Ashour 16 PH D

UNIVERSITY OF BIRMINGHAM
THE RESPONSE OF STONE COLUMNS

UNDER CYCLIC LOADING
By
SAMIR ASHOUR
A Thesis submitted to
The University of Birmingham
For the degree of
DOCTOR OF PHILOSOPHY
School of CIVIL ENGINEERING

College of ENGINEERING AND PHYSICAL SCIENCE
The University of Birmingham
September 2015
University of Birmingham Research Archive
e-theses repository
This unpublished thesis/dissertation is copyright of the author and/or third

parties. The intellectual property rights of the author or third parties in respect
of this work are as defined by The Copyright Designs and Patents Act 1988 or
as modified by any successor legislation.
Any use made of information contained in this thesis/dissertation must be in

accordance with that legislation and must be properly acknowledged. Further
distribution or reproduction in any format is prohibited without the permission
of the copyright holder.
Samir Ashour Abstract
ABSTRACT
Soft clay soils in railway track can be problematic as, unless they are treated, they can result
in increased deformation of track. This will inheritably mean lower track speed. A number of
techniques including stone columns are available for improving strength of weak soils.
The use of stone columns, for improving both bearing capacity and settlement is well
rehearsed for static loading; little is understood about their response, when subjected to cyclic
loading. This study is focused on investigating the behaviour of stone columns when
subjected to cyclic loading as in railway tracks based on a laboratory scale investigation.
A series of monotonic and cyclic loading conditions were undertaken on two laboratory
models undrained triaxial (diameter 100 mm, and 200 mm height) and large scale model
(diameter 300 mm, and 300 mm height). Tests were conducted on both soft soils (no column)
and soil/ stone column composite. All tests were performed on normally consolidated
specimens of soft clay (undrained shear strength of ≈ 12 kPa) and for treated soils they were
reinforced with 28 mm diameter stone columns. Three cyclic stress (50, 60 and 70 kPa) on
subgrade level and three loading frequencies (0.5, 1 and 3 Hz) simulating different train
speeds (35, 70 and 225 km/hr) were used to study the performance of both soft soils (with and
without column). The effect of both cyclic stresses and loading frequency on the permanent
deformation, soils stiffness and pore water pressure generation were investigated.
Generally, soft clay bed reinforced with stone column showed a significant improvement in
terms of load/deformation characteristics; in monotonic loading condition, there was
approximately 30% increase in failure load of soil with the stone column compared to that soil
only.
i
Samir Ashour Abstract
It was found that threshold dynamic stress of soil for cyclic loading increased from 50 kPa for
soil only to 60 kPa for soil with stone column. This is equivalent to CSR of 0.7.
Changes in frequencies from 0.5 to 3 Hz did not significantly influence the permanent strain
of reinforced soil, but these changes do affect the stiffness. Stiffness of the soil with the stone
column was about 25% higher at 3.0 Hz compared to that at 0.5 Hz.
Stone columns also helped reduce pore water pressure build up under cyclic loading by
providing a drainage path. This resulted in an increase in cyclic stress ratio from 0.6 to 0.7. It
also decreases the permanent deformation by about 70% in most tests, when compared to
situation without stone column.
ii
Samir Ashour Dedication
DEDICATION
This thesis work is dedicated to my parents, who have always loved me unconditionally and
whose good examples have taught me to work hard for the things that I aspire to achieve.
Also to my brother who always has been there for me.
This work is also dedicated to my wife and children, who have been a constant source of
support and encouragement during the challenges of life. I am truly thankful for having you in
my life.
And above all
This thesis is dedicated to the sake of Allah, my Creator and my Master, My great teacher and
messenger, Mohammed (May Allah bless and grant him)
iii
Samir Ashour Acknowledgments
ACKNOWLEDGMENTS
I would like to acknowledge gratefully for the assistance received from supervisors Dr
Gurmel Ghataora and Professor Ian Jefferson for their constant supervision, guidance and
encouragement throughout the research.
I would also express my sincere thanks to Lap technicians Mr Sebastian Ballard and Mr Jim
White for their help and support.
I am very much grateful to Libyan Ministry of Higher Education for granting the scholarship
for this research.
Finally, I thank my family for their patience and understanding to make this piece work
possible.
iv
Samir Ashour Table of Contents
TABLE OF CONTENTS
ABSTRACT ............................................................................................................................................. i
DEDICATION ....................................................................................................................................... iii
ACKNOWLEDGMENTS...................................................................................................................... iv
TABLE OF CONTENTS ........................................................................................................................ v
LIST OF FIGURES.............................................................................................................................. xiv
LIST OF TABLES ............................................................................................................................... xiv
LIST OF ABBREVIATION ................................................................................................................. xv
CHAPTER 1............................................................................................................................................ 1
1 INTRODUCTION ........................................................................................................................... 1
1.1 Background ...................................................................................................................... 1
1.2 Research problem ............................................................................................................ 3
1.3 Research Aim and Objectives.......................................................................................... 4
1.4 Outline of Thesis ............................................................................................................. 5
2 CHAPTER 2.................................................................................................................................... 6
2 LITERATURE REVIEW ................................................................................................................ 6
2.1 Introduction ..................................................................................................................... 6
2.2 Introduction to Ground Improvement.............................................................................. 6
2.3 Stone Column Technique: Historical Development ........................................................ 8
2.4 Stone Column Construction Methods ............................................................................. 8
2.4.1 Vibro Equipment (Vibro-float) ....................................................................................... 9

2.4.2 Vibro Replacement (Wet Method) ................................................................................ 11
2.4.3 Vibro Displacement (Dry Method) ............................................................................... 13
2.4.4 Monitoring, Quality Control and Assessment of Stone Columns Performance ............ 15
2.4.5 Material Used for Stone Columns Constructions .......................................................... 16
2.5 Application and Limitation of Stone Column ............................................................... 18
2.5.1 Limitations .................................................................................................................... 19
v
2.6 Design Approaches of Stone Column ........................................................................... 19
2.6.1 Basic Design Parameters ............................................................................................... 20

2.6.2 Consolidation and Settlement ........................................................................................ 25
2.6.3 Bearing Capacity Prediction Methods ........................................................................... 35
2.7 Mechanism and Performance of Stone Columns .......................................................... 39
2.7.1 Stone Columns Investigation under Monotonic Loading Condition ............................. 39

2.7.2 Dynamic Loading Condition ......................................................................................... 47
2.8 Concluding Remarks ..................................................................................................... 55
3 CHAPTER 3.................................................................................................................................. 57
3. METHODOLOGY ........................................................................................................................ 57
3.1 Introduction ................................................................................................................... 57
3.2 Materials ........................................................................................................................ 57
3.3 Physical Properties of the Model Stone Column and Surrounding Soil Material ......... 59
3.3.1 Index Properties............................................................................................................. 59

3.3.2 Particle Size Distribution .............................................................................................. 59
3.3.3 Specific Gravity............................................................................................................. 60
3.3.4 Dry Density/Moisture Content Relationship of Kaolin Clay and Crushed Aggregate .. 60
3.3.5 Shear Strength ............................................................................................................... 61
3.3.6 Compressibility Properties ............................................................................................ 65
3.4 Test Requirement and Apparatus Design ...................................................................... 69
3.4.1 Scale Effect ................................................................................................................... 69

3.4.2 Geometrical Dimension of Column .............................................................................. 71
3.4.3 General Test Requirements ........................................................................................... 72
3.5 Sample Preparation ........................................................................................................ 73
3.6 Test Apparatus ............................................................................................................... 75
3.6.1 Cyclic Load Frame ........................................................................................................ 75

3.6.2 Instrumentation.............................................................................................................. 76
3.7 Test Variables ................................................................................................................ 76
vi
3.7.1 Loading Frequency (f) ................................................................................................... 76

3.7.2 Amplitude and Dynamic Stress ..................................................................................... 78
3.8 Experimental Procedure ................................................................................................ 81
3.8.1 Clay Bed Preparation .................................................................................................... 81

3.8.2 Column Installation ....................................................................................................... 88
3.8.3 Loading Test .................................................................................................................. 90
3.8.4 Preloading Investigation ................................................................................................ 92
4 CHAPTER 4.................................................................................................................................. 96
4 MONOTONIC LOADING RESULTS AND DISCUSSION ....................................................... 96
4.1 Introduction ................................................................................................................... 96
4.2 Model I (Triaxial) .......................................................................................................... 97
4.2.1 Saturation ...................................................................................................................... 97

4.2.2 Consolidation ................................................................................................................ 98
4.2.3 Stress - Strain Behaviour ............................................................................................... 99
4.2.4 Effect of Strain Rate .................................................................................................... 104
4.2.5 Effect of Column Density ........................................................................................... 113
4.3 Model II (Large Scale Test: Foundation Test) ............................................................ 117
4.3.1 Load - Displacement Relationship .............................................................................. 117

4.3.2 Change in Pore Water Pressure ................................................................................... 125
4.3.3 Post Testing Investigation ........................................................................................... 127
4.4 Concluding Remarks ................................................................................................... 129
5. CHAPTER 5................................................................................................................................ 131

5. CYCLIC LOADING RESULTS AND DISCUSSION ............................................................... 131
5.1 Introduction ................................................................................................................. 131
5.2 Model I (Triaxial Tests) ............................................................................................... 132
5.2.1 Permanent Strain during Cyclic Loading .................................................................... 132

5.2.2 Cyclic Deformation ..................................................................................................... 140
5.2.3 Resilient Modulus ....................................................................................................... 142
5.2.4 Pore Water Pressure Response .................................................................................... 145
5.2.5 Deformation Pattern .................................................................................................... 148
vii
5.3 Model II Test Results (Large Scale Test: Foundation)................................................ 150
5.3.1 Permanent Strain during Cyclic Loading .................................................................... 150

5.3.2 Pore Water Pressure Changes...................................................................................... 153
5.3.3 Soil stiffness ................................................................................................................ 154
5.4 Implication for Stone Column Performance ................................................................ 156
5.5 Concluding Remarks ................................................................................................... 157
5. CHAPTER 6................................................................................................................................ 159

6 CONCLUSION ........................................................................................................................... 159
6.1 General......................................................................................................................... 159
6.2 Monotonic Loading ..................................................................................................... 160
6.2.1 General ........................................................................................................................ 160

6.2.2 Effect of Strain Rate .................................................................................................... 160
6.2.3 Effect of Column Material Density ............................................................................. 161
6.3 Cyclic Loading ............................................................................................................ 161
6.3.1 Effect of Stress Level .................................................................................................. 161

6.3.2 Effect of Frequency ..................................................................................................... 162
6.4 Recommendations for Future Research Works ........................................................... 163
REFERENCE ...................................................................................................................................... 164

APPENDIX A ..................................................................................................................................... 172
APPENDIX B ..................................................................................................................................... 186
APPENDIX C ..................................................................................................................................... 190
APPENDIX D ..................................................................................................................................... 210
APPENDIX F ...................................................................................................................................... 231
APPENDIX G ..................................................................................................................................... 235
viii
Samir Ashour List of Figures
LIST OF FIGURES
Figure ‎2.1 Deep vibrator details (Moseley and Kirsch, 2004) ................................................. 11
Figure ‎2.2 Vibro-replacement method (Keller GmbH, 2005) .................................................. 12
Figure ‎2.3 Vibro-displacement (Keller GmbH, 2005) ............................................................. 13
Figure ‎2.4 Dry bottom feed (Keller GmbH, 2005) ................................................................... 14
Figure ‎2.5 Unit cell concept (Barksdale and Bachus, 1983) .................................................... 21
Figure ‎2.6 Stone Column arrangements in (a) Triangular (b) Square (Balaam & Booker, 1981)
.................................................................................................................................................. 22
Figure ‎2.7 Settlement estimation conditions ............................................................................ 25
Figure ‎2.8 Settlement diagram for stone column in uniform soft clay (Greenwood, 1970) .... 27
Figure ‎2.9 Improvement factor (Priebe, 1995) ......................................................................... 32
Figure ‎2.10 Load – settlement curves for stone column (Balaam et al., 1977) ........................ 34
Figure ‎2.11 Vesic’s cylindrical cavity expansion factors (Barksdale and Bachus 1983)......... 37
Figure ‎2.12 Failure mechanism of single stone column (Barksdale and Bachus, 1983) .......... 40
Figure ‎2.13 Photographs of sand columns beneath circular footing at beginning, middle and
end of foundation loading process: (a) TS-01, 150 mm; (b) TS-02, 250 mm ......................... 42
Figure ‎2.14 Effect of undrained shear strength of the surrounding soil on the ultimate bearing
capacity of the treated ground .................................................................................................. 45
Figure ‎2.15 Effect area replacement ratio on the ultimate bearing capacity of the treated
ground ....................................................................................................................................... 46
Figure ‎2.16 Classification of dynamic problems (Ishihara; 1996) ........................................... 49
Figure ‎2.17 Design chart for vibro replacement (Priebe; 1998) ............................................... 51
Figure ‎2.18 Cyclic loading sequences for railway embankment (Kempfert et al., 1990). ....... 53
Figure ‎3.1 Typical particle size distribution of Kaolin clay and crushed basalt ...................... 60
xiv
Figure ‎3.2 Dry density-water content relationship for Kaolin clay .......................................... 61
Figure ‎3.3 Relationship between undrained shear strength and water content for Kaolin clay62
Figure ‎3.4 The undrained shear strength – liquidity index relationship ................................... 63
Figure ‎3.5 Typical stress-strain, direct shear box test (a) wet condition and (b) dry condition64
Figure ‎3.6 Shear stress verses normal stress at peak ................................................................ 65
Figure ‎3.7 Consolidation curve square root of time step 50 to 100 kPa................................... 67
Figure ‎3.8 Void ratio / pressure increment relationship ........................................................... 68
Figure ‎3.9 Cofficients of consolidation and compressibilty and stress relationship ................ 68
Figure 3.10 Model I consolidation process for 100 mm dia. Samples .................................... 74
Figure 3.11 Model II consolidation chamber for preparing 300 mm dia. Samples ................. 74
Figure ‎3.12 Testing equipment ................................................................................................. 75
Figure ‎3.13 Dynamic measurement of subgrade stress (Yoo and Selig, 1979)........................ 79
Figure ‎3.14 Dynamic stresses mesurments at subgrade level (under 300mm of ballast) ......... 80
Figure ‎3.15 Typical settlement /√time relationship for 100 mm dia. triaxial sample .............. 83
Figure ‎3.16 Settlement and total vertical stress relationship for 100 mm dia. triaxial sample. 84
Figure ‎3.17 Void ratio/ log σv relationship for 100 mm dia. triaxial sample ........................... 84
Figure ‎3.18 Typical settlement / time relationship for 300 mm specimens ............................. 86
Figure ‎3.19 Settlement and total vertical stress relationship for 300 mm specimes ................ 86
Figure ‎3.20 Variation of excess pore water pressure at PPT1, PPT2, and PPT 3 positions
during consolidation ................................................................................................................. 87
Figure ‎3.21 Sample preparation for Model I ............................................................................ 89
Figure ‎3.22 Sample prepartion for Model II ............................................................................ 89
Figure ‎3.23 Cyclic stress state .................................................................................................. 90
Figure ‎3.24 Large test set up .................................................................................................... 91
xv
Figure ‎3.25 Variation in water content before testing for both Models (I and II) .................... 93
Figure ‎3.26 Shear strength determinations in test bed (pilot stage) ......................................... 94
Figure ‎4.1 Consolidation characteristics .................................................................................. 99
Figure ‎4.2 Stress strain behaviour of soft soil (no column) and soil/stone column composite.
................................................................................................................................................ 101
Figure ‎4.3 Improvement ratio versus area replacement ratio ................................................. 101
Figure ‎4.4: Excess pore water pressure of soil (no column) and soil/ stone column composite.
................................................................................................................................................ 102
Figure ‎4.5 Shape of failure under static loading condition..................................................... 104
Figure ‎4.6 Typical stress-strain curves: (a) soil (no column); (b) soil/stone column composite
................................................................................................................................................ 106
Figure ‎4.7(b) Typical change in pore water presures-strain curves for soil/ stone column
composit ................................................................................................................................. 108
Figure ‎4.8 Normalised excess pore water pressure – strain rate relasionship ........................ 108
Figure ‎4.9 Soil modulus of both soils only and soil/ stone column composit ........................ 110
Figure ‎4.10 The subgrade modulus of reaction (𝑘𝑠) for both soils (with no column and soil/
stone column composit) at a) at settlement of 1.25 mm; b) at settlement of 6 mm; and c) at
settlement of 10 mm ............................................................................................................... 112
Figure ‎4.11 The stress-strain for different column densities .................................................. 114
Figure ‎4.12 Improvement ratio versus the relative density .................................................... 115
Figure ‎4.13 Change of pore pressure with column density .................................................... 115
Figure ‎4.14 normalised deviator stress against column density ............................................. 116
Figure ‎4.15 stress-strain relationship ...................................................................................... 118
Figure ‎4.16 Scatter plot of all the variables ............................................................................ 121
xvi
Figure ‎4.17 Relationship between qratio and As in terms of ∅ ................................................. 123
Figure ‎4.18 Typical results for changes in pore water pressure for soil (no column) specimens
................................................................................................................................................ 126
Figure ‎4.19 Typical results for changes in pore water pressure for soil/stone column
specimens ............................................................................................................................... 126
Figure ‎4.20 Hand vane shear test results before and after testing .......................................... 127
Figure ‎4.21 Water content variation before and after testing ................................................. 128
Figure ‎4.22 Deformation pattern ............................................................................................ 129
Figure ‎5.1 Pemanent axial strain and number of cycles relationship: at 0.5Hz loading
frequency ................................................................................................................................ 133
Figure ‎5.2 Pemanent axial strain and number of cycles relationship: at 1 Hz loading frequency
................................................................................................................................................ 133
................................................................................................................................................ 134
Figure ‎5.4 Effect of loading application and cyclic stress leve at frequency of 1 Hz ............ 136
Figure ‎5.5 Effect of loading frequency and number of cycles under cyclic stress of 50 kPa. 137
Figure ‎5.6 Cyclic deformation during cyclic triaxial test: at 0.5 Hz loading frequency ........ 140
Figure ‎5.7 Cyclic deformation during cyclic triaxial test: at 1 Hz loading frequency ........... 140
Figure ‎5.8 Cyclic deformation during cyclic triaxial test: at 3 Hz loading frequency ........... 141
Figure ‎5.9 Variation of resilient modulus with number of cycles at loading frequncy of 0.5 Hz
................................................................................................................................................ 142
Figure ‎5.10 Variation of resilient modulus with number of cycles at loading frequncy of 1 Hz
................................................................................................................................................ 143
xvii
................................................................................................................................................ 143
Figure ‎5.12 Variation of resilient modulus with respect to dynamic stress and loading
frequency ................................................................................................................................ 144
Figure ‎5.13 Accumulation of excess pore water pressure at loading frequency of 0.5 Hz .... 145
Figure ‎5.14 Accumulation of excess pore water pressure at loading frequency of 1Hz ........ 145
Figure ‎5.15 Accumulation of excess pore water pressure at loading frequency of 3Hz ........ 146
Figure ‎5.16 Effect of cyclic stress on pore water pressure ..................................................... 147
Figure ‎5.17 Deformed shape for soil only specimes .............................................................. 149
Figure ‎5.18 Deformed shape for the soil/column specimen ................................................... 149
Figure ‎5.19 Permanent strain results/ number of cycles ......................................................... 151
Figure ‎5.20 Deformation pattern after cyclic loading ............................................................ 153
Figure ‎5.21 Pore water pressure measurment at the centre of the specime ............................ 153
Figure ‎5.22 Pore water pressure measurment at 50 mm from the centre of the specime ....... 154
Figure ‎5.23 Stifness results..................................................................................................... 156
xviii
Samir Ashour List of Tables
LIST OF TABLES
Table ‎2.1 Summary of stone column installation methods ........................................................ 9
Table ‎2.2 Equivalent unit cell diameter (adopted from Balaam & Booker, (1981)) ................ 21
Table ‎2.3 Methods of estimation of settlement ........................................................................ 26
Table ‎2.4 Comparison between predictions of tank’s settlement (Ellouze et al., 2010) .......... 33
Table ‎2.5 Estimation of ultimate bearing capacity (Bergado et al., 1991) ............................... 36
Table ‎3.1 Materials used in previous studies............................................................................ 58
Table ‎4.1 Test programme ........................................................................................................ 96
Table ‎4.2 Typical triaxial saturation data ................................................................................. 98
Table ‎4.3 Measured and calculated vertical stress obtained from different studies ............... 103
Table ‎4.4 Modulus of subgrade reaction ................................................................................ 111
Table ‎4.5 Published database ................................................................................................. 120
Table ‎4.6 Deviance, BIC, R-Squared and Adjusted R2 values .............................................. 122
Table ‎5.1 Summary varibles investigated using cyclic triaxial test ....................................... 132
Table ‎5.2 Material parameters (after Li and Selig (1996)) ..................................................... 139
Table ‎5.3 Model II test program ............................................................................................. 150
xiv
Samir Ashour List of Abbreviation
LIST OF ABBREVIATION
As Area replacement ratio
Cc Compression index
cg Constant dependent upon the pattern of stone columns
CSR Cyclic stress ratio
cu Undrained shear strength of the soil
cv Coefficient of consolidation
dc Column diameter
De Equivalent diameter
dg Average particle diameter of stone column material
E Deformation modulus of soil
e0 Initial void ratio
f Frequency
Fc and Fq Cavity expansion factors
Gc Elastic shear modulus of the clay
Gs Specific gravity
Gs Elastic shear modulus of the column material
h Depth of ballast
Ir Rigidity index
Kac Coefficient of active pressure of the column
KP Coefficient of the passive pressure of the column
Ks The modulus of subgrade reaction
xv
L Column length
Lc Critical column length
L.L Liquid limit
LI Liquidity index
Mr Resilient modulus
mv Coefficient of compressibility
n Stress concentration factor
N Number of cycles
Nc Bearing capacity factor
p Mean stress
PC Subgrade pressure
Pm Applied stress on ballast
Pult Ultimate column load
q Deviator stress
S Spacing between the columns
Tv Time factor of 90% consolidation
u Pore water pressure
Δu Change in pore water pressure
w Water content
Δe Change in void ratio
σ3’ Effective confining pressure
σc Stress in the surrounding soil
σv Vertical stress
xvi
σr Lateral confining pressure
σs Stress in the stone column.
ϒc Unit weight of the clay
ϒs Unit weight of the column material

μc
Stress ratios in the soil
μs Stress ratios stone column
φ Angle of internal friction
𝐶̅ Average shaft cohesion
εp Permanent strain
xvii
Samir Ashour Chapter 1: Introduction
CHAPTER 1
1 INTRODUCTION
1.1 Background
Soil improvement methods have received much attention over the last few decades, as a
result of shortage of good quality land for development, together with the need to develop
transport infrastructure in areas which were previously regarded as being unsuitable. These
often included soft soil deposits. In many instances, the proposed structures on improved/
stabilised soft soils are considered to be viable on economic grounds due to the high cost of
virgin land and the environmental need to site infrastructure away from developed areas.
Soft soil deposits are characterised as having low bearing capacity, and high compressibility
leading to potential instability and large settlement. Therefore, in order to reduce these
problems, either the structure or the underlying soils or both needs to be modified. Most often,
it is cheaper to undertake ground improvement than to modify the structure. Although in some
cases, this is a possible solution.
A range of ground improvement methods can be used. They can be categorised as:
densification, consolidation, reinforcement, chemical stabilization, thermal stabilization, and
biotechnical stabilization. Sometimes it may be more efficient to use more than one type of
treatment.
Stone columns are constructed by using the same equipment as vibro-compaction. They
are considered as a densification method and also can act as a vertical drain that can lead to
speedier the consolidation progress of the soil. It is also considered as a ground reinforcement
1
technique due to the granular nature of the column material. Granular columns are installed
into the soft soil by means of compacting gravel or crushed rock into the cylindrical void
created by a vibrating poker. This technique has been used extensively over the last few
decades in many parts of the world. The design and construction techniques have been
improved to an extent that stone columns could be used in some cases as an alternative to the
traditional foundations such as piling (McCabe et al., 2009; Barksdale & Bachus, 1983;
Greenwood & Kirsch, 1984; Mitchell and Huber 1985; Priebe, 1995; Sivakumar et al., 2010).
Many full-scale, laboratory experimental and analytical investigations have been undertaken,
which have led to improvement of the stone column technique and methods of analysis,
providing better prediction for both bearing capacity and settlement (Bergado and Lam, 1987;
Greenwood, 1991). These studies demonstrate that the behaviour of the stone columns was
affected considerably by a number of parameters, including: column length to diameter ratio;
the area replacement ratio; column spacing; column and surrounding soil stiffness; the stress
ratio of both the column and surrounding soil; and the method of installation. Also they have
shown that columns can fail by bulging, bending, punching and shearing, depending on the
length of the column and the method of applying the loads on top of the column.
Despite of the fact that, the previous investigations have provided a wide understanding on
using stone columns to improve the properties of soft soils, these investigations were focused
the behaviour of stone columns under the application of static and monotonic loading and
there was very limited information on their behaviour when subjected to cyclic loading. The
key difference in this study and the previous ones is the mode of loading, where the applied
loads will vary with time. Thus this research will provide a better understanding of the
behaviour of stabilised stone columns under static and dynamic loading conditions. This
2
includes evaluation of deformation under loading and mechanisms of stress transfer in both
treated and untreated soil.
1.2 Research problem
There is increasing need for higher performance track substructure systems (ballast, sub-
ballast and subgrade) in order to provide constant and uniform support for the sleepers and the
rail allowing them to cope with increased demand for high speed trains. Without any
upgrades, tracks built over soft ground can be problematic, leading to increase in maintenance
in order to maintain line and level (Raju, 2003).
Many researchers (Seed et al., 1955; Brown et al., 1975; Li and Selig, 1996; Miller et al.,
2000; Li-Zhong Wang et al., 2011) have studied the effect of cyclic loading on the
compressibility and strength characterisation of soft soil. They demonstrated that such soils
have high compressibility and low bearing capacity. Therefore, under cyclic stresses above
the threshold stress ratio (i.e. ratio between the dynamic stresses to the soils static strength,
which is typically between 0.4 and 0.65 depending on the soil type and conditions), it is
expected to have a high residual settlement resulting in a reduction in strength. Hence leading
to a reduction in track speed limit.
In order to overcome this problem, it is generally better (i.e cost-effective) to improve the
ground rather than import materials of suitable quality to replace poor soils. This has
additional environmental benefits. A number of treatment options including the construction
of vibro stone columns can be considered for improving subgrade soil (Raju, 2003; Fatahi et
al., 2012).
Vibro stone columns have been widely used to reinforce and stabilise railway tracks built on
soft soils to control the permanent deformation and the generation of excess pore pressure
3
(Abdullah et al., 2009; Fatahi et al., 2012). However, studies on the dynamic performance of
stone columns are mostly limited to liquefaction mitigation potential in silty soils (Munfakh,
1984; Priebe, 1995; Munfakh, 2003; Rollins et al., 2009). Additionally, very limited research
has been carried out on the performance of stone columns subjected to repeated vertical
loading. Therefore, aspects such as identifying failure mechanisms and quantifying the
amount of settlement, especially at low replacement ratio (under 10%) require investigation.
For this study, vibro stone column were modelled in both triaxial tests and small scale tank
tests. Stone columns were constructed in soft soils (cu <15 kPa) and loaded to failure under
both monotonic and dynamic loading. The main objective of the testing program was to
examine the effect of loading rate, loading frequency, and cyclic stress ratio on the behaviour
of both soft soils and stone column reinforced soil. Thus, allowing for a better understanding
of the response of stone column under dynamic loading application can be achieved.
1.3 Research Aim and Objectives
This research aim is to assess the behaviour of stone column subjected to cyclic loading as a
precursor to their application as a railway subgrade improvement technique. This can be
approached by developing the appropriate loading frequency and dynamic stresses on both
treated and untreated soft soils in order to investigate the stiffness, permanent deformation
and pore water pressures. In order to achieve the above aim, the following objectives were set:
i. Develop a suitable physical, laboratory scale model
ii. Examine the effect of loading strain rate during the monotonic conditions on the
degree of improvement of soft ground;
iii. Examine the effect of relative density of the column material on column bearing
capacity;
4
iv. Investigate the effect of loading frequency on the behaviour of stone column;
v. Investigate the effect of dynamic stress level on the behaviour of stone column.
1.4 Outline of Thesis
Brief descriptions of the remaining chapters of this thesis are outlined below:
Chapter 2 provides a brief background of stone column foundations and reviews previous
research work relevant to the subject.
The properties of the materials used; the design and manufacture of testing apparatus and
instrumentation used in this laboratory base study; and the specimen preparation and general
procedures used for the model testing are presented in Chapter 3.
The monotonic testing programme and results of the effect of varying strain rate and column
density are described in Chapter 4. Results from current study are analysed and compared
with other research findings from both laboratory tests and field studies. In addition to this,
the failure mechanisms for a single stone column reinforced foundation in both models
(triaxial and large scale) are discussed.
Chapter 5 shows results and discussion of tests on stone columns, subjected to cyclic loading.
(i.e the effect of both frequency and cyclic stress level) in order to provide a further
understanding to the behaviour of stone columns.
Finally, both conclusions and recommendations are presented in Chapter 6.
5
Samir Ashour Chapter 2: Literature Review
2 CHAPTER 2
2 LITERATURE REVIEW
2.1 Introduction
Vibro stone column method has been widely used to improve soft soils over the last four
decades. This chapter reviews and the design concepts, installation methods, failure
mechanisms, current practice, and relevant experimental procedures.
2.2 Introduction to Ground Improvement
Geotechnical engineers are often faced with the challenges of building on poor ground and
have to match the level of ground improvement with project requirements. Apart from
abandoning the project, Mitchell and Jardine, (2002) suggested four alternatives:
- replace poor ground with more suitable material;
- bypass the area of poor ground laterally by relocating the facility, or vertically by
using piles or deep foundations;
- redesign the structure to meet the ground limitations; or
- modify the natural condition of the poor ground to meet the project requirements.
The rising cost of land and increasing awareness of the impact that construction has on the
environment have been major contributors to the increasing use of ground improvement
techniques.
Ground improvement techniques generally aim to modify some of the soil characteristics in
order to meet the requirements of a construction project. These improvements include
increasing the density and shear strength in order to: improve bearing capacity; reduce soil
6
compressibility; influence permeability in order to control ground water flow or to increase
the rate of consolidation; and improve the homogeneity of soil. Therefore, ground
improvement can be defined as “the controlled alteration of the state, nature or mass
behaviour of ground materials in order to achieve an intended satisfactory response to existing
or projected environmental and engineering actions” (Mitchell and Jardine, 2002).
A wide range of ground treatments are available. Different classifications are proposed
depending on: equipment used; required period of treatment (temporary or permanent), or the
purpose of use (Munfakh, 1997; Van Impe et al., 1997; Mitchell and Jardine, 2002; Munfakh,
2003). Munfakh (1997) divides ground improvement methods into the following categories:
 Densification methods include vibro-compaction; dynamic compaction; blasting;

and compaction grouting.
 Consolidation method: preloading and vertical drains.
 Reinforcement methods include mechanically stabilised backfill; soil nailing;

element wall; micro-piles; and stone columns.
 Chemical stabilization: permeation grouting; jet grouting; deep soil mixing; and
lime/ cement columns.
Vibro stone columns can be considered one of the most effective methods to improve
performance of the soft ground. stone columns have the advantages of accelerating
consolidation, increasing the bearing capacity of poor ground, and improving the slope
stability and mitigate liquefaction by providing a sufficient reinforcement and providing a
drainage path allowing the pore water pressure to dissipate. Therefore, they have been used in
a verity of applications, such as road and railway embankments, tanks and marine structures
(Unnikrishnan et al., 2002; Fatahi et al., 2012).
7
2.3 Stone Column Technique: Historical Development
In 1937 there was the first reported application of vibro compaction used to densify a loose
sand deposit under a building in Berlin by penetrating a vibrating poker to the ground
(Slocombe et al., 2000).
In the 1950’s this technique was adopted in the UK, however, its application was limited due
to the geological nature of the ground (soft soil; shear strength < 15 kPa). This led to the
development of vibro compaction technique to cover the treatment of fine grained and
cohesive soils by introducing a coarse granular backfill into a vertical void in soft ground
formed using a vibratory poker. This early method of stone column construction received
further improvement in Europe and the United State of America until it became a well-
established technique (Greenwood, 1975; Mitchell, 1981).
Over the last three decades this technique has been widely used for both compaction of
cohesion-less soils and for reinforcing soft soils with granular columns. The method has been
shown to be an effective and reliable ground improvement method, especially, in cases, where
high sensitivity to settlement is not critical. It has proven to be a good alternative to the
traditional deep foundation methods such as piles (Serridge, 2006; McCabe et al., 2009).
2.4 Stone Column Construction Methods
Soil type and condition are the key factors that govern the method of installation of stone
columns to ensure the required quality of improvement (McCabe et al., 2009). The commonly
applicable installation techniques for soft soils (and the equipment used) are discussed below.
A brief summary of the installation methods are shown in Table 2.1.
8
Table ‎2.1 Summary of stone column installation methods
cu L dc
Installation method Commentary
(kPa) (m) (mm)
- Suitable for very soft soil, and

deep treatments;
Replacement
(wet top feed) 15 to 35 5-20 600-1200 - Applicable blow ground water

Section 2.4.2.1 level
- Has an environmental issue.
- Suitable only for stable and stiff

Displacement
soils;
(dry top feed) > 30 10 400-800
Section 2.4.3.1 - Shallow and medium treatment

depth.
- Suitable for very soft soil;
- Applicable below ground water

Displacement
level;
(dry bottom feed) 15 to50 15 400-800
Section 2.4.3.2 - Borehole stability is insured
- Clean stone, column diameter and

column length are assured.
cu: undrained shear strength of the soil;

L: column length;
dc: column diameter.
2.4.1 Vibro Equipment (Vibro-float)
Stone columns are constructed first by creating a vertical void in soft ground using a vibratory
poker (Figure 2.1). As the poker is slowly withdrawn, stone is inserted in the void in typical
9
stages of 300 to 500 mm thickness and then compacted by the vibrating poker for 30 to 60
seconds. This process is continued upward to form a column (Watts, 2000).
The equipment used to form the stone column typically comprises a hollow cylindrical steel
pipe with diameter ranging between 300 mm and 450 mm, and length ranges from 2.00 to
3.50 meter. The vibrator body unit consists of two main parts: the upper part known as the
extension or follower tube as shown in Figure 2.1. The main function of the follower tube is
to allow deep penetration into the ground (Watts, 2000; Moseley and Kirsch, 2004).
The lower section of the vibrating unit contains an eccentric weight which is powered by a
motor at the top of an axial shaft (typical motor power capacity ranges between 50 to 150 kW
and operating frequency ranging from 10 to 30Hz). This eccentric weight typically ranges
between 15 and 40 kN depending on the level of improvement and the condition of the
ground. There is a vibratory force of 150 to 700 kN can be transmitted from the vibrator
casing to the surrounding soil. (Greenwood and Kirsch, 1984; Raju and Sondermann, 2005).
In order to improve the installation process with a high level of quality control, the equipment
was modified by including useful features such as bottom feed delivery system, and
computerised monitoring system that could transfer data from site (Slocombe et al., 2000).
10
Figure ‎2.1 Deep vibrator details (Moseley and Kirsch, 2004)
2.4.2 Vibro Replacement (Wet Method)
2.4.2.1 Wet top feed
This technique has been used successfully in soft soils (fines content over 20 %; and
undrained shear strength of soil, cu, in range of 15 to 35 kPa), it can also be used for deep
ground treatments below water level (Moseley and Kirsch, 2004; McCabe et al., 2009).
In order to increase the bearing capacity and the stiffness of soft soil, a 10 to 35 % of the soil
may need to be replaced with uniformly graded stone, column diameters ranging between 0.6
11
and 1.2 meter and column depths between 5 and 15 meters. In this method it is preferable to
build columns that extend to end bearing stratum (Barksdale and Bachus, 1983).
The installation sequence of this technique is illustrated in Figure 2.2. The vibrator is inserted
into the soil with the assistance of jetted water. Both the flow of flushing water and the
vibration can reduce soil resistance and allow the vibrator to penetrate the soil under its own
weight. Once the required depth is achieved, granular backfill is placed from the bottom of the
hole to the top where gravel falls against a continuous upward flow of water. As the gravel is
accumulated at the bottom of the column, the vibrator is withdrawn gradually in
approximately 0.50 meter intervals. A continuous water flow protects the bore from the
collapsing. As the vibrational energy is dissipated in radial waves through the backfill
materials into the surrounding soil and additional expansion of the column is prevented by the
passive resistance of the cohesive soil. (Greenwood and Kirsch, 1984; Watts, 2000).
However, due to environmental issues relating to disposal of flush arising, application of this
method is limited currently to developments where large bearing capacity is required
(McCabe et al., 2009).
Figure ‎2.2 Vibro-replacement method (Keller GmbH, 2005)
12
2.4.3 Vibro Displacement (Dry Method)
2.4.3.1 Dry top feed
Dry top feed method is considered to be the simplest installation method, and most preferred
in the case of shallow to medium treatment depths of stable soils with undraind shear strength
larger than 30 kPa. This method is not applicable for treatment of soft clay soils due to the
fact that the borehole will collapse when the poker withdrawn due to lack of lateral support
(McKelvey, 2002; McCabe et al., 2009).
In this method, (Figure 2.3), once the required depth is reached the poker is lifted and the
granular backfill of material introduced from the ground surface into the hole. The vibro-float
is reinserted and stone in the column is compacted. Typical geometries of formed stone
columns are usually 400 to 800 mm diameter and 10 to 15 meters in length (Munfakh, 1984;
McKelvey, 2002).
Figure ‎2.3 Vibro-displacement (Keller GmbH, 2005)
13
2.4.3.2 Dry bottom feed
This method was introduced into the UK in the 1980’s and currently considered the most
commonly adopted installation method for stone column (McCabe et al., 2009). In terms of
application it is used in soils (cu ranges between 15 and 50 kPa), and unaffected by the
presence of ground water. During the backfill process the vibro-float remains inside the bore
providing more stability and preventing unwanted inclusion. The stone material is supplied
using a hopper to a pipe fixed to the side of the vibro-float shown in Figure 2.4. This can
allow the column to be constructed clean with the same back fill material. As with the top
feed method the stone is then compacted by repeated withdrawal and insertion of the poker
(McCabe et al., 2009). The average treatment depth achieved by this method is 15 meter
(Watts, 2000).
Figure ‎2.4 Dry bottom feed (Keller GmbH, 2005)
14
2.4.4 Monitoring, Quality Control and Assessment of Stone Columns Performance
As a result of the variety of equipment and construction methodologies and due to the lack of
availability of a fixed procedure to design vibro replacement method, construction of the
stone column is controlled by recording a range of parameters, including penetration depth,
energy consumption of the vibrator, and air/water pressure (Moseley and Kirsch, 2004).
Continuous recording of parameters as a function of time can provide reliable data that can be
used to control stone column construction and help prevent problems. This includes the
possibility of having to alter the design of columns and ensures column uniformity over entire
length.
In addition to the above, cone penetration test (CPT) and standard penetration test (SPT) are
generally used to confirm the achievement of a satisfactory level of vibro compaction (Bell,
2004; McCabe et al., 2009). However, it should be noted that, whilst the data collected during
column construction are for quality control, the penetration tests are used to ensure the degree
of improvement achieved post-construction.
In order to evaluate the compressibility of treated soil, it is better to conduct compaction tests
within at least one week of the compaction work to prevent the effect of ageing factor, which
influences the soil properties and causes an increase to its strength ranging from 50 to 100%.
This increase in strength occurs as a result of the reduction in water pressure, and also could
be due to the rearrangement of the physical and chemical bonding forces between the soil
particles within the column.
Stone columns are primarily used to reduce settlement. Therefore, large-scale loading tests,
which can be carried out by loading a rigid plate on top of one or more columns are used to
assess the degree of improvement in terms of settlement. However, as this test is costly and
15
time-consuming, it cannot be used as a control of workmanship, so it is only used in case of
large projects (Greenwood and Kirsch, 1984; Greenwood, 1991).
Recently, geophysical methods including continuous surface wave have been used to assess
settlement improvement (Madun et al., 2012). Although there is no large database to validate
the data collected by geophysical investigation, this method has an advantage of covering
large areas and could be used for both short and long term investigations, which could make it
cost effective (Redgers et al., 2008; Madun, 2012; Madun et al., 2012).
2.4.5 Material Used for Stone Columns Constructions
The key parameter in stone column design is the angle of shearing resistance, which is a
function of several parameters such as degree of compaction, grading and material strength
(McKelvey and Sivakumar, 2000; Jefferson et al., 2010). Most of the design methods
developed for predicting the bearing capacity of soils improved by stone column
reinforcement (Hughes and Withers, 1974; Greenwood and Kirsch, 1984) are related to the
undrained shear strength of surrounding soil and the internal angle of friction of the stone
materials; a reduction of 10o in the internal angle of friction can reduce the bearing capacity
by 50% (McKelvey et al., 2002).
In general, 45o can considered to be the maximum friction angle, while in the UK 40o
considered to be the most typical value used in the design (Serridge, 2006). Therefore, it is
important that the stone aggregate within the column have sufficient shear resistance whereby
the particles are strong enough to withstand local stress concentrations which occur during
construction. In addition aggregate used must be durable in the long-term (Watts, 2000).
16
There are three main sources of aggregate, which can be used in stone column construction.
Until recently natural or primary stone aggregate was the main source. However, increasing
demands for sustainable construction has led to the use of recycled aggregate as a second
source (e.g. construction demolitions and recycled railway ballast) and when possible the
secondary aggregates (i.e. industrial processed products such as steel slag, rock waste, and
china clay waste) (Jefferson et al., 2010).
Stone columns were installed in closely spaced groups beneath pad and strip foundations to
provide bearing pressures of 100 kN/m2 (associated with the main portal frame structure) and
on a square grid pattern beneath ground-bearing floor slab areas to provide bearing pressures
in the range 30–50 kN/m2
Serridge (2005), provides a detailed case history of a project in Coatbridge, Scotland, where
recycled (crushed concrete) aggregate was used successfully for stone column construction.
Columns with diameter of 600 mm and varied in lengths from 2.5 to 6 meters were installed
in a square pattern beneath pad and strip foundations to provide bearing pressure of 100
kN/m2 (for the main frame structure) and 30 kN/m2 under floor slab areas. The average
deformation modulus of 48 MN/m2 was achieved when assessed using plate load tests. He
suggested that, this value was comparable with typical results if natural aggregate was used in
similar ground condition.
McKelvey et al. (2002) examined the shear strength of different recycled aggregates (i.e.
quarry waste, crushed concrete, and building debris). The performance of these materials
were compared with the performance of crushed rock primary aggregate, in terms of the
influence of dry and wet conditions and the effect of the fines content. They found that the
primary aggregate performed better than the other recycled aggregate in all test conditions.
17
For example, in the dry condition, the angle of internal friction for the primary aggregate was
51o, whereas for the secondary aggregate was ranging from (46 o to 37o). These values were
decreased by about 6 % in the wet condition and by about 30 % when 20 % of Kaolin slurry
was add to the aggregate. This study indicated that the internal angle of friction of natural
aggregate had been influenced by all the tested conditions while the quarry waste showed
independency toward the slurry content and affected by the water content, where it showed
different trend in case of crushed concrete aggregate, where it was influenced by the slurry
content but not with the water content.
This could indicate that, regardless of the stone source used, the achievable quality of the
stone column is mostly dependent on the quality of aggregate adopted. The ICE (1987) and
BSI (2005) suggested that in order to achieve a good interlock between the stone particles and
to allow an adequate level of drainage through the column, the grading of aggregates should
range between 20 to 75 mm, also the percentage of fine content should not exceed 10 %.
2.5 Application and Limitation of Stone Column
Stone column technique has been used successfully in many projects around the world for
ground improvement purposes (e.g. reduce differential settlement, accelerate the
consolidation process; increase the bearing capacity and to mitigate liquefaction) in various
types of soil from coarse gravel to fine sand, silt and clay (Woodward, 2004). The same
benefits have been found when this technique was also applied to improve the characterisation
of soft marine clay, non-engineering fills and layered soils (McKelvey, 2002; Raju and
Sondermann, 2005; Han, 2015).
18
2.5.1 Limitations
Although stone columns have been installed successfully in soils with undrained shear
strengths ranging between 10 kPa and 50 kPa (Hu, 1995; McKelvey, 2002), they are not
suitable for soils with very low undrained shear strength and in soils with thick layers of peat.
This is due to insufficient lateral support that these materials provide and excessive settlement
that could occur due to the loss of water content during consolidation (Greenwood and Kirsch,
1984). Barksdale and Bachus (1983) suggest that stone columns can be more effective when
used for soil stabilisation rather than structural foundations. Most recommended treatment
depths are in the range of 6 to 15 m although columns have been constructed to a depth
greater than 30 m (Watts, 2000). Floating columns are not recommended to be used in weak
soils and therefore, a competent end bearing for stone columns is generally specified (Killeen
and McCabe, 2014).
2.6 Design Approaches of Stone Column
As the column and the surrounding soils work together and share the stresses subjected to
them, the bearing response of the soil/ column system is influenced by the properties of both
materials and their interaction between each other. In order to theoretically solve this complex
problem a certain level of idealisation (unit cell idealisation) is used. Most existing theories
consider stone and clay as perfect elastic or elastic-plastic materials (Hughes and Withers,
1974; Barksdale and Bachus, 1983; Priebe, 1995). Therefore the design of stone columns
requires cognisance of parameters such as used grid pattern, area replacement ratio and stress
concentration factor (Sections 2.6.1.1 to 2.6.1.3). The typical design procedure suggested by
Watts (2000) can be summarised as follow:
19
1. Predict the ultimate bearing capacity of the single column using the undrained shear
strength of the soil and the internal angle of friction,
2. Determine the column diameter and the allowable spacing between the column,
3. Predict the level of settlement.
2.6.1 Basic Design Parameters
2.6.1.1 Unit Cell Idealization
Balaam et al. (1977) stated that the unit cell idealises for the behaviour of single column and
the soil around it, and the behaviour of one unit cells within a group is the same. This concept
is used to determine the area within which vertical stresses are considered to be distributed
(Barksdale and Bachus, 1983). Stone columns can be installed in three main patterns: triangle,
square and rectangular patterns. The equilateral triangle pattern is considered the most usable
and efficient arrangement with regards to large areas and uniform rate of densification, while
the square and rectangular pattern is applied in condition of isolated spread footing (Watts,
2000; BSI, 2005).
Typical layouts of stone columns in different patterns are shown in Figure 2.5. It is
convenient to associate the tributary area of soil surrounding each stone column. The tributary
area can be closely approximated as an equivalent circle (unit cell) with equivalent diameter
(De) having approximately similar total area.
The spacing (S) between the columns is a critical factor since it influences the degree of
improvement. The spacing is generally determined in order to provide an overlapping zone to
cover a wide area of treated soils. Greenwood (1970) stated that, in practice, a narrow spacing
is preferential under isolated foundation compared to beneath large raft foundations.
20
However, the spacing is generally ranges from 2 to 3 times the column diameter (Hughes and
Withers, 1974).
Depending on the stone column arrangements, spacing and the equivalent diameter can be
related to each other as shown in Table 2.2
Table ‎2.2 Equivalent unit cell diameter (adopted from Balaam & Booker, (1981))
Arrangement Equivalent unit cell diameter (De); (S) centre-to-centre spacing
12 0.25
Triangular ( 2) 𝑆 ≅ 1.05 𝑆
𝜋
16 0.25
Square ( 2) 𝑆 ≅ 1.13 𝑆
𝜋
De
S
D
Plan view
De Stone column
Rigid
frictionless side
L
Unit cell
Figure ‎2.5 Unit cell concept (Barksdale and Bachus, 1983)
21
1⁄
12 4
De 𝐷𝑒 = ( ) 𝑆
𝜋2
=1.05 S
1⁄
16 4
𝐷𝑒 = ( ) 𝑆
De 𝜋2
=1.13 S
Figure ‎2.6 Stone Column arrangements in (a) Triangular (b) Square (Balaam & Booker, 1981)
2.6.1.2 Area Replacement Ratio
Based on the concept of unit cell, it is possible to use the geometry of a mesh of stone
columns in order to determine the amount of soil replaced and the area replacement ratio As,
which is defined as the ratio between the area of each column Ac and soil area A (Barksdale
and Bachus, 1983):
2
𝐴𝑐⁄ 𝐷𝑐
𝐴𝑠 = 𝐴 = [ ⁄𝐷𝑒 ] ‎2.1
Where: Dc is the compacted stone column diameter; and De is the equivalent unit cell
diameter.
The area replacement ratio, also can be presented in terms of the diameter and spacing of the
stone column as follows:
22
𝐷
𝐴𝑠 = 𝑐𝑔 〈 𝑆𝑐 〉2 2.2
Where: Dc is the diameter of the compacted stone column; S is centre-to-centre spacing; cg is
a constant dependent upon the pattern of stone columns used.
In general, increasing the area replacement ratio leads to improved behaviour of the
composite soil (i.e. increased bearing capacity and reduced settlement). For improvement in
bearing capacity larger than 30 %, Wood et al. (2000) indicated that the area replacement ratio
should be 25% or above.
2.6.1.3 Stress Concentration Factor (n)
A large portion of the stress concentration is transferred from the loaded foundation to the
stone column. The remaining stresses are transferred to the soil as it is weaker than the
column material. The vertical stress distribution within the unit cell (concentration factor) is
defined as the ratio of the stress on the stone column to that on the surrounding soil within a
unit cell (Han & Ye, 1991). The magnitude of the concentration factor generally ranges from
2 to 6 (Aboshi et al., 1979; Barksdale and Bachus, 1983).
𝜎𝑠
𝑛= ⁄𝜎𝑐 ‎2.3
Where: σs is stress in the stone column; and σc is the stress in the surrounding soil.
There are several factors that could influence the magnitude of stress concentration such as
type of foundation, length of the column and time of consolidation. Juran and Guermazi
(1988) stated that the magnitude of (n) increases with time of consolidation and decreases
with the length of the column. in addition, loading the soil/column system through a rigid
23
foundation would lead to larger stress concentration than if flexible a foundation was used
(Barksdale and Bachus, 1983).
Ambily and Gandhi (2007) indicated that the modular ratio had a direct impact on the stress
concentration, they found that the n factor increased with any increase in modular ratio; and
reduced with any increase in the strength of the surrounding soil.
The total vertical stresses σ in stone columns and the surrounding soil over the unit cell area
and corresponding to a given area replacement ratio, can be expressed as:
𝜎 = 𝜎𝑠 . 𝐴𝑠 + 𝜎𝑐 (1 − 𝐴𝑠 ) 2.4
Where the terms have been previously defined.
Thus the stress concentration factor can be estimated (as follow) assuming equal vertical
displacement would occur and using the elastic theory as a function of the modular ratio of
the column and the soil (Barksdale and Bachus, 1983; Babu et al., 2013).
𝜎
𝜎𝑐 ≤ 𝜇𝑐 𝜎 = 𝜇𝑐 ( 𝑠⁄𝜇𝑠 )‎2 ‎2.5
Where: 𝜇𝑐 𝑎𝑛𝑑 𝜇𝑠 are the stress ratios in the clay and the stone column respectively.
𝜇𝑠 = 𝑛⁄[1 ‎2.6
+ (𝑛 − 1)𝐴𝑠 ]
𝜇𝑐 = 1⁄[1 + (𝑛 − 1)𝐴 ] ‎2.7

𝑠
24
2.6.2 Consolidation and Settlement
One of the main purposes of using stone columns is to reduce the overall settlement of the
treated soil. This reduction is governed by the magnitude of applied load distribution that
occurs between the column and surrounding soil.
In order to satisfy the unit cell concept condition (Figure 2.7), most of the design approaches
for predicting settlement of soil/column system assume an infinitely wide loaded area
reinforced with granular column with a constant diameter and spacing. The column is
assumed to be loaded via a rigid plate and ending in an undeformable bearing layer (Van
Impe et al., 1997; Babu et al., 2013).
There are various methods for estimating the settlement of the stone column/ soil system.
These are summarised in Table 2.3.
Figure ‎2.7 Settlement estimation conditions
25
Table ‎2.3 Methods of estimation of settlement
Method Reference Equation comment
Greenwood’s Limited to cu between 20

Greenwood Empirical design curve
method and 40kPa ; underestimate
(1970) (Figure 2.8)
(2.6.2.1) settlement when S>2.5m
Equilibrium 𝑆𝑡 = 𝑚𝑣 (𝜇𝑐 𝜎)𝐻 Dependent on the n and As;

Aboshi et al.
method 1 Might be sufficient for
(1979) 𝑅 = 𝜇𝑐 =
(2.6.2.2) 1 + (𝑛 − 1)𝐴𝑠 preliminary design;
1 0.5 + 𝑓(𝜇, 𝐴𝑠 )
= 1 + 𝐴𝑠 [ − 1]
𝑅 (𝑘𝐴 )𝑠 𝑓(𝜇, 𝐴𝑠 )
Priebe Priebe
1 − 𝜇2 (1 − 2𝜇)(1 − 𝐴𝑠 )
method (1976, 𝑓(𝜇, 𝐴𝑠 ) = [ 2
][ ] (𝑘𝐴 )𝑠 Overestimated results
1 − 𝜇 − 2𝜇 1 − 2𝜇 + 𝐴𝑠
(2.6.2.3) 1995)
∅
= tan2 (45 − )
2
Finite
Balaam et Its accuracy dependent on
element [𝐾𝐸 ]{∆𝜎 (𝑚−1) } = {{∆𝐹𝐸 } + {𝐾𝑐(𝑚) } {∆𝜎 (𝑚) } + {∆𝐹𝐷𝑁
(𝑚)
}}
al. (1977) the input parameters
method
2.6.2.1 Greenwood Method.
Greenwood (1970) was one of the first to introduce empirical design curves to estimate the
settlement of stone column/ soil systems under widespread loading. The empirical curves
showed in Figure 2.8, represents the settlement reduction as a function of column spacing and
the undrained shear strength of the surrounding soil, which was limited between 20 and 40
kPa. The shaded zone in the curve represents the expected settlement reductions when
applying a wet process of construction. Greenwood (1970) suggested that these charts should
be used with caution within the indicated range. Although Balaam and Booker (1985)
indicated that these curves showed a close agreement with their finite element analysis
26
approach for settlement estimation, they suggested that at column spacing of over 2.5 m, the
Greenwood’s method might underestimate the degree of improvement.
Spacing of stone column, (m)

Settlement of treated ground as % of
settlement of untreated ground
 Curve neglect immediate settlement and shear displacement.

 Columns assumed resting on firm clay, sand or harder ground.
Figure ‎2.8 Settlement diagram for stone column in uniform soft clay (Greenwood, 1970)
2.6.2.2 The equilibrium method
This method has been used (mainly in Japan) to estimate the settlement of sand compaction
piles (Barksdale and Bachus, 1983). Aboshi et al. (1979) presented this simple approach for
estimating the reduction in settlement of ground reinforced by stone columns based on the
unit cell assumptions and the one dimensional consolidation theory. This approach require
estimation of stress concentration factor n using previous work experience and the past results
from field measurements of stresses.
27
The variation of the vertical stress in the clay (σc) due to the applied external stress can be
expressed by:
𝜎𝑐 = 𝜇𝑐 𝜎 2.8
Where: σ is the average external applied stress; and µc is the ratio of stresses in the clay (Eq.
2.7).
By applying the one dimensional consolidation theory the primary consolidation settlement
can be estimated using the following equation.
𝐶 𝜎0′ + 𝜎𝑐
𝑆𝑡 = (1+𝑒𝑐 ) . log10 ( ). 𝐻 2.9
0 𝜎0 ′
Where: St is the primary consolidation settlement over distance H of stone column treated
ground; H is the height of stone column; e0 is the initial void ratio; Cc is the compression
index; σ0 is the average initial effective stress; and σc is the changing in stress in the clay
layer.
Additionally, the following equation may be used to estimate the settlement of unreinforced
and reinforced soils.
𝑆 = 𝑚𝑣 . 𝜎. 𝐻 ‎2.10
𝑆𝑡 = 𝑚𝑣 (𝜇𝑐 𝜎)𝐻 ‎2.11
Where, S and St, as defined previously, are the consolidation settlements of unreinforced and
reinforced soils respectively; mv if the coefficient of compressibility; and H is the thickness of
the soil layer.
28
The ratio of the settlement of the stone column (i.e. improved soil to the unimproved soils)
can be expressed as:
𝑆𝑡⁄ 1
𝑅= 𝑆 = [1 + (𝑛 − 1)𝐴𝑠 ] ‎2. 12
Where n is the stress concentration factor and As is the area replacement ratio.
The above equation showed that the settlement ratio is a function of the area replacement ratio
and the stress concentration factor, and the settlement of the treated soil decreases as those
parameters increases. Therefore, as suggested by Barksdale and Bachus (1983) the
equilibrium method would be sufficient for preliminary design if the stress concentration
factor was reasonably estimated.
2.6.2.3 Priebe method
Priebe method is considered one of the most widely adopted semi-empirical methods used to
obtain the improvement factor (estimating settlement reduction due to soil treatment with
stone column). This method has undergone a series of improvements and modifications to its
current version proposed (Priebe, 1995). These modifications considered a number of
additional factors such as the effect of compressibility of the column material and
confinement from overburden, which considers the effect of unit weight of both column and
soil materials by adding the depth factor to the design calculation (Barksdale and Bachus,
1983; Ellouze et al., 2010; Babu et al., 2013).
The original solution proposed contained a number of simplifying assumptions that are listed
below:
29
 equal vertical settlement of the stone and soil;
 uniform stresses in the column/soil system;
 stone column is overlying a rigid soil layer;
 incompressible stone column, so changing in volume within the soil associated to

the axial reduction of the cylindrical column;
 long elastic hollow cylinder solution is assumed in order to find out the radial
deformation of the soil and;
 a uniform internal pressure within the unit cell boundary.
According to these assumptions it can be concluded that the column is designed to only fail in
a bulging mode and the settlement can be controlled.
Based on these assumptions, Priebe (1995) presented in Figure 2.9 the ratio of settlement of
untreated to treated ground S/St as a function of the area replacement ratio As and internal
angle of friction of the stone column φs.
Therefore, the radial deformation of the column/soil system can be determined from the
solution of expanding cylindrical cavity in half elastic. This expression gives the radial
deformation of the column based on vertical stress acting on it and on the ground (Priebe,
1995). Establishing equality of vertical deformations in the soil and the column and by the
equilibrium condition of vertical stresses, the following improvement factor n0 expression can
be obtained:
𝐴𝑐 1⁄ +𝑓(𝜇 ,𝐴 /𝐴)
𝑠 𝑐
𝑛0 = 1 + . ( 𝐾 2 .𝑓(𝜇 − 1) 2.13
𝐴 𝑎𝑐 𝑠 ,𝐴𝑐 /𝐴)
(1−𝜇𝑠 ) . (1−𝐴𝑐⁄𝐴)
Where: 𝑓(𝜇𝑠 , 𝐴𝑐 /𝐴) = 𝐴 ; 2.14
1−2𝜇𝑠 + 𝑐⁄𝐴
30
µs is the Poisson’s ratio; and (𝐾𝑎𝑐 ) coefficient of active pressure of the column is
defined as:
∅𝑐⁄
(𝐾𝑎𝑐 ) = 𝑡𝑎𝑛2 (45𝑜 − 2) 2.15
Although this method has been widely used in ground improvement industry, it has
limitations in settlement prediction, such as not being suitable for estimating the behaviour of
floating columns (Babu et al., 2013).
Ellouze et al. (2010) reported that some of the assumptions in Preibe’s method, such as that
the unit weights of column material and initial soil are neglected, are questionable and not
well defined. Additionally, they highlight that the design procedure is not connected
mathematically and inconsistence. For example, in the first step of the design, Priebe (1995)
considers a cylindrical cavity subjected to lateral expansion during which zero vertical
deformation is assumed in order to give a solution expressed in plane stress condition. In a
second step, this solution is incorporated in the soil/column model for which there is a
distribution of vertical stress generating non-null vertical deformation, and, consequently, the
settlement is assumed constant.
31
Figure ‎2.9 Improvement factor (Priebe, 1995)
Using a case study of 54 m diameter tank built on reclaimed ground reinforced by stone
columns to ensure the stability of the tank at Zarzis terminal (Tunisia), Ellouze et al. (2010)
demonstrated that Priebe’s method was overestimated the settlement of reinforced soil by
stone columns comparing with other linear elastic models (Balaam and Booker (1981); Chow
(1996) Bouassida et al. (2003) and the French recommendations (Françaises, 2005)).
Working load of the tank was approximated as a quasi-uniform stress of 120 kPa. The
reinforcement was performed along an average depth of 7 meters with a columns diameter of
1.2 meter installed in a triangular pattern. The soil was improved on a circular area with a
diameter of 62 meters, which implies an improved area ratio of about 32 %.
The settlement of unreinforced soil was estimated to be about 230 mm underneath the centre-
line of the tank, whereas the settlement at the edge of tank was estimated to be about 60 mm.
32
The settlement predictions of the reinforced soil at the centre-line of the tank and its edge that
were obtained using the software columns (Bouassida et al., 2009) and Priebe’s method are
presented in Table 2.4.
Table ‎2.4 Comparison between predictions of tank’s settlement (Ellouze et al., 2010)
Settlement of Settlement of
Settlement reduction Settlement reduction
reinforced soil at reinforced soil at
Method factor at the centre factor at the edge of
the centre line of the edge of the
line of the tank the tank
the tank, (mm) tank, (mm)
Recorded - 30 - 2.00
Bouassida et al. (2003) 58 28 3.96 2.14
French recommendation
55 26 4.18 2.30
(Françaises, 2005)
Balaam and Booker (1981) 51 24 4.50 2.50
Priebe (1995) 61 21 3.77 2.85
2.6.2.4 Finite Element Method
The finite element method could provide an appropriate theoretically approach to model stone
column reinforced ground, where nonlinear material properties, the column/ soil interface and
the boundary conditions can all be sufficiently modelled (Gniel and Bouazza, 2007; Killeen,
2012; Babu et al., 2013; Killeen and McCabe, 2014; Mohanty and Samanta, 2015). Most of
the finite elements studies have utilized the axisymmetric unit cell model to analyse the
conditions of either a uniform load on a large group of stone columns or a single stone
column. Balaam et al. (1977) used finite element method to investigate the behaviour of stone
column, and developed a design curves for predicting settlement reduction. Figure 2.10
showed an example of these curves. Balaam et al. (1977) highlited that column diameter and
33
ratio of column length to depth of the soil layer had significant influence on settlement
amount, whereas modular ratio of column to soil was less important.
q ̸ cu
δ ̸d
Figure ‎2.10 Load – settlement curves for stone column (Balaam et al., 1977)
Barksdale and Bachus (1983) developed a design curves for estimating settlement in both low
compressibility and compressible soils using a nonlinear finite element method combined
with the idealisation of unit cell to predict the primary consolidation settlement of column/soil
system. In this approach the soil modulus of elasticity was assumed to be constant with the
depth of the soil. In addition to this, different area replacement ratios (10, 20, 25, and 35 %) at
different L/D ratios (5, 10, and 20) were considered and each case has a different design chart.
Poorooshasb and Meyerhof (1997) introduced an elastic soil model to predict the settlement
reduction of raft foundation resting on end bearing stone column reinforced soft soils. This
model indicated that the design pattern, spacing and the degree of compaction of the column
material were the most viable parameters to control the level of settlement.
34
Generally, finite element solution claimed by authors (Ambily and Gandhi, 2007; Killeen,
2012; Killeen and McCabe, 2014) to provide a good agreement with the full scale and site
investigation results; however, their utility is governed by the accuracy of the input
parameters.
2.6.3 Bearing Capacity Prediction Methods
Stone columns are often constructed penetrating through soft soil to end bearing layers with a
critical length of 4 to 6 times the column diameter. Thus the most possible mode of failure
that might develop for an isolated stone column over a depth of 2 to 3 diameters from the
surface is bulging failure (Barksdale and Bachus, 1983; Bergado et al., 1991). Several
approaches have been developed to predict the ultimate capacity of an isolated stone column
surrounded by soft soils. Table 2.5 summaries the most common of these corresponding to the
mode of failure as presented by Bergado et al. (1991).
These approaches were developed on the basis that the lateral confining stress supporting the
column is usually considered as the ultimate passive resistance which the surrounding soil can
mobilize as the column bulges. Since then the ultimate vertical stress(𝜎𝑣 ) can be predicted as
the coefficient of the passive pressure of the column (𝑘𝑝 ) times the lateral confining pressure
(𝜎𝑟 ).
𝜎𝑣 = 𝜎𝑟 ∗ 𝑘𝑝
1  sin ΄
= 𝜎𝑟 ( ) 2.16
1  sin ΄
35
Table ‎2.5 Estimation of ultimate bearing capacity (Bergado et al., 1991)
Mode of failure Design equation Reference
1 + sin ∅𝑠
𝑞𝑢𝑙𝑡 = (𝛾𝑐 𝑧𝑘𝑝𝑐 + 2𝑐𝑜 𝑧√𝑘𝑝𝑐 ) Greenwood (1970)
1 − sin ∅𝑠
1 + sin ∅𝑠
𝑞𝑢𝑙𝑡 = (𝐹𝑐, 𝐶𝑜 + 𝐹𝑞′ 𝑄𝑜 ) Vesic (1972)
1 − sin ∅𝑠
Bulging
1 + sin ∅𝑠
𝑞𝑢𝑙𝑡 = (𝜎𝑟𝑜 + 4𝐶𝑢 ) Hughes and Withers (1974)
1 − sin ∅𝑠
𝑞𝑢𝑙𝑡 = (𝑁𝑐 𝐶𝑢 ) Barksdale and Bachus (1984)
𝑞𝑢𝑙𝑡 = 𝐶𝑢 𝑁𝑐 + 0.5𝛾𝑐 𝑁𝛾 𝐵 + 𝛾𝑐 𝐷𝑓 𝑁𝑞 Madhav and Vitkar (1978)
3 2
General shear 𝑞𝑢𝑙𝑡 = 0.5𝛾𝑐 𝐵𝑡𝑎𝑛𝜓 + 2𝐶𝑜 𝑡𝑎𝑛𝜓 + 2 (1 − 𝑎𝑠 )𝐶𝑜 𝑡𝑎𝑛𝜓
Barksdale and Bachus (1983)
tan(𝜇𝑠 𝑎𝑠 tan ∅𝑠 )
𝜓 = 45𝑜 +
2
Greenwood (1970) noted that there was no exact mathematical method to estimate the bearing
capacity of cohesive soils treated by stone columns, because of the dilation that occurs within
the column and the resulting lateral stress to the surrounding soil which can be resisted by
passive pressure; (Greenwood, 1970) hypothesed that the column will behave as if it was in a
triaxial chamber, and the degree of improvement in the bearing capacity would be governed
by the lateral support from the surrounded clay to the column and the internal angle of
friction. In addition, Greenwood (1970) highlighted that the carrying capacity of the column
increases until either a local shear failure in the clay or end bearing failure at the bottom of the
column accords.
Hughes and Withers (1974) used a laboratory based model and radiography device, to study
the behaviour of both sand columns and surrounding clay by tracking the deformations
occurred within and outside the column. They concluded that the cylindrical cavity expansion
theory can be used to define the column behaviour. They proposed the following equation for
estimating the ultimate vertical stress (𝜎𝑣 ) in a stone column:
36
1+sin ∅′ ′
𝜎𝑣 = (𝜎𝑟𝑜 + 4𝑐) ‎2.16
1−sin ∅′
Vesic (1972), based on the theory of cylindrical cavity expansion, developed an expression
for estimating lateral resistance where the ultimate lateral resistance is:
σ3 = c Fc + q Fq 2.17
where: c is the cohesion of the soil, q is the mean stress at the equivalent failure depth and
Fc, Fq are cavity expansion factors which determine from Figure 2.11 using the rigidity index,
Ir,:
E
Ir = ‎2.18
2(1  v)(c  q tan ΄ )
Where: E is the modulus of elasticity of the soil in which cavity expansion occur, c is the
cohesion of the soil, v is the Poisson’s ratio of the soil, ΄ is the effective angle of friction of
the soil and q is the mean stress within the zone of failure.
Figure ‎2.11 Vesic’s cylindrical cavity expansion factors (Barksdale and Bachus 1983).
37
Therefore assuming (𝜎3 = 𝜎𝑟 ) substituting (2.18) in (2.16) the ultimate stress that can be
applied to a stone column is:
1  sin ΄
𝜎𝑣 = [c Fc + q Fq] ( ) ‎2.19
1  sin ΄
The mean stress q used in the above equations should be taken as stress at the average depth
of the bulge taking into account the initial and the final stresses in the ground. Vesic (1972)
expressions can be used both for short and long term calculation.
Barksdale and Bachus (1983) developed a simple approach for estimating the bearing
capacity of a single column. (Equation 2.21) based on authors ‘past experience and utilising
good engineering judgment’.
𝑞𝑢𝑙𝑡 = 𝑐𝑢 ∗ 𝑁𝑐 2.20
Where, 𝑞𝑢𝑙𝑡 is the ultimate bearing capacity of the stone column; 𝑐𝑢 is the undrained shear
strength of the surrounding soil; and 𝑁𝑐 is the bearing capacity factor for the stone column
usually ranging between 18 and 22, and 5 for estimating the bearing cabacity of the untreated
soil. The value of 𝑁𝑐 is highly dependent on the compressibility of the soil surrounding the
column, where it is increased with the increase of soil stiffness (Barksdale and Bachus, 1983;
McKelvey, 2002).
To conclude, determination of load bearing capacity of a single stone column is complex and
there is no exact mathematical solution to predict it. This is due to the uncertainty of the
interaction behaviour between the stone column and the surrounding soils. However, the
relationship based on the laboratory model tests proposed by Hughes and Withers (1974) is
38
widely considered realistic for vibro stone columns design (Greenwood, 1991; Babu et al.,
2013; Najjar, 2013).
2.7 Mechanism and Performance of Stone Columns
2.7.1 Stone Columns Investigation under Monotonic Loading Condition
2.7.1.1 Failure mechanism
Generally, stone columns are created using high friction granular material where the stiffness
of the column is dependent on the lateral support given by the surrounded soil. If this support
is not sufficient the column will fail (Barksdale and Bachus, 1983; Bergado et al., 1991).
The failure mechanism of a single column can be a function of two main factors in addition to
its length whether it is considered to be long or short (i.e. the length of the column is greater
𝐿
or shorter than the critical length(𝐷 ≈ 6)).
The first one is the method of column construction as either end bearing on a firm competent
layer of soil, or as floating columns.
The second factor is the load application on the column, which could be either directly to the
top of the column or through a rigid cap over an area larger than the diameter of the stone
column (Barksdale and Bachus, 1983; Sivakumar et al., 2004a). Figures 2.12 shows the
expected mod of failures, which can be summarised as:
39
Rigid foundation
Side Friction
2-3D
End Bearing
D
a. Long stone column with rigid base, loaded b. Long stone column with rigid base c. Short column with rigid base d. Short floating column
to area larger than the column diameter  Bulging failure  Shear failure  Punching failure
 Bulging failure
Figure ‎2.12 Failure mechanism of single stone column (Barksdale and Bachus, 1983)
 Bulging failure: for columns overlying on rigid base, having length over their critical
length and loaded directly over the column area, a bulging failure would occur over 2
to 3 diameter of the depth. However, if the load was applied over larger area, then
there would be an increase in the vertical and lateral stresses as in the surrounding soft
soil, which would then affect the response of the column to loading. Normally this
leads to smaller bulging and higher ultimate bearing capacity (Barksdale and Bachus,
1983).
𝐿
 Shear and punching failure: columns shorter than the critical length (𝐷 ≈ 6) were
likely to fail in shear failure if they were end bearing on rigid base, or in punching
failure if they were floating columns (Barksdale and Bachus, 1983).
Bulging of the column is more noticeable in the upper portion of the column as indicated by
Greenwood (1970), Hughes and Withers (1974), Barksdale and Bachus (1983), Charles and
Watts (1983), Greenwood (1991), and Sivakumar et al. (2004a). Barksdale and Bachus (1983)
suggested that bulging would occur within 2 to 3 times the column diameter whereas Hughes
and Withers (1974) and Sivakumar et al. (2004a) observed that the column bulged at depth of
approximately four times the diameter of the column. Bae et al. (2002), suggested that the
column diameter could be the main parameter affecting the depth of this bulging zone, while
40
other factors such as soil strength and the depth ratio could be considered to have less
influence. on the other hand Sivakumar et al. (2007) noted that the degree of bulging is
largely dependent on in situ shear strength of the surrounding soil.
The mode of failure of stone columns in group has been investigated by a number of
researchers (e.g. Barksdale and Bachus (1983); Hu (1995); McKelvey (2002); Sivakumar et
al. (2004a)). In general they indicated that stone columns constructed in groups showed
different failure behaviour than the isolated columns; in groups, each column can interact and
restrain the expansion of the neighbouring column leading to increase the bearing capacity.
Hu (1995) studied the behaviour of stone column groups and suggested that there are three
different modes of failures (bulging, shearing and lateral deflection) which depend on the
geometric configuration of the columns. Different geometry parameters (L/r0 = 2, 3.2 and 3.4;
rc = 5.5 and 8.75 mm; As = 24 and 30 %) were tested in group of 5 and 7 sand column
reinforced by a footing with a radius (r0) of 50 mm. Hu (1995) showed that the shear planes
through the columns would occur towards the edge of the footing, while bulging could be
seen deep directly under the foundation. The depth of the bulging was found to increase as the
area replacement ratio As increases. Additionally he observed that short columns tend to
penetrate underlying clay, however, as column length increases penetration decreases.
Sivakumar et al. (2004a) noted that the central columns beneath the foundation bulge
uniformly, whereas the others at the edge they bulge away toward the surrounding soils.
A range of methods can be used to examine the deformation and failure patterns of stone
columns. For example, Hughes and Withers (1974) used an X ray technique to monitor the
deformation of isolated stone column. Whereas (Hu (1995), Wood et al. (2000), Ambily and
41
Gandhi (2007)) successfully formed a plaster-cast of the vacuumed holes after exhuming the
columns material.
McKelvey (2002) and Sivakumar et al. (2004a) used a transparent medium as can be seen in
Figure 2.13 (with shear strength properties similar to the soft clay) allowing constant
monitoring to the column deformation.
Figure ‎2.13 Photographs of sand columns beneath circular footing at beginning, middle and end
of foundation loading process: (a) TS-01, 150 mm; (b) TS-02, 250 mm
(Sivakumar et al., 2004a)
42
Sivakumar et al. (2004b) investigated the behaviour of single stone column under the
condition of undrained triaxial, they examined the mode of failure by splitting the specimen
(100 mm diameter X 200 mm length) vertically along its centre axis.
Another method that can be used for this purpose by grouting the stone column by pouring the
concentrated cement slurry after testing and allowing it to set for some time depending on the
concentration of the slurry and then carefully removing the surrounding soils (Sivakumar et
al., 2010).
2.7.1.2 Critical length
Hughes et al. (1976) defined the critical length of the column as the shortest length that allows
the column to carry the designed ultimate load regardless of the deformation that would occur.
𝑃𝑢𝑙𝑡 = 𝑐̅(𝜋𝐷𝐿𝑐 ) + 𝑁𝑐 𝑐 𝐴𝑐 2.21
Where, 𝑃𝑢𝑙𝑡 is the ultimate column load; 𝑐̅ 𝑎𝑛𝑑 𝑐 are respectively the average shaft cohesion
and the cohesion of the soil at the bottom of the critical length; 𝐴𝑐 is the column area with
diameter D; and (𝜋𝐷𝐿𝑐 ) representing the surface area of the side of the column with critical
length 𝐿𝑐 .
Hughes and Withers (1974) introduced the L/D ratio and they found that there is no
improvement in the bearing capacity of the column beyond (L/D = 4.1). Samadhiya et al.
(2008) studied the effect of the column length on the column bearing capacity, they identified
that the critical length was at 4.5D and found that any increase of the bearing capacity was
marginal after this length. Black et al. (2011) presented results obtained from a large physical
triaxial model, they investigated the influence of L/D ratio and the area replacement ratio on
the settlement improvement factor. Black et al. (2011) suggested that the L/D ratio greater
43
than 6 is required to provide the full limiting axial stress on the column. in addition they
found that an L/D ratio beyond 8 offers insignificant improvement in terms of settlement
control (i.e less than 5%). They also suggested that settlement can be controlled in short
columns (L/D < 6) using a relatively high area replacement ratio (> 30 %).
In conclusion, column length larger than the critical length may not provide further
improvement in the column bearing capacity, but it may be used to control the level of
settlement
2.7.1.3 Undrained shear strength of the surrounding soil
As explained in previous sections (2.7.2 and 2.7.3), surrounding soils provide the lateral
confinement to stone columns, therefore when designing stone column it is important to have
an appropriate knowledge of the soils undrained shear strength. Figure 2.14 shows the
influence of this factor on the ultimate carrying capacity of the stone column according to
both an empirical design relation presented in Equation 2.16 and different laboratory
investigation results carried out by Hughes and Withers (1974), Ambily and Gandhi (2004),
Kim and Lee (2005), Ambily and Gandhi (2007), Zahmatkesh and Choobbasti (2010) , Black
et al. (2011), and Ali et al. (2014). Although the obtained results did not show a perfect fit
with the prediction lines, they show that the column bearing capacity increases with the
increase of the soil shear strength. However, the degree of improvement at similar soil
strength is dependent on other factors such as the area replacement ratio and the internal angle
of friction. For example, Ambily and Gandhi (2007) investigated the effect of the soil shear
strength cu= 7, 14, and 30 kPa) and the area replacement ratio (As = 5, 10, and 19 %) on the
behaviour of single stone column, they noted that the ultimate bearing capacity of the
44
improved soil increased in linearly from 175 kPa to 740 kPa when increasing the undrained
shear strength from 7 to 30 kPa.
1200 Hughes and Withers (1974); (cu:19kPa,As:40%,φ:35o)

qult Prediction ∅′ 43o
Ultimate bearing capacity of the treated
Kim and Lee (2005); (cu:4kPa,As:30%,φ:38o)

Ali et al. (2014); (cu:7kPa,As:25%,φ:38o)
1000 Zahmatkesh and Choobbasti (2012); (cu:5kPa,As:20%,φ:43o)
Ambily and Gandhi (2007); (cu:7-30kPa,As:5-19%,φ:43o)
Bergado et al. (1987); (cu:20kPa,As:6%,φ:43o)
800 Black et al. (2011); (cu:35kPa,As:17-28%,φ:43o)
soil, (kPa)
600
400
qult Prediction ∅′ 35o
200
0
0 5 10 15 20 25 30 35 40
Undrained shear strength, (kPa)
Figure ‎2.14 Effect of undrained shear strength of the surrounding soil on the ultimate bearing
capacity of the treated ground
2.7.1.4 Area replacement ratio
Data from previous research studies Wood et al. (2000), Kim and Lee (2005), Ambily and
Gandhi (2007), Najjar et al. (2010), Zahmatkesh and Choobbasti (2010), Black et al. (2011),
were collected and plotted together in Figure 2.15 in order to observe the impact of the area
replacement factor on the bearing capacity of the improved ground. From Figure 2.15 it can
be seen that as spacing between the columns increases, the area replacement ratio decreases,
leading to a decrease in the axial capacity of the column which may lead to increase degree of
settlement. An area replacement ratio from 5 to 25 % could improve the column capacity by
about 30 to 45 % depending on the surrounding soil characterisation; significant improvement
45
occurs when the area replacement ratio between 0 and 20 % and beyond As of 20 % the
increment became very small.
1000
Ultimate bearing capacity of the treated
cu = 30 kPa
800 cu=35 kPa
Kim and Lee (2005); cu=4kPa
Zahmatkesh and Choobbasti (2012); cu=5kPa
ground, (kPa)
600 Wood et al. (2000); cu=10kPa

Black et al. (2011); cu:35kPa
Ambily and Gandhi (2007); cu=30kPa
400 cu=14 kPa
Ambily and Gandhi (2007) cu=14kPa
cu = 5 kPa
200 cu = 4kPa
cu = 10 kPa; Group columns

0
0 20 40 60 80
Area replacement ratio, (%)
Figure ‎2.15 Effect area replacement ratio on the ultimate bearing capacity of the treated ground
The shear strength of the surrounding soil has a large impact on the improved ground bearing
capacity, where at As of 20 %, for example, the column capacity was increased from 170 to
770 kPa by increasing the undrained shear strength from 5 to 30 kPa.
For significant improvement in bearing capacity, Wood et al. (2000) indicated that the area
replacement ratio should be 25% or above, and Ambily and Gandhi (2007) suggested that the
improvement beyond S/D = 3 (i.e. As < 7 %) is negligible.
Black et al. (2011) examined the influence of area replacement ratio and the column length on
the performance of a footing supported on stone column using a developed large triaxial cell.
The cell accommodated a consolidated specimen (cu =35 kPa) of 300 mm diameter by 400
mm height. Granular columns with different diameters of 25, 32, and 38 mm and various
lengths 125, 250, and 400 mm were tested. Results indicated that the settlement improvement
factor any increased with the increase in the area replacement ratio; however, the
46
improvement appeared to have a threshold of about 30–40 % area replacement ratio. For
foundations supported on column groups, the pressure-settlement response was found to be
similar to the individual columns at the same area replacement ratio. Therein settlement
improvement factors for the area replacement ratios of 28 and 40 % were 3.2 and 3.8
respectively. The settlement improvement factors for the corresponding single columns were
about 6.5–7.5 indicating that the performance of the group is not as good as that of the single
columns. Limited results for pressures recorded in the column and in the clay in the group
indicate a stress concentration factor of 1.5.
Bergado et al. (1987) conducted a full scale investigation on six footings by applying load on
single rammed aggregate piers (crushed gravel) with a diameter of 0.3 m and installed to a
depth of 8 m in a site characterized by a 2 m thick layer of overconsolidated clay overlying a
layer of very soft clay with 6 meter thickness. The undrained shear strength ranged from 30 to
40 kPa for the upper clay layer, and from 15 to 25 kPa for the soft clay. Columns were loaded
in maintained stress increments with footings with diameters of 0.3, 0.45, 0.60, 0.75, 0.9, and
1.2 m, representing replacement ratios of 100, 44, 25, 16, 11, and 6 %, respectively. Results
indicated that the settlement improvement factor decreased from 8 to 1.5 as the area
replacement ratio decreased from 100 to 6 %.
2.7.2 Dynamic Loading Condition
2.7.2.1 General soil behaviour under cyclic loading
Dynamic loading applications on the subsurface of soft soil layers induced by vibration
sources (e.g. earthquakes, traffic loads, or offshore waves) can cause a complex stress field, in
which these stresses vary with time (Gu et al., 2012). The stress strain behaviour of clay soils
47
under cyclic loading is dependent on factors such as stress level, drainage condition, type and
rate of loading (Brown, 1996).
Undrained long-term cyclic loading for normally consolidated clays may lead to cyclic
failure. This failure can be identified by number of loading applications in which an arbitrary
predetermined double amplitude failure axial strain is reached (Andersen et al., 1980; Jianhua
Wang et al., 2006; Li-Zhong Wang et al., 2011). Yasuhara et al. (1992) considered residual
pore pressure as an indication of cyclic failure.
Loading frequency is also considered an important issue in cyclic loading. For example,
Ishihara (1996) classified the dynamic problems according to the time of loading as shown in
Figure 2.16. in other wards instances where a load application stays for more than tens of
seconds are considered as static whereas those with a shorter time of loading are considered as
dynamic loading. For example, a shaking wave during earthquakes involves 10 to 20 times
repetition of loads with different amplitudes, and the period of each impulse ranges between
0.1 and 3.0 seconds, which gives a time of loading of 0.02 to 1.0 seconds. On the other hand,
in case of repetitive loads induced by traffic, the soils in the subgrade underneath railways or
road embankments are subjected to a large number of load cycles during the life span. The
time of loading may be deemed on the order of 0.1 second to a few seconds.
48
Figure ‎2.16 Classification of dynamic problems (Ishihara; 1996)
The influence of loading frequency on soft clay behaviour is not well understood. Some
published studies show that both the accumulated pore pressure and shear strain induced by
cyclic loading increases with the reduction in loading frequency (Matsui et al., 1980; Procter
and Khaffaf, 1984; Wang et al., 1998). However, Ansal and Erken, (1989) and Hyde et al.,
(1993) reported different results, indicating that frequency has negligible influence on cyclic
strength and deformation of soils.
2.7.2.2 Stone column studies under dynamic loading
Stone columns have been used to provide the efficient support for different type of
foundations in different soils. They have been used to support infrastructure projects, where
dynamic loading is imposed such as railway and road embankments. Stone columns also have
been used to mitigate liquefaction induced by earthquakes, by increasing the density of the
surrounding soil, allowing drainage, which controls the level of pore water pressure under the
foundation. Also the presents of the stone column will increase the total load carrying
49
capacity and at the same time the stress level on the surrounding soil will decrease (Munfakh,
1984; 1997; Van Impe et al., 1997; Raju, 2003; Adalier and Elgamal, 2004).
Liquefaction mitigation
Liquefaction occurs in loose to medium dense, saturated soils with fairly uniform grain size
distribution, covering the silty sandy range. In this soil condition, the dynamic forces imposed
by earthquake lead to a rearrangement of the grain structure to a denser state. Therefore, if the
drainage is insufficient, pore water pressure will increase and shear resistance of the soil will
reduce (Seed and Booker, 1977; Salahi et al., 2015).
Priebe (1998) presented a method for evaluating the potential impact of liquefaction with
stone columns. This method is similar to some extent to the design of vibro replacement in
static condition which was reported by him in 1976 and 1995.
In the dynamic condition, in order to be more realistic and to simplify the problem, Priebe
(1998) considers deformation of soil with the constant volume to calculate with Poisson’s
ratio µs = 0.5. In this method, the improvement factor n0 is determined using some
simplifications and approximations shown in Figure 2.17
𝐴𝑐 1
𝑛0 = 1 + ∗[ 𝐴 − 1] 3.14
𝐴 𝐾𝑎𝑐 ∗(1− 𝑐⁄𝐴)
∅′𝑐⁄
𝐾𝑎𝑐 = 𝑡𝑎𝑛2 (45𝑜 − 2) 3.15
Where: A is the area within the compaction grid; Ac is cross section of stone column; and ∅′𝑐
is friction angle of column material.
50
Figure ‎2.17 Design chart for vibro replacement (Priebe; 1998)
Baez and Martin (1992) reviewed experimental data on the use of stone column for
earthquake liquefaction mitigation and indicated that a linear mechanism of consolidation is
valid only if the pore pressure ratio remains below 0.5. In addition, they noted that pore
pressure within the gravel drain is not constant, contrary to the assumption that they remain
essentially unchanged.
A case study presented by Rudolph et al. (2011) looked at the effectiveness of using rammed
aggregate piers (RAP) in mitigating the liquefaction potential in a site containing artificial
fills over liquefiable sandy clay soils; and a groundwater level was approximately 3.4 meters
below ground surface. RAP used had a 0.40 meter diameter, 8.5 meters depth and spaced at
2.1 meters centre to centre. The study focused on the densification of the soil surrounding the
RAPs. pre- and post-RAP CPTs measurements were used to calculate the Liquefaction
Potential Index (LPI) of the matrix soil and residual liquefaction and seismic settlement
potential, as it relates to tolerable settlement of the constructed project.
51
Their results indicated that using rammed aggregate piers improved the soil properties of the
site and has been effective at mitigating liquefaction potential to acceptable levels, for
example, within the spread footing area, the reduction in LPI from the pre-RAP CPT versus
post-RAP CPT shows a reduction from very high risk (LPI of approximately 17.8) to low risk
(LPI of approximately 3.2).
Rollins et al. (2006) presented a case history in which prefabricated vertical drains and stone
columns were used in combination to provide treatment for a 4 meters thick layer of
liquefiable silty and sandy silt. They suggested that this method can only be suitable for the
soils with fines percentage lower than 20%, otherwise less effectiveness of this method will
be achieved. Also they reported that using stone column together with wick drain will
significantly improve the penetration resistance by an average of 94 %, where the average
(N1)60-cs increased from a pre-treatment value of 17 to an average post-treatment value of 33.
Embankments:
Building embankments on soft clay soils is usually a very challenging geotechnical task due
to the potential of bearing failure, excessive settlement, and local and global instability under
dynamic load (Mitchell and Jardine, 2002). Vibro stone column is one of the most commonly
adopted soil improvement method, it has been utilized to increase the bearing capacity and
reduce the settlement of superstructures (Raju, 2003).
Kempfert et al., (1999) investigated the bearing and deformation behaviours of a geotextile
encased sand column foundation at area ratios of 4; 12 and 16 % under static and cyclic using
both small scale models with scales of 1:6.5; 1:11.5 and 1:13 and full scale models with scale
of 1 : 1. The cyclic loading was divided into two stages as shown in Figure 2.18, the first
sequence was simulating an embankment with low height (i.e low preloading pressure (25
52
kPa) and high cyclic stresses expected on the soil (62.4 kPa)), whereas in the second cyclic
loading sequence a deep embankment was simulated by applying a low cyclic stresses with an
amplitude of 30.5 kPa and high preloading pressure 100 kPa). At the end of the cyclic loading
stages, the sample was subjected to increment of static loads to failure.
Results showed that at the same level of stresses the system developed a higher settlement
during the cyclic loading compared with the static; also they indicated that during the first
loading sequence the settlement was developed very quickly in comparison with the second
loading sequence. They showed that at a constant Area ratio of 12 %, 100 kN/m2 of loading
stress and a scale model of 1:6.5, using a geotextile coated sand columns can reduce
settlement by more than 30 % compared with a non-coated columns.
Figure ‎2.18 Cyclic loading sequences for railway embankment (Kempfert et al., 1990).
Kolekar et al. (2011) presented results from an experimental study investigating the behaviour
of stone columns under repeated loading. The cyclic test consists of two stages, in stage one
35 % of the static load failure was applied in a form of sinusoidal wave at a frequency of 0.1
Hz for a number of cycles of 500, whereas in the second stage the load was increased to about
53
100 % of the load at the static failure. The study was focused on the vertical stiffness ratio.
Kolekar et al. (2011) claimed that the stiffness ratio was increased dramatically (from 7 to 80
kPa/mm) by moving from stage one to stage two loading condition and suggest that the first
loading stage compacted the column making it more dense, however, they defined the
stiffness ratio as the ratio between the change in the vertical stress to the average vertical
settlement and as the deformation is much smaller in stage one compared with that in stage
two the material stiffness should be larger than that in the second stage.
Fatahi et al. (2012) employed a finite element model using PLAXIS to assess the relation between
the column position beneath the train track and overall settlement of the ballast rail formation. The
study also includes the effect of using geogrid reinforcement (where two layers of geogrid were placed
at the interface between the subgrade and sub-ballast; and sub-ballast and ballast). The model
geometry considers the typical track cross section with concrete sleepers as recommended on the NSW
rail network (i.e. layers depth from top to bottom including sleeper, ballast, sub-ballast, are 150, 300
and 150 mm). The subgrade layer depth was assumed to be 10 m and the gauge length of the track
was 1.4 m. Also train load was considered as 125 kN/m. Stone columns (1 m diameter) were
arranged in rectangular grid pattern at with a spacing of 1.5 m along the rail track.
Results indicated that the overall settlement was reduced by the presence of stone columns spaced
more closely at the centre of the track and not just under the rail (i.e. when the offset distance of the
columns from the track centreline increases, the maximum horizontal displacement of the improved
sub-grade also increases). Also indicated that the optimum column pattern will occur, when stone
columns just overlap in the cross section and located right in the centreline of the track.
In addition, Fatahi et al. (2012) showed that the use of stone columns (no geogrids reinforcement) was
less effective in reducing vertical settlements than geogrids only (no column), when distance of
columns from the centre line exceeds 1 m. Furthermore, the use of two stone columns with geogrids
54
has limited benefits over the use of just geogrids when the distance of columns from the track
centreline is greater than 1.6 m
2.8 Concluding Remarks
Soft clay deposits usually have a low bearing capacity, low permeability, and high
compressibility. Thus it is important that these soils receive sufficient improvement before
construction activities commence. Whilst there are a range of ground improvement techniques
available, the use of stone columns is still regarded as one of the most popular and effective
techniques. Stone columns provide an adequate level of improvement to soft soils by
increasing its shear strength, reducing excessive and differential settlement whilst speeding up
the consolidation progress by shortening horizontal drainage paths. Thus, it has been
successfully adopted for projects such as highway embankment, industrial and residential
structures.
Over the last three decades a considerable amount of research has been conducted to
investigate the behaviour of stone column in soft soils. This chapter reviewed different
aspects (i.e. critical column length, area replacement ratio, surrounding soil strength, internal
angle of friction of the column material, different column material, and column mode of
failure) that have been examined by authors to identify their influence on the behaviour of the
stone column. In addition, this chapter discussed design approaches and construction methods
that are commonly adopted. Almost all the previous study parameters mentioned above were
based on the assumptions that the external loading is applied monotonically with a constant
strain rate during the whole duration of the test. However, in many engineering projects, such
as tank filling and discharging, highway embankments, ocean banks, etc., the surcharge
loading is not applied instantaneously, but changes with time. Therefore, it must be
55
acknowledge that the exact behaviour of foundation on unimproved soft soil and on stone
column/ soft soil composite is not fully understood.
high demands for high speed railway upgrade became a topic of discussion. Modern railway
infrastructure demands have increased not only to provide a high level of performance in
terms of settlements and stability of the railway track but also to reduce maintenance cost. In
areas where loose or soft cohesive deposits are found, ground improvement methods are often
required to ensure the required level of performance. Therefore, is it still viable to consider
stone column as one of these methods?
Bulging or stone column expanding under applied load is important to understand the general
behaviour of vibro stone column foundation, whilst some researchers have studied this topic
under the application of static loading, the question here is would a column bulge in the same
way if the loading sequence was changed (i.e. cyclic loading)? In addition to this, what are the
implications for the threshold stress such improved soils can carry when compared to the
static stress failure? And what is the influence of the loading frequency on the overall
behaviour of improved soils?
The current research was undertaken in an attempt to investigate the above questions and to
provide a better understanding to the behaviour of stone column under cyclic loading
application.
56
Samir Ashour Chapter 3: Methodology
3 CHAPTER 3
3. METHODOLOGY
3.1 Introduction
As noted in the previous chapter, laboratory investigations have played an important part in
development of vibro-stone column technique. Findings of laboratory studies were advanced
trough full scale trials. Thus, mindful of the cost of full scale trials; and in order to further
advance the development of stone columns for cyclic loading a laboratory investigation was
undertaken. This study includes investigation of behaviour of stone columns under both static
and cyclic loading.
Properties of material used together with both the description of apparatus used and procedure
are described in this chapter. It also includes a programme of the laboratory investigation
undertaken.
3.2 Materials
This laboratory scale investigation focused mainly on the fundamental mechanical behaviour
of stone column foundations subjected to cyclic loading. Twenty two laboratory based studies
(see Table 3.1) were reviewed in order to identify best practice and to identify suitable
materials for both the stone column construction and that for the surrounding soil. It was
found that over 50% of the studies were based on Kaolin clay bed because of its high rate of
consolidation compared to that of other fine soils derived from natural deposits. It could also
be used to produce high quality repeatable samples as Kaolin is a processed clay. In terms of
stone column construction, crushed rocks were found to give the highest angle of internal
57
friction (over 40o) and were used by about 50 % of the researchers. Others used sand and river
gravel. Therefore a crushed basalt and kaolin clay were chosen to simulate the column and the
clay bed materials.
Table ‎3.1 Materials used in previous studies
Surrounding
Column material φ' (o) cu (kPa) Reference
material
1 Leighton buzzard sand Kaolin clay 35 19 Hughes and Withers (1974)
2 River gravel Grey silty clay 38 22 Hughes et al. (1975)
3 Gravel Boulder clay - 4.4 Greenwood (1975)
4 Uniform gravel Boulder clay 47, 51, 53 30 Charles and Watts (1983)
5 Quartz Kaolin clay - 14.4-19.1 Barksdal and Bachus (1984)
6 Sand/gravel Bangkok marine clay 35, 43 20-45 Bergado and Lam (1987)
7 River sand Silt 38 - Juran and Guermazi (1988)
8 Loch Aline fine sand Kaolin clay 30 - Hu (1995).
9 River sand - 40.5 - Rajagopal et al. (1999)
10 Stone Ash/ clay 45 40 Watts et al. (2000)
11 - Kaolin clay 35 28 Sivakumar et al. (2004)
Crushed rock, Quarry Transparent clay-type
12 51, 46, 39 17.5-21.5, 32 McKelvey (2002)
waste, Crushed concrete material, Kaolin clay
13 Coarse sand Sand/clay 44 - Ayadat and Hanna (2005)
14 Sand/gravel CH 38, 49 12 Kim and Lee (2005)
15 Crushed basalt Kaolin clay 35 Black et al. (2006)
16 Crushed limestone CL clay 43 30 White et al. (2007)
17 stones CH 43 7, 14, 30 Ambily and Gandhi (2007)
18 Granite chips Lake clay 41 2.5 Murugesan and Rajagopal (2008)
19 Commercial sand Kaolin clay 35 5 Gniel and Bouazza (2009)
20 Quartz Kaolin clay 33 - Najjar et al. (2010)
21 Crushed basalt Kaolin clay 35 - Black et al. (2011)
22 Crushed basalt Kaolin clay 35 - Sivakumar et al. (2011)
23 Gravel Kaolin clay - 41 Cimentada et al. (2011)
24 Gravely sand Oxford clay - 16 Madun et al. (2012)
25 Aggregate Soft clay 41 5 Vekli et al. (2012)
26 Sand Clay 38-45 10.5-16.5 Hanna et al. (2013)
27 Hostun sand and Gravel Kaolin clay - - Frikha et al. (2014)
28 Stone chips Kaolin clay - 6-6.8 Ali et al. (2014)
29 Aggregate Pulverized clay 45 54, 15 Mohanty and Samanta (2015)
58
Physical properties of Kaolin clay were determined in accordance with British Standard BS
1377 (1990) (The tests data available in Appendix A). These properties are summarised in
Table 3.2.
Table ‎3.2 Physical properties of Kaolin clay
Material Property Value

Kaolin clay Specific gravity 2.63
Liquid limit (%) 56
Plastic limit (%) 27
Plasticity index (%) 29
Undrained shear strength (kPa) 12
Crushed basalt aggregate Particle size: (mm) 1.18-2.0
The peak internal angle of shearing, ɸ (degree) 48
Specific gravity 2.70
3.3 Physical Properties of the Model Stone Column and Surrounding Soil Material
3.3.1 Index Properties
Both the Liquid and Plastic limits for the Kaolin clay were determined according to the
procedures in Clauses 4.3 and 5.0 of BS1377: Part 2 (1990) (BSI, 1990). Liquid Limit of
Kaolin was 56%, and the Plastic Limit was 27%. These values are considered typical for
Kaolin clay (John, 2011).
3.3.2 Particle Size Distribution
The particle size distribution of the Kaolin clay material was determined, using a hydrometer
in accordance with Clause 9.5 of BS1377: Part 2 (1990) (BSI, 1990). Results showed that
about of 50 % of the material was in the fine silt range and 50 % was clay (sub 0.002 mm).
The particle size distribution of the crushed rock was determined by using a dry sieve analysis
method in accordance with Clause 9.3 of BS1377: Part 2 (1990) (BSI, 1990), all the material
59
used was between 1.18 and 2 mm. Typical grading curves for both the Kaolin clay and the
crushed basalt are shown in Figure 3.1.
100
Crushed Basalt
Percentage of passing (%)
80 Kaolin clay
60
40
20
0
0.001 0.01 0.1 1 10
Particale size (mm)
Figure ‎3.1 Typical particle size distribution of Kaolin clay and crushed basalt
3.3.3 Specific Gravity
Specific gravity was determined in accordance with Clause 8.3 of BS1377: Part 2 (1990)
(BSI, 1990). The specific gravity of the Kaolin clay ranged from 2.61 to 2.65 with an average
of 2.63, and the values for crushed rock used to form the column, ranged from 2.70 to 2.73
with an average of 2.71. These values fall within the range typical of these material types.
3.3.4 Dry Density/Moisture Content Relationship of Kaolin Clay and Crushed

Aggregate
The compaction test for Kaolin clay was conducted using a 2.5 kg rammer compaction falling
from 300mm (light compaction method). The test was carried according to BS 1377: Part 4
(BSI, 1990). The dry - moisture content relationship is given in Figure 3.2. The maximum dry
density was estimated as 1410 kg/m3 which occurred at an optimum moisture content of 29%.
60
The maximum density of the crushed aggregate was determined in accordance with the
procedure explained in BS 1377: Part 4. 3 kg of saturated crushed materials were compacted
in a mould under a controlled water level using a vibrating hammer. This procedure was
repeated three times to ensure the results repeatability. The maximum density was found to be
1892 kg/m3 with variation of ± 1% which considered acceptable.
The minimum density was determined by pouring the material from 0.5 m height into the
mould without disturbing the soil (BSI, 1990). The minimum density for the crushed basalt
was 1494.5 kg/m3 with less than 0.01kg/m3 variation.
1.50
1.45 Air voids lines
1.40
Dry Density (Mg/m3)
1.35
1.30
0%
1.25
5%
1.20
10 %
1.15
1.10
10 20 30 40 50
Moisture Content (%)
Figure ‎3.2 Dry density-water content relationship for Kaolin clay
3.3.5 Shear Strength
In order to simulate very soft ground that has shear strength below 15 kPa, a series of quick
undrained triaxial tests and hand vane shear tests were conducted on compacted Kaolin clay,
at different water contents ranging from 30 % to 48 %. Shear strength of soil ranged between
61
53 kPa and 5 kPa for moisture content of 33 % to 48 % respectively. Correlation between the
water content and undrained shear strength is shown in Figure 3.3. From these test result, it
can be concluded that soil shear strength of 15 kPa can be achieved with a water content of
42%.
60
Vane shear test

50
Undrained shear strength (kPa)
Quick undrained triaxial test
40
y = 9569.1e-0.159x
30
20
10
0
30.00 35.00 40.00 45.00 50.00
Water content (%)
Figure ‎3.3 Relationship between undrained shear strength and water content for Kaolin clay
Undrained shear strength can be related to liquidity index of soil (Skempton, 1957), as shown
in Figure 3.4, this relationship generally follows the following trend.
𝐿𝐼 ≅ 1 − 0.192 ln 𝑐𝑢 ‎3.1
Vardanega and Haigh (2014) examined a database of 641 fall cone tests in different soil types
in order to determine a statistical relationship between the undrained shear strength and the
liquidity index. Although they used a different method (fall cone tests) to carry out the
undrained shear strength of the soil, a considerable agreement between their statistical model
(Eq. 3.2) and the one presented in the current study was found.
62
𝐿𝐼 ≅ 1.12 − 0.226 ln 𝑐𝑢 ‎3.2
0.8
y=-0.226ln(cu) + 1.12
0.7 R2 = 0.949
0.6
Liquidity index, LI
0.5 y = -0.192ln(cu) + 0.9885

R² = 0.9237
0.4
0.3
Quick undrained triaxial
0.2 Vane shear test
(Varadanega and Haigh, 2014)
0.1
0
1 10 100
Undrained shear strength, cu, (kPa)
Figure ‎3.4 The undrained shear strength – liquidity index relationship
A series of direct shear tests were carried out on dry and wet samples of crushed basalt. In
order to simulate the condition where stone columns are installed in wet ground, the crushed
basalt was left to soak in water for 24 hr before testing. Each test was repeated three times to
ensure result repeatability. All tests were performed in a 100 mm x 100 mm direct shear box,
under vertical stresses of 25 kPa, 50 kPa, 100 kPa and 200 kPa. These normal stresses were
chosen to represent the typical application in the field. The material was placed into the shear
box in three layers; each one was compacted using a small wooden block. The average
density of the compacted material was 1540.6 kg/m3, with variation of about ± 20 kg/m3.
Shearing rate of 1 mm/min was used for all the tests, as suggested in the BS 1377: Part 7
(BSI, 1990), for the aggregate material. During the test, horizontal; vertical displacements and
the shear force measurements were recorded. Typical test results are presented in Figure 3.5.
63
The general behaviour of the crushed basalt was followed a typical pattern for dense granular
material where by a small initial contraction occurs and is followed by dilation.
The peak value of the shear stress was plotted against the corresponding normal stress in order
to determine the failure envelop (see Figure 3.6). for example at 100 kPa effective pressure,
under dry condition, the angle of internal friction was found to be 50o, whilst in wet condition
the internal angle of friction reduced to 48o.
300 300
(a) Wet condition (b) Dry condition
250 200 kPa 250 200 kPa
Shear stress, (kPa)
200 Shear stress, kPa 200
150 100 kPa 150 100 kPa
100 100
50 kPa 50 kPa
50 50
25 kPa 25 kPa
0 0
0 5 10 0 5 10
Horizontal displacement, (mm) Horizontal displacement, (mm)
4.5 4.5
4 (b) Wet condition 4 (b) Dry condition 25 kPa
3.5 3.5 50 kPa

Vertical displacement, (mm)
Vertical displacement, (mm)
25 kPa
3 3 100 kPa
2.5 50 kPa 2.5
2 2
1.5 1.5
100 kP
1 200 kPa 1
0.5 0.5
0 0
0 5 10 15 0 5 10 15
-0.5 -0.5
Horizontal displacement, (mm) Horizontal displacement, (mm)
Figure ‎3.5 Typical stress-strain, direct shear box test (a) wet condition and (b) dry condition
64
300
Wet condition
250
Dry condition
200
Shear stress; kPa
150
100
50
0
0 50 100 150 200 250
Normal stress, (kPa)
Figure ‎3.6 Shear stress verses normal stress at peak
3.3.6 Compressibility Properties
In order to determine the primary consolidation behaviour of Kaolin clay within tests, the
consolidation time, settlement and required load were needed. One dimensional consolidation
tests were carried out using the oedometer apparatus. A small quantity of slurry (1.5 * L.L
water content (Head, 1996; Sivakumar et al., 2004b)) was prepared then placed in the 75 mm
diameter consolidation ring. Considering the condition of the soil (slurry), the specimens were
firstly consolidated under the load of the apparatus hunger only for 24 hour, and then they
were monotonically loaded to pressures of 25 kPa, 50 kPa, 100 kPa, 200 kPa and 400 kPa. At
this stage the samples were unloaded to 50 kPa then reloaded to 800 kPa. Each pressure
maintained for a period of 24 h. During the increment loading period, the settlement of the
sample was carefully measured and the coefficient of consolidation, cv, was determined using
Equation 3.3:
𝑇𝑣 𝑑 2
𝑐𝑣 = 𝑡90
‎3.3
65
Where: Tv is time factor of 90% consolidation and equals 0.848;
D is the drainage path length;
t90 is the time taken for 90% consolidation.
The square root time method was used to determine t90 (Head, 1998). A typical result for the
pressure increment 50 kPa to 100 kPa is presented in Figure 3.7.
The relationship between the void ratio and applied consolidation stresses is presented in
Figure 3.8. The void ratio, e, at the end of each increment period was calculated using
Equation 3.4:
∆e 1 + e0
= ‎3. 4
∆H H0
where: H0 is the initial thickness of the specimen;
ΔH is the change in thickness during each loading increment (H0 - H1);
Δe is the change in void ratio during each loading increment (e0-e1).
e0 is initial void ratio and can be found using the relationship between the water content and
void ratio in the condition of saturated soils,
Thus: 𝑒0 = 𝑤𝐺𝑠 ‎3. 5
Where, w is the water content, and Gs is the specific gravity of the clay.
The coefficient of compressibility, mv, at each pressure increment was determined using the
following equation:
1 𝑒 −𝑒
𝑚𝑣 = ( 0 1)
1+𝑒0 𝜎1 − 𝜎0
‎3.6
66
In Equation 3.6, e0 and e1 are the initial and final void ratio respectively; and σ0, and σ1 are the
initial and final applied stresses.
The coefficient of consolidation, cv, ranged from 7.8 to 27.2 m2/minute, whereas the
coefficient of compressibility, mv, ranged between 0.51 and 3.62 (m2/MN). Both of these
coefficient values are comparable with other typical values of normally consolidated clays
(Head, 1998). The compressibility and consolidation coefficients for each loading increment
are presented in Figure 3.9.
Square root of time, (minutes)
0 5 10 15 20 25 30 35 40
0.00
0.10
Settlement, (mm)
0.20
0.30
0.40
0.50
0.60
0.70
Figure ‎3.7 Consolidation curve square root of time step 50 to 100 kPa
67
1.3
Cc= 0.275,
1.2 Cr= 0.069
1.1
1.0
void ratio, e , (%)
0.9
0.8
0.7
10 100 1000
log σv, (kPa)
Figure ‎3.8 Void ratio / pressure increment relationship
3 30
Consolidation coefficient, (m2/minute)

Compressibility coefficient, (m2/MN)
2 25
(m2/MN)
mv, (m2/MN)
mv,
cv, (mm2/minute)
cv, (mm2/minute)
1 20
0 15
10 100 1000
Δσ́, (kPa)
Figure ‎3.9 Cofficients of consolidation and compressibilty and stress relationship
68
3.4 Test Requirement and Apparatus Design
3.4.1 Scale Effect
An appropriate scale factor should be selected in order to simulate field situations in
laboratory scale test. However, it was not possible in physical model tests undertaken to
maintain a suitable scale factor that satisfies completely all the parameter that governs the
prototype response. In spite of this, using a geotechnical centrifuge, it may be possible to
maintain the prototype stress level (Wood, 2004); however, this approach was unavailable for
this current study due to laboratory limitation.
Hu (1995) noted that parameters affecting response of soil-stone column systems and
influencing the load and settlement relationship can be divided into two categories of major
and minor influence as representing in equation 3.7.
q ult G G γ d γ
⁄cu = f1 (S⁄D , L⁄D , L⁄d , c⁄G , As , ∅′ ) . f2 ( s⁄cu , s⁄γc , g⁄d , DB , c⁄cu ) ‎3.7
s c
Where:
S: the spacing between the column and represented by the area replacement ratio As
d
As = (mS)2 ‎3.8
cu: the undrained shear strength of soil

d: the diameter of the footing
L: the length of the column
d: the diameter of the column
dg: the average particle diameter of stone column material
D: diameter of the unit cell
∅’: the angle of the internal friction of the stone column materials
Gs: elastic shear modulus of the column material
Gc: elastic shear modulus of the clay
ϒs: unit weight of the column material
69
ϒc: unit weight of the clay

m: shape factor which depends on the pattern of columns arrangements.
According to this, some of the main aspects of the field situation such as L/D, L/d, As, and ϕ’
were maintained in the laboratory scale testing. It was considered that this could provide an
adequate response and would generate high enough quality data so that it can be compared
with other studies of similar conditions and would suggest expected behaviour from full scale
columns.
On the other hand, adopting the unit cell concept, and using the relationship between area of
the column to the area of the treated soil then refer it to the diameter ratio (N) as in equation
3.9.
2
A d
As = col⁄A = [ c⁄d ] = 1⁄ 2 ‎3.9
e N
Where: Acol and A is the area of the column and the soft soil area, respectively. Typical value
of this ratio is 5 to 30%. This range results in approximate values of diameter ratio (N)
between 2 and 5. This shows an agreement with (Hughes and Withers, 1974) indicated that
the column can improve the strength of the surrounding soil that falls within a diameter of 2.5
times the diameter of the column. Also Bowels (1988) identified the failure zone extended
radially to a distance of 1.5 times the diameter of the foundation and over a depth of
approximately 2 times the diameter of the column.
To satisfy the above conditions and taking account of the facilities available, the maximum
size of 100 mm diameter foundation could be used. Therefore, N could be used as 3.5. This
leads to the diameter of the column of 28 mm. In terms of area replacement, these values
70
correspond to replacement rates of 7%. With this, assumption the horizontal scale of the
model from reality was about 1 / 20.
3.4.2 Geometrical Dimension of Column
The diameter of the stone column (28 mm) and the spacing (100 mm) were selected on the
basis of a pre-selected area replacement ratio of 7 % as explained before.
The length of sample was determined by taking account of the typical value of length/ column
diameter ratio (L/d), which is generally between 6 and 20. This ratio may reduce to 4 in case
of an isolated stone column (Hughes and Withers, 1974) .
After selecting the diameter and length of stone column, it is important to select an
appropriate particle size of material to be used. Typical ratios of column diameter to average
particle diameter range from 20 to 40, thus granular materials of 2 mm particle sizes were
adopted. The stone column was subsequently tested in two sizes of clay beds as described
below.
Model I: A single 28 mm diameter stone column was installed in a clay specimen of 100 mm
diameter and 200 mm height and subjected to triaxial loading. This set simulated of constant
pressure surrounding the clay. This approach was first used by Hughes and Withers (1974)
and later by Sivakumar et.al. (2004).
Model II: The test is conducted on a single column of 28 mm diameter constructed in clay
bed of 300 mm diameter and 300 mm height. The column was loaded through a 70 mm
diameter steel plate which was 2.5 times the diameter of a single column. The lateral
dimension of the tank (300 mm) was chosen according to the minimum free distance between
the side-line of the column and the wall of the tank.
71
The values of the main parameters relating to columns used in the field and those used in this
study are summarised in Table 3.4. Comparing the typical stone column diameter to the
model diameter implies the scale factor of the model column was about 1/20.
Table ‎3.4 Typical and model values of stone column parameters
Parameter As L d dg
L/d
Unit (%) (m) (mm) (mm)
Typical values 5-35 Up to 20 600-1200 6-20 20-50

Model I 7 0.2 28 7.14 2
Model II 16 0.3 28 10.7 2
3.4.3 General Test Requirements
The test programme involved application of both static and cyclic loading application, in two
models (3.4.2). Investigating the behaviour of soft soil /stone column system under cyclic
loading requires examining the influence of loading frequencies, in addition to the effect of
the cyclic stress ratios. Therefore, in order to fulfil this aim, the test apparatus and procedures
was developed and described below:
 The clay bed for all tests was prepared by consolidating slurry with an initial moisture
content of 84 %. This allowed for a uniform and homogenous soil sample in which
stone columns were constructed.
 Target strength of the clay bed was 12 kPa.
 In order to eliminate the boundary effect of the tank walls on the specimen, a tank
diameter of 300 mm was used (i.e. 4.25 times the model footing diameter (70 mm).
72
3.5 Sample Preparation
The clay bed was prepared by placing kaolin slurry into five plastic tanks. Two tanks were
designed to prepare 100 mm triaxial specimens. The containers were 100 mm in diameter and
400 mm height and the set up was shown in Figure 3.10. Pressures were applied by adding
weights to a load hanger connected to a piston in top of the slurry in four stages giving
pressures equivalent to 12, 25, 50, and 100 kPa. The piston was perforated to allow water to
escape from the specimen.
Three tanks measuring 300 mm in diameter with 550 mm depth were used to create the larger
specimens of 300 mm diameter by 300 mm height and the set up was shown in Figure 3.11.
For safety reasons the large quantity of weights that would be necessary in order to generate
the required pressure could not be used. Therefore, loading was applied via hydraulic jack
attached to a pump connected to air compressor. Proving rings were fitted between the
hydraulic jack and the perforated plate placed on top of the slurry sample to monitor and
control the amount of loading. Two dial gauges were placed on top of each sample to monitor
and record the settlement of the sample as consolidation progressed, also to recognise a
possible plate tilting at an early stage.
73
Dial gage
Frame
Top
drainage
Perforated
Piston
Slurry
Filter
Bench
Bottom
drainage
Weight
Figure 3.10 Model I consolidation process for 100 mm dia. Samples
25 mm thick perforated plate
20mm thick tank wall
Slurry
550 mm
15 mm thick base plate
PPT 3
300 mm
400 mm
Ø10mm
PPT1
PPT2 PPT3
50 mm
100 mm
150 mm
300 mm
Figure 3.11 Model II consolidation chamber for preparing 300 mm dia. Samples
74
3.6 Test Apparatus
3.6.1 Cyclic Load Frame
A 50 kN capacity load frame (Figure 3.12) was used to apply both static and cyclic loading.
Although it is specified that this machine can provide a range of cyclic loading frequency
between 10 Hz and 0.0001 Hz, several trials were conducted to check capacity of the load
frame. Loading frequencies adopted ranged from 0.5 Hz to 3 Hz. The minimum applied load
that was 0.25 kN with minimum amplitude of 0.1 kN.
Cyclic load actuator
Triaxial
Soil specimen
chamber
Data logger
Pores disk
PWP
transeducer
Computer
Soil
Pores disk
Stone column
Filter papers
Figure ‎3.12 Testing equipment
75
3.6.2 Instrumentation
 Load cell: A 10 kN capacity load cell was used for this study. It was calibrated using
SANS electromechanical universal testing machine (CMT 5000 series). A dynamic
test was run in this cell three times for both compression and tension, in order to
ascertain margins of error, which was found to be less than 1 %.
 Pore water pressure transducers: Three WF17060 (Wykeham Farrance) pressure

transducers with 1000 kPa capacity were fitted to the base of the tank for measuring
the pore pressures. These transducers were calibrated before each test using a
cylindrical triaxial chamber full of water. Compressed air was used to pressurise the
water.
 Data acquisition: A logging system supplied by Servocon Digital Control Ltd was
used. to control and monitor the load and pressure transducers. The signals were read
up to 0.001 accuracy, which was considered acceptable for this research. A logging
interval of 0.1 second was used. This leads to a very large number of data, especially
for the static loading test which lasted up to 10 hours at 0.03 mm/min. For cyclic
loading test of data can was recorded every 10 cycles. The large quantity of data was
considered important as it would help to better define the behaviour of specimen under
loading.
3.7 Test Variables
Several factors affect the behaviour of soft subgrade soil subjected to train induced cyclic
loading. these include load frequency, level of applied cyclic stress and stress condition, and
the physical condition of the soil (Brown et al., 1975; Li and Selig, 1998). These are discussed
below.
3.7.1 Loading Frequency (f)
It has been suggested that loading frequencies and number of loading application have a small
influence on the deformation behaviour of the granular materials (Peacock and Bolton, 1968;
76
Brown and Hyde, 1975). However, effect can be significant in case of soft cohesive soils
(Brown et al., 1975; Yasuhara et al., 1982; Zhou and Gong, 2001).
Table 3.5 shows the type of soil and test loading frequency applied by other researchers. This
shows that all the cyclic tests were conducted under triaxial conditions with loading
frequencies ranging between of 0.1 to 5 Hz to simulate the loading condition of traffic and
passing trains. The number of load applications ranged between 500 and 1,000,000 depending
on the material.
Table ‎3.5 Cyclic load testing conditions for previous laboratory investigation
Test condition
Sample
Material Sample σ3’ Frequency Number of Application Reference
size (mm)
preparation (kPa) f, (Hz) cycles
DXL
1 clay 100X100 OCR=(2-20) 38-380 10 1x106 - (Brown et al., 1975)

2 marine clays 35X87.5 CSR=1 0.1-1 - (Yasuhara et al., 1982)
3 marine clays 35X87.5 CSR=1 σa = 200 0.1-3 3600-538200 - (Yasuhara et al., 1992)
σa = 75-120
4 clay 39.1X80 OCR=(1-4) 110-240 0.01- 1 200-3000 - (Zhou and Gong, 2001)
5 Ballast 300X600 compaction 1-240 20 500000 High-speed (Lackenby et al., 2007)
trains
6 Kaolin clay 300X600 anisotropic - 5 3500 PVD (Indraratna et al., 2009)
K0 = 0.6 installation
beneath rail
tracks
7 Marine clay - CSR=1 41 0.01, 0.1, 1 1000 - (Li-Zhong Wang et al.,
2011)
8 Clay 50X100 - 100, 200 0.1, 0.5 400 Traffic loading (Gu et al., 2012)
9 Silt (ML) 76X152 - 100-250 1 100 Traffic loading (Ng et al., 2013)
(25% Kaolin (2 m × 2.2 - - 0.77 100 Subgrade (Sun et al., 2014)
+ 75% m × 2m) stabilisation
Kansas River
sand)
10 Nanjing fine 50X200 - 50-150 0.1 10000 Traffic loading (Cai et al., 2015)
sand
77
Load frequency of a train, at the subgrade level is related to train speed (𝑣), and carriage
length (𝐿), Sekine (1996) developed a relationship to estimate the loading frequency at the
subgrade level:
𝑉
𝑓 = 3.6𝐿 ‎3.10
In this study, a sinusoidal loading wave form was used to simulate train movement with
loading frequencies of 0.5, 1.0 and 3.0 Hz representing train speed of 35, 70 and 225 km/hr
for a typical carriage length of 20 metres.
3.7.2 Amplitude and Dynamic Stress
Based on the depth of ballast and the axle load, there are several theoretical relationships for
determining the pressure at the subgrade level. Perhaps the most well-known is the Talbot
equation (Selig and Waters, 1994).
16.8𝑃𝑚
𝑃𝐶 = ‎3.11
ℎ1.25
where PC is subgrade pressure in psi, Pm is the applied stress on ballast in (psi) and h is the
depth of ballast in (inches).
Japanese National Railway also used similar formula to estimate the stresses at the subgrade
level.
50𝑃
𝑃𝐶 = 10+ℎ𝑚
1.35
‎3.12
where PC and Pm are as defined in equation 3.11 and h is the depth of ballast in cm.
78
Field observation of dynamic stresses conducted by Yoo and Selig (1979) using a four axle
hopper cars each weighing 131 tons, with ballast layer thickness of 300 mm, showed that the
maximum stress on the subgrade level ranged between 80 and 110 kPa (Figure ‎3.13).
Compressive stress
Compressive stress
(kPa)
(psi)
Figure ‎3.13 Dynamic measurement of subgrade stress (Yoo and Selig, 1979)
However, there are other empirical equations where the speed of the train is also taken it
account together with axle load. Zhou, (1996) derived an empirical formula to predict the
dynamic stress at the subgrade level:
𝜎𝑑 = 0.26 𝑃(1 + 𝛼𝑉) ‎3.13
where 𝜎𝑑 =dynamic vertical stress on the subgrade surface (kPa); 𝑃 is the axel load of the
train (kN); 𝑉 is the train speed (km/hr) and α is speed coefficient and equals to 0.005, 0.004
and 0.003 for ordinary grade railways speed, quasi-high speed and high speed railway
respectively.
Furthermore, field measurements in existing railways in China indicates that the average
magnitude of the vertical dynamic stress on the subgrade level for passenger trains is about 46
kPa (Chinese Academy of Railway Science, CARS., 2006). Therefore, for a 225 km/hr train
speed and 100 kN axle load the cyclic stress will be about 43 kPa.
79
A large scale test at the University of Birmingham (unpublished work) included application of
125 kN wheel load applied at 2 Hz for 2 million cycles on rail across three sleepers (moor
information include in Appendix G) indicated that the stresses subgrade level were
approximately 80 kPa (Figure 3.14) showing good agreement with Yoo and Selig (1979)
study.
100
PT1
PT2
Stresses at the subgrade level, (kPa)
PT3
80
60
40
20
0
0 50 100 150
Time, (minute)
Figure ‎3.14 Dynamic stresses mesurments at subgrade level (under 300mm of ballast)
Based on the above, cyclic stresses of 50 kPa, 60 kPa and 70 kPa were used in this study.
These values are equivalent to a cyclic stress ratio (CSR) of 0.6, 0.7 and 0.8 (CSR = the cyclic
deviator stress (qcyclic)/ the static deviator stress of reinforced soil at failure (qfailure)).
80
3.8 Experimental Procedure
3.8.1 Clay Bed Preparation
3.8.1.1 Mixing process
In order to create homogenous bed of soft clay, Kaolin clay was mixed with water content of
1.5 times the Liquid Limit (i.e. water content of 84 %). A Hobart A-120 mixer with capacity
of 12 litres was used for mixing 5 kg of Kaolin with 4.2 litre of de-aired water for each
required batch. Soil and water was mixed a constant speed of 60 rpm for 60 minutes. At the
middle of the mixing process, in order to check the quality of mixing, the mixer was stopped
and the slurry was checked by hand and freed from any created masses (lumps). Once fully
mixed the slurry mixture was stored in sealed plastic containers for 24 hours before using it to
insure saturation. Each batch was just enough to prepare two samples for triaxial testing. For
the large sample five more batches were required to prepare one sample.
3.8.1.2 Consolidation
A one directional consolidation method is considered to be the most commonly adopted
approach by many researchers to create homogeneous clay bed (Juran and Guermazi, 1988;
Sivakumar et al., 2004b; Maakaroun et al., 2009). However, whilst this technique can provide
satisfactory results in short consolidation chambers (< 200 mm) such as the Rowe
consolidation cell (Rowe and Barden, 1966), some difficulties have been reported by a
number of researcher (Anderson et al., 1991; McKelvey, 2002; Ahmadi and Robertson, 2004;
Black et al., 2011) when using this method for preparing larger depth samples. This because
friction occurs between the soil and the walls of the consolidation chamber which could lead
clay particles to arrange themselves into an internal concave structure. In addition, there might
be a loss in consolidation pressure with the depth of the samples. (Valls-Marquez, 2009;
81
Black et al., 2011). To mitigate these problem, as proposed by the previous researchers inner
walls were carefully lubricated with silicon grease which also aids sample extrusion after
consolidation. A hydraulic pump was adopted in order to keep the consolidation pressure
constant when preparing the larger samples.
Preparation of specimens for Model I (triaxial test):
Approximately 4000 g of the prepared slurry was poured into each consolidation chamber to
an initial height of 294 mm in three stages. The cell was vibrated for 30 seconds to eliminate
trapped air.
According to BS 1377-5:1990, under the fully saturation conditions the initial void ratio (e0)
and the initial height (H0) were calculated using equations 3.2 and 3.3, they were 2.209 and
294 mm respectively. Before and after placing the soil, a 100 mm diameter saturated porous
plastic filter (Vyon sheet) with 1.2 mm thickness was placed on top of the slurry surface to
ensure that only water could drain out.
Every load step was held for a minimum of 24 hours and a maximum of 72 hours to ensure
that primary consolidation was complete. Typical consolidation plots are shown in Figures
3.15 and 3.16 (see Appendix A for more information).
The void ratio, e, at the end of each increment period was calculated using equation 3.2. The
relationship between void ratio and applied consolidation stresses is presented in Figure 3.17.
After completion of consolidation, the sample was either directly extruded from the
consolidation chamber (no column installed) or the stone column was first installed and then
extruded. A 100 mm diameter Piston extruder was used to remove the samples in the same
direction as consolidation loading. A steel ruler was used to trim the surface of the specimen
to ensure the flatness and height of 200 mm. In each test samples were taken to measure water
82
content. Filter papers were fitted around the sample and then a rubber membrane was placed
around the sample, before being mounted in the triaxial apparatus.
Square root of time √time, (minute)

0 10 20 30 40 50 60
0
6.45 kPa
Settlement, (mm) 10
20
30
40
50
60
Square root of time √Time, (minute) Square root of time √Time, (minute)
0 10 20 30 40 50 0 10 20 30 40 50
0 0
Settelement, (mm)
12 kPa 25 kPa
Settelement, (mm)
5 5
10
10
15
15
Square root of time √Time, (minute) Square root of time √Time, (minute)
0 10 20 30 40 0 10 20 30 40
0 0
100 kPa
Settlement, (mm)
50 kPa
Settelement, (mm)
5 5
10
10
15
15
Figure ‎3.15 Typical settlement /√time relationship for 100 mm dia. triaxial sample
83
Vertical stress, σv (kPa)
0 20 40 60 80 100 120
0
20
Settlement (mm)
40
60
80
100
Figure ‎3.16 Settlement and total vertical stress relationship for 100 mm dia. triaxial sample
1.8
1.6
Void ratio (e)
1.4
1.2
1.0
1 10 100
log σv (kPa)
Figure ‎3.17 Void ratio/ log σv relationship for 100 mm dia. triaxial sample
84
Preparation of Model II (300 mm specimens):
A similar procedure was followed when preparing larger specimens, with the exception of
drainage which was allowed only from the top in order to measure pore water pressure from
the bottom of the sample at different positions (PPT1, PPT2, and PPT3) as shown in Figure
3.11. As mentioned in Section 3.5 three samples were consolidated simultaneously,
approximately 45 kg of slurry to be poured into each tank giving an initial slurry height of
448 mm. A saturated plastic porous filter was placed on top surface of the slurry. A perforated
25 mm thickness plate was placed at the top of the filter, and used as a consolidation loading
plate. A similar loading sequence was used then adopted to give pressures of 6, 12, 25, 50,
and 100 kPa. After the completion of the consolidation stage, the specimens surface was
flattened using metal ruler with 298 mm length, samples were taken for determining water
content. Stone column was installed in the clay bed (as explained in the following section
3.8.2), and the tank was moved to the loading frame for testing.
Figure 3.18 shows typical consolidation curves for each loading sequence. Accumulated
settlement against the associated loading stress is presented in Figure 3.19.
85
Square root of time, √(minute) Square root of time, √(minute)

0 20 40 60 80 0 20 40 60 80
0 0
Settlement, (mm)
Settlement, (mm)
12 kpa 25 kpa
5
5
10
10 15
Square root of time, √(minute)

0 20 40 60 80 Square root of time, √(minute)
0 0 20 40 60 80
Settlement, (mm)
Settlement, (mm)
50 kpa 0
100 kpa
5
5
10 10
15 15
Figure ‎3.18 Typical settlement / time relationship for 300 mm specimens
Vertical stress, σv (kPa)

0 20 40 60 80 100
0
20
Settlement (mm)
40
60
80
100
120
Figure ‎3.19 Settlement and total vertical stress relationship for 300 mm specimes
86
In the case of 300 mm diameter specimens, settlement during the consolidation stage was
measured together with change in pore water pressure. Load was only increased when pore
pressure had reached an equilibrium value as there was a minimal increase in settlement
beyond this value as can be observed in Figures 3.18. Figure 3.20 shows typical response of
pore water pressure during loading stages during consolidation of slurry. It was difficult to
take measurements in the initial loading stage (7 kPa) where the data logger was showing un-
stabilised readings, this was probably as a result of trapped air in the system. During the
subsequent loading stages, pore water pressure increased to approximately the increment in
pressure. Pore water pressure dissipated and stabilised at about 3 kPa.
50
100 kPa
40
Pore water pressure (kPa)
PPT1
30
12.5 kPa 50 kPa PPT2
PPT3
20 25 kPa
10
0
0 5000 10000 15000 20000
Time (minute)
Figure ‎3.20 Variation of excess pore water pressure at PPT1, PPT2, and PPT 3 positions during
consolidation
87
3.8.2 Column Installation
As stated in the literature review (section 2.4) stone columns can be constructed in the field
either by replacement or displacement method. Wood et al. (2000) and Sivakumar et al.
(2004b) investigated various methods for installing model stone columns in laboratory. They
used methods such as forming the column by freezing the stone material in a mould of the
same size of the required columns, another method was by displacing the soil using a rod with
the same size of the column and then filling the formed cavity by the column material, and
also the most common used method which was done by replacing a part of the soil and
compacting the stones in the cavity. A reduction in the density of the column upon thawing
was found in case of preformed frozen columns, while it was difficult to implement the
displacement technique in a small scale model due to the generated suction that may occur
during the removal of used rod.
In this study the replacement method was adopted because it proved to be repeatable. Thus a
28 mm diameter cavity at the centre of the sample was created by inserting a pipe with this
diameter to the clay bed and removing the soil included in the tube (Figures 3.21 and 3.22).
After creating the cavity, the granular aggregate was introduced in four stages and compacted
using a 1.5 Kg metal rod free falling through a fixed distance of 50 mm. 15 bowls were
applied to each layer. Aggregate between 1.18 mm and 2.1 mm in the size was used to form
the stone column.
The average bulk density of the column was calculated in each test. It was 1900 ± 45 kg/m3. It
is suggested that some of the variation in density may have been due to small increases in
diameter of the column as it was formed.
88
Making a cavity in the soil Introdusing aggregate Extroding the sample
Figure ‎3.21 Sample preparation for Model I
Making cavity and colum Installation
Figure ‎3.22 Sample prepartion for Model II
89
3.8.3 Loading Test
3.8.3.1 Triaxial model test
The prepared sample was located in the triaxial cell, and saturated using increments of cell
pressure and back pressure. A Skempton 𝐵̅ parameter of over 0.95 was achieved. The sample
was then consolidated at an effective confining pressure of 100 kPa with a back pressure of
300 kPa.
Two loading conditions were examined:
i) Static loading: where the specimens were sheared at a constant axial strain rate of 0.03 mm/
minute.
ii) Cyclic loading: load was applied at different cyclic stress ratios (CSR) 0.6, 0.7, and 0.8,
each at three frequencies 0.5, 1 and 3 Hz. The cyclic stress was applied in sine wave form as
shown in Figure 3.23.
In both loading conditions, pore water pressure was measured at the bottom of the specimen.
T = 0.3,1, and 2 (s)

qcyclic =50, 60, and 70 kPa
Deviator stress, q, (kPa)
qmax
q0=20 kPa
Time (s)
Figure ‎3.23 Cyclic stress state
90
3.8.3.2 Large scale model test
Following installation of the column, the specimen was saturated slowly from bottom to top
using piezometer. This process took about 4 hours, in order to ensure that the stone column
was saturated and the air trapped between the particles was forced out, so as not to affect the
quality of pore water pressure measurement beneath the column. Then a sand blanket of 3 mm
thickness was placed covering the entire top surface of both soil and column to ensure that the
stress imposed from the surcharge load will be uniformly distributed to cover the entire area.
The footing was then located at the centre using a guide plate (see Figure 3.24). the sample
was then reconsolidated for 48 hours using a slightly higher pressure (120 kPa) than that used
to create samples (100 kPa), the extra pressure simulated the surcharge pressure on top of the
footing. Following this, in similar manner to the triaxial model, both static and cyclic loading
condition was applied.
Loading ram
Plastic plate
Inflatable rubber ring
Loading foundation
3 mm Sand blanket
Clay bed
300 mm
Stone column
Pore pressure transducer
300 mm
400 mm
Figure ‎3.24 Large test set up
91
3.8.4 Preloading Investigation
3.8.4.1 Water content
Water content in the clay bed can provide a good indication of its strength and uniformity.
This was carried before testing (Figure 3.25) in order to ensure the quality and the general
undrained shear strength of the soil.
A plastic pipe (similar to the one used to install the column) with 28 mm diameter and 45 mm
height was used to take the soil samples for water content test. Soil within the pipe was
loosened and a vacuum was used to lift out the soil from within. This technique can provide
samples along the depth of the clay bed and can be applied in different locations.
As observed in Figure 3.25 the water contents through the clay bed were consistent and
ranged between about 43 and 42.5 %. This range of water content was very close to that
estimated using the relationship of undrained strength and liquidity index (3.3.5). At about 43
% water content, strength was estimated to be around 12 kPa.
Results showed that in most of the tests the variation in water content was comparatively
larger at the middle height of the specimen (± 0.5 %); this could be related to length of the
drainage path.
92
Water content (%)

42 42.5 43 43.5
Depth of clay bed (mm) 0
50
Model I
100
Model II
150
200
Figure ‎3.25 Variation in water content before testing for both Models (I and II)
3.8.4.2 Shear strength of clay bed
In order to investigate the quality of the prepared clay bed, a hand vane was used to determine
the variation in shear strength by testing clay at a range of positions near where samples were
taken for water content determinations. Hand vane tests were conducted in accordance with
BS 1377:1990. A 12.7 mm hand vane was used to limit clay disturbance to minimum. For the
100 mm diameter specimens, due to size of the specimen it was difficult to perform the hand
vane test at more than one location; however, three pilot tests in three specimens were carried
at different depth. In the large sample, the undrained shear strength was determined at four
(50, 100, 150, and 200 mm) depth and four horizontal positions (80 mm around the centre).
This procedure was carried out in three pilot tests only. After that the shear vane test was taken
in each sample before testing only in the centre and after testing in various locations to collect
as much data as possible to observe the changes that occurred before and after testing.
93
Figure 3.26 shows the undrained shear strength values obtained from these tests including
estimated values using Equation 3.2 (3.3.5). The average undrained shear strength was 11.5
kPa with a variation of ± 0.5 kPa. This result was less than the estimated values with a
difference of less than 2 kPa.
The results from both water content and undrained shear strength demonstrated excellent
repeatability of both the mixing and consolidation process.
Shear strength, (kPa)

9.5 10 10.5 11 11.5 12 12.5
0
50
Depth of clay bed, (mm)
100
150
200
Measured undrained shear strength (Hand vane)
Estimated undrained shear strength using Eq.3.1
250
Figure ‎3.26 Shear strength determinations in test bed (pilot stage)
3.8.4.3 Column deformation
Investigating the shape of the column after testing can provide a good indication of the failure
mechanisms and the interaction between the soil and the stone column. As it is not easy to
monitor changes during the test, observation at the end of the test is the only available guide
about the behaviour of the materials. Several techniques were investigated to assess condition
and shape of the column at the end of the test. These included grouting the column material,
94
freezing, using both wax and resin to hold the aggregate material together and section slicing
method.
The simplest was slicing the specimen in half vertically through a selected section. This
method appeared to be a very effective, quick and showed the shape of the column. It was
adopted in this study.
95
Samir Ashour Chapter 4: Monotonic Loading Results and Discussion
4 CHAPTER 4
4 MONOTONIC LOADING RESULTS AND DISCUSSION
4.1 Introduction
This chapter presents the results of both 100 mm diameter triaxial specimens and foundation
loading tests on 300 mm diameter specimens according to the methodology described in
Sections 3.6. The general behaviour of both soft clay (no column) and clay/stone column
composite system are discussed together with the influence of loading rate, and column stone
density on the bearing capacity. The discussion includes results from previous studies with
similar testing conditions in order to validate the current study work. The test programme
adopted for this study is shown in Table 4.1
Table ‎4.1 Test programme
Pc Cu 𝝈′𝟑 As Strain rate Column density

Test type Test No. Hc/Hs comment
(kPa) (kPa) (kPa) (%) (mm/min) (kg/m3)
TRI-C-01 0.003
TRI-C-02 - 0 0.03 -
TRI-C-03 0.3
TRI-C.C-04 0.003
Triaxial Model TRI-C.C-05 100 12 100 0.03

1805
TRI-C.C-06
0.3
TRI-C.C-07 7 1
TRI-C.C-08 1805
TRI-C.C-09 1640
0.03
TRI-C.C-10 1500
LM-C-01 - 0
- 70 mm Footing
LM-C-02 - 0
Large Model 100 12 100 0.03
LM-C.C-03 16 1
1805 70 mm Footing
LM-C.C-04 16 1
96
4.2 Model I (Triaxial)

Two types of samples were prepared one representing the soft soil (no column) and another
representing soil/ stone column composite. These two types of specimens were tested under
the same testing conditions. After extruding the specimen from the consolidation chamber,
they were installed in a triaxial cell for loading. Each sample was first saturated and then
isotopically consolidated under effective confining pressure of 100 kPa, with back pressure of
300 kPa. At this stage drainage was allowed from the bottom of the specimen, as such a
volume change device was connected, allowing volume changes to be recorded during
consolidation. After consolidation, the entire top of the sample was loaded under undrained
condition.
4.2.1 Saturation
The specimen was saturated by increasing the pore water pressure so that any air in the void
space can be eliminated. This allows for reasonable, reliable and repeatable readings of pore
water pressure changes during loading stage (Head, 1998). As described on BS1377: Part
8(1990), there are several established methods for saturation of samples, the one applied in
this study is saturating samples by using the application of back pressures, so that air in the
system is forced into solution. To achieve over 95 % of saturation, the process were carried
out by raising the confining pressure and back pressure in alternate increments as shown in
Table 4.2.
As the tested material is characterized by its low permeability and the size of the sample is
considered large, therefore the time required to reach 95 % of saturation was quite long
between 3 and 5 hours. The stone column reinforced samples took a shorter time to saturate
97
Table ‎4.2 Typical triaxial saturation data
Soil only Soil/ column composite
Cell Back Cell Back

PWP PWP difference PWP PWP difference
pressure pressure B value pressure pressure B value
(kPa) (kPa) (kPa) (kPa)
(kPa) (kPa) (kPa) (kPa)
0 - -1.2 - - 0 - -2.5 - -
50 - 27 28.2 0.56 50 - 35 37.5 0.75
50 40 38 - - 50 40 39.6 - -
100 - 69 31 0.62 100 - 80 40.4 0.81
100 90 89 - - 100 90 89.7 - -
200 - 170 81 0.81 200 - 178.5 88.8 0.9
200 190 188 - - 200 190 188 - -
300 - 283 95 0.95 300 - 284 96 0.96
300 290 288 - - 300 290 289 - -
400 - 384 96 0.96 400 - 387 98 0.98
4.2.2 Consolidation
After the saturation stage was complete the cell and back pressures were adjusted to allow the
sample to be isotopically consolidated under effective confining pressure of 100 kPa and back
pressure of 300 kPa. By allowing the water to drain from the sample to the back pressure
system, the pore water pressure was gradually decreased until it reached approximately the
value of the back pressure. 95% dissipation of pore water pressure was achieved after 24
hours.
The degree of consolidation was determined by expressing volume change at a given time as a
function of the final volume change. As expected consolidation took place quicker when stone
columns were installed compared to soil only (without stone column), where 90%
consolidation was achieved seven times faster, see results presented in terms of degree of
consolidation versus the square root of time in Figure 4.1. This result showed similar
98
behaviour to previous research by Sivakumar et al. (2004b) under similar test conditions.
However, as shown in Figure 4.1, there was a difference in approaching 100 % of
consolidation between this current study and the one conducted by Sivakumar et al. (2004b),
this difference could be due to using the filter papers around specimens used in this current
study as there was no indication that they used such method in their study. In addition to this,
in case of reinforced specimens, the size of the column material particles used and the
resulting void ratio could have a significant influence on rate of consolidation.
100
90
80
Degree of consolidation, (%)
70
60
50
40
30
20 Soil/stone column (Current study)"

Soil_no column (current study)
10 Soil_no columnl (Sivakumar (2004b))
Soil/stone column (Sivakumar (2004b))
0
0 10 20 30 40 50 60
Square root Time (Minutes)
Figure ‎4.1 Consolidation characteristics
4.2.3 Stress - Strain Behaviour
The unit cell approach was used to analyse the load – deformation data, where the column and
the surrounding soil were assumed to act as a single element with a homogeneous distribution
of stresses and strains as proposed by Balaam et al. (1977). Whilst this assumes that the
sample is homogeneous, this approach has proved satisfactory for many practical design
99
applications (Lee and Pande, 1998; Maakaroun et al., 2009; Shahu et al., 2000; Sivakumar et
al., 2010, 2004b).
The influence of the inclusion of stone column in soft clay bed was observed in all tests
curried out. A typical stress – strain relationship is presented in Figure 4.2. The results from
reinforced clay tests are compared with tests carried out on soft clay specimens only. It was
observed that the reinforced samples developed higher bearing capacity than the soil only
samples. The deviator stress at failure for soil samples was approximately 60 kPa, while with
the presence of full length stone column this increased to about 90 kPa. This is an increase of
about 30 % for an area ratio of 7 %. This can considered to be as a substantial improvement,
especially; when many previous researches have reported that the area over 15 % is required
to show significant improvement in the bearing capacity. This significant improvement can be
attributed to higher internal angle of friction of the material used in this study for building the
stone column (48o in wet condition) compared to other studies.
Figure 4.3 shows results of tests conducted by Sivakumar et al. (2004b); Black et al. (2006);
Andreou et al. (2008); Najjar et al. (2010) at similar testing condition (undrained triaxial
condition at the same effective confining pressure) are shown together with results determined
in the current study. In terms of the degree of improvement of bearing capacity, results show
that a higher angle of internal friction would provide a greater improvement ratio at a lower
area replacement ratio. For example, using material with an angle of friction at 48o in this
study (comparing with 35o at the other studies) enhanced the degree of improvement by about
35 % at the same area replacement ratio (7 %).
100
Despite of the limited number of studies on stone columns using undrained triaxial conditions,
it is possible to obtain a relationship between the improvement ratio and the area replacement
ratio as presented in Equation 4.1.
𝐼𝑚𝑝𝑟𝑜𝑣𝑒𝑚𝑒𝑛𝑡 𝑟𝑎𝑡𝑖𝑜 = 4.003𝐴𝑠 − 10.831 𝐴2𝑠 ‎4.1
100
80
60
40 Reinforced soil
Soil/column (current
(current study)
study)
Reinforced
Soil/columnsoil (Sivakumar
(Sivakumar et al,
et al, 2004)
2004)
Soilonly
Soil only(Sivakumar,
(Sivakumar 2004)
et al, 2004)
20
Soil only (current study)
0
0 5 10 15
Vertical Strain, (%)
Figure ‎4.2 Stress strain behaviour of soft soil (no column) and soil/stone column composite.
0.4
0.35
Improvement ratio
0.3
Y = 4.0034(As) - 10.831(As)2
0.25
R² = 0.7338
0.2 Sivakumar et al. (2004)

Black et al. (2006)
Andereou et al. (2008)
0.15
Najjar et al (2010)
Current study
0.1
0 0.05 0.1 0.15 0.2
Area ratio
Figure ‎4.3 Improvement ratio versus area replacement ratio
101
In all tests, pore water pressures were measured, and expected application of stress on low
permeability soils induces excess pore water pressure. Through the process of consolidation,
pore water is reduced resulting in a gradual increase in effective stress. The speed of
consolidation amongst other factors depends on drainage path. Installation of stone column
leads to a reduction in drainage path resulting in faster consolidation.
Figure 4.4 shows that the pore water pressure in all cases increased tending towards their
highest value as the deviator stresses reaches their maximum value at axial strain of about 5
%. There was a reduction in pore water pressure at about 15 % in case of the reinforced soils.
This reduction during the undrained loading stage is possibly due to the dilation of the stone
column.
80
Increase in pore water pressure, (kPa)
60
40
Reinforced soil
Soil/column (current
(current study)
study)
20 Soil only (current study)
Reinforced soil
Soil/column (Sivakumar
(Sivakumar et al, 2004)
et al,2004)
Soil only (Sivakumar, 2004)
(Sivakumar et al, 2004)
0
0 5 10 15
Figure ‎4.4: Excess pore water pressure of soil (no column) and soil/ stone column composite.
The vertical deviator stresses at failure for both unreinforced and reinforced specimens under
undrained conditions are presented in Table 4.3, for tests under similar condition to this study.
The improvement occurred in terms of bearing capacity, which ranged between 27 and 56 %,
102
and was mainly dependent on the area ratio and the internal angle of friction of the column
materials.
The estimation of vertical stress at failure of a single column is included in Table 4.3. using
the expression proposed by Hughes and Withers (1974):
1+sin ∅′
σv = ( ) (4c + σr′0 ) ‎4.2
1−sin ∅′
In the above equation σr′0 was considered firstly equivalent to 2𝑐𝑢 as suggested by Hughes
and Withers (1974), and secondly was considered to be equivalent to the effective confining
stress in the test (100 kPa). These calculated values are presented in Table 4.3 as
σv(calculated) (1) and (2) respectively. Whilst both assumption provide an over estimation of
values for all tests, it shows that the first assumption (σr′0 =2𝑐𝑢 ) would provide a closer
prediction (within an average of 25 %) than considering the amount of the effective stress.
Table ‎4.3 Measured and calculated vertical stress obtained from different studies
As σ3 σ3' quntreated qtreated uuntraeted utratead σv(measured) σv (calculated), (kPa)

Reference Φ(o) qratio
(%) kPa kPa kPa kPa kPa kPa kPa Using
(1) (2)
Eq 4.1
Sivakumar et
35 0.10 400 100 60 83 56 47 1.38 483 664 812 478
al. (2004)
Black et al.
35 0.17 400 100 160 254 - - 1.59 654 775 1550 619
(2006)
Andereou et
35 0.04 300 100 51 65 65 51 1.27 368 554 738 357
al. (2008)
Najjar et al
33 0.08 410 100 64 74 61 50 1.16 484 651 773 490
(2010)
Current study 48 0.07 400 100 60 88 53 44 1.46 488 529 1004 474
After each test, the specimens were removed from the cell and split vertically along their
diameter to examine the mode of failure. It was expected that the columns would bulge at a
depth of 2 to 3 times the column diameter as suggested by Hughes and Withers, (1974). In all
cases the column bulging was found to be relatively uniform with the depth of the sample as
shown in Figure 4.5. This probably was due to the constant confining pressure along the depth
103
of the specimen (triaxial test condition), whereas in the field condition the confining stress
increases as the depth of the soil increases.
200
150
100
50
0
-50 0 50 100 150
Figure ‎4.5 Shape of failure under static loading condition
4.2.4 Effect of Strain Rate
4.2.4.1 Stress-strain characteristics
Strain rate has been recognised as having a significant influence on the undrained shear
strength of saturated soils (Kimura and Saitoh, 1983; Nakase and Kamei, 1986). Skempton
and Bishop (1954) summarised that the strength of cohesive soil are generally sensitive to the
change in strain rate and any reduction in the strain rate would decrease the strength of the
soil, whereas the granular soils are generally independent of this factor.
The strain rate for shearing saturated soils under undrained condition can be determined using
the following expression:
𝜀𝑓 𝐿
𝑅𝑎𝑡𝑒 𝑜𝑓 𝑠𝑡𝑟𝑎𝑖𝑛 (𝑚𝑚⁄𝑚𝑖𝑛𝑢𝑡𝑒) = 100𝑡 ‎4.3
𝑓
104
Where: 𝜀𝑓 is the strain at failure and typically ranging from 15 to 20 % for normally
consolidated clay; and 𝐿 is the specimen length (mm); and finally 𝑡𝑓 (minute) is the time to
failure which is calculated from 𝑡100 which is the time required to ensure at least 95 %
equalisation of pore water pressure in the specimen (Head, 1996).
The axial strain rate for consolidated undrained triaxial test with pore water pressure
measurement was carried out using the above expression and it was found to be 0.03 mm/
minute. This study included investigation of the influence of the strain rate on the undrained
shear strength of both soft soils and stone column reinforced soft soil.
Three strain rate values were applied to cover the range recommended for cohesive soils (0.3,
0.03, and 0.003 mm/minutes). Each test was repeated at least two times to ensure repeatability
in results.
Figure 4.6 (a) and (b) show the typical stress-strain curves of both treated and untreated soils.
Considering the load capacity of the both soils to be at 5 % of the specimen strain (equivalent
to 10 % of the foundation width) as suggested by Bowles (1996). Results show that there is
about increase of about 10 % change in deviator stress associated with increasing the
magnitude of rate of strain from 0.003 to 0.3 mm/minute in unreinforced specimens. In case
of reinforced specimens the trend was in reverse order where the deviator stress showed an
increase by 8 % with the reduction of the strain rate.
Conversely, if failure zone was considering to be at 8 % of the deformation (i.e. 58% of stone
column diameter (Hughes and Withers, 1974)), the deviator stress in the reinforced
specimens was almost fall toward same point, this could be due to the drainage path provided
by the presents of the stone column which results in reduction in the effect of pore water
105
pressures. Whereas in case of unreinforced soils the deviator stress was continued to increase
with the increase of strain rate.
This suggests that with or without stone column results are not very sensitive to rate of strain.
This may be due to the fact that the soil examined is kaolin clay which has a high percentage
of silt content (about 50%). In addition to this it is worth noting that the soil with stone
column is less sensitive to change in rate of strain.
80
60
Deviator stress, (kPa)
40
0.003 mm/min
20 0.03 mm/min
0.3 mm/min
0
0 2 4 6 8 10 12 14
Vertical strain, (%)
(a) Soil (no column)
100
80
60
40
0.003 mm/min
20 0.03 mm/min
0.3 mm/min
0
0 2 4 6 8 10 12 14
(b) Soil/ Stone column composite
Figure ‎4.6 Typical stress-strain curves: (a) soil (no column); (b) soil/stone column composite
106
4.2.4.2 Excess Pore Water Pressure-Strain Characteristics
The typical observed relationships of the change in pore water pressures and the axial strains
for both soil only and stone column/soil specimens for strain rates of 0.3 to 0.003 mm/minute
are given in Figure 4.7 (a) and (b). Results show that there is an approximate 20 % increase in
pore water pressure for reduction in speed of second order of magnitude for clay only
specimens. For the case of stone column reinforced soil, change in pore water pressure was
about 8 %. This was most likely due to shorter drainage path. Thus, the effect of rate of
testing stone column reinforced soil can be discounted when assessing their behaviour in case
of silty soil such as kaolin.
Figure 4.8 shows the change in pore water pressure- strain rate at the peak deviator stress
level. The pore water pressure was normalised in terms of the effective confining pressure
(𝜎3′ ). Findings tend to confirm that reduction in pore water pressure at tested strain rate as
observed in this study. It also showed the difference in pore water pressure between treated
and untreated soils.
60
Change in pore water pressure, (kPa)
40
20 0.003 mm/min
0.03 mm/min
0.3 mm/min
0
0 2 4 6 8 10 12 14
(a) Soil (no column)
Figure 4.7(a) Typical change in pore water pressures-strain curves for soil only
107
60

40
0.003 mm/min
20
0.03 mm/min
0.3 mm/min
0
0 2 4 6 8 10 12 14
(b) Soil/Stone column composite
Figure ‎4.7(b) Typical change in pore water presures-strain curves for soil/ stone column
composit
0.6
0.5
Δu/σ'₃
0.4
Soil only
soil/stone column
SivaKumar et al. (2004)
SivaKumar et al. (2004)
Andreou at al. (2008)
Maakaroun et al. (2009)
0.3
0.001 0.01 0.1 1
Strain rate, (mm/minutes)
Figure ‎4.8 Normalised excess pore water pressure – strain rate relasionship
108
4.2.4.3 Rate Effect on Soil Modulus
Young’s modulus of soil (E) is the proportion between the changes in deviator stress to
change in axial strain (Powrie, 2013).
∆𝜎
𝐸= ∆𝜀
‎4.4
Secant modulus, which is defined as the slope of the secant drawn from the origin to a
particular point on the stress-strain curve (Schanz et al., 1999), can provide a convenient
measure of soil stiffness. This modulus is an important parameter for the design of projects
such as roads and railways, where it is used to predict settlement (Selig and Waters, 1994).
During this study the effect of rate of strain and inclusion of stone column on soil stiffness
were investigated by tracking the secant modulus behaviour of both reinforced and
unreinforced soils at different stress between 50 and 100% of the soils shear strength.
Figure 4.9 shows the relationship between the secant modulus of both soil only and soil
reinforced by stone column and the strain at different strain rates. Other data from different
studies are also included in this graph. Generally the deformation modulus of soil showed a
independency on the tested strain range for both treated and untreated soils specimens. In
addition to this, there is a trend of reduction in deformation modulus with increase in strain to
about 4 %. After that a very small reduction was noted for both soil and soil with column.
In case of soil/stone column composite, the secant modulus was increased by about 30%
compared to soil (no column). This improvement was higher between 1 and 3 % strain. For
strains higher than about 4 % the secant modulus decreases and the degree of improvement in
modulus decreased by about 17%. These trends showed a very close agreement with results
109
conducted by Sivakumar et al. (2004b) and Andreou et al. (2008) in terms of percentage of
improvement in the deformation modulus, but in lower modulus value.
10 Esoil -0.003 mm/min

Esoil/col -0.003 mm/min
9 Esoil -0.03 mm/min
8 Esoil -0.3 mm/min
Esoil _0.005 mm/min_SivaKumar et al. (2004)
Secant modulus, (Mpa)
7
Esoil/col _0.005 mm/min_SivaKumar et al. (2004)
6 Esoil _0.03 mm/min_Andreou at al. (2008)
Esoil/col _0.03 mm/min_Andreou at al. (2008)
5
4
3
2
1
0
0 1 2 3 4 5 6
Axial Strain, (%)
Figure ‎4.9 Soil modulus of both soils only and soil/ stone column composit
In order to relate soil stiffness to road and railway applications results were used for
evaluation of subgrade reaction, which is used mainly in the design of foundation under wheel
and concentrated loads. The modulus of subgrade reaction (𝑘𝑠 ) can be defined as the relation
between soil pressure (𝑞) and deflection(𝛿) (Bowles, 1996).
𝑞
𝑘𝑠 = 𝛿
‎4.5
Bowles (1996) provided an empirical equation to estimate the modulus of subgrade reaction
using the allowable bearing capacity qa.
𝑘𝑠 = (40, 50, 83, 𝑜𝑟 160) ∗ (𝑆𝐹) ∗ 𝑞𝑎 ‎4.6
Where: Factors (40, 50, 83, and 160) are dependent on the settlement(∆𝐻) at 25, 20, 12 and 6
mm respectively; and
110
𝑞𝑢𝑙𝑡
𝑞𝑎 = ⁄𝑆𝐹 ‎4.7
Therefore, the subgrade modulus of reaction for both soil (with no column) and soil/ stone
column composit were compared for a settlement of 1.25 mm as suggested by the UK
Department of Transport (as for pavement foundations the modulus should be limited to small
strains in the elastic range) and at 6 mm and 10 mm as suggested by Bowles (1996) for
foundation design. Results illustrated in Figure 4.10 (a, b, and c) show that the strain rate has
no significant effect on modulus of subgrade reaction for both soils (with and with no
column) for each level of modulus deformation. However, the installation of the stone column
increased 𝑘𝑠 by about 30 % compared with the soil only specimens; this improvement was
observed in all deformation levels. It was also observed that the subgrade modulus of reaction
was decreased with the increase of deformation level.
A comparison between test results and the calculated modulus of subgrade reaction using
equation (4.3) can be seen in Table 4.4 which shows good agreement as the empirical
equation is dependent only on the ultimate stresses.
Table ‎4.4 Modulus of subgrade reaction
Calculated Lab result Calculated Lab. Result

Settlement 𝒒𝒖𝒍𝒕−𝒔𝒐𝒊𝒍 𝒒𝒖𝒍𝒕−𝒔𝒐𝒊𝒍/𝒄𝒐𝒍
(ks)soil (ks)soil (ks)soil/col (ks)soil/col
(mm) (kPa) (kPa)
(kN/m2/m) (kN/m2/m) (kN/m2/m) (kN/m2/m)
6 61 9760 10000 90 14400 14500
10 61 5063 6200 90 7470 9000
111
35 Soil only
at1.25 mm deformation Thousands

Soil-Column composit
30
25
Ks, (MN/m2/m)
20
15
10
0
0.003 0.03 0.3
(a)
Strain rate, (mm/minutes)
35
at 6 mm deformation Thousands
30
25
Ks, (MN/m2/m)
20
15
10
0
0.003 0.03 0.3
(b) Strain rate, (mm/minutes)
35
at 10 mm deformationThousands
30
25
Ks, (MN/m2/m)
20
15
10
0
0.003 0.03 0.3
(c) Strain rate, (mm/minutes)
Figure ‎4.10 The subgrade modulus of reaction (𝑘𝑠 ) for both soils (with no column and soil/ stone
column composit) at a) at settlement of 1.25 mm; b) at settlement of 6 mm; and c) at settlement of
10 mm
112
4.2.5 Effect of Column Density
Most of the widely adopted design methods used for predicting bearing capacity of stone
columns are mainly dependent on two factors, shear strength of the surrounding soil and the
internal angle of friction of the column material (as discussed in Section 2.6). In the specific
case of constant surrounding soil properties, the friction angle of a fill material (i.e. granular
soil) plays a crucial role in estimating stresses that the soil/ stone column composite can carry.
This friction angle is dependent on many factors such as the magnitude of stresses; relative
density; and on the size and shape of soil particles (Bolton, 1991).
The impact of density on the shear strength of granular material has been widely investigated,
however, very limited information on the influence of this factor on the stone column
application is available. This might be due to that the final diameter of the column is only
roughly known which makes the estimation of density unreliable.
Generally, for a given material, the internal angle of friction decreases with the reduction in
density, which in itself is substantially influenced by the particle size distribution.
In this study, columns were constructed with three different densities 1805, 1642, and 1500
kg/m3 representing relative densities of 80, 42, and 2% respectively. All tests were conducted
at the same effective confining pressure of 100 kPa and same strain rate of 0.03 mm/minute.
The stress-strain results are shown in Figure 4.11. These results show that the degree of
improvement was influenced by the increase column density.
Firstly, for columns made at the loose density (1500 kg/m3) (columns were installed by
pouring the aggregate from 500 mm height with no compaction was applied (simulating a
condition close to the minimum dry density of 1490 kg/m3, (maximum void ratio))) there was
about 25 % improvement.
113
At the second test where the density was increased by applying compaction to the column (the
material was filled in two layers and 10 tamps were applied each time by 1.5 kg rod) the
improvement was increased to 28 %.
At the final test the column material was compacted to achieve the a density close to the
maximum density, the deviator stress reach its maximum at about 90 kPa which is about 31 %
improvement in bearing capacity.
The variation of bearing improvement ratio qt/qunt versus the relative density Dr is shown in
Figure 4.12. This indicates that the relative density may have a significant influence in
increasing the strength of soil/column system and accordingly the bearing capacity of the
ground.
100
80
60
40
Soil only
1805kg/m3
1642kg/m3
20
1500kg/m3
0
0 2 4 6 8 10 12 14
Figure ‎4.11 The stress-strain for different column densities
114
1.8
Improvement ratio, (qt/qunt) y = 0.0037x + 1.2714

1.6
R² = 0.972
1.4
1.2
1
0 10 20 30 40 50 60 70 80 90
Relative density, (%)
Figure ‎4.12 Improvement ratio versus the relative density
The pore pressure behaviour showed independency from the change in column density as can
be seen in Figure 4.13 which shows the measurements of pore water pressures against the
axial strain.
60
40
1805 kg/m3
20
1642 kg/m3
1500 kg/m3
Soil only
0
0 2 4 6 8 10 12 14
Figure ‎4.13 Change of pore pressure with column density
115
Results of this study together with these of other researchers are shown in Figure 4.14 in
terms of maximum deviator stress to effective confining pressure ratio (q/𝜎3′ ). Results show a
general trend represented by the following equation:
𝑞
𝜎3′
= 0.056 (𝛾𝑑 ) + 0.0972 ‎4.8
y = 0.056x - 0.0972
R² = 0.9836
0.8
y = 0.0475x - 0.0124
R² = 0.5157
q/σ'3
0.6
Current study
Juran et al. (1988)
0.4 SivaKumar et al. (2004)
Najjar et al. (2009)
0.2
13 15 17 19
Column density, kN/m3
Figure ‎4.14 normalised deviator stress against column density
To conclude any reduction in relative density of the column leads to a reduction in the bearing
capacity of about 14 % (midpoint of upper and lower bounds). This could be due to a
reduction in the peak internal angle of friction of the column material and this might be
related to the increase in the dilation angle of the column particle. Therefore, it might be ideal
when designing stone columns to use the critical angle of friction which is function of both
the peak shear angle and the angle of dilatancy.
116
4.3 Model II (Large Scale Test: Foundation Test)

Specimens were prepared and tested under the conditions explained in Sections 3.8.1 to 3.8.3.
The specimens were tested under a confined pressure of 120 kPa which was maintained
during the whole test procedure, aiming to simulate the surcharge on the ground in the field.
The model foundation was loaded at a constant strain rate of 0.03 mm/min, because It was
concluded from the previous tests (i.e. triaxial model) that this rate was sufficient to allow
excess pore water pressures to dissipate. The test was stopped when the footing displacement
reached about 40 mm (15 % of deformation).
Three pore water pressures transducers placed in different locations at the base of the
specimen as (see Section 3.5) monitored readings throughout.
The main aim of these large tests was to investigate the condition within which the lateral
pressure distributed along the depth of the column (typically increase with the increase in
depth (Bowles, 1996)). Thus the mode of failure is investigated in more detail than that in the
triaxial condition.
4.3.1 Load - Displacement Relationship
Typical load-displacement relationship for both the reinforced and unreinforced soils are
presented in Figure 4.15. The maximum load carrying capacity of soil (no column) specimen
was found to be about 0.7 kN (equivalent to bearing pressure of 200 kPa). Penetration of 35
mm was required to achieve this. This bearing capacity was close to the predicted value of
171.6 kPa (16 % under estimated) based on the following relationship:
qult = cu Nc + qNq ‎4.9
117
Where cu is the undrained shear strength of the soft soil (12 kPa), and Nc, Nq are the bearing
capacity factor (value of 5.14 and 1 respectively for zero internal angle of friction (φ = 0)) and
q is the surcharge pressure.
This value was approximately doubled when the 28 mm diameter column was installed
beneath the foundation of 70 mm diameter. Bearing capacity of the stone column reinforced
foundation was approximately 435 kPa (1.7 kN load capacity).
500
400
Foundation stress, (kPa)
300
200
100
Soil/column
Soil only
0
0 2 4 6 8 10 12 14
Axial strain, (%)
Figure ‎4.15 stress-strain relationship
Bearing capacity estimation
A reasonable agreement (within 13 %) between measured and predicted values was achieved,
which using the expression proposed by (Hughes and Withers, 1974) (Equation 4.10). This
was not surprising due to the fact that the empirical method considered only the influence of
strength of the surrounding soil and the degree of the internal friction, but not the effect of
118
area replacement ratio and the column density as discussed in chapter two (sections 2.7.1.2-
2.7.1.4).
1+sin ∅′
σv = (1−sin ∅′ ) (4𝑐𝑢 + σr′0 ) ‎4.10
Where ∅′ is the friction angle of the stone column material; σr′0 is the effective in situ lateral
stress and assumed to be 2𝑐𝑢 ; and 𝑐𝑢 is the undrained shear strength of the soil and equals to
12 kPa. The predicted value was 485 kPa assuming a factor of safety of 1 to allow the
comparison.
In addition to this, the degree of improvement was comparable with other researcher’s results
as can be noted in Table 4.5. The load bearing capacity ratios for before and after
reinforcement over similar test conditions ranged between 2 and 4.
Results included laboratory, field, and numerical studies on the improvement provided by
stone column are summarised in Table 4.5. Most of methods used to predict the bearing
capacity of a stone column considered the soil shear strength and the angle of shearing
resistance of the column material and ignoring the influence of column diameter (area
replacement ratio).
This database was used to developed a simple linear model to predict the ultimate bearing
capacity of reinforced soil as a function of cu, As, and ∅. (more information available in
Appendix D)
Figure 4.16 shows a scatter plot matrix between all variables. The bottom row of this scatter
plot matrix gives the scatter plots of qult against each of the other three input variables.
Additionally, the univariate relationship between our outcome variable qult and the input
variables cu, As, and ∅.
119
Table ‎4.5 Published database
cu, Φ Dc AS, q untreated q treated qtreated/

Reference (kP (o) Sample size mm
mm (%) (kPa) (kPa) quntreated
a)
Wood et al. (2000) 10.5 25 17.5 D= 300 30 84 105 1.25
Wood et al. (2000) 10 25 11 D= 300 24 84 90 1.07
Wood et al. (2000) 10 25 11 D= 300 10 84 86 1.02
Andereou et al. (2008) 20 32 20 D= 100, H= 200 4 55 62 1.13
Najjar et al. (2010) 20 33 20 D= 70.9, H= 141 7.9 62 75 1.21
Najjar et al. (2010) 20 33 30 D= 70.9, H= 141 17.8 67 100 1.49
Najjar et al. (2010) 20 33 20 D= 70.9, H= 141 7.9 84 101 1.20
Najjar et al. (2010) 20 33 30 D= 70.9, H= 141 17.8 84 148 1.76
Sivakumar et al. (2004) 25 35 32 D= 100, H= 200 10.2 60 83 1.38
Black et al. (2006) 25 35 25 D= 300, H= 400 17 1.25 2 1.60
Bergado et al. (1987) 20 35 300 Field Test 6.25 175 221 1.26
Hughes and Withers (1974) 19 35 38 225*160*150 40 171 418 2.44
Juran and Guermazi (1988) 30 38 20 D= 100, H= 200 4 120 154 1.28
Ali et al. (2014) 7 38 30 D= 300, H= 550 25 25 70 2.80
Kim and Lee (2005) 4 38 22 250*100*250 30 30 75 2.50
Kim and Lee (2005) 4 38 22 250*100*250 40 45 135 3.00
Kim and Lee (2005) 4 38 22 250*100*250 50 45 155 3.44
Bergado et al. (1987); st. Helens 30 42 600 Field Test 65 60 270 4.50
Bergado et al. (1987); Canvey 20 42 - Field Test 28 95 240 2.53
Humber Bridge 25 42 - Field Test 11 115 270 2.35
Island
Black et al. (2011) 35 43 25 D= 300, H= 400 17 240 680 2.83
Black et al. (2011) 35 43 23 D= 300, H= 400 28 230 750 3.26
Black et al. (2011) 35 43 38 D= 300, H= 400 40 230 820 3.57
Bergado et al. (1987) 20 43 - Field Test 6.25 175 370 2.11
Ambily and Gandhi (2007) 7 43 100 D= 830, H= 450 19 75 160 2.13
Zahmatkesh and Choobbasti (2012) 5 43 800 F.E. PLAXIS 10 80 165 2.06
Watts et al. (2000) 40 45 600 Field Test 44 12 248 20.67
Current study (Model I) 12 48 28 D= 100, H= 200 5 55 90 1.63
Current study (Model II) 12 48 28 D= 300, H= 300 13 140 428 3.06
Kim and Lee (2005) 4 49 22 250*100*250 30 40 150 3.75
Kim and Lee (2005) 4 49 22 250*100*250 40 40 165 4.13
Kim and Lee (2005) 4 49 22 250*100*250 50 40 215 5.38
120
φ qult
cu Cu
φ
φ
As AS
qult qult
Cu AS
Figure ‎4.16 Scatter plot of all the variables
Regression analysis is a powerful technique that can be used to address various research
questions. In this report, this was used to check how q levels are affected by As; cu and ∅. In
particular, the type of regression we are going to use is
Multiple linear regressions were used to process the data presented database, in order to
determine the best-fitting line through the data points (this line is sometimes referred to as the
regression line). Multiple means we have more than one input variable (also known as
predictor); hence, we are trying to fit a plane or hyper-plane rather than a line (i.e. Combining
all the input variables by conducting a coefficient multiplied by each variable, and then
summing these variables). The idea was to use a linear combination of input variables to
121
model their relationship with an output variable (qult). After using R's lm() function, the model
looks as follows:
𝑞𝑢𝑙𝑡 = −609.893 + 11.313 ∗ 𝑐𝑢 + 3.167 ∗ 𝐴𝑠 + 14.629 ∗ ∅ ‎4.11
By examining equation 4.11, we observe that when fixing ∅ and AS, an increase, or decrease,
of cu by one unit, causes an increase, or decrease, in qult by 11.313 units. Similarly, when
fixing cu and AS, an increase, or decrease, of ∅ by one unit, causes an increase, or decrease, in
𝑞𝑢𝑙𝑡 by 14.629 units. Also, when fixing cu and ∅, an increase, or decrease, of AS by one unit,
causes an increase, or decrease, in qult by 3.167 units.
Table 4.6 shows the values of four diagnostics that were used to examine the goodness of fit
of a model. Namely, these were: Deviance, BIC, R2, Adjusted R2 and the Residuals (these
factors were defined in Appendix D). As shown in this table the regression model has a
coefficient of determination R2 of 0.44, in other words 44% of the variation in the data can be
explained using the best fitting represented by Equation 4.11.
Table ‎4.6 Deviance, BIC, R-Squared and Adjusted R2 values
Deviance BIC Value R2 Adj R2
682279.8 415.1196 0.4448661 0.3831845
These data is also presented in Figure 4.17. The ultimate bearing capacity of the reinforced
soil measurements were normalised by dividing it by the soils bearing capacity before
reinforcement and plotted against the area replacement in terms of internal angle of friction of
the column material creating a linear relationships between the bearing capacity ratio and the
area replacement ratio at different internal angle of friction as presented in equations 4.12-
4.15:
122
𝑓𝑜𝑟 ∅ 𝑏𝑒𝑡𝑤𝑒𝑒𝑛 (33𝑜 − 35𝑜 ); 𝑞𝑟𝑎𝑡𝑖𝑜 = 3.6 ∗ 𝐴𝑠 + 0.91 ‎4.12

q treated/q untreated, (ratio)
Φ=(33-35o), q ratio = 3.6* As + 0.91
Φ=(37-39o), q ratio = 4.1* As + 1.4
Φ=(42-43o), q ratio = 4* As + 2
Φ=(48-49o), q ratio = 4.3* As + 2.5
As, (%)
Figure ‎4.17 Relationship between qratio and As in terms of ∅
Settlement estimation:
Various design approaches used for estimating settlement of a stone column – soil system, as
explained in the literature (section 2.6.2). They are based on the unit cell assumption. Priebe’s
method for settlement prediction is one of these approaches and considered to be the most
123
commonly used in practice. In order to compare test results, from this study firstly, one
dimensional settlement for the soil only was calculated using the following formula:
𝑆0 = 𝑚𝑣 ∆𝜎𝑧 𝐻 ‎4.16
Where mv is the coefficient of volume compressibility of clay; Δσz is the vertical stress at
depth z due to the applied pressure at the surface and equals to (Ig P0);( Ig is the Influence
facto for the increase in vertical stress, P0 is the imposed load) and H is the thickness of the
soil layer.
Secondly, the settlement improvement factor was determined using the area replacement and
angle of shearing resistance of column material relationship proposed in Priebe’s method. As
illustrated in chapter two (section 2.6.2.3). The internal angle of shearing resistance ranged
between 35o and 45o, thus a value of 45o was used to predict the improvement factor.
For example, based on the test conducted in this study, settlement estimation was based on
Equation 4.16, where Ig = 0.64644, and P0 = 100 kPa, leading to Δσz = 64.64 kPa, also using
mv as 9.803* 10-4 m2/kN (Section (3.3.6)). Settlement was estimated to be about 18 mm for
the clay only; and then the settlement of the reinforced soil can be estimated by dividing the
estimated settlement of the unreinforced soil by the improvement factor (n0) of 2.8 (obtained
from Priebe’s charts, Figure 2.9) resulting in a value of 6.4 mm, which was about six times
higher than the measured one. Although Priebe’s method demonstrated an overestimation to
the settlement of the reinforced soil, which confirmed the findings of Bouassida et al. (2003)
and Ellouze et al. (2010), the measured improvement factor of 2.1 showed reasonable
agreement with similar studies (Ambily and Gandhi, 2007; Black et al., 2011).
124
4.3.2 Change in Pore Water Pressure
Pore water pressure transducers PPT1,PPT2, and PPT3 were fitted at the centre of base, 50
mm from the centre and 100 mm from the centre respectively. Figures 4.18 and 4.19 show
results during the foundation loading on the soil (only) and soil/ stone column composite
specimens. When soils (no column) were tested, pore water pressure was seen to increase
only at PPT1 (the midpoint) when loading the foundation to about 3 kPa, while in the other
two (PPT2 and PPT3) transducers there were no observed change during the test.
On the other hand, when the column was installed, the PPT3 (close to the tank wall) did not
show any change in pore water pressure measurements during the loading. There was gradual
increase by about 5 kPa in pore water pressure at PPT2 (50 mm from the column); and
beneath the column at the centre of the specimen (PPT1), the pore water pressure build-up
rapidly after 20 mm (7 %) of footing settlement to approach approximately 20 kPa, then it
started dissipating. This confirms that the column carried larger part of the load compared to
the surrounding soil. Also when the pore water pressure stopped building up under the
column and still increasing beneath the soil could indicate that the failure of the column had
occurred.
125
4
PPT1
Pore water pressure, (kPa) PPT2
0
0 5 10 15
Axial strain, (%)
Figure ‎4.18 Typical results for changes in pore water pressure for soil (no column) specimens
30
PPT1
PPT2
Pore water pressure, (kPa)
20
10
0
0 5 10 15
Axial strain, (%)
Figure ‎4.19 Typical results for changes in pore water pressure for soil/stone column specimens
126
4.3.3 Post Testing Investigation
In order to investigate the quality of the prepared clay bed, a series of shear vane tests were
carried out at different positions to determine the undrained shear strength of the clay
surrounding the column. Strength measurements are shown in Figure 4.20. The undrained
shear strength was determined at (50, 100, 150, and 200 mm) depth and at four horizontal
positions (around the midway between the column and the tank wall). In addition to this, water
content measurements were made at the same levels as shown in Figure 4.21. Comparing both
before and after testing for the shear vane results and water content, it can be observed that at
the first 50 mm thickness of the soil there was an increase in the soil strength of about 12%
this increase was accompanied by a reduction in water content of almost 1%.
Vane shear strength (kPa)

10.5 11 11.5 12 12.5 13 13.5
0
Before testing
After testing
50
Depth of clay bed (kPa)
100
150
200
Figure ‎4.20 Hand vane shear test results before and after testing
127
Water content (%)

42 42.5 43 43.5
0
Before testing
Depth of clay bed (mm)
50 After testing
100
150
200
Figure ‎4.21 Water content variation before and after testing
After each test, the specimens were removed from the cell and split vertically along their
diameter to examine the mode of failure. A typical example is shown in Figure 4.22. In most
cases column bulge, took place at approximately at 120 mm below the surface, which is about
1.7 times the column diameter. This was very close to the expected value of 2 to 3 times the
column diameter suggested by Hughes and Withers (1974).
128
300
250
200
150
100
50
0
0 50 100 150 200 250 300
Figure ‎4.22 Deformation pattern

A series of undrained triaxial and reduced scale foundation model tests were carried out in
order to investigate the influence of strain rate and column material density on the behaviour
and performance of soft clay deposits reinforced with a stone column.
Results showed that there was a considerable improvement in load carrying capacity and
stiffness of soft clay when stone column was installed. For example, replacing a small part of
soft soil (7 %) with granular materials that have a comparably a high angle of shear resistance
(48o) leads to an increase in bearing capacity and soil stiffness of approximately 30 % and
reduces the settlement by about 80 % compared with the settlement of unreinforced soils.
Although results from this study showed reasonable agreement with empirical estimations and
other published works in terms of load bearing capacity, observed settlement behaviour
129
showed a conservative figure not least when compared with the settlement estimated using
Priebe’s method.
Strain rate did not appear to have a significant influence on the level of improvements, where
results show that there is a decrease of less than 10 % in deviator stress for second order of
magnitude in change in rate of strain.
There was a strong indication suggests that the relative density of the column material may
have a significant influence on column bearing capacity. For example, increasing the column
density by about 25 % leads to an increase in soil/ stone column composite capacity of about
17 %.
130
Samir Ashour Chapter 5: Cyclic Loading Results and Discussion
5. CHAPTER 5
5. CYCLIC LOADING RESULTS AND DISCUSSION
5.1 Introduction
This chapter presents results of both the undrained cyclic triaxial tests (Model I) and cyclic
foundation tests on large scale model (Model II) carried out on both soft clay and soft clay
/stone column composite. Dynamic stress levels ranging between 50 to 70 kPa (equivalent to
cyclic stress ratio (CSR) from 0.6 to 0.8) and frequencies ranging between 0.5 to 3Hz were
used, as described in Sections 3.7.1 and 3.7.2. All tests were conducted at an effective
confining pressure of 100 kPa. Comparison is made between the behaviour of the soft clay
soils with and without stone column installation.
Change in pore water pressure and deformation with number of cycles were examined. The
latter was used to identify threshold value of stress ratio. The variables examined during this
stage of the study are shown in Table 5.1
131
Table ‎5.1 Summary varibles investigated using cyclic triaxial test
Pmax Pmin σ'cyclic Loading frequency

Soil type Test CSR
(kN) (kN) ( kPa) (Hz)
TRI-C.C-12 0.8 0.2 70 0.8
TRI-C.C-13 0.7 0.2 60 0.7 0.5
TRI-C.C-14 0.6 0.2 50 0.6

Stone column / soil system
TRI-C.C-15 0.8 0.2 70 0.8
TRI-C.C-16 0.7 0.2 60 0.7 1.0
TRI-C.C-17 0.6 0.2 50 0.6
TRI-C.C-18 0.8 0.2 70 0.8
TRI-C.C-19 0.7 0.2 60 0.7 3.0
TRI-C.C-20 0.6 0.2 50 0.6
TRI-C-21 0.7 0.2 60 0.7

0.5
TRI-C-22 0.6 0.2 50 0.6
TRI-C-23 0.7 0.2 60 0.7

Soil only
1
TRI-C-24 0.6 0.2 50 0.6
TRI-C-25 0.7 0.2 60 0.7

3
TRI-C-26 0.6 0.2 50 0.6
5.2 Model I (Triaxial Tests)
5.2.1 Permanent Strain during Cyclic Loading
Typical results of cyclic undrained triaxial tests are shown in Figures 5.1-5.3, which shows
permanent axial strain against number of cycles. Results show the influence of cyclic stress
levels (50, 60 and 70 kPa) at different loading frequency (0.5, 1 and 3 Hz) on the deformation
pattern for soft soil specimens (no column) and soft soil/ stone column specimens. All
specimens were subject to a number of cycles of 10000 cycles, apart from those whom failed
before completing the test.
132
16 CSR=0.6_Soil
qcyclic=50kPa_Soilonly
only
14 CSR=0.7_soil only
qcyclic=60 kPa_Soil only
Permanent axial strain, (%) 12

CSR=0.6_Soil/column
qcyclic=50 kPa_Soil/column
CSR=0.7_Soil/column
10 qcyclic=70 kPa_Soil/column
CSR=0.8_Soil/column
0
10 100 1000 10000
Number of cycles
Figure ‎5.1 Pemanent axial strain and number of cycles relationship: at 0.5Hz loading frequency
16 qcyclic=50 kPa_Soil
qcyclic=50kPa_Soil only only
14 qcyclic=60 kPa_soil
qcyclic=60 kPa_Soil onlyonly
Permanent axial strain, (%)
12
10 qcyclic=70 kPa_Soil/column
0
10 100 1000 10000
Number of cycles
133
16
CSR=0.6_Soil only
qcyclic=50kPa_Soil only
14
qcyclic=60 kPa_Soil
CSR=0.7_soil onlyonly
Permanent axial strain, (%) 12 qcyclic=50 kPa_Soil/column
CSR=0.6_Soil/column
CSR=0.7_Soil/column
10
CSR=0.8_Soil/column
8
0
10 100 1000 10000
Number of cycles
It can be noted that, for both soil (no column) and soil/ stone column composite, the progress
of cumulative strain was highly influenced by increases in cyclic deviator stress applied hence
CSR. However, in the range of the loading frequencies examined, soils (no column)
specimens were more sensitive to the change of frequency than the soil/ stone column
composite.
Results in Figures 5.1 to 5.3 showed that the permanent axial strains for soils specimens (no
column) were larger at lower frequency (0.5 Hz) and decreases with the increase in loading
frequency, for instance, when specimens were subjected to a cyclic stress of 50 kPa, the
permanent axial strains after 10000 cycle were 6.3, 4.4 and 6.3 % for frequencies 0.5, 1 and 3
Hz respectively. However, in soil/stone column system specimens the effect of frequency on
permanent axial strain was very small (at 50 kPa cyclic stress they were 1.9, 1.8 and 1.6 % for
frequencies 0.5, 1 and 3 Hz respectively).
134
In addition, the effect of the cyclic stress on the development of permanent strain was
significant for both types of specimens. For example, when tests were performed at loading
frequency of 1 Hz, results (Figure 5.2) showed that soil specimens (no column) at an applied
cyclic stress of 50 kPa stabilised after 900 cycles at a strain level of 4.4 %, when the cyclic
stress was increased to 60 kPa a shear failure occurred at less than 100 cycles of load
applications. On the other hand, when the stone column was installed and specimens tested
under a 50 kPa cyclic stress application, the progress of the permanent strain stabilised after
1000 cycle at a value of 1.8 %. This value increases to reach 5.6 % by increasing the cyclic
stress to 60 kPa, and when the test was performed at a cyclic stress of 70 kPa the specimen
was failed in a shear failure after 30 cycles. Similar behaviour was shown when specimens
tested under the condition of other loading frequencies (0.5 and 3 Hz).
These results indicate that there can be a threshold stresses of 50 kPa and 60 kPa for soil (no
column) specimens and soil/ stone column composite specimens respectively above which the
total permanent strain increased rapidly with repeat load application; and failure can occur
during the first few cycles. Normalising these dynamic stresses in terms of static vertical
stress at failure for both types of specimens (with and without column) leads to have a CSR of
about 0.75. From literature this value varies from 0.4 to 0.8 and often is assumed to be 0.5.
(Seed et al., 1955; Heath et al., 1972; Brown et al., 1975; Brown, 1996; Andersen, 2004; Jiang
et al., 2010; Cai et al., 2013) This variation in the value of the threshold stress could be due
to the difference in the soil types and their physical properties.
5.2.1.1 Effect of Cyclic Deviator Stress on Permanent Strain
Figure 5.4 shows the influence of cyclic deviator stress on the permanent strain of both soil
only and soil/column composite specimens at load applications of 100 and 1000 cycle and
loading frequency of 1 Hz. Results showed that below the threshold stress level in soil (no
135
column) specimens were more sensitive to number of loading application than the reinforced
samples. For example, at 50 kPa cyclic stress, the permanent strain increased from 1.7 to 4.4
% by increasing number of cycles from 100 to 1000 cycle. On the other hand, the influence of
number of cycles were much smaller in case of soil/column specimens, where difference in
permanent strain between 100 and 1000 cycles at similar stress level was about 0.9 %. This is
probably due to the fact that the column materials were compacted due to load application,
increasing the density of the column leading to greater resistance with lower settlement.
Cyclic deviator stress, (kPa)

40 50 60 70 80 90
0
Soil/column_100 cycle
5 Soil only_100 cycle
Soil/column_1000 cycle
Permanent strain, (%)
Soil only _1000 cycle
10
15 Failure
20
25
Figure ‎5.4 Effect of loading application and cyclic stress leve at frequency of 1 Hz
5.2.1.2 Effect of Loading Frequency on Permanent Strain
Frequency of cyclic load is believed to have an important influence on the dynamic properties
of the saturated soft clay (Andersen et al., 1980), however, its effect is still attracting debate.
Jian-hua Wang et al. (1998), Zhou and Gong (2001), and Jiang et al. (2010) indicated that
136
both accumulative permanent strain and pore water pressure induced by cyclic loading will
increase with reduction in loading frequency. Whereas other published work indicated a
different observation, where they reported that the change in frequency has little or no
influence on the strength and deformation properties of soil (Ansal and Erken, 1989; Hyde et
al., 1993).
In this current study, an increase in permanent deformation for the soil only specimens was
observed with reduction in frequency of loading, as shown in Figure 5.5, for both soil
specimens (with and without column) at load applications of 100, 1000 and 10000 cycles. All
the test were undertaken at the same cyclic stress level (50 kPa) which was less than the
threshold stress. For stone column reinforced specimens, the influence of loading frequency
on the permanent stain was less pronounced compared with that for unreinforced specimens.
Also the effect of number of cycles was negligible after 1000 cycle.
Frequency, (Hz)
0 0.5 1 1.5 2 2.5 3 3.5
0
2
6 Soil/column_100 cycle Soil/column_1000 cycle Soil/column_10000 cycle
7 Soil only_100cycle Soil only_1000 cycle Soil only _10000cycle
Figure ‎5.5 Effect of loading frequency and number of cycles under cyclic stress of 50 kPa.
137
5.2.1.3 Prediction of Permanent Strain
There are several models to predict permanent strain in soils subjected to cyclic loading
conditions. The most commonly used is the power model which takes into account a number
of factors such as number of loading cycles and both static and dynamic stresses (Li and
Selig, 1996). This model is shown as follow:
𝜀𝑝 = 𝐴𝑁 𝑏 5.1
Where 𝜀𝑝 is the permanent strain (cumulative strain) in %; N is the number of cycles; and A
and b are material parameters (dependent in the soil properties and stress level). Li and Selig
(1996), quantified the coefficient A by relating it to the physical state of the soil and the
applied deviator stress, so it could be written as:
𝜎
𝐴 = 𝑎( 𝑑⁄𝜎𝑠 )𝑚 5.2
Where a and m are the correlation parameters; 𝜎𝑑 𝑎𝑛𝑑 𝜎𝑠 are the deviator stress and static
strength respectively. Therefore, parameter (A) is highly dependent on the applied deviator
stress and the maximum deviator stress that the soil can take at static test condition.
Parameters a, m, and b can be determined from a cyclic triaxial tests.
Li and Selig (1996), recommend values for the above parameters for a number of soil types
based upon back calculation from a number of different studies (Table 5.2). Values for the
soft soils (no column) and the soil/ stone column composite parameters calculated in this
study through regression analysis of cyclic triaxial results are included (see Appendix F for
more information on regression analysis).
138
Table ‎5.2 Material parameters (after Li and Selig (1996))
Soil classification
Kaolin/stone
Model parameters Kaolin clay
ML MH CL CH column
(test condition)
(test condition)
Average 0.64 0.84 1.1 1.2 4.346 23.392

a
Range - - 0.3-3.5 0.82-1.5 3.960-4.733 22.519-24.264
Average 0.1 0.13 0.16 0.18 0.096 0.019

b
Range 0.06-0.17 0.08-0.19 0.08-0.34 0.12-0.27 0.085-0.107 0.015-0.024
Average 1.7 2.0 2.0 2.4 2.469 5.081

m
Range 1.4-2.0 1.3-4.2 1.0-2.6 1.3-3.9 2.136-2.803 5.004-5.157
As explained by Li and Selig (1996), that exponent b reflects the accumulative strain rate
under cyclic loading (i.e. that the higher the value of b the more the soil affected by number of
cycles). Whereas exponents a and m are related to the deformation occurred at the first cycle
and its relation to the cyclic stress ratio applied and also to the degree of softening of the
deviator on the soil.
As can be seen in Table 5.2 the parameters value of the used soil (no column) is comparable
with the CL values, the reason for this is that the kaolin clay considered as a clay soil with
high plasticity. However, for soil/ stone column composite the exponent b showed a smaller
value than other type of soils; this might indicate that the stone column reinforced soils are
likely to be less sensitive to number of loading application. Whereas a value showed a higher
level (23.39) than the other soils, which indicates that, the soil/ column system could have a
higher degree of softening.
139
5.2.2 Cyclic Deformation
Figures 5.6 to 5.8 show the cyclic deformation behaviour for soil (no column) and soil/ stone
column composite. The influence of loading frequency; dynamic stress level and number of
loading application were included in these Figures.
0.9
qCSR=0.6_Soil
cyclic=50kPa_Soilonly
only
0.8 qcyclic=60 kPa_Soil only
Series4
Cyclic deformation, (mm)
0.7 qcyclic=50 kPa_Soil/column

CSR=0.6_Soil/column
CSR=0.7_Soil/column
0.5 Series7
0.4
0.3
0.2
0.1
10 100 1000 10000
Number of cycles
Figure ‎5.6 Cyclic deformation during cyclic triaxial test: at 0.5 Hz loading frequency
0.7 CSR=0.6_Soil onlyonly

qcyclic=50kPa_Soil
0.6 CSR=0.7_soil onlyonly

qcyclic=60 kPa_Soil
Cyclic deformation, (mm)
CSR=0.6_Soil/column
CSR=0.7_Soil/column
q =70 kPa_Soil/column
cyclic
0.4 CSR=0.8_Soil/column
0.3
0.2
0.1
10 100 1000 10000
Number of cycles
Figure ‎5.7 Cyclic deformation during cyclic triaxial test: at 1 Hz loading frequency
140
0.7
CSR=0.6_Soil only
0.6
Elastic deformation, (mm)
CSR=0.7_soil onlyonly
qcyclic=60 kPa_Soil
CSR=0.6_Soil/column
0.5
CSR=0.7_Soil/column
0.4 CSR=0.8_Soil/column
0.3
0.2
0.1
10 100 1000 10000
Number of cycles
Figure ‎5.8 Cyclic deformation during cyclic triaxial test: at 3 Hz loading frequency
In all specimens tested, the cyclic deformation decreases as the number of cycles increases.
This was due to hardening of soil specimens. This phenomenon occurs only at stresses below
the threshold level (i.e. 50 kPa for soil (no column) and 60 kPa for soil/ stone column
composite) above these stresses the cyclic strain sharply increases due to the shear failure
occurred.
Figures 5.6 to 5.8 have also indicated that the cyclic strain in both soils (with and without
stone column) are slightly influenced by the change in loading frequency as there was a small
decrease in the cyclic deformation with the increase in the frequency.
For instance, when the specimens subjected to 50 kPa cyclic stress, the values of cyclic
deformation for soil (no column) specimens after 10000 cycle were 0.24, 0.22 and 0.18 mm
for frequencies 0.5, 1 and 3 Hz respectively. Whereas in the soil/ stone column specimens the
cyclic deformations were 0.23, 0.22, and 0.17 mm for frequencies 0.5, 1 and 3 Hz
respectively. This would indicate that there is a threshold elastic strain of 0.2 % for soil (no
column) and about 0.1 % for soil/ stone column composite.
141
5.2.3 Resilient Modulus
Soil resilient modulus (Mr) can be defined as the ratio of the dynamic deviator stress (𝑞𝑑 ) to
the elastic (recoverable) strain (𝜀𝑟 ) (Brown et al., 1975; Selig and Waters, 1994; Brown,
1996).
𝑞𝑑
𝑀𝑟 = 5.3
𝜀𝑟
This parameter is considered to be one of the fundamental factors of understanding the elastic
behaviour of soil foundations subjected to dynamic stresses. For instance, in railway track
analysis, resilient modulus is important for determining stresses and settlement in the track
foundation system (Loh and Nikraz, 2012).
Typical results of resilient modulus determination of both specimens type (soil (no column)
and soil/ stone column composite) under dynamic stresses of 50 and 60 kPa at different
loading frequencies (i.e. 0.5, 1 and 3.0 Hz) are shown in Figure 5.9 to 5.11 respectively.
45
40
Resilient modulus, (MPa)
35
30
CSR=0.6_Soil onlyonly
qcyclic=50kPa_Soil
25 CSR=0.6_Soil/column
CSR=0.7_Soil/column
20
10 100 1000 10000
Number of cycles
Figure ‎5.9 Variation of resilient modulus with number of cycles at loading frequncy of 0.5 Hz
142
50
45
40
35
30
CSR=0.6_Soil only
CSR=0.7_Soil/column
20
10 100 1000 10000
Number of cycles
60
55
50
45
40
35
CSR=0.6_Soil only
30
CSR=0.6_Soil/column
20
10 100 1000 10000
Number of cycles
Results indicated that the resilient modulus increased with frequency in both specimen types
(20 % for soil/ stone column specimens and about 12 % for soil (no column) specimens). For
example, during the application of 50 kPa dynamic stress, the increase was from 40.6 MPa to
143
49.4 MPa in case of soil/ stone column composite specimens, while it was from 33.5 MPa to
38 MPa in the soil specimens (no column).
In addition to this, Figure 5.12 shows that any increase in the dynamic deviator stress has a
negative impact on resilient modulus, which decreases as dynamic stress increases. For soil/
stone column specimens, when the dynamic deviator stress was increased from 50 to 60 kPa,
there was a general reduction in the amount of resilient modulus this reduction influenced by
the change of frequency as well (i.e. the higher the frequency the smaller the reduction). For
example, there was a reduction percentage in resilient modulus of 6.5, 7.6 and 13.3% when
the tests were conducted at frequencies of 0.5, 1 and 3 Hz respectively.
Comparing the resilient modulus results with the secant modulus obtained from the static
triaxial test in the previous chapter (Section 4.2.4.3), which was about 9.5 and 8 MPa for
soil/column specimens and soil only specimens respectively; it was found that the resilient
modulus could reach up to four times the secant modulus value when the cyclic deviator stress
was 50 kPa.
55 Soil/column_3 Hz
Soil/column_1 Hz
50
Soil/column_0.5 Hz
45 Soil only_3Hz
Soil only_1 Hz
40 Soil only_0.5 Hz
35
30
25
20
15
40 50 60 70 80
Dynamic deviator stress, (kPa)
Figure ‎5.12 Variation of resilient modulus with respect to dynamic stress and loading frequency
144
5.2.4 Pore Water Pressure Response
Under continued application of cyclic loading, excess pore water pressure gradually built up
over each cycle. Figures 5.13 to 5.15 show changes of pore water pressure corresponding to
the cyclic deviator stress at the three loading frequencies for both soil (no column) and soil/
stone column composite specimens.
90 CSR=0.6_Soil onlyonly
qcyclic=50kPa_Soil
CSR=0.7_Soil onlyonly
qcyclic=60 kPa_Soil
80
Increase in pore water pressure, kPa
CSR=0.6_Soil/column
70
CSR=0.7_Soil/column
50
40
30
20
10
10 100 1000 10000
Number of cycles
Figure ‎5.13 Accumulation of excess pore water pressure at loading frequency of 0.5 Hz
qcyclic=50kPa_Soil
qcyclic=60 kPa_Soil
CSR=0.6_Soil/column
70 cyclic
CSR=0.7_Soil/column
cyclic
cyclic
50
40
30
20
10
10 100 1000 10000
Number of cycles
Figure ‎5.14 Accumulation of excess pore water pressure at loading frequency of 1Hz
145
100 CSR=0.6_Soil
qcyclic=50kPa_Soilonly
only

90 CSR=0.7_Soil only
qcyclic=60 kPa_Soil only
CSR=0.7_Soil/column
70
CSR=0.8_Soil/column
60
50
40
30
20
10
10 100 1000 10000
Number of cycles
Figure ‎5.15 Accumulation of excess pore water pressure at loading frequency of 3Hz
As expected, the excess of pore water pressure increased with cyclic deviator stress at all
loading frequencies examined. Results show that in the case of most specimens, the pore
water pressure increased significantly at the initial stage of loading application (i.e. number of
cycle less than 150) then there was a gradual increase up to 1000 cycle before it started to
stabilise.
Figures 5.13 to 5.15 Also indicated that, the frequencies of loading did not show a noticeable
influence on the excess of pore water pressure in the soil /column specimens. Tests carried out
at 50 kPa cyclic stress the pore water pressures after 10000 cycles were 38, 37, and37 kPa for
frequencies 0.5, 1 and 3 Hz. At higher cyclic stress level (70 kPa) the specimen failed before
the pore water pressure could stabilise. Before failure it had increased rapidly to over than 75
kPa.
In contrast, for soil only samples results (Figures 5.13-5.15) showed that at a cyclic stress of
60 kPa, the excess pore water pressure increases rapidly to reach 70 kPa in few cycles leading
to a rapid failure in the sample, however, there was a small effect on the soil only specimens,
146
where the build-up of pore water pressure decreased with increase in the loading frequency
especially at the early stage of loading application (in the application of 50 kPa cyclic stress
and after 10000 loading cycle, pore water pressures were 51, 45.4 and 42.5 kPa for
frequencies 0.5, 1 and 3 Hz).
Generally, in most tests conducted below the threshold stress levels pore water pressure
stabilises after 2000 cycles of loading applications. This behaviour is comparable with the
general behaviour of soft clay under cyclic loading condition (Miller et al., 2000; Indraratna et
al., 2009).
Figure 5.16, indicates that there was a critical excess of pore pressure and found at around 60
kPa for the soil specimens (no column) and about 70 kPa for the soil/column specimens ( pore
water pressure measurements at failure). These pore water pressure measures can be
normalised in terms of effective confining pressure (𝜎3′ = 100 𝑘𝑃𝑎) leading to ratios of 0.6
for soil (no column) and 0.7 for soil/ stone column composite. This also indicates that the
failure occurred when the pore water pressure approached the level of the dynamic stress.
80
70
Pore water pressure, (kPa)
60
50
Soil/column_0.5 Hz
Soil/column_1 Hz
40 Soil/column_3 Hz
Soil only_0.5 Hz
Soil only_1Hz
Soil only_3 Hz
30
40 50 60 70 80
Cyclic deviator stress, (kPa)
Figure ‎5.16 Effect of cyclic stress on pore water pressure
147
5.2.5 Deformation Pattern
Observation indicates that both soil specimens (no column) and soil/ stone column composite
specimens fail in shear failure mode when subjected to cyclic stress above the threshold cyclic
stress 50 kPa for soil and 60 kPa for soil/column specimens; this was irrespective of
frequency of loading. Also below these stresses both specimens did not fail at the end of the
test and behaved in a stable manner.
Figures 5.17 and 5.18 show the mode of failure for both reinforced and unreinforced
specimens. Soil specimens (no column) specimens they failed within a range from 60 to 100
cycles of loading application (Figure 5.17) a slip surface was clear along weakest plan.
A combined mode of failure was observed in case of reinforced specimens (bulging and shear
failure (Figure 5.18)). Due to the large strain happened bulging failure took place first and
then by the increase in loading cycle shear failure accrued. The shear surface for the column
was at approximately the bottom 50 mm resulting in L/D ratio of 5.4 which is close to the
identified critical column length (L/D = 6).
148
12
10
Sample height, (cm)
0
0 10 20
Horizontal measurment, (cm)
Figure ‎5.17 Deformed shape for soil only specimes
Horizontal measurment, (cm)
-5 5 15 25 35
20
15
Sample height, (cm)
10
Figure ‎5.18 Deformed shape for the soil/column specimen
149
5.3 Model II Test Results (Large Scale Test: Foundation)
The procedure for undertaking these tests is described in Sections 3.6.7 and 4.3. As
summarized in Table 5.3, the testing program included a series of cyclic zone load tests (Load
applied to column and portion of the surrounding soil in case of soil/column test). These tests
were conducted for both treated and untreated soils under two loading frequencies (1 and 3
Hz). Loading was applied in a sinusoidal wave form using two dynamic stress levels of 165
and 235 kPa, which were equivalent to cyclic stress ratios of 0.5 and 0.7 respectively.
The range of stress applied was based on the results of triaxial tests; particularly findings
relating to the threshold value, which was 70 % of the reinforced soil strength at failure. Thus
cyclic stress ratios of 0.5 and 0.7 were chosen for large scale test to evaluate this observation.
Table ‎5.3 Model II test program
Stress at Loading
Pmax Pmin σmax σmin
Soil type Test failure CSR frequency
(kN) (kN) ( kPa) ( kPa)
(kPa) (f)
LM-C-05 0.8 0.2 210 45 - 1

200
only
Soil
LM-C-06 0.8 0.2 210 45 - 3
LM-C.C-07 0.8 0.2 210 45 0.5 1

Soil/column
LM-C.C-08 0.8 0.2 210 45 0.5 3

system
410
LM-C.C-09 1.05 0.2 280 45 0.7 1
LM-C.C-10 1.05 0.2 280 45 0.7 3
5.3.1 Permanent Strain during Cyclic Loading
Investigating the behaviour of permanent deformation of the reinforced soil on large scale
specimens indicated that the permanent stain was independent from the change in frequency
of loading; also the permanent deformation is highly dependent on the amount of the dynamic
150
stress applied and its relation to monotonic stress at failure. This confirms to the results
obtained from triaxial test described earlier in Section 5.2.
Figure 5.19 shows the effect of number of cycles; loading frequency, and cyclic stress ratio on
the measured permanent strain for both soil (no column) and soil/ column composite
specimens. It was observed when testing soil only specimens (no column) at a dynamic
loading stress of 165 kPa and loading frequency of 1 and 3 Hz, that the soil failed after 20
cycles of loading application; this was the cases at both loading frequencies. This was
expected as the loading stress was larger than the threshold ratio of the soil obtained in
Section 5.2.2. (As discussed in the methodology chapter this load was chosen due to the
limitation of the apparatus used in this study). On the other hand, soils/ stone columns
composite showed increased resistance to load, and the permanent deformation did not go
beyond 2 % at both loading frequencies. This increased to about 6 % when the dynamic stress
load was increased from 165 to 235 kPa.
12
A Soil/column_CSR = 0.5; f=1 Hz
Soil/column _CSR = 0.5; f = 3 Hz
10 Soil/column_ CSR = 0.7; f = 1 Hz
Soil only_CSR = 0.5; f = 1 Hz
B Predicted permanent strain

8
Predicted Permanent Strain.
0
1 2001 4001 6001 8001
Number of cycles
Figure ‎5.19 Permanent strain results/ number of cycles
151
Predicted values of permanent strain obtained from Equation 5.2 are also shown in Figure
5.19. Material parameters a, b, and m, determined from the trixial test and explained in
Section 5.2.1, are used.
During the application of dynamic stress at 165 kPa, permanent strain was predicted to be
increasing with the increase of number of applications reaching approximetly 1.6 % at 1500
cycles and then maintaining this level. Comparing this with the laboratory results, at the same
stress level the permanent strain increased to 1.8% at 1500 cycles and then continued at that
level to the end of the test (10000 cycles). A similar trend was observed when the stress level
was increased to 235 kPa. The estimated permanent strain was supposed to reach a maximum
of 4.6 % and stabilize after 1500 cycles, however, test results showed that the development of
permanent strain reached a maximum of 5.3 % and stabilised at this level after 4000 cycles.
Although the predicted permanent strain was underestimated compared with the measured
values, the difference seems to be considerable where it was within a range of 15 %. This
variation could be due to the difference in test conditions, as the parameters used in the
prediction were calculated from the triaxial test and for a limited CSR.
Figure 5.20 shows the deformation pattern after the cyclic loading application of 10000 cycles
at a cyclic stress of 165 kPa and loading frequency of 3 Hz. Although it was expected that the
columns will bulge at a depth of 2 to 3 times the column diameter as suggested by Hughes
and Withers, (1974), it appears that the column bulging under dynamic loading was different
from that of monotonic loading. In other words it was observed that the column bulge in two
sections (double bulging). These bulges were at 20 and 60 mm depth from the top, this can
expressed as 0.7 and 2 times the column diameter.
152
Double Bulging
300
280
260
240
Sample depth, (mm)
220
200
180
160
140
120
100
80
60
40
20
0
0 50 100 150 200 250 300
Sample diameter, (mm)
Figure ‎5.20 Deformation pattern after cyclic loading
5.3.2 Pore Water Pressure Changes
During cyclic loading pore water pressures were measured in two positions, at the centre
(PPT1) of the specimen base and at a distance of 50 mm from the centre (PPT2). The pore
water pressure measurements are illustrated in Figures 5.21 and 5.22; both show the pore
pressure changes that occurred under peak cyclic loading on the foundation plate.
12
Soil/column_CSR = 0.5; f=1 Hz
Soil/column _CSR = 0.5; f = 3 Hz C PPT1
10 Soil/column_ CSR = 0.7; f = 1 Hz

8
D
6
0
10 100 1000 10000
Number of cycles
Figure ‎5.21 Pore water pressure measurment at the centre of the specime
153
4
Soil/column _CSR = 0.5; f = 3 Hz PPT2
Soil/column_ CSR = 0.7; f = 1 Hz
3
0
10 100 1000 10000
Number of cycles
Figure ‎5.22 Pore water pressure measurment at 50 mm from the centre of the specime
As was the case for triaxial tests, under continues application of cyclic loading, excess pore
water pressure was initially built up rapidly before it start to decrease (Figures 5.13 to 5.15).
Figure 5.21 shows the change in pore water pressure under the stone column at CSR of 0.5
was increased from about 1.5 kPa at the initial stage to about 8.7 kPa during the loading
application of 1000 cycle, after this the pore water pressure start to decrease gradually.
Similar behaviour was observed using CSR of 0.7 with larger magnitude of pore water
pressure (about 10 kPa). However, this behaviour was different at the other position (PPT2);
the change in pore water pressure was almost steady during the first 2000 cycle then started to
rapidly increase.
5.3.3 Soil stiffness
Soil stiffness (K) is defined as the ratio of the dynamic applied load (𝑃𝑑 ) on the foundation
area to the recoverable displacement experienced.
154
Stiffness of the treated and untreated soil tested at CSR of 0.5 and 0.7 at loading frequencies
of 1 and 3.0 Hz are shown in Figure 5.23 . Results show that after applying 10000 cycles the
stiffness for soil/ stone column composite increased from 1770 to 2050 kN/m as frequency
increased in from 1 Hz to 3 Hz (i.e. about 15 %).
Soil specimens (no column) at both frequencies 1 and 3 Hz showed low stiffness (745 and
960 kN/m respectively) before failure occurred after 100 and 330 cycles respectively. This
was as a result of the high deformation occurred as shown before in Figure 5.19 (Points A and
B).
In addition, it was noticeable that the stiffness at soil/ stone column composite, at CSR of 0.5,
was decreased gradually (from1700 to 1630 kN/m at frequency of 3 Hz and from 1540 to
1450 kN/m at frequency of 1 Hz) during the first 100 cycles then increased by 20 % before it
stabilised after 1500 cycles. This is likely due to the stiffening effect of stone column and its
provision as a drainage path leading to reduction in pore water pressure (Figure 5.21 points C
and D). As pore water pressures start decreasing from 8.7 to around 6 kPa after 1000 cycles
for f = 1 Hz after 3000 cycle and at f = 3 Hz.
This behaviour was different when the dynamic stresses was increased (CSR = 0.7), where
the stiffness of the reinforced soil was almost constant during the first 100 cycles at 1320
kN/m then it started to decrease slowly to reach 1100 kN/m at the end of the test (10000
cycles). This could be due to softening of the composite system.
155
2500
Stiffness, (kN/m) 2000
1500
1000
Soil/column _CSR = 0.5; f = 3 Hz
Soil/column_ CSR = 0.7; f = 1 Hz
500
Failure
0
10 100 1000 10000
Number of cycles
Figure ‎5.23 Stifness results
5.4 Implication for Stone Column Performance
The applied dynamic stress through the rail is transmitted through the ballast layer to the
subgrade level, played the main role in controlling the amount of permanent deformation of
the subgrade. Therefore, as suggested by Heath et al. (1972); Brown (1996) and Frost et al.
(2004) identifying the threshold stress would help for determining the thickness of unbound
layer.
From looking at literature and monotonic triaxal test results presented here it can be seen that
the bearing capacity of soft clay subgrade soils can be improved by (15 to 40 %) if the soil
was replaced by granular material having an angle of internal friction between 35 and 48 o
formed in columns with an area replacement ratio between 5 and 15 %. However, in this
current study using an angular crushed aggregate with an internal angle of friction and area
replacement ratio of 7 % have led to improved the bearing capacity of the soil by 30 % under
the monotonic loading (Figure 4.3).
156
This improvement in the bearing capacity under the monotonic loading showed an increase in
the threshold cyclic stress ( under the cyclic loading condition) of about 15 % and 70 %
reduction in settlement; also showed an increase by 25 % in the resilient modulus of the soil.
The threshold relation between reinforced and reinforced soils can be related to limiting the
allowable stress applied on the soft subgrade via the relationship between the internal angle of
friction, area replacement ration and the undrained shear strength of the soil (Section 4.3.1).
This limit then can be applied in the design procedure of, for instance, railway tracks.
For example, this improvement in the soil properties could help to reduce the thickness of the
ballast layer. As one of the important performing functions of this layer is to reduce stresses
transferred through the sleepers from the train to the acceptable stress levels that the subgrade
soils can carry. Thus for a 100 kN axel wheel load using Talbot’s equation (Section 3.7.2), the
ballast depth (h) can be reduced by 15 % from 500 mm to 425mm. This can save about 225m3
of ballast per one kilometre of railway track which can cost around 2000 Pounds sterling per
one kilometre. This can be considered as a cost effective impact.
Another impact is that, by simulating train movement with loading frequencies of 0.5, 1.0 and
3.0 Hz representing train speed of 35, 70 and 225 km/hr, stone columns reinforced subgrades
showed stability and less pronounced influence toward both the change in loading frequency
and long term loading application. This can increase the period of maintenance and provide
the stability required to upgrade the track to higher speed level.
A comparison between the behaviour of both reinforced and unreinforced soils under the
application of cyclic loading was proposed in this chapter.
157
Results indicated that the presence of stone column enhanced the threshold cyclic stress of the
soft soil by about 15 % and reduced the amount of permanent strain by more than double
when testing the soils below their threshold cyclic stress. Stone columns also help reduce pore
water pressure by providing a drainage path. This might be one of the reasons that allow the
reinforced soil to tackle higher cyclic stress level.
Changes in frequencies had no significant influence on the total strain of reinforced soil, but
these changes do affect the stiffness. Stiffness of the soil with the stone column was about
25% higher at 3.0 Hz compared to that at 0.5 Hz. Resilient modulus of the reinforced soil
increased by 13 to 20 % depending on the frequency of loading where the resilient modulus
tend to increase by the increase of the frequency.
158
Samir Ashour Chapter 6: Conclusion
5. CHAPTER 6
6 CONCLUSION
6.1 General
The vibro stone column technique is commonly used to improve bearing capacity and reduce
settlement of soft soils. They have been used worldwide to provide economic, flexible
(applicable to various types of week soils) and sustainable solution. More recently, stone
columns have been used in the UK and Europe to improve soft subgrade soils subjected to
cyclic loading such as railway track.
Performance and failure mechanisms of both isolated and grouped stone columns have been
investigated under the application of monotonic loading but not under cyclic loading. This
study was aimed to provide improved understanding of this behaviour.
This research focused on the load/deformation behaviour of single stone column subjected to
monotonic and cyclic loading in both small scale models (triaxial specimen 100 mm diameter)
and large scale models (300 mm diameter * 300 mm height). In both cases tests were
performed on normally consolidated specimens of soft clay (undrained shear strength of ≈ 12
kPa) reinforced with 28 mm diameter stone columns. Effects of loading frequencies (0.5-3Hz)
at a range of cyclic stress ratios (0.5-0.8) were examined. In addition to the influence of rate
of strain on shear strength on both soils (no column) and soil/ stone column composite in the
monotonic loading conditions (section 6.2) and cyclic loading conditions (Section 6.3). The
following correlations were drawn from this study.
159
6.2 Monotonic Loading
6.2.1 General
Replacing 7 % of soft clay with granular materials, which had an internal of friction angle of
48o, led to increase the bearing capacity and stiffness of the soil by a factor of 1.3. Compared
to predicted values and previously reported results reasonable agreement was found in terms
of bearing capacity, However, measured value for settlement were about 83 % lower than
values predicted based on Priebe’s method.
6.2.2 Effect of Strain Rate
 Both soils (no column) and soil/ stone column composite specimens showed a small
influence from rate of strain; for soils (no column) there was an increase of about 10
% deviator stress associated with increasing the magnitude of rate of strain from 0.003
to 0.3 mm/minute. However, in the case of reinforced specimens the trend was
reversed order where the deviator stress showed an increase of 8 % with the reduction
of the strain rate (0.3 to 0.003).
 The presence of stone column helped reduce build-up of pore water pressure and there
was limited influence from rate of strain on pore pressure as compared to soils (no
column). The latter showed pore water pressure reduction with increase in strain rate.
 In case of soil/stone column composite, the secant modulus increased by about 30%
compared to soil (no column). This improvement was higher between 1 and 3 % strain
and for strains higher than about 4 % the secant modulus decreases and the degree of
improvement in modulus decreased by about 17%.
160
 The secant modulus of soil was seen to be independent of strain across the range
examined (0.3 to 0.003) for both soils (no column) and soil/ stone column composite
specimens.
 In model I (Triaxial tests), the column bulging was found to be relatively uniform with
the depth of the sample, this was probably due to the constant confining pressure
along the depth. However, in model II (Large scale tests), bulging took place at
approximately 120 mm below the surface, which was about 1.7 times the column
diameter. This was very close to the expected value of 2 to 3 times the column
diameter
6.2.3 Effect of Column Material Density
 A reduction in relative density of the column from 80 to 42 % decreases the bearing
capacity of the soil/ stone column composite by about 14 %, this could be due to a
reduction in the peak internal angle of friction of the column material and this might
be related to the increase in the dilation angle of the column particle.
 A change in aggregate density of the column has showed negligible effect on pore
pressure measurement.
6.3 Cyclic Loading
6.3.1 Effect of Stress Level
 Below the threshold stress level, reinforced soils were less sensitive to the number of
loading application than soils (no column). This is likely due to compaction of the
column materials that occurs during load application, leading to increase in density
and resulting in greater resistance to deformation hence lower settlement.
161
 Presence of a stone column enhanced the threshold cyclic stress of the soil by about 15
% and reduced the amount of permanent strain by half relative to soil without column.
 Stone columns helped reduce pore water pressure by providing a drainage path. This
may be one of the reasons that allows the reinforced soil to support higher cyclic stress
level (since pore water pressure would also be reduced).
 Both reinforced and unreinforced soils failed in shear mode when subjected cyclic
stress above the threshold cyclic stress. This failure mode was not affected by
frequency of loading and below threshold stress both specimens did not fail
throughout the load regime (10000 cycles).
 Large scale tests showed that the column bulging that forms under dynamic loading
was different from that of monotonic loading. It was observed that the column bulge
in two sections (double bulging) under cyclic loading. These bulges were at a depth of
0.7 and 2 times the column diameter below the surface of the samples.
6.3.2 Effect of Frequency
 In terms of permanent deformation, soil/ stone column composite showed little
influence by the change in frequency of loading. However, in the case of soil (no
column), larger permanent strains were measured at lower frequencies compared to
those at higher frequency, about 55 % increase when frequency was changed from 3 to
0.5 Hz.
 Below the threshold stress, the resilient modulus of both reinforced and unreinforced
soils increase with the increase of loading frequency. The value for soils/ stone
column composite soils was twice that of soils (no column).
162
6.4 Recommendations for Future Research Works

Whilst this study has highlighted important aspects in the behaviour of stone column
reinforced soil subjected to cyclic loading there are still important areas requiring further
investigation. These are suggested below:
 Future work also needed to be carried out to investigate the effectiveness of stone
columns under the condition of fluctuating ground water table associated with the
application cyclic loading.
 It is anticipated that fines from surrounding soil may ingress into the stone column.
Investigating the rate of movement of fine and their effect on durability of the stone
column.
 Effect of embankment height on the general behaviour of the soil/ stone column
composite soils.
 Study the influence of stone columns grid pattern and spacing on deformation of track
due to cyclic loading.
 Relative density of the column has an influence on the bearing capacity of the soil/
stone composite soils. Therefore, it might be ideal to investigate the effect dilatancy of
granular material on both settlement and bearing capacity.
163
Samir Ashour Reference
REFERENCE
Aboshi, H., Ichimoto, E., Harada, K., et al. (1979) "The composer: A method to improve the
characteristics of soft clays by inclusion of large diameter sand columns". Proc., Int. Conf.
on Soil Reinforcement., E.N.P.C.,. Paris.
Adalier, K. and Elgamal, A. (2004) Mitigation of liquefaction and associated ground
deformations by stone columns. Engineering Geology, 72: (3): 275-291.
Ahmadi, M. and Robertson, P. (2004) Calibration chamber size and boundary effect for CPT
q (c) measurements. Geotechnical and Geophysical Site Characterization. vols, 1: 829-
833.
Ali, K., Shahu, J.T. and Sharma, K.G. (2014) Model tests on single and groups of stone
columns with different geosynthetic reinforcement arrangement. Geosynthetics
International, 21: 103-118(115).
Ambily, A. and Gandhi, S. (2004) Experimental and theoretical evaluation of stone column in
soft clay. ICGGE-2004, 201-206.
Ambily, A. and Gandhi, S. (2007) Behavior of Stone Columns Based on Experimental and
FEM Analysis. Journal of Geotechnical and Geoenvironmental Engineering, 133: 405-
415.
Andersen, K.H. (2004) "Cyclic clay data for foundation design of structures subjected to wave
loading". Proceedings of the International Conference on Cyclic Behaviour of Soils and
Liquefaction Phenomena, CBS04, Bochum, Germany.
Andersen, K.H., Rosenbrand, W.F., Brown, S.F., et al. (1980) Cyclic and static laboratory
tests on Drammen clay. Journal of the Geotechnical Engineering Division, 106: (5): 499-
529.
Anderson, W., Pyrah, I. and Fryer, S. (1991) A clay calibration chamber for testing field
devices. Geotechnical Testing Journal, 14: (4).
Andreou, P., Frank, R., Frikha, W., et al. (2008) Experimental study on sand and gravel
columns in clay. Proceedings of the ICE - Ground Improvement, 161: 189-198.
Ansal, A. and Erken, A. (1989) Undrained Behavior of Clay Under Cyclic Shear Stresses.
Journal of Geotechnical Engineering, 115: 968-983.
Babu, M.D., Nayak, S. and Shivashankar, R. (2013) A critical review of construction, analysis
and behaviour of stone columns. Geotechnical and Geological Engineering, 31: (1): 1-22.
Bae, W.-S., Shin, B.-W., An, B.-C., et al. (2002) "Behaviors of Foundation System Improved
with Stone Columns". The Twelfth International Offshore and Polar Engineering
Conference. Kitakyushu, Japan.
Balaam, N.P. and Booker, J.R. (1981) Analysis of rigid rafts supported by granular piles.
International Journal for Numerical and Analytical Methods in Geomechanics, 5: (4):
379-403.
Balaam, N.P., Brown, P.T., Brown, H.G., et al. (1977) Settlement Analysis of Soft Clays
Reinforced with Granular Piles. University of Sydney, School of Civil Engineering, Civil
Engineering Laboratories.
164
Barksdale, R.D. and Bachus, R.C. (1983) "Design and construction of stone column ". Federal
Highway Administration, RD-83/026.
Bell, A. (2004) The development and importance of construction technique in deep vibratory
ground improvement. Ground and Soil Improvement, 103-111.
Bergado, D., Alfaro, M. and Chai, J. (1991) The granular pile: its present state and future
prospects for improvement of soft Bangkok clay. Geotechnical Engineering, 22: 143-175.
Bergado, D., Huat, S.H. and Kalvade, S. (1987) "Improvement of soft Bangkok clay using
granular piles in subsiding environment". In: Proceedings, 5th international geotechnical
seminar on case histories in soft clay. Singapore.
Bergado, D. and Lam, F.L. (1987) Full scale load test of granular piles with different densities
and different proportions of gravel and sand on soft Bangkok clay. Soil Found, 27: (1): 86-
93.
Black, J., Sivakumar, V. and Bell, A. (2011) The settlement performance of stone column
foundations. Geotechnique, 61: 909-922.
Black, J., Sivakumar, V., Madhav, M., et al. (2006) An Improved Experimental Test Set-up to
Study the Performance of Granular Columns. Geotechnical Testing Journal, 29: 14195.
Bolton, M.D. (1991) A guide to soil mechanics. Universities Press.
Bouassida, M., Guetif, Z., De Buhan, P., et al. (2003) Variational approach for settlement
estimation of a foundation on soil reinforced by columns. Revue Française de
Géotechnique, 102: (1): 21-29.
Bowles, J.E. (1996) "Foundation analysis and design". Fifth ed., McGraw-Hill, New York.
Brough, M., Ghataora, G., Stirling, A., et al. (2003) Investigation of railway track subgrade. I:
In-situ assessment. Proceedings of the ICE-Transport, 156: (3): 145-154.
Brown, S.F. (1996) Soil mechanics in pavement engineering. Geotechnique, 46: 383-426.
Brown, S.F. and Hyde, A. (1975) Significance of cyclic confining stress in repeated-load
triaxial testing of granular material. Transportation Research Record, (537).
Brown, S.F., Lashine, A.K.F. and Hyde, A.F.L. (1975) Repeated load triaxial testing of a silty
clay. 25: 95 - 114.
BSI, B.S.I. (2005) "Execution of special geotechnical works - Ground treatment by deep
vibration". London, BS EN 14731:2005, BSI.
Cai, Y., Gu, C., Wang, J., et al. (2013) One-Way Cyclic Triaxial Behavior of Saturated Clay:
Comparison between Constant and Variable Confining Pressure. Journal of Geotechnical
and Geoenvironmental Engineering, 139: 797-809.
Cai, Y., Sun, Q., Guo, L., et al. (2015) Permanent deformation characteristics of saturated
sand under cyclic loading. Canadian Geotechnical Journal, 52: (6): 1-13.
Charles, J. and Watts, K. (1983) "Compressibility of soft clay reinforced with granular
columns". Proceedings of the 8th European conference on soil mechanics and foundation
Engineering, Helsinki.
Chow, Y. (1996) Settlement Analysis of Sand Compaction Pile. Soils and Foundations, 36:
(1): 111-113.
165
Ellouze, S., Bouassida, M., Hazzar, L., et al. (2010) On settlement of stone column foundation
by Priebe's method. Proceedings of the ICE-Ground Improvement, 163: (2): 101-107.
Fatahi, B., Khabbaz, H. and Le, T.M. (2012) Improvement of rail track subgrade using stone
columns combined with geosynthetics. Advances in Transportation Geotechnics Ii, 202-
206.
Françaises, N. (2005) Recommendations for the design, calculation, installation and control of
stone column foundations under buildings and structures sensitive to settlement. Revue
Française de Géotechnique, 111: (2): 3-16.
Frikha, W., Bouassida, M. and Canou, J. (2014) Parametric Study of a Clayey Specimen
Reinforced by a Granular Column. International Journal of Geomechanics.
Frost, M.W., Fleming, P.R. and Rogers, C.D. (2004) Cyclic triaxial tests on clay subgrades
for analytical pavement design. Journal of transportation Engineering, 130: (3): 378-386.
Gniel, J. and Bouazza, A. (2007) "Methods used for the design of conventional and
geosynthetic reinforced stone columns". 10th Australia New Zealand Conf. on
Geomechanics (Australian Geomechanics Society and New Zealand Geomechanics
Society 21/10/2007 to 24/10/2008). Geomechanics Society.
Gniel, J. and Bouazza, A. (2009) Improvement of soft soils using geogrid encased stone
columns. Geotextiles and Geomembranes, 27: (3): 167-175.
Greenwood, D. (1970) "Mechanical improvement of soils below ground surface".
Proceedings, Ground Engineering Conference, Institution of Civil Engineers London,
UK.
Greenwood, D. (1991) Load tests on stone columns. Deep Foundation Improvements:
Design, Construction, and Testing, 1089: 148.
Greenwood, D. and Kirsch, K. (1984) "Specialist ground treatment by vibratory and dynamic
methods".
Gu, C., Wang, J., Cai, Y., et al. (2012) Undrained cyclic triaxial behavior of saturated clays
under variable confining pressure. Soil Dynamics and Earthquake Engineering, 40: 118-
128.
Han, J. (2015) Recent research and development of ground column technologies.
Proceedings of the Institution of Civil Engineers - Ground Improvement, 168: (4): 246-
264.
Hanna, A., Etezad, M. and Ayadat, T. (2013) Mode of failure of a group of stone columns in
soft soil. International Journal of Geomechanics.
Head, K.H. (1996) Manual of Soil Laboratory Testing, Permeability, Shear Strength and
Compressibility Tests. Wiley.
Head, K.H. (1998) Manual of Soil Laboratory Testing, Effective Stress Tests. Wiley.
Heath, D., Waters, J., Shenton, M., et al. (1972) "Design of conventional rail track
foundations". ICE Proceedings. Thomas Telford.
Hu, W. (1995) Physical modelling of group behaviour of stone column foundations.
University of Glasgow.
166
Hughes, J.M.O. and Withers, N.J. (1974) Reinforcing of soft cohesive soils with stone
columns. GROUND ENGINEERING, 7.
Hughes, J.M.O., Withers, N.J. and Greenwood, D.A. (1976) A field trial of the reinforcing
effect of a stone column in soil. Ground treatment by deep compaction.
Hyde, A., Yasuhara, K. and Hirao, K. (1993) Stability Criteria for Marine Clay under One‐
Way Cyclic Loading. Journal of Geotechnical Engineering, 119: 1771-1789.
ICE (1987) "Specification for Ground Treatment". Thomas Telford.
Indraratna, B., Attya, A. and Rujikiatkamjorn, C. (2009) Experimental Investigation on
Effectiveness of a Vertical Drain under Cyclic Loads. Journal of Geotechnical and
Geoenvironmental Engineering, 135: 835-839.
Jefferson, I., Gaterell, M., Thomas, A.M., et al. (2010) Emissions assessment related to vibro
stone columns. Proceedings of the ICE-Ground Improvement, 163: (1): 71-77.
Jiang, M., Cai, Z., Cao, P., et al. (2010) "Effect of Cyclic Loading Frequency on Dynamic
Properties of Marine Clay". Soil Dynamics and Earthquake Engineering. American Society
of Civil Engineers 240-245.
Juran, I. and Guermazi, A. (1988) Settlement Response of Soft Soils Reinforced by
Compacted Sand Columns. Journal of Geotechnical Engineering, 114: 930-943.
Killeen, M. (2012) Numerical modelling of small groups of stone columns. PhD, National
University of Ireland, Galway.
Killeen, M. and McCabe, B.A. (2014) Settlement performance of pad footings on soft clay
supported by stone columns: A numerical study. Soils and Foundations, 54: (4): 760-776.
Kim, B.-I. and Lee, S.-H. (2005) Comparison of bearing capacity characteristics of sand and
gravel compaction pile treated ground. KSCE Journal of Civil Engineering, 9: (3): 197-203.
Kimura, T. and Saitoh, K. (1983) The influence of strain rate on pore pressures in
consolidated undrained triaxial tests on cohesive soils. Soils and Foundations, 23: (1): 80-90.
Kolekar, Y.A., Mir, O.S. and Dasaka, S.M. (2011) " Behaviour of stone column reinforced
marine clay under static and cyclic loading". Proceedings of Indian Geotechnical
Conference. December 15-17, Kochi.
Lackenby, J., Indraratna, B., McDowell, G., et al. (2007) Effect of confining pressure on
ballast degradation and deformation under cyclic triaxial loading. Geotechnique, 57: (6):
527-536.
Li, D. and Selig, E. (1996) Cumulative Plastic Deformation for Fine-Grained Subgrade Soils.
Journal of Geotechnical Engineering, 122: 1006-1013.
Loh, R.B. and Nikraz, H.R. (2012) Compressibility and Strength of Clay Subgrade of
Railroad Foundation in Highly Saturated Condition. Journal of GeoEngineering, 7: (2): 043-
051.
Maakaroun, T., Najjar, S. and Sadek, S. (2009) "Effect of Sand Columns on the Load
Response of Soft Clays". Contemporary Topics in Ground Modification, Problem Soils,
and Geo-Support. American Society of Civil Engineers 217-224.
Madun, A. (2012) Seismic evaluation of vibro-stone column. PhD, University of
Birmingham.
167
Madun, A., Jefferson, I., Foo, K., et al. (2012) Characterization and quality control of stone
columns using surface wave testing. Canadian Geotechnical Journal, 49: (12): 1357-1368.
McCabe, B.A., Nimmons, G.J. and Egan, D. (2009) A review of field performance of stone
columns in soft soils. Proceedings of the Institution of Civil Engineers-Geotechnical
engineering, 162: 323–334.
McKelvey, D. (2002) The performance of vibro stone column reinforced foundations in
deep soft ground. PhD, Queen's University of Belfast.
McKelvey, D. and Sivakumar, V. (2000) "A review of the performance of vibro stone column
foundations". Proceedings of the 3rd International Conference on Ground Improvement
Techniques, Singapore.
McKelvey, D., Sivakumar, V., Bell, A., et al. (2002) Shear strength of recycled construction
materials intended for use in vibro ground improvement. Proceedings of the ICE-Ground
Improvement, 6: (2): 59-68.
Miller, G.A., Teh, S.Y., Li, D., et al. (2000) Cyclic Shear Strength of Soft Railroad Subgrade.
Journal of Geotechnical and Geoenvironmental Engineering, 126: 139-147.
Mitchell, J.M. and Jardine, F.M. (2002) A Guide to Ground Treatment. Construction
Industry Research & Information Association.
Mohanty, P. and Samanta, M. (2015) Experimental and Numerical Studies on Response of the
Stone Column in Layered Soil. International Journal of Geosynthetics and Ground
Engineering, 1: (3): 1-14.
Moseley, M.P. and Kirsch, K. (2004) Ground improvement. London; New York: Spon
Press.
Munfakh, G.A. (1984) "Soil Reinforcement by stone columns–varied case applications".
Munfakh, G.A. (1997) Ground improvement engineering–the state of the US practice: part 1.
Methods. Ground Improvement, 1: 193–214.
Munfakh, G.A. (2003) Ground improvement in transportation projects: from old visions to
innovative applications. Proceedings of the ICE - Ground Improvement, 7: 47-60.
Najjar, S. (2013) A State-of-the-Art Review of Stone/Sand-Column Reinforced Clay Systems.
Geotechnical and Geological Engineering, 31: 355-386.
Najjar, S., Sadek, S. and Maakaroun, T. (2010) Effect of sand columns on the undrained load
response of soft clays. Journal of Geotechnical and Geoenvironmental Engineering.
Nakase, A. and Kamei, T. (1986) Influence of strain rate on undrained shear characteristics of
K0-consolidated cohesive soils. Japanese Socity of Soil Mechanics and Foundation
Engineering, 26: (1): 85-95.
Ng, C., Zhou, C., Yuan, Q., et al. (2013) Resilient modulus of unsaturated subgrade soil:
experimental and theoretical investigations. Canadian Geotechnical Journal, 50: (2): 223-
232.
Peacock, W. and Bolton, S. (1968) Sand liquefaction under cyclic loading simple shear
conditions. Journal of Soil Mechanics & Foundations Div.
Poorooshasb, H.B. and Meyerhof, G.G. (1997) Analysis of behavior of stone columns and
lime columns. Computers and Geotechnics, 20: 47–70.
168
Powrie, W. (2013) Soil mechanics: concepts and applications. CRC Press.

Priebe, H.J. (1995) The design of vibro replacement. GROUND ENGINEERING, 28: 31–
37.
Raju, V. (2003) "Ground Improvement Techniques for Railway Embankments". Railtech
Conference.
Raju, V. and Sondermann, W. (2005) Ground improvement using deep vibro techniques.
Elsevier Geo-Engineering Book Series, 3: 601-638.
Redgers, J.D., Moxhay, A.L., Ghataora, G.S., et al. (2008) Case Histories of Settlement
Performance Comparisons on Ground Improvement Using Soil Stiffness Seismic Wave and
Traditional Methods. Proceedings of The International Conference On Case Histories In
Geotechnical Engineering.
Rollins, K., Price, B., Dibb, E., et al. (2006) Liquefaction mitigation of silty sands in Utah
using stone columns with wick drains. Ground Modification and Seismic Mitigation, 343-
348.
Rowe, P.W. and Barden, L. (1966) A New Consolidation Cell. Geotechnique, 16: (2): 162-
&.
Rudolph, R., Serna, B. and Farrell, T. (2011) Mitigation of Liquefaction Potential Using
Rammed Aggregate Piers. Advances in Geotechnical Engineering, 557-566.
Salahi, A., Niroumand, H. and Kassim, K.A. (2015) Evaluation of Stone Columns versus
Liquefaction Phenomenon. Electronic Journal of Geotechnical Engineering (EJGE), 20:
(2).
Samadhiya, N., Maheswari, P., Basu, P., et al. (2008) Load-settlement characteristics of
granular piles with randomly mixed fibres. Indian Geotech. J, 38: (3): 345-354.
Schanz, T., Vermeer, P. and Bonnier, P. (1999) The hardening soil model: formulation and
verification. Beyond 2000 in computational geotechnics, 281-296.
Seed, H.B. and Booker, J.R. (1977) Stabilization of potentially liquefiable sand deposits using
gravel drain systems. ASCE Journal of Geotechnical Engineering, 103: (7): 757-768.
Seed, H.B., Chan, C.K. and Monismith, C.L. (1955) "Effects of repeated loading on the
strength and deformation of compacted clay". Highway Research Board Proceedings.
Sekine, E. (1996) Bearing Capacity of Roadbed and its Reinforcement. Department of
Civil Engineering, Nagaoka University of Technology for the degree of Doctor.
Selig, E.T. and Waters, J.M. (1994) Track geotechnology and substructure management.
Thomas Telford.
Serridge, C.J. (2005) Achieving sustainability in vibro stone column techniques. Proceedings
of the ICE-Engineering Sustainability, 158: (4): 211-222.
Serridge, C.J. (2006) Some applications of ground improvement techniques in the urban
environment. Engineering Geology for Tomorrow's Cities, Geological Society, London,
Engineering Geology Special Publications, 22.
Sivakumar, V., Bell, A., McKelvey, D., et al. (2004a) Modelling vibrated stone columns in
soft clay. Proceedings of the ICE - Geotechnical Engineering, 157: 137-149.
169
Sivakumar, V., Boyd, J.L., Black, J.A., et al. (2010) Effects of granular columns in compacted
fills. Proceedings of the ICE - Geotechnical Engineering, 163: 189-196.
Sivakumar, V., Glynn, D., Black, J., et al. (2007) "A laboratory model study of the
performance of vibrated stone columns in soft clay". Proceedings of the 14th European
Conference on Soil Mechanics and Geotechnical Engineering, Madrid, Spain.
Sivakumar, V., McKelvey, D., Graham, J., et al. (2004b) Triaxial tests on model sand
columns in clay. Canadian Geotechnical Journal, 41: 299-312.
Skempton, A. (1957) Discussion: Further data on the c/p ratio in normally consolidated clays.
Proceedings of the Institution of Civil Engineers, 7: 305-307.
Slocombe, B., Bell, A. and Baez, J. (2000) The densification of granular soils using vibro
methods. Geotechnique, 50: (6): 11.
Sun, X., Han, J., Wayne, M.H., et al. (2014) "Quantifying the Benefit of Triaxial Geogrid in
Stabilizing Granular Bases over Soft Subgrade under Cyclic Loading at Different Intensities".
Geo-Congress 2014 Technical Papers at Geo-characterization and Modeling for
Sustainability. ASCE.
Unnikrishnan, N., Rajagopal, K. and Krishnaswamy, N.R. (2002) Behaviour of reinforced
clay under monotonic and cyclic loading. Geotextiles and Geomembranes, 20: 117–133.
Valls-Marquez, M. (2009) Evaluating the capabilities of some constitutive models in
reproducing the experimental behaviour of stiff clay subjected to tunnelling stress paths.
The University of Birmingham.
Van Impe, W.F., De Cock, F., Van der Cruyssen, J.P., et al. (1997) Soil improvement
experiences in Belgium: part II. Vibrocompaction and stone columns. Proceedings of the
ICE-Ground Improvement, 1: 157–168.
Vardanega, P. and Haigh, S. (2014) The undrained strength–liquidity index relationship.
Canadian Geotechnical Journal, 51: (9): 1073-1086.
Vekli, M., Aytekin, M., Ikizler, S.B., et al. (2012) Experimental and numerical investigation
of slope stabilization by stone columns. Natural hazards, 64: (1): 797-820.
Wang, J.-h., Liu, Y.-f. and Xing, Y. (1998) Estimation of undrained bearing capacity for
offshore soft foundations with cyclic load. China Ocean Engineering, 12: (2): 213-222.
Wang, J., Li, C. and Shi, M. (2006) Experimental study on cyclic mechanics characteristic of
saturated soft clay strata. Trans. Tianjin Univ., 12: (2).
Wang, L.-Z., Li, L.-L. and Dan, H.-B. (2011) Undrained behavior of natural marine clay
under cyclic loading. Ocean Engineering, 38: 1792-1805.
Watts, F. (2000) Specifying vibro stone columns. British Research Establishment (BRE)
Report, Watford, UK.
Wood, D.M., Hu, W. and Nash, D.F.T. (2000) Group effects in stone column foundations:
model tests. Geotechnique, 50: 689–698.
Woodward, J. (2004) An Introduction to Geotechnical Processes. CRC Press.
Yasuhara, Hirao, K. and Hyde, A.F.L. (1992) Effects of cyclic loading on undrained strength
and compressibility of clay. Soils and Foundations, 32: (1): 100-116.
170
Yasuhara, Yamanouchi, T. and Hirao, K. (1982) Cyclic strength and deformation of normally
consolidated clay. Soils and Foundations.
Yoo, T. and Selig, E. (1979) Field observations of ballast and subgrade deformations in track.
Transportation Research Record, (733).
Zahmatkesh, A. and Choobbasti, A.J. (2010) Settlement evaluation of soft clay reinforced
with stone columns using the equivalent secant modulus. Arabian Journal of Geosciences.
Zhou, J. and Gong, X. (2001) Strain degradation of saturated clay under cyclic loading.
Canadian Geotechnical Journal, 38: 208-212.
171
Samir Ashour Appendix A
APPENDIX A
Material Properties
 Data of Sieve Analysis for Aggregate Used To Construct Stone Column
Test (1) Test (2) Test (3)

Sieve Retain Retain Retain
(mm) Retain Passin Retain Passin Retain Passin
weight weight weight
(%) g (%) (%) g (%) (%) g (%)
(g) (g) (g)
5 1.1 0.22 99.78 0.8 0.16 99.84 1.00 0.19 99.81
3.35 106.9 21.36 78.42 102.5 20.36 79.48 109.94 20.86 78.95
2.36 153 30.57 47.85 145 28.80 50.69 156.45 29.68 49.27
2 140 27.97 19.88 150 29.79 20.89 152.25 28.88 20.39
1.18 99.2 19.82 0.06 105 20.85 0.04 107.21 20.34 0.05
0.6 0.2 0.04 0.02 0.1 0.02 0.02 0.16 0.03 0.02
Pan 0.1 0.02 0.00 0.1 0.02 0.00 0.11 0.02 0.00
Total
weight 500.5 503.5 527.10
(g)
100.00
90.00 Test (1)
Percentage of passing (%)
80.00 Test (2)

70.00 Test (3)
60.00
50.00
40.00
30.00
20.00
10.00
0.00
0.1 1 10
Particale size (mm)
172
 China clay particle size distribution (hydrometer sedimentation)
Elapsed Rh' + Effective Particle

Temperature Reading Rh' - Ro' Percentage
time, Cm depth Hr diameter
C° Rh' =Rd finer (%)
(min) =Rh (mm) D (mm)
0 25 25 25.5 97.07 24.5
0.5 25 19 19.5 121.41 0.06 18.5 99.0
1 25 19 19.5 121.41 0.04 18.5 99.0
2 25 19 19.5 121.41 0.03 18.5 99.0
4 25 19 19.5 121.41 0.02 18.5 99.0
8 25 18.5 19 123.44 0.02 18 96.4
15 25 18 18.5 125.46 0.01 17.5 93.7
30 25 17.5 18 127.49 0.01 17 91.0
60 25 16 16.5 133.58 0.01 15.5 83.0
120 25 14 14.5 141.69 0.00 13.5 72.3
240 25 12.5 13 147.77 0.00 12 64.2
450 25 10.5 11 155.88 0.00 10 53.5
1420 25 8.5 9 164.00 0.00 8 42.8
Particle size supplied by manufacturer WBB
size size Percentage

(micron) (mm) (%)
1 0.001 37
2 0.002 49
5 0.005 76
10 0.01 94
20 0.02 99
100
90
80
Percentage passing, %
70
60
50
40
30
20 Hydrometer test
10 Supplied by manufacturer WBB
0
0.00 0.01 0.10
Particle size, mm
173
 Soil Index
 Liquid Limit
Test No.: (1) (2) (3) (4)
dial gage reading 14.00 13.80 14.10 19.40 19.00 18.90 19.90 20.50 20.30 23.20 22.90 23.50
Average
13.97 19.10 20.23 23.20
Penetration
Container # p6 pla3 y11 LLL m45 120 E 342y w17 8A mc51 mc8
Mass of container
3.15 3.99 3.19 7.84 5.49 4.41 3.34 3.39 4.72 4.73 4.76 5.40
g
Mass of container +
16.31 19.24 23.67 21.59 26.45 20.52 17.11 18.34 25.51 20.64 28.46 26.06
Wet Soil (g)
Mass of container +
11.75 13.95 16.56 16.70 19.00 14.79 12.17 12.98 18.06 14.83 19.80 18.51
Dry Soil (g)
Mass of Dry Soil

8.60 9.96 13.37 8.86 13.51 10.38 8.83 9.59 13.34 10.10 15.04 13.11
(g)
Mass of Moisture
4.56 5.29 7.11 4.89 7.45 5.73 4.94 5.36 7.45 5.81 8.66 7.55
(g)
Moisture content
53.02 53.11 53.18 55.19 55.14 55.20 55.95 55.89 55.85 57.52 57.58 57.59
(%)
Ave. moisture
53.10 55.18 55.89 57.56
content (%)
174
25.00
Penetration of cone, (mm)
20.00
15.00
10.00
50.00 52.00 54.00 56.00 58.00 60.00
Water content, (%)
 Plastic Limit
Test No.: (1) (2) (3) (4)

Mass of container ( g) 3.22 3.38 4.43 3.99
Mass of container + Wet Soil ( g) 3.97 4.47 4.96 4.82
Mass of container + Dry Soil (g) 3.81 4.24 4.85 4.64
Mass of Dry Soil ( g) 0.59 0.86 0.42 0.65
Mass of Moisture (g) 0.16 0.23 0.11 0.18
Moisture content (%) 27.12 27.09 27.14 27.08
Ave. water content (%) 27.11
175
 Compaction Test Data
Location: Civil Engineering

BS 1377:1990 Test NO.: Lab
No. of layers: 3 Rammer: 2.5 kg
Soil description: English china
Blow per layer : 25 Drop: 300mm clay
Compacted by: Sample preparation: Sample type:
Proctor cylinder NO.: 1 No. of separate batches:
Density : Volume of cylinder "V": (Dia: 104.903mm, length : 115.470mm):

997.952mm3
Test No.: (1) (2) (3)
Cylinder & soil, A (g) 6679 6785 6801

Cylinder, B (g) 4978.5 4978.5 4978.5
Wet soil, A-B (g) 1700.5 1806.5 1822.5
Volume, cm3 0.998 0.998 0.998
3
Wet density, ( Mg/m ) 1.70 1.81 1.83
Vane shear strength over 120 over 120 53
Moisture content
Container No.: f28t31 P12 66 S10 B22-8 212ms2 m9t2 T25 M7T
Wet soil & container, g 8.43 8.64 7.76 14.67 12.19 21.01 11.97 17.55 13.51
Dry soil & container, g 7.41 7.72 6.98 12.20 10.66 17.33 9.84 14.43 11.41
Container,( g) 3.17 3.92 3.67 3.39 5.32 4.22 3.36 4.70 4.78
Dry soil , (g) 4.24 3.80 3.31 8.81 5.34 13.11 6.48 9.73 6.63
Moisture loss, (g) 1.02 0.92 0.78 2.47 1.53 3.68 2.13 3.12 2.10
Moisture content,( %) 24.06 24.21 23.56 28.04 28.65 28.07 32.87 32.07 31.67
Average moisture, (%) 23.94 28.25 32.20
Dry density, (Mg/m3) - 1.41 1.38
Test NO.: 4 5 6
Cylinder & soil, A g 6738 6677 6593
Cylinder, B g 4978.5 4978.5 4979.3
Wet soil, A-B g 1759.5 1698.5 1613.7
Volume, cm3 0.998 0.998 0.998
Wet density, ρ Mg/m3 1.76 1.70 1.62
Vane shear strength 23 14 -
176
Moisture content
Container No.: f28t 39s CL17 SA01 SA02 SA03 m90u f28t31 39s
Wet soil & container, g 18.65 8.65 21.11 12.59 12.14 18.65 19.50 11.17 13.18
Dry soil & container, g 14.96 7.19 17.22 9.93 9.65 14.22 14.46 8.33 9.64
Container, g 4.77 3.18 6.54 3.29 3.50 3.19 5.32 3.18 3.18
Dry soil , g 10.19 4.01 10.68 6.64 6.15 11.03 9.14 5.15 6.46
Moisture loss, g 3.69 1.46 3.89 2.66 2.49 4.43 5.04 2.84 3.54
Moisture content, % 36.21 36.41 36.42 40.06 40.49 40.16 55.14 55.15 54.80
Average moisture, % 36.35 40.24 55.03
Dry density ρd Mg/m3 1.29 1.21 1.04
1.90
1.80
Dry Density (Mg/m3)
1.70
1.60 5%
10%
1.50 0%
Dry Density
1.40 Wet Density
1.30
1.20
1.10
1.00
10 20 30 40 50
Moisture Content %
 Undrained Shear Strength for Kaolin Clay
 Hand Vane Shear Test
Test No.: Water content (%) Shear strength (kPa)
A B C Average A B C Average
Test (1) 32.87 32.07 31.67 32.2 52 53 53 52.6
Test (2) 36.21 36.41 36.42 36.35 25 24 25 24.6
Test (3) 40.06 40.49 41.16 40.24 15 14 14 14.3
Test (4) 42.2 42.15 41.85 42.06 9 9 10 9.3
Test (5) 44.5 44.7 44.1 44.43 9 8 7 8
Test (6) 55.14 55.15 54.80 47.31 4 5 6 5
177
 Quick Undrained Triaxial Test
Test Average Average deviator stress results Shear strength
No.: water content σ3= 50 σ3= 100 σ3= 150 (kPa)
(%) kPa kPa kPa
Test (1) 35 108 109 112 54.4
Test (2) 38 37 38 39 18.5
Test (3) 42.5 30 31 33 15.5
Test (4) 44 25.56 23.5 25.32 12.3
80
70 Vane shear test

Quick undrained triaxial test
60
Shear strength; kPa
50
40
30
20
10
0
30.00 35.00 40.00 45.00 50.00
Water content; %
178
4.1 Shear Box Data Results for Column Material
Shear stress results (wet condition) Average internal

Test
angle of friction ∅
No.: 25 kPa 50 kPa 100 kPa 200 kPa
(o)
Test (1) 37 64.13 121.63 217.66

48
Test (2) 38.20 63.21 120.1 210
Shear stress results (dry condition) Average internal

Test
angle of friction ∅
No.: 25 kPa 50 kPa 100 kPa 200 kPa
(o)
Test (1) 48 75.11 141.57 239.1

50
Test (2) 52.25 71.40 143.54 242.44
250
Wet condition 200kPa
200 Test 1_200kPa
Test 2_200kPa
Shear stress, (kPa)
Test 1_100kPa
150 100kPa
Test 2_100kPa
Test 1_50kPa
100 Test 2_50kPa
50kPa Test 1_25kPa
Test 2_25kPa
50
25kPa
0
0 2 4 6 8 10
Horisontal displacement, (mm)
179
300
Dry condition
250 200 kPa
Test 1_ 25 kPa
200
Shear stress, (kPa)
Test 1_50 kPa

100 kPa Test 1_100 kPa
150 Test 1_200 kPa
Test 2_25kPa
100 50 kPa Test 2_50kPa
Test 2_100kPa
50 Test 2_200kPa
25 kPa
0
0 2 4 6 8 10
Horizontal displacement, (mm)
300
Wet condition
250
Dry condition
200
Shear stress; kPa
150
100
50
0
0 50 100 150 200 250
Normal stress, (kPa)
180
 Specific Gravity Test Data
 Specific Gravity for Kaolin Clay
Test No.: 1 2 3 4
Mass of bottle (g) m1 48.16 46.88 49.48 48.57
Mass of bottle & soil (g) m2 49.97 49.51 51.59 50.74
Mass of bottle water & soil (g) m3 104.05 108.38 108.49 105.51
Mass of bottle full water (g) m4 102.93 106.75 107.18 104.17
Specific gravity 2.62 2.64 2.64 2.61
average 2.63
 Specific Gravity for Column Material
Test No.: 1 2 3 4
Mass of bottle (g) m1 48.16 46.88 49.48 48.57
Mass of bottle & soil (g) m2 60.98 61.03 59.62 58.94
Mass of bottle water & soil (g) m3 110.97 115.72 113.60 110.67
Mass of bottle full water (g) m4 102.90 106.78 107.19 104.15
Specific gravity 2.70 2.72 2.72 2.69
average 2.71
 Clay Bed Consolidation
Pressure Test (1) Test (2) Test (3)

Root Root Root
Consolidation pressure (6.45 kPa)
Time Deformation Time Deformation Time Deformation

Square Square Square
(minute) (mm) (minute) (mm) (minute) (mm)
Time Time Time
0 0.00 0.00 0 0.00 0.00 0 0.00 0.00
1 1.00 0.15 1 1.00 0.19 1 1.00 0.20
2 1.41 0.30 2 1.41 0.35 2 1.41 0.34

4 2.00 0.55 4 2.00 0.63 4 2.00 0.40
8 2.83 0.98 8 2.83 1.15 8 2.83 1.03
15 3.87 1.60 15 3.87 1.95 20 4.47 2.14
30 5.48 2.64 30 5.48 3.18 33 5.74 3.10
60 7.75 4.30 60 7.75 5.25 60 7.75 4.80
120 10.95 9.05 120 10.95 8.50 240 15.49 16.85
240 15.49 14.82 240 15.49 17.00 1486 38.55 38.57
181
480 21.91 22.30 432 20.78 23.00 1526 39.06 39.09

1440 37.95 36.50 1476 38.42 39.95 2788 52.80 51.15
2880 53.66 49.81 2827 53.17 45.95 2936 54.18 51.52
3600 60.00 50.10
Root Root Root
Time Time Time
Consolidation pressure (12 kPa)
0 0.00 0.00 0 0.00 0.00 0 0.00 0.00

1 1.00 0.63 1 1.00 0.33 1 1.00 0.27
2 1.41 0.81 2 1.41 0.62 2 1.41 0.33
4 2.00 1.07 4 2.00 1.07 4 2.00 0.48
8 2.83 1.45 8 2.83 1.50 8 2.83 0.70
15 3.87 1.95 15 3.87 2.10 15 3.87 1.13
30 5.48 2.75 30 5.48 2.67 30 5.48 1.70
60 7.75 3.89 60 7.75 3.50 65 8.06 2.47
120 10.95 4.84 120 10.95 4.68 115 10.72 3.50
240 15.49 6.54 240 15.49 6.78 1322 36.36 12.84
450 21.21 8.75 480 21.91 9.42 1460 38.21 13.25
1329 36.46 10.90 1440 37.95 11.65
1440 37.95 11.00
1635 40.44 11.40
Root Root Root
Time Time Time
0 0.00 0.00 0 0.00 0.00 0 0.00 0.00

1 1.00 0.60 1 1.00 0.61 1 1.00 0.61
2 1.41 0.79 2 1.41 0.76 2 1.41 0.76
4 2.00 0.98 4 2.00 0.97 4 2.00 0.97
9 3.00 1.35 9 3.00 1.25 9 3.00 1.25
21 4.58 1.95 15 3.87 1.67 15 3.87 1.67
30 5.48 2.32 30 5.48 2.29 30 5.48 2.29
60 7.75 3.26 53 7.28 3.05 53 7.28 3.05
120 10.95 4.55 116 10.77 4.45 116 10.77 4.45
200 14.14 5.77 200 14.14 6.37 200 14.14 6.37
462 21.49 7.95 456 21.35 8.16 456 21.35 8.16
1448 38.05 8.95 1460 38.21 9.30 1460 38.21 9.30

Root Root Root
Consolidation pressure

Time Time Time
0 0.00 0.00 0 0.00 0.00 0 0.00 0.00
(50 kPa)
1 1.00 1.00 1 1.00 0.80 1 1.00 0.80

2 1.41 1.20 2 1.41 0.97 2 1.41 0.97
4 2.00 1.51 4 2.00 1.21 4 2.00 1.21
8 2.83 1.97 8 2.83 1.58 8 2.83 1.58
15 3.87 2.70 15 3.87 2.04 15 3.87 2.04
30 5.48 3.40 30 5.48 2.84 30 5.48 2.84
182
60 7.75 4.61 63 7.94 3.91 63 7.94 3.91

120 10.95 6.33 123 11.09 5.60 123 11.09 5.90
240 15.49 8.31 240 15.49 7.57 240 15.49 7.87
480 21.91 9.65 360 18.97 8.55 360 18.97 8.85
1440 37.95 10.02 440 20.98 8.86 420 20.49 9.06
1440 37.95 9.42 1414 37.60 9.62
Root Root Root

Time Time Time
0 0.00 0.00 0 0.00 0.00 0 0.00 0.00
1 1.00 0.35 1 1.00 0.35 1 1.00 0.35
2 1.41 0.50 2 1.41 0.55 2 1.41 0.55
4 2.00 0.82 4 2.00 0.88 4 2.00 0.88
8 2.83 1.25 8 2.83 1.30 8 2.83 1.30
15 3.87 1.87 15 3.87 1.90 15 3.87 1.70
30 5.48 2.88 30 5.48 3.20 30 5.48 2.40
60 7.75 4.25 60 7.75 5.10 60 7.75 4.10
120 10.95 6.02 120 10.95 7.33 120 10.95 6.33
240 15.49 7.50 240 15.49 9.25 240 15.49 8.25
485 22.02 8.17 449 21.19 9.84 449 21.19 8.84
1440 37.95 8.40 1380 37.15 10.14 1380 37.15 9.14
Square root of time

0 20 40 60
0.00
Test (1)
Test (2)
20.00
Test (3)
Settlement (mm)
40.00
60.00
Consolidation pressure 6.45 kPa
183
Square root of time

0 20 40
0.00
Test (1)
Test (2)
Test (3)
Settlement (mm)
10.00
20.00
Consolidation pressure 12 kPa
Square root of time

0 20 40
0.00
Test (1)
Test (2)
Test (3)
Settlement (mm)
5.00
10.00
184
Square root of time

0 20 40
0.00
Test (1)
Test (2)
5.00 Test (3)
Settlement (mm)
10.00
15.00
Square root of time

0 20 40
0.00
Test (1)
Test (2)
5.00 Test (3)
Settlement (mm)
10.00
15.00
185
Samir Ashour Appendix B
APPENDIX B
1. Test Procedure
 Triaxial Test Procedure
 Consolidaing the sample
186
 Installing the stone colum
 Extroding the sample
 Placing the sample at the Mand Machine
187
 Large Scale Model Procedure
 Sample consolidation
 Stone colum Installation
188
 Positioning and placing the foundation
 Placing the sample at the Mand Machine
189
Samir Ashour Appendix C
APPENDIX C
 Example of the data of the monotonic loading test on reinforce soil
Date 21/08/2013 Test Number (5)

Time 08:30:00 Water content (%) 42.86
Area of specimen
sample Dia.(mm) 101.00 0.01
(m^2)
sample Length (mm) 199.50
Cell pressure (kPa) 400.67
Back pressure (kPa) 295.00 Strain rate (mm/min) 0.03
stress Change
Load Pk Stress corr. Disp. Strain PWP PWP σ3
Posn Pk (kN) (kPa) (kPa) (mm) (%) (kPa) (kPa) (kPa)
-23.11 0.00 0.00 0.00 0.00 0.00 294.77 0 400.58
-23.05 0.04 4.46 4.45 0.38 0.18 295.15 0.54 400.67
-22.99 0.03 3.69 3.68 0.45 0.22 295.32 0.72 400.66
-22.92 0.02 2.67 2.67 0.52 0.25 295.54 0.93 400.69
-22.86 0.02 3.06 3.05 0.58 0.28 296.03 1.42 400.64
-22.79 0.09 10.95 10.92 0.65 0.31 297.80 3.20 400.77
-22.73 0.10 12.73 12.69 0.71 0.34 300.62 6.02 400.79
-22.67 0.16 19.74 19.66 0.77 0.37 302.96 8.36 400.53
-22.60 0.19 24.70 24.60 0.84 0.40 304.98 10.38 400.60
-22.54 0.19 24.70 24.59 0.91 0.44 306.82 12.22 400.56
-22.47 0.21 26.87 26.74 0.97 0.47 308.27 13.67 400.61
-22.40 0.24 31.07 30.91 1.04 0.50 309.80 15.20 400.85
-22.34 0.22 27.88 27.74 1.10 0.53 311.01 16.41 400.84
-22.27 0.25 32.21 32.03 1.17 0.56 312.15 17.54 400.49
-22.21 0.24 31.07 30.88 1.23 0.59 313.54 18.94 400.26
-22.14 0.29 37.31 37.07 1.30 0.62 314.67 20.07 400.63
-22.08 0.28 36.16 35.92 1.36 0.66 315.88 21.27 400.40
-22.01 0.32 40.74 40.46 1.43 0.69 316.77 22.17 400.81
-21.95 0.31 39.22 38.93 1.49 0.72 317.77 23.17 400.63
-21.88 0.30 38.45 38.16 1.56 0.75 318.45 23.84 400.57
-21.82 0.30 37.94 37.65 1.63 0.78 318.70 24.10 400.52
-21.75 0.34 43.29 42.94 1.69 0.81 319.69 25.09 400.40
-21.69 0.36 46.09 45.70 1.76 0.84 320.74 26.14 400.65
-21.62 0.38 47.75 47.33 1.82 0.87 321.64 27.03 400.67
-21.56 0.36 45.84 45.42 1.89 0.91 322.10 27.50 400.66
-21.49 0.41 52.71 52.22 1.95 0.94 322.87 28.27 400.82
190
-21.43 0.38 48.51 48.04 2.02 0.97 323.43 28.83 400.39

-21.36 0.38 48.26 47.77 2.08 1.00 323.92 29.32 400.46
-21.29 0.36 45.84 45.36 2.15 1.03 324.57 29.96 400.23
-21.23 0.39 49.27 48.75 2.21 1.06 325.14 30.54 400.59
-21.16 0.43 54.11 53.52 2.28 1.09 325.45 30.85 400.77
-21.10 0.41 52.08 51.49 2.35 1.13 325.98 31.38 400.18
-21.03 0.39 50.17 49.58 2.41 1.16 326.14 31.54 400.43
-20.97 0.41 51.57 50.95 2.47 1.19 326.99 32.39 400.26
-20.90 0.41 52.33 51.69 2.54 1.22 327.29 32.69 400.29
-20.83 0.42 53.73 53.06 2.61 1.25 327.80 33.19 400.51
-20.77 0.46 59.08 58.32 2.67 1.28 328.02 33.42 401.11
-20.70 0.46 58.82 58.05 2.74 1.32 328.46 33.85 400.64
-20.64 0.44 55.39 54.64 2.80 1.35 328.86 34.26 400.35
-20.57 0.47 59.59 58.77 2.87 1.38 329.18 34.58 400.65
-20.51 0.43 54.62 53.85 2.93 1.41 329.53 34.93 400.21
-20.44 0.46 58.82 57.98 3.00 1.44 330.00 35.40 400.63
-20.38 0.50 63.79 62.85 3.06 1.47 330.44 35.84 400.73
-20.31 0.42 53.86 53.05 3.13 1.50 330.28 35.68 399.96
-20.25 0.47 59.59 58.67 3.20 1.54 330.77 36.17 400.20
-20.18 0.47 59.21 58.28 3.26 1.57 330.94 36.34 400.61
-20.11 0.50 63.03 62.02 3.33 1.60 331.35 36.74 400.66
-20.05 0.47 60.35 59.37 3.39 1.63 331.42 36.82 400.51
-19.98 0.50 63.66 62.60 3.46 1.66 331.75 37.15 400.69
-19.92 0.48 60.61 59.58 3.52 1.69 332.01 37.41 400.24
-19.85 0.50 63.03 61.94 3.59 1.73 332.29 37.68 400.67
-19.79 0.54 69.01 67.80 3.65 1.76 332.46 37.86 400.74
-19.72 0.54 68.37 67.15 3.72 1.79 332.74 38.13 400.97
-19.65 0.52 66.72 65.50 3.78 1.82 332.95 38.35 400.79
-19.59 0.50 63.53 62.36 3.85 1.85 333.08 38.47 400.56
-19.52 0.51 64.94 63.71 3.92 1.88 333.40 38.79 400.46
-19.46 0.53 67.48 66.19 3.98 1.92 333.55 38.94 400.47
-19.39 0.41 52.71 51.69 4.05 1.95 333.20 38.60 400.29
-19.32 0.53 67.86 66.52 4.11 1.98 333.74 39.13 400.58
-19.26 0.49 62.64 61.38 4.18 2.01 333.63 39.03 400.21
-19.19 0.52 66.34 64.98 4.24 2.04 333.94 39.33 400.44
-19.13 0.56 70.92 69.45 4.31 2.07 334.34 39.74 400.60
-19.06 0.56 71.05 69.55 4.37 2.10 334.40 39.79 400.47
-19.00 0.47 59.33 58.07 4.44 2.14 334.34 39.74 399.87
-18.93 0.51 65.19 63.78 4.51 2.17 334.68 40.07 400.41
-18.87 0.54 68.63 67.12 4.57 2.20 334.90 40.30 400.48
-18.80 0.56 70.92 69.34 4.64 2.23 335.12 40.52 400.45
-18.74 0.55 70.03 68.44 4.70 2.26 335.31 40.70 400.30
-18.67 0.55 70.03 68.42 4.77 2.29 335.27 40.66 400.33
191
-18.61 0.59 74.74 73.00 4.84 2.33 335.53 40.92 400.65

-18.54 0.57 72.45 70.74 4.90 2.36 335.52 40.92 400.61
-18.47 0.56 70.79 69.10 4.97 2.39 335.57 40.97 400.37
-18.41 0.53 67.10 65.48 5.03 2.42 335.72 41.12 400.13
-18.34 0.56 71.68 69.93 5.10 2.45 335.73 41.13 400.06
-18.28 0.58 74.23 72.39 5.16 2.48 336.09 41.49 400.26
-18.21 0.54 68.88 67.15 5.23 2.52 335.85 41.25 400.28
-18.14 0.55 69.39 67.62 5.30 2.55 335.98 41.38 400.20
-18.08 0.61 77.41 75.42 5.36 2.58 336.20 41.59 400.63
-18.01 0.62 79.32 77.25 5.43 2.61 336.39 41.79 400.57
-17.95 0.56 71.68 69.79 5.49 2.64 336.29 41.68 400.20
-17.88 0.56 70.79 68.90 5.56 2.67 336.48 41.87 400.05
-17.82 0.58 73.21 71.23 5.62 2.70 336.66 42.06 400.24
-17.75 0.61 77.03 74.92 5.69 2.74 336.50 41.90 400.96
-17.68 0.58 74.10 72.05 5.75 2.77 336.70 42.10 400.39
-17.62 0.61 78.05 75.87 5.82 2.80 336.79 42.19 400.32
-17.55 0.60 76.14 73.98 5.89 2.83 336.82 42.21 400.46
-17.49 0.59 75.12 72.97 5.95 2.86 336.96 42.36 400.18
-17.42 0.61 77.03 74.80 6.02 2.89 336.81 42.20 400.61
-17.36 0.60 75.89 73.67 6.08 2.93 336.94 42.34 400.36
-17.29 0.56 70.92 68.82 6.15 2.96 336.99 42.39 399.93
-17.22 0.62 78.43 76.09 6.22 2.99 337.24 42.64 400.21
-17.16 0.59 75.63 73.35 6.28 3.02 337.29 42.69 400.13
-17.09 0.60 76.14 73.82 6.35 3.05 337.11 42.50 399.92
-17.03 0.61 78.18 75.77 6.41 3.08 337.56 42.96 400.46
-16.96 0.61 77.41 75.00 6.48 3.11 337.48 42.88 400.42
-16.90 0.61 78.18 75.72 6.54 3.15 337.54 42.93 400.45
-16.83 0.64 80.85 78.28 6.61 3.18 337.67 43.06 400.12
-16.76 0.60 75.76 73.33 6.68 3.21 337.61 43.01 400.15
-16.70 0.66 83.78 81.06 6.74 3.24 337.78 43.18 400.65
-16.63 0.57 72.32 69.95 6.80 3.27 337.57 42.97 400.02
-16.57 0.62 78.69 76.09 6.87 3.30 337.76 43.16 400.35
-16.50 0.64 81.61 78.89 6.94 3.33 337.66 43.06 400.36
-16.44 0.65 83.02 80.22 7.00 3.37 337.95 43.35 400.58
-16.37 0.66 83.52 80.69 7.07 3.40 338.01 43.41 400.55
-16.31 0.64 81.61 78.82 7.13 3.43 338.01 43.40 400.46
-16.24 0.61 78.18 75.47 7.20 3.46 337.90 43.30 400.37
-16.18 0.68 87.09 84.05 7.26 3.49 338.12 43.52 400.71
-16.11 0.61 77.16 74.44 7.33 3.52 338.11 43.50 400.20
-16.04 0.63 80.09 77.24 7.40 3.56 338.18 43.58 400.27
-15.98 0.66 83.40 80.41 7.46 3.59 338.25 43.65 400.66
-15.91 0.62 78.81 75.96 7.53 3.62 338.12 43.52 400.22
-15.85 0.61 78.05 75.20 7.59 3.65 338.17 43.56 400.21
192
-15.78 0.62 79.32 76.40 7.66 3.68 338.01 43.41 400.04

-15.72 0.65 82.63 79.56 7.72 3.71 338.14 43.54 400.48
-15.65 0.64 80.98 77.95 7.79 3.75 338.61 44.01 400.18
-15.58 0.65 82.63 79.51 7.86 3.78 338.47 43.87 400.69
-15.52 0.63 79.83 76.79 7.92 3.81 338.49 43.88 400.13
-15.46 0.67 84.67 81.42 7.99 3.84 338.57 43.97 400.72
-15.39 0.61 77.54 74.54 8.05 3.87 338.41 43.80 400.04
-15.32 0.66 83.65 80.39 8.12 3.90 338.52 43.92 400.37
-15.26 0.64 81.23 78.04 8.18 3.93 338.22 43.62 400.67
-15.19 0.64 80.85 77.64 8.25 3.97 338.67 44.07 400.53
-15.13 0.63 80.72 77.50 8.31 4.00 338.63 44.03 400.51
-15.06 0.69 87.73 84.19 8.38 4.03 338.82 44.21 400.72
-15.00 0.62 78.43 75.25 8.44 4.06 338.65 44.05 399.96
-14.93 0.63 80.34 77.05 8.51 4.09 338.75 44.15 399.96
-14.87 0.63 80.21 76.91 8.58 4.12 338.78 44.18 400.06
-14.80 0.66 84.16 80.67 8.64 4.15 338.77 44.16 400.58
-14.74 0.66 84.16 80.64 8.71 4.19 338.72 44.11 400.57
-14.67 0.63 80.34 76.95 8.77 4.22 338.27 43.66 400.17
-14.60 0.64 81.87 78.39 8.84 4.25 338.74 44.13 400.01
-14.54 0.69 87.22 83.48 8.90 4.28 338.85 44.24 400.46
-14.47 0.69 87.85 84.07 8.97 4.31 338.88 44.28 400.79
-14.41 0.68 86.71 82.94 9.03 4.34 339.11 44.50 400.46
-14.34 0.67 85.31 81.58 9.10 4.37 339.07 44.46 400.58
-14.28 0.62 78.43 74.98 9.17 4.41 339.01 44.41 399.95
-14.21 0.70 88.49 84.56 9.23 4.44 339.02 44.42 400.47
-14.14 0.61 77.54 74.07 9.30 4.47 338.76 44.15 399.88
-14.08 0.66 84.54 80.74 9.36 4.50 338.94 44.34 400.38
-14.01 0.68 86.96 83.02 9.43 4.53 339.10 44.50 400.56
-13.95 0.65 82.63 78.86 9.49 4.56 339.05 44.45 400.44
-13.88 0.67 85.56 81.63 9.56 4.60 339.04 44.43 400.23
-13.82 0.66 83.52 79.66 9.63 4.63 338.90 44.29 400.15
-13.75 0.74 93.84 89.47 9.69 4.66 339.11 44.51 400.72
-13.68 0.69 88.36 84.22 9.76 4.69 339.21 44.60 400.50
-13.62 0.69 87.34 83.22 9.82 4.72 339.19 44.59 400.41
-13.55 0.67 84.93 80.89 9.89 4.75 338.93 44.32 400.31
-13.49 0.68 85.94 81.83 9.95 4.78 339.02 44.42 400.44
-13.42 0.67 84.80 80.71 10.02 4.82 338.96 44.36 399.97
-13.36 0.70 89.00 84.69 10.08 4.85 339.00 44.40 400.28
-13.29 0.71 90.02 85.63 10.15 4.88 339.25 44.65 400.65
-13.23 0.67 85.56 81.36 10.21 4.91 339.09 44.49 400.31
-13.16 0.72 91.29 86.78 10.28 4.94 339.37 44.76 400.47
-13.10 0.65 82.76 78.64 10.35 4.97 339.12 44.52 400.06
-13.03 0.71 90.65 86.12 10.41 5.01 339.15 44.55 400.16
193
-12.96 0.69 87.47 83.07 10.48 5.04 339.26 44.66 400.12

-12.90 0.72 92.06 87.39 10.54 5.07 339.24 44.64 400.68
-12.83 0.66 84.54 80.23 10.61 5.10 339.23 44.63 400.28
-12.77 0.70 88.49 83.95 10.67 5.13 339.36 44.76 400.60
-12.70 0.71 89.76 85.13 10.74 5.16 339.35 44.74 400.36
-12.64 0.76 97.15 92.10 10.80 5.19 339.34 44.73 400.94
-12.57 0.76 96.38 91.35 10.87 5.23 339.28 44.68 400.63
-12.50 0.70 89.13 84.44 10.93 5.26 339.24 44.63 400.12
-12.44 0.68 87.09 82.48 11.00 5.29 339.24 44.64 400.35
-12.37 0.66 84.29 79.80 11.07 5.32 339.10 44.49 400.22
-12.31 0.71 89.89 85.08 11.13 5.35 339.16 44.56 400.13
-12.24 0.70 89.38 84.57 11.20 5.38 339.09 44.49 400.08
-12.18 0.77 98.17 92.85 11.26 5.41 339.47 44.87 400.54
-12.11 0.74 93.97 88.85 11.33 5.45 339.20 44.59 400.68
-12.04 0.78 99.06 93.63 11.39 5.48 339.22 44.62 400.77
-11.98 0.72 92.06 86.98 11.46 5.51 339.29 44.69 400.44
-11.91 0.67 85.05 80.34 11.53 5.54 339.25 44.64 399.58
-11.85 0.64 81.61 77.07 11.59 5.57 339.21 44.61 399.50
-11.78 0.70 89.51 84.49 11.66 5.60 339.23 44.63 399.99
-11.72 0.65 82.25 77.62 11.72 5.64 339.26 44.66 399.99
-11.65 0.73 92.95 87.68 11.79 5.67 339.23 44.62 400.57
-11.59 0.70 89.51 84.41 11.85 5.70 339.27 44.67 400.06
-11.52 0.74 94.35 88.94 11.92 5.73 339.17 44.57 400.40
-11.46 0.68 86.58 81.59 11.99 5.76 339.08 44.48 399.86
-11.39 0.66 83.78 78.93 12.05 5.79 339.22 44.61 399.74
-11.33 0.71 90.02 84.77 12.12 5.82 339.19 44.59 400.27
-11.26 0.75 94.86 89.30 12.18 5.86 339.34 44.73 400.53
-11.19 0.76 96.13 90.47 12.25 5.89 339.32 44.72 400.84
-11.13 0.73 93.33 87.80 12.31 5.92 339.25 44.64 400.28
-11.06 0.72 91.29 85.86 12.38 5.95 339.03 44.43 400.16
-11.00 0.70 88.62 83.32 12.45 5.98 339.13 44.53 400.00
-10.93 0.74 94.73 89.03 12.51 6.01 339.13 44.53 400.77
-10.86 0.73 92.95 87.33 12.57 6.05 339.39 44.78 400.35
-10.80 0.73 92.56 86.94 12.64 6.08 339.28 44.68 400.64
-10.73 0.66 83.91 78.78 12.71 6.11 338.80 44.20 399.84
-10.67 0.66 84.54 79.35 12.77 6.14 339.33 44.73 399.40
-10.60 0.73 92.31 86.61 12.84 6.17 339.31 44.71 400.10
-10.53 0.72 91.93 86.23 12.90 6.20 339.35 44.75 400.30
-10.47 0.72 91.55 85.84 12.97 6.23 339.21 44.61 400.50
-10.41 0.74 93.84 87.96 13.03 6.27 339.26 44.65 400.47
-10.34 0.71 90.40 84.71 13.10 6.30 339.24 44.64 399.90
-10.28 0.73 92.56 86.71 13.17 6.33 339.50 44.90 400.13
-10.21 0.71 90.65 84.89 13.23 6.36 339.36 44.75 400.18
194
-10.15 0.79 100.59 94.15 13.30 6.39 339.56 44.96 400.99

-10.08 0.66 84.16 78.75 13.36 6.42 338.92 44.31 399.36
-10.01 0.72 92.06 86.11 13.43 6.46 339.24 44.63 400.10
-9.95 0.75 95.37 89.18 13.49 6.49 339.49 44.88 400.14
-9.88 0.70 89.38 83.55 13.56 6.52 339.00 44.40 399.87
-9.82 0.72 91.29 85.31 13.63 6.55 339.32 44.71 399.94
-9.75 0.68 86.96 81.24 13.69 6.58 339.14 44.54 399.55
-9.68 0.71 90.78 84.78 13.75 6.61 339.33 44.73 399.77
-9.62 0.77 98.42 91.88 13.82 6.64 339.29 44.68 400.36
-9.55 0.74 94.35 88.05 13.89 6.68 339.34 44.73 400.01
-9.49 0.71 90.02 83.98 13.96 6.71 338.98 44.38 399.84
-9.42 0.78 99.57 92.86 14.02 6.74 339.42 44.81 400.70
-9.36 0.71 90.40 84.28 14.08 6.77 339.33 44.72 400.08
-9.29 0.71 90.91 84.73 14.15 6.80 339.02 44.42 399.83
-9.23 0.74 94.35 87.90 14.22 6.83 339.21 44.61 399.91
-9.16 0.78 99.06 92.26 14.28 6.87 339.12 44.52 400.16
-9.09 0.73 92.69 86.30 14.34 6.90 339.07 44.47 400.00
-9.03 0.67 85.82 79.87 14.41 6.93 338.87 44.27 399.75
-8.96 0.69 87.22 81.15 14.48 6.96 339.13 44.53 399.76
-8.90 0.74 93.84 87.28 14.54 6.99 339.29 44.68 399.70
-8.83 0.68 86.83 80.74 14.61 7.02 339.14 44.53 399.39
-8.77 0.80 102.24 95.03 14.67 7.05 339.60 44.99 400.54
-8.70 0.76 96.51 89.67 14.74 7.09 339.52 44.91 400.20
-8.64 0.79 100.97 93.78 14.80 7.12 339.44 44.83 400.43
-8.57 0.72 91.16 84.65 14.87 7.15 339.33 44.72 399.69
-8.51 0.80 101.35 94.07 14.93 7.18 339.40 44.79 400.27
-8.44 0.71 90.15 83.64 15.00 7.21 339.09 44.49 399.70
-8.37 0.74 93.97 87.16 15.07 7.24 339.35 44.74 399.94
-8.31 0.74 94.60 87.72 15.13 7.27 339.21 44.61 400.31
-8.24 0.74 94.60 87.69 15.20 7.31 339.31 44.70 400.23
-8.18 0.81 103.26 95.68 15.26 7.34 339.28 44.68 400.64
-8.11 0.71 89.76 83.15 15.33 7.37 339.13 44.53 399.73
-8.05 0.69 87.47 81.00 15.40 7.40 339.08 44.47 399.37
-7.98 0.77 97.66 90.40 15.46 7.43 339.26 44.66 399.72
-7.91 0.76 96.26 89.07 15.53 7.46 339.32 44.72 399.84
-7.85 0.73 93.46 86.45 15.59 7.50 339.18 44.58 400.37
-7.78 0.75 95.37 88.19 15.66 7.53 339.22 44.61 399.92
-7.72 0.76 96.26 88.98 15.72 7.56 339.27 44.66 399.88
-7.65 0.80 101.35 93.66 15.79 7.59 339.32 44.72 400.07
-7.58 0.75 95.62 88.33 15.85 7.62 339.12 44.52 399.67
-7.52 0.72 91.04 84.07 15.92 7.65 339.24 44.63 399.73
-7.46 0.76 96.64 89.21 15.99 7.69 339.28 44.68 399.96
-7.39 0.73 93.20 86.01 16.05 7.72 339.24 44.63 399.41
195
-7.32 0.74 93.71 86.45 16.11 7.75 339.07 44.47 399.57

-7.26 0.72 91.67 84.54 16.18 7.78 339.16 44.56 399.56
-7.19 0.78 99.06 91.32 16.25 7.81 339.33 44.72 400.15
-7.13 0.75 94.86 87.42 16.31 7.84 339.28 44.68 400.02
-7.06 0.75 95.11 87.62 16.38 7.87 339.20 44.60 399.99
-7.00 0.80 102.24 94.16 16.44 7.90 339.23 44.63 400.17
-6.93 0.77 97.53 89.79 16.51 7.94 339.17 44.56 399.87
-6.86 0.82 104.66 96.32 16.57 7.97 339.36 44.75 400.31
-6.80 0.74 93.71 86.21 16.64 8.00 338.93 44.32 399.87
-6.73 0.75 94.86 87.24 16.71 8.03 338.88 44.28 400.02
-6.67 0.76 96.51 88.73 16.77 8.06 339.05 44.45 399.84
-6.60 0.78 98.80 90.81 16.84 8.09 339.02 44.42 399.77
-6.54 0.70 88.49 81.30 16.90 8.13 338.84 44.24 399.14
-6.47 0.77 97.91 89.92 16.97 8.16 339.00 44.39 400.32
-6.41 0.70 89.64 82.29 17.04 8.19 339.00 44.40 399.42
-6.34 0.81 103.13 94.65 17.10 8.22 338.97 44.37 400.50
-6.27 0.77 97.78 89.72 17.16 8.25 339.06 44.46 400.04
-6.21 0.77 97.78 89.68 17.23 8.28 339.09 44.49 399.91
-6.14 0.76 96.77 88.72 17.30 8.31 339.10 44.49 399.55
-6.08 0.72 91.29 83.67 17.36 8.35 338.95 44.35 399.37
-6.01 0.78 99.31 90.99 17.43 8.38 338.96 44.35 399.93
-5.95 0.78 99.06 90.73 17.50 8.41 339.16 44.55 399.92
-5.88 0.76 96.26 88.13 17.56 8.44 339.05 44.44 399.64
-5.82 0.74 94.47 86.47 17.63 8.47 338.95 44.34 399.61
-5.75 0.77 98.04 89.70 17.69 8.51 338.97 44.37 399.63
-5.68 0.78 99.19 90.72 17.76 8.54 338.68 44.07 399.98
-5.62 0.71 90.53 82.77 17.82 8.57 338.78 44.17 399.36
-5.55 0.78 99.57 91.01 17.89 8.60 338.92 44.32 400.30
-5.49 0.77 98.55 90.04 17.95 8.63 338.97 44.36 399.75
-5.42 0.77 97.91 89.43 18.02 8.66 338.97 44.36 399.34
-5.35 0.79 99.95 91.26 18.08 8.69 338.94 44.34 399.85
-5.29 0.74 93.84 85.65 18.15 8.73 338.96 44.35 399.59
-5.22 0.73 92.82 84.69 18.22 8.76 339.03 44.43 399.27
-5.16 0.82 104.02 94.88 18.28 8.79 339.03 44.42 400.36
-5.09 0.80 102.11 93.11 18.35 8.82 338.99 44.38 399.99
-5.03 0.77 98.29 89.59 18.41 8.85 338.86 44.26 399.83
-4.96 0.76 96.51 87.94 18.48 8.88 338.85 44.24 399.59
-4.90 0.78 99.19 90.34 18.54 8.91 338.75 44.15 399.81
-4.83 0.73 92.56 84.28 18.61 8.95 338.72 44.11 399.46
-4.77 0.78 99.19 90.28 18.67 8.98 338.99 44.39 399.68
-4.70 0.80 101.35 92.22 18.74 9.01 338.94 44.34 399.65
-4.64 0.78 98.68 89.75 18.81 9.04 338.87 44.27 400.03
-4.57 0.75 95.37 86.71 18.87 9.07 338.73 44.13 399.60
196
-4.50 0.71 90.53 82.29 18.94 9.10 338.69 44.08 399.48

-4.44 0.77 98.04 89.08 19.00 9.13 338.73 44.12 399.68
-4.37 0.76 96.26 87.43 19.07 9.17 338.84 44.24 399.49
-4.31 0.77 97.53 88.56 19.13 9.20 338.88 44.27 399.72
-4.24 0.78 99.31 90.15 19.20 9.23 338.88 44.27 399.69
-4.18 0.80 101.22 91.85 19.27 9.26 338.74 44.14 399.58
-4.11 0.79 100.20 90.89 19.33 9.29 338.76 44.15 399.78
-4.05 0.86 109.88 99.63 19.40 9.33 338.92 44.31 400.41
-3.98 0.82 104.91 95.10 19.46 9.36 339.01 44.40 400.01
-3.91 0.78 99.19 89.87 19.53 9.39 338.78 44.17 399.67
-3.85 0.76 96.89 87.77 19.59 9.42 338.78 44.17 399.20
-3.78 0.80 101.60 92.00 19.66 9.45 338.96 44.36 399.62
-3.72 0.76 96.26 87.13 19.72 9.48 338.80 44.20 399.80
-3.65 0.77 98.55 89.17 19.79 9.51 338.80 44.19 399.89
-3.59 0.87 110.77 100.20 19.85 9.55 339.07 44.47 400.22
-3.52 0.82 104.02 94.06 19.92 9.58 338.83 44.22 399.74
-3.45 0.81 102.75 92.88 19.99 9.61 338.89 44.28 399.77
-3.39 0.74 94.22 85.14 20.05 9.64 338.60 44.00 399.38
-3.32 0.79 100.46 90.74 20.12 9.67 338.78 44.18 400.03
-3.26 0.83 105.42 95.20 20.18 9.70 338.67 44.06 400.36
-3.19 0.74 93.84 84.70 20.25 9.73 338.66 44.06 399.22
-3.13 0.83 105.30 95.01 20.31 9.77 338.77 44.17 399.75
-3.06 0.77 97.53 87.97 20.38 9.80 338.59 43.98 399.63
-2.99 0.82 103.90 93.68 20.45 9.83 338.82 44.22 399.93
-2.93 0.78 99.57 89.75 20.51 9.86 338.83 44.23 399.80
-2.87 0.76 97.28 87.65 20.57 9.89 338.60 43.99 399.79
-2.80 0.75 95.75 86.25 20.64 9.92 338.72 44.12 399.61
-2.73 0.76 97.28 87.59 20.71 9.96 338.79 44.18 399.46
-2.67 0.79 100.33 90.31 20.77 9.99 338.87 44.27 399.71
-2.60 0.81 102.75 92.46 20.84 10.02 338.94 44.34 400.10
-2.54 0.74 94.22 84.75 20.90 10.05 338.38 43.78 399.52
-2.47 0.82 103.77 93.31 20.97 10.08 339.04 44.43 399.84
-2.41 0.86 109.50 98.43 21.03 10.11 338.98 44.37 400.19
-2.34 0.74 94.35 84.78 21.10 10.14 338.80 44.20 399.28
-2.28 0.78 98.68 88.63 21.17 10.18 338.72 44.11 399.88
-2.21 0.79 100.33 90.09 21.23 10.21 338.55 43.94 399.89
-2.14 0.87 110.64 99.32 21.30 10.24 338.89 44.29 400.79
-2.08 0.81 103.01 92.43 21.36 10.27 338.76 44.16 399.81
-2.01 0.79 100.97 90.57 21.43 10.30 338.69 44.09 399.76
-1.95 0.83 106.06 95.10 21.49 10.33 338.75 44.14 400.02
-1.88 0.82 104.53 93.70 21.56 10.36 338.80 44.19 399.95
-1.82 0.76 96.89 86.82 21.63 10.40 338.77 44.17 399.66
-1.75 0.81 103.39 92.61 21.69 10.43 338.68 44.08 400.19
197
-1.68 0.76 96.89 86.76 21.76 10.46 338.71 44.10 399.27

-1.62 0.73 93.20 83.42 21.82 10.49 338.57 43.96 399.09
-1.55 0.84 107.33 96.04 21.89 10.52 338.62 44.02 400.25
-1.49 0.77 97.53 87.24 21.95 10.55 338.56 43.95 399.73
-1.42 0.73 93.07 83.22 22.02 10.59 338.58 43.97 399.21
-1.36 0.79 100.84 90.13 22.09 10.62 338.53 43.93 399.72
-1.29 0.75 94.98 84.87 22.15 10.65 338.47 43.87 399.20
-1.22 0.80 101.35 90.53 22.21 10.68 338.38 43.78 399.44
-1.16 0.79 100.46 89.70 22.28 10.71 338.69 44.08 399.84
-1.09 0.76 96.13 85.80 22.35 10.74 338.54 43.93 399.66
-1.03 0.79 99.95 89.18 22.41 10.77 338.62 44.02 399.63
-0.96 0.87 110.39 98.46 22.48 10.81 338.76 44.15 399.97
-0.90 0.87 110.14 98.20 22.54 10.84 338.85 44.24 399.97
-0.83 0.83 105.30 93.85 22.61 10.87 338.79 44.19 399.52
-0.76 0.87 110.77 98.70 22.67 10.90 338.87 44.27 400.04
-0.70 0.80 101.22 90.16 22.74 10.93 338.69 44.08 399.66
-0.63 0.86 108.86 96.93 22.80 10.96 338.72 44.11 400.16
-0.57 0.79 100.20 89.19 22.87 10.99 338.52 43.91 399.70
-0.51 0.84 107.21 95.38 22.94 11.03 338.65 44.05 399.56
-0.44 0.73 93.33 83.01 23.00 11.06 338.35 43.75 399.14
-0.37 0.80 102.37 91.01 23.07 11.09 338.60 44.00 399.74
-0.31 0.82 104.02 92.45 23.13 11.12 338.52 43.91 399.98
-0.24 0.81 103.01 91.52 23.20 11.15 338.68 44.07 399.70
-0.18 0.77 97.78 86.85 23.26 11.18 338.40 43.80 399.22
-0.11 0.77 98.17 87.16 23.33 11.22 338.46 43.85 399.42
-0.04 0.82 103.90 92.21 23.40 11.25 338.81 44.20 399.44
0.02 0.84 106.32 94.32 23.46 11.28 338.51 43.91 399.68
0.09 0.76 96.64 85.71 23.53 11.31 338.14 43.54 399.10
0.15 0.75 95.37 84.55 23.59 11.34 338.30 43.70 399.12
0.22 0.87 111.28 98.62 23.66 11.37 338.49 43.89 400.01
0.28 0.78 99.57 88.21 23.73 11.41 337.94 43.34 399.42
0.35 0.84 106.44 94.27 23.79 11.44 338.18 43.58 399.70
0.42 0.84 106.95 94.69 23.86 11.47 338.68 44.07 399.83
0.48 0.81 103.64 91.72 23.92 11.50 338.60 44.00 399.34
0.55 0.75 95.62 84.59 23.99 11.53 338.31 43.71 399.13
0.61 0.79 100.46 88.84 24.05 11.56 338.29 43.69 399.43
0.68 0.80 101.22 89.49 24.12 11.60 338.40 43.80 399.11
0.74 0.77 97.78 86.42 24.18 11.63 338.14 43.53 398.92
0.81 0.83 105.81 93.47 24.25 11.66 338.42 43.81 399.49
0.87 0.81 102.75 90.74 24.32 11.69 338.34 43.73 399.54
0.94 0.87 110.26 97.34 24.38 11.72 338.55 43.94 400.00
1.01 0.77 98.29 86.74 24.45 11.75 338.05 43.45 399.23
1.07 0.83 106.19 93.67 24.51 11.78 338.30 43.69 399.43
198
1.14 0.82 104.15 91.84 24.58 11.82 338.46 43.86 399.09

1.21 0.86 109.50 96.53 24.64 11.85 338.49 43.89 399.82
1.27 0.75 94.98 83.70 24.71 11.88 338.26 43.66 399.12
1.34 0.80 101.73 89.62 24.77 11.91 338.41 43.80 399.19
1.40 0.78 99.82 87.90 24.84 11.94 338.28 43.67 398.99
1.47 0.78 99.82 87.87 24.90 11.97 338.09 43.48 399.10
1.53 0.83 105.93 93.22 24.97 12.00 338.53 43.93 399.58
1.59 0.78 98.80 86.91 25.03 12.03 338.22 43.61 399.22
1.66 0.81 102.75 90.35 25.10 12.07 338.58 43.97 399.47
1.73 0.79 100.84 88.64 25.17 12.10 338.19 43.58 399.21
1.79 0.78 99.06 87.04 25.23 12.13 338.23 43.62 399.34
1.86 0.85 107.84 94.73 25.30 12.16 338.25 43.64 399.52
1.93 0.88 111.66 98.05 25.36 12.19 338.42 43.82 399.72
1.99 0.82 104.66 91.87 25.43 12.23 338.17 43.57 399.53
2.05 0.83 105.30 92.39 25.50 12.26 338.40 43.80 399.64
2.12 0.75 95.75 83.98 25.56 12.29 338.11 43.51 399.25
2.19 0.83 105.42 92.44 25.62 12.32 338.47 43.86 399.23
2.25 0.81 103.01 90.28 25.69 12.35 338.05 43.45 399.28
2.32 0.78 98.80 86.57 25.76 12.38 337.96 43.35 399.23
2.39 0.85 108.61 95.12 25.82 12.41 338.08 43.48 399.94
2.45 0.81 103.26 90.41 25.89 12.45 338.17 43.57 399.41
2.52 0.80 101.60 88.93 25.96 12.48 338.40 43.80 399.44
2.58 0.87 110.52 96.69 26.02 12.51 338.29 43.68 399.99
2.65 0.79 100.08 87.53 26.09 12.54 338.09 43.48 399.26
2.71 0.84 107.46 93.95 26.15 12.57 338.46 43.85 399.45
2.78 0.79 100.59 87.91 26.22 12.60 338.25 43.64 399.05
2.84 0.82 104.91 91.66 26.28 12.64 338.33 43.73 399.02
2.91 0.81 103.26 90.18 26.35 12.67 338.15 43.55 399.32
2.97 0.80 101.86 88.93 26.41 12.70 338.17 43.57 399.29
3.04 0.78 98.93 86.34 26.48 12.73 338.31 43.70 399.09
3.11 0.83 105.04 91.64 26.55 12.76 338.13 43.52 399.39
3.17 0.81 102.75 89.60 26.61 12.79 338.23 43.62 399.29
3.24 0.82 103.77 90.46 26.67 12.82 338.15 43.55 399.35
3.30 0.82 104.91 91.43 26.74 12.86 338.44 43.84 399.67
3.37 0.79 100.20 87.29 26.81 12.89 338.23 43.62 399.32
3.43 0.79 100.46 87.48 26.87 12.92 338.23 43.62 398.91
3.50 0.85 108.10 94.10 26.94 12.95 338.34 43.74 399.34
3.56 0.83 106.19 92.40 27.00 12.98 338.36 43.75 399.54
3.63 0.84 107.08 93.14 27.07 13.01 338.39 43.79 399.61
3.70 0.82 104.66 91.01 27.13 13.04 338.09 43.49 399.61
3.76 0.79 100.97 87.76 27.20 13.08 338.15 43.55 399.07
3.83 0.84 106.82 92.82 27.27 13.11 338.36 43.76 399.33
3.89 0.81 102.88 89.36 27.33 13.14 338.14 43.54 399.15
199
3.96 0.82 104.91 91.10 27.40 13.17 338.32 43.71 399.42

4.02 0.79 100.20 86.97 27.46 13.20 338.22 43.62 399.11
4.09 0.80 102.11 88.60 27.53 13.23 338.31 43.70 399.22
4.15 0.82 104.41 90.56 27.59 13.27 338.33 43.73 399.30
4.22 0.78 99.19 86.00 27.66 13.30 338.10 43.50 399.09
4.29 0.79 100.84 87.40 27.73 13.33 338.22 43.62 399.07
4.35 0.85 107.72 93.32 27.79 13.36 338.50 43.90 399.42
4.42 0.80 102.37 88.66 27.86 13.39 338.21 43.60 399.34
4.48 0.91 116.25 100.64 27.92 13.42 338.36 43.75 399.90
4.55 0.84 106.57 92.23 27.99 13.45 338.17 43.56 399.72
4.61 0.85 108.48 93.85 28.05 13.49 338.51 43.90 399.70
4.68 0.80 102.11 88.31 28.12 13.52 338.26 43.66 398.88
4.74 0.85 108.73 94.00 28.19 13.55 338.42 43.82 399.29
4.81 0.85 107.72 93.09 28.25 13.58 338.18 43.58 399.20
4.88 0.83 106.06 91.62 28.32 13.61 338.17 43.57 399.51
4.94 0.84 107.33 92.69 28.38 13.64 338.37 43.77 399.33
5.01 0.88 111.66 96.39 28.45 13.68 338.41 43.81 399.61
5.07 0.85 107.97 93.17 28.51 13.71 338.39 43.79 399.51
5.14 0.77 98.55 85.01 28.58 13.74 338.12 43.51 398.88
5.20 0.73 93.07 80.26 28.64 13.77 337.94 43.34 398.41
5.27 0.78 98.93 85.28 28.71 13.80 338.22 43.61 399.10
5.34 0.80 101.35 87.33 28.77 13.83 338.05 43.45 399.46
5.40 0.82 104.91 90.37 28.84 13.86 338.19 43.59 399.39
5.47 0.80 101.73 87.60 28.90 13.90 338.24 43.63 399.15
5.53 0.88 111.92 96.33 28.97 13.93 338.31 43.70 399.66
5.60 0.82 104.66 90.05 29.04 13.96 338.11 43.50 399.13
5.66 0.92 117.65 101.19 29.10 13.99 338.32 43.72 399.81
5.73 0.80 101.73 87.46 29.17 14.02 338.07 43.47 399.09
5.80 0.83 105.81 90.94 29.23 14.05 337.99 43.39 399.59
5.86 0.84 106.32 91.34 29.30 14.09 338.42 43.82 399.28
5.93 0.83 106.19 91.20 29.36 14.12 338.20 43.60 399.20
5.99 0.84 107.46 92.26 29.43 14.15 338.18 43.58 399.53
6.05 0.82 103.90 89.16 29.49 14.18 338.06 43.46 399.10
6.12 0.85 108.61 93.17 29.56 14.21 338.12 43.52 399.53
6.19 0.82 104.41 89.54 29.63 14.24 337.95 43.34 399.32
6.25 0.81 103.13 88.41 29.69 14.28 338.15 43.55 399.20
6.32 0.83 105.55 90.45 29.76 14.31 338.13 43.52 399.27
6.38 0.86 109.75 94.02 29.82 14.34 338.27 43.67 399.62
6.45 0.84 106.57 91.25 29.89 14.37 338.05 43.45 399.54
6.52 0.77 97.53 83.48 29.96 14.40 337.75 43.14 398.76
6.58 0.86 109.63 93.80 30.02 14.43 338.08 43.47 399.60
6.65 0.81 102.75 87.89 30.09 14.46 337.95 43.35 398.94
6.71 0.83 105.55 90.25 30.15 14.50 338.15 43.55 399.12
200
6.78 0.86 109.12 93.26 30.22 14.53 338.16 43.55 399.63

6.85 0.84 106.95 91.38 30.28 14.56 338.09 43.48 399.61
6.91 0.87 110.14 94.07 30.35 14.59 338.32 43.72 399.59
6.98 0.85 108.35 92.51 30.42 14.62 338.12 43.52 399.60
7.04 0.88 112.55 96.06 30.48 14.65 338.46 43.86 399.66
7.10 0.83 105.30 89.83 30.55 14.69 337.99 43.39 399.36
7.17 0.87 110.77 94.47 30.61 14.72 338.26 43.65 399.43
7.24 0.87 111.28 94.87 30.68 14.75 338.22 43.62 399.57
7.30 0.81 103.51 88.21 30.74 14.78 337.73 43.13 399.59
7.37 0.81 102.88 87.64 30.81 14.81 337.91 43.31 399.35
7.43 0.83 105.17 89.56 30.87 14.84 338.14 43.54 399.27
7.50 0.85 107.72 91.69 30.94 14.87 338.26 43.66 399.25
7.56 0.78 99.06 84.29 31.00 14.91 338.10 43.50 398.93
7.63 0.85 108.23 92.06 31.07 14.94 338.38 43.77 399.56
7.70 0.84 107.33 91.27 31.13 14.97 338.00 43.40 399.72
7.76 0.88 112.55 95.67 31.20 15.00 338.48 43.88 399.72
7.83 0.86 109.75 93.26 31.26 15.03 338.23 43.63 399.40
7.89 0.79 100.46 85.33 31.33 15.06 338.04 43.44 398.95
7.96 0.82 104.28 88.54 31.40 15.09 338.16 43.55 399.18
8.02 0.81 103.64 87.97 31.46 15.13 338.05 43.44 399.32
8.09 0.78 99.19 84.15 31.53 15.16 338.12 43.52 399.24
8.15 0.76 97.28 82.50 31.60 15.19 337.96 43.35 398.91
8.22 0.83 105.17 89.16 31.66 15.22 338.26 43.65 399.27
8.29 0.86 109.75 93.01 31.73 15.25 338.03 43.42 399.67
8.35 0.83 106.19 89.96 31.79 15.28 338.09 43.49 399.26
8.42 0.84 107.21 90.79 31.86 15.32 338.00 43.39 399.57
8.48 0.80 102.24 86.55 31.92 15.35 337.91 43.30 399.30
8.55 0.84 107.46 90.93 31.99 15.38 338.17 43.56 399.26
8.62 0.90 114.34 96.72 32.06 15.41 338.17 43.57 399.56
8.68 0.81 103.13 87.21 32.12 15.44 337.88 43.28 399.26
8.75 0.90 114.21 96.54 32.19 15.47 338.18 43.57 399.63
8.81 0.78 98.80 83.48 32.25 15.51 337.93 43.32 398.85
8.88 0.82 103.77 87.65 32.32 15.54 337.93 43.33 399.42
8.94 0.86 109.75 92.67 32.38 15.57 338.21 43.60 399.42
9.01 0.85 107.59 90.81 32.45 15.60 338.00 43.39 399.18
9.07 0.81 103.26 87.12 32.51 15.63 338.06 43.45 399.06
9.14 0.83 105.55 89.02 32.58 15.66 338.09 43.48 399.04
9.20 0.79 100.71 84.91 32.65 15.70 337.66 43.06 399.21
9.27 0.88 111.79 94.21 32.71 15.73 338.29 43.68 399.53
9.34 0.90 114.08 96.11 32.78 15.76 338.04 43.44 399.79
9.40 0.83 105.17 88.56 32.84 15.79 338.04 43.43 399.08
9.47 0.88 111.41 93.78 32.91 15.82 338.20 43.60 399.28
9.53 0.87 111.15 93.54 32.97 15.85 338.36 43.76 399.56
201
9.60 0.84 106.82 89.86 33.04 15.88 338.09 43.48 399.54

9.66 0.85 108.35 91.11 33.10 15.91 338.13 43.53 399.48
9.73 0.89 113.70 95.57 33.17 15.95 337.96 43.36 399.52
9.79 0.91 115.61 97.14 33.23 15.98 338.17 43.57 399.74
9.86 0.89 112.81 94.75 33.30 16.01 338.23 43.62 399.43
9.93 0.84 107.33 90.12 33.37 16.04 338.08 43.48 399.22
9.99 0.85 108.48 91.04 33.43 16.07 338.03 43.42 399.76
10.06 0.83 106.19 89.09 33.50 16.10 338.10 43.49 399.44
10.12 0.89 112.81 94.61 33.56 16.14 338.11 43.50 399.61
10.19 0.87 110.39 92.54 33.63 16.17 338.23 43.63 399.30
10.26 0.87 110.26 92.40 33.69 16.20 338.07 43.46 399.06
10.32 0.81 102.75 86.07 33.76 16.23 338.22 43.62 398.90
10.38 0.85 108.10 90.52 33.82 16.26 338.09 43.49 399.65
10.45 0.84 106.57 89.21 33.89 16.29 338.08 43.47 399.41
10.52 0.86 110.01 92.05 33.96 16.33 337.88 43.28 399.44
10.58 0.82 104.79 87.65 34.02 16.36 337.84 43.23 399.25
10.65 0.91 115.86 96.88 34.09 16.39 337.98 43.38 399.74
10.71 0.86 109.37 91.41 34.15 16.42 337.96 43.35 399.74
10.78 0.84 106.32 88.83 34.22 16.45 338.05 43.44 399.43
10.84 0.90 114.46 95.60 34.28 16.48 338.08 43.48 399.69
10.91 0.75 96.00 80.15 34.35 16.51 337.83 43.22 398.39
10.97 0.78 98.93 82.56 34.41 16.54 337.88 43.28 399.10
11.04 0.87 110.26 91.99 34.48 16.58 338.15 43.55 399.72
11.10 0.85 107.59 89.72 34.54 16.61 338.02 43.41 399.64
11.17 0.85 108.10 90.11 34.61 16.64 338.21 43.61 399.79
11.24 0.88 111.66 93.05 34.68 16.67 338.08 43.48 399.80
11.30 0.78 98.80 82.30 34.74 16.70 337.86 43.25 399.13
11.37 0.81 102.50 85.34 34.81 16.74 337.92 43.32 399.03
11.44 0.84 106.57 88.70 34.87 16.77 338.06 43.46 399.27
11.50 0.83 105.68 87.93 34.94 16.80 338.17 43.56 399.24
11.56 0.82 104.02 86.52 35.01 16.83 338.17 43.57 399.36
11.63 0.83 105.81 87.97 35.07 16.86 338.18 43.57 399.10
11.70 0.87 110.14 91.53 35.14 16.89 338.13 43.53 399.77
11.76 0.85 108.23 89.91 35.20 16.92 337.88 43.28 399.53
11.83 0.88 112.43 93.37 35.27 16.95 338.08 43.47 399.81
11.89 0.82 104.53 86.78 35.33 16.99 338.40 43.80 399.48
11.96 0.90 114.21 94.77 35.40 17.02 338.44 43.84 399.74
12.03 0.81 103.01 85.44 35.47 17.05 338.01 43.41 399.10
12.09 0.84 106.70 88.47 35.53 17.08 337.98 43.38 399.57
12.15 0.86 109.63 90.87 35.60 17.11 338.31 43.70 399.71
12.22 0.82 104.91 86.93 35.66 17.14 338.25 43.65 399.81
12.29 0.87 110.14 91.22 35.73 17.18 338.19 43.59 399.91
12.35 0.88 112.55 93.19 35.79 17.21 338.41 43.81 399.55
202
12.42 0.83 105.30 87.15 35.86 17.24 338.09 43.49 399.37

12.48 0.80 101.73 84.16 35.93 17.27 338.07 43.47 399.21
12.55 0.82 104.91 86.76 35.99 17.30 337.85 43.25 399.62
12.62 0.92 117.27 96.94 36.06 17.33 338.58 43.97 399.97
12.68 0.88 111.79 92.38 36.12 17.36 338.42 43.82 399.78
12.74 0.82 104.91 86.66 36.18 17.40 338.31 43.71 399.48
12.81 0.86 108.99 90.00 36.25 17.43 338.26 43.65 400.04
12.88 0.78 99.44 82.08 36.32 17.46 338.11 43.50 399.19
12.94 0.88 112.55 92.87 36.38 17.49 338.34 43.74 400.36
13.01 0.85 107.59 88.74 36.45 17.52 338.44 43.83 399.36
13.07 0.84 106.32 87.65 36.51 17.55 338.38 43.78 399.68
13.14 0.82 103.77 85.52 36.58 17.59 338.27 43.67 398.92
13.21 0.82 104.15 85.80 36.64 17.62 338.10 43.49 399.32
13.27 0.80 101.99 83.99 36.71 17.65 338.23 43.63 399.11
13.34 0.85 108.35 89.20 36.78 17.68 338.07 43.46 399.72
13.40 0.85 108.10 88.95 36.84 17.71 338.34 43.73 399.86
13.47 0.80 101.86 83.79 36.91 17.74 338.09 43.48 399.22
13.53 0.86 109.63 90.14 36.97 17.77 337.92 43.31 399.88
13.60 0.77 97.78 80.37 37.04 17.81 338.23 43.63 398.95
13.66 0.81 103.26 84.84 37.10 17.84 337.87 43.27 399.60
13.73 0.78 98.80 81.15 37.17 17.87 337.96 43.36 399.27
13.79 0.85 108.48 89.06 37.23 17.90 338.44 43.83 399.79
13.86 0.90 114.08 93.63 37.30 17.93 338.43 43.82 400.05
13.92 0.83 105.30 86.38 37.37 17.96 338.34 43.73 399.09
13.99 0.83 105.17 86.24 37.43 18.00 338.45 43.84 399.01
14.06 0.83 106.19 87.05 37.50 18.03 338.15 43.55 399.60
14.12 0.84 107.21 87.85 37.56 18.06 338.40 43.79 399.47
14.19 0.92 116.76 95.64 37.63 18.09 338.63 44.02 399.79
14.25 0.82 103.90 85.07 37.69 18.12 338.29 43.69 399.69
14.32 0.82 104.15 85.24 37.76 18.15 338.24 43.63 398.98
14.39 0.80 101.22 82.82 37.82 18.18 338.36 43.75 398.93
14.45 0.88 111.79 91.43 37.89 18.22 338.43 43.83 399.59
14.52 0.81 103.39 84.52 37.95 18.25 338.38 43.78 399.24
14.58 0.89 113.45 92.71 38.02 18.28 338.52 43.92 400.58
14.64 0.85 107.84 88.10 38.08 18.31 338.41 43.80 400.00
14.71 0.78 99.44 81.20 38.15 18.34 338.38 43.78 399.37
14.78 0.88 111.66 91.15 38.22 18.38 338.42 43.81 399.87
14.85 0.81 102.88 83.94 38.28 18.41 337.94 43.33 399.45
14.91 0.85 107.97 88.07 38.35 18.44 338.33 43.72 399.65
14.97 0.79 100.08 81.59 38.41 18.47 338.54 43.94 399.35
15.04 0.91 116.37 94.84 38.48 18.50 338.45 43.84 400.52
203
100
90
80
70
60
50
(a) Stone column test results
40
30 Test 5-a (clay/column_0.03mm/min)
Test 5-b(clay/column_0.03mm/min)
20
Test 5-c (clay/column_0.03mm/min)
10 Test 5-d (clay/column_0.03mm/min)
0
0 2 4 6 8 10 12 14 16 18
Axial strain, (%)
70.00
60.00
50.00
40.00
(b) Soil only test results
30.00
Test 1_a (soil only_0.03mm/min)
20.00 Test 1_b (Soil only_0.03mm/min)
Test 1_c (Soil only_0.03mm/min)
10.00
Test 1_d (Soil only_0.03mm/min)
0.00
0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00
204
100
80
60
40
Reinforced soil (current study)
Soil only (current study)

20
0
0 5 10 15
205
 Example of the data of the cyclic loading test on reinforce soil
Load Load Cyclic Strain

Stress PK Stress Tr dP ( PK-Trgh) Posn Posn Dis. Pk Dis. Tr Strain ΔU U ΔU
Cycles Pk Tr. def. stiffness Trgh U Pk
(kPa) (kPa) (kN) Pk Trgh (mm) (mm) PK % pk Trgh Trgh
(kN) (kN) (mm) %
1 0.35 0.35 44.69 44.69 0.00 -22.47 -22.47 0.86 0.86 0.241 0.40 0.40 328.54 20.54 328.54 20.54
10 0.47 0.15 59.87 19.52 0.32 -21.88 -22.13 0.59 0.35 0.23 1315.15 0.28 0.16 332.53 24.53 334.73 26.73
20 0.48 0.15 61.45 19.07 0.33 -21.65 -21.88 0.82 0.59 0.25 1446.96 0.38 0.28 336.18 28.18 338.46 30.46
30 0.47 0.15 60.36 18.89 0.33 -21.53 -21.78 0.95 0.70 0.256 1302.60 0.44 0.33 338.41 30.41 340.93 32.93
40 0.48 0.15 61.20 18.89 0.33 -21.43 -21.68 1.05 0.79 0.238 1298.05 0.49 0.37 340.06 32.06 342.42 34.42
50 0.48 0.15 61.57 18.98 0.33 -21.38 -21.62 1.09 0.86 0.229 1405.25 0.51 0.40 341.08 33.08 343.58 35.58
60 0.53 0.18 66.89 23.26 0.34 -21.15 -21.38 1.33 1.10 0.246 1496.29 0.62 0.51 343.42 35.42 345.82 37.82
70 0.53 0.17 67.13 22.18 0.35 -21.00 -21.24 1.48 1.23 0.25 1435.16 0.69 0.57 344.78 36.78 347.14 39.14
80 0.53 0.17 67.74 22.09 0.36 -20.87 -21.12 1.60 1.35 0.229 1433.88 0.75 0.63 345.73 37.73 348.31 40.31
90 0.53 0.18 67.13 22.35 0.35 -20.80 -21.03 1.67 1.44 0.242 1535.85 0.78 0.67 346.34 38.34 348.94 40.94
100 0.53 0.17 67.25 21.75 0.36 -20.73 -20.97 1.75 1.50 0.235 1476.65 0.82 0.70 347.31 39.31 349.69 41.69
110 0.55 0.19 69.91 24.48 0.36 -20.60 -20.84 1.87 1.64 0.254 1518.34 0.87 0.76 348.30 40.30 350.55 42.55
120 0.55 0.19 69.55 24.31 0.36 -20.50 -20.75 1.98 1.72 0.255 1398.82 0.92 0.80 348.93 40.93 350.97 42.97
130 0.55 0.20 69.67 24.91 0.35 -20.42 -20.68 2.05 1.80 0.236 1378.67 0.96 0.84 349.32 41.32 351.63 43.63
140 0.55 0.19 70.03 24.40 0.36 -20.38 -20.62 2.09 1.86 0.258 1518.77 0.98 0.87 349.88 41.88 352.06 44.06
150 0.55 0.19 69.91 24.31 0.36 -20.30 -20.56 2.17 1.91 0.256 1388.18 1.01 0.89 350.13 42.13 352.63 44.63
160 0.56 0.20 71.73 25.51 0.36 -20.22 -20.47 2.26 2.00 0.25 1418.05 1.05 0.93 350.92 42.92 353.25 45.25
170 0.57 0.20 72.45 25.25 0.37 -20.14 -20.39 2.34 2.09 0.234 1482.92 1.09 0.98 351.23 43.23 353.55 45.55
180 0.57 0.20 72.21 25.08 0.37 -20.07 -20.31 2.40 2.17 0.257 1581.92 1.12 1.01 351.51 43.51 353.75 45.75
190 0.57 0.20 72.09 24.91 0.37 -19.99 -20.25 2.48 2.23 0.238 1441.87 1.16 1.04 351.89 43.89 354.19 46.19
206
200 0.56 0.19 71.73 24.82 0.37 -19.95 -20.18 2.53 2.29 0.237 1547.82 1.18 1.07 352.07 44.07 354.47 46.47
210 0.57 0.20 73.06 25.59 0.37 -19.87 -20.11 2.60 2.36 0.235 1573.00 1.21 1.10 352.31 44.31 355.00 47.00
220 0.57 0.20 73.18 25.85 0.37 -19.83 -20.07 2.64 2.40 0.258 1581.87 1.23 1.12 352.34 44.34 354.85 46.85
230 0.58 0.20 73.30 25.51 0.38 -19.77 -20.03 2.70 2.45 0.246 1454.92 1.26 1.14 352.79 44.79 355.06 47.06
240 0.56 0.20 71.85 26.02 0.36 -19.72 -19.97 2.75 2.51 0.236 1463.21 1.29 1.17 352.94 44.94 355.38 47.38
250 0.58 0.20 73.30 25.76 0.37 -19.67 -19.91 2.80 2.56 0.239 1582.03 1.31 1.20 353.17 45.17 355.62 47.62
260 0.58 0.20 73.42 26.02 0.37 -19.64 -19.87 2.84 2.60 0.241 1557.74 1.33 1.21 353.56 45.56 355.87 47.87
270 0.58 0.21 73.54 26.27 0.37 -19.59 -19.83 2.89 2.64 0.227 1540.41 1.35 1.24 353.75 45.75 355.92 47.92
280 0.57 0.21 73.06 26.36 0.37 -19.55 -19.78 2.92 2.69 0.247 1615.73 1.36 1.26 353.61 45.61 356.14 48.14
290 0.57 0.20 72.09 26.02 0.36 -19.51 -19.76 2.96 2.72 0.223 1464.98 1.38 1.27 353.85 45.85 356.14 48.14
300 0.57 0.20 72.70 26.02 0.37 -19.52 -19.74 2.96 2.73 0.24 1643.95 1.38 1.28 353.92 45.92 356.15 48.15
310 0.58 0.21 73.54 26.53 0.37 -19.47 -19.71 3.00 2.76 0.247 1538.46 1.40 1.29 354.04 46.04 356.55 48.55
320 0.58 0.21 73.42 26.10 0.37 -19.44 -19.69 3.03 2.78 0.227 1504.57 1.42 1.30 353.99 45.99 356.42 48.42
330 0.58 0.20 74.03 26.02 0.38 -19.41 -19.64 3.06 2.84 0.247 1661.01 1.43 1.33 354.28 46.28 356.56 48.56
340 0.57 0.21 73.06 26.45 0.37 -19.39 -19.64 3.08 2.84 0.223 1482.19 1.44 1.32 354.29 46.29 356.64 48.64
350 0.58 0.21 73.42 26.19 0.37 -19.37 -19.60 3.10 2.88 0.237 1663.50 1.45 1.34 354.69 46.69 356.79 48.79
360 0.58 0.20 73.42 26.02 0.37 -19.34 -19.58 3.13 2.89 0.226 1570.89 1.46 1.35 354.57 46.57 357.07 49.07
370 0.58 0.21 73.42 26.19 0.37 -19.32 -19.54 3.16 2.93 0.255 1641.42 1.48 1.37 354.77 46.77 357.04 49.04
380 0.57 0.21 73.06 26.62 0.36 -19.28 -19.54 3.19 2.93 0.236 1430.43 1.49 1.37 354.77 46.77 357.13 49.13
390 0.58 0.21 73.91 26.62 0.37 -19.27 -19.51 3.20 2.97 0.235 1573.77 1.50 1.39 354.66 46.66 356.98 48.98
400 0.58 0.21 74.03 26.79 0.37 -19.25 -19.49 3.22 2.98 0.249 1578.81 1.50 1.39 354.61 46.61 357.19 49.19
410 0.57 0.21 73.06 26.27 0.37 -19.22 -19.47 3.25 3.01 0.258 1475.66 1.52 1.40 354.65 46.65 357.16 49.16
420 0.58 0.20 73.66 25.85 0.38 -19.19 -19.45 3.28 3.02 0.226 1455.58 1.53 1.41 354.73 46.73 357.27 49.27
430 0.58 0.21 73.78 26.27 0.37 -19.20 -19.43 3.27 3.04 0.248 1651.06 1.53 1.42 354.98 46.98 357.25 49.25
440 0.58 0.21 74.03 26.27 0.38 -19.17 -19.41 3.31 3.06 0.24 1512.26 1.55 1.43 355.18 47.18 357.28 49.28
450 0.57 0.21 72.94 26.10 0.37 -19.15 -19.39 3.32 3.08 0.236 1532.63 1.55 1.44 355.21 47.21 357.56 49.56
460 0.58 0.21 73.42 26.53 0.37 -19.14 -19.38 3.33 3.10 0.233 1560.51 1.56 1.45 355.36 47.36 357.57 49.57
207
470 0.58 0.21 73.66 26.36 0.37 -19.11 -19.35 3.36 3.13 0.245 1594.51 1.57 1.46 355.26 47.26 357.70 49.70
480 0.58 0.20 73.91 26.02 0.38 -19.10 -19.34 3.37 3.13 0.241 1535.10 1.58 1.46 355.21 47.21 357.69 49.69
490 0.57 0.21 73.18 26.10 0.37 -19.08 -19.32 3.39 3.15 0.23 1534.15 1.59 1.47 355.32 47.32 357.61 49.61
500 0.58 0.20 73.91 25.93 0.38 -19.07 -19.30 3.41 3.18 0.226 1638.13 1.59 1.48 355.32 47.32 357.64 49.64
510 0.58 0.21 73.42 26.53 0.37 -19.07 -19.29 3.41 3.18 0.249 1629.56 1.59 1.49 355.35 47.35 357.66 49.66
520 0.58 0.21 73.66 26.10 0.37 -19.04 -19.29 3.43 3.19 0.229 1500.12 1.60 1.49 355.50 47.50 357.82 49.82
530 0.58 0.21 73.30 26.36 0.37 -19.03 -19.26 3.45 3.22 0.259 1609.91 1.61 1.50 355.75 47.75 357.89 49.89
540 0.58 0.21 73.66 26.27 0.37 -19.01 -19.27 3.46 3.20 0.231 1437.03 1.62 1.50 355.71 47.71 357.85 49.85
550 0.58 0.21 73.30 26.45 0.37 -19.01 -19.24 3.46 3.23 0.25 1593.07 1.62 1.51 355.57 47.57 357.97 49.97
560 0.58 0.20 73.54 25.93 0.37 -18.99 -19.24 3.49 3.24 0.235 1495.68 1.63 1.51 355.69 47.69 358.12 50.12
570 0.58 0.21 73.78 26.27 0.37 -18.98 -19.21 3.49 3.26 0.234 1587.83 1.63 1.52 355.65 47.65 357.90 49.90
580 0.58 0.21 73.54 26.45 0.37 -18.98 -19.21 3.50 3.26 0.238 1580.77 1.63 1.53 355.59 47.59 357.92 49.92
590 0.58 0.21 73.66 26.27 0.37 -18.96 -19.20 3.51 3.27 0.227 1563.82 1.64 1.53 355.71 47.71 358.08 50.08
600 0.58 0.21 73.91 26.10 0.38 -18.95 -19.18 3.52 3.29 0.234 1653.88 1.64 1.54 355.70 47.70 357.87 49.87
610 0.57 0.21 72.94 26.53 0.36 -18.95 -19.18 3.53 3.29 0.236 1557.61 1.65 1.54 355.83 47.83 357.95 49.95
620 0.57 0.21 73.18 26.36 0.37 -18.92 -19.16 3.55 3.31 0.237 1558.14 1.66 1.55 355.82 47.82 358.05 50.05
630 0.58 0.21 74.03 26.53 0.37 -18.92 -19.16 3.55 3.31 0.247 1573.97 1.66 1.55 355.98 47.98 358.16 50.16
640 0.58 0.21 73.66 26.27 0.37 -18.91 -19.16 3.56 3.31 0.237 1506.84 1.66 1.55 355.86 47.86 358.24 50.24
650 0.58 0.21 74.03 26.27 0.38 -18.91 -19.15 3.56 3.33 0.221 1582.45 1.66 1.55 355.99 47.99 358.32 50.32
660 0.58 0.21 73.30 26.19 0.37 -18.89 -19.12 3.58 3.36 0.24 1674.25 1.67 1.57 355.94 47.94 358.47 50.47
670 0.59 0.21 74.51 26.62 0.38 -18.88 -19.12 3.59 3.35 0.235 1567.33 1.68 1.56 355.87 47.87 358.21 50.21
680 0.58 0.21 74.39 26.96 0.37 -18.89 -19.12 3.59 3.35 0.241 1585.23 1.68 1.57 355.96 47.96 358.28 50.28
690 0.58 0.21 73.91 26.70 0.37 -18.87 -19.11 3.61 3.36 0.242 1538.34 1.68 1.57 356.02 48.02 358.16 50.16
700 0.58 0.21 73.91 27.04 0.37 -18.86 -19.10 3.61 3.37 0.247 1520.91 1.69 1.57 356.02 48.02 358.39 50.39
710 0.58 0.21 73.54 26.27 0.37 -18.85 -19.09 3.63 3.38 0.237 1503.00 1.70 1.58 356.08 48.08 358.41 50.41
720 0.59 0.21 74.75 26.45 0.38 -18.84 -19.08 3.63 3.40 0.243 1600.84 1.70 1.59 356.04 48.04 358.56 50.56
730 0.58 0.21 74.15 26.27 0.38 -18.84 -19.08 3.64 3.40 0.225 1547.28 1.70 1.59 356.05 48.05 358.40 50.40
208
740 0.58 0.21 74.15 26.27 0.38 -18.83 -19.06 3.64 3.41 0.218 1671.07 1.70 1.60 356.14 48.14 358.47 50.47
750 0.58 0.21 74.15 26.19 0.38 -18.83 -19.05 3.64 3.43 0.255 1727.80 1.70 1.60 356.19 48.19 358.62 50.62
760 0.58 0.21 73.54 26.53 0.37 -18.81 -19.07 3.66 3.41 0.236 1447.96 1.71 1.59 356.23 48.23 358.41 50.41
770 0.58 0.21 73.78 26.19 0.37 -18.82 -19.06 3.65 3.42 0.232 1583.94 1.71 1.60 356.09 48.09 358.35 50.35
780 0.58 0.21 73.42 26.36 0.37 -18.80 -19.03 3.67 3.44 0.237 1593.19 1.72 1.61 356.08 48.08 358.40 50.40
790 0.57 0.21 73.18 26.19 0.37 -18.80 -19.03 3.68 3.44 0.229 1557.22 1.72 1.61 356.07 48.07 358.48 50.48
800 0.58 0.21 73.42 26.19 0.37 -18.79 -19.02 3.68 3.45 0.245 1619.91 1.72 1.61 356.19 48.19 358.55 50.55
810 0.57 0.21 73.18 26.79 0.36 -18.79 -19.03 3.69 3.44 0.245 1487.22 1.72 1.61 356.23 48.23 358.82 50.82
820 0.58 0.21 74.15 26.87 0.37 -18.77 -19.02 3.70 3.46 0.247 1515.51 1.73 1.61 356.31 48.31 358.96 50.96
830 0.58 0.21 74.15 26.19 0.38 -18.76 -19.01 3.71 3.46 0.234 1524.94 1.73 1.62 356.34 48.34 358.67 50.67
840 0.58 0.21 73.91 27.04 0.37 -18.76 -19.00 3.71 3.48 0.23 1572.91 1.73 1.62 356.32 48.32 358.68 50.68
850 0.58 0.21 73.78 26.19 0.37 -18.76 -18.99 3.71 3.48 0.247 1625.26 1.73 1.63 356.41 48.41 358.66 50.66
860 0.58 0.21 73.54 26.45 0.37 -18.75 -19.00 3.72 3.48 0.223 1497.57 1.74 1.62 356.30 48.30 358.61 50.61
870 0.58 0.21 73.66 27.04 0.37 -18.75 -18.98 3.72 3.50 0.242 1641.97 1.74 1.63 356.23 48.23 358.56 50.56
880 0.57 0.21 73.06 26.53 0.37 -18.74 -18.98 3.74 3.49 0.255 1510.04 1.75 1.63 356.35 48.35 358.69 50.69
890 0.58 0.21 74.27 26.53 0.37 -18.72 -18.98 3.75 3.50 0.23 1470.31 1.75 1.63 356.31 48.31 358.87 50.87
900 0.58 0.21 73.54 26.79 0.37 -18.73 -18.96 3.74 3.51 0.232 1596.61 1.75 1.64 356.24 48.24 358.65 50.65
910 0.58 0.20 73.66 26.02 0.37 -18.72 -18.96 3.75 3.52 0.235 1612.93 1.75 1.64 356.37 48.37 358.79 50.79
920 0.58 0.21 74.27 26.36 0.38 -18.73 -18.96 3.75 3.51 0.236 1601.15 1.75 1.64 356.51 48.51 358.87 50.87
930 0.57 0.21 72.94 26.36 0.37 -18.72 -18.95 3.76 3.52 0.238 1550.08 1.76 1.65 356.73 48.73 358.99 50.99
940 0.58 0.21 74.15 26.27 0.38 -18.71 -18.95 3.76 3.52 0.251 1579.79 1.76 1.65 356.37 48.37 358.91 50.91
950 0.58 0.21 73.42 26.10 0.37 -18.70 -18.95 3.77 3.52 0.232 1480.60 1.76 1.65 356.57 48.57 358.70 50.70
960 0.58 0.21 74.03 26.62 0.37 -18.69 -18.92 3.79 3.55 0.246 1605.00 1.77 1.66 356.30 48.30 358.65 50.65
970 0.58 0.21 73.78 26.45 0.37 -18.70 -18.95 3.77 3.53 0.245 1511.38 1.76 1.65 356.43 48.43 358.72 50.72
980 0.58 0.21 74.15 26.36 0.38 -18.69 -18.94 3.78 3.54 0.221 1531.92 1.77 1.65 356.41 48.41 358.76 50.76
990 0.58 0.21 73.78 26.10 0.37 -18.70 -18.92 3.78 3.56 0.231 1694.48 1.76 1.66 356.53 48.53 358.66 50.66
1000 0.58 0.21 74.15 26.62 0.37 -18.69 -18.92 3.79 3.56 0.241 1616.06 1.77 1.66 356.65 48.65 359.02 51.02
209
Samir Ashour Appendix D
APPENDIX D
1 Linear Regression Analysis
1.1 Introduction
Regression analysis is a powerful technique that can be used to address various research
questions. In this report, we are going to use it to check how qult levels are affected by cu
and φ.
In particular, the type of regression we are going to use is ”Multiple Linear Regression”.
Linear regression is the process of finding the best-fitting straight line through data points
(this line is sometimes referred to as the regression line). Multiple means we have more
than one input variable (also known as predictor), hence, we are trying to fit a plane or
hyper-plane rather than a line. The input variables in our case are cu, phi and as. Linear
means that we are trying to find a combination of the input variables such that each
variable is multiplied by a coefficient and then we sum the products. The idea is to use this
linear combination of input variables to model their relationship with an output variable (in
our case, this is qult).
1.2 The Modelling Tool

In order to fit models, we are going to use R [ ] which is a powerful and easy to use tool
for statistical computing and graphics. R makes it easy to manipulate data and perform
calculations as well as display information graphically. It also facilitates modelling (linear
and nonlinear) and other statistical processes.
1.3 Diagnostics for Examining a Fit

There are several diagnostics that can be used to explore the goodness of fit of a model.
The following are going to be used:
210
 R-squared
This value calculates the percentage of variation of the output explained by the input
variables in the model. This means the higher the value of R-squared the better the model.
 R-squared adjusted
This value is similar to R-squared but it accounts for the number of input variables in the
model, hence, it is sometimes preferred to R-squared.
 Residuals
A residual is the difference between the actual value and predicted value for each point (or
record) in the data. Histograms are often used to check the distribution of residuals. Also,
they are plotted against each input variable. If a model fits well, the residuals will be small
and will be no pattern of their distribution around zero (i.e. they should be evenly spread
around zero).
 Deviance
The deviance is a statistic that is used to determine the quality of fit for a model. It is a
measure of how much better a model with more parameters fits the data. It is used to
compare nested models. A nested model is a model which is a subset of another model.
For example, if we have two models, the first describes the relationship between one input
variable x and an output variable y and the second describes the relationship between two
input variables x and z and an output variable y, then the first model is nested within the
second.
211
 Bayesian Information Criterion (BIC)
When there are a limited number of models, the BIC is often used for model selection.
Often, the model with the smallest BIC is preferred. The BIC is a penalised version of the
deviance where the penalty is relevant to the number of input variables.
 P-values
It is common to provide p-values when conducting statistical analysis. P-values are
propabilities (i.e. their values always lie between zero to 1) and they show how likely
certain situations are. P-values which are close to zero (usually <= 0.05) are more likely to
occur if the study has shown something positive. It is said that the result is significant if
the p-value is close to zero. On the other hand, the result is said to be non-significant if the
p-value is away from zero (usually > 0.05).
 Confidence Intervals
It is known that in statistics we use the sample data at hand to draw inferences about the
entire population (i.e. all the data) and make an estimate of the value(s) we are trying to
measure or predict. It is very important to present such estimate with a measure of
precision. This measure of precision depends on the sample size and normally takes the
form of a 95% Confidence Interval or a standard error value (the former is calculated from
the latter). The 95% confidence interval gives the range of poulation parameters that the
sample leads us to believe are possible. The 95% confidence interval is presented as a
range of two values (a, b) and is interpreted as: we can be 95% confident that the
result/effect we are trying to measure what will happen by an overage of at least a and
maybe as much as b
212
Data base
Reference cu, Φ, AS, q untreated, q treated, q treated/q

(o)
Wood et al. (2000) (kPa)
10.5 25 (%)
30 (kPa)
84 (kPa)
105 1.25
untreated
Wood et al. (2000) 10 25 24 84 90 1.07
Wood et al. (2000) 10 25 10 84 86 1.02
Andereou et al. (2008) 20 32 4 55 62 1.13
Najjar et al. (2010) 20 33 7.9 62 75 1.21
Najjar et al. (2010) 20 33 17.8 67 100 1.49
Najjar et al. (2010) 20 33 7.9 84 101 1.20
Najjar et al. (2010) 20 33 17.8 84 148 1.76
Sivakumar et al. (2004) 25 35 10.2 60 83 1.38
Black et al. (2006) 25 35 17 1.25 2 1.60
Bergado et al. (1987) 20 35 6.25 175 221 1.26
Hughes and Withers 19 35 40 171 418 2.44
Bergado et al. (1987) 20 37 6.25 175 314 1.79
(1974)and Guermazi
Juran 30 38 4 120 154 1.28
Ali et al. (2014) 7 38 25 25 70 2.80
(1988)
Kim and Lee (2005) 4 38 30 30 75 2.50
Kim and Lee (2005) 4 38 40 45 135 3.00
Kim and Lee (2005) 4 38 50 45 155 3.44
Bergado et al. (1987) 20 39 6.25 175 320 1.83
Bergado et al. (1987); st. 30 42 65 60 270 4.50
Bergado et al. (1987); 20 42 28 95 240 2.53
Helens Bridge
Humber 25 42 11 115 270 2.35
CanveyetIsland
Black al. (2011) 35 43 17 240 680 2.83
Black et al. (2011) 35 43 28 230 750 3.26
Black et al. (2011) 35 43 40 230 820 3.57
Black et al. (2011) 35 43 1.6 29 65 2.24
Bergado et al. (1987) 20 43 6.25 175 370 2.11
Ambily and Gandhi 7 43 19 75 160 2.13
(2007)
(2007) and Gandhi
Ambily 30 43 5 150 600 4.00
(2007) and Gandhi
Ambily 30 43 9 150 680 4.53
(2007) and Gandhi
Ambily 30 43 19 150 770 5.13
(2007)
Zahmatkesh and 5 43 10 80 165 2.06
(2007)
Zahmatkesh and 5 43 20 80 170 2.13
Choobbasti (2012)
Zahmatkesh and 5 43 25 80 250 3.13
Choobbasti
Zahmatkesh and(2012) 5 43 30 80 280 3.50
Choobbasti
Watts (2012)
et al. (2000) 40 45 44 12 248 20.67
Choobbasti
Current study (2012) 12 48 5 55 100 1.82
Current study 12 48 13 140 428 3.06
Kim and Lee (2005) 4 49 30 40 150 3.75
Kim and Lee (2005) 4 49 40 40 165 4.13
Kim and Lee (2005) 4 49 50 40 215 5.38
213
1.4 Examining the Variables

As there are three numerical input variables, it is appropriate to examine their statistical
summaries. This is what we show in table 1.
Table 1: Statistical Summaries of the Variables with 95% Confidence Interval for the
Mean values
Property cu ∅ As qult
Valid 31 31 31 31
Missing - - - -
Mean 17.23 40.31 21.55 240.3
Median 20.00 39.00 17.00 165.0
Std. deviation 10.87101 5.168765 16.41788 202.4052
Minimum 4.00 33.00 1.60 2.0
Maximum 35.00 49.00 65.00 820.00
By examining the table, we observe that we have data for all the points (i.e. no missing
values). Also, we notice that there is no big deference between the mean and median
values of each variable, hence outliers are unlikely (outliers usually have a big influence
on the deference between the mean and the median). Another values that can analyse from
the table are the minimum and maximum values for each input variable. They appear to be
within possible ranges for all of the three input variables. Finally, as standard deviation is
an indicator of how spread out the data is, we can check the validity of our data by going 2
standard deviations on each side of the mean for the outcome variable Q. We notice that
more than 95% of all values of this variable lie within that range.
214
1.5 Histograms of the Variables

Histograms of the three input variables and the output variable are displayed in Figure 1. A
close look at these histograms indicate some "bunching" around multiples of 5 for the cu
input variable (Figure 1a) and around multiples of 10 for the AS variable (Figure 1c).
Figure 1: Histograms of the Variables
1.6 Relationships between the Variables

A scatter plot matrix between all variables in is shown Figure 2. The top row of this scatter
plot matrix gives the scatter plots of qult against each of the other three input variables.
Additionally, we show the univariate relationship between our outcome variable qult and
the input variables in Table 2. From this table, we can notice that the correlation between
cu and qult and between φ and qult is significant at the p < 0.05 level.
215
Figure 2: Scatter plot of all the variables
Table 2: Correlations between the Input Variables and qult
Influence factor Correlation (p-value) with qult
Pearson Spearman
cu 0.4595897 (0.009293) 0.1952728 (0.2925)
As 0.1964417 (0.2895) 0.2026105 (0.2743)
∅ 0.3082577 (0.09158) 0.3982587 (0.02649)
Information also are provided about the correlations between our input variables in table 3.
From this table we can observe that none of the correlations between the input variables is
significant at the p < 0.05 level
216
Table 3 correlation between input variables
(a) Pearson correlation
As cu ∅
As 1 -0.2596521 0.2605386
(0.1584) (0.1569)
cu -0.2596521 1 -0.2176558
(0.1584) (0.2395)
∅ 0.2605386 -0.2176558 1
(0.1569) (0.2395)
(b) Spearman correlation
As cu ∅
As 1 -0.3944457 0.204013
(0.0281) (0.271)
cu -0.3944457 1 -0.2161602
(0.0281) (0.2428)
∅ 0.204013 -0.2161602 1
(0.271) (0.2428)
1.7 Fitting the Models

As the aim is to determine the relation between VARIABLE(s) and qult. Therefore, it
would seem reasonable to model CONFOUNDER FIRST and then enter cu into the model
and see whether it is associated with qult. We can do this in the opposite way by entering cu
into the model first and then follow it by φ and AS in turn to see what impact each of them
would have on the model.
217
1.7.1 Modelling the Relationship between cu and qult (Model 1)
In this section, a simple linear regression model is built using just cu as input and qult as
output. After using R's lm() function, the model looks as follows:
𝑞𝑢𝑙𝑡 = 92.890 + 8.557 ∗ 𝑐𝑢 (1)
By examining Equation 1, we observe that an increase, or decrease, of cu by one unit,
causes an increase, or decrease, in qult by 8.557 units
Figure 3: The Relationship between cu and qult
Now after building the model that describes the relationship between cu and qult, let us plot
the input variable cu against the residuals. As Figure 4 shows, the residual values do not
seem to have a particular pattern and they are randomly scattered around zero.
218
Figure 4: cu vs Residuals for Model 1
1.7.2 Modelling the Relationship between AS and qult (Model 2)
In this section, we are going to build a simple linear regression model using just AS as
input and Q as output. After using R's lm() function, the model looks as follows:
𝑞𝑢𝑙𝑡 = 188.089 + 2.422 ∗ 𝐴𝑠 (2)
By examining equation 2, we observe that an increase, or decrease, of AS by one unit,
causes an increase, or decrease, in qult by 2.422 units
219
Figure 5: The Relationship between AS and qult
Now after building the model that describes the relationship between AS and qult, let us plot
the input variable AS against the residuals. As Figure 6 shows, the residual values do not
seem to have a particular pattern and they are randomly scattered around zero (as we stated
in Section 4.3).
220
Figure 6: AS vs Residuals for Model 2
1.7.3 Modelling the Relationship between φ and qult (Model 3)
In this section, a simple linear regression model was built using just φ as input and qult as
output. After using R's lm() function, the model looks as follows:
𝑞𝑢𝑙𝑡 = −244.11 + 12.07 ∗ ∅ (3)
By examining equation 3, we observe that an increase, or decrease, of PHI by one unit,
causes an increase, or decrease, in Q by 12.07 units
221
Figure 7: The Relationship between φ and qult
Now after building the model that describes the relationship between φ and qult, let us plot
the input variable PHI against the residuals. As Figure 8 shows, the residual values do not
seem to have a particular pattern and they are randomly scattered around zero.
Figure 8: PHI vs Residuals for Model 3
222
1.7.4 Modelling the Relationship between cu, AS and qult (Model 4)
In this section, we are going to build a simple linear regression model using CU and AS as
inputs and Q as output. After using R's lm() function, the model looks as follows:
𝑞𝑢𝑙𝑡 = −25.286 + 10.194 ∗ 𝑐𝑢 + 4.174 ∗ 𝐴𝑠 (4)
By examining Equation 4, we observe that when fixing AS, an increase, or decrease, of cu
by one unit, causes an increase, or decrease, in qult by 10.194 units. Similarly, when fixing
cu, an increase, or decrease, of AS by one unit, causes an increase, or decrease, in qult by
4.174 units.
Now after building the model that describes the relationship between cu, AS and qult, let us
plot the input variables cu and AS against the residuals. As Figure 9 shows, the residual
values do not seem to have a particular pattern and they are randomly scattered around
zero.
(a) cu vs Residuals for Model 4
223
(b) AS vs Residuals for Model 4
Figure 9: Input Variables vs Residuals for Model 4
1.7.5 Modelling the Relationship between cu, φ and qult (Model 5)
In this section, we are going to build a simple linear regression model using cu and φ as
inputs and qult as output. After using R's lm() function, the model looks as follows:
𝑞𝑢𝑙𝑡 = −610.53 + 10.29 ∗ 𝑐𝑢 + 16.78 ∗ ∅ (5)
By examining equation 5, we observe that when fixing φ, an increase, or decrease, of cu by
one unit, causes an increase, or decrease, in qult by 10.29 units. Similarly, when fixing CU,
an increase, or decrease, of φ by one unit, causes an increase, or decrease, in qult by 16.78
units. Now after building the model that describes the relationship between cu, PHI and
qult, let us plot the input variables cu and φ against the residuals. As Figure 10 shows, the
residual values do not seem to have a particular pattern and they are randomly scattered
around zero.
224
(b) φ vs Residuals for Model 5
1.7.6 Modelling the Relationship between AS, φ and qult (Model 6)
In this section, we are going to build a simple linear regression model using AS and φ as
inputs and qult as output. After using R's lm() function, the model looks as follows:
𝑞𝑢𝑙𝑡 = −226.213 + 1.536 ∗ 𝐴𝑠 + 10.8 ∗ ∅ (6)
225
By examining equation 6, we observe that when fixing φ, an increase, or decrease, of A S
by one unit, causes an increase, or decrease, in qult by 1.536 units. Similarly, when fixing
AS, an increase, or decrease, of φ by one unit, causes an increase, or decrease, in qult by
10.8 units. Now after building the model that describes the relationship between AS, φ and
qult, let us plot the input variables AS and φ against the residuals. As Figure 11 shows, the
around zero.
226
(a) AS vs Residuals for Model 6
(b) φvs Residuals for Model 6
1.7.7 Modelling the Relationship between cu, AS, φ and qult (Model 7)
In this section, we are going to build a simple linear regression model using CU, AS and
PHI as inputs and Q as output (recall this is the purpose of this study). After using R's lm()
function, the model looks as follows:
𝑞𝑢𝑙𝑡 = −609.893 + 11.313 ∗ 𝑐𝑢 + 3.167 ∗ 𝐴𝑠 + 14.629 ∗ ∅ (7)
By examining Equation 7, we observe that when fixing φ and AS, an increase, or decrease,
of cu by one unit, causes an increase, or decrease, in qult by 11.313 units. Similarly, when
fixing cu and AS, an increase, or decrease, of φ by one unit, causes an increase, or decrease,
in qult by 14.629 units. Also, when fixing cu and φ, an increase, or decrease, of AS by one
unit, causes an increase, or decrease, in qult by 3.167 units.
227
Now after building the model that describes the relationship between cu, AS, φ and qult, let
us plot the input variables cu, AS and φ against the residuals. As Figure 12 shows, the
around zero.
(b) AS vs Residuals for Model 7
228
φ vs Residuals for Model 7
1.7.8 Checking the Model Fit
Five diagnostics that can be used to examine the goodness of fit of a model were
mentioned In Section 1.3. Namely, these were: Deviance, BIC, R-Squared, Adjusted R-
Squared and the Residuals. Table 4 shows the values of the remaining four diagnostics for
the seven models.
Table 4: Deviance, BIC, R-Squared and Adjusted R-Squared of the Seven models
Model Deviance BIC Value R-Squared Adj R-Squared
Model (1) 969436 419.1411 0.2112227 0.1840235
Model (2) 1181609 425.2766 0.03858933 0.005437238
Model (3) 1112250 423.4013 0.09502283 0.06381672
Model (4) 838024.3 418.0594 0.3181453 0.2694414
Model (5) 754365.9 414.7992 0.3862136 0.3423717
Model (6) 1094468 426.3357 0.1094908 0.04588297
Model (7) 682279.8 415.1196 0.4448661 0.3831845
229
As we mentioned before, lower BIC indicates a better fitting model. By analysing table 4
we observe that Model 5 has the lowest BIC with Model 7 in second. Howerver, when we
examine the value of R-Squared, we realise that Model 7 has the highest R-Squared
amongst all the models (see Section 4.1). This gives us confidence that from amongst the
seven models that we created using various combinations of the input variables, Model 7 is
the best model that describes the relationship between the input variables CU, AS and PHI
and the output variable Q.
230
Samir Ashour Appendix F
APPENDIX F
Nonlinear Regression Analysis
 Nonlinear Regression Analysis_ soil only
Iteration Historyb
Residual Parameter
Iteration Sum of
a
Number Squares a b m
1.0 71288.417 .000 .000 .000
1.1 29769.136 1.807 .000 .000
2.0 29769.136 1.807 .000 .000
2.1 23243.121 3.354 .202 3.196
3.0 23243.121 3.354 .202 3.196
3.1 4830.984 3.625 .145 2.993
4.0 4830.984 3.625 .145 2.993
4.1 4333.676 4.377 .091 2.411
5.0 4333.676 4.377 .091 2.411
5.1 4316.993 4.332 .097 2.477
6.0 4316.993 4.332 .097 2.477
6.1 4316.966 4.348 .096 2.468
7.0 4316.966 4.348 .096 2.468
7.1 4316.965 4.346 .096 2.470
8.0 4316.965 4.346 .096 2.470
8.1 4316.965 4.346 .096 2.469
Derivatives are calculated numerically.
a. Major iteration number is displayed to the left of the decimal,
and minor iteration number is to the right of the decimal.
b. Run stopped after 16 model evaluations and 8 derivative
evaluations because the relative reduction between successive
residual sums of squares is at most SSCON = 1.000E-8.
231
Parameter Estimates
95% Confidence Interval
Std. Lower Upper
Parameter Estimate Error Bound Bound
a 4.346 .197 3.960 4.733
b .096 .005 .085 .107
m 2.469 .170 2.136 2.803
Correlations of Parameter
Estimates
a b m
a 1.000 -.400 .550
b -.400 1.000 .540
m .550 .540 1.000
ANOVAa
Sum of Mean
Source Squares df Squares
Regression 66971.451 3 22323.817
Residual 4316.965 3058 1.412
Uncorrected
71288.417 3061
Total
Corrected Total 4932.510 3060
Dependent variable: VAR00005
a. R squared = 1 - (Residual Sum of Squares) / (Corrected
Sum of Squares) = .125.
232
 Nonlinear Regression Analysis_ soil/stone column
Iteration Historyb
Residual Parameter
Iteration Sum of
a
Number Squares a b m
1.0 105048.595 .000 .000 .000
1.1 59039.952 1.285 .000 .000
2.0 59039.952 1.285 .000 .000
2.1 14058.800 3.055 .145 2.207
3.0 14058.800 3.055 .145 2.207
3.1 15227.522 6.360 .098 4.206
3.2 9802.091 4.171 .146 3.403
4.0 9802.091 4.171 .146 3.403
4.1 10647.892 7.604 .105 4.544
4.2 8249.791 5.485 .145 4.075
5.0 8249.791 5.485 .145 4.075
5.1 7672.376 8.405 .108 4.564
6.0 7672.376 8.405 .108 4.564
6.1 6139.380 11.429 .082 4.654
7.0 6139.380 11.429 .082 4.654
7.1 6122.613 17.495 .034 4.895
8.0 6122.613 17.495 .034 4.895
8.1 5059.439 20.596 .030 4.991
9.0 5059.439 20.596 .030 4.991
9.1 5023.065 23.229 .019 5.066
10.0 5023.065 23.229 .019 5.066
10.1 5017.609 23.381 .019 5.082
11.0 5017.609 23.381 .019 5.082
11.1 5017.607 23.393 .019 5.080
12.0 5017.607 23.393 .019 5.080
12.1 5017.607 23.392 .019 5.081
Derivatives are calculated numerically.
a. Major iteration number is displayed to the left of the decimal,
and minor iteration number is to the right of the decimal.
b. Run stopped after 26 model evaluations and 12 derivative
evaluations because the relative reduction between successive
residual sums of squares is at most SSCON = 1.000E-8.
233
Parameter Estimates
95% Confidence Interval
Std. Lower Upper
Parameter Estimate Error Bound Bound
a 23.392 .445 22.519 24.264
b .019 .002 .015 .024
m 5.081 .039 5.004 5.157
Correlations of Parameter
Estimates
a b m
a 1.000 -.751 .335
b -.751 1.000 .355
m .335 .355 1.000
ANOVAa
Sum of Mean
Source Squares df Squares
Regression 100030.989 3 33343.663
Residual 5017.607 6050 .829
Uncorrected
105048.595 6053
Total
Corrected Total 26624.392 6052
Dependent variable: permanent deformation
a. R squared = 1 - (Residual Sum of Squares) / (Corrected
Sum of Squares) = .812.
234
Samir Ashour Appendix G
APPENDIX G
Subgrade Stress Investigation
A large scale test at the University of Birmingham (unpublished work) included
application of 125 kN wheel load on rail across three sleepers as shown in Figure 1a. The
cross section of the setup is shown in Figure 1b. Load was applied at 2 Hz for 2 million
cycles. Transducer positioned at top of subgrade under the central sleeper (where load was
applied) in order to monitor the stresses. Figure 2 showed that the dynamic stresses at the
subgrade level were approximately 80 kPa showing good agreement with Yoo and Selig
(1979) study.
a. Cyclic load application
235
Rail Pressure
215 mm Load ram Transducer
Rail conneted to three sleepers Sleeper
Pressure Transducer
Half sleeper PT
Ballast 2
600mm
X
300mm
X
1800mm
PT
570mm
Subgrade material
1
600mm
225mm
100mm
Building sand
730mm
PT
3
3000.00mm
Load
application
b. Test cross section layout
Figure 1 Large scale test set up
100
PT1
PT2
Stresses at the subgrade level, (kPa)
PT3
80
60
40
20
0
0 50 100 150
Time, (minute)
Figure 2 Dynamic stresses measurements at subgrade level (under 300mm of ballast)
236
Based on the above, cyclic stresses of 50 kPa, 60 kPa and 70 kPa were used in this study.
These values are equivalent to a cyclic stress ratio (CSR) of 0.6, 0.7 and 0.8 (CSR = the
cyclic deviator stress (qcyclic)/ the static deviator stress of reinforced soil at failure (qfailure)).
Surcharge Pressure
Surcharge pressure, equivalent to the weight of the track component (see Figure 3) was
applied as the lowest pressure level during cyclic loading. this could be ranged between 10
and 25 kPa depending on the rail and sleeper types and the depth of the used ballast layer
(Selig and Waters, 1994; Brough et al., 2003). Using ballast thickness of 300 mm with a
density of 19 kN/m3, rail weight of 60 kg/m, and a concrete sleeper of 300 kg with typical
dimensions of 285 mm X 2420 (l sleeper) mm laying in distance (S sleeper) of 600 mm centre
to centre, the surcharge pressure should be equivalent to 9.4 kPa. However, due to
limitation of the test apparatus minimum of 20 kPa only could be applied
SSleeper
lSleeper
Figure 3 Structure of a rail track system used to calculate the surcharge pressure.
237

Ashour 16 PH D

Uploaded by

Copyright:

Available Formats

Ashour 16 PH D

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ashour 16 PH D

Uploaded by

Copyright:

Available Formats

UNIVERSITY OF BIRMINGHAM

THE RESPONSE OF STONE COLUMNS

The University of Birmingham

For the degree of

School of CIVIL ENGINEERING

This unpublished thesis/dissertation is copyright of the author and/or third

Any use made of information contained in this thesis/dissertation must be in

subjected to cyclic loading as in railway tracks based on a laboratory scale investigation.

terms of load/deformation characteristics; in monotonic loading condition, there was

situation without stone column.

Also to my brother who always has been there for me.

And above all

messenger, Mohammed (May Allah bless and grant him)

encouragement throughout the research.

White for their help and support.

for this research.

1.2 Research problem ............................................................................................................ 3

1.3 Research Aim and Objectives.......................................................................................... 4

1.4 Outline of Thesis ............................................................................................................. 5

2.2 Introduction to Ground Improvement.............................................................................. 6

2.3 Stone Column Technique: Historical Development ........................................................ 8

2.4 Stone Column Construction Methods ............................................................................. 8

2.4.1 Vibro Equipment (Vibro-float) ....................................................................................... 9

2.5.1 Limitations .................................................................................................................... 19

2.6 Design Approaches of Stone Column ........................................................................... 19

2.6.1 Basic Design Parameters ............................................................................................... 20

2.7.1 Stone Columns Investigation under Monotonic Loading Condition ............................. 39

3.2 Materials ........................................................................................................................ 57

3.3.1 Index Properties............................................................................................................. 59

3.4.1 Scale Effect ................................................................................................................... 69

3.6 Test Apparatus ............................................................................................................... 75

3.6.1 Cyclic Load Frame ........................................................................................................ 75

3.7.1 Loading Frequency (f) ................................................................................................... 76

3.8.1 Clay Bed Preparation .................................................................................................... 81

4.2 Model I (Triaxial) .......................................................................................................... 97

4.2.1 Saturation ...................................................................................................................... 97

4.3.1 Load - Displacement Relationship .............................................................................. 117

5. CHAPTER 5................................................................................................................................ 131

5.2 Model I (Triaxial Tests) ............................................................................................... 132

5.2.1 Permanent Strain during Cyclic Loading .................................................................... 132

5.3 Model II Test Results (Large Scale Test: Foundation)................................................ 150

5.3.1 Permanent Strain during Cyclic Loading .................................................................... 150

5.5 Concluding Remarks ................................................................................................... 157

5. CHAPTER 6................................................................................................................................ 159

6.2 Monotonic Loading ..................................................................................................... 160

6.2.1 General ........................................................................................................................ 160

6.3.1 Effect of Stress Level .................................................................................................. 161

REFERENCE ...................................................................................................................................... 164

Figure ‎2.2 Vibro-replacement method (Keller GmbH, 2005) .................................................. 12

Figure ‎2.3 Vibro-displacement (Keller GmbH, 2005) ............................................................. 13

Figure ‎2.4 Dry bottom feed (Keller GmbH, 2005) ................................................................... 14

Figure ‎2.7 Settlement estimation conditions ............................................................................ 25

Figure ‎2.9 Improvement factor (Priebe, 1995) ......................................................................... 32

capacity of the treated ground .................................................................................................. 45

Figure ‎2.16 Classification of dynamic problems (Ishihara; 1996) ........................................... 49

Figure ‎3.6 Shear stress verses normal stress at peak ................................................................ 65

Figure ‎3.8 Void ratio / pressure increment relationship ........................................................... 68

Figure ‎3.12 Testing equipment ................................................................................................. 75

during consolidation ................................................................................................................. 87