Analysis of 3D-Structures Resting On Piled Raft PDF

Suez Canal University
Faculty of Engineering
Analysis of 3D-Structures Resting on Piled Raft
BY
Waleed Hamdy Mohamed El Kamash
M.Sc. Civil Engineering, 2004

A Thesis
Submitted in Partial Fulfillment
For the Requirements of the Degree of
Doctor of Philosophy
In
Civil Engineering
Supervised by
Prof. Dr. Faisal Khalil Ibrahim

Professor of Structural Engineering,
Faculty of Engineering,
Dr. Mohamed Mosaad El-Gendy Dr. Rafeek Wadieh Salib

Assoc. Prof. Civil Eng. Dept., Assist. Prof. Civil Eng. Dept.,
Faculty of Engineering, Faculty of Engineering,
Suez Canal University Suez Canal University
2009

Supervision Committee
Name, Title & Affiliation Signature
1- Prof. Dr. Faisal Khalil Ibrahim ..
Prof. of Structural Engineering

2- Dr. Mohamed Mosaad El-Gendy ..
Assoc. Prof. Civil Eng. Dept.

3- Dr. Rafeek Wadieh Abdel Missih Salib ..
Assist. Prof. Civil Eng. Dept.

i
Approval Sheet
Analysis of 3D-Structures Resting on Piled Raft
BY
Waleed Hamdy Mohamed El Kamash
Name, Title & Affiliation Signature
3- Prof. Dr. Faisal Khalil Ibrahim .
Prof. of Structural Engineering
4- Dr. Mohamed Mosaad El-Gendy .
Assoc. Prof. Civil Eng. Dept.
ii
Acknowledgements
I wish to express my deepest gratitude and sincere appreciation to Prof. Dr.
Faisal Khalil Ibrahim, Professor of Structural Engineering, Suez Canal
University, for his kind supervision, generous support, valuable suggestions
and constant encouragement during all phases of this research work.
I wish to express my deep thanks and gratitude to Prof. Dr. Mohamed Mosaad
El- Gendy, Associate Professor of Soil Mechanics and Foundation
Engineering, Suez Canal University, for his kind supervision and useful
suggestions during this work.
I am deeply indebted to Dr. Rafeek Wadieh Abdel Missih Salib, Assistant
Professor, Civil Eng. Dept., Suez Canal University, for his constructive
criticisms, precious and generous support that have truly helped in bringing
this work to a successful end.
Words, however ample, would be insufficient to express my gratitude to my
great mother soul who she is in my heart all the time and also to my family
specially my dear wife for their continuos sacrifice and fruitful care.
iii
ABSTRACT
SYNOPSIS
In the present study, there is a numerical model of a numerical analysis

based on the elastic theory is developed for both axially and laterally loaded
3D-space structures with piled raft which are embedded into a non-
homogenous soil medium. Such model is employed in which the raft is
presented as plates using the Framework Analogy; piles are modeled by
rigid beam elements, while the soil is treated as continuum medium. Full
interactions among structural members pilesoilpile, pilesoilraft and
raftsoilraft are taken into account The main objective of this thesis is to
develop a perfect system of structural analysis that makes engineers able to
design a 3D-space multi-storey structure on piled raft foundation under
conditions of space forces. The effect of superstructure and non-
homogeneity of the soil is considered in order to add more reality to the
model. The nonlinear analysis of the piled raft on continuum model of soil
using Equivalent stiffness Technique is employed to be applicable for piled
raft-superstructures in order to reduce the stiffness matrix size and
consequently saving time in the iteration solution. Herein, a version of
computer program ASTNIII is edited by the author to simulate the field
problem against vertical and lateral forces. Moreover, another version of
computer program ASTNII is edited by the author to simulate the large piled
raft-superstructure under vertical loads using Composed Coefficient
Technique. Such technique makes the FEM has the ability to solve large
problems more rapidly in a high accuracy. Verification models are used to
approve new versions of the programs ASTNII and ASTNIII. Two parametric
studies of 3D-space structures with piled raft on Port-Said soil medium
iv
ABSTRACT
exposed to vertical loads and seismic loads are carried out by ASTNII and
ASTNIII respectively. Discussion for the numerical results is performed and
conclusions are drawn.
v
TABLE OF CONTENTS
TABLE OF CONTENTS
Page
CHAPTER 1: INTRODUCTION
1.1. Background....... 1
1.2. Objective of Thesis.... 3
1.3. Layout of Thesis........ 4
CHAPTER 2: LITERATURE REVIEW

2.1. Introduction....... 7
2.2. Rafts...... 7
2.2.1. Analytical methods . 8
2.2.2 Boundary Element methods..... 9
2.3.2. Finite Element method.... 7
2.2.3 Finite Element methods... 10
2.2.4 Hybrid approach.. 12
2.2.5 Resisting of Rafts against lateral loads.. 13
2.3 Pile Group . 14
2.3.1 Simplified analytical method 14
2.3.2 Hybrid method... 15
2.3.3 Boundary Element Method... 17
2.3.4 Finite Element Method.. 20
2.3.5 Infinite Layer Method... 20
2.3.6 Finite Layer Method.. 21
2.4 Piled Raft ... 24
2.4.1 Approximation Method 24
2.4.2 Boundary Element Method.. 25
v
TABLE OF CONTENTS
Page
2.4.3 The 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 30
2.4.6 Variational Approach.. 30
2.5. Lateral Load-Resisting Systems 31
2.5.1. Types of lateral loads. 31
2.5.2 Seismic loads... 31
2.5.3 Distribution of shear forces among stories 34
2.5.4 Types of resisting systems... 35
CHAPTER 3: PORT-SAID SOIL PROPERTIES AND

PARAMETERS
3.1. Introduction. 37
3.2. Port-Said soil properties.. 38
3.2.1. Investigation area 38
3.2.2. Soil log 38
3.3. Soil parameter. 44
CHAPTER 4: ANALYSIS OF PILED RAFT IN LAYERED SOIL

4.1. Introduction... 50
4.2. Analysis of Rafts... 51
4.2.1. Flexure of a rectangular element of a plate. 52
4.2.2. Flexure of the equivalent grid framework model 54
4.2.3. Evaluation of the Cross-Sectional Properties of Members. 55
4.2.4. Raft Subjected to Bending Action... 56
4.2.5. Rafts on half-space Soil... 62
vi
TABLE OF CONTENTS
Page
4.2.6. Modeling the layered Soil... 63
4.2.7. Influence Matrix for the Soil... 64
4.2.8. Linear analysis of the raft 65
4.3. Analysis single pile and pile groups.. 66
4.3.1. Modeling single pile in half-space soil (vertical loads).. 66
4.3.2. Modeling the layered soil 68
4.3.3. Soil flexibility for a single pile 68
4.3.4. Linear analysis. 68
4.3.5. Nonlinear analysis... 70
4.3.6. Modeling pile groups... 72
4.3.7. Soil stiffness for freestanding raft (pile groups under vertical
loads)... 72
4.4. Modeling piled raft 75
4.4.1. Linear analysis 79
4.4.1.I. Analysis of piled flexible raft.. 79
4.4.1.II. Analysis of piled rigid raft. 79
4.4.1.III. Analysis of piled elastic raft... 80
4 4.4.2. Nonlinear analysis. 82
4 4.4.3. Iterative Procedure 82
4.5. The pile model against lateral loads.. 84
4.5.1. Modeling of the layered soil... 85
4.5.2. Soil stiffness for pile groups against lateral
forces. 86
4.5.3. Linear analysis 87
4.5.4. Nonlinear analysis.. 89
vii
TABLE OF CONTENTS
Page
4.5.6. Limiting pile-soil stresses... 89
4.5.6.1. Cohesive soil. 90
4.5.6.2. Cohesionless soil 90
4.6. Soil-structure interaction... 90
4.6.1. Effect of superstructure on foundations.. 91
4.6.2. Piled raft-3D superstructure interaction.. 93
CHAPTER 5: COMPUTER PROGRAM AND VALIDITY OF

PROCEDURE
5.1. Introduction 102
5.2. Solution Algorithm and the Computer Program 103
5.2.1. The program ASTNII... 103
5.2.2. Subprogram READ DATA. 104
5.2.3 The graphic screen 108
5.2.4. Subprogram OUTPUT DATA.. 108
5.2.5. The program MAIN PROGRAM.. 108
5.2.6. Subprogram SAVE FORCES... 112
5.2.7. Subprogram RUN TIME.. 112
5.3. Numerical analysis and results.. 126
5.3.1. Test problem 1: verification of axially piled raft considering
nonlinear soil medium (Rabiei,M.,2009)... 126
5.3.2. Test problem 2: Distribution of shear resistance and
settlement curves for an axially single pile (Poulus,1980)... 130
5.3.3. Test problem 3: Single pile response (Basile,2003)... 131
5.3.4. Test problem 4: Pile group settlement(Basile,2003)... 132
5.3.5. Test problem 5 Case study: Piled raft of Torhaus (El
viii
TABLE OF CONTENTS
Page
Gendy,2007). 134
5.3.5.1. Soil properties.. 135
5.3.5.2. Raft and pile material.. 136
5.3.5.3. Analysis of the piled raft. 136
5.3.5.4. Comparison with three-dimensional finite element
analysis and field measurements.. 136
5.3.5.5. Comparing among different analysis types. 138
5.3.6. Test problem 6: Pile Group under General Loading
Conditions (Basile, 2003).. 142
5.3.7. Test problem 7: Lateral load distribution (Basile,2003). 150
5.4. conclusion.. 151
CHAPTER 6: AXIALLY LOADED BUILDINGS ON PORT

SAID SOIL MEDUIM
6.1. Introduction 151
6.2. Mathematical Model. 152
6.3. Parametric study 152
6.3.1. Description, geometry and properties of models 152
6.3.2. Soil properties. 155
6.3.3. Modulus of compressibility. 156
6.3.4. Limited shaft friction... 158
6.3.5. Concrete material properties... 159
6.3.6. Analysis... 159
6.4. Results and discussion... 160
6.4.1. Effect of pile length 161
6.4.1.1. Settlement and differential settlement. 161
ix
TABLE OF CONTENTS
Page
6.4.1.2. Moment 162
6.4.1.3. Pile load... 162
6.4.1.4. Bearing factor.. 162
6.4.2. Effect of pile diameter 163
6.4.2.2. Moment 164
6.4.2.3. Pile load... 164
6.4.2.4. Bearing factor 164
6.4.3. Effect of varying pile length. 165
6.4.3.2. Moment 166
6.4.3.3. Pile load... 166
6.4.4. Effect of superstructure.. 167
6.4.4.1. Settlement and differential settlement 167
6.4.4.2. Moment 168
6.4.4.3. Pile load... 168
6.5. Conclusions 183
CHAPTER 7: BUILDINGS ON PORT SAID SOIL MEDUIM

UNDER SEISMIC LOADS
7.1. Introduction... 184
7.2. Mathematical Model. 185
7.3. Parametric study 186
x
TABLE OF CONTENTS
Page
7.3.1. Description, geometry and properties of models... 186
7.3.2. Soil properties 186
7.3.3. Modulus of compressibility... 187
7.3.4. Limited shaft friction. 188
7.3.5. Concrete material properties.. 188
7.4. Analysis. 188
7.5. Results and discussion... 190
7.5.1. Effect of pile length 190
7.5.1.1. Vertical settlement and differential settlement. 190
7.5.1.2. Lateral settlement and drift ratio... 191
7.5.1.3. Moment. 193
7.5.1.4. Vertical pile load... 193
7.5.1.5. Lateral pile load 194
7.5.1.6. Bearing factor... 194
7.5.2. Effect of pile diameter... 194
7.5.2.3. Moment. 197
7.5.2.6. Bearing factor 198
7.5.3. Effect of varying in pile length.. 198
7.5.3.3. Moment. 201
7.5.3.4. Vertical pile load 202
xi
TABLE OF CONTENTS
Page
7.5.4. Effect of raft thickness... 203
7.5.4.1. Vertical settlement and differential settlement 203
7.5.4.3. Moment. 204
7.5.4.5. Lateral pile load.... 205
7.6. Conclusions... 206
CHAPTER 8: SUMMARY AND CONCLUSIONS

7.1. Summary... 233
7.2. Conclusions... 234
7.3. Suggestions for the Future Research. 239
REFERENCES... 240
APPENDIX I Displacement at the center due to a cylindrical
surface stress..... 252
APPENDIX II Displacement at the center due to a circular uniform
load 257
APPENDIX III Lateral displacement at the center due to a
rectangular uniform load........................ 259
xii
LIST OF TABLES
LIST OF TABLES
Table No. Title Page
Table (2.1) Reduction coefficient 33
Table (2.2) Soil coefficient factors 34
Table (3.1) Thickness of soil layers in the different zones in meter 40
Table (3.2) Soil parameters 47
Table (5.1) Comparison between results obtained by 3D FE- 137

Analysis, ElGendy analysis and field measurements
with those obtained by the present analysis
Table (5.2) Comparison between results of different analysis types 139
Table (5.3) Input data 142
Table (6.1) Models and dimensions of columns 156
Table (6.2) Soil parameters 159
xiii
LIST OF FIGURES
LIST OF FIGURES
Figure No. Title Page
Figure (2.1) Stresses on piles and the soil for a two pile group
under the axial load 36
Figure (2.2) Stresses acting on piles and soil for a two pile
group under lateral loads 36
Figure (3.1) Port-Said zones 39
Figure (3.2) Main soil properties 43
Figure (3.3) Modulus of compressibility with depth for all
46
zones
Figure (3.4) Main soil profile of Port-Said 48
Figure (3.5) Typical soil properties of Port-Said area 49
Figure (4.1) Framework representation of plate (Lattice of
96
plate)
Figure (4.2) Plate element and equivalent Framework model 96
Figure (4.3) Member Axes and freedoms for a 3D-space

97
frame
Figure (4.4) Calculation of the direction cosines for the
member x member x axis 97
Figure (4.5) Principal axes 97
Figure (4.6) 3D-Space framework analogy of raft-pile-soil
98
modeling
Figure (4.7) Pile geometry and elements 98
xiv
LIST OF FIGURES

Figure (4.8) Modeling freestanding raft on pile groups 99
Figure (4.9) Modeling piled raft 100
Figure (4.10) Load-settlement curve of a single pile (hyperbolic
relation) 101
Figure (4.11) Interface lateral forces of pile group in layered
101
soil
Figure (5.1) The Master program ASTNII & ASTNIII 113
Figure (5.2) The flow chart of READING DATA 116
Figure (5.3) Flow chart of MAIN PROGRAM of ASTNII 118
Figure (5.4) Flow chart of the adding stiffness process 119
Figure (5.5) Flow chart of the iteration process of ASTNII 121
Figure (5.6) Flow chart of MAIN PROGRAM of ASTNIII 123
Figure (5.7) Flow chart of the iteration process of ASTNIII

125
Figure (5.8-a) Load capacity of each pile = 875 [kN] Pile and
127
load configuration
Figure (5.8-b) Model condition and material properties 127
Figure (5.9-a) Comparison of various methods for load-settlement
128
analysis
Figure (5.9-b) Effect of raft thickness on maximum bending
128
moment
Figure (5.9-c) Effect of raft thickness on maximum settlement 129
Figure (5.9-d) Effect of raft thickness on diffrential settlement 129
Figure (5.10-a) Distribution of shear stress along 130
xv
LIST OF FIGURES
Figure No. Title

Page
Figure (5.10.b) Settlement for varies pile lengths

131
Figure (5.11) Comparison of load-settlement response for single

pile (EP = 30,000GPa) 132
Figure (5.12) Comparison of different pile group analysis

methods for s/d=2.5 133
Figure (5.13) Layout of Tourhous with piled raft 140

Figure (5.14) Mesh of piled raft with piles 140
Figure (5.15) Comparison between pile loads obtained by 3D FE
analysis, El Gendy analysis and field
measurments and those obtained by the 141
present analysis (Piles 3, 4 and 5).
Figure (5.16) Comparison between pile loads obtained by 3D FE

analysis, El Gendy analysis and field
measurments and those obtained by the 141
Figure (5.17) Group of 3 piles considered in comparison of

method 143
Figure (5.18) Comparison between pile loads V1 obtained by

Equivelant bent analysis, DEFPIG, GEPAN,
PGROUPN, PIGLET and those analyzed by
ASTNIII 143
xvi
LIST OF FIGURES
Figure (5.19) Comparison between pile loads V2 obtained by

144
ASTNIII
Figure (5.20) Fig. 5.19. Comparison between pile loads V3
obtained by Equivelant bent analysis,
DEFPIG, GEPAN, PGROUPN, PIGLET 144
and those analyzed by ASTNIII
Figure (5.21) Comparison between lateral pile loads H1 obtained

by Equivelant bent analysis, DEFPIG, GEPAN,
145
ASTNIII
by Equivelant bent analysis, DEFPIG,
GEPAN, PGROUPN, PIGLET and those 145
analyzed by ASTNIII

by Equivelant bent analysis, DEFPIG,
GEPAN, PGROUPN, PIGLET and those 146
analyzed by ASTNIII
Figure (5.24) Comparison between pile moment M1 obtained by

146
ASTNIII
xvii
LIST OF FIGURES

147
ASTNIII
147
ASTNIII
Figure (5.27) Comparison between settlement w obtained by
148
ASTNIII
Figure (5.28) Comparison between lateral displacement u
obtained by Equivelant bent analysis, DEFPIG,
148
GEPAN, PGROUPN, PIGLET and those analyzed
by ASTNIII
Figure (5.29) Comparison between pile rotation obtained by
PGROUPN, PIGLET and those analyzed by 149
ASTNIII
Figure (5.30) Comparison of lateral load distribution to individual
piles in 5 5 pile group 153
Figure (6.1) 3D-12 storeys-space structures of square and
rectangular typical floors 150
Figure (6.2) Typical floors for the two models with column
arrangement 154
xviii
LIST OF FIGURES
Figure No Title Page
Figure (6.3) Piled rafts of the two shape models

155
Figure (6.4) Modulus of compressibility with depth for all

zones
157
Figure (6.5) Settlement at the center of the building for pile

diameter = 0.8 [m]
169
Figure (6.6) Settlement at the center of the building for pile

length = 20 [m] 169
Figure (6.7) Settlement at the center of the building when
varying pile length 170
Figure (6.8) Settlement at the center of the building with and
without superstructure effect 170
Figure (6.9) Settlement at the center of the building without
superstructure effect 171
Figure (6.10) Differential settlement between central and
corner points for pile diameter = 0.8 [m] 171
corner points for pile length = 20 [m] 172
corner points when varying pile length 172
Figure (6.13) Moment in the raft at section I-I for square-shape
model for pile diameter = 0.8 [m] 173
xix
LIST OF FIGURES
Figure No. Title

Page

model for pile length = 20 [m]
173

model when varying pile length 174
model with and without superstructure effect
174
Figure (6.17) Moment in the raft at section II-II for
rectangular-shape model for pile diameter = 0.8
[m] 175
rectangular-shape model for pile length = 20 [m]
175
Figure (6.19) Moment in the raft at section II-II for rectangular-
shape model when varying pile length 176
rectangular-shape model with and without
176
superstructure effect
Figure (6.21) Pile load for square-shape model for pile
diameter = 0.8 [m] 177
Figure (6.22) Pile load for square-shape model for pile length =
20 [m] 177
xx
LIST OF FIGURES

Figure (6.23) Pile load for square-shape model when varying pile
length 178
Figure (6.24) Pile load for square-shape model with and
Figure (6.25) Pile load for rectangular-shape model for pile
Figure (6.26) Pile load for rectangular-shape model for pile length
= 20 [m] 179
Figure (6.27) Pile load for rectangular-shape model when varying
pile length 180
Figure (6.28) Pile load for rectangular-shape model with and
Figure (6.29) Bearing factor for pile diameter = 0.8 [m] 181
Figure (6.30) Bearing factor for pile length = 20 [m] 181
Figure (7.1) Vertical settlement at the center of the building for

pile diameter = 0.8 [m] 207
Figure (7.2) Lateral settlement x-direction at the center of the
building for pile diameter = 0.8 [m]
207
Figure (7.3) Lateral settlement y-direction at the center of the

building for pile diameter = 0.8 [m] 208
Figure (7.4) Vertical settlement at the center of the building for
pile length =20 [m]
208
xxi
LIST OF FIGURES
Title Page
Figure No.
building for pile length =20 [m] 209
building for pile length =20 [m] 209
Figure (7.7) Vertical settlement at the center of the building
when varying pile length 210
building when varying pile length
210
building when varying pile length
211
Figure (7.10) Vertical settlement at the center of the building
when varying raft thickness 211
building when varying raft thickness
212
building when varying raft thickness
212
Figure (7.13) Vertical differential settlement of the building for
pile diameter =0.8 [m] 213
Figure (7.14) Drift ratio in x-direction of the building for pile
diameter =0.8 [m] 213
Figure (7.15) Drift ratio in y-direction of the building for pile
diameter =0.8 [m] 214
Figure (7.16) Vertical differential settlement of the building for
pile length =20 [m] 214
xxii
LIST OF FIGURES
Page
Figure No. Title
Figure (7.17) Drift ratio in x-direction of the building for pile
length =20 [m] 215
Figure (7.18) Drift ratio in y-direction of the building for pile
length =20 [m] 215
Figure (7.19) Vertical differential settlement of the building with
Figure (7.20) Drift ratio in x-direction of the building with
Figure (7.21) Drift ratio in y-direction of the building with
Figure (7.22) Vertical differential settlement of the building with
varying raft thickness 217
Figure (7.23) Drift ratio in x-direction of the building with
Figure (7.24) Drift ratio in y-direction of the building with
model for pile length = 20 [m] 219
model with varying in pile length 220
model with varying in raft thickness 220
xxiii
LIST OF FIGURES
Figure (7.29) Moment in the raft at section I-I for rectangular-

shape model for pile diameter = 0.8 [m] 221
shape model for pile length = 20 [m] 221
shape model with varying in pile length 222
shape model with varying in raft thickness 222
length = 20 [m] 223
Figure (7.35) Pile load for rectangular-shape model with
varying in pile length 224
Figure (7.36) Pile load for rectangular-shape model with
varying in raft thickness 224
Figure (7.37) Pile load for square-shape model for pile
diameter= 0.8 [m] 225
Figure (7.38) Pile load for square-shape model for pile length
= 20 [m] 225
Figure (7.39) Pile load for square-shape model with varying in
pile length
226
Figure (7.40) Pile load for square-shape model with varying in

raft thickness 226
xxiv
LIST OF FIGURES
Figure (7.41) Lateral pile load in x-direction for rectangular-

shape model for pile diameter = 0.8 [m] 227
shape model for pile length = 20 [m] 227
shape model with varying in pile length 228
shape model with varying in raft thickness 228
Figure (7.45) Lateral pile load in x-direction for square-shape
model for pile length = 20 [m] 229
model with varying in pile length 230
model with varying in raft thickness 230
Figure (7.49) Bearing factor for pile diameter = 0.8 [m] 231
Figure (7.50) Bearing factor for pile length = 20 [m] 231
Figure (7.51) Bearing factor for varying in pile length 232

Figure (7.52) Bearing factor for varying in raft thickness 232
xxv
CHAPTER 1 INTRODUCTION
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 behavior 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 behavior of the raft which includes the
settlements, bending moments and the proportion of load carried by the raft,
and (ii) the behavior 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.
Analyzing piled raft is a complex task because it is a three-dimensional
problem including many capabilities. The main capabilities that must be
considered in the analysis are: the interaction between all piled raft and soil
elements; taking into account the actual loading and geometry of the piled
raft; representing the soil by a real model and treating the problem as
nonlinear analysis. Considering all these capabilities requires great
experience and effort. Besides such a problem requires long computational
time where huge size soil matrix is required for a large piled raft due to
discretized nodes along piles and under the raft. For these reasons many
1
authors suggested simplified methods in recent years to reduce the size of

analysis.
Clancy & Randolph (1993) and (1994) developed the hybrid layer method to
reduce the computing effort. Ta & Small (1997) approximated the surface
displacement of the soil by a polynomial instead of generating flexibility
factors, but the raft have to be square and of equal size. Russo (1998)
presented an approximate numerical method for the analysis of piled raft
where piles were modeled as interactive linear or non-linear springs. He used
the interaction factor method and a preliminary BEM to model pile to pile
interaction. Poulos (1999) described an approximate analysis for the
response of a pile group. The analysis uses a simplified form of boundary
element analysis to obtain single pile responses and interaction factors, and
employs various simplifying assumptions to facilitate the computational
process. Lee & Xiao (2001) presented a simplified analytical method for
nonlinear analysis of the behavior of pile groups using a hyperbolic
approach to describe the nonlinear relation between the shaft stress and
displacement. They developed the method for pile groups under both rigid
and flexible pile cap based on the load-transfer function. Kitiyodom &
Matsumoto (2002) and (2003) developed a simplified method of numerical
analysis of piled raft using a hybrid mode. Raft is modeled as thin plate, the
piles as elastic beams and the soil as springs. Mendona & Paiva (2003)
presented BEM/ FEM formulation for the analysis of piled raft in which
each pile is represented by a single element with three nodal points and the
shear force along the shaft is approximated by a quadratic function. The soil
is considered as half-space medium. Jeong et al. (2003) proposed a simple
algorithm to analyze laterally loaded three-dimensional pile groups using
beam column method. Liang & Chen (2004) presented a modified
2
variational approach for analyzing piled raft by a simplified analytical

solution to evaluate the pile-soil interaction. They applied the approach on
piled rigid and flexible rafts resting on homogeneous soil. Wong & Poulos
(2005) developed approximations for the settlement interaction factors
between dissimilar piles via an extensive parametric study. Lutz et al. (2006)
presented a simple method to estimate the load settlement behavior of piled
raft based on the theory of elasticity and solutions for calculation of ultimate
limit state. Most of the simplified analyses carried out by the methods
mentioned previously approximated the soil model. However, several
methods are available for analyzing this complex problem by a full three-
dimensional analysis but they are time consuming even for fast computers of
today.
In standard methods of analyzing piled raft based on elasticity theory, the
entire soil stiffness matrix of the piled raft is assembled due to all elements
of piles and raft. Then, settlements of piled raft elements are obtained
directly by solving the global equations. This paper presents more efficient
analysis of piled raft based on elasticity theory by using composed
coefficient technique to reduce the size of entire soil stiffness matrix. A full
interaction among piled raft elements is taken into account by generating the
entire flexibility matrix of the piled raft. The method may be also applied for
single pile or pile groups.
1.2. Objectives of the Thesis

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
3
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 and the raft. This new method has the
following features:
1) Both horizontal and vertical loadings can be applied to the piled raft.
2) The piled raft is analyzed 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) This model has the ability to save the time and save the memory of the
computer.
Effects of the dimensions of piles on different kinds of interaction and on the
overall behavior 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. Layout of Thesis

The thesis contains eight chapters as follows: -
4
Chapter one: A summary of objectives and a general scope of the present

study are introduced in this chapter.
Chapter two: includes a historical review of the available literature
concerning the problem under investigation.
Chapter three: states the used soil properties of Port-Said and the method
of deducing the modulus of soil elasticity and modulus of the soil
compressibility .
Chapter four: presents the numerical models used in the present study to
analyze the piled raft-superstructure. Finite Element models, Framework
Analogy method, Composed coefficient technique and Equivelant stiffness
technique which are used to simulate the different elements.
Chapter five: describes the designed algorithm to handle the proposed

method of analysis. Two computer programs are coded to analyze 3D-space
structures with piled raft resting on a nonlinear and non homogeneous soil
and using the Finite Element technique. Flow charts are clarified in details.
In addition, Results are verified by the available literature data.
Chapter six: a parametric study using the proposed method of analysis for
different construction models on Port-Said soil meduim is carried out. It
demonstrates the behavior of piled raft-structure resting on nonlinear elastic
medium and subjected to vertical loads using Composed coefficient
technique.
5
Chapter seven: a parametric study using the proposed method of analysis

for different construction models on Port-Said soil meduim is carried out. It
demonstrates the behavior of piled raft-structure resting on layered soil
medium and subjected to vertical and lateral loads simultaneously.
Nonlinearity of the soil is considered by Equivelant stiffness technique.
Chapter eight: a summary of the obtained results and conclusions drawn

from this investigation are stated. Also, suggestions for the future work are
introduced.
6
CHAPTER 2 LITERATURE REVIEW
CHAPTER 2
LITERATURE REVIEW
2.1. Introduction
Piled raft has attracted many researchers during the last four decades. Early
researches focused on hand calculation techniques with the help of empirical
charts and formulas for single piles and pile groups. With the advent of
computer and numerical procedures, Finite Element techniques were
developed to solve piled foundation. The different analysis methods 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 are presented. Also, a short brief
about seismic forces has been introduced.
2.2. Rafts
In the design of rafts, the soil may be treated as one of the following
methods.
a) A series of individual springs known as Winkler model.
b) A continuum model.
The Winkler model provides sufficient accuracy with reasonable time of
solution and storage requirements. The soil is represented by an infinite
number of isolated elastic springs. The deflection of the soil medium si at the
point i on the surface is directly proportional to the soil contact pressure qi at
that point regardless of the pressure at other locations. However, it neglects
the interaction between individual springs and therefore, the soil medium is
not modeled as a continuum soil. The other approach, which considers the
7
supporting soil as elastic continuum, can represent the soil more accurate
than the first one. The soil parameters used in this approach depend on the
field stress state as Hain and Lee (1974).
2.2.1. Analytical methods

Numerous researchers have investigated the use of analytical methods for
the analysis of rafts on elastic foundations. However, this approach is valid
for simple shapes and homogeneous soils. Zhemochkin and Sinitsyn (1962)
obtained the analytical solution by assuming that the contact pressures
between the raft and the 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 at a
number of points beneath the raft. A rigid circular raft subjected to
symmetrical loading was examined by Poulos (1994). Brown (1969) firstly
put forward the solutions for vertically loaded circular rafts resting on an
isotropic elastic foundation layer of finite depth. Subsequently, Brown
examined a circular raft resting on a deep foundation. The relative raft
stiffness was defined to reflect the raft-soil stiffness characteristics, and was
expressed as
E (1 2 ) t r
3
K= r ( 2.1)
Es a
where
Er denotes Young's modulus of the raft.
Es is the soil modulus.
vs is Poisson's ratio of the soil.
t, is the raft thickness.
a is the raft radius.
8
Hankel transforms were used to express the vertical surface displacement

and stress. The reaction force distributions and bending moment
distributions for uniformly loaded smooth circular rafts with any flexibility
were analyzed. He found that for relative raft stiffness K in the range 0.1-2,
as K increases, maximum bending moment increases rapidly and the
differential displacement reduces rapidly. Moreover, Kay and Cvagnaro
(1983) presented a method for the prediction of settlement for rafts by using
of field parameters in which the soil can have numerous sub-layers having
different properties.
2.2.2. Boundary Element methods

The boundary element method is a numerical computational method of
solving linear partial differential equations, which have been formulated as
integral equations (i.e. in boundary integral form). It can be applied in many
areas of engineering and science. The integral equation may be regarded as
an exact solution of the governing partial differential equation. The
boundary element method attempts to use the given boundary conditions to
fit boundary values into the integral equation, rather than values throughout
the space defined by a partial differential equation. Once this is done, in the
post-processing stage, the integral equation can then be used again to
calculate numerically the solution directly at any desired point in the interior
of the solution domain. The boundary element method is often more
efficient than other methods. However, for many problems boundary
element methods are significantly less efficient than volume-discretisation
methods (Finite element method, Finite difference method, Finite volume
method). BEM is used by Kastsikadelis and Armenkas (1984) by integral
equation method for analysing plates resting on Winkler soil medium. This
method is developed by Bezin(1988) into a new boundary element method
9
for bending of plates on a Winkler soil medium which employed an original

boundary integral equation method. Kirchhoff's theory was used to solve
flexural problems. In this method, the boundary is discretized into a number
of straight elements and the domain is divided into a number of rectangular
panels. The solution is obtained by Gauss-Legendre integration over the
elements on the boundary and the panels on the domain. A direct boundary
element method was developed which considers the interaction between an
elastic plate and an elastic half-space. The contact area was divided into a set
of contact elements or quadratic elements. The interface between the plate
and the soil is divided into triangular elements. 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 finite element method (FEM) is a numerical technique for finding
approximate solutions of partial differential equations as well as of integral
equations. The solution approach is based either on eliminating the
differential equation completely. The first solution which employed FEM for
the analysis of raft structures on an elastic half space was obtained by
Cheung and Zienkieicz (1965). In this technique, the raft is divided into a
number of rectangular elements joined at a discrete number of nodal points.
The soil is modeled either by Winkler model in which interactions between
springs are not considered or by elastic continuum model in which
separation between the raft and the soil is not allowed in case of negative
reactions. The stiffness matrix for whole system is formed by combining the
stiffness of soil with the stiffness of the plate bending elements. Contact
10
stresses were represented by equivalent force applied at nodal points of the

finite element mesh. El-Gendy (2007) developed an analysis of a thin elastic
plate on an elastic half space considering a nonlinear soil medium. El-
Kamash (2003) used Framework Analogy method in simulating a raft rested
on nonlinear Winkler soil. Svec and Gladwell (1973)developed an improved
method for the analysis of a thin plastic plated on elastic half space using
triangular elements. The contact pressure was represented by a cubic
polynomial and the displacements at the surface was determined from
Boussinesq equation. Wardle and Fraser (1974) developed an approach of a
multi-layered system with isotropic or cross-anisotropic properties. The
layered soil system was divided into a number of horizontal layers of
uniform thickness with infinite lateral extend. The loaded surface of the soil
mass was discretized into surface elements corresponding to the raft
elements, and the raft was modeled by conventional finite elements. 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).
Hain and Lee (1974) suggested that in the analysis of rafts, the structure,
rafts 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 was analysed by the
Substructure method developed by Przemieniecki. The supporting soil was
modeled by both Winkler model and the linear elastic model. Also, Three
Dimensional Finite Element method used by El Kamash for analyzing 3D-
space structure with the raft which is resting on nonlinear soil. Results have
revealed that there were significant differences in the behavior of the raft
predicted by the use of different soil models for the supporting soil.
Moreover, results have shown that the nonlinear soil analysis provided a
11
redistribution of contact pressures. Also, El Kamash employed the last

method for the analysis of rafts with different shapes resting on Winkler soil.
Sharma et al. used Finite Element method to analyze rafts of any shape
resting on an elastic half space. The raft was modeled by using eight nodded
isoperimetric 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 Boussinesq solution.
2.2.4. Hybrid approach

Zhang and Small (2000) presented a method for the analysis of soil-raft
interaction. This method employed the finite layer technique to determine
the behavior 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 the plan and subjected to
uniform, concentrated, or eccentric loads. The elastic soil can be anisotropic
or non-homogeneous. Rashed (2005) developed a new Boundary/Domain
element method for the analysis an elastic raft. Shear deformable plate
bending theory was used to model the raft. The developed formulation is
then used in analysing building raft foundations. For more practical
representation, the considered raft plate is treated as thick plate with free
edge boundary conditions. The soil or the elastic foundation is represented
as continuous media (follows the Winkler assumption). The boundary
element method is employed to model the raft plate; whereas the soil is
12
modelled using constant domain cells or elements. Therefore, in this

formulation both the domain and the boundary of the raft plate are
discretized. The associate soil domain integral is replaced by equivalent
boundary integrals along each cell contour. The main advantage of this
formulation is the ability of analysing rafts on non-homogenous soils.
2.2.5. Resisting of Rafts against lateral loads

Regarding to experimental and analytical studies pertaining to the lateral
resistance of pile cap, there is a rarity of published information available in
the subject area of pile cap lateral resistance. The results from studies show
that the lateral load resistance provided by pile caps can be very significant,
and that in some cases the cap resistance is as large as the resistance
provided by the piles themselves. Beatty tested two 6-pile groups of step-
tapered piles and determined that approximately 50 percent of the applied
lateral load was resisted by passive pressure on the pile cap. Kim and Singh
tested three 6-pile groups of piles and found that removal of soil beneath the
pile caps significantly increased the measured deflections, rotations, and
bending moments. This effect increased as the load increased.
Rollins et al. performed a lateral testing on a group of 9 piles and determined
the lateral load resistance of the pile cap was greater than the lateral
resistance provided by the piles themselves. The tests performed by Rollins
et al. considered only the passive resistance at the front of the cap, and only
dynamic loads. Kim and Singh considered only the soil in contact with the
bottom of the pile cap. The pile caps in Kim and Singhs study were
constructed on the ground surface, and thus results do not include any
passive resistance at the front of the cap or frictional resistance of soil along
the sides of the cap. The tests by Beatty only involved the passive resistance
at the front of the cap. The tests by Zafir and Vanderpool were performed on
13
an atypical pile cap, which consisted of a large, deep circular embedded cap.
These studies indicate that the lateral resistance of pile caps can be quite
significant, especially when the pile cap is embedded beneath the ground
surface. There is clearly a need for a rational method to evaluate the
magnitude of the pile cap resistance, and for including this resistance in the
design of pile groups to resist lateral loads.
2.3. Pile Group

2.3.1. Simplified analytical method
Randolph and Worth (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 lower layer in which the base of the upper
layer corresponds 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
modeled as shearing of concentric cylinders. For the lower layer, the base of
the pile was assumed to act as a rigid punch on the surface of the layer. This
approach was then extended to the analysis of pile groups by the same
authors by incorporating the interaction between loaded piles. The
interaction factors were computed using an iterative procedure. Lee (1993)
presented an approach which was modified from expressions used by
Randolph and Worth (1979). The expression for approximating the load-
settlement behavior of a pile was modified for a compressible pile embedded
in a soil with the stiffness increasing linearly with the 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. Guo and Randolph (1999)
14
developed a program GASGROUP that employed an exact closed form

solution for predicting the settlement of pile groups in nonhomogeneous soil
by using a Load Transfer approach. The stiffness of the soil was assumed to
increase with some power of the depth. The pile-soil interaction was
presented 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 Poisson's 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 consider the presence of neighboring 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 nonlinear
analysis of pile groups in multi-layered soils. The approach employed the
solution presented by Randolph and Worth to simulate the interaction
between two piles and hyperbolic function as the load transfer function to
model the nonlinear behavior between the shear stress and displacements of
the pile shaft along the pile-soil interface. The nonlinear displacement was
approximated as displacement developed at the distributed soil around the
pile shaft. The interaction effect between two identical piles was assumed to
be linearly elastic and the shaft and the 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 nonlinear 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 O'Neill et al. (1977) for the analysis of pile
groups. The response of the pile was modeled by the load transfer method
15
and the pile-soilpile interaction was determined based on Mindlin's

solution. In this approach, by using the 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 Mindlin's 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 Worth and the nonlinear behavior
of the soil was approximated by hyperbolic function. The interaction
between piles based on Mindlin's solution was computed by replacing the
continuous distributed loads at the pile shaft and pile base by point loads at
the nodes at the piles. The nonlinear response of an individual pile was
approximated by introducing a discontinuity in the soil. Since the
nonlinearity 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 to analyze laterally loaded pile groups. For
lateral loading, the soil response was modeled by the modulus of subgrade
reaction approach. The pile was divided into a number of discrete beam
elements and the soil was modeled by nonlinear 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 Mindlin's solution for a unit horizontal point load in a
16
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 (1980) presented the solution for a single pile by using
Mindlin's equations. Poulos 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 the
following:-
additional settlement due to an adjacent pile

= ( 2.2)
settlement of a pile under its own load
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 Fig. 2.1. The shaft of the pile was assumed to be perfectly rough while the
base was assumed to be perfectly smooth such that the shear stresses were
not developed on the base. The displacement factors were obtained by
integration of the Mindlin equation for vertical displacement due to a point
load within a semi-infinite soil mass. By considering the compatibility of the
vertical displacement The distribution of shear stress can be obtained 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 piles, the
displacement of an individual pile in the group was obtained by
superposition. Poulos further extended the method to laterally loaded pile
groups by introducing the interaction factor for rotation, which is defined
as following:-
17
additional rotation due to adjacent pile

= ( 2.3)
rotation of a pile under its own load
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 by replacing the vertical loads with
horizontal loads . The displacement of the pile can be expressed as
Butterfield and Banerjee (1971) who presented the analysis of
compressible piles and pile groups with a rigid floating pile cap based on
the integration of the Mindlins 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 by Butterfield and Banerjee (1971). The capsoil and pilesoil
interfaces were divided into discrete elements. The displacements of the
soil due to direct stresses acting on the interface s were obtained from
Mindlin's equations.
Chin et al. (1990) presented a method based on 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 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. while for the nodes on the same pile, the
18
displacement was obtained from the Gaussian quadrature method by

applying a uniformly distributed force over the element. Pile groups with
different radii, lengths, and moduli can be analyzed by this approach. Lee
and Poulos (1990) developed an approach for the analysis of pile groups in
non-homogeneous soil. This method 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 loadings and ground movements. The analysis was based
on the principal of the three-dimensional boundary element method. The
piles were assumed circular in cross-section and each pile was divided into a
series of cylindrical elements on the shaft and ring elements on the bases.
The cylindrical elements were then divided into several sub-elements. In the
analysis the cylindrical or annular boundary elements were transformed into
rectangular elements by mathematical transforms. The rectangular elements
were divided into a number of smaller rectangles and interaction between the
elements was obtained from the integration of Mindlin's equation. The
program GEPAN can be applied to a variety of pile problems.
Basile (2003) extended the approach proposed by Butterfield and Banerjee
(1971) and developed a computer program PGROUPN which was based on
a complete non-linear boundary element formulation. The program
PGROUPN accounts for the interactions between piles, group stiffening
effects, load deformation coupling and non-linearity of the soil. A Sub-
structuring 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 Mindlins functions. The piles were
19
modeled by simple beam-columns and Bernoulli-Eular beam theory was

applied to obtain the flexibility matrix. Slip was allowed in the computation
by limiting the stresses at the pile-soil interfaces and the pile base. The non-
linear soil behavior 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. Chow
(1987) 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
uncoupled. The piles were discretized into discrete elements and the soil was
discretized into constant elements at the pile-soil interface. The load
deformation relationship of the soil was determined using the flexibility
approach in which the soil flexibility coefficients were evaluated using the
Finite approach method. Non-homogeneity of the soil was taken into
account by incorporating continuously varying soil stiffness into the
numerical integration process during the formulation of the element stiffness
matrices. The soil nonlinearity was assumed elasto-plastic and was analyzed
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 analyze pile groups in cross-anisotropic
soils. Results have shown that the effect of soil anisotropy on small pile
groups embedded in homogeneous soil was small; however, the effect on
large pile groups in nonhomogeneous soils was significant.
20
2.3.5. Infinite Layer Method

Guo et al. (1987) presented an infinite 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.
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 a solid bar. Interaction
between the soil and the 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 nodded isoperimetric bending elements
and analysed by Mindlin thick plate bending theory. The connection between
the cap and piles was assumed a sliding ball joint such that only the vertical
forces were transmitted from the cap to the pile heads.
2.3.6. Finite Layer Method

The finite layer method developed by Small and Booker (1984) was first
introduced into the analysis of axially loaded piles embedded in isotropic
and cross-anisotropic layered soils by Lee and Small (1991). 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 nodded 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
21
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
from 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 (1991)
modified this approach for the analysis of laterally loaded piles by replacing
the vertical loads acting along the shaft and base with horizontal loads.
Southcott and Small (1996) extended the approach of Lee and Small (1991)
to the analysis of vertically loaded pile group but the loads applied to soil
nodes, which were assumed to be a series of uniform annular nodes rather
than ring loads.
Zhang and Small (1991) proposed two methods based on the finite layer
theory to analyze axially and laterally loaded pile groups embedded in
homogeneous and nonhomogeneous soils. The principle of the methods is
similar to that used by Lee and Small (1991). 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 was divided into
corresponding layers. Interaction and stiffness methods were developed to
generate the influence of 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 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
22
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 pile.
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 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 iterative procedure presented by Zhang and Small (2000). By considering
the equilibrium of interaction forces and compatibility of displacements
23
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

2.4.1. Approximation Method
One of approaches that treated the raft as a thin plate, the piles as springs and
the soil as an elastic continuum, was used by Hongladaromp et al. (1973) in
which the interaction effects between the piles were ignored. Poulos (1980)
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. Russo
(1998) employed a similar method where the piles and soil were modeled by
linear or non-linear interacting springs. The soil displacements were
calculated using Boussinesqs solution thus yielding a closed form solution.
The non-linear behavior of the piles was modeled by the assumption of a
hyperbolic load-settlement curve for a single pile.
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 modeled by two-dimensional thin plate finite
elements, the piles were modeled by one-dimensional rod finite elements
and the soil response was calculated by using an analytical solution. This
method took into account the non-linearity the behavior of the pile and the
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
24
modeled 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 (1971) employed the
boundary element method to study the behavior 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 the analysis.
Mindlins solution was used to describe the soil response and the interaction
effects. Kuwabara (1989) described a boundary element analysis based on
elastic theory to examine the behavior of a piled raft foundation in a
homogeneous elastic soil mass. In the analysis, the raft was assumed rigid
but the compressibility of piles was considered. 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
25
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 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 analyzed by dividing the interface into triangular elements
and the subgrade reaction was assumed to vary linearly across each element.
2.4.3. The 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. Hooper (1973), in which a symmetric model with eight
nodded isoparametric elements was used, gave an early example of the
analysis of a piled raft (the Hyde Park Barracks). In the analysis,
approximation of the equivalent stiffness of the pile group was made such
that each concentric row of piles was modeled by a stiffness that was
equivalent to the sum of the stiffnesses of individual piles. The soil was
assumed a linear elastic isotropic material with the modulus which is
increasing linearly with the 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
26
the contribution of the stiffening effect of the superstructure on the behavior

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. 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.
Liu and Novak (1971) employed the Finite Element method to examine the
behavior 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 nodded
isoperimetric 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
modeled by finite elements. A weak zone around the pile with lower strength
and the 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. 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 modeled 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 consider interaction, the
reciprocal theorem was applied to the pile, and influence factors were
calculated based on elastic theory. Non-linear behavior of piles and soil were
considered by comparing the reaction forces with the limiting values during
27
the iterative process. Katzenbach and Reul (1997) described a structural

model which employed the Finite Element method for the geometrical
modeling of the continuum, an elastoplastic constitutive model to describe
the soil behavior and a step-by-step analysis for numerical simulation. The
piles were modeled by 3-dimensional isoparametric Finite Elements and the
raft was modeled by shell elements. Katzenbach et al. (2000) used the same
structural model to carry out parametric studies on the behavior 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 half space.
Prakoso and Kulhawy (2001) analyzed 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 and six nodded triangular elements were used to model the
piled raft and the soil.
Reul and Randolph (2004) 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. The circular piles were modeled by square piles with
the same shaft circumference. The interfaces between the raft and soil, and
between the pile and the soil were modeled by thin solid continuum
elements and were assumed to be perfectly rough. The soil was modeled by
a cap model to simulate the non-linear behavior. 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. 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
28
Newton-Raphson method to solve the non-linear equations involved in a

plasticity analysis. The raft, pile and soil were discretized into eight nodded
brick elements. The raft and piles were assumed to be linearly elastic and the
non-linear behavior of the soil was modeled by the Drucker-Prager (1952)
yield criterion.
2.4.4. Combined Boundary Element and Finite Element Method

Hain and Lee (1978) combined the Finite Element method for the analysis of
the raft and the boundary element method for the analysis of piles and the
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. In the analysis, they suggested
that the behavior of the piled raft would depend on the relative flexibility of
the raft and the relative stiffness of the pile to the soil. 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 behavior of a
piled raft foundation on soft ground. Mandolini and Viggiani (1997)
presented an analysis to predict the settlement of piled raft foundations. The
method is capable of considering the soil-structure interaction and non-linear
behavior at the pile-soil interface. To stimulate the non-linear behavior, a
stepwise linear incremental procedure was used and a hyperbolic load
settlement relationship for a single pile was assumed.
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 analyzed as a plate resting
on an elastic soil medium and was discretised into four nodded rectangular
elements. The pile was discretised into cylindrical elements and analyzed by
29
the boundary element method, and the soil was assumed to be a

homogeneous elastic soil mass. Non-linear behavior 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.
2.4.5. Combined Finite Layer and Finite Element Method

Ta and Small presented an approach based on the Finite Layer technique
developed by Small and Booker to compute the behavior of piled rafts
subjected to vertical loads in layered soils. Ta (1996) proposed two
approximation methods (Type I and II) which may be used to compute
interactions between the piles or piles and the 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 consider 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.
2.4.6. Variational Approach

This approach makes use of the principle of minimum potential energy to
simulate the response of the foundation system. The same approach was
30
used by Shen and Teh (2000) which has incorporated lateral loading into the
analysis. The piles were modeled as flexible beams and the non-linear
behavior 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.
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. Interactions between the raft, piles and soil were described by an
approximate closed-form analytical method.
2.5. Lateral Load-Resisting Systems

2.5.1. Types of lateral loads
In general, there are two types of lateral loads acting on buildings; seismic
loads and wind loads. Only the first one shall be addressed in this study.
Seismic ground motion is the result of combination of direct and reflected
waves. It is consequently greatly affected by the geologic configuration of
the subsoil at a particular site. Although both wind and seismic loads are
dynamic in nature, a major difference is noticed in the manner at which these
loads are included in the structure.
2.5.2. Seismic loads

Seismic loads are defined as inertia loads that are a function of the mass of
structure. The Egyptian Building code (2008) mentioned two methods for
determining seismic loads. These methods are mentioned briefly in the
31
following paragraphs.
I) Dynamic analysis (time history analysis)

In this method, the structure is idealized as an assembly of masses
interconnected by springs and dash pots. It is generally used in buildings of
height more than 100 [ms] and having irregular geometry.
II) Simplified modal response spectrum method (equivalent static load)

This approach utilizes a set of statically applied horizontal forces to simulate
the effect of seismic loading. The distribution of these forces along the
building height is similar to that of the maximum forces obtained from the
dynamic analysis. The Egyptian Building code (2008) specifies the total
lateral seismic force acting on a building Fb according to the following
Equation:
Fb = Sd(T1). W / g (2.3)
Where, W is the weight of the structure used to calculate the seismic load.
Sd(T1) is the horizontal design spectrum for elastic analysis, is the
corrective factor and g is the earth acceleration and equals to 9.81 [m/s2].
The type of soil is an effective factor affects on Fb as well as other factors
like the type of the construction, the geographic zone according to the
earthquake map,
a The coefficient of importance 1

This coefficient is considered to provide a greater factor of safety against
damage of important facilities that must continue to work during and after
32
the earthquake such as hospitals, power plant, nuclear facilities, fire and
police stations. Values given in the Egyptian Building code are listed below:
- Essential facilities I=1.2~1.4
- All others I=.8~1.2
b The coefficient of force reduction R

Structures with different framing systems have different periods of response
and consequently different seismic loads. The different systems and their
corresponding coefficients are listed in Table 2.1.
Table 2.1. Reduction coefficient
Lateral load resistance system R Coefficient
Space frame resisting 100% of load

i) Ductile 7.0
ii) Non-ductile 5.0
Shear wall-frame building with frames 5.0~6.0
resisting at least 25% of load
Shear wall building 2.0~4.5
Other buildings like towers and chimney 3.0~3.5
c The period of buildings T1

The response of a structure to the earthquake loads depends upon the nature
of the structure. A structure which has a small period will have a higher
force coefficient and will be subjected to higher seismic loads. The Egyptian
Building code uses the following Equation to calculate the response
coefficient T1 till 60 [m]
33
T1 = Ct H 3 / 4 (2.4)
where, H is the total height of the loaded construction building and Ct is a

coefficient depends on the structural system and the material.
d The effect of soil coefficient

The influence of the soil under the building is very significant in
determining the attached force to the structure. The value of wave velocity
related the soil coefficient may be determined from Table (2.2)
Table 2.2. Soil coefficient factors

Vs,30 Types depth of soil layer
[m/s]
800< Rock ,dense soil greater than 15 meters in depth or medium dense
soil of less than 15 meters in depth overlying a stratum of better
properties
180~800 Medium and dense soil greater than 15 m in depth or loose soil of
depth less than 15 m overlying a stratum of better properties.
180< Loose soil greater than 15 meters in depth
2.5.3. Distribution of shear forces among stories

The horizontal force Fi for each floor mi can be carried out according the
following formula:-
34

zW
Fi = n .Fb
i
(2.5)

z jW j
j =1
Where, zi, zj are the height of mi, mj calculated from the level of foundation.
2.5.4. Types of resisting systems

Mainly, there are three systems that may be used to resist lateral loads that
are frames, walls and tubes. A frame is a structure that provides resistance to
lateral loading from the rigidity of members and connections as well. Shear
walls are introduced in high-rise buildings to stand against lateral loads
when the frame system is inadequate. Walls may be solid or with openings.
Connection between the wall sections is provided by either connecting
beams or floors slabs. Tubes may be in the shape of an internal core of shear
walls that behave in the same way as walls. However, externally perforated
tubes could be used which has a closely spaced exterior column system and
spandrel beams. This is called a framed tube system. The behavior of such
system gathers between those of the frame and the wall.
35
P P
S
Pj
Pb Pb
Pb Pb
Pile 1 Pile 2
(a) Stresses on piles (b) Stresses on soil
Fig.2.1 Stresses on piles and the soil for a two pile group under the
axial load

H, M d
d
S H, M
Fig.2.2 Stresses acting on piles and soil for a two pile group under
lateral loads
36
CHAPTER 3 PORT-SAID SOIL PROPERTIES AND PARAMETERS
Chapter 3
PORT-SAID SOIL PROPERTIES AND PARAMETERS
3.1. Introduction
Port-Said lies on the eastern side of the Nile Delta at the north end of the
Suez Canal on the Mediterranean Sea. Most dry land in Port-Said has been
reclaimed except for a narrow beach which separated Lake Manzala from
the sea. The material used in reclamation came from the Suez Canal and
harbor excogitations and raised the ground surface to its present height of
some 1 to 2 m above sea level. This thesis will focus on the study of piled
raft 3D-space buildings with different shapes exposed to both vertical loads
and seismic loads due to construction in Port-Said, taking into account the
soil data in the reclaimed zones as source data. This source data has been
based upon the extensive geotechnical study performed by Golder
Associates (1979).
This chapter will focus on available Port-Said soil properties in different

zones and how to get soil parameters that will be required for analyzing
single pile, raft, piled raft and pile group such as modules of compressibility.
Then, values are verified by comparing with ECP (2005) limits and compare
analytically settlement of selected buildings with the measured values.
3.2. Port-Said soil properties
3.2.1. Investigation area

The location of Port-Said zones, which were reclaimed, is shown in Fig.3.1.
Zone 1 contains Al Qabuti and Al Salam Regions, zone 2 contains PortSaid
37
water treatment station, few apartment buildings, agricultural and industrial

development, zones 3 and 4 contain El Zohor Region, zone 5 contains
industrial zone, zone 6 still unused.
3.2.2. Soil log

Fig.3.2 shows a five longitudinal sections in PortSaid reclaimed zones,
which were carried out by Golder Associates (1979) according to several
investigated points up to 60 [m] under the ground surface. These figures
show that the surface soil conditions are relatively uniform but in some
places the ground surface is underlain by fill. The whole areas have a thin
layer of very soft surface clay with an average thickness of 0.2 [m] in the
northern part of the zones to 2 [m] in the south. Below the surface clay there
is compact dense fine sand with an average thickness of about 7 [m], the
sand grades downward through a transition zone into firm clay. The clay
extends to an average depth of about 50 [m] below the ground surface, the
clay resting on basal deposits of very hard clay and dense sand.
38
Fig.3.1. Port-Said zones
Subsoil of reclaimed zones may be classified in 6 main layers as follows.
39
- Fill
- Surface clay
- Sand stratum
- Transition zone
- Lower clay
- Basal deposits
Table 3-1 shows the thickness of each main layer in the reclaimed zones.
The next paragraphs describe the soil layers in details according to EL Azab
(1988).
Table 3-1 Thickness of soil layers in the different zones in meter.

Layer Zone 1 Zone 2 Zone 3 Zone 4 Zone 5 Zone 6
Surface
1.2 1.9 1.0 1.0 1.9 1.7
clay
Sand
6.2 6.4 7.7 8.1 5.6 5.9
stratum
Transition
3.6 3.7 2.8 5.4 4 2.9
zone
Lower
36 29.5 36 32 40 42
clay
Surface clay
40
The surface clay is typically grey, very soft and highly plastic. Sand and silt
are present in varying proportions such that in some areas it could be
classified as sandy silty clay and fine sand. It has high silt content that when
dry it appears to be silt.
The surface clay contains occasional layers of fine sand ranging up to 0.20
[m] in thickness. Shell fragments embedded in clay are frequently
concentrated in brands up to 0.15 [m] thick. The sand is micaceous. In some
areas there are thin horizontally stratified sand and silt layers.
The thickness of the surface clay increases in a south direction. It is thinnest

in the north direction in zones 3 and 4 and thickest in the south and west in
zones 5 and 6.
Sand stratum
The sand stratum is under the surface clay. The thickness of the sand varies
between 4.2 [m] and 10 m with an average for whole site of 6.5 [m].
Transition zone
The transition zone is under the sand stratum, at an elevation of about 7 m.
thin layers of silt and clay occur in the sand stratum. This marks the start of
the gradational change from the sand stratum to the lower clay. The layers of
clay increases in frequency until the soil are predominantly clay with some
sand layers. They increase in thickness and frequency towards the base of
the stratum. These bands of sand and clay have been termed the transition
zone.
Lower clay
41
The lower clay is firm becoming stiff, dark grey, organic and micaceous. Its
structure characteristics are laminations and fissuring. The laminar structure
is more dominant at the top of the clay deposit than at the base. To about
elevation 22 [m], there are frequent thin bands and parting of sand and silt.
At greater depth there are occasional pockets of very finely laminated clay.
The bottom of the lower clay was reached at elevation ranging between 40
[m] to 52 [m].
42
CHAPTER 3
Zone 1 Zone 2 Zone 4 Zone 5 Zone 6

A Typically grey, very soft and highly plastic A A A A
A could be classified as sandy silty clay and fine sand A A A A
1.20 has a high silt content that when dry it appears to be a silt 1.90 1.00 1.90 t 1.70 5.00
Sand stratum
mS mS mS mS mS
10.00
7.40 8.30 9.10 7.50 7.60
The layers of clay increases in frequency until the soil is
U a clay with some sand layer U U U U 15.00
11.00 12.00 14.50 11.50 10.50
20.00
25.00
30.00
35.00
T
40.00
41.50
Firm becoming stiff, dark grey, organic and micaceous
T T
structure characteristics are laminations and fissuring 45.00
47.00 46.50
T
T 50.00
51.50
52.50
43
55.00
PORT-SAID SOIL PROPERTIES AND PARAMETERS
Fig.3.2. Main soil properties

3.3. Soil parameters
Modules of compressibility:
Analysis of foundation using continuum model required to obtain the
modules of compressibility of the clay Es or modulus of Elasticity E as a
main soil parameter. Considering the available water content for clay layers
from Golder Associates (1979), this modules can be presented in general
equations related to the depth.
The relationship between compression index and water content according to

Nishida (1956) may be expressed as:
Cc = 0.0054(2.6W 35) (3.1)

where:
W Water content, [%].
Cc Compression index, [-].
while the void ratio eo as a function in compression index Cc can be

expressed as:
Cc = 0.54(eo 0.35) (3.2)

where:
eo Void ratio, [-].
Now, The modulus of compressibility can be determined by equating

settlement obtained from both compression index Cc with void ratio eo and
that obtained from Es.
44
1 Cc +
H = log o H (3.3)
Es (1 + eo ) o
where:
Es Modulus of compressibility of Port-Said clay, [kN/m2].
Average vertical stress increase in clay, [kN/m2].
H Layer thickness, [m].
o Initial overburden pressure in a layer, [kN/m2].
Reda (2009) had determined the modulus of compressibility for each zone
and verified it. He had found that the difference in Es for all zones is small
and may be neglected. Values of Es for all zones in Port-Said are plotted in
the 0. Fitting values of Es is shown in the same figure and indicates that Es
increases with depth and can be approximated by the following linear
relation:
E s = E so (1+ 0.06 z ) (3.4)

where:
Es Modulus of compressibility, [MN/m2].
Eso Initial modulus of compressibility, Eso=2.7 [MN/m2].
z Depth measured from the clay surface, [m].
45
2
Modulus of compressibility E s [kN/m ]
0 2000 4000 6000 8000 10000
0
10
Depth under the surfacez [m]
20
30
Zone 1
Zone 2
40
Zone 3
Zone 4
50 Zone 5
Fitting
60
Fig.3.3. Modulus of compressibility with depth for all zones

Poissons ratio:
Poissons ratio of the soil is taken to be s=0.25 [-].
Modules of Elasticity:
Modulus of elasticity E [kN/m2] of the soil is given by:
1 s 2 s2
E= Es = 0.83Es (3.5)
1 s
Limited shaft friction:

To carry out the nonlinear analysis for piles, it is required to assess the limit
shaft friction of the pile ql. Russo (1998) suggested a limited shaft friction
ql=180 [kN/m2] meeting a value of undrained cohesion cu = 200 [kN/m2].
This value of ql may be used in the nonlinear analysis as the undrained
cohesion cu of Port-Said lower clay does not exceed 200.
46
Groundwater:
Groundwater in Port-Said lies in within 2 [m] from the ground surface. The
groundwater level is assumed to be lie directly below the raft.
Typical soil profile and soil properties:

Typical soil profile and soil properties of Port-Said used in the analysis in
this study are shown in Fig.3.4, Fig.3.5 and Table 3-2.
Table 3-2 Soil parameters

Type of soil Depth Modulus Undrained Poissons Unit
Layer under of cohesion ratio weight
No. ground Elasticity of the
surface soil
z cu s
[m] E [kN/m2] [-]
[kN/m2] s
[kN/m3]
1 Fill or 2.0 1750 - 18
surface clay 0.25
2 Sand 8.0 60000 - 8
3 Silt 12.0 6500 - 8
4 Clay 41.5 Eq. 3.4 200 8
47
2.0 A Fill or surface clay with

A typically grey, very soft and highly plastic
2.0 5.00
Sand stratum
mS Fine to medium sand
8.0 10.00
Transition zone
U Silt followed by clay increases in frequency until
A, Filling 12.0 the soil becomes a clay with some sand bockts 15.00
A
mS, Medium sand
U, Silt 20.00
T, Clay
25.00
30.00
35.00
Lower clay
T firm becoming stiff, dark grey, organic and micaceous 50.00
50.0
Fig.3.4. Main soil profile of Port-Said
48
B1
A
A
GW 2.0 2.0
5.00
Es = 72289[kN/m2]
mS Ws=72289 [kN/m2]
Gam=8[kN/m2]
8.00
Es = 7831[kN/m2]
U Ws=7831 [kN/m2] 10.00
Gam=8[kN/m2]
12.00
Es = 2866[kN/m2]
T Ws=2866 [kN/m2]
14.50 Gam=8[kN/m2]
15.00
A A, Filling
Es = 3339[kN/m2]
T Ws=3339 [kN/m2]
17.50 Gam=8[kN/m2]
mS, Medium sand
Es = 3810[kN/m2]
T Ws=3810 [kN/m2] 20.00
U, Silt 20.50 Gam=8[kN/m2]
Es = 4281[kN/m2]
T Ws=4281 [kN/m2]
T, Clay 23.50 Gam=8[kN/m2]
Es = 4753[kN/m2]
T Ws=4753 [kN/m2] 25.00
26.50 Gam=8[kN/m2]
Es = 5224[kN/m2]
T Ws=5224 [kN/m2]
29.50 Gam=8[kN/m2]
Es = 5695[kN/m2] 30.00
T Ws=5695 [kN/m2]
32.50 Gam=8[kN/m2]
Es = 6167[kN/m2]
T Ws=6167 [kN/m2]
Gam=8[kN/m2]
35.00
35.50
Es = 6639[kN/m2]
T Ws=6639 [kN/m2]
38.50 Gam=8[kN/m2]
40.00
Es = 7110[kN/m2]
Ws=7110 [kN/m2]
Gam=8[kN/m2]
T
50.00 50.00
Fig.3.5. Typical soil properties of Port-Said area.
49
CHAPTER 4 ANALYSIS OF PILED RAFT IN LAYERED SOIL
CHAPTER 4
ANALYSIS OF PILED RAFT 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 the 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 behavior 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 analyze 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 behavior 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)
50
to calculate the displacements 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 (Ottaviani, 1975; Valliappan, et al. 1974; Pressley and Poulos,
1986; Chow, 1987a) have been used for settlement analysis. 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. Moreover,
Steinbrenner employed an approximation technique to solve such problems
of finite layer by Finite element method based on Mindlins equations.
In this chapter, the analysis of raft and pile groups through the use of
Composed coefficients technique and Finite element methods will be
presented. Loadings and moments in different directions can be applied to
the foundation. The soil behavior is modeled by the use of Finite layer
techniques and the raft and piles are modeled by Finite element techniques.
4.2. Analysis of Rafts

Several analytical and numerical methods are available to analyze the plate
were introduced by many investigators. One of these numerical methods is
the grid framework analogy by Baz (1987). Using the lattice analogy, the
familiar concept of the frame analysis may be extended to two-dimensional
continua. Consequently, it becomes possible to solve various plate
configurations of different boundary conditions using this method and also
determination of stress and strain distributions in a continuous elastic
medium could be achieved. The behavior of plate element can be simulated
by an equivalent framework to become possible to a high degree of
accuracy. In this method, a basic requirement is that, the distribution of
stiffness throughout the analyzed structure should be represented as close as
51
possible to that of an equivalent frame. To achieve such demand, it is

necessary to create the equivalent framework model for each element in a
continuous medium.
In the grid framework model which had been introduced by Yettram (1965)
and Hosain (1983), a rectangular grid framework model applicable to the
analysis of plates with any Poissons ratio had been introduced. The model is
consisting of side beams and diagonal beams as shown in Fig. 4.1. The
cross-sectional properties of members are obtained by equating the rotations
of the grid nodes with those of a similar element, when both are subjected to
statically equivalent moments and torque. A rectangular grid model with five
cross sectional properties will define uniquely a rectangular element in a
plate. These properties are flexural and torsional rigidities of side beams and
the flexural rigidity of the diagonals.
The half-space soil is presented by a continuum model at discrete nodes. The
interaction between such elements is needed to achieve the continuum
model. Such interaction is on the bases of Boussinesq equation and Mindlin's
solution form by Finite element method. Steinbrenners technique is also
used for nonhomogeneous soil.
4.2.1. Flexure of a rectangular element of a plate

Fig. 4.2.a. illustrates the rectangular plate element. When a plate element is
subjected to a unidirectional moments M1, as illustrated in Fig. 4.2.c., the
rotation of the element is given by the following Equation.
52
k a M1
1 = 3
(4.1)
h
E( )
12
where, a and k a are the side lengths of the element as shown in Fig. 4.2.a., h
is its thickness, and E is the elastic modulus of the material. The rotation of
the element in the orthogonal direction is given as follows:
a M1
2 = 3
(4.2)
h
E( )
12
where, is Poissons ratio. Similarly, when the element is subjected to

unidirectional moment intensities M2 as shown in Fig. 4.2.e., the associated
rotations 3 and 4 about x and z directions respectively are given as follows:
a M2
3 = 3
(4.3)
h
E( )
12
and,
k a M2
4 = (4.4)
h3
E( )
12
In Fig. 4.2.g., the element which is subjected to self-equilibrating torque

intensities H along the edge will distort in such a manner that one corner
deflects with respect to the other three and all edges remain straight.
Therefore, the angle of twist 5 is:
k a H (1 + )
5 = (4.5)
h3
E( )
12
53
4.2.2. Flexure of the equivalent grid framework model

In the present work, the equivalent grid framework model is used to
represent flat slabs and the raft foundation. In this model, six beams are
arranged as illustrated in Fig. 4.2.b. The side beams of length a have the
same moment of inertia Is and the same torsional stiffness G Js. While, the
side beams of length ka have the moment of inertia Ic and the torsional
stiffness G Jc. The diagonals of length r a have the second moment of area Id
and no torsional stiffness.
When the model is subjected to a nodal moment about z-direction M1 /2, Fig.
4.2.d, the rotations in main and transverse directions are:
M1 a 2 K r 3Is + I d
6 = (4.6)
2E r 3Is Ic + I d Ic + K 3Id Is
and,
M1 a 2 K2 Id
7 = (4.7)
2E r Is I c + Id Ic + K 3Id Is
3
Similarly, when the model is subjected at each node to a moment K a M2 / 2

about the x-axis, Fig. 4.2.f., the rotations are:
M 2a 2 K r 3Ic + K 3Id
8 = (4.8)
2E r 3Is I c + Id Ic + K 3Id Is
and
M 2a 2 K 3 Id
9 = (4.9)
2E r Is I c + I d I c + K 3I d Is
3
In addition, if the grid model is suffering from a system of eight self

equilibrating torque at nodes as shown in Fig. 4.2.h, the corresponding
rotation is:
54
H r3 a2 K
10 = (4.10)
GJ
2E r 3 c + 2KI d
E
w4 = -a 10 (4.11)
where, w4 is the lateral deflection at any point.
4.2.3. Evaluation of the Cross-Sectional Properties of Members

In order to achieve the simulation between the plate element and the grid
framework model, the corresponding rotations of the two models should be
equaled, thus:
1 = 6 (4.12.a)
2 = 7 (4.12.b)
3 = 8 (4.12.c)
4 = 9 (4.12.d)
5 = 10 (4.12.e)
Substituting by values in the above Equations and solving, the inertia of

grid members are developed, thus:
(K 2 ) a h 3
IS = . (4.13)
2K (1 2 ) 12
(1 K 2) a h 3
IC = . (4.14)
2 (1 2 ) 12
r3 a h3
Id = . (4.15)
2K (1 2 ) 12
55
The torsional stiffness is obtained from Equation 4.10 and may be given by
the following Equations:
GJ c (1 3 ) a h 3
= . (4.16)
E 2 (1 2 ) 12
GJ s K(1 3 ) a h 3
= . (4.17)
E 2 (1 2 ) 12
4.2.4. Raft Subjected to Bending Action

Considering the frame element shown in Fig. 4.3., the nodal displacements
are given as follows.

{} = i (4.18)
j
Displacements vectors at i and j nodes are given as follows:
{} T ={ui, vi, wi, xi, yi, zi, uj, vj, wj, xj, yj, zj} (4.19)
By expressing the displacements at any point within the element as

polynomials in x, y and z with a number of constants equal to the number of
freedom per element. Therefore, the element i-j has six degrees of freedom
at each node which means, it has totally twelve constants.
u = a1 + a2 X (4.20.a)
v = a3 + a4 X + a5 X2 + a6 X3 (4.20.b)
w = a7 + a8 X + a9 X2 + a10 X3 (4.20.c)
56
x = a11 + a12 X (4.20.d)
w
y = = a8 + 2a9X + 3a10X2 (4.20.e)
x
v
z = = a4 + 2a5X + 3a6X2 (4.20.f)
x
The previous Equations can be put in the following form.
{} = [A] {} (4.21)
The axial strain may be given in terms of the displacement.
{} = u / x = a2 = (uj ui) /L (4.22.a)
The flexural strain (curvature) is obtained from Equations 4.20 (b) and (c) in
terms of the displacement, thus:
2 w/x 2
{} = 2
(4.22.b)
v/x
2
From Equations 4.21 and 4.22 (b) and (c), the strain {} can be formed in
terms of {} as follows.
{} = [C]{} (4.22.c)
By substituting for {} from Equation 4.22.c into the previous Equation;

thus:
57
{} = [B]{} (4.23)
where,
[B] = [C] [A]-1 (4.24)
The relationship between stresses and strains can be expressed as follows: -

{f} = [D] [] (4.25.a)
where, {f} is the stress resulted from the axial force. The moment M is
obtained from Equations 4.25(b) and (c)
{
M y = [D] 2 w/x 2 } (4.25.b)
M z = [D]{ 2 v/x 2 } (4.25.c)
where, D represents the modulus of elasticity in Equation 4.25.a., while it

represents EIy and EIz in Equation 4.25 (b) and (c) respectively. Therefore,
the stress may be given by the following Equation:
{} = [D]{} (4.25.d)
By equating the external and the internal work done in the frame element ,
thus:
L
{} {F} = (1/2){}T[B]T[D][B]{}x
T
(4.26)
0
Where, {F} is the nodal force vector which is given by the following
equation.
58
{F} = [K]{} (4.27)
Where, [K] is the local stiffness matrix.
L
[K] = (1/2) [B]T [D][B]dx (4.28)
0
The element stiffness matrix of the 3D-space frame [K] is always square and
symmetric.
After deriving the local stiffness matrix of the frame element, the
transformation matrix of forces and displacements is needed to obtain the
global stiffness matrix of the mentioned element. The member stiffness in
the local system of coordinates x, y and z is transferred to the global system
of coordinates X/, Y/ and Z/ using the transformation matrix. The nodal
forces in local coordinates can be expressed in terms of those in the global
coordinates as follows.
[F] = [T] [F\] (4.29)
where,
[F] and [F\] are column matrices representing the nodal forces in local and
global coordinates respectively and [T] is the transformation matrix and
given by following forms.
R 0 0 0
0 R 0 0
[T] = (4.30)
0 0 R 0

0 0 0 R
59
and
l x mx nx
[R ] = ly my n y (4.31)
l z mz n z
where, lx, mx, nx are direction cosines of the angles between x-axis and each
X\, Y\ and Z\ axes respectively as shown in Fig. 4.4. Ly, my, ny are direction
cosines of the angles between y axis and each X\, Y\ and Z\ axes respectively.
Lz, mz, nz are direction cosines of the angles between z-axis and each X\, Y\
and Z\ axes respectively. The orientation of the member x-axis should be
defined to find the remaining elements of [T] as lx, mx and nx (i.e. both
members y and z-axes rotate about the member x-axis to take up any
orientation) as shown in Fig. 4.5. These axes should be defined to proceed
the orientation of one of them.
If, however the member x axis lies along the global Z\ axis then, the
condition that the member y axis must lies in the global X\ and Y\ plane is
automatically satisfied and a further condition. For members not laying
along the global Z\ axis, the evaluation of the direction cosines of members y
and z axes is less straightforward, the problem is best tackled using vector
cross products. Therefore, the cross product C of two vectors A and B is
defined as follows.
A = i Xa + j Ya + k Za (4.32.a)
B = i Xb + j Yb + k Zb (4.32.b)
and
i j k
C = A x B = Xa Xb Xc (4.32.c)
Ya Yb Yc
60
where, C is a third vector normal to the plane of A and B and is directly

equivalent to |A| |B| |sin(A, B)| and sin(A, B) is the absolute value of the sine
of the angle between A and B. Also, i, j, k are unite vectors in the direction
of the x, y and z axes respectively. Moreover, Xa, Ya, Za are the magnitude of
the components vector A in directions of x, y and z axes respectively and Xb,
Yb, Zb are the magnitude of the components vector B in directions of x, y and
z axes respectively. As all of the direction cosines have been expressed in
terms of lx, mx and nx, the subscript x is no longer necessary and the
transformation matrix is affected by the following matrix.
T1 T4 T7
R = T2 T5 T8 (4.33)
T3 T6 T9
where, D = (1- n2 ), T1 = l, T2 = -(m cos + ln sin)/D, T3 = (m sin - ln

cos) /D, T4 = m, T5 = (l cos - mn sin)/D, T6 = -(l sin + mn cos)/D,
T7 = n, T8 = D sin , T9 = D cos.
Similarly the nodal displacements in local coordinates [] and those in
global coordinates [\] are related as the following form.
[] = [T] [\] (4.34)
Denoting the member stiffness in local coordinates by [K] and in global

coordinates by [K\]. It is noticed that [T] is orthogonal matrix; therefore, the
inverse of matrix is equaled to the transported matrix. thus, the following
form illustrate the relation between local and global systems for the stiffness
matrix.
61
[K\] = [T] T [K] [T] (4.35)
where, [K\] is the member stiffness matrix in the system of global

coordinates.
4.2.5. Rafts on half-space Soil

The half-space soil is presented by a continuum model at discrete nodes. The
interaction between elements is needed to achieve the continuum model.
Such interaction is on the bases of Boussinesq equation, Mindlin's solution
and the following form.
wi = fij Pj (4.36)
The service area of each element is a rectangle or a square area. For the
nodes i in Fig. 4.6, the flexible factor fii is determined using Bousinesque
concept as follows
qB
S e (corner ) = (1 2 )[I 5 ] (4.37)
2Es
Se is the elastic settlement beneath the corner of the rectangular area (B*L), q
is the contact pressure, Es is the modulus of elasticity of the soil and is
Poisson ratio of the soil. The factor m is a ratio between the length and the
width of the service area.
I5 in equation (4.37) can be written in the following manner.
1 1 + m 2 + m 1 + m 2 + 1
I5 = ln + m ln (4.38)
1 + m 2 m 1 + m 2 1

62
The flexible factor fij is determined approximately relating to Mindlin's form

for the elastic settlement due to a point load in half-space soil as follows.
1 8(1 ) 2
f ij = (4.39)
16G (1 ) r
where, is the Radial distance between points i and j and G is the Shear
modulus of the soil.
4.2.6. Modeling the layered Soil
The Mindlin's solutions which are employed in this work are derived for
homogeneous soil medium. In the fact, the practical problems are almost
nonhomogeneous soil. It leads to a need to promote this model to be able to
treat with the said kind of problems. Mindlins solution has been used to
obtain approximate solutions for a layer of finite thickness by employing the
Steinbrenners approximation (Steinbrenner, 1934) to allow for the effect of
the underlying rigid base in reducing the soil displacements (Poulos, 1989;
Poulos & Davis, 1980). For n layers under the i which are shown in Fig. 4.6,
the layer which contains i is denoted by "". The settlement in each layer k
due to a point load at j can be calculated from the following equations (4.40)
and (4.41).
fkj = f ( ztop
k
) f ( zBottom
k
) (4.40)
n
f + ( fij f ( zBottom ))
\
fij = kj (4.41)
k=1, where k
where, fij\ is the modified flexible factor for the case of layered soil.
63
4.2.7. Influence Matrix for the Soil

When the vertical 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 analyzed by the finite layer method. The
entries of the soil influence matrix [Is] are formed by applying a unit
uniform block of pressure in the z-directions to each rectangular element on
the soil in turn and calculating the soil displacements at the centers of all
elements at the interface of the raft and soil. The soil influence matrix can be
generated as follows:
- Apply a uniform pressure in the z-direction to the first element to obtain
the associated column of [Is].
- Repeat the above steps by applying uniform pressure to each element in
turn to construct the influence matrix [Is].
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 the above equation can be written as
wz1 I z11 I z12 . I z1n p z1

w I I z 22 . I z11 p z 2
z 2 = z 21
. . . . . .

wzn I zn1 I zn 2 . I znn p zn
(4.42)
The above equation can be written as
[s] = [Is] [ps] (4.43)
64
where s = ( z1,z2, , zn) is the vector of displacements in the z-

directions respectively.
ps = ( pz1, pz2, , pzn) is the force vector of pressure on the elements in the
z-directions.
4.2.8. Linear analysis of the raft

If the raft has nr nodes, the relation between the settlement of the raft and the
contact pressure force on the raft is expressed as Eq.4.43. The total
flexibility matrix in Eq.4.43 can be inverted to give the relation between
contact forces and nodal settlements as follows:
Pj = Kij wi (4.44)
where, ki,j is the stiffness coefficient of the soil stiffness. Accordingly, the
relationship between contact forces under raft nodes and settlements is
expressed as follows:
{w}=[Cr] {Q} (4.45)
where, [Cr] is nr*nr flexibility matrix of the raft, [Cr]=[kr]-1.

It is possible to treat the raft as an elastic plate on elastic medium soil. From
the Finite Element Analysis, the equilibrium of the raft is expressed as:
[kr] {} = {P}-{Q} (4.46)
where:
{P} 3*nr vector of applied loads and moments on the raft nodes.
[kr] 3nr*3nr piled plate stiffness matrix.
65
{} 3*nr deformation vector of the raft.
The forces on the raft nodes are given by:
{Q} = [kr][X]T {} (4.47)
where:
[X] 3*nr matrix xi, yi and zi are coordinates of node i.
{} 3 vector of translations wc and rotations tan y and tan x.
4.3. Analysis single pile and pile groups
To carry out the analysis, a composed coefficient or modulus ks [kN/m]

representing the linear soil stiffness of the pile is determined. The modulus
ks is a parameter used in both linear and nonlinear analysis of the pile. It is
defined as the ratio between the applied force on the pile head Ph [kN] and
the pile settlement wo [m]. The modulus ks is not a soil constant, it depends
on pile load, pile geometry and stratification of the soil. It is analogous to the
modulus of subgrade reaction of the foundation on Winklers soil medium,
which is the ratio between the average contact pressure and the settlement
under the characteristic point on the foundation. In this thesis, a method to
obtain the modulus ks from the rigid analysis of the pile is proposed, and
then a nonlinear analysis is presented.
4.3.1. Modeling single pile in half-space soil (vertical loads)

Estimating deformations and load distributions in a group of piles subjected
to typical loading conditions usually requires the use of computer-based
methods of analysis. Piles are modeled using the column element, while, the
soil is presented by elements attached to each pile at discrete nodes.
66
Elements which are distributed along the shaft present the shear strength of
the soil. The bearing resistance is presented by an element at the base of the
pile as shown in Fig. 4.6. The flexible factors for shear resistance and
bearing are fijsh and fijb respectively which may be determined according to
equations (4.48) and (4.49) for bearing will be determined approximately
related to Eqs. (4.49) and (4.51). The vertical displacement w of the node i
caused by the vertical load P at j for any depth beneath the surface, may be
expressed as follows:
w(i) = G (i,j) P(j) (4.48)
1 3 4 8 (1 ) 2 ( 3 4 ) ( Z i Z j )
2
G (i, j ) = + +
16 G (1 ) R 1 R2 R1
3

1 ( 3 4 )( Z i + Z j ) 2 2 Z i Z 6Z i Z j (Z i + Z j ) 2
+ j
+ (4.49)
16 G (1 ) R2
3
R2
5

For uniform distributed tractions t over the surface area of the shaft denoted
by S, the vertical displacement is obtained as follows:
w(i ) = t ( j )
S
G (i, j ) dS ( j ) (4.50)
The flexible factor fiish can be gotten subsequently from the above Eq. 4.50.
By the integration over the area of the pile base which has a radius ro, while
fiib can be gotten from the following form.
2 ro
w(i ) = q f ij r dr d (4.51)
0 0
The integrations for fiish and fiisb are illustrated in details in Annex1.
67
4.3.2. Modeling the layered soil

Like the raft on nonhomogeneous soil Mindlins solution has been used to
obtain approximate solutions for a layer of finite thickness by employing the
Steinbrenners approximation (Steinbrenner, 1934) to allow for the effect of
the underlying rigid base in reducing the soil displacements (Poulos, 1989;
Poulos & Davis, 1980). For n layers under the i which are shown in Fig. 4.6,
the layer which contains i is denoted by "". The settlement in each layer k
due to a point load at j can be calculated from the following equations (4.40)
and (4.41).
4.3.3. Soil flexibility for a single pile

In the analysis, the pile is divided into a number of shaft elements with m
nodes, each one acted upon by a uniform shear stress qsj [kN/m2] and a
circular base having a uniform stress qb [kN/m2] as shown in Fig. 4.7.a To
carry out, the analysis, pile shaft elements are represented by line elements
as indicated in Fig. 4.7.b All stresses acting on shaft elements and on the
base are replaced by a series of concentrated forces Qi acting on line nodes.
4.3.4. Linear analysis
Considering a typical node i as shown in Fig. 4.7.b, the settlement of the
soil wi adjacent to n nodes of the pile is expressed as:
n
wi = I i , j Q j (4.52)
j =1
Where:
Qj Contact force on node j, [kN]
68
Ii,j Flexibility coefficient of node i due to a unite force on node j, [m/kN]

If the pile is completely rigid, the pile will settle as a rigid body and the
settlement will be the same at all pile nodes. Therefore, the unknowns of the
problem are n contact forces Qj and the rigid body translation wo. The
derivation of the settlement for the rigid pile can be carried out by equating
the expanding Eq. (4.52) for all nodes yields to :
w1 = wo = I1,1Q1 + I 2,1Q2 + I31Q3 + ...+ I1,nQn

w2 = wo = I 2,1Q1 + I 2,2Q2 + I3,2Q3 + ...+ I 2,nQn

w3 = wo = I3,1Q1 + I3,2Q2 + I3,3Q3 + ...+ I3,nQn
(4.53)
........

wn = wo = I n,1Q1 + I n,2Q2 + I n,3Q3 + ...+ I n,nQn
Let the matrix [Is] represents the flexibility coefficients Ii,j and the matrix
[ks] represents the inverse of it, which its coefficients are ki,j. then, contact
forces can be written as a function in terms ki,j of the inverted matrix as
follows:
Q1 = k1,1wo+ k2,1wo+ k31wo + ...+ k1,n wo

Q2 = k2,1wo+ k2,2 wo+ k3,2 wo+ ...+ k2,n wo

Q3 = k3,1wo+ k3,2 wo+ k3,3wo+ ...+ k3,n wo
(4.54)
........

Qn = kn,1wo+ kn,2 wo+ kn,3wo+ ...+ kn,n wo
Carrying out the summation of all contact forces leads to:
n n n
Q
i =1
i = wo
i =1
k
j =1
i, j (4.55)
69
Equation (4.55) may be rewritten as:
Ph= ks wo (4.56)
In Eq. (4.56), the force on the pile head is the sum of all contact forces,
n
ph= Qi , while the composed coefficient is the sum of all stiffness matrix
i =1
n n
coefficients of the soil, ks = k i, j .
i =1 j =1
Equation (4.56) gives the linear relation between the applied load on the pile
head and the uniform settlement wo, which is analogous to Hooks low.
Therefore, the composed coefficient ks may be used to determine the soil
stiffness adjacent to the pile. In case of analysis of a single pile, it is easy to
determine the contact forces Qi. Substituting the value of wo from Eq. (4.56)
in Eq. (4.54) gives Eq. (4.57) in n unknown contact forces Qi as:
n
ph k i , j
j =1
Qi = (4.57)
ks
Equation (4.57) of contact forces on the rigid pile is found to be independent

on the Modulus of elasticity of the soil Es in case of isotropic elastic half-
space soil medium.
4.3.5. Nonlinear analysis

Nonlinear analysis is an important consideration since piles may be loaded
close to their full capacity, even under working condition. The nonlinear
relation between the load and settlement of pile is determined by considering
a hyperbolic relation between the load and the settlement, Fig. 4.10 shows a
typical nonlinear curve of load-settlement for a wide range of soils. The
70
curve can be approximated through a hyperbolic interpolation formula where

several equation forms are available to verify this curve.
Many methods were developed to study both pile-soil systems with
nonlinear response using a hyperbolic relation between the load and the
settlement. Fleming (1992) developed a method to analyze and predict load-
deformation behavior of a single pile using two hyperbolic functions
describing the shaft and the base performance individually under applied
load. Analyzing nonlinear behavior by hyperbolic function was used by
Mandolini & Viggiani (1997) for pile groups and was used by Russo (1998)
for piled raft. They considered piles as nonlinear interacting springs based
on the method of interaction factors. Basile (1999) assumed soil Youngs
modulus varies with the stress level at the pile-soil interface using a
hyperbolic stress-strain relationship.
Available nonlinear analysis of foundation on Winklers soil medium was
presented by Baz (1987) for a grid and Hasnien (1993) for a raft. El Gendy
(1999) extended this analysis to be applicable for a raft on continuum soil
medium. The composed coefficient technique enables to apply this analysis
on pile problems. The nonlinear behavior of the pile head force-settlement at
the piled raft-soil interface may be represented as:
wn
ph = (4.58)
1 wn
+
ks Ql
where:
wn Nonlinear settlement of the pile, [m]
Ql Limit pile load, [kN].
71
In Fig. 4.8 and Eq. 4.58, the initial tangent modulus for single pile is easily
obtained from linear analysis of the pile, which is equal to the modulus of
soil stiffness ks. The limit pile load Ql is a geometrical parameter of the
hyperbolic relation. In some cases the value of Ql is different from the actual
ultimate pile load. For a single pile, the force on the pile head ph is known.
Therefore, Eq. 4.58 gives the nonlinear settlement of the pile wn.
4.3.6. Modeling pile groups

The coefficient technique is first used to perform a linear analysis of pile
groups and piled raft. Then it is extended to include the nonlinearity effect.
The next paragraphs present the formulation of composed coefficients for
pile groups to generate a soil stiffness matrix of composed coefficients.
4.3.7. Soil stiffness for freestanding raft (pile groups under vertical
loads)
Deriving equations for piled raft foundation requires taking into account the
interaction effect among the pile groups. For doing that, the simple
freestanding raft on pile groups shown in Fig. 4.8 as an example is
considered, which having np = 4 piles and total nodes of n=23. The relation
between pile settlement and contact force on pile groups shown in Fig. 4.8
can be expressed in matrix form as:
72
I1,1 ... I1,6 I1,7 ... I1,12 I1,13 ... ... I1,17 I1,18 ... I1, 20 ... ... I1, 23
w1 ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... Q1
....
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ....
.... I 6,1 ....
... I 6,6 I 6,7 ... I 6,12 I 6,13 ... ... I 6,17 I 6,18 ... I 6, 20 ... ... I 6, 23
w6 1 Q
w I 7,1 ... I 7,6 I 7,7 ... I 7, 23 6 1
7 .... Q7
... ... ... ... ....
....
.... ... ... ... ...
....
w I ... ... ... ... I12,13 I12, 23 Q12 2
12 2 = 12,1
w13 I13,1 ... ... ... ... I13, 23 Q13
....
... ... ... ... ....
.... ....
w17 3
... ... ... ... Q
w ... ... ... ... 17 3
Q18
18 I18,1 ... ... ... ... I18, 23 ...
....

.... ... ... ... ...
...
w ... ... ... ... Q23 4
23 4
I ... ... ... ... I 23, 23
23,1
(4.59)
The total flexibility matrix in Eq. (4.59) can be inverted to give the
relationship between contact forces and nodal settlements as follows:
k1,1 ... k1,6 k1,7 ... k1,12 I1,13 ... ... k1,17 k1,18 ... k1,20 ... ... k1, 23
Q1 ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... w1
....
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ....

.... k 6,1 ... k 6,6 I 6,7 ... k 6,12 k 6,13 ... ... k 6,17 k 6,18 ... k 6, 20 ... ... k 6, 23 ....
Q6 1 w
Q k 7,1 ... k 7,6 k 7,7 ... k 7, 23 6 1
7 .... ... ... ... ...
w7
....
....
.... ... ... ... ...
....
Q k ... ... ... ... k12,13 k12, 23 w12 2
12 2 = 12,1
Q13 k13,1 ... ... ... ... k13, 23 w13
....
... ... ... ... ....
.... ....
Q17 3
... ... ... ... w
Q ... ... ... ... 17 3
w18
18 k18,1 ... ... ... ... k18, 23 ....
...

... ... ... ... ...
....
Q ... ... ... ... w23 4
23 4
k ... ... ... ... k 23, 23
23,1
73
(4.60)
where, ki,j is the stiffness coefficient of the soil stiffness matrix,[kN/m]. Due
to the high rigidity of the pile in its length direction, the settlement in every
pile itself is considered as a uniform. This assumption can establish the
relationship between the uniform pile settlement and the force on the pile
head in the pile groups. It can be done by equating all settlements in each
pile by a uniform settlement. Carrying out the summation of rows and
columns corresponding to the pile i in Eq. (4.60) leads to:
6 6 6
6 12 6 17 6 23 6
Q i k i , j k i, j k i, j k i, j wo1
i =1 1 i =1 j =1 i =1 j = 7 i =1 j =13 i =1 j =18
12 i =1 1
12 12 6 12 12 12 17 12 23

Q i k i , j k i, j k i, j k i , j wo 2

i17=7 2 = i = 7 j =1

i =7 j =7 i = 7 j =13 i = 7 j =18
17
i =7 2 (4.61)
17 6 17 12 17 17 17 23

Qi i k i, j k k k i , j wo 3

=13 j =1 i =13 j = 7
i, j
i =13 j =13
i, j
i =13 j =18 i =13 3
23 23 6 23
i =13 3
23 12 23 17 23 23
Q
k i, j k k k i , j wo 4

i i

i, j i, j
i =18 4
=18 4
i =18 j =1 i =18 j = 7 i =18 j =13 i =18 j =18
Accordingly, Eq. (4.61) can be rewritten for the simple pile groups in Fig.
4.8 in composed coefficients as:
ph1 K 1,1 K 1, 2 K 1,3 K 1, 4 wo1

ph K K 2, 2 K 2 ,3 K 2, 4 wo 2
2 2 ,1 (4.62)
=
ph3 K 3,1 K 3, 2 K 3, 3 K 3, 4 wo 3

ph4 K 4 ,1 K 4, 2 K 4 ,3 K 4, 4 wo 4
where:
woi Settlement in pile i, [m]
74
n2 m2
Ki,j Composed coefficient, [kN/m]. In general Ki,j= k
n = n1 m = m1
n,m
Phi Force on the head of pile i, which is equal to the summation of all
n2
contact forces in that pile, [kN]. In general Phi= Qn
n = n1
i 1 i j 1 j
n1= 1+ nn(l ), n2 =
l =1
nn(l ) , m1= 1+
l =1
nn(l ), m2 =
l =1
nn(l ) ,
l =1
nn(l) Number of nodes in pile l.
4.4. Modeling piled raft

For a complete analysis of piled raft foundation, pile-soil-raft and raft-soil-
raft interactions must be taken into account in addition to pile-soil-pile
interaction. To illustrate how to formulate the composed coefficient
technique for piled raft, the simple piled raft shown in Fig. 4.9 is considered,
which having np=4 and a total npr =33 contact nodes of raft and piles with
the soil. If the raft analyzed alone without piles, the number of its nodes will
be nr=14. In the analysis, the contact area is divided for the raft into
triangular and/or rectangular elements, while that for pile shafts into
cylindrical elements and that for pile bases into circular elements. The
contact pressure under the raft, on pile shafts or on pile bases is represented
by a series of contact forces on nodes. For the set of 33 nodes of the piled
raft, the relation between soil settlements and contact forces is expressed as:
75
w1 I 1,1 I 1, 2 I 1,3 ... I 1,8 I 1,9 I 1,10 I 1,11 ... I 1,16 ... I 1,33 Q
I 2 ,1 I 2,33 Q
1
w I 2, 2 I 2,3 ... I 2 ,8 I 2,9 I 2 ,10 I 2,11 ... I 2,16 ...
2 2
3 w I 3,1 I 3, 2 I 3, 3 ... I 3, 8 I 3, 9 I 3,10 I 3,11 ... I 3,16 ... I 3,33 Q3
... ... ... ... ... ... ... ... ... ... ... ...

... ...

w8 p1 I 8,1 I 8, 2 I 8, 3 ... I 8,8 I 8,9 I 8,10 I 8,11 ... I 8,16 ... I 8,33 Q8 p1
w I I 9, 2 I 9,3 ... I 9 ,8 I 9,9 I 9 ,10 I 9,11 ... I 9,16 ...

I 9,33 Q9
9 = 9,1
w10 I 10 ,1 I 10 , 2 I 10 ,3 ... I 10 ,8 I 10 ,9 I 10 ,10 I 10 ,11 ... I 10 ,16 ... I 10 ,33 Q10
w
11 I 11,1 I 11, 2 I 11,3 ... I 11,8 I 11,9 I 11,10 I 11,11 ... I 11,16 ... I 11,33 Q11

... ... ... ... ... ... ... ... ... ... ... ... ... ...
w
16 p 2 I 16 ,1 I 16 , 2 I 16 ,3 ... I 16 ,8 I 16 ,9 I 16 ,10 I 16 ,11 ... I 16 ,16 ... I 16 ,33 Q16 p 2
... ... ... ... ... ... ... ... ... ... ... ...

... ...

w33 I 33,1 I 33, 2 I 33,3 ... I 33,8 I 33,9 I 33,10 I 33,11 ... I 33,16 ... I 33,33 Q33
(4.63)
where p1, p2, .. are number of the piles.

The total flexibility matrix in Eq. (4.63) can be inverted to give the
relationship between contact forces and nodal settlements as follows:
Q1 k1,1 k1, 2 k1,3 ... k1,8 k1,9 k1,10 k1,11 ... k1,16 ... k1,33 w
k 2 ,1 k 2,33 w
1
Q k 2, 2 k 2,3 ... k 2 ,8 k 2,9 k 2,10 k 2 ,11 ... k 2,16 ...
2 2
Q3 k 3,1 k 3, 2 k 3, 3 ... k 3, 8 k 3, 9 k 3,10 k 3,11 ... k 3,16 ... k 3,33 w3
... ... ... ... ... ... ... ... ... ... ... ...

... ...

Q8 p1 k 8,1 k 8, 2 k 8,3 ... k 8,8 k 8,9 k 8,10 k 8,11 ... k 8,16 ... k 8,33 w8 p1
Q k k 9, 2 k 9,3 ... k 9 ,8 k 9 ,9 k 9,10 k 9,11 ... k 9 ,16 ...

k 9 ,33 w9
9 = 9 ,1
Q10 k10 ,1 k10 , 2 k10 ,3 ... k10 ,8 k10 ,9 k10 ,10 k10 ,11 ... k10 ,16 ... k10 ,33 w10

Q11 k11,1 k11, 2 k11,3 ... k11,8 k11,9 k11,10 k11,11 ... k11,16 ... k11,33 w11

... ... ... ... ... ... ... ... ... ... ... ... ... ...

Q16 p 2 k16 ,1 k16 , 2 k16 ,3 ... k16 ,8 k16 ,9 k16 ,10 k16 ,11 ... k16 ,16 ... k16 ,33 w16 p 2
... ... ... ... ... ... ... ... ... ... ... ...

... ...

Q33 k 33,1 k 33, 2 k 33,3 ... k 33,8 k 33,9 k 33,10 k 33,11 ... k 33,16 ... k 33,33 w33
(4.64)
As indicated before, equating settlements of all nodes on each pile by a

uniform settlement gives the composed coefficients with the force on the
pile head phi and its corresponding settlement woi as follows:
76
8 16

k1,1 k1,2 k1, j k1,9 k1,10 k 1, j k1,33
j =3
8
j =11

k k2,33
2,1
k2,2 k
j =3
1, j k2,9 k2,10 k2,11

8 8 8 8 8 8 8 16 8
Q1 k ki,33
k k k k k w1
Q i =3
i ,1
i =3
i,2
i =3 i =3
i, j
i =3
i ,9
i =3
i ,10
i =3 i =11
i, j
i =3
8 2 w2
8

Qi k9,1 k8,33
8
woi
i =3 p1
k9,2 k
j =3
1, j k9,9 k9,10 k8,11
i =3 p1
Q9 8 w9
Q = k10,1 k10,2 k 1, j k10,9 k10,10 k9,11 k9,33 w
10
j =3 16 10
16 16
ki,33 j woi
16 8 16 16 16 16 16
Qi ki,1 j k10,2 k k k k
i =11 p 2
i, j i ,9 j i ,10 j i, j
i =11 p 2 j =11 i =11 i =3 j =11 j =11 i =11i =11 j =11
... ...
Q w
33 33

8 16

k33,1 k33,2 k 33, jj k33,9 k33,10 k 33, jj k33,33
j =3 j =11
(4.65)
Accordingly, the total stiffness matrix of the piled raft of size (npr * npr) is
reduced to the size (nr * nr). It means that the composed coefficients
technique makes the size of the soil stiffness matrix of the piled raft problem
equivalent to that of the raft problem alone without piles. Now, Eq. (4.65)
can be rewritten in composed coefficients as:
77
Q1 K1,1 K1,2 K1, p1 K1,9 K1,10 K1, p2 ... K1,33 w1

Q K K2,2 K2, p1 K2,9 K2,10 K2, p2 ... K2,33 w2
2 2,1
ph1 K p1,1 K p1,2 K p1, p1 K p1,9 K p1,10 K p1, p2 ... K p1,33 wo1

Q9 = K9,1 K9,2 K9, p1 K9,9 K9,10 K9, p2 ... K9,33 w9
Q10 K10,1 K10,2 K10, p1 K10,9 K10,10 K10, p2 ... K10,33 w10

ph2 K p2,1 K p2,2 K p2, p1 K p2,9 K p2,10 K p2, p2 ... K p2,33 wo2
... ... ... ... ... ... ... ... ... ...

Q33 K33,1 K33,2 K33, p1 K33,9 K33,10 K33, p2 ... K33,33 w33
(4.66)
Where Kpi,pj, Ki,pj and Kpi,j are composed coefficients of the piled raft,
[kN/m].
Based on Eq. (4.66), the relationship between settlements and contact forces
of the piled raft can be written in general compacted matrix form as:
{Q}= [kb]{w} (4.67)

where:
{w} nr settlement vector.
{Q} nr contact force vector.
[kb] nr * nr soil stiffness matrix of the piled raft.
For simplicity of the formulation, the settlement on either raft node or pile
head is denoted by wi, while the contact force on either raft node or pile head
is denoted by Qi.
78
I) Analysis of piled flexible raft

In case of analyzing full flexible raft, the contact force vector {Q} on raft
nodes is known. Only settlements are required. The advantage of the
composed coefficient technique is that the composed soil stiffness matrix
can be inverted to get a composed flexibility matrix. Accordingly, a
relationship between contact forces under the flexible raft besides forces on
pile heads and nodal settlements is expressed as:
{w}=[Cpr] {Q} (4.68)
where, [Cpr] is npr*npr flexibility matrix of the piled raft, [Cpr]=[kpr]-1.
II) Analysis of piled rigid raft

For piled rigid raft, unknowns of the interaction problem are nr contact
forces Qi, the rigid body translation of the piled raft wc, and the rigid body
rotations x and y of the piled raft about axes of geometry centroid. These
are determined by considering nr compatibility equations of rigid piled raft
deflection and the displacement of subsoil at nr nodal points in addition to
the three equations of overall equilibrium.
The linear system of equations of the piled rigid raft may be expressed as:
{N }= [X ][kb][X ]T {} (4.69)
where:
{} 3 vector of translation wc and rotations tan y and tan x
79
[X] 3*nr matrix of {1, xi, yi }. xi, yi are coordinates of node i.

{N} 3 vector of resultant and moments of applied loads acting on the piled
raft.
while forces on pile heads and raft is given by:
{Q}= [kb][X ]T {} (4.70)
III) Analysis of piled elastic raft

It is possible to treat the raft as an elastic plate on rigid piles. From the finite
element analysis of the plate, the equilibrium of the raft is expressed as:
[kr ] {}= {P} - {Q} (4.71)
where:
{p} 3*nr vector of applied loads and moments on the raft nodes.
[kr] 3 nr*3 nr plate stiffness matrix.
{} 3*nr deformation vector of the raft.
In the case of analyzing an elastic raft on pile groups, the elastic shortening
of the pile may be added to the pile settlement in Eq. (4.70). The elastic
shortening of the pile i is expressed as:
Phi l i
i = (4.72)
Epi Api
where:
i Elastic shortening of pile i, [m].
li Length of pile i, [m].
80
Api Cross-section area of pile i, [m2].

Epi Modulus of elasticity of the material of pile i, [kN/m2].
Equation (4.72) is written for the entire piled raft in matrix form as:
{wp}= [Cp ]{Ph} (4.73)
where:
{wp} Elastic shortening vector.
[Cp] Elastic pile matrix, which is a diagonal matrix.
{Ph} Vector of forces on pile heads.
To take the effect of pile shortening into account, the elastic coefficient of
the pile i in the matrix [Cp] is added to the flexibility coefficient of that pile
in the matrix [Cb] in Eq. (4.72) as follows:
{w}= [ [Cb]+ [Cp ] ]{Q} (4.74)
Inverting the total flexibility matrix [[Cb]+[Cp]] gives also the total stiffness
matrix of piled raft [kp] with the effect of pile stiffness due to its elastic
material.
{Q}= [kp]{w} (4.75)
where [kp] is nr*nr stiffness matrix of the piled raft with the effect of pile
elastic material, [kp]=[[Cb]+[Cp]]-1.
Considering compatibility between soil settlement and piled raft

81
displacement and substituting Eq. (4.75) in Eq. (4.71) gives the following
linear system of equations of the piled elastic raft:
[ [kp]+ [kr ] ]{}= {P} (4.76)
Once settlements on piles woi are determined in the above three cases of
piled rafts, the individual forces along the pile shaft and on the pile base can
be obtained from Eq. (4.64).

The nonlinear analysis of the piled raft is also based on the hyperbolic
relation presented in section 4.3.5. The initial tangent modulus of the
hyperbolic relation may be obtained from the linear analysis of the piled raft
as:
Phio
ksi = (4.77)
woio
where:
Phio Force on the pile head obtained from the linear analysis, [kN].
woio Pile settlement obtained from the linear analysis, [m].
4.2.3. Iterative Procedure

An iteration method is presented to solve the system of nonlinear equations
of the piled raft. The main idea of this method is that the full stiffness matrix
[kb] for rigid raft or [kp] for elastic raft is converted to a diagonal stiffness
matrix [ke]. Stiffness coefficients of this matrix, which represent nodal raft
stiffness and pile stiffness coefficients, are determined from the contact force
82
and its corresponding settlement. Only the pile stiffness is modified at each
cycle from the iteration process. Using the equivalent diagonal matrix,
equations of the piled raft are solved for each iteration cycle until the
compatibility between raft, piles and soil is achieved.
Figure 5 shows the iteration cycle and the flow chart of the iteration process.
The iteration process can be described in the following steps:
1 Carry out the linear analysis of the piled raft by solving system of
linear Eqns (4.69) or (4.76) whichever is applicable, to get the
settlements {w}.
2 Find the nodal contact forces {Q} due to the computed settlements
from Eq. (4.70) for rigid raft and from Eq. (4.75) for elastic raft.
3 From the computed settlements and contact forces, determine the
nodal stiffness at all nodes on the raft and on pile heads from:
Qi
kei = (4.78)
wi
4 Modify the pile stiffness by:
1
kei = (4.79)
1 w
+ i
ksi Qu i
5 Convert the full soil stiffness matrix (matrix [kb] or matrix [kp]) to
equivalent diagonal stiffness matrix [ke]. This matrix can be generated
from nodal raft stiffness computed in step 3 and pile stiffness
83
computed in step 4.
6 Replace the full matrix by diagonal matrix [ke]. Then, carry out the
nonlinear analysis of the piled raft to get the settlements {w}.
7 Compute the contact force under the raft and force on pile head by:
Qi = kei wi (4.80)
8 Compare the settlement from cycle i with that of cycle i-1 to find the
accuracy of the solution.
The steps 3 to 8 are repeated until the accuracy reaches a specified tolerance
, which means that a sufficient compatibility between settlements of piles,
raft and soil are achieved at the piles-raft-soil interface. However, in this
analysis the nonlinear response is applied on the raft and piles as indicated
by El Gendy (1999) to the piled raft system.
4.5. The pile model against lateral loads

Estimating lateral displacements of piles subjected to horizontal loads
usually requires the use of computer-based methods of analysis based on an
elastic continuum model of the soil as described by Douglas, Banerjee,
Davis and Poulos (1980). Piles are modeled to be thin rectangular vertical
strips of width d, length Lp, and constant flexibility EpIp. In case of using
circular pile, d can be taken as the pile diameter. Possible horizontal shear
stress developed between the soil and the sides of the pile are not taken into
account. The soil is presented by a group of elements attached to each pile at
discrete nodes. Elements which are distributed along the shaft present the
passive pressure of the soil as shown in Fig. 4.11. The flexibility coefficients
for horizontal resistance are fhii and fhij. The flexibility coefficient fhij may be
84
determined according to Eqns. 4.81 and 4.82, while the other one fhii is
gotten by integrating Eq. 4.82 over the surface Sp of the pile exposed to the
passive pressure of the soil (d*lp), where lp is the length of the pile element
in Eq. 4.81.
w p (i ) = t p ( j )
Sp
G (i, j ) dS p ( j ) (4.81)

1 3 4 1 x2 x 2 ( 3 4 )
G (i , j ) = + + +
16 G (1 )

x2 + y2 x2 + y2 ( x2 + y2 ) (
3
)
x 2 + y 2
3

1 4 (1 )( 1 2 ) 4 (1 )( 1 2 ) x 2 (4.82)
+
16 G (1 )

x2 + y2 x2 + y2
3
( )

4.5.1. Modeling of the layered soil

Mindlin's solutions on which the present work is based on are derived for
homogeneous soil mediums. In fact, most of practical problems usually deal
with nonhomogeneous soil. Therefore, a considerable demand to develop
Mindlin's model to be applicable for nonhomogeneous soil layers is strongly
recommended. The averaging technique suggested by Poulos (1979) is
incorporated into the analysis to approximate the interaction between the
structures' members of a piled raft foundation embedded in non-
homogeneous soils against lateral loads. As shown in Fig. 4.6, averaged
Youngs modulus, Es(ij); used in the approximation for pilesoilpile, raft
soilpile and raftsoilraft interactions can be calculated using Equations
(4.83)(4.85), respectively.
( Es (i ) + Es ( i +1) ) + ( Es ( j ) + Es ( j +1) )
Es (ij ) = (4.83)
4
85
2 Es (1) + ( Es ( j ) + Es ( j +1) )
Es (1 j ) = (4.84)
4
Es (11) = Es (11) (4.85)
where Es(ij)
is the averaged Youngs modulus of the soil at the depths Zi and
Zj; Es(i) and Es(j) are the Youngs modulus for soil layer numbers i and j.
4.5.2. Soil stiffness for pile groups against lateral forces

When the horizontal load is applied to the piles, the load is transferred from
piles to the soil through the contact stresses. The soil may consist of several
different material layers and can be analyzed by the finite layer method. The
entries of the soil influence matrix [Is] are formed by applying a unit
uniform block of pressure in the x- and y-directions to each rectangular
element on the soil in turn and calculating the soil displacements at the
centers of all elements at the interface of the piles and the soil. The soil
influence matrix can be generated as follows:
- Apply a uniform pressure in the x- and y-direction to the first element to
obtain the associated column of [Is].
- Repeat the above steps by applying uniform pressure to each element in
turn to construct the influence matrix [Is].
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 the above equation can be written as
86
wx1 I x11 I x12 . I x1n p x1

w I I x 22 . I x11 p x 2
x 2 = x 21
. . . . . . (4.86)

wxn I xn1 I xn 2 . I xnn p xn
and
wy1 I y11 I y12 . I y1n p y1

w I I y 22 . I y11 p y 2
y 2 = y 21
. . . . . . (4.87)

wyn I yn1 I yn 2 . I ynn p yn
The above equations can be written as
[s] = [Is] [ps] (4.88)
where s = ( x1,x2, , xn, y1, y2,, yn) is the vector of

displacements in the x- and y-directions respectively.
ps = ( px1, px2, , pxn,py1,py2,,pyn) is the force vector of pressure on the
elements in x- and y-directions.

From the finite element analysis of the piles, the equilibrium of the piles
against lateral loads is expressed as:
[kph ] {h }= {Ph } - {Qh } (4.89)
where:
{ph} 4*np vector of applied loads and moments on the piles nodes in x-
87
and y-directions.
[kph] 4 np*4 np pile stiffness matrix.
{h} 4*np deformation vector of the pile in x- and y-directions.
To express the displacement due to the elasticity of piles Eq. (4.90) can be
written as follows.
{wph}= [Cp h ]{Ph h } (4.90)
where:
{wph} Elastic deformation vector.
[Cph] Elastic pile matrix.
{Phh} Vector of forces on pile heads.
To take the effect of pile deformation due to the elasticity into account, the
elastic coefficient of the pile i in the matrix [Cph] is added to the flexibility
coefficient of that pile in the matrix [Cbh] in Eq. (4.90) as follows:
{wh }= [ [Cb h ]+ [Cp h ] ]{Q h } (4.91)
Inverting the total flexibility matrix [[Cbh]+[Cph]] gives also the total
stiffness matrix of piled raft [kph] with the effect of pile stiffness due to its
elastic material.
{Q h }= [kph ]{wh } (4.91)
where [kph] is 2np*2np stiffness matrix of the piled raft with the effect of pile
88
elastic material, [kph]=[[Cbh]+[Cph]]-1.
Considering compatibility between soil settlement and piles displacements

and substituting Eq. (4.91) in Eq. (4.87) gives the following linear system of
equations of the elastic piles against lateral forces:
[ kph ]{ h }= {Ph } (4.92)
Once displacements on piles whi are determined in the above three cases of
piles, the individual forces along the pile shaft can be obtained from Eq.
(4.93).
Phj = Khij whi (4.93)
where khi,j is the stiffness coefficient of the soil stiffness.

The nonlinear analysis of the piles against lateral forces is based on the
hyperbolic relation as Eq. (4.58). The initial tangent modulus of the
hyperbolic relation may be obtained from the linear analysis of the whole
system considering the limitations of stresses as clarified in the next section.
4.5.5. Limiting pile-soil stresses

It is essential to ensure that the stress state at the pile-soil interface does not
violate the yield criteria. This can be achieved by specifying the limiting
stresses for the soil.
89
4.5.5.1. Cohesive soil

For cohesive soils, a total stress approach is adopted. The limiting bearing
stress on the pile shaft for the lateral response is calculated as:
tss = Nc Cu (4.95)
where Cu is the undrained shear strength of the soil and Nc is a bearing

capacity factor increasing linearly from 2 at the surface to a constant value
of 9 at a depth of three pile diameters and below, much as was originally
suggested by Broms (1964) and Fleming et al.(1992).
4.5.5.2. Cohesionless soil

For cohesionless soils, an effective stress approach is adopted. The limiting
bearing stress on the pile shaft for the lateral response is calculated as
Fleming et al.(1992):
tsc = Kp2 'v (4.96)
where is the effective vertical stress and Kp is the passive earth pressure
coefficient, equal to (1+sin) / (1-sin), where is the friction angle of the
soil.
4.6. Soil-structure interaction

Soil-structure interaction refers to the interaction between a superstructure,
its foundation and the subsoil. As the term soil-structure interaction covers a
broad field, the work presented in this chapter will be focused to the analysis
of pile groups or piled raft foundations with the superstructure under static
load. Traditionally, the design work for structures has been separated from
90
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. In the next sections, the
interaction mechanism for piled raft against vertical and lateral loads is
presented.
4.6.1. 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 that the maximum
moment in the foundation decreases with increasing flexibility of the
foundation. Fraser and Wardle (1975) used the finite element method to
analyze a two bay portal frame on a layered cross-anisotropic elastic
continuum in which the frame was modeled 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
91
displacements in the raft were dependent on the relative stiffness of the

structure. Zhao and Cao (1985) used the substructure method to analyze a
twelve storey frame structure with two basements founded on soft clay. The
structure, raft and the soil were analyzed 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 behavior 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).
Zhang and Small (1994) employed the finite layer technique to analyze
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
92
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 behavior 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.
4.6.2. Piled raft-3D superstructure interaction

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.
Interaction between piles group and the superstructure is an important
consideration in the analysis of pile groups and piled rafts. In this analysis,
the superstructure including all structural elements as well as the
substructure including the piled raft is analyzed simultaneously. Slabs and
the raft are simulated as plates using the Framework Analogy, while beams,
columns and piles are presented by beam elements. The interaction between
different elements is created to achieve raft-soil-raft, pile-soil-pile and pile-
soil-raft interactions as shown in Fig. 4.6. If the raft has number of nodes nr
and n piles have np nodes with total number of nodes npp and the
superstructure has total number of nodes nst. The relation between the
settlement of pile and contact pressure force on pile groups is expressed as
Eqns. (4.71) to (4.76) and Eqns. (4.89) to (4.93), the relation between the
piled raft against vertical loads and the soil in the 3D-space structure can be
expressed as follows.
93
[kprv] {v} = {Pv}-{Qv} (4.97.a)
where:
{Pv} 3*nst vector of applied vertical loads and moments on the 3D-space
structure nodes.
[kprv] npr*npr piled plate stiffness matrix.
{v} 3*npr deformation vector of the piled raft.
The forces on piled raft nodes are given by:
{Qv} = [kprv][Xv]T {v} (4.97.b)
where:
[Xv] 3*npr matrix xi, yi and zi are coordinates of node i.
{v} 3*npr vector of translations wc and rotations tan y and tan x.
while, the relation between the piles group against horizontal loads and the
soil in the 3D-space structure can be expressed as follows:
[kph] {v} = {Ph}-{Qh} (4.98.a)
where:
{Ph} 3*nst vector of applied lateral loads and torsion on the 3D-space
structure nodes.
[kph] 2npp*2npp piles stiffness matrix.
{h} 3*npp deformation vector of the pile group.
The forces on piles nodes are given by:
94
{Qh} = [kph][Xh]T {h} (4.98.b)
where:
[Xh] 3*npp matrix xi, yi and zi are coordinates of node i.
{h} 3*npp vector of translations x, y and the rotations tan z.
The linear system of equations of the whole superstructure and the piled raft
stiffness due to the elasticity of system can be expressed as:
{P} = [X] [kst][X]T {} (4.99.a)
Considering the compatibility between the soil settlement and the piled raft
displacement, the following equations of 3D- space structure with elastic
piled raft against vertical and horizontal loads is given as follows.
[[kpr] + [kst]+ [kph]]{} = {P} (4.99.b)
where:
{P} 6*nst vector of applied vertical and horizontal loads, moments and
the torsion on the 3D-space structure nodes.
{} 6*nst deformation vector of the whole system in 3D-space.
95
diagonal beam
P1
side beam nodal point
P2
Z
Y
framework cell
Fig.4.1 Framework representation of plate (Lattice of plate)
K.a K.a
3 4
3 Ic 4
Is Id Is
Id
a
a
r.a
1 2 Z
1 2
(a) (b)
X
1 2 6 7
aM1/2 aM1/2
M1 M1 M1
aM1/2 aM1/2
(c) (d)
M2
3
kaM2/2
8 KaM2/2
KaM2/2 KaM2/2
M2 4 9
(e) (f)
H KaH/2 KaH/2
H aH/2
aH/2
5 10
H
aH/2 aH/2
KaH/2 KaH/2
H (g) (h)
Fig.4.2 Plate element and equivalent Framework model

96
y y 2 8
4 10
5 1
11
x x 7
z i z j i4 j
3
6 9 12
Fig.4.3 Member axes and freedoms for a 3D-space frame Element
Y\
member
j x axis
yj - yi
i X\
xj - xi zj - zi
Z\
Fig.4.4 Calculation of the direction cosines for the member x Axis
Y\ y

Z\
z
Fig.4.5 Principal axes
97
y
z
n Raft-soil-raft
interaction
x
Layer 1 j i
Pile-soil-raft
l interaction
k
Layer 2
Pile-soil-pile
Layer n interaction
Fig.4.6 3D-Space framework analogy of raft-pile-soil modeling

ph
Ph
a Ground surface
a
1
1 qs1 Q1 1
1 2
zj-1 2 Q2 2
zj 3
2
i
j-1
4
j qsj lj
j (lj-1+lj)/2
j
Qj j
b 2ro 6
n
qb
Qn
a) Pile shaft elements b) Pile line elements
Fig.4.7 Pile geometry and elements
98
P2 P1
p
Ground surface
1 7 13 18
2
8 14 19
3 20
9 Q9 15
pile 1 pile np=4
4 21
10 16
5
17 22
11
6
23
Q6 12
a) Free-standing raft on piles y
N
ex
Ground surface
Ph1 y Phnp
b) Equivalent statical system
pile 1 pile np
C.L.
Ground surface
wo1 wonp
2y wc
c) Soil settlement Pile displacement
Fig.4.8 Modeling freestanding raft on pile groups
99
P2 Ground surface
P1
p
1 2 9 10 17 18 24 25 32 npr = 33
3 11 19 26
Q1 Q32
4
12 20 27
5 28
13 Q13 21
6 29
14 22
7
23 30
15
8
31
Q8 16
pile 4
pile 1
a) Piled raft with node numbering of piled raft
P2 Ground surface
P1
p
1 2 3 4 5 6 7 8 9 10 11 12 13 nr = 14
Q1 Q32
Ph1 Ph4
pile 1 pile 4
b) Equivalent statical system with node numbering of raft alone
Ground surface
w1 w32
wo1 wo4
c) Soil settlement
Fig.4.9 Modeling piled raft
100
Linear analysis
Limit pile load Ql [kN]

Pile load Ph [kN]
Nonlinear analysis
ks
Nonlinear settlement wn [m]
Fig.4.10 Load-settlement curve of a single pile (hyperbolic relation)
Pile-soil-pile
Py1 interaction in x-
Px1 direction
1
Pyn+2
Pxn+2
2
Py3
lp Px3
3
Pyn+4 Pile-soil-pile
Pxn+4 interaction in y-
Pyn direction
Pxn
a) Forces acting on soil interface b) Forces acting on the pile group
Fig.4.11 Interface lateral forces of pile group in layered soil
101
CHAPTER 5 COMPUTER PROGRAM AND VALIDITY OF PROCEDURE
CHAPTER 5
COMPUTER PROGRAM AND VALIDITY OF PROCEDURE
5.1. Introduction
In this chapter a designed algorithm is introduced to handle the proposed
analysis. A computer program is edited to study the piled raft behavior
against own weight, live loads and seismic loads with 3D-space
superstructure considering the nonlinearity and the non-homogeneity of the
supporting soil medium using the proposed method of analysis. The program
is called ASTNII (Analysis of Piled Raft with Space Structure Resting on
Nonlinear Soil using Composed coefficient technique) and ASTNIII
(Analysis of Piled Raft with Space Structure Resting on Nonlinear Soil using
against vertical and horizontal loads). This program includes mainly the
effect of the continuity of the soil through a continuum soil model and the
nonlinear relation between contact pressures and soil deformations.
Moreover, ASTNII&ASTNIII can deal with a mat of variable thickness, a
mat with openings, and a mat with inverted beams. Besides, it can deal with
many types of loads such as concentrated vertical and horizontal forces
and/or moments. In addition, partial uniform vertical loads within the frame
element length could be analyzed. ASTNII&ASTNIII can deal with slabs of
variable thickness and variable properties. The effect of the variety of piles
length, diameter and geometry arrangement has been considered. Also, they
can deal with soil bores considering different horizontal layers of the soil
with different geological characteristics. They can also analyze the 3D-space
structure with its boundary conditions taking into account a linear soil or a
nonlinear soil. To include the effect of seismic loads on construction
buildings ASTNIII has been developed to resist the said forces by the pile
102
group. It has the ability to reflect the effect of earthquake forces on the both
superstructure and substructure elements. Composed coefficient technique
employed in ASTNII & ASTNIII by some way to save in memory during
solving equations and consequently to reduce the run time. In spite of taking
the time factor in the consideration, the accuracy is considered also to be in
accepted limits. However, the interface plays an important role in such
analysis codes. It can be considered the gate which controls input and output
data especially huge problems. To confirm the validity of problems analyzed
by ASTNII&ASTNIII, they have been supported by a strong interface using
data base tables. Moreover, a graphic show has been added to the code in
order to enable the user to check up input data as 2D slights and 3D
isometric. ASTNII&ASTNIII are written by Visual Basic which has the
ability to use the extend memory on the hard disk to be promoted easy and
subsequently introduces a more comfortable program. Fig. 5.1 illustrates the
relation between the master program and subprograms. Brief details and
clarifying of master and subprograms are evidently explained in the
following sections.
5.2. Solution Algorithm and the Computer Program

5.2.1. The program ASTNII
It is the master program that controls a group of subprograms to analyze the
3D-space structures with piled raft embedded into a non-homogenous soil
medium for axially loads. The raft is presented as plates using the
Framework Analogy; piles are modeled by rigid beam elements, while the
soil is treated as continuum medium. Full interactions among structural
members; pilesoilpile, pilesoilraft and raftsoilraft are taken into
account. A real simulation model for 3D-space structures resting on piled
103
raft is introduced. Herein, the effect of superstructure and non-homogeneity

of soil is considered to add more reality to the model. The nonlinear analysis
of raft on continuum model of soil using Composed Coefficient Technique is
employed to be applicable for piled raft structures in order to reduce the
stiffness matrix size and consequently saving time.
A general flow chart for each accompanied subprogram is illustrated in
Fig.5.2 and clarified in the following sections.
5.2.2. Subprogram READ DATA

This subroutine reads all the data required for analysis from a preformatted
file data. It contains a group of icons created in an interface screen and
stored in data base files. Each icon defines a part of data by data base tables
which haves the ability to be modified. Data are given successively in
reliable sheets. They are summarized as shown in Fig. 5.2 as follows:
5.2.2.i. The title data block

It contains the data that related to the project title, the date and any
comments.
5.2.2.ii. The main data block

It contains the main data that control the method of calculations in a check
list form as follows:
1- The method of calculation is Winkler or continuum model.
2- The elements which are needed to be in a contact with the soil in
resisting loads as (raft only, pile group in x-direction only, pile group
in y-direction only, pile group in z-direction only and piled raft). One
element or more than one can be selected together.
104
3- The type of analysis is (linear/nonlinear) and the accuracy of

nonlinearity may be determined by allowable error percent or limited
No. of iterations.
5.2.2.iii. The nodes block:

This block defines the coordinates of each node in 3D-space coordinates X\,
Y\ and Z\. Nodes are classified to two types; foundation nodes and
superstructure nodes. Nodes numbers are not restricted to any restrictions in
selecting numbers. Also, the gab is allowed in numbering.
5.2.2.iv. The material block:

This block defines the properties of each element. The properties of each
kind of material and the members are listed in the program using the
following characteristics for frame element type:
a : the area of the cross section,
iyy : the second moment of inertia about Y\-axis in the plane of the cross
section,
izz : the second moment of inertia about Z\-axis in the plane of the cross
section,
j : the polar moment of inertia about the centroidal axis perpendicular to
the plan of the cross section,
beta: the clarified angle in Fig. 4.5.,
GG : the modulus of rigidity of the material and
EE : the modulus of elasticity of the material.
while, for the plate element, the properties of each element is defined by the
thickness, the modulus of elasticity and Poisson ration.
105
Each element should be defined by its number, number of nodes of the two
ends and the kind of its material.
5.2.2.v. The restraints block:

This block defines the boundary conditions of each node by inserting three
numbers. The first number denotes the first node number of restrained
nodes, while the second number represents the end node number of
restrained nodes. The third number represents the increment of nodes.
Degrees of freedom that are restrained can be inserted by selecting restrained
directions from a check list has a six degrees as icons.
5.2.2.vi. The additional springs block:

This block defines the spring stiffnesses that represent the stiffness of soil
for nodes of the foundation. This block is defined by inserting three
numbers. The first number denotes the first node number of springs nodes,
while the second node number represents the last node number of springs
nodes. The third number represents the increment of nodes. Degrees of
freedom that are related to the springs can be inserted by selecting restrained
directions from a check list has a three degrees as icons x-, y- and z-
directions. The stiffness value can be inserted in a text box.
5.2.2.vii. The block of concentrated loads:

This block defines the values, directions and locations of concentrated forces
and moments at nodes. This block is defined by inserting three numbers. The
first number denotes the first node number of loaded nodes, while the second
node number represents the last one. The third number represents the
increment of nodes. Direction of loads that are concentrated can be inserted
106
by selecting directions from a check list has a six degrees as icons. The load
value can be inserted in a text box.
5.2.2.viii. The block of frame element

In this block, the frame members are classified into substructure and
superstructure elements and each type can be defined in tables by the
following items:
- The number of each member, gab in the numbering is allowable.
- The start node number and the end node number.
- The code number of the material type.
5.2.2. ix. The block of plate members

In this block, the plate members are classified into substructure and
superstructure elements and each type can be defined in tables by the
following items:
- The number of each element
- The four nodes which define the plate element
- The code number of the material type.
5.2.2. x. The block of pile data

In this block the data of pile members can be input in flexible grids by
defining the following data:
- The code number of the pile head node.
- The diameter of the pile, length and No. of segments.
The code has the ability to generate the required piles nodes and great
the associated members.
107
5.2.2. xi. The block of soil data

In this block the soil data can be input as half-space or as multi-layers.
Each layer of soil may be described by the following data:
- The start level and the end level of each layer.
- The modulus of elasticity Es, Poisson ratio s of each layer.
- The ultimate bearing stress resisting axially and laterally loads.
5.2.3 The graphic screen

To validate the data, tables are designed to check missed data and illogic
data by interactive messages. However, the code is supported by a
graphic screen has the ability to explore the data as graphics. The screen
has the ability to explore 2D and 3D elements and can avail plans in any
direction and at the desirable level. Moreover, node numbers, member
numbers can be showed on the drawings.
5.2.4. Subprogram OUTPUT DATA

This subprogram prints out the data in a sheet file to check and correct these
data if necessary. The output files can be opened and printed in Excel
format.
5.2.5. The program MAIN PROGRAM

This program is divided into a group of subprograms to get the
displacements in soil and all elements of the structure. Besides, internal
forces under the applied loads for both linear and nonlinear stress-settlement
relationships could be calculated. Figs. 5.3 and 5.6 show the flow chart of
the subprogram MAIN PROGRAM for ASTNII & ASTNIII The groups of
108
subprograms of the main program are evidently explained in the following

sub sections.
5.2.5.i. Subprogram FREEDOM VECTOR

This subprogram generates the freedom vector. Each element in the matrix
has a characterized number related to six degrees of freedom and the node
number. In the first loop, the matrix is equated by zero for all subscripts of
the matrix. In the second loop, subscripts of restrained degrees of freedom
only are equated by 1. This subprogram has the ability to carefully
distinguish between free and restrained nodes.
5.2.5.ii. Subprogram FREEDOMS NUMBER

Herein, the restraints which are equal to 1 will be equated to zero. Therefore,
the free directions of nodes which were equated to zero will be numbered
successively. Each free direction of node will have the number that
represents its arrangement in the vector related to its free directions number
and its nodes number.
5.2.5.iii. Subprogram SOIL STIFFNESS MATRIX

This program generates the flexible matrix for the soil elements of the piled
raft. Pile-soil-pile, pile-soil-raft and raft-soil-raft interactions have been
carried out. In ASTNII, the flexible matrix may be inverted to create the
stiffness matrix. By using Composed coefficient technique the coefficients
of the matrix have been reduced and can be ready to be added to the whole
stiffness matrix of the system as clarified in the next section. In ASTNIII, the
stiffness matrix of the soil is formed for x-, y- and z- directions.
109
5.2.5.iiiv. Subprogram WHOLE STIFFNESS MATRIX

This program is divided into a group of subroutines through which the
factors in each stiffness matrix are formed as shown in Fig. 5.4. The first
subroutine is called CALCULATING ELEMENT. It calculates the length
of element and the direction cosines of angles that the member makes with
X\-axis.
The second subroutine is called SOIL STIFFNESS. Herein, the full matrix
of the soil stiffness can be generated for each element considering the
number of intersecting elements at any node. The third subroutine is called
STRUCTURE STIFFNESS. The upper triangle of the stiffness matrix of
the element in local coordinates can be generated by calculating its length,
its direction and properties in the input data file. The final local matrix of
each element is the summation of the two matrices of the local frame
element and the soil. The fourth subroutine is called ELEMENT CODE. It
sets up the element code number by forming a special matrix of twelve
subscripts considering the same number of degrees of freedom for the two
nodes of each element. The number of the degrees of freedom of any node is
equal to the number of node number multiplied by six, that is the number of
degrees of freedom for any node.
The fifth subroutine is termed ELEMENT STIFFNESS. At first, this
subroutine assigns all elements stiffness matrices in global coordinates to be
zero. Using this assignment, the subroutine can add the elements stiffness
matrices in the global coordinates to the global matrix. Displacements of
same nodes of any two neighboring elements should be added in the global
matrix and stored as half band width by transporting the diagonals as
columns. The half band width matrix is suitable for the cases of
comparatively small memory. It also reduces the calculation time.
110
5.2.5.ix. Subprogram LOADING VECTOR

It forms the load vector of all types of loads for each case of loading.
5.2.5.x. Subprogram LINEAR EQUATION SOLUTION

This subprogram utilizes Gauss Elimination method that leads to the nodal
displacements. The subroutine saves the nodal displacement vector in the
corresponding load vector to minimize the required computer storage
capacity. To increase the solution accuracy, the algorithm has been planned
to decrease the total number of arithmetic operations, the matter which leads
to decrease the error by using the half band width.
5.2.5.xi Subprogram CHECK ACCURACY

In this subprogram, the linear contact pressure calculated by Equation 4.37
and the nonlinear contact pressure calculated by Equation 4.79 are
determined for all nodes. At each node, the error percentage is calculated as
follows.
Ewi = |win - win-1| (5.1)
where, wi is the vertical displacement at any node of the foundation

and n refers to the number of iterations.
%ERROR = MAX |Ewi| / MIN |wi| (5.2)
If the maximum percentage of error exceeds the allowable limit (5%), the
modules of subgrade reaction of all nodes are modified to be the slopes of
111
secants of the nonlinear stresssettlement curve corresponding to the

displacements at every node as illustrated in Figs. 5.5 and 5.7.
5.2.5.xii. Subprogram SAVE DISPLACEMENTS

This program is constructed to save a part of the output results in an output
file. It stores the final nodal displacements (rotations & vertical translations)
in a selected file.
5.2.6. Subprogram SAVE FORCES

This subprogram is divided to many subroutines. The target is to calculate
the soil contact pressure at the mat nodes and the straining actions
(moments, shearing forces, and normal forces) at all nodes of the structure
using nodal displacements. Besides, it calculates the reactions and the
maximum contact pressure underneath the mat.
5.2.7. Subprogram RUN TIME

It measures the time from the start of execution of the program up to the end.
112
START
5.1. CALL READ DATA 1
5.1. CHECK VALIDITY OF 2

DATA
5.1. CALL OUTPUT DATA 3
5.1. CALL MAIN PROGRAM 4
5.1. CALL SAVE FORCES 5

AND DISPLACEMENTS
5.1. CALL RUN TIME 6
END
Fig. 5.1. The Master program ASTNII & ASTNIII
113
START
5.1.1. INPUT TILE DATA 1
5.1.1. INPUT MAIN DATA 2
5.1.1. INPUT PILED RAFT NODES 3

DATA
5.1.1. INPUT SUPER STRUCTURE 4

NODES
5.1.1. INPUT MATERIAL FOR 5

FRAME ELEMENT
5.1.1. INPUT MATERIAL FOR PLATE 6

ELEMENT
114
5.1.1. INPUT FOUNDATION FRAME 7

ELEMENTS
5.1.1. INPUT SUPERSTRUCTURE 8

FRAME ELEMENTS
5.1.1. INPUT FOUNDATION PLATE 9

ELEMENTS
5.1.1. INPUT SUPERSTRUCTURE 10

PLATE ELEMENTS
5.1.1. INPUT PILES DATA 11
5.1.1. GENERATING PILES NODES 12
5.1.1. GENERATING PILES MEMBERS 13
115
5.1.1. INPUT SOIL DATA 14
5.1.1. INPUT ADDITIONAL SPRINGS 15
5.1.1. INPUT RESTRAINED NODES 16
5.1.1. INPUT CONCENTRATED LOADS 17
5.1.1. INPUT DISTRIBUTED FRAME 18

LOADS
5.1.1. INPUT DISTRIBUTED PLATE 19

LOADS
RETURN
Fig. 5.2 The flow chart of READING DATA
116
l No. of Iterations
START
l=0
5.2.4. CALL GENERATING SOIL- a

INTERFACE ELEMENTS
5.2.4. CALL SOIL FLEXABLE b

MATRIX
5.2.4. CALL SOIL STIFFNESS c

MATRIX
5.2.4. CALL COMPOSED c

COEFFICIENT TECHNIQUE
l= l+1
5.2.4. CALL FREEDOM VECTOR d
5.2.4 CALL WHOLE STIFFNESS e

MATRIX
5.2.4. CALL LOADING VECTOR f
5.2.4. LINEAR EQUATION SOLUTION e
X
Y
117
Y
X
5.2.4. CALL SOLVING NONLINEAR 7

EQUATIONS
5.2.4. CALL CHECK ACCURACY 7
IF NO
ACCURACY
IS ACCEPTED
YES
5.2.4. CALL SAVE DISPLACEMENTS h
RETURN
Fig. 5.3 Flow chart of MAIN PROGRAM of ASTNII
118
START
FOR K=1 TO NMEMB
5.2.4.c. CALL CALCULATING ELEMENT 1
5.2.4.c. CALL SOIL STIFFNESS 2
5.2.4.c. CALL STRUCTURE STIFFNESS 3
5.2.4.c. CALL ELEMENT CODE 4
5.2.4.d. CALL ELEMENT STIFFNESS 5
NEXT
NMEMB= Number of total
RETURN
Fig. 5.4. Flow chart of the adding stiffness process
119
START l=Iteration cycle No.
l=0
a Carry out linear analysis to get {w}

Eq. (4.69) or (4.76)
b Find the nodal contact forces {Q}

From Eq. (4.70) or (4.75)
c Determine the nodal stiffness at all

nodes
Kei = Qi / wi
d Modify the piled raft stiffness
1
kei =
1 w
+ i
ksi Qli
e Generate the stiffness matrix [ke]
X Y
120
YES Y
Rigid raft
Find the vector {} of the rigid

body translation and rotations
about x-and y-axes
{N}=[Xs][Ke][Xs]T{}
Compute contact
forces Qi=kei wi
Find the vector {} from FE-method

[[kpr] + [kst]]{} = {P}
l=l+1
Convergence NO
satisfied
{}=|{w}l-{w}l-1|
YES
END
Fig. 5.5 Flow chart of the iteration process of ASTNII
121
l No. of Iterations
START
l=0
5.2.4. CALL GENERATING SOIL- a

INTERFACE ELEMENTS
5.2.4. CALL SOIL FLEXABLE b

MATRIX X-,Y-, Z-DIRECTIONS
5.2.4. CALL SOIL STIFFNESS c

MATRIX X-,Y-, Z-DIRECTIONS
l= l+1
5.2.4. CALL FREEDOM VECTOR d
5.2.4 CALL WHOLE STIFFNESS e

MATRIX
5.2.4. CALL LOADING VECTOR f
5.2.4. LINEAR EQUATION SOLUTION e
X
Y
122
Y
X
5.2.4. CALL SOLVING NONLINEAR 7

EQUATIONS X-,Y-, Z-DIRECTIONS
5.2.4. CALL CHECK ACCURACY 7
IF NO
ACCURACY
IS ACCEPTED
YES
5.2.4. CALL SAVE DISPLACEMENTS h
RETURN
Fig. 5.6 Flow chart of MAIN PROGRAM of ASTNIII
123
START l=Iteration cycle No.
l=0
a Carry out linear analysis to get {w}
b Find the nodal contact forces {Q}
c Determine the nodal stiffness at all
nodes
d Modify the piled raft stiffness in x-, y-, z-directions
1
kei =
1 w
+ i
ksi Qli
e Generate the stiffness matrix [ke] by
Composed Coefficient technique
X Y
124
Compute contact
forces Qi=kei wi
Find the vector {} from FE-method

[[kpr]+ [kph] + [kst]]{} = {P}
l=l+1
NO
Convergence
satisfied
YES
END
Fig. 5.7 Flow chart of the iteration process of ASTNIII
125
5.3. Numerical analysis and results

To verify the validity of the previous procedure of ASTNII & ASTNIII, some
test problems discussed previously by researchers using varied methodes of
analysis and approved programs are compared herein to results carried out
by the present approach. The parametric study presented in next chapters
was carried out with a computer program ASTNII & ASTNIII. ASTNII &
ASTNIII are programs for analyzing 3D superstructure-piled raft foundations
of arbitrary shape with the real subsoil model under axially and horizontally
loads. The mathematical solution of the piled raft is based on the Finite
Element method. These programs can analyze different types of subsoil
models. Continuum model that considers any number of irregular layers has
been achieved. A good advantage of these codes is the capability to handle
the 3D buildings with their piled raft considering the real superstructure in
the analyses.
5.3.1. Test problem 1: verification of axially piled raft considering

nonlinear soil medium (Rabiei,M.,2009)
This model verification was carried out for example represented in Figure
5.8. Figure 5.9 compares the computed load-settlement relationships (up to a
total load of 18 MN) computed from ASTNII with various methods
especially the one done by ELPLA for the centre of the raft with 9 identical
piles (Figure 5.8-b), one under each column. Variation of maximum positive
bending moments, maximum settlement with raft thickness are respectively
illustrated on Fig. 5.9-a, Fig. 5.9-b. Fig. 5.9-c and Fig. 5.9-d. There is
reasonably good agreement between the computed results in this research
with results obtained from other researchers.
126
P2 = P1
Piles are located below P1 & P2
1m
P1 P2
P1
2
P1
P2 P1
2
P1
P2 P1
1
1m 2 2 2 2
Fig. 5.8-a. Load capacity of each pile = 875 [kN] Pile and load configuration
P2
Ep = ER = tR = 0.5 m
P1 P1
30000 MPa
P = P =0.2
l = 10 m
E = 20 MPa
= 0.3
H = 20 m
2 2 2 2
d = 0.5 m
Fig.5.8-b. Model condition and material properties
127
20
Total load [MN] 15
10 ASTNII
ELPLA
5 Van Imp(2001)
GARP
0
0 20 40 60 80 100 120 140
Central settlement [mm]
Fig. 5.9-a Comparison of various methods for load-settlement analysis
1200
Max Moment [kNm/m]
900
600
ASTNII
300
ELPLA
Van Impe (2001)
0
0 0.5 1 1.5 2 2.5
Raft thickness [m]
Fig. 5.9-b Effect of raft thickness on maximum bending moment
128
8
7
6
Max. settlement [cm]
5
4
3
2 ASTNII
1 ELPLA
Van Impe (2001)
0
0 0.5 1 1.5 2 2.5
Raft thickness [m]
Fig. 5.9-c Effect of raft thickness on maximum settlement
2
Diffrential settlement [cm
ASTNII
ELPLA
1.5 Van Impe (2001)
0.5
0
0 0.5 1 1.5 2 2.5
Raft thickness [m]
Fig. 5.9-d Effect of raft thickness on diffrential settlement
129
5.3.2. Test problem 2: Distribution of shear resistance and settlement

curves for an axially single pile (Poulus,1980)
In Fig. 5.10-a, the distribution of shear strength in the single pile which is
carried out by Poulus (1980) has been compared to results analyzed by
ASTNII. The force is normalized as pdl/P, where, p is the shear sress in the
pile, d is the diameter of the pile, l is the length of the pile and P is the total
force concentrated on the pile. The shear stress is drawn againist the depth of
node at the pile as a ratio of the total length.
In Fig. 5.10-b, the settlement is normalized as Ip=Esl/P, where is the
settlement of nodes at the pile and Es id the modulus of elasticity of the soil.
Ip is plotted againist the ratio of pile length to the pile diameter.Results
gotten by ASTNII has been compared to others by Poulos (1980).
3.5
l /d =25, K=50
3 poisson=0
2.5
2
z /l
1.5
ASTNII
1 Polus
0.5
0
0 0.2 0.4 0.6 0.8 1
p (pai) dl /P
Fig. 5.10-a Distribution of shear stress along compressible pile
130
l /d
2 4 6 8 10 12 14 16 18 20
0
0.05
0.1
Ip=settlement *Es *l /P
0.15
0.2
0.25
0.3
0.35
0.4 Polus FEM
Polus Elastic approach
0.45
FEM by ASTNII
0.5
Fig. 5.10-b Settlement for varies pile lengths
5.3.3. Test problem 3: Single pile response (Basile,2003)

In Fig. 5.11, the settlement response of the single pile is plotted againist
changing in the pile load. The results analyzed by ASTNII has been
compared to those done by other approved programs, where, EP is the
modulus of elasticity of the pile.
131
20
15
Pile head load [MN]
10
ASTNII
Jarden et al (1986)
PGROUPN RF shaft=0, base=0
5 Poulos 1986 curve a
Poulos 1986 curve b
PGROUPN RF shaft=.9, base=.9
0
0 1 2 3 4 5 6 7 8 9 10
Head settlement s [mm]
Fig. 5.11 Comparison of load-settlement response for single pile (EP =
30,000GPa)
5.3.4. Test problem 4: Pile group settlement(Basile,2003)

In order to investigate pile group settlement predictions in the linear range,
Figure 5.12 compares ASTNII results with those obtained by some of the
computer programs. Results are expressed in terms of the normalized group
stiffness kp/( n sG) of square groups of piles at different spacing (where kP is
the ratio of the total vertical load acting on the group to the average
settlement of the group, n is the number of piles in the group, s is the pile
spacing and G is the soil shear modulus). In the analyses of
PIGLET and GRUPPALO, it is assumed that axial interaction effects
between piles become insignificant for a pile spacing greater than a limiting
value smax equal to (Randolph & Wroth,1979):
Smax=2.5L (1- s ) + rg (5.1)
132
where s is the soil Poissons ratio and rg, for rectangular pile group
configurations, may be taken as the radius of the circle of equivalent area to
that covered by the pile group. In the analyses of ASTNII, DEFPIG,
PGROUP and PGROUPN, no limiting value for axial interaction effects has
been adopted.
15
ASTNII
L/d=25 DEFPIG
)
GROUPPALO
0.5
Ep/G=1000
Normalized group stiffnessKp /s (Gn
s/d=2.5 PGROUP
PIGLET
10 PGROUPN
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Square root of number of piles in group
Fig. 5.12 Comparison of different pile group analysis methods for s/d=2.5
133
5.3.5. Test problem 5 Case study: Piled raft of Torhaus (El Gendy,2007)
Torhaus is the first building in Germany with a foundation designed as a
piled raft. The building lies in Frankfurt city in Germany. It is 130 [m] high
and rests on two separate piled rafts, where a street passes under the
building. Measured instruments were installed inside the foundation to
record piled raft settlement and stress. Many authors studied the foundation
of the Torhaus and applied their analysis methods on piled raft. Some of
them are Sommer et al. (1985), Sommer (1989) and Reul & Randolf (2003).
Figure 5.13 shows a layout of Torhaus with piled rafts. The building has no
underground floors. The foundation is two separate equal piled rafts with
rectangular shape areas, each of 17.5 [m] * 24.5 [m] sides. The distance
between the two rafts is 10 [m]. Rafts are founded at a depth 3.0 [m] under
the ground surface. The estimated total load on each raft is 200 [MN]. Raft
thickness is 2.5 [m]. A total of 42 bored piles with a length of l = 20 [m] and
diameter of D = 0.9 [m] are located under each raft. The pile spacing varies
from 3.5 D to 3.0 D. The subsoil at the location of the building consists of
gravel and sand up to 2.5 [m] below the ground surface, followed by layers
of Frankfurt clay extending to great depth. The groundwater level lies below
rafts.
The building is constructed between 1983 and 1986, the recorded maximum
settlement at the raft meddle in 1988 was about 12 [cm] according to
Sommer (1989). If Torhaus stands on a raft only, the expected settlement
would be about 26 [cm] based on geotechnical studies according to Sommer
et al. (1985). Therefore, to reduce the settlement, piled rafts were
considered. Using available data and results of Torhaus piled rafts, which
have been discussed in details in the previous references, the present piled
raft analysis is evaluated and verified for analyzing a piled raft.
134
5.3.5.1. Soil properties

According to Reul & Randolf (2003), Youngs modulus of the sand with
gravel layer under the rafts is E = 75000 [kN/m2]. The Youngs modulus for
reloading is taken to be W = 3 E. Based on the back analysis after Amann et
al. (1975), the distribution of modulus of compressibility for loading of
Frankfurt clay with depth is defined by the following empirical formula:
E s = E so (1+ 0.35 z ) (5.2)
while that for reloading is:
[
W s = 70 MN/ m
2
] (5.3)
where:
Es Modulus of compressibility for loading, [MN/m2].
Ws Modulus of compressibility for reloading, [MN/m2].
Eso Initial modulus of compressibility, Eso=7 [MN/m2].
z Depth measured from the clay surface, [m].
The undrained cohesion cu of Frankfurt clay increases with depth from

cu=100 [kN/m2] to cu=400 [kN/m2] in 70 [m] depth under the clay surface
according to Sommer & Katzenbach (1990). Russo (1998) suggested a
limiting shaft friction not less than 180 [kN/m2] meeting undrained shear
strength of 200 [kN/m2]. To carry out the present analysis a limit shaft
friction of ql = 180 [kN/m2] is assumed, which gives a limiting pile load of
Ql = 10 [MN].
Poissons ratio of the soil is taken to be s=0.25 [-].
135
5.3.5.2. Raft and pile material
Raft has the following material parameters:
Young's modulus Eb = 3.4 * 107 [kN/m kN/m2]
Poisson's ratio b = 0.2 [-]
Unit weight b = 25 [kN/m3]
while piles have the following material parameters:
Young's modulus Eb = 2.35 * 107 [kN/m2]
Unit weight b = 25 [kN/m3]
5.3.5.3. Analysis of the piled raft

Comparisons are carried out to evaluate the present nonlinear analysis of
piled elastic raft using composed coefficient technique. In which, results of
three-dimensional finite element analysis and field measurements are
compared with those obtained by the present analysis. The raft is divided
into rectangular elements as shown in Fig. 5.14 and piles are divided into
line elements with 2.0 [m] in length. The raft is considered to be elastic plate
supported on rigid piles. The effective depth of the soil layers under the raft
is taken to be H = 110 [m] as assumed by three-dimensional finite element
analysis.
5.3.5.4. Comparison with three-dimensional finite element analysis and

field measurements
Reul & Randolf (2003) analyzed Torhaus piled rafts using three dimensional
finite element model and compared their results with those obtained by field
measurements according to Sommer (1989). For reducing the computational
effort and time, they took the advantage of the symmetry in shape, soil and
136
load geometry about both x- and y-axes to carry out the analysis for a quarter
of a piled raft. In the present analysis, seven cycles in few minutes are
required to obtain the nonlinear analysis of the piled rafts together. This is
related to that using composed coefficient technique reduced the size of soil
stiffness matrix from [1314*1314] to [390*390]. Accordingly, the total
number of equations was reduced to 1170, where npr =1314, nr = 390 and
number of unknown per node is 3 (3 nr = 1170).
Table 5.1 lists results of central settlement and bearing factor of piled raft
obtained by the present analysis, El Gendy (2007) and those obtained by
Reul & Randolf (2003) using three-dimensional finite element analysis.
Also, the table includes the measured results presented by Sommer (1989).
Figures 5.15 and 5.16 compare loads on piles 1 to 6 (Fig. 5.14) obtained by
the present analysis with those obtained by Reul & Randolf (2003) using
three-dimensional finite element analysis, El Gendy Analysis and with
measured pile loads presented by Sommer (1989).
Table 5.1 Comparison between results obtained by 3D FE-

Analysis, ElGendy analysis and field
measurements with those obtained by the present
analysis
3D FE- El Gendy Present analysis
Type of analysis Measurement
Analysis Analysis
Central settlement 12.4 9.6 11.2 14.9
Bearing factor kpp 67 76 64 77
[%]
This case study shows that the present analysis is not only an acceptable
method to analyze piled raft but also a practical one for analyzing large piled
raft problems. Besides the analysis gives good agreement with measured
137
results, it takes less computational time and less effort for generating input
data compared with other complicated models using three dimensional finite
element analysis.
5.3.5.5. Comparing among different analysis types

To show the difference between results when analyzing piled raft of Torhous
linearly and nonlinearly as piled elastic raft or piled rigid raft, piled raft of
Torhous is analyzed four times as follows:
- Linear piled rigid raft.
- Nonlinear piled rigid raft.
- Linear piled elastic raft.
- Nonlinear piled elastic raft.
- For the four analysis types, Table 5.2 shows central settlement and
bearing factor of piled raft.
138
Table 5.2 Comparison between results of different analysis types
Piled rigid raft by Piled elastic raft by Present study

Type of El Gendy El Gendy
analysis
Linear Nonlinear Linear Nonlinear Linear Nonlinear
Central 7.0 13.4 8.0 11.2 9.2 14.9

settlement
scenter [cm]
Bearing 88 77 75 64 88 77
factor kpp
[%]
Applying different analysis types on piled raft of Torhous shows that the
nonlinear analysis of piled elastic raft is the acceptable analysis type, which
its results are agreement with measured values.
139
(130m)
(100m)
(0.0m)
Sand with gravel
Frankfurt clay
Fig. 5.13 Layout of Tourhous with piled raft
6 1 2 24.5 [m]
5 4 3
17.5 [m] 10 [m] 17.5 [m]
G1.dwg
Fig. 5.14 Mesh of piled raft with piles
140
8
Measurement
Present study
3D-FE analysis
El Gendy analysis
6
Pile load Ph [MN]
0
Pile 5 Pile 4 Pile 3
Fig. 5.15 Comparison between pile loads obtained by 3D FE analysis, El

Gendy analysis and field measurments and those obtained by the
8
Measurement
Present study
3D-FE analysis
El Gendy analysis
6
Pile load Ph [MN]
0
Pile 6 Pile 1 Pile 2
Fig. 5.16 Comparison between pile loads obtained by 3D FE analysis, El

Gendy analysis and field measurments and those obtained by the
141
5.3.6. Test problem 6: Pile Group under General Loading Conditions

(Basile, 2003)
The deformations and load distribution in a 3-pile group under a
combination of axial load, lateral load and moment are examined in the
linear range (refer to Figure 5.16). Results from selected numerical analyses
are compared in figures 5.17~ 5.28 in which w3, u and are the vertical
head displacement of pile no. 3, the horizontal cap displacement and the
rotation of the cap, respectively. The value of the input data is illustrated in
the table 5.3. There is a good agreement between the solutions which
consider pile soil-pile interaction and those analyzed by ASTNIII in this
present study as shown in following figures.
Table 5.3 Input data
Item value unite

The diameter of pile d 0.4 [m]
The space between piles s 1.2 [m]
The modulus of elasticity for the soil Es 7 [MPa]
The modulus of elasticity for the piles Ep 21 [GPa]
Length of piles lp 10 [m]
Vertical load V 600 [kN]
Lateral force H 200 [kN]
Bending moment M 300 [kN.m]
142
V
M
H
.
1 2 3
d lp
s s
Fig.5.17. Group of 3 piles considered in comparison of method
400
300
Equivelant-bent analysis
DEFPIG
Pile load V 1 [kN]
ASTNIII
200 GEPAN
PGROUPN
PIGLET
100
Fig. 5.18. Comparison between pile loads V1 obtained by Equivelant bent

analysis, DEFPIG, GEPAN, PGROUPN, PIGLET and those
analyzed by ASTNIII
143
400
DEFPIG
300 ASTNIII
GEPAN
Pile load V 2 [kN]
PGROUPN
PIGLET
200
100

analyzed by ASTNIII
400
300
Equivelant-
Pile load V 3 [kN]
bent analysis
DEFPIG
200
ASTNIII
GEPAN
100 PGROUPN
PIGLET

analyzed by ASTNIII
144
400
300
DEFPIG
Pile load H1 [kN]
ASTNIII
200 GEPAN
PGROUPN
PIGLET
100
Fig. 5.21. Comparison between lateral pile loads H1 obtained by Equivelant

bent analysis, DEFPIG, GEPAN, PGROUPN, PIGLET and those
analyzed by ASTNIII
400
300
DEFPIG
Pile load H2 [kN]
ASTNIII
200 GEPAN
PGROUPN
PIGLET
100

analyzed by ASTNIII
145
400
300
DEFPIG
Pile load H3 [kN]
ASTNIII
200 GEPAN
PGROUPN
PIGLET
100

analyzed by ASTNIII
-50
Equivelant -bent
analysis
-40 DEFPIG
Pile momentM1 [kN.m]
ASTNIII
-30 GEPAN
PGROUPN
-20 PIGLET
-10
Fig. 5.24. Comparison between pile moment M1 obtained by Equivelant bent

analyzed by ASTNIII
146
-50
Equivelant -bent
analysis
-40 DEFPIG
Pile momentM2 [kN.m]
ASTNIII
-30 GEPAN
PGROUPN
-20 PIGLET
-10

analyzed by ASTNIII
-50
Equivelant -bent
analysis
-40 DEFPIG
Pile moment M3 [kN.m]
ASTNIII
-30 GEPAN
PGROUPN
-20 PIGLET
-10

analyzed by ASTNIII
147
30
Equivelant-bent
analysis
25 DEFPIG
ASTNIII
20
GEPAN
Settlement w [mm]
PGROUPN
15
PIGLET
10
Fig. 5.27. Comparison between settlement w obtained by Equivelant bent

analyzed by ASTNIII
30
Equivelant-bent
analysis
25 DEFPIG
ASTNIII
20
Lateral displacement u [mm]
GEPAN
PGROUPN
15
PIGLET
10
Fig.5.28. Comparison between lateral displacement u obtained by Equivelant

analyzed by ASTNIII
148
10
8 DEFPIG
Pile rotation * 10 -3 [rad]
ASTNIII
6 GEPAN
PGROUPN
PIGLET
4
Fig.5.29. Comparison between pile rotation obtained by Equivelant bent

analyzed by ASTNIII
5.3.7. Test problem 7: Lateral load distribution (Basile,2003)

Model verification was carried out to be represented by an individual piles 5
x 5 pile group as shown in Figure 5.29. compares the distribution of lateral
pile load which is normalized as a ratio of Hav/H and the space between
piles which normalized as a ratio between s/d computed from ASTNIII with
various methods especially the one carried out by PGROUPN developed by
Basile (2003). H is the load acting on the individual pile head, Hav is the
average load acting on each pile head as a function of the normalized pile
spacing (s/d), s is the space between piles and d is the diameter of the pile.
There is reasonably good agreement between the computed results in this
research with results obtained from other researchers.
149
2
1.8 Corner pile curve
1.6
MPILE
1.4
1.2 DEFPIG
H/Hav
1
PGROUPN(nonlin)
0.8
0.6 PGROUPN(lin)
0.4
Center pile curve ASTNIII
0.2
0
3 6 9 12 15
s/d
Fig. 5.30. Comparison of lateral load distribution to individual piles in 5 5

pile group
5.4. conclusion
In this chapter, a new two codes edited by the author have been introduced
using flow charts. ASTNII and ASTNIII which presented in this chapter have
the ability to analyze 3D buildings resting on piled raft considering a
continuum model for the soil. This model takes in the account the interaction
effect between different elements pile-soil-pile, raft-soil-raft and pile-soil-
raft. Moreover, loads in the lateral directions besides the vertical direction
are considered and non-homogenuity and non-linearity of the soil are
considered also. The valadity of those codes are examined in some test
problems which are presented in this chapter also. Curves which show the
comparison between results carried out by ASTNII and ASTNIII and those
results analyzed by other approved programs reflected a good agreement
between ASTNII and ASTNIII from a side and other approved programs from
another side.
150
REFERENCES
REFERENCES
References
1. Ahmad, J., Durrani, Mau, S. T. and Amr, A. A. Hashish,

(1994): "Earthquake Response of Flat-Slab Buildings",
Journal of Structural Engineering, Vol. 120, No. 3, pp. 947-
964.
2. Banerjee, P. K. and Davies, T. G. (1977): Analysis of Pile
Groups Embedded in Gibson Soil, Proc. 9th International
Conference on Soil Mechanics and Foundation Engineering,
Tokyo, pp. 381-386.
3. Basile,F. (2003): Analysis and Design of Pile Groups,
Numerical Analysis and Modelling in Geomechanics, edited by
Bull, London, Ch. 10, pp. 278-315.
4. Baz, M. (1987): Plates on Nonlinear Subgrade , M. Sc.
Thesis, El-Mansoura University, Egypt.
5. 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.
6. Broms, B. B. (1964): Lateral resistance of piles in cohesive
soils. J. Soil Mechs Fdn Division, Am. Soc. Civ. Engrs 90, No.
SM2, 27-63.
7. 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.
240
REFERENCES
8. 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.
9. Brown, P. T. (1975). The Significance of Structure-
Foundation Interaction, Proc. 2nd Australia-New Zealand
Conference on Geomechanics, Brisbane, IEQust, Vol. 1, No. 1,
pp. 79-82.
10. Butterfield, R. and Banerjee, P. K. (1971a): The Elastic
Analysis of Compressible Piles and Pile Groups,
11. Butterfield , R. and Banerjee, P. K. (1971b): The Problem of
Pile Group Pile Cap Interaction, Gotechnique, Vol. 21, No.
2, pp. 135-142.
12. Cheung, Y, K. and Zienkiewicz, O. C. (1965): Plates and
Tanks on Elastic FoundationsAn Application of Finite
Element Method, Int. Jl. of Solids Sturctures, Vol. 1, pp. 451-
461.
13. Chow, Y. K. (1986): Analysis of Vertically Loaded Pile
Groups, Int. Jl. For Numerical and Analytical Methods in
Geomechanics, Vol. 10, pp. 59-72.
14. 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.
15. 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.
16. Chow, Y. K. (1987a): Axial and Lateral Response of Pile
Groups Embedded in Nonhomogeneous Soil, Int. Jl. For
241
REFERENCES
Numerical and Analytical Methods in Geomechanics, Vol. 11,

no. 6, pp. 621-638.
17. Chow, Y. K. (1989): Axially Loaded Piles and Pile Groups
Embedded in a Crossanisotropic Soil, Gotechnique, Vol. 39,
No. 2, pp. 203-211.
18. 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.
19. Cheung, Y. K. (1976): Finite Strip Method in Structural
Analysis, Pergamon Press, Oxford.
20. 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.
21. Cheung, Y. K. and Nag, D. K. (1968): Plates and Beams on
Elastic Foundations Linear and Nonlinear Behaviour,
Gotechnique, Vol. 18, pp. 250-260.
22. 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.
23. 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.
24. Das,B. (1999): "Shallow Foundations Bearing Capacity and
Settlement", California State University, Sacramento, USA.
242
REFERENCES
25. Drucker, D. C. and Prager, W. (1952): Soil mechanics and

Plastic Analysis of Limit Design, Quart. Applied
Mathematics, Vol. 10, No. 2.
26. Douglas, D. J. & Davis, E. H. (1964): The Movement of
Buried Footings Due to Moment and Horizontal Load and the
Movement of Anchor Plates Geot., vol. 14: 115-132.
27. Egyptian code (2008): Calculation Loads and Forces in the
Construction and Building Works, Egypt
28. El Azab, M. (1986): "Engineering properties of geotechnical
properties of soil", M.Sc. Thesis, Faculty of Engineering, Suez
Canal University, Port-said, Egypt.
29. El Gendy,M. (2007): "Formulation of a Composed Coefficient
Technique for Analyzing Large Piled Raft", Ain Shams
university, Vol.42, No.1, March 31.
30. El kamash,W. (2004): "Analysis of Structures Resting on
Nonlinear Soil", M. Sc. Thesis, Suez Canal University, Egypt.
31. Fleming, W. (1992): A new method for single pile settlement
predication and analysis. Gotechnique, Vol. 42, No. 3, 411-
425.
32. Fleming, W. G. K., Weltman, A. J., Randolph, M. F. & Elson,
W. K. (1992): Piling Engineering (2nd edn). Blackie
Academic & professional, Glasgow.
33. 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.
34. Golder Associates, (1979): "Geotechnical Report for Port-Said
Area", Port-Said, Egypt.
243
REFERENCES
35. 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.
36. Hain, S. J. and Lee, I. K. (1974): Rational Analysis of Raft
Foundation, Jl. Of Geotechnical Engineering, Vol. 100, No.
GT7, pp. 843-860.
37. Honghua,Z. (2000): "Finite Layer Method for Analysis of
Piled Raft Foundation", Sedney University, Australia.
38. Helensze,W. (2007): "Analysis of Piled-Raft Foundations with
Piles of Different Lengths and Diameters", Sedney University,
Australia
39. 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.
40. Hain, S. J. and Lee, I. K. (1978): The Analysis of Flexible
Raft-Pile Systems, Gotechnique, Vol. 28, No. 1, pp. 65-83.
41. 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.
42. Hussin, M. (1983): Finite Element Analysis of Mat Resting
On Nonlinear Elastic Medium , M. Sc. Thesis, Ain Shams
University, Egypt.
43. Kay, J. N and Cavagnaro, R. L. (1983): Settlement of Raft
Foundations, Jl. of Geotechnical Engineering, Vol. 109, No.
11, pp. 1367-1382.
244
REFERENCES
44. Kuwabara, F. (1989): An Elastic Analysis for Piled Raft

Foundations in a Homogeneous Soil, Soils and Foundations,
Vol. 29, No. 1, pp. 82-92.
45. Kitiyodom, P. and Matsumoto, T. (2003): A Simplified
Analysis Method for Piled Raft Foundations in Non-
homogeneous Soils, Int Jl. for Numerical and Analytical
Methodsin Geomechanics, Vol. 27, No. 2, pp. 88-109.
46. Katsikadelis, J. T and Armenkas, A. E. (1984): Plates on
Elastic Foundation by BIE Method, Jl. of Engineering
Mechanics, Vol. 110, No. 7, pp. 1086-1105.
47. Katzenbach, R. and Reul, O. (1997): Design and Performance
of Piled Rafts, Proc.XIVth ICSMFE 97, Hamburg, Vol. 4, pp.
2253-2256.
48. 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.
49. Kakurai, M., Yamashita, K. and Tomono, M. (1987):
Settlement Behaviour of Piled Raft Foundation on Soft
Ground, Proc. 8th Asian Regional Conference on Soil
Mechanics and Foundation Engineering, Kyoto, Japan, Vol. 1,
pp. 373-376.
50. Katsikadelis, J. T and Armenkas, A. E. (1984): Analysis of
Clamped Plates on Elastic Foundation by the Boundary
Integral Equation Method, Jl. of Applied Mechanics, Vol. 51,
pp. 574-580.
245
REFERENCES
51. Lee, C. Y. (1993): Settlement of Pile Groups Practical

Approach, Jl. of Geotechnical Engineering, Vol. 119, No. 9,
pp. 1449-1461.
52. Guo, W. D. and Randolph, M. F. (1999): An Efficient
Approach for Settlement Prediction of Pile Groups,
53. 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.
54. 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.
55. 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.
56. 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.
57. 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.
58. 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.
59. 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.
246
REFERENCES
60. 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.
61. Mandolini, A. and Viggiani, C. (1997): Settlement of Piled
Foundations, Gotechnique, Vol. 47, No. 4, pp. 791-816.
62. 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.
63. 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.
64. Mindlin, R. D. (1936): Force at a point in the Interior of Semi-
Infinite Solid. Physics 7:195.
65. Nishida, Y. (1956): A brief note on compression index of
soils. Journal of SMFE Div., ASCE, July, p 1027 (1-14).
66. 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.
67. 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. 9th
Offshore Technology Conference, Vol. 2, pp. 245-256.
68. Ottaviani, M. (1975): Three-dimensional Finite Element
Analysis of Vertically Loaded Pile Groups, Gotechnique,
Vol. 25, No. 2, pp. 159-174.
247
REFERENCES
69. Poulos HG. (1994): "An Approximate Numerical Analysis of

PileRaft Interaction", International Journal for Numerical and
Analytical Methods in Geomechanics; 18:7392.
70. Poulos, HG. and Davis (1980): "Pile Foundation Analysis and
Design", Jhon Wiley & Sons Inc., New York.
71. Poulos, H. G. (1993): Piled Rafts in Swelling or
Consolidating Soils, Jl. Geotech. Engng, ASCE, Vol. 119, No.
2, pp. 374-380.
72. 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.
73. Sinha, J. (1997): Piled raft foundations subjected to swelling
and shrinking soils. Ph.D. Thesis, Univ. Sydney, Australia.
74. 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.
75. 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.
76. 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.
77. 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.
248
REFERENCES
78. Small, J. C. (2001): Practical Solutions to Soil-Structure

Interaction Problems, Prog. Struct. Engng. Mater., Vol. 3, pp.
305-314.
79. Small, J. C. (2002): Soil-Structure Interaction, E. H. Davis
Memorial Lecture, Australian Geomechanics, March, pp. 1-16.
80. Somaprasad, R. H. and Ahmet, E. A. (1994): "Seismic
Vulnerability of Flat Slab-Core Buildings", Journal of
Structural Engineering, Vol. 120, No. 2, pp. 339-359.
81. 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.
82. Rabiei, M. (2009): Parametric Study for Piled Raft
Foundations, Journal of Geotechnical Engineering, EJGE;
Vol. 14, Bundle A, 11 p.
83. 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.
84. 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.
85. 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.
86. Reda, A. (2009): Optimization of piled raft in Port-Said
M.Sc. Thesis, Faculty of Engineering, Suez Canal University,
Port-said, Egypt.
249
REFERENCES
87. Russo, G. (1998). Numerical Analysis of Piled rafts, Int Jl.

for Numerical and Analytical Methods in Geomechanics, Vol.
22, No. 6, pp. 477-493.
88. 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.
89. 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.
90. Reul, O. (1998): Soil-structure-interaction of a Piled Raft
Foundation of a 121 m high Office Building in Berlin, Proc.
12th European Young Geotechnical Engineers Conference,
Tallinn, Estonia, pp. 1-12.
91. 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..
92. 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.
93. 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.
94. Wardle, L. J. and Fraser, R. A. (1974): Finite Element
Analysis of a Plate on Layered Cross-Anisotropic Foundation,
250
REFERENCES
Proc. 1st International Conference for Finite Element Method

Engineering, University of NSW, pp. 565-578.
95. Wiesner, T. J. (1991): Various Applications of Piled Raft
Analysis, Computer Methods and Advances in Geomechanics,
Beer, Booker & Carter (eds), Balkema, Rotterdam, pp. 1035-
1039.
96. 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.
97. 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.
98. Yao, Z. E. and Zhang, J. R. (1985): An Assessment of the
Effects of Structure/Raft/Soil Interaction, Proc. 5th Int.
Conference on Numerical Methods in Geomechanics, 1-5
April, Nagoya, Japan, Vol. 2, pp. 813-819.
99. Yettram, L./ Husain, M. (1965): Grid Framework Method for
Plates in Flexure ASCE, Mechanics Division, June, pp. 53-64.
100. 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.
101. Zhang, B. Q and Small, J. C. (1994): Finite Layer
Analysis of Soil-Raft-Structure Interaction, Proc. XIIIth
ICSMFE 94, 5 10 January, New Delhi, Vol. 2, pp. 587-590.
102. 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.
251
REFERENCES
103. Zhang, H. H. (2000): Finite Layer Method for Analysis

of Piled Raft Foundations. Ph.D. Thesis, Univ. Sydney,
Australia.
104. Zhao, X. and Cao, M. (1985): Soil-Structure Interaction
Analysis and its Application to Foundation Design in
Shanghai, Proc. 5th Int. Conference on Numerical Methods in
Geomechanics, 1-5 April, Nagoya, Japan, Vol. 2, pp. 805-811.
105. Zhemochkin, B. N. and Sinitsyn, A. P. (1962):.Practical
Methods of Designing Foundation Beams and Slabs Resting on
an Elastic Foundation, 2nd ed., State Publishing House for
Lit. on Struct., Archit. and Struct. Matl., Moscow.
252
APPENDIX I DISPLACEMENT AT THE CENTER DUE TO A CYLINDRICAL SURFACE STRESS
APPENDIX I
DISPLACEMENT AT THE CENTER DUE TO A CYLINDRICAL

SURFACE STRESS
The derivation of the integrated form of the singular Mindlins solution over
cylinder surface is presented.
R2
ZB
A
ZA ro = x2 + y2
R1
ZB
B
R1 = ro2 + ( z A z B ) 2
ro
R2 = ro2 + ( z A + z B ) 2
P
y
x
Z
The vertical displacement w at the nodal point A caused by a vertical point

load P at B may be expressed as:
w( A) = G ( A, B) P( B) (I.1)
252
where G(A,B)P(B) is given by Mindlins solution:
3 4 s 8 (1 s ) 2 ( 3 4 s ) ( z A z B ) 2
+ +
1 R1 R2 R1 3

G ( A, B ) = 2
16 G s (1 s ) ( 3 4 s )( z A + z B ) 2 z A z B
2
6z A zB (z A + zB )
+ 3
+
R 25
R2
(I.2)
For distributed (constant ) traction t over a surface S, the following equation
can be obtained:
w( A) = t ( B ) ( A, B )dS ( B ) (I.3)
s
where the axial traction t are distributed over a cylindrical shaft area of the
pile-soil interface. The singular part G*(A,B) of Mindlins solution may be
expressed as:
3 4 s ( z A z B )
G * ( A, B ) = C + (I.4)
R1 R13
1
Where C is a constant equal to
16G s (1 s ) . In order to integrate Eq.
(I.4), it is convenient to choose the origin of coordinates at the nodal point A,
ie zA =0, and, for simplicity, zB=z. Therefore, R1 may be expressed as
R1 = ro2 + z 2 , where ro represents the pile radius. Thus, Eq. (I.4)

becomes:
253
3 4
z2
G ( A, B ) = C
* s
+ 2
r +z
2 2 (ro + z )
2 3/ 2 (I.5)
o
and the following surface integral has to be evaluated:
3 4 2
z
s G ( A, B)dS ( B) = C s r 2 + z 2 + (ro2 + z 2 ) 3 / 2 dS ( B)
* s
(I.6)
o
where S is the cylindrical surface of the pile shaft element. The two integrals
in Eq. (I.6) may be evaluated in polar coordinates) separately. The first
integral yields:
2 h / 2
1 1
(3 4 s ) dS ( B ) = (3 4 s ) ro dzd
s r +z
o
2 2
0 h / 2 r +z
o
2 2
2
h/2 h + h 2 + 4ro2
= 2ro (3 4 s ) ln z + z + ro
2
= 2ro (3 4 s ) ln
h / 2 h + h 2 + 4r02
h + h2 + d 2
= d (3 4 s ) ln
h + h2 + d 2
where h is the height of the pile shaft element and d is the pile shaft
diameter. The second integral yields:
2 h / 2
z2 z2
s (ro2 + z 2 ) 3 / 2 dS ( B) = 0 h/ 2 (ro2 + z 2 ) 3 / 2 ro dzd
254
h/2
z2
= 2 ro dz
h / 2 ( ro
2
+ z 2 3/ 2
)
This integral may be evaluated by substitution-let z = ro tan , then:
dz = ro sec 2 d
Thus:
ro2 tan 2
h/2 arctan h / 2 r
z2 o

h / 2 ( ro + z )
2 2 3/ 2
dz =
arctan h / 2 ro ( ro tan + ro )
2 2
r sec 2 d
2 3/ 2 o
arctan h / d o
=
arctan h / d o
tan sin d
The above integral may be evaluated by parts- let:
u = tan , dv = sin d , du = sec 2 d , v = cos

Thus:
arctan h / d o arctan h / d
= tan sin d = [ cos tan ] + cos sec d
arctan h / d 2
arctan h / d
arctan h / d o arctan h / d
h h
arctan h / d + sec arctan
= [ sin ]arctan h / d
h
+ sec d = 2 sin arctan + ln d d
arctan h / d
d h h
arctan h / d sec arctan
d d
Thus, the final expression of Eq. (I.6) is:
255
G ( A, B )dS ( B ) =
*
h h
sec arctan +
h+ h +d
2 2
h d d
= Cd (3 4 s ) ln 2 sin arctan + ln
+ + h h
sec arctan
h h 2
d 2 d
d d
256
APPENDIX II DISPLACEMENT AT THE CENTER DUE TO A CIRCULAR UNIFORM LOAD
APPENDIX II
DISPLACEMENT AT THE CENTER DUE TO A CIRCULAR

UNIFORM LOAD
The displacement wi [m] at the edge i of a circular loaded area of radius ro

[m] and a uniform load q [kN/m2] beneath the surface of a semi-infinite
mass, Figure II.1, can be obtained from:

2 ro
wi = q f ij rdrd (II.1)
0 0
Substituting Eq. (I.2) into Eq. (II.1) and carrying out the integration:

[
(3 4 ) r 2 + ( z c) 2 z c ]

][ ]
s o

[
+ 8(1 ) 2 (3 4 ) r 2 + ( z c) 2 ( z + c)
s s o

1
wi =
q + ( z c) 2

1

8G s (1 s ) z c ro + ( z c)
2 2

[
+ (3 4 )( z + c) 2 2cz 1 ] 1

z + c
s
ro2 + ( z + c) 2

+ 2cz ( z + c) 2 1
1

( z + c ) 3
(
ro + ( z + c)
2
)
2 3/ 2

(II.2)
257
APPENDIX II DISPLACEMENT AT THE CENTER DUE TO A CIRCULAR UNIFORM LOAD
Fig. II.1 Geometry of circular loaded area for finding displacements at center
At depth z = c, the displacement at the center of a circular loaded area is

given by:
2cz ( z + c) 2
8([
1 ) 2
(3 4 ) ]
( z + c )
( z + c)
s s

3
q
wi =
16G s (1 s ) (II.3)
(2 4 ) z c + (3 4 s )( z + c) 2cz
2
z+c
s
258
APPENDIX III ELASTIC EQUATIONS USED FOR LATERALLY-LOADED PILE ANALYSIS
APPENDIX III
LATERAL DISPLACEMENT AT THE CENTER DUE TO A

RECTANGULAR UNIFORM LOAD
The basic equation from which the required solution is derived that obtained
by Mindlin (1936), for the horizontal displacements caused by a horizontal
point load embedded in the interior of a semi-infinite elastic-isotropic
homogeneous mass. Referring to Fig. III.1, this solution is as follows:
(3 4 ) 1 x 2 (3 4 ) x 2 2cz 3 x 2
+ + 3
+ 3
+ 3 1 2 +
Q R R R R R2 R2
x =
1 2 1 2
(III.1)
16G (1 ) 4(1 )(1 2 ) x2
1
R2 + z + c R2 ( R2 + z + c )
Douglas and Davis (1964) integrated this equation over rectangular area, and
obtained the following solution.
At the upper corners A and B as shown in Fig. III.2, for a uniform horizontal
pressure p,
259
c
Surface
R2= [r2 + (z+c)2]1/2

c
Q
z
R1= [r2 + (z-c)2]1/2
r = (x2+y2)1/2
(Loading in x direction)
Fig. III.1. Definition of Point-load problem.
x =
pb
[(3 4 ) F1 + F4 + 4(1 2 ) (1 ) F5 ] (III.3)
32G (1 )
At the lower corners D and C,
x =
pb
[(3 4 ) F1 + F2 + 4(1 2 ) (1 ) F3 ]
32G (1 )
where
2C1
K1 =
b
2C2
K2 =
b
( K1 K 2 ) 2
F1 = ( K1 + K 2 ) ln 2 ln
2 + 4 + ( K1 K 2 ) 2 ( K1 K 2 ) + 4 + ( K1 K 2 ) 2
260
2( K1 + 1 + K12 )
F2 = 2 ln + ( K1 K 2 )
( K1 + K 2 ) + 4 + ( K1 + K 2 ) 2
4 + ( K1 + K 2 ) 2 4 + ( K + K )2 1 + K12
ln 2 + K1
2 1 2

( K1 + K 2 ) ( K1 + K 2 ) K1

( K1 + K 2 )
F3 = 2 K1 ln + ( K + K ) ln
K1

1+ 1+ K 2 1 2
2 + 4 + ( K + K ) 2

1 1 2
( K + K ) + 4 + ( K + K )2 ( K + K )
ln 1

2 1
2( K1 + 1 + K1 )
2
2
+ 1
4
2
[ 4 + (K + K )
1 2
2
]
( K1 + K 2 )
K1 ( 1 + K12 K1 )
2( K 2 + 1 + K 22 ) 2 + 4 + ( K + K )2
F4 = 2 ln + ( K1 K 2 ) ln 1 2
+
( K1 + K 2 ) + 4 + ( K1 + K 2 )
2
( K1 + K 2 )
4 + ( K + K )2 1 + K 22
K
2 1 2

( K1 + K 2 )
2
K2

K1 + K 2
F5 = 2 K 2 ln ( K + K ) ln
K2
+
1+ 1+ K 2 1 2
2 + 4 + ( K + K ) 2

2 1 2
( K + K ) + 4 + ( K + K )2 K + K
ln 1

2 1
2( K 2 + 1 + K 2 )
2
2
1

2
4
[ 4 + (K + K )
1 2
2
( K1 + K 2 ) ]
(
K 2 K 2 1 + K 22 )
For the displacement at other points in the same plane, the principle of
superposition may be employed.
261
Surface z = 0
C2
A B
C1
D b C
Fig. III.2. Definition of rectangular area
262

Analysis of 3D-Structures Resting On Piled Raft PDF

Caricato da

Informazioni sul documento

Titolo originale

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

Analysis of 3D-Structures Resting On Piled Raft PDF

Caricato da

Copyright:

Formati disponibili

Suez Canal University

Analysis of 3D-Structures Resting on Piled Raft

Waleed Hamdy Mohamed El Kamash

M.Sc. Civil Engineering, 2004

Prof. Dr. Faisal Khalil Ibrahim

Dr. Mohamed Mosaad El-Gendy Dr. Rafeek Wadieh Salib

Name, Title & Affiliation Signature

1- Prof. Dr. Faisal Khalil Ibrahim ..

Prof. of Structural Engineering

2- Dr. Mohamed Mosaad El-Gendy ..

Assoc. Prof. Civil Eng. Dept.

3- Dr. Rafeek Wadieh Abdel Missih Salib ..

Assist. Prof. Civil Eng. Dept.

Analysis of 3D-Structures Resting on Piled Raft

Waleed Hamdy Mohamed El Kamash

Name, Title & Affiliation Signature

3- Prof. Dr. Faisal Khalil Ibrahim .

Prof. of Structural Engineering

Suez Canal University

4- Dr. Mohamed Mosaad El-Gendy .

Assoc. Prof. Civil Eng. Dept.

Suez Canal University

I wish to express my deepest gratitude and sincere appreciation to Prof. Dr.

Faisal Khalil Ibrahim, Professor of Structural Engineering, Suez Canal

University, for his kind supervision, generous support, valuable suggestions

and constant encouragement during all phases of this research work.

El- Gendy, Associate Professor of Soil Mechanics and Foundation

suggestions during this work.

I am deeply indebted to Dr. Rafeek Wadieh Abdel Missih Salib, Assistant

this work to a successful end.

Words, however ample, would be insufficient to express my gratitude to my

In the present study, there is a numerical model of a numerical analysis

CHAPTER 2: LITERATURE REVIEW

CHAPTER 3: PORT-SAID SOIL PROPERTIES AND

CHAPTER 4: ANALYSIS OF PILED RAFT IN LAYERED SOIL

CHAPTER 5: COMPUTER PROGRAM AND VALIDITY OF

CHAPTER 6: AXIALLY LOADED BUILDINGS ON PORT

CHAPTER 7: BUILDINGS ON PORT SAID SOIL MEDUIM

CHAPTER 8: SUMMARY AND CONCLUSIONS

Table No. Title Page

Table (2.1) Reduction coefficient 33

Table (2.2) Soil coefficient factors 34

Table (3.1) Thickness of soil layers in the different zones in meter 40

Table (3.2) Soil parameters 47

Table (5.1) Comparison between results obtained by 3D FE- 137

Table (5.2) Comparison between results of different analysis types 139

Table (5.3) Input data 142

Table (6.1) Models and dimensions of columns 156

Table (6.2) Soil parameters 159

Figure No. Title Page

Figure (4.3) Member Axes and freedoms for a 3D-space

Figure No. Title Page

Figure (5.7) Flow chart of the iteration process of ASTNIII

Figure No. Title

Figure (5.10.b) Settlement for varies pile lengths

Figure (5.11) Comparison of load-settlement response for single

Figure (5.12) Comparison of different pile group analysis

Figure (5.13) Layout of Tourhous with piled raft 140

Figure (5.16) Comparison between pile loads obtained by 3D FE

Figure (5.17) Group of 3 piles considered in comparison of

Figure (5.18) Comparison between pile loads V1 obtained by

Figure (5.19) Comparison between pile loads V2 obtained by

Figure (5.21) Comparison between lateral pile loads H1 obtained