Experimental Studies of Model Piled Raft PDF

EXPERIMENTAL STUDIES OF MODEL PILED
RAFTS ON SAND AND FIELD STUDY OF

PROTOTYPE BEHAVIOUR
A THESIS
Submitted by
V. BALAKUMAR
in fulfillment for the award of the degree

of
DOCTOR OF PHILOSOPHY
FACULTY OF CIVIL ENGINEERING

ANNA UNIVERSITY
CHENNAI – 600 025
JULY 2008
ii
ANNA UNIVERSITY : CHENNAI 600 025
BONAFIDE CERTIFICATE
Certified that this thesis titled “EXPERIMENTAL STUDIES
OF MODEL PILED RAFTS ON SAND AND FIELD STUDY OF
PROTOTYPE BEHAVIOUR” is the bonafide work of
Mr. V. BALAKUMAR who carried out the research under my
supervision. Certified further that to the best of my knowledge the work
reported herein does not form part of any other thesis or dissertation on
the basis of which a degree or award was conferred on an earlier
occasion of this or any other candidate.
Dr. K. ILAMPARUTHI
Place: Chennai (SUPERVISOR)
Date : 31.12.2007 Professor and Head
Division of Soil Mechanics and
Foundation Engineering
Department of Civil Engineering
Anna University
Chennai – 600 025
iii
6
7ABSTRACT
Piled raft foundation is a fairly recent technique, in which the load

is shared by its constituting elements namely the raft, piles and the supporting
soil. This is yet to become a choice in the minds of designers because of the
complexity involved in the analysis. The works reported in the literature are
mostly on overconsolidated clay. Further its applicability to support
moderately loaded buildings on sand and predominantly sandy soils are yet to
be established. Keeping all the above in mind, a series of small scale 1g
model test were conducted on piled raft founded on sand and the results were
compared with three dimensional nonlinear analyses. Apart from model tests,
a prototype piled raft foundation supporting a twelve storied structure was
designed, constructed, and instrumented with settlement gauges and
monitored during and after the construction.
In the case of small scale model studies, nearly 110 tests were
carried out in all on the circular, square and strip piled raft. Test on the
unpiled raft was also performed along with a series of tests on free standing
pile group (pile group with cap/raft not in contact with soil). The load
settlement behavior of the plain raft, free standing pile group and the piled raft
were studied in detail under uniformly distributed load. From the results, the
settlement reduction and load share between the raft and the pile group was
obtained. It was found that the pile group of piled raft was exhibiting an
elastic and work hardening behavior and was resisting a higher load than the
free standing pile group. It was found that at a settlement of 2mm, which is
1% of the raft diameter (or width of the raft) used, the pile group starts losing
its stiffness rapidly. In the case of piled raft while the settlement was same,
the magnitude of frictional resistance of piles was much higher than free
iv
standing pile group for a given settlement. This settlement has been termed as
critical settlement and the magnitude of load taken, has been termed as
limiting friction. All the tests were performed upto a settlement of 20mm, and
the settlement reduction was arrived with reference to plain raft settlement.
It was found that the piled raft exhibited three phase behaviour. In
the beginning the major part of the load was shared by the pile group. As the
settlement increased the stiffness of the piled raft reduced and it was found
that beyond a settlement level of 3% of the diameter / the width of the raft, the
stiffness approached the magnitude of plain raft and at this stage the load
shared by the pile group remained constant. This response of piled raft in sand
is characterised as three phase behaviour and is termed as multi-linear strain
hardening behaviour. The relative stiffness of the piledraft tested is higher
than the raft stiffness irrespective of the pile parameters and densities of sand
tested.
The parametric analyses carried out through 1g tests in this study

established that the addition of even a smaller number of piles could reduce
the raft settlement by an appreciable amount, though the load shared by them
was small. It was found that an addition of piles of area ratio of 5.2% to the
raft could produce a settlement reduction of 37% corresponding to 20mm
settlement of the plain raft. It was also found that the load sharing behavior
was depending upon the settlement. As the settlement increased the load
sharing ratio decreased and for settlements more than 3% to 4% of the lateral
dimension of the raft used, the magnitude of load sharing ratio, αpr remained
constant. It was found that the number and length of the piles had a noticeable
influence on the settlement reduction and load sharing behavior of the piled
raft. Further, it was seen that increasing the pile length more than the L/B
ratio (L = Length of pile and B = diameter or width of square raft) of 0.8 the
contribution towards load sharing and settlement reduction is not appreciable.
v
However the length of pile of 1B is preferable for effective performance of

the piled raft. Also the diameter to raft thickness ratio (d/t ratio) of unity will
be ideal for obtaining a maximum settlement reduction and load sharing. It
was also found that the performance of piled rafts were similar in all the three
bed densities irrespective of their shapes.
The numerical analyses performed using ANSYS FEM Code

showed a stiffer response when soil was treated as elastic material. The 3D
nonlinear analyses with MISO material model for soil continuum predicted
the behaviour closely with the 1g model tests. The head load and the tip load
obtained from the numerical analysis indicated that the tip load was of the
order of 10% of the applied load on the piled raft indicating the pile group
functions purely as friction pile. The load sharing ratios obtained from the
numerical analyses and the model tests agreed very closely.
Concurrently a twelve storied structure designed and supported on

the piled raft was monitored over a period of 796 days including the post
construction period of 436 days. The structure was supported on 93 piles of
500mm diameter having a length of 14m below the raft of 600mm thick
founded in a deposit which is predominantly sand. The maximum settlement
of the piled raft under the combination of structural load on completion and
load on occupation of building (145kN/m2) was found to vary 9mm to 14mm
at the location of settlement markers, which had showed that practically there
was no differential settlement. From the settlements observed, the contact
pressure below the raft was back computed using elastic theory for each stage
of settlement; it was found that the load shared by the raft progressively
increased to share 43% of the load at the end. The three dimensional linear
analysis performed indicated that the contact pressure distribution below the
raft was almost uniform. Also the pile head load and tip load distribution
indicated that the piles behaved as friction piles which is one of the essential
vi
requirements of the settlement reducing piles. Also it was found that the load
sharing between the piles and the raft was of 62% and 38% respectively and is
in close agreement with the values obtained from the observed settlements
adopting elastic theory.
The entire study had established that the piled raft can perform
satisfactorily in sand. The introduction of piles having a small ratio of area
around 5% can produce a settlement reduction of 20% to 65% depending on
the diameter and the length of pile adopted. The performance of piled raft was
found to be almost similar in all the three bed densities. It was also found that
length and diameter of the pile had more influence on the piled raft behaviour.
The study on the prototype piled raft established that the piled raft with
smaller diameter piles and relatively thinner raft is effective in supporting
moderately loaded structure founded in predominantly sandy strata. In other
words the study established the applicability of piled raft as foundation
system in sandy soil to support moderately loaded structure.
vii
8ACKNOWLEDGEMENTS
I wish to place on record my deep and profound sense of gratitude

to my Guide Prof. K. Ilamparuthi, Ph.D., Head, Division of Soil Mechanics
and Foundation Engineering, Department of Civil Engineering, Anna
University, Chennai-600 025 for his inspiring guidance and involvement
through out the period of this work till the submission of this thesis.
I express my sincere thanks to Dr. A. M Thirumurthy, Professor and
Head, Department of Civil Engineering for providing all the departmental
facilities. The support provided by Dr. K. Premalatha, Dr. V. K. Stalin, and
Dr. Mutharam, Assistant Professors of this division is gratefully
acknowledged. The inspiring guidance an encouragement provided by
Dr. Bh. Nagabhushana Rao and Dr. K. Muthukrishnaiah, my doctoral
Committee members, and Dr. V. Ramamoorthy, Director Research, Anna
University, are acknowledged with thanks. The assistance of all the lab staff,
Anna University, is accepted with thanks. Thanks are to Shri. B.D. Mundra,
Chairman & Managing Director, and Mr. Shankar Guha, Sr.Tech. director,
M/S Simplex Infrastructure Limited, for the moral and financial support
provided during the entire period. The help provided by Mrs. Karpagam
needs a special mention. Thanks to Mr. Gokulpathija of M/s Palace
Investments who permitted me to adopt the complex for this research.
The inspiring support and untiring efforts put in by

Mr. I.V. Anirudhan, Geo-Technical solutions, in preparation of this thesis
need a special mention here and I place on record my deep sense of gratitude
to him. The hard work and inspiring efforts of my wife Mrs. B.Gomathy
during this period is acknowledged with deep sense of gratitude. The
encouragement and support provided by my sons B. Pradeep and B.Prashant
is highly appreciated. I am indebted for the blessings of my Parents but for
which this work could not have been completed.
V. Balakumar
viii
TABLE OF CONTENTS
CHAPTER NO. TITLE PAGE NO.
ABSTRACT iii
LIST OF TABLES xv
LIST OF FIGURES xix
LIST OF PLATES xxxi
LIST OF SYMBOLS AND
ABBREVIATIONS xxxii
1 INTRODUCTION 1
1.1 GENERAL 1
1.2 BASIC CONCEPT 3
1.3 AIM AND SCOPE OF THE WORKS 5
1.4 THESIS COMPOSITION 6
2 LITERATURE REVIEW 8
2.1 INTRODUCTION 8
2.2 SMALL SCALE MODEL STUDIES 9
2.2.1 1g Model Studies 10
2.2.2 Centrifuge Model 14
2.3 STUDY ON THE PROTOTYPE PILED
RAFT 17
2.4 ANALYTICAL AND NUMERICAL
MODELING 36
2.4.1 Simplified Analytical Methods 37
ix
2.4.1.1 Equivalent pier method 38

2.4.1.2 Flexibility matrix approach 39
2.4.2 Simplified Numerical Methods 41
2.4.2.1 Strip on springs approach 41
2.4.2.2 Plate on springs approach 42
2.4.3 Rigorous Numerical Methods 46
2.4.3.1 Boundary element method 47
2.4.3.2 Combined boundary element
and finite element method 48
2.4.3.3 Finite element analysis 50
2.5 SUMMARY 57
3 EXPERIMENTAL SET UP AND MATERIAL

PROPERTIES 60
3.1 INTRODUCTION 60
3.2 TEST FACILITY 61
3.2.1 Loading Frame 62
3.2.2 Test Box 64
3.2.3 Raft And Pile Models 64
3.2.3.1 Model raft 64
3.2.3.2 Model piles 65
3.2.3.3 Piled raft models 66
3.2.4 Instrumentation 69
3.3 TEST MEDIUM 69
3.4 BED CALIBRATION 72
3.5 INSTALLATION OF PILES AND PILED
RAFT 73
3.6 SEQUENCE OF TESTING 76
x
3.7 LOAD TEST PROCEDURE 77

3.7.1 Tests on Plain Raft 77
3.7.2 Test on Free Standing Pile Group 78
3.7.3 Load Tests on Piled Raft 78
4 EXPERIMENTAL RESULTS AND DISCUSSION 83

4.1 GENERAL 83
4.2 EFFECT OF PILE ROUGHNESS 84
4.3 SELECTION OF PILE ARRANGEMENT
FOR CIRCULAR PILED RAFT 85
4.4 BEHAVIOUR OF CIRCULAR PILED RAFT 87
4.4.1 Behaviour of Unpiled Raft
(Plain Raft) 87
4.4.2 Comparison of Piled Raft Behaviour
with Unpiled (Plain) Raft 90
4.4.3 Behaviour of Pile Group of Piled Raft 96
4.4.4 Load - Settlement Behaviour 100
4.4.4.1 Effect of pile length 102
4.4.4.2 Effect of pile diameter 109
4.4.4.3 Effect of raft thickness 113
4.4.4.4 Effect of pile raft area ratio 115
4.4.4.5 Variation in bed density 119
4.4.5 Settlement Reduction Behaviour 123
4.4.5.1 effect of variation in length 124
4.4.5.3 Effect of area ratio 127
4.4.6 Load Sharing Behaviour 128
4.4.6.1 αpr versus bed density 129
xi
4.4.6.2 αpr versus pile lengths 131

4.4.6.3 αpr versus pile diameter 134
4.4.6.4 αpr versus pile raft area ratio 136
4.4.7 Effect of pile configuration on sr
and αpr 138
4.5 BEHAVIOUR OF SQUARE PILED RAFT 141
4.5.1 Behaviour of Unpiled Raft 141
4.5.2 Comparison of Piled Raft Behaviour
with Unpiled Raft (Plain Raft) 145
4.5.3 Behaviour of Pile Group of Piled Raft 150
4.5.4 Load-settlement Behaviour of Square
Piled Raft 153
4.5.4.1 Effect of pile spacing 153
4.5.4.3 The effect of pile length 161
4.5.4.4 Effect of bed density 165
4.5.5.4 Effect of pile configuration 174
4.6 BEHAVIOUR OF RECTANGULAR
PILED RAFT 177
4.6.1 Behaviour of Plain Raft, Piled Raft
and Pile Group of Piled Raft 178
xii

4.7 SUMMARY 184
4.8 LIMITATIONS AND ADVANTAGES /
USE FULLNESS OF 1g MODEL TEST 186
5 NUMERICAL ANALYSES OF 1g MODEL

TESTS OF PILED RAFT 188
5.1 GENERAL 188
5.2 SLECTION OF SOFTWARE 190
5.3 VALIDATION OF NUMERICAL TOOL
ANSYS 192
5.3.1 Material Model 192
5.3.2 Linear Elastic Model 192
5.3.3 Miso Model 193
5.3.4 Loading Options 194
5.3.5 The Continuum Element 194
5.3.5.1 Plane 42 194
5.3.5.2 Solid 45 196
5.4 VALIDATION OF SOLID 45 ELEMENT
OF ANSYS 197
5.5 VALIDATION FOR PLANE STRAIN
ANALYSIS 201
5.6 ANALYSES OF CIRCULAR PILEDRAFT 204
5.6.1 3d Linear Analysis of Piled Raft 204
5.6.2 Axisymmetric Analysis (MISO) 207
5.7 PLANE STRAIN ANALYSIS FOR
RECTANGULAR PILED RAFT 213
xiii
5.7.1 3d Nonlinear Analysis (MISO Model) 216

5.7.1.1 Load- settlement behaviour 219
5.7.1.2 Load share between raft and
piles 221
5.8 ANALYSES OF SQUARE PILED RAFT 230
5.8.1 Linear Analysis 231
5.8.2 Nonlinear Analysis 233
5.8.2.1 Load- settlement behaviour 235
5.8.2.2 Load share between raft and
piles 237
5.9 SUMMARY 249
6 BEHAVIOUR OF PROTOTYPE PILED RAFT 251

6.1 GENERAL 251
6.2 PALACE REGENCY BUILDING,
CHENNAI 252
6.2.1 Soil Profiles of The Building Site 253
6.2.2 Selection of Design Parameters and
Design of Foundation System 256
6.2.3 Instrumentation of the Piled Raft and
Measurements 260
6.2.4 Load-Settlement and Load Sharing
Behaviour 263
6.3 NUMERICAL ANALYSIS 268
6.3.1 Material Property 270
6.3.2 Application of Load 271
6.3.3 Settlement of Piled Raft 271
6.3.4 Load Sharing Between the Raft
and Piles 274
xiv
6.4 SUMMARY 283
7 CONCLUSIONS 284
7.1 GENERAL 284
7.2 SMALL SCALE MODEL STUDIES 284
7.3 NUMERICAL MODELLING 288
7.4 BEHAVIOUR OF PROTO TYPE PILED
RAFT 289
7.5 SCOPE FOR FURTHER STUDY 291
APPENDIX 1: TYPICAL LCPT (LIGHT CONE

PENETRATION TEST) RESULTS 292
APPENDIX 2: MORI’S CHART 293
REFERENCES 294
LIST OF PUBLICATIONS 303
VITAE 304
xv
LIST OF TABLES
TABLE NO. TITLE PAGE NO.
2.1 Details of piled raft foundation of the buildings

in Germany (Katzanbach et al 2000) 32
3.1 Details of Model Raft 65
3.2 Dimensions of Model piles 65
3.3 Properties of sand used in the test bed 70
3.4 Strength properties of bed material at different densities 72
3.5 Method adopted for preparation of sand bed 73
3.6 Details of test done on circular piled raft 80
3.7 Details of test done on square piled raft 81
3.8 Details of test done on rectangular piled raft 82
4.1 Load corresponding to settlement of 20mm for plain
circular raft of 200mm diameter 88
4.2 Comparison of load carried by plain and piled raft
for different settlements (D = 200mm, t = 8mm,
d = 10mm, L = 160mm and N = 21) 91
4.3 Comparison of stiffness of plain and piled raft for
various settlements (D = 200mm, t = 8mm, L = 160mm
and d = 10mm, and N = 21) 94
4.4 Effect of pile length on capacity and stiffness
(N/mm) in medium dense sand (D = 200mm,
t = 8mm and N = 21) 104
4.5 The variation of stiffness (N/mm) in loose sand
(D = 200 d = 10mm Bed = Loose N = 21 t = 8mm) 105
xvi
4.6 The variation of stiffness (N/mm) (D = 200mm,

d = 10mm, Bed = Medium dense, N = 21, t = 8mm) 107
4.7 Effect of pile diameter on capacity and stiffness
(N/mm) in medium dense sand (D = 200mm t = 8mm
N = 21 and L = 160mm) 109
4.8 The variation of stiffness (N/mm) of piled raft with
piles of different diameters (D = 200m l = 160mm
Bed = Medium dense N = 21 t = 8mm) 112
4.9 Effect of raft thickness on load taken by piled raft and
stiffness (N/mm) in medium dense sand 115
4.10 Effect of pile raft area ratio on load taken by piled
raft and stiffness (N/mm) (D = 200mm, t = 8mm,
d = 10mm, L = 160mm) 117
4.11 The variation of stiffness (N/mm) 118
4.12 The variation of stiffness (N/mm) of piled
raft for different bed densities 121
4.13 Variation of load and stiffness (N/mm) of plain
square raft 143
4.14 Variation of stiffness (N/mm) of square raft for
different bed densities and raft thicknesses 145
4.15 Comparison load response between square raft and
square piled raft (B=200mm, d=10mm, L=160mm
and N=25) 147
4.16 Comparison of stiffness (N/mm) between plain
& piled raft (B=200mm, t=8mm, d=10mm, S= 4d,
L=160mm and N=25) 149
4.17 Effect of spacing on load carried and stiffness (N/mm)
(B = 200mm, t = 8mm, L = 160mm, d = 10mm) 157
xvii
4.18 Effect of pile diameter on load and stiffness

(N/mm) for the square piled raft (t = 8mm, S = 6d
and L = 160mm) 159
4.19 Effect of pile length on load taken and stiffness
(N/mm) (a) loose (b) medium dense (c) Dense
(B = 200mm sq. t = 8mm d = 10mm N = 9 S = 6d ) 163
4.20 Effect of bed density on load taken and stiffness
(N/mm) (a) 4d Spacing and (b) 6d spacing
(L = 160mm, t = 6mm & d = 10mm) 166
4.21 Details of pile, raft and bed 178
4.22 Details of pile configuration 178
4.23 Comparison of load and stiffness (N/mm) for
Plain raft and piled raft (Size = 70mm x 200mm,
d = 10mm, Ar= 3.75%, N = 6, L=160mm, t = 8mm) 179
4.24 Effect of pile-raft area ratio on load taken by the piled
raft and settlement reduction 181
4.25 Effect of pile diameter on the load taken by the
piled raft and settlement reduction 181
4.26 Effect of pile-raft area ratio on load taken by the
piled raft and settlement reduction 181
5.1 Loads Applicable in a Structural Static Analysis 195
5.2 SOLID 45 Input data 196
5.3 Iw for settlement of soil surface subjected to
uniform over square area loading 197
5.4 Analytical cases 198
5.5 Normalised settlement of loaded area 2 m x 2 m for
conditions analysed 199
xviii
5.6 Variation between ANSYS and closed form

solution for homogeneous soil 199
5.7 Effect of side boundaries on the normalised
settlement of loaded area 2 m x 2 m 200
5.8 Properties of circular piled raft 204
5.9 Properties of bed material 204
5.10 Details of piled raft and bed material 213
5.11 Properties of square piled raft and bed material 217
5.12 Properties of square piled raft 230
5.13 Properties of sand used in ANSYS analyses. 230
6.1a Observed settlements – Longitudinal section 265
6.1b Observed settlements – Transverse section 265
A 1.1 Typical results of LCPT in dense sand 292
xix
LIST OF FIGURES
FIGURE NO. TITLE PAGE NO.
1.1 General concept of piled raft 4

2.1 Configurations considered (Kim et al 2002) 12
2.2 Load ratio vs settlement (after Turek and
Katzanbach, 2003) 13
2.3 Centrifuge Package (Horikoshi 1995) 16
2.4 Burland’s Model (Burland et al 1977) 18
2.5a Messe-Torhaus, Frankfurt (Katzenbach 2000) 22
2.5b Observed time dependant load-settlement behaviour
and load sharing for Messe-Torhaus, Frankfurt
(Katzenbach 2000) 23
2.6 Instrumented foundation system for Messeturm tower
(Katzenbzch 2000) 24
2.7 Settlement of raft and load shared by piles for
Messeturm Tower 26
2.8 Load shared by piles and contact pressure variation for
Messeturm Tower (Katzanbach et al 2000) 27
2.9 Load settlement response – DG Building
(After Katzanbach 2000) 28
2.10 Treptowers building (after Katzanbach 1998) 34
2.11 Concept of Equivalent Pier (Poulos 1968) 38
2.12 HyPR Model (Clancy 1993) 43
2.13 Comparison of results from numerical methods
(Poulos 2001) 45
2.14 Plane Strain idealisation of piled raft
(Prokoso and Kulhawy 2001) 50
xx
2.15 Influence of pile-raft interaction on contact pressure

Frankfurt Clay model (Katzanbach et al 1994) 52
2.16 FE mesh of the foundation – Messeturm (Reul 2000) 54
2.17 FE mesh of the foundation – Westend (Reul 2000) 55
2.18 FE mesh of the foundation – Torhaus (Reul 2000) 56
3.1 General arrangement of experimental set up 61
3.2 Layout of piles for Circular piled raft 67
3.3 Layout of piles for Rectangular piled raft 68
3.4 Layout of piles for Square piled raft 68
3.5 Variation of angle of internal friction with unit weight 71
3.6 Typical stress - strain curve of test sand (Medium dense) 71
4.1 Effect of roughness on load - settlement response of
piled raft in dense sand 84
4.2 Layout of piles in Circular piled raft 85
4.3 Comparison of load settlement response of plain
and piled raft –Radial and square grid 86
4.4 Load - settlement response of plain circular raft of
200mm diameter 87
4.5 Characterisation of load - settlement response of
circular raft (D=200mm and t=8mm) 89
4.6 Load – settlement response of plain raft, piled raft and
pile group of piled raft in medium dense sand 90
4.7 Characteristic response of Plain raft and piled raft
(loose sand) 92
(medium sand) 93
(dense sand) 94
xxi
FIGURE NO. TITLE PAGE NO,
4.10 Characteristic response of plain raft and piled raft

(Area ratio 2.75%) 95
4.11 Comparison of load-settlement response of free
standing pile group and pile group of piled raft 96
4.12 Characteristic response of freestanding pile group
and pile group of piled raft 99
4.13 Characteristic response of plain raft and raft of piled raft 100
4.14 Load - settlement response of circular piled raft
with 10mm dia piles for various pile lengths 103
4.15 Load - settlement response of circular piled raft for
various pile lengths with piles of 8mm diameters 105
4.16 Characteristic response of piled raft for various
pile lengths 106
4.17 Non-dimensional plots for various lengths 108
4.18 Hyperbolic plots for various pile lengths 108
4.19 Load-settlement response of circular piled raft for
various pile diameters 110
4.20 Characterisation curves for various pile diameter 111
4.21 Non-dimensional plots for various diameter 112
4.22 Hyperbolic plots for various diameter 113
4.23 Load - settlement response of piled raft for various
raft thickness for pile diameter 6mm 114
4.24 Load - settlement response of piled raft for various
raft thickness for pile diameter 10mm 114
4.25 Load-settlement response of piled raft for various
radial angles (area ratio) 116
4.26 Characterisation curves for various area ratios 118
xxii
4.27 Load- settlement response of piled raft for various

bed densities 119
4.28 Characterisation curves for various bed densities 120
4.29 Non-dimensional plots for various bed densities 122
4.30 Hyperbolic plots for various bed densities 123
4.31a Settlement reduction ratio versus L/D ratio for
pile diameter 10mm 124
4.31b Settlement reduction ratio versus L/D ratio for
4.31c Settlement reduction ratio versus L/D ratio for
4.32 Variation of SR value with pile diameter 126
4.33 Variation of SR value with piled raft area ratio in
medium dense sand 127
4.34 Variation of αpr with settlement for various densities 130
4.35 αpr v/s Krs(e) based on Es 131
4.36 Settlement v/s LS ratio αPR for 10mm dia pile 132
4.37a Variation of αpr with L/D ratio, pile diameter 10mm 133
4.37b Variation of αpr with L/D ratio, pile diameter 8mm 134
4.37c Variation of αpr with L/D ratio, pile diameter 6mm 134
4.38 Settlement v/s LS ratio αpr for various pile diameter 135
4.39 Variation of αpr value with d/t ratio 136
4.40 Settlement v/s LS ratio αpr for different area ratios 137
4.41 Variation of αpr with piled raft area ratio 138
4.42 Configurations studied (with 11 piles) 138
4.43 Configuration of 21 piles studied 139
4.44 Effect of configuration on the variation of
SR and αpr with Settlement 140
xxiii
4.45 Load-settlement response of plain square raft of

8mm thick 141
4.46 Load-settlement response of square plain raft of
10mm thick 142
4.47 Load-settlement response of square plain raft of
12mm thick 142
4.48 Characterised load-settlement response of
plain square raft of 8mm thick 143
4.49 Characterised load-settlement response of
square raft of 10mm thick 144
4.50 Characterized load-settlement response of
square raft 12mm thick 144
4.51 Load-settlement response of Plain raft, piled raft and
pile group of piled raft (Square 200 x 200 mm) 146
4.52 Characterised load-settlement response of plain raft
and piled raft in medium dense sand 148
4.53 Characterised load-settlement response of plain raft
and piled raft in medium dense sand 149
4.54 Load-settlement response of pile group of piled raft
and free standing group (square shape) 150
4.55 Characterisation of load-settlement response of
free standing pile group and pile group of piled raft 152
plain raft and raft of piled raft 152
4.57 Load-settlement response of piled raft for
various spacing - 8mm thick square raft 153
piled raft with different pile spacing 155
xxiv
4.59 Non-dimensional plot for piled raft with different

pile spacing 156
4.60 Hyperbolic plot for piled raft with different pile spacing 156
pile diameters in loose sand bed 158
pile diameters in dense sand bed 159
pile diameters 160
4.64 Hyperbolic plot for piled raft with different pile diameter 161
lengths in medium dense bed 161
pile length 164
bed densities (L = 160mm) 165
4.68 Characterisation of load-settlement response of piled raft
with different bed densities 166
4.69 Variation in αpr with settlement for various spacing
(a) 8mm raft (b) 10mm raft 168
4.70a Variation of αpr with pile spacing for 8mm raft 169
4.70b Variation of αpr with pile spacing for 10mm raft 169
4.71 Variation of αpr with settlement for various
pile lengths at 6d spacing in loose sand 170
4.72 Variation in αpr with Settlement for various lengths
at 6d spacing in dense sand 171
4.73a Variation of αpr with pile length for loose bed 171
xxv
4.73b Variation of αpr with pile length for dense bed 172
4.74 Variation of αpr with settlement for various
pile diameters in loose sand bed and dense sand bed 173
4.75 Variation of αpr with d/t ratio for loose bed 173
4.76 Variation of αpr with d/t for dense bed 173
4.77 Configurations studied 174
4.78 Variation of SR and αpr with Settlement for
various configuration for medium dense sand 175
4.79 Variation of SR with pile spacing (a) t =10mm
(b) t = 8 mm 176
4.80 Variation of SR with pile diameter for different
bed densities 176
4.81 Variation of SR with pile length for different
bed densities 177
4.82 Load settlement response of raft, pile and piled raft 179
4.83 Characterisation curve of load - settlement response 179
4.84 Variation of SR with pile length for various
pile diameters 182
4.85 Variation of SR with pile diameter 183
4.86 Variation of SR with Area Ratio 183
4.87 Variation of αpr v/s Settlement for10mm dia pile for
various lengths and spacing 184
5.1 Two dimensional structural solid element (Plane 42) 195
5.2 SOLID 45 – Element Geometry 196
5.3 Finite element discretization of soil continuum 198
5.4 Finite element mesh for piled raft (plane strain analysis) 202
5.5 Reference displacement ratio v/s Pile depth 203
xxvi
FIGURE NO TITLE PAGE NO
5.6 Differential displacement ratio v/s Pile depth 203

5.7 Finite element mesh for piled raft ( linear analysis) 205
5.8 Settlement contour for the load of 7.80kN 206
5.9 Comparison of load – settlement response between
ANSYS (linear) and model test results for circular
piled raft 207
5.10 Axisymmetric model and mesh used in ANSYS analyses 209
5.11 Settlement contour for the load of 8.70kN for
circular piled raft in medium dense sand 209
5.12 Stress contour for the load of 8.70kN for the circular
piled raft 210
5.13 Comparison of load-settlement behaviour between
numerical and 1g model test data (D=200mm, t=8mm,
d=10mm and l=100mm) 211
5.14 Characterisation curves for experimental and
numerical (Axisymmetric) 211
5.15 Rectangular piled raft model with finite element mesh
used in plane strain analysis of ANSYS 214
5.16 Displacement contour for rectangular piled raft for
the load of 1.55 kN in medium sand 214
5.17 Settlement of piled raft at the load of 1.55 kN 215
5.18 Comparison of load- settlement response between
ANSYS and test data for rectangular piled raft 216
5.19 Finite element mesh of a circular piled raft
(Quarter model) used in ANSYS analysis 218
5.20 Settlement contour for a circular piled raft for
the load of 8.1 kN 219
xxvii
5.21 Comparison of load- settlement behaviour between

ANSYS and model test data (Circular Raft) 220
5.22 Characterisation curves for experimental and
numerical – 3D Circular 220
5.23 Raft contact stress at typical locations of the raft for
the load of 2.1kN (settlement =1.80mm) 222
5.24 Vertical stress at typical locations of the raft for
the load of 8.10 kN (settlement = 17.80mm) 223
5.25 Pile head stress for the load of 2.1 kN
(settlement = 1.8mm) 224
5.26 Pile Head stress for the load of 8.1 kN
(settlement =17.8mm) 224
5.27 Stresses in pile tips for the load of 2.1 kN
5.28 Stresses in pile tips for the load of 8.1 kN
5.29a Variation of stress along the shaft of typical piles
along the centre line of raft for 2.10kN 227
5.29b Variation of stress along the shaft of typical piles
along the centre line of raft for 8.10kN 227
5.30a Load distribution between raft and pile of the piled raft
at different settlement levels in terms of load 228
5.30b Load distribution between raft and pile of the piled raft
at different settlement levels in terms of load ratio 229
5.31 Comparison of settlement vs αpr for experimental
and numerical studies 230
5.32 Quarter model with finite element mesh for square
piled raft (200mm × 200mm) for linear analysis 231
xxviii
5.33 Settlement contour for the load of 6.2kN for square raft
with 4d pile spacing 232
ANSYS (linear) and on square piled raft in medium sand 233
5.35 Quarter model and finite element mesh adopted for
square piled raft in ANSYS nonlinear analysis 234
5.36 Settlement contour for the load of 8.70kN
(settlement 18.90mm) 234
ANSYS and test data for square piled raft
with 4d pile spacing 235
5.38 Comparison of load -settlement response between
ANSYS and test data for square piled raft with
6d pile spacing 236
5.39 Characterisation curves for experimental and numerical
Square piled raft, 4d spacing 236
5.40 Vertical stress in the square piled raft for the load
of 2.8kN and 4d pile spacing. 238
5.41 Vertical stress in the square piled raft with piles at
4d spacing for the load of 8.7kN. 238
5.42 Pile head stress for the load of 2.80kN (4d pile spacing) 239
5.43 Pile head stress for the load of 8.70kN (4d pile spacing) 239
5.44 Pile tip stress for the load of 2.80kN (4d pile spacing) 241
5.45 Pile tip stress for the load of 8.7kN (4d pile spacing) 241
5.46 Vertical stress in the piled raft with piles at 6d spacing
for the load of 1.8kN. 242
5.47 Vertical stress in the piled raft with piles at 6d spacing
for the load of 7.7kN 243
xxix
5.48 Pile head stress for the load of 7.70kN – 6d pile

spacing 243
5.49 Pile tip stress for the load of 7.7 kN (6d pile spacing) 244
5.50 Variation of stress over the length of pile of square
piled raft for the load of 8.7kN (No. of piles 25 at 4d
spacing) 245
5.51a Stress distribution along the pile of square piled raft
with 6d pile spacing for 1.8kN 245
5.51b Stress distribution along the pile of square piled raft
with 6d pile spacing for 7.7kN 246
5.52a Load share between raft and piles of square piled
raft (25 piles at 4d spacing) in terms of load 246
5.52b Load share between raft and piles of square piled
raft (25 piles at 4d spacing) in terms of load ratio 247
5.53 Comparison of settlement vs αpr for experimental
and numerical studies 248
5.54a Load share between raft and piles of square piled raft
with 9 piles in medium dense sand in terms of load 248
5.54b Load share between raft and piles of square piled raft
with 9 piles in medium dense sand in terms of load
ratio 249
6.1 Typical soil profile at Palace Regency site, Chennai 254
6.2 Sectional elevation with geotechnical data 255
6.3 Layout of piles and settlement markers 258
6.4 Time dependent load settlement curves 264
6.5 Percentage load taken by the raft at various stages
of construction period 267
6.6 Finite element simulation and meshing of piled raft 270
xxx
6.7 Elastic modulus of various layers 271

6.8 Settlement contour 272
6.9 Observed settlement vs computed value at
various section 273
6.10 Raft contact stress along grid P 275
6.11 Raft contact stress along grid G 278
6.12 Raft contact stress along grid B 275
6.13 Contact stress at specific points of the raft 277
6.14 Contact stress between the rows of piles and
transverse sections 279
6.15 Typical pile numbering 279
6.16 Typical head stress values 280
6.17 Typical tip stress values 280
6.18 The head load – tip load distribution with the
column load 281
A2.1 Relation between N Value and modulus of
deformation (Es) 293
xxxi
LIST OF PLATES
PLATE NO. TITLE PAGE NO.
3.1 Loading assembly 63

3.2 Test set up 63
3.3 Circular piled raft 66
3.4 Square piled raft 66
3.5 Rectangular piled raft 66
3.6 Displacement Measuring Arrangement 69
3.7 Pile installation assembly 74
3.8 Pile driving 74
3.9 Test pile installation 75
3.10 Model piled raft in position 75
6.1 Elevation of Palace Regency Building, Chennai 253
6.2 Construction of pile and raft in progress at
Palace Regency site, Chennai 260
6.3 Typical settlement marker in position 261
6.4 Settlement measurement at one of the settlement
markers 262
6.5 Settlement measurement using precision level 262
9
xxxii
10
11LIST OF SYMBOLS AND ABBREVIATIONS
αpr Load sharing ratio

φ° Angle of shearing resistance
δr Settlement ratio
σ Stress
γ Unit weight (kN/m3
γdmax Maximum dry unit weight
γdmin Minimum dry unit weight
νr Poison’s ratio of raft
νr Poisson’s ratio of the raft
νs Poisson’s ratio of the soil
νs Poison’s ratio of sand
A Area of raft
Ag Gross area of pile group
Ap Area of pile crosses section
AR Area ratio of piled raft
At Total cross sectional area of piles
B Pile diameter
B, L Breadth and length of raft respectively for
rectangular raft
Cc Coefficient of curvature
Cu Uniformity coefficient
d Diameter of pile
D Diameter of raft
D10 Effective grain size
deq Equivalent diameter
xxxiii
E Modulus of elasticity
emax Maximum void ratio
emin Minimum void ratio
Ep Pile Young’s modulus
Epa Equivalent pier (area) modulus
Epa Equivalent pier (area) modulus
Epp Equivalent pile soil pier modulus
Epp Equivalent pier pile modulus.
Eps Equivalent pier soil modulus
Er Elastic modulus of the raft
Es Elastic modulus of the soil
F Factor of safety against block failure
G Specific gravity
GASP Geotechnical Analysis of Strip with Piles
kp Overall stiffness of pile group in isolation
kr Overall stiffness of raft in isolation
Kre Raft soil stiffness ratio
L Length of the pile
Le Effective length
Lr Raft length in plane
N Number of piles
N Standard Penetration Number
np-row Number of piles in a row
P Total structural load
pi Individual pile capacity
pp Total load carried by pile group
PR Load ratio
pr Total load carried by raft
Pre Relative pier length
Pw Bearing capacity of raft
xxxiv
Qr Load on the raft at a given settlement

Qpr Load on the piled raft at the same settlement
Qp Total load on the pile at a given settlement
r Radius of the raft
rm Maximum radius of influence of each individual piles
rp Pile radius
rr Radius of raft (cap)
Spr Settlement of the piled raft at the same load considered
for the plain raft settlement
SR Settlement reduction
Sr Settlement of the plain raft at any given load
SR Settlement reduction ratio
t Raft thickness
tr Thickness of the raft
tr Thickness of the raft
wp Average settlement of pile group
wr Average settlement of raft
αrp Interaction factor of pile group on raft
ε Strain
ή Group efficiency
λp Pile slenderness ratio
1
4
5CHAPTER 1
INTRODUCTION
1.1 GENERAL
The concept in foundation design has always been tradition bound.

Tall and heavily loaded structures such as office complex, shopping complex,
residential apartments, large storage tanks, etc., are supported either on raft or
on pile foundations. When both bearing capacity and settlement requirements
are satisfied by the supporting soil mass which is near the ground, the obvious
choice of the foundation system is the raft, otherwise pile foundation becomes
the automatic choice. The main aim in introducing the pile foundation thus
becomes twofold; one to provide sufficient safety against bearing failure, and
another to eliminate the movement of the structure, otherwise called
settlement of the structure. By convention even when adequate bearing
capacity is available, if settlement poses a problem, piling is resorted to.
The traditional design of pile foundation is largely through

empirical or semi empirical approaches. When the piles pass through
particularly sandy strata, the pile design entirely depends on the observed N
value and the parameters derived based on N value. In some cases, even
dynamic formula is used and the piles are eventually terminated in the hard
strata. The ultimate load on the pile is expected to be purely from the bearing
and in many cases even the shaft friction is not accounted for, leaving this as
an additional factor of safety.
When the piles pass through clayey strata, the ambiguities

prevalent in taking the adhesion factor and computation of negative friction
make the designer resort to a highly conservative design. This is mainly due
to the complicated interactions between the constituents and the three
2
dimensional nature of the problem. The difficulties faced in sampling at site

make the evaluation of in situ properties of the strata, a questionable issue.
The above limitations have led to the permissible stress and deformation to be
within the elastic limits.
By default, the design of piles and pile group are prepared in such
a way that the entire structural load is taken by the piles and the load is
transferred to the hard strata. The presence of the pile cap, or the raft as the
case may be, and its contribution is completely ignored. Even when the cap or
the raft is placed on a ground with a reasonable bearing capacity, its
contribution in sharing the load is ignored. Even though this provides a very
high overall safety factor, invariably the number of piles provided becomes
more than what is required. The estimation of settlement in such cases
becomes a routine, as it is going to be far less than the permissible value. This
makes the foundation designers feel more confident although the design is
uneconomical.
The geotechnical design of foundation has following three steps:
1. Estimation of bearing capacity of the soil.

2. Computation of settlement.
3. Evaluation of stress in the various components of the
foundation element.
In the conventional approach of foundation design, the bearing

capacity governs the design and the settlement is completely eliminated.
Foundation design becomes economical when both the criteria of bearing
capacity and settlement are satisfied in an optimum way. It will always be
more economical when the settlement is controlled and brought to an
acceptable limit rather than altogether eliminating it. This involves the
computation of settlements and differential settlement more accurately and
evaluation of the parameters required such as Es of the soil with a reasonable
level of reliability.
3
Considerable amount of work has been done on pile and pile group
settlement over the last three decades by authors like Poulos from 1968,
although most of the earlier part of work were with pile group having the cap
not in contact with the ground. Simultaneously authors like Whitaker (1961)
and Butterfield and Banerjee (1971) had established that the contribution of
pile cap or the raft in sharing the load is significant. Subsequent attempts by
Zeavert (1957) Burland et al (1977) had established the use of piles as
settlement reducing elements with raft led to the generation of the concept of
piled raft. In this case an optimum design of foundation is evolved with a
simple raft, and piles installed below the raft to reduce the raft settlement.
1.2 BASIC CONCEPT
The basic concept in the piled raft make the piles behave more as
settlement reducing elements than as load bearing members. In these cases the
spacing of the piles can be larger resulting in the use of smaller number of
piles. The piles are expected to undergo settlements, sufficient enough to
generate their full capacity by friction.
The new concept of design is schematically shown in the

Figure 1.1. Figure 1.1a indicates the contact pressure beneath a rigid raft,
arrived based on the elastic solution. Figure 1.1 b shows a flexible raft with
uniform contact pressure, but can undergo non-uniform settlement
(i.e. differential settlement). By introduction of a small number of piles spread
over the entire raft, or over a specific area of raft, where intensity of load is
more and if the contact pressure can be converted to a uniform contact
pressure with the overall settlement within the permissible limit; then the
foundation design becomes economical.
4
Figure 1.1 General concept of piled raft (Poulos, 2001)
It is to be noted here that settlement problem is not unique for clay

alone, but is also applicable for sand. In the case of sand the permissible
settlement is lesser than clay. Hence controlling of settlement in sand also
gains paramount importance, particularly in the case of storage tanks and
slender, but heavily loaded structures. The concept of introducing pile
elements as settlement reducer is applicable for raft placed on sandy strata
also when it undergoes larger settlement. One important aspect to be noted
here is that installation of piles, particularly driven type compacts the sand
and this aspect adds further fillup to the necessity of studying the interaction
of pile-soil-raft in the sandy strata.
5
1.3 AIM AND SCOPE OF THE WORKS
The recent developments in the numerical techniques and

application softwares have facilitated the study on the complex soil-structure
interaction problems such as piled raft foundations. Real interactions can be
considered with some of the advanced techniques similar to what had been
developed by Clancy (1993). Studies have also been conducted on small scale
models and centrifuge models and the outcome had provided a lot of
information on the behavior of piled raft pertaining to load-settlement and
load sharing behavior. Field observations of the performance of piled raft are
on the rise. In spite of all the above, the piled raft is yet to become an alternate
choice among the practising geotechnical engineers. Further, most of the
studies reported are in clay, its suitability on sand needs to be studied in detail
to establish that this system can be adopted in all types of soil. Also the use of
piled raft for moderately loaded structure with relatively smaller dia piles and
thinner raft placed on an actual site conditions need to be studied. The utility
of a commonly available software in generating the numerical model by
validating the model with the data need to be established.
The present scope of the work is aimed at establishing the

applicability of piled raft on granular soil and the effect of various parameters
relating to pile, raft and the supporting strata on the overall settlement
reduction and load sharing behavior of piled raft. In order to achieve this, a
series of small scale 1g model tests were planned and conducted on circular,
square and rectangular shaped piled raft with loose, medium dense and dense
sand beds. The performance of piled raft was analysed using ANSYS finite
element code with MISO (Multi-linear ISO tropic hardening) nonlinear
material model and verified the results of numerical analyses with
experiments. Simultaneously a twelve storied building was designed with
piled raft as supporting system and was constructed; the settlement was
monitored for a period of 796 days, 360 days of construction period and 436
days after construction. The monitoring was done for a period of nearly 2
months after occupation. Through the study of these results the applicability
6
of the piled raft with smaller diameter piles and thinner raft supporting a
moderately loaded building has been examined. The behavior has been
studied adopting the ANSYS, FEA code and linear material model (elastic
behaviour) to bring out its suitability for modelling real size piled raft under
working load in layered strata. The results in the form of contact pressure and
load sharing are studied to establish the performance of piled raft for a
moderately loaded building.
The results of all the three studies are compiled to understand the
behaviour of piled raft in sand and also to that a simple concept such as
equivalent pier theory can be used for the design of piled raft.
1.4 THESIS COMPOSITION
The present thesis consists of the following seven chapters:
Chapter one presents basic concept of piled raft along with aim and
scope of the proposed research.
The work done so far and the present level of understanding on the
behaviour of piled raft has been presented in the Chapter 2. The review has
been presented under three heads namely: (i) Small scale model studies (ii)
Observational study on the proto type piled raft, and (iii) Numerical
modelling or Numerical models. The limitations of the above works have
been indicated to establish the need of the present study.
The details of the experimental set up and other facilities

developed for conducting 1g test on models are discussed in Chapter 3 The
procedure for bed preparation and installation of model piles and raft are out
lined and the test adopted producers are explained.
In the Chapter 4, the results obtained from the small scale model
tests are presented in the form of load-settlement curves and discussed by
comparing with the plain raft behaviour. The settlement reduction achieved
7
by the introduction of piles is presented for various cases; also the load shared
behaviour of the raft and the piles are presented. The load transfer mechanism
is discussed and the effect of various parameters relating to pile, raft and the
bed on load sharing and settlement reduction are brought out in this chapter.
Chapter 5 presents the details on ANSYS FEM code used in this

study to numerically model the tests conducted on the laboratory models of
piled raft foundation. Typical results of linear and nonlinear analyses are
presented and validated with the experimental results. Results of numerical
studies are analysed further and the load shared by the piles of the piled raft
for various settlement are compared with the load taken by the raft.
Chapter 6 presents the field study performed by the observations

made on an instrumented piled raft supporting a twelve storied structure. The
settlements observed are used to compute the load taken by the raft. The
results of the study conducted on the numerical model are used to establish
the contact pressure at various sections and the load shared by the raft and the
pile group.
The conclusions drawn in this study based on tests on lab models,

numerical analyses and field test results; and the scope for further study are
presented in the Chapter 7.
8
2
3CHAPTER 2
LITERATURE REVIEW
2.1 INTRODUCTION
Raft and pile are two independent foundation systems, which are
used extensively to support variety of civil engineering structures.
Understanding on the behaviour of these two foundations are established
through extensive studies on physical models (lab models), analytical models
and also through limited field studies. It is a common practice that the pile
groups are provided with pile cap (or raft). The distinction between the pile
groups with and without ground contacting cap has been brought out by few
researchers (Poulos 1968; Butterfield and Banerjee 1971; Davis and Poulos
1972) while analysing pile group behaviour.
Though the foundation engineers have long recognized the use of

piles along with raft, the contribution of raft towards load sharing and
settlement is ignored in most of the design. It is only in the recent past the
foundation engineers have realised that when competent soil is present
beneath the raft the traditional design of pile group without ground contacting
cap (or raft) is tend to become uneconomical. Even then many tall structures
with basements and storage tanks supported on piled raft foundations are
designed conventionally (i.e., without considering the ground contacting raft).
This is due to three dimensional nature of the problem, which imposes
complicated interaction between the elements of piled raft and soil. These
interactions make the quantitative analysis very difficult and expensive.
Despite the limitations, many conceptual design procedures have been
developed and improved periodically.
9
Zeavert (1957) was one of the earliest to prepare the concept of

using deep foundation elements particularly piles to reduce the raft settlement.
Such a combination of raft with piles was termed as piled raft. A notable early
example of the use of settlement-reducing piles in foundation design was
given by Zeevaert (1957) for La Azteca office building in Mexico City
constructed during 1954-57. The foundation comprised of 41m × 16m raft
with 83 concrete piles of length of 24m and diameter 0.4m. Partially
compensated friction-pile foundation was adopted in the design. The
calculated immediate raft settlement was about 200mm, compatible with the
observed settlement at the end of construction and the predicted differential
settlement was about 30mm. A comparison of observed and computed
settlements was given in an attempt to predict the future behavior of the
foundation of the building.
Numerous researchers (Poulos 1994; Padfield and Sharrock 1983;

Randolph 1983; Poulos et al 1997) have studied the behaviour of piled raft
and evolved design procedures for piled raft system. Many tall structures have
been designed and constructed based on these design procedures. A large
number of field data related to performance of tall structures supported on
piled raft have been published during the past few years. As the present study
is focused on 1g model tests and based on the monitoring of real size piled
raft of multi-storied structure, the literature on small scale model studies and
studies on prototype foundations are discussed first and analytical model
studies are discussed subsequently in the literature review. Finally the present
level of understanding on the load settlement behaviour of the piled raft and
the need for this study are outlined.
2.2 SMALL SCALE MODEL STUDIES
Even though the use of piles and raft either independently or in

combination is extensive, experimental studies on piled raft appear to be
limited. The limited studies reported in the literature based on the small scale
10
model studies relating to the behavior of piled raft are grouped under
following heads and discussed:
(i) 1g model studies, and
(ii) Studies on centrifuge models
2.2.1 1g Model Studies
Tests on small scale (1g model test) models are popular in the field
of Geotechnical Engineering. Results of model tests were effectively verified
to understand complete behavior of soil-structure interaction problems and
influence of various parameters on load sharing and settlement were also
established. Some of the classical theories were also verified experimentally,
through small scale tests. However small scale test on piled raft foundation is
very limited. The limited tests reported in literature both in clay and sand are
reviewed in this section.
Whitaker (1957 and 1961) conducted series of tests on 1g models

of free standing pile groups and pile groups with rigid caps (piled raft) resting
directly on a soil bed and examined the influence of number of piles, length of
piles and spacing between piles on bearing capacity and settlement of pile
group. The rigid cap of piled raft system showed block failure and efficiency
of pile group was higher than 100% particularly for wider pile spacing. Since
the failure was essentially block failure, a method was proposed for
estimating bearing capacity of piled raft based on it. No differential settlement
was reported since the raft used was relatively rigid.
Weisner and Brown (1978) performed tests on piled raft models

installed in overconsolidated Kaolin clay bed and demonstrated the
applicability of linear elastic continuum theory for predicting behaviour of
piled raft with both rigid and flexible rafts. In these experiments settlement of
raft and strains in the longitudinal and transverse directions of the raft were
measured at selected points using displacement transducers and semi
11
conductor strain gauges. The settlements and moments thus measured for
various vertical loads were compared. Satisfactory agreement were reported
between the experimental results and results of elastic analyses extended from
Hain (1975). Though the tests were conducted on rafts of two different shapes
and different soil-raft stiffness (both flexible and rigid rafts), the role of pile
group was not analyzed. Concentration was more on differential settlement
and variation of moment in the raft rather than loadsharing between the raft
and piles.
Cooke (1986) conducted elaborate model tests on rafts (unpiled),

free standing pile groups and piled rafts on overconsolidated clay bed. The
tests were conducted till the settlement was of the order of 1% of the width of
the raft. Cooke established that very little advantage could be obtained by
designing the piled raft with spacing lesser than 4d and also indicated that the
block behaviour occurred at even much wider spacing (i.e. 6d to 8d) than
what was being traditionally accepted for piled raft design. Hence fewer piles
also could produce considerable settlement reduction. However this
observation needs to be strengthened either through detailed model studies or
observations on real structures supported on small number of piles. Cooke
also established that the ratio of length of the pile to width of the raft
influenced the behaviour. He indicated that the length of the pile must be such
that the shaft friction should influence the settlement than the tip resistance,
meaning that the settlement reducing piles must be friction piles.
Kim et al (2002) proposed a new approach for the optimum design

of piled raft foundations. In this approach genetic algorithm without gradient
requirements was used and the analysis of piled raft foundation were
performed based on hybrid approach of Clancy (1993). In order to validate
the method proposed, small scaled laboratory tests were conducted. Three
series of tests were conducted in their study to bring out effects of stiffness of
raft, spacing between the piles and arrangement of piles. All these tests were
conducted on standard sand (Jumujin sand) with relative density of 68.3%. In
12
the first series of tests raft (180mm×180mm) with four piles (2×2) were tested
to understand effect of raft thickness on load sharing between the piles and
the raft. Tests were conducted for the raft thickness of 3mm and 8mm. It was
reported that the load sharing ratio, αpr (= ∑ Rpile/Rtot where ∑ Rpile = sum of
loads carried by piles and Rpile= total load applied) decreased gradually with
settlement ratio, ßpr (=špr /Lp, where špr = settlement of piled raft and Lp=
length of the pile). However for the ßpr higher than 0.5 in the case of raft
thickness of 3mm and higher than 0.1 in the case of 8mm thick raft, the load
sharing ratio (αpr) of the raft was greater than that of piles. The maximum load
shared by the rafts were 71% and 55% for the raft thickness of 8mm and 3mm
respectively and the corresponding ßpr was more than 1%. In the second series
of tests (2d, 3d and 4d where d = diameter of the pile) influence of spacing
between the piles was studied. Tests were conducted on 2×2 and 3×3 pile
groups with raft thickness of 3mm and 6mm. From the analyses of test results,
it was concluded that the total number of piles almost had no effect on the
variation in the load sharing ratio. The load sharing ratio of the raft increased
with thickness and size of the raft. The third series of tests were on pile
configuration. Six different configurations were considered, which are as
shown in Figure 2.1. Among the configurations tested, the second
configuration (con-2) was the most efficient arrangement wherein average
settlement was minimum and bearing capacity was maximum. Results of the
experiments were utilized to validate the genetic algorithm based optimum
design and both the results agreed well with each other.
Con-1 Con-2 Con-3 Con-4 Con-5 Con-6
Figure 2.1 Configurations considered (Kim et al 2002)

13
Turek and Katzenbach (2003) investigated the bearing behaviour of a

combined pile-raft foundation on sand. A number of small scale model test at
1g level were performed. The geometric dimensions of the model was arrived
based on typical dimensions of piled raft in Frankfurt am Main and a model
scale of 1/50 was adopted. Tests were carried out not only with piled rafts but
also with pile groups without raft-soil contact and raft foundations.
A total of 12 tests were performed, two for each foundation on loose sand and
two on dense sand. Comparison of test results of piled raft and plain raft
showed a settlement reduction of 30% in loose sand and 50% in dense sand.
However, much higher settlement was reported for the pile group. The load
shared between the bearing elements (piles and raft) of piled raft was a
function of settlement of piled raft as shown in Figure 2.2. It was also found
that the central pile of the five pile group they had used behaved slightly
stiffer than the outer piles.
Loose sand Dense sand
Figure 2.2 Load ratio vs settlement (after Turek and Katzanbach, 2003)
The study showed that at a settlement level of 4mm the pile group
mobilized its ultimate shaft resistance in the case of loose sand. The increase
in the stress level below the raft influenced the shaft resistance of pile group
more in the case of dense sand than in the case of loose sand. Also it was
found that the piles of piled raft in loose sand did not reach the ultimate shaft
resistance due to the increase in the stress level under the raft. Although the
number of tests reported was limited, the study established that the
14
performance of piled raft in settlement reduction was better in dense sand than
in loose sand.
As seen from the discussions the small scale model studies carried
out on model piled rafts were mostly on overconsolidated clay bed. The
concentration was mostly on the reduction of differential settlement and load
carried by the piles. Even though Cooke’s study produced important
conclusions on spacing of piles and length of piles, the effect of many other
parameters appeared to have not been given adequate importance. The
behaviour of piled raft on sand needs further study in detail to establish its
effectiveness under various parameters of the constituting elements of piled
raft.
2.2.2 Centrifuge Model
Centrifuge testing concerns the study of geotechnical events using

small scale models subjected to acceleration fields of many times Earth’s
gravity. With this technique, self weight measures and gravity dependent
process are correctly reported and observations from small scale models can
be related to the full scale proto type situation using well established scaling
laws. Centrifuge model tests have proved to be useful in displaying
mechanisms of deformation and collapse and in providing valuable data for
validation of numerical analyses. However very limited centrifuge tests on
piled raft foundation are reported in literature. The limited studies reported are
reviewed here.
Thaher and Jessberger (1991) investigated the effect of pile length,

pile number and pile diameter on the load bearing behaviour of piled rafts
through centrifuge model tests conducted on 1/150 scale models. The models
were tested in overconsolidated clay under centrifuge acceleration of 150g. In
their study the raft used was 15mm thick aluminium alloy plate which
represented the rigid behaviour of the raft. However the piles were placed
only on the external periphery of the rigid raft. Besides the normal basic
15
models, the foundation of an actual tower building (Messe Turm) was also
modelled in which longer piles were installed near the centre of the square
raft and shorter in the edges as suggested by Padfield and Sharrock (1983).
The important conclusions drawn by Thaher and Jessberger (1991)

were
(1) The peak contact pressure under the piled raft with regularly
spaced pile group was similar to that of conventional raft
(unpiled raft).
(2) The consolidation of clay equivalent to 13 years at proposed

scale increased the total pile load up to 10% of the structural
load.
(3) The pile spacing and the number of piles influenced the load
carried by the pile group. However the effect of length of the
piles in the group was much less than the spacing of the piles
and the number of piles.
(4) The pile spacing ratio (s/d, s-spacing and d-diameter of pile)
played a key role in the behaviour.
Horikoshi and Randolph (1996) conducted a series of centrifuge

model studies on piled raft supported on overconsolidated clay. A centrifuge
acceleration of 100g was applied to 1/100 scale models. A small pile group
was installed only beneath the central area of a flexible raft. The raft
settlement and load transferred to the pile group were monitored. A fully piled
raft designed by conventional method was also tested. The centrifuge model
test set up used by them was as shown in Figure 2.3.
Tests were conducted to establish the fact that even a small pile
group in the central area of flexible piled raft could reduce the differential
settlement to a large extent and a cap of smaller size could increase the total
bearing capacity also to a large extent, because of the load transferred through
16
the cap. An extensive parametric study was performed in order to develop a

rational design method for piled rafts. In addition to fully elastic analyses, the
centrifuge model tests were analysed using the hybrid method (HyPR), a
numerical model developed by Clancy (1993) and compared the results. The
applicability of equivalent pier theory in studying the load settlement
behaviour of piled raft was also examined by generating nondimensional pier
parameters. This study showed that piled rafts could be designed for
negligible differential settlements by introducing a pile group over the central
16-25% area of the raft and the piles could share about 40-70% of the total
load, depending on the pile group area ratio and Poisson’s ratio of the soil.
The study established the effectiveness of a small centrally located pile group
in reducing the differential settlement to a considerable extent. However the
overall settlement reduction was not given much importance. Further it was
established that the equivalent pier approach overestimated the stiffness (i.e.)
underestimated the average settlement.
a) General Plan
b) General Section
Figure 2.3 Centrifuge Package (Horikoshi 1995)

17
The degree of overestimation according to the study depended upon

the depth of the layer compared to pile length, rigidity of the piles and
spacing. The study was further extended to nonhomogeneous soil of
increasing elastic modulus and showed that the variation in results was within
10% of the results produced by other approaches.
2.3 STUDY ON THE PROTOTYPE PILED RAFT
Even though 1g model studies provided very valuable data on the

behaviour of piled raft, these are more appropriate under homogeneous soil
conditions. Observational methods, even though are time consuming, they
provide more factual results. The above facts made the engineers of those
who have involved in the design of piled raft foundation to monitor the
response of foundation during construction as well as for some period after
construction. This has helped the geotechnical engineers to understand the
mechanism of load sharing between the components of piled raft and in turn
to improve the piled raft design. The data from field monitoring, thus obtained
are published periodically. Some of them have been presented and discussed
below.
Morton and Au (1975) presented and discussed the settlement data

of eight structures in London monitored over a period of eight years. The
settlement at the end of construction period was 60% of the anticipated
maximum settlement irrespective of the type of foundation. Differential
settlements were about 25% of the maximum settlements for all the structures
examined. Finally the authors concluded that the maximum settlements of
piled raft were ⅓ of the corresponding maximum settlements of rafts.
Burland et al (1977) was one of the earliest researchers to adopt the

concept of settlement reducing pile. Figure 2.4 refers to the design concept of
Burland. The foundation system comprised of 2.0m dia 16m long bored cast
in situ piles placed under each column.
18
(a) (b)
Figure 2.4 Burland’s Model (Burland, et al 1977)
It was proposed in the study that the behaviour of such piles should
be ductile in their load-settlement response to take the raft settlement,
meaning in general the settlement reducing piles should be, floating piles or
mainly friction piles. Since the number of piles to reduce the raft settlement
would be small, the spacing of such piles would be quite larger than the
spacing used in the traditional design.
These piles were installed on 3m thick medium to dense sand with

an individual capacity of 2850kN. In his concept the pile mobilized the
ultimate shaft capacity, resulting in a reduced column load acting on the raft.
The pile loads and the raft settlements were monitored and subsequently
results of the study were further used by other designers and researchers like
Yamashita et al (1994).
Hooper (1974) reported the behaviour of piled raft supporting a

twin block for which the field measurements were taken for nearly 7 years
from 1967 to 1973. The raft was 1.52m thick located at 8.8m below the
average ground level. The piles were underreamed pile of 0.91m shaft dia and
2.44m bulb dia reinforced in the top 6m only. Even though the gross
foundation load was 368kN/m2, the net pressure was 196kN/m2, due to the
deep excavations involved. The tower was 90m tall having 31 floors. The
maximum settlement observed was 21.2mm. The numerical analysis was
19
performed by adopting axisymmetric condition and treating each concentric

row of piles as an annulus with an overall stiffness equal to the sum of
stiffness of the individual piles. It was found that in the initial stages the
applied load increased the contact pressure very slowly due to the fact that in
the initial stages the excavation induced large uplift forces. At the final stages
the ratio of distribution of load between raft and the pile was about 0.3. This
computation indicates that the long term consolidation could reduce the raft
contact pressure increasing the pile loads. The study further concluded that
most of the settlement was taken place during the construction period itself.
The observed behaviour of a piled raft foundation of stone bridge

park tower block, London was reported by Cooke et al (1981). The building is
43m high supported on 0.9m thick raft connected to 351 bored piles each
0.45m diameter and 13m long located on an almost square grid of 1.6m
spacing. From the measurements using the borehole extensometers, the
average raft settlement increased from about 10mm at the end of construction
to around 18mm after a further period of four years and precise observations
indicated that differential raft settlements were very small. Using the
computer program NAPRA, Viggiani (1998) was able to obtain good
agreement between the calculated and observed settlements. Viggiani also
showed that the number and layout of piles were significant in terms of
differential settlement. While reducing the number of piles tended to increase
the differential settlement, concentrating the piles towards the centre of the
foundation led to a marked reduction in the differential settlement. Almost
similar conclusion was drawn through centrifuge model tests, which were
conducted nearly 15 years later by Horikoshi and Randolph (1996).
Padfield and Sharrock (1983) analysed the piled raft system of an

actual building which was constructed on London clay based on conventional
design method. After confirming a good agreement between their elastic
analysis and measured data, they proposed an alternative design to reduce the
differential settlement and suggested that the shaft resistance of piles should
20
nearly be fully mobilized in the range of raft settlement. This would facilitate
the role of settlement reducing piles. Even though the pile capacities would be
nearly the maximum, the pile might not fail since the raft would have the
capability of absorbing the additional load or any load shed from the pile.
Padfield and Sharrock (1983) further proposed a stiff response in the central
area of the raft and much softer response in the edges.
An extensive review of the settlements of actual structures was

conducted by Cooke during 1986 with an intention to understand the factors
affecting the overall and differential settlements under working condition and
also to know how the structural loads were transferred to the supporting soil
by the components of foundation. He concluded that, for unpiled rafts, the
immediate settlements were found to be in the range of 0.2% to 0.3% of the
equivalent breadth of the raft, when foundations were designed
conventionally with a factor of safety 3. The settlement of structures on deep
foundations was found to be ⅓ of those for comparable structures on shallow
foundations. Almost similar observation was reported by Morton and Au
(1975) based on their studies on prototype structures. The load shared by the
raft was 30% of the structural loads even through the piles were designed to
carry full structural loads. Finally Cooke (1986) concluded that the
distribution of load between the piles of piled raft foundation was influenced
by the stiffness of the structure-foundation system and the structural loading.
However for highly rigid structures, the load distribution between the piles
depends essentially on the number of piles and their spacing.
A coal silo of 11.6m supported on a circular piled raft was

monitored to study the behaviour of piled raft embedded in a soft ground
(Kulhawy et al 1987). The thickness of raft was 0.6m; it was supported by
five pile group, with an annular ring beam of 1m wide and 1.3m thickness
provided at the bottom of the raft. One pile was placed at the centre of the raft
and the remaining four piles were located at a radial distance of 4.3m from the
centre of the raft. From the observations reported, it was found that the safety
21
factor of piles against bearing failure found to vary between 1.6 and 1.8.
Further good agreements between observed results and the results of elastic
analysis were reported.
Franke (1991) from the review of the published results of three

major buildings namely Torhause, Messe Turm and West End Street Tower I,
provided certain guidelines for the design. A brief description of each
structure and their performance are given below:
The Messe-Torhause (Figure 2.5a) building which is a 30 storied

structure was supported on twin piled raft system measuring 17.5m×24.5m
each (Figure 2.5b). The rafts were located at a depth of 3m and each one was
supported on 42 piles of 20m long and 500mm dia each. The raft thickness
was 2.5m. Each raft was loaded with an effective load of 200MN. Being the
first building on the piled raft, this was designed conventionally.
The piles were assured to be utilized to its bearing capacity, while

the remaining part of load would have to be taken by the raft and transmitted
directly to the soil. The observed settlement was 150 mm over a period of 4
years and the load sharing ratio by the pile group, αpr was 0.8 (Figure 2.5b).
The maximum load shared by the raft was 20% of the total load. This piled
raft is an example for first generation design.
22
(a) Isometric view (b) Site plan
(c) Instrumentation in piled raft
(a)
Figure 2.5 (a) Messe-Torhaus, Frankfurt (Katzenbach, 2000)

23
(b)
Figure 2.5(b) Observed time-dependent load –settlement behaviour and

load sharing for Messe-Torhaus, Frankfurt (Katzenbach,
2000)
24
The 256.5m tall Messeturm tower has a basement with two under
ground floors of each 58.8m2 in plan and a 60 storey core shaft of size
41m×41m (Figure 2.6). The estimated total load was 1880MN. The raft was
located at a depth of 14m below the ground level in a deposit of gravel and
sand of 8m thick followed by Frankfurt clay of 100m thick or more. The raft
was 6m thick at the centre and reduced to 3m at the edges. The bored cast in-
situ piles were of 1.3m diameter arranged in 3 concentric circles below the
raft. The length of piles in the outer middle and inner circles were 26.9m,
30.0m and 34.9m respectively.
(a) Plan and cross-section (b) location of instrumentation
Figure 2.6 Instrumented foundation system for Messeturm tower

(Katzenbach, 2000)
The design was done by assuming a load sharing ratio. The piled
raft was proportioned for the following two cases:
25
(i) In the first case, the piles were assumed to carry only 30% of
the structural load and the remaining load carried by the raft
via contact pressure.
(ii) In the second case, the piles were assumed to carry 55% of
the load and the remaining load on the raft
[
Further it was stated that the foundation system was instrumented

elaborately (Figure 2.6) to measure settlements, contact pressures and load on
piles. These instruments were monitored and measurements were made
regularly over a period of about four years including the construction period
of one and a half years.
The settlement of raft, load shared by the piles, and variation of

contact pressure on raft with time are presented in Figures 2.7 and 2.8. The
settlement of the raft at the end of construction was 85mm at the center and
48mm at the edges. The load shared by the piles at the end of construction
was 55%. The average contact pressure was 160kN/m2 and it was reported
that the measured contact pressure at the centre of raft was about 20% higher
than at the edges. Finally it was concluded that, the assumption made in the
design was not complied with the field observation.
The West end tower is also known as DG bank building which was
constructed during the period between 1990 and1993. The structure is 53
storied 208m tall office building in L shaped form. The total structural load
was 1420MN. This load was transferred to Frankfurt clay through a piled raft
of raft thickness of 3 to 4.5m and 40 bored piles of 1.3m dia with a constant
length of 30m. The raft was founded at a depth of 14.5m which is 9.5m below
the water table. As stated in the earlier cases, this structure was also
instrumented and monitored. The typical variation of load-settlement response
was reported as presented in Figure 2.9. In this case the load was equally
shared by the piles and raft.
26
Figure 2.7 Settlement of raft and load shared by piles for Messeturm
Tower (After Katzenbach et al 2000)
27
Figure 2.8 Load shared by piles and contact pressure variation for
Messeturm Tower, (Katzenbach et al 2000)
28
Figure 2.9 Load settlement response -DG Building (After Katzenbach,

2000)
From the performance of piled raft foundation of three tall

structures, Frank (1991) concluded that the raft contact pressure would
increase only when the base resistance of the piles was relatively smaller
when compared to the shaft friction and also suggested that a skill full layout
of piles would reduce the raft bending moment and the internal stresses in the
raft. The study further concluded that a balancing of load share must be done
in such a way that the raft and piles share the load equally for a given
settlement. This was achieved in the West end tower. This would warrant a
repeated analyses to fix the length of the pile after selecting the diameter.
Frank (1991) further gave more importance for the load shared by the raft and
indicated that finite element and boundary element methods must take into
account the bilinear elastic / plastic shaft resistance behaviour, and simple
design calculation must be developed for the design office. Poulos (1991)
used the computer program GASP (Geotechnical Analysis of Strip with Piles)
and analysed various foundation options for the printing press building,
Sydney. As a result of these analyses, it was decided to use a strip foundation
29
of 5m wide and 0.8m thick, with pairs of bored piles of 0.9m diameter,
typically extending about 7.5m into moderately weathered shale.
Schwab et al (1991) used electronic devices for monitoring the

performance of raft and measured raft contact pressure, settlement, pile head
and tip loads. According to them the measurement of deformation below the
raft has to be done for the assurance of structural safety and serve as the basis
of design for future projects.
Yamashita et al (1994) instrumented and monitored a five storied

structure support on piled raft in Urava, Japan and analyzed numerically. Here
the piles were steel “H” piles embedded in predrilled holes and grouted. The
raft was 300mm thick and located at 2.4m below the ground level. The
foundation load was 47.5MN. The settlement was monitored for a period of
300 days and the magnitude of settlement was 12mm at the end of 300 days.
The measured pile load in the corner piles was less than the peripheral piles
contrary to the general opinion and also the view expressed by Cooke et al
(1981). According to the study, this was due to a certain amount of load
transferred through the edge beam of the raft to the soil directly, thus reducing
the load carried by the pile. In the numerical model, the raft was modelled as
plate and beam elements and the piles were idealized as springs. A portion of
the load was applied as concentrated load on the column location and the
balance load as uniformly distributed load on the tributary area of the raft.
However no specific mention about the magnitude of the load ratio was
specified. The results were in agreement when the load transfer through the
sides of the foundation was considered.
The growing demand for office space in Frankfurt, Germany and

the prevailing soil conditions of this area lead to the development of tall
structures and cost saving foundation system. While the conventional design
lead to choose deep piles, the concept of piled raft gained importance as an
economical alternative foundation system for tall buildings. Numerous
structures were built in Germany on piled raft foundations since 1983. The
30
first few German experiences with piled raft foundations started in the over-
consolidated clay with the construction of high-rise Messe-Torhaus building.
The second piled raft structure was the Messeturm building in Frankfurt,
illustrated the advantages of this foundation and lead to the acceptance of the
concept of piled raft foundation by geotechnical experts. Following this
concept several other skyscrapers were built on piled raft and monitored.
More recently for some of the high-rise buildings on loose sand in Berlin,
piled raft foundation was adopted and monitored. It was referred that piled
raft performed efficiently in sand and reduced the settlement of structures
appreciably. The experience gained in each building was efficiently utilized to
improve the design of piled raft. The improved design methods offer the
possibility for an optimal design of piled raft foundations both in clay and
sand.
In order to refine the design concept of piled raft keeping in mind

its extensive use, Katzenbach and his team have been researching
continuously on the piled raft foundation for more than 15 years (Katzenbach
1993; Katzenbach and Reul, 1997; Katzenbach and Moormann 1997;
Katzenbach et al, 1998, 2000 & 2002 etc.) and validated their analyses results
with the field performance data of piled raft foundations of several buildings
in clay and sand. In all their works, the complex soil- structure interaction
(soil-pile, pile-pile, soil-raft and pile-raft) was modelled numerically and the
influence of each foundation element (i.e., soil, pile and raft) over the other
was demonstrated and discussed. Finally, in the year 2000, they came out
with recommendations for the design and construction of piled rafts wherein
safety concept was also included. In book on “Design Application of Raft
Foundation” edited by Helmsly (2000), Katzenbach and his research team
contributed a chapter on piled raft foundation projects in Germany. In this
chapter, the authors discussed about the concept of piled raft foundation,
numerical analyses of piled raft foundation, design considerations of piled
raft, safety concept of piled raft and piled rafts in Germany. In the chapter on
piled rafts in Germany, the piled raft projects already built or in progress in
31
Frankfurt am Main during 1983 to 2001 were well documented. The authors
classified them as first, second and third generation piled rafts. The first
generation projects were Messe-Torhaus and Messeturm buildings for which
conventional methods combined with engineering judgment were used. Piled
raft foundations of the second third generation were designed using
increasingly improved and verified calculation methods, based on analytical
models or three dimensional finite element analysis with elasto - plastic
constitutive laws for the soil. In Table 2.1 details of piled raft foundation of
all the three generations including their performance are presented as reported
by Katzenbach et al (2000).
From the table presented it can be seen that piled rafts provided for
various structures composed of thick raft (maximum thickness 6m) with large
diameter bored piles. Pile diameters were found to vary between 0.9m and
1.5m. These piles were normally spread at 3d (d-diameter of pile) distance.
However maximum spacing adopted was 6d. Pile lengths provided were
found to vary between 20m and 35m and ensured that the piles were sharing
the load mostly through frictional resistance. The settlement and load sharing
ratio, αpr of piled raft foundation were also reported. The settlement reported
was found to vary between 32mm and 150mm and the minimum and
maximum load sharing ratios reported were 0.3 and 0.8 respectively.
Table 2.1 Details of piled raft foundation of the buildings in Germany (Katzanbach,et.al., 2000)
32
33
The settlement and load-sharing ratio of piled raft foundation of

Eurotheum tower of 110m tall with total load 500MN were 32mm and 0.3
respectively. This structure was constructed in 1997 in Frankfurt clay of
Frankfurt am Main and it was known as structure on piled raft of third
generation. In the same area a structure of 130m tall with 400MN (Messe-
Torhaus building) was constructed in the year 1983. The settlement and αpr
values recorded on its piled raft foundation were 150mm and 0.8 respectively.
Despite the foundations of both the structures were on almost

identical soil deposits and on piled raft, the structure with higher load
recorded very less settlement, which was 1/5th of the settlement of first
generation piled raft. Further the raft of Eurotheum tower shared 70% of the
total load while Torhaus raft resisted only 30% of its total load. This is a good
example for the improvement on the design concept of piled raft foundation
of first and third generation.
Katzenbach et al (1998) have reported the performance of some of

the recent structures built on sand. For example in Berlin, the 121m tall
Treptowers building was supported on piled raft foundation to transmit part of
total building load of 670MN through the loose sand below the raft to dense
sand at deeper depth. Thickness of the raft provided to the structure was found
to vary between 2m and 3m. The raft of least thickness was provided in the
area of elevator.
The raft was supported on 54 bored piles of 0.9m diameter each.

The length of piles were varied between 12.5m and 16m to match with the
founding level of the raft. However all the piles were terminated at the same
level, which is as shown in Figure 2.10a.
In the Figure 2.10b the settlement of piled raft foundation measured

over the construction period of three years was shown. The settlement
reported was 65mm at the end of construction. Further the authors have
reported the results of numerical analysis of piled raft foundation and
34
compared with results of field measurements. The measured and calculated

load-settlement behaviour of piled raft was in good agreement. The calculated
settlement of piled raft foundation was 57% of the settlement of shallow
foundation arrived through numerical simulation.
(a) plan and cross section of piled raft foundation
(b) measured time-dependent settlements
Figure 2.10 Treptowers building (after Katzenbach, 1998)
The 103m high Sony - Centre building is a part of new construction

at the Potsdamer Platz in Berlin. The plan area of building is 2600m2. It was
supported on a piled raft comprising of 44 bored piles of 1.5m diameter and
lengths of piles of 20-25m and raft of 1.2 to 2.5m thick. The foundation
35
underwent a settlement of 20mm at the centre of the raft on construction of 18

of the 26 storey.
In 1998 the new office building for Deutche Messe AG was

constructed in Hannover, where the deposit is multi - layered system consists
of semi stiff to stiff glacial drift of 12-15m thick followed by a 2-4m thick
layer of fine sand which is under lain by dense sand and gravel with a
thickness of 5-7m. Then stiff clay and till is encountered, which is underlain
by marly clay rock at a depth of 30m. The structure constructed was 68m
high office tower over an area of 24m×24m. The foundation for this structure
was built as a piled raft with 32 bored piles, each 1.2m diameter and length
17.5m. The performance of the building was monitored over a period and the
maximum settlement at the centre of raft was 14mm. The settlement at the
corners of stair case towers was 10mm. From the performance of tower it was
concluded that the concept of piled raft can also be applied successfully in
multi-layered and non-homogeneous deposits in order to reduce total and
differential settlements.
Reul and Randolph (2003) presented a comparison of in situ

measurements and numerical analyses for three piled raft foundations on
over-consolidated clay. Comparison of overall settlement, differential
settlement and load carried by the piles showed reasonably good agreement
even though numerical methods show higher pile loads than predicted from
the field measurements. Three different structures studied were Westend,
Messeturm and Torhaus. The features of the buildings are presented
elsewhere (Table 2.1).
The design of Canary Wharf office tower, England was discussed

briefly by Jones (1998) and in more detail by Bergmann and Campbell
(1991). It is a building of 236m high 50 storey steel framed office tower
supported on a 65m square raft of 4m thick connected to 222 concrete bored
piles, each 1.5m in diameter. The piled raft had a substantial end bearing
component, thus reduced the load shared by the raft.
36
Hong et al (1997) reported the behaviour of a piled cellular raft

foundation of 24 storey office building constructed at Shanghai, China. The
building is supported on 1.5m thick cellular raft. The raft is supported on 233
pre-cast concrete driven piles of 0.45m × 0.45m size. Building settlements
were measured during the construction period of two years. At the end of
construction, the average measured settlement of 31mm reflected the presence
of end bearing piles.
A number of field data reported in literature on the performance of

piled raft were on overconsolidated clays and surprisingly major part of the
research was on the Frankfurt clay. In most of the cases, the rafts provided
were 3m thick or more and seated at deeper depths (around 10m depth). Early
days piled raft designs followed adhoc methods and data reported were not
adequately validated. Very few cases only presented with detailed analytical
interpretations.
There seems to be paucity of published data on moderately loaded

and flexible piled raft. Performance of piled raft with flexible raft and slender
piles (smaller diameter) is also scarce. Very little is discussed in the literature
on pile arrangements in real field conditions.
2.4 ANALYTICAL AND NUMERICAL MODELING
The design of any three dimensional foundation systems including

complicated interactions among the constituting elements can be done only
through a powerful numerical model which will reasonably represent the field
conditions. Piled raft foundation is one such complex soil structure interaction
problem which requires rigorous analysis for understanding the mechanism of
piled raft and for the reliable design. Moreover the use of this foundation for
supporting high-rise buildings is in increasing recognition for the past few
years, which has forced the geotechnical engineers to develop a suitable
method for the optimal design of piled raft. This was recognized long back
and the Technical committee (TC 18) of the International Society had focused
37
its attention since 1994 towards piled raft foundation and has collected
considerable information on methods of analysis and design including case
histories.
Van Impe and Lungue (1996) and O’Neill et al (1996) have

prepared the comprehensive reports on several activities of piled raft. A
treatise on numerical modeling of piled raft foundation has also been
presented by El-Mossallamy and Franke in the year 1997. Apart from this, the
concept of piled raft foundation had been described by numerous researchers
including Zeevart (1957), Davis and Poulos (1972), Hooper (1974), Burland
et al (1984), Randolph (1994), Sommer et al (1985), Price and Wardle (1986),
Franke et al (1994), Katzenbach et al (1998) and Reul (2000) among many
others. Various methods of analysis of piled raft reported in literature have
been presented under following heads.
(i) Simplified analytical methods.
(ii) Simplified numerical methods.
(iii) Rigorous numerical methods
2.4.1 Simplified Analytical Methods
The simple analytical methods, though were not directly aimed to

handle piled raft, they were aiming at solving the complicated pile group
settlement analysis. In such cases certain approximations were done to make
the computational procedure simpler. Subsequent development on
computational facilities made the researchers to adopt alternate numerical
modeling methods to solve the same problem. Equivalent pier and equivalent
raft theory fall in the first category; whereas stiffness matrix method falls in
the second category. However the equivalent raft method is meant only for
pile group settlement where the raft would not be in contact with the ground.
38
2.4.1.1 Equivalent pier method
Poulos (1968) established that it might be useful to replace the pile

group by an equivalent pier as shown in Figure 2.11 of the same gross area as
in the group. The length Le of the pier may be obtained by comparing the
solutions for the settlement of pile group with the settlement of a single pier
(Davies and Poulos 1972) and evaluating the required value of Le in terms of
L to give equal settlement of the pier and the group. He further established
that the Le / L is influenced more by pile spacing than by the number of piles
and L/d ratio.
Figure 2.11 Concept of Equivalent Pier ( Poulos, 1968)
Extending this, Poulos and Davis (1980) presented two methods;

(i) An equivalent single pier of the same circumscribed plan area of the group
(Figure 2.11) and of some equivalent length Le, and (ii) An equivalent single
pier of same length L as the piles but having an equivalent diameter deq.
While the second method is good in computing the overall average settlement,
this does not offer any solution for differential settlement. The equivalent pier
approach of Poulos and Davis (1980) was further analysed by Poulos (1993).
39
In this approach Poulos (1993) replaced a portion of pile group by an

equivalent pier and the stiffness of pier was predicted based on the elastic
solution of single compressible pile proposed by Poulos and Davis (1980).
The diameter of equivalent pier for friction and bearing piles was worked
using following relations.
For friction piles

4
d eq = Ag (2.1)
π
For bearing piles

2
d eq = Ag (2.2)
π
where Ag is the plan area of the pile group as a block. These two equations
are derived using total peripheral length of the block and plan area of the
block respectively. Randolph (1994) reported that Equation 2.1 might offer a
more accurate stiffness even for a friction pile group.
2.4.1.2 Flexibility matrix approach
Randolph (1983) proposed an approximate method for the analysis

of piled raft foundation. In the analysis two interaction parameters were
introduced and related to settlement of pile group and raft as indicated in
equation 2.3.
(2.3)
wp : Average settlement of pile group

wr : Average settlement of raft
Pp : Total load carried by pile group
Pr : Total load carried by raft
kp : Overall stiffness of pile group in isolation
kr : Overall stiffness of raft in isolation
αrp : Interaction factor of pile group on raft
αpr : Interaction factor of raft on pile group
40
From the reciprocal theorem, the interaction factors were related as
kr
α pr = α rp (2.4)
kp
Since the average settlement of the piles and raft were assumed
equal and from the above equations, the stiffness of the piled raft the
proportion of the load taken by the raft were given by following expressions
(2.5)
(2.6)
(2.7)
rr : Radius of raft / pile cap

rp : Pile radius
rm : Maximum radius of influence of each individual pile
Randolph compared these results with the results obtained by

Poulos and Davis (1980) and found them to be in good agreement. Clancy and
Randolph (1992) and Clancy (1993) found that the value of αrp approached 0.8
as the number of piles increased. Clancy and Randolph (1993) introduced a
parameter known as aspect ratio (R), which is expressed as √ns/Lp where, n is
the number of piles in the group, s is the spacing between the piles and Lp is
the length of the piles. This parameter is used to understand the behaviour of
pile group. It was reported that for an aspect ratio greater than 4, the
behaviour of pile group (group settlement pattern) was similar to that of a raft
foundation. Therefore an equivalent raft approach is ideal choice for analysis.
For values of R less than 2, it was recommended to adopt equivalent pier
approach, at least for estimating average settlement of pile group.
41
2.4.2 Simplified Numerical Methods
In this method of analysis, the soil- structure interaction between

the foundation element (raft or strip foundation) and the soil was
accomplished by modeling soil continuum as a series of spring elements
(Winkler model) and the foundation as plate or strip element. This also was
known as simplified computer based analysis. It is an approximate method,
because of inadequate representation of continuum response of soil. More
over infinite extent of semi-infinite medium is also not represented by this
model. Despite the limitations, this method is widely used in the design of raft
and piled raft foundations. Based on the idealization of foundation element,
this method has been identified as
(i) Strip on springs and
(ii) Plate on springs.
2.4.2.1 Strip on springs approach
A typical method in this group was that presented by Poulos (1991)

in which a section of the raft was represented by a strip and the supporting
piles by springs. Approximate allowance was made for all four components of
interaction and the effect of the parts of the raft outside the section was
accounted by computing free-field soil movements due to these parts. It has
been shown that this method agreed reasonably with the results of more
rigorous analysis. However this technique has following limitations:
(i) It cannot consider torsional resistance of the raft, and
(ii) Settlements at a point may not be consistent, if strip in two

directions through that point are analyzed.
Brown and Wiesner (1975) and Wiesner and Brown (1976) have
developed almost identical method of analysis as explained above for applied
strip and suggested the way to extend this analysis to a piled raft.
42
2.4.2.2 Plate on springs approach
In this method, the raft is modelled as a plate, piles are modelled as

interacting springs and the soil is represented by an elastic continuum (Poulos,
1994). In certain cases the soil is represented by springs and the pile as an one
– dimensional element (Clancy and Randolph, 1993). Poulos (1994) presented
an approximate analysis of piled raft in which the raft was modelled as a thin
plate and the piles as interacting springs of appropriate stiffness. Allowance
was made for the development of bearing failure below the raft, the piles
reaching their ultimate capacity and free - field soil settlements. This method
adopted boundary element approach as the basis and was similar to load cut
off procedure explained by Hain and Lee (1978). Here, from the first set of
analysis, in the nodes where the contact pressure exceeded the limiting value,
the limiting value was applied on them; the analysis was repeated to obtain
the constant value not exceeding the limiting value. The analysis was
implemented through a package GARP and was compared with other
published results. This method of numerical analysis involves a trial and error
procedure and assumes limiting values of pile capacity and raft, hence has
limitations for practical design.
O’Neill et al (1977) developed the so called the hybrid approach for

the analysis of offshore pile group by modifying the boundary element
approach. A load transfer analysis was used to compute the individual pile
behaviour and Mindlin’s theory was used for the soil to consider the influence
of the adjacent piles. Chow (1986) presented a hybrid analysis for linear and
nonlinear response of the pile group subjected to vertical loads. The procedure
employed the method developed by Randolph and Wroth (1978) to arrive at
the nonlinear t-z relation with Mindlin solution for pile- soil and soil- pile
interaction to be used for the formation of stiffness matrix. Chow (1987 b)
further extended his work for the pile group with ground contacting cap. The
soil non- homogeneity was considered by using finite element analysis.
43
Chow and Teh (1991) used Chow’s method to study the piled raft
on non-homogeneous soil with a finite depth. Here the raft was assumed rigid
raft and modelled as sub-elements and the piles were modelled as two nodded
one dimensional elements with an axial node of deformation. Finally they
reported that the effect of rigid raft contacting the ground had a little influence
on the stiffness of the group compared with the free standing pile group. The
analysis based on the non-homogeneous soil condition resulted very low load
on piles.
Clancy and Randolph (1993) extended the hybrid model generated

by Chow (1986) and Chow and Teh (1991) in which the pile group and the
interaction between piles and the raft element were based on Mindlin’s
solution. But they have used the load transfer model of Randolph and Wroth
(1979) for the single pile behaviour.
Figure 2.12 shows the model developed by Clancy (1993) wherein

idealization of different elements of piled raft foundation is presented. The
method allowed for slip along the pile shaft. The model was named as HyPR
(Hybrid piled raft analysis).
1. One dimensional pile element

2. Lumped soil response at each pile node: Load transfer
springs
3. Two dimensional plate bending finite element raft mesh
4. Lumped soil response at each raft node: Direct solution
5. Lumped soil response at each rigid disc node
6. Pile-soil-pile interaction effects calculated between pairs
of nodes - Mindlin’s Equation
7. Raft soil-raft-interaction
8. Rigid disc-soil-rigid disc interaction
9. Pile-soil-raft interaction
10. Pile-soil-rigid disc interaction
11. Raft-soil-rigid disc interaction
Figure 2.12 HyPR Model (Clancy, 1993)

44
This model was further used by Yamashita et al (1998) in

validating their observations made on a prototype piled raft of a five storied
building. The only difference in their approach was both the soil and piles
were represented by appropriate springs. Russo (1998) and Russo and
Viggiani (1998) described a method wherein various interactions (soil-pile-
raft) were obtained from elastic theory and non-linear response of piles was
modelled using hyperbolic load settlement relation. Interaction between the
piles was applied only to the elastic settlement of pile, while non-linear
component of settlement of pile was assumed to arise only from the load of
the particular pile. However this model can be used only for linear elastic
analysis. Since the soil-pile interaction is represented through springs, the
method becomes very sensitive to soil modulus (Es). Poulos (1994)
formulated a numerical procedure for the analysis of piled raft wherein raft is
modelled as a thin plate and piles are modelled as springs. The method
permits the analysis to limit the base pressure of the raft and ultimate
capacity. Poulos (1998) used GARP program for the analysis of piled raft and
concluded that increasing the number of piles while generally benefit, did not
always produce the best foundation performance. But no quantitative
indication or limit was given for the number of piles to be used. The raft
thickness influenced the bending moment and differential settlement of raft
and not on load sharing or maximum settlement. Finally the author concluded
that the piled raft designed by utilizing the full capacity of individual piles
was economical.
Poulos et al (1997) reviewed a number of available methods for

analyzing piled raft behaviour and applied few methods to analyse a
hypothetical problem involving three different condition of piled raft viz.,
(i) a raft with 15 piles and a total load of 12MN, (ii) a raft with 15 piles and a
total load of 15MN (iii) a raft with 9 piles and a total load of 12MN and an
actual case history involving a tall building (Westend building) built in
Frankfurt. The main objective of the analysis was to examine the capability of
each method in predicting the behaviour of piled raft. Six different methods
45
have been used. They were Poulos and Davis (1980), Randolph (1983), strips
on springs (GASP), plate on springs (GARP), finite element method of Ta and
Small (1996) and finite element and boundary element of Sinha (1996).
The results of average settlement, differential settlement, maximum

bending moment and proportion of load carried by piles obtained from the
6 methods for all the cases chosen for this study were compared
independently. The results of analyses of Westend building from the six
methods considered were compared in Figure 2.13.
(a) Comparison of settlement (b) Comparison of % Pile
(c) Pile load in kN (d) Comparison of Pile load in
Figure 2.13 Comparison of results from numerical methods (Poulos,

2001)
From the results compared, Poulos et al (1997) concluded the following:

(a) The differences in the results of analyses between six
different methods were significant.
(b) At lower loads on the raft, the axial capacity of pile was not
utilized fully thus the average settlement predicted by the six
46
methods was similar. When the pile capacity of a significant

number of piles was fully mobilized, there was a significant
variation in the predicted behaviour of the piled raft
foundation.
(c) There was a considerable difference between the computed

maximum bending moments of the raft depending on the
method of analysis employed.
Further Poulos et al (1997) concluded that though the methods

appear to provide reasonable basis for estimating overall behaviour of a piled
raft foundations, there appears to be broad scope for a further study to
improve the methods to predict the entire load settlement behaviour of piled
raft reliabily.
2.4.3 Rigorous Numerical Methods
The rapid developments in computing systems and powerful

numerical methods have encouraged the researchers to develop various
numerical methods to study the load-settlement and load sharing behaviour of
piled raft. The various numerical approaches can be grouped as detailed
below:
(a) Boundary element method (Butterfield and Banerjee 1971;

Kuwabara 1989; Sinha 1997).
(b) Methods combining boundary element for the piles and

finite element analysis for the raft (Hain and Lee 1978;
Frank et al 1994; Ta and Small 1996).
(c) Simplified finite element analysis adopting a plane strain

approach or axi symmetric idealization (Hooper 1974;
Prokoso and Kulhawy 2001).
47
(d) Detailed finite element analysis involving three dimensional

modeling (Zhang et al 1991; Lee, 1993; Gandhi and Maharaj
(1996); Maharaj 1996; Katzenbach et al 1998; Reul 2000).
(e) Methods employing strip on spring approach with a series of

strips representing the raft and springs of appropriate
stiffness representing the piles (Poulos 1991).
(f) Plate on spring approach in which the raft is represented by a

plate and piles by springs (Clancy and Randolph 1993;
Poulos 1994; Russo and Viggianni 1998; Yamashita et al
1998).
2.4.3.1 Boundary element method
Butterfield and Banerjee (1971) studied two problems considering

the interaction with the supporting ground of an arbitrarily spaced group of
piles with a ground contacting pile cap. They studied the load- displacement
behaviour and load sharing behaviour. In the analysis they adopted Mindlin’s
solution for a point load in an elastic half space. The parameters considered in
the analysis were lengths to diameter ratio of pile, pile cap size,
compressibility of the pile and the soil medium. They found that the floating
and ground contacting caps behave differently. The ground contacting cap
exhibited 5 to 15% higher stiffness than the floating cap, which is the function
of group size and spacing. The proportion of the load taken by the ground
contacting cap varied from 20% to 60% depending on the group size and pile
spacing. However no specific mention about the settlement behaviour or
differential settlement was observed. This study was restricted only to small
group of 2 and 4 piles.
Kuwabara (1989) performed boundary element analysis based on

elastic theory to study the settlement and load transfer behaviour of piled raft.
The soil was treated as homogeneous elastic and isotropic half space medium.
The behaviour of the pile group of piled raft was compared with free standing
pile group and termed the length beyond which there was no further decrease
48
in the settlement as critical length. Kuwabara’s analyses showed that 20% to

40% of the load was shared by the raft resting on the soil directly and
concluded further that the critical length of the pile group was larger than the
single pile when the dimensions were the same as the piles in the group.
Accordingly the critical length would not get affected by the spacing of the
pile so long as the spacing was restricted to 10 times the pile diameter. The
author also found that the ratio of long term settlement to total settlement of
pile group was higher than that of single piles but, the presence of raft did not
alter the situation except for short pile groups. However, the study was
restricted to smaller square pile groups.
Sinha and Poulos (1997) studied the effect of ground movement

caused by the expansive soil on the behavior of piled raft considering two
types of raft namely totally flexible and rigid. The raft was idealized as elastic
plate, the pile as circular element each acted upon by shear stress and circular
ring having uniform stress. From the study it was established that the
movement caused redistribution of stresses and resulted higher load on piles.
But swelling caused a reduction in the load taken by the piles.
2.4.3.2 Combined boundary element and finite element method
Hain and Lee (1976) in their study on raft pile system, considered
the piled raft system as a flexible elastic plate supported on a group of
compressible pile and the supporting soil was represented as an elastic
homogeneous or non - homogeneous material. In order to consider the
ultimate capacity of piles, Hain and Lee (1976) adopted a load cut off
procedure, wherein the piles that were subjected to reactions more than their
capacity were deleted from the compatibility equations and their loads were
held constant in the equilibrium equations. Hence, any excess load on that pile
got redistributed to the adjacent piles. This resulted in an analysis which took
the finite load capacity of the pile into account.
49
Hain and Lee’s analysis incorporated two case studies

(La Azatec and Cavalry Barracks buildings). In the analytical procedure
applied to the building La Azatec, it was found that only 17% of the load was
taken by the raft. Around 40% of the piles carried only 0.1MN against the
design load of 0.26MN, 8% of the piles exceeded the ultimate load of
0.69MN. The measured differential settlement after 20 months was only
30mm, but the data available was insufficient to make a comparison of long
term predicted and measured settlement. In the case of the second building
namely Cavalry Barracks, it was shown that practically all the loads were
carried by the piles because of the closer spacing. The paper however
indicated the importance of the use of appropriate value for the modulus of
variation of soil. It was predicted that the use of homogeneous soil model
would overestimate the settlement by 50% whereas in the case of non -
homogeneous model it was underestimated by more than 100%.
Ta and Small (1995, 1996, & 1998) performed a more rigorous

analysis considering layered soil, adopting the finite layer theory in which the
soil was treated as a series of horizontal isotropic or cross isotropic elastic
layers of infinite extent in the lateral directions. For the case of load on
circular area, Henckel transformations were applied to all field quantities
namely loads and displacements. Solution was obtained for transformed
quantities and these were finally converted to obtain actual field quantities.
The study was done by Ta and Small for pile group by extending the method
of Lee et al (1991) for settlement of single pile using finite layer theory. It
was found that this method was advantageous in studying the behaviour in
layered soil only; but a very limited number of layers only be considered to
save computer time. They extended the study to piled raft. The raft was
modelled as thin elastic plate and the soil was modelled using finite layer
method. They could estimate the raft bending moment and pile head loads and
found that the load distributions along the shaft of the pile in layered soils
were affected by relative thickness and stiffness of the soil layers.
50
Combining the finite element method for raft and finite layer
analysis for soil and pile group, Ta and Small (1998) analyzed the results of
several case studies including the centrifuge models and prototype piled rafts
and found that the results were in agreement. However under normal working
load conditions which is mostly elastic in nature, the method proposed
appears to predict the results with a reasonable accuracy even though this
method did not include the contact forces and the material non- homogeneity.
Xu and Poulos (2002) developed a computer program GEPAN to

study the behaviour of piles and pile groups and compared the results with the
output of PIGLET and DEFPIG and found to be in agreement. In spite of the
fact that was able to give various interaction factors and appear to be very
versatile, no further application was developed mainly due to the
mathematical complexity.
2.4.3.3 Finite element analysis
Prokoso and Kulhawy (2001) studied the effect of various

parameters related to pile and raft on the normalized settlement of piled raft
adopting two dimensional plane strain model (Figure 2.14).
Figure 2.14 Plane Strain idealisation of piled raft (Prokoso and

Kulhawy, 2001)
51
The modulus of pile wall was computed from the term equivalent
modulus which was a function of number of piles, width or dia of the pile and
soil modulus. The study concluded that the ratio between the width of the raft
and the length of the pile played an important role on settlement behaviour of
piled raft. The piled raft with ratio equal to unity was very effective in
reducing over all settlement, where as a ratio 0.5 was very effective in
minimizing the differential settlement. Further it was concluded that a pile to
raft area ratio of 5% to 6% was adequate to reduce the overall settlement. The
results were mostly in the form of non-dimensional parameters. While the
contribution was very useful as a parametric study, it has only a very limited
application. The procedure would be ideal for a single group of large number
of piles, in a row.
Three dimensional analyses were carried out by Reul (2000) and

Katzenbach et al (1998) in analysing the observations made on the proto type
piled raft supporting large structures on Frankfurt clay. In the analysis the
long term behaviour of the soil was considered with the drained shear
parameters c and ‘φ’. The nonlinear material behaviour was considered by the
elasto - plastic cap model. They developed a new design concept for piled raft
foundations. This concept was developed based on the results of geotechnical
measurements made on Frankfurt clay and numerical model that was
developed at the Institute of Geotechnics, Darmstadt University of
Technology, Germany. This numerical model consists of three parts: (i) the
geotechnical model for continuum (i.e. 3 dimensional finite element mesh)
(ii) elasto-plastic constitutive law to described the soil behaviour (i.e.
constitutive relation), and (iii) numerical computer simulation in a step by
step analysis. The constitutive law is based on a modified Drucker-Prager
model taking into account inelastic soil behaviour depending upon the stress
path and previous stress history. The ability of the model was verified by
numerical simulation of in-situ pile load test and by the back analysis of
existing piled raft. In the numerical analysis, single pile and a single pile-raft
unit were used.
52
Figure 2.15 Influence of pile-raft interaction on contact pressure

Frankfurt Clay model (Katzenbach et al 1994)
It was concluded that the frictional resistance of pile was not only a
function of soil strength but was also a function of residual stress in the soil,
besides relative movement between pile and soil. The influence of pile-raft
interaction on the contact pressure has been brought out in Figure 2.15. This
indicates that the pile leads to a significant reduction of contact stress in the
raft next to the pile shaft.
Further, the load-settlement behaviour of piled raft in Frankfurt

clay was examined by varying the number of piles and drawn the following
conclusions :
Settlement of piled raft plays a significant role in sharing the load

between the piles and the raft but the load sharing ratio (αpr) is not a linear
function of settlement.
53
The load bearing behaviour of the pile in a piled raft is different

from the behaviour of a comparable single free standing pile and it depends
on the position of the pile in the group and its spacing.
Based on the observations made through detailed numerical

analysis, the requirements for the safe and economical design of piled raft was
suggested and the proposed concept was applied for the design of the
foundations of Main Tower and Eurotheum building, which are founded on
piled rafts in the Frankfurt clay.
Katzenbach et al (2006) compared the load bearing response

between a conventional pile foundation and piled raft foundation through a
three dimensional finite element analysis which was developed in the
Darmstadt University. This three dimensional finite element model was
validated further by analysing the high-rise buildings of Messeturm and
Eurotheum towers in Frankfurt am main. The measured settlements of these
two buildings were compared with numerically computed values. The close
agreement between the measured and numerically computed settlements,
emphasized the importance of rigorous finite element method for numerical
modelling of complex structures.
Reul (2000) in his study, on the piled raft on overconsolidated clay

incorporated the effect of consolidation by a coupled pore pressure and stress
finite element analysis. It was shown that with every load, increase the piled
raft coefficient increased with time. Further it was shown through the analysis
that there was no difference in the results for both one phase model (without
pore pressure) and two phase model (with pore pressure). The contact
between the structure and the soil was taken as ideal contact meaning that no
relative motion between the nodes of structure and the soil. Reul first
established the constitutive model required for Frankfurt clay through a series
of tests as Drucker – Prager cap model and this was further used for all the
studies. The study also considered the effect of mesh refinement wherein it
was shown that relatively coarser mesh also produced results within an
54
acceptable level of accuracy. The analytical model was calibrated by means of

back analysis of the results of field measurements monitored in the following
three buildings: Messe Tower, Westend and Torhaus (Figures 2.16 to 2.18).
It was reported that a settlement reduction coefficient of 0.51 to 0.63 could be
achieved. Further the author established that under normal load, the piles do
not reach their ultimate bearing capacity.
(a)Plan of the foundation
(c)Finite element nodes for
(d) Finite element mesh for soil and combined

piled raft
Figure 2.16 FE mesh of the foundation – Messeturm (Ruel, 2000)

55
(a) Plan of the foundation (b) Finite element nodes for

raft
(c) Finite element mesh for soil and combined piled raft
Figure 2.17 FE mesh of the foundation – Westend (Ruel, 2000)

56
(a) Plan of the foundation (b) Finite element nodes for raft
(c) Finite element mesh for soil and combined piled raft
Figure 2.18 FE mesh of the foundation – Torhaus (Ruel, 2000)

57
One of the very important conclusions drawn was, that stiffness of

the raft or the stiffness of the super structures had no influence in the load
share between the piles and the raft. But the depth of the soil layer below the
pile tip influenced the settlement of the piled raft. Reul and Randolph (2003)
further showed that the measured values and analytical values agree well for
the cases analysed. It was shown that by changing the layout marginally,
improved value of settlement reduction could be achieved. In this analysis the
difference between the results of FEA and measured values ranged from 20%
to 30%.
In their subsequent study (Reul and Randolf, 2004) on the effect of

non-uniform loading, they showed that the length of piles rather than number
of piles had more influence on the load taken by the piles. The overall
stiffness of the piled raft decreased with increasing load level. But differential
settlement was more sensitive to raft soil stiffness ratio. For uniform loading,
a small central group was found to be efficient in reducing the differential
settlement.
2.5 SUMMARY
The literature on piled raft foundation was reviewed under

following heads: Laboratory model studies, studies on prototype piled rafts
and analytical methods.
It appears that the laboratory model studies reported both on 1g and

centrifuge models have aimed at to validate analytical models developed for
soil-structure interaction. 1g model tests conducted are very limited and in the
limited work, the concentration is on piled raft in overconsolidated clay.
Though it is well known that model tests conducted in a geotechnical
centrifuge provide more relevant information than 1g model tests, the
research on piled raft foundation using centrifuge technique is also scarce,
which is suppose to the best alternative method to field test on prototype
foundation.
58
Field monitoring studies on piled raft foundation are more in

number than laboratory studies on 1g models. Most of the studies are on
deeply seated piled raft foundation in Frankfurt clay. The intention of
providing piles below the raft is to reduce both total and different settlements
of tall and heavy structures. The data are obtained by monitoring the
performance of foundation during construction and some period after the
completion of structure. These data are used improving design philosophy and
validating analytical models. However, parametric analyses are almost nil and
are nearly impossible to carry out such studies in prototype models. It is also
true that its suitability for buildings of medium size and load with relatively
smaller diameter piles and thinner raft is not been studied and validated.
Further they are time consuming and expensive. Despite some of the
limitations, the field data are considered highly valuable in the sense that they
represent real behaviour, which will be very useful for improving design
method and for validating the analytical methods.
A number of versatile analytical methods have been developed over

the past two decades for analyzing complex behaviour of piled raft. Relative
performance between various methods is also been discussed. Few analytical
methods have been validated with limited experimental data. Some
researchers have specifically concentrated on the analysis of piled raft and
developed suitable constitutive model and validated with the limited field
data. It is also true to certain extent that there are few field data to validate
and is mostly on one soil type. Further no comparison has been made in the
field between structures designed both by the conventional and the latest
design philosophy such as Katzenbach et al (2000) and Clancy and Randolph
(1993).
If piled raft is to be an automatic choice in locations where pile is

required to reduce the settlement of raft, the existing design methods need to
be improved by concentrating on minimizing the number of piles and pile
locations. This may end to economy in piled raft design as well as meaningful
59
performance of piled raft. Further the design based on observational data that
was validated by a suitable analytical method is considered more realistic. For
example the results of observations made on the piled raft of Eurotheum
tower has been used to design the piled raft of the Maxtower (Katzenbach
et al 2002). This implies that adequate data based on the observations must be
available along with analytical validation so that the results can be used as a
base for the design. Therefore, it is felt essential that such data base must be
available for different soil types. From the reasons presented above, it was
decided to study the behavior of piled raft through 1g model tests, analysing
the test results through numerical model and monitoring a real size piled raft
foundation supporting a twelve storied building constructed in a deposit
where the soil is dominantly sandy.
60
3
4CHAPTER 3
EXPERIMENTAL SET UP AND MATERIAL PROPERTIES
3.1 INTRODUCTION
The behaviour of piled rafts involves complex mutual interactions

between the pile group, the raft and the soil. These interactions depend on
many factors such as pile parameters (diameter, lengths, spacing stiffness etc.)
raft parameters (size, thickness, stiffness etc) and soil parameters (nature of
deposit, stress history, relative density/compactness etc.). Apart from these
variables, there are other parameters like shape of raft, arrangement of pile,
load position, location of raft, stress level etc. affect the performance.
Therefore, in this research it is aimed at to study the behaviour of piled raft
foundation in sand through 1g model tests. Tests were conducted to
understand the load settlement and load sharing behaviour of piled raft as well
as to establish the influence of various parameters related to pile, raft and soil
on the settlement reduction and load sharing. In the piled raft test series, the
variables analysed were diameter and length of piles, number of piles, spacing
between piles, pile arrangement and shape of raft (circular, square and
rectangular). Variation of raft thickness and bed density were the other
parameters considered for the raft and the supporting medium (sand)
respectively. A series of 1g tests were also conducted on unpiled raft (plain
raft) and free standing pile group wherein pile cap/raft is not resting on the
surface of the sand bed. The purpose was to compare load-settlement response
of unpiled raft with piled raft and in the case of free standing pile group, the
effort was put to know the settlement at which total friction mobilized.
61
In all the tests a model scale of 1/100 was used. It should be noted
that the models selected for the present work were not intended to simulate
behaviour of a specific prototype foundation at a specific location. The results
from the 1g model tests are analysed to understand the role of various pile
parameters in reducing settlement of the raft and in sharing the applied load
and also to investigate the applicability of existing numerical methods. In this
chapter, the details of testing facility, the details of various models, properties
of the test bed and testing procedures are described.
3.2 TEST FACILITY
The test facility developed in this study consists of a loading

frame, a steel box, sand-raining and controlled tamping device, piled raft and
raft models and loading arrangement. The details on above facilities are
presented in this section. The schematic view of the experimental set up is
given in Figure 3.1.
Figure 3.1 General arrangement of experimental set up

62
3.2.1 Loading Frame
A loading frame of reaction capacity of 50kN was designed and

fabricated in the division of soil mechanics and foundation engineering for the
proposed study. The dimensions of the frame was arrived by keeping in mind
a test box of 1m×1m×0.6m size with sufficient height to support loading
arrangement as well as to change the position of loading whenever required.
The set up consists of a portal frame arrangement made up of two threaded
steel rods of 40mm diameter and length of 2500mm and connected together
by means of a steel girder. The steel girder was a built up section made out of
two ISMC 200 channel sections arranged back to back and welded together
with stiffener plates. Suitable arrangement was made in the girder to support
loading arrangement such as hydraulic jack. The girder was secured with the
vertical column members using clamps provided in the threaded rod. The
threaded rod was provided with arrangement and facilities to change the
position of the girder to a required height.
The frame was secured firmly on to the floor of the laboratory by

means of base plates and anchor bolts. A platform was made covering the
frame which had the facility to support the sand raining frame and space for
supporting dial gauge frame. A schematic line diagram of the test frame is
also shown in Figure 3.1.
The vertical compressive load was applied to pile raft model

through a hydraulic jack of 100kN capacity, which was supported centrally at
the bottom flange of the steel girder made of channel sections. It was operated
by means of a manually operated hydraulic pump. The load was transferred to
the model through a proving ring, which was fixed on the head of jack and the
rigid loading platen was connected to the proving ring as shown in plate 3.1.
The over all view of the test set up is presented in plate 3.2.
63
Jack
Proving Ring
Dial gauge
Platform
Loading platen
Load transfer plate with
Piled Raft
buttons
Plate 3.1 Loading assembly
Plate 3.2 Test set up

64
3.2.2 Test Box
A rectangular rigid steel test box of inner dimensions

1m×1m×0.6m with 10mm wall thickness was fabricated and used for
conducting the experiments on the model piled raft in the laboratory. With a
wall thickness of 10mm and several stiffening beams on the outer side, the
box is considered to be rigid. One of the faces of the box was fitted with plain
glass of 12mm thickness with suitable steel flat stiffeners. This facilitates
calibration and physical examination during bed preparation. One side
(1m×0.6m) of the box was provided with a rigid sliding door so that at the
end of each test, the tank can be emptied easily. The inside face of the tank
was graduated at every 50mm depth intervals to aid preparation of sand bed in
layers by sand raining technique.
3.2.3 Raft And Pile Models
Piled raft models of three different shapes namely, circular, square

and rectangular models were chosen. They represented the behaviour of
foundation of circular tank, square raft of multi-storied building and
rectangular raft of ware houses respectively.
3.2.3.1 Model Raft
Rafts of three different shapes with model to prototype scale of

1/100 were fabricated from a single Perspex sheet. The size of the model was
chosen to represent the prototype foundation of 20m wide/diameter for
square/circular shape, and 7m×20m size for rectangular raft. The thickness of
model raft was maintained uniform, though thickness of raft of real size piled
raft foundations were in general varied from the edge to centre. The basic
properties of model raft are shown in Table 3.1.
65
Table 3.1 Details of Model Raft
Description Raft Details

Shape Circular Square Rectangular
Material Perspex Perspex Perspex
Size 200mm dia 200mm×200mm 75mm×200mm
Thickness 6, 8 &10mm 8, 10 & 12mm 8mm
3.2.3.2 Model Piles
The model piles were of Perspex rods, circular in shape. In order to

represent commonly used pile diameters of 600mm, 800mm and1000mm,
Perspex rods of 6mm, 8mm and 10mm dia respectively were used as model
piles. The maximum and minimum lengths of piles tested were 200mm and
75mm respectively. Padfield and Sharock (1983) had indicated that the piles
have to be compressible, meaning they have to be friction piles. Hence the
length of the piles were chosen in such a way that the depth of the bed below
the tip was sufficiently thick so that, bottom rigidity does not affect the pile
behavior and pile functions purely as friction pile. However, in one or two
cases length larger than the raft width was tested. Threads were provided at
the top end of piles to facilitate the proper connection and to generate
monolithic action between the piles and the raft. Details of piles used in the
model tests are presented in Table 3.2. For test on free standing pile groups a
set of piles were made 20mm longer than the piles of piled raft.
Table 3.2 Dimensions of Model piles
Diameter Lengths of piles (mm)

6mm 100,120,160
8mm 100,120,160,200
10mm 75,100,120,160,200
66
3.2.3.3 Piled Raft Models
Raft and pile models of required diameter and length were

assembled by fitting the piles in the holes of the raft and connecting them
together with the help of stainless steel screws. Plates 3.3, 3.4 and 3.5 show
typical circular, square and rectangular piled raft models respectively.
Plate 3.3 Circular piled raft
Plate 3.4 Square piled raft
Plate 3.5 Rectangular piled raft

67
Piled raft models were tested for different arrangements of piles.

The purpose was to examine the influence of pile configuration and number
of piles on load sharing and settlement reduction. Various pile configurations
tried in this study for circular, rectangular and square shapes are presented in
Figures 3.2, 3.3 and 3.4 respectively.
The perspex material used for making raft and pile models was
tested in the laboratory as per ASTM D638 and Young’s modulus of material
was determined. The modulus value thus determined is 3000N/mm2. Its
Poission’s ratio is taken as 0.33. Though Young’s modulus of model material
is one eighth of the concrete, the material is chosen keeping in mind the
modulus value of the test sand (Es), the ratio between the modulus of perspex
and the sand is maintained in the range between 5 and 25, so that the
foundation system represents neither too stiff nor too flexible behaviour.
More ever some of the model tests reported on piled raft (ex. Weisner and
Brown, 1978) were conducted using pespex materials.
NO OF PILES 21 NO OF PILES 25 NO OF PILES 37

RADIAL ANGLE 36° RADIAL ANGLE 30° RADIAL ANGLE 20°
NO OF PILES 17 NO OF PILES 21 Raft diameter – 200mm

RADIAL ANGLE 45° SQ GRID SPACVING 4d
Figure 3.2 Layout of piles for Circular piled raft

68
NO OF PILES 14 NO OF PILES 10
SPACING 3d SPACING 4d
SPACING 6d SPACING 7.5d
Raft size – 70mm x 200mm
Figure 3.3 Layout of piles for Rectangular piled raft
SPACING 4d SAPCING 6d
Figure 3.4 Layout of piles for Square piled raft

69
3.2.4 Instrumentation
The instrumentation part of the set up comprised of a 20kN

capacity proving ring to measure the applied load and two dial gauges to
measure the piled raft settlement. Each division in the proving ring was
equivalent to 2.48kN. This was calibrated before put to use in the test. Each
time when one series of tests were completed, the calibration was rechecked.
The settlement was measured with two dial gauges of 0.02mm accuracy,
having a 50mm plunger. The loading platen had a specific provision to
position dial gauge plunger so that during the test the plunger would not get
disturbed. Plate 3.6 shows the arrangement of dial gauges and proving ring for
measuring settlement and corresponding load respectively.
Dial gauge
Arm
Dial gauge
Plate 3.6 Displacement Measuring Arrangement
3.3 TEST MEDIUM
Uniformly graded clean river sand was used in all the experiments
as a test medium. Laboratory tests were conducted on the representative sand
samples for gradation, specific gravity, maximum and minimum dry unit
weights and angle of sharing resistance for different densities as per relevant
70
ASTM standards. The index properties obtained are presented in Table 3.3
and the sand is classified as poorly graded sand (SP). The angle of shearing
resistance obtained was used in the validation of analytical model for various
densities and are also presented in Figure 3.5.
Triaxial tests were also conducted at low confining pressures to

represent the stresses in test condition. Typical stress strain curve is shown in
Figure 3.6. The angle of shearing resistance obtained from the triaxial shear
and direct shear tests were compared and found that the direct shear test
values were higher by 2 degrees maximum. The strength properties of sand
bed for densities at which tests conducted are presented in Table 3.4. From
the stress – strain curves of triaxial test, the initial tangent modulus of sand for
densities of test bed was determined and these values were used in the
analysis.
Table 3.3 Properties of sand used in the test bed
Parameter Symbol Value

Maximum dry unit weight γdmax 18.19 kN/m3
Minimum dry unit weight γdmin 14.41 kN/m3
Maximum void ratio emax 0.84
Minimum void ratio emin 0.46
Specific gravity G 2.65
Coarse sand 3%
Medium sand 85%
Fine sand 12%
Effective grain size D10 0.38
Uniformity coefficient Cu 2.63
Coefficient of curvature Cc 1.22
Classification SP – Uniformly graded sand
71
ANGLE OF INTERNAL FRICTION, deg
3
UNIT WEIGHT, kN/m
Figure 3.5 Variation of angle of internal friction with unit weight
0.40
2
70N/mm2
0.070N/mm
0.35
0.035N/mm2
0.30
2
DEVIATOR STRESS N/mm
0.25
0.20
0.15
0.10
0.05
0.00
0.00 0.05 0.10 0.15
AXIAL STRAIN
Figure 3.6 Typical stress - strain curve of test sand (Medium dense)
72
Table 3.4 Strength properties of bed material at different densities
Sl State of Unit weight, γ Angle of shearing

No. Compaction (kN/m )3
resistance, φ °
1 Dense 16.2 41
2 Medium 15.5 37
3 Loose 14.8 34
3.4 BED CALIBRATION
One factor, which has strong influence on the load-settlement

response of model foundation is the density of packing of the sand grains. The
test series conducted in this research covers tests with loose (ID = 0.35)
medium dense (ID = 0.45) and dense (ID = 0.65) packing. Therefore a method
of forming artificial beds of sand that are homogeneous and reproducible over
a wide range of densities were required (Turek and Katzenbach 2003). This
was achieved by adopting controlled pouring and tamping method, which was
successfully adopted for several model tests conducted (Ilamparuthi and
Muthukrishnaiah 2001).
The controlled pouring of sand was carried out using sand

depositing apparatus which is a funnel in shape with a pipe of diameter 25mm
and length 600mm. Sand was poured into the test tank with a constant height
of drop. The height of drop used for the test with loose density was 50mm.
For medium dense and dense conditions, both controlled pouring and tamping
was adopted. The sand deposited in layers of 50mm thick was completed
using a tamper of square steel plate (75mm×75mm) as a base and steel rod of
10mm diameter and 500mm long as a guide rod. One end of the steel rod was
welded at the centre of the base plate. A drop weight 23.6N, cylinder in shape
was used for compacting the sand to reproduce the required placement
density.
73
The test tank was calibrated by adopting the procedure outlined

above to achieve sand packing of different densities. For each density used in
the investigation, the weight of sand used for each layer of 50mm thickness
was determined. The total weight of sand used for filling the entire height of
tank was determined and average density was worked by knowing the volume
of the container. The container was calibrated three times for each density.
The difference in density between each trial was negligible; hence it was
considered that the method adopted for the preparation of bed is consistent
and reproducible. As a method checking measure, container of known volume
was placed in the tank and the weight of sand filled in the container was
measured and density was arrived. The methods adopted for obtaining loose,
medium dense and dense conditions are presented in Table 3.5.
Table 3.5 Method adopted for preparation of sand bed
No. of blows Average

State of Ht. of fall
Method for 5cm weight unit
density (cm)
thick layer (kN/m3)
Loose Sand raining 5 Nil 14.8
Medium Sand raining 10 110 15.5
dense & tamping
Dense Sand raining 15 165 16.2
& tamping
In order to verify uniformity of sand bed, light cone penetration

tests were conducted at different locations of sand bed. Details of the cone test
and typical results obtained are presented in Appendix - 1.
3.5 INSTALLATION OF PILES AND PILED RAFT
The sand bed of proposed density was prepared by adopting the

procedure as explained in the section 3.4. After preparing the bed to the
specific level of the tank, the template of pile group configuration to be
analyzed was placed on the sand bed by matching the centre of the template
74
with the centre of the tank. A guide tube which can accommodate specific pile
diameter was positioned on the template and the pile was inserted through the
guide tube and made to touch the sand bed (plate-3.7).
The pile thus positioned was driven into the sand as shown in plates
3.7, 3.8 and 3.9 using a wooden mallet to a depth 20mm short of its length.
All the piles were installed by adopting the procedure as explained keeping all
the piles to project for a height of 20mm above the bed as shown in plate 3.10.
Model pile
Guide pipe
T l t f il l t
Plate 3.7 Pile installation assembly
Plate 3.8 Pile driving

75
Plate 3.9 Test pile installation
Plate 3.10 Model piled raft in position
The top level of the piles was checked to ensure that all the piles
were projecting above the bed to the same height. The raft was placed on the
piles and connected to them with the help of stainless steel screws. The entire
assembly was pushed slowly into the sand bed with the help of loading
assembly; the resistance offered by the piles was recorded. Once the raft
touched the sand bed, the loading assembly was removed, and the levels were
checked again to make sure that the contact of the raft with the bed was
uniform. The same procedure was repeated for all the tests conducted.
76
The templates and the guide tubes were made as per the
requirements of the pile group layouts proposed, and dimensions of piles to be
tested. Guide pipes were made from Perspex tubes with inner diameter of tube
to be marginally higher than the diameter of pile to avoid friction between the
pile and the tube. Length of tube was 50% of the length of longer pile tested.
The piles were driven in a sequence. In the case of circular piled

raft the piles were driven from the centre and proceeded radially towards the
edge. For square and rectangular piled rafts, the driving was from one edge of
the raft and proceeded towards the other edge. While driving the piles, there
was marginal disturbance to the sand in the near vicinity of the piles. A small
depression of sand around each pile to the extent of 2.5D and depth of 3 to
4mm was noticed. The maximum depression of sand was noticed particularly
in the loose sand bed. At the end of driving all the depressed areas were filled
with sand and the surface was levelled. The maximum weight of sand used for
levelling the surface of sand bed was around 120gms, and hence variation in
density was considered insignificant.
Solid rods were used as model piles, and therefore the pile tip
would offer some resistance. The effect of pile tip resistance was considered
not to influence the behaviour pile, since the length of pile was many fold
higher than its diameter. Moreover, model pile was designed as primarily a
friction pile.
3.6 SEQUENCE OF TESTING
Initially test on plain raft was carried out and the load settlement
curve for the plain raft was established. This was followed by tests on free
standing pile group in which the raft was not in contact with the sand.
Subsequently test on piled raft was carried out. This sequence was followed
for all the shapes and densities of sand except for rectangular raft. In the case
of rectangular raft free standing pile group was not tested. From the tests
conducted on circular and square rafts, it was clearly established that the
77
piled raft takes much higher load than the free standing pile group. The
parametric study was conducted for each shape of piled raft in the same
sequence.
3.7 LOAD TEST PROCEDURE
The following sections outline the procedure of the tests carried out
in the same sequence.
3.7.1 Tests on Plain Raft
Tests were performed initially on raft without piles for all the three
shapes of rafts considered namely circular, square and rectangular shapes.
While tests were done on loose, medium dense and dense beds in the case of
circular and square raft, rectangular raft was tested on medium dense bed
only. The bed was prepared as explained earlier and the raft without piles,
namely plain raft was placed on the bed. While placing the model raft,
sufficient care was taken so as to ensure that the model was placed perfectly
horizontal. This was checked on either directions of the raft using sensitive
sprit level. On the raft a loading pad was placed to apply uniformly distributed
load on the raft through a rigid loading platen. The size and shape of loading
pad is same as that of the raft to be tested. But it was provided with holes at
regular spacing of 10mm to accommodate bullet shaped loading nipples. The
loading platen was set on the loading pad and the seating load of 5kN/m2 was
applied. This load was kept for a period of 15minutes. The readings in the
proving ring and the two dial gauges set on loading platen were reset to zero
before commencing the actual test.
The load was applied in equal increments on the model raft by

operating hydraulic pump. A constant load- time rate method was adopted.
Each load increment was maintained constant on the plate for a duration of 30
minutes and the settlement readings were recorded at every 5minutes interval
and before applying the next load increment. The duration of loading period
78
was arrived based on trial tests conducted wherein most of the settlement
under a given load was completed within a loading period of 20minutes. The
test was continued until raft settlement was 20mm or the load at which
settlement was uncontrollable.
3.7.2 Test on Free Standing Pile Group
The tests on the free standing pile group provide frictional capacity
of the piles in the group when the cap was not in contact with ground. This
magnitude of friction forms an important parameter to study the effect of raft
when it was in contact with the ground as the enhancement in the friction
value forms a measure on the performance of the piled raft in load sharing. As
described earlier, the pile group was installed in such a way that all the piles
project above the top level of sand bed to a length of 20mm. The raft was
placed on the piles and fixed with stainless steel screws. After checking the
levels of the raft, loading pad was positioned on the raft and the load was
applied through loading platen by operating the hydraulic pump connected to
the jack. The load was applied in small increments. Each increment was
retained till there was no further settlement. The application of the load and
recording of the settlement was continued till the pile group reached their
ultimate resistance. In all the tests on free standing pile groups, the piles
reached ultimate resistance at the settlement around 2mm to 2.5 mm
irrespective of pile group configuration and number of piles.
3.7.3 Load Tests on Piled Raft
For load test on piled raft, piles were installed as explained in the
section 3.5 and raft was fixed on the piles. Entire assembly was pushed into
the soil and the raft was made to sit on the top of the sand bed. The piled raft
so installed was load tested by adopting the procedure explained. The details
of tests conducted on circular, square and rectangular shaped piled raft rafts
are presented in Table 3.6, Table 3.7 and Table 3.8.
79
As described earlier after assembling the piled raft and the

mandatory check were completed, the loading plate with ball arrangement
was placed on the raft. The loading platten was fixed to the bottom of the
proving ring and the assembly was lowered on the plate. A small seating load
was applied to ensure proper seating of the raft on the bed. The dial gauges
were fixed in its position and was set to zero. At this stage the assembly was
ready to commence the test. The loading increment was kept 10% of the
estimated capacity of the piled raft. Each increment of load was kept on the
piled raft for a maximum period of 30 minutes. The test was continued till the
piled raft settlement was 20mm.
Table 3.6 Details of tests done on circular piled- rafts
System details
Nature of
Tests conducted on Raft Pile Bed Parameters studied
output
Size Thickness L mm d mm N Nos. Ar % density
Plain raft Load settlement 200mm DIA. 6mm, 8mm, - - - - L, MD, D Load settlement
10mm behaviour
Free standing pile group Load settlement 200mm DIA. 8mm 200 10 21 5.2 MD Pile group capacity
Piled raft Load settlement 200mm DIA. 8mm 200, 160, 120, 100, 75 10 21 5.2 MD Variation in length
200mm DIA. 8mm 200, 160, 120, 100 8 21 5.2 MD Variation in length
200mm DIA. 8mm 160, 120, 100 6 21 5.2 MD Variation in length
200mm DIA. 8mm 160 10, 8, 6 21 5.2 MD Variation in diameter
200mm DIA. 8mm 200, 160, 120, 100 10 21 5.2 L Variation in length
200mm DIA. 8mm 160 10 37, 25, 21, 11 MD Area ratio
200mm DIA. 8mm 200 10 21 5.2 MD Square grid
200mm DIA. 8mm 160 10 11 2.2 MD Configuration
200mm DIA. 8mm 160 10 21 5.2 L, MD, D Bed density
200mm DIA. 6, 8, 10mm 160 6 21 5.2 MD Raft thickness
200mm DIA. 6, 8, 10mm 160 10 21 5.2 MD Raft thickness
80
Table 3.7 Details of tests done on square piled- rafts
System details
Nature of
Tests conducted on Raft Pile Bed Parameters studied
output
Size Thickness L mm d mm N Nos. Spacing density
Plain raft Load settlement 200 x 200 8, 10 ,12 - - - - L, MD, D
Free standing pile group Load settlement 200 x 200 8 160 10 9 MD
Piled raft Load settlement 200 x 200 8 160 10 25, 9, 9 4D, 6D, 7.5D MD Variation of spacing
200 x 200 8 160 10 25, 9, 9 4D, 6D, 7.5D L Variation of spacing
200 x 200 8 160 10 25, 9, 9 4D, 6D, 7.5D D Variation of spacing
200 x 200 10 160 10 9 4D, 6D, 7.5D MD Variation of spacing
200 x 200 12 16 10 9 4D, 6D L Variation of spacing
200 x 200 12 16 10 9 4D, 6D MD Variation of spacing
200 x 200 12 16 10 9 4D, 6D D Variation of spacing
200 x 200 8 160 6, 8, 10 9 6D L Pile diameter
200 x 200 8 160 6, 8, 10 9 6D D Pile diameter
200 x 200 8 200, 160, 120 10 9 6D L Pile length
200 x 200 8 200, 160, 120 10 9 6D MD Pile length
200 x 200 8 200, 160, 120 10 9 6D D Pile length
200 x 200 8 160 10 9 6D L, MD, D Effect of bed density
200 x 200 8, 10, 12 160 10 9 6D MD Eff. of raft thickness
200 x 200 8, 10, 12 160 10 9 6D MD Configuration
81
Table 3.8 Details of tests done on rectangular piled- rafts
System details
Tests conducted on Nature of output Raft Pile Bed Parameters studied
Size Thickness L mm d mm N Nos. Ar % density
Plain raft Load settlement 70x200 8 - - - - MD Behaviour of plain raft
Piled raft Load settlement 70x200 8 200, 160, 120, 100 10 8 4.4 MD Variation in length
70x200 8 200, 160, 120, 100 8 8 4.4 MD Variation in length
70x200 8 160, 120, 100 6 8 4.4 MD Variation in length
70x200 8 75 10 14, 8, 6, 4 7.8, 4.4, 3.8, 2.2 MD Effect of pile area ratio
70x200 8 120 6, 8, 10 8 4.4 MD Effect of pile diameter
82
83
CHAPTER 4
EXPERIMENTAL RESULTS AND DISCUSSION
4.1 GENERAL
The experimental results of 1g model tests on piled rafts are

presented and discussed in detail in this chapter. The main focus in this
chapter is comparison of the behaviour of piled raft with that of the unpiled
raft (plain raft). In this study, reduction in the settlement of the raft due to
addition of piles and the load shared between the pile group and raft at
different settlements of the piled raft are focussed. Accordingly, 1g model test
results are analysed, and the effect of various parameters related to pile, raft
and the soil on the load - settlement and load sharing behaviour are brought
out. In this chapter results of all the three shapes of piled raft namely, circular,
square and rectangular rafts are presented and discussed. Among the
parameters investigated, surface roughness of piled raft and layout of piles
(size, pile arrangement) of circular piled raft are analysed initially, in order to
limit the number of parameters to be investigated. The parameters for which
tests conducted, have shown very consistent results. Results of tests are
presented and discussed in the following sequence.
• Effect of surface roughness of piled raft
• Selection of pile layout for circular piled raft
• Behaviour of circular piled raft
• Behaviour of square piled raft
• Behaviour of rectangular piled raft

84
In all the above cases the influence of various parameters in load

settlement response including settlement reduction, and load sharing
behaviour between the components of piled raft are discussed independently.
4.2 EFFECT OF PILE ROUGHNESS
In the model tests, Perspex was used as model material. In order to

bring out the effect of surface roughness, limited tests were conducted on
piled raft using piles, and raft with rough surface. The surfaces of piles and
raft were made rough by gluing sand of particle sizes between 0.10mm and
0.20mm on their surfaces. The technique adopted here is similar to the
method adopted by Leung and Dickin (1990) in their research on uplift
behaviour of belled piles.
Typical load – settlement curves of piled raft of diameter (D)

200mm, and thickness (t) 8mm, with piles of rough and smooth surfaces are
compared in Figure 4.1 for dense sand. Pile diameter (d), length (L), and
number of piles (N) adopted in the tests are 10mm, 160mm and 21
respectively.
LOAD, kN
SETTLEMENT, mm
Figure 4.1 Effect of roughness on load - settlement response of piled

raft in dense sand
85
While the load settlement behaviour remains the same, the piled
raft with rough surface shows marginally higher load than smooth surfaced
piled raft for a given settlement. Similar response is seen in the tests
conducted on loose and medium dense sand conditions. The roughness
increases the resistance of piled raft by a maximum of 8% in dense sand.
Since the increase in the piled raft capacity was only 8%, studies were
continued on piled rafts with natural surface of the Perspex material (i.e.
surface of raft and pile are not made rough).
4.3 SELECTION OF PILE ARRANGEMENT FOR CIRCULAR

PILED RAFT
In order to select the arrangement of piles to be used in the study,

the load-settlement response of piled raft with two independent arrangements
of piles of radial and square grid configurations as shown in Figure 4.2 were
analysed by keeping the number of piles and spacing between the piles are the
same for both the configurations.
No: OF PILES = 21 No: OF PILES = 21

RADIAL ANGLE 36 ° & 4d spacing SQUARE GRID SPACING 4d
Figure 4.2 Layout of piles in Circular piled raft
The first arrangement has piles placed radially with a radial angle
(RA) of 360, with one pile at the centre of the raft. The total number of piles
(N) is 21 with a spacing of 4d (d = diameter of pile) along the radial direction.
The second arrangement is termed as square grid layout, in which piles are
placed at a spacing of 4d and here also the number of piles is 21. The model
tested is a circular piled raft of 200mm diameter and 8mm thick.
86
In both the cases, the parameters relating to the piles and raft were
kept the same, except the arrangement of piles. The diameter and length of the
pile were 8mm and 160mm respectively and the area ratio (Ar) of piled raft
(Ar = Ap/A, where Ap is total cross sectional area of piles and A is the area of
raft) was kept as 5.2%.
Figure 4.3 presents the load - settlement curves of plain and piled
rafts. A study of the curves indicates that the settlement of piled raft is lesser
than the plain raft for a given load irrespective of the pile arrangement.
LOAD, kN
0 1 2 3 4 5 6 7 8 9 10
0
d = 10mm
2 t = 8 mm
4 L = 160 mm
N = 21
6
R.A. = 36º
8 D = 200 mm
SETTLEMENT, mm
Bed = Dense
10
12
14
16
18
PLAIN RAFT
20
RADIAL GRID
Piled raft
22
SQUARE GRID
24
Figure 4.3 Comparison of load settlement response of plain and piled

raft –Radial and square grid
It was found that for the given area ratio both the radial and square
grid arrangements exhibited almost identical load- settlement behaviour.
However the square grid layout carried a marginally higher load, when the
settlement was approaching 20mm particularly from the settlement level of
14mm. The maximum variation in the resistance between the two pile layouts
is 3% for the maximum piled raft settlement of 20mm. Tests were also
conducted on piled raft with square and radial arrangement for piles in other
densities and compared. The results of tests on other two densities showed
87
virtually no difference in load – settlement response between the two pile

arrangements of piled raft. Since the radial arrangement is more commonly
used in practice for tank pads, it was decided to carry out rest of the studies on
piled raft with the piles arranged in the radial directions.
4.4 BEHAVIOR OF CIRCULAR PILED RAFT
Circular rafts are commonly used for tanks and storage structures.
Therefore tests were conducted on plain circular rafts and circular piled rafts
with piles arranged radially. The load – settlement response of plain raft
(unpiled raft) and piled raft are compared and presented in this section in the
beginning, followed by the other aspects of the study.
4.4.1 Behaviour of Unpiled Raft (Plain Raft)
Figure 4.4 presents the behaviour of plain raft of 200mm diameter

and 8mm thick tested in loose, medium dense and dense sand bed. The results
are presented upto a settlement of 20mm for all the three densities.
LOAD IN kN
LOAD, kN
0 1 2 3 4 5 6 7 8 9 10
0
2 MEDIUM DENSE
LOOSE
4
DENSE
6
IN M
SETTLEMENTmm
10
SETTLEMENT,
12
14
16
D = 200mm
18 t = 8mm
20
22
Figure 4.4 Load - settlement response of plain circular raft of

200mm diameter
As expected, much higher settlements occurred in loose sand than

in medium dense and dense sand at any given load level. Further, the rate of
88
settlement is also higher in loose sand. The stiffest response is seen in dense
sand condition, and the raft – soil stiffness in dense sand is more than 3 times
higher when compared to loose sand for the reference settlement of 20mm.
The shape of load – settlement curves is almost similar for all the three
densities studied and also showed strain hardening behaviour. However soil
yielding occurred at higher load and lesser settlement in denser medium.
Further, the plain raft offered negligible resistance beyond 20mm settlement
even in dense sand, which indicates deformation of soil (large deformation
even for a small increase of load).
The effect of thickness of raft on the load – settlement response is

studied by conducting tests on raft having thickness of 6mm and 10mm, apart
from 8mm. It was found that the response is almost similar to that of 8mm
thick raft for a given density of sand. However, the resistance offered by the
raft is higher for larger thickness of raft irrespective of the settlement.
The load corresponds to 20mm settlement of the raft is presented in

Table 4.1 for various thickness of raft and densities of sand bed.
Table 4.1 Load corresponding to settlement of 20mm for plain

circular raft of 200mm diameter
Raft thickness, Load, kN

t
Loose Medium Dense
(mm)
6mm 1.67 4.31 5.88
8mm 1.96 5.10 6.28
10mm 2.06 5.29 6.77
For the raft thickness of 10mm the load corresponds to 20mm

settlement is 23% higher than the raft of 6mm thick in loose sand and in dense
sand it is 15% higher. Thus the carrying capacity of raft is higher for larger
thickness of the raft. From the results presented in Table 4.1, it can be inferred
89
that the raft thickness influences the load carrying capacity in loose and
medium sand than in dense sand. Also as the raft thickness increases, the rate
of increase in the load resistance reduces. Thus relative stiffness of raft – soil
system is more influenced by the density of sand bed, rather than raft
thickness. One reason for this response is perhaps the method adopted for
loading the model (ie, the raft is loaded through an arrangement which causes
uniform settlement irrespective of the thickness of raft). The experimental
data of plain raft (un-piled raft) is analysed further by plotting the data in
semi-log scale as shown in Figure 4.5.
10.0
C
B
LOAD kN
LOAD,
1.0
A
MEDIUM DENSE
LOOSE
DENSE
0.1 O
0 2 4 6 8 10 12 14 16 18 20 22
SETTLEMENT,
SETTLEMENTmm
mm
Figure 4.5 Characterisation of load - settlement response of circular

raft (D=200mm and t=8mm)
This figure presents the results of raft of 200mm diameter and 8mm
thickness, tested in loose, medium dense and dense sand conditions. The load-
settlement response presented in semi-log scale has distinctly shown the
change in slope of the curve. In the case of loose sand the change in slope of
the curve is seen at settlements of 3mm and 8mm and the corresponding loads
are 0.6kN and 1.1 kN respectively. Similar response is seen in medium dense
and dense sand conditions also. The slopes of the curve are changed at two
distinct displacements, which are1mm and 6mm for the raft tested in medium
sand and the corresponding settlements are 1.6mm and 6mm respectively in
dense sand. From the results presented above, the load – settlement response
90
of plain raft can be characterised as multi-linear response as shown in

Figure 4.5.
This multi-linear response comprised of three phases. The first

phase (0A part) represents elastic behaviour of raft-soil system, which has
higher stiffness than other two phases. Second phase (AB part) and third
phase (BC part) represent elasto-plastic behaviour of the raft-soil system and
they exhibit strain hardening behaviour. This response is seen in all the three
densities of sand. Thus the load-settlement response of model raft in sand can
be characterised as multi-linear strain hardening behaviour, which consists of
three phases.
4.4.2 Comparison of Piled Raft Behaviour with Unpiled (Plain) Raft
Figure 4.6 compares the load- settlement behaviour of plain raft and
the piled raft tested in medium dense sand for the raft thickness of 8mm.
LOAD, kN
0 1 2 3 4 5 6 7
0
RAFT
2
PILE
4 PILED RAFT
6
, mm
8
SETTLEMENT,
10
12 d= 10mm
14 t=
d =8mm
8mm
LL == 120mm
200mm
16 N = 21
18 R.A. = 36º
D = 200mm
20
Bed = Mediumdense
22
Figure 4.6 Load – settlement response of plain raft, piled raft and
pile group of piled raft in medium dense sand
Similar tests were carried out in the case of loose and dense sand
also. In all the tests, the raft diameter was kept as 200mm and the thickness
was 8mm. The length and the diameter of the piles were 120mm and 10mm
respectively. Twenty one piles were provided at a radial angle of 360. The pile
91
diameter was kept as 10mm, and the length was 120mm.It can be seen that at
any given settlement, the piled raft takes much higher load than the plain raft.
Table 4.2 presents the comparison of the load taken by the piled raft and the
plain raft for three different settlements and the densities tested.
Table 4.2 Comparison of load carried by plain and piled raft for
different settlements (D = 200mm, t = 8mm, d = 10mm,
L = 120mm and N = 21)
Load in kN @ Load in kN @ Load in kN @

Bed density 2mm sett 6mm sett 20mm sett
Plain Piled Plain Piled Plain Piled
Loose 0.34 0.72 0.84 1.30 2.00 2.70
Medium
1.30 2.30 2.80 4.05 5.10 6.80
dense
Dense 1.60 2.90 3.70 4.90 6.30 8.20
In all the three densities of sand, the load taken by the piled raft is
higher than the plain raft irrespective of the magnitude of settlement. The
settlement of piled raft in the initial stages of loading is much lesser than plain
raft. For the settlement of 2mm in loose sand the load on piled raft is 0.72kN
which is 110% higher than the plain raft load, whereas for the settlement of
20mm, the piled raft load is 35% higher. In other words the ratio between the
load taken by the piled raft and the plain raft progressively reduces with the
settlement. At the settlement of 20mm the excess load taken by the piled raft
varies between 30% and 35% for the three densities of sand. The trend seen
above in load sharing by the piled raft indicates that, in the initial stages of
loading, the addition of piles makes the system stiffer. The piles function as
settlement reducer, and the combined interaction between pile–raft-soils
makes the raft to take higher load under reduced settlement. However as the
load increases, the settlement of piled raft increases; this is due to the
reduction in soil- pile stiffness. This indicates that the provision of piles to the
raft is very effective when the settlements are less, and in particular settlement
92
less than 2% of the raft size tested. The load- settlement response explained
above is seen in sand of all the three densities tested. The characterisation
curve for the piled raft along with the corresponding plain raft for the three
densities have been presented in Figures 4.7 to 4.9. The results presented are
for identical test conditions (D = 200mm, t = 8mm, d = 10mm, L = 160mm
and N = 21) except the bed density.
1000
C Piled raft
Raft
-2
LOAD, kN x 10
100
B
A
CIRCULAR
LOOSE
10 O
0 5 10 15 20 25
SETTLEMENT, mm
Figure 4.7 Characteristic response of Plain raft and piled raft (loose
sand)
In the three phase behaviour, up to a load level of 0.6kN in the case

of loose sand, (the first phase) the combined system stiffness is high. Both the
piles and the raft are in the elastic phase, and the increase in the settlement is
very small. In the second phase reduction on the pile-soil stiffness becomes
high; hence the rate of change in the settlement with the load is higher. In the
third phase, the stiffness of the piled raft – soil reduced drastically and even
for a small increase in the load the piled raft settles rapidly. Beyond 20mm
settlement the piled raft settlement was very high even for a small increase in
the load. Similar trend was observed in the case of medium and dense sand
(Figures 4.8 and 4.9) although the load levels vary with the densities in all the
three phases.
93
Comparing the behaviour of the plain raft and the piled raft through
the load- settlement characterization curves, it can be seen that, an addition of
a small area (5.2%) of the piles to the raft, enhances the performance of the
foundation system. The addition of piles to the raft enhances the stiffness of
the combined system and it becomes far higher than the plain raft at any
particular level of settlement. The variation of stiffness of plain raft and the
piled raft is compared in the Table 4.3. This comparison indicates that the
stiffness of the combined foundation system reduces as the settlement
increases. It is seen that at the maximum settlement of 20mm (all the tests
were conducted upto the settlement of 20mm), the stiffness of the combined
system is very close to that of plain raft indicating that at higher settlement the
piles tend to behave as settlement reducer and not primarily a load bearing
member.
1000
Piled raft
Plain raft
C
-2
x210
B
x 10
kNkN
A
LOAD,
100
LOAD
CIRCULAR
MEDIUM DENSE
O
10
0 5 10 15 20 25
SETTLEMENT, mm
SETTLEMENT MM
Figure 4.8 Characteristic response of Plain raft and piled raft

(medium sand)
94
1000
Piled raft
Plain raft
B
C
A
-2
LOAD, kN x 10
100
CIRCULAR
O DENSE
10
0 5 10 15 20 25
SETTLEMENT, mm
Figure 4.9 Characteristic response of Plain raft and piled raft (dense
sand)
Table 4.3 Comparison of stiffness of plain and piled raft for various
settlements (D = 200mm, t = 8mm, L = 160mm and
d = 10mm, and N = 21)
Phase OA Phase AB Phase BC

Bed (N/mm) (N/mm) (N/mm)
Loose 195 380 137 197 98 130
Medium 600 1100 467 633 255 345
Dense 800 1700 617 800 314 410
The comparison was also done for the piled raft with area ratio of
2.75%. Nearly 50% of the piles were reduced in this case. However to
maintain symmetry in pile arrangement, 11 piles were provided, which has
resulted in the area ratio as 2.75%. Figure 4.10 presents the characterization
curves of plain raft and a piled raft with area ratio of 2.75%. The parameters
relating to the raft and the bed are kept as same in the tests on plain and piled
raft. It can be seen from the figure that with an addition of 11 piles to the raft,
95
the load carrying capacity of the combined system is higher than the plain raft
and the behaviour pattern remains same as that of the previous case (i.e.
Ar=5.2%) with three well defined phases. It is also seen in this case that as the
loading increases the difference between the capacity of the plain raft and the
piled raft reduces indicating that the influence of piles provided reduces at
higher load and remains constant beyond a particular level. Even though the
number of piles added is less in number, the behaviour pattern of the piled raft
remains almost the same. Here also, it can be seen that beyond a settlement
level of around 3% of the raft dimension, the soil piled raft stiffness reduced
to low value and the piled raft settles as that of plain raft. This indicates that,
even when the area ratio is quite small the pile group enhances the load
carrying capacity of the raft as a combined system. However, at higher
settlement even though the enhancement in the capacity is less, the pile group
contributes to settlement reduction.
1000
C
B d = 8mm
-2
LOAD, kN x 10
L = 200mm
A
N = 11
100
R.A. = 45º
D = 200mm
Bed = Medium dense
11 Piles
Piled raft
Plain
Plainraft
Raft
O
10
0 2 4 6 8 10 12 14 16 18 20 22 24
SETTLEMENT, mm
Figure 4.10 Characteristic response of plain raft and piled raft

(Area ratio 2.75%)
The first phase of the curve up to a settlement level of around 2mm

represents the elastic behaviour of the entire system. The second phase shows
(upto 6mm settlement) gradual loss of system stiffness (the pile group loses
its elastic behaviour) and beyond this stage the loss of stiffness is rapid and at
96
20mm (the maximum settlement at which all the tests were terminated) the
stiffness is close to that of plain raft. In other words, beyond a settlement level
of 3% of the least lateral dimension of the raft, the piled raft system behaves
more like plain raft.
4.4.3 Behaviour of Pile Group of Piled Raft
As seen from the previous discussion the introduction of a small

number of piles to the raft enhances the performance of the piled raft, and for
any given load, the settlement of the piled raft is lesser than that of the plain
raft. In order to understand the contribution of the pile group in enhancing the
stiffness behaviour of the piled raft, the behaviour of the pile group of piled
raft need to be studied by comparing it, with the behaviour of free standing
pile group. In the case of free standing pile group the raft would not be in
contact with the sand bed. Figure 4.11 compares the typical behaviour of pile
group of piled raft and free standing pile group tested in medium dense sand.
The dimensions of raft and piles are presented in the figure. Initially the
stiffness is high, indicating that the pile group resists the applied load by the
shaft friction. Once the friction is fully mobilized, the pile group settlement is
rapid indicating rapid reduction in the stiffness of pile-soil system.
LOAD, kN
0 1 2 3 4
0
2 FREE STANDING GROUP
4 PILED RAFT GROUP
6
SETTLEMENT, mm
8
10
d = 10mm
12
t == 8mm
d 8mm
14 L = 200mm
16 N = 21
R.A. = 36º
18
D = 200mm
20 Bed = Mediumdense
22
Figure 4.11 Comparison of load-settlement response of free standing

pile group and pile group of piled raft
97
Based on the results presented above, the settlement corresponding

to the load of 1.2kN was the yield point of the free standing pile group, which
is termed here as critical settlement (Scr) of the free standing pile group. The
load taken by the pile group corresponds to this settlement is known as
limiting friction (Fl). The magnitude of limiting friction is represented in
terms of the group capacity, corresponding to the reference settlement of 20
mm, which is 10% of the diameter of the raft and the critical settlement is
represented in terms of the pile length. The magnitude of limiting friction for
the free standing pile group was of the order of 75% of the final load taken by
the pile group, indicating that the load transfer was predominately by friction
generated by the interaction between the piles and the soil. The magnitude of
the critical settlement was observed as 1.3 mm which was 0.65% of the pile
length. It is seen that, in the case of free standing pile group, up to a load of
1.2kN the settlement is 1.3mm which is 0.65% of the pile length. Beyond this
load, even for a small increase in the load, the increase in the settlement is
rapid. This indicates that entire load applied up to that limit was resisted by
friction. Once the frictional resistance of piles mobilised fully, the pile group
loses its stiffness rapidly and settles, indicating the failure of the group. Thus
it can be said that in the case of free standing pile group on sand, the load
settlement response has two phases, which is similar to elastic strain
hardening behaviour. Initially the stiffness is higher indicating the pile group
resists the applied load by shaft friction. On mobilisation of complete friction,
the rate of settlement of pile group is rapid which indicates the reduction in
the stiffness of pile-soil system.
The load taken by the pile group of piled raft for any given
magnitude of settlement was computed from the load-settlement response of
piled raft and raft and is presented in Figure 4.11. The piles of piled raft
(where the raft was in contact with the bed) exhibited rapid increase in
resistance with settlement. This behaviour is almost same as free standing
piles. But the yield in the load-settlement curve occurred at a load of 1.8kN,
which is 50% higher than the yield load of free standing pile group. The load
98
of 1.8kN was the limiting frictional resistance of the piles of piled raft tested
in medium dense sand and the corresponding settlement of 2mm was the
critical settlement.
Further, the increase in load above 1.8kN induced a higher rate of

settlement. But this rate remained almost constant until the load which has
caused the piled raft settlement of 20mm. Further the rate of settlement with
load was lesser than the free standing piles for the load greater than the yield
load of piles of piled raft. In other words the pile-soil stiffness of piled raft
was higher than the free standing pile group for the loads higher than the yield
load of free standing piles. It can also be seen that the group capacity of piles
of piled raft is 140% higher than the piles of free standing piles. The increase
in capacity at critical settlement is 50%. Thus the piles of piled raft are
capable of carrying additional load even at settlements higher than the critical
settlement. The additional resistance offered by the piles even after the
mobilization of full friction (i.e. limiting friction) is attributed to increase in
confining stress on piles (radial stress) due to the load transferred to the soil
through the raft. The increase in load on the raft, increases the shaft resistance
of piles of piled raft indicating that the frictional resistance of piles beyond
limiting friction increases almost linearly with the load applied on the raft.
Similar observation was seen in the case of square and rectangular piled rafts
tested in the study.
The test results of free standing pile group and pile group of piled
raft are presented in a semi-log plot as shown in Figure 4.12. Both the curves
show distinct change in the slope of load-displacement response. The change
in shape is seen at the load corresponds to limiting friction. For the load
beyond limiting friction, the slope of the curve is constant. However the slope
of the curve of piles of piled raft is higher than the curve corresponding to
free standing pile.
99
10
LOAD, kN
1
d=8mm
d = 10mm
L=200mm
t = 8mm
N=21
L = 200mm
R.A. = 36deg
N = 11
FREE STANDING D=200mm
R.A. = 45º
PILED RAFT Bed Medium dense
D = 200mm
0.1 B d M di d
0 2 4 6 8 10 12 14 16 18 20 22 24
SETTLEMENT, m m
Figure 4.12 Characteristic response of freestanding pile group and

pile group of piled raft
From the Figure 4.12 it can be recognised that the load-settlement

response of piles of piled raft is two phase behaviour, which is similar to
elastic-work hardening behaviour. From the characteristic response of piles of
piled raft, it is clear that the pile-soil stiffness is high initially and it reduces
appreciably after the load on piles is more than the limiting friction. Thus the
piles involve effectively in load sharing in the initial part of load on the piled
raft. When the settlement of piled raft becomes more than the critical
settlement, they effectively act as settlement reducer, though they continue to
share some part of applied load. However the load shared by the piles reduces
with the settlement.
The behaviour of raft of piled raft is compared with the plain raft in
Figure 4.13. Both the plain raft and the raft of piled raft showed three phase
behaviour as reported in the section 4.4.2. However the raft – soil stiffness of
piled raft is higher than the stiffness of plain raft except in the first phase of
the characteristic behaviour of the raft. In the first phase, the raft-soil stiffness
of piled raft is lesser than the plain raft, whereas in the second and third
phases of curve show higher stiffness than the plain raft. This variation is due
to the interaction between raft and pile. Introduction of pile enhances the
overall stiffness of soil, and piles also share major portion in the initial part of
100
loading. On increasing the load, the raft shares major part of the load and this
is evident from the Figure 4.13, which shows higher stiffness for the raft of
piled raft. Thus the performance of raft confirms that the piles at higher load
performs as settlement reducer, which in term enhanced the stiffness of raft-
soil system.
1000
C
2
B
LOAD, kN x 10
100 A
RA FT OF PILED RA FT
PLA IN RA FT
10
O
0 2 4 6 8 10 12 14 16 18 20 22 24
SETTLEMENT, mm
Figure 4.13 Characteristic response of plain raft and raft of piled raft
4.4.4 Load - Settlement Behaviour
The previous sections have clearly established that the addition of

piles to the plain raft enhances the load carrying capacity of the raft as a
combined system at any given settlement. It was also seen that the rate of
enhancement of the magnitude of the load is settlement dependent, and this
causes the reduction of the stiffness of the combined system with the
settlement. Three phase behaviour was observed and the stiffness of the
system reduces progressively in each phase. The reduction of the stiffness
with the enhancement of load is mainly due to the load- settlement behaviour.
As the economics of the foundation system in reducing the settlement
depends upon the various parameters of the pile and raft for a given density
condition, the effect of various parameters relating to the pile and raft are
studied under different bed densities and presented in this section. The main
101
aspects of the load- settlement behaviour are the limiting friction and the
critical settlement. The critical settlement was found to be a function of the
length of the pile. The load corresponding to the critical settlement is the
limiting friction. The magnitude of the limiting friction is compared with the
load taken by the foundation system corresponding to a settlement of 20mm
which is 1% of the diameter of the raft.
Further the load- settlement curves obtained from the experiments

have been plotted in an non-dimensional form. In this, instead load Vs
settlement; the load ratio vs. settlement ratio has been plotted, where
Load at a particular settlement,Q

Load ratio, PR =
Load corresponding to 20 mmsettlement,Q u
Settlement corresponding to any load, δ

Settlement ratio δR =
Finalsettlement 20 mm, δu
The curves thus obtained are analysed and the relation between the
two non- dimensional parameters is arrived based on Chin’s method (Chin,
1970, and 1972) which is as follows
(δ R ) /(PR ) = mδ R + C (4.1)
where δR is the settlement ratio and PR is the load ratio. The inverse of the
slope l/m gives the limiting value of load ratio and constant c gives the
inverse of the initial stiffness ratio.
The above equation had been developed typically for all the cases
namely variation in density, length, pile diameter and pile – raft area ratio.
The effect of various parameters relating to pile, raft and the bed
density on the variation of different features of the load settlement behaviour,
102
namely the stiffness, critical settlement and the limiting friction are discussed.
The parameters studied are
• Bed density,
• Thickness of the raft
• Length of the pile,
• Diameter of the pile,
• Pile raft area ratio,
• Configuration of piles,
The load-settlement behaviour of the circular piled raft was studied

and compared by taking the load taken by the piled raft at the three reference
settlement levels. Accordingly the variation of stiffness at the three different
settlement levels is compared.
4.4.4.1 Effect of pile length
Figures 4.14 and 4.15 present the load settlement response for piled
raft with various lengths of the piles for two different diameters of the piles
namely 10mm and 8mm respectively. In both the cases the raft dia, and
thickness were kept as 200mm and 8mm respectively. Five different lengths
were used in the case of 10mm dia piles, namely 200mm, 160mm, 120mm,
100mm and 75mm. In the other case namely 8mm dia piles, four different
lengths have been used, namely 200mm, 160mm, 120mm, and 100mm. It can
be seen that the load settlement response of the piled raft with piles of various
lengths is similar, although the load taken at any settlement varied with the
length. In all the cases it is observed that for a given load, the settlement of
piled raft is smaller than that of the plain raft. It is seen from the Figure 4.14,
that introduction of piles whose length is as small as 75mm (0.375 times the
dia of the raft) makes the piled raft to take a higher load at any given
settlement.
103
LOAD, kN
0 1 2 3 4 5 6 7 8 9 10
0
d = 10mm
2 t = 8mm
N = 21
4
R.A. = 36º
6 D = 200mm
Bed = MD
SETTLEMENT, mm
10
12
14 PLAIN RAFT
200 mm
16
120 mm Piled raft
18 100 mm Pile length
20 75 mm
22
Figure 4.14 Load - settlement response of circular piled raft with

10mm dia piles for various pile lengths
Table 4.4 presents the load taken and the stiffness of the piled raft
with various length of the piles (Table 4.4a for 10mm dia, 4.4b for 8mm dia
and 4.4c for 6mm dia piles). The table presented above clearly indicates that
irrespective of the pile diameter, at lower level of settlement the stiffness of
the piled raft is far higher than that of the plain raft.
For example when the pile length is 100mm (0.5 times of the raft
dia), the stiffness of the piled raft is 850N/mm, 650N/mm and 500N/mm for
pile diameter 10mm, 8mm and 6mm respectively at 2mm settlement. This is
higher than the corresponding plain raft stiffness at 2mm settlement by 233%,
155% and 96% respectively. For the settlement of 20mm, the corresponding
increase in the stiffness values are 19%, 7.8% and close to 0% respectively.
When the length of the pile was 120mm, at 20mm settlement the increase in
the stiffness was 25%, 17.6% and 7.8% respectively. This indicates that as the
settlement increases, the piled raft stiffness approaches the plain raft stiffness
irrespective of the length and diameter of the pile.
104
Table 4.4 Effect of pile length on capacity and stiffness (N/mm) in

medium dense sand (D = 200mm, t = 8mm and N = 21)
(a) Pile dia 10mm

@ 2mm sett @ 6mm sett @ 20mm sett
Pile
Load in Load in Load in
Length Stiffness Stiffness Stiffness
kN kN kN
200 mm 2.80 1400 4.80 - 800 8.80 440
120 mm 2.20 1100 3.40 567 6.00 320
100 mm 1.70 850 3.10 517 6.10 310
75 mm 1.60 800 3.00 500 5.50 275
(b) Pile dia 8mm
Pile
kN kN kN
200 mm 1.90 950 3.45 575 6.80 340
160 mm 1.80 900 2.80 467 6.30 365
120 mm 1.70 850 3.20 517 6.00 300
100 mm 1.30 650 2.90 484 5.50 275
(c) Pile dia 6mm
Pile
kN kN kN
160 mm 1.40 700 3.00 500 5.70 285
120 mm 1.30 650 2.80 467 5.50 275
100 mm 1.00 500 2.10 350 5.00 250
It is also interesting to note that at a settlement of 3% of the raft

diameter, increase in stiffness varies from 96% to 148%, confirming the
above view. Hence it can be said that the effect of increase in the confining
stress due to the raft- soil interaction on the pile group exists up to a
settlement level of say 3 to 6% of the raft diameter and the thereafter its effect
is insignificant.
105
LOAD, kN
0 1 2 3 4 5 6 7 8
0
2 d = 8mm
t = 8mm
4 N = 21
R.A. = 36º
6
SETTLEMENT, mm
D = 200mm
Bed = Medium dense
8
10
12
14 PLAIN RAFT
16 200mm
160mm Piled raf t
18 120mm Pile length
20 100mm
22
Figure 4.15 Load - settlement response of circular piled raft for various
pile lengths with piles of 8mm diameters
It can also be seen that the load taken by the combined system has
the maximum value when the length of the pile is equal to the diameter of the
raft. Similar tests were carried out by changing the diameter of the pile to
8mm and 6mm.
Figure 4.15 presents the comparison for 8mm dia piles and similar
behaviour was observed in the case of 8mm dia piles also. As a matter of
confirmation the tests are repeated in loose sand and the results are presented
in the Table 4.5.
Table 4.5 The variation of stiffness (N/mm) in loose sand

(D = 200 d = 10mm Bed = Loose N = 21 t = 8mm)

Pile
kN kN kN
200 mm 0.80 400 1.35 225 2.70 135
160 mm 0.68 340 1.15 192 2.50 125
120 mm 0.62 310 1.08 180 2.00 100
106
It can be seen that the variation of stiffness occurs in the same

manner as in the case of medium dense sand. Hence it can be concluded that
the load response of piled raft with piles of different lengths is similar in all
the bed density conditions. The typical characterisation curve of the piled raft
shown with various pile lengths are given in the Figure 4.16 for a pile
diameter of 10mm.
10
B
A
LOAD, kN
1
D = 200mm
Pile length
t = 8mm
200mm
d = 10mm
120mm N = 21
O 75mm Medium dense
0.1
0.0 5.0 10.0 15.0 20.0 25.0

SETTLMENT, mm
Figure 4.16 Characteristic response of piled raft for various pile lengths
The characterisation curve indicates that irrespective of the pile

length, the behaviour has three phases. Although the settlement up to which
the linear elastic stage (portion OA of the curve) remains same as 1mm, the
load corresponding to this varies. In the case of 200mm it is 2.7kN, for
120mm it is 2.0kN, and for 75mm it is 1.3kN. In other words the values are
30%, 28% and 22% of the load corresponding to 20mm settlement of the
piled raft for the corresponding length. As can be seen at higher length the
linear behaviour extends nearly upto 30% of the load taken by the piled raft
corresponding to settlement level corresponding to 10% of the pile length.
The second stage of the curve AB is the stage where the behaviour tends to
become elasto- plastic, which extends up to a settlement level 9mm for
107
200mm long pile 7.5mm for 120mm long pile and 4.5mm for 75mm long pile.
The variation in stiffness for different lengths is presented in Table 4.6.
Table 4.6 The variation of stiffness (N/mm) (D = 200mm, d = 10mm,

Bed = Medium dense, N = 21, t = 8mm)
Stiffness at various phases

Pile length mm
200 2800 471 239
120 2100 343 217
75 833 213 96
The reduction in the stiffness in the region AB is very high as seen

from the Table 4.6 and beyond the level of B the foundation stiffness is close
to that of plain raft. This indicates that at a settlement level of 4.5% to 6% of
the length of the pile, the piled raft losses its stiffness indicating that the
piles do not involve effectively in load sharing and functions mainly as
settlement reducer.
Figure 4.17 presents the non dimensional plot where in the load
ratio PR and settlement ratio δR has been plotted for various lengths of piles of
piled raft. The plot indicates that irrespective of the length, the settlement
ratio remains same for any given load ratio. The relation between load ratio
and settlement ratio appears close to rectangular hyperbolic relation. Thus the
non linear part of the curve has been taken and plotted as hyperbolic curve in
Figure 4.18 to arrive relationship between load ratio and settlement ratio and
the relationship between them is given below.
δR
PR = (4.2)
mδR + C
where, m = 0.689 and C = 0.35.

108
LOAD RATIO
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00
0.00
0.10 d = 10mm
t = 8mm
0.20 N = 21
R.A. = 36º
SETTLEMENT RATIO
0.30 D = 200mm
0.40 Bed - Medium dense
0.50
0.60
0.70
0.80 200mm
0.90 120mm
75mm
1.00
Figure 4.17 Non-dimensional plots for various lengths
1.00
0.90
SETTLEMENT RATIO/LOAD RATIO
0.80
d = 10mm
0.70 t = 8mm
0.60 N = 21
R.A. = 36º
0.50 D = 200mm
Bed -Medium dense
0.40
0.30
200mm
0.20 120mm
75mm y = 0.6889x + 0.35
0.10 Average
0.00
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00
SETTLEMENT RATIO
Figure 4.18 Hyperbolic plots for various pile lengths
It can be seen from the Table 4.4 that in the initial stage
(settlement 2mm) the load taken by the piled raft varied from 23% (pile
length 75mm) to 115% (pile length 200mm) in the case of 10mm dia pile. It is
to be seen here that the load taken by the piled raft is not linearly proportional
to pile length even though the diameter is the same. At the final stage, the
109
load taken varies from 8% (pile length 75mm) to 72% (pile length 200mm).
Similar trend was observed in the case of 8mm dia and 6mm dia piles also.
Although there is an indication of enhancement of confining stress to the piles
due to raft-soil interaction, it is not contributed effectively to the pile group
resistance when the length is shorter than 0.8 D where D is the diameter of the
raft.
4.4.4.2 Effect of pile diameter
The load settlement behaviour of piled raft with 160mm long pile
and 8mm thick raft for the three different diameters namely 10mm, 8mm and
6mm were studied and presented in Figure 4.19. The characterization curve is
given in Figure 4.20. The Table 4.7 presents the load taken by the piled raft
for various diameters used, corresponding to the three different settlements,
along with the corresponding stiffness.
Table 4.7 Effect of pile diameter on capacity and stiffness (N/mm) in

medium dense sand (D = 200mm t = 8mm N = 21 and
L = 160mm)
Pile @ 2mm sett @ 6mm sett @ 20mm sett

d/t Load in Load in Load in
Dia Stiffness Stiffness Stiffness
ratio kN kN kN
6 mm 0.75 1.40 700 3.00 233 5.70 285
8 mm 1.0 1.80 900 2.80 667 6.30 315
10 mm 1.25 2.20 1100 3.80 900 6.40 320
As seen in the case of length variation, here also the stiffness

reduces with settlement and at 20mm settlement, the stiffness of the system
approaches the magnitude of the plain raft stiffness irrespective of the pile
diameter. In the initial stages of settlement, the stiffness variation is
proportional to the diameter, but as the settlement increases this
proportionality does not exist and the reduction is steeper. The increase is
110
27% in the case of 10mm dia piles and 14% in the case of 8mm dia piles at
20mm settlement. In the case of piles of 6mm diameter, it is almost close to
that of plain raft, which indicates that the piled raft with smaller diameter
piles loses its stiffness much faster than large diameter piles particularly at
higher settlements.
LOAD, kN
0 1 2 3 4 5 6 7 8
0
2 t = 8mm
L = 160 mm
4 N = 21
R.A. = 36º
6 D = 200mm
Bed = Medium dense
8
SETTLEMENT, mm
10
12
14
16
PLAIN RAFT
18
10mm
Piled raft
20 8mm
pile diameter
6mm
22
Figure 4.19 Load-settlement response of circular piled raft for various

pile diameters
For a better representation, the diameter (d) is considered in relation

with the thickness of the raft (t) as d/t ratio. It is seen that as the d/t ratio
increases, the load taken by the piled raft increases. A study of the load-
settlement curve indicates that in the initial stages of settlement (i.e. at 2mm)
the difference in the load taken by the piled raft having pile group of d/t ratio
0.75 and 1.0 is 28% and 0.75 to 1.25 is 57%. Further with increase in the
settlement, the effect of d/t ratio gets pronounced and offers additional
resistance. However it can be seen that even though the change in the
diameter is 66% between 6mm and 10mm, the increase in the load taken by
the piled raft is only 12.3% at a settlement level of 20mm. At the reference
settlement of 2mm, the increase in the load taken by the piled raft is 57%, at
6mm settlement it is 26.7%. This indicates that as the settlement level
111
increases, the contribution of the pile group in taking the load decreases
irrespective of the diameter. Similar trend was observed in the case of other
lengths and other bed densities also.
10
B C
A
LOAD, kN
D = 200mm
Pile diameter
t = 8mm
10mm
L = 160mm
8mm N = 21
O 6mm Medium dense
0.1
0.0 5.0 10.0 15.0 20.0 25.0

SETTLMENT, m m
Figure 4.20 Characterisation curves for various pile diameter
Studying the characterization curve shown in Figure 4.20, it

indicates that the initial elastic stage exists up to a settlement level of 1.0mm
practically for all the three diameters studied. However the corresponding
load varies as 1.1kN, 1.5kN and 1.8kN for 6mm, 8mm and 10mm diameter
piles respectively. The second stage namely the elasto- plastic stage extends
up to 6mm and at this stage the loading is 3.0kN, 3.6kN and 4.20kN. The
Table 4.8 presents the variation in the stiffness corresponding to the three
stages.
It can be seen from the stiffness variation, that as the settlement

increases, the stiffness decreases rapidly and beyond the settlement of 6mm
the stiffness of the system reduces rapidly and approaches to the stiffness
value of plain raft. Figure 4.21 presents the non-dimensional plot of load
ratio Vs settlement ratio. A unique relation is seen between them irrespective
of the diameter.
112
Table 4.8 The variation of stiffness (N/mm) of piled raft with

piles of different diameters (D = 200m l = 160mm Bed =
Medium dense N = 21 t = 8mm)
Pile diameter Stiffness at various phases

mm Phase OA Phase AB Phase BC
10mm 1800 660 270
8mm 1500 400 275
6mm 1100 360 290
LOAD RATIO
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00
0.00
0.10 d = 10mm
t = 8mm
0.20 L = 160
N = 21
SETTLEMENT RATIO
0.30 R.A. = 36º

0.40 D = 200mm
0.50
0.60
0.70
0.80 8mm
0.90 10mm
6mm
1.00
Figure 4.21 Non-dimensional plots for various diameter
The nonlinear response of the curve has been plotted as a

hyperbolic plot in Figure 4.22 and the relation between them is given in
Equation 4.3.
δR
PR = (4.3)
0.674δ R + 0.36
The m and c values obtained with reference to the pile diameter are
0.36 and 0.674 respectively, which are almost equal to the respective values
of m and c obtained with reference to the length.
113
This indicates that the load-settlement response of piled rafts tested

in this study is almost identical and is independent of pile length and pile
diameter.
1.00
0.90
0.80 d = 10mm
t = 8mm
0.70 L = 160
N = 21
0.60
R.A. = 36º
0.50 D = 200mm
0.40
0.30
8mm
0.20 10mm
6mm y = 0.6737x + 0.36
0.10 Average
0.00
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00
SETTLEMENT RATIO
Figure 4.22 Hyperbolic plots for various diameter
4.4.4.3 Effect of raft thickness
The effect of raft thickness on the load settlement behaviour has

been compared for two different diameters of piles of 6mm and 10mm. Figure
4.23 presents the load-settlement curves for piled raft with 6mm dia piles and
Figure 4.24 presents the load- settlement curves for piled raft with 10mm dia
piles.
Both the Figures 4.23 and 4.24 indicate that the raft thickness does
not have significant influence on the behaviour of the piled raft. However
when larger diameter pile is used, the piled raft takes higher load indicating
that higher diameter piles with higher raft thickness takes more load. This
behaviour was observed particularly with the raft thickness equal to the pile
diameter. It is also seen from both the figures that up to a settlement level of
6mm, the load taken by the piled raft remains more or less same irrespective
of the pile diameter used.
114
LOAD, kN
0 1 2 3 4 5 6 7 8
0
d = 6mm
2
L = 160
4 N = 21
R.A. = 36º
6 D = 200mm
SETTLEMENT, mm
10
12
14
Piled raft
16 Raft thickness
10mm raft
18
8mm raft
20
6mm raft
22
Figure 4.23 Load - settlement response of piled raft for various raft
thickness for pile diameter 6mm
LOAD, kN
0 1 2 3 4 5 6 7 8
0
d = 10mm
2 L = 160
N = 21
4
R.A. = 36º
6 D = 200mm
Bed = Medium dense
8
SETTLEMENT, mm
SETTLEMENT, MM
10
12
14
16
6mm raft
18 Piled raf t
8mm raft
Raft thickness
20 10mm raf t
22
Figure 4.24 Load-settlement response of piled raft for various raft

thickness for pile diameter 10mm
115
Table 4.9 Effect of raft thickness on load taken by piled raft and
stiffness (N/mm) in medium dense sand
(a) Pile diameter 6 mm

@ 2mm settlement @ 6mm settlement @ 20mm settlement
Raft
thickness Stiffness Stiffness Stiffness
kN kN kN
6mm 1.40 720 3.10 425 5.60 178
8mm 1.50 750 3.20 425 5.80 186
10mm 1.50 750 3.10 400 5.80 193
(b) Pile diameter 10 mm
Raft
thickness Stiffness Stiffness Stiffness
kN kN kN
6mm 2.00 1000 3.55 388 6.70 225
8mm 2.20 1100 3.60 350 6.80 228
10mm 2.30 1150 4.10 450 6.90 200
Table 4.9 presents the stiffness variation of the piled raft. It is seen
that in the case of piled raft with 6mm dia pile, the stiffness is marginally
lesser than the piled raft with 10mm dia piles at any given settlement level.
The stiffness of the piled raft with 10mm thick raft and 10mm dia pile is
marginally higher than the other cases indicating that the influence of the raft
thickness is not significant for the piled raft and loading system considered in
this research.. Therefore most of the studies are carried out with piled raft
thickness of 8mm and d/t ratio of 1.25.
4.4.4.4 Effect of pile raft area ratio
The load-settlement behaviour of the piled raft under various pile-

raft area ratios was studied. The pile –raft area ratio Ar, which is defined as
the ratio of pile group area (sum of the area of the piles Ap) to the raft area A
116
and was varied from 4.25% to 9.25%. The corresponding radial angles are
changed from 200 to 450 respectively. However the diameter of the pile
(10mm), length (160mm) and the raft thickness (8mm) were maintained
constant for all the cases.
The load-settlement curve for all the cases is presented in the

Figure 4.25. The figure indicates that up to a settlement level of 2mm the
variation in the load is small in the case of piled rafts with area ratio varying
from 4.25% to 6.25%. When the area ratio is 9.25%, the variation in the load
taken increases.
LOAD, kN
0 1 2 3 4 5 6 7 8
0
d = 10mm
2 t = 8mm
L = 160mm
4 D = 200mm
Bed = Medium dense
6
SETTLEMENT, mm
10
12
14
PLAIN RAFT
16 20 deg 37 piles
30 deg 25 piles Piled raft
18 Radial angle/
36 deg 21 piles
20 No. of piles
45 deg 17 piles
22
Figure 4.25 Load-settlement response of piled raft for various radial

angles (area ratio)
The area ratio adopted are presented in Table 4.10 along with the
load taken by the piled raft at the three different settlements.
117
Table 4.10 Effect of pile raft area ratio on load taken by piled raft and
stiffness (N/mm) (D = 200mm, t = 8mm, d = 10mm,
L = 160mm)
Pile – raft @ 2mm sett @ 6mm sett @ 20mm sett

Area ratio Load in Load in Load in
% Stiffness Stiffness Stiffness
kN kN kN
9.25 3.00 1500 4.70 783 7.90 395
6.25 2.70 1350 4.40 733 7.00 350
5.25 2.20 1100 3.70 617 6.10 305
4.20 1.70 850 3.30 550 5.70 285
It is seen from the table that as the area ratio increases the load
taken by the pile group increases. However this increases is not uniform. The
proportion of the increase in the load taken reduces as the area ratio increases.
For example, when the area ratio increased up to 4.2%, the increase in the
load taken is almost 100%, whereas from 4.2% to 5.2% it is 29%. From 5.2%
to 6.25% it is 22% indicating that the increase in the pile-raft capacity reduces
as the area ratio increases. This trend is the same at all the settlement levels.
The characterization curve for the behaviour of the piled raft for
three area ratios has been plotted in Figure 4.26. It is seen from the curves that
irrespective of the area ratio, the linear behaviour exists up to a settlement
level around 1.5mm. In other words in the initial stages the relative stiffness
of the piled raft is very high. As the loading increases up to a settlement level
of 6mm, the loss of stiffness is gradual, and after 6mm even for a small
increase in the load, the increase in the settlement is very high; in other words
the loss of stiffness in the case of piled raft is very rapid.
Table 4.10 present the magnitude of stiffness at the three reference

settlements of 2mm, 6mm and 20mm. It is seen from the table that the
118
stiffness of the foundation system is far high compared to that of the plain raft
up to a settlement level of 1.0mm, for all the cases.
10
B
LOAD, kN
A
1
Area ratio D = 200mm

t = 8mm
9.25%
L = 160mm
6.25% d = 10mm
O 4.25% Medium dense
0.1
0.0 5.0 10.0 15.0 20.0 25.0

SETTLMENT, m m
Figure 4.26 Characterisation curves for various area ratios
But as the settlement increases the stiffness reduces rapidly and at

20mm, even for the high area ratio of 9.25% the stiffness of the combined
system is higher only by 56%. In the case of area ratio 6.25%, the increase is
35%. The higher stiffness indicates that the system behaves more close to that
of fully piled system. The variation in the piled raft stiffness at the different
stages of behaviour is compared in Table 4.11.
Table 4.11 The variation of stiffness (N/mm)
Stiffness at various phases

Area ratio %
9.25 2900 420 280
6.25 2600 390 220
4.25 1600 340 170
It can be seen that in the linear elastic stage (unit OA) the stiffness
is very high compared to that of the plain raft. The line AB presents the
119
reduction in the stiffness in the elasto-plastic region. However when the pile-
raft area ratio is 9.25% the stiffness remains far higher than the plain raft. The
line BC indicates the plastic behaviour and the stiffness is close to that of the
plain raft. However the stiffness of the piled raft with pile-raft area ratio of
9.25%, still remained higher by more than 50% indicating that increasing the
number of piles will tend to convert the piled raft to a fully piled foundation.
Table 4.10 clearly indicates that at higher load as the settlement increases the
stiffness of the system approaches the stiffness value of the plain raft. The
higher increase in the stiffness for the area ratio of 9.25% indicates that the
system perhaps behaves as fully piled, as the stiffness is almost 57% more
than the plain raft unlike when the area ratio is 5.2%, it is only 12% more.
4.4.4.5 Variation in bed density
The load-settlement behaviour of piled raft was studied under three

different densities viz loose (φ=35), medium dense (φ=37.5) and dense
(φ=42.5). The number of piles was kept as 21(area ratio 5.2%) and the raft
thickness was kept as 8mm. The length and dia of the piles were kept as
160mm and 10mm respectively. The Figure 4.27 presents the load settlement
behaviour for the three densities tested.
LOAD, kN
0 1 2 3 4 5 6 7 8 9 10
0
2 d = 10mm
4 t = 8mm
L = 160mm
6
N = 21
8 R.A. = 36º
SETTLEMENT, mm
10 D = 200mm
12
14
16
Bed density
18
Dense
20
Medium Dense
22 Loose
24
Figure 4.27 Load-settlement response of piled raft for various bed

densities
120
In the case of loose sand for 10mm dia pile the limiting friction was
of the order of 0.50kN which is 18% of the piled raft capacity corresponding
to 20mm settlement, whereas in the case of medium dense sand it was found
to be 32% of the corresponding piled raft capacity. In dense sand it was 33%.
However the piled raft capacity at 20mm settlement was much higher in
medium dense and dense sand than in the case of loose sand. The magnitude
of critical settlement (0.6mm) in the case of loose sand was much lower than
the value corresponding to medium dense and dense sand (1 to1.2mm)
although the variation of limiting friction magnitude between medium dense
and dense sand was smaller. Figure 4.28. presents the characterization curve
for the load settlement behaviour of piled raft for the various density
condition. The three stages of the behaviour is quite distinct and it can be
noted that the loss of stiffness in the case of loose sand is very rapid. Table
4.12 presents the variation of the stiffness under the three stages of load-
settlement response.
10
C
B
LOAD, kN
A
1
D = 200mm
Bed density
t = 8mm
Loose
d = 10mm
Medium L = 160mm
O Dense N = 21
0.1
0.0 5.0 10.0 15.0 20.0 25.0

SETTLMENT, m m
Figure 4.28 Characterisation curves for various bed densities

121
Table 4.12 The variation of stiffness (N/mm) of piled raft for different
bed densities
Stiffness in N/mm
Bed density
Loose 600 200 160
Medium dense 1800 620 210
Dense 2700 640 250
It can be seen that although the settlement level for the change in
the behaviour remains same in the case of medium dense and dense sand, the
load level varies resulting in the variation in stiffness value. In all the three
cases of densities the stiffness reduces with the increase in the settlement.
Beyond the settlement level of 3% of the raft diameter the stiffness of the
piled raft reduces to that of plain raft, in all the three bed density conditions.
Figure 4.27 indicates that at any given settlement the load taken by
the piled raft on medium dense and dense sand bed is higher than the loose
sand. The characterisation curve presented in Figure 4.28 indicates that the
piled raft on loose sand losses its stiffness at a very early stage of loading
itself. In the case of medium dense and dense sand, the linear elastic stage
limits up to a level of 1mm. While the elasto- plastic stage exist up to 6 to
6.5mm settlement in medium dense sand, in the case of dense sand, this stage
extends up to 7.5mm. Thereafter in both the cases of medium dense and dense
sand, the stiffness reduces rapidly and at 20mm the stiffness level remains
very close to that of plain raft of respective density.
122
LOAD RATIO
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00
0.00
0.10 d = 10mm
t = 8mm
0.20 L = 160mm
N = 21
SETTLEMENT RATIO
0.30 R.A. = 36º

0.40 D = 200mm
0.50
0.60
0.70
0.80 Dense
0.90 Medium Dense
Loose
1.00
Figure 4.29 Non-dimensional plots for various bed densities
The non- dimensional plot presented in Figure 4.29 indicates that at

any given load ratio, the settlement ratio is same for the medium dense and
dense sand but in the case of loose sand the settlement ratio remains lesser
than the medium dense and dense sand. The hyperbolic plot of the load –
settlement response of the piled raft tested in sand of three densities are
presented in Figure 4.30.The plot shows better fitting for the medium dense
and dense sand and the relation is as follows:
δR
PR =
0.678 δ R + 0.36
(4.4)
From the hyperbolic plot of load-displacement response of piled

raft, it is recognised that, the load displacement response of piled raft is
almost identical irrespective of the pile length, pile diameter and the bed
density.
123
1.00
0.90

0.80 d = 10mm
t = 8mm
0.70 L = 160mm
N = 21
0.60
R.A. = 36º
0.50 D = 200mm
0.40
0.30
Dense
0.20 Medium Dense
Loose y = 0.6778x + 0.36
0.10 Average
0.00
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00
SETTLEMENT RATIO
Figure 4.30 Hyperbolic plots for various bed densities
4.4.5 Settlement Reduction Behaviour
The piled raft as a foundation system aims at reducing the

settlement of the foundation rather than eliminating it in total. The optimum
design of piled raft must be such that with minimum number of piles with the
required thickness of raft must reduce the raft settlement to the acceptable
limit for the given loading. This settlement reduction, that can be achieved
will be influenced by the number, diameter, and the length of the pile for the
given raft thickness and bed density. The effect of various parameters
associated with the pile, raft and the bed material have been analysed
independently. The settlement reduction behaviour is quantified by a factor
called settlement reduction ratio SR and is defined as below:
(4.5)
S r − S pr
SR =
Sr
where SR = Settlement reduction ratio

Sr = Settlement of the plain raft at any given load.
Spr = Settlement of the piled raft at the same load considered
for the plain raft Settlement.
124
The settlement reduction ratio is computed for two different levels

of piled raft, which are 2mm (less than critical settlement) and 20mm.These
are presented for piles of various lengths and diameter, area ratio of piled raft
and density of sand.
4.4.5.1 Effect of variation in length
In order to study the influence of pile length on settlement

reduction ratio, SR, the pile length is normalised by the diameter of the raft
(D) and plotted as SR v/s L/D ratio. A typical non-dimensional relationship
between SR and L/D ratio is presented in Figure 4.31 for the piled rafts tested
with piles of 10mm, 8mm, and 6mm diameters in medium dense sand. The SR
value increases with the increase in L/D ratio for all diameters of pile
irrespective of the settlement of piled raft. Among the two settlement values
compared, the SR value is higher for lesser settlement irrespective of the L/D
ratio and pile diameter. For a given settlement of piled raft, the rate of
increase of SR value with L/D is not constant, and the rate is decreased
beyond certain L/D ratio. This trend is seen for the pile diameter of 10mm and
8mm. In the case of 6mm dia piles the relation between SR and L/D is linear
and the rate of increase is constant for the range of L/D ratios of piles tested.
1.00
t = 8mm
0.80 d = 10mm
N = 21
SRSVALUE
0.60 D = 200mm
Bed = Medium dense
R
0.40
0.20
2mm 20mm
0.00
0.2 0.4 0.6 0.8 1.0 1.2
L/D
Figure 4.31a Settlement reduction ratio versus L/D ratio for pile
diameter 10mm
125
0.80
t = 8mm
d = 8mm
0.60 N = 21
SR VALUE
D = 200mm
Bed = Medium dense
SR
0.40
0.20
2mm 20mm
0.00
0.2 0.4 0.6 0.8 1.0 1.2
L/D
Figure 4.31b Settlement reduction ratio versus L/D ratio for

pile diameter 8mm
R
0.40
t = 8mm
d = 6mm
0.30 N = 21
SRSVALUE
D = 200mm
0.20 Bed = Medium dense
R
0.10
2mm 20mm
0.00
0.4 0.5 0.6 0.7 0.8 0.9
L/D
Figure 4.31c Settlement reduction ratio versus L/D ratio for pile
diameter 6mm
Further the SR value is higher for the piled raft with piles of higher
diameter which is independent of length of pile and magnitude of piled raft
settlement. This is attributed to higher surface area of number of piles
provided and higher pile-soil stiffness. Similar observations are made for the
tests on soils of other two densities.
From the results presented above it can be said that the relation
between SR and L/D is bilinear. The piles of L/D ratio less than 0.6, follows
one relation and the piles with L/D ratio greater than 0.6 follows the other
relation. The rate of increase of SR with L/D is higher for piles with L/D less
126
than 0.6 than the piles with L/D ratio greater than 0.6. Higher rate of SR for
piled raft with shorter piles (ie less than 0.6) is attributed to the confining
effect due to the surcharge load from the raft, which is higher in the soil close
to the base of the raft.
This observation shows that minimum length of pile required to

control the settlement effectively in medium dense sand for the raft subjected
to uniformly distributed load is 0.6D (D=diameter of the raft. The length of
pile more than 0.6D certainly has better control on settlement reduction, but
its contribution towards settlement reduction is not appreciable on increasing
the pile length more than 0.6D in the case of medium dense sand. However
this condition is not satisfied for the smaller diameter of the pile (ie.pile of
6mm diameter). The critical length of the pile for the dense and loose
conditions also is in the similar range.
The effect of pile diameter on the settlement reduction ratio has

been shown in Figure 4.32.
R
1.00
t = 8mm
0.80 L = 160mm
N = 21
VALUE
0.60 D = 200mm
Bed = Medium dense
SR S R
0.40
0.20
2mm 20mm
0.00
0.50 0.75 1.00 1.25 1.50
d/t RATIO
Figure 4.32 Variation of SR value with pile diameter
The pile diameter variation has been represented as d/t ratio namely
ratio of pile diameter (d) to raft thickness (t). The variation of SR value has
been presented for the piled raft settlements of 2mm and 20mm. From the
curve it is seen that SR value increases with the increase in the d/t ratio
127
irrespective of the level of the settlement, and the rate of increase decreases
with increase in the d/t ratio. At higher diameter of the pile, the relation
between SR and the d/t ratio tends to show asymptotic response, which
indicates that for a given thickness of raft, increasing the diameter of the pile
higher than 1.25 times the thickness of the raft may not be very effective in
controlling the settlement.
It is also seen that in the initial stages of loads, the diameter of the
pile has pronounced effect on controlling the settlement due to the fact that
the load transfer to the pile is significant. As the load increases the raft shares
major part of the load and hence SR is lesser for higher settlement and the rate
of increase in SR with d/t is lesser because of the above said reason. It is
evident that in the settlement reduction, the role of diameter is only in the
initial stages of load at which the pile group takes the maximum load. The
response indicated above is seen in the case of piled raft with other pile
lengths and the densities of the sand bed.
4.4.5.3 Effect of area ratio
Figure 4.33 presents the effect of pile-raft area ratio on the SR value
for the tests conducted in medium dense sand. The plot provides the variation
of SR value at two different settlement level as in the previous cases, namely
at 2mm and 20mm.
1.00
2MM 20MM
0.80
0.60
SR
0.40
0.20
0.00
4.00 5.00 6.00 7.00 8.00 9.00 10.00
Area Ratio A R
Figure 4.33 Variation of SR value with piled raft area ratio in medium
dense sand
128
The SR value increases rapidly as the area ratio increases from

4.25% to 6%. However the increase in the SR value when the area ratio
increased from 6% to 9.25% is smaller than the earlier variation. From the
Figure 4.33 it is seen that an area ratio of 5.25% to 6% offers good control on
settlement reduction, and increasing the number of piles beyond this limit
increases the SR. But the relative improvement is not commensurate with the
increase in the number of the piles. This response clearly indicates that, if the
area ratio of the piled raft is more than 6%, the increase in the efficiency of
the piled raft is reducing with the settlement. Tests conducted on dense sand
and loose sand also exhibited almost identical behaviour.
4.4.6 Load Sharing Behaviour
While the performance of piled raft is assessed based on reduction

in settlement of combined system, the economy in design is arrived based on
number of piles provided. This is largely decided on load share permitted by
the designer between the piles and the raft. For the assumed share of load,
number of piles are decided based on the individual capacity of pile and
analysed to know the reduction in settlement as well as to validate the
assumption made in the initial design with reference to load share. The load
share between the two bearing elements (piles and raft) of piled raft is
influenced by the parameters of pile and raft including settlement of piled raft.
But the load share is influenced much by the level of settlement of piled raft.
In the initial stages of loading, the overall settlement of piled raft is small;
hence major portion of applied load is shared by the piles. On increasing the
load, settlement is increased and the load share of raft is also increased. At
higher values of settlement (i.e. settlement due to plastic deformation of
sand), the percentage share of load by the piles remains almost constant. This
aspect of piled raft is analysed further in this section in terms of load sharing
ratio, αpr. The ratio αpr is defined as the ratio between the amount of load
(shaft resistance + base resistance of all piles) shared by the piles at a given
settlement of piled raft (Qp) to the total load on the piled raft causing same
settlement (Qpr).
129
Qp
α pr = (4.6)
Q pr
where Qp = Qpr – Qr and Qr = load shared by the raft at the same settlement.
It is to be referred that the value of αpr = 0 for a raft and equal to

unity for a pile group, but for the piled raft the αpr may take any value
between zero and one. It depends on density of sand bed, length of piles,
diameter of piles, thickness of raft, settlement of raft etc. The load sharing
ratios derived from the experiments conducted are presented for various
settlements, and influence of pile and raft parameters on αpr for a given
settlement is analysed. In the analysis of αpr with settlement, total pile
resistance is considered as stated above. In general, pile derives its resistance
both from frictional resistance on shaft and base resistance, therefore
information on share of pile load between shaft and base of pile is very much
useful for the design of pile. However the study is not focused on base
resistance of the pile. The model pile is assumed as friction pile since the
surface area of pile is manifold higher than its base area and is resting on a
stratum, which is not hard. Thus the contribution from base resistance to pile
capacity will be far less than the contribution from shaft resistance.
4.4.6.1 αpr Versus Bed Density
In order to understand the influence of density of sand on load

sharing ratio (αpr) the results of tests conducted on model piled raft of
diameter 200mm with pile of diameter 10mm and length of 160mm
embedded in loose (14.5 kN/m3), medium dense (15.5 kN/m3) and dense sand
(16.5kN/m3) conditions are compared in Figure 4.34.
The figure presents the variation of αpr with settlement of piled raft.
The value of αpr of pile group is high at initial loads (ie at low settlement) and
falls down with the settlement. This trend indicates that the proportion of load
shared by the piles reduces as settlement of piled raft increases due to
130
continuous loading. Similar trend was observed by Horikoshi and Randolph

(1996) when performing tests on piled raft in centrifuge and 1g model test on
piled raft in sand by Turek and Katzenbach (2003).
∝ pr
0.00 0.20 0.40 0.60 0.80 1.00
0
4 d=10mm
L=160mm
6 D=200mm
t=8mm
SETTLEMENT, mm
10
12
14
16
18 Loose
medium
20 Dense
Figure 4.34 Variation of αpr with settlement for various densities
The reduction in αpr with settlement is rapid initially and attains

almost constant value for the piled raft settlements more than 8mm, which is
equal to 5% of the pile length used in this set of test. Further the reduction is
rapid till the settlement close to 3mm and is equal to 1.5% of the diameter of
the raft. The trend indicated above is seen in all the three densities of sand
(see Figure 4.34), which indicates that the load sharing ratio of pile group of
piled raft is almost same for a given settlement. However marginal variation
is seen between the densities, especially at higher settlements. Overall it can
be said that, for a given set of conditions of piled raft, the effect of density on
load share between the piles of piled raft and the raft is almost same in the
case of circular piled raft tested under uniformly distributed load.
131
0.80
d=10mm
n=21
0.60 l=160mm
Bed = Medium dense
∝ pr
0.40
0.20
2mm 20mm
0.00
0.00 0.02 0.04 0.06 0.08 0.10 0.12
Krs(e)
Figure 4.35 αpr v/s Krs(e) based on Es
The variation of αpr with raft-soil stiffness ratio (Krs(e)) is shown in

Figure 4.35 for the settlements of 2mm and 20mm. It is seen from the figure
that the raft – soil stiffness has marginal influence on αpr values particularly
for the settlement less than critical settlement (ie < 3mm). For the settlement
of 2mm, the piles carried 50% of the total load irrespective of the bed density,
whereas for the settlement of 20mm αpr is reduced to 25%. However reduction
in αpr is slightly higher in loose sand and the load share of piles is 20% of the
load corresponds to settlement of 20mm. From the results presented and
analysed, it is clear that density of sand has very little effect on load share
between piles and raft hence, effect of other parameters (length of pile,
diameter of pile, thickness of raft etc.) on αpr are presented only for the
medium dense condition.
4.4.6.2 αpr versus pile lengths
As stated in the previous section the load share ratio (αpr) is

required for the design of piled raft. The relative influence of length of pile
and diameter is of primary interest to a design engineer to arrive at an optimal
design. In this section effort of pile length on αpr are presented and discussed.
The typical variation of αpr with settlement for the piled raft of
200mm diameter with pile lengths of 200mm, 160mm, 120mm, 100mm and
75mm embedded in medium dense sand (φ = 37.5ο) is presented in Figure
132
4.36. The results presented are for the pile group of 21 piles with diameter of
pile of 10mm and raft thickness of 8mm.
α pr
0.0 0.2 0.4 0.6 0.8 1.0
0
4
t = 8mm
6
d= 10mm
SETTLEMENT, mm
8 N = 21
R.A. = 36º
10
D = 200mm
12 Bed = Medium dense
14 Pile length
16 200 mm
160 mm
18 120 mm
100 mm
20
75 mm
22
Figure 4.36 Settlement v/s LS ratio αPR for 10mm dia pile
From the Figure 4.36 it can be seen that for a given length of pile,
αpr value is maximum at the initial stage of loading where, the settlement of
the piled raft is less and elastic. On increasing the load on the piled raft the
settlement increases and in turn the load share ratio of piled raft system
decreases. The decrease in load share ratio is rapid till the settlement around
3mm, and the reduction is gradual for the settlements between 3mm and
7mm. In the above two phases, the interaction between raft and pile is
effective, thus the rapid transfer of load to the raft is taking place.
This behaviour shows reduction in the pile-soil stiffness with

increase in settlement. Although the raft-soil stiffness reduces, it is not to the
level of pile-soil stiffness. In the third phase (settlement more than 7mm) the
load share ratio is almost constant with settlement for a given pile length. This
phase indicates that the piles at higher settlement of the piled raft do not serve
effectively as load sharing element. Otherwise, they perform the role of
settlement reducer. Thus raft of piled raft has marginally higher stiffness than
133
the stiffness of plain raft at higher settlements. The relation between αpr and
settlement of piled raft explained above is seen in all the tests irrespective of
the pile length. However the magnitude of αpr is different for different pile
lengths at a given settlement of piled raft. For example αpr value for the pile
length of 200mm is 0.7, 0.45 and 0.4 for the piled raft settlement of 2mm,
7mm and 20mm respectively. The corresponding values for the pile length of
75mm are 0.3, 0.1 and 0.095. Thus it is clearly evident from the results that
though the piles of different lengths show identical trend in load share, the
magnitude of load share at a given settlement is less for piles of shorter
length. This indicates reduction in load share becomes more significant for a
pile group with shorter piles and is attributed to overall reduction in the pile
group-soil stiffness for the piles of shorter length. Further almost identical
behaviour is seen in the tests on piles with 8mm and 6mm diameters and also
in sand beds of different densities.
In Figure 4.37a, 4.37b and 4.37c the variation of αpr with pile length
normalised with the diameter of raft (L/D) is presented for the settlement of
2mm and 20mm for the diameter of pile of 10mm and 8mm respectively.
1.00
t = 8mm
0.80 d = 10mm
N = 21
∝ pr VALUE
0.60 D = 200mm
αpr
Bed = Medium dense

0.40
0.20
2mm 20mm
0.00
0.2 0.4 0.6 0.8 1.0 1.2
L/D
Figure 4.37 (a) Variation of αpr with L/D ratio, pile diameter 10mm
The proportion of load transferred to the piles increases with

increase in L/D ratio of pile irrespective of the settlement as well as diameter
of the piles. However the rate of increase of αpr value is higher for piled raft
with L/D ratios more than 0.6. This trend is seen invariably in piles of
134
different diameters (10mm, 8mm and 6mm) tested in this study. Thus longer
the pile length higher is the load share irrespective of the settlement of piled
raft and diameter of pile used. However, the advantage derived on increase in
the pile length more then 0.6D is not very significant especially for the piles
of smaller diameter (Figure 4.37c).
0.80
t = 8mm
d = 8mm
0.60
N = 21
VALUE
D = 200mm
αpr
∝ pr
0.20
2mm 20mm
0.00
0.2 0.4 0.6 0.8 1.0 1.2
L/D
Figure 4.37 (b) Variation of αpr with L/D ratio, pile diameter 8mm
p
0.40
t = 8mm
d = 6mm
0.30 N = 21
∝ pr VALUE
D = 200mm
αpr
0.10
2mm 20mm
0.00
0.4 0.5 0.6 0.7 0.8 0.9
L/D
Figure 4.37 (c) Variation of αpr with L/D ratio, pile diameter 6 mm
4.4.6.3 αpr versus pile diameter
The αpr values arrived from the load-settlement response of piled

raft for various diameters of pile with pile length of 160mm (L/D =0.8) have
been analysed to evaluate the influence of diameter of pile on αpr. Figure 4.38
135
shows the effect of pile diameter on αpr ratio for various settlement of piled
raft.
The response of αpr versus settlement is almost identical for the

diameters of pile used in the model tests. Irrespective of the settlement, the αpr
value for the piled raft with pile of higher diameter is high for the settlement
of 2mm, the αpr values are 0.49 and 0.18 for pile diameters of 10mm and 6mm
respectively. Similarly for 20mm settlement the values are 0.25 and 0.08
respectively.
α pr
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
0
4 N = 21
L = 160mm
D = 200mm
8 t = 8mm
SETTLEMENT,mm
Bed = Medium dense
12
16 Pile dia
10mm
20 8mm
6mm
24
Figure 4.38 Settlement v/s LS ratio αpr for various pile diameter
Further the difference in αpr value between any two diameters of

pile is also decreases with the settlement. However this difference at higher
settlement remains almost constant between them. For example, the
difference in αpr between the pile of diameter 10mm and 6mm is 0.31 at 2mm
settlement and 0.16 at settlement around 20mm. Overall the αpr of larger
diameter pile (i.e. 10mm dia) is 2.75 times the smaller diameter pile
irrespective of the settlement of piled raft, which is almost equal to the ratio
between the area of respective pile groups.
136
Further analysis on αpr is carried out by presenting this value based

on ratio of the pile diameter and thickness of raft (d/t). The variation of αpr for
the settlement of 2mm and 20mm is presented in Figure 4.39 for different d/t
ratios. In general, the αpr value increases with d/t ratio irrespective of the
settlement of piled raft. The increase in αpr is almost linear with the d/t ratio
for the range of d/t ratios adopted in this study. Further, higher the d/t ratio
larger is the load shared by the piles. The pile of larger diameter has higher
axial stiffness and higher surface area of pile, hence it offers higher load
resistance.
0.80
t = 8mm 2mm 20mm
L =160mm
0.60 N = 21
D = 200mm
αpr
0.20
0.00
0.50 0.75 1.00 1.25 1.50
d/t RATIO
Figure 4.39 Variation of αPR value with d/t ratio
4.4.6.4 αpr versus pile raft area ratio
The load transferred to the piles are analysed by varying number of

piles of the piled raft. Since the tests were conducted on piles of different
diameters and spacing, a parameter named pile raft area ratio (AR) is used.
The load share ratio, αpr with settlement is plotted in Figure 4.40 for the area
ratios of 9.25%, 5.25% and 4.25%. The number of piles (N) used for the area
ratios indicated above are 37, 21 and 17 respectively.
The load shared by the piles at the beginning of application of the

load is high irrespective of the number of piles. As the applied load increases,
the proportion of load carried by the piles decreases; this is associated with
increase in settlement of the piled raft. The rate of decrease in αpr varies with
137
settlement and becomes constant after the settlement of around 7mm. Though
the αpr value for the piled raft with area ratio 4.25% is the lowest irrespective
of the settlement among the AR ratios studied in this research, it is of interest
to note that the curves of αpr versus settlement are very similar for all the piled
rafts of different area ratios. The reduction in αpr with load and lowest αpr
value for the piled raft with less number of piles are attributed to initial
stiffness of respective pile-soil system and non-linearity of the pile behaviour
(two-phase response) makes the relative stiffness of the pile group to the soil
smaller. This is evident from sharing minimum load of the order of 10% and
20% of the total applied load for the area ratio of 4.25% and 9.25%
respectively at the final settlement of 20mm.
α PR
0.0 0.2 0.4 0.6 0.8 1.0
0
4
t = 8mm
6
d= 8mm
SETTLEMENT,mm
8 N = 21
R.A. = 36º
10 D = 200mm
14 Area
Pile Ratio
length
16
9.25%
18 5.25%
20 4.25%
22
Figure 4.40 Settlement v/s LS ratio αpr for different area ratios
The load sharing ratio, αpr is further analysed by presenting the

value for various pile raft area ratio as shown in Figure 4.41. The results are
presented for the settlement less than critical settlement and settlement of
20mm. The response between αpr and AR is similar to the response of
settlement vs AR values presented in the Figure 4.33. The αpr value increases
rapidly with area ratio and the increase is negligibly small for the area ratio
more than 6%. It appears from the study, increasing the pile raft area ratio
more than 6% is not gaining much advantage either in sharing the load or
reducing the piled raft settlement. The value of αpr decreases rapidly until the
138
load, which is causing the settlement around 3mm. This settlement is known
as critical settlement, which is required for mobilizing frictional resistance of
piles (limiting fiction) fully. Thereafter, though the piles share more load, its
contribution is not significant, which is evident from constant αpr value at
higher settlement of piled raft. The trend indicated above has been seen in all
the tests conducted.
1.00
2MM 20MM
0.80
0.60
αpr
0.40
0.20
0.00
4.00 5.00 6.00 7.00 8.00 9.00 10.00
Area Ratio AR
Figure 4.41 Variation of αPR with piled raft area ratio
4.4.7 Effect of Pile Configuration on Sr and αpr
The influence of pile configuration (i.e. arrangement of piles) on

piled raft on load sharing αpr and settlement reduction SR was arrived for the
piled raft area ratio of 2.75% wherein the number of piles provided are 11 and
are arranged in three different configurations as shown in Figure 4.42a to
4.42c.
(a) Configuration A (b) Configuration B (c) Configuration C
Figure 4.42 (a), (b) and (c) Configurations studied with 11 piles
139
In all the configuration one pile is provided commonly at the centre

of the raft. Rest of the piles are arranged radially as shown in Figure 4.42 and
are identified as configuration A, configuration B and configuration C. In the
configuration A, B and C, ten piles are arranged in outer ring, inner ring and
both inner and outer ring (alternate ring) respectively and one pile in the
centre.
(d) Configuration D
Figure 4.43 Configuration of 21 piles studied
The results of the piled raft with area ratio 2.75% is compared with
the piled raft of area ratio 5.25% having 21 piles uniformly arranged as shown
in Figure 4.43 (Configuration D).
The variation of αpr with the settlement is compared for all the four
configurations of piles in Figure 4.44a. In the case of piled raft with the piles
distributed symmetrically and equally to the entire area of the raft (i.e. in
configuration D), the αpr value is high irrespective of the settlement when
compared with the other configurations namely A, B and C.
This response is obvious and is attributed to reduction in the

number of piles. Further it can be observed from the figure that the
configurations adopted in this study show some difference in the αpr value
despite all the configurations (A, B, and C) have same number of piles.
Among the configurations A, B and C the configuration C shows higher αpr
value than the other two configurations irrespective of the magnitude of
settlement excepting for settlements less than or close to critical settlement.
140
The least value is for the configuration B (piles in the inner ring), which is
two to four times lesser than the αpr value of the configuration D (Area ratio
5.25%). The higher reduction in the αpr value is for the settlement more than
10mm. For the configuration B the αpr values lie between the pile
configurations of B and C. However for the settlements less than the critical
settlement the difference in αpr value between the three configurations (A, B,
and C) is small.
1.00 1.00
OUTER
OUTER
INNER
INNER
0.80 0.80 A LT.
ALT.
FULLY
FULLY
0.60 0.60
αpr
SR
0.40 0.40
0.20 0.20
0.00 0.00
0 4 8 12 16 20 24 0 4 8 12 16 20 24
SETTLEMENT, mm SETTLEMENT, mm
(a) (b)
Figure 4.44 Effect of configuration on the variation of SR and αpr with
Settlement
In Figure 4.44 (b) the settlement reduction, SR is compared for the

configurations of A, B, C, and D. Settlement reductions are compared for the
given settlement of the raft. This figure also shows the trend as seen in the
case of relation between load sharing ratio (αpr) and settlement.
From the comparison made it can be said that among the three
configurations (A, B and C) the configuration C (piles arranged in alternate
ring) has performed better in load sharing and settlement reduction. Thus the
piles distributed evenly over the entire area are best choice than concentrating
them over a specific area in the case of piled raft subjected to uniformly
distributed load.
141
4.5 BEHAVIOR OF SQUARE PILED RAFT
In this section, the results of 1g model tests conducted on a square

shaped piled raft are discussed. In the case of square piled raft also the results
are compared at three different settlement levels, namely, 2mm (0.01B), 6mm
(0.03B) and 20mm (0.10B) of the plain raft. In order to compare the
behaviour of square piled raft with the circular piled raft (Ar=5.2%), were
conducted on square raft with AR = 4.9% (i.e. pile spacing = 4d). The small
difference in area ratio is attributed to the difference in pile configurations
adopted between the square and circular piled rafts.
4.5.1 Behaviour of Unpiled Raft
The load-settlement curves of 8mm, 10mm and 12mm thick (plain

raft), unpiled raft conducted on beds of different densities are presented in
Figures 4.45 to 4.47. In the case of 8mm and 12mm thick rafts, the tests have
been carried out in all the three densities namely loose, medium dense and
dense and, the test was done on medium dense sand only for the raft thickness
of 10mm.
LOAD, kN
0 1 2 3 4 5 6 7 8
0
2
4
6
MM
SETTLEMENT, mm
8
10
12
14
16
Plain raft
18
B=200mm
20 t=8mm Medium Dense
22
Loose
24
Figure 4.45 Load-settlement response of plain square raft of 8mm thick

142
LOAD, kN
0 1 2 3 4 5 6 7 8
0
2
4
6
MM
SETTLEMENT, mm
8
10
12
14
16
18 Plain raft
B=200mm
20 Medium
t=10mm
22
24
Figure 4.46 Load-settlement response of square plain raft of 10mm thick

LOAD, kN
0 1 2 3 4 5 6 7 8
0
2
4
6
MM
SETTLEMENT, mm
8
10
12
14
16
Plain raft
18
B=200mm
20 t=12mm Dense
Loose Medium
22
24
Figure 4.47 Load-settlement response of square plain raft of 12mm

thick
The load taken by the plain raft on all the three densities at the
settlement of 20mm is presented in Table 4.13 along with the corresponding
stiffness. It can be seen that the variation of load taken at any particular
settlement level increases marginally with the raft thickness in square raft also
as seen in circular piled raft.
143
Table 4.13 Variation of load and stiffness (N/mm) of plain square raft
t = 8mm t = 10mm t = 12mm

Bed Load Load Load
Stiffness Stiffness Stiffness
(kN) (kN) (kN)
Loose 1.90 95 - - 1.90 95
Medium 5.20 260 5.20 260 5.40 270
Dense 6.60 330 - - 6.70 335
The behaviour of plain raft has been characterised in Figures 4.48

to 4.50 by plotting the curve in semi-log plot for the raft thicknesses of
8mm,10mm and 12mm respectively.
10
B
LOAD, kN
DENSE
Plain raft
B=200mm MEDIUM DESNE
t=8mm LOOSE
O
0.1
0
0 2 4 6 8 10 12 14 16 18 20 22
SETTLEMENT, mm
Figure 4.48 Characterised load-settlement response of plain square raft

of 8mm thick
The characterisation curve clearly indicates the load-settlement

behaviour has three phases as in the case of circular piled raft. The slope
changes at two settlement levels, namely 2mm and 8mm in loose sand, close
to 2mm and 6mm, in the medium dense sand, 1.8mm and 6mm in dense sand,
for raft thickness of 8mm. For 12mm thick raft, while the first stage is close to
144
2mm settlement in, the second stage slope change occurs at 5.5mm in the
medium and dense sand condition.
10.0
B
LOAD, kN
1.0
Plain raft
B=200mm
t=10mm MEDIUM DENSE
O
0.1
0.1
0 2 4 6 8 10 12 14 16 18 20 22
SETTLEMENT, mm
Figure 4.49 Characterised load-settlement response of square raft of

10mm thick
10
B
LOAD, kN
1
A
DENSE
Plain raft
MEDIUM DENSE
B=200mm
t=12mm LOOSE
O
0.1
0
0 2 4 6 8 10 12 14 16 18 20 22
SETTLEMENT, mm
Figure 4.50 Characterized load-settlement response of square raft 12mm

thick
In the case of 10mm thick raft, for medium dense sand these
changes take place at 1.6mm and 6mm. Thus in overall, it can be said that the
change in first slope is around the settlement of 1.6mm second slope is at a
settlement of 6mm which are almost close to the settlement values at which
145
changes in slopes are seen in circular raft of 8mm and 10mm thick. The
stiffness variations for the three stages are given in Table 4.14.
Table 4.14 Variation of stiffness (N/mm) of square raft for different

bed densities and raft thicknesses
t = 8mm t = 10mm t = 12mm

Bed
OA AB BC OA AB BC OA AB BC
Loose 200 100 83 - - 210 140 110
Medium 714 319 225 730 450 240 760 340 240
Dense 750 575 245 790 610 240
The Table 4.14 gives the stiffness of the plain raft at the three
stages of behaviour. The characterisation curve indicates that the plain raft on
the sand exhibits multi linear isotropic hardening behaviour. This behaviour
was found to be identical for all the raft thicknesses tested. The variation of
stiffness as seen from the above depends more on the bed density than on the
raft thickness. Hence, the study was conducted only on 8mm thick square raft.
Also based on the performance of circular piled raft it was decided to keep the
d/t ratio (pile diameter versus raft thickness ratio) more than 1 for most of the
studies on square piled raft.
4.5.2 Comparison of Piled Raft Behaviour with Unpiled Raft

(Plain Raft)
Figure 4.51 demonstrates the load-settlement behaviour of plain raft

and piled raft, along with the load-settlement curve derived for the pile group
of piled raft typically for medium dense sand. Similar tests were carried out in
the case of loose and dense sand also. In all the tests the parameters relating to
the pile and raft were kept constant. The raft size was kept as 200mm x
200mm and the thickness was 8mm. The test was carried out with 25 piles
placed at 4d spacing. The pile - raft area ratio was kept as 4.9%. The load-
settlement response of the square piled raft and the square raft was found to
146
be almost similar to the respective response of the circular piled raft and
circular raft. At any given load, the settlement of the piled raft was far less
than that of the plain raft. Further, it can be seen that in the case of piled raft,
upto a settlement of 2.0mm (load=2.8kN), the rate of increase in the
settlement was small. From 2mm settlement to a settlement of 6mm
(load=4.7kN), the increase was gradual with load. Beyond this level, the
settlement increased rapidly with load. However for any given load, the
settlement of piled raft was much smaller than that of the plain raft. For
20mm settlement the load taken by the piled raft was higher than the plain raft
by 46% .
LOAD, kN
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0
0
N = 25
L = 200mm
4 d = 10mm
S = 4d
SETTLEMENT, mm
8 B = 200mm
t = 8mm
φ = 37.5 °
12
16
20
RAFT PILE PILED RAFT
24
Figure 4.51 Load-settlement response of Plain raft, piled raft and pile
group of piled raft (Square 200 x 200 mm)
For the load of plain raft corresponding to settlement of 20mm, the

settlement of the piled raft was 8mm and the reduction in the settlement was
60%. The trend indicated by this curve is very similar to that of the circular
piled raft. The addition of piles in this case also, increases the stiffness of the
system. It can be seen here that although the area ratio is marginally smaller
than the circular piled raft, the load taken by the square piled raft is more than
the circular piled raft. This is because of larger raft area and more number of
147
piles than circular piled raft. The combined interaction of the pile- raft and the
soil in this case generates a marginally high capacity mainly because of the
square raft area is more by 27% and the number of piles is more by 19%.
Interestingly, the square raft on an average takes 27% higher load than the
circular raft. Surprisingly this is equivalent to the ratio between the areas of
square and circular raft. At any given settlement, the square piled raft was
found to take more load than circular piled raft. This variation was found to
be in the range of 21 to 34%. The load taken by the plain raft and piled raft
with different raft thickness and bed densities are given in Table 4.15.
Table 4.15 Comparison load response between square raft and square
piled raft (B=200mm, d=10mm, L=160mm and N=25)
(a) Loose
Load in kN @ Load in kN @ 6mm Load in kN @
Raft 2mm settlement settlement 20mm settlement
thickness
8mm 0.35 1.20 0.80 1.95 2.00 5.25
12mm 0.40 1.00 0.90 1.70 1.95 4.20
(b) Medium dense
thickness
8mm 1.40 2.80 3.10 4.90 5.70 9.10
10mm 1.40 2.80 3.10 4.70 5.60 8.70
12mm 1.60 2.90 3.00 4.85 6.90 9.00
(c) Dense
thickness
8mm 1.75 3.60 3.30 5.60 7.00 10.20
12mm 1.32 3.80 3.20 6.00 6.60 10.80
148
The characterisation curve presented in Figure 4.52 typically for

piled raft in medium dense sand, whose parameters are given in the figure
establishes the effect of adding piles on the three phase behaviour of the plain
raft. It is seen in the characterisation curve that the provision of piles in the
raft has shifted the linear elastic behaviour of the foundation system to a load
level of 2.8kN. The second phase commences from 2.8kN and extends up to
4.9kN, and the third stage is from 4.9kN to till the load corresponds to 20mm
settlement. Comparing this to the plain raft, for the settlements of 2mm, 6mm,
and 20mm the addition of piles has enhanced the piled raft resistance by
100%, 58%, and 39% respectively. This is marginally more than the
enhancement generated in the circular piled raft by the piles which are 77%,
45%,and 33%.
1000
Piled raft
C
Raft
B
102-2
LOAD, kN X10
100
A d = 10mm
L= 160mm
t = 8mm
S=4d
B = 200 mm
O
10
0 5 10 15 20 25
SETTLEMENT, mm
MM
Figure 4.52 Characterised load-settlement response of plain raft & piled

raft in medium dense sand
The first phase represented by the curve is the stage where the
major part of the applied load is taken by the pile group-soil interaction. This
condition is seen up to a settlement level of 1.5mm which is very close to
0.8% of the least lateral dimension of the raft. At this stage the entire system
is under elastic stage. Beyond this, the stiffness of the combined system
reduces, and the rate of change in the settlement increases. In the second
phase the pile group exhibits reduction in its stiffness; this leads to rapid
increase in the settlement. This stage exists upto a settlement level of around
149
6mm as per characterisation curve and this is 3% of the lateral dimension of

the raft. The increase in the settlement is very high even for a small increase
in the load after the second phase of the load settlement response. In other
words at higher settlement, say beyond 3% of the lateral dimension of the raft,
the interaction between pile and the soil reduces and, resulting in the
reduction of overall stiffness. However combined interaction makes the
system to take a higher load than the plain raft.
The test was conducted on the piled raft with 10mm thick raft,
keeping the other parameters same. It was found that the behaviour
(Figure 4.53) was same as in the case of piled raft with 8mm thick raft.
1000
Piled raft
C
Raft
B
-2
102
LOAD, kN x10
A
100
d = 10mm
L= 160mm
t = 10mm
S=4d
B = 200 mm
10 O
0 5 10 15 20 25
SETTLEMENT, mm
MM
Figure 4.53 Characterised load-settlement response of plain raft & piled

Table 4.16 Comparison of stiffness (N/mm) between plain & piled raft
(B=200mm, t=8mm, d=10mm, S= 4d, L=160mm and N=25)
Stiffness in N/mm at different stages

OA AB BC
8mm thick raft
Plain raft 1000 500 194
Piled raft 2400 675 266
10mm thick raft
Plain raft 1000 550 194
Piled raft 2400 600 300
150
Table 4.16 presents the stiffness at different stages of the

characterisation curve. It is clearly seen that, in the case of square piled raft
also, as the settlement increases, the stiffness of the system approaches to that
of plain raft.
4.5.3 Behaviour of Pile Group of Piled Raft
As in the case of circular piled raft, here also the load-settlement

pattern of the free standing pile group was compared with the load-settlement
curve of the pile group of piled raft. The results presented here are typically
for medium dense bed condition. The free standing pile group and the pile
group of piled raft of square piled raft are provided with 9 piles and are
distributed uniformly over the raft. The spacing of the piles was 6d and the
piles were 10mm dia, 160mm long. Figure 4.54 presents the comparative
behaviour of the free standing pile group and the pile group of the piled raft in
medium dense sand.
LOAD, kN
0.0 0.5 1.0 1.5 2.0
0
6 d Spacing free standing
4 6 d Spacing piled raf t

SETTLEMENT, mm
12
d = 10mm
L = 160mm
16
t = 8mm
S = 6d
20 B = 200mm
Bed = Medium dense
24
Figure 4.54 Load-settlement response of pile group of piled raft and free
standing group (square shape)
As seen from the curve in the case of free standing pile group, the
magnitude of critical settlement was around 1.5mm upto which the pile group
resisted the applied load. Beyond this level of settlement, the pile group lost
151
its relative stiffness rapidly. The limiting friction namely the load
corresponding to critical settlement was 0.45kN, and beyond this load level,
the pile group took hardly any load. The loss of pile group stiffness was
instantaneous. This indicates that the free standing pile group-soil interaction
mechanism fails instantaneously beyond the load of limiting friction. The
resistance of piles due to limiting friction was found to be 82% of the final
load corresponding to a settlement of 20mm of the free standing pile group.
In the case of pile group of piled raft, the magnitude of the critical
settlement was also close to 1.5mm and the limiting friction was 0.6kN,
which is higher than that of the free standing pile group limiting friction value
by 33%, indicating that the ground contacting pile cap, pile group – soil
interaction enhances the frictional capacity of the pile group, although the
area ratio of piles is very small. Beyond the critical settlement the pile group
of piled raft continued to take additional load on further loading of the piled
raft. At the final settlement of 20mm, the pile group of piled raft shared a load
of 1.9kN which is far in excess of the load taken by the free standing pile
group. The gradual increase in the load taken by the pile group of piled raft
indicate that the elastic-work hardening behaviour exists in the case of pile
group of piled raft, however small the group size may be. This additional
resistance may be attributed to the increase in the stress in the soil around the
piles due to the load transfer between raft and the soil.
The characterisation curve (Figure 4.55) indicates that the

behaviour of the free standing pile group and pile group of piled raft is
similar manner and their responses are identical to the responses demonstrated
in circular piled raft. However the corresponding load level is enhanced in the
case of pile group of piled raft. In the second stage of the characteristic
behaviour represented by the portion AB, the pile group of piled raft
continues to take load, whereas, the free standing pile group fails to take any
load. This behaviour indicates that the shape of the raft has no influence on
the behaviour of pile group of piled raft.
152
1000
6d SPACING FREE STANDING
6d SPACING PILE GROUP OF PILED RAFT
LOAD, kN x 10-2
C
100
B
A
10 O
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
SETTLEMENT, m m
Figure 4.55 Characterisation of load-settlement response of free

standing pile group and pile group of piled raft
Figure 4.56 presents the typical characterisation curves of plain raft

and the raft of piled raft. A comparison indicates that the addition of the pile
group, however small it be, improve the performance of the piled raft. The
curves show at any given settlement, the load taken by the raft of piled raft is
higher than the plain raft. It indicates that the raft of piled raft has a higher
stiffness than that of the plain raft. However at higher settlement (third phase),
the stiffness of raft of the piled raft is close to the stiffness of plain raft.
1000
Raft of piled raft

Plain raft
C
B
kNx x10-
100
A
LO AD kN
100
2
LOAD,
200mm sq raft
t = 8mm
Bed = Medium dense
N=9
O
10
0.0 5.0 10.0 15.0 20.0 25.0

SETTLMENT MM
SETTLEMENT, mm
Figure 4.56 Characterisation of load-settlement response of plain raft

and raft of piled raft
153
4.5.4 Load-settlement Behaviour of Square Piled Raft
As explained in the case of circular piled raft, the economics of

piled raft design lies in the reduction of settlement to the permissible level by
providing the piles in an appropriate layout. As seen earlier in the case of
circular piled raft, the load-settlement behaviour depends essentially on the
parameters relating to the pile and the bed. However, the square piled raft
behaviour under various parametric conditions was studied and the results are
discussed below.
4.5.4.1 Effect of pile spacing
Figure 4.57 gives the load-settlement curve for piled raft with three
different pile spacing namely 4d (25 piles), 6d (9 piles) and 7.5d (9 piles).
The study has been carried out for two different raft thicknesses. This figure
presents the curve for piled raft having 8mm thick raft typically for medium
dense sand. The pile diameter was kept as 10mm and the length was160mm.
LOAD, kN
0 1 2 3 4 5 6 7 8 9 10
0
2 d = 10mm
t = 8mm
4 L = 160mm
B = 200mm
MM
SETTLEMENT, mm
10
12
14
16
18 PLAIN RAFT
4d Piled raft
20 6d
Pile spacing
7.5 d
22
Figure 4.57 Load-settlement response of piled raft for various spacing -

8mm thick square raft
154
It can be seen that irrespective of spacing, the rate of increase in the

settlement upto a settlement level of 2mm is very small. However the load
corresponding to this settlement shows very little variation between the pile
spacing of 6d and 7.5d. The magnitude of the loading is 1.7kN; but for the
piled raft with 4d spacing, the load at the settlement level of 2mm is 2.6kN.
However, in all these cases the magnitude of this load works out to 22% of
the load corresponding to 20mm settlement of the respective piled raft. At
6mm settlement the load taken by the piled raft with nine piles (spacing 6d) is
2.2kN but with twenty-five piles it is 4.0kN.
It is to be noted here that the load taken by the piled raft is not
directly proportional to the area or perimeter of the pile group provided.
Further it is to be noticed that although the number of piles remains the same
between the spacing of 6d and 7.5d, beyond 6mm settlement the load taken by
the piled raft with 6d pile spacing is more than the piled raft with 7.5d pile
spacing. This indicates that as the pile spacing increases, the level of
interaction between the pile and the soil decreases. In the case of 4d spacing,
upto the settlement of 8mm, although the raft settlement increases, it is not
rapid. But beyond this settlement the loss of stiffness is rapid. However
irrespective of pile spacing, beyond 20mm settlement, the piled raft did not
take any further load. At final settlement of 20mm, even though the number
of piles are the same, the load taken by the piled raft with 6d spacing is more
than 7.5d spacing by 17% whereas between 6d (9piles) and 4d (25 piles) the
variation is 22.9%. It is seen that even when the number of piles are more, the
increase in the load carrying capacity is only 22.9% indicating that the
number of piles do not have much influence at higher settlement. This is
attributed to reduction in pile – soil stiffness when the settlement is more than
the settlement corresponds to limiting friction of pile group.
Figure 4.58 presents the characterisation curve for 4d and 6d

spacing for medium dense sand. The change in spacing between the piles
(ie change in number of piles) does not alter the characteristic response of
155
square piled raft except the difference in the stiffness of piled raft system. The
stiffness of piled raft with higher number of piles (i.e. piles at closer spacing)
is higher than the piled raft with 9 piles. But the difference in the third phase
(i.e. BC part) is not appreciable as seen in the case of circular piled raft.
Although the change of slope occurs in a way similar to that of circular piled
raft; the stiffness of square piled raft in all the three phases are more than the
circular piled raft of almost identical test condition.
10
B
A
LOAD, kN
B = 200mm
t = 8mm
d = 10mm
L = 160mm
4d
N=9
O 6d Medium dense
0.1
0.0 5.0 10.0 15.0 20.0 25.0

SETTLMENT, m m
Figure 4.58 Characterisation of load-settlement response of piled raft

with different pile spacing
The non-dimensional plot (Figure 4.59) shows some divergence

mainly due to the wider variation in the number of piles, namely 25 for 4d
spacing and 9 for 6d spacing. However, the load ratio and settlement ratio can
be said to be in close agreement.
The equation (4.7) of the hyperbolic plot (Figure 4.60) shows a

marginally flatter slope compared to that of the circular piled raft. The values
of ‘m’ and ‘c’ are 0.60 and 0.42 respectively.
δR (4.7)
PR =
0.6δ R + 0.42
156
LOAD RATIO
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00
0.00
d = 10mm
0.10
t = 8mm
0.20 L = 160mm
B = 200mm
SETTLEMENT RATIO
0.30 Medium dense

6d
0.40
4d
0.50 7d
0.60
0.70
0.80
0.90
1.00
Figure 4.59 Non-dimensional plot for piled raft with different pile
spacing
1.00
0.90
d = 10mm
0.80
t = 8mm
0.70 L = 160mm
B = 200mm
0.60 Medium dense
0.50 6d
4d
0.40 7d
Average
0.30
δR
PR =
0.20
m δR + c
0.10 y =and
m=0.60 0.6xc=0.42
+ 0.42
0.00
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00
SETTLEMENT RATIO
Figure 4.60 Hyperbolic plot for piled raft with different pile spacing
The Table 4.17 presents the comparison of the stiffness at three

different settlements for 10mm and 8mm thick piled rafts tested with piles of
different spacing in all the three densities of sand.
157
Table 4.17 Effect of spacing on load carried and stiffness (N/mm)

(B = 200mm, t = 8mm, L = 160mm, d = 10mm)
(a) Loose
Spacing
Load, kN Stiffness Load, kN Stiffness Load, kN Stiffness
4d 1.20 600 1.95 325 5.30 265
6d 0.80 400 1.60 267 3.80 190
7.5d 0.80 400 1.50 250 4.00 200
(b) Medium dense
Spacing
4d 1.45 725 4.90 817 9.10 455
6d 0.88 437 3.55 592 7.40 390
7.5d 0.88 437 3.75 625 6.80 340
(c) Dense
Spacing
4d 3.60 1800 5.70 950 10.20 510
6d 2.10 1050 4.60 767 8.80 440
7.5d 2.40 1200 4.45 741 7.90 345
Note : 4d = 25 piles, 6d = 9 piles and 7.5d = 9 piles
It can be seen that irrespective of the raft thickness, the stiffness

reduces as the settlement increases. Further the stiffness of piled raft
approaches the value of plain raft stiffness in the third phase (BC). While in
the case of pile spacing of 6d and 7.5d, the stiffness of the piled raft is 25%
higher than the plain raft stiffness, in the case of 4d spacing, it is 60%
indicating that at spacing less than around 6d, the tendency is to behave like a
fully piled system.
158
The effect of pile diameter on the load-settlement behaviour was

studied for the bed densities of loose and dense condition. In both the cases
the raft thickness and the pile length and the pile diameter were maintained
constant as 8mm, 160mm and 10mm respectively. Figure 4.61 presents the
load-settlement curve for loose sand and Figure 4.62 presents the curve for
dense sand.
The difference in resistance of piled raft against settlement between

the rafts with different pile diameters is not appreciable upto the settlement
around 2 to 3mm both in loose and dense sand conditions. The trend as
indicated above was seen in circular piled raft also although the test was done
with sand of medium dense state. But as the load increases the rate of increase
in the settlement was more in the rafts with 6mm and 8mm diameter piles
than in the raft with 10mm diameter piles. This indicates that, the loss of piled
raft stiffness was high and rapid with load for the rafts with smaller diameter
piles (the ratio d/t < 1 and d/t =1) irrespective of the shape of the raft.
However, the variation in the load taken between the piled rafts with 6mm
and 8mm diameter piles was less for the entire range of settlement both in
loose and dense sand states.
LOAD, kN
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
0
L = 160mm
t = 8mm
4
S=6d
B = 200mm
8 Bed = Loose
MM
SETTLEMENT, mm
12
16
20
10mm
24 8mm Pile dia

6mm
28
Figure 4.61 Load-settlement response of piled raft for various pile

diameters in loose sand bed
159
LOAD, k N
0 1 2 3 4 5 6 7 8 9 10 11 12
0
2 t = 8mm
L = 160 mm
4
N = 25
6 B = 200mm
Bed = Dense
MM
SETTLEMENT, mm
10
12
14
16
10mm
18
8mm Pile dia
20 6mm
22
Figure 4.62 Load-settlement response of piled raft for various pile

diameters in dense sand bed
The characteristics curves for the piled rafts with piles at different
diameters are established and found that they show three phase response as
established in the other cases. The stiffnesses of piled rafts in all the three
phases are arrived from the response curves for loose and dense conditions
and are compared in Table 4.18.
Table 4.18 Effect of pile diameter on load and stiffness (N/mm) for the
square piled raft (t = 8mm, S = 6d and L = 160mm)
(a) Loose sand

Pile @ 2mm settlement @ 6mm settlement @ 20mm settlement
diameter Load, kN Stiffness Load, kN Stiffness Load, kN Stiffness
6 mm 0.55 275 1.05 175 2.20 110
8 mm 0.65 325 1.20 200 2.40 120
10 mm 0.85 425 1.50 250 3.40 170
(b) Dense sand
diameter Load, kN Stiffness Load, kN Stiffness Load, kN Stiffness
6 mm 1.80 900 3.60 600 7.00 385
8 mm 1.85 925 3.90 650 7.30 365
10 mm 1.90 950 4.50 750 9.30 465
160
The stiffness values compared in the Table 4.18 indicate that

irrespective of the density, the stiffness increases with the increase in the
diameter of piles in sands of loose and dense compactness. However in the
loose sand, the stiffness of response curve in all the three phases is less than
the dense sand and the effect of density is more pronounced in the care of the
raft with pile of large diameter. As the diameter increases, the rate of increase
in the load resistance of square piled raft follows the trend almost same as in
the case of circular piled raft. This observation indicates that neither the shape
of the piled raft nor the number of piles has influence on the characteristics
response of piled raft.
The non-dimensional and hyperbolic plots demonstrated in Figures

4.63 and 4.64 show identical trends individually as explained in the previous
section. The relation between the load ratio and the settlement ratio is
δR (4.8)
PR =
0.611δR + 0.4
The m and c values are 0.611 and 0.4 respectively, which are
almost same as that of the values arrived for the condition of different pile
spacing.
LOAD RATIO
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00
0.00
L = 160mm
0.10
t = 8mm
0.20 S = 6d
B = 200mm
SETTLEMENT RATIO
0.30 Dense sand

0.40 8mm
0.50 10mm
6mm
0.60
0.70
0.80
0.90
1.00
Figure 4.63 Non-dimensional plot for piled raft with different pile
diameters
161
1.00
0.90

L = 160mm
0.80
t = 8mm
0.70 S = 6d
B = 200mm
0.60 Dense sand
0.50
8mm
0.40 10mm
6mm
0.30 Average
0.20
0.10 y = 0.6111x + 0.4
0.00
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00
SETTLEMENT RATIO
Figure 4.64 Hyperbolic plot for piled raft with different pile diameter
4.5.4.3 The effect of pile length
The effect of the length of the pile on the load-settlement behaviour

of square piled raft was studied with piles placed at 6d spacing. The number
of piles, the raft thickness, and the pile diameter were kept as 9, 8mm and 10
mm respectively. The study was carried out in sands of all the three bed
densities.
LOAD, kN
0 1 2 3 4 5 6 7 8 9 10
0
2 d = 10mm
t = 8mm
4 S=6d
B = 200mm
MM
6
SETTLEMENT, mm
Bed = Medium dense

8
10
12
14
16
PLAIN RAFT
18 200mm
160mm pile length
20 120mm
22
Figure 4.65 Load-settlement response of piled raft for various lengths

in medium dense bed
162
Figure 4.65 provides the load-settlement curves for pile the lengths
tested in the case of medium dense sand.
Typically in the case of medium dense and dense sand, up to the

load of 2.8kN, the difference in the settlement between the piled rafts tested
with three different pile lengths was small. But at higher loads the lengths of
pile is effective in load sharing and settlement reduction particularly in denser
sand beds. Although the rate of increase in the load is not in direct proportion
with the length, it can be seen that the increase in the stiffness when the length
increased from 120 mm to 160 mm was 42%, but when the length increased
from 160 mm to 200 mm the increase was 17.6% in the first phase of the
curve in medium sand. The corresponding stiffness values in the third phase
are 5.9 % and 5.5 % respectively. This clearly indicates that at higher
settlement irrespective of the length the load taken by the piled raft reduces as
observed in the case of circular piled raft, although the pile raft area ratio was
higher. It is also seen that at 20 mm settlement, the stiffness of the piled raft
is reduced significantly irrespective of the density and the pile length. Slightly
higher stiffness of the piled raft then the plain raft in the third phase of
response curve indicates that the piles behave more as settlement reducer
rather than effective load sharing member, when the settlement of piled raft is
higher.
The load taken by the piled raft with different pile lengths for the
three different densities is presented in the Table 4.19. It can be seen that
irrespective of the densities, as the L/B (length of the pile to the width of the
raft) ratio increases, the load taken by the piled raft increases but as the ratio
becomes unity the rate of increase in the load resistance of piled raft reduces.
When the L/B ratio increased from 0.6 to 0.8, the increase was around 23%,
but when the ratio changed from 0.8 to unity, it was 18% in the first phase
response curve of loose sand. Similar response is seen in other phases and
163
also in other densities of sand. This indicates that the increase in the L/B ratio
of pile of the piled raft radius, the rate of increase in piled raft capacity. The
same trend was seen in the case of circular piled raft also.
Table 4.19 Effect of pile length on load taken and stiffness (N/mm)
(a) loose (b) medium dense (c) Dense (B = 200mm sq.
t = 8mm d = 10mm N = 9 S = 6d )
(a) Loose sand

Length Load, kN Stiffness Load, kN Stiffness Load, kN Stiffness
200 mm 0.95 475 1.50 250 3.90 195
160 mm 0.80 400 1.60 267 3.80 190
120 mm 0.65 325 1.25 208 3.20 160
(b) Medium dense sand

200 mm 2.00 1000 4.20 700 7.60 380
160 mm 1.70 850 3.85 642 7.20 360
120 mm 1.20 600 3.40 566 6.80 340
(c) Dense sand
200 mm 3.05 1525 5.20 867 9.30 465
160 mm 2.10 1050 4.60 766 8.80 440
120 mm 1.90 950 4.10 683 8.10 405
164
LOAD RATIO
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00
0.00
d = 10mm
0.10 t = 8mm
S = 6d
0.20
B = 200mm
SETTLEMENT RATIO
0.30 Medium dense
0.40 160mm
0.50 200mm
120mm
0.60
0.70
0.80
0.90
1.00
Figure 4.66 Non-dimensional plot for piled raft with different pile length
The relation between the dimensionless parameters namely load

ratio (PR) and settlement ratio (δR) for the piled raft with piles of different
length is demonstrated in Figure 4.66. A unique relation is obtained between
the two parameters despite the results of piled rafts tested with pile groups of
different lengths. The relation thus obtained is as follows
δR (4.9)
PR =
0.6δR + 0.41
The Equation presented above (Equation 4.9) is same as that of

other two equations (Equation 4.7 and 4.8) arrived for the conditions of
different pile spacing and piles of different diameters. There is a marginal
difference between them with their slopes (m) and vertical intercepts (c). The
slopes and vertical intercepts for the three conditions are found to vary in the
ranges of 0.6 to 0.61 and 0.4 to 0.42 respectively. These values are compared
with the values of m and c obtained for different conditions of circular piled
raft and they found to vary in the ranges of 0.674 to 0.689 and 0.35 to 0.37 for
m and c respectively. The slope m arrived for the square piled raft is lesser
than the circular piled raft for all the conditions of piled raft and densities of
165
sand. This is due to the fact that square piled raft resists higher load than
circular piled raft tested under identical test conditions with diameter of
circular raft equal to the width of square raft.
4.5.4.4 Effect of bed density
The load-settlement behaviour for the square piled raft tested in

three different densities of sand was analysed. Figure 4.67 presents the load-
settlement curves for the three densities and the raft thickness of 8m. The pile
spacing adopted was 6d and the number of piles was 9.
The load-settlement curves indicate that for any given settlement of

the piled raft the load taken by the piled raft is tested in dense condition is
higher than other two densities. This indicates that the relative stiffness of raft
– pile – soil stiffness in dense condition is higher, which can be seen from the
characterisation curves compared in Figure 4.68 for all the three densities.
Table 4.20 provides the load taken by the piled raft under various bed
densities, for the pile spacing of 4d and 6d.
LOAD, kN
0 1 2 3 4 5 6 7 8 9 10
0
d = 10mm
4 t = 8mm
S=6d
B = 200 mm
mm
8
SETTLEMENT, MM
SETTLEMENT,
12
16
20 Loose
Med dense Bed density
Dense
24
Figure 4.67 Load-settlement response of piled raft for various bed

densities (L = 160mm)
166
Table 4.20 Effect of bed density on load taken and stiffness (N/mm)
(a) 4d Spacing and (b) 6d spacing (L = 160mm, t = 6mm &
d = 10mm)
(a) 4d Spacing
Bed
Load kN Stiffness Load kN Stiffness Load kN Stiffness
Loose 1.2 600 1.95 325 5.3 265
Medium 2.9 1450 4.9 814 9.1 455
Dense 3.6 1800 5.7 950 10.2 510
(b) 6d Spacing
Bed
Load kN Stiffness Load kN Stiffness Load kN Stiffness
Loose 0.8 400 1.6 267 3.8 190
Medium 1.7 850 3.85 642 7.2 360
Dense 2.1 1050 4.6 766 8.8 440
It is seen that in the case of loose sand, the stiffness at any

settlement level (in all the three phases) is less than medium dense and dense
sand conditions. The stiffness of piled raft in dense condition in all the three
phases is more than two times the stiffness of piled raft in loose condition.
10
B
LOAD, kN
A
1
B = 200mm
t = 8mm
d = 10mm
Loose
L = 160mm
Medium S = 6d
Dense N=9
O
0.1
0.0 5.0 10.0 15.0 20.0 25.0

SETTLMENT, m m
Figure 4.68 Characterisation of load-settlement response of piled raft

with different bed densities
167
The importance of the study on the load sharing behaviour has been
dealt with, while discussing the behaviour of circular piled raft. In the
foregoing sections the load share between the two components (raft and piles)
of square piled raft under various parametric conditions are discussed in terms
of load sharing ratio, αpr . The value of αpr is already defined in section 4.46
(Eq 4b) and its value can range from zero for a plain raft foundation to one for
a pile group. However for piled raft the αpr value may lie between the value of
zero and one, which depends on the settlement caused by the applied load
apart from other parameters such as spacing of pile group, pile length, pile
diameter, density etc.
The values of αpr derived from the tests conducted on piled rafts
with raft thickness of 8mm and 10mm are presented in Figure 4.69 for piles
arranged at different spacings. In this figure the variation of αpr with
settlement is presented for the square piled raft tested in medium dense sand.
Starting with a value of approximately 0.6 (i.e close to settlement of 1mm) the
value of αpr decreases to 0.35 with increase in settlement for the pile spacing
of 4d. On increasing the pile spacing of a pile group, the αpr value is reduced
irrespective of the magnitude of settlement (or load) of piled raft with raft
thickness of 8mm and 10mm. Reduction in αpr value is attributed to lesser
pile-soil stiffness since number of piles of higher spacings (6d and 7.5d) is
nine while comparing with the 21 piles for the spacing of 4d.
168
∝ pr ∝ pr
0.00 0.20 0.40 0.60 0.80 1.00 0.00 0.20 0.40 0.60 0.80 1.00
0 0
4 4 d = 10mm
d = 10mm L=160mm
L=160mm t = 10mm
t = 8mm S = 6d
S = 6d 8
SETTLEMENT, mm
8 B = 200mm
SETTLEMENT, mm
MM
MM
B = 200mm Bed = Medium dense

Bed = Medium dense
12 12
16 16
20 4d
20 4d Piled raft Piled raft
6d pile 6d pile
7.5d spacing 7.5d spacing
24
24
(a) 8mm raft (b) 10mm raft
Figure 4.69 Variation in αpr with settlement for various spacing

(a) 8mm raft (b) 10mm raft
Initially the αpr value is high because the load, which is producing
smaller settlement of piled raft system is effectively carried by the piles. On
increasing the load, the settlement of the piled raft increased and the raft is
involved effectively in sharing the load thus αpr reduces. However the piles
share major part of the applied load which is causing settlement lesser than
2mm and raft shares major part of the load for which settlement is more than
6mm.
The general trend in load sharing ratio with settlement is same as

the trend seen in circular piled raft. From the study on square and circular
piled raft, the characteristic response of αpr with settlement is unique which is
independent of pile lengths, density of sand, diameter of pile and number of
piles. However, the actual value of αpr of a pile raft is influenced by all the
parameters indicated above.
The effect of spacing of the piles on αpr is presented in Figure 4.70a

for 8mm thick raft and Figure 4.70b for raft thickness of 8mm and 10mm
respectively.
169
0.80
2mm 20mm
0.60
∝ pr
0.40
t = 8mm
0.20 d = 10mm
L = 160mm
Bed = Medium dense
0.00
2 4 6 8 10
S/d
Figure 4.70 (a) Variation of αpr with pile spacing for 8mm raft
1.00
2mm 20mm
0.80
0.60
∝ pr
0.40
t = 10mm
d = 10mm
0.20
L = 160mm
Bed = Medium dense
0.00
2 4 6 8 10
S/d
Figure 4.70 (b) Variation of αpr with pile spacing for 10mm raft
As can be seen from the figures that the variation of αpr between
pile group of 7.5d and 6d spacings is much smaller than the variation of αpr
with 6d and 4d spacings at all the settlement levels. Even though the number
of piles is the same for 6d and 7.5d spacings (nine piles). In the case of 7.5d,
the pile to raft interaction could not be effective as in closer spacings. Hence
providing piles beyond 6d spacing may not produce any added advantage
though interaction between the piles at wider spacing is almost nil. Further it
is to be noted that the reduction in αpr between the pile spacing of 4d and 6d is
rapid, which indicates pile – soil stiffness is important parameter in sharing
the load particularly in the initial stages of load.
170
The plot given in the Figures 4.71 and 4.72 represent the variation
of αpr with settlement for piled raft with various pile lengths in loose and
dense conditions of sand.
∝ pr
0.00 0.20 0.40 0.60 0.80 1.00
0
d = 10mm
4 t = 8mm
S=6d
B = 200mm
Bed = Loose
8
mm
SETTLEMENT, MM
12
16
20 200mm Piled raf t

160mm pile
length
120mm
24
Figure 4.71 Variation of αpr with settlement for various pile lengths at
6d spacing in loose sand
The response of αpr with settlement is same for various lengths of

pile in both densities of bed. This response is reinforcing the earlier
observations made in circular piled raft as well as square piled raft. As long as
the settlement caused by the load is less than 1% of the lateral dimension of
the raft, the piles share more load whereas for the load causing settlement
more than 3%, the raft shares more load. Moreover the pile share remains
almost constant for the later case indicating that piles serve as settlement
reducer.
Figure 4.73 presents the variation of αpr with length. It is seen that
irrespective of the densities the αpr values are more for longer piles for the
settlement of 2mm and 20mm. Thus longer is the pile, higher is the αpr value,
171
which is independent of settlement of piled raft. However the increase in αpr

with pile length is not direct proportion of pile length provided, particularly
for the settlement of 2mm. For the pile length more than 0.8B the rate of
increase in αpr is marginally reduced which is pointing towards that piles of
longer length may not be advantageous. This may be attributed to arching in
the case of longer piles in sand and reduction in the increase of additional
stress due to load transfer between the raft and the soil at deeper depths.
∝ pr
0.00 0.20 0.40 0.60 0.80 1.00
0
4 d = 10mm
t = 8mm
6 S = 6d
8 B = 200mm
SETTLEMENT, mm
Bed = Dense
10
12
14
16
18
200mm
20 Piled raft
160mm Pile length
22 120mm
24
Figure 4.72 Variation in αpr with Settlement for various lengths at

6d spacing in dense sand
p
0.80
0.60
∝ pr VALUE
0.40
t = 8mm
0.20 d = 10mm
S = 6d
Bed = Loose 2mm 20mm
0.00
0.4 0.6 0.8 1.0 1.2
L/B
Figure 4.73 (a) Variation of αpr with pile length for loose bed
172
0.80
2mm 20mm
0.60
∝ pr VALUE
0.40
t = 8mm
0.20 d = 10mm
S = 6d
Bed = Dense
0.00
0.4 0.6 0.8 1.0 1.2
L/B
[ Figure 4.73 (b) Variation of αpr with pile length for dense bed
4.5.5.3 Effect of Pile Diameter
The variation of αpr with settlement for the piled raft with various
diameters of pile is presented in Figure 4.74 both for loose and dense sand
conditions. From the figures it can be said that the variation of αpr with
settlement is identical for the piles of different diameters. Though the trend is
the same αpr value is higher for the pile group with piles of larger diameter in
both the densities. However in dense sand the reduction in αpr is rapid with
settlement and showed almost constant αpr value for the settlement more than
6mm. This indicates that the pile group is more stiff and rigid in dense sand
than in loose sand. The effect of diameter of pile on load sharing ratio is
analysed in terms of d/t ratio (d = diameter of pile and t = thickness of raft)
and is presented in Figures 4.75 and 4.76 for loose and dense sand conditions.
For the settlement values compared in figures, the αpr value is increased with
d/t ratio both in loose and dense sand. It increases almost linearly with d/t
ratio, which indicates that the load sharing ratio has linear relation with
surface area of piles of pile group. The total surface area of piles is higher for
the piles of larger diameter and is directly proportional to pile diameter. Thus
the frictional resistance of pile is higher for larger diameter pile, which has
contributed for higher αpr value.
173
∝ pr VALUE ∝ pr VALUE
0.00 0.20 0.40 0.60 0.80 1.00 0.00 0.20 0.40 0.60 0.80 1.00
0 0
4 4
L = 160 mm
L = 160 mm
t = 8mm
t = 8mm
S = 6d
8 S = 6d
SETTLEMENT, mm
8
SETTLEMENT, mm
B = 200mm
B = 200mm
Bed = Loose
Bed = dense
12 12
16 16
20 20
6mm Piled raft 6mm Piled raft
8mm pile dia pile dia
10mm 10mm
24 24
Figure 4.74 Variation of αpr with settlement for various pile diameters in
loose sand bed and dense sand bed
1.00
t = 8mm
0.80 S = 6d
L = 160mm
∝pr VALUE
0.60 Bed = Loose

αpr
0.40
0.20
2mm 20mm
0.00
0.50 0.75 1.00 1.25 1.50
d/t RATIO
Figure 4.75 Variation of αpr with d/t ratio for loose bed
0.60
t = 8mm
S = 6d
L = 160mm
0.40
∝ pr VALUE
Bed = Dense
0.20
2mm 20mm
0.00
0.50 0.75 1.00 1.25 1.50
d/t RATIO
Figure 4.76 Variation of αpr with d/t for dense bed

174
4.5.5.4 Effect of pile configuration
In outer to bring out the arrangement of piles on settlement

reduction (SR) and load sharing ration (αpr), two different pile configurations
are considered as shown in Figure 4.77 and their response is compared with
the piled raft arranged uniformly (fully piled) over entire area. All the
parameters are kept constant expect number of piles and pile spacing. The
piled raft with piles arranged only in the inner and outer rows are 9 number
whereas fully piled raft consists of 25 numbers.
CONFIGURATION A CONFIGURATION B NO OF PILES 25

NO OF PILES 16 NO OF PILES 9 SPACING 4d
Figure 4.77 Configurations studied
In Figures 4.78 (a) and (b) the variation of SR and αpr respectively
are compared between three pile configurations. For the fully piled raft
(configuration C) system the SR and αpr values are much higher than other two
configurations (A and B) since, number of piles are more than twice the piles
provided in configurations A and B. Between the two configurations of A and
B, the SR value is marginally higher for configuration B for the settlement less
than 6mm, whereas for higher settlement (or higher load), the SR value is
higher for configuration A. This may be due to the piles provided at the outer
periphery of the piled raft are providing support to the soil at the edges of the
raft against early yielding as it would happen in the configuration B. Slightly
higher αpr for the configuration A also supports this observation. The αpr value
175
for the configuration A is for less than the configuration C but higher than the
αpr value of configuration B (inner row piles) irrespective of the magnitude of
settlement. Between the two configurations A and B tested, the configuration
A (ie piles in the outer periphery) is slightly efficient in load sharing and
controlling settlement of piled raft subjected to uniformly distributed load.
1.00 0.80 n=9 t = 8mm

n=9 t = 8mm FULLY PILED
L = 160mm B = 200mm FULLY PILED
L = 160mm B = 200mm
0.80 OUTER d = 10mm
d = 10mm 0.60 OUTER
INNER
0.60 INNER
∝ pr
SR
0.40
0.40
0.20
0.20
0.00 0.00
0 4 8 12 16 20 24 0 4 8 12 16 20 24
SETTLEMENT, mm SETTLEMENT, mm
Figure 4.78 Variation of SR and αpr with Settlement for various

configuration for medium dense sand
As discussed earlier, the settlement reduction is the main

requirement of a piled raft as a foundation system and this is true irrespective
of the shape of the raft. In this particular study, it is to be noted that in many
of the cases the number of piles is 9 and the area ratio being as small as
1.75%. Hence the present study gains considerable importance, that even
when the area ratio is as small as 1.75%, it is shown that the piled raft is
capable of producing considerable amount of settlement reduction. The
foregoing section presents the effect of various parameters on the settlement
reduction behaviour of the square piled raft.
Figure 4.79 presents the variation of SR with the spacing. In general

the SR decreases with increasing spacing between the piles of a given pile
group. The change in SR value between the pile spacings of 4d and 6d is
176
higher than the difference in SR between the spacings of 6d and 7.5d. Trend as
stated above is seen in tests conducted on sand beds of other two densities.
The reduction in settlement around 60% for the s/d ratio of 4 is attributed to
more number of piles. But for the pile spacing of 6d and 7.5d, the difference
in SR value between them is not appreciable since number of piles provided is
same for both the pile spacings. However thickness of the raft also adds to the
spacing of piles in settlement. For thinner raft SR is lesser and the difference
in SR between the spacings of 6d and 7.5 d is also higher.
R
1.00 0.80
2mm 20mm 2mm 20mm
0.80
0.60
0.60
SR
SR 0.40
0.40 t = 10mm t = 8mm
d = 10mm d = 10mm
0.20
0.20 L = 160mm
L = 160mm
Bed = Medium dense
Bed = Medium dense
0.00 0.00
2 4 6 8 10 2 4 6 8 10
S/d S/d
(a) 10mm thick raft (b) 8mm thick raft
Figure 4.79 Variation of SR with pile spacing (a) t =10mm (b) t = 8 mm
Figure 4.80 provides the variation of SR with the d/t ratio of piled
raft with raft thickness of 8mm. From the figures, it can be said that the SR
value increases with d/t ratio both in loose and dense sand beds. The SR value
is higher in loose sand than dense condition for the entire range of settlement
recorded in the raft.
R
1.00 0.60
t = 8mm t = 8mm
0.80 S = 6d S = 6d
L = 160mm L = 160mm
SR VALUE
0.40
SR VALUE
0.60 Bed = Loose Bed = Dense
0.40
0.20
0.20
2mm 20mm
2mm 20mm
0.00 0.00
0.50 0.75 1.00 1.25 1.50 0.50 0.75 1.00 1.25 1.50
d/t RATIO d/t RATIO
Figure 4.80 Variation of SR with pile diameter for different bed densities
177
The variation of SR value with length has been plotted by

representing the length as L/B ratio (pile length to raft width ratio) in Figure
4.81, for loose and medium bed densities. In both the density conditions, the
SR increases nonlinearly with L/B ratio but at decreasing rate. For the
settlement of 2mm, the reduction in settlement of piled raft is higher
irrespective of pile length and density of sand. However, reduction in
settlement is lesser for the pile lengths more than 0.8B. Thus, increasing the
pile length more than 0.8B is not contributing appreciably in reducing the
settlement particularly in loose sand.
0.80 0.80
t = 8mm 2mm 20mm
d = 10mm
0.60 0.60 S = 6d
SR VALUE
SR VALUE
Bed = Medium dense

0.40 0.40
t = 8mm
0.20 d = 10mm 0.20
S = 6d
2mm 20mm
Bed = Loose
0.00 0.00
0.4 0.6 0.8 1.0 1.2 0.4 0.6 0.8 1.0 1.2
L/B L/B
Figure 4.81 Variation of SR with pile length for different bed densities
4.6 BEHAVIOR OF RECTANGULAR PILED RAFT
Although the circular and square piled rafts have given adequate
information on the behaviour of piled raft, it was still felt necessary to study
the behaviour of strip piled raft. Hence tests were conducted on a 70mmx
200mm strip piled raft and the Table 4.21 presents the details of the pile and
the raft used in the study.
178
Table 4.21 Details of pile, raft and bed
Raft model size Pile dia Pile length

(mm3) (mm) (mm)
10 120,160,100,75
70 x 200 x 8 8 160,120,100
6 160,120,100
The effect of pile length, diameter and the piled raft area ratio were
studied. In studying the pile area ratio, the spacing of the piles were varied
from 4d to 7.5d. The number of piles and the area ratio considered in the
study are presented in Table 4.22.
Table 4.22 Details of pile configuration
No. of piles Spacing Area ratio %

14 4d 7.80
8 5d 4.40
6 6d 3.80
4 7.5d 2.20
Since it was established that the pile group of piled raft takes more
load than the free standing pile group both in the case of circular and square
raft, it was felt that such a study is not required in the case of rectangular raft.
Tests were conducted on medium dense sand bed with 8mm thick raft.
4.6.1 Behaviour of Plain Raft, Piled Raft and Pile Group of Piled
Raft
Figure 4.82 presents the load settlement response of plain raft, piled
raft and pile group of piled raft. Figure 4.83 presents the characterization
curves for plain and piled raft. Both plain and piled raft show three phase
behaviour as observed in tests on square and circular piled rafts. The load
taken by the plain raft and the piled raft along with the stiffness are presented
in Table 4.23.
179
Table 4.23 Comparison of load and stiffness (N/mm) for Plain raft and
piled raft (Size = 70mm x 200mm, d = 10mm, Ar= 3.75%,
N = 6, L=160mm, t = 8mm)
OA AB BC
System Load in Load in Load in
Stiffness Stiffness Stiffness
kN kN kN
Plain raft 0.41 171 0.8 150 1.5 100
Piled raft 0.69 530 1.08 219 2.2 100
LO A D IN k N
0 20 40 60 80LOAD,
100kN120 140 160 180 200
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
00
RAFT
22
PILE
44
SETTLEMENT, mm
PILED RAFT
66
8
8
1010 t=8mm, d=10mm
l=75mm, s = 6d
1212 Raft 70mm x200mm
Bed = Medium dense
1414
Figure 4.82 Load settlement response of raft, pile and piled raft
1000
t=8mm, d=10mm
PILED RAFT l=200mm, s = 6d
PLAIN RAFT Raf t 70mm x200mm
Bed = Medium dense
LOAD, kN X 10-2
C
100
B
10
O
0 2 4 6 8 10 12 14
SETTLEMENT, m m
Figure 4.83 Characterisation curve of load - settlement response

180
It can be seen from the table that addition of a pile group having a
pile-raft area ratio of 3.8% could enhance the carrying capacity of the raft soil
system by 43% and at the same time reducing the settlement by 37.5%. In the
case of plain raft, the rate of increase in the settlement was gradual up to a
load level of 0.45kN, beyond which the settlement rate was higher; 12mm
was the maximum settlement beyond which the plain raft did not take any
further load. However in the case of piled raft the rate of increase in the
settlement was very small till the load of 0.6kN. From there and up to the load
of 1.6kN, the rate of increase in the settlement is gradual; but at any
settlement level, the piled raft was taking more load than the plain raft, At
2mm settlement, the load taken by the piled raft was almost 100% higher than
the raft and at 12mm the load taken by the piled raft was 43% higher than the
load taken by the plain raft. It is to be noted here that at higher settlement the
difference in the load taken by the plain raft and the piled raft is reduces.
The load settlement curve derived for the pile group of piled raft,
indicates that after a load level of 0.15kN, even though the pile group
settlement was rapid, it took additional load indicating a similar behaviour as
in the case of circular and square piled rafts. The characterization curve
presented in Figure 4.83 clearly indicates that the plain raft and the pile group
of piled raft present two phase behaviour. The loss of stiffness in the case of
plain raft and the pile group of piled raft is very rapid than the piled raft. The
three phase behaviour of the load settlement curve of rectangular piled raft
indicate that the behaviour of piled raft is similar irrespective of the shape of
the piled raft and beyond a settlement level of 7% of the least lateral
dimension of the raft the stiffness of the piled raft is almost equal to that of
the plain raft.
The tests were done with various pile lengths, pile diameters and
pile-raft area ratio. The behaviour was found to be identical. The load taken
and settlement reduction achieved for the piled raft with 10mm dia pile for
various lengths are presented in Table 4.24. The effect of pile diameter on the
181
load taken and settlement reduction achieved for the piled raft with piles of
100mm length illustrated in Table 4.25. Table 4.26 gives the above values for
various area ratios.
Table 4.24 Effect of pile-raft area ratio on load taken by the piled raft
and settlement reduction

Length Load kN SR Load kN SR Load kN SR
200 mm 0.76 0.70 1.54 0.61 2.12 0.61
160 mm 0.54 0.65 1.40 0.50 2.04 0.58
120 mm 0.70 0.60 1.40 0.45 1.80 0.53
100 mm 0.70 0.53 1.34 0.38 1.76 0.48
Table 4.25 Effect of pile diameter on the load taken by the piled raft

Diameter Load kN SR Load kN SR Load kN SR
10 mm 0.70 0.65 1.34 0.48 1.76 0.48
8 mm 0.72 0.60 1.33 0.41 1.70 0.45
6 mm 0.40 0.50 0.80 0.30 1.10 0.28
Table 4.26 Effect of pile-raft area ratio on load taken by the piled raft
Pile–raft @ 2mm sett @ 6mm sett @ 12mm sett

area ratio Load kN SR Load kN SR Load kN SR
2.2% 0.50 0.36 1.15 0.30 1.58 0.22
3.8% 0.50 0.48 1.16 0.40 1.60 0.34
4.4% 0.54 0.56 1.30 0.46 1.64 0.40
7.8% 0.76 0.60 1.60 0.53 1.72 0.43
182
The settlement reduction behaviour of the rectangular piled raft was

found to be similar to that of the circular and square piled raft. The effect of
pile length, pile diameter and the pile raft area ratio on the settlement
reduction behaviour of the piled raft are presented below:
Figure 4.84 presents the variation of SR value with length. As can

be seen from the figure the SR value increases with the length, beyond the
length of 120mm the rate of increase in the SR value with the length reduces.
The tests conducted on 8mm and 6mm dia piles also showed similar trend.
Vs SR
1.00 1.00
0.80 0.80
SR VALUE
SR VALUE
0.60 0.60
0.40 0.40
0.20 0.20
2MM 12MM 2MM 12MM
0.00 0.00
25 50 75 100 125 150 175 200 225 80 100 120 140 160 180
LENGTH, mm
LENGTH (m m) LENGTH
LENGTH,(m m
mm)
(a) 10mm dia pile (b) 8mm dia pile
Figure 4.84 Variation of SR with pile length for various pile diameters
Figure 4.85 presents the effect of variation of SR value with the pile
diameter. It is seen from the figure that the behaviour is bilinear at all
settlement levels. The rate of increase in the SR value is high for the piles of
lesser diameter when compared with higher diameter. With the raft thickness
being 8mm, it indicates that when the d/t ratio increases beyond unity, the
increase in the SR value reduces.
Figure 4.86 presents the effect of pile-raft area ratio on the SR

value. Here also the behaviour is bilinear. Further as seen in the case of
circular piled raft increasing the area ratio beyond 6% is not effective in
reducing the settlement in the case of rectangular piled raft also.
183
0.8 0.8
0.7 0.7
0.6 0.6
SR VALUE
SR VALUE
0.5 0.5
0.4 0.4
0.3 0.3
0.2 0.2
0.1 2 MM 12 MM 0.1 2 MM 12 MM
0.0 0.0
4 6 8 10 12 4 6 8 10 12
PILE DIAMETER, mm PILE DIAMETER, mm
(a) Pile length 120mm (b) pile length 100mm
Figure 4.85 Variation of SR with pile diameter
1.00
0.80
SR VALUE
0.60
0.40
0.20
2MM 12MM
0.00
0.00 2.00 4.00 6.00 8.00 10.00
AREA RATIO
Figure 4.86 Variation of SR with Area Ratio
Figure 4.87a presents typically the load sharing behaviour of the

rectangular piled raft which has a pile-raft area ratio of 3.75%. The raft is
8mm thick and the piles are 10mm diameter. The figure presents the load
sharing behaviour for various lengths. The load sharing ratio αPR is plotted
against the settlement. It is seen that the behaviour is identical to that of
circular and square piled raft. Irrespective of the length and the pile-raft area
ratio, the load sharing ratio αPR reduces rapidly with settlement upto a
settlement level of 6mm when the length is more than 75mm which is the
least lateral dimension of the raft. When the pile length is 75mm (Refer
Figure 4.87b) the αPR value reduces rapidly from the beginning. Upto a
settlement level of 3mm. This trend is same as that of circular and square raft.
But beyond this level, there is a further decrease in the value and only after
184
8mm the value remains constant. This indicates that in the case of strip raft,
when the load sharing becomes a criteria, the pile length has to be larger.
αPR αPR
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.000 0.100 0.200 0.300 0.400 0.500 0.600 0.700 0.800
0 0
1 1
2 2
3 3
4
N=8
4 L=75mm
S = 6D
5 5 D=10mm
SETTLEMENT, mm
B = MD
SE T T L EM E N T mm
MM
6 t = 8mm 6 B = MD
SETTLEMENT,
7 t = 8mm
7
8 8
9 9
10 10
200mm
11 11
160mm 7.5D
12 12
120mm 6.0D
13 100mm 13 5.0D
14 14 4.0D
15 15
(a) for various lengths (b) for various spacing
Figure 4.87 Variation of αpr Vs Settlement for 10mm dia pile for
various lengths and spacings
4.7 SUMMARY
From the small scale model studies it was seen that irrespective of
the shape of the raft the load settlement response of the plain raft and the piled
raft was the same. The response was found to be three phased, within the
settlement range of 10% of the least lateral dimension of the raft. It was found
that upto a settlement level of 1% of the least lateral dimension of the raft,
which is close to the critical settlement; the load settlement response was
linear. The load settlement response of the piled raft is characterized as multi
linear strain hardening behavior. Through the characteristic response, it was
seen that although the pile-raft-soil stiffness is higher than the raft soil
stiffness in all the three phases of the load settlement response, in the third
185
phase the pile raft soil stiffness was close to the raft soil stiffness indicating
that the contribution of pile group is reduced drastically and its share on the
load applied remained constant.
It was found that the pile length beyond 80% of the least lateral
dimension of the raft did not have any appreciable influence on the pile soil
stiffness. Similarly increasing the pile- raft area ratio beyond 6% was not
contributing to the stiffness of the piled raft appreciably. In all the cases
studied, the pile group shared major part of the load upto a settlement level of
1% of the least lateral dimension of the raft. As the settlement increased the
load sharing ratio decreased and remained constant beyond a settlement level
equal to 3% of the least lateral dimension of the raft. The performance of the
piled raft was found to be better from settlement reduction and load sharing
point of view in loose and medium dense sand till a settlement level of 3% of
the least lateral dimension of the raft. But at higher settlement level the
settlement reduction was found to be lesser in loose sand than in medium
dense sand. The parametric study has indicated that the pile length and
diameter have a predominant role on the behavior of the piled raft; the
number of piles in the form of pile raft area ratio beyond 5% to 6% does not
have any pronounced influence on the behavior.
Comparing the behavior of square and circular raft, under identical

conditions the square piled raft took more load than the circular piled raft as
the square piled raft could accommodate more number of piles for the same
area ratio. Significantly in either case at 20mm settlement of the plain raft the
settlement reduction ratio was found to be around 0.4 and the load sharing
ratio was around 0.35 to 0.38. It was also found that in either case the piles
have to be placed evenly on the total area of the raft for better performance.
186
4.8 LIMITATIONS AND ADVANTAGES/USE FULLNESS OF

1g MODEL TEST
The small scale model studies facilitates having a close control on

the variables met in the practice and study their influence on the behavior of
piled raft. Although these results cannot be compared directly with full scale
behavior, they are very useful in providing an understanding on the behavior
pattern, as a guide to full scale performance when examined in conjunction
with the theoretical solutions developed. Similar view has been expressed by
many researchers such as Murray and Geddes (1989). In this study the
concentration was on the non dimensional parameters namely settlement
reduction ratio load sharing ratio as the piled raft behavior is influenced by
the settlement. While full scale tests are ideal, they are time consuming and
prohibitively costly. Also the boundary conditions for the tests may not be as
clear as that can be created in the laboratory. This can lead to the complexity
in the analyses of the observed data. Also full scale studies are not amenable
for parametric study.
In spite of the fact that stresses of the field conditions particularly

in sand cannot be modeled, the model tests have the following advantages:
1) It is possible to model very flexible raft
2) Parametric studies of the foundation are relatively easier and

thus more economical than the full scale tests.
3) Boundary conditions of the model are generally clear than

those found for the full scale tests.
The main disadvantages are :
1) The existence of the rigid side walls and rigid base may
affect the behavior of the model. However this was
eliminated by providing the tank of adequate size.
187
2) In 1g model tests, the soil used in prototype is normally used

as test medium. This does not satisfy the law of similarity as
the grain size and the ratio of particle size and the foundation
size is not scaled, leading to scale effect. This effect is
significant in cohesion less material. But this effect has been
studied by many and proved insignificant if the ratio of the
foundation width or length to the D50 is less than 30.
188
CHAPTER 5
NUMERICAL ANALYSES OF
1g MODEL TESTS OF PILED RAFT
5.1 GENERAL
In the chapter 4 results of tests conducted on model piled rafts were

presented and discussed. Influence of various piled raft parameters on load
sharing between pile and raft were also discussed. However it is very
important to compare the results of the model tests quantitatively, either with
the results from closed form techniques or from numerical approaches. For
any geotechnical analysis closed form techniques provide the best results.
Closed form solutions generally satisfy both compatibility and equilibrium
conditions. But this technique becomes very difficult for complicated
problems like piled raft foundation. Under such conditions, if it is possible to
establish a numerical model comprising of appropriate material model and
constitutive relations, combining them with equations of equilibrium and
compatibility realistic solution can be obtained. Even though the numerical
solution may not be exact, it is still appropriate and acceptable for solving real
time problems. The inaccuracy is due to various assumptions made relating to
geometry, idealization of material behaviour, solution technique, boundary
conditions etc. Yet it is possible to obtain reasonably accurate solutions for
the varieties of geotechnical problems through numerical analysis.
The present problem of piled raft is associated with movement of

the foundation system due to the structural loading and interaction between
three independent components of foundation system namely raft, pile and soil.
In most of the structures, construction / design loads generate stresses which
189
are normally within the elastic range; hence idealization of the material as
elastic becomes an acceptable assumption. But in the case of soil this is not
always true; therefore proper idealisation for its non-linear behaviour is very
important in numerical analysis. Even though the prediction of the behaviour
from numerical analysis may not be very realistic, (Potts and Zdrakovic,
2000) if the predicted and the observed results agree to a reasonable degree, it
can be concluded that the particular model can be used for real problem.
Among the various tools available for solving the numerical model,
finite element is one of the versatile tools available for solving varieties of
engineering problems, since the elements can be arranged in any fashion and
they can be used to model any shapes. The method is based on simple
algebraic equations devoid of any high level mathematics. Varieties of
interactive packages have been made available and this has made the FEM a
very popular tool for engineering analyses.
The two main aspects of finite element analysis are element

discretisation and mesh refinement. In all the cases, a mesh of rectangular
shaped elements provides the best results. In the present analysis this aspect
has been given due importance and consideration. In the numerical study of
soil structure interaction the other important aspect to be considered is the
interface idealization. The interface idealization is to be done based on the
output required. For example in the analysis of laterally loaded pile the
interface idealisation plays an important role. Presently the concentration is
more on the settlement reduction and load sharing. Unlike the study of
laterally loaded pile, in the present study the provision of interface elements
will increase the stiffness of the pile and hence as indicated by Potts and
Zdrakovick (2000) it was felt that the interface element need not be
considered between the pile and soil. However contact was taken as perfect
contact between the pile and the soil as assumed by Reul (2000).
190
5.2 SELECTION OF SOFTWARE
The effectiveness of the finite element software depends upon the

use of efficient programming methods and the sophisticated use of hardware.
In order to understand the behaviour of piled raft and to compare the
analytical results with experiments, the FE package ANSYS was chosen as
the tool of analysis. The ANSYS is a total research package with more than
200 elements. With these elements different problems in the engineering field
can be solved.
The package finds its application in many engineering fields such as,
• Structural
• Thermal
• Fluids and
• Electro magnetic
The present analysis focuses on an interaction problem, wherein

three interacting components are involved in the load transfer mechanism.
The three elements are soil, raft and piles of the pile group. The interaction
between the raft –soil, soil- pile; pile-soil and finally in total, the raft-soil-pile
interaction is analysed to generate the displacement and the load shared
between the raft and the pile at a particular level of displacement. In order to
achieve this, among the methods available in ANSYS, the “static analysis”
from structural model is adopted. The primary unknown calculated in static
structural analysis is displacement. The other values associated with the
output such as strains, stresses and reaction forces are then derived from the
actual displacements. The pre- processor generates the model, the solution
module solves the problem and the post processor exhibits the results.
Generating the model in ANSYS is done through the pre-processor

taking the preference as “structural” in the first step. The first essential
191
requirement in modelling is selection of element, and the boundary conditions

to be used. The second level requirements are the material model and the
selection of type of analysis, namely linear or nonlinear. The selected problem
can also be solved as large deformation problem. The third level requirement
will be the meshing and the extent of mesh refinement required. Depending
upon the type of problem, the above aspects are selected before executing the
numerical model under the solution command. In this particular study, the
following analysis were carried out and the results are compared with the 1g
model tests for the validation of the numerical model.
• Linear analysis
• Axisymmetric analysis
• Plane strain analysis
• Nonlinear analysis
The linear analysis has been carried out in the case of circular and
square piled raft to serve as a preliminary study and also to check and
compare the settlement level up to which the results of 1g model studies and
the numerical analysis agree with each other. The circular piled raft has been
subjected to axisymmetric analysis based on the principles outlined by
Hooper (1977) in his study on piled raft foundation of Cavalry Barracks
building. This provides an idea for adopting certain idealisations in the
numerical modelling of piled raft. The rectangular piled raft has been
modelled as plane strain problem. These studies indicate the applicability of
simpler numerical models to study the behaviour of piled raft, both circular
and non circular shapes. The three dimensional nonlinear analysis conducted
on the circular and square piled rafts have provided the load -settlement and
also load sharing behaviour of the piled raft and by comparing these results
with the 1g model test results, the applicability of numerical model generated
with ANSYS is established.
192
5.3 VALIDATION OF NUMERICAL TOOL ANSYS
The package selected for performing the numerical analysis needs

to be tested for its applicability to the problem of soil- structure interaction.
The problem of soil structure interaction is more complicated as this involves
two different elements in contact with each other. The problem can be a pure
deformation problem or deformation combined with bending. For example,
the problems of raft can be a pure settlement problem or settlement cum
bending. These problems have been solved in a simpler way by assuming soil
continuum as springs of appropriate stiffness. But piled raft has one more
component namely load sharing. This is due to the addition of piles. Hence
piled raft problem becomes more complex. Thus ANSYS finite element code
chosen for this study needs validation for its applicability to analyse piled raft
foundation.
5.3.1 Material Model
The ANSYS material table contains more than 40 material models.

Out of these, the linear elastic and the multi-linear isotropic hardening
material (MISO) model are used to simulate the linear elastic and the
nonlinear behaviour of the soil respectively.
5.3.2 Linear Elastic Model
The basic assumption of the elastic behaviour is that the stress is

proportional to the strain.
σ=Eε (5.1)
where σ - stress
E - modulus of elasticity
ε - strain
193
The elastic constitutive model provided in ANSYS can take

different forms, namely isotropic, anisotropic and so on. An isotropic material
is one that has point symmetry and requires two elastic constants. They are
Young’s modulus, E and Poisson’s ratio, μ.
5.3.3 Miso Model
The soil medium below the raft was modeled using an eight-node
brick element (SOLID 45) having three degrees of freedom of translations in
the respective co-ordinate directions at each node. The soil was idealized as
an isotropic homogenous half-space. The nonlinear behaviour of the soil is
modeled using the multi-linear isotropic hardening (MISO) material model of
ANSYS. This model incorporates the Von-Mises yield criterion with
associated flow rule and isotropic work hardening. The equivalent stress,
σe = [1.5{S}T[M]{S}]0.5 (5.2)
where, {S} - the deviator stress and
[M] - the Mass matrix.
The material is assumed to satisfy the yield condition when σe is

equal to the yield stress.
The yield criterion is,
F = [1.5{S}T[M]{S}]0.5- σk = 0 (5.3)
For work hardening, σk is a function of the amount of the plastic

work done. For the case of the isotropic plasticity assumed here, σk is the
output as the equivalent stress parameter. The material properties to be given
as the input in addition to the stress-strain values are the Young’s modulus
(E), the Poisson’s ratio (μ), and the shear modulus (G).
To provide the required parameters as the input for the MISO

model triaxial tests were conducted on dry Palar river sand used in the
194
experiments. The test was conducted at an average unit weight of 15.5kN/m3

(15.5 + 0.1kN/m3) under different confining pressures. A value of 0.35 was
used in computation for Poisson’s ratio.
5.3.4 Loading Options
The main goal of a finite element analysis is to examine how a

structure or component responds to certain loading conditions. Specifying the
proper loading condition is therefore a key step in the analysis. Load can be
applied on the model in a variety of ways in the ANSYS program with the
help of the load control option. For the present study the load was applied
as a pressure.
5.3.5 The Continuum Element
The finite element method treats any continuum (soil and structure)
as an assemblage of finite elements whose boundaries are defined by the
nodal points, and assumes that the response of the continuum can be
described by the response of the nodal points. In ANSYS library, a variety of
continuum elements are included to represent the 2D and the 3D structural
analysis. The features and capabilities of plane 42 and solid 45 elements used
in 2D and 3D analyses respectively for the idealization of the soil and the raft
foundation are described in this section.
5.3.5.1 Plane 42
Two dimensional structural solid (Figure 5.1) is used for modelling

two dimensional solid structures. This element can be used as a plane element
(plane stress / plane strain) or as an axisymmetry element. It is defined by
four nodes having two degrees of freedom at each node namely translations in
the nodal x and y directions. The element has plasticity, creep, swelling stress,
stiffening, large deflection and large strain capabilities. In the load point data,
the nodes at which the loads are applied, number of load cases and the
magnitude of load or moment applied along with the directions at the nodes
195
are given. The details of load applicable in structural static analysis are shown
in Table 5.1.
Figure 5.1 Two dimensional structural solid element (Plane 42)
Table 5.1 Loads Applicable in a Structural Static Analysis
S. No. Load Name Load Category

i Displacement Constraints
(UX,UY,UZ,ROTX,ROTY,ROTZ)
ii Forces/moments Forces
(FX,FY,FZ,MX.MY,MZ)
iii Pressure(Pres) Surface loads
iv Temperature(Temp) Body loads
Fluences (Flue)
v Gravity, Spinning, etc. Inertia loads
Displacements and degrees of freedom constraints usually specified

at model boundaries to specify or define rigid support points. They are used to
indicate symmetric boundary conditions and points of known motion.
Forces/moments are concentrated loads usually specified on the model
exterior. Pressures are surface forces usually applied to model exterior.
Positive values of pressure act towards the element face.
196
5.3.5.2 Solid 45
Solid 45 is used for the 3D modelling of solid structures. The

element is defined by eight nodes having three degrees of freedom at each
node. They are translations in the nodal x, y, and z directions. The elements
have plasticity, creep, swelling, stress stiffening, large deflection and large
strain capabilities. The geometry, the node locations and the co-ordinate
system for this element are shown in Figure 5.2. The input data required for
the above element is as detailed in Table 5.2. The SOLID45 is used for
modelling the soil continuum, raft and pile elements.
Figure 5.2 SOLID 45 – Element Geometry
Table 5.2 SOLID 45 Input data
Element Name SOLID 45

Nodes I, J, K, L, M, N, O, P
Degree of Freedom UX, UY, UZ
Real Constants NIL
Material Properties EX, EY, EZ, PRXY, PRYZ, PRXZ, ALPX, ALPY,
ALPZ, DENS, GXY, GYZ, GXZ, DAMP
Surface Loads PRESSURE
197
5.4 VALIDATION OF SOLID 45 ELEMENT OF ANSYS
The soil continuum is modelled using eight noded Solid45 element

of ANSYS in the proposed finite element analysis of piled raft. Details of the
element and input parameters required are explained already in the section
5.3.5.2. The suitability of this element for the idealization of soil behaviour is
verified with closed form solution for uniform loading on deep soil. The
surface settlement of a uniform square loading on homogeneous soil is given
by the following expression:
qB (1 − μ s )
2
(5.4)
w = Iw
Es
where, w is the settlement at any point of loaded area, q is the applied uniform
pressure, B width of the loaded area, Es is the modulus soil, μs is the Poisson’s
ratio and Iw is the influence factor for settlement. The non-dimensional value
Iw is a function of the position in the loaded area. The settlement in the above
equation can be normalised as shown below:
w Es
Iw = 2
(5.5)
qB (1 − μ s )
For a typical location, values of Iw for the elastic semi-infinite

medium subjected to uniform square loading are as shown in Table 5.3.
Table 5.3 Iw for settlement of soil surface subjected to uniform over

square area loading
Method Value of Iw at different position

Closed Centre Corner Edge Centre-Corner Average
form 1.12 0.56 0.76 0.56 0.95
Since closed form solution is available for surface settlement of

uniformly loaded area, it is proposed to analyse the same through ANSYS
programme. The data taken for the analysis are as reported by Daniel
Thangaraj (2006) and are as detailed below:
Size of loaded area - 2mx2m

198
Intensity of load - 100 kN/m2

Modulus of soil - 10, 000 kN/m2
Poisson’s ratio of soil - 0.35
The Table 5.4 shows the details of cases analyzed. Idealization of

semi infinite medium like soil medium requires rigid boundaries both in the
lateral and vertical directions. This is required in the finite element analysis
to reduce memory space and time needed for computation without
compromising much on the accuracy of results. In the ANSYS three
dimensional analysis, the following boundary distance are assumed for the
loaded half width of the raft as 1 m.
Table 5.4 Analytical cases
Distance to boundaries (m)

Identification Number of elements
Width (m) Depth (m)
BD1 5 6 1204
BD2 5 10 2000
BD3 5 17 3400
The typical meshes used for the model are shown in Figure 5.3. The
calculated settlements from the ANSYS are presented in Table 5.5 along with
the closed form solution.
SOIL
PRESSURE
ON THE SOIL
10 m 10 m
10 m
10 m
1m 1m
Figure 5.3 Finite element discretization of soil continuum

199
Table 5.5 Normalised settlement of loaded area 2 m × 2 m for

conditions analysed
Width Depth Iw Centre - Iw at Average

Case
(m) (m) Centre Corner Corner Edge Iw
BD1 5 6 0.951 0.395 0.556 0.587 0.764
BD2 5 10 1.014 0.456 0.558 0.661 0.838
BD3 5 17 1.117 0.558 0.559 0.757 0.937
Closed form 1.12 0.56 0.56 0.76 0.95
According to the results of deep soil condition the variation

between the ANSYS results and the closed form solutions are shown in Table
5.6 for different boundary distances.
Table 5.6 Variation between ANSYS and closed form solution for
homogeneous soil
Differential settlement Average settlement

Analytical Case
(%) (%)
BD1 0.71 16.3
BD2 0.36 11.8
BD3 0.18 1.48
Results presented above show that the differential settlements

calculated from the finite element analysis are very close to those calculated
for deep soil condition. However average settlement is underestimated by
16.3% for the BD1 condition (width 5m and depth 6m) , compared with that
for a deep soil condition. The least variation is for the condition BD3 (width =
5 m and depth = 17 m), which is equal to 1.4%. This shows that the
magnitude of underestimation of average settlement becomes much smaller
with increase in thickness of the soil and the effect of thickness of soil on the
average settlement becomes negligible. Both the differential and the average
settlement are very close to the results of deep soil condition for the boundary
200
condition of BD3. The ANSYS results are considered sufficiently accurate,

particularly in the respect of low errors for the differential settlement.
Therefore SOLID 45 element of ANSYS is very much suitable for the
idealization of elastic behaviour of the soil. The effect of soil lateral
boundaries on the settlement of loaded area are also analysed as detailed in
Table 5.7 for the constant thickness of soil. The thickness of soil considered
in the analysis is 5 m which is 2.5 times the width of loaded area. From the
results presented in Table 5.7 it can be considered that the effect of the side
boundary distance on the differential settlement is negligible and the
maximum variation in average settlement between SB1 and SB5 is around
6%. Thus the increase in boundary distances more than 2.5 times the width of
the loaded area exerts little influence on the differential settlement. Further,
the calculated differential settlements are very close to those calculated for an
infinite boundary condition.
Table 5.7 Effect of side boundaries on the normalised settlement of

loaded area 2 m × 2 m
Analytical Width Depth Iw Centre- Iw at Average

Case (m) (m) Centre Corner Corner edge Iw
SB1 4 5 0.923 0.368 0.555 0.558 0.741
SB2 5 5 0.94 0.384 0.556 0.586 0.763
SB3 6 5 0.951 0.394 0.557 0.598 0.775
SB4 8 5 0.963 0.406 0.557 0.609 0.786
SB5 10 5 0.968 0.411 0.557 0.615 0.792
From the discussions presented above it is concluded that the

settlement is less sensitive to the adjacent boundaries (both along the depth
and width direction) particularly if the boundary distances are more than two
times the width of the loaded area. Further, the calculated differential
settlements (centre-corner) are close with one another for the boundary
conditions considered in this study. A similar observation was reported by
Brown (1969b) based on the numerical analyses of uniformly loaded circular
201
rafts on elastic layers of finite depth. The results and discussion presented
above indicate that the proximity of the boundary has least influence on the
settlement as long as it is away by more than twice the width of the loaded
area.
5.5 VALIDATION FOR PLANE STRAIN ANALYSIS
In order to assess the capability of the software in solving certain

idealized problem, it is required to test the software by solving similar
problem for which solution is available. Prakoso and Kulhawy (2001)
performed a piled raft problem by assuming plane strain condition. The same
problem was taken up and was analysed in ANSYS. The plane strain
condition involves in a fundamental simplification wherein a finite sized piled
raft is simplified into a strip piled raft. Desai et al., (1974) had established that
this type of model can provide good results and can be used to analyse a
relatively large piled raft without any excessive modelling and computing
time. Here each row of piles in the plane is converted into a plane strain pile
wall of an equivalent pile Young’s modulus as indicated in Equation 5.6.
n p −row Ap E p
Eeq = (5.6)
Lr B
where,
np-row : number of piles in a row
Ap : area of pile crosses section
Ep : pile Young’s modulus
Lr : raft length in plane
B : pile diameter.
Since the response of vertically loaded piles in the group is mainly

contributed by its axial stiffness, the plane strain pile model is represented in
ANSYS with the value of Eeq. The plane strain system geometry, material
202
properties, and the load are kept unchanged. In modelling this problem with
ANSYS, the members are represented by plane 42 elements. The properties of
the soil, raft, and the pile group are as detailed below.
Soil:
Width =100m
Height = 80m
Young’s modulus of the soil, Es = 20MN/m2
Poisson’s ratio, μs = 0.495
Raft:
Thickness = 2m
Width, Brr = 42m
Young’s modulus of raft Er = 20GN/m2
Poisson’s ratio, μr = 0.25
Pile group:
Pile Diameter, B = 1m
Width, Bg = 12.5m
Spacing, C = 4m
Length = 54.0m
21.0m
80m
100m
Figure 5.4 Finite element mesh for piled raft (plane strain analysis)
203
Figure 5.4 presents the FEA mesh of the problem chosen for the
comparative study. In ANSYS automatic mesh generation was adopted after
fixing the boundary for fine mesh. Symmetric boundary condition was
adopted along the centre line of the piled raft and displacements at two other
boundaries were arrested. The loading was applied as pressure on the raft and
the analysis was performed for various lengths of pile. The results were
plotted and compared with the output of the PLAXIS as presented by Prokoso
and Kulhawy (2001).
RAFT REFERENCE DISPLACEMENT
PILED RAFT DISPLACEMENT TO
DEPTH OF PILE, m
Figure 5.5 Reference displacement ratio v/s Pile depth

DIFFERENTIAL DISPLACEMENT
RATIO
DEPTH OF PILE, m
Figure 5.6 Differential displacement ratio v/s Pile depth
Figure 5.5 and Figure 5.6 present the comparison between the
results of PLAXIS and ANSYS in terms of reference displacement ratio and
differential displacement ratio respectively. The results were found to agree
very closely confirming the applicability of the ANSYS software.
204
5.6 ANALYSES OF CIRCULAR PILEDRAFT
The following analyses were carried out for circular piled raft.
• Linear Analysis
• Nonlinear Analysis
• Axisymmetric
• 3D - Analysis
The linear and axisymmetric (nonlinear) analyses are performed to

study the load settlement pattern alone but in the nonlinear analysis, in
addition to load- settlement response load sharing behaviour between raft and
piles have also been studied for a particular density of sand.
5.6.1 3d Linear Analysis of Piled Raft
Three dimensional analysis was performed on the circular piled raft

with 21 piles provided in the radial directions as shown in Figure 4.2. The
analysis was carried out to obtain the load settlement behaviour. The details
of piled raft and the bed material are presented in Table 5.8 and 5.9
respectively.
Table 5.8 Properties of circular piled raft
Raft Pile
Er μr
Dia Thickness Dia Length Area ratio
200mm 8mm 10mm 160mm 5.2% 3000 N/mm2 0.30
[
Table 5.9 Properties of bed material
Material γ’ φ Es μs State of compaction

Poorly graded
15.5kN/m3 37.50 35 N/mm2 0.30 Medium dense
sand
205
For the circular raft half model was used. The continuum and the
piled raft were represented by solid45 elements. Mesh refinement was done so
that fine mesh is formed near the pile and coarse mesh is formed in the
continuum away from the piles. Appropriate boundary conditions were
applied on the edges of the soil continuum and symmetric condition was
applied on the symmetry line. Figure 5.7 presents the model and the meshing.
The load was applied as pressure and the settlement was taken as
the resultant of all the deformation. In the linear analysis the load settlement
behaviour was studied and compared with the 1g model test results. Figure
5.8 presents the deformation contour for the load of 7.8kN, which is the
maximum test load of model piled raft. From the contour it can be seen that
the raft had undergone almost uniform settlement. Maximum settlement of
piled raft is 6.4mm for the load of 7.8kN as against 15.0 mm measured in the
model test. Settlement at the edge of the raft is 5.7mm, which shows
differential settlement of 0.7mm. This difference can be considered as small.
However, the settlement measured in the experiment is uniform, since the
load was applied through a rigid plate.
1
ELEMENTS
MAT NUM AUG 7 2007
23:46:36
Z
X
Y
CIRCULAR RAFT PROFILE WITH 21 PILES-HALF MODEL
Figure 5.7 Finite element mesh for piled raft ( linear analysis)
206
1
NODAL SOLUTION
STEP=1 OCT 7 2007
SUB =1 12:55:39
TIME=1
USUM (AVG)
RSYS=0
DMX =6.414
SMX =6.414
MN
XZ
Y
MX
0 1.425 2.851 4.276 5.701

.712678 2.138 3.563 4.989 6.414
CIRCULAR RAFT PROFILE WITH 21 PILES-HALF MODEL
Figure 5.8 Settlement contour for the load of 7.80kN
The load-settlement curve obtained from the 1g model test and the
linear analysis are compared in Figure 5.9. A study of the curve indicates that
up to a load level of 3kN, which is around 38% of the maximum test load, the
results agree closely. As seen from the characterization curve presented
earlier, this is the load up to which the pile group shares major part of the
applied load. The magnitude of settlement at this load is 2.2mm, which is
close to the critical settlement at which frictional resistance of the pile is fully
mobilized. However the linear analysis grossly under estimated the settlement
of piled raft (stiffer response) for the load more than 3kN, indicating that the
load share ratio between the pile and the raft is independent of the stress level.
This observation is different from whatever observed from 1g model tests of
this study. Model tests showed nonlinear response beyond critical settlement,
hence linear analysis is not applicable to capture the response of piled raft
over entire load range.
207
LOAD, kN
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0

0.0
2.0
4.0 t=8mm
SETTLEMENT, mm
d=10mm
6.0
D=200mm
8.0 l=200mm
Bed = Medium dense
10.0
12.0 EXPERIMENTAL
3D LINEAR
14.0
16.0
Figure 5.9 Comparison of load – settlement response between ANSYS

(linear) and model test results for circular piled raft
5.6.2 Axisymmetric Analysis (MISO)
Hooper (1977) studied the behaviour of piled raft supporting the

Hyde Park Cavalry Barracks tower in London by adopting axisymmetric
analysis. Performing axisymmetric analysis of piled raft requires some
amount of approximation, because in the real foundation the piles are located
at discrete points. The method adopted was to simulate each concentric row of
piles by a continuous annulus with an overall stiffness (ie. displacement per
unit force} equal to the sum of the stiffness of the individual piles.
Adopting the same principle, the circular piled raft was modeled
with appropriate boundary condition in this study. The bed dimension was
taken as that used in the model raft. Analysis was carried for the circular piled
raft of 200mm diameter and 21 piles arranged radially. The length of the piles
was 200mm. For other details and material properties adopted in the analysis,
refer Tables 5.8 and 5.9. Plane 42 element was used and automatic mesh
generation was adopted. The problem was taken as large deformation
208
problem. The elastic modulus Eeq was computed as out lined by Desai (1974)
for plane strain wall and subsequently used by Prakoso and Kulhawy (2001)
for piled raft analysis. The expression used for computing Eeq is as given in
equation 5.7.
In the analysis, the problem was treated as large deformation

problem and a nonlinear analysis was performed. The soil was idealized by
MISO material model. The material properties given for the model as the
input are stress-strain values (Figure 3.6) in addition Young’s modulus
(35000kN/m2) and Poission’s ratio (0.33). As the loading increased the
number of sub steps were also increased so that the iterations will be
sufficient for the convergence of the results. The loading was applied in small
increments as pressure on the surface of raft. Figure 5.10 shows the model
and the meshing.
Figure 5.11 gives the deformation pattern for the load of 9.2kN. It
is seen from the contour that the maximum settlement of the raft is 18.9mm
and is reduced radially. The average settlement of piled raft is 18mm. Further
the settlement of soil is almost negligible at a radial distance close to two
times the diameter of the raft, which shows that the influence zone is well
within the dimensions of the test tank. Similarly the applicable settlement in
soil below the raft is also seen only to a depth of two and half times the
diameter of the raft. This observation confirms that the tank size adopted in
this study has negligible influence on the test results.
209
100mm
200mm
600mm
500mm
Figure 5.10 Axisymmetric model and mesh used in ANSYS analyses
Figure 5.11 Settlement contour for the load of 8.70kN for circular piled
210
The stress contour as presented in Figure 5.12 indicates that the

stress intensity in the raft is almost uniform except at pile locations. Among
the piles, the pile at the centre is taking more load than piles at other
locations. Also the magnitude of the tip stress on piles (ie pile walls) is much
smaller than the head stress indicating that the major part of the load has been
transferred by friction, and the plane strain pile wall has behaved as friction
pile.
Figure 5.12 Stress contour for the load of 8.70kN for the circular piled
raft
The results obtained from the numerical analysis and the 1g model
tests are compared in Figure 5.13. The results agree closely up to the load of
2.3kN; where the major part of the applied load is taken by the pile group.
Beyond the load of 2.3kN, the difference between the results of 1g model test
and the numerical model increases with the load. The maximum difference in
settlement is around 2mm for the maximum test load of 8.7kN. This indicates
that the MISO model under estimated the settlement by 10%. The MISO
material model showed nonlinear behaviour as observed in 1g model test,
however the lesser settlement <10% at higher loads than experimentally
measured values indicates stiffer behaviour of the model. This may be
attributed to idealisation of pile group as a pile wall.
211
LOAD, kN
0 1 2 3 4 5 6 7 8 9 10
0
4
SETTLEMENT, mm
12
16
EXPERIMENTAL
20 NUMERICAL - MISO
24
Figure 5.13 Comparison of load-settlement behaviour between

numerical and 1g model test data (D=200mm, t=8mm,
d=10mm and l=100mm)
The comparison of load-settlement of numerical analysis with the three

phase behaviour of experimental curve (Figure 5.14) resulted the following:
In the phase I the variation between the results of numerical and experimental
analysis is almost negligible indicating good agreement between them. In the
phase II of the load - settlement behaviour wherein the pile-soil stiffness
(ie friction stiffness) decreases with the load rapidly than raft-soil stiffness.
10
LOAD, kN
EXPERIMENTAL
NUMERICAL - MISO
1
0 2 4 6 8 10 12 14 16 18 20 22 24
SETTLEMENT, m m
Figure 5.14 Characterisation curves for experimental and numerical

(Axisymmetric)
212
The numerical model also showed similar response; however the

results of numerical analysis are deviating from the experimental results. The
deviation is slightly higher in the phase III, wherein the piled raft shows
plastic behaviour, and raft – soil stiffness decreases rapidly. For example the
settlement is 8mm for the load of 6kN and is increased to 18mm for the load
of 8.7kN. At these loads, the difference in settlement between the test and
numerical analysis is 8% and 10% respectively. Randolph (1994) as explained
by Ruso and Viggiani (2001) indicated that the accuracy of the available
computer programme is probably not better than + or- 20%.
Katzenbach (2005) has made observations that the implementation

of linear and nonlinear soil modulus largely depends upon the cases under
study as the results vary to an extent of 20% to 30%. However in the present
analysis, the results agreed reasonably well within a variation of 10%. It is to
be noted that the change in slope of the load settlement curve indicates the
change in load sharing process between the piles and raft. The variation in
load for a given settlement becomes higher at settlement around 8mm, at
which the load on the piled raft is 70% of the maximum test load. In the third
phase, the load shared by the raft is higher than the pile group.
In the axisymmetric analysis the settlement of the piled raft was

found to be less than the experimental values. This may be attributed to the
assumption made in the computation of equivalent modulus of pile-soil
system. The computation of equivalent modulus ignoring the presence of soil
prism (as the annulus mainly comprises of the soil) might have caused an
increase in the stiffness of the raft-pile-soil system causing lesser settlement.
However under working load condition the numerical modelling

with axisymmetric idealization is quite adequate in predicting the
performance of circular piled raft. Further, it can be said that the
axisymmetric nonlinear model performs better than linear model.
213
5.7 PLANE STRAIN ANALYSIS FOR RECTANGULAR

PILED RAFT
There are instances where strip rafts (rectangular raft, where length
is greater than width) are used to support a row of columns. But if the
settlement is more, the columns are supported on isolated pile groups. If the
strip raft settlement can be reduced by providing few piles in a row, it can
prove more economical. With this aim 1g model tests were conducted on
rectangular piled raft and the results were presented in chapter 4. Here the
numerical study done on the piled strip raft (rectangular piled raft) is
discussed. The analysis was carried out for medium dense condition. The
details of the piled raft and the bed material are presented in Table 5.10.
Table 5.10 Details of piled raft and bed material
Raft piles Sand bed

Bed
Size Thickness Length Area ratio Es
N density νs
B×L mm mm % KN/m2
kN/m3
70×200 8 14 75 7.85 15.5 35 0.30
The model comprised of 2 rows of piles with 4d spacing on either

directions. The problem was idealised as large deformation problem and the
equivalent pile modulus was computed as outlined in section 5.5. The plane
42 element was used and the problem was defined as plane strain. The value
of Es and the density of soil were as adopted in the axisymmetric analysis of
circular piled raft. The bed dimension was taken same as that used in the
small scale model tests. The loading was applied as a pressure in small
increments and response of piled raft was obtained. Figure 5.15 shows the
finite element model of piled raft.
214
70mm
600mm
600mm
Figure 5.15 Rectangular piled raft model with finite element mesh used
in plane strain analysis of ANSYS
Figure 5.16 Displacement contour for rectangular piled raft for the load
of 1.55 kN in medium sand
215
In Figure 5.16 settlement of the piled raft and bed material are
shown for the load of 1.55kN. The settlement of soil close to the raft is higher
and it has reduced with distance both in the lateral and vertical directions. In
the lateral direction, the settlement is close to zero at a distance beyond two
times the width of the raft. Similarly settlement is close to zero at the depth of
5 times the width of the raft.
Figure 5.17 presents the settlement of the raft. The raft has
undergone a settlement of 12.5mm and 11.15mm at its centre and edge
respectively. At the pile locations the settlements are 11.3mm and 11.4mm.
The difference in the settlement is marginal hence construed as uniform
settlement. The pile tip has undergone a settlement of 8.4mm under the load
of 1.55kN.
Figure 5.17 Settlement of piled raft at the load of 1.55 kN
Figure 5.18 presents the load settlement curves of small scale test
and the numerical analysis. The load-settlement response of plain strain
analysis compares reasonably with the model test curve. However the
settlement obtained for a given intensity of load from the nonlinear analysis is
216
lesser than the settlement measured from the model test. Difference in
settlement between the test and numerical analysis increased with load and
the maximum difference is 1.8mm for the load of 1.55kN, which is 15%
lesser than the experimental value of 12mm. Assumption of equivalent
modulus under estimated the settlement as observed in the case of
axisymmetric analysis of circular piled raft. Thus the idealization of piles in a
row in to plane stain pile wall of an equivalent pile Young’s modulus shows
stiffer behaviour.
LOAD, kN
0.00 0.40 0.80 1.20 1.60 2.00
0.00
2.00
4.00
t=8mm
SETTLEMENT, mm
6.00 d=10mm
l=75mm
8.00 Size =70 x 200mm
Bed = Medium dense
10.00
12.00
NUMERICAL - MISO
14.00 EXPERIMENTAL
16.00
Figure 5.18 Comparison of load-settlement response between ANSYS

and test data for rectangular piled raft
5.7.1 3d Nonlinear Analysis (MISO Model)
The results of 3D linear analysis and nonlinear analysis

(axisymmertric and plane strain conditions) are presented in the section 5.6.
Among the two methods, the nonlinear analysis of axisymmetric and plane
strain conditions predicted the behaviour of piled raft better than linear
analysis, but showed stiffer behaviour with maximum variation of 15% in
settlement. To overcome the limitation of assumption made in the plane strain
analysis and to have better understanding on load sharing between the raft and
pile group of piled raft, three dimensional nonlinear analysis was carried out.
217
The detailed study on full model becomes computationally expensive due to

the large number of elements, and meshing becomes complicated; hence only
quarter model of piled raft was analysed taking advantage of the symmetry.
The various parameters pertaining to the piled raft are given in Table 5.11.
Table 5.11 Properties of square piled raft and bed material
Raft Pile Bed

Spacing
Size Thickness Shape Dia Length Number AR Density
200mm 8mm Circular 10mm 200mm 21 5.2% 4D MD
The three dimensional nonlinear analysis was carried out keeping

the same layout and pile- raft area ratio for the piled raft, for which the linear
and axisymmetric analysis were performed. The bed density was kept as
medium dense with φ=37.50 and unit weight=15.5kN/m3. As said earlier
MISO material model was used for the soil. The continuum was modeled
using solid 45 elements with three degrees of freedom at each node.
In selecting the elements for the various structural components of

piled raft, certain important aspects as given in the manual were considered.
When elements having different degrees of freedom are selected, there will be
inconsistency at the interface. When elements are not consistent to each other,
at the interface, the solution may not transfer appropriate forces or moments
among the different connecting nodes of the various elements at the interface.
To ensure compatibility between the elements used for modelling different
structural elements, they must have the same degrees of freedom. The DOFs
must overlay each other and must be continuous across the element
boundaries at the interface. For example if a solid 45 element is joined either
to shell 63 or beam 4 element, the nodal forces corresponding to displacement
DOFS will be transmitted to the solid element. But the nodal moments
corresponding to the rotational DOFS of the shell element will not be
transmitted to solid 45. Although these conditions may not invalidate the
218
analysis, it is appropriate to select compatible elements for various

components of structure. Solid 45 element was used to model raft, pile and
soil continuum.
In the analysis the bed dimensions were kept same as that of the lab
model tested in the laboratory. The raft and pile were modelled as solid 45
element in order to maintain the element compatibility. Reasonable mesh
refinement was done with an achieved aspect ratio of 5. Required checks were
made for element continuity and continuity at nodes. The material property
was fed in the form of stress-strain data. The mandatory check for proper
meshing at various levels, element continuities etc. were made and then the
solution command was activated to solve the model after applying the load.
The load was applied as pressure in small increments till the load on the raft
equal to the final test load. Figure 5.19 shows the quarter model including
finite element meshing adopted in the analysis.
Length/2 = 500mm
Width/2 = 500mm
Depth = 600mm
Figure 5.19 Finite element mesh of a circular piled raft (Quarter model)
used in ANSYS analysis
219
Figure 5.20 presents the deformation contour of piled raft and

surrounding soil for a total load of 8.1kN on the raft. The maximum
settlement of the raft is 17.7mm for the load of 8.1kN and the maximum
difference between central and edge settlements is 1.5mm only. This indicates
that the settlement of raft is almost uniform. Piles also settled equally and the
magnitude of tip settlement is around 15mm. The settlement in soil around the
piled raft and below the raft decreased with depth and the influence of load in
terms of settlement is seen to a depth of 500mm which is 2.5 times the
diameter of the raft.
Figure 5.20 Settlement contour for a circular piled raft for the load of
8.1 kN
5.7.1.1 Load- settlement behaviour
Figure 5.21 presents the load settlement curves of circular piled raft
obtained from 1g model test and the numerical model. Figure 5.22 presents
comparison of characteristic load-settlement response of circular piled raft
220
between experiment and numerical analyses. The results obtained from the 1g
model test and numerical model agree very closely, till the settlement
level of 4mm.
LOAD, kN
0 1 2 3 4 5 6 7 8 9 10
0
4
SETTLEMENT, mm
t=8mm
8
d=10mm
D=200mm
l=200mm
12
Bed = Medium dense
16 EXPERIMENTAL
NUMERICAL - MISO
20
Figure 5.21 Comparison of load-settlement behaviour between

ANSYS and model test data (Circular Raft)
10
LOAD, kN
EXPERIMENTAL
NUMERICAL - MISO
1
0 2 4 6 8 10 12 14 16 18 20 22 24
SETTLEMENT, m m
Figure 5.22 Characterisation curves for experimental and numerical –

3D Circular
221
As the load increases the difference in settlement between the 1g

model and the numerical model increases marginally. In other words as the
load sharing get enhanced with the raft taking more load, the variation
between the small scale model test results and numerical model increased.
However the maximum variation in the load between the numerical model
and the small scale model results was less than of 5%. The maximum
settlement is 17.7mm as against the measured settlement of 20mm in the 1g
test. This comparison indicates a close agreement between the numerical
model and 1g model test. Similar observation is made in the analyses of piled
raft in loose and dense sand. Thus the nonlinear analysis using MISO model
idealization for the soil predicts the performance of piled raft reasonably well.
5.7.1.2 Load share between raft and piles
The piles are used in the piled raft foundations either for settlement
reduction or for load transfer. Whatever be the purpose of piles, the
knowledge of load share between the raft and the piles of piled raft is essential
for the design of piled raft foundation. The ratio of load carried by piles
depends on many parameters such as number of piles, disposition of piles,
dimensions of raft, relative stiffness between the component of piled raft and
surrounding soil and the magnitude of settlement. In this numerical study, the
loads carried by the piles and raft are analysed for different settlements of
model circular piled raft supported on 21 piles.
Figure 5.23 shows the vertical stress on the typical nodes in the
bottom face of the piled raft for the load of 2.1kN on the piled raft. The
corresponding pressure and the settlement of the piled raft are 0.067N/mm2
and 1.8mm respectively.
In the figure MN and MX indicate the points of maximum and

minimum compressive stress values and these values are -0.8N/mm2 and
-0.003N/mm2 respectively. These values are on negligibly smaller area and
hence not considered. From the figure it can be seen that the stress at the pile
222
locations are several folds higher than the other locations of the raft. The
contact stress is 0.021N/mm2 along the edge of the raft and near the centre
line of the raft, the stress is 0.04N/mm2. But the stress at most of the
locations, and in between the piles are uniform (intensity of colour is almost
uniform in the major part of the raft). Based on the average stress on the raft,
the load shared by the raft is worked out by ignoring the small area of higher
stresses The load shared by the raft is 35% of the applied load of 2.1kN. The
remaining part of the load is taken by the piles, which indicates that the piles
share a major part of the load in the initial part of the loading.
MN
MX
Inverted surface of the raft of piled raft
Figure 5.23 Raft contact stress at typical locations of the raft for the load
of 2.1kN (settlement =1.80mm)
Similarly the contact stresses on the raft is presented in Figure 5.24

for the load of 8.1kN (0.258N/mm2), which is the maximum load for which
the 1g model is tested. The maximum settlement of the raft is 17.7mm. The
stress values at the pile locations are higher than the stress at other locations
of the raft as observed for the load of 2.1kN. The maximum stress is in the
pile locations of outer row. Further the stress in the piles increases with
distance from the raft centre. This shows that the outer piles carry more load
223
than inner piles (both piles in the inner ring and centre pile) in a group.
Similar observation was reported by Horikoshi (1995) through centrifuge test
on piled raft.
MN
MX
Inverted surface of the raft of piled

ft
Figure 5.24 Vertical stress at typical locations of the raft for the load of
8.10 kN (settlement = 17.80mm)
The stress at typical locations of the raft is also shown in the

Figure 5.24. The stress at these points found to vary between 0.162N/mm2 to
0.169N/mm2. Moreover the uniform colour over the entire area of the raft
except at few points indicates that the contact stress is more or less uniform
from the edge to the centre of the raft. The load shared by the raft is worked
based on average contact stress and the percentage load shared by the raft is
64% of final load applied on the raft. This shows that the raft shares major
part of the load at higher settlements where the soil deformation is dominantly
plastic.
Figures 5.25 and 5.26 present the pile head stresses for the raft
loads of 2.10kN,5.5kN and 8.1kN respectively. The stress values indicated
224
are the vertical stresses at the nodes close to the centre of the respective piles.
Stress variation in one of the outer pile head is also shown in Figure 5.25.
Average stress on the pile head is determined to arrive load shared by the
piles. Among the pile head stresses, the stress in the centre pile is lowest and
the maximum stress is in the piles of outer ring. This indicates that the centre
pile carry lowest load and the load carried by the other piles increases with
radial distance from the raft centre.
Sy = -0 13606
Figure 5.25 Pile head stress for the load of 2.1 kN (settlement = 1.8mm)
Figure 5.26 Pile Head stress for the load of 8.1 kN (settlement =17.8mm)
225
Almost similar behaviour is seen for other loads of the raft.

However, increase in pile head stress is not directly proportional to the
respective loads. Though the pile head stress increases with the load, the
percentage share of the load by the pile decreases. This is due to reduction in
relative stiffness of pile-soil due to nonlinear behaviour of soil.
Pile tip stress variation of the pile group of circular piled raft is
shown in Figure 5.27 for the raft load of 2.1kN.
Figure 5.27 Stresses in pile tips for the load of 2.1 kN (settlement
=1.8mm)
The pile tip stresses are almost uniform over the cross section of the
pile than the head stresses. In the same figure pile tip stress values at points
close to the centre of the pile are also shown. The tip stresses are lower than
the respective pile head stresses. The pile tip stress of centre pile is
0.058N/mm2, whereas its pile head stress at the same point is 0.187N/mm2,
which is 31% of the pile head stress. Tip stress values of piles are compared
with their respective pile head stresses and load sharing ratios are calculated.
226
For the piles of inner row the ratio is found to vary between 22% to 24%
whereas for the outer row piles the ratio is in the range of 30 to 33%. From
the results presented above it can be inferred that the pile is sharing major part
of the load by frictional resistance and the load shared by bearing resistance
of piles is less.
In Figure 5.28, tip stresses in piles for the total load of 8.1kN on the
piled raft are presented. The tip stresses are higher than the tip stresses of
piled raft load of 2.1kN. This indicates that the tip stress of pile increases with
load on the raft but not in direct proportion of the load. Similar trend has been
seen in the case of pile head stresses also. The load sharing ratio between tip
and head stresses are obtained and are found to be 11%, 9% to 10% and 17%
to19% for the centre pile, piles in inner ring and piles of outer ring
respectively. These values are lesser than the load sharing ratio corresponds to
the load of 2.1kN. The variation of stress along the pile shaft for the centre,
inner and outer piles located along the centre line of the raft is compared in
Figure 5.29 for the piled raft load of 2,1kN and 8.1kN.
Figure 5.28 Stresses in pile tips for the load of 8.1 kN (settlement
=17.8mm)
227
SHAFT STRESS, N/m m 2

0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60
0.00
0.20
LENGTH RATIO 1 0.40
0.60
0.80 Central pile

Inner pile
1.00 Outer pile
1.20
Figure 5.29(a) Variation of stress along the shaft of typical piles along the
centre line of raft for 2.10kN

0.00 0.40 0.80 1.20 1.60 2.00 2.40 2.80
0.00
0.11
0.22
LENGTH RATIO 1
0.33
0.44
0.55
0.66
0.77 Central pile
0.88 Inner pile
0.99 Outer pile
1.10
Figure 5.29 (b) Variation of stress along the shaft of typical piles along
the centre line of raft for 8.10kN
The difference between the head and tip load (head load=average
pile head stress × area) shows the variation of frictional resistance along the
length of the pile shaft. The frictional resistance of the piles increases with
increase in load irrespective of its location. The increase in frictional
resistance with load even for the settlement more than the critical
settlement(refer chapter 4) is because of enhanced confining stress generated
in the soil between the piles due to load transferred from the raft to the soil
(i.e raft – soil interaction). However frictional resistance is more in the outer
ring piles irrespective of the intensity of the load. It is also observed that piles
228
in the outer ring resist higher load than inner ring and centre piles. This is
attributed to tributary area of the raft for the outer row piles being higher than
inner row and centre piles.
Figure 5.30 presents the load shared by the pile group and the raft
of the piled raft for various loads. In the same figure tip resistance of piles is
also compared. The applied load on the piled raft is shared by the piles and
the raft and the load shared by both the components (raft and piles) of piled
raft increased with the load (settlement), but with higher rate in the case of
raft than piles.
8400
7700 d=8mm
D=200mm
7000 Ar=5.2%
6300
5600
4900
LOAD, N
4200
3500
2800
2100
1400
700
0
1.8mm 7.8mm 17.8mm
SETTLEMENT LEVEL
Total Load Raf t Pile Head Pile Tip
Figure 5.30 (a) Load distribution between raft and pile of the piled raft
at different settlement levels in terms of load
In the initial stages of loading, pile group shares more load than the
raft and its share is around 65% for the load of 2.1kN. Its share is decreased
with the load and minimum share is 36% (Figure 5.30 b) for the load of
8.1kN, which indicates that as the load increased, the proportion of load
transferred to the piles decreased. However, reduction in load sharing ratio of
pile group is insignificant for the load on the piled raft at which deformation
in soil is dominantly plastic. The performance of raft in load sharing is vice
229
versa while comparing with the performance of pile group of the piled raft.
The raft shares more load when the load is increased and particularly for the
load causing yielding of soil and the maximum share is 65% for the circular
piled raft with area ratio of pile of 5.2% in medium dense sand.
1.00
d=8mm
D=200mm
0.80 Ar=5.2%
LOAD RATIO 1
0.60
0.40
0.20
0.00
1.8mm 7.8mm 17.8mm
SETTLEMENT LEVEL
Raft Pile Head Pile Tip
Figure 5.30 (b) Load distribution between raft and pile of the piled raft
at different settlement levels in terms of load ratio
The load sharing ratio of pile group obtained from numerical

analysis is compared with model test in Figure 5.31 for different settlements.
From the figure it can be seen that the load sharing ratio of pile group
compares well between experiment and numerical analyses. The study of the
load sharing behaviour further indicates that the tip resistance of pile
increases with increase in load at a very slow rate. This confirms that both tip
and frictional resistance of pile mobilised simultaneously, but the rate of
mobilisation of frictional resistance with settlement is much higher than tip
resistance. For the problem analysed here the contribution by the tip
resistance of piles is close to 6% irrespective of the load or settlement level.
This response confirms that the piles behaved essentially as friction piles.
230
0.80
0.60
∝ pr
0.40
0.20 EXPERIMENTAL
NUMERICAL
0.00
0.0 4.0 8.0 12.0 16.0 20.0
SETTLEMENT, m m
Figure 5.31 Comparison of settlement vs αpr for experimental and

numerical studies
5.8 ANALYSES OF SQUARE PILED RAFT
In the case of square piled raft linear and nonlinear analyses were
carried out. The geometric features of the piled raft and the bed properties are
provided in Tables 5.12 and 5.13 respectively. Analysis was performed for
two pile spacings of 4d and 6d. The procedure for modelling, element
selections, etc. are the same as circular piled raft, both in the case of linear
and nonlinear analyses.
Table 5.12 Properties of square piled raft
Raft size Raft thickness Pile dia Spacing Length Area ratio
200mm 8m 10m 4d 160m 4.9%
200mm 8m 10m 6d 160m 2.25%
Er and μr are the same as circular piled raft
Table 5.13 Properties of sand used in ANSYS analyses
Material γ’ φ Es State of compaction

Poorly
15.5kN/m3 37.50 35N/mm2 Medium dense
graded sand
231
5.8.1 Linear Analysis
In the case of square piled raft quarter model was used. The load
was applied in the form of pressure as done in the circular piled raft. Figure
5.32 shows the model and the mesh. Figure 5.33 presents the settlement
contour for a load of 6.2kN and the maximum settlement obtained from the
analysis is 3.9mm. The settlement in the edge is 3.45mm; the differential
settlement is 0.45mm. The deformation contour indicates that the raft has
undergone some differential settlement which is close to 10% of the
maximum settlement of the raft. The differential settlement of the raft is
small, hence the raft behaviour is considered as rigid. However, model test
showed uniform settlement of the raft. This is due to the method adopted for
loading the piled raft in the experiment.
1
ELEMENTS
MAT NUM AUG 7 2007
23:57:08
Z
Y X
SQUARE RAFT PROFILE WITH 25 PILES - 4D SPACING
Figure 5.32 Quarter model with finite element mesh for square piled
raft (200mm × 200mm) for linear analysis
232
1
NODAL SOLUTION
STEP=1 OCT 28 2007
SUB =1 19:11:16
TIME=1
USUM (AVG)
RSYS=0
DMX =3.886
SMX =3.886 Z X
Y
MX
0 .863507 1.727 2.591 3.454

.431753 1.295 2.159 3.022 3.886
SQUARE RAFT PROFILE WITH 25 PILES - 4D SPACING
Figure 5.33 Settlement contour for the load of 6.2kN for square raft with
4d pile spacing
The load – settlement curves of 1g model test of square piled raft

and ANSYS linear analysis are compared in Figure 5.34 for the medium
dense condition of sand. The load-settlement curve of linear analysis
compares reasonably with experiment up to the load of 2.5kN (ie initial part
of load curve). The settlement of piled-raft corresponding to this load is
around 1.6mm. This settlement is close to critical settlement observed in the
experiment. For the load more than 2.5kN, the difference in settlement
between experiment and linear analysis increases with the load. Thus, the
linear analysis underestimated the settlement of piled raft on increase in load
as observed in the case of circular piled raft.
The linear analysis in the case of square raft is grossly under

estimated the settlement indicating that the load share between piles and raft
is independent of the intensity of applied load. The settlement of square raft
for a given load is lesser than circular raft having diameter equal to the width
of the square raft. This is due to raft area, more number of piles and uniform
233
arrangement of piles (square pattern) over the area of square raft. Hence the
stiffness of square piled raft is more than the circular piled raft. However, for
the settlement higher than the critical settlement., 1g model studies showed a
nonlinear behaviour, and the linear analysis is not capturing this behaviour.
LOAD, kN
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0
0.0
2.0
SETTLEMENT, mm
t=8mm
4.0
d=10mm
B=200mm
6.0 l=200mm
Bed = Medium dense
8.0
EXPERIMENTAL
10.0
3D LINEAR
12.0
Figure 5.34 Comparison of load- settlement response between ANSYS

(linear) and on square piled raft in medium sand
5.8.2 Nonlinear Analysis
The results of the linear analysis presented in the earlier section

have clearly established its inadequacy in predicting the behaviour of piled
raft beyond the critical settlement of pile. In as much as in the case of circular
piled raft the load sharing behaviour has to be studied in the case of square
piled raft also. Hence, nonlinear analysis is performed on the square piled raft
using MISO soil model.
The nonlinear analysis was performed for 4d and 6d pile spacings

and their area ratios are 4.9% and 2.25% respectively. Figure 5.35 presents
the typical model and the meshing that were adopted.
234
Figure 5.35 Quarter model and finite element mesh adopted for
square piled raft in ANSYS nonlinear analysis
Figure 5.36 presents the deformation contour for the maximum load
of 8.7kN of piled raft with 25 piles.
Figure 5.36 Settlement contour for the load of 8.70kN (settlement

18.90mm)
235
The maximum settlement measured in the analytical model is

18.3mm and the differential settlement is 2.5mm for the applied load of
8.7kN. The differential settlement is 13% of the maximum settlement, which
can be ignored and said that the raft settlement is uniform. Further all the piles
settled almost equally and the tip settlement is closed to 14mm. The
settlement in the soil around the piled raft reduced uniformly with depth and
lateral distance. At a depth around 500mm (2.5 times the width) the
settlement of the soil is around 2mm. Similar is the case in the lateral
direction also. Thus the dimensions of test bed adopted in this study have the
least influence on the performance of model piled raft.
5.8.2.1 Load- settlement behaviour
Figures 5.37 and 5.38 present the load settlement curves of square
raft with piles spaced at 4d and 6d as obtained from 1g model tests and
numerical analyses. The three phase characteristic response of load-settlement
is also compared in Figure 5.39 for the square piled raft with piles at 4d
spacing.
LOAD, kN
0 1 2 3 4 5 6 7 8 9 10
0
4 t=8mm
SETTLEMENT, mm
d=10mm
B=200mm
8
l=160mm
S=4d
NUMERICAL - MISO
16
EXPERIMENTAL
20
Figure 5.37 Comparison of load- settlement response between ANSYS

and test data for square piled raft with 4d pile spacing
236
In the case of 4d spacing the results agree very well up to the

settlement of piled raft of 8mm, which is 40% of the final settlement achieved
in 1g model test. The corresponding load is 5.5kN which is 60% of the load
corresponding to the maximum settlement of 20mm. Beyond this load a
marginal difference between the 1g model and numerical analysis is seen and
at the final load the variation is less than 5%. The maximum settlement
reached in the analytical study was 18.3mm at a load level of 8.7kN which is
marginally lesser than the experimental value indicating negligibly stiffer
behaviour of the numerical model.
LOAD, kN
0 1 2 3 4 5 6 7 8 9 10
0
4
t=8mm
SETTLEMENT, mm
8 d=10mm
B=200mm
l=160mm
12
S=6d
Bed = Medium dense
16
EXPERIMENTAL
20
NUMERICAL - MISO
24
Figure 5.38 Comparison of load -settlement response between ANSYS

and test data for square piled raft with 6d pile spacing
10
LOAD, kN
EXPERIMENTAL
NUMERICAL - MISO
1
0 2 4 6 8 10 12 14 16 18 20 22 24
SETTLEMENT, m m
Figure 5.39 Characterisation curves for experimental and numerical

Square piled raft, 4d spacing
237
In the case of piled raft with piles at 6d spacing, the results of

numerical analysis match very closely with experiment till the settlement
reaches a magnitude of 16mm. The load corresponds to this settlement is 6kN,
which is 78% of the final test load. Beyond this load, the difference between
the results of 1g-model test and numerical model increases gradually and at
the final load, the variation in settlement is around 0.8mm, which is 4%
higher than the experimental value.
In the case of square piled raft, the numerical analysis with MISO
model idealisation for soil predicts the behaviour very closely with 1g model
test results. The comparison is better in square piled raft than circular piled
raft with radial arrangement of piles. This may be attributed to the tributary
area of the raft related to the pile is uniform in the case of square piled raft.
From the results presented it may be said that 3D nonlinear analysis with
MISO model idealisation of the soil predicts the performance of piled raft
reasonably well.
5.8.2.2 Load share between raft and piles
Figure 5.40 presents the stresses on the raft at a load level of 2.8kN
(the applied stress 0.07 N/mm2). At the pile locations the stresses are in the
range of 0.32N/mm2 to 0.47N/mm2. But in general on the surface of the raft
the stress varies dominantly from 0.021 to 0.025N/mm2. The colour coding
for stress obtained from the analysis confirms the uniform distribution of the
stress on the entire raft surface except at the locations of piles. For the load of
2.8kN, the raft shares around 33% of the applied load and the balance 67% is
shared by the piles. Hence it can be said that at the earlier part of the loading,
the piles share more load. Figure 5.41 presents the raft stress value for the
load of 8.7kN at which settlement of piled raft was 18.3mm. At the locations
of pile the magnitude of stresses varies from 1.67N/mm2 to 1.97N/mm2.
A variation of 6% in stress between the centre and edge is seen. The stress at
the edge is higher but its influence is seen over a smaller area of the raft. The
average stress at the base of raft is 0.134N/mm2 as against applied pressure of
238
0.217N/mm2. Hence the raft shares a load of 62%. Accordingly the load taken
by the pile group is 38%. This indicates that the raft shares more load on
increasing the load.
Figure 5.40 Vertical stress in the square piled raft for the load of 2.8kN
and 4d pile spacing
Figure 5.41 Vertical stress in the square piled raft with piles at 4d
spacing for the load of 8.7kN
239
Figure 5.42 Pile head stress for the load of 2.80kN (4d pile spacing)
Figure 5.43 Pile head stress for the load of 8.70kN (4d pile spacing)
240
In the Figures 5.42 and 5.43, the head stresses on piles for the load
of 2.8kN and 8.7kN respectively are presented. The head stress on the pile is
not uniform as observed in the case of circular piled raft. However average
stress on the pile head is determined by eliminating sharp point of high
stresses. The average stresses are shown on the pile head for piles at different
locations and the values found to vary between 0.80N/mm2 and 0.96N/mm2
for the load of 2.8kN on the piled raft. The maximum difference in head stress
between the piles is around 20%.
It can be seen from the Figure 5.43 that piles in the inner row
(along the centre line of the raft) carry slightly higher load than piles in the
outer row. Average head stress in the centre pile is 2.2N/mm2 where as in the
corner pile the value is 1.78N/mm2, which is 80% of the stress of centre pile.
This variation indicates that the stress on the piles decreases with distance
from the centre of the raft as obtained in circular piled raft. Further, the stress
on piles at symmetric locations is almost equal. The head stress in centre pile
is more than piles of other locations. This may be attributed to higher
confinement.
Pile tip stresses obtained from the analysis for the piled raft load of
2.8kN and 8.7kN are presented in Figures 5.44 and 5.45 respectively.
Distribution of tip stress in each pile is more uniform than head stress
irrespective of the intensity of load on raft and location of pile. In the figures
average tip stresses are presented. As observed in the case of head stresses,
the tip stress is higher for centre pile than other piles for the loads of 2.8kN
and 8.7kN. Among the pile tip stresses, the stress in the corner pile is the
lowest and is close to 80% of the tip stress of centre pile. The tip resistance of
all the piles are compared with the pile head resistance and is close to 15%
and 30% of respective pile head resistances for the load of 2.8kN and 8.7kN
respectively. This indicates tip resistance of the pile is increased with increase
in load. Further, the tip resistance of 15% to 30% of head resistance shows
that the pile group transfers major part of the load acting on it by friction.
241
However, the tip resistance of the pile group is between 10% and 14% of the
load on the piled raft in medium dense sand
Figure 5.44 Pile tip stress for the load of 2.80kN (4d pile spacing)
Figure 5.45 Pile tip stress for the load of 8.7kN (4d pile spacing)
242
In Figures 5.46 and 5.47, the stress distribution on the raft contact
surface with the soil is presented for the piled raft supported on 9 piles (ie pile
spacing of 6d) in medium sand for the load of 1.8kN and 7.7kN respectively.
For these loads the settlements were 1.7mm and 19.2mm respectively.
Stresses on the raft is found to vary uniformly from the edge as seen in the
case of piled raft with 21 piles.
Stresses along the edges are more than other part of the raft.
However, stress concentration is more at pile locations, where the stresses are
many times higher than raft stresses. The raft stress at ten points are shown in
the figure. Based on the average stress on the raft, the load shared by the raft
is worked out, which is 56% and 61% of the applied load of 1.8kN and 7.7kN
respectively. For both the loads the load shared by the pile group is lesser than
the raft.
Figure 5.46 Vertical stress in the piled raft with piles at 6d spacing for
the load of 1.8kN
243
Figure 5.47 Vertical stress in the piled raft with piles at 6d spacing for
the load of 7.7kN
The pile head stress and tip stress distribution for the 9 pile group
of square pile raft are presented in Figures 5.48 and 5.49 respectively for the
load of 7.7kN.
Figure 5.48 Pile head stress for the load of 7.70kN – 6d pile spacing
244
The head stress is not uniform on each pile as observed in the case
of 25 pile group of piled raft. But tip stress is uniform as in the other case.
The intensity of head and tip stresses are higher than 25 piles even for a load
of 7.7kN, which are lesser than the maximum load of 8.8kN of piled raft with
21 piles. However load shared by the 9 pile group is lesser than 25 piles.
Further percentage share is around 40% which is close to the value of 25 pile
group. Average stresses of head and tip are compared and found that the
stresses are close to 20% of head stresses for the load of 7.7kN.
Figure 5.49 Pile tip stress for the load of 7.7 kN (6d pile spacing)
Variation of vertical stress in pile along its length is presented in

Figure 5.50 for the piles of centre row of piled raft. The stresses presented are
for the pile spacing of 4d (No. of piles25) and the load of 8.7kN.
As stated earlier the product of area of pile and difference in head

and tip stresses is the frictional resistance of the pile. Frictional resistance of
the pile is higher than the base resistance (bearing resistance). The stress
variation along the pile length is gradual up to the length ratio of 0.7 from the
pile head and thereafter reduction in the stress is increased but at constant rate
over the remaining length of the pile. The transition in rate of change of stress
245
occurred at the pile depth around 12 times the pile diameter. Surprisingly this
value is close to the critical depth of pile in sand as reported by Vesic (1967)
and others. Piles at other locations also showed trend as reported.

0.00 0.40 0.80 1.20 1.60 2.00 2.40 2.80
0.00
0.20
LENGTH RATIO 3
0.40
0.60
0.80 Central pile

Inner pile
1.00
Outer pile
1.20
Figure 5.50 Variation of stress over the length of pile of square piled raft
for the load of 8.7kN (No. of piles 25 at 4d spacing)
Figure 5.51 presents the stress variation along the length of pile for
the piled raft with 6d pile spacing for the piled raft load of 1.8kN and 7.7kN.
The stress on pile reduced with depth, but the rate of reduction is higher at the
lower part of the pile than support part. This trend is seen in both the piles of
centre row irrespective of intensity of the load on the raft. As observed in the
case of pile group with 4d spacing, the change in the rate of reduction of
stress is seen at the depth ratio close to 0.7, which is equal to 11times the pile
diameter.
0.00 0.20 0.40 0.60 0.80
0.00
0.11
0.22
LENGTH RATIO 1
0.33
0.44
0.55
0.66
0.77
0.88 Centre pile
0.99 Outer pile
1.10
Figure 5.51 (a) Stress distribution along the pile of square piled raft with
6d pile spacing for 1.8kN
246

0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00
0.00
0.11
LENGTH RATIO 1
0.22
0.33
0.44
0.55
0.66
0.77
0.88 Centre pile
0.99 Outer pile
1.10
Figure 5.51 (b) Stress distribution along the pile of square piled raft
with 6d pile spacing for 7.7kN
The load carried by the pile group of 25 piles is compared with the
raft load for different settlements of the piled raft in Figure 5.52. These values
are compared in terms of absolute load and load ratio in Figures 5.52(a) and
5.52(b) respectively. Load shared by the raft increased with settlement
(intensity of load on the piled raft) of the piled raft.
9000
t=8mm
8000 d=10mm
B=200mm
7000 Ar= 4.9%
6000
LOAD, N
5000
4000
3000
2000
1000
0
1.8mm 7.8mm 18.2mm
SETTLEMENT LEVEL
Total Load Raft Pile Head Pile Tip
Figure 5.52(a) Load share between raft and piles of square piled raft
(25 piles at 4d spacing) in terms of load
247
1.00
t=8mm
d=10mm
B=200mm
0.80 Ar= 4.9%
LOAD RATIO 1
0.60
0.40
0.20
0.00
1.8mm 7.8mm 18.2mm
SETTLEMENT LEVEL
Figure 5.52 (b) Load share between raft and piles of square piled raft
(25 piles at 4d spacing) in terms of load ratio
Piles also showed the trend same as that of the raft but at different
rate. Rate of increase of load on the piles is lesser than the rate which raft
shared the load. Further the load sharing ratio of pile group is decreased with
the load on the piled raft whereas reverse trend is seen in the case of load
share of the raft. The load sharing ratio of the raft is increased with the load
but at decreasing rate. However load sharing ratio of raft is higher than pile
group for higher loads on the piled raft.
The load sharing ratio of pile group of 25 piles obtained from the
nonlinear analyses for different settlements are compared with the values
obtained from 1g model test in Figure 5.51 for piled raft with 25 piles. The
values obtained from the numerical analyses compare reasonably well with
experimentally predicted values except for the settlements close to the critical
settlement of the piled raft. Numerical analyses however predicted higher load
sharing ratios than experiment.
248
0.80
0.60
∝ pr
0.40
0.20
EXPERIMENTAL
NUMERICAL
0.00
0.0 4.0 8.0 12.0 16.0 20.0
SETTLEMENT LEVEL, m m
Figure 5.53 Comparison of settlement vs αpr for experimental and

numerical studies
Similarly the load distribution between raft and piles of piled raft
with 9 piles (pile spacing=6d) is compared in Figure 5.54. Piled raft with
9 pile group showed almost similar response as observed in the case of piled
raft with 25 piles except that the load sharing ratio of raft is higher than the
pile group even for the loads lesser than 20% of the maximum test load. This
is due to lesser number of piles and higher contact area of the raft than the
piled raft with 25 piles.
9000
8000 t=8mm d=10mm

7000 B=200mm
A r= 1.80%
6000
LOAD, N
5000
4000
3000
2000
1000
0
1.80mm 17.20mm
SETTLEMENT LEVEL
Tot al Load Raf t Pile Head Pile Tip
Figure 5.54 (a) Load share between raft and piles of square piled raft
with 9 piles in medium dense sand in terms of load
249
1.00
t=8mm d=10mm
0.80 B=200mm
A r= 1.80%
LOAD RATIO 1
0.60
0.40
0.20
0.00
1.80mm 17.20mm
SETTLEMENT LEVEL
Figure 5.54 (b) Load share between raft and piles of square piled raft
with 9 piles in medium dense sand in terms of load ratio
The observations made on load share between the raft and pile is
attributed to reduction in stiffness of pile-soil and raft-soil with increase in
deformation of soil. Further pile-soil stiffness is reduced at a rate higher than
raft-soil stiffness and is increased further on reduction in number of piles.
5.9 SUMMARY
The behaviour of piled rafts of circular, square and rectangular raft

with piles of different spacings were analysis numerically and compared with
the results of 1g model tests. Both linear and nonlinear analysis were
performed. In the nonlinear analysis, soil continuum was modelled using
MISO material off ANSYS. The profile of the load-settlement relationship
was better calculated in linear analysis particularly in the first stage of the
loading. As the total load increased, the discrepancy between the measured
and calculated results increased. Since no account was taken of the non-
linearity of the soil behaviour in the analysis. Nonlinear analysis with MISO
model predicted the load-settlement relationship well for entire range of load.
The maximum discrepancy between the measured and calculated loads was
less than 10%.
250
Nonlinear analysis predicted some differential settlement (<15%) in

the rafts, whereas in the experiments differential settlements were zero
because models were loaded using rigid plate. The load shared by the pile
group was determined from the numerical model based on average stress on
each pile of the pile group and compared with values obtained experimentally.
The calculated total load which was transferred to the piles agreed reasonably
well with the experimental values. According to the ANSYS results, the piled
raft system behaved elastic in the initial stages of load and on further loading,
the pile-soil and the raft-soil interactions became plastic gradually. Thus the
load sharing ratio between the pile group and the raft was not constant and
load shared by the raft was higher than the pile group particularly for the piled
raft loads (total loads) that caused predominantly plastic deformations in the
soil.
251
7CHAPTER 6
BEHAVIOUR OF PROTOTYPE PILED RAFT
6.1 GENERAL
Although small scale model studies and numerical modelling have

provided considerable amount of data on the performance of piled raft, they
are subjected to certain limitations and assumptions. Therefore, German code
advocates monitoring the behaviour of instrumented real size piled raft
foundation. Further the development on the design philosophy of piled raft as
brought out in the difference is based more on the observational methods.
Here the performance of instrumented prototype piled raft was monitored
during the construction and post construction period and the results are back
analyzed to improve upon the design philosophy and efficiency of piled raft.
An ideal design of piled raft would be the one wherein the number of piles
would be minimum with the raft sharing more loads under reduced settlement.
Even though in few cases as reported by Katzenbach (2000) this has been
achieved, they are mostly with raft thickness of 1.5m to 3m founded on deep
deposits of over consolidated clay.
If the piled raft has to be used as foundation system with more

freedom and confidence, its suitability has to be established for medium sized
structure also with relatively smaller diameter piles and thinner raft. Its
effectiveness in overall settlement reduction must be adequately studied and
validated. An settlement oriented design will be more effective it is based on
observational data, duly validated by a suitable analytical test. For example
the results of observations made on the piled raft of Eurotheum tower was
used to design the piled raft of Max Tower (Katzenbach et al 2002). This
252
implies that adequate data on the performance of real size foundation must be
available along with analytical validation so that the data can be used as a
base for the design. It is felt very much essential that such data bank must be
available for every regional deposit. The study on prototype piled raft has
been carried out with the above as one of the main objective of this research.
The structure monitored in this study is a twelve-storied

commercial cum residential complex which is supported on piled raft
foundation. The raft is reasonably thin with the piles of smaller diameter. The
piles pass through soil layers that are predominantly sandy and rest on
medium dense to dense sand. The observations made in the field study are
analysed to understand settlement behaviour and load sharing between the raft
and piles. The results thus obtained are compared with the results of
numerical analysis on prototype piled raft.
6.2 PALACE REGENCY BUILDING, CHENNAI
The building adopted for this study is located at Purasawakkam in

Chennai, India. The structure is a twelve storied reinforced concrete
construction with a basement floor. Plate 6.1 shows the elevation of the
structure, as completed. The weight of this concrete building on the raft is
99000kN. The ground area of the building measures 32m x 25m. The height
of the building is 36m. The slenderness ratio of the building if the lower
dimension alone is considered works out to 3; but if overall dimension is
considered it is 1.5.
The basement, ground floor and the first floor are for commercial
use and the other upper floors are for residential requirement. The maximum
and minimum loads on columns are 2870kN and 1050kN respectively at the
raft level. The structural design has been done in accordance with the
provisions of the latest Indian Standard code of practice. The minimum grade
of concrete is M20 and all the reinforcements were Fe 415 grade bars. The
entire building was completed in 460 days and was fully occupied shortly
253
thereafter. The structure was analysed assuming that it is resting on an

unyielding support using STADD PRO, a versatile package for structural
analysis and design. The support reactions from the analysis have been taken
for the design of foundation.
6.2.1 Soil Profiles of The Building Site
Palace Regency building site is located in the busy commercial

area of Purasawakkam covering over an area of roughly 4000m2. This area is
populated with multi-storeyed structures and most of them were supported on
pile foundations. Therefore, it is decided to explore this site through three
deep boreholes in order to obtain the knowledge on soil condition. Soil
investigation work was carried out as per BIS specifications under the
supervision of the author. In order to obtain relative strength of the strata, SPT
was conducted at regular intervals.
Plate 6.1 Elevation of Palace Regency Building, Chennai

254
Figure 6.1 presents the typical soil profile of the site. The soil
profile at the site comprised of medium stiff sandy clay layer of CI type for
the top 7m depth. This layer recorded low N 2 to 4 value particularly at depth
between 4m and 7m. From 7m to 14m the layer is clayey sand SC/SM type
with sand content more than 65%, and the state of compaction is medium
dense. Beyond 14m the percentage of sand content increases with depth and
so the state of compaction. This layer extends up to 24m. The stratum beyond
24m is disintegrated rock. The bore hole was terminated in this layer, after
ascertaining that the disintegrated rock extends for a larger depth. Figure 6.2
presents the sectional elevation of the structure with the soil profile and the
essential soil parameters.
DEPTH IN METRES
Figure 6.1 Typical soil profile at Palace Regency site, Chennai

255
39.55m
BASEMENT GROUND LEVEL

RAFT
-3.00
-4.00
N=~5 γ = 1.6 t/cum MI Sandy Clayey Silt
-7.00 20% 40% 34%
Clayey Silty Sand

γ = 1.8 t/cum
SC 12% 24% 66%
φ = 27°
-14.00
600mm dia pile (Typ)
-17.00
N=38 γ = 1.9 t/cum Clayey Sand
φ = 34° (Percentage of sand increases
with depth)
-24.00
Very Soft Disintegrated Rock
N=61
Figure 6.2 Sectional Elevation with Geotechnical Data

256
6.2.2 Selection of Design Parameters and Design of Foundation

System
Since the structure has a basement it was proposed initially to

support it on raft. Computations showed that the raft will undergo a total
settlement of 300mm including long term settlement which is more than the
permissible value as per the code IS1904-1986. Further the structure has a
central courtyard and at the column loads are less in this area. This has
resulted in higher differential settlement. Hence keeping the variations of
column loads and the magnitude of differential settlement in mind, it was
decided to support the structure on deep piles. Further few structures in the
neighbouring area within 100m carrying load lesser than this structure are
supported on piles. However at this stage the option of piled raft was thought
of keeping the basement and the thickness of the basement raft required in
mind
To the best of our knowledge no clear cut procedure was available

relating to the load sharing behaviour of piled raft, of the type with thinner
raft and smaller dia piles. Moreover the piles pass through soil layers that are
predominantly sand. However, the principles outlined by Burland et al (1977),
Padfield and Sharock (1983), and Franke (1993), were taken as the basic
guidelines for the design. According to Burland (1977) the piles have to be
ductile. Franke (1993) had indicated that the base resistance of the pile has to
be much smaller than the shaft friction and the load shared by the raft has to
be considerable. Padfield and Sharock (1983) suggested that shaft friction has
to be fully mobilized for settlement reducing piles. Accordingly the main
principles considered in the design are
(i) the applied load would be shared by the raft and the piles
equally.
(ii) the piles have to be dominantly floating and
(iii) the settlement level must be such that the pile must mobilize
the friction entirely.
257
While the first two aspects could be taken care of, the movement
required for the mobilization of the entire friction could not be assumed.
Hence it was considered that the piled raft settlement shall be more than
12mm for the complete mobilization of frictional resistance, which is the
limiting settlement for individual piles as per IS 2911 (Part IV) 1985.
As large diameter bored piles were becoming uneconomical for the

intensities of column load, smaller diameter piles were used. The layout of
piles was made in such a way that it follows a known pattern. Among the
various patterns, it was decided to locate the piles below the column as
adopted by Yamshita et al (1994). It was also decided to adopt the load
sharing data available under identical circumstances. To the best of our
knowledge no published data was available to establish the load settlement or
load sharing behaviour of piled raft, particularly for relatively medium sized
building with flexible raft.
Yamashita et al (1994) established the load sharing behaviour by

adopting thinner raft with smaller dia piles and showed that the load sharing
between raft and the pile was in the ratio of 49% and 51% respectively for the
raft with piles below the column. Based on this, the load shared by the piles
and the raft was assumed as 50% of the total load in the present design. The
piles were designed to have a factor of safety of 1.75, keeping in mind the
likely variations of the loading in the commercial floors. The raft was
designed as a flat slab treating the piles as column; 600mm was the raft
thickness required from the bending moment and shear considerations for the
above assumption. The layout of piles is given in the Figure 6.3.
The most important parameter in the design of piled raft is the Es

value of the supporting strata. When the stratum becomes predominantly
cohesionless the most reliable method to assess the soil parameters at its in-
situ conditions is the N value. While a lot of empirical expressions are
available to relate N and Es, the charts published by Mori (1965) for clay and
sand were used in this study to obtain the Es values of various layers of the
deposits of this area (Appendix 2).
Figure 6.3 Lay out of piles and settlement markers
258
259
The other important factor is the load sharing ratio between the
piles and the raft which was taken as 50% as stated earlier. In designing the
pile group the factor of safety against block failure was computed as given by
Poulos (2001)
Pw + Np iη
F = . (6.1)
P
where
F = Factor of safety against block failure
Pw = bearing capacity of raft
N = number of piles
pi = individual pile capacity
ή = group efficiency
P = total structural load
Accordingly the number of piles required were computed as 86.

However the column layout and optimization of column load for pile
grouping necessitated the provision of 93 piles as shown in the layout Figure
6.3. The piles provided were of 600 mm diameter under each column. Under
many columns two piles were installed. The piles were installed adopting the
rotary drilling equipment and were terminated in sand layer at a depth of 17m
below the existing ground level. It was ensured that the bottom of the bore
was clean so that the seating of the pile would be done on the natural soil. The
piles were terminated in sand layer where the observed N-value was around
40. The construction of pile foundation and raft elements are presented in
Plate 6.2.
260
Plate 6.2 Construction of pile and raft in progress at Palace Regency

site, Chennai
6.2.3 Instrumentation of the Piled Raft and Measurements
Since the primary aim was to study the load settlement behaviour
of the piled raft, importance was given to obtain the settlement values at
various locations at every stages of loading. Hence settlement markers were
placed at 15 points as shown in Figure 6.3. The settlement markers comprised
of 75mm × 75mm × 6mm plate two numbers separated by a distance of
600mm and was made to form an open box by welding the plates with 4 bars
of 12mm diameter. This box was welded to the bottom layer of the raft
reinforcement. The verticality of the marker and the level of the top surface
were checked using mercury levels and plumb bob. The selection of the
location for the settlement markers was done in such a way that the settlement
profile of the raft can be plotted in both the directions at various sections. In
order to measure the settlement a standard bench mark was established such
261
that it can be viewed from any point and will not undergo any movement.
Plate 6.3 presents a typical settlement marker in position.
Plate 6.3 Typical settlement marker in position
Plates 6.4 and 6.5 illustrate the field settlement measurements

using precision level. The initial reading was taken as soon as the raft was
cast. Subsequent readings were taken each time the slab was cast (i.e.
immediately after the deshuttering of the slab was done). The settlement
recorded by all the markers was monitored for the entire construction period
of 360 days. The top most slab namely the Lift Machine Room was cast after
402 days pointing the completion of the structure, and by the time 50% of the
brick work and partition along with the flooring was completed. Settlement of
the raft was monitored continuously and reading were taken at an interval of
every three months spread over a period of one year. During this period the
rate of increase of the settlement was very small and gradual. Thereafter the
balance work relating to other interior and occupational loadings were
completed.
The observations were continued and completed on 796th day from

the date of commencement of construction. The period from 360th day and
796th day has been termed here as post construction period. The increase in
the settlement was recorded at various locations regularly as far as possible,
even though the construction procedure and interior work caused a lot of
disturbances in the observations of few gauges. The settlement readings were
262
recorded with a high precision levelling instrument. The settlements recorded

at various locations during the entire construction and post construction
period are presented and discussed in the investigation.
Plate 6.4 Settlement measurement at one of the settlement markers
Plate 6.5 Settlement measurement using precision level

263
6.2.4 Load-Settlement and Load Sharing Behaviour
The readings recorded at the locations of settlement gauges are

reduced corresponding to the level of permanent bench mark and the
difference in level between the initial and subsequent readings are reported as
settlement. Fig. 6.4 indicates the sequence of loading with time and observed
settlement pattern in all the gauges representing the four tower and the central
courtyard areas. The settlement readings are also presented in Tables 6.1 a
and 6.1 b for three different sections. Section 1 is along the grid p, which is
designed as top row. The gauges installed in this row are designed as a, f, g, m
and n. Section 2 is middle row, which is close to grid ‘G ‘. Gauges installed
along this section are 6 in numbers. They are e, b, j, g, l and o. Along the
section 3 (bottom row) number of gauges are 4 (c, d, k and p).
As stated earlier the construction was completed in 360 days, and

whatever settlements recorded within this period was due to self-weight of the
structure inclusive of construction load. The settlement recorded after this
period was post construction settlement which includes the effect of all the
loads existed at that point of time. However the load variation during the post
construction period is considered small.
Settlement of the raft at all the three sections followed almost

identical trend with loading sequence. Among the sections, the settlement
along the middle row is shifting either than other two rows. Further variations
in settlement between the gauges of top and bottom rows are lesser than the
gauges of middle row.
264
DAYS Vs PERCENTAGE SUPERSTRUCTURE LOAD

120
100
80
LOAD, %
60
40
20
0
0 100 200 300 400 500 600 700 800 900
DAYS
0
4
SETTLEMENT, mm
12 GAUGE-A
GAUGE-F
16 GAUGE-G
GAUGE-M
20 GAUGE-N TOP ROW
24
0
4
SETTLEMENT, mm
12 GAUGE-B
GAUGE-E
16 GAUGE-J
GAUGE-H
20 GAUGE-L
GAUGE-O CENTRAL ROW
24
0
4
SETTLEMENT, mm
12 GAUGE-C
GAUGE-D
16
GAUGE-K
GAUGE-P
20 BOTTOM ROW
24
Figure 6.4 Time dependent load settlement curves

265
i
i
4
266
During the initial sixty days practically there was no settlement or

the magnitude was not measurable. This may be due to the fact that the
intensity of loading was inadequate to cause any measurable movement. But
when the third floor slab was cast and deshuttered, the settlement at various
locations as shown by the gauges varied from 1mm to 3mm. Measurable
variation in settlement occurred after the completion of the sixth floor. The
corresponding load on the foundation system was around 45% of the total
load. Thereafter the settlement rate was more with load and at the time of
completion of structure the maximum settlement in the tower area was 11mm.
Typically the gauges b, e, c & d representing one of the tower areas

showed a settlement of 11mm, 10mm, 9mm and 10mm respectively,
indicating almost uniform settlement. In the central courtyard represented by
the gauges “h, j and g” the recorded settlements were 6mm,8mm, and 8mm
respectively. However a gradual increase in settlement was observed and
recorded after the construction period of 360 days at all the locations. The
maximum and minimum settlement recorded is 11mm and 6mm at the end of
construction, which are at locations b and h (central row) respectively. The
gauge b represents the centre of the tower area whereas the gauge h represents
the court yard area which is the least loaded area of the structure. Moreover
piles provided in this area are identical to the piles provided at other areas of
the raft, which apparently has reduced the settlement. Study on settlement
profile showed that the settlement at centre part of the raft in each of the
tower area is higher than at the edges.
The higher settlement at the middle portion is attributed to heavier

column loads and the flexible nature of the raft. Measurement of settlement
was continued after the completion of building. The maximum increase in
settlement during the post construction period of 436 days was 3mm. The rate
of increase in the settlement during post construction period is very small and
gradual.
267
Further it can be seen from the Figure 6.4 that the rate of
settlement increased during the construction period between 200 to 360 days,
where the percentage of loading increased steadily from 45% to 95%. In other
words we can infer that only after 50% of the structural load acted on the
piled raft, the total friction got mobilized causing a larger movement of the
system. This increase in the settlement indicates that the load shed by the piles
is taken by the raft. It is at this stage the piles take up the role as settlement
reducer.
Figure 6.5 presents the load taken by the raft at various stages of
construction. Even though no instrumentation could be done for measuring
the load, the load taken by the raft could be computed from the settlement
observed during various stages of construction. The load shared by the raft
was back calculated from the settlements measured using elastic equations
(Hemsley, 2000). The tributary area of the particular column of influence of
the load was computed and the stresses were calculated using observed
settlement adopting elastic theory. This construction was carried out for
various column sections. Since the structure is symmetry and the settlement is
uniform, calculation on various sections showed almost identical trend in load
sharing. For an average settlement of 3mm the load taken by the raft is of the
order of 7.6%. This indicates that in the initial stages the piles take the applied
load.
50
Percent load taken by the
40
30
raft
20
10
0
0 200 400 600 800 1000
Tim e in days
Figure 6.5 Percentage Load Taken by the Raft at Various Stages of

Construction period
268
The steep increase occurred in the load taken by the raft occurred
when the applied load was 60% of the total load. At 50% and 80% of the total
load the raft was found to share 17.5% and 33% of the total load. At the final
loading the raft takes 43% of the applied load.
During the initial stages of loading, most part of the load has been
taken by the piles, but as loading increased the raft started sharing the load.
This is clearly seen from the Figure 6.5. Sharing of load by the raft increases
with increase in settlement of the piled raft system. When the load is 80%, the
raft shared around 33% of the total load. However at the final load, the load
taken by the raft was 43%, which is close to the assumed load share of 50%
considered in the design of piled raft. Yamashita et al(1994) have also
reported that the raft carry 49% of the total load which is in close agreement
with the observations made in this study.
6.3 NUMERICAL ANALYSIS
The piled raft foundation of Palace Regency building was modeled

numerically and the results obtained were compared with the field test results.
The numerical analysis was performed by using program code ANSYS and
the details of which have already been introduced. The analysis shown in the
present work are restricted to elastic condition for the following reasons:
(i) From practical point of view, elastic analysis is simple and most
convenient for the piled raft system which involves complex interaction
between the raft, pile and soil, in addition to the three dimensional nature of
foundation.
(ii) The strain level around piled foundations is small, particularly

under working loads. Thus majority of piled raft foundations can be delt with
using elastic approaches with moduli chosen with due consideration of small
strain condition.
269
(iii) Elastic methods (ex: Poulos and Davis, 1980) contributed

significantly to me practical analysis of different foundations. Further most of
the references available on the behaviour of piled are based on elastic
approaches.
Three dimensional finite element analysis was preformed using

ANSYS program. In order to generate the model, the principles of solid
modelling have been used. The soil has been modelled with eight nodded
brick elements with each node having three degrees of freedom. The raft and
the piles were modelled with Solid45, which is a eight nodded brick element
having three degrees of freedom at each node. The features of the solid45
elements are discussed elsewhere in the previous chapter. They are nodal
translations in the x, y and z directions. The interface characteristics between
the raft and soil have been represented by the element Targe 170 and
Conta 174.
In the analysis perfect contact between the raft and soil is ensured
through default option available in the program. Half model has been taken to
reduce the computing time. Certain amount of marginal variation found
during the idealization has been ignored while generating the model, as this
will not affect the performance of the model during the analysis and accuracy
of the results. Appropriate boundary condition has been imposed on the edges
of the model. In order to generate the mesh, map meshing technique has been
adopted. Reul and Randolph (2002) have studied the effect of mesh
refinement on the quality and accuracy of the results and have proved that
mesh refinement beyond a certain extent does not enhance the quality of the
results. However in our analysis the maximum aspect ratio adopted was 5.
The simulated finite element model of piled raft is shown in Figure 6.6.
The extent of soil medium and element sizes were chosen by trial
and error to suit the required accuracy and computing time. It was found that
for the breadth of soil medium more than 2 times that of the least width of
foundation the variation in settlement and contact pressures are maximum.
270
Therefore soil medium considered in the half model extends 30m in the x
direction and 30m in y direction. The model has 93,000 elements and 108761
nodes.
Figure 6.6 Finite Element Simulation and Meshing of Piled Raft
6.3.1 Material Property
There are only two materials namely concrete and the soil. The
property of the concrete namely the elastic modulus was taken as per the
recommendation of I S 456-2000. In the case of soil the elastic modulus of the
soil Es was found from the N values. The properties of materials used in the
ANSYS analysis are as presented in Figure 6.7. The Poisson’s ratio of 0.35
was used in the analysis for all the soil layers and for the concrete the value
adopted is 0.20.
271
Figure 6.7 Elastic modulus of various layers
6.3.2 Application of Load
Since the structure is a tall frame, a three dimensional frame

analysis was performed for various load combinations and support reactions
at column points were arrived at. For the foundation design these support
reactions were taken. The column forces (axial, lateral and moment)
computed from the three dimensional frame analysis in the form of support
reactions were applied at the respective column locations. Since the
contribution of transient loading namely wind and earthquake in the form of
horizontal loads are beyond the scope of present study, these forces were not
accounted for in this analysis. Even though the raft was placed at a depth of
3m below the original ground level, the change in the insitu stresses due to
excavation was not simulated in the present study.
6.3.3 Settlement of Piled Raft
The output from the analysis pertaining to the settlement behaviour

was studied from the settlement contour and the results are shown in
Figure 6.8.
272
Figure 6.8 Settlement contour
The nodes representing each settlement marker was located and the
magnitude of settlement was picked up. These values were compared with the
final settlement measured along the longitudinal (grids P,B,&G) and
transverse sections (B14-P14) 10 Figure 6.9. Since the structure has symmetry
in transverse direction one section was considered adequate to represent the
transverse behaviour. The measured settlement along any section shows that
the settlement at edges is more than the centre and the maximum difference is
around 4mm. However it is to be noted the central area is courtyard which is
only 3 storied. The column load is relatively smaller which justifies the
reduced settlement This indicates that the differential settlement is far less
than the permissible value. Further the settlement at the edge and centre of the
piled raft are 6% and 7% respectively of the plain raft elastic settlement
(computed). Thus the piled raft reduced the settlement effectively despite
adopting a raft of 600mm thick (flexible raft).
273
OUTER ROW, GRID P INNER ROW, GRID G
16 16
14 14
SETTLEMENT, mm
SETTLEMENT, mm
12 12
10 10
8 8
6 6
4 Settlement observed 4 Settlement observed
2 Computed 2 Computed
0 0
0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35
DISTANCE, m DISTANCE, m
BOTTOM GRID B TRANSVERSE SECTION

16 16
14 14
SETTLEMENT, mm
SETTLEMENT, mm
12 12
10 10
8 8
6 6
4 Settlement observed 4 Settlement observed
2 Computed 2 Computed
0 0
0 5 10 15 20 25 30 35 0 4 8 12 16 20 24
Figure 6.9 Observed Settlement Vs Computed Value at Various Section
A comparison of the observed and computed settlement profile

indicates that the results agree well. However, the computed values are
marginally higher in general. The computed settlement is higher by 2.5mm
than the observed values in the edges and the same is lesser by 3.5mm in
the mid portion of the raft. This is perhaps in reality the structure had a
retaining wall which was not modeled in the numerical model. The presence
of retaining wall increased the stiffness of the raft causing a reduction in the
settlement. In the central portion the computed settlement was less than the
observed value. This is perhaps the contribution of super structure rigidity
was not taken in the numerical modelling. Also in elastic analysis irrespective
of the relative stiffness of soil-foundation system and types of soil, the contact
pressure is always higher (manifolds) at the edges than at central part of the
274
raft. But this difference reduces in the central grid B. In grid G the computed
values are lower than the observed value.
The results in the central portion where the courtyard has been
located agree very closely. Here the loadings are relatively smaller, indicating
that at lower level of load the analytical and the observed values agree very
well. The computed and the observed results agree to a reasonable extent in
the transverse sections also. Looking at the degree of agreement it can be
stated that the numerical and prototype result compare closely for linear
elastic condition of soil. Since most of the settlement has taken place during
the construction and after post construction increase in settlement is very less,
we can say that under working load the load settlement relation is close to
elastic behaviour of soil.
6.3.4 Load Sharing Between the Raft and Piles
In Figures 6.10, 6.11 and 6.12 the contact stress at selected points
are presented. The stress thus obtained along specific sections are presented
and discussed below. Figure 6.13 presents the contact pressure along the
grids P, G and B. These are typical grids having the piles. The stress has been
picked up close to the pile location and in between the pile location along the
grids. Each grid presents contact pressure over half section of the raft as the
either half is symmetrical with each other. The peak values indicate the stress
near the piles and the lower values indicate the stress in between the piles. In
the case of grid P the raft stress is maximum at the edge; the magnitude being
0.12N/mm2. The peak value varies from 0.10N/mm2 to 0.12N/mm2. In the
grid G the edge stress is 0.07N/mm2 and the peak values various from
0.09N/mm2 to 0.10N/mm2. In grid B the edge and the peak value are more or
less uniform with the magnitude being 0.1N/mm2. The low values which are
in between the pile group indicate an average value of 0.05N/mm2 to
0.06N/mm2. The grid P as can be seen from the pile lay out has a tributary
275
area of raft much lesser compared to the other two grids. This is the outer
most grid having a higher load due to the presence of the wall for the entire
height and RCC retaining wall from the ground to the basement level. Hence
with a higher loading and lesser raft tributary area, higher percentage of load
gets transferred to the piles and the load taken by the raft becomes less.
Figure 6.10 Raft contact stress along grid P
In the case of grid G and grid B, the tributary area of the raft is
higher and the raft shares a higher amount of load. A study of the curves
clearly indicates that the magnitude of the lower peak remains more or less
equal.
The peak value near the pile location ranges from 0.1N /mm2 and
the lower peak value ranges between 0.04 to 0.6kN/mm2. This indicates that
irrespective of the location the magnitude of the contact stress is nearly equal
indicating that the contact pressure is uniform.
276
Figure 6.11 Raft contact stress along grid G
Figure 6.12 Raft contact stress along grid B

277
0.20
Grid line P
0.16
2
CONTACT STRESS, N/mm
0.12
0.08
0.04
0.00
0 2 4 6 8 10 12 14 16
DISTANCE, m
0.16
Grid line G
2
0.12
0.08
0.04
0.00
0 2 4 6 8 10 12 14 16
S C m( )
DISTANCE,
0.16
Grid line B
2
0.12
0.08
0.04
0.00
0 2 4 6 8 10 12 14 16
DISTANCE, m
Figure 6.13 Contact stress at specific points of the raft

278
0.20 0.20
IN BETWEEN GRIDS P AND L IN BETWEEN GRIDS G AND H
RAFT STRESS, N/mm2

RAFT STRESS, N/mm2
0.16 0.16
0.12 0.12
0.08 0.08
0.04 0.04
0.00 0.00
0 4 8 12 16 0 4 8 12 16
0.20 0.20
IN BETWEEN GRIDS B AND C TRANSVERSE SECTION

RAFT STRESS, N/mm2
0.16
RAFT STRESS, N/mm2

0.16
0.12 0.12
0.08 0.08
0.04 0.04
0.00 0.00
0 4 8 12 16 0 4 8 12 16 20 24
0.20 0.20
TRANSVERSE SECTION TRANSVERSE SECTION

RAFT STRESS, N/mm2
RAFT STRESS, N/mm2
0.16 0.16
0.12 0.12
0.08 0.08
0.04 0.04
0.00 0.00
0 4 8 12 16 20 0 4 8 12 16 20
Figure 6.14 Contact stress between the rows of piles in transverse

sections
Figures 6.14 presents the raft – soil contact stress variation in other
raft sections, which are between the pile grids and along the transverse
section. The first section marked, is section between grids P and L taken
approximately at a distance of 1.5m down from the grid P. The raft has an
edge stress of 0.06N/mm2. The average contact stress on the raft found out to
0.06N/mm2, with a variation being from 0.04N/mm2 to 0.07N/mm2. The
279
second section marked between grid G and H lies exactly in the middle of
grid G and H. In this case the edge stress is in the order of 0.055N/mm2 and
the stress along this section is uniform than other sections. The third section is
between B and C grids which is located at a distance of around 1.4m from the
grid B appears towards C. At this section, the edge stress is higher which is of
the order of 0.061N/mm2 and the average stress is 0.055N/mm2. The
transverse section taken in between the grids 13 and 15 indicates a similar
trend with relatively uniform stress of 0.045Nmm2. As can be seen in all these
grids, the raft contact stress is fairly uniform and of the order of 0.055N/mm2
which is 38% of the applied load.
Figure 6.15 Typical Pile numbering
The load shared by the pile group has been arrived from the
average vertical stress values of pile elements. Figure 6.15 presents the pile
layout considered for picking up the stress on the pile to compute the head
280
load and tip load of the pile. Very small areas of stress concentrations have
been ignored, stresses at the head and tip of certain piles are presented in
Figure 6.16 and 6.17 respectively. From the figures it can be seen that the
stress in the pile is decreasing towards the positive gradually. The decrease in
stress with depth on the pile is due to frictional resistance offered by the soil
around the pile. The average of remaining values of stress multiplied by the
area has been taken as head load and tip load.
Figure 6.16 Typical head stress values
Figure 6.17 Typical tip stress values

281
The head load and tip load distribution with the column load has
been plotted in the form bar chart as shown in Figure 6.18. In general the
percentage of tip load varies from 20% to 40% but in majority of the cases the
variation is 20% to 30%. Further it is noticed that the tip resistance of piles in
the central part of tower sections is more. It is because of more confinement
to these piles than at the edges. This is also quite evident from the section B,
where the tip resistance of the piles are the lowest. The magnitude of the tip
load clearly indicates the ductile behaviour of the piles as expected.
Correspondingly the raft has taken a reasonable share of the load.
PILE HEAD LOAD AND TIP LOAD - GRID P

200
160
LOAD, t
120
80
40
0
0 3.81 9.32 14.04
DISTANCE, m
PILE HEAD LOAD AND TIP LOAD - GRID G

320
280
240
LOAD, t
200
160
120
80
40
0
0 4.3 8.6 12.9
DISTANCE, m
PILE HEAD LOAD AND TIP LOAD - GRID B

240
200
160
LOAD, t
120
80
40
0
0 4.3 8.6 12.9
DISTANCE, m
Figure 6.18 The Head Load – Tip Load Distribution with the column
load (continue…)
282
ALONG GRID 'L'

520
480
440
column load head load tip load
400
360
LOAD, t
320
280
240
200
160
120
80
40
0
0 4.3 9.3 12.7 (LIFT
DISTANCE,
DISTANCE (mm) PIT
COLUMNS)
ALONG GRID J / H
520
480
440
column load head load tip load
400
360
320
280
LOAD, t
240
200
160
120
80
40
0
0 4.3 8.6 13.6
DISTANCE, m
Figure 6.19 Contd. The Head Load – Tip Load Distribution with the
column load
Hence it can be conclusively said that the behaviour of the

foundation system has been as expected with the piles behaving as a flexible
group and settlement reducer. Though the piles have been located below the
columns, the load sharing of the raft has been 43% from the observed
settlement value and 37% as per the numerical modeling with the piles
283
behaving as a settlement reducer sharing 57% to 63% of the total applied

load.
6.4 SUMMARY
The study done on the prototype piled raft has indicated that the
foundation system can be used successfully for moderately loaded structures
also. It has shown that the major portion of the settlement takes place during
the construction period itself. The settlements measured at field and the values
obtained from the numerical analysis agreed closely indicating this, even
linear analysis can predict the behaviour of the piled raft so long as the
settlement is elastic settlement. The load sharing between the pile and the raft
had been 57% and 43% as observed. This indicates that the design compares
well with the third generation piled raft as qualified by Katzenbach. The
numerical model predicted relatively higher settlement and lower value for
the load shared by the pile group.
284
7
8CHAPTER 7
CONCLUSIONS
7.1 GENERAL
The performance of piled raft foundation on sand bed was studied

by conducting a series of 1g model tests. The results were examined to study
the influence of various parameters relating to the foundation and the
supporting soil on the settlement reduction and load sharing behavior. The
results of the small scale model studies were then compared with the results
of the numerical analyses carried out using ANSYS FEM code. Concurrently
the behaviour of a piled raft foundation supporting a twelve storied building
was monitored by installing settlement gauges over a period of 796 days
including post construction period of 436 days. A three dimensional linear
analysis was performed and the results were compared.
7.2 SMALL SCALE MODEL STUDIES
In all 110 tests including plain and piled rafts were conducted on
loose, medium dense and dense sand. Three shapes of piled raft namely
circular, square and rectangular were considered in this research. A series of
tests on free standing pile groups were also performed and compared with the
behaviour of piled raft to know the contribution of ground contacting cap (raft
of piled raft) on load sharing as well as reduction in settlement of piled raft.
The load–settlement response of piled raft and unpiled raft is

almost similar. The response is three phase within the settlement range of
10% of the least lateral dimensions of the model raft. This response is seen in
all the three shapes of the piled raft and is independent of density of sand, pile
285
parameters such as diameter, length, spacing (area ratio), configuration

(layout) and thickness of raft. Further, non-dimensional plots between load
ratio and settlement ratio also confirm that load–settlement responses of piled
raft with piles of different diameter and lengths are similar.
The three phase behaviour consists of elastic phase followed by

elasto –plastic behaviour with strain hardening response. This response is
characterised as multi linear strain hardening behaviour. The elastic response
is seen till the settlement of piled raft is around 1% of the dimension of the
raft, which is more or less equal to the critical settlement of free standing pile
group (ie, settlement requires to mobilise frictional resistance fully in the free
standing pile group).
Though the characteristic response of raft and piled raft is the

same, the piles-raft-soil stiffness is higher than raft-soil stiffness in all the
three phases of the load–settlement response irrespective of density of sand.
However presence of piles in the ground contacting raft (ie, piled raft) is
much pronounced in the elastic phase by recording high stiffness than plain
raft. In the second and third phase, the piled raft–soil stiffness is reduced and
the stiffness in the third phase is close to the stiffness of plain raft–soil
indicating contribution by the pile group is reduced drastically and its share
on applied load remains almost constant in the third phase.
The piled raft stiffness is increased with increase in diameter, pile

length and area ratio of pile group. For a given area ratio and density of sand,
the stiffness of piled raft is higher for longer piles. However, pile lengths
more than 80% of the lateral dimension of the raft (width or diameter), the
increase in stiffness of pile - soil is not appreciable. Further, increase in area
ratio of piles more than 6% is not contributing to stiffness of piled raft
appreciably.
286
Free standing pile groups exhibited two phase behaviour

irrespective of density of sand and number of piles. In the present study
ultimate capacity was reached between the settlement of 2mm and 3mm for
the dimensions of piles adopted in this study, which is close to 1% of the
lateral dimension of the circular and square raft tested in this research.
The increase in stiffness of piled raft (or capacity of piled raft) is

due to increase in capacity of piles of piled raft than free standing piles. The
capacity of piles installed beneath the raft is larger than that observed from
free standing piles. This is attributed to additional confining stress to the piles
due to the load transferred from the raft to the soil within the group.
In all the cases, the pile group shares major portion of the applied
load till the settlement around 1% of the least lateral dimension of the raft. As
the settlement increased, the load sharing ratio decreased till the settlement
closed to 3% of the lateral dimension of the raft and then remained almost
constant. Irrespective of the shape of the raft, the load sharing ratio of piles is
higher for larger diameter, longer piles and higher area ratio. The maximum
load sharing ratio is close to 70% for the area ratio of 9.25% and is reduced to
25% for the area ratio of 4.25% in the elastic phase of the load–settlement
response, whereas the same is reduced to 35% and 10% respectively in the
third phase of load–settlement response. Thus, the piles are effective in load
sharing as long as the settlement due to applied load is less and is elastic.
Similarly settlement reduction ratio of piled raft is increased with increase in
the diameter, the length and the area ratio of piles. Further, settlement
reduction ratio is higher in the initial stages of load (ie, elastic phase of load –
settlement) and reduces with increase in load.
The performance of piled raft by way of settlement reduction and

load sharing is better in loose and medium dense sand in the initial stages of
load and till the settlement level is around 3% of the dimension of the raft.
However when the intensity of load is higher the reduction in settlement is
287
found to decrease in loose sand when compared to medium and dense sand
conditions.
The parametric study on the circular and non circular shaped piled
rafts indicated that among the parameters analysed, the length and the
diameter of piles have the maximum influence. However the pile length more
than 0.8 times the least lateral dimension of the circular and square rafts tested
does not have appreciable enhancement on the settlement reduction or load
sharing. Similarly the d/t ratio (d = diameter of pile and t = thickness of raft)
close to unity provides effective performance of the piled raft. The raft
thickness by far, to the extent studied, has no major influence in the
settlement reduction and load sharing. Also it was found that the performance
of piled raft was more effective in the case of loose and medium dense sand
than in dense sand.
The effect of the number of piles in the form of pile raft area ratio
indicated that even a small number of piles can reduce the settlement.
However it is found that a pile–raft area ratio between 5% and 6% appears to
produce an optimum level of settlement reduction and load sharing. It is also
found that increase in the number of piles beyond 6% was not producing
appreciable increase in the load sharing and settlement reduction as indicated
by Cooke (1983) and Poulos (2001) based on their studies on over-
consolidated clay. Further the above observation of present study is in
conformity with the conclusion of Kim et al. (2001).
In the case of circular piled raft the radial and square grid
arrangement of piles produced identical results. However the study on the
effect of configuration, both on the circular and square shaped piled raft
showed that the piles have to be evenly spread over the entire area of the raft
to produce effective overall settlement reduction in the case of uniform
distribution of load on the foundation.
288
7.3 NUMERICAL MODELLING
The load-settlement response of circular, square and rectangular

piled rafts was studied numerically adopting ANSYS FEA code. Both linear
and nonlinear analyses were performed. The nonlinear behaviour of soil was
idealised using multi – linear isotropic hardening material model (MISO).
The load–settlement relationship of piled raft was better predicted

in the linear analysis particularly in the first phase of load–settlement
response. The 3D nonlinear analyses with MISO model predicted the load–
settlement response well for the entire range of loads applied in the
experiments on circular and square piled rafts. The maximum difference
between the measured and calculated loads was less than 10%. Thus the
MISO material model predicted the three phase behaviour of piled raft as
observed in the experiments. The piled raft system exhibited elastic behaviour
in the initial stages of load. On increase of load the pile–soil and raft–soil
interfaces became plastic gradually.
The raft contact stresses obtained from the numerical analyses

showed uniform distribution except at the edges and pile locations.
The load on piles increased with increase in applied load on pile

but not in direct proportion. It was also observed that load on centre pile of
circular raft and centre row of square piledraft were higher than piles at other
locations. The load on piles decreased with distance as reported by Turek and
Katzenbach (2003).
The shaft stress distribution obtained from the numerical analysis

indicated that the shaft stress is higher near the pile head and reduces towards
the pile tip. The reduction is gradual till the length of pile of 12 times the pile
diameter and the reduction is increased below this depth. The depth at which
the change in rate of stress distribution occurred is close to critical depths of
pile in sand as reported by Vesic (1967).
289
The load sharing behavior predicted in the form of load sharing

ratio αpr by the numerical model agreed reasonably with the results produced
by the model test. In both the cases irrespective of the shape of the raft, the
pile group shared a higher proportion of the load initially and beyond the
settlement level of around 3% of the least lateral dimension of the raft the
load shared by the pile group remains constant. This confirms that the pile
group enables the system to take a higher load initially and further confirms
the behavior of pile group as settlement reducer and sharing a reasonable
amount of load at higher settlements. However, the load sharing ratio mainly
depends on the area ratio of pile group.
It was also found from the analyses that the tip stresses in piles
were lesser than the head stresses indicating that the shaft friction was fully
mobilized and the piles were dominantly friction piles. This is in conformity
with the design requirement of piles of piled raft as suggested by Padfield and
Sharrock (1983). Also the mobilization of higher raft stress as the load
increased and low pile tip stress at the final settlement confirmed the
behaviour of the pile group as settlement reducer. This is in agreement with
the conclusions of Franke (1991).
7.4 BEHAVIOUR OF PROTO TYPE PILED RAFT
In order to study the suitability of this foundation system to support

a moderately loaded structure, the piled raft foundation of a twelve storied
residential cum commercial complex was monitored with settlement gauges
for a period of 796 days including post construction period of 436 days. The
observed settlements were used to compute the load shared by the raft with
time and loading sequence from elastic theory. The piled raft foundation
system was numerically modeled and a three dimensional linear analysis was
performed. The settlement profile and the contact stress below the raft were
computed. The pile head load and the tip load were taken from the analysis
and the load sharing was evaluated. It was found that the observed and the
290
analytical results agree closely even though the analytical model predicted a
higher load share for the piles.
The maximum settlement at the end of the post construction period

and immediately after constructions indicate that the major part of settlement
took place during construction. The total settlement is mostly elastic in nature
as the pressure at the base of the raft is close to the existing overburden
pressure at the level of raft.
The settlements predicted by numerical model using ANSYS

compare well with the observed values indicating that linear modeling is
adequate to predict the settlement and load sharing between piles and raft as
long as settlement caused by the load is elastic.
The load sharing behavior as predicted by the numerical analysis

indicates that the raft stress at the edges is higher than that in the centre. But
the contact pressure in between the pile rows in both the directions is uniform.
The introduction of piles as observed has not only made the contact pressure
uniform but also reduced the overall settlement. The load share is of the order
of 62% and 38% for pile and raft respectively. However in the central portion
where the raft area between the piles is larger, the raft share increases, which
is about 41% of the total load. The load share predicted by the elastic analysis
using the tributary area of the raft and observed settlement is 43%. The
numerical model predicts a higher load sharing for piles and lower for raft, as
compared to what has been based on the tributary area and elastic analysis
The tip stress in the pile was found to be of the order of 25% of the
head stress; as a ratio of the total load on the piled raft, the tip stress was of
the order of 10 to 15%. This indicates that the transfer of the load by the pile
group was mostly by friction (mobilization of shaft friction was very nearly
complete) as observed in the case of 1g model tests in sand. Further the
percentage load shared by the raft is also in agreement with the observations
made in the 1g model tests.
291
7.5 SCOPE FOR FURTHER STUDY
The present study has covered a reasonably wide range of

parameters on the behaviour of piled raft foundation in sand which is an
acceptable alternative foundation for the settlement reduction. Further studies
can be concentrated out on the following aspects.
• The optimum area ratio of pile found by the 1g model tests

should be confirmed through centrifuge modelling.
• The studies can be extended to the piled raft wherein the raft
is located at a deeper depth accounting for the surcharge and
the presence of basement retaining wall.
• The effect of super structure rigidity can be taken into

account for further study along with the large lateral loads
associated with tall structures in addition to the vertical load.
• The effect of parameters like relative raft-pile stiffness and

raft-soil stiffness needs further study.
• The effect of pile installation technique on the load sharing

behaviour of the piled raft can be included in the study.
• Large pile groups are associated with equipment foundations

for turbo generators and compressors etc. These are
subjected to large dynamic forces. The applicability of piled
raft subjected to dynamic loads will be an useful study.
Studies on the behaviour of piled rafts with raker piles subjected to

lateral loads can be of considerable importance keeping in mind large tank
foundations.
The study on the behaviour of piled raft in sand adopting

centrifuge models will be of much interest, since the behaviour of foundation
in sand is dependent on the applied stress and the stress history.
292
6APPENDIX 1
TYPICAL LCPT (LIGHT CONE PENETRATION TEST)

RESULTS
Specifications
Weight of hammer - 100N Location1 Location2
Height of drop - 500 mm
200mmdia
Diameter of cone - 25 mm Raft
Apex angle - 60°
Location4
Material - EN8 Stainless Steel
Location3
Dia of driving rods - 200mm
Test tank 1m X 1mx 0.60m deep
Length of driving rods - 1000mm
Table A1.1 Typical results of LCPT in dense sand
Depth of Cumulative number of blows

penetration in mm
Location 1 Location 2 Location 3 Location 4
100 2 2 2 2
200 5 4 4 4
300 8 8 9 8
400 13 13 13 13
500 18 18 18 18
7APPENDIX 2
MORI’S CHART
Figure A2.1 Relation between N value and modulus of deformation (Es)

(Mori, 1965)
29
294
REFERENCES
1. Brown P.T. and Weisner T.J. (1975), ‘Behaviour of Uniformly Loaded

Piled Strip Footings’, Soils and Foundations, Vol. 15, pp. 13-21.
2. Burland J.B. (1995), ‘Piles as Settlement Reducer’ 18th Italian

Congress on Soil Mech., Pavia.
3. Burland J.B., Broms B.B. and de Mello V.F.B. (1977), ‘Behaviour of

Foundations and Structures’, Proc. 9 ICSMFE Tokyo 2, pp. 495 – 546.
4. Butterfield R. and Banerjee P.K. (1971), ‘The Problem of Pile Group –

Pile Cap Interaction’, Geotechnique 21, No. 2, pp. 135 – 142.
5. Chin F.K. (1970), ‘Estimation of the Ultimate Load of Piles from Tests
carried to Failure’, Proc. 2nd S.E. Asian Conf. on Soil Engg, Singapore,
pp. 81-92.
6. Chin F.K. (1972), ‘The Inverse Slope as a Prediction of Ultimate

Capacity of Piles’, Proc. 3rd S. E. Asian Conf. on Soil Engg,
Hongkong, pp. 83-91.
7. Clancy P. (1993), Numerical Analysis of Piled Raft Foundations,

University of Western Australia, PhD Thesis.
8. Clancy P. and Randolph M.F. (1992), ‘Analysis and Design of Piled

Raft Foundations’, Research Report No. G 1062, Department of Civil
Engineering, The University of Western Australia.
9. Clancy P. and Randolph M.F. (1993), ‘Simple Design Tests for Piled
Raft Foundations’, Geotechnique, Vol. 36, No. 2, pp. 169-203.
10. Chow Y.K. (1986), ‘Analysis of Vertically Loaded Pile Groups’, Intnl.
Jnl. for Numerical and Analytical Methods in Geomechanics, Vol. 10,
No.1, pp. 59-72.
11. Chow Y.K. (1987b), ‘Axial and Lateral Response of Pile Groups
Embedded in Non-homogeneous Soils’, Intnl. Jnl. for Numerical and
Analytical Methods in Geomechanics, Vol. 11, No. 6, pp. 621-638.
295
12. Chow Y.K. and The C.I. (1991), ‘Pile-Cap-Pile-Group Interaction in

Nonhomogeneous Soil’, Journal of Geotechnical Engineering, ASCE,
Vol. 117, No. 11, pp. 1655-1668.
13. Cooke R.W. (1986), ‘Piled Raft Foundation on Stiff Clays – A

Contribution to Design Philosophy’, Geotechnique, Vol. 36, No. 2,
pp. 169 -203.
14. Cooke R.W., Bryden-Smith D.W., Gooch M.N. and Sillett D.F. (1981),
‘Some Observations of the Foundation Loading and Settlement of a
Multi- story Building on a Piled Raft Foundation in London Clay’,
proc. Institution of Civil Engineers, Part 1, Vol. 70, pp. 433-460.
15. Cunha R.P., Poulos H.G. and Small J.C. (2001), ‘Investigation of
Design Alternative for a Piled Raft Case History’, Journal of
Geotechnical and Geoenvironmental Engineering, ASCE, pp. 635-641.
16. Daniel Thangaraj (2006), ‘Interaction analysis of soil-raft-space frame:

parametric analysis and laboratory models studies, Ph.D. thesis
submitted to Anna University.
17. Davis E.H. and Poulos H.G. (1972), ‘The Analysis of Piled Raft
System’, Aust. Geomech. Jnl., Vol. 2, pp. 21-27.
18. Desai C.S., Johnson L.D. and Hergett (1974), ‘Analysis of Pile
Supp.orted Gravity Lock’, Jnl. Geotech. Engg., Vol. 100, No. 9,
pp. 1009-1029.
19. Dickin and Leung (1990), ‘Performance of Piles with Enlarged Bases
subject to Uplift Forces’, Canadian Geotechnical Jnl, Vol. 27, No. 5,
pp. 546-556.
20. Dickin and Leung (1992), ‘The Influence of Foundation Geometry on

the Uplift Behaviour of the Piles with Enlarged Bases in Sand’,
Canadian Geotechnical Jnl, Vol. 29, No. 3, pp. 498-505.
21. El’Mossallamy Y. and Frank E. (1997), Piled Raft: Numerical

Modelling to Simulate the Behaviour of Piled Raft Foundations,
Darmstadt: the authors.
22. Fleming W.G.K., Weltman A.J., Randolph M.F. and Elson W.K.
(1992), Piling Engineering, 2nd Edition, John Wiley and Sons.
296
23. Frank E. (1991), ‘Measurements Beneath Piled Rafts’, Keynote

lecture, ENPC Conference Paris, pp. 1 –21.
24. Frank E., Lutz B. and El-Mossallamy Y. (1994), ‘Measurements and

Numerical Modelling of High Rise Building Foundations on Frankfurt
Clay’, Geotechnical Special Publication 40, American Society of Civil
Engineers, New York, pp. 1325 – 1336.
25. Fraser R.A. and Wardle L.J. (1976), ‘Numerical Analysis of

Rectangular Rafts on Layered Foundations’, Geotechnique, Vol. 26,
No. 4, pp. 613 -630.
26. Gandhi S.R and Maharaj D.K. (1996), ‘Analysis of Piled Raft
Foundations’, 6th International Conference on Piling and Deep
Foundations, Bombay, pp. 1.11.1-1.11.7.
27. Hain S.J. (1975), ‘Analysis of Rafts and Raft-Pile Foundations’, Proc.
of Symposium on Recent Development In Soil Mechanics, Univ.
NSW. Australia, pp. 213 – 253.
28. Hain S.J. and Lee I.K. (1978), ‘The Analysis of Flexible Pile- raft
Systems’, Geotechnique, Vol. 28, No. 1, pp. 65-83.
29. Hemsley J.A. (2000), Design App.lication of Raft Foundations,

Thomas Telford.
30. Hooper J.A. (1974), ‘Observations on the Behavior of a Piled Raft

Foundation on London Clay’, Proc. Institution of Civil Engineers, Part
2, Vol. 55, pp. 855-877.
31. Horikoshi K. (1995), ‘Optimum Design of Piled Raft Foundations’,

Dissertation submitted for the degree of Doctor of Philosophy,
University of Western Australia.
32. Horikoshi K. and Randolph M.F. (1996), ‘Centrifuge Modeling of

Piled Raft Foundations on Clay’, Geotechnique, Vol. 46, No. 4,
pp. 741-752.
33. Horikoshi K. and Randolph M.F. (1998), ‘A Contribution to Optimal

Design of Piled Raft’, Geotechnique, Vol. 48, No. 3, pp. 301 – 317.
34. Ilamparuthi, K. (1991), Experimental Investigation on the Pullout

Capacity of Plate Anchors in Sand, PhD Thesis, IIT Madras, Chennai
297
35. IS:1904 (1986) Code of Practice for Design and Construction of

Foundations in Soils: General requirements.
36. IS:2911 (Part 1-Section2) (1979), Code of Practice for Design and
Construction of Bored Cast in situ Piles.
37. IS:2911 (Part 4) (1985), Code of Practice for Design and Construction
of Pile Foundations, Load Test on Piles.
38. Katzenbach R. (1993), ‘The Technical and Economical Importance of

the Combined Piled Raft Foundation Represented by Example of High
rise Buildings’, Bautechniks, 70-3, pp. 161-170.
39. Katzenbach R., Arslan V. and Moorman ch (1998), ‘Design and Safety
Concept of Piled Raft Foundations’, Proc. of 3rd Int. Conference on
Deep Foundations on Bored and Auger Piles, pp. 439-448.
40. Katzenbach R., Arslan V. and Moorman ch (2000a), ‘Numerical

Stimulations of Combined Piled Raft Foundations for the New High
Rise Building, Max in Frankfurt am main’, Proc. 2nd Int. Conf. on Soil
Structure Interaction in Urban Civil Engineering.
41. Katzenbach R., Arslan V. and Moorman ch. (2000b), ‘Piled Raft
Foundation Projects in Germany’, Design Application of Raft
Foundations, Ed, by J.A. Hemsley Thomas Telford Ltd, pp. 323 – 391.
42. Katzenbach R. and Reul O. (1997), ‘Design and Performance of Piled

Rafts’, Proc. XIV Inst. Conference on Soil Mechanics and Foundation
Engineering, pp. 2253-2256.
43. Kim K.N., Lee S.K., Chung C.K., Kim M.N. and Lee H. (2001),
‘Optimal Pile Arrangement for Minimizing Differential Settlements in
Piled Raft Foundations’, Computers and Geotechiques, Vol. 28,
pp. 235 – 253.
44. Kim H.T., Yoo H.K. and Kang I.K. (2002), ‘Genetic Algorithm
Optimum Design of Piled Raft Foundations with Model Tests’, Journal
of South East Asian Geotechnical society, pp. 1-9.
45. Kuwabara F. (1989), ‘An Elastic Analysis for a Piled Raft Foundation
in a Homogeneous Soil’, Soils and Foundations, Vol. 28, No. 1,
pp. 82-92.
298
46. Lee I.K. (1993), ‘Analysis and Performance of Raft and Raft–pile
Systems’, Keynote Lecture, 3rd Int. Conf. on Case Histories in
Geotechnical Engineering, St Louis ADFA UNIV. NSW. – Australia.
47. Maharaj D.K. (1996), ‘App.lication of Elastic and Elasto-plastic

Analysis for Piled Raft Foundations’, PhD Thesis, IIT Madras,
Chennai
48. Mori (1965), Discussion, Proc. 6th Intl. Conf. Soil Mech. Foundn.
Engg. Montreal, pp. 501-504
49. Morton K. and Au E. (1975), ‘Settlement Observations of Eight

Structures in London’, Proc. Conf. of Settlement of Structures, Pentech
Press, London, pp. 182-203.
50. Murray E.J. and Geddes J.D. (1989), ‘Resistance of Passive Inclined
Anchors in Cohesionless Medium’, Geotechnique, Vol. 39, No. 3,
pp. 417-431.
51. O’Neill M.W., Caputo V., DeCook F., Hartikeinen J. and Mets M.
(1996), Case Histories of Pile Supp.orted Raft – Report for ISSME
Technical Committee TC-18, Houstan, Tx: University of Houstan.
52. O’Neill M.W., Ghazzaly O.I. and Ha H.B. (1977), ‘Analysis of Three-
dimensional Pile Groups with Non-Linear Soil Response and Pile-Soil-
Pile Interaction’, Proc. 9th Annual Offshore Technology Conference,
Houston, Paper OTC 2838, pp. 245-256.
53. Padfield C.J. and Sharrock M.J. (1983), ‘Settlement of Structures on

Clay Soils’, Construction Industry Research and Information Institute,
Special Publication 27.
54. Poulos H.G. (1968), ‘Analysis of the Settlement of Pile Groups’,

Geotechnique, Vol. 18, pp. 449 - 471.
55. Poulos H.G. (1972), ‘Load Settlement Prediction for Piles and
Friction’, Journal of the Soil Mechanics and Foundations Division,
Proc. of the ASCE Sept. 1972, pp. 879 -897.
56. Poulos H.G. (1989), ‘Pile Behaviour – Theory and Applications’,

Geotechnique, Vol. 39, No. 3, pp. 365 - 415.
57. Poulos H.G. (1991), ‘Analysis of Piled Raft Foundations’, Computer

Methods and Advances in Geotechniques, Ed. Beer et al., Balkema,
Rotterdam, pp. 153–191.
299
58. Poulos H.G. (1993), ‘Piled Raft in Swelling or Consolidating Soil’, Jnl.
Geotech. Divn. ASCE-119, No.2, pp. 374–380.
59. Poulos H.G. (1994), ‘An App.roximate Numerical Analysis Pile Raft
Interaction’, International Journal of Numerical and Analytical
Methods in Geotechniques, Vol. I, No. 18, pp. 73-92.
60. Poulos H.G. (1994b), ‘Alternate Design Strategies for Piled Raft
Foundations’, Proc. 3rd Int. Conf. Deep Foundations – Singapore,
pp. 239-244.
61. Poulos H.G. (2001), ‘Piled Raft Foundation: Design and App.lication’,
Geotechnique, Vol. 51, No. 2, pp. 111-113.
62. Poulos H.G., Small J.C. Ta L.D., Sinha J. and Chen L. (1997),
‘Comparison of Some Methods for Analysis of Piled Rafts’, Proc. 14th
Int. Conf. Soil Mech. Foundn. Engg- Hamburg – 2, pp. 1119-1124.
63. Poulos H.G. and Davis E.H. (1968), ‘The Settlement Behaviour of
Single Axially Loaded Incompressible Piles and Piers’, Geotechnique,
Vol. 18, pp. 351-337.
64. Poulos H.G. and Davis E.H. (1974), Elastic Solutions for Soil and
Rock Mechanics, John Wiley and Sons, New York.
65. Poulos H.G. and Davis E.H. (1980), Pile Foundation Analysis and
Design, John Wiley and Sons, New York.
66. Price G. and Wardle I.F. (1986), ‘Queen Elizabeth II Conference

Centre- Monitoring of Load Sharing between Piles and Raft’, Proc.
Inst. Civil Engers. 80-No: 1 pp. 1505-1518.
67. Prokoso W.A. and Kulhawy F.H. (2001), ‘Contribution of Piled Raft
Foundation’, Journal of Geotechnical and Geoenvironmental
Engineering, ASCE, pp. 17-24.
68. Randolph M.F. (1983), ‘Design of Piled raft Foundation’, Proceedings

of Recent Developments in Laboratory and Filed Tests and Analysis of
Geotechnical Problems, Bangkok, pp. 525-537.
69. Randolph M.F. (1994), ‘Design Methods for Pile Group and Piled
Rafts’, Proc. 13th International Conference on Soil Mechanics and
Foundation Engineering, New Delhi, pp. 5,61-82.
300
70. Randolph M.F. and Worth C.P. (1978), ‘Analysis of Deformation of

Vertically Loaded Piles’, J .Geotechnical Engineering Division, ASCE,
Vol. 104, No. GT12, pp. 1465-1488.
71. Randolph M.F. and Worth C.P. (1979), ‘An Analysis of the Vertical
Deformation of Pile Groups’, Geotechnique, Vol. 29, No. 4,
pp. 423–439.
72. Reul O. (2000), ‘In-Situ Messungen Und Numerische Studien Zum

Tragverhalten Der Kombinierten Pfahl-Plattngrundung’, Mitteillungen
Desinstituts Und Der Versuchanstilt Fur Geotechnique Der
Technischen Universitadt. Darmstadt. Heft.53.
73. Reul O. and Randolph M.F. (2002), ‘Study on the Influence of Finite
Element Mesh Refinement on the Calculated Behavior of Piled Raft’,
Proc. of 8th Int. Symp. Numer. Models. Geomech Rome, pp.259-264.
74. Reul O. and Randolph M.F. (2003), ‘Piled Rafts in Over consolidated
Clay: Comparison of In-situ Measurements and Numerical Analyses’,
Geotechnique, Vol. 53, No.3, pp.301-315.
75. Reul O. and Randolph M.F. (2004), ‘Design Strategies for Piled Rafts
Subjected to Non Uniform Vertical Loading’, Jnl. Geotech. and
Geoenviron. Engg. ASCE, Jan. 2004, pp.1-13.
76. Russo G. (1998), ‘Numerical Analyses of Piled Rafts’, Intl. Jnl. Num.
and Anl. Methods in Geomech, Vol. 22, pp. 477-493.
77. Russo G. and Viggiani C. (1998), ‘Factors Controlling the Soil

Structure Interaction of Piled Rafts’, Darmstadt, Geotechniques, No. 4,
pp. 297-322.
78. Sinha J. (1996), Analyses of Piles and Piled Raft in Swelling and
Shrinkable Soils, PhD Thesis, University of Sydney, Australia.
79. Sinha J. and Poulos H.G. (1999), ‘Piled Raft System and Free Standing
Pile Groups on Expansive Soils’, Proc. 8th Aust. and New Zealand
Conference, Geomech. Huburt, 1, pp. 207-212.
80. Schwab H., Gundling N. and Lutz B. (1991), Field Measurements in

Geotechnique (Ed), Sorum, Balkema.
81. Thaher M. and Jessberger H.L. (1991), ‘The Behaviour of Pile-raft

Foundations Investigated in Centrifuge Model Tests’, Proc. Centrifuge
91, Colorado, Balkema, pp. 225-234.
301
82. Ta L.D. and Small J.C. (1995), ‘Finite Layer Analysis of Pile Groups
in Layered Soils’, Int. Symp. on Numerical Models in Geomechanics,
Davoos, Switzerland.
83. Ta L.D. and Small J.C. (1996), ‘Analysis of Piled Raft System in
Layered Soil’, International Journal of Numerical and Analytical
Methods in Geotechniques, Vol. 20, pp.57-72.
84. Ta L.D. and Small J.C. (1998), ‘Analysis and Performance of Piled
Raft Foundations on Layered Soils-Case studies’, Soils and
Foundations, Vol. 38, No. 4, pp. 145-150.
85. Turek J. and Katzenbach R. (2003), ‘Small Scale Model Tests with
Combined Piled Raft Foundations’, Proceedings of the 4th International
Seminar on Deep foundations on Bored and Augured piles, Ghent,
Belgium, pp. 409-413.
86. Van Impe W.F. (1991), ‘Developments in Pile Design’, Proc. 4Th Int.
Conference on Piling and Deep Foundations, Strena.
87. Van Impe W.F. (1991), ‘Developments of Deep Foundations’, Proc.

10th European Conference on Soil Mech. and Foundation Engineering,
Florence, 3, pp. 1031-1062.
88. Van Impe W.F. and Lounge I. (1996), ‘Technical Report on Settlement
Prediction Methods for Piled Raft Foundations’, Ghent, Univ,
Belgium.
89. Vesic A.S. (1969), ‘Experiments with Pile Groups on Sand -

Performance of Deep Foundations’, ASTM, STP, 444, pp.177-222.
90. Viggiani C. (1998), ‘Pile Groups and Piled raft Behaviour’, Deep
Foundations on Bored and Auger Piles, Ed. Van Impe, W.F. and
Haegman, Balkema, Roterdam, pp. 77-90.
91. Weisner T.J. and Brown P.T. (1976), ‘Behaviour of Piled Strip
Footings Subjected to Concentrated Loads’, Aust. Geomechanics Jnl.
G6, pp.1-5.
92. Weisner T.J. and Brown P.T. (1978), ‘Laboratory Tests on Model
Piled Raft Foundations’, Research Report 318, Sydney University.
93. Whitaker T. (1957), ‘Experiments with Model Piles in Groups’

Geotechnique, Vol. 7, No. 4, pp. 147-167.
302
94. Whitaker T. (1961), ‘Some Experiments on Model Piled Foundations

in Clay’, Proc. Symp. Pile foundations, 6th International Congress of
the International Association for Bridge and Structural Engineering,
Stockholm, pp. 124–139.
95. Xu K.J. and Poulos H.G. (2002), ‘General Elastic Analysis of Piles and
Pile Groups.
96. Yamashita K., Kakurai M. and Yamada T. (1994), ‘Investigation of a

Piled Raft Foundation on Stiff Clay’, Proc. of 13th International
Conference on Soil Mechanics and Foundation Engineering, New
Delhi, pp. 543-546.
97. Yamashita K., Yamada T. and Kakurai M. (1998), ‘Simplified Method

for Analysing Piled Raft Foundations’, Deep Foundations on Bored
and Auger Piles, Ed. Van Impe, W.F, Balkema, Roterdam, pp. 457-
454.
98. Zeevaert L. (1957), ‘Compensated Friction Pile Foundation to Reduce

the Settlement of Buildings on Highly Compressible Volcanic Clay of
Mexico City’, Proc. 4th Int. Conf. Soil Mechanics and Found. Engg.
London 2.
99. Zhang G.M., Lee I.K. and Zhao X.H. (1991), ‘Interactive Analysis
Behavior of Raft-pile Foundations’, Proc. Geo coast 91 – Yokahama 2,
pp. 719-764.
303
LIST OF PUBLICATIONS
AWARDS
1. Balakumar V. and Ilamparuthi K. (2002), IGS Afcons Biennial prize
for the best paper on case histories for the year 2002-2003 for the
paper ‘Settlement of Piled Raft A case Study’, Proceedings of IGC
2002, Allahabad, India, pp. 651-654.
JOURNAL
1. Balakumar V. and Ilamparuthi K. (2007), ‘Performance Monitoring
of a Piled Raft Foundation of Twelve Storied Building and Analytical
Validation,’ Indian geotechnical Journal, Vol. 37, No. 2, pp. 94-115.
INTERNATIONAL CONFERENCE
1. Balakumar V. and Ilamparuthi K. (2003), ‘Field Study on Piled Raft
Foundation of Twelve Storied Building in Chennai,’ Proc.
International conference on Field measurements in Geotechnical, 2003
Oslo Norway, pp. 17-22.
2. Balakumr V. and Ilamparuthi K. (2004), ‘Laboratory Study on the
Behaviour of Piled Raft on Granular Soils,’ Proc. 15th South East Asian
Geotechnical Society Conference 2004, Bangkok, Thailand, pp. 293-
298.
3. Balakumar V., Kalaiarasi V. and Ilamparuthi K. (2005).
‘Experimental and Analytical Study on The Behaviour of Circular
Piled Raft on Sand,’ Proc. 16th International Conference on Soil
Mechanics and Geotechnical Engineering-2005, Osaka, Japan,
pp. 1943-1947.
NATIONAL CONFERENCE
1. Balakumar V. and Ilamparuthi K. (2002), ‘Settlement of Piled Raft –

A Case Study,’ Proceedings of Indian Geotechnical Conference, 2002
– Allahabad, India, pp. 651-654.
2. Balkumar V. Ilamparuthi K.and Kalaiarasi V. (2004), ‘Study on the
Behaviour Of Circular Piled Raft in Sand,’ Proc. Indian Geotechnical
Conference 2004, Warangal, India, pp. 362-365.
3. Balakumar V. Ilamparuthi K.and Kalaiarasi V. (2005), ‘Study on
Square Piled Raft on Granular Soil, ‘Proc. Indian Geotechnical
Conference 2005. Ahmedabad, India, Vol. 2, pp. 11-14.
4. Balakumar V. and Ilamparuthi K. (2006), ‘Performance of Model
Piled Raft on Sand,’ Proc. Indian Geotechnical Conference 2006,
Chennai, India, pp. 463-466.
304
VITAE
Sri. V. Balakumar got graduated from I.I.T., Madras, Chennai in

Civil Engineering after obtaining his degree in Physics from Madras
University. He took his M.Sc., Engineering degree from College of
Engineering Guindy, Chennai (presently – Anna University). He started his
carrier in M/S Simplex Infra Structure Limited as a trainee. Subsequent to this
till 1987 he worked in various organizations in different capacities. During his
carrier he was involved in the design and construction of pile foundations,
ground improvement technique and soil investigation for various types of
projects. Tuticorin Thermal Power Project, Chennai Doordharshan Tower,
IFFCO, Kantla, GCDA Marine Drive Scheme, etc. are few of the note worthy
projects wherein he was involved.
In 1987 he started his own practice as Structural & Geotechnical and is

in the same capacity till now. As a consultant, he was been involved in the
design and construction of numerous Residential apartments, Commercial
complexes and Industrial facilities. M/S Simplex Infra Structure Limited,
M/S C.R.Narayana Rao Consultants, M/S Doshi Housing Limited,
M/S Sundaram Clayton Limited, etc. are some of the prestigious Clients for
him. Numerous Multi Storied Residential apartments have been designed by
him, one of which has been taken for the study in the present thesis.
Presently he is a whole time consultant for M/S Simplex Infrastructure

Limited.
He is a member of Indian Geotechnical Society and International

Society of Soil Mechanics and Geotechnical Engineering. He is also a
member in Tc38 in the ISSMGE involved in soil structure interaction study.
He has travelled to Norway, Germany, Bangkok, Japan and Australia for
attending conferences. Some of the papers were presented by him personally.
ANNA UNIVERSITY- CHENNAI :: CHENNAI – 600 025.
PH.D. THESIS: EXAMINER’S EVALUTION REPORT.
Name of the Candidate : Mr. V.Balakumar

a) Reference code : 98 . 47 . 7 / ph. D/ AR6
Title of the Thesis: “EXPERIMENTAL STUDIES ON MODEL

PILED RAFTS OF SAND AND FIELD STUDY OF PROTO TYPE
BEHAVIOUR”
Chapter 2 Literature Review
2.1 A comprehensive literature review on piled raft has been conducted

covering laboratory model studies, prototype field monitoring as well as
analytical and numerical methods on piled raft. Limitations and inadequacies
of existing studies on piled raft have been provided. If there is any drawback
for this chapter, it is a general lack of critical evaluations of previous
research studies to elicit any conflicts or controversies among various
research findings regarding fundamental issues on piled raft.
Ans:-
The critical evaluation of the previous studies has been summed up at

the end of each section and the necessity for the present study has been
presented as conclusion to the chapter 2. Further, from the literature reviewed,
we are unable to find any controversial findings in the fundamental
understanding on the behavior of piled raft, except for the limitations pointed
out by the scholar in the chapter 2 of the thesis.
Chapter 3 Experimental Investigation on Piled Raft Models
3.1 It is suggested to amend the title of this chapter to “Experimental

Setup and Material Properties.
Ans:-
The Title of the chapter has been amended as suggested namely “
Experimental Setup and Material Properties”.
3.2 Since the study involves 1-g model tests without correct simulation of
realistic soil stress levels and the properties of the piles and rafts have not
been properly scaled (using Perspex material), the reference to the prototype
scale dimension appears superfluous.
Ans:-
We do agree that 1g model tests cannot simulate the insitu stress
conditions of real time situation. However 1g model technique is a well
accepted method in the geotechnical engineering research. To arrive at a
reasonable size of the model alone, a storage tank of 20m dia was considered
and a scale of 1 in 100 was followed. Even to day quite a number of research
works are carried out with 1g models since they enhance the understanding of
geotechnical problems despite centrifuge technique is considered to be a
better testing technique. The aim of the present study is not intended to
simulate any specific proto type foundation. The intention is to study the
effect of various parameters on the load sharing and the settlement reduction
behavior of the piled raft.
In 1g model tests, the soil used in prototype is normally used as test

medium. This does not satisfy the law of similarity as the grain size and the
ratio of particle size and the foundation size is not scaled, leading to scale
effect. This effect is significant in cohesionless material. But this effect has
been studied by many and proved insignificant if the ratio of the foundation
width or length to the D50 size of the sand is less than 30.
To fabricate the model of raft and the pile Perspex material was chosen
since some of the earlier researchers (Barden, 1962; Mondal et al., 1998;
Weisner and Brown, 1978) have used Perspex material for raft models in their
research works. Fabrication of the models with Perspex is easy and can be
made with great accuracy. The purpose and reason for choosing the material
is already given in Para 2 of page 67.
3.3 The Poisson ratio value of 0.33 for Perspex material adopted appears
too high. Please check.
Ans:-
The Poisson’s ratio of 0.33 adopted in this study is based on the
following references and they have been added in the reference list.
1. Daniel Thangaraj (2007) Interaction analysis of soil-raft-space

frame: parametric analysis and laboratory model studies. PhD
thesis submitted to Anna University
2. Model analysis of structures by T.P. Ganesan pp.123. Table 8.1.
The reported values are in the range of 0.33 to 0.35. In this study the
lowest value of 0.33 has been taken.
3.4 The numbering of some figures and plates has been mixed up. As an
example, Figure 3.7 is mentioned in page 70 but there appears no such figure.
Shear resistance is not mentioned in Figure 3.6, although it is stated on page
70. Plate 3.6 does not show any pile installation guide tube as claimed on
page 74. Please check the numbering of all the figures and plates in this
chapter.
Ans:-
In page 70 para 2 line 3 the word Figure 3.7 is corrected as Figure
3.6.In the above page, first para last line the word Figure 3.6 is amended as
Figure 3.5.In page 74, last line the word Plate 3.6 is amended as Plate 3.7
3.5 On page 78, it is stated that it takes 20 minutes for the settlement of
sand to complete. This is unusual as it is expected that settlement of sand
should be instantaneous. The candidate should present the settlement – time
history plot under a given applied to support the above statement.
Ans:-
It is agreed that the sand will settle immediately. However it will show
some creep with the applied load, particularly when the stress on the sand is
high. Moreover in the study, constant load- time rate method has been
adopted in the tests conducted on the piled rafts. In order to maintain the
constant load, the jack has to be operated frequently due to the settlement of
the plate. It took around 20 minutes for the settlement to stabilize under the
applied load. A typical load settlement curve is appended. Consequent to this
the sentence in the page 78 Para 1 line3 has been amended as “The loading
period was arrived based on trial tests conducted, and the maximum loading
period was decided as 30 minutes.”
Chapter 4 Results and Discussions

4.1 A large number of 1-g model tests were conducted to study the
behavior of piled raft foundation in sand, especially on the settlement
reduction and load sharing of piled raft. The 110 tests include plain and piled
rafts, free standing pile groups on loose, medium dense and dense sand.
Three shapes of piled raft (circular, square and rectangular) were considered
in the tests. The 1-g model tests cover a wide range of parameters and
meaningful conclusions have been drawn from the tests. However, the
drawback of 1-g model tests is that any conclusion drawn is behavioral only
and may not be applicable to large-scale prototype problems due to the heavy
dependence of sand behavior on actual stress levels. The candidate should
address this issue in the thesis and to highlight the limitations and
advantages/usefulness of 1-g model tests.
The small scale model studies facilitates having a close control on the
variables met in the practice and study their influence on the behavior of piled
raft. Although these results cannot be compared directly with full scale
behavior, they are very useful in providing an understanding on the behavior
pattern, as a guide to full scale performance when examined in conjunction
with the theoretical solutions developed. Similar view has been expressed by
many researchers such as Murray and Geddes (1989). In this study the
concentration was on the non dimensional parameters namely settlement
reduction ratio load sharing ratio as the piled raft behavior is influenced by
the settlement. While full scale tests are ideal, they are time consuming and
prohibitively costly. Also the boundary conditions for the tests may not be as
clear as that can be created in the laboratory. This can lead to the complexity
in the analyses of the observed data. Also full scale studies are not amenable
for parametric study.
In spite of the fact that stresses of the field conditions particularly in

sand cannot be modeled, the model tests have the following advantages:
1) It is possible to model very flexible raft

2) Parametric studies of the foundation are relatively easier and
thus more economical than the full scale tests.
3) Boundary conditions of the model are generally clear than those
found for the full scale tests.
The main disadvantages are :
1) The existence of the rigid side walls and rigid base may affect the
behavior of the model. However this was eliminated by providing the
tank of adequate size.
2) In 1g model tests, the soil used in prototype is normally used as test
medium. This does not satisfy the law of similarity as the grain size
and the ratio of particle size and the foundation size is not scaled,
leading to scale effect. This effect is significant in cohesion less
material. But this effect has been studied by many and proved
insignificant if the ratio of the foundation width or length to the D50 is
less than 30.
4.2 The title of this chapter may be amended to “Experimental Results and
Discussions”.
4.3 The suggestion has been accepted and the title has been amended as
“Experimental Results and Discussions”.
The legend in Figure 4.3 needs improvement. This suggestion is also

applicable to many other figures in this chapter.
The legend in Figure 4.3 and similar deficiencies have been amended.
4.4 On page 88, it is stated that beyond 20mm settlement, the plain raft
exhibits uncontrollable settlement even in dense sand. However, an
examination of the curve for dense sand shows that the curve is still
increasing but with a decreasing rate of increment.
In page 88 para 1 line 8: The word “uncontrollable” has been removed.

The sentence has been amended as “Further the plain raft offered less
resistance beyond 20mm settlement even in dense sand, which indicates
progressive deformation.
4.5 It is not clear (no explanation was given) why semi-log plot of the load
settlement curves will yield the three linear segments. The fitting of the three
linear segments in Fig. 4.5 appears arbitrary. If one removes the linear lines
in Fig. 4.5, the test data still exhibit trend of highly non-linear behavior
rather than three linear segments. The candidate should elaborate on this.
There are many ways to linearise a non linear load settlement curve
(log-log method, semi log method, rectangular hyperbolic method etc). In this
study semi- log method has been adopted to represent the non-linear response
of load settlement behavior. This representation helped to identify with a
reasonable level of accuracy at which the change in the slope occurred. As
expected the plot presents three slopes distinctly. They represent elastic,
elasto-plastic response with work hardening.
4.6 The symbols shown in Figure 4.10 clearly show that for the initial
linear portion, the stiffness of piled raft is higher than the plain raft
(consistent with Figures 4.7 to 4.9) As such, the blue line and red line should
not be drawn parallel for the initial linear portion of Figure 4.10.
The Figure 4.10 is corrected and the corrected curve is presented with
better clarity as Figure 4.10.
4.7 For Figure 4.13, since no strain gauges have been installed for the
piles in the piled raft system, only total loads applied can be measured. As
such, clear explanation should be given on how to derive the load shared by
the raft presented in Figure 4.13.
The load shared by the raft of piled raft has been obtained by deducting
the load taken by the pile group from the total load on the piled raft at a given
settlement.
4.8 On page 110, since both pile diameter d and raft thickness t have the
same effect of increasing the load resistance and stiffness of piled raft, it
appears unusual to use d/t as the normalized factor. Any rationale for using
this factor?
The function d/t dominates the response of piled raft. It is seen from
the chapter 2, table 2.1 that many of the piled rafts have the raft thickness
varying from 2.5m to 6.0 m indicating that many of the studies have been
done on piled rafts having higher raft thickness. In all these cases the pile
diameter also had been more than 1.0m.The primary aim of this study being
piled rafts with relatively thinner raft proportional to the pile diameter, the raft
thickness was kept close to the pile diameter value. It was also seen that in the
case of piled raft, raft thickness did not have any appreciable influence on the
load settlement behaviour.Hence keeping the raft thickness constant the pile
diameter was varied in the study and as expected the d/t ratio was found to
influence the settlement reduction ratio and load sharing ratio as shown in
fig.4.32 and 4.39 respectively.
4.9 For Fig. 4.31, SR is the ratio of settlement between piled raft and plain
raft at some given load. Which load is adopted for the plot of Fig. 3.1? What
is the meaning of 2mm and 20mm in the figure? This is not clear.
In the definition of SR the settlement of 2mm and 20mm corresponds to

2mm settlement of plain raft and 20mm settlement of plain raft. Further 2mm
settlement represents the critical settlement of the piled raft and 20 mm
corresponds to the assumed failure settlement of piled raft.
4.10 For Figure 4.34, how do you measure the load taken by the piles only?
The candidates should explain clearly as strain gauges were installed in the
piles to measure the load acting on the piles in the piled raft system.
Since the concentration was on the total load taken by the pile group
and not on the load taken by the individual piles, the piles were not provided
with strain gauges. The load on the pile group is obtained by deducting the
load taken by the unpiled raft (plain raft) for any given settlement from the
load taken by the piled raft for the same settlement from the test conducted
under identical bed conditions.
4.11 It is suggested to give a summary of the major findings arising from

the present 1-g model tests in Section 4.7.
From the small scale model studies it was seen that irrespective of the
shape of the raft the load settlement response of the plain raft and the piled
raft was the same. The response was found to be three phased, within the
settlement range of 10% of the least lateral dimension of the raft. It was found
that upto a settlement level of 1% of the least lateral dimension of the raft,
which is close to the critical settlement; the load settlement response was
linear. The load settlement response of the piled raft is characterized as multi
linear strain hardening behavior. Through the characteristic response, it was
seen that although the pile-raft-soil stiffness is higher than the raft soil
stiffness in all the three phases of the load settlement response, in the third
phase the pile raft soil stiffness was close to the raft soil stiffness indicating
that the contribution of pile group is reduced drastically and its share on the
load applied remained constant.
It was found that the pile length beyond 80% of the least lateral
dimension of the raft did not have any appreciable influence on the pile soil
stiffness. Similarly increasing the pile- raft area ratio beyond 6% was not
contributing to the stiffness of the piled raft appreciably. In all the cases
studied, the pile group shared major part of the load upto a settlement level of
1% of the least lateral dimension of the raft. As the settlement increased the
load sharing ratio decreased and remained constant beyond a settlement level
equal to 3% of the least lateral dimension of the raft. The performance of the
piled raft was found to be better from settlement reduction and load sharing
point of view in loose and medium dense sand till a settlement level of 3% of
the least lateral dimension of the raft. But at higher settlement level the
settlement reduction was found to be lesser in loose sand than in medium
dense sand. The parametric study has indicated that the pile length and
diameter have a predominant role on the behavior of the piled raft; the
number of piles in the form of pile raft area ratio beyond 5% to 6% does not
have any pronounced influence on the behavior.
Comparing the behavior of square and circular raft, under identical

conditions the square piled raft took more load than the circular piled raft as
the square piled raft could accommodate more number of piles for the same
area ratio. Significantly in either case at 20mm settlement of the plain raft the
settlement reduction ratio was found to be around 0.4 and the load sharing
ratio was around 0.35 to 0.38. It was also found that in either case the piles
have to be placed evenly on the total area of the raft for better performance.
Chapter 5 Numerical Analysis of 1 g Model tests of piled raft
5.1 On page 186, the statements that “a mesh of rectangular shaped

elements provides the best results” and that “ the mesh refinement does not
improve the accuracy of the results in any appreciable manner” are
controversial and may have been taken out of context from some references.
The point raised by the examiner is valid and well taken. In the page
186 para3, the lines from 4 to 6 is removed.
5.2 On page 186, the statement “the interface element need not be
considered between the pile and soil” may be fundamentally wrong,
especially for the case of full mobilization of the pile shaft friction where the
use of interface elements along the pile-soil interface becomes very critical in
the numerical simulation of pile-soil interaction. The candidate needs to
defend or revise this.
In the present study interface elements were used in the numerical

modeling of prototype piled raft. In the proto type piled raft analysis default
values were used. In order to validate the default values, a plain raft was
analyzed using the data given by Viladkar et al (1993). The table presented
below compares the contact pressure distribution below the footing for
various tangential stiffness values. From the table it can be inferred as the
tangential stiffness has negligible influence on the contact pressure except at
the edges.
Variation of contact pressure in kN/m2 below the strip footing Normal
stiffness = 106 kN/m3 (Viladkar et al 1993)
Tangential Tangential Tangential

Default
Dist., m stiffness stiffness stiffness
values
8000kN/m3 16000 kN/m3 24000 kN/m3
0.0 172 203 203 203
0.5 73 70 70 70
1.0 69 61 61 61
1.5 62 57 57 57
2.0 60 56 56 56
Similarly the settlement variation is also negligible. Hence the

influence of normal and tangential stiffness on the performance of strip
footing is insignificant when the loads are vertical. The earlier researchers
who have worked on the performance of raft also have confirmed that
roughness idealization through interface element is not advantageous if the
foundation structure is subject to vertical load alone (King and Chandrasekhar
1975; Viladkar et al 1983; Pots and Zdrakovick 2000).
In the present case of model tests the material used was Perspex. The
interfacial friction between the foundation model and the soil is not that of
real foundation. Normally concrete is used as material for foundation which
has the interfacial friction value in the order of 0.6 to 0.75, where φ is the
angle of frictional resistance of the soil. The friction value below the soil and
the Perspex was determined in the laboratory; the value was 12 degrees in
sand with φ is equal to 38degrees which is close to 50% of the wall friction
against concrete surface. Hence the frictional resistance between the soil and
the foundation is idealized through the perfect contact and not using interface
element. Further the effect of foundation roughness investigated by earlier
researchers (King and Chandrasekhar 1977; Mcway 1984; Kim and Yo 2005;
Kong et al 2007) and reported show that friction included through interface
element to represent the rough surface of interface and buried structures did
not provide any advantage when they are subjected to vertical load. Based on
the above observations the interface elements were not considered in the
numerical validation of small scale model studies.
Unlike the study of laterally loaded pile, in the present study the
provision of interface elements will increase the stiffness of the pile and hence
as indicated by Potts and Zdrakovick (2000) it was felt that the interface
element need not be considered between the pile and soil for the axial load
condition. However contact was taken as perfect contact between the pile and
soil as assumed by Reul (2000).
5.3 In general, Von-Mises yield criterion is not a good model for soil. As
such, the choice of MISO material in the numerical analysis may underscore
some fundamental disadvantage for the present numerical study. The
candidate should comment on this aspect.
The elasto-plastic behavior of the soil is idealized by any one of the

following elasto-plastic models:
A) Tresca
B) Von Mises
C) Mohr- Colulomb
D) Drucker-Prager
Among the models, Mohr-Colulomb model is widely used to represent

elastic-perfectly plastic material. The implicit assumption in the Mohr-
Coulomb model is that the yield and the strength are independent of the
intermediate principal stress. However the performance of the sand depends
upon the intermediate principal stress also. Though the Von Mises material is
essentially meant for undrained conditions the yield and the strength are
dependent on the intermediate principal stress. Hence the MISO model with
Von Mises yield criterion is considered suitable for idealizing the behavior of
sand. Moreover the tests were conducted on dry sand wherein the total and the
effective stresses are the same.
The load settlement response of the model piled raft on sand is

observed as non linear which has been characterized as multi linear hardening
behavior. Therefore the material behavior is idealized as a multilinear strain
hardening behavior (MISO). The constitutive relationship for the model is
included in the material library of the Ansys package. The input parameters
required for this model are stress- strain curve of the soil and the initial
tangent modulus. The MISO model uses Von Mises yield criterion, which
determines the stress level at which the yielding is initiated.
5.4 The upper corner of Fig. 5.3 is truncated.
The Figure 5.3 in page 195 has been amended.
5.5 In Table 5.8, the use of Poisson ratio of 0.3 for the piled raft appears
too high.
The Poisson’s ratio of 0.33 adopted in this study is based on the

following references and they have been added in the reference list.
3. Daniel Thangaraj (2007) Interaction analysis of soil-raft-space frame:

parametric analysis and laboratory model studies. PhD thesis submitted
to Anna University
4. Model analysis of structures by T.P. Ganesan pp.123. Table 8.1
The reported values are in the range of 0.33 to 0.35. In this study the
lowest value of 0.33 has been taken.
5.6 In section 5.6.2, it is stated that the “problem was taken as large
deformation problem” which is unusual and unnecessary for a problem with
a maximum settlement of only about 20mm. The candidate needs to defend or
revise this.
In the model study the magnitude of settlement up to which the test has
been conducted is 20mm which is 10% of the least lateral dimension of the
raft. The 20mm is equal to 10% of the least lateral dimension of the model
raft. This is very high in terms of real scale. In reality such high lateral
deformations represents large deformation. This has influence not only on the
material but also on geometry of soil elements. In order to account for
geometric non linearity of the soil elements the large deformation command is
invoked in the analysis.
5.7 On page 205, it is stated that the stress-strain values in Fig. 3.7 was
used. But where is Fig. 3.7??
Figure number is corrected as 3.6.in the page 205.
5.8 In Table 5.11, you list mainly dimensions only instead of “properties of
piled raft and sand”.
Table 5.11 is corrected as properties of piled raft.
It is suggested to give a full view of the 3D FEM mesh for Fig. 5.19
instead of the truncated view as it is now.
Figure 5.19 is amended and the amended figure is added

6.1 Chapter 6 Behavior of prototype Piled raft
The prototype field study is very interesting with many meaningful

field measurements and observations. The description of the project and
instrumentation is very clear. It provides a good case history of piled raft
foundation with relatively thin raft and relative small diameter piles.
6.2 On page 254, the empirical charts by Mori (1965) and Schultz (1966)
to drive the Young’s modulus of clay and sand should be given.
Accepted and the chart has been added in the appendix.
6.3 In the numerical simulation, the 3-m thick soil layer above the raft has
been ignored. A more realistic simulation may be to consider such 3-m top
soil as surcharge applied around the piled raft.
2. The raft is not backfilled. Also primarily the study is on load

settlement behavior and not on the bearing capacity behavior of the piled raft.
Also the strata are considered as normally consolidated and as such it would
have under gone the deformation under the existing surcharge load. Since the
bearing capacity is not the aim of the study, the surcharge is not taken into
account.
General
The entire thesis contains many minor typographical and grammatical

errors. The candidate should go through his thesis carefully to fix all these
errors. These are attended in the thesis.
The candidate may furnish satisfactory clarifications on the queries
raised in my report below during the oral examination
1. Page 69 – In section 3.3, the author highlights that uniformly graded

river sand was used. However, further on it is demonstrated in Table
3.3, the sand used is poorly graded. This needs to be explained!
Ans:-
The description of the sand used is corrected as uniformly graded sand.
2. Page 73-While preparing sand beds in the test tank, the thickness of
each layer seems to be much more than the diameter of the rammer
used. Will this effect the compaction of the layers to achieve
appropriate densities or not?
Ans:-
The thickness of the sand layer is 50mm. The ram diameter was 50mm
and the blow for compaction was transferred through a 8mm plate of size
75mmx75mm.Hence this will not affect the compaction of the layer.
3. Why the Differential settlement of raft (as indicated on page 228 and in
Figure 5.33) is observed in the numerical simulations (of 0.45mm) and
what may be the reason for such a behavior. Can it have been reduced
if interface elements have been considered in the simulations?
Ans:-
In the numerical analysis of square piled raft a differential settlement
of 0.45mm is observed. This is equal to 10% of the total settlement. In the
experimental study the differential settlement of the raft was zero, mainly
because the load was applied through a specially made loading system which
prevents the unequal settlement of the model raft. The loading system adopted
cannot be modeled in the numerical analysis. Moreover the numerical
analysis always show some order of differential settlement even if the
foundation element is rigid because of the accuracy with which the numerical
computation is performed.
4. Why the author has not considered the interface elements for the
interface of the rafts and the soils and the piles and soil while
simulating the model rafts? What would be the influence of the same
on the settlements observed? Further, author has considered interface
elements while simulating prototype piled raft model (see page 266).
Comment
Ans:-
In the present study interface elements were used in the numerical
modeling of prototype piled raft. In the proto type piled raft analysis default
values were used. In order to validate the default values, a plain raft was
analyzed using the data given by Viladkar et al (1993). The table presented
below compares the contact pressure distribution below the footing for
various tangential stiffness values. From the table it can be inferred as the
tangential stiffness has negligible influence on the contact pressure except at
the edges. Similarly the settlement variation is also negligible. Hence the
influence of normal and tangential stiffness on the performance of strip
footing is insignificant when the loads are vertical. The earlier researchers
who have worked on the performance of raft also have also confirmed the
roughness idealization through interface element is not advantageous if the
foundation structure is subject to vertical load alone (King and
Chandrasekhar, 1975; Viladkar et al 1983; Pots and Zdrakovick 2000).
In the case of model tests the material used was Perspex. The
interfacial friction between the foundation model and the soil is not that of
real foundation. Normally concrete is used as material for foundation which
has the interfacial friction value in the order of 0.6 to 0.75, where PHI is the
angle of frictional resistance of the soil. The friction value below the soil and
the Perspex determined in the laboratory; the value was 12 degrees in sand
with PHI is equal to 38degrees which is close to 50% of the wall friction
against concrete surface. Hence the frictional resistance between the soil and
the foundation is idealized through the perfect contact and not using interface
element. Further the effect of foundation roughness investigated by earlier
researchers (King and Chandrasekhar,1977; Mcway,1984; Kim and Yo,2005;
Kong etal., 2007) and reported that friction included through interface
element to represent the rough surface of interface and buried pipes did not
provide any advantage when they are subjected to vertical load.Based on the
above observations the interface elements were not considered in the
numerical validation of small scale model studies.
Unlike the study of laterally loaded pile, in the present study the
provision of interface elements will increase the stiffness of the pile and hence
as indicated by Potts and Zdrakovick (2000) it was felt that the interface
element need not be considered between the pile and soil. However contact
was taken as perfect contact between the pile and soil as assumed by Reul
(2000).
5. What are the interface properties adopted for the interface elements in
the prototype models and how the stiffness values were selected for the
elements Targe 170 and Conta 174 (refer page 266) Comment.
Ans:-
Default values have been used and the reasons are furnished in the
response of the foreign examiner.
6. Figure 4.6 (page 90) and Figure 4.11 ( page 96 also Figure 4.12 on
page 99) – Load settlement response for plain raft, piled raft for
medium dense sand looks different and also has small values of
applied loads. Is this correct or there is some problem? Kindly clarify
Ans:-
The curve in the Figure 4.6 is meant for the pile length of
120mm.Accordingly the legend of Figure 4.6 has been corrected. Text is also
edited
7) Similarly, Figure 4.54 (page 150 and Figure 4.55 on page 152)
Explain?
Ans:-
The scale in the Y-Axis in the Figure 4.55 shall be corrected as kNx10-2
Suggested typographical corrections in the thesis
1. Pg 70-line 2 and 7- “Angle of sharing” modify as “Angle of Shearing”

2. Page 70 – Figure 3.6 shall be corrected as Figure 3.5 (line no 4) and
Figure 3.7 as Figure 3.6 (line no 7)??
3. Page 14- line 1 – “nodded” shall be “noded”
4. Reference is missing in the list of references- page 194- Daniel
Thangaraj (2006)
5. Page 291 – reference no 20- Incomplete reference
The suggested typographical errors as given above have been carried

out and incorporated.

Experimental Studies of Model Piled Raft PDF

Загружено:

Сведения о документе

Исходное описание:

Оригинальное название

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Experimental Studies of Model Piled Raft PDF

Загружено:

Авторское право:

Доступные форматы

EXPERIMENTAL STUDIES OF MODEL PILED

RAFTS ON SAND AND FIELD STUDY OF

in fulfillment for the award of the degree

FACULTY OF CIVIL ENGINEERING

ANNA UNIVERSITY : CHENNAI 600 025

Certified that this thesis titled “EXPERIMENTAL STUDIES

OF MODEL PILED RAFTS ON SAND AND FIELD STUDY OF

PROTOTYPE BEHAVIOUR” is the bonafide work of

Mr. V. BALAKUMAR who carried out the research under my

supervision. Certified further that to the best of my knowledge the work

the basis of which a degree or award was conferred on an earlier

occasion of this or any other candidate.

Piled raft foundation is a fairly recent technique, in which the load

The parametric analyses carried out through 1g tests in this study

However the length of pile of 1B is preferable for effective performance of

The numerical analyses performed using ANSYS FEM Code

Concurrently a twelve storied structure designed and supported on

I wish to place on record my deep and profound sense of gratitude

The inspiring support and untiring efforts put in by

CHAPTER NO. TITLE PAGE NO.

CHAPTER NO. TITLE PAGE NO.

2.4.1.1 Equivalent pier method 38

3 EXPERIMENTAL SET UP AND MATERIAL

CHAPTER NO. TITLE PAGE NO.

3.7 LOAD TEST PROCEDURE 77

4 EXPERIMENTAL RESULTS AND DISCUSSION 83

CHAPTER NO. TITLE PAGE NO.

4.4.6.2 αpr versus pile lengths 131

CHAPTER NO. TITLE PAGE NO.

4.6.2 Settlement Reduction Behaviour 182

5 NUMERICAL ANALYSES OF 1g MODEL

CHAPTER NO. TITLE PAGE NO.

5.7.1 3d Nonlinear Analysis (MISO Model) 216

6 BEHAVIOUR OF PROTOTYPE PILED RAFT 251

CHAPTER NO. TITLE PAGE NO.

6.4 SUMMARY 283

APPENDIX 1: TYPICAL LCPT (LIGHT CONE

APPENDIX 2: MORI’S CHART 293

LIST OF PUBLICATIONS 303

TABLE NO. TITLE PAGE NO.

2.1 Details of piled raft foundation of the buildings

TABLE NO. TITLE PAGE NO.

4.6 The variation of stiffness (N/mm) (D = 200mm,

TABLE NO. TITLE PAGE NO.

4.18 Effect of pile diameter on load and stiffness

TABLE NO. TITLE PAGE NO.

5.6 Variation between ANSYS and closed form

FIGURE NO. TITLE PAGE NO.

1.1 General concept of piled raft 4

FIGURE NO. TITLE PAGE NO.

2.15 Influence of pile-raft interaction on contact pressure

FIGURE NO. TITLE PAGE NO,

4.10 Characteristic response of plain raft and piled raft

FIGURE NO. TITLE PAGE NO.

4.27 Load- settlement response of piled raft for various

FIGURE NO. TITLE PAGE NO.

4.45 Load-settlement response of plain square raft of

FIGURE NO. TITLE PAGE NO.

4.59 Non-dimensional plot for piled raft with different

FIGURE NO. TITLE PAGE NO.

FIGURE NO TITLE PAGE NO

5.6 Differential displacement ratio v/s Pile depth 203