILS was used for the analysis of the tests, and distributions of contact pressures and
pile load were compared with the measured values and the results from GARP.

Figure 8.2 shows the comparison of the measured and theoretical distribution of contact
pressures across two sections of the unpiled raft foundation (Test Series I, 1). The results
obtained from APRILS are in good agreement with the measured values (Thaher and
Jessberger, 1991b) and GARP (Poulos, 1994). According to the measured values, the
contact pressures ranged between 300kPa at the centre and 900kPa at the edges of the raft.
Due to the high contact stresses at the edges of the raft, the supporting soil might have
yielded. The elastic solution obtained from GARP and APRILS, (in this case these
Chapter 8 Application of Program to Piled Raft Founddations
287
programs were not used to limit the contact pressures) predicted a pressure of over
1000kPa along the edges of the raft that are higher than the measured values.

The effect of the number of piles on the settlement and the proportion of load carried by
the piles are shown in Figures 8.3a and b. Results from GARP and APRILS are in good
agreement with the measurements. As the number of piles supporting the raft increases,
the settlement will decrease and this is accompanied by an increase in the proportion of
load carried by the piles. Results from GARP and APRILS have shown that the effect of
the number of piles on the settlement and on the load carried by the piles becomes less
significant after the optimum number of piles have been used. This could be due to the
interaction effects of the piles in the pile group which causes the soil to deform as a block.
Increasing the number of piles without considering the pile arrangement would not be
economical.

Figure 8.4 shows the effect of pile length on the settlement and the load carried by the
piles in a group of 8 piles. Both of the results from GARP and APRILS are in good
agreement with the measured values. It is obvious that the settlement of the raft can be
reduced by the use of longer piles. However, the load carried by the piles may not be
improved much (i.e. increased) after the piles have reached the critical pile length
(Thaher and Jessberger, 1991b).

The effect of pile diameter on the behaviour of the piled raft is shown in Figure 8.5. If the
pile is assumed to be elastic, both GARP and APRILS underpredicted the raft settlement
and overpredicted the load carried by the piles. Poulos (1994) suggested the development
of an ultimate skin friction of 80kPa along the pile shaft would give a better prediction of
the settlement and pile load as shown in Figure 8.5. APRILS models the slip along the
pile-soil interface by limiting the ring load applied on each soil interface. If slip is
allowed along the pile-soil interface of the small diameter piles, the predicted and
measured results are in much better agreement.
Chapter 8 Application of Program to Piled Raft Founddations
288
8.2.2 Case 2: Messe-Torhaus Building, Frankfurt, Germany
The Messe-Torhaus in Frankfurt was the first building in Germany with a piled-raft
foundation and was constructed between 1983 and 1986. The 30 storey building with a
height of 130m is surrounded by apartment blocks and railway bridges are in close
proximity. Due to the existing road and large box culvert that run through the foundation
area, the foundation of the building consists of two 17.5m x 24.5m rafts which are 10m
apart. Each raft has a thickness of 2.5 m and is supported by 42 piles with a diameter of
0.9m and a length of 20m. The rafts are founded on 3m below the ground level and each
of them carries an effective structural load of 200MN. The soil profile is comprised of
quaternary sand and gravel with a thickness of 5m, followed by the Frankfurt clay
(Figure 8.6a). Instrumentation including earth pressure cells, strain gauges and load cells
at the pile bases were used to monitor the behaviour of the foundation (Figure 8.6b). The
modulus of piles and raft were 23500MPa and 34000MPa respectively. The modulus of
the soil was estimated by the emprical formulation presented by Reul (2000):
z
z
E 7 . 0 1
15
30
tanh 45

+ = (8.1)
where E = Youngs modulus (MPa)
z = Depth below surface of the tertiary layer (m)

As the Messe-Torhaus was the first building in Germany to use a piled raft as the
foundation, the design was based on a conventional piled foundation approach in which
all piles are assumed to be loaded to their ultimate bearing capacities and the raft
assumed to transmit the remaining part of the load to the soil (Katzenbach et al., 2000).

APRILS was used for the analysis of the foundation and the results are compared with
the measurements. In the analysis, the whole foundation (Figure 8.7) was analysed in
order to take into account the interaction between the two separate rafts. Thin raft
elements were used in the region between the two rafts to simulate the gap. The load was
applied to the raft in a stepwise incremental procedure in order to model the non-linear
behaviour in which slip is allowed at the pile-soil interfaces.

Chapter 8 Application of Program to Piled Raft Founddations
289
The settlement of the raft was measured as 70mm at the end of construction but this
increased to 150mm with time due to the consolidation and creep of the Frankfurt Clay.
Figure 8.8 shows the settlement contours obtained from APRILS, the calculated
settlement of the raft at the centre is about 62mm which is in good agreement with the
measured settlement prior to the consolidation. Figure 8.9 shows good agreement
between the measured and calculated load-settlement behaviour of the northern piled raft
except for the raft settlements. The initial load increase for the raft was due to the
placement of the concrete for the raft. This was not specifically modelled as the load of
the structure was just applied in equal increments. Both the measured and calculated piled
raft coefficient
pr(resist)
are 0.8 where
pr(resist)
= the total pile resistance divided by the
total effective resistance of the foundation, and this agrees with the measured value
(Katzenbach et al., 2000). Reul and Randolph (2003) described the piled raft coefficient,

pr(load)
, as the ratio of the sum of all the pile loads to the total load on the foundation.
Based on this relationship, the calculated piled raft coefficient is 0.71 which lies between
the measured piled raft coefficient of 0.67 (Sommer, 1991) and 0.76 as obtained from the
finite element analysis of Reul and Randolph (2003).

Figure 8.10 shows the comparison of measured and calculated pile loads under the
northern raft obtained from the finite element method (Reul and Randolph, 2003) and the
present method. The pile loads increase from the centre pile (P1), to the edge piles (P2,
P4 and P6) then to the corner piles (P3 and P5). Due to the interaction between the two
rafts and load eccentricity (Katzenbachet al., 2000), the pile load distribution on an
individual raft is not symmetrical. Figure 8.11 shows the comparisons of measured and
calculated distributions of pile load and skin friction along the pile shaft for the corner
and inner piles. The corner pile mobilised a higher skin friction than the inner pile, and
this is predicted fairly well.

8.2.3 Case 3: Liquid Gas Terminal, Gdansk, Poland
A liquid gas terminal in Gdansk, Poland was constructed on a piled raft foundation. The
foundation consists of two rectangular rafts 71.0m x 61.2m in plan which are 13m apart.
Chapter 8 Application of Program to Piled Raft Founddations
290
Each raft has a thickness of 0.9m and is supported by 180 bored piles. Two sizes of bored
piles were used, 1.0m diameter and 0.62m diameter. The smaller piles are located along
three edges of the raft. Figure 8.12 shows the pile layout underneath the raft. The piles
were 26.5m in length and were embedded in alluvium and sand layers. Table 8.2
summarises the properties of the soil layers as provided by Tejechman (2000). The
Youngs modulus of the piles was taken as 25,000MPa. The raft is subjected to a total
load of 357MN vertically. During construction, settlements around the perimeter of the
raft were monitored.

The present program APRILS was used to predict the settlement of the raft. The edges of
the raft were assumed to be about 1m from the centres of the piles, therefore, the size of
the raft used for the analysis was 73.0m x 63.2m in plan. A vertical uniform load of
77kPa was applied to the raft. The prediction of settlements along the edge of the raft by
the present program APRILS is compared with the measured settlement as shown in
Figure 8.13. The computed settlements are slightly larger than the measured values. The
maximum predicted settlement along the edge of the raft is 20mm which is about 2mm
larger than the measured settlement.

Figure 8.14 shows the contours of the vertical displacement of the raft. The maximum
settlement occurring at the centre of the raft is found to be about 29mm which is slightly
less than the predictions made by other investigators as reported by Van Impe (2001).
Table 8.3 summaries the settlement predictions by different methods.

Number
of layer
Lower level
of layer (m)
Type of soil
Youngs Modulus,
E
s
(MPa)
Poissons ratio,

I +2.9 Fine/Medium sand 32 0.28
II 0.0 Fine/Medium sand 52 0.28
III -2.5 Sandy alluvium 0.9 0.37
IV -7.0 Fine/Medium sand 52 0.28
V -9.7 Medium sand 76.4 0.25
Chapter 8 Application of Program to Piled Raft Founddations
291
VI -11.9 Alluvium 1.3 0.37
VII -18.0 Medium sand 76.4 0.25
VIII -22.3 Alluvium 1.3 0.37
IX -23.6 Medium sand 76.4 0.25
X -24.0 Sandy clay 36.4 0.32
XI Medium sand 76.4 0.25
Table 8.2: Properties of Soil Layers

Method Settlement (mm)
Equivalent foundation according to Polish Code PN-69 32.7
Equivalent foundation according to Tomlinson 30.9
Equivalent foundation according to Van Impe 33.4
Randolph formula 34.0 Equivalent column according to
Poulos, calculated by: Chow method 38.9
Finite layer technique APRILS 28.9
Table 8.3: Settlement Predictions by Different Methods
8.2.4 Case 4: Roosevelt Bridge, Florida
Ruesta and Townsend (1997) performed field tests on an isolated single pile and a large-
scale pile group. The pile group consisted of 16 free-headed prestressed concrete piles
which were spaced at 2.3m. Each pile has a square section with an area of 0.76m
2
and a
length of 16.5m. The modulus of concrete was given as 34475MPa. The piles were
subjected to static lateral loadings using a fixed-headed pile group for the reaction.
Instrumentation including strain gauges and inclinometers were installed in the isolated
single pile and 10 piles in the pile group. Deflections at the pile head were measured by
potentiometers. Figure 8.15 shows the pile layout and instrumentation arrangement.
Incremental loads were applied to the isolated pile and pile group during the load tests.
The loads were applied at about 2.1m and 2.4m above the soil surface for the isolated
single pile and the pile group respectively.

Chapter 8 Application of Program to Piled Raft Founddations
292
The subsurface conditions at the test site consisted of two layers of cohesionless soil: a
layer of 4m thick silty fine sand underlain by a 10m thick layer of cemented sand. In-situ
tests, including Standard Penetration Tests (SPT), Cone Penetration Tests (CPT),
Dilatometer Tests (DMT) and Pressuremeter Tests (PMT) were performed to obtain the
soil properties.

The moduli of the soil layers were estimated by the dilatometer and cone penetrometer
test results and back-analysis of the single pile. Based on the dilatometer test results, the
modulus of the different soils can be estimated from the relationship with the dilatometer
modulus (Robertson et al., 1989):
E
i
= F

E
D
(8.1)
where F

= empirical stiffness factor, and
E
D
= dilatometer modulus

Moduli of soil layers can also be estimated from the cone resistance, Q
c
, obtained from
the cone penetrometer test. The modulus and cone resistance relationship is given by:
E = Q
c
(8.2)
where = modulus factor

Results from the back-analysis have shown that the empirical stiffness factor, F

, and the
modulus factor, , are 0.7 and 2 respectively. The moduli for the soil layers obtained
from both tests are listed in Table 8.4. An average value of the modulus for each 1m thick
layer of soil was computed and used in the analysis. The modulus of the silty fine sand
layer increases with depth and the cemented sand layer is assumed to have a constant
modulus on average.

The drained soil parameters obtained from both in-situ tests were used to compute the
displacements and bending moments by using the present program APRILS. As there
was no pile cap used for the test, the raft was assumed to be flexible with an extremely
low modulus (i.e. E
r
/E
p
= 1x10
-6
was used for the analysis). A soft layer of 2.4 m with a
modulus of 0.01 MPa was inserted between the raft and soil surface to model the off-
Chapter 8 Application of Program to Piled Raft Founddations
293
ground effect. The overall depth of the soil was assumed to be 50 m from the ground
surface. The load was applied to each of the pile heads in the test group, and a linear
analysis was carried out.

The comparisons of the predicted and measured load-deflection behaviour for each row
are shown in Figure 8.16. As loads of the same magnitudes were applied to each of the
piles (although in reality the loads were applied through a steel frame and may not be
equal), the load-deflection curves for the leading and trailing rows were similar. The
predictions with the use of the dilatometer modulus and cone resistance are slightly
different but both of them are reasonably close to the measurements over the initial
loading portion of the load-deflection curve.



Dilatometer
Test
E
i
= F

E
D
F

= 0.7

Cone
Penetrometer
Test
E
i
= a

E
D
a = 2.0

Depth
from
Surface
(m)
Dilatometer
Modulus,
E
d
(bar)
Youngs
Modulus,
E
i
(MPa)
Average
Modulus, E
s

(MPa)
Cone
Resistance,
Q
c
(MPa)
Youngs
Modulus,
E
i
(MPa)
Average
Modulus, E
s

(MPa)
0.5 63 4.4 (4.4+6.2)/2
1.89
3.78 (3.78+3.34)/2
1.0 87 6.1 5.25 1.67 3.34 3.56
1.5 109 7.6 (7.6+6.7)/2 4.79 9.56 (9.58+11.08)/2
2.0 96 6.7 7.15 5.54 11.08 10.33
2.5 108 7.6 (7.6+9.3)/2
3.26
6.52 (6.52+10.74)/2
3.0 133 9.3 8.45 5.37 10.74 8.63
3.5 206 14.4 (14.4+37.5)/2 4.9 9.8 (9.8+41.08)/2
4.0 536 37.5 25.95 20.54 41.08 25.44
4.5 578 40.5 (40.5+42.4)/2
33.79
67.58 (67.58+51.16)/2
5.0 605 42.4 41.45
25.58
51.16 59.37
Table 8-4: Modulus of Soil Layers from Dilatometer Test and Cone Penetrometer Test
Chapter 8 Application of Program to Piled Raft Founddations
294

Figures 8.17 and 8.18 show the comparisons of the predicted and measured average
moments along the piles of each row. The total load applied to the pile groups was
2230kN (139.4 kN on each pile). The predicted moments in the leading and trailing rows
are about the same, however, the measured moments in the leading rows are higher than
the moments in the trailing rows. Figure 8.17 shows the differences between the
predicted and measured moments for the trailing and leading rows where the elastic
moduli have been obtained from the CPT data or DMT data. The differences with the
measured moments for the trailing and leading rows are about 1.4% and 14.2%
respectively. Figure 8.18 shows that the differences between the predicted and measured
moments for the middle trailing rows is about 11.5% and about 7.7% for the middle
leading rows. The predicted moments for each row obtained from the DMT and CPT test
results are in good agreement.

8.2.5 Case 5: Bridge Foundation, Portugal
de Sousa Coutinho (2006) investigated this bridge foundation that was subjected to
horizontal loads and developed data reduction methods for the results. Mathematical
functions were used to fit the curves obtained from the tests.

Full scale tests on pile groups were performed at different test sites. Test 1 was located at
the River Arade, Portimo where the foundation supports the tower of the bridge over the
river. It consisted of two square pile caps (one at the upstream and one at the down
stream) with a size in plan of 10.5m x 10.5m. Each cap was supported by 16 piles spaced
at 3m with a diameter of 1.1m and a length of 48m. The pile caps were tied by a braking
beam and loads were applied by stretching the prestressed cables of the braking beam.
The soil profile for the test site was comprised of a 12m thick layer of silty clay underlain
by a 32m thick layer of clayey and coarse sand which was underlain by the bedrock.
Figure 8.19 shows the pile arrangement and soil profile for Test 1. The piles were loaded
up to a level lower than that at which cracking would begin.

Chapter 8 Application of Program to Piled Raft Founddations
295
Test 2 was performed on a 2 pile group located at the River Sado, Alccer do Sal. The
diameter of the piles was 1.0 m and the piles were spaced at 3 pile diameters apart with a
rectangular cap as shown in Figure 8.20a. Loads were applied by a hydraulic jack and the
piles were loaded up to failure. The soil profile was comprised of thin layers of silty clay
and coarse sand underlain by thick layers of soft clay and clayey sand as shown in Figure
8.20b.

Table 8.5 summarises the details of the piled raft used for the tests. The pile caps were off
the ground surface such that the applied loads were resisted by the piles only. The piles
were bored into the bedrock and were instrumented with strain gauges. Based on the
measured strains, mathematical functions were fitted to the test results to obtain the
curvature of each pile section (de Sousa Coutinho, 2006). Bending moments,
displacements and shear forces along the pile shaft were then computed from the
curvatures.

Test
No.
Place Pile
Arrangement
Pile
Length
(m)
Pile
Diameter
(m)
Longitudinal
Reinforcement
(mm)
Helical
Reinforcement
1 River
Arade,
Portimo
16 piles
(Figure
8.19a)
48 1.1 1225 Double
8 mm spaced
at 30 cm pitch
2 River
Sado,
Alccer
do Sal
2 piles
(Figure
8.20a)
27 1.0 3020 Double
10 mm
spaced at 20
cm pitch
Table 8-5: Details of the Piled Raft for the Tests

Properties of the soil profile are shown in Table 8.6. The undrained modulus of clay was
determined from the empirical correlation of E
s
/s
u
~ 180 450 presented by Poulos (1973)
and the modulus of sand was determined from the empirical correlation of E
s
= 1.6 x SPT
value presented by Kishida and Nakai (1977).
Chapter 8 Application of Program to Piled Raft Founddations
296

Program APRILS was used for the analyses of the full scale tests for result comparisons.
For test 1, a total load of 4800 kN was applied to the piled raft as shown in Figure 8.21
and the contours of horizontal displacement are shown in Figure 8.22. The raft deformed
uniformly in the direction of the applied load. Figure 8.23a shows the comparisons of the
bending moments for Test 1 obtained from the measured strains and program APRILS.
The bending moments at the pile head are in good agreement with those computed from
the measured strains. As stated by de Sousa Coutinho (2006), the bending moments were
obtained from the curvature by fitting the B-Splines least squares function to the
measured strains with the boundary conditions such that the piles in the group were
assumed to have null rotation, displacement and soil reaction at a depth of 20 m and null
soil reaction at the pile head. The bending moments computed from the measured strains
have shown that the curvatures of the pile approached zero at a depth of 14 m from the
pile head, however, the predictions from APRILS have shown that zero curvature of the
pile occurs at a depth of 20 m. The bending moments for Pile 16 (pile closest to the
applied load) are slightly higher than the moments for Pile 3 (pile furthest away from the
applied load).

Test
No.
Soil Type

(kN/m
3
)
c
u

(kPa)
Q
c

(MPa)
N
SPT
E
s
(MPa)
Soft, Silty Clay 18 50 - - 10 (200 c
u
)
Soft, Clayey Sand 18 - - 10-12 19.2 (N
SPT
= 12)
1
Coarse sand with
Pebbles
18 -
-
20-60 96 ((N
SPT
= 60)
Soft, Silty Clay 16 20 - - 4 (200 c
u
)
Medium to Coarse
Sand
19 2 - 6 8 - 14
22.4 (N
SPT
= 14)
Soft Clay 16 32 - - 6.4 (200 c
u
)
2
Soft, Clayey Sand 19 - 2 - 8 5 - 20 32 (N
SPT
= 20)
Table 8-6: Properties of the Soil Profile

Chapter 8 Application of Program to Piled Raft Founddations
297
de Sousa Coutinho (2006) suggested that the shear forces computed by derivation of
bending moments along the pile would not give reliable distributions. Instead, the shear
forces are computed by first solving an integral equation to obtain the soil reaction and
then by integrating the solution with the conditions of null soil reaction at the surface and
at a depth of 20 m. Comparison of the shear forces for Test 1 between the integral
solution and program APRILS are shown in Figure 8.23b. The shear forces obtained from
program APRILS were computed from the derivatives of bending moments. As for the
bending moments, the shear forces for Pile 16 were higher than for Pile 3 at the pile head
but were approaching the same magnitude down towards the pile base.
For Test 2, a total load of 500 kN was applied to the smaller two pile group as shown in
Figure 8.24 and the calculated contours of horizontal displacement are shown in Figure
8.25. The displacement along Pile 2 for Test 2 predicted by program APRILS was
compared with the measured result and showed good agreement with the displacement
computed by de Sousa Coutinho (2006) from the curvatures at the pile head as shown in
Figure 8.26 but predicted displacements were larger elsewhere. de Sousa Coutinho (2006)
stated that the displacements were computed from curvature integration with the
assumptions of null rotations and displacements at a depth of 20 m from the pile head.
The displacements of the pile approached zero at a depth of 8 m from the pile head.
However, the prediction from APRILS shows that the displacements of the pile
approached zero where the piles were embedded in the clayey sand layer which has a
high Youngs modulus.

As shown in Figure 8.27a the bending moments obtained from APRILS were in good
agreement with de Sousa Coutinhos curvature computation at the pile head, however,
large differences were observed down the pile. Such differences can be explained by the
curvature along the pile which is related to the bending moment. The displacements
computed from the measured strain in Figure 8.26 have shown that the curvatures along
the pile were larger than those obtained from APRILS which leads to larger bending
moments. The shear forces along the pile as illustrated in Figure 8.27b showed
differences in the lower part of the pile (as did the moments) as the shear forces obtained
from APRILS were computed from moment derivatives.
Chapter 8 Application of Program to Piled Raft Founddations
298

8.3 CONCLUSIONS
Results obtained from the program APRILS have been compared with the actual
measurements obtained from case studies presented in the literature. The results for field
performance can be classed as Class C predictions, as they are mode subsequent to the
known performance of the foundations. Based on the comparisons, the following
conclusions can be made:

(1) APRILS could provide reasonably accurate predictions of the behaviour of unpiled
and piled rafts used in a centrifuge model test. The effects of the number of piles and
pile length or diameter on the behaviour of a piled raft were also examined. The
results have shown that when the optimum pile number has been used and critical
pile lengths or diameters have been reached, further increases in the number of piles
and pile sizes would not have a significant effect on the performance of the piled raft.
(2) APRILS could model the non-linear behaviour of the foundation for the Messe-
Torhaus in Frankfurt. Settlements prior to consolidation and pile load distribution
along the shaft obtained from APRILS are in good agreement with the measured
results. However, the program did not accurately predict the settlement of the raft at
an early construction stage, because the rapid placement of the raft concrete was not
modelled.
(3) Based on the modulus estimated from in-situ test results and back analysis of a single
pile, APRILS could predict the linear behaviour (horizontal displacements and
moments) of a large scale test (16 pile group) with fair accuracy. This was due to the
fact that the load applied to the pile group was above the soil surface and a thin soft
layer was assumed in the analysis. The predictions based on the modulus estimated
from different in-situ tests were in good agreement with measured values. The soil
modulus correlation factors for the laterally loaded pile were obtained from the in-
situ tests.
(4) For Case 5, a full scale test on a bridge foundation in Portugal, the results obtained
from APRILS were only in fair agreement with the results computed from the
Chapter 8 Application of Program to Piled Raft Founddations
299
measured strains. This could have been due to the assumptions of null rotation and
displacements in the computation of the displacements, bending moments and shear
forces from the measured strain that were used by de Sousa Coutinho (2006). The
soil moduli used in APRILS for the analysis were computed from correlations with
the in-situ tests.
Chapter 8 Application of Program to Piled Raft Founddations
300

180 mm
1
8
0

m
m

s s s = 45d
180 mm
s = 67.5d s
s


s

=

6
7
.
5
d

180 mm
1
8
0

m
m

s = 45d s s
s

=

4
5
d

s


s


Figure 8.1 Pile arrangement for model tests
0
500
1000
1500
0 20 40 60 80 100
Di stance from Edge (mm)
C
o
n
t
a
c
t

P
r
e
s
s
u
r
e

(
k
P
a
)
0
500
1000
1500
2000
0 20 40 60 80 100 120 140
Di stance from Edge (mm)
C
o
n
t
a
c
t

P
r
e
s
s
u
r
e

(
k
P
a
)
Measured
GARP
APRILS
Figure 8.2 Comparisons between measured and theoretical contact
pressures on the raft foundation (Thaher and Jessberger,
1991a and b)
(a) Section A-A (b) Section B-B
B
A
B
A
8.5
27
75
90 7.4 21 53
Chapter 8 Application of Program to Piled Raft Founddations
301

0
50
100
150
0 4 8 12 16
Number of Pi l es
%

o
f

L
o
a
d

c
a
r
r
i
e
d

b
y

P
i
l
e
s
MEASURED
GARP
APRILS
0
2
4
6
8
0 5 10 15 20
Number of Pi l es
S
e
t
t
l
e
m
e
n
t

(
m
m
)
L = 135 mm
d = 16.7 mm
(a) Settlement (b) Load carried by Piles
Figure 8.3 Effect of number of piles on piled raft behaviour
(Thaher and Jessberger, 1991a and b)
0
50
100
150
0 100 200 300 400
Pi l e Length (mm)
%

o
f

L
o
a
d

c
a
r
r
i
e
d

b
y

P
i
l
e
s
MEASURED
GARP
APRILS
0
2
4
6
8
0 100 200 300 400
Pi l e Length (mm)
S
e
t
t
l
e
m
e
n
t

(
m
m
)
(a) Settlement (b) Load carried by Piles
Figure 8.4 Effect of pile length on piled raft behaviour
(Thaher and Jessberger, 1991a and b)
Chapter 8 Application of Program to Piled Raft Founddations
302
0
25
50
75
100
0 5 10 15 20
Pi l e Di ameter (mm)
%

o
f

L
o
a
d

c
a
r
r
i
e
d

b
y

P
i
l
e
s
MEASURED
LINEAR GARP
NON-LINEAR GARP
LINEAR APRILS
NON-LINEAR APRILS
0
2
4
6
8
0 5 10 15 20
Pi l e Di ameter (mm)
S
e
t
t
l
e
m
e
n
t

(
m
m
)
(a) Settlement (b) Load carried by Piles
Figure 8.5 Effect of pile diameter on piled raft behaviour
(Thaher and Jessberger, 1991a and b)

Chapter 8 Application of Program to Piled Raft Founddations
303
TP3
TP5 TP4
TP6 TP1 TP2
EX1 EX2
EX3
Northern Raft Southern Raft
(a) Subsoil condition for the foundation
(b) Pile arrangement and Instrumentation
Figure 8.6 Soil Profile and piled raft layout for the Messe-Torhaus
Chapter 8 Application of Program to Piled Raft Founddations
304
PLOT OF FINITE ELEMENT MESH
ANALYSIS OF MESSE-TORHAUS, FRANKFURT
Figure 8.7 Mesh of the piled raft for Messe-Torhaus
CONTOURS OF VERTICAL DISPLACEMENT
ANALYSIS OF MESSE-TORHAUS, FRANKFURT
C o nto ur L e ge nd
1.500E - 02
2.000E - 02
2.500E - 02
3.000E - 02
3.500E - 02
4.000E - 02
4.500E - 02
5.000E - 02
5.500E - 02
6.000E - 02
6.500E - 02
7.000E - 02
Figure 8.8 Contours of vertical settlement of the raft
Chapter 8 Application of Program to Piled Raft Founddations
305
Figure 8.9 Load-settlement behaviour of the northern piled raft of the Messe-Torhaus
(Katzenbach et al., 2000)
0
10
20
30
40
50
60
70
80
0 20 40 60 80 100 120 140 160 180 200
LOAD (MN)
S
E
T
T
L
E
M
E
N
T

(
m
m
)
MEASURED TOTAL APRILS - TOTAL
MEASURED PILE APRILS - PILE
MEASURED RAFT APRILS - RAFT
Chapter 8 Application of Program to Piled Raft Founddations
306
0
1
2
3
4
5
6
7
6 6A 1 1A 2A 2
Pi l e No.
P
i
l
e

L
o
a
d

(
M
N
)
0
1
2
3
4
5
6
7
8
5 5A 4 4A 3A 3
Pi l e No.
P
i
l
e

L
o
a
d

(
M
N
)
MEASURED
REUL
APRILS
A A
2
B B
6
5 3
6A 1 1A 2A
5A 4 4A 3A
(a) Section A (b) Section B
Figure 8.10 Comparison of measured and calculated pile loads under the northern raft
(Reul and Randolph, 2003)
Chapter 8 Application of Program to Piled Raft Founddations
307
0
5
10
15
20
25
0 2 4 6 8
Pi l e Load (MN)
D
e
p
t
h

(
m
)
MEASURED - INNER
APRILS - INNER
MEASURED - CORNER
APRILS - CORNER
0
5
10
15
20
25
0 50 100 150 200
Ski n Fri cti on (kPa)
D
e
p
t
h

(
m
)
Figure 8.11 Comparison of measured and calculated distributions of
pile load and skin friction along the pile shaft for Pile 1 and
Pile 5 (Katzenbach et al., 2000)
Chapter 8 Application of Program to Piled Raft Founddations
308

Figure 8.13 Comparison of measured and computed settlement
along the edge of the raft (Tejechman, 2000)
-0.025
-0.02
-0.015
-0.01
-0.005
0
0 10 20 30 40 50 60 70 80
DISTANCE (M)
D
I
S
P
L
A
C
E
M
E
N
T

(
M
)
MEASURED
APRILS

400 400 400 400 400 400 400 400 400 400 400 400 400 400 400 400 400
550 550 550 550 550 550 550 550 550 550 550 550 550
6
8
0

6
8
0

6
8
0

6
8
0

6
8
0

6
8
0

6
8
0

6
8
0

6
8
0

6980 cm
400
6
1
2
0

c
m

Figure 8.12 Pile layout of raft for Liquid Gas Terminal
Chapter 8 Application of Program to Piled Raft Founddations
309





CONTOURS OF VERTICAL DISPLACEMENT
LIQUID GAS HOUSE
SET T L EMEN T
P R ED I C T I ON
C o ntou r L e ge nd
8. 000E- 03
1. 000E- 02
1. 200E- 02
1. 400E- 02
1. 600E- 02
1. 800E- 02
2. 000E- 02
2. 200E- 02
2. 400E- 02
2. 600E- 02
2. 800E- 02
3. 000E- 02
Figure 8.14 Contours of the vertical displacement of the raft
Figure 8.15 Pile layout and instrumentation arrangement for field tests at Roosevelt Bridge
1
5
9
13
2
6
10
14
3
7
11
15
4
8
12
16
1
5
9
13
5
6
10
14
3
7
11
15
4
8
12
16
Reaction Group
Jack and
Load Cell
POT 4
POT 5
POT 3
POT 1
POT 2
Load Cells
Single Pile
POT 1 & 2
Reference Pile
Inclinometers
Test Group
Chapter 8 Application of Program to Piled Raft Founddations
310








Figure 8.16 Comparison of the measured and predicted load-deflection curves for the
piles (Ruesta and Townsend, 1997)
0
20
40
60
80
100
120
140
0 0.01 0.02 0.03
Hori zontal defl ecti on (m)
L
o
a
d

(
k
N
)
MEASURED - TRAILING
ROW
DMT - TRAILING ROW
CPT - TRAILING ROW
MEASURED - MIDDLE
TRAILING ROW
DMT - MIDDLE TRAILING
ROW
CPT - MIDDLE TRAILING
ROW
MEASURED - MIDDLE
LEADING ROW
DMT - MIDDLE LEADING
ROW
CPT - MIDDLE LEADING
ROW
MEASURED - LEADING
ROW
DMT - LEADING ROW
CPT - LEADING ROW
Chapter 8 Application of Program to Piled Raft Founddations
311
Figure 8.17 Measured and predicted moment in piles of trailing and leading rows
(Ruesta and Townsend, 1997)
0
2
4
6
8
10
12
14
16
18
0 100 200 300 400 500 600 700
Moment (kN-m)
D
e
p
t
h

f
r
o
m

p
o
i
n
t

o
f

l
o
a
d

(
m
)
MEASURED -
TRAILING ROW
DMT - TRAILING
ROW
CPT - TRAILING
ROW
MEASURED -
LEADING ROW
DMT - LEADING
ROW
CPT - LEADING
ROW
Chapter 8 Application of Program to Piled Raft Founddations
312
Figure 8.18 Measured and predicted moment in piles of middle trailing and
leading rows (Ruesta and Townsend, 1997)
0
2
4
6
8
10
12
14
16
18
0 100 200 300 400 500 600 700
Moment (kN-m)
D
e
p
t
h

f
r
o
m

p
o
i
n
t

o
f

l
o
a
d

(
m
)
MEASURED - MIDDLE
TRAILING ROW
DMT - MIDDLE
TRAILING ROW
CPT - MIDDLE
TRAILING ROW
MEASURED - MIDDLE
LEADING ROW
DMT - MIDDLE
LEADING ROW
CPT - MIDDLE
LEADING ROW
Chapter 8 Application of Program to Piled Raft Founddations
313
Figure 8.19 Pile layout and soil profile for Test 1 (de Sousa Coutinho, 2006)
(a) Pile layout
(b) Soil Profile for River Arade, Portimo

10.5 m
1
0
.
5

m

d = 1.1 m
Upstream Downstream
3.0 m
Braking beam
Instrumented pile
1
2
3
4
7
8
11
12
15
16
Soft, silty clay
= 18kN/m
3
N
SPT
= 0-2
c
u
= 50 kPa
Soft, clayey sand
Coarse sand
with pebbles
Bed rock
= 18kN/m
3

N
SPT
= 10-12
= 18kN/m
3

N
SPT
= 20-60
z (m)
0
12
26
44
1
.
7
0

m

1.70 m 8.00 m 5.00 m
3d
d = 1.1 m
Pile 1 Pile 2
Pile 3
Soft silty clay
Medium to coarse sand
Soft clay
Soft clayey sand
Bed rock
= 18kN/m
3

c
u
= 20 kPa
= 18kN/m
3

N
SPT
= 8 - 14
q
c
= 2 6 MPa
= 16kN/m
3

c
u
= 32 kPa
= 19kN/m
3

N
SPT
= 5 - 20
q
c
= 2 8 MPa
z (m)
0
3
5
24
14
(a) Pile layout
(b) Soil Profile for River Sado, Alccer do Sal
Figure 8.20 Pile layout and soil profile for Test 2(de Sousa Coutinho, 2006)
Chapter 8 Application of Program to Piled Raft Founddations
314
POINT & DISTRIBUTED LOADS
Test 1 - Foundation of Tower of the Bridge over River Arade, Portimao
Uniform loads
Scale
2.00E+00
Figure 8.21 Mesh of the piled raft for Test 1
Figure 8.22 Contours of horizontal displacement for Test 1
.0
.0
.0
.0
.0
0.0
2.0
CONTOURS OF X-DISPLACEMENT
Test 1 - Foundation of Tower of the Bridge over River Arade, Portimao
Contour Legend
2.045E-02
2.050E-02
2.055E-02
2.060E-02
2.065E-02
2.070E-02
2.075E-02
2.080E-02
2.085E-02
2.090E-02
Chapter 8 Application of Program to Piled Raft Founddations
315

(a) Comparison of bending moments along the pile shaft for Test 1
(b) Comparison of shear forces along the pile shaft for Test 1
Figure 8.23 Comparison of moments and shear forces along the pile
shaft for Test 1 (de Sousa Coutinho, 2006)
0
2
4
6
8
10
12
14
16
18
20
-1.2 -0.9 -0.6 -0.3 0 0.3 0.6 0.9 1.2
Moment (kN-m)
P
i
l
e

D
e
p
t
h

(
m
)
MEASURED - PILE 3
APRILS - PILE 3
MEASURED - PILE 16
APRILS - PILE 16
0
2
4
6
8
10
12
14
16
18
20
-0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4
Shear Force (kN)
P
i
l
e

D
e
p
t
h

(
m
)
MEASURED - PILE 3
APRILS - PILE 3
MEASURED - PILE 16
APRILS - PILE 16
Chapter 8 Application of Program to Piled Raft Founddations
316
POINT & DISTRIBUTED LOADS
Test 2 - River Arade, Portimao
Uniform loads
Scale
5.00E-01
Figure 8.24 Mesh of piled raft for Test 2
CONTOURS OF X-DISPLACEMENT
Test 2 - River Arade, Portimao
Contour Legend
6.020E-03
6.030E-03
6.040E-03
6.050E-03
6.060E-03
6.070E-03
6.080E-03
6.090E-03
6.100E-03
6.110E-03
Figure 8.25 Contours of horizontal displacement for Test 2
Chapter 8 Application of Program to Piled Raft Founddations
317

Figure 8.26 Comparison of displacement along the pile for Pile 2 (Test 2)
(de Sousa Coutinho, 2006)
0
2
4
6
8
10
12
14
16
18
20
-8 -6 -4 -2 0 2 4 6 8
Di spl acement (mm)
P
i
l
e

D
e
p
t
h

(
m
)
MEASURED
APRILS
Chapter 8 Application of Program to Piled Raft Founddations
318

(a) Comparison of bending moments along the pile shaft for Test 2
(b) Comparison of shear forces along the pile shaft for Test 2
Figure 8.27 Comparison of moments and shear forces along the pile
shaft for Test 2 (de Sousa Coutinho, 2006)
0
2
4
6
8
10
12
14
16
18
20
-0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4
Shear Force (kN)
P
i
l
e

D
e
p
t
h

(
m
)
MEASURED - PILE 2
APRILS - PILE 2
MEASURED - PILE 3
APRILS - PILE 3
0
2
4
6
8
10
12
14
16
18
20
-0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8
Moment (kN-m)
P
i
l
e

D
e
p
t
h

(
m
)
MEASURED - PILE 2
APRILS - PILE 2
MEASURED - PILE 3
APRILS - PILE 3
Chapter 9 Conclusions and Recommendations for Future Research
319
CHAPTER 9
CONCLUSIONS AND
RECOMMENDATIONS FOR FUTURE
RESEARCH

9.1 CONCLUSIONS
The primary objective of this thesis was to develop a numerical method to study the
behaviour of piled rafts with non-identical piles which are subjected to horizontal and
vertical loadings. The method combined the finite layer method which was used for the
analysis of layered soil systems and the finite element method which was used for the
analyses of the rafts and piles. Full interactions between raft, piles and soil have been
taken into account in the analysis. Unlike the existing methods based on the same
approach that separates the piles from the raft in the analysis, the present method
analysed the piled raft as a complete structure. This method led to a better simulation of
loads acting at the pile heads for large piled rafts and a reduction in the number of raft
elements required for the analysis.

Based on this method, computer programs APRILS and APRILS-V were developed.
APRILS treats the raft base as being rough and is used for analysing piled rafts subjected
to horizontal and vertical loadings. APRILS-V treats the raft base as being smooth and is
used for analysing vertically loaded piled rafts. In both computer programs, the soil
layers can have different material properties, the raft can be of any shape and the raft
elements can have different sizes and thicknesses, while the piles can have different
dimensions and properties. Non-linear analysis of piled rafts can be performed by
limiting the load acting on the pile-soil interface or on the raft elements through the
incremental iterative process. These two programs can also be used for analysing rafts
and pile groups. Solutions including the displacement, shear forces and bending moments
in the raft and displacement and load distribution along the piles can be obtained directly
Chapter 9 Conclusions and Recommendations for Future Research
320
from the programs. Extensive testing of the programs showed that solutions from existing
methods such as the three-dimensional finite element method can be reproduced
accurately. This method has advantages over the three-dimensional finite element method
as it requires less memory for computer storage and less time for data preparation and
computation.

In Chapter 3, the application of the finite layer theory in solving problems with different
shapes of loadings was described and through the numerical examples, the following
conclusions may be drawn:
(1) For different shaped loadings in both horizontal and vertical directions, the finite
layer method can provide accurate and reliable solutions.
(2) The accuracy of the solution is dependent on the numerical integration scheme used
for the inversion of transformed field quantities. The integration range is inversely
proportion to the size of the loaded area and the distance of the solutions from the
loaded centre.
(3) The displacements due to different shaped loadings will be fairly close to the
displacement due to a point load if the loaded area is small enough and the point of
interest is far enough from the loaded centre.

Numerical methods for the analyses of rafts and pile group foundations on layered soil
were presented in Chapter 4. The methods involved a combination of the finite element
method for the analysis of the rafts and piles and the finite layer method for the analysis
of the layered soil system. The following conclusions may be drawn from this chapter:
(1) For the analysis of a raft subjected to concentrated loads, a refined finite element
mesh where the loads are applied is necessary to obtain more accurate solutions.
(2) For a horizontally loaded raft, the displacement of the raft in the direction of the
applied load is larger than the displacements in the other two directions.
(3) For a square pile group subjected to equal magnitude of either horizontal or vertical
loads at the pile heads, maximum displacement occurs at the centre pile. The corner
piles will have the minimum displacement for vertically loaded pile group and the
maximum bending moments for a horizontally loaded pile group.
Chapter 9 Conclusions and Recommendations for Future Research
321
(4) For pile groups embedded in non-homogeneous soil, the load along the pile is
decreasing non-linearly for vertically loaded pile groups while for horizontally
loaded pile groups, the soil stiffness from the soil surface to about half of the
embedment length of the pile have effect on the behaviour of the pile group.
(5) For small pile groups embedded in an infinite layer, the pile head stiffness increases
as the ratio of pile length to radius increases but decreases as the number of piles in
the group increases.

Chapter 5 presented four different types of interaction that were considered in the
analysis of piled rafts. Among the four types of interaction, the interaction between the
piles plays an important role in the behaviour of piled rafts, and so the effect of
interaction factors between identical and non-identical piles were presented. The
following conclusions may be drawn from this chapter:
(1) If the stiffness of the superstructure is taken into account in the analysis of rafts or
piled rafts, there will be a reduction in the differential settlement of the raft.
Neglecting the stiffness of the superstructure in the analysis can be conservative for
very flexible structures.
(2) The present method can provide accurate solutions for foundations embedded in
non-homogeneous soils without the use of averaging techniques as in the other
existing methods.
(3) For two vertically loaded identical piles embedded in a soil with a stiff layer beneath
the pile bases, the interaction factors decrease with increasing thickness of the stiff
layer.
(4) For piles subjected to horizontal loads, the interaction factors between two piles
located in the same direction as the applied load are higher than that when the
unloaded pile is located perpendicular to the direction of the applied load.
(5) For non-identical piles, the diameters of the piles have significant effect on the
interaction factors when the piles are loaded horizontally. For a pair of piles under
consideration, the reciprocal theorem holds when the diameter or length of one of the
piles is a multiple of the diameter or length of another pile.
Chapter 9 Conclusions and Recommendations for Future Research
322
(6) For interaction between a pile and the soil surface, the dimensions of the piles has
little effect on the interaction factors. However, for the interaction between the soil
surface and a pile, the interaction factors for a short and small pile are higher than
those for a long and large diameter pile and the factors approach zero as the distance
between the soil surface and the pile increases.

In Chapter 6, a new method (Method II) for the analysis of piled rafts was presented.
Computer programs APRILS-V and APRILS were developed based on this method in
which the piled raft is analysed as a complete structure. The following conclusions may
be drawn from this chapter:
(1) In Method II, the piles are attached to the nodes on the raft and considered as part of
the complete structure in the analysis. The load acting on the pile head is treated as a
concentrated load, therefore, the loads acting on the pile heads for large piled rafts
can be better simulated. With the use of Method II rather than Method I, the number
of elements used for the analysis can be reduced leading to a reduction in the number
of equations that need to be solved.
(2) Even though different fixity conditions were used at the pile heads in Methods I and
II, solutions for vertical loading from both methods were found to have little
difference.
(3) The use of a pile fixed to a node of the raft resulted in high bending moments in the
raft where the piles are located. However, as the distances from the pile head
increase, the bending moments decrease significantly.
(4) For vertically loaded piled rafts, the overall settlement and bending moments of the
piled raft with a rough base are smaller than those of a raft with a smooth base as the
lateral resistance of the piles leads to a reduction in the lateral movement of the piles
and therefore the vertical movement of the raft.
(5) For the 5x5 square piled raft examined in this chapter, when the piled raft is loaded
horizontally, the raft and piles are largely displaced in the same direction as the
applied load. Maximum shear forces and bending moments are observed at the
corner piles.
Chapter 9 Conclusions and Recommendations for Future Research
323
(6) Analysis of the example of a 2x2 piled raft in non-homogenous soil illustrated that
for a vertically loaded piled raft with the pile base embedded in the stiffest layer,
least vertical settlement is observed. For a horizontally loaded piled raft with the pile
head embedded in the stiffest layer, the least horizontal displacement is observed.
(7) In the non-linear analysis for which the pile-soil interface is allowed to slip, the pile
is carrying its ultimate load and this results in increasing the overall and differential
settlements and bending moment of the foundation.
(8) The settlement of the raft reduces significantly by increasing the number of
supporting piles to an optimum number. One of the ways to improve the overall
performance of a piled raft is by increasing the length of the piles which is more
effective than increasing the number of piles or increasing the raft thickness. This
can reduce the overall and differential settlements, bending moments of the raft and
increase the percentage of total load carried by the piles.
(9) The example of the 9 pile group illustrated that increasing the raft thickness leads to
a reduction in the overall settlement and a significant increase in the bending
moment. The differential settlement approaches zero as the thickness of the raft
increases. For very thick rafts, further increasing the thickness would have no effect
on the overall settlement. The percentage of load carried by the piles is not affected
by the raft thickness for such a pile group.

Chapter 7 presented several factors that have an effect on the performance of a piled raft
foundation. Factors such as the type of foundation, the soil conditions and the dimensions
of the piles were examined and the following conclusions may be drawn:
(1) The skin friction near the pile head of a single piled raft is significantly reduced
compared with the skin friction in a single isolated pile.
(2) Under vertical loadings, piled rafts can be used to reduce the overall and differential
displacement and bending moments of the raft, while under horizontal loading the
piles underneath the raft can reduce the lateral displacement of the raft. However,
larger bending moments are induced in the raft.
(3) Under vertical loadings, the weighted average technique provides a good
approximation of the behaviour of piled rafts embedded in layered soils if the moduli
Chapter 9 Conclusions and Recommendations for Future Research
324
of the layers are not greatly different. However, this technique would underestimate
the behaviour of a piled raft under horizontal loadings as the stiffness of the soil near
the pile head has significant effect on the lateral movement of piled rafts.
(4) For both horizontally and vertically loaded piled rafts, increasing the pile-soil
stiffness leads to a decrease in the displacement of a piled raft associated with an
increase in the proportion of load carried by the piles. Furthermore, under horizontal
and vertical loads, an increase in pile spacings (accompanied by an increase in the
size of the raft) results in a decrease in the displacement of the piled raft and in the
proportion of load carried by the piles.
(5) The use of longer piles underneath the heavily loaded region in a vertically loaded
piled raft results in a significant reduction in the overall and differential
displacements and bending moments of a piled raft. There is a significant effect on
the centre pile but less effect on the corner ones.
(6) The proportion of load carried by the piles in a horizontally loaded piled raft is lower
than that carried by the piles for vertical loading. This indicates that for piled raft
subjected to horiztonal loads, the resistance of the soil at the raft-soil interface is
proportionally greater than for a vertical load.
(7) For a horizontally loaded piled raft, the critical length of the piles governs the lateral
resistance of the piles, therefore, using piles longer than the critical length would
have only a small effect on the behaviour of the piled raft.
(8) The length of the piles has a significant effect on vertically loaded piled rafts while
the diameter of the piles has the most significant effect on horizontally loaded piled
rafts.
(9) The example of a large piled raft subjected to non-uniform vertical loadings
illustrated that when using longer piles underneath the heavily loaded region is
effectively in reducing the overall and differential settelement and bending moments.
Further increasing the diameter of piles could result in a further small reduction in
the overall and differential settlement.

Several case studies which were published previously were examined in Chapter 8. These
case studies included centrifuge model tests for rafts and piled rafts, piled rafts for
Chapter 9 Conclusions and Recommendations for Future Research
325
commercial buildings in Europe and full scale load tests on piled rafts for bridges.
Comparisons between the results from APRILS and the field measurements lead to the
following conclusions:
(1) The centrifuge model tests have shown that as the optimum number of piles and
critical lengths or diameters of the piles used in the piled raft have been reached,
further increasing the pile number or pile sizes would have no significant effect on
the performance of the piled raft.
(2) The analysis of the piled raft for the Messe-Torhaus has shown that APRILS can
model the non-linear behaviour of the foundation by allowing slip along the pile-soil
interface. APRILS accurately predicted the load distribution along the pile shaft and
settlements prior to consolidation. Pile skin friction that was measured and predicted
showed that the presence of the raft caused a reduction in the friction at the top of the
pile.
(3) From the in-situ test results, soil modulus correlation factors for laterally loaded piles
were derived. Based on the modulus estimated from the in-situ test results and back
analysed from single piles, APRILS could predict the linear response of full scale
laterally loaded piles with respect to the horizontal displacement and moment with
fair accuracy.

9.2 RECOMMENDATIONS FOR FUTURE RESEARCH
The method presented in this thesis can accurately predict the linear or non-linear
behaviour of horizontally and vertically loaded conventional piled rafts with piles of non-
identical sizes with a raft of different thicknesses. With an increasing number of
structures being built with deep basements, the rafts are embedded at the base of an
excavation. This type of raft if used in conjunction with piles, is called a compensated
piled raft foundation. The research presented here may be extended to the analysis of
compensated piled rafts with the following suggestions:
(1) When the stucture is applying the initial load, the soil will undergo recompression
and a reload modulus should be used for the soil.
Chapter 9 Conclusions and Recommendations for Future Research
326
(2) The pressure at the base of the raft should be monitored to determine when the pre-
compression load is reached, and for loads in excess of this pre-compression, the
modulus for normal compression is used.
(3) Water pressures also apply uplift to the raft and the effects of an upward water
pressure need to be taken into account.

Appendix I Exact Solution for Auxiliary Function
327
APPENDIX I
EXACT SOLUTION FOR AUXILIARY
FUNCTION

The exact solution for the auxiliary function can be obtained from the fourth-order
differential equation which is formed by using equations (3.14a) to (3.14c) in Chapter 3.
Elimination of U
r
and U
z
leads to the following fourth-order differential equation in
0 ) 2 (
4
2
2
2
4
4
= +

C
z
F B
z
A (I.1)
Suppose that the solution of the differential equation has the form = Le
z
, where L and
are constants. Substituting the auxiliary function into equation (I.1) gives the
following characteristic equation
A
2
+
2
(2B-F)
2
+
4
C = 0 (I.2)
and the roots of the above equation are given by
( )
A
AC F B B F
2
4 2 ) 2 (
2 2 2
2

=

(I.3)
Let = p and = q
( )
A
AC F B B F p
2
4 2 ) 2 (
2 2
2
+
=

(I.4a)
( )
A
AC F B B F q
2
4 2 ) 2 (
2 2
2

=

(I.4b)
The general solution for the differential equation (I.1) may be written in the form
= L
1
e
-pz
+L
2
e
pz
+L
3
e
-qz
+L
4
e
qz
(I.5)
where L
1
, L
2
, L
3
and L
4
are constants. The above equation can be rewritten in terms of
hyperbolic functions as
= L
a
cosh(pz) + M
a
cosh(qz) + L
b
sinh(pz) + M
b
sinh(qz) (I.6)
where L
a
, L
b
, M
a
and M
b
are constants.
Appendix II Flexibility Matrix for a Single Layer of Material
328
APPENDIX II
FLEXIBILITY MATRIX FOR A SINGLE
LAYER OF MATERIAL

The flexibility relationship between the transformed stresses and displacements may be
expressed as

m
m
p
p
rm
zm
rp
zp
T
N
T
N
F F F F
F F F F
F F F F
F F F F
U
U
U
U
44 43 42 41
34 33 32 31
24 23 22 21
14 13 12 11
(II.1)
The flexibility matrix F
i
is a 4x4 symmetrical matrix (i.e F
ij
= F
ji
) with the elements as
follows
22 44
12 34
11 33
24
14 23
22
14
13
12
11
) (
2
) (
2
) (
2
) (
2
) (
2
) (
2
F F
F F
F F
A
F
F F
A
F
A
F
A
F
B A
F
A
F
a b
b a
a b
b a
b a
b a
=
=
=
=
=
+ =
=
=
+ =
+ =

(II.2)
Appendix II Flexibility Matrix for a Single Layer of Material
329
where

b
q p
b
b
a
b
b
q p
b
a
q p
a
a
b
a
a
q p
a
D
S S q p
D
pqD
D
C C q p pq
D
C C q p
D
pqD
D
S S q p pq
) (
) (
) (
) (
2 2 2
2
2 2
2 2 2
2
2 2

=
=

=
=

and
D
a
=
3
(pS
p
C
q
- qS
q
C
p
)
D
b
=
3
(pC
p
S
q
- qC
q
S
p
)
The notation C
p
, C
q
, S
p
and S
q
are defined as
C
p
= cosh(ph), C
q
= cosh(qh)
S
p
= sinh(ph), S
q
= sinh(qh)




(II.3)
Appendix III Boundary Conditions for Circular and Rectangular Loadings Applied to
Layered System
330
APPENDIX III
BOUNDARY CONDITIONS FOR
CIRCULAR AND RECTANGULAR
LOADINGS APPLIED TO A LAYERED
SYSTEM



Circular loadings:
At the base, z = h
(a) for a rough rigid base, the horizontal displacement is zero, i.e.
(U
p
) = 0
(b) for a smooth rigid base, the shear stress is zero, i.e.
(S
zp
) = 0

Rectangular loadings:
At the base, z = h
(a) for a rough rigid base, the horizontal displacement is zero, i.e.
(U
p
) = 0
(b) for a smooth rigid base, the shear stress is zero, i.e.
(S
zp
) = 0


Appendix IV Shape Function for Raft Elements (8 noded Isoparametric Element)
331
APPENDIX IV
SHAPE FUNCTION FOR RAFT
ELEMENTS (8 NODED ISOPARAMETRIC
ELEMENT)





















N
1
= 0.25(1-s)(1-t)(-s-t-1)
N
2
= 0.50(1-s
2
)(1-t)
N
3
= 0.25(1+s)(1-t)(s-t-1)
N
4
= 0.50(1-t
2
)(1+s)
N
5
= 0.25(1+s)(1+t)(s+t-1)
N
6
= 0.50(1-s
2
)(1+t)
N
7
= 0.25(1-s)(1+t)(-s+t-1)
N
8
= 0.50(1-t
2
)(1-s)


s
t
1
8
7
6
5
3
3
2
References
332
REFERENCES


Banerjee, P. K. and Davies, T. G. (1977). Analysis of Pile Groups Embedded in Gibson
Soil, Proc. 9
th
International Conference on Soil Mechanics and Foundation Engineering,
Tokyo, pp. 381-386.

Basile, F. (2003). Analysis and Design of Pile Groups, Numerical Analysis and
Modelling in Geomechanics, edited by Bull, London, Ch. 10, pp. 278-315.

Bezerra, J. E, Cunha, R. P. and Sales, M. M. (2005), Optimization Concepts for the
Design of Piled Raft Foundation Systems, Proc. 16
th
International Conference on Soil
Mechanics and Geotechnical Engineering, Osaka, Japan, pp. 1947-1950.

Bezine, G. (1988). A New Boundary Element Method for Bending of Plates on Elastic
Foundations, Int. Jl. of Solids Structures, Vol. 24, No. 6, pp. 557-565.

Bogner, F. K., Fox, R. L. and Schmit, L. A. (1965). The Generation of Inter-element
compatible Stiffness and Mass Matrices by the use of Interpolation Formulas, Proc.
Conference on Matrix Methods in Structural Mechanics, AFFEL-TR-66-80, pp. 397-443.

Booker, J. R. and Small, J. C. (1983). The Analysis of Liquid Storage Tanks on Deep
Foundations, Int. Jl. for Numerical and Analytical Methods in Geomechanics, Vol. 7, pp.
187-207.

Booker, J. R, Balaam, N. and Davis, E. H. (1985). The Behaviour of an Elastic Non-
homogeneous Half-space, Part I & II, Int. Jl. for Numerical and Analytical Methods in
Geomechanics, Vol. 9, No. 4, pp. 353-367, 369-381.

Brown, P. T. (1969a) Numerical Analysis of Uniformly Loaded Circular Rafts on
Elastic Layers of Finite Depth, Gotechnique, Vol. 19, No. 2, pp. 301-306.

Brown, P. T. (1969b) Numerical Analysis of Uniformly Loaded Circular Rafts on Deep
Foundations, Gotechnique, Vol. 19, No. 3, pp. 399-404.

Brown, P. T. (1975). The Significance of Structure-Foundation Interaction, Proc. 2
nd

Australia-New Zealand Conference on Geomechanics, Brisbane, IEQust, Vol. 1, No. 1,
pp. 79-82.

Brown, P. T. and Yu, S. K. R. (1986). Load Sequence and Structure-Foundation
Interaction, Jl. of Structural Engineering, ASCE, Vol. 112, No. 1, pp. 481-488.

Brown, P. T. and Wiesner, T. J. (1975). The Behaviour of Uniformly Loaded Piled Strip
Footings, Soils and Foundations, Vol. 15, No. 4, pp. 13-21.

References
333
Burmister, D. M. (1956). Stress and Displacement Charateristics of a Two Layer Rigid
Base Soil System: Influence Diagrams and Practical Application, Proc. Highway
Research Board., Vol. 35, pp. 773-814.

Butterfield , R. and Banerjee, P. K. (1971a). The Elastic Analysis of Compressible Piles
and Pile Groups, Gotechnique, Vol. 21, No. 1, pp. 43-60.

Butterfield , R. and Banerjee, P. K. (1971b). The Problem of Pile Group - Pile Cap
Interaction, Gotechnique, Vol. 21, No. 2, pp. 135-142.

Chan, K. S., Karasudhi, P. and Lee, S. L. (1974). Force at a Point in the Interior of
Layered Elastic Half-space, Int. Jl. of Solids Sturctures, Vol. 10, No. 11, pp. 1179-1199.

Cheung, Y. K. (1976). Finite Strip Method in Structural Analysis, Pergamon Press,
Oxford.

Cheung, Y. K. and Fan, S. C. (1979). Analysis of Pavement and Layered Foundations
by Finite Layer Method, Proc. 3
rd
International Conference on Numerical Methods in
Geomechanics, pp. 1129-1135.

Cheung, Y. K., Tham, L. G. and Guo, D. J. (1988). Analysis of Pile Group by Infinite
Layer Method, Gotechnique, Vol. 38, pp. 415-431.

Cheung, Y. K. and Nag, D. K. (1968). Plates and Beams on Elastic Foundations
Linear and Nonlinear Behaviour, Gotechnique, Vol. 18, pp. 250-260.

Cheung, Y, K. and Zienkiewicz, O. C. (1965) Plates and Tanks on Elastic Foundations
An Application of Finite Element Method, Int. Jl. of Solids Sturctures, Vol. 1, pp. 451-
461.

Chin, J. T., Chow, Y. K. and Poulos, H. G. (1990). Numerical Analysis of Axially
loaded Vertical Piles and Pile Groups, Computers and Geotechnics, Vol. 9, pp. 273-290.

Chow, Y. K. (1986). Analysis of Vertically Loaded Pile Groups, Int. Jl. For Numerical
and Analytical Methods in Geomechanics, Vol. 10, pp. 59-72.

Chow, Y. K. (1987a). Axial and Lateral Response of Pile Groups Embedded in
Nonhomogeneous Soil, Int. Jl. For Numerical and Analytical Methods in Geomechanics,
Vol. 11, no. 6, pp. 621-638.

Chow, Y. K. (1987b). Vertical Deformation of Rigid Foundations of Arbitrary Shape on
Layered Soil Media, Int. Jl. For Numerical and Analytical Methods in Geomechanics,
Vol. 11, No. 1, pp. 1-15.

Chow, Y. K. (1989). Axially Loaded Piles and Pile Groups Embedded in a Cross-
anisotropic Soil, Gotechnique, Vol. 39, No. 2, pp. 203-211.
References
334

Chow, Y. K. and Teh, C. I. (1991). Pile-Cap-Pile-Group Interaction in Non-
homogeneous Soil, Jl. of Geotechnical Engineering, Vol. 117, No. 11, pp. 1655-1667.

Chow, Y. K., Yong, K. Y. and Shen, W. Y. (2001). Analysis of Piled Raft Foundations
using a Variational Approach, The Int. Jl. of Geomechanics, Vol. 1, No. 2, pp. 129-147.

Clancy, P. and Randolph, M. F. (1993). An Approximate Analysis Procedure for Piled
Raft Foundations, Int. Jl. for Numerical and Analytical Methods in Geomechanics, Vol.
17, No. 12, pp. 849-869.

Cooke, R. W. (1974). The Settlement of Friction Pile Foundations, Proc. Conf. on Tall
Buildings, Kuala Lumper, pp. 7-19.

Costa, J. A. and Brebbia, C. A. (1985). The Boundary Element Method Applied to
Plates on Elastic Foundations, Engineering Analysis, Vol. 2, No. 4, pp. 174-183.

Costa, J. A. and Brebbia, C. A. (1986). On the Reduction of Domain Integrals to the
Boundary for the BEM Formulation of Plates on Elastic Foundation, Engineering
Analysis, Vol. 3, No. 2, pp. 123-126.

Damjanic, F., and Owen, D. R. J. (1984). Mapped Infinite Elements in Transient
Thermal Analysis, Computers and Structures, Vol. 19, pp. 673-687.

Davis, E. H. and Poulos. H. G. (1972). The Analysis of Piled Raft Systems, Australian
Geomechanics Jl., G2, No. 1, pp. 21-27.

de Sousa Coutinho, A. G. F. (2006). Data Reduction of Horizontal Load Full-Scale
Tests on Bored Concrete Piles and Pile Groups, Jl. of Geotechnical and
GeoEnvironmental Engineering, Vol. 132 No. 6, pp. 752-769.

de Paiva, J. B. and Butterfield, R. (1997). Boundary Element Analysis of Plate-soil
Interaction, Computer and Structures, Vol. 64, No.1-4, pp. 319-328.

Drucker, D. C. and Prager, W. (1952). Soil mechanics and Plastic Analysis of Limit
Design, Quart. Applied Mathematics, Vol. 10, No. 2.

El-Mossallamy, Y, Lutz, B. and Richter, Th. (2006). Innovative Application of Piled
Raft Foundation to Optimize the Design of High-Rise Buildings and Bridge Foundations,
Proc. 10
th
DFI/EFFE International Conference of Piling and Deep Foundations,
Amsterdam, The Netherlands, pp. 269-278.

El Sharnouby, B. and Novak, M. (1990). Stiffness Constants and Interaction Factors for
Vertical Response of Pile Groups, Canadian Geotechnical Jl., Vol. 27, pp. 813-822.

References
335
Franke, E., El-Mossallamy, Y. and Wittmann, P. (2000). Calculation Methods for Raft
Foundation in Germany, Design Applications of Raft Foundation, edited by Hemsley,
Thomas Telford, pp. 283-322.

Fraser, R. A and Wardle, L. J. (1975). A Rational Analysis of Shallow Foundations
Considering Soil-Structure Interaction, Australian Geomechanics Journal, pp. 20-25.

Fraser, R. A and Wardle, L. J. (1976). Numerical Analysis of Rectangular Rafts on
Layered Foundations, Gotechnique, Vol. 26, No. 4, pp. 613-630.

Gerrard, C. M. (1967). Stresses and Displacements in Layered, Cross-anisotropic,
Elastic Systems, Proc. 5
th
Australia-New Zealand Conference on Soil Mechanics and
Foundations, pp. 187-197.

Gerrard, C. M. and Harrison, W. J. (1971). The Analysis of a Loaded Half-space
Comprised of Anisotropic Layers, Technical Paper No. 10. Melbourne: CSIRO Division
of Applied Geomechanics.

Giroud, J. P. (1968). Settlement of a Linearly Loaded Rectangular Area, Jl. of Soil
Mechanics Foundations Division, ASCE, Vol. 94, No. SM4, pp. 813-831.

Giroud, J. P. (1969). Dplacement horizontal d'une droite particulire de la surface d'un
massif lastique semi-infini lineairement charg C.R. Acad. Sc. Paris, t. 268, pp. 252-
255, Srie A.

Guo, D. J., Tham, L. G. and Cheung, Y. K. (1987). Infinite Layer for the Analysis of a
Single Pile, Computers and Geotechnics, Vol. 3, pp. 229-249.

Guo, W. D. and Randolph, M. F. (1996). Settlement of Pile Groups in Non-
homogeneous Soil, Proc. 7
th
Australia-New Zealand Conference on Soil Mechanics and
Foundations, Vol. 1, pp. 631-636.

Guo, W. D. and Randolph, M. F. (1999). An Efficient Approach for Settlement
Prediction of Pile Groups, Gotechnique, Vol. 49, No. 2, pp. 161-179.

Hain, S. J. and Lee, I. K. (1974). Rational Analysis of Raft Foundation, Jl. of
Geotechnical Engineering, Vol. 100, No. GT7, pp. 843-860.

Hain, S. J. and Lee, I. K. (1978). The Analysis of Flexible Raft-Pile Systems,
Gotechnique, Vol. 28, No. 1, pp. 65-83.

Harr, M. E. (1966). Foundations of Theoretical Soil Mechanics. McGraw Hill, New York.

Hong, D. C., Chow, Y. K. and Yong, K. Y. (1999). A Method for the Analysis of Large
Vertically Loaded Pile Groups, Int. Jl. for Numerical and Analytical Methods in
Geomechanics, Vol. 23, pp. 243-262.
References
336

Hongladaromp, T., Chen, N. J. and Lee, S. L. (1973). Load Distributions in Rectangular
Footings on Piles, Geotechnical Engineering, Vol. 4, No.2, pp. 77-90.

Hooper, J. A. (1973). Observations on the Behaviour of a Piled-Raft Foundation on
London Clay, Proc. Instn. Civil Engineers, Vol. 55, No. 2, pp. 855-877.

Horikoshi, K. and Randolph, M. F. (1998). A Contribution to the Optimum Design of
Piled Rafts, Gotechnique, Vol. 48, No. 2, pp. 301-317.

Hu, C. and Hartley, G. A. (1994). Analysis of a Thin Plate on an Elastic Half-space,
Computers and Structures, Vol. 52, No. 2, pp. 227-235.

Kakurai, M., Yamashita, K. and Tomono, M. (1987). Settlement Behaviour of Piled Raft
Foundation on Soft Ground, Proc. 8
th
Asian Regional Conference on Soil Mechanics
and Foundation Engineering, Kyoto, Japan, Vol. 1, pp. 373-376.

Katsikadelis, J. T and Armenkas, A. E. (1984a). Plates on Elastic Foundation by BIE
Method, Jl. of Engineering Mechanics, Vol. 110, No. 7, pp. 1086-1105.

Katsikadelis, J. T and Armenkas, A. E. (1984b). Analysis of Clamped Plates on Elastic
Foundation by the Boundary Integral Equation Method, Jl. of Applied Mechanics, Vol.
51, pp. 574-580.

Katzenbach, R. and Reul, O. (1997). Design and Performance of Piled Rafts, Proc.
XIV
th
ICSMFE 97, Hamburg, Vol. 4, pp. 2253-2256.

Katzenbach, R., Arslan, U. and Moormann, C. (2000). Piled Raft Foundation Projects in
Germany, Design Applications of Raft Foundation, edited by Hemsley, Thomas Telford
pp. 323-391.

Kay, J. N and Cavagnaro, R. L. (1983). Settlement of Raft Foundations, Jl. of
Geotechnical Engineering, Vol. 109, No. 11, pp. 1367-1382.

Kim, K. N., Lee, S. H., Kim, K. S., Chung, C. K., Kim, M. M. and Lee, H. S. (2001).
Optimal Pile Arrangement for Minimizing Differential Settlements in Piled Raft
Foundations, Computers and Geotechnics, Vol. 28, pp. 235-253.

Kishida, H. and Nakai, S. (1977). Large Deflection of Single Pile under Horizontal
Load, Proc, 9
th
International Conference on Soil Mechanics & Foundation Engineering,
Tokyo, pp. 87-92.

Kitiyodom, P. and Matsumoto, T. (2003). A Simplified Analysis Method for Piled Raft
Foundations in Non-homogeneous Soils, Int Jl. for Numerical and Analytical Methods
in Geomechanics, Vol. 27, No. 2, pp. 88-109

References
337
Koning, H. (1957). Stress Distribution in a Homogeneous, Anisotropic, Elastic Semi-
infinite Solid, Proc. of 4
th
International Conference on Soil Mechanics and Foundation
Engineering, pp. 335-338.

Kuwabara, F. (1989). An Elastic Analysis for Piled Raft Foundations in a Homogeneous
Soil, Soils and Foundations, Vol. 29, No. 1, pp. 82-92.

Lee, C. Y. (1991). Discrete Layer Analysis of Axially Loaded Piles and Pile Groups,
Computers and Geotechnics, Vol. 11, pp. 295-313.

Lee, C. Y. (1993). Settlement of Pile Groups Practical Approach, Jl. of Geotechnical
Engineering, Vol. 119, No. 9, pp. 1449-1461.

Lee, C. Y. and Poulos, H. G. (1990). Axial Response Analysis of Piles in Vertically and
Horizontally Non-homogeneous Soils, Computers and Geotechnics, Vol. 9, pp. 133-148.

Lee, C. Y. and Small, J. C. (1991a). Finite Layer Analysis of Axially Loaded Piles, Jl.
of Geotechnical Engineering, Vol. 117, No. 11, pp. 1706-1722.

Lee, C. Y. and Small, J. C. (1991b). Finite Layer Analysis of Laterally Loaded Piles in
Cross-anisotropic Soils, Int Jl. for Numerical and Analytical Methods in Geomechanics,
Vol. 15, pp. 785-808.

Lee, I. K. and Brown, P. T. (1972). Structure-Foundation Interaction Analysis, Jl. of
the Structural Division, ASCE, Vol. 98, No. ST11, pp. 2413-2431.

Lee, K. M. and Xiao, Z. R. (2001). A Simplified Non-linear Approach for Pile Group
Settlement Analysis in Multilayered Soils, Canadian Geotechnical Jl., Vol. 38, pp.
1063-1080.

Leung, C. F. and Chow, Y. K. (1987). Response of Pile groups Subjected to Lateral
Loads, Int. Jl. for Numerical and Analytical Methods in Geomechanics, Vol. 11, pp.
307-314.

Liang, F. Y. and Chen, L. Z. (2004). A Modified Variational Approach for the Analysis
of Piled Raft Foundation, Mechanics Research Communications, Vol. 31, pp. 593-604.

Liu, W, and Novak, M. (1991) Soil-pile-cap Static Interaction Analysis by Finite and
Infinite Elements, Canadian Geotechnical Jl., Vol. 28, pp. 771-783.

Maharaj, D. K. and Gandhi, S. R. (2004). Non-linear Finite Element Analysis of Piled
Raft Foundations, Proc. Instn. Civil Engineers, Geotechnical Engineering, No. 157, pp.
107-113.

References
338
Mandal, J. J. and Ghosh, D. P. (1999). Prediction of Elastic Settlement of Rectangular
Raft Foundation A Coupled FE-BE Approach, Int. Jl. for Numerical and Analytical
Methods in Geomechanics, Vol. 23, pp. 263-273.

Mandolini, A. and Viggiani, C. (1997). Settlement of Piled Foundations, Gotechnique,
Vol. 47, No. 4, pp. 791-816.

Mendona, A. V. and de Paiva, J. B. (2000). A Boundary Element Method for the Static
Analysis of Raft Foundations on Piles, Engineering Analysis with Boundary Elements,
Vol. 24, pp. 237-247.

Mendona, A. V. and de Paiva, J. B. (2003). An Elastostatic FEM/BEM Analysis of
Vertically Loaded Raft and Piled Raft Foundation, Engineering Analysis with Boundary
Elements, Vol. 27 pp. 919-933.

Nishida, Y. (1966). Vertical Stress and Vertical Deformation of Ground Under a Deep
Circular Uniform Pressure in Semi-infinite, Proc. 1
st
Congress of the International
Society of Rock Mechanics, Vol.2, pp. 493-498.

Noorzaei, J., Viladkar, M. N. and Godbole, P. N. (1991). Soil-Structure Interaction of
Space Frame-Raft-Soil System A Parametric Study, Computers and Structures, Vol.
40, No. 5, pp. 1235-1247.

ONeill, M. W., Ghazzaly, O. I. and Ha, H. B. (1977). Analysis of Three-dimensional
Pile Groups with Nonlinear Soil Response and Pile-Soil-Pile Interaction, Proc. 9
th

Offshore Technology Conference, Vol. 2, pp. 245-256.

Ottaviani, M. (1975). Three-dimensional Finite Element Analysis of Vertically Loaded
Pile Groups, Gotechnique, Vol. 25, No. 2, pp. 159-174.

Poulos, H. G. (1968). Analysis of the Settlement of Pile Groups, Gotechnique, Vol. 18,
pp. 449-471.

Poulos, H. G. (1973). Load-deflection Prediction for Laterally Loaded Piles, Australian
Geomechanics Jl., Vol. 3, No. 1, pp. 1-8.

Poulos, H. G. (1971). Behaviour of Laterally Loaded Piles: II Pile Groups, Jl. of the
Soil Mechanics and Foundations Division, ASCE, Vol. 97 (SM5), pp. 733-751.

Poulos, H. G. (1979). Settlement of Single Piles in Non-homogeneous Soil, Jl. of the
Geotechnical Engineering Division, ASCE, Vol. 105, No. GT5, pp. 627-641.

Poulos, H. G. (1990). DEFPIG Users Manual. Centre for Geotechnical Research,
University of Sydney.

References
339
Poulos, H. G. (1993). Piled Rafts in Swelling or Consolidating Soils, Jl. Geotech.
Engng, ASCE, Vol. 119, No. 2, pp. 374-380.

Poulos, H. G. (1994). An Approximate Numerical Analysis of Pile-raft Interaction, Int.
Jl. for Numerical and Analytical Methods in Geomechanics, Vol. 20, pp. 57-72.

Poulos, H. G. and Davis, E. H. (1968). The Settlement Behaviour of Single Axially
Loaded Incompressible Piles and Piers, Gotechnique, Vol. 18, pp. 351-371.

Poulos, H. G. and Davis, E. H. (1974). Elastic Solutions for Soil and Rock Mechanics,
John Wiley & Sons Inc., Sydney.

Poulos, H. G. and Davis, E. H. (1980). Pile Foundation Analysis and Design, John Wiley
& Sons Inc., New York.

Prakoso, W. A. and Kulhawy, F. H. (2001). Contribution to Piled Raft Foundation
Design, Jl. of Geotechnical and GeoEnvironmental Engineering, Vol. 127, No. 1, pp.
17-24.

Pressley, J. S. and Poulos, H. G. (1986). Finite Element Analysis of Mechanisms of Pile
Group Behaviour, Int. Jl. for Numerical and Analytical Methods in Geomechanics, Vol.
10, No. 2, pp. 213-221.

Przemieniecki, J. S. (1968). Theory of Matrix Structural Analysis, McGraw-Hill Book
Co., Inc., New York, N. Y.

Randolph, M. F. and Wroth, C. P. (1978). Analysis of Deformation of Vertically Loaded
Piles, Jl. of the Geotechnical Engineering Division, ASCE, Vol. 104(GT12), pp. 1462-
1488.

Randolph, M. F. and Wroth, C. P. (1979). An Analysis of the Vertical Deformation of
Pile Groups, Gotechnique, Vol. 29, No. 4, pp. 423-439.

Randolph, M. F. (1981). The Response of Flexible Piles to Lateral Loading,
Gotechnique, Vol. 31, No. 2, pp. 247-259.

Randolph, M. F. (1983). Design of piled raft foundations, Proceedings of the
international symposium on recent developments in laboratory and field tests and
analysis of geotechnical problems, Bangkok, pp. 525537.

Randolph, M. F. (1994). Design Methods for Pile Groups and Piled Rafts, Proc. 13
th

International Conference on Soil Mechanics and Foundation Engineering, New Delhi,
Vol. 5, pp. 61-82.

Rashed, Y. F. (2005). A Boundary/Domain Element Method for Analysis of Building
Raft Foundations, Engineering Analysis with Boundary Elements, Vol. 29, pp. 859-877.
References
340

Reul, O. (1998). Soil-structure-interaction of a Piled Raft Foundation of a 121 m high
Office Building in Berlin, Proc. 12
th
European Young Geotechnical Engineers
Conference, Tallinn, Estonia, pp. 1-12.

Reul, O. and Randolph, M. F. (2003). Piled Rafts in Overconsolidated Clay:
Comparison of In situ Measurements and Numerical Analyses, Gotechnique, Vol. 53,
No. 3, pp. 301-315.

Reul, O. and Randolph, M. F. (2004). Design Strategies for Piled Rafts Subjected to
Nonuniform Vertical Loading, Jl. of Geotechnical and GeoEnvironmental Engineering,
Vol. 130, No. 1, pp. 1-10.

Robertson, P. K., Davies, M. P., and Campanella, R. G. (1989). Design of Laterally
Loaded Driven Piles using the Flat Dilatometer., Geotechnical Testing Jl., Vol. 12, No.
1, pp. 30-39.

Ruesta, P. F. and Townsend, F. C. (1997). Evaluation of Laterally Loaded Pile Group at
Roosevelt Bridge, Jl. of Geotechnical and GeoEnvironmental Engineering, ASCE, Vol.
123, No. 12, pp. 1153 - 1161.

Russo, G. (1998). Numerical Analysis of Piled rafts, Int Jl. for Numerical and
Analytical Methods in Geomechanics, Vol. 22, No. 6, pp. 477-493.

Sapountzakis, E. J. and Katsikadelis, J. T. (1992). Unilaterally Supported Plates on
Elastic Foundations by the Boundary Element Method, Jl. of Applied Mechanics, Vol.
59, pp. 580-586.

Sharma, K. G., Nagpal, A. K. and Garg, M. K. (1984). Finite Element Analysis of Rafts
Resting on Elastic Half Space, Indian Geotechnical Jl., Vol. 14, No. 1, pp. 28-39.

Shen, W. Y., Chow, Y. K. and Yong, K. Y. (1999). Variational Solution for Vertically
Loaded Pile Groups in an Elastic Half-space, Gotechnique, Vol. 49, No. 2, pp. 199-213.

Shen, W. Y., Chow, Y. K. and Yong, K. Y. (2000). A Variational Approach for the
Analysis of Pile Group-Pile Cap Interaction, Gotechnique, Vol. 50, No. 4, pp. 349-357.

Shen, W. Y. and Teh, C. I. (2002). Analysis of Laterally Loaded Pile Groups using a
Variational Approach, Gotechnique, Vol. 52, No. 3, pp. 201-208.

Sinha, J. Piled raft foundations subjected to swelling and shrinking soils. Ph.D. Thesis,
Univ. Sydney, Australia, 1997.

Small, J. C. (2001). Practical Solutions to Soil-Structure Interaction Problems, Prog.
Struct. Engng. Mater., Vol. 3, pp. 305-314.

References
341
Small, J. C. (2002). Soil-Structure Interaction, E. H. Davis Memorial Lecture,
Australian Geomechanics, March, pp. 1-16.

Small, J. C. and Booker, J. R. (1984). Finite Layer Analysis of Layered Elastic Materials
Using Flexibility Approach. Part I. Strip Loadings, Int. Jl. for Numerical Methods in
Engineering, Vol. 20, pp. 1025-1037.

Small, J. C. and Booker, J. R. (1986). Finite Layer Analysis of Layered Elastic Materials
Using Flexibility Approach. Part II. Circular and Rectangular Loadings, Int. Jl. for
Numerical Methods in Engineering, Vol. 23, pp. 959-978.

Smith, I. M. and Wang, A. (1998). Analysis of Piled Rafts, Int. Jl. for Numerical and
Analytical Methods in Geomechanics, Vol. 22, pp. 777-790.

Sneddon, I. N. (1951). Fourier Transforms, McGraw-Hill, New York.

Sommer, H. (1991), Entwicklung der Hochhausgrndungen in Frankfurt/Main,
Festkolloquium 20 Jahre Grundbauinstitut Prof. Dr.-Ing. J. Sommer und Partner, pp. 47-
62, Germany.

Southcott, P. H. and Small, J. C. (1996). Finite Layer Analysis of Vertically Loaded
Piles and Pile Groups, Computers and Geotechnics, Vol. 18, No. 1, pp. 47-63.

Svec, O. J. and Gladwell, G. M. L. (1973). A Triangular Plate Bending Element for
Contact Problems, Int. Jl. of Solids Structures, Vol. 9, pp. 433-446.

Ta, L. D. (1996). A Finite Layer Method for Analysis of Pile Groups, Rafts and Piled Raft
Foundations in Layered Soils. Ph.D. Thesis, Univ. Sydney, Australia, 1996.

Ta, L. D. and Small, J. C. (1995). Finite Layer Analysis of Pile Groups in Layered
Soils, Proc. the 5
th
International Symposium on Numerical Models in Geomechanics
NUMOG V, Davos, Switzerland, 6-8 September, pp. 577-582.

Ta, L. D. and Small, J. C. (1996). Analysis of Piled Raft Systems in Layered Soils, Int.
Jl. for Numerical and Analytical Methods in Geomechanics, Vol. 20, pp. 57-72.

Tandijiria, V., Valliappan, S. and Khalili, N. (1999). Optimization Approach Applied to
Non-linear Analysis of Raft-Pile Foundations, Structural Engineering and Mechanics,
Vol. 7, No. 6.

Tejchman, A. (2000). Calculations and Masurements for Three Piled Foundations,
Internal ISSMGE TC-18 Memorandum, October, Technical University of Gdansk.

Thaher, M. and Jessberger, H. L. (1991a). The Behaviour of Pile-Raft Foundations,
Investigated in Centrifuge Model Tests, Centrifuge 91, Boulder, Colorado, USA, pp.
101-106.
References
342

Thaher, M. and Jessberger, H. L. (1991b). Investigation of the Behaviour of Pile-Raft
Foundations by Centrifuge Modelling, Proc. 10
th
European Conference on Soil
Mechanics and Foundation Engineering, Vol. 2, pp. 597-603.

Valliappan, S. Lee, I. K. and Boonlualohr, P. (1974). Settlement Analysis of Piles in
Layered Soil, Proc. 7
th
Conference of the Australia Road Research Board, Vol. 7, Part.
7, pp. 144-153.

Valliappan, S., Tandjiria, V. and Khalili, N. (1999). Design of Raft-Pile Foundtion
Using Combined Optimization and Finite Element Approach, Int. Jl. for Numerical and
Analytical Methods in Geomechanics, Vol. 23, pp. 1043-1065.

Van Impe, W. F. (2001). Methods of Analysis of Piled Raft Foundations, International
Society of Soil Mechanics and Geotechnical Engineering, Technical Committee TC18 on
Piled Foundations.

Vermeer, P. A. and Brinkgreve, R. B. J. (1995). PLAXIS users manual version 6.1,
Balkema, Rotterdam, The Netherlands.

Viladkar, M. N., Godbole, P. N. and Noorzaei, J. (1992). Space Frame-Raft-Soil
Interaction Including Effect of Slab Stiffness, Computers and Structures, Vol. 43, No. 1,
pp. 93-106.

Wardle, L. J. and Fraser, R. A. (1974). Finite Element Analysis of a Plate on Layered
Cross-Anisotropic Foundation, Proc. 1
st
International Conference for Finite Element
Method Engineering, University of NSW, pp. 565-578.

Westmann, R. A. (1963). Layered Systems subjected to Surface Shears, Jl. of
Engineering Mechanics Division, ASCE, Vol. 89, No. EM6, pp. 177-191.

Wiesner, T. J. (1991). Various Applications of Piled Raft Analysis, Computer Methods
and Advances in Geomechanics, Beer, Booker & Carter (eds), 1991 Balkema, Rotterdam,
pp. 1035-1039.

Winkler, E. (1867). Die Lehre von der Elastizitaz und Festigkeit, Prague.

Wood, L. A. and Larnach, W. J. (1974). The Effects of Soil-structure Interaction on Raft
Foundations, Proc. Conf. Settlement of Structures, Cambridge, Pentech Press, pp. 460-
470.

Wood, L. A. and Larnach, W. J. (1975). The Interactive Behaviour of a Soil-structure
System and its Effect on Settlements, Int. Symp. Geot. Struct., Univ. N. S. W., Australia,
pp. 75-88.

References
343
Wood, L. A. (1977). The Economic Analysis of Raft Foundations, Int. Jl. for
Numerical and Analytical Methods in Geomechanics, Vol. 1, pp. 397-405.

Xu, K. J. and Poulos, H. G. (2000). General Elastic Analysis of Piles and Pile Groups,
Int. Jl. for Numerical and Analytical Methods in Geomechanics, Vol. 24, pp. 1109-1138.

Yamashita, K. Tomono, M. and Kakurai, M. (1987). A Method for Estimating
Immediate Settlement of Piles and Pile Groups, Soils and Foundations, Vol. 27, No. 1,
pp. 61-76.

Yao, Z. E. and Zhang, J. R. (1985). An Assessment of the Effects of Structure/Raft/Soil
Interaction, Proc. 5
th
Int. Conference on Numerical Methods in Geomechanics, 1-5 April,
Nagoya, Japan, Vol. 2, pp. 813-819.

Zhang, B. Q and Small, J. C. (1991). Finite Layer Analysis of Soil-Raft Interaction,
Research Report No. 642, School of Civil and Mining Engineering, The University of
Sydney.

Zhang, B. Q and Small, J. C. (1994). Finite Layer Analysis of Soil-Raft-Structure
Interaction, Proc. XIII
th
ICSMFE 94, 5 10 January, New Delhi, Vol. 2, pp. 587-590.

Zhang, H. H. (2000). Finite Layer Method for Analysis of Piled Raft Foundations. Ph.D.
Thesis, Univ. Sydney, Australia, 2000.

Zhang, H. H. and Small, J. C. (1999). Analysis of Axially and Laterally Loaded Pile
Groups Embedded in Layered Soils, Proc. 8
th
Australia New Zealand Conference on
Geomechanics, 15-17 Feburary, Hobart, Tasmania, Edited by N. Vitharana and R.
Coleman, Vol. 1, pp. 475-483.

Zhang, H. H. and Small, J. C. (2000). Analysis of Capped Pile Groups subjected to
Horizontal and Vertical Loads, Computers and Geotechnics, Vol. 26, pp. 1-21.

Zhao, X. and Cao, M. (1985). Soil-Structure Interaction Analysis and its Application to
Foundation Design in Shanghai, Proc. 5
th
Int. Conference on Numerical Methods in
Geomechanics, 1-5 April, Nagoya, Japan, Vol. 2, pp. 805-811.

Zhemochkin, B. N. and Sinitsyn, A. P. (1962). Practical Methods of Designing
Foundation Beams and Slabs Resting on an Elastic Foundation, 2
nd
ed., State Publishing
House for Lit. on Struct., Archit. and Struct. Matl., Moscow.