Analysing Ground Deformation Data to Predict

Characteristics of Smear Zone Induced by Vertical Drain

Installation for Soft Soil Improvement

A thesis in fulfilment of the requirement

for the award of the degree

Doctor of Philosophy

from

University of Technology, Sydney

by

Ali Parsa-Pajouh, BSc Eng, Msc Eng

School of Civil and Environmental Engineering,

Faculty of Engineering and Information Technology

February 2014
 CERTIFICATION

I, Ali Parsa-Pajouh, declare that this thesis, submitted in fulfilment of the

requirements for the award of Doctor of Philosophy, in the Department of Civil and
Environmental Engineering, University of Technology, Sydney, is wholly my own
work unless otherwise referenced or acknowledged. The document has not been
submitted for qualification at any other academic institution.

Ali Parsa-Pajouh
February 2014

II
 DEDICATION

This work is sincerely dedicated to the following special people:

To My Lovely Wife, Neda

For her endless encouragement and patience, without her love and support, I could
have never completed this journey.

To My Father, Davoud
He was my role-model of hard work, persistence and personal sacrifices. He inspired
me how to be strong, how to be honest.

To My Late Beloved Mother, Maryam

For all the sacrifices she has made to ensure that I obtain the best education possible,
for her beautiful mind and unconditional love.

III
 ABSTRACT

The use of prefabricated vertical drain (PVD) assisted preloading has been
recognised over the last two decades as a very efficient method of ground
improvement for sites with deposits of deep soft soil. One of the major parameters
influencing the PVD assisted consolidation process, and consequently the required
preloading time, is the formation of a smear zone around the vertical drains, and the
corresponding soil properties. In this research a systematic procedure integrated with
a developed numerical code is proposed to accurately back calculate the properties of
the smear zone based on the consolidation data collected in the laboratory and in the
field. Furthermore, an expanded back calculation method is developed to determine
the minimum required degree of consolidation and corresponding time after the
construction of the trial embankment that would result in accurately predicted smear
zone characteristics. The explicit finite difference program FLAC 2D was used to
develop the numerical code, simulate the laboratory testing and PVD assisted
preloading case histories. Furthermore a comprehensive parametric study was
conducted to investigate the effect of smear zone properties variations on the
preloading process, and back calculated characteristics of the smear zone.
A large and fully instrumented Rowe cell apparatus was used to investigate
the effect of the smear zone on the consolidation process and verify the developed
numerical code. The Rowe cell was filled with the intact zone, smear zone, and
vertical drain materials to evaluate the permeability and extent ratios of kh/ks=4 and
rs/rm=3, respectively. The back calculation procedure was used to conduct the
parametric study and predict the properties of the smear zone. According to the
results, the predicted properties of the smear zone were similar to the properties of
the applied soil, proving that the proposed back calculation procedure integrated with
the developed numerical simulation can successfully predict these properties.
The developed numerical code was used to simulate five PVD assisted
preloading case studies, including four trial embankments and a large scale
consolidometer, while the back calculation procedure was used to conduct a
parametric study to determine the extent and permeability of the smear zone.
According to the results, integration of the back calculation procedure in the

IV
numerical code can be used as a reliable tool to make an accurate prediction of the
smear zone characteristics in PVD and vacuum assisted preloading projects.
The developed method in this research can be considered as a practical,
accurate and cost effective tool, due to its capability in precise estimation of the
extent and permeability of the smear zone in the early stages of constructing the trial
embankment. In this study, the proposed systematic back calculation procedure was
extended to determine the minimum degree of consolidation (i.e. the minimum
waiting time after constructing the trial embankment), and accurately predict the
properties of the smear zone. The numerical results of the simulated case studies
were used to conduct the analyses. Accordingly, it is found that the extent and
permeability of the smear zone can be predicted very well with the proposed
calculation procedure when at least 33% of predicted final settlement has been
reached (i.e. 33% of the degree of consolidation).

V
 ACKNOWLEDGEMENT

Working as a PhD student at University of Technology, Sydney was a magnificent as

well as challenging experience to me. In all these years, many people were involved
directly or indirectly in shaping up my academic achievements. It was hardly
possible for me to succeed in my doctoral work without the precious support of these
kind personalities. Here is a small tribute to all those people.
First of all, I would like to express my deepest gratitude to my supervisor,
Dr. Behzad Fatahi for his brilliant guidance, caring, patience, and providing me with
an excellent atmosphere for doing research. It has been an honour to be his first PhD
student. His numerous comments, criticisms and suggestions during the preparation
of this thesis are gratefully acknowledged.
I would like to thank my co-supervisor, A/Prof. Hadi Khabbaz, who was
abundantly helpful and offered invaluable assistance, support and guidance. His
encouragement, support and critical review of the thesis, as well as his
comprehensive suggestions at various stages of this study are greatly appreciated. I
would also like to acknowledge Bill Clayton for his critical review of the thesis and
helpful comments.
My appreciation is also extended to UTS laboratory and workshop staff
Antonio Reyno and Laurence Stonard as well as the PhD students Babak Azari, and
Thu Minh Le for their valuable assistance and contribution in running and
conducting the laboratory test. I am also thankful to University of Technology,
Sydeny and its staff for their financial, academic and technical support. The library
and computer facilities of the University have been indispensable. I very much
appreciate the Centre for Built Infrastructure Research (CBIR), particularly, Prof.
Bijan Samali for their support and assistance since the start of my PhD study.
I gratefully acknowledge the funding sources that made my PhD work
possible. This research was funded by Australian Research Council and supported by
the Menard-Bachy Pty Ltd as the industry partner, which are so appreciated.
I would also like to acknowledge Road and Maritime Services (RMS) for
providing the field measurements and site investigation results for Ballina Bypass
and Cumbalam trial embankments.

VI
 I have been in charge of tutoring different subjects in last four years, which
was a priceless experience for me. I would like to take this opportunity to express my
special gratitude towards academic staff of the School of Civil and Environmental
Engineering, Dr. Behzad Fatahi, A/Prof. Hadi Khabbaz, Dr. Shami Nejadi, Dr.
Rezaul Karim, Anne Gardner, Chris Wilkinson and Dr. Rijun Shrestha, who
supported me in this journey.
Gratitude is also expressed to my friends and fellow students at the
University of Technology, Sydney, particularly, Amir Zad, Babak Azari, Behnam
Fatahi, Binod Shresta, Hamid Tabatabaei, Thu minh Le and Yashar Maali for
keeping the study atmosphere enjoyable and pleasant.
I would like to express heartfelt acknowledgement to my wife, Neda
Khodadoust, for her constant love, sacrifices, support, encouragement, and patience.
She was always there cheering me up and stood by me through the happy and
hardships times. Without her guidance and support, I may never have come this far
in my study.
Lastly, I would like to thank my parents, who raised me with a love of
science and supported me in all my pursuits. Special thanks to my two elder sisters,
and my younger brother, who were always supporting and encouraging me with their
best wishes.

VII
 TABLE OF CONTENT

ABSTRACT IV
ACKNOWLEDGEMENT VI
TABLE OF CONTENT VIII
LIST OF FIGURES XIV
LIST OF TABLES XXIII
LIST OF SYMBOLS XXV

1. INTRODUCTION
1.1 General 1
1.2 Accelerating the Consolidation Process 4
1.3 Installation of Vertical Drains and Smear Zone 7
1.4 Trial Embankment Monitoring to Obtain the Smear zone Properties 9
1.5 Objectives and Scope of Present Study 9
1.6 Organisation of the Thesis 11

2. LITERATURE REVIEW
2.1 History and Development of Vertical Drain Assisted Preloading 13
2.2 Vacuum Preloading via Prefabricated Vertical Drains 14
2.2.1 History and Developments of Vacuum Preloading 15
2.2.2 Vacuum Preloading Using Membrane 17
2.2.3 Vacuum Preloading (Membrane Free Techniques) 18
2.3 Factors Affecting Consolidation of Clay with PVDs 20
2.3.1 Equivalent Diameter 20
2.3.2 Filter and Apparent Opening Size (AOS) 22
2.3.3 Tensile Strengths 23
2.3.4 Discharge Capacity and Well Resistance 24
2.3.5 Smear Zone 26
2.3.6 Soil Macro Fibre 26
2.3.7 Mandrel Size and Shape 27
2.3.8 Installation Procedure 28
2.3.9 Drain Spacing and Influence Zone 29
2.4 Smearing Effect 29

VIII
 2.4.1 Smear Zone Generation 30
2.4.1.1 Soil Remoulding Concept 30
2.4.1.2 The Reconsolidation Theory 31
2.4.2 Smear Zone Extent & Permeability Variation 32
2.4.3 Estimation of the Smear Zone Properties 35
2.4.3.1 Experimental Methods 35
2.4.3.2 Cavity Expansion Theory 43
2.4.3.3 Finite Element Methods 48
2.4.3.4 Back Calculation Methods 50
2.4.4 Relationship between Experimental and Practical Results for Smear
Zone Properties 54
2.5 Development of Consolidation Theory 55
2.5.1 Vertical Consolidation 55
2.5.2 Radial (or Horizontal) Consolidation Considering Smear Zone
Characteristics 57
2.5.2.1 Conversion of Axi-symmetric to Plane-Strain Condition 62
2.5.3 Combined Vertical and Radial Consolidation Theory 67
2.5.3.1 Single Layer Consolidation (Rigorous Solutions) 67
2.5.3.2 Single Layer Consolidation (Approximate Solutions) 68
2.5.3.3 Multi-Layered Consolidation 69
2.5.4 Theoretical Solutions for Vacuum Consolidation 72
2.6 Numerical Simulation of PVD Assisted Preloading 74
2.7 Summary 78

3. PVD ASSISTED PRELOADING SIMULATION AND BACK

CALCULATION PROCEDURE USING FLAC
3.1 General 80
3.2 Numerical Modelling 81
3.2.1 Adopted Numerical Simulation Software 81
3.2.2 Explicit Finite Difference Method and Lagrangian Analysis 82
3.2.3 The Grid and Mixed-Discretization Zoning Technique 85
3.2.4 Continuum Expression of the Governing Equations 87
3.2.4.1 Water Flow Equation 90

IX
 3.2.4.2 Balance Laws 90
3.2.4.4 Constitutive Laws 91
3.2.4.4 Compatibility Equation 92
3.2.5 Numerical Fluid Flow Formulation 92
3.2.5.1 Basic Scheme 92
3.2.5.2 Constitutive Law: Derivation of Element “Stiffness Matrix”
93
3.2.5.3 Continuity Equation 95
3.2.5.4 Numerical Stability: Fluid Time-step 96
3.2.6 Optimisation of the Mechanical and Fluid Time steps 97
3.2.6.1 Manual Method 98
3.2.6.2 Automatic Method 98
3.2.7 Modified Cam-Clay Model 99
3.2.7.1 Virgin Consolidation Line and Swelling Lines 99
3.2.7.2 Yield and Potential Functions 100
3.2.7.3 Determination of the Input Parameters 101
3.3 Numerical Code Development 102
3.3.1 General 102
3.3.2 Numerical Code Structure 103
3.3.2.1 Input data: Geometry and properties of the Materials 104
3.3.2.2 Grid and Mesh Generation 106
3.3.2.3 Layering and Assigning Material Properties 106
3.3.2.4 Defining location of instrumentations and transducers 107
3.3.2.5 Boundary Conditions, Initial Stresses, and Undrained
Analysis 107
3.3.2.6 Simulation of Preconsolidation Stage 108
3.3.2.7 Vertical Drains and Smear Zone 109
3.3.2.8 Vacuum Pressure 110
3.3.2.9 Construction of Trial Embankment and Consolidation
Process 111
3.4 Systematic Back Calculation Procedure 111
3.5 Summary 115

X
4. LABORATORY STUDY TO INVESTIGATE THE SMEARING EFFECT
ON THE PERFORMANCE OF PVD ASSISTED PRELOADING
4.1 General 117
4.2 Testing Apparatus and Experimental Procedure 118
4.2.1 Apparatus 118
4.2.2 Material Properties 123
4.2.2.1 Soil samples 123
4.2.2.2 Consolidation test on reconstituted samples 126
4.2.3 Preparation of Rowe cell and initial sample 131
4.2.4 The pre-consolidation process and preparation of final sample 133
4.2.5 Initial drainage and de-airing of the Rowe cell system 136
4.2.6 Vertical drain assisted consolidation test procedure 138
4.3 Test Results 138
4.4 Verification of the Numerical Code 144
4.5 Evaluation of Axi-symmetric to Plane-Strain Conversion Methods 149
4.6 Summary 154

5. CASE STUDIES AND FURTHER VALIDATION EXERCISES

5.1 General 156
5.2 Cumbalum Trial Embankment 157
5.2.1 Introduction 157
5.2.2 Geological Model and Subsoil Condition 159
5.2.3 Installation of Vertical Drains and Embankment Construction 161
5.2.4 Numerical Simulation 163
5.3 Case Study 2: Ballina Bypass Trial Embankment 167
5.3.1 Introduction 167
5.3.2 Geological Model and Subsoil Condition 168
5.3.3 Vacuum Consolidation and Embankment Construction 169
5.3.4 Numerical Simulation 171
5.4 Case Study 3: Sunshine Motorway Trail Embankment 176
5.4.1 Introduction 176
5.4.2 Subsoil Condition 177
5.4.3 Installation of Vertical Drains and Embankment Construction 178

XI
 5.4.4 Numerical Simulation 179
5.5 Chittagong Airport Trial Embankment 182
5.5.1 Introduction 182
5.5.2 Subsurface Conditions 183
5.2.3 Soil Properties 185
5.5.4 Installation of Vertical Drain and Embankment Construction 186
5.5.4.1 Monitoring of Settlement 189
5.5.5 Numerical Simulation 190
5.6 Case Study 5: Large-scale Consolidometer 193
5.6.1 Test Apparatus 193
5.6.2 Test Sample 195
5.6.3 Test Procedure 196
5.6.4 Verification of the Developed Numerical Code 197
5.7 Summary 200

6. DETERMINING THE MINIMUM PERIOD OF MONITORING FOR

PREDICTING SMEAR ZONE PROPERTIES OF A TRIAL
EMBANKMENT
6.1 General 203
6.2 Step I: Estimating the Primary Consolidation Settlement 205
6.3 Step II: Conducting Parametric Studies 206
6.3.1 Chittagong Sea Port Trial Embankment 206
6.3.2 Cumbalum Trial Embankment 213
6.3.3 Ballina Bypass trial Embankment 216
6.3.4 Sunshine Trial Embankment 219
6.3.5 Large Scale Consolidometer 220
6.4 Step III: Determining the Error 221
6.5 Step IV: Determining the Minimum Required Monitoring Time 224
6.6 Summary 228

7. CONCLUSIONS AND RECOMMENDATIONS

7.1 Summary 230
7.2 Conclusions 231
7.3 Recommendations for Future Research 235
XII
REFERENCES 237
APPENDIX A: Developed FLAC Code to Simulate Trial Embankment 252
APPENDIX B: Developed FLAC Code to Simulate Rowe Cell Test 274

XIII
 LIST OF FIGURES

Figure 1.1 Procedure to select the appropriate technique for ground improvement
(after Arulrajah et al., 2003) 3
Figure 1.2 Typical consolidation settlement 4
Figure 1.3 Preloading method (a) without vertical drains, (b) with vertical drains 6
Figure 1.4 Prefabricated vertical drain, (a) Circular, (b) & (c) Band shape 6
Figure 1.5 Typical Effect of vertical drains on consolidation settlement rate of soft
clay 6
Figure 1.6 PVD installation (a) crane mounted installation rig, (b) drain delivery
arrangement, (c) cross section of mandrel and drain (after Koerner, 1987),
and (d) schematic installation process 7
Figure 1.7 3D schematic diagram of the installation of vertical drains in square patter
8
Figure 2.1Vacuum preloading system 17
Figure 2.2 Horizontal pipe used for vacuum preloading (a) corrugate flexible pipes,
(b) and (c) other types of geo-composites (after Chu et al. 2008) 18
Figure 2.3 Vacuum preloading-membrane free technique (a) PVD and tubing for
vacuum preloading, and (b) cross section of vacuum-PVDs method 19
Figure 2.4 Low level vacuum preloading with no membrane (after Chu et al. 2008)20
Figure 2.5 Equivalent diameter, (a) vertical band shaped drain, and (b) PVD
equivalent diameter 20
Figure 2.6 Schematic comparison of different PVD equivalent diameter calculation
approach 22
Figure 2.7 Examples of mandrel shapes, (a) rectangular, (b) rhombic, and (c) circular
27
Figure 2.8 PVD installation equipment, (a) crane and drain delivery arrangement,
and (b) vertical drain surrounded by hollow mandrel and attached to the
anchor plate at bottom 28
Figure 2.9 Influence zone of PVD, (a) square patter and (b) triangular pattern 29
Figure 2.10 PVD surrounding by smear zone (remoulding theory), (a) installed drain,
(b) profile A-A, and (c) cross section B-B 31
Figure 2.11 PVD surrounding by disturbed soil (reconsolidation theory), (a) installed
drain, (b) profile A-A, and (c) cross section B-B 32

XIV
Figure 2.12 PVD surrounded by smear zone, (a) for two zones hypothesis profile, (b)
for three zones hypothesis profile, (c) cross section A-A for two zones
hypothesis, and (d) cross section B-B for three zones hypothesis 33
Figure 2.13 Variation of permeability in the disturbed zone, (a) two zones hypothesis
and (b) three zones hypothesis 34
Figure 2.14 Large-scale consolidation apparatus (after Indraratna and Redana 1998)
36
Figure 2.15 Large scale consolidation apparatus (after Sharma and Xiao 2000) 37
Figure 2.16 Schematic diagram of sampling locations for the Oedometer test
specimen (after Sharma and Xiao 2000) 38
Figure 2.17 Radial penetration test, (a) Large-scale consolidometer, and (b) Micro-
cone penetrometer (after Shine et al. 2009) 39
Figure 2.18 Directions of MCPs and ERPs horizontal penetration (after Sine et al.
2009) 40
Figure 2.19 Schematic design, (a) Large consolidometer, and (b) radial positions
(plannar view) of fast response pore pressure transducers (Ts) relative to
the centre of the cell at levels identified in (a) (after Ghandeharioon et
al., 2012) 42
Figure 2.20 Expansion of a cavity (after Yu 2000) 44
Figure 2.21 Distribution pattern for the ratio of the plastic shear strain to the rigidity
index in relation to the radial distance normalized by the equivalent
elliptical radius of the mandrel characterising the disturbed soil
surrounding a PVD (after Ghandeharioon et al. 2010) 47
Figure 2.22 Mesh and boundary conditions for the large-scale test (after
Rujikiatkamjorn et al. 2009) 49
Figure 2.23 Proposed values for smear zone characteristics 53
Figure 2.24 Average degree of consolidation versus time factor based on Equation
2.9 (after Das, 2008) 56
Figure 2.25 Profile prefabricated vertical drain and smear zone, (a) axisymmetric, (b)
plane-strain 65
Figure 2.26 Proposed equivalent plane-strain unit cells, (a) Indraratna et al. (1997),
and (b) Tran and Mitachi (2008) 67
Figure 2.27 Partially penetrating drain based on DPM model (after Wang and Jiao,
2004) 70
Figure 2.28 Down drag effects due to mandrel installation in layered soil (after
Rujikiatkamjorn and Indraratna 2010) 71

XV
Figure 2.29 A schematic diagram of a vacuum preloading system: (a) vacuum and
surcharge combining load, (b) vacuum preloading, and (c) surcharge
preloading (after Mohamedelhassan and Shang 2002) 72
Figure 2.30 The distribution patterns of vacuum pressure in the horizontal and
vertical directions (after Indraratna et al. 2005a) 73
Figure 3.1 Basic explicit calculation cycle (after Itasca, 2008) 83
Figure 3.2 Finite difference mesh, (a) FLAC 2D zone composed of overlaid
triangular elements and (b) typical triangular element 86
Figure 3.3 Normal consolidation line swelling lines for an isotropic compression test
100
Figure 3.4 Yield surface of the Modified Cam-Clay model in p᦯-q plane (after
Roscoe and Borland, 1968) 101
Figure 3.5 Generated FLAC mesh, (a) sample of discretised finite-difference mesh of
trial embankment and (b) the pattern of meshes in the smear zone and
undisturbed region 103
Figure 3.6 Structure of the developed numerical code 104
Figure 3.7 Typical input parameters for numerical simulation
105
Figure 3.8 Input parameters in terms of different groups for numerical simulation107
Figure 3.9 Boundary conditions of the numerical model 108
Figure 3.10 Schematic model for the preconsolidation stage 109
Figure 3.11 Sample of the simulated vertical drain and adjacent disturbed area, (a)
Discretised finite-difference mesh of the trial embankment, (b) Variation
of the hydrostatic pressure along the vertical drain and (c) the pattern of
meshes in the smear zone and undisturbed region elements 110
Figure 3.12 Simulation of vacuum pressure, (a) variation of vacuum pressure along
the vertical drain, and (b) the pattern of meshes in the smear zone and
undisturbed region elements 111
Figure 3.13 Back calculation flowchart for smear zone characteristics and the
minimum required monitoring time for trial embankment 112
Figure 4.1 Large scale Rowe cell apparatus (a) schematic diagram of the cell and (b)
locations of the pore pressure transducers at the base of the cell 119
Figure 4.2 Pressure/volume controller device, (a) a photographic view of the GDS
controller and (b) an operational schematic diagram of the instrument 120
Figure 4.3 Infinite volume controller instrument 121
Figure 4.4 Schematic diagram of Rowe cell set-up 122

XVI
Figure 4.5 Established setup in the laboratory 123
Figure 4.6 Grain size distribution curve for vertical drain sand 125
Figure 4.7 Pre-consolidation process prior to the oedometer test; (a) cylinder
contacting reconstituted sample and (b) samples under pre-consolidation
pressure 127
Figure 4.8 Preparing the samples for the oedometer test, (a) placing the oedometer
ring, (b) cutting the extra top part, (c) cutting the extra bottom part, and
(d) the final sample 127
Figure 4.9 Consolidation test, (a) placing the prepared sample and (b) oedometer
apparatus connected to the data logger 128
Figure 4.10 Variation of permeability against void ratio (sample S1) 129
Figure 4.11 Variation of permeability against void ratio (sample S3) 129
Figure 4.12 Variation of void ratio versus effective stress (sample S1) 130
Figure 4.13 Variation of void ratio versus effective stress (sample S3) 130
Figure 4.14 Placing of PVC and brass pipes as the smear zone boundary and the
vertical drain border, (a) top view, (b) side view and (c) a typical cross
section of the Rowe cell 132
Figure 4.15 Sample placement, (a) filling the undisturbed area (intact zone) with the
prepared soil and (b) the setup after placing PVC and Brass pipes as the
smear zone boundary and vertical drain border 133
Figure 4.16 Rig set up, (a) geotextile filters, (b) pre-consolidation loading rings, (c)
the first two loading rings with drainage grooves and holes, (d) placing
of the first loading ring and (e) full loading condition 134
Figure 4.17 Testing procedures, (a) Pouring the vertical drain material and (b)
Pulling out the outer pipe 135
Figure 4.18 Testing procedures, (a) pulling out the inner pipe and (b) cutting the
extra part of the filter paper 135
Figure 4.19 Testing procedures, (a) leveling the top surface and (b) placing the
geotextile on top surface 136
Figure 4.20 Testing procedures, (a) filling the cell with water and (b) placing the cell
top 136
Figure 4.21 Schematic diagram of the de-airing process 137
Figure 4.22 Schematic diagram of the instrumentation plan, (a) plan view of the body
of Rowe cell and (b) the cross section of bottom of the Rowe cell 139
Figure 4.23 Settlement and corresponding surcharge versus consolidation time 140

XVII
Figure 4.24 Measured excess pore water pressure at transducers located on the
bottom of the cell (r is measured from the centre of the drain) 141
Figure 4.25 Measured excess pore water pressures from transducers located on the
sides of the cell (h is measured from the bottom of the impervious
boundary of the cell) 141
Figure 4.26 Variation of excess pore water pressures with the vertical distance from
the bottom of the impermeable boundary 142
Figure 4.27 Variation of excess pore water pressures with the radial distance from
the centre of the drain 143
Figure 4.28 Sample of the grid pattern for the simulated Rowe cell using the
developed code (rs= smear zone extent, and R= width of the cell), and
the condition of the pore water pressure distribution boundary along the
vertical drain. 145
Figure 4.29 Comparison between predicted numerical settlements and laboratory
measurements 146
Figure 4.30 Calculated final cumulative error for the selected cases by using the
proposed back calculation procedure 147
Figure 4.31 Variation of excess pore water pressures at PWPT A1 versus
consolidation time 148
Figure 4.32 Variation of excess pore water pressures at PWPT B3 versus
consolidation time 148
Figure 4.33 Variation of excess pore water pressure versus consolidation time for
different equations used to convert permeability from an axi-symmetric
to a plane-strain condition 152
Figure 4.34 Cumulative error between plane-strain and axisymmetric results in each
loading stage 152
Figure 4.35 Variation of predicted excess pore water pressure versus consolidation
time for the adopted equations for permeability conversion from
axisymmetric to plane-strain condition 153
Figure 5.1 A map of the Ballina Bypass upgrade route and surrounding surface
features (modified after Bishop, 2004) 158
Figure 5.2 Location of the Cumbalum trial embankment (courtesy of Google Maps)
159
Figure 5.3 Interpreted section for Cumbalum trial embankment study area (after
Bishop, 2004) 160
Figure 5.4 Moisture content, liquid limit and plastic limit (after RTA, 2000) 161

XVIII
Figure 5.5 Construction history of Cumbalum trial embankment (after Kelly 2008)
162
Figure 5.6 Layout of the Cumbalum trial embankment (after RTA, 2000) 162
Figure 5.7 Cross section of the Cumbalum trial embankment and subsoil profile 163
Figure 5.8 Finite-difference meshes for the plane-strain analysis of Cumbalum trial
embankment 164
Figure 5.9 Numerical parametric study results; Cumbalum trial embankment at SP9
165
Figure 5.10 Variation of excess pore pressure over time for PC2 at a depth 5.8m 166
Figure 5.11 Location of the critical section and trial embankment of the Ballina
Bypass (courtesy of Google Maps) 167
Figure 5.12 Interpreted section for Ballina Bypass trial embankment study area
(after Bishop 2004) 169
Figure 5.13 Layout of instrumentation for the trial embankment at Ballina Bypass
(modified after Kelly and Wong 2009) 170
Figure 5.14 Cross section of the Ballina Bypass trial embankment and subsoil
profile 170
Figure 5.15 Construction history of Ballina Bypass trial embankment 171
Figure 5.16 (a) Discretised finite-difference mesh of Ballina Bypass trial
embankment, (b) Variation of vacuum pressure along the vertical drain,
(c) The pattern of meshes in the smear zone and undisturbed region, (d)
FLAC zone composed of overlaid triangular elements 172
Figure 5.17 Results of numerical parametric study; Ballina Bypass trial
embankment at SP12 173
Figure 5.18 Variation of excess pore pressure over time for P3 at depth 11.8m 174
Figure 5.19 Location of the Sunshine Motorway study area, (courtesy of Google
Maps) 176
Figure 5.20 Profile of the Geotechnical characteristics (Sunshine Motorway Stage 2
Interim Report, 1992) 177
Figure 5.21 Plan view of Sunshine Motorway trial embankment 178
Figure 5.22 Typical cross-section of embankment with selected instrumentation
points (after Sathananthan et al. 2008) 179
Figure 5.23 Cross section of the Sunshine trial embankment and the subsoil profile
179
Figure 5.24 Sample of mesh grid pattern for an embankment considering the smear
zone and undisturbed region, (rs= smear zone extent; LI= intact zone

XIX
 extent; ht= depth of the soil profile; and hd= length of the vertical drain,
sd= drain spacing) 180
Figure 5.25 Construction history of the trial embankment 180
Figure 5.26 Comparison of numerical results with filed data (for settlement plate
P1) 182
Figure 5.27 Location of the container yard at Chittagong Port (courtesy of Google
Maps) 183
Figure 5.28 Approximate locations of the boreholes, Chittagong Airport trial
embankmnet (after Dhar et al., 2011) 184
Figure 5.29 General ground profile along with SPT N-values (after Dhar et al.,
2011) 185
Figure 5.30 Settlement monitoring program, (a) Settlement gauge and (b) Points of
settlement measurement (schematic plan) (after Dhar et al. 2011) 188
Figure 5.31 Schematic details of the ground improvement works (after Dhar et al.,
2011) 189
Figure 5.32 Variation of ground settlement with time 190
Figure 5.33 Cross section of constructed embankment at Chittagong Port site 190
Figure 5.34 Construction history (Chittagong Port embankment) 191
Figure 5.35 Sample of mesh grid pattern for Chittagong Port embankment
considering the smear 191
Figure 5.36 Comparison of numerical results with filed data at Chittagong Port site
192
Figure 5.37 Numerical variation of excess pore water pressure against consolidation
time 193
Figure 5.38 Large scale consolidometer (after Bamunawita 2004) 194
Figure 5.39 Cross section of the large consolidometer (modified after Indraratna et
al. 2004a) 195
Figure 5.40 Loading history for the large scale consolidometer 197
Figure 5.41 Sample of grid pattern for the large consolidometer applying developed
code (rs= smear zone extent; li= intact zone extent; sd= drain space, hc=
length of the cell, w= width of the cell), (a) boundary conditions applied
to the simulated cell and (b) pore pressure distribution boundary
condition along the vertical drain 198
Figure 5.42 Results of large consolidometer cell: Settlement versus consolidation
time 199

XX
 Figure 5.43 Results of large consolidometer cell: Excess pore pressure versus
consolidation time 199
Figure 6.1 Flowchart of the systematic procedure to determine the minimum degree
of consolidation resulting in an accurate estimation of smear zone
characteristics 204
Figure 6.2 Parametric study results for Chittagong port case history at point G1; (a)
rs/rm =2, (b) rs/rm =3, (c) rs/rm =4, and (d) rs/rm =5 208
Figure 6.3 Results of parametric study using the FLAC code developed to investigate
the influence of the smear zone properties on the dissipation of excess
pore water pressure for Chittagong port case history at point G2, (a)
rs/rm=2, (b) rs/rm =3, (c) rs/rm =4, and (d) rs/rm =5 210
Figure 6.4 Predicted time to obtain 90% degree of consolidation (Chittagong trial
embankment) 211
Figure 6.5 Results of FLAC analysis for points in Table 7 using the Chittagong port
case history, (a) Settlement variation, and (b) dissipation of excess pore
water pressure
212
Figure 6.6 Settlement variations against consolidation time for the Chittagong trial
embankment 213
Figure 6.7 Results of numerical parametric study: Cumbalum trial embankment at
SP9 214
Figure 6.8 Variation of excess pore pressure with time for the Cumbalum trial
embankment at PC2 (depth 5.8m) 215
Figure 6.9 Results of numerical parametric study: Ballina Bypass trial embankment
at SP12 216
Figure 6.10 Variation of excess pore pressure variation with time for Ballina Bypass
trial embankment at P3 (depth 11.8m) 217
Figure 6.11 Results of numerical parametric study: Sunshine trial embankment at P1
219
Figure 6.12 Results of numerical parametric study: Large scale consolidometer test
220
Figure 6.13 Normalised cumulative error versus degree of consolidation for different
smear zone properties, (a) Cumbalum trial embankment at SP9, (b)
Ballina Bypass trial embankment at SP12, (c) Sunshine trial
embankment at P1, (d) Chittagong Port trial embankment, and (e) large-
scale consolidometer 223
Figure 6.14 Total cumulative error (for the smear zone properties resulting in
minimum cumulative error) versus degree of consolidation, (a)
XXI
Cumbalum trial embankment at SP9, (b) Ballina Bypass trial
embankment at SP12, (c) Sunshine trial embankment at P1, and (d)
Chittagong Port trial embankment at G1, (e) Large-scale consolidometer
test 226

XXII
 LIST OF TABLES

Table 1.1 Applicability of ground improvement for different soil types (after Kamon
and Bergado, 1992) 2
Table 2.1 PVD assisted preloading projects 14
Table 2.2 Vacuum preloading projects 16
Table 2.3 Suggested equations for equivalent diameter of PVD 21
Table 2.4 Apparatus opening size requirements of PVD 23
Table 2.5 Suggested values for discharge capacity of PVD 26
Table 2.6 Back calculated smear zone properties (after Chai and Miura, 1999) 51
Table 2.7 Back calculated kh/ks ratio (reported by Saowapakpiboon et al., 2010) 52
Table 2.8 Proposed values for Cf (after Chai and Miura, 1999) 55
Table 2.9 Proposed solutions for radial consolidation considering constant smear
zone properties 59
Table 2.10 Proposed solutions for radial consolidation considering variable smear
zone properties 60
Table 2.11 Summary of conducted numerical studies to simulate PVD assisted
preloading process 76
Table 3.1 Comparison of explicit and implicit solution methods (after Itasca, 2008)84
Table 4.1 Properties of the adopted soil samples in this study 124
Table 4.2 Important sizes 125
Table 4.3 Mix design for the reconstituted samples 126
Table 4.4 Properties of the reconstituted samples 126
Table 4.5 Permeability of mixtures (Surcharge = 20 kPa) 129
Table 4.6 Properties of the intact zone, the smear zone, and drain 130
Table 4.7 Adopted properties for the numerical simulation 131
Table 4.8 Details of consolidation loading stages 138
Table 4.9 Surface settlement at the end of each loading stage 140
Table 4.10 The EPWP increase rate from the centre of the drain to the boundary of
smear zone for selected consolidation times 144
Table 4.11 Applied combinations of smear zone permeability and extent in numerical
analyses 146
Table 4.12 Adopted equations for converting permeability from axi-symmetric to
plane-strain condition 150
XXIII
Table 4.13 Adopted permeability coefficients for 2D analyses 151
Table 5.1 Summary of simulated case studies 157
Table 5.2 Adopted properties for subsoil layers for Cumbalum trial embankment near
SP9 164
Table 5.3 Adopted properties for the layers of subsoil for Ballina Bypass trial
embankment near SP12 173
Table 5.4 Adopted properties for the numerical simulation (after Sathananthan et al.
2008) 181
Table 5.5 Applied Properties for Sand Layer (after Sathananthan et al. 2008) 181
Table 5.6 Properties of the cohesive soil samples (Dhar et al., 2011) 186
Table 5.7 Properties of the applied PVD (Dhar et al., 2011) 187
Table 5.8 Adopted soil properties in FLAC simulation (after Dhar et al. 2011) 192
Table 5.9 Soil properties of the reconstituted sample of Moruya clay (after
Bamunawita 2004) 196
Table 5.10 Soil properties for Modified Cam-Clay (after Indraratna et al. 2004a) 198
Table 6.1 Primary consolidation settlements and the corresponding degree of
consolidations 205
Table 6.2 Back calculated smear zone properties to achieve t90%= 34 days
(Chittagong trial embankment) 212
Table 6.3 The final cumulative errors for different combinations of smear zone
properties 224
Table 6.4 Minimum required degree of consolidation and corresponding time
resulting in predicting reliable smear zone properties 227

XXIV
 LIST OF SYMBOLS

aᩳ initial radius of cavity

B equivalent plane-strain radius of the influence zone
BP back Pressure
bs equivalent plane-strain radius of the smear zone
bw equivalent plane-strain radius of the drain
Cc compression index
hydraulic conductivity ratio between the field and laboratory
Cf
values
CP cell Pressure
Cs swelling index
cƲ cohesion in drained condition
cf hydraulic conductivity ratio between field and laboratory values
ch coefficient of horizontal consolidation
ck permeability change index
cv vertical coefficient of consolidation
coefficient of consolidation for the combined vacuum and
cvc
surcharge preloading
D10 effective particle size
D15 grain diameter at 15% passing
D30 grain diameter at 30% passing
D50 grain diameter at 50% passing
D60 grain diameter at 60% passing
D85 grain diameter at 85% passing
DL data logger
dc diameter of unit cell
de equivalent drain diameter
dm mandrel diameter
ds smear zone diameter
dw diameter of drain
E Young’s modulus
Ef final cumulative error
(Ef)i cumulative error at step i

XXV
(Ef)s corresponding cumulative error to the extent ratio s
error between numerical predictions and field measurements at
Ei
step i
(Emin)s minimum error corresponding to step s
(Et)n normalised cumulative error at time t and step number n
(Et)n normalised cumulative error at time t and step number n
(EU%)i corresponding error to U% at step i
(EU%)min minimum error
(EU%)s corresponding error to the extent ratio s
EPWP excess pore water pressure
e void ratio
ecs void ratio at the critical state
eᩳ initial void ratio
Fk field measurements
Fs reduction factor
G shear modulus
Gs Specific garvity
gi or gk gravity vector
Hd vertical drainage length
Ir normalised by the rigidity index
IVC definite volume controller
Kw water bulk modulus
k coefficient of isotropic mobility
(kh/ks)i permeability ratio at step i
(kh/ks)max maximum permeability ratio
(kh/ks)min minimum permeability ratio
(kh/ks)ᩳ assumed initial permeability ratio
(kh/ks)opt predicted permeability ratio at U%min
permeability ratio corresponding to step s that gives the minimum
(kh/ks)s
error
݇෠(s) relative permeability
ratio between the vacuum pressure at the top and bottom of the
k1
drain
ke,ax equivalent horizontal permeability for the axi-symmetric unit cell

XXVI
ke,pl equivalent horizontal permeability for the plane-strain unit cell
horizontal permeability coefficient in the equivalent zone of plane-
kep
strain unit
kf permeability of the filter
kh intact zone horizontal permeability
kh/ks permeability ratio
khp horizontal permeability of intact zone in plane-strain condition
kij tensor of the coefficient of permeability
ks smear zone permeability
ksp horizontal permeability of smear zone in plane-strain condition
kv intact zone vertical permeability
kve equivalent coefficient of vertical permeability
kw coefficient of permeability of PVD
LL liquid limit
LVDT linear vertical displacement Transducer
l drainage length
lm the length of the vertical drain
[M] stiffness matrix
Mb biot modulus
M slope of the critical state line
MCC modified cam clay
N total number of observation points
n porosity
n= R/rw ratio of the influence radius to the drain radius
n= kh/ks permeability ratio
np isotropic over consolidation ratio
opening size of the filter which is larger than 50% of the fabric
O50
pore
O95 the apparent opening size of the filter
OCR over consolidated ratio
P/V C pressure/volume controller
PL plastic limit
PVD prefabricated vertical drain

XXVII
PWPT pore water Pressure Transducer
pԢ mean effective stress
PԢ1 reference pressure
PԢc preconsolidation pressure
PI plasticity index
Pk numerical predictions
pᩳ initial mean pressure
pԢᩳ initial effective mean pressure
Pvο vacuum pressure applied at the top of the drain
pԢyᩳ maximum isotropic preconsolidation stress
{Q} nodal flow rate
q deviator stress
qi specific discharge vector
qreq required discharge capacity
qv intensity of the volumetric water source
qw discharge capacity of PVD
qwp equivalent plane-strain discharge capacity
R radius of the influence zone
(rs/rm)i extent ratio at step i
(rs/rm)max maximum extent ratio
(rs/rm)min minimum extent ratio
(rs/rm)ᩳ assumed initial extent ratio
(rs/rm)opt predicted extent ratio at U%min
(rs/rm)s extent ratio corresponding to step s that gives the minimum error
r1 instantaneous radius of an elliptical cavity
rm mandrel radius
initial radius (in the direction of the semi-major axis) of an
rᩳ
elliptical cavity
rp radial distance of the plastic zone around the cavity
rs smear zone radius
rs/rm & rs/rw extent ratio
rtr radius of transition zone

XXVIII
rw drain radius
S drain spacing (centre to centre)
Sf final primary consolidation settlement
Spr field settlement at the end of preloading time
St field settlement at time t
Stp predicted settlement at time t
s extent ratio (rs/rw) or (rs/rm)
Thp radial consolidation time factor
Tr or Th radial consolidation time factor
Tv vertical consolidation time factor
Tvc time factor for the combined vacuum and surcharge preloading
t90% corresponding time at 90% degree of consolidation
tmin corresponding consolidation time to the U%min
tpr preloading time
ܷ total degree of consolidation in plane-strain condition
U% degree of consolidation at time t
U%min minimum required degree of consolidation
U%n degree of consolidation at step n
ܷh average degree of radial consolidation in axi-symmetric condition
ܷhp average degree of radial consolidation in plane-strain condition
Upr% degree of consolidation at the end of preloading
ܷv average degree of vertical consolidation
‫ݑ‬ᩳ initial excess pore water pressure
V total volume associated with the node
w water Content
z depth
Greek Letters
Ƚ biot’s coefficient
β pore pressure coefficient
ߛ௤௣ plastic shear strain
γs unit weight
ΔR permeability ratio incremental rate

XXIX
ΔS extent ratio incremental rate
equivalent increase in the nodal volume arising from
ΔVmech
mechanical deformation of the grid
Δσoct octahedral normal stress
Δτoct octahedral shear stress
Ԗ volumetric strain
Ƀ ariation of water volume per unit volume of porous material
Ʉ stress ratio
κ slope of the specific volume versus ln(p’) curve for swelling
Ȧ plastic volumetric strain ratio
λ slope of the specific volume versus ln(p’) curve for compression
Ɋ Poisson’s ratio
ɋ specific volume
ɋɉ specific volume at the reference pressure
ρd bulk density of the dry matrix
ɏ• density of the solid phase
ɏ™ mass density of the water
σc the lateral pressure
σᩳ initial cavity internal pressure
ɐr total stress
ɐrp total radial stress at the elastic-plastic boundary
φƲ friction angle in drained condition

XXX
 CHAPTER ONE

1 INTRODUCTION

1.1 GENERAL

According to the Australian Bureau of Statistics (2008), the population of Australia

would experience a growth of 60%, from 22 million to 35 million by 2056, and the
population of Queensland and New South Wales would boom by about 110% and
50% by 2056, respectively. This rapid growth in population and development in
urbanisation have increased the need to build and expand infrastructure such as
highways, railways, and ports. In Australia, the major cities lie along coastlines that
consist of deep soft clay deposits, especially in Northern Queensland and New South
Wales. These deposits can be categorised as problematic soils due to their
characteristic low bearing capacity and high compressibility.
Various ground improvement methods are available to improve the strength
and stiffness of soft soil, shorten construction time, and reduce costs by considering
the environmental aspects. To select the most appropriate method, various factors
such as the significance of the structure, the loading and site conditions, the
construction time limitation, and the type of soil, should be considered. Table 1.1
summarises the applicability of various ground improvement methods for different
types of soils, and Figure 1.1 shows the procedure for selecting an appropriate
ground improvement technique based on site conditions.
According to Table 1.1 and Figure 1.1, dewatering or preloading can be used
efficiently to strengthen the soft clay deposits and reduces their compressibility.
Preloading is a successful technique for stabilising soft clays. In this
method, a surcharge is applied onto the surface of the ground until most of the
ground settlement, expected to occur due to future structural loads, is achieved. In
soft and compressible soils that are relatively thick, consolidation takes a long time
to complete due to the very low permeability of the soil. Hence, preloading is not an
appropriate ground improvement method for projects with limited construction time.

1
Over the past decades, a system of prefabricated vertical drains has been used to
accelerate the consolidation process.

Table 1.1 Applicability of ground improvement for different soil types (after Kamon
and Bergado, 1992)

Improvement Admixture
Reinforcement Compacting Dewatering
Mechanism or grouting
Organic soil D D D
Volcanic clay
D D D
soil
Highly plastic
D D D
clay
Lowly plastic
D D D
clay
Silty soil D D D D

Sandy soil D D

Gravel soil D D
Interaction High density High density
Improved state
between soil Cementation by decreasing by decreasing
of soil
and inclusion void ratio void ratio

2
 Deep Ground
Improvement

Do you improve
YES NO
properties of
soft soil?
Structural
Can you deal Do you reinforce NO supporting e. g.,
YES
with the ground the soil? pile, retaining
water in soil? YES wall, etc.
NO
Can you
Do need the NO
decrease the Do you have NO permanent effect?
water level? plenty of time
for improving? YES
YES NO
YES Anchor or micro
pile method Do you use
Is there a risk Grouting grout?
Is it enough? Do you use NO
for liquefaction? method
admixture?
YES NO
YES NO YES NO YES

Vibrio floating Forced Preloading Preloading with Deep mixture

Compacted sand Grouting Freezing
method or dewatering without prefabricated method e. g., D.
pile method method method
gravel drain method e. g. prefabricated vertical drain M. M. jet
method well point, deep vertical drain method grouting method
well, etc. method

Figure 1.1 Procedure to select the appropriate technique for ground improvement

3
1.2 ACCELERATING THE CONSOLIDATION PROCESS

Soil may consist of two or three phases namely, solid, liquid, and gas. The voids
between the solid particles are filled with water, air, or a combination of both. Soil
deformation involves the reduction of voids under loads, compaction is the removal
of air from the soil, and consolidation is removal of water from the soil by applying a
surcharge. Figure 1.2 illustrates the settlement stages due to a reduction in the void
ratio during consolidation; this process includes, (i) initial compression, (ii) primary
consolidation, and (iii) secondary compression.

1 Initial
Compression
Settlement (mm)

2 Primary
Consolidation
3

4
Secondary
Compression
5
0.1 0 10 100 1000 10000
Time (min)

Figure 1.2 Typical consolidation settlement

Initial settlement in clayey soils occurs immediately after a load has been applied,
with no change in void ratio, so that only the shape changes. In saturated soils (i.e. no
free air in the voids), an increase in the stress induced by the external load is
immediately taken by water, which can be reasonably assumed to be incompressible.
The excess pore water pressure gradually dissipates as water seeps out of the soil
through drainage boundaries, and the pressure is transferred to the soil skeleton; this
is known as primary consolidation. Secondary compression might be determined as a
continuation of the mechanism of a change in volume following primary
consolidation (Mesri, 1973). This mechanism consists of the deformation of
individual particles and their relative movements due to normal stresses or shear
displacement where the particles contact each other, induced by shear stresses that
exceed the shear resistance.

4
The average degree of radial consolidation at time t can be calculated by using the
following equation (Hansbo 1981):

ഥ௛ ൌ ͳ െ ݁‫ ݌ݔ‬ቀି଼்೓ ቁ
ܷ (1.1)
ி

௖೓ ௧
ܶ௛ ൌ ோమ
(1.2)

ഥ௛ is the average degree of radial consolidation, Th is the radial drainage time

where, ܷ
factor, ch is the radial consolidation coefficient, R is the radius of the influence zone,
and F is a function related to factors such as the coefficient of permeability, the
drain spacing, the smear effect, and the well resistance.
It is seen that the time needed to attain a certain degree of consolidation is
governed by the permeability of the soil and the distance to the drainage boundaries.
There a two approaches that can be used to accelerate the rate of natural settlement;
(i) increasing the applied stress, or (ii) reducing the drainage length. Constructing a
higher embankment can increase the applied stress and accelerate the rate of
settlement, but in some cases, high embankments may cause stability problems. The
installation of prefabricated vertical drains (PVDs) is highly recommended to
accelerate the rate of radial drainage and consolidation by reducing the length of the
drainage paths (Figure 1.3). The consolidation time is inversely proportional to the
square of the length of the drainage path.
A prefabricated vertical drain (PVD) is a composite geosynthetic system
consisting of a polymeric inner core with formed flow path grooves on both sides
and an outer non-woven geotextile filter jacket (Figure 1.4). The main function of the
filter in a vertical drain is to ensure that fine particles cannot pass through and clog
the drainage channels in the core (Hansbo, 1981). Figure 1.5 shows how vertical
drains can accelerate the rate of settlement. As observed, PVD assisted preloading
reduces the time required to achieve a certain settlement compared to traditional
preloading (excluding PVD).

5
 Embankment
Embankment
Vertical drain

(a) (b)
Figure 1.3 Preloading method (a) without vertical drains, (b) with vertical drains

(a) (b) (c)

Figure 1.4 Prefabricated vertical drain, (a) Circular, (b) & (c) Band shape

500
450
Vertical Displacement (mm)

400
350
300
250
200
150 without drain

100
with drain
50
0
0 200 400 600 800 1000 1200 1400 1600 1800 2000
Time (Day)

Figure 1.5 Typical effect of vertical drains on consolidation settlement rate of soft clay

A system of vertical drains combined with vacuum pressure and surcharge

preloading can be used as an alternative method of ground improvement to increase
the settlement rate and reduce construction time. This technique accelerates

6
consolidation by applying negative pore water pressures along the vertical drains and
on the ground surface, which subsequently reduces the lateral pressure and increases
the effective stress. This technique reduces the height of the surcharge embankment
to prevent any instability and lateral movement in the soil.

1.3 INSTALLATION OF VERTICAL DRAINS AND SMEAR

ZONE

Prefabricated vertical drains are installed by either vibratory or static crowd methods.
In both cases, the drain is enclosed in a tubular steel mandrel with a small cross
sectional area, and then a small steel anchor plate is attached to the drain at the
bottom of the mandrel. The mandrel is then driven into the soil with a static crowd or
vibratory rig. When the design depth is reached, the mandrel is extracted and the
anchored plate holds the drain in the soil. Figure 1.6 shows the PVD installation
process.
Upper
main
roller
PVD
delivery Static
route guide
casing Steel
PVD on PVD on
Mandrel PVD mandrel
reel reel
Roller
Lower
main
roller

(a) (b) (c)

(d)
Figure 1.6 PVD installation (a) crane mounted installation rig, (b) drain delivery
arrangement, (c) cross section of mandrel and drain (after Koerner, 1987), and (d)
schematic installation process

7
Installing prefabricated vertical drains with a mandrel causes a significant
disturbance of the subsoil, especially in the area immediately surrounding the
mandrel. This disturbed zone is an area of reduced permeability that is called the
smear zone (Figure 1.7). Typically, two parameters should be addressed to
characterise this region, they include: (i) the permeability ratio, and (ii) the extent
ratio. The permeability ratio is the ratio between the permeability of the intact zone
and the permeability of the smear zone (kh/ks), and the extent ratio is equal to the
radius of the smear zone over the radius of the drain or mandrel (rs/rw or rs/rm). The
emerging smear zone and characteristics of the soil in this region can have a
significant influence over the consolidation process. Therefore, an accurate
estimation of the properties of the smear zone based on the type of soil and ground
conditions is vital for ground improvement projects when using the prefabricated
vertical drains assisted preloading method. Current literature indicates that further
research is needed to quantify the properties of the smear zone based on the
influencing factors, to help design engineers and clients optimise their designs and
minimise construction costs, respectively.

PVD S Smear
zone

rs

rw

Figure 1.7 3D schematic diagram of the installation of vertical drains in a square

patter

8
1.4 TRIAL EMBANKMENT MONITORING TO OBTAIN THE
SMEAR ZONE PROPERTIES

Simulating the PVD assisted preloading process for complex ground conditions in a
laboratory in order to estimate the characteristics of the smear zone is a challenging
task due to the disturbed nature of the soil in the lab and simplifying the testing
procedure. This means that a laboratory estimation of the properties of the smear
zone may not be reliable for practical design purposes. Furthermore, most of the
analytical solutions were developed on an assumption of a single axisymmetric
drainage system and therefore cannot be used to directly analyse the behaviour of a
soft soil deposit improved with multiple vertical drains. These limitations could well
be magnified when the subsurface soil consists of a multi-layer soil profile. Assumed
properties for the smear zone may result in inaccurate predictions of ground
behaviour, and lead to an early removal of the surcharge in the construction process
with subsequent excessive post construction settlements or construction time
(required time to complete the construction project), and increasing the project cost.
Field monitoring the actual preloading projects combined with a numerical analysis
gives an opportunity to investigate the consolidation behaviour of the soft soil and
back calculate the smear zone properties precisely.
The construction of a fully instrumented trial embankment has been used
extensively as a reliable method to determine the feasibility of preloading with
vertical drains, and to estimate the properties of the smear zone by applying a back
calculation procedure. However, trial embankments take a long time to construct,
which is the major challenge in using this method to estimate the extent and
permeability of the smear zone, and in many cases may cause a considerable delay in
the construction of the actual embankment and a significant increase in the project
cost. Estimating the extent and permeability of the smear zone in the early stages of
constructing a trial embankment can convert this method into a very practical,
accurate, and cost effective approach, so in this study, it is investigated in depth.

1.5 OBJECTIVES AND SCOPE OF PRESENT STUDY

The ultimate goals of this study are to propose a systematic back calculation
procedure to accurately estimate the properties of the smear zone and determine the

9
minimum degree of consolidation required, and consequently, the minimum time
needed to wait after constructing a trial embankment, to obtain reliable smear zone
properties. This research work consists of the following parts:

(i) Developing a numerical code and proposing a systematic back calculation

procedure to obtain the properties of the smear zone,

(ii) Conducting laboratory testing to validate the designed back calculation

procedure,

(iii) Simulating several case studies combined with parametric studies, and

(iv) Proposing an expanded back calculation procedure to determine the minimum

required degree of consolidation that results in predictable smear zone
properties.

FLAC software and its built-in programming language FISH have been used to
develop a numerical code to be integrated into the proposed systematic back
calculation procedure. A fully instrumented large Rowe cell has been used to conduct
the laboratory test. The intact and smear zones have been simulated by filling the cell
with clay that has a different permeability. A column of sand covered by geotextile
filter has been placed at the centre of the cell to act as a vertical drain. The validity of
the proposed back calculation procedure has been evaluated by comparing the
numerical results with the laboratory measurements.
Then, the numerical code has been used to simulate the selected case
histories and back calculate the smear zone properties using the proposed procedure.
The numerical results of the simulated case studies have been used to determine the
minimum degree of consolidation (after constructing a trial embankment) required,
and the corresponding time that resulted in accurate smear zone properties by using
an expanded back calculation procedure.

The specific objectives of this study are:

x Investigating the effects of the variability of smear zone characteristics on the

PVD assisted consolidation process and preloading design,

10
x Investigating the variations of pore water pressure with vertical and radial
distances from the drainage boundaries, and

x Evaluating how effectively the proposed equations are for converting the soil
permeability from an axisymmetric condition to a plane-strain condition.

A systematic parametric study has been conducted on several case histories using the
proposed back calculation procedure to investigate the variability of the smear zone
properties on the preloading process. The laboratory test process has been simulated
numerically in the axisymmetric condition, as well as the plane-strain condition, by
adopting the proposed axisymmetric to plane-strain conversion equations. The
numerical results for plane-strain and axisymmetric conditions have been compared
to evaluate the accuracy of the conversion methods now available.
Finally, a minimum preloading time has been proposed for practicing
engineers to predict the smear zone properties, which converts the construction of a
trial embankment into a very efficient and practical method.

1.6 ORGANISATION OF THE THESIS

In Chapter 1, a brief introduction is presented where the aim and scope of the present
research are highlighted. Chapter 2 presents a comprehensive survey of the literature
review associated with the present work. The history of vertical drains and factors
that affect the efficacy of consolidation by vertical drains are reviewed, with the
focus directed towards existing experimental, analytical, and numerical approaches
to estimate smear zone properties.
Chapter 3 presents the developed numerical code and the systematic back
calculation procedure adopted in this research. FLAC software (v6.0.287) and its
built-in programing language FISH are introduced for the development of numerical
code, and the governing equations and adopted soil constitutive model are described.
In Chapter 4, the procedure for a laboratory consolidation test is described
and the numerical code and back calculation procedure developed in Chapter 3 is
verified using the experimental results. Furthermore, an evaluation of the available
axisymmetric to plane-strain conversion equations is presented. A large Rowe cell
with a diameter of 250 mm and a height of 200 mm was used to conduct the PVD
assisted consolidation test procedure. The cell was fully instrumented and included a
11
vertical displacement gauge at the surface level and nine pore water pressure
transducers on the sides and at the base of the cell.
In Chapter 5, five case studies are numerically simulated, including four
trial embankments stabilised with vertical drains and one large consolidometer test.
The systematic back calculation procedure was used to predict the characteristics of
the smear zone. Chapter 6 presents the results of a parametric study where the
simulated case studies in Chapter 5 were used to investigate the effect of changes in
the property of the smear zone on the preloading process. Furthermore, the extended
back calculation procedure was used to determine the minimum waiting time
required after the construction of a trial embankment to reliably predict and report
the smear zone properties to be used by practicing geotechnical engineers.
Chapter 7 draws conclusions from the current research and provides
recommendations for future work. Following Chapter 7 are the list of references and
the appendices.

12
 CHAPTER TWO

2 LITERATURE REVIEW

2.1 HISTORY AND DEVELOPMENT OF VERTICAL DRAIN

ASSISTED PRELOADING

To improve soft ground, different types of vertical drainage systems (e.g. sand
compaction piles and prefabricated vertical drains) have been used extensively over
the past few decades. According to Johnson (1970), the application of a vertical
drainage system was first proposed around the 1920’s and was patented in 1926 by
Daniel J. Moran, an American engineer. Moran suggested the first practical
application of sand drains to stabilise the mud soil beneath the roadway approach to
the San Francisco Oakland Bay Bridge.
This solution led to comprehensive laboratory and field-testing by the
California Division of Highways in 1933 on the effectiveness of sand drains on the
rate of consolidation, and in 1936 Porter described some successful trials and
contributed to further studies and developments. Since World War Two, the
application of sand drains has undergone enormous development, largely due to
better methods of installation and greater knowledge of the principles controlling
their performance in different types of soft clays (Jamiolkowski et al., 1983). In
Japan, during the 1940’s, the behaviour of vertical sand drains was not understood
very well because although the bearing capacity of the foundation was considered to
be sufficient for a full load immediately after installation, frequent failures of
foundations had been reported (Aboshi, 1992).
Walter Kjellman installed the first prefabricated drain system in a field test
in 1937 using tubes made from a wood/fibre material, but after realising this material
was inappropriate and too expensive, Kjellman invented and patented a band shaped
cardboard drain in 1939, and a method for driving it into the ground. This cardboard
drain consisted of two cardboard sheets glued together with an external cross section
that was 100 mm wide by 3 mm thick, with ten 3-mm wide by 1-mm thick
longitudinal internal channels. The efficiency of cardboard wicks was first

13
investigated at Lilla Mellosa in Sweden, in a full-scale test, after which several types
of prefabricated band drains such as Geodrain (Sweden), Alidrain (England), and
Mebradrain (Netherlands), were developed. Basically, prefabricated vertical drains
(PVDs) have a rectangular cross section consisting of a filter fabric sleeve or jacket
surrounding a plastic core. The sleeve acts as a physical barrier separating the core
and the surrounding soil but permits pore water to enter the drain. It is made from
non-woven polyester geotextiles, polypropylene geotextiles, or synthetic papers. The
plastic core has grooved channels, which act as flow paths and supports for the filter
sleeve (Bergado et al., 1996).
In the last 20 years, the PVD assisted preloading has been widely used as an
efficient ground improvement technique. A selected number of successful PVD
assisted projects are summarised in Table 2.1.

Table 2.1 PVD assisted preloading projects

Project Location Reference

Ningbo international airport China Zhu et al. (1993)
Bhosle and Vaishampayan
Coal & iron ore stack yard India
(2009)
Changi east reclamation
Singapore Bo et al. (2007)
project
Sunshine motorway Australia Sathananthan et al. (2008)
Chittagong airport Bangladesh Dhar et al. (2011)
Sarapui trial embankmnet Brazil Almeida et al. (2005)
Barra da Tijuca embankmnet Brazil Almeida et al. (2005)
Cumbulum trial embankment Australia Kelly (2008)
Muar test embankments Malaysia Balasubramaniam et al. (2007)
Haarajoki trial embankment Finland Yildiz and Karstunen (2009)

2.2 VACUUM PRELOADING AIDED PREFABRICATED

VERTICAL DRAINS

Sometimes it is not feasible to build a high embankment over soft soil, as it might be
extremely weak so that even a common 1.5m high embankment might cause a
stability problem. Hence, it can be suitable to use vacuum preloading, because it will
rectify the stability problems due to the reduced embankment height. A prefabricated

14
vertical drain combined with vacuum pressure is used to further enhance the rate of
consolidation and shorten the construction periods (Chu et al., 2000; Chai et al.,
2005; Liu and Chu, 2009; Liu et al., 2009). It can be noted that the prefabricated
vertical drains used in vacuum preloading commonly have a circular cross section
that facilitates the transfer of vacuum pressure to deeper soil layers with greater
efficiency.

2.2.1 History and Developments of Vacuum Preloading

The concept of a vacuum preloading method (VPM) was first proposed by W.
Kjellman of the Swedish Geotechnical Institute in 1952 (Kjellman, 1952). Since
then, the vacuum preloading method has evolved into a mature and efficient
technique for treating soft clays. This method has been successfully used for soil
improvement or land reclamation projects in a number of countries (Bergado et al.,
1990; Chu et al., 2000; Cao et al., 2001; Indraratna et al., 2007a). With the merging
of new materials and new technologies, this method has been further improved in
recent years. In adopting this technique, sand drains and recently prefabricated
vertical drains have often been used to distribute the vacuum pressure and discharge
pore water. A nominal vacuum load of 80 kPa is normally used in design, although a
higher vacuum pressure of up to 90 kPa may be achieved. When a surcharge load
higher than 80 kPa is required, a combined vacuum and fill surcharge can be applied.
For the treatment of very soft ground, the vacuum preloading method is faster than
the fill surcharge method, because an 80-kPa vacuum pressure can be applied almost
instantly, without causing stability problems.
In terms of cost, the vacuum preloading is a very competitive option when the
required embankment fill is high requiring large berms or heavy reinforcement.
However, according to a comparison made by TPEI (1995), the cost of soil
improvement using vacuum preloading is only two-thirds of that by fill surcharge.
The vacuum preloading method has also been incorporated in the land reclamation
process when clay slurry dredged from the seabed is used as fill material for land
reclamation. As the dredged clay slurry is too soft to withstand the fill surcharge
placed on top, the vacuum preloading method can be ideally used to consolidate the
clay slurry. Table 2.2 shows a selected number of vacuum preloading projects
successfully completed in different countries.

15
 Table 2.2 Vacuum preloading projects
Project Location Reference
Tianjin Port China Yan & Chu (2005)
Survarnabhumi Airport Thailand Saowapakpiboon et al. (2009)
Port of Brisbane Australia Indraratna et al. (2011)
Ballina Bypass Highway Australia Kelly and Wong (2009)
Yaoqiang Airport China Tang and Sang (2000)
Kimhae Port Airport
South Korea Masse et al. (2002)
Highway
Second Bangkok
Thailand Indraratna et al. (2004a)
International Airport
Jang-U,
Sewage Disposal Plant Songa and Kim (2004)
Korea
Huanghua Port China Gao (2004)
Roadway embankment France Spaulding and Porbaha (2004)
Saga Road Construction Japan Chai et al. (2006)
Reclaimed land Japan Chai et al. (2008)
Maizuru Wakasa
Japan Kawaida et al. (2012)
Expressway
Nakorn Sri Thammarat
Thailand Teparaksa and Ngo (2012)
Airport

The Vacuum pressure technique has several advantages over the embankment
loading, and they are summarised below:

(i) Less fill material needs to be used,

(ii) Construction period is generally shorter,

(iii) No need for heavy machinery for heavy preloading,

(iv) No need to add any chemical admixtures into the ground, and

(v) Environmentally friendly ground improvement method.

Indraratna et al. (2010) reported that the carbon emissions from vacuum technology
with PVDs are much lower than the pile foundations because the production of
concrete and steel requires significant levels of energy. Two different methods are
proposed to conduct the vacuum preloading process; (i) Vacuum preloading using
membrane, and (ii) Membrane free technique.

16
2.2.2 Vacuum Preloading Using Membrane
In this method, vacuum consolidation consists of a system of vertical drains and a
drainage sand blanket on top sealed from the atmosphere by an impervious
membrane on top. Horizontal drains are installed in the drainage layer and are
connected to a vacuum pump. To maintain air tightness, the ends of the membrane
are placed at the bottom of a peripheral trench filled with bentonite. Negative
pressure is created in the drainage layer by the vacuum pump. The negative pressure
generates negative pore water pressures, resulting in an increase in the effective
stresses in the soil, which in turn leads to an accelerated consolidation process (Qian
et al, 2003). This system is illustrated in Figure 2.1.

Embankment
Possible geofabric layer
Protection
sand layer PVC membrane Vacuum pumping
station
Horizontal drains

Primary fill

Drainage layer

Peripheral trenches filled with

bentonite and polyacrylate

Sealing sheet pile

Vertical drains Soft soil deposit

Figure 2.1Vacuum preloading system

In recent years, new materials have been developed for horizontal drain pipes to
improve the vacuum preloading process (Figure 2.2). The drain panels can be used
instead of pipes to ensure the drainage channels will still function correctly under a
high surcharge pressure, as in the case of combined fill and vacuum preloading. The
drainage panels also provide better channels for distributing vacuum pressure and
water discharging. Some drainage panels also have slots for a direct connection with
PVDs, which also improves the efficiency of the system.

17
 2.1.1.1.1.1.1.1

(a) (b) (c)

Figure 2.2 Horizontal pipe used for vacuum preloading (a) corrugate flexible pipes,
(b) and (c) other types of geo-composites (after Chu et al. 2008)

2.2.3 Vacuum Preloading: Membrane Free Techniques

When the total area must be sub-divided into a number of sections to facilitate the
installation of the membrane, vacuum preloading can only be carried out in one
section at a time. This may not be efficient when vacuum preloading is used for land
reclamation over a large area. A common method to overcome this problem is to
connect the vacuum channel directly to each individual drain using a tubing system,
as shown in Figure 2.3a. In this way, the channel from the top of the PVD to the
vacuum line is sealed. Hence, a sand blanket and membranes are not required (Figure
2.3b). This system was used for constructing the new Bangkok International Airport
(Seah, 2006). However, as such a system does not provide an airtight condition for
the entire area, its efficiency can be low, and the vacuum pressure applied may only
be 50 kPa or lower (Seah, 2006). This method only works when the layer of soil to
be improved is predominantly clay with a very low permeability.

18
 Surcharge PVD-Tube Connector Vacuum
if required pump
To vacuum

PVD-Tube
PVD-Tube
To Next
Connector Connector
PVD
Vertical drain

PVD

(a) (b)

Figure 2.3 Vacuum preloading-membrane free technique (a) PVD and tubing for
vacuum preloading, and (b) cross section of vacuum-PVDs method

Another method to exclude the membrane is to use the so-called low level vacuum
preloading method (Yan and Cao, 2005). This method is schematically illustrated in
Figure 2.4. When clay slurry is used as fill for land reclamation, the vacuum pipes
can be installed at the seabed or a few metres below the ground surface, and the clay
slurry fill can be placed on top of the vacuum pipes. As clay has a low permeability,
the fill material will provide a good sealing cap and membranes will not be required.
However, this method is not free of problems. Tension cracks will develop in the top
layer when dried under the sunlight (Chu et al., 2008). The vacuum pressure may not
be distributed properly unless a drainage blanket is used at the level where the
drainage pipes are installed, or where the individual drains are directly connected to
the vacuum pipes. It is also difficult to install drainage pipes or panels underwater.
Nevertheless, this method does not require the construction of inner dikes for a sub-
division and thus cuts down the project costs substantially (Chu et al., 2008).

19
 Vacuum
Pump PVD Tube Dredge Clay Slurry Sand Blanket

2-3 m

Dike

PVD Tube

Figure 2.4 Low level vacuum preloading with no membrane (after Chu et al. 2008)

2.3 FACTORS AFFECTING CONSOLIDATION OF CLAY

WITH PVDS

The design of prefabricated vertical drains varies according to the specific

application. The key parameters and factors influencing the PVD assisted preloading
design are the equivalent diameter, the filter and apparent opening size, the tensile
strength, the discharge capacity and well resistance, smear zone, soil macro fibre,
mandrel size and shape, installation procedure, and the drain spacing and influence
zone, all of which are explained in detail in the following section.

2.3.1 Equivalent Diameter

In the equivalent two dimensional design, it is necessary to convert the band shaped
drain into an equivalent circular diameter called the “equivalent diameter,” as shown
in Figure 2.5. This dimension dictates the size of the inflow surface.

Polypropylene core

Geotextile filter de

a
b
(a) (b)

Figure 2.5 Equivalent diameter, (a) vertical band shaped drain, and (b) PVD
equivalent diameter

20
Different equations for the equivalent drain diameter (de) of band shaped PVD have
been proposed and are presented in Table 2.3. The dimensions a=100 mm and b=4
mm are considered to compare the results of different equations.

Table 2.3 Suggested equations for equivalent diameter of PVD

Equivalent
diameter
References Suggested equation
of drain (mm)
(PVD 100×4 mm)
2×(a+b)
Hansbo (1979) de = 66.2
π
a+b
Atkinson and Eldred (1981) de = 52.0
2

4ab 0.5
Fellenius and Castonguay (1985) de = ൬ ൰ 22.6
π

de =de -2ට൫sҧ2 ൯+b

Pradhan et al. (1993)
1 1 2a 44.1
sҧ2 = de 2 + a2 - 2 de
4 12 π
Long and Covo (1994) de =0.5a+0.7b 52.8
de =0.4546a
Abuel-Naga and Bouazza (2009) 0.0333൑b/a൑ 0.0875 45.4

Ideally, it is proper to adopt Hansbo’s method (1979, 1981), wherein the de value is
the diameter of an annulus having the equivalent circumference of a PVD. This
method gives rise to the largest value (6.62 cm) among the listed values, which is
exclusively used in practice. Holtz et al. (1991) reported that Hansbo’s method is
reasonable. However, the other methods present reduced values compared to the
above value, possibly due to the effectiveness of the size of the PVD.
A numerical study was performed by Abuel-Naga and Bouazza (2009) to
establish the equivalent diameter of a PVD well by considering the equal flow
condition. According to this research, for a rectangular section (0.0333 ≤ b/a ≤
0.0875), the equivalent diameter is only a function of the width of the PVD. The
equivalent diameter equation presented in this study is close to the equations
proposed by Atkinson and Eldred (1981), Pradhan et al. (1993), and Long and Covo
(1994), and is in agreement with the experimental results reported in literature. It is

21
therefore suggested that the proposed equation, together with Long and Covo’s
(1994) equation, can be used to evaluate the PVD equivalent PVD diameter. Figure
2.6 shows a schematic comparison of de derived from three different equations.

de=(a+b)/2
de=2×(a+b)/π

Band-Drain de=0.5a+0.7b

b
a
Pradhan et al.
de=(4 (a×b)/π) 0.5 (1993)

Figure 2.6 Schematic comparison of different PVD equivalent diameter calculation

approach

Welker and Herdin (2003) conducted five experiments with both an injection and an
extraction component to evaluate four of the most commonly used equivalent
diameter formulations for PVDs. The injection and extraction tests were performed
applying constant head and constant vacuum pressure, respectively. Based on the
results, there is little difference between the Hansbo (1979), Long and Covo (1994),
and Atkinson and Eldred’s (1981) formulations based on their corresponding
injection flow rates. All three of those equations had injection and extraction rates
that were similar to those obtained for the PVD both with and without considering
the effects of the smear zone. The Fellenius and Castonguay (1985) formulation held
the least correlation to the PVD under injection conditions.

2.3.2 Filter and Apparent Opening Size (AOS)

As explained by Indratana et al. (2002), the drain material (sand drain) and the filter
jacket of PVD have two basic but contrasting roles, which are: (i) retaining the soil
particles, and (ii) at the same time allowing the pore water to pass through. The
characteristics of the filter that affect the functionality of the vertical drain are
reported as follows:

22
(i) Permeability of the filter jacket,
(ii) Filtration (effective filtration can minimise movement of soil particles through
the filter),
(iii) Retention ability of the filter, and
(iv) Filter clogging (filter material may be clogged when the soil particles are
trapped within the filter fabric)
To cover the PVD functions, the apparent opening size (AOS) should satisfy the
proposed requirements given in Table 2.4. The parameter O95 indicates the
approximate largest particle that would effectively pass through the filter.

Table 2.4 Apparatus opening size requirements of PVD

Function Requirement Reference

Drainage kf > 2ksoil Bo et al. (2003)

O95
Filtration ≤2-3 Caroll (1983)
D85
O50
Retention ability ≤10-12 Caroll (1983)
D50
O95
Clog preventing ≥3 Christopher and Holtz, 1985
D15
Note: kf is the permeability of the filter and ksoil is the permeability of the soil, O95 is the
AOS of the filter, O50 is the size which is larger than 50% of the fabric pore, D85 is the size
for 85% of passing of the soil particles, D50 is the size for 50% of passing of the soil
particles, and D15 is the size for 15% of passing the soil particles.

Bo et al. (2003) reported that the permeability of the filter is normally required to be
at least one order of magnitude higher than the soil. Considering the clogging effect,
a much higher permeability should be required for the filter. Nevertheless, even more
stringent requirements on the permeability of the filter can be met easily because
most of the PVDs have a filter permeability that is higher than 10-4 m/s, which is far
more than what is required. The thickness of the filter is another consideration.
Normally, the thicker the filter, the better it becomes, while keeping the other
conditions the same. Based on Wang and Chen (1996), the mass to area ratio should
be generally larger than 90 g/m2.

2.3.3 Tensile Strengths

Prefabricated vertical drains should be strong enough to sustain the tensile stresses
subjected to it during installation, and therefore the strength of the core, the filter,
23
and the strength of the entire drain and the spliced drain should be specified in both
wet and dry conditions (Bo et al., 2003). Kremer et al. (1983) reported that a drain
should be able to withstand at least 0.5 kN of tensile force without exceeding 10% of
elongation. It is common nowadays to specify the tensile strength of the entire drain
at wet and dry conditions to be larger than 1 kN at a tensile strain of 10% (Bo et al.,
2003). A spliced vertical drain should also be required to have a tensile strength
comparable to an unspliced drain.

2.3.4 Discharge Capacity and Well Resistance

Prefabricated vertical drains are used to dissipate excess pore water pressure in the
soil and discharge water, and therefore the higher their discharge capacity the better
they perform. The discharge capacity of a band drain, denoted as qw, is defined as
the rate of flow through the drain at a hydraulic gradient of unity. Bergado et al.
(1996) reported that the discharge capacity can be expressed as the product of the
longitudinal permeability of the drain and its cross sectional area. The discharge
capacity of vertical drains can be affected by factors such as the consolidation stress,
deformation of the drain, time, clogging, and the hydraulic gradient and temperature.
When the discharge capacity of a vertical drain is smaller than the amount
of water that needs to be discharged, well resistance will occur. Ideally, the discharge
capacity of the drain should be large enough to ignore the well resistance in the
design because the well resistance increases as the length of the drain increases, and
reduces the rate of consolidation. Well resistance retards the dissipation of pore
pressure, which means that settlement is retarded. According to Bamunawita (2004),
the three main factors that increase the well resistance are as follows:

(i) Deterioration of the drain filter (reduction of the cross section of the drain),

(ii) Passing of fine soil particles through the filter (reduction of the cross section of
the drain), and

(iii) The drain folds due to large settlement or lateral movement.

Xie (1987) and Wang and Chen (1996) reported that the following condition must be
met in order to maintain the well resistance at an insignificant level:

24
గ ௞೓ ଶ
݈
ସ ௤ೢ ௠
൏ ͲǤͳ (2.1)

where, kh is the horizontal hydraulic conductivity of soil (m/s) and lm is the length of
the vertical drain (m).
To satisfy Equation (2.1), the discharge factor (DF) must be held by the
following equation:

௤ೢ
‫ܦ‬ி ൌ ௞ మ ൒ ͹Ǥͺͷ (2.2)
೓ ௟೘

The required discharge capacity after applying a reduction factor to consider all the
factors influencing the discharge capacity is reduced to:

‫ݍ‬௥௘௤ ൒ ͹Ǥͺͷ‫ܨ‬௦ ݇௛ ݈௠ ଶ (2.3)

where, qreq is the required discharge capacity (m3/s) and Fs is the reduction factor
with a value between 4 and 6.
Bo et al. (2003) states that it is unnecessary to use an excessively high
reduction factor for the discharge capacity because, although the discharge capacity
reduces with the deformation of the vertical drain and time, the permeability of the
soil reduces with consolidation, so the amount of water discharged also reduces over
time. Different studies have been carried out to determine the value of qw, and they
are summarised in Table 2.5.
It can be observed that the discharge capacity is affected by the lateral
confining pressure. According to Indraratna and Bamunawita (2002), in the absence
of laboratory test data, the discharge capacity can be conservatively assumed to be
100 m3/year. The results of the back analysis showed that the discharge capacity can
fall below this desired value due to the drain kinking and clogging (e.g. 25-100
m3/year) (Indraratna and Bamunawita, 2002). Clearly, the ‘clogged’ drains are
associated with qw values approaching zero.

25
 Table 2.5 Suggested values for discharge capacity of PVD
References Discharge capacity Remarks
3
Den Hoedt (1981) 95 Lab. Test (σc=50-300 kPa)
Kremer et al. (1982) 160 Lab. Test (σc=100 kPa)
Jamiolkowski et al. (1983) 10~15 Lab. Test (σc=300-500 kPa)
Oostveen (1986) 150 Back analysis
Rixner et al. (1986) 100 Lab. test
Hansbo (1987) 50~100 Lab. test
Koda et al. (1984) 100 Lab. test
Holtz et al. (1991) 100~150 Lab. Test (σc=300-500 kPa)
Mesri and Lo (1991) 2~80 Back analysis
Koerner & Ko (1992) 50~150 Back analysis
Indraratna (2002) 100 Back analysis
Bo (2004) 27~405 Back analysis
Kamon and Suwa (2006) 48~125 Lab. Test (σc= 300 kPa)
Note: σc is the lateral pressure

A new type of PVD, the so called integrated PVD where the core and filter adhered
together due to heat melting, was introduced by Liu and Chu (2009). The integrated
PVD offers a number of advantages over the separable PVD. Both the tensile
strength and the discharge capacity of the integrated PVD were higher than the
separable one made from the same material.

2.3.5 Smear Zone

Installing vertical drains with a mandrel causes a significant disturbance of the sub-
soil, especially near the mandrel. This disturbed zone is called the smear zone.
According to Sharma and Xiao (2000), the permeability of the soil around the drain
can be decreased significantly by smearing and retarding the rate of consolidation.
Two parameters are needed to characterise the smear effects, namely the smear zone
radius (rs), and the permeability of the smear zone. The smearing effect is
comprehensively discussed in Section 2.4.

26
2.3.6 Soil Macro Fibre
For soils with pronounced macro fabrics, the ratio of horizontal permeability to
vertical permeability (kh/kv) can be very high, whereas the kh/kv ratio becomes unity
within the disturbed (smear) zone. Vertical drains are very efficient where the layers
of clay contain many micro layers of horizontal sand or silt lenses. However, if these
micro layers are continuous in a horizontal direction, installing vertical drains may be
ineffective because rapid drainage of the pore water out of the layers of soil may
occur irrespective of whether drains are installed or not (Sharma and Xiao 2000).

2.3.7 Mandrel Size and Shape

Prefabricated vertical drains are installed by pushing a steel mandrel, approximately
5 inches (127 mm) by 2 inches (51 mm) in section, vertically into the ground. The
material of the PVDs is protected inside the mandrel and is affixed to an anchor plate
or bar at the base of the mandrel. Generally, any disturbance increases with the total
cross sectional area of the mandrel, which means the size of the mandrel should be as
close as possible to the drain to minimise the smearing effect. The common types of
mandrel cross sections are shown in Figure 2.7.

Mandrel Mandrel
Drain Drain

(b)
(a)
Mandrel
Drain

(c)

Figure 2.7 Examples of mandrel shapes, (a) rectangular, (b) rhombic and (c) circular

The effect of the shape of the mandrel was investigated analytically by Basu et. al
(2010) using four different size mandrels: 125 mm × 50 mm, 150 mm × 50 mm, 120

27
mm × 120 mm, and 150 mm × 150 mm. The results clearly showed that the rate of
consolidation decreased as the size of the mandrel increased. Furthermore, square
mandrels disturb a much larger area than rectangular mandrels, which means that
PVDs installed with square mandrels are less effective.

2.3.8 Installation Procedure

Prefabricated vertical drains should be installed using approved equipment which can
insert the material into the ground with the minimum possible disturbance to the soils
while keeping the mandrel vertical during installation. The PVD is forced into the
soil with the hollow mandrel, and once a firm layer is encountered or the target depth
is reached, the mandrel is retracted and the PVD, which is held in the ground by the
anchor, stays in place. Once the bottom of the mandrel reaches the surface, the PVD
is cut and the anchor for the next drain is attached. The mandrel is part of a mast
assembly, which is supported and powered by a hydraulic excavator or crane, and it
is then removed leaving the PVD in place. Figure 2.8 illustrates the installation rig.

Vertical drain
on reel

Vertical drain

Mandrel

Anchor plate

(a) (b)

Figure 2.8 PVD installation equipment, (a) crane and drain delivery arrangement,
and (b) vertical drain surrounded by hollow mandrel and attached to the anchor
plate at bottom

28
2.3.9 Drain Spacing and Influence Zone
Vertical drains may be installed at regular intervals in a square, rectangular, or
triangular pattern (Bergado et al. 1996), with centre to centre distance (S) varying
from about 1.0 to 3.5 m (Holtz 1987). As illustrated in Figure 2.9, the radius of the
influence zone (R) is a controlled variable because it is a function of the drain
spacing (S). The zone of influence can be determined using following equations:

R=0.546S ሺsquare patternሻ (2.4)

R=0.525S ሺtriangular patternሻ (2.5)

S S

S
S
R
R
Drains
R=0.546S R=0.525S
Square pattern Triangular pattern

Figure 2.9 Influence zone of PVD, (a) square patter and (b) triangular pattern

A square pattern of drains is often easier to lay out during field installation, although
a triangular pattern is usually preferred because it provides more uniform
consolidation between the drains (Indraratna and Bamunawita, 2002). Drain spacing
can be determined based on the required degree of consolidation and work timeframe
by adopting analytical solutions. According to the practical recommendations the
drain spacing should usually be more than 0.8 m.

2.4 SMEARING EFFECT

According to Sharma and Xiao (2000), the installation of prefabricated vertical

drains (PVDs) using a mandrel disturbs the soil around the drain, which results in a
smear zone of reduced permeability that adversely affects the consolidation process.
Predicting the behaviour of soil surrounding the drain requires an accurate estimation

29
of the smear zone properties. Generally, two major parameters are proposed to
characterise the smear zone, the permeability (ks), and the extent (rs) of the smear
zone.
Factors affecting the properties of the smear zone are explained by Bergado
et al. (1991) as follows:

(i) PVD installation procedure,

(ii) Specifications of the mandrel the type of anchor at the bottom of the PVD, and

(iii) Type of soil.

Hird and Moseley (2000) mentioned that the surface roughness of the mandrel and
the inward movement of the soil when the mandrel sheath is removed are the key
parameters affecting the characteristics of the smear zone. According to their study,
the smearing effects increase in severity as the thickness of the clay layer or the
driving rate is reduced. It was indicated that the angle of the tip of the mandrel does
not necessarily influence the severity of smearing in layered soils.
According to the experimental results reported by Hird and Sangtian (2002),
the smearing effects can be reduced substantially by changing the cross section of the
driving mandrel from a circular shape to a slim rectangular one.

2.4.1 Smear Zone Generation

Generally, two main reasons and theories have been proposed to explain the
generation of the smear zone: (i) the concept of soil remoulding, and (ii) the
reconsolidation theory.

2.4.1.1 Soil Remoulding Concept

When drain wells are installed by driving cased holes, which are back filled as the
casing is withdrawn, driving and pulling the casing would distort and remould the
adjacent soil (e.g. Barron 1948). According to the remoulding concept restated by
several other researchers (Barron 1948; Hansbo 1981; Indraratna and Redana 1997),
for PVD assisted preloading design, the soil surrounding the drain is divided into two
sections: (i) the disturbed region near the drain, and (ii) the intact or undisturbed zone

30
beyond the undisturbed zone. Figure 2.10 shows the smear zone near a vertical drain
based on the remoulding concept.

rw

kh
Intact zone Intact zone
ks
ks
Smear zone
kv rw kh
PVD kv

Smear zone rs

PVD R

rs R
(a) (b) (c)
Figure 2.10 PVD surrounding by smear zone (remoulding theory), (a) installed drain,
(b) profile A-A, and (c) cross section B-B
Note: rs is the radius of the smear zone, rw is the radius of the vertical drain, kh is the
horizontal permeability of the intact zone, kv is the vertical permeability of the intact zone, ks
is the permeability of the smear zone, and R is radius of drain influence zone.

2.4.1.2 The Reconsolidation Theory

Sharma and Xiao (2000) reported that the change in the properties of the clay layer
due to mandrel insertion is brought about by a combination of reconsolidation due to
the dissipation of excess pore pressure and remoulding of the clay layer due to the
shear stresses applied by the outer surface of the mandrel. According to the above
mentioned study, the process of reconsolidation has a much greater influence on the
properties of the clay layer than the process of remoulding, even for sensitive,
structured clays. The so-called smear zone around the vertical drains may be further
divided into two zones: (a) a remoulded zone of rather limited extent located close to
the drain, and (b) a reconsolidated zone of much wider extent located between the
remoulded zone and the intact clay (Figure 2.11).

31
 rw

kh
Intact zone
Intact zone
PVD
kv Reconsolidated ks
zone rw kh
PVD kv
Reconsolidated
zone
rs
Remoulded zone
Remoulded
R
zone
rs R
(a) (b) (c)
Figure 2.11 PVD surrounding by disturbed soil (reconsolidation theory), (a) installed
drain, (b) profile A-A, and (c) cross section B-B

In recent years, a few researchers have used the cavity expansion theory to analyse
the reconsolidation of soil associated with the installation of mandrel driven
prefabricated vertical drains with a mandrel (e.g. Sathananthan et al. 2008;
Ghandeharioon et al. 2010). Furthermore, Shin et al. (2009) used a micro-cone and
an electrical resistance probe to investigate the smearing effect before and after
reconsolidation.

2.4.2 Smear Zone Extent & Permeability Variation

The combined effects of permeability and compressibility within the smear zone
result in a different smeared soil behaviour compared to undisturbed soil. Predicting
the behaviour of soil surrounding the drain requires an accurate estimation of the
properties of the smear zone (Sathananthan, 2005).
There are two main schools of thought to determine the characteristics of
the soil surrounding the drain:

(a) Two zones hypothesis, which divides the surrounding soil into the smear zone
and undisturbed (intact) zone.

(b) Three zones hypothesis which considers three zones; the smear zone (inner
smear zone) in the immediate vicinity of the drain, the transition zone (outer
smear zone), and the undisturbed (intact) zone.

In the two zone hypothesis, two major parameters are proposed to characterise the
smear zone; the extent ratio (s) and the permeability ratio (n):

32
 ௥ ௥
‫ ݏ‬ൌ ௥ ೞ ‫ ݎ݋‬௥ ೞ (2.6)
೘ ೢ

௞೓
݊ൌ (2.7)
௞ೞ

where, rs is the radius of the smear zone, rm is the radius of the mandrel, rw is the
radius of the vertical drain, kh is the horizontal permeability of the intact zone, and ks
is the permeability of the smear zone. It should be noted that it is usually assumed
that the vertical and horizontal permeability values in the smear zone are the same
(Indraratna and Redana, 1998).
Figure 2.12 illustrates the cross section of the ground near a vertical drain
for both hypotheses.

rw rw R

Intact zone kh kh Intact zone

Transition zone
ks kv ks
Smear zone Smear zone
kv

PVD PVD

rs R rs rtr
(a) (b)

Intact zone
Intact zone

ks rtr Transition Zone

Smear zone rs
rw kh kh PVD
PVD kv kv rw
Smear zone
rs ks

R R

(c) (d)

Figure 2.12 PVD surrounded by smear zone, (a) for two zones hypothesis profile, (b)
for three zones hypothesis profile, (c) cross section A-A for two zones hypothesis,
and (d) cross section B-B for three zones hypothesis

According to the literature (e.g. Barron, 1948; Hansbo, 1981; Jamiolkowski et al.,
1983; Onoue et al., 1991; Madhav et al., 1993; Chai and Miura, 1999; Hawlader et
al., 2002; Walker and Indraratna, 2006; Basu et al., 2006; Rujikiatkamjorn and

33
Indraratna, 2009) different patterns are proposed for the permeability variation in the
disturbed zone (Figures 2.13a and 2.13b).

Intact Zone
Intact Zone
Transition Zone
Smear rtr Smear
Zone Zone
kh r s rw kh rw
kv kv rs
ks ks

Vertical
Drain
R Vertical
Drain R

k/kh k/kh
1 1
Case A Case D
F0 F0
k/kh rs r rs rtr r
k/kh
1 1
Case B F1
Case E
F0 F0
rs r rs rtr r
k/kh
k/kh
1
Case C 1
Case F F1
F0
rs F0
r
rs rtr r
(a) k/kh

1
Case G F1
F0
rs rp rtr r
(b)

Figure 2.13 Variation of permeability in the disturbed zone, (a) two zones
hypothesis and (b) three zones hypothesis

Following the two zones hypothesis, a number of researchers (e.g. Barron 1948;
Holtz and Holm 1973; Hansbo 1981; Jamiolkowski et al. 1983; Chai and Miura
1999) have proposed a constant smear zone permeability (Figure 2.13a, Case A).
Rujikiatkamjorn and Indraratna (2009) considered a linear variation for the
permeability of the smear zone with the radial distance (Figure 2.13a, Case B),
whereas Walker and Indraratna (2006) assumed a parabolic distribution for
permeability in the smear zone (Figure 2.13a, Case C).

34
Following the three zones hypothesis, Madhav et al. (1993) suggested that soil has a
constant permeability (ks) in the smear zone, while in the transition zone surrounding
the smear zone, the permeability linearly decreases from a value equal to ks at the
smear zone boundary to the initial in situ value kh at the boundary of the transition
zone (Figure 2.13b, Case D). Onoue et al. (1991) proposed two equations for the
variation of the permeability coefficients in the smear zone and transition zone
(Figure 2.13b, Case E). Hawlader et al. (2002) proposed a linear variation of
permeability within the smear zone changing from ks to kh and a constant
permeability of kh for the transition zone (Figure 2.13b, Case F). It was indicated that
the overall progress of consolidation is mainly controlled by the smear zone
surrounding the drains, while the transition zone has almost no influence. In addition,
Basu et al. (2006) reported that permeability remains constant at the value of ks
within the smear zone and increases in the transition zone following a bi-linear curve
with an initial slope from ks at the smear zone boundary to kp at an intermediate point
(at r=rp), and a secondary slope from kp (at r=rp) to kh at the transition zone boundary
(Figure 2.13b, Case G).

2.4.3 Estimation of the Smear Zone Properties

2.4.3.1 Experimental Methods

Estimating the radius and permeability of the smear zone has been the subject of
intense discussion in the geotechnical engineering industry and research. In many
classical theories the influence of the smear zone is described based on the two zone
hypotheses, i.e., an undisturbed zone with natural permeability and a smear zone
with reduced permeability. The accuracy of the predictions using classical theories
depends on a correct assessment of the horizontal permeability and extent of the
smear zone. Typically, the properties of a constant smear zone were assumed or back
calculated by geotechnical experts (e.g. Baron 1948; Casagrande and Poulos 1969;
Holtz and Holm 1973; Akagi 1976; Hanso 1981; Jamiolkowski et al. 1983; and
Bergado et al. 1991) to be used in practical designs, and no laboratory tests to
evaluate the characteristics of the smear zone were conducted until 1991.
Since 1991, comprehensive experimental studies have been conducted to
determine the extent and permeability of the smear zone (Bergado 1991; Onoue et al.

35
1991; Madhav et al. 1993; Indraratna and Redana 1998; Hird 2000; Sharma and Xiao
2000; Sathanantan and Indraratna 2006; Shin et al. 2009; Ghandeharioon 2009).
Indraratna and Redana (1998) investigated the effect of smearing due to the
installation of a vertical sand compaction pile in the laboratory, using a large scale
consolidometer, which is illustrated in Figure 2.14.

Pully

hoist Line Connecting

to Mandrel Volumetric Burette
LVDT
Axial
Displacement Pressure
Readout Unit gauge Loading Piston
Top Plate
Pore Pressure
Readout Unit

Pressure
Chamber
Moveable
Plate
Porous
material
Steel Cell (2 half cylinder)
Soil
Vertical 2 mm Teflon sheet
Drain Pressure gauge
Pressure regulator
Piezometer Release valve
Points
bottom
drainage line Mobile Compressor Unit

Figure 2.14 Large-scale consolidation apparatus (after Indraratna and Redana 1998)

The clay was thoroughly mixed with water before placing into the large cylinder, and
then the soil was placed inside and compacted in layers. A vertical drain was
installed using a purpose designed “pipe” mandrel and hoist (pulley system). The
mandrel was 50 mm in diameter by 2mm thick. After the sand was poured into the
pipe, it was gradually withdrawn with the hoist while the pipe was lightly vibrated
and the sand was compacted with an external rod.
After installing the sand compaction pile into the large cell, small specimens
were collected from different locations within the cell at known radial distances,

36
using a specially designed tube sampler. The samples were subjected to one-
dimensional consolidation using conventional oedometers in order to study the
variation in the properties of the soil close to and away from the central drain. It was
observed that the compressibility and permeability of the smear zone near the sand
drain would be quite different from the clay unaffected by the installation of the sand
compaction pile. The radius of the smear zone (rs) was estimated to be a factor of 4
to 5 times the radius of the mandrel (rm).
A large scale test apparatus was developed by Sharma and Xiao (2000) to
examine the extent and characteristics of the smear zone surrounding the
prefabricated vertical drains. The apparatus consists of three main components: a
consolidation tank, a motorised drain installation machine, and instrumentation. A
schematic diagram of the test apparatus is illustrated in Figure 2.15.

Clamp Nut

Screw
To computer Mandrel

Motro & Gearbox

Data Speed Controller
Logger Miniature PPTs

Lower Cylinder
0.4 m

20 mm

1.0 m
Clay

Base Plate
30 mm
Total Stress Cell

Figure 2.15 Large scale consolidation apparatus (after Sharma and Xiao 2000)

A motorised drain installation machine was developed to install the PVD by pushing
or pulling the mandrel at speeds ranging from 0.1 to 4 m/min. The miniature PPTs
were inserted from the base plate through the cable adapters and fixed at
predetermined radial and vertical distances using 5 mm diameter bronze rods.

37
A typical test was carried out in three stages: preparation of the clay deposit,
installation of the drain, and collection of the oedometer and moisture content
samples. After the excess pore water pressures induced by installing the PVD were
dissipated, the test was stopped and the entire clay deposit was removed from the
lower cylinder. Undisturbed vertical and horizontal samples were taken from the
middle section of the clay deposit at seven different locations, as shown in Figure
2.16. Oedometer tests were performed on these samples to evaluate the permeability
and compressibility of the clay deposit.

Drain
400 mm

Zone of Sampling
1000 mm
Smear Zone

Horizontal Specimen

Vertical Specimen

Miniature PPTs

Figure 2.16 Schematic diagram of sampling locations for the Oedometer test
specimen (after Sharma and Xiao 2000)

The test results indicated that the radial extent of the smear zone was 4 times greater
than the vertical drain radius. The average permeability of the clay in the smear zone
was 1.3 times smaller than in the intact zone.
Shin et al. (2009) conducted a series of radial penetration tests, including a
micro-cone penetrometer (MCP) and an electrical resistance probe (ERP), into the
specimen of reconstituted clay prepared in a large scale consolidometer to evaluate
the extent of the smear zone induced by the installation of a prefabricated vertical
drain with a rectangular mandrel and an anchor shoe. The schematics of large scale
consolidometer and micro-cone penetrometer are shown in Figure 2.17.

38
 To drainage line
Sand mat

Filter textile
Adapters for 0.4
penetrometer

1.0 m
Main body of 0.4
consolidometer

Filter textile 0.7 m

Sand mat Strain gauges

2.6

21.7
Friction sleeve
Hydraulic
Pressure line

Silicon seal
Cone tip
Bottom
drainage line (apex angle: 60°)
5.0
(a) (b)
Figure 2.17 Radial penetration test, (a) Large-scale consolidometer, and (b) Micro-
cone penetrometer (after Shine et al. 2009)

The main test procedure consisted of six stages; (i) preparing the clay specimen, (ii)
horizontal penetration tests of the micro-cone and ERP to obtain the reference
profiles of the resistance of the cone tip, and the electrical resistivity, (iii) installation
of the PVD, (iv) horizontal penetration tests of the micro-cone and ERP to detect the
smear zone, (v) a 7 day consolidation with PVD, and (vi) horizontal penetration tests
of the micro-cone and ERP to observe any changes in the smear zone. Both
penetration tests were conducted at 0.3 m below the top of the specimen. Each
penetration test was conducted at a rate of 1 mm/s to 300-340 mm from the chamber
wall. Figure 2.18 shows a plan view indicating the directions of each horizontal
penetration test.

39
 180°
EL
225°
CR 135°
Mandrel EI

270° 90°
ES CS

315° 45°
CI ER
0°
CL

Micro-cone Penetrometer Electrical Resistance Probe

(MCP) (ERP)
Reference CI EI
After PVD CL EL
installation CS ES
After re-consolidation
CR ER
with PVD

Figure 2.18 Directions of MCPs and ERPs horizontal penetration (after Sine et al.
2009)

As reported by Shine et al. (2009) after 32 days of consolidation under 200 kPa
pressure, horizontal penetration tests of the micro-cone and ERP were performed in
315° and 135° directions to obtain the reference profiles of the resistance of the cone
tip and the electrical resistivity (CI and EI, respectively). As there was no disturbed
zone in the specimen before the installation of PVD these profiles were expected to
be referred to when a comparison was made with those following the installation of
PVD. A modified double core PVD was driven into the soil specimen prepared in
the consolidometer with a rectangular steel mandrel. A rectangular shaped shoe was
attached to the end of the PVD and the mandrel was inserted to a depth of 0.6 m at a
constant penetration rate of 20 mm/s. Immediately after the PVD was installed,
micro-cone penetration tests were performed in 0° and 90° directions to determine
the extent of the smear zone (CL and CS, respectively).
While the PVD was being installed, an undrained condition was maintained
and excess pore water pressure built up in the smear zone (Onoue et al. 1991; Hird

40
and Moseley 2000; Sharma and Xiao 2000). It was therefore expected that the void
ratio or the electrical resistivity was the same throughout the specimen if the ERP
tests were carried out immediately after the installation of PVD. The ERP penetration
tests were performed in 180° and 270° directions (EL and ES, respectively) about 6 h
after the micro-cone penetration tests. After the clay specimen was re-consolidated
with the PVD for 7 days and the applied pressure of 200 kPa was maintained, both
penetration tests were performed in 225° and 45° directions (CR and ER,
respectively).
According to the continuous profiles of cone resistance and electrical
resistivity obtained from the penetrations of the micro-cone and ERP (Shin et al.,
2009), in the intact zone where the specimen was not affected by the installation of
PVD, both measurements reached constant values once the boundary effect of the
cylinder wall had disappeared. The resistance of the cone tip and electrical resistivity
started to deviate from constant values at certain locations, suspected to be the outer
boundary of the smear zone, and the magnitudes of the deviation continued to
increase towards the area where the PVD was installed. After allowing 7 days of
consolidation with the PVD, the resistance of the cone tip within the smear zone
increased significantly, while a slight increase in the electrical resistivity occurred
throughout the whole specimen. It was found that the size of the smear zone was
approximately 3.2~3.4 times the half-length and 5.3~5.4 times the half-width of the
mandrel. This implied that the shape of the smear zone was elliptical. In other words
the smear zone was about 4.0~4.2 times that of the equivalent radius in the longer
axis of the mandrel and 3.3~3.4 times that of the equivalent radius in the shorter axis
of the mandrel.
The installation of mandrel driven prefabricated vertical drains was
simulated in large scale laboratory conditions by Ghandeharioon et al. (2012) (Figure
2.19). The laboratory experiments were carried out with a consolidometer
specifically designed for that purpose, and a machine capable of driving mandrels at
realistic rates. The variations of pore water pressure at different locations during the
installation of the mandrel driven PVD, and while the mandrel was being withdrawn,
were also investigated.

41
 664 mm
Top flange

390 mm Predetermined positions of pore

pressure transducers (T1-T11)
A A

260 mm 1070 mm
B B

Consolidometer
260 mm
Base
120 mm

440mm

(a)

Mandrel/PVD

T12 T3
T10 T2
T9
T11
T6
T7 T4 T1
T5
T8
Mandrel/PVD

Section B-B Section A-A

(b)

Figure 2.19 Schematic design, (a) Large consolidometer, and (b) radial positions
(plannar view) of fast response pore pressure transducers (Ts) relative to the centre
of the cell at levels identified in (a) (after Ghandeharioon et al., 2012)

The extent of the smear zone in the large-scale consolidometer was determined using
the results from moisture content tests on samples, which in relation to the installed
PVD, were cored from various locations along different polar axes. The results
verified the concept of an elliptical smear zone around drains driven by rectangular
mandrels, and showed that for a given type of soft soil subjected to a particular rate
of installation, the size of the smear zone around the mandrel decreased as the in situ
effective stresses increased. On the basis of these results, the average radius of the

42
smear zone around each PVD was predicted to be about 3.5 times the equivalent
radius of the mandrel.
According to the current literature, the proposed experimental values for the
smear zone permeability ratio (kh/ks) and extent ratio (rs/rm or rs/rw) are between 1.34
to 3 and 1.6 to 7, respectively. The properties of the smear zone obtained from
experimental studies and laboratory measurements were compiled and are
summarised in Figure 2.23 in section 2.4.3.4.

2.4.3.2 Cavity Expansion Theory

It can be observed that a wide range is proposed for the characteristics of the smear
zone based on laboratory investigations and no comprehensive method is
recommended to precisely predict their properties.
In recent years a few researchers have used the cavity expansion theory
(CET) to analyse the reconsolidation of soil associated with mandrel driven
prefabricated vertical drain installation (Sathananthan et al. 2008 and Ghandeharioon
et al. 2010). Using this theory, the installation process is modelled as the expansion of
a cavity with a final radius equal to that of the mandrel (rm). The shape of the mandrel
can be rectangular, circular, or rhombic; hence, an equivalent mandrel radius is
evaluated by equating the area between the assumed circular cross section and the
actual shape of the mandrel.
The analytical cavity expansion solutions capturing the critical state theories
were developed by Collins and Yu (1996), and comprehensively described by Yu
(2000). The expansion of a cavity is illustrated in Figure 2.20. This figure shows a
cavity with an initial radius a0 and an internal pressure σ0. The cavity expands to a
radius of a when the internal pressure increases from σ0 to σa, whereas an element
initially at a radial distance r0 from the centre of the cavity moves to a new radial
position r from the centre. The soil on the cavity wall will yield when the pressure in
the cavity is large enough, whereas further increases in this pressure will lead to the
formation of a plastic zone around the cavity. The radial distance of the plastic zone
around the cavity is denoted by rp, whereas the soil beyond this would remain in a
state of elastic equilibrium.

43
 σr

Plastic zone Elastic zone

Figure 2.20 Expansion of a cavity (after Yu 2000)

Developing an analytical framework to analyse cavity expansion can be based on the

assumption that soil obeys Hooke’s elasticity until the onset of yielding (elastic
zone), and the yielding of soil (plastic zone) is described by the MCC model
(Sathananthan et al., 2008).
According to Sathananthan et al. (2008), the stress ratio, ߟ ൌ ‫ݍ‬Ȁ‫݌‬ᇱ , and the
radial distance from the centre of cavity ሺ‫ݎ‬ሻ can be related by:

௥೘ మ ଶሺଵାఓሻ ఑ ఑ஃ
݈݊ ቀͳ െ ቁ ൌ െ ଷξଷሺଵିଶఓሻ ఔ ߟʹξ͵ జெ ݂ሺ‫ܯ‬ǡ ߟǡ ܴଵ ሻ (2.8)
௥మ

where

ଵ ሺெାఎሻ൫ଵିඥோభ ିଵ൯ ఎ
݂ሺ‫ܯ‬ǡ ߟǡ ܴଵ ሻ ൌ ଶ ݈݊ ൤ሺெିఎሻ൫ଵା ൨ െ ‫ି݊ܽݐ‬ଵ ቀெቁ ൅ ‫ି݊ܽݐ‬ଵ ൫ඥܴଵ െ ͳ൯
ඥோభ ିଵ൯

where, ߥ=specific volume; ߤ=Poisson’s ratio; ‫݌‬ᇱ =effective mean pressure;

q=deviator stress; Ȧ ൌ ͳ െ ߢȀߣ=plastic volumetric strain ratio; ߣ=gradient of the
virgin compression line on ߭ െ ݈݊‫݌‬ᇱ space; ߢ=slope of swelling line on ߭ െ ݈݊‫݌‬ᇱ
space; M=slope of the critical state line; and the isotropic over-consolidation ratio
(R1) can be related to the conventional over consolidation ratio (OCR) as follows:

ଷ൫ସହିଵଶெାெమ ൯ை஼ோ
ܴଵ ൌ ሺ଺ିெሻ൫଺ାெାଶሺ଺ିெሻை஼ோሺయಾȀలశಾሻ൯ (2.9)

44
The radius of the plastic zone (rp) could be determined by substituting ߟ ൌ ‫ܯ‬ξܴ െ ͳ
in Equation (2.8) and the stress ratio (ߟ) and the effective mean pressure (‫݌‬ᇱ ) in the
plastic zone can be calculated by the Equation (2.10) (Sathananthan et al. 2008).

ஃ
ᇱ ᇱ ோ
‫ ݌‬ൌ ‫݌‬଴ ቈ ആ మ ቉ (2.10)
ଵାቀ ቁ
ಾ

where, ‫݌‬଴ᇱ =initial effective mean pressure.

A series representation of the effective mean pressure and the deviatoric
stress with radius can be found using Equation (2.8) and (2.10). As proposed by
Sathananthan et al. (2008), the variation of the pore water pressure (u) with radial
distance can be obtained from the Equation (2.11), when the cavity radius becomes
equal to the equivalent mandrel radius:

ெ௣బᇲ ௤ ଶ ௥೛ ௤
‫ ݑ‬ൌ ‫݌‬଴ ൅ ξܴ െ ͳ െ ൅ ‫׬‬ ݀‫ݎ‬ െ ‫݌‬ᇱ (2.11)
ξଷ ξଷ ξଷ ௥ ௥

where, ‫݌‬଴ =initial mean pressure.

The cavity expansion theory (CET) was used by Sathananthan et al. (2008)
by incorporating the Modified Cam-Clay model to predict the extent of the smear
zone (rm). It was assumed that within the smear zone the normalised pore pressure
u/σv0, (i.e., pore pressure/initial overburden stress) exceeded unity, but according to
the results, the size of the smear zone was at least 2.5 times the equivalent mandrel
radius and the permeability of the smear zone varied from 61% to 92% (average of
75%) of the intact zone permeability. The predicted radius of the smear zone was
verified using the large scale consolidometer. The test was conducted on the basis of
the development of excess pore water pressure while the mandrel was being
installed, the variation of normalised permeability (kh /khu) (khu is the maximum
undisturbed zone horizontal permeability), and a reduction of the normalised water
content. The results of this verification indicated there was a good agreement
between the analytical predictions and the laboratory measurements.
Ghandeharioon et al. (2010) used a new elliptical cavity expansion theory to
characterise the disturbed zone around rectangular and rhomboidal mandrels, while
incorporating a Modified Cam-Clay model for the soft soil. They assumed that the

45
installation rate of PVDs was more than 0.15 m/s so mandrel penetration was
considered as undrained, which may not be applicable in every case. Moreover, the
variation of shear strain and excess pore pressure were determined just after the
mandrel was inserted and the extent of the smear zone was estimated based on these
changes, even though in reality the properties of the smear zone should be defined
after reconsolidation and complete dissipation of any excess pore pressure. In this
elliptical solution, the changes in the ratio of the semi-major axis to the semi-minor
axis were not considered and a constant ratio was assumed.
Ghandeharioon et al. (2010) determined the excess pore water pressure at a
certain radial distance from the centre as follows:

ȟ‫ ݑ‬ൌ ȟ‫ ݌‬െ ȟ‫݌‬ᇱ (2.12)

Equations (2.13)-(2.15) were used by Ghandeharioon et al. (2010) to calculate the

total stress (ߪ௥ ሻ, effective stress (‫݌‬ᇱ ) and the deviator stress (q) at any soil element
inside the plastic region:

ଶ ௥ ௗ௥
ߪ௥ ൌ ߪ௥೛ െ ‫ݍ ׬‬ (2.13)
ξଷ ௥ ೛ ௥

௥భమ ି௥బమ ௤ ఑ஃ ఎ
݈݊ ቀͳ െ ቁൌ െ ʹξ͵ ఔெ ቂߞ െ ‫ି݊ܽݐ‬ଵ ቀெቁ ൅ ‫ି݊ܽݐ‬ଵ ඥ݊௣ െ ͳቃ (2.14)
௥మ ξଷீ

௣ᇲ ିଵȀஃ
‫ ݍ‬ൌ ‫݌ܯ‬ᇱ ඨ݊௣ ቀ௣ᇲ ቁ െͳ (2.15)
బ

ଵ ሺఎାெሻ൫ඥ௡೛ ିଵିଵ൯ ௤
where, ߞ ൌ ଶ ݈݊ ሺఎାெሻ൫ and ߟ ൌ ௣ᇲ .
ඥ௡೛ ିଵାଵ൯

In the above equations, ‫ݎ‬௣ is the initial radius of the plastic zone measured
from the centre of the cavity, ‫ݎ‬଴ is the initial radius (in the direction of the semi-
major axis) of an elliptical cavity, ‫ݎ‬ଵ is the instantaneous radius of an elliptical
cavity, ݊௣ ൌ ‫݌‬௬ᇱ బ Ȁ‫݌‬଴ᇱ is the isotropic over consolidation ratio (‫݌‬௬ᇱ బ is the maximum
isotropic preconsolidation stress),ߪ௥೛ is the total radial stress at the elastic-plastic
boundary (kPa), Ȧ ൌ ͳ െ ሺߢȀߣሻ is the plastic volumetric strain ratio,ߣ is the slope of
the normal compression line in the ߥ െ ݈݊‫݌‬ᇱ space, ߢ is slope of the elastic swelling

46
line in the ߥ െ ݈݊‫݌‬ᇱ space, ߥ is the soil specific volume (1+e), ‫ ܯ‬is the slope of the
critical state line in the ‫݌‬ᇱ െ ‫ ݍ‬plane, r is the instantaneous position of a soil element
measured from the centre of the cavity, and ‫݌‬଴ᇱ is the initial effective mean stress.
According to the results reported by Ghandeharioon et al. (2010), the
proposed relationship for determining the excess pore water pressure (Equation 2.12)
based on elliptical CET captures the actual position of the soil elements (polar
coordinates) and therefore provides a better match for the actual disturbance.
The plastic shear strain (ߛ௤௣ ) normalised by the rigidity index (‫ܫ‬௥ ) may also
be used to characterise the disturbed soil surrounding the mandrel (Figure 2.21).
Ghandeharioon et al. (2010) reported that ߛ௤௣ Ȁ‫ܫ‬௥ is as large as 0.86–1.05% at the
boundary of the failed soil (Figure 2.21). The smear zone propagates outward where
the normalised plastic shear strain within the smear zone is from 0.10 to 0.17%. At
the boundary of the marginally disturbed zone ߛ௤௣ Ȁ‫ܫ‬௥ is about 0.01– 0.05%. The
region in which ߛ௤௣ is greater than zero constitutes the plastic zone, while the outer
elastic zone (zero plastic shear strain) is not affected by any change in the pore water
pressure before or after driving the mandrel.

Undisturbed Zone
Smear Zone
Marginally
Cavity
Failed Zone Disturbed Zone

0.86% - 1.05%

0.10% - 0.17%
0.01% - 0.09%

Figure 2.21 Distribution pattern for the ratio of the plastic shear strain to the rigidity
index in relation to the radial distance normalized by the equivalent elliptical radius
of the mandrel characterising the disturbed soil surrounding a PVD (after
Ghandeharioon et al. 2010)

47
The proposed analytical solution by Ghandeharioon et al. (2010) was successfully
verified by using a case history from the Muar clay region in Malaysia. According to
the results of this verification, the radius of the smear zone was about 3.1 times the
equivalent radius of the mandrel, a value that is close to the radius suggested by
previous studies.

2.4.3.3 Finite Element Methods

The advancements in finite element codes have assisted geotechnical researchers to
simulate mandrel insertion numerically and investigate the load deformation during
installation of the mandrel by incorporating a large strain condition. This method is
typically used to simulate the installation of piles (Sheng et al. 2005; Walker and Yu,
2006).
Small et al. (1976) stated that the behaviour of saturated clays under rapid
loading occurred under fully undrained conditions and therefore the study of mandrel
penetration in saturated clays is essentially an undrained analysis. During
installation, excess pore water pressure can be induced as a result of a change in the
octahedral normal stress (Δσoct) and the octahedral shear stress (Δτoct) (Burns and
Maynes, 1998). While the volumetric strain remains constant during undrained
conditions, as does the octahedral normal stress, variations in the shear strain
adjacent to the side of the mandrel alters the octahedral shear stress. This zone of
disturbance depends on the size and shape of the mandrel, the properties of the soil,
the speed of installation, and the geometry of the cone (Kim et al. 2007).
Rujikiatkamjorn et al. (2009) used the ABAQUS finite element code for a
numerical simulation of mandrel penetration in normally consolidated soils. The
geometry, mesh, and boundary condition of the simulated large scale laboratory
model is illustrated in Figure 2.22. The Axisymmetric 4-node bi-linear displacement
and pore pressure elements (CAX4P), and the axisymmetric 4-node bi-linear
displacement elements (CAX4), were used to simulate the soil and the mandrel
respectively. A coupled analysis with large strain frictional contact was used to
simulate the soil-mandrel interface during mandrel penetration. The modified Cam
Clay model was used to represent the soft soil. The numerical predictions were then
compared with the excess pore pressures obtained from the large scale laboratory

48
tests to predict the extent of the smear zone. According to this research, the smear
zone was approximately 2.5 times the diameter of the mandrel.

k0p
1.0m

0.325m

Figure 2.22 Mesh and boundary conditions for the large-scale test (after
Rujikiatkamjorn et al. 2009)

In addition, ABAQUS software package was used by Ghandeharioon et al. (2012) to

numerically model the installation of mandrel driven prefabricated vertical drains. It
was assumed that the mandrel would be installed into a homogeneous soil. The
behaviour of saturated clays under rapid installation was analysed in undrained
conditions. It was also assumed that the elastic behaviour of saturated soft clays
could be described by Hook’s law of elasticity, and the Modified Cam-Clay model
would be suitable for studying the plastic behaviour of cohesive soils. The master-
slave concept was used to simulate driving a mandrel into soft clay, where the
periphery of the mandrel was specified as the master surface, and the top and lateral
surfaces of the soil (in an axisymmetric configuration) were identified as the slave
surfaces. This technique allowed the mandrel to develop a cavity and accommodate
its volume in the soil by displacing the soil during installation.
The numerical model showed how a smooth corner at the bottom of the
drain shoe, and the penalty parameter, affected the deformation, stresses, and the
constraints of physical contact within the soil. The variations of excess pore water
pressure at different locations during the installation of a mandrel driven PVD, and
while the mandrel was removed, were also investigated. According to the outcomes

49
of this verification, there was a good agreement between the pore pressures measured
in the laboratory and the finite element predictions.
According to Ghandeharioon et al. (2012), the laboratory observations
indicated that the soil near the mandrel moves radially and downward as the mandrel
is installed. The magnitude of displacement experienced by the soil surrounding the
mandrel justifies the large-strain formulation used for the frictional contact between
the mandrel and the soft clay. The extent of the smear zone in the large scale
consolidometer was determined using the results from the moisture content tests on
samples, which in relation to the installed PVDs, were cored from various locations
along different polar axes.
The results obtained by Ghandeharioon et al. (2012) verified the concept of
an elliptical smear zone around drains driven by rectangular mandrels, and showed
that for a given type of soft soil subjected to a particular rate of installation, the size
of the smear zone around the mandrel decreased as the effective in situ stresses
increased. Furthermore, the normalised plastic shear strain associated with the
boundary of the smear zone was almost constant for a given type of soil, regardless
of the effective in situ stresses under a specified rate of mandrel driving. The
numerical model of mandrel installation was applied to a case history from the
Second Bangkok International Airport. On the basis of these results, the average
radius of the smear zone around each PVD was predicted to be about 3.5 times the
equivalent radius of the mandrel, a value that is in agreement with the size suggested
in previous studies. It was shown that the smear zone expands when the over
consolidation ratio of the soil decreased.

2.4.3.4 Back Calculation Methods

The measurements obtained from the large scale consolidation tests and the
construction of trial embankments have been widely used to back calculate the
required parameters for the design of PVD assisted preloading, including the
characteristics of the smear zone. The back calculation procedure can be conducted
through a parametric study by applying numerical or analytical approaches. Different
values for the properties of the smear zone may be used to carry out the parametric
study, including the characteristics of the smear zone, which result in the best

50
agreement between the field measurements and numerical/analytical predictions, and
can be reported as reasonable values for design purposes.
Bergado et al. (1991) used the back calculation method based on the
observed time-settlement relationship of the large scale laboratory consolidation test
to determine the extent and permeability of the smear zone. According to the results
of the back analysis the diameter of the smear zone was 2 times the equivalent
diameter of the mandrel, and the permeability of the smear zone (ks) was
approximated to be equal to the vertical permeability coefficient of the soil in the
intact zone (kv).
Bergado et al. (1993) used the computer program VERDRN, which was
written in FORTRAN 77, to conduct numerical simulations of a large scale
consolidation test and a full scale test embankment, and then to back calculate the
properties of the smear zone. The laboratory settlement measurements were
compared with the finite element predictions, which indicated that adopting kh/ks=2
and rs/rm=2 resulted in the most accurate predictions. According to the numerical
results obtained for the full-scale test embankment, the performance of the vertical
drains was predicted quite well when the permeability and extent ratios were
considered as kh/ks=10 and rs/rm=2, respectively. In addition, the effect of the kh/ks
ratio on the settlement rate was more than the rs/rm ratio.
The trial embankment in Saga Airport (Japan) was simulated by Chai and
Miura (1999) using the finite element program CRISP (Britto and Gunn 1987) to
conduct a back analysis process and predict the characteristics of the smear zone.
Two different approaches were proposed to carry out the parametric study; (i)
considering the kh/ks and rs/rm as the constant values and back calculate qw
(discharge capacity), (ii) considering the rs/rm and qw as the constant values and back
calculate kh/ks. The results of the back calculation procedure conducted by Chai and
Miura (1999) are summarised in Table 2.6.

Table 2.6 Back calculated smear zone properties (after Chai and Miura, 1999)

Case kh/ks rs/rm qw (m3/yr) Description

1 10 3 85 Fix kh/ks and rs/rm, and then back calculate qw
2 11 3 150 Fix rs/rm and qw, and then back calculate kh/ks

51
Saowapakpiboon et al. (2010) conducted a large scale consolidometer test with
reconstituted specimens installed with prefabricated vertical drains with and without
vacuum preloading. Additionally, field measurements were collected from the
Second Bangkok International Airport site improved by PVD, with and without
vacuum pressures. The Hansbo (1979) method was applied to conduct the analytical
analyses for both cases and predict the consolidation settlement by adopting different
smear zone permeability coefficients. Also, the finite element software called
PVDCON was used by Saowapakpiboon et al. (2010) for numerical simulation of
PVD improved soft clay with and without vacuum preloading to back calculate the
smear zone permeability. Table 2.7 indicates the back calculated kh/ks ratios.

Table 2.7 Back calculated kh/ks ratio (reported by Saowapakpiboon et al., 2010)

Method of kh/ks kh/ks rs/rm

Measurements
simulation PVD only PVD-vacuum Constant
Large-scale
2.7 2.5 2
Analytical consolidometer
(Hansbo, 1979)
Trial embankment 7.2 6.6 2

Finite element Trial embankment

4-6 4-8 2
(PVDCON) OCR=1.2

In the current literature, the properties of the smear zone were back calculated by
comparing the field measurements and numerical results for at least 90% of
consolidation. In many cases, a long period of time is required to obtain 90% of
consolidation after constructing a trial embankment, which may increase the cost of
the project quite considerably. Determining the minimum degree of consolidation
that results in predicable smear zone properties can convert the construction of a trial
embankment to a very practical and cost effective method in design of PVD assisted
preloading projects.
Researchers have used different methods to investigate the characteristics of
the smear zone and very diverse values have been reported regarding the extent and
permeability of the smear zone. A summary of the proposed values for the kh/ks and
rs/rm ratios are illustrated in Figure 2.23.

52
 Extent ratio (rs /rm) Permeability ratio (kh /ks )
Barron (1948) Lower bound Barron (1948) Lower bound
Casagrande and Poulos (1969)
Holtz and Holm (1973) Upper bound Hansbo (1981) Upper bound
Akagi (1976) Hansbo et al.(1981)
Hansbo (1981)
Hansbo et al.(1981) Bergado et al. (1991)
Jamiolkowski et al. (1983)
Onoue et al. (1991)
Bergado et al. (1991)
Onoue et al. (1991) Bergado et al. (1993)
Bergado et al. (1993)
Almedia et al. (1993) Almedia et al. (1993)
Hansbo (1994) Hansbo (1997)
Mesri et al. (1994)
Hansbo (1997) Chai and Miura (1999)
(rs/rw )
Indraratna and Redana (1998) Eriksson et al. (2000)
Chai and Miura (1999)
Eriksson et al. (2000) Hird and Moseley (2000)
Sharma and Xiao (2000) Sharma and Xiao (2000)
(rs/rw )
Hird and Moseley (2000) (rs/rw )
Bo et al. (2003) Bo et al. (2003)
Indraratna et al. (2005b) (rs/rw ) Indraratna et al. (2005b)
Indraratna et al. (2005c)
Sathananthan and Inraratna (2006) Indraratna et al. (2005c)
Sathananthan et al. (2008)
Ghandeharioon et al. (2009) Sathananthan and Inraratna (2006)
Kim et al. (2010) Sathananthan et al. (2008)
Tran-Nguyen and Edil (2011)
Ghandeharioon et al. (2012) Tran-Nguyen and Edil (2011)

0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8 9 10 11

Applied method: Back-Analysis Experimental Analytical FEM Assumed

Figure 2.23 Proposed values for smear zone characteristics

Figure 2.23 indicates that the extent ratio (rs/rm) or (rs/rw) is generally assumed as 2
(Holtz and Holm, 1973; Akagi, 1976; Bergado et al., 1991; Mersi et al. 1994;
Hansbo 1997). According to some recent experimental results the smear zone radius
(rs) may be 4 times the radius of the mandrel (Sharma and Xiao, 2000; Sathananthan
et al. 2008). The proposed range shows that the extent of the smear zone (rs) may
vary from 1.6 to 7 times the radius of the drain (rw) or 1 to 6 times the equivalent
diameter of the mandrel (rm). The maximum values for rs/rw and rs/rm were proposed
by Bo et al. (2003) and Sathananthan (2008), respectively. The results reported in
Figure 2.23 indicate that the proposed range for the permeability ratio (kh/ks) is 1.03
to 10. The minimum and maximum values for kh/ks were reported by Tran-Nguyen
and Edil (2011) and Bo et al. (2003), respectively.
According to Figure 2.23, wide ranges were proposed for kh/ks and rs/rm and
there is no specific recommended method that practising engineers can use to predict
these parameters precisely. This can lead to an early removal of the surcharge in the
construction process and result in excessive post construction settlement or excessive
construction time, which would increase the project cost.

53
2.4.4 Relationship between Experimental and Practical Results for
Smear Zone Properties
There are many uncertainties regarding the kh/ks value. As Chai and Miura (1999)
and Shen et al. (2005) mentioned, the experimental results for the permeability of the
smear zone are underestimated values and a coefficient is required to convert these
values for practical design purposes. Chai and Miura (1999) suggested the following
equation as the relationship between the experimental and practical results for kh/ks:

kh k
= ቀ kh ቁ .Cf (2.16)
ks s l

where, the subscript l represents the value determined in the laboratory, and Cf is the
hydraulic conductivity ratio between the field and laboratory values. In some cases
(kh/ks) = (kh/kv), which means that the horizontal permeability in the smear zone is
equal to the vertical permeability in the intact zone. One way to estimate the value of
field hydraulic conductivity is to back analyse the local case histories, such as the
measured embankment settlement, or to measure hydraulic conductivity in the field
using a self-boring permeameter, if this sophisticated piece of equipment is available
(Tavenas et al. 1986). According Chai and Miura (1999), the most important factors
affecting the value of Cf are considered to be the deposit stratifications. For a
homogeneous deposit the Cf value can be close to 1.0, but for stratified deposits,
even those with thin layers of sand and seams of sand that cannot be clearly
identified from the borehole record, the Cf value can be much larger than 1.0. The Cf
values of a few clay deposits are listed in Table 2.8.

54
 Table 2.8 Proposed values for Cf (after Chai and Miura, 1999)
Method for
evaluating
Cf
Deposit field value of Remarks
Value
hydraulic
conductivity
Bangkok clay at Asian
Chai et al.
Institute of Technology 25 Back analysis
(1995)
campus
Bangkok clay at Nong Ngu Chai et al.
4 Back analysis
Hao (1996)
Chai and
Malaysia Muar clay deposit 2 Back analysis Bergado
(1993)
Chai and
Ariake clay (close to sea) 4 Back analysis
Miura (1999)
Self-boring Tavenas et al.
Louiseville (Canada) Around 1*
permeameter (1986)
Self-boring Tavenas et al.
St-Alban (Canada) Around 3*
permeameter (1986)
*The laboratory value was determined by direct measurements. For other cases, laboratory
values were deducted from the coefficient of consolidation, Cv.)

2.5 DEVELOPMENT OF CONSOLIDATION THEORY

2.5.1 Vertical Consolidation

The basic theory for vertical consolidation was proposed by Terzaghi (1925). This
theory is based on the following assumptions:
(i) Soil is homogeneous and fully saturated,

(ii) Pore water flow is solely in the vertical direction,

(iii) Pore water flow is governed by Darcy’s law,

(iv) Relationship between the void ratio and effective stress is linear,

(v) The permeability coefficient of the soil is constant during consolidation,

(vi) No creep occurs during soil settlement, and

(vii) The effect of geometric changes caused by the soil compressing is

insignificant (i. e. small strain theory)

ഥ௩ ) as a function of
Terzaghi (1925) presented the average degree of consolidation (ܷ
dimensionless time factor Tv:

55
 ଶ
ഥ௩ ൌ ͳ െ σ௠ୀஶ
ܷ ଶ
௠ୀ଴ ெమ ݁‫ ݌ݔ‬ሺ‫ܶ ܯ‬௩ ሻ (2.17)

where

గ
‫ ܯ‬ൌ ቀଶ ቁ ሺʹ݉ ൅ ͳሻ (2.18)

௖ೡ ௧
ܶ௩ ൌ (2.19)
ு೏మ

where,cv is the vertical coefficient of consolidation, and Hd is the vertical drainage

distance. The average degree of consolidation versus the time factor curves are
plotted in Figure 2.24 based on Equation (2.17) for different distributions of pore
water pressure while considering one way and two-way drainage boundary
conditions.

u=0

2d
20
Average Degree of Consolidation (%)

u=ui
C3
Open layers

40 u=0
C1
d

u=ui
60 Curve C1 Curve C3 Curve C2
Half-closed layers

80
C2

100
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
Time Factor Tv

Figure 2.24 Average degree of consolidation versus time factor based on Equation
2.9 (after Das, 2008)

According to Figure 2.24, the distribution of pore water pressure is constant in curve
C1, while the curves C2 and C3 represent the linear variation of pore water pressure.
In curve C2 the pore water pressure changes from zero at the top of the drainage
boundary to the maximum value at the bottom of the clay layer, while the condition

56
is vice versa in curve C3. Among the closed form approximations of Eq. (2.17), the
following equation provides almost the same results as Equation (2.17) for 0≤ Uv ≤
0.997, or 0 ≤ Tv ≤ 6, as reported by Sivaram and Swamee (1977).

ସ்ೡൗ బǤఱ
ቀ గቁ
ഥ௩ ൌ
ܷ బǤభళవ (2.20)
ସ்ೡൗ మǤఴ
൤ଵାቀ గቁ ൨

According to the literature, a number of solutions have been developed, relaxing

some of the assumptions in Terzaghi’s theory (1925). Morris (2005) assumed a finite
strain to develop analytical solutions for one dimensional consolidation. Finite
difference techniques with a piecewise linear formulation for large strain
consolidation were used by Fox and Berles (1997). Fox and Qui (2004) applied the
finite difference method to include compressible pore fluid. Xie and Leo (2004)
considered one dimensional large strain consolidation with variable compressibility
and permeability. According to Xie and Leo (2004), predicting small strain
settlement is more difficult than predicting large strains. Small strain pore pressure
and settlement evolution is slower than large strain evolution. Zhu and Yin (1998)
considered the consolidation of soil under depth dependant ramp load. Furthermore,
the consolidation of unsaturated ground was investigated by Vaziri and Christian
(1994).

2.5.2 Radial (or Horizontal) Consolidation Considering Smear

Zone Characteristics
Barron (1948) introduced the first classical solutions (rigorous and approximate) to
estimate the radial consolidation based on the Terzaghi (1925) consolidation theory.
This rigorous solution was developed based on ‘free strain hypothesis,’ and the
approximate solution was established based on ‘equal strain hypothesis.’ The
difference between free and equal strain cases was found to be negligible so an equal
strain case was used. Barron (1948) also included the effect of the smear zone
properties. The constant smear zone properties were considered following the two
zone hypothesis (Figure 2.9a, Case A). The solutions involved Bessel functions and
were time consuming to perform, so the effects of smear and well resistance were
often ignored to simplify the calculation (e.g. Hansbo, 1981).

57
 A number of solutions were developed to improve the Barron (1948) theory
for radial consolidation (Scott, 1963; Hansbo, 1981; Onoue, 1988; and Lo 1991). The
solutions proposed for radial consolidation are summarised in Table 2.9. The
equations proposed in Table 2.9 indicate that all the solutions can be expressed in a
general form as shown below:

ഥ௛ ൌ ͳ െ ݁‫ ݌ݔ‬ቀି଼்೓ ቁ
ܷ (2.21)
ி

where, Th=cht/S2 is the radial drainage time factor, and F is a function related to
factors such as drain spacing, the smear effect, and well resistance.
A rigorous approach was used in a number of studies to develop an
analytical solution for axisymmetric radial drainage that considers both the smear
effect and well resistance (Barron, 1948; Onoue, 1988; Zeng and Xie, 1989; Lo,
1991). As mentioned earlier, Hansbo (1981) developed an approximate equation for
horizontal drainage based on “equal strain hypothesis” to take both the smear zone
with a reduced permeability and the well resistance effect into consideration.
Function F in Hansbo’s (1981) equation that only considers the smear effect
(ignoring well resistance) is then reduced to:

n k 3
‫ ܨ‬ൌ lnሺ s ሻ+ kh ln(n)- 4 (2.22)
s

In the case of a perfect drain (neglecting the effects of both smear and well
resistance), the function F is as follows:

3
‫ ܨ‬ൌ ln(n)- (2.23)
4

Scott (1963) proposed an approximate solution for radial consolidation by

considering the smear effect and ignoring the well resistance factor. In the proposed
equation by Yoshikuni and Nakanodo (1974), only the well resistance factor was
considered. According to Table 2.9, Indraratna and Redanna (1997) developed an
equation to determine the average degree of horizontal consolidation for the plane-
strain condition by converting the radius of the smear zone and its permeability
(axisymmetric) into equivalent plane-strain parameters.

58
Table 2.9 Proposed solutions for radial consolidation considering constant smear
zone properties
Reference Suggested theoretical solution for radial consolidation
-8Tr n2 n 3n2 -s2 kh 3n2 -s2 n 3 kh
ഥ h =1-expሺ
U ሻ, F(n+S) = 2 ln ቀ ቁ - + × 2 lnሺsሻ≈ln ቀ ቁ - + lnሺsሻ
Barron F(n+s) n -1 s 4n2 ks n s 4 ks
2
(1948) 1 2ு೏
-8Tr expሾbሺz-2‫ܪ‬ௗ ሻሿ +exp(-bz) 2kh ൫n2 -s2 ൯ de
ഥ h=
U න ሺ1-expሺ ×fሺzሻdzሻሻ , F(z) = , b= ቈ × ቉ , R=
2‫ܪ‬ௗ 0 F(n+s) 1+exp(-2b‫ܪ‬ௗ ) kw R2 .F(n+s) 2

Scott -8Tr n2 3n2 -1 kh n2 -1 n2 3n2 -1 (s-1)kh n2 -1

ഥ h =1-expሺ
U ሻ, F(n+S) = 2 lnሺnሻ- + × 2 lnሺsሻ= 2 lnሺnሻ- + × 2
(1963) F(n+s) n -1 4n 2 rw kw n n -1 4n2 ks n

Yoshikuni Ǧͺ” Ǧͺ” ʹ ͵ʹ Ǧͳ Š ͵ʹ Ǧ• ʹ ͵

ഥ Š ൌͳǦ‡š’ሺ
 ሻൌͳǦ‡š’ሺ ሻǡ ሺሻ ൌ ŽሺሻǦ ൎ ൈ Žሺ•ሻൎŽሺሻǦ ,
& ሺሻ ൅ͲǤͺൈ ሺሻ ൅ʹǤ͸
ʹ Ǧͳ Ͷʹ • ʹ Ͷ
Nakanodo ଷଶ ௞೓ ௟ ଶ kh 1
‫ܮ‬ൌ ൈ ቀ ቁ ǡG= ሺ ሻ2
(1974) గమ ௞ೢ ௗ ೢ kw dw

-8Tr n2 n kh 3 s2 s2
ഥ h =1-expሺ
U ሻ F= 2 lnሺ + lnሺsሻ- ሻ+ 2 ሺ1- 2 ሻ+
Hansbo F n -1 s ks 4 n -1 4n
(1981) kh 1 s4 -1 2 kh 1 n kh 3 kh
. ሺ -s +1ሻ+πzሺ2l-zሻ ሺ1- 2 ሻ≈lnሺ ሻ+ ln(n)- +πzሺ2l-zሻ
ks n2 -1 4n2 qw n s ks 4 qw

-8Tr n2 n kh 3 s2 s2 kh 1 s4 -1
Rixner et ഥ h =1-expሺ
U ሻ F= 2 lnሺ + lnሺsሻ- ሻ+ 2 ሺ1- 2 ሻ+ . 2 ሺ 2 -s2 +1ሻ+
F n -1 s ks 4 n -1 4n ks n -1 4n
al.
2 kh .l2 1
(1986) π ሺ1- 2 ሻ
3 qw n
2 2
Onoue -8Tr ൫n' ൯ 3൫n' ൯ -1 ' kh 32 kh 1
ഥ h =1-expሺ
U ሻ, Fn' = ' 2 ×ln൫n' ൯- , n =n. sη-1 , η= , L* =× 2 × ሺ ሻ
(1988) Fn' +0.8×L* ሺn ሻ -1 4ሺn ሻ' 2 ks π kw dw
Xie
(1987) ∞
2 -8Tr -8Tr n kh 3 kh 1
Zeng & ഥ h =1- ෍
U exp ቆ ቇ =1- exp ቆ ቇ , Fሺn+Sሻ =ln ቀ ቁ + - and G= ሺ ሻ2
2 Fሺn+sሻ +݀݁ Fሺn+sሻ +πG s ks 4 kw dw
Xie m=0
(1989)
-8Tr n2 n kh 3 s2 s2 kh 1 s4 -1
Lo (1991) ഥ h =1-expሺ
U ሻ, F(n+S) = 2 lnሺ + lnሺsሻ- ሻ+ 2 ሺ1- 2 ሻ+ . 2 ሺ 2 -s2 +1ሻ
Fሺn+sሻ +2.5G n -1 s ks 4 n -1 4n ks n -1 4n

-8Thp kh’ ʹ ʹܾ௦ ܾ௦ ʹܾ௦ ଶ

Indraratna ഥ h’ =1-expሺ
U ሻǡ Ɋp ൌ ቈߙ ൅ ሺߚሻ ൅ ሺߠሻሺʹ݈‫ ݖ‬െ ‫ ݖ‬ଶ ሻ቉ ǡ ߙ ൌ െ ቆͳ െ ൅ ቇǡ
Ɋp k̵h’ ͵ ‫ܤ‬ ‫ܤ͵ ܤ‬ଶ
& Redana
௕ೞ మ ଶ௞೓೛ మ
(1997) ߚ ൌ
ଵ
ሺܾ௦ െ ܾ௪ ሻଶ ൅ ൫͵ܾ௪ ଶ െ ܾ௦ ଶ ൯ǡ ߠ ൌ ቀͳ െ
௕ೢ
ቁ ǡ ‫ݍ‬௭ ൌ
ଶ
‫ݍ‬
஻మ ଷ஻య k̵h’ ௤ೢ೛ ஻ ஻ గ஻ ௪

Note: ഥ ୦ and ഥ ୦୮ are the average degree of radial consolidation in axisymmetric and plane-
strain condition respectively; Tr is the radial consolidation time factor; Hd is the vertical
drainage distance; n is the ratio of the influence radius to the drain radius (R/rw); s is the
extent ratio (rs/rw); kw is the coefficient of permeability of PVD; qw is the discharge capacity
of PVD; z is depth; l is drainage length; khp and k'hp are horizontal permeability of intact zone
and smear zone in plane-strain condition, respectively; bw and bs are the equivalent plane-
strain radius of the drain and smear zone, respectively; B is the equivalent plane-strain radius
of the influence zone; and qwp is the equivalent plane-strain discharge capacity.

Walker and Indraratna (2006) proposed a vertical drain radial consolidation equation
based on the parabolic reduction of permeability in the smear zone towards the
vertical drain, thus modifying Hansbo’s (1981) well-known radial consolidation

59
equations (where a constant coefficient of consolidation is assumed), without
increasing the number of variables (Figure 2.9 case C). The validity of this method
was examined by comparing it with settlement data using a large scale
consolidometer.
Basu et al. (2006) developed analytical solutions to estimate the average
degree of consolidation for the cases based on the two & three zone hypotheses
(Figure 2.13b cases D, E, F, and G) that are shown in Table 2.10.

Table 2.10 Proposed solutions for radial consolidation considering variable smear
zone properties

Permeability variation Analytical solution for radial consolidation

Case
pattern (Basu et al. 2006)
k/kh
െͺܶ௥ ݊ ͵ ߢሺ‫ ݏ‬െ ͳሻଶ ‫ݏ‬
ഥ௛ ൌ ͳ െ ‡š’ ൬
ܷ ൰ ǡ ‫ ܨ‬ൌ Ž ቀ ቁ െ ൅ ଶ Ž ൬ ൰
1 ‫ܨ‬ ‫ݏ‬ Ͷ ሺ‫ ݏ‬െ ʹߢ‫ ݏ‬൅ ߢሻ ξߢ
C ‫ݏ‬ሺ‫ ݏ‬െ ͳሻඥߢሺ݇ െ ͳሻ ξߢ ൅ ξߢ െ ͳ ݇
F0
െ ݈݊ ቆ ቇ൅ ߨ‫ݖ‬ሺʹ݈ െ ‫ݖ‬ሻ
rs r ʹሺ‫ ݏ‬ଶ െ ʹߢ‫ ݏ‬൅ ߢሻ ξߢ െ ξߢ െ ͳ ‫ݍ‬௪

k/kh
െͺܶ௥ ݊ ͳ
1
ഥ௛ ൌ ͳ െ ‡š’ ൬
ܷ ൰ ǡ ‫ ܨ‬ൌ Ž ൬ ൰ ൅ Žሺ‫ݏ‬ሻ
‫ܨ‬ ‫ݍ‬ ߚ
D ሺ‫ ݍ‬െ ‫ݏ‬ሻ ߚ‫ݍ‬ ͵
F0 ൅ ݈݊ ൬ ൰ െ
rs rtr r ሺߚ‫ ݍ‬െ ‫ݏ‬ሻ ‫ݏ‬ Ͷ

k/kh
െͺܶ௥ ݊ ሺ‫ ݏ‬െ ͳሻ ߚ‫ݏ‬
1
ഥ௛ ൌ ͳ െ ‡š’ ൬
ܷ ൰ ǡ ‫ ܨ‬ൌ Ž ൬ ൰ ൅ Ž ൬ ൰
‫ܨ‬ ‫ݍ‬ ሺߚ‫ݏ‬ െ ߚ௧ ሻ ߚ௧
E F1
ሺ‫ ݍ‬െ ‫ݏ‬ሻ ߚ௧ ‫ݍ‬ ͵
F0 ൅ ݈݊ ൬ ൰ െ
rs rtr r ሺߚ௧ ‫ ݍ‬െ ‫ݏ‬ሻ ‫ݏ‬ Ͷ

k/kh

1 െͺܶ௥ ݊ ሺ‫ ݍ‬െ ͳሻ ͵
F F1 ഥ௛ ൌ ͳ െ ݁‫݌ݔ‬ሺ
ܷ ሻǡ ‫ ܨ‬ൌ ݈݊ሺ ሻ ൅ ݈݊ሺߚ‫ݍ‬ሻ െ
F0 ‫ܨ‬ ‫ݍ‬ ሺߚ‫ ݍ‬െ ͳሻ Ͷ
rs rtr r

k/kh
െͺܶ௥ ݊ ͳ
ഥ௛ ൌ ͳ െ ‡š’ ൬
ܷ ൰ ǡ ‫ ܨ‬ൌ Ž ൬ ൰ ൅ Žሺ‫ݏ‬ሻ
1 ‫ܨ‬ ‫ݍ‬ ߚ
G F1
ሺ‫ ݌‬െ ‫ݏ‬ሻ ߚ‫݌‬ ሺ‫ ݍ‬െ ‫݌‬ሻ ߚ௣ ‫ݍ‬ ͵
F0 ൅ ݈݊ ቆ ቇ൅ ݈݊ ቆ ቇെ
rs r p rtr r ൫ߚ‫ ݌‬െ ߚ௣ ‫ݏ‬൯ ߚ ௣ ‫ݏ‬ ൫ߚ௣ ‫ ݍ‬െ ‫݌‬൯ ‫݌‬ Ͷ

݇௛ ݇௦ ݇௧ ݇௣ ‫ݎ‬௧௥ ܴ ‫ݎ‬௦ ‫ݎ‬௣

ߢൌ ǡߚ ൌ ǡ ߚ௧ ൌ ǡ ߚ௣ ൌ ǡ ‫ ݍ‬ൌ ǡ ݊ ൌ ǡ ‫ ݏ‬ൌ ǡ ‫ ݌‬ൌ
݇௦ ݇௛ ݇௛ ݇௛ ‫ݎ‬௪ ‫ݎ‬௪ ‫ݎ‬௪ ‫ݎ‬௪

Note:  ഥ ୦ is the average degree of radial consolidation; Tr is the radial consolidation time
factor; R is the influence radius; rw is the drain radius; rs is the smear zone radius; rp is the
distance from the centre of the PVD to any intermediate point within the transition zone; rtr is
the radius of the transition zone; ks is the permeability at the drain boundary; kt is the
permeability at the smear zone boundary; kp is the permeability at any intermediate point
within the transition zone; kh is the permeability at the transition zone boundary or intact
zone permeability; l is the length of the drainage path; z is depth of horizontal plane, and qw
is the discharge capacity of the drain.

60
According to Basu et al. (2006), the transition zone has a definite impact in slowing
the consolidation process down. Moreover, the rate of consolidation depends a great
deal on the variations in hydraulic conductivity within the transition zone. According
to their study, an accurate definition of the most likely hydraulic conductivity profile
cannot be reached due to the limited amount of experimental data, so all possible
hydraulic conductivity profiles, as outlined in Figure 2.13, should be considered
before final design decisions are made.
A simple but practical procedure for designing prefabricated vertical drains
was proposed by Bellezza and Fentini (2008) which included the effects of smear
and well resistance. Based on Hansbo’s well-known solution, a simplified design
equation was obtained that can be effectively used for design purposes by
considering the drain spacing ratio (n) as an explicit function of the drain size, the
degree of consolidation required, the time available, the geotechnical properties of
the compressible layer, and the smear and well resistance parameters. According to
the procedure proposed by Bellezza and Fentini (2008), the drain spacing can be
designed using the following equations:

்ೢ
ߣൌ ഥሻሿ
ഥೡ ሻȀሺଵି௎
(2.24)
௟௡ሾሺଵି௎

ఒ ஻
݊ ൌ ቀ஺ቁ (2.25)

௘௫௣ሺିଷǤସହ௚೎ ሻ
‫ܣ‬ൌ ଶ௚೎
(2.26)

‫ ܤ‬ൌ ሺʹ ൅ ݃௖ ሻିଵ (2.27)

݃௖ ൌ ሺ‫ܨ‬ௌ ൅ ‫ܨ‬ோ ൅ ʹǤ͹ሻିଵ (2.28)

஼೓ ௧
where, ܶ௪ ൌ మ
௥ೢ
is the time factor in terms of the equivalent radius of the vertical

ഥ௩ is the average degree of vertical consolidation, ܷ

drain (rw), ܷ ഥ is the overall degree
of consolidation due to a combination of drainage in the horizontal and vertical
௞ ଶగ௞೓ಽమ
directions; n=R/rw is the drain spacing ratio, ‫ܨ‬ௌ ൌ ቀ ௞೓ െ ͳቁ ݈݊‫ ݏ‬and ‫ܨ‬ோ ൌ are
ೞ ଷ௤ೢ

dimensionless parameters.

61
The first stage of the above procedure is to calculate the parameter ߣ using Equation
(2.24). Then, Equations (2.25) to (2.28) can be used to determine the drain spacing
ratio (n=R/rw) and subsequently, the radius of the influence zone (R) can be obtained
for a given rw. The drain spacing can be determined as the final stage by considering
the installation pattern to be applied (i.e. square or triangular).
Chung et al. (2009) developed a method to predict the performance of
prefabricated vertical drains based on the Baron’s solution and the hyperbolic
settlement-time relationship. The method proposed by Chung et al. (2009) was
validated using three case studies. According to the analytical analyses, the
hyperbolic method developed is suitable when the degrees of consolidation are
between 60 and 90%.
Kianfar et al. (2013) proposed a non-linear relationship between the flow
velocity and the hydraulic gradient for the entire consolidation process, including the
smear effect. The proposed relationship was developed based on the laboratory
measurements and by conducting radial consolidation tests that were subjected to
vacuum and surcharge loading using a modified Rowe cell. Kianfar et al. (2013)
summarised the advantages of the developed flow relationship as follows: (i) it
provides more realistic flow behaviour during consolidation, and (ii) in contrast to
the relationship proposed by Hansbo (1960), the threshold hydraulic gradient was not
required to differentiate between the linear and non-linear flow relationships.

2.5.2.1 Conversion of Axisymmetric to Plane-Strain Condition

The axisymmetric classical solutions proposed by Barron (1948) and Hansbo (1981)
can be applied to analyse the performance of a vertical drain under a conventional
embankment (along the centreline of the embankment). Although single drain
analysis is often enough to model the behaviour of soil along the embankment
centreline, multi-drain analysis is needed to incorporate the effect of changing the
gravity load along the embankment to accurately predict settlement and lateral
displacement (Indraratna and Redana, 2000). Due to the rapid development of
computer technology in recent years, it is possible to conduct a numerical simulation
of the multiple vertical drains system under an embankment and analyse the
consolidation of the now stabilised soft soil using developed numerical programs.
For this purpose, a 3D full scale numerical model of the ground improved by PVD

62
should be simulated with a lot of cubic elements. Yildiz (2009) reported that a full
scale 3D analysis of an embankment on PVD improved soft clays is very time
consuming and not practical for engineers. The 2D model can be applied instead of
the 3D one to facilitate and accelerate the numerical simulation of the multiple
vertical drains or complex tree root system adopting the equivalent plane-strain
permeability for the intact region and the smear zone (Indraratna et al., 2005a; Fatahi
et al, 2009 and 2010).
Hird (1992) proposed three procedures based on Hansbo’s (1981) solution
to match the average degree of consolidation, at any time and depth, in the
axisymmetric and plane-strain unit cells, including; (i) geometric mapping – the
drain spacing is matched while maintaining the same permeability coefficient; (ii)
permeability mapping – the coefficient of permeability is matched while keeping the
same drain spacing; and (iii) a combination of (i) and (ii), with the plane-strain
permeability calculated for a convenient drain spacing. Yildiz et al. (2006) showed
that three proposed matching procedures by Hird et al. (1992) resulted in similar
settlement predictions. According to the combined procedure proposed by Hird et al.
(1992), the equivalent plane-strain permeability (khp) can be determined as follows:

௞೓೛ ଶ஻మ
ൌ ೃ ೖ య (2.29)
௞೓ ଷோ మ ቂ௟௡ቀ ቁାቀ ೓ ቁ௟௡ሺ௦ሻି ቃ
ೝೞ ೖೞ ర

where, B is the half-width of the unit cell, kh and khp are the axisymmetric and plane-
strain horizontal permeability values of the intact zone respectively, and ks is the
axisymmetric coefficient of permeability in the smear zone.
Equation (2.30) indicates the reduced form of Equation (2.29) by ignoring
the effects of smear and well resistance.

௞೓೛ ଴Ǥ଺଻
ൌ ሾ௟௡ሺ௡ሻି଴Ǥ଻ହሿ (2.30)
௞೓

Indraratna and Redana (1997) proposed a solution to estimate the degree of radial
consolidation for the plane-strain condition by considering the effects of smear and
well resistance. For this purpose, an analytical solution was developed to calculate
the equivalent plane-strain permeability for both the smear zone and undisturbed
soils. The relationship between the coefficient of permeability in the intact zone (khp)

63
and the smear zone (ksp) for the equivalent plane-strain condition can be determined
with Equation (2.31) by considering both the smear effect and well resistance.

ೖ೓೛
௞೓೛ ൤ఈାఉ ାఏ൫ଶ௟௭ି௭ మ ൯൨
ೖೞ೛
ൌ ೙ ೖ ೖ (2.31)
௞೓ ቂ௟௡ቀ ቁାቀ ೓ ቁ௟௡ሺ௦ሻି଴Ǥ଻ହగሺଶ௟௭ି௭ మ ሻ ೓ ቃ
ೞ ೖೞ ೜ೢ

Equation (2.31) can be simplified to Equation (2.30) by ignoring the smear effect and
the well resistance factor. If the effect of well resistance is ignored, the final terms in
Equation (2.31) can be neglected and the equation would be reduced to:

௞ೞ೛ ఉ
ൌ ೖ೓೛ ೖ
(2.32)
௞೓೛ ೙
ቂ௟௡ቀ ቁାቀ ೓ ቁ௟௡ሺ௦ሻି଴Ǥ଻ହቃିఈ
ೖ೓ ೞ ೖೞ

where

ଶ ଶ௕ೞ ௕ೞ ௕మ
ߙ ൌଷെ ቀͳ െ ൅ ଷ஻ೞమ ቁ (2.33)
஻ ஻

ଵ ௕ೞ
ߚ ൌ ஻మ ሺܾ௦ െ ܾ௪ ሻଶ ൅ ଶ
ሺ͵ܾ௪ െ ܾ௦ଶ ሻ (2.34)
ଷ஻య

మ
ଶ ௞೓೛ ௕ೢ
ߠൌ௞ ஻௤ೢ೛
ቀͳ െ ஻
ቁ (2.35)
ೞ೛

where, ksp is the plane-strain permeability coefficient of the smear zone, α, β, and θ
are geometric coefficients, n is the spacing ratio equal to R/rw=B/bw, s=rs/rw, qwp is
the equivalent plane-strain discharge capacity of the drain, z is the depth of the drain,
B, bs and bw are the drain dimensions (Figure 2.25). For plane-strain analysis, B, bs,
and bw are assumed to be equal to R (radius of the influence zone), rs (radius of the
smear zone), and rw (radius of the drain well), respectively. The value of kh needs to
be determined first (laboratory or field), then khp can be calculated using Equation
(2.30). When khp is known, ksp can be obtained from Equations (2.32). Figure 2.25
shows the axisymmetric and plane-strain profiles of the PVD and the surrounding
ground.

64
 rw bw

Smear zone
Intact zone
kh khp

Smear zone
Intact zone
ks ksp

kv kv

PVD PVD

rs R bs B
(a) (b)

Figure 2.25 Profile of prefabricated vertical drain and smear zone, (a) axisymmetric,
(b) plane-strain

Lin et al. (2000) developed analytical solutions by proposing equivalent horizontal

permeability coefficients for both axisymmetric and plane-strain unit cells to
consider the smear effect, which can then be adopted in finite element analyses. The
equivalent horizontal permeability for the axisymmetric unit cell (ke,ax) can be
determined by the following equation:

௞೓ ௟௡ሺ௡ሻ
݇௘ǡ௔௫ ൌ ೖ (2.36)
௟௡ሺ௡ሻାቀ ೓ ቁ௟௡ሺ௦ሻ
ೖೞ

The equivalent horizontal permeability for the plane-strain unit cell (ke,pl) can be
calculated by Equation 2.37:

௞೓ గ
݇௘ǡ௣௟ ൌ ೙ ೖ య (2.37)
଺ቂ௟௡ቀ ቁାቀ ೓ ቁ௟௡ሺ௦ሻି ቃ
ೞ ೖೞ ర

According to Chai et al. (2001), the installation of PVDs increases the mass
permeability of the improved subsoil in a vertical direction and an equivalent vertical
permeability can be established that represents the effect of both vertical drainage of
the natural subsoil and the radial flow due to the installation of vertical drains. The
proposed equivalent value of the vertical coefficient of permeability (kve) was derived
based on the equal average degree of consolidation under the 1D condition proposed
by Carrillo (1942), which can be expressed as:

65
 ଶǤହ௟మ ௞೓
݇ve ൌ ቀͳ ൅ ఋ ௗ೎ మ
ቁ ݇௛ (2.38)

where, kv is the vertical coefficient of permeability in the undisturbed soil, l is the

length of the drainage, and dc is the diameter of unit cell. The value of ߜ can be
expressed as

௡ ௞೓ ଷ ଶ௟మ ௞೓
ߜ ൌ ݈݊ ൅ Žሺ‫ݏ‬ሻ െ ൅ ߨ (2.39)
௦ ௞ೞ ସ ଷ௤ೢ

This method is more practical than the other methods since both the vertical drain
and the associated smear zone do not need to be represented in the equivalent plane-
strain analyses (Yildiz, 2009).
Indraratna et al. (2005a) derived permeability and vacuum pressure
relationships between the axisymmetric and equivalent plane-strain conditions by
extending the theory developed by Indraratna and Redana (1997), and assuming that
the average excess pore pressures for both axisymmetric and plane-strain conditions
are equal. According to Indraratna et al. (2005a), the equivalent permeability under
plane-strain condition can be determined using the following equations:

ೖ೓೛
௞೓೛ ൤ఈାఉ ାఏ൨
ೖೞ೛
ൌ ೙ ೖ ೖ (2.40)
௞೓ ቂ௟௡ቀ ቁାቀ ೓ ቁ௟௡ሺ௦ሻି଴Ǥ଻ହగሺଶ௟௭ି௭ మ ሻ ೓ ቃ
ೞ ೖೞ ೜ೢ

ଶ ሺ௡ି௦ሻయ
ߙൌ (2.41)
ଷ ௡మ ሺ௡ିଵሻ

ଶሺ௦ିଵሻ ଵ
ߚ ൌ ௡మ ሺ௡ିଵሻ ቂ݊ሺ݊ െ ‫ ݏ‬െ ͳሻ ൅ ଷ ሺ‫ ݏ‬ଶ ൅ ‫ ݏ‬൅ ͳሻቃ (2.42)

ସ௞ ଵ
ߠ ൌ ଷ஻௤೓೛ ቀͳ െ ௡ቁ ݈ ଶ (2.43)
ೢ೛

where, l is the length of the drain, and qw and qwp are the axisymmetric and plane-
strain discharge capacities, respectively.
Equation (2.44) shows the ratio of equivalent plane-strain permeability to
axisymmetric permeability in the intact zone by neglecting the effects of smear and
well resistance (Indraratna et al., 2005a).

66
 మሺ೙షభሻమ
௞೓೛ య ೙మ
ൌ ሾ௟௡ሺ௡ሻି଴Ǥ଻ହሿ (2.44)
௞೓

Equation (2.32) can be used the effect of well resistance is neglected.

Tran and Mitachi (2008) proposed a method to convert the axisymmetric
unit cell to an equivalent plane-strain unit cell under vacuum-surcharge preloading,
which can be expressed as:

ଶ஻మ ௞೓
݇௘௣ ൌ ଷோమ ೙ య ೖ೓ (2.45)
௟௡ ି ା ௟௡௦
ೞ ర ೖೞ

ଶ஻
‫ݍ‬௪௣ ൌ గோమ ‫ݍ‬௪ (2.46)

where, kep is the horizontal permeability coefficient in the equivalent zone of plane-
strain unit, and n=R/rw. The proposed equivalent plane-strain models by Indraratna et
al. (1997), and Tran and Mitachi (2008) are compared in Figure 2.26.

Plane strain undisturbed zone Equivalent zone

Drain wall
Drain wall
Plane strain smear zone

bw B
B bw
bs
(a) (b)

Figure 2.26 Proposed equivalent plane-strain unit cells, (a) Indraratna et al. (1997),
and (b) Tran and Mitachi (2008)

2.5.3 Combined Vertical and Radial Consolidation Theory

2.5.3.1 Single Layer Consolidation (Rigorous Solutions)

In vertical drain consolidation problems, the radial component of flow is often much
larger than the vertical component. Thus, consolidation due to vertical flow in the

67
soil is ignored in many cases, especially for long drains. Obviously, when vertical
drainage does become significant, it must be included in the analysis.
Yoshikuni and Nakanodo (1974) reported a solution to free strain
consolidation by combined vertical and radial drainage, including the effects of well
resistance. Zhu and Yin (2001) proposed design charts on the same problem under
ramp loading by ignoring well resistance. Both solutions are lengthy and involve
double summation series solutions. By using the separation of variables, the radial
drainage is solved with Bessel functions, while the vertical drainage is solved with
Fourier sine series. The coupled problem is simplified quite significantly if flow in
the vertical direction is assumed to occur due to the average hydraulic gradient across
a radial cross section. This approach was taken by Tang and Onitsuka (2000), who
produced a solution with a single Fourier series. Leo (2004) determined that a closed
form solution could be found to the equal strain problem by using modified Bessel
functions for both instantaneous and ramp loadings. The advantage of the proposed
closed form solution (similar to Terzaghi’s one dimensional consolidation solution)
is that each term in the series summation is a simple expression rather than the zeros
of a transcendental equation, as is the case for Yoshikuni and Nakanodo (1974) and
Zhu and Yin (2001).

2.5.3.2 Single Layer Consolidation (Approximate Solutions)

Researchers developed a number of approximate solutions due to the difficulty of
implementing rigorous solutions. The simplest approximate equation was proposed
by Carillo (1942), which relates the total degree of consolidation to the separately
considered radial and vertical degrees of consolidation as follows:

ഥ ൌ ሺͳ െ ܷ
ͳെܷ ഥ௩ ሻሺͳ െ ܷ
ഥ௥ ሻ (2.47)

ഥ is the average degree of consolidation of PVD improved subsoil, ܷ

where, ܷ ഥ௥ is the
ഥ௩ is the average degree
average degree of consolidation due to radial drainage, and ܷ
of consolidation due to vertical drainage. It should be noted that the above
relationship is only valid for homogeneous soil conditions.
Tang and Onitsuka (2000) proposed a solution for consolidation by vertical
drains with well resistance under a time dependent loading. It was indicated that
Carillo’s solution was not strictly applicable for ramped loading, even though the

68
discrepancy was small. Chai et al. (2001) developed a simple approximate method to
analyse soft soil improved by PVD in the same way as the unimproved case, by
proposing an equivalent vertical hydraulic conductivity which represents the effect of
both vertical and radial hydraulic conductivity, coefficients (Equations 2.33 and
2.34). According to Chai et al. (2001), the proposed method can be used for 1D, 2D,
and 3D analyses.

2.5.3.3 Multi-Layered Consolidation

As explained earlier, many analytical solutions have been developed for
consolidating homogeneous single layered soils with vertical drains. Usually, the
ground consists of several layers that must be considered in PVD consolidation
analyses to obtain more realistic results. A number of solutions were proposed to
predict the consolidation behaviour of layered soils. Chai et al. (2001) proposed a
simple approximate method for analysing PVD improved multi-layered subsoils by
considering an equivalent vertical hydraulic conductivity kve for the PVD improved
subsoil.
Tang and Onitsuka (2001) developed an analytical solution for the
consolidation of double-layered ground with vertical drains by considering the
effects of well resistance and smear. It was assumed that the layer of soil consisted of
two parts. The upper part of the layer with the vertical drains was considered to be
homogenous ground with both vertical and horizontal flow, while the lower part was
considered to be single layered ground with only vertical flow. In other words, in the
model proposed by Tang and Onitsuka (2001), horizontal flow in the lower part of
the soil was neglected.
An analytical solution for calculating the overall degree of consolidation for
partially penetrating vertical drains was developed after Tang and Onitsuka’s (2001)
work, by Wang and Jiao (2004) using a one dimensional double porosity model
(DPM). The partially penetrating drain was approximated by a fully penetrating drain
and two segments were considered; (i) the upper with a real drain, and (ii) the lower
one with an imaginary drain, with the same radius as the upper one and the same
properties as the undisturbed soil (Figure 2.27). In the study conducted by Wang and
Jiao (2004), horizontal flow in the lower part and the vertical compressibility of the
drain (neglected by Tang and Onitsuka, 2001) were considered.

69
 Upper soil

Drain wall

Seepage direction

Imaginary drain well

with soil property Lower soil

Figure 2.27 Partially penetrating drain based on DPM model (after Wang and Jiao,
2004)

Zhu and Yin (2005) developed design charts for vertical drainage with two layers.
Xie et al. (1999) solved the same problem with partially drained boundaries, while
Xie et al. (2002) incorporated small strain theory and non-linear soil properties where
the decrease in permeability was assumed to be proportional to the decrease in
compressibility (i.e. the coefficient of consolidation was assumed to be constant).
Walker et al. (2009) developed a solution to the consolidation of multi-
layered soils based on the spectral method by considering combined vertical and
radial drainage under instantaneous or single ramp loading, and by ignoring well
resistance. The most important advantage of the spectral methods is its ease of
implementation as the profile of excess pore water pressure across all the layers of
soil is conveniently described by a single expression, whereas existing solutions to
multi-layered consolidation problems have a separate equation for each layer.
Calculating the average excess pore pressures within or across any number of layers
would be far easier with a single expression. This solution includes the smear effect
by basing flow in a vertical direction on the average hydraulic gradient at a particular
depth.
Rujikiatkamjorn and Indraratna (2010) reported that the consolidation of multi-layer
deposits of soil depends on the characteristics of the smear zone, the permeability
ratio between the upper and lower layers of soil, the depth of penetration, and the

70
drain spacing. When the horizontal permeability in the upper layer is less than the
underlying layer, the duration of consolidation of the underlying layer can be
retarded depending on the depth of penetration of the upper layer of soil. On the
other hand, the consolidation process can be accelerated further when the down
dragged upper layer of soil has a higher permeability. The down drag effect due to
mandrel installation in layered soil is illustrated in Figure 2.28.

Vertical drain

Soil layer 1

Soil layer 2
Downdrag effect

i th Soil layer

Smear zone

Figure 2.28 Down drag effects due to mandrel installation in layered soil (after
Rujikiatkamjorn and Indraratna 2010)

In the study conducted by Rujikiatkamjorn and Indraratna (2010), a radial

consolidation model using a piecewise technique was proposed to investigate the
smear effect in layered soils. It was reported that the intrusion of the upper layer of
soil into the underlying layer of soil creates an additional zone where the remoulded
permeability can be increased or decreased depending on the initial permeability of
the upper layer. In the intrusion zone located in the lower layer of soil the variations
in permeability can be divided into three zones, including (i) a smear zone due to the
down drag of the upper layer of soil, (ii) a smear zone due to a remoulding of the
underlying soil, and (iii) the undisturbed lower layer of soil.

71
2.5.4 Theoretical Solutions for Vacuum Consolidation
A combined vacuum and surcharge one dimensional consolidation model was
proposed by Mohamedelhassan and Shang (2002) based on Terzaghi’s consolidation
theory. The mechanism of a combined vacuum and surcharge loading may be
determined by the law of superposition as a combination of surcharge preloading and
vacuum preloading (Figure 2.29).

Vacuum Surcharge Load (∆p) Vacuum Surcharge Load (∆p)

Preloading Preloading
(-p0) (-p0)

Z Impermeable Z Impermeable Z Impermeable

membrane membrane membrane
Permeable = Permeable + Permeable
H H H
layer layer layer
Impermeable Impermeable Impermeable

((a)) ((b)) (c)

Figure 2.29 A schematic diagram of a vacuum preloading system: (a) vacuum and
surcharge combining load, (b) vacuum preloading, and (c) surcharge preloading
(after Mohamedelhassan and Shang 2002)

According to Mohamedelhassan and Shang (2002), the average degree of

consolidation for combined vacuum and surcharge preloading can be expressed by:

ଶ
ܷ௩ ൌ ͳ െ σ௠ୀஶ ଶ
௠ୀ଴ ெ ݁‫ ݌ݔ‬ሺെ‫ܶ ܯ‬௩௖ ሻ (2.48)

ܶ௩ ൌ ܿ௩௖ ‫ݐ‬Ȁ‫ܪ‬ௗଶ (2.49)

where, Tvc is a time factor for the combined vacuum and surcharge preloading, and
cvc is the coefficient of consolidation for the combined vacuum and surcharge
preloading.
According to Indraratna et al. (2004b), when a vacuum pressure is applied
in the field through PVDs, the suction head along the length of the drain may
decrease with depth, thereby reducing its efficiency. Laboratory measurements taken
at a few points along PVDs installed in a large scale consolidometer clearly indicated
that the vacuum pressure propagates immediately, but a slight reduction in suction
pressure might be found down the length of the drain. The rate at which a vacuum

72
pressure develops within the drain may depend on the length and type of PVDs (core
and filter properties), but some field studies suggest that the vacuum pressure
develops rapidly even if the PVDs are up to 30 m long (Bo et al., 2003; Indraratna et
al., 2005a; Basu et al., 2007; Tran and Mitachi, 2008; Walker and Indraratna, 2007
and 2009; Walker et al., 2009b).
Based on laboratory observations, Indraratna et al. (2005a) developed a
modified radial consolidation theory to include different patterns of vacuum pressure
distribution (Figure 2.30). The results indicate that the efficiency of vertical drains
depends on both the magnitude of the vacuum pressure and its distribution. In order
to study the effect of the loss of vacuum, a trapezoidal distribution of vacuum
pressure may be assumed.

CL

-p0
Smear Zone Z

Vacuum Pressure
distribution
ks kh H

Undisturbed
Zone

-k1p0
ds/2
de/2

Figure 2.30 The distribution patterns of vacuum pressure in the horizontal and
vertical directions (after Indraratna et al. 2005a)

Indraratna et al. (2005b) assumed that the average excess pore pressure ratio
(ܴ௨ ൌ ο‫݌‬Ȁ‫ݑ‬ത଴ ) of the soil cylinder for radial drainage incorporating vacuum
preloading can be given by:

௣౬బ ሺଵା௞భ ሻ ଼்೓ ௣౬బ ሺଵା௞భ ሻ

ܴ௨ ൌ ቀͳ ൅ ഥబ
ቁ ‡š’ ቀെ ቁെ ഥబ
(2.50)
௨ ଶ ఓ ௨ ଶ

73
 ೖ೓
௡ ௞೓ గ௭ሺଶ௟ି௭ሻ௞೓ ିଵ
ೖೞ
ߤ ൌ Ž ቀ ௦ ቁ ൅ ቀ ௞ ቁ Žሺ‫ݏ‬ሻ െ ͲǤ͹ͷ ൅ ௤ೢ
ቊͳ െ ೖ೓ ೙ మ ቋ (2.51)
ೞ ቀ ቁሺ ሻ
ೖೞ ೞ

where, pv0 is the vacuum pressure applied at the top of the drain, k1 is the ratio
between the vacuum pressure at the top and bottom of the drain, u0 is the initial
excess pore water pressure, kh is the horizontal coefficient of permeability of soil in
the intact zone, ks is the horizontal coefficient of permeability of soil in the smear
zone, Th is time factor, n is ratio R/rw (R is the radius of the equivalent soil cylinder),
s is the ratio rs/rw , z is the depth, l is the equivalent length of the drain, and qw is the
discharge capacity of the well.
Geng et al. (2012) developed analytical solutions for vertical drains with
vacuum preloading for membrane and membrane-less systems under time-dependent
surcharge preloading by considering both vertical and horizontal drainage. The
properties of the constant smear zone (permeability and extent) and well resistance
were also considered in both models. According to the Laboratory results and field
measurements, the solution proposed by Geng et al. (2012) for the membrane-less
system includes a loss of vacuum along the length of the drain. The Laplace
transform technique was applied to derive the general solutions for pore water
pressure, settlement, and the degree of consolidation. Time dependent surcharge
preloading was considered by using the Laplace transform to simulate the history of
the embankment construction and the changes in vacuum pressure over time.

2.6 NUMERICAL SIMULATION OF PVD ASSISTED

PRELOADING

Different methods of numerical analysis have been widely used by geotechnical

engineers to rectify the limitations of analytical approaches in simulating complex
PVD assisted preloading projects to predict the ground behaviour, as well as conduct
parametric studies and back calculate the properties of the smear zone. Different
numerical programs such as CRISP, PLAXIS, ABAQUS, and FLAC have been used
by researchers to conduct numerical analyses. A number of numerical investigations
already performed are summarised in Table 2.11.
Indraratna and Redana (2000) used the finite element code CRISP92 (Britto and
Gunn 1987) to simulate PVD improved ground under the load of an embankment by

74
adopting the equivalent coefficient of permeability for the undisturbed soil and the
smear zone, and by incorporating Equations (2.32) and (2.33). The Modified Cam-
Clay model was applied as the soil constitutive model to conduct the numerical
analysis. The equivalent plane-strain width of the drain and smear zone were taken to
be similar to the axisymmetric radius. The discretised finite element mesh was
generated by linear strain quadrilateral elements with pore pressure nodes at the
corners. The embankment surcharge was modelled by applying incremental vertical
loads to the upper boundary. The smear zone was assumed to be three to four times
the radius of the mandrel while assuming a linearly decreasing permeability within
the smear zone. The numerical simulation conducted by Indraratna and Redana
(2000) indicated that including the smearing effects can significantly increase the
accuracy of the predicted settlements.
Arulrajah et al. (2005) used the PLAXIS 2D V.8 program to finite element
model vertical drains that simulated the unit cell and the full scale embankment. The
equivalent coefficients of horizontal permeability for the unit cells (axisymmetric to
plane-strain) were adopted to conduct the numerical simulations with the conversion
method proposed by Lin et al. (2000). The equivalent horizontal permeability of the
surrounding soil was taken as twice that of the equivalent vertical permeability.
According to Arulrajah et al. (2005), a good agreement was achieved between the
numerical results and the field measurements. Settlement predictions obtained from
the axisymmetric unit cell and the full scale analysis of vertical drains were found to
be in a good agreement with each other, as well as with the actual field
measurements.

75
Table 2.11 Summary of conducted numerical studies to simulate PVD assisted
preloading process
Applied
Simulated
Reference Objective of the Numerical Study Numerical
Case Study
Program
To evaluate the performance of soft Muar clay
Indraratna & clay foundation beneath embankmen CRISP92
Redana (2000) embankments stabilised with vertical t 2D
drains considering smear effects Malaysia
To investigate the effect of
unsaturation at a drain boundary on Large-scale
Indraratna et ABAQUS
the behaviour of a single PVD consolidom
al. (2004a) 2D
subjected to vacuum preloading eter test
applying 2D plane-strain model
To study the performance of a full- Second
Indraratna et scale test embankment constructed Bangkok ABAQUS
al. (2005a) on PVD improved soft clay international 2D
incorporating vacuum preloading airport
To model a single vertical drain
Indraratna et incorporating the effects of vacuum ABAQUS
-
al. (2005b) preloading and compare the results 2D
with proposed analytical solutions
To investigate the deformation
behaviour of marine clay under
Singapore
Arulrajah et al. reclamation fills conducting full- PLAXIS
marine clay
(2005) scale analysis of PVDs and compare 2D
at Changi
the results with the axisymmetric
approach
Vacuum
To determine the optimum CRISP-
Chai et al. consolidatio
penetration depth of PVDs in AIT
(2006) n Saga,
combined vacuum-surcharge loading 2D
Japan
To conduct 2D and 3D numerical
storage yard
Rujikiatkamjor modelling of combined surcharge ABAQUS
at Tianjin
et al. (2008) and vacuum preloading with vertical 2D & 3D
Port, China
drains
To evaluate the consolidation
Indraratna et Skå-Edeby ABAQUS
behaviour of PVD improved soft soil
al. (2008) Sweden 2D
beneath a circular embankment
To evaluate the efficiency of a
proposed conversion method
Tran & converting from an axisymmetric CRISP
-
Mitachi (2008) unit cell to an equivalent plane-strain 2D
unit cell under embankment loading
combined with vacuum preloading

76
 Table 2.12 (cont.)
To study the performance of the
matching procedures proposed by
Yildiz and Haarajoki
Hird et al. (1991 & 1995) when PLAXIS
Karstunen embankment
complex elasto-plastic models are 3D
(2009) Finland
used in the plane-strain analyses of
vertical drains
To evaluate the accuracy of three
different matching methods for Haarajoki
PLAXIS
Yildiz (2009) conversion of axisymmetric to plane- embankment
3D
strain conditions by comparing Finland
results of 2D and 3D analyses
To investigate the drainage Second
Lin and Chang behaviour of the PVD unit cell and Bangkok
FLAC 3D
(2009) the full-scale PVD improved ground international
conducting 3D numerical simulation airport

A number of researchers applied the numerical approach to analyse the combined

vacuum-surcharge preloading process by considering the smearing effects
(Indraratna et al., 2005a; Chai et al., 2006; Rujikiatkamjorn et al. 2008). 2D and 3D
multi-drain finite element analyses were conducted by Rujikiatkamjorn et al. (2008)
using ABAQUS software coupled with Biot consolidation theory to predict the
behaviour of PVD improved soft soil under combined vacuum and surcharge
(embankment) loading. The Modified Cam-Clay theory was adopted as the soil
constitutive model in these numerical analyses. The finite element mesh was
generated by C3D8RP solid elements (eight node tri-linear displacement and pore
pressure). The negative pore pressure was specified along the length of the PVD to
simulate the vacuum pressure. The equivalent coefficients of permeability in the
smear zone and undisturbed region were determined using the 2D conversion method
proposed by Indraratna et al. (2005a), and the extent of the smear zone extent was
considered to be twice the equivalent radius of the mandrel. The numerical study
conducted by Rujikiatkamjorn et al. (2008) indicated that the results of 2D and 3D
analyses were in good agreement, in terms of settlements, excess pore pressures, and
lateral displacements. Rujikiatkamjorn et al. (2008) reported that the application of
the equivalent plane-strain approach can result in accurate 2D numerical predictions
for multi-drain systems in large projects. According to Rujikiatkamjorn et al. (2008),
the height of the surcharge fill can be reduced by a vacuum preloading to achieve the
same desired rate of consolidation.

77
Yildiz (2009) conducted 2D and 3D finite element simulations to model PVD
improved soft soil under the embankment loading by adopting a recently developed
elasto-plastic S-CLAY1S constitutive model to represent the soft soil. The three
different conversion methods proposed by Hird et al. (1992), Indraratna and Redana
(1997), and Chai et al. (2001) were used to convert the 3D behaviour of the PVDs
into the equivalent plane-strain condition. A comparison between 3D and 2D
analyses showed that all the conversion methods applied resulted in acceptable
predictions, however, the method proposed by Indraratna and Redana (1997)
produced the best agreement with the 3D results, with a maximum difference of 3%
in the predicted settlements.

2.7 SUMMARY

As elaborated in this Chapter, the PVD assisted preloading system has been widely
used as a soft soil ground improvement technique to accelerate consolidation and
improve the strength of the soil, including the bearing capacity and the shear
strength. The installation of vertical drains disturbs the soil near the drains, reduces
the permeability of the smear zone, and retards the rate of consolidation quite
significantly.
Being able to accurately predict the characteristics of the smear zone is
essential in any analysis of the soil surrounding the drain, and to carry out reliable
practical designs. Two major parameters were proposed to characterise the smear
zone; , including the extent ratio (s= rs/rm) and the permeability ratio (n= kh/ks).
Determining the radius and permeability of the smear zone has been a subject of
intense discussion in literature. Different approaches have been proposed to evaluate
the extent and permeability of the smear zone, including analytical solutions,
laboratory tests, finite element methods, and back calculation techniques.
According to the available literature, rs/rm may vary from 1.6 to 7, while the
permeability ratio (kh/ks) changes from 1 to 10. It can be noted that wide ranges are
proposed for kh/ks and rs/rm and there is no meticulous method for being able to
predict these parameters precisely that can be used by practising engineers. The
assumed properties of the smear zone can lead to an early removal of the surcharge
during construction, which would result in excessive post-construction settlement or
excessive construction time and a subsequent increase in the project cost.

78
In this Chapter, the development of a consolidation theory (for vertical and radial
conditions) was explained by focusing on the solutions proposed for radial
consolidation and by considering the constant and variable properties of the smear
zone. Furthermore, the solutions proposed for the combined vertical and radial
consolidation theory in a single layer and multi-layered soil profiles were discussed.
Generally, 2D analysis with plane-strain permeability was used for the
numerical simulation of real PVD assisted case studies. The approaches available to
convert permeability from axisymmetric to plane-strain conditions were described in
this Chapter. Moreover, the application of numerical programs in simulating PVD
assisted preloading was discussed by presenting a summary of numerical studies
previously conducted in this area.
Construction of a trial embankment on the top of ground improved by PVD
is one of the most reliable techniques for accurately estimating the characteristics of
the smear zone in conjunction with the back analysis method. In many cases, a long
period of time is required to complete the consolidation process or to obtain 90% of
consolidation after constructing the trial embankment, which may delay the project
quite considerably, and also increase the costs significantly. Although back
calculation is used in practice, there is no indication of the minimum monitoring time
required. Determining the minimum degree of consolidation required resulting in
predicable properties of the smear zone can convert the construction of the trial
embankment into a very practical and cost effective method in the design of PVD
assisted preloading projects.
In this research, a numerical based attempt has been made to estimate the
minimum required degree of consolidation and consequently the minimum required
preloading time that would result in a reliable prediction of the extent and
permeability of the smear zone. In addition, an experimental program has been
conducted to validate the proposed back calculation procedure to obtain the smear
zone properties.

79
 CHAPTER THREE

3 PVD ASSISTED PRELOADING SIMULATION AND

BACK CALCULATION PROCEDURE USING FLAC

3.1 GENERAL

Numerical analysis methods have been widely used by geotechnical engineers to

rectify the limitations that analytical approaches have in simulating complex
prefabricated vertical drain (PVD) assisted preloading projects to predict ground
behaviour, conduct parametric studies, and back calculate the properties of the smear
zone. Bergado et al. (1993) applied the numerical program VERDRN, developed in
FORTRAN 77, to model and back calculate the characteristics of the smear zone in a
single vertical drain. The finite element program CRISP (Britto and Gunn, 1987) was
used by Chai et al. (1999) to simulate a trial embankment at Saga Airport in Japan by
adopting a one-dimensional finite element model. The smear zone properties were
simulated by Indraratna et al. (2005b) through a back calculation process simulating
the Suvarnabhumi Airport in Bangkok using ABAQUS software. Lin and Chang
(2009) applied FLAC 3D to investigate the consolidation of a full scale PVD
improved ground by simulating the Second Bangkok International Airport case
history. Lin and Chang (2009) multiplied the permeability ratio obtained in the
laboratory ((kh/ks)laboratory) by the modification coefficient of 4 (Cf =4) to determine
the permeability ratio in field conditions ((kh/ks)field). The horizontal permeability of
the undisturbed soil was then measured and the permeability ratio ((kh/ks)field) was
used to determine the permeability of the smear zone. Lin and Chang (2009) applied
the equation proposed by Bergado and Long (1993) to determine the equivalent
horizontal permeability by taking the smearing effect into consideration.
Literature shows that CRISP, ABAQUS, and PLAXIS are the most
commonly used software packages by researchers conducting numerical simulations
(e.g. Indraratna and Redana, 2000; Saowapakpiboon et al., 2011; Sathananthan et al.,
2008; Stapelfeldt et al., 2008).

80
The main purpose of this chapter is to back calculate the properties of the smear zone
using the available consolidation data, which needs a coded numerical program with
the capability of conducting an efficient parametric study. Therefore, a systematic
back calculation procedure was proposed to obtain accurate smear zone properties.
The finite difference program FLAC 2D, was introduced and selected as the
appropriate numerical code to conduct consolidation analyses for PVD assisted
preloading case histories with complex geometries, and by adopting the proposed
procedure for back analysis.

3.2 NUMERICAL MODELLING

3.2.1 Adopted Numerical Simulation Software

In this study the capabilities of different numerical programs (e.g. PLAXIS, FLAC,
ABAQUS, and CRISP) were investigated to select suitable software embracing the
requirements of this research.
PLAXIS and CRISP are finite element packages that are intended for a two
dimensional analysis of deformation and stability in geotechnical engineering. The
convenient CAD-style drawing option and automatic generation process of
unstructured 2D finite element meshes converted these codes into practical tools
which are being used for geotechnical analysis by geotechnical engineers who are
not necessarily numerical specialists. However, the lack of a built-in programming
language and the automatic mesh generation can be addressed as the limitations of
these two programs in conducting parametric studies and sensitive analysis for
research purposes. It should be noted that the manual process in CRISP to generate
meshes, is time consuming and tedious.
For example, when simulating PVD assisted preloading to investigate the
characteristics of the smear zone on consolidation, the geometry of the model should
be changed manually every time. There are new values for the extent and
permeability of the smear zone, to be implemented in the program, which is a tedious
and time consuming process and may increase the risk of having faults and errors.
Furthermore, the size of meshes cannot be defined manually in an appropriate way to
suit the specific needs. Another limitation, associated with PLAXIS to model the
preloading projects, is lack of a procedure to assign and fix certain values, such as

81
the pore water pressure at the vertical drain boundaries, to properly model the
vertical drain and vacuum pressure. Therefore, numerical software with more
advanced features is needed to conduct geotechnical research studies, especially
when simulating prefabricated vertical drain preloading projects.
ABAQUS 2D is a general finite element code, which is also an appropriate
numerical code for simulating mandrel driven PVD performance. This software suite
incorporated many civil and structural features, but not developed exclusively for
geotechnical engineering design purposes. FLAC 2D is specifically designed for 2D
geotechnical analyses, which is more popular in the geotechnical industry than
ABAQUS. However, limited researchers have used FLAC 2D to model PVD
performance and verify its accuracy for potential users. Therefore, FLAC 2D
software was selected for numerical simulation in this study because of its ability to
conduct systematic parametric studies and back calculation procedures.
FLAC 2D v6.0 was used to model the PVD assisted preloading process that
focuses on the variability of the smear zone. FLAC 2D is a two-dimensional explicit
finite difference program for engineering mechanics computation. In comparison to
the common commercial finite element numerical modelling software, the mixed
discretisation scheme developed by Marti and Cundall (1982) was used to model
plastic deformation and flow accurately because it was believed to be physically
more justifiable than the reduced integration scheme that is commonly used with
finite element analysis. In addition, the explicit solution scheme used in FLAC 2D,
unlike the common implicit methods, can compute any material non-linear
behaviour in almost the same computation time as a linear law, whereas an implicit
solution can take significantly longer to solve non-linear problems such as fully
coupled time dependent stress-deformation analysis capturing soil consolidation. In
addition, large strain simulation (e.g. deformation of deep soft layers of clay due to
preloading) in FLAC 2D is hardly more time consuming than a small strain case.

3.2.2 Explicit Finite Difference Method and Lagrangian Analysis

FLAC 2D is an explicit, finite difference program that performs a Lagrangian

analysis. The finite difference method is a numerical technique, which solves sets of
differential equations, and given initial values and/or boundary values (Atkinson
and Han, 2005). In the finite difference method every derivative in the set of

82
governing equations is replaced directly by an algebraic expression, written in terms
of the field variables (e.g., stress or displacement) at discrete points in space, which
are undefined within the elements. In contrast, the finite element method has a
central requirement that the field quantities (stress, displacement) vary in each
element in a prescribed fashion, using specific functions. It can be noted that finite
element programs often combine the matrices of the element into a large global
stiffness matrix (Logan, 2007), whereas this is not normally conducted in the finite
difference approach because it is relatively efficient to regenerate the finite difference
equations at each step. FLAC 2D uses an explicit, time-marching method to solve
the algebraic equations.
The general calculation sequence embodied in FLAC 2D is illustrated in
Figure 3.1. This procedure first invokes the equations of motion to derive new
velocities and displacements from stresses and forces, and the strain rates are
derived from velocities and new stresses are derived from the strain rates. One time-
step is taken for every cycle around the loop. It should be remembered that each
box in Figure 3.1 updates all of its grid variables from known values that remain
fixed while control is within the box. The selected time-step is so small that
information cannot physically pass from one element to another in that interval, and
since one loop of the cycle occupies one time-step, an assumption of “frozen”
velocities (i.e., the newly calculated stresses do not affect the velocities) is justified
(neighboring elements really cannot affect one another during the period of
calculation) (Itasca, 2008).

Equiblirium Equation
(Equation of Motion)

new new
velocities and stresses
displacements or forces

Stress/Strain Relation
(Constitutive Equation)

Figure 3.1 Basic explicit calculation cycle (after Itasca, 2008)

83
The central mathematical concept of the explicit method is that the calculated
“wave speed” always keeps ahead of the physical wave speed, so that the equations
always operate on known values that are fixed for the duration of the calculation.
The most important advantage of this concept is that no iteration process is required
while the stresses are being computed from strains in an element, even if the
constitutive law is significantly non-linear. In an implicit method (which is
commonly used in finite element programs), every element communicates with
every other element during one solution step, thus, several cycles of iteration are
necessary before compatibility and equilibrium are obtained. Table 3.1 compares
the explicit and implicit methods. The key disadvantage of the explicit method is
that small time-steps should be used which means that a large number of steps must
be taken. Overall, the explicit methods are best for ill-behaved systems (e.g., non-
linear, large strain, and physical instability).

Table 3.1 Comparison of explicit and implicit solution methods (after Itasca, 2008)

Explicit Implicit

Time-step must be smaller than a critical Time-step can be arbitrarily large, with
value for stability unconditionally stable schemes
Small amount of computational effort per Large amount of computational effort per
time-step time-step
No significant numerical damping Numerical damping dependent on time-step
introduced for dynamic solution present with unconditionally stable schemes

No iterations necessary to follow nonlinear Iterative procedure necessary to follow

constitutive law nonlinear constitutive law
Always necessary to demonstrate that the
Provided that the time-step criterion is
above- mentioned procedure is: (a) stable;
always satisfied, nonlinear laws are always
and (b) follows the physically correct path
followed in a valid physical way
(for path-sensitive problems)
Stiffness matrices must be stored. Ways
Matrices are never formed. Memory
must be found to overcome associated
requirements are always at a minimum. No
problems such as bandwidth. Memory
bandwidth limitations
requirements tend to be large
Since matrices are never formed, large
Additional computing effort needed to
displacements and strains are accommodated
follow large displacements and strains
without additional computing effort

Since there is no need to form a global stiffness matrix, it is a small matter to update
coordinates at each time-step in the large-strain mode. These incremental

84
displacements are added to the coordinates so that the grid moves and deforms with
the material it represents. This is termed a “Lagrangian” formulation, in contrast to
an “Eulerian” formulation where the material moves and deforms relative to a fixed
grid. The constitutive formulation at each step is a small-strain one, but is equivalent
to a large-strain formulation over many steps.
FEM codes usually represent steady plastic flow by a series of static
equilibrium solutions (Zienkiewicz et al., 2005). The quality of the solution for
increasing applied displacements depends on the nature of the algorithm used to
return stresses to the yield surface, following an initial estimate using linear stiffness
matrices. The best FEM codes will give a limit load (for a perfectly plastic material)
that remains constant with an increasing applied displacement. The solution provided
by these codes will be similar to that provided by FLAC 2D. However, FLAC’s
formulation is simpler because no algorithm is needed to bring the stress of each
element to the yield surface, and therefore the plasticity equations are solved exactly
in one step. Therefore, FLAC 2D may be more efficient than some FEM codes for
modelling steady plastic flow. Moreover, FLAC 2D is robust in a sense that it can
handle any constitutive model with no adjustment to the solution algorithm, while
many FEM codes need different solution techniques for different constitutive
models.

3.2.3 The Grid and Mixed-Discretization Zoning Technique

In this study a solid body is divided into a finite difference mesh consisting of
quadrilateral elements in order to carry out numerical modelling. Internally, FLAC
2D sub-divides each element into two overlaid sets of constant-strain triangular
elements, as shown in Figure 3.2.

85
 a
b
ni(2) ni(1)

S(2) S(1)

c
d

(a) (b)
Figure 3.2 Finite difference mesh, (a) FLAC 2D zone composed of overlaid
triangular elements and (b) typical triangular element

The four triangular sub-elements are called a, b, c and d (see Figure 3.2). The
deviatoric stress components of each triangle are maintained independently, requiring
sixteen stress components to be stored for each quadrilateral (4 × σxx , σyy , σzz , σxy ).
The force vector exerted on each node is taken to be the mean of the two force vectors
exerted by the two overlaid quadrilaterals so that the response of the composite
element is symmetrical for a symmetric loading. If one pair of triangles becomes
significantly distorted (e.g., if the area of one triangle becomes much smaller than
the area of its companion), then the corresponding quadrilateral is not used, and only
nodal forces from other (more reasonably shaped) quadrilaterals are used. It can be
noted that if both overlaid sets of triangles are badly distorted, FLAC 2D complains
with an error message.
The use of triangular elements eliminates the problem of hourglass
deformation, which may occur with constant-strain finite difference quadrilaterals.
The term “hourglassing” comes from the shape of the deformation pattern of
elements within a mesh. A common problem, which occurs when modelling
materials that will undergo yielding, is the incompressibility of plastic flow (volume
change is zero). The use of plane-strain or axisymmetric geometries introduces a
kinematic restraint in the out-of-plane direction which often gives rise to an over
prediction of the collapse load. This condition is sometimes referred to as “mesh-
locking” or “excessively stiff” elements, and is discussed in detail by Nagtegaal et
al. (1974). The problem arises as a condition of local mesh incompressibility which
must be satisfied during flow, and which results in over constrained elements. To
86
 overcome this problem the isotropic stress and strain components are taken to be
constant over the whole quadrilateral element, while the deviatoric components are
treated separately for each triangular sub-element. This procedure, referred to as
mixed discretisation, was described by Marti and Cundall (1982). The term mixed
discretisation arises from the different discretisations for the isotropic and deviatoric
parts of the stress and stain tensors.

3.2.4 Continuum Expression of the Governing Equations

Applying a surcharge on top of the ground will result in a sudden change in the pore
water pressure along the saturated soil profile (e.g. Terzaghi and Frohlich, 1936;
Jacob, 1940; Skempton, 1954). This change in the pore water pressure is often
known as the excess pore water pressure, and the difference between the applied total
stress and the pore water pressure is referred to as the effective stress (Terzaghi,
1925). Since soil is a compressible porous medium with interconnected pores linked
to drainage boundaries, the excess pore water pressure dissipates gradually over time
and causes a time dependent deformation of the soil body, called “consolidation”
(Terzaghi, 1925). First, changes in the pore water pressure induce changes in the
effective stress, which affects the response of the solid (for example, a reduction in
the effective stress may cause plastic yield). Second, a change in the pore water
pressure alters the volume of soil in a zone.
In this study, a coupled fluid-mechanical approach was used to simulate
consolidation by adopting a basic fluid flow model with the following characteristics:

i. The fluid flow law corresponds to both isotropic and anisotropic permeability,

ii. Different zones may have different fluid-flow properties,

iii. Fluid pressure, flux, and impermeable boundary conditions may be prescribed,

iv. Fluid sources (wells) may be inserted into the material as either point sources
(interior discharge) or volume source (interior well), and

v. Any of the mechanical models may be used with the fluid flow models.

87
The fluid and mechanical loops are implemented alternately to conduct a coupled
fluid-mechanical modelling. Saturated transient fluid flow through the soil is
modelled in the fluid loops to simulate the generation and dissipation of excess pore
water pressure, while the stress-displacement development processes are modelled in
the mechanical loops. The Biot theory of consolidation (Biot, 1941) was used in
FLAC 2D to formulate coupled fluid-deformation mechanisms. Biot (1941)
developed four differential equations as general equations governing consolidation in
a three-dimensional environment. They are satisfied by four unknowns, including the
components of displacement in x, y, and z directions (denoted by u, Ԃ, and w,
respectively) and the increment of water pressure (ߪ). The equations proposed by
Biot (1941) are shown as follows:

ீ డఢ డఙ
‫׏ܩ‬ଶ ‫ ݑ‬൅ െ ߙ డ௫ ൌ Ͳ (3.1)
ଵିଶఓ డ௫

ீ డఢ డఙ
‫׏ܩ‬ଶ ߴ ൅ ଵିଶఓ డ௬ െ ߙ డ௬ ൌ Ͳ (3.2)

ீ డఢ డఙ
‫׏ܩ‬ଶ ‫ ݓ‬൅ ଵିଶఓ డ௭ െ ߙ డ௭ ൌ Ͳ (3.3)

డఢ ଵ డఙ
݇‫׏‬ଶ ߪ ൌ ߙ డ௧ ൅ ொ డ௧ (3.4)

where,

డమ డమ డమ
‫׏‬ଶ ൌ డ௫ మ ൅ డ௬ మ ൅ డ௭ మ (3.5)

ଶሺଵାఓሻ ீ
ߙ ൌ ଷሺଵିଶఓሻ ு (3.6)

ଵ ଵ ఈ
ொ
ൌ ோെு (3.7)

ఙ
ߞ ൌ െோ (3.8)

ఙ
߳ ൌ െு (3.9)

88
where, G is the shear modulus, ߤ is Poisson’s ratio, ߳ is the volumetric strain, ߙ is
Biot’s coefficient, ߞ denotes the variation of water content (variation of water
volume per unit volume of porous material), k is the coefficient of permeability, 1/H
is a measure of the compressibility of the soil due to a change in water pressure, 1/R
is a measure of the change in water content due to a given change in water pressure,
and 1/Q is a measure of the amount of water, which can be forced into the soil under
pressure while the volume of the soil is kept constant.
Biot (1941) reported that the constants ߙ and Q are dependent on the degree
of saturation and incompressibility of the system, respectively, and can be taken as 1
and λ (ߙ ൌ ͳ and ܳ ൌ λ) by assuming that water is an incompressible material
(with no air bubbles), and by considering the soil to be a fully saturated material.
Furthermore, the change in volume of the saturated soil is equal to the amount of
water squeezed out. Based on these assumptions, the fundamental equations
governing the consolidation of completely saturated clay were proposed by Biot
(1941) in the following forms:

ீ డఢ డఙ
‫׏ܩ‬ଶ ‫ ݑ‬൅ െ డ௫ ൌ Ͳ (3.10)
ଵିଶఓ డ௫

ீ డఢ డఙ
‫׏ܩ‬ଶ ߴ ൅ ଵିଶఓ డ௬ െ డ௬ ൌ Ͳ (3.11)

ீ డఢ డఙ
‫׏ܩ‬ଶ ‫ ݓ‬൅ ଵିଶఓ డ௭ െ డ௭ ൌ Ͳ (3.12)

డఢ
݇‫׏‬ଶ ߪ ൌ డ௧
(3.13)

where

ଶீሺଵାఓሻ
‫ܪ‬ൌܴൌ ଷሺଵିଶఓሻ
(3.14)

ߞൌ߳ (3.15)

The coupled calculation process starts with a fluid loop that evaluates pore water
pressure and specific discharge using the transport law (Darcy’s law), the fluid
balance laws, and the compatibility law. Then the change of pore water pressure is

89
passed to the mechanical loop to update the effective stress, which is used to check
against failure and calculate the volumetric strain based on the selected constitutive
model. The calculated volumetric strain is passed back to the fluid loop to account
for the pore water pressure and specific discharge in the new fluid loop on the basis
of the linear quasi-static Biot’s theory (Biot, 1941), which was formulated in an
incremental format. Because these two loops were implemented alternately, pore
water pressures and volumetric strains are kept exchanging and updating. Modelling
this process may require thousands of cycles, which is terminated when a specific
criterion (such as the maximum unbalanced force, the flow time, the stress ratio and
so on) is reached.
The governing differential equations corresponding to FLAC numerical
implementation are discussed below.

3.2.4.1 Water Flow Equation

Darcy’s law was used in FLAC 2D to describe the water flow as follows:

డ
‫ݍ‬௜ ൌ െ݇௜௝ ݇෠ሺ‫ݏ‬ሻ డ௫ ሺܲ െ ߩ௪ g௞ ‫ݔ‬௞ ሻ (3.16)
೔

where, ‫ݍ‬௜ is the specific discharge vector (i.e. volumetric flux or the filtration
velocity per unit volume), ݇௜௝ is the tensor of the coefficient of permeability (FLAC
permeability tensor), ݇෠ሺ‫ݏ‬ሻ is the relative permeability (when soil is fully saturated
݇෠ሺ‫ݏ‬ሻ ൌ ͳ), P is the water pressure, ߩ௪ is the mass density of the water, and g௞ is the
gravity vector (0,0,g). For saturated/unsaturated flow in FLAC 2D, the air pressure is
assumed to be constant and equal to zero.

3.2.4.2 Balance Laws

Biot (1941) applied Darcy’s law to derive the water mass balance equation based on
the fact that the rate of the water content of an element of soil must be equal to the
volume of water entering per unit time through the surface of the element, assuming
the water to be incompressible. The equation proposed by Biot (1941) was used in
FLAC 2D for the water mass balance as follows:

90
డ఍ డ௤
డ௧
ൌ െ డ௫೔ ൅ ‫ݍ‬௩ (3.17)
೔

where, ‫ݍ‬௩ is the intensity of the volumetric water source. According to Bear (1972),
the balance of momentum has the following form:

డఙ೔ೕ ௗ௨ሶ ೔
൅ ߩg௜ ൌ ߩ (3.18)
డ௫೔ ௗ௧

where, ɐ୧୨ is the effective stress, ߩ ൌ ሺͳ െ ݊ሻߩ௦ ൅ ݊ߩ௪ is the solid bulk density, ߩ௦
and ߩ௪ are the densities of the solid and water phases, respectively, and n is porosity.
Note that (1 − n)ρs corresponds to the bulk density of the dry matrix (ρd) (i.e., ρ = ρd
+ nρw ).

3.2.4.3 Constitutive Laws

In order to have a complete set of differential equations for the transient phenomenon
of consolidation (i.e., those equations governing the distribution of stress, water
content, and settlement as a function of time in a soil under given loads), Dracy’s law
was used by Biot (1941) to develop an extra equation which relates the variations of
water pressure (ܲ), water content (ߞ), and volume changes (߳) over time to each
other (Equation (3.4)). This relationship is called the response equation for the pore
water and depends on the degree of saturation. In FLAC 2D, the following form of
the response equation is used:

డ௉ డ఍ డఢ
డ௧
ൌ ‫ܯ‬௕ ቀడ௧ െ ߙ డ௧ ቁ (3.19)

where, Mb is the Biot modulus, α is the Biot coefficient and ߳ is the volumetric
strain. The Biot modulus is related to the drained bulk modulus of the porous medium
(K) and the water bulk modulus, (Kw), via n and α, as follows:

௄ೢ
‫ܯ‬௕ ൌ ಼ (3.20)
௡ାሺఈି௡ሻሺଵିఈሻ ೢ
಼

For saturated clay when the compressibility of grains is negligible compared to the
drained bulk material (α = 1), the Equation (3.20) is converted to:

91
 ௄ೢ
‫ܯ‬௕ ൌ ௡
(3.21)

The small-strain constitutive response for the porous solid is described by:

ௗ
ௗ௧
൫ߪ௜௝ ൅ ߙܲߜ௜௝ ൯ ൌ ‫ܪ‬൫ߪ௜௝ ǡ ߳ሶ௜௝ ǡ ߢ൯ (3.22)

where, H is the functional form of the constitutive law (e.g. modified Cam-clay,
Mohr-Coulomb, Hoek-Brown, strain-hardening/softening) and κ is a history
parameter. In particular, the elastic relationships that relate effective stresses to
strains have the following form:

ଶ
ߪ௜௝ െ ߪ௜௝ఖ ൅ ߙሺܲ െ ܲఖ ሻߜ௜௝ ൌ ʹ‫߳ܩ‬ሶ௜௝ ൅ ሺ‫ ܭ‬െ ଷ ‫ܩ‬ሻ߳ሶ௞௞ ߜ௜௝ (3.23)

where, the superscript “ o ” refers to the initial state and ߳ሶ௞௞ is the volumetric strain
increment.

3.2.4.4 Compatibility Equation

In numerical analysis approaches, the equations of equilibrium satisfying the strain

compatibility equations can be solved by using a certain constitutive law under
prescribed boundary conditions. In FLAC 2D, Equation (3.24) was used as the strain
compatibility equation because it relates the strain rate to the velocity gradient.

ଵ డ௨ሶ ೕ డ௨ሶ
߳ሶ௜௝ ൌ ଶ ൤డ௫ ೔ ൅ డ௫ ൨ (3.24)
ೕ ೔

3.2.5 Numerical Fluid Flow Formulation

3.2.5.1 Basic Scheme

The basic fluid flow scheme can be applied for fully saturated transient or steady
state fluid flow, run independently or coupled to a mechanical calculation.
Substituting Equation (3.17) into Equation (3.19) yields expressions for the fluid
continuity equations, which are then solved in FLAC 2D using a finite-difference
approach based on a discretisation of the medium into zones composed of two
overlays of triangles. T h e pore water pressure varies linearly in a triangle,

92
assuming a uniform specific discharge in the triangle. Figure 3.2 should be consulted
whenever reference is made to the triangles that make up FLAC’s quadrilateral
elements. In the related equations, pressures and saturation values are assumed to be
located at grid points (or “nodes”). Zone pressures are derived from the surrounding
nodal values by simple averaging.
Commencing from a state of mechanical equilibrium, a coupled hydro-
mechanical static simulation using the basic scheme in FLAC 2D involves a series of
steps (Itasca, 2008). Each step includes one or more flow steps (flow loop), followed by
enough mechanical steps (mechanical loop) to maintain a quasi-static equilibrium. The
increment of the pore water pressure due to water flow was evaluated in the flow loop,
while the contribution from volumetric strain was evaluated in the mechanical loop as
a zone value, which is then distributed to the nodes. The total stress correction due to
the change in pore water pressure arising from the mechanical volume strain is
performed in the mechanical loop, and that arising from the flow of water in the flow
loop. The total value of the pore water pressure was used to evaluate effective
stresses and detect failure in plastic materials (Itasca, 2008).

3.2.5.2 Constitutive Law: Derivation of Element “Stiffness Matrix”

Neglecting gravity for the moment (it will be incorporated later), Darcy’s Equation
for an anisotropic porous medium is reduced to the following equation:

డ௉
‫ݍ‬௜ ൌ െ݇௜௝ డ௫ (3.25)
೔

Each quadrilateral element is divided into triangles in two different ways (see Figure
3.2(a)). The pore water pressure was assumed to vary linearly in a triangle, and the
vector ‫ݍ‬௜ was derived for a generic triangle of area A by applying the Gauss
divergence theorem (Itasca, 2008). For a triangular sub-element, Equation (3.25)
has the following finite difference form:

௞೔ೕ
‫ݍ‬௜ ؆ σ ܲ݊௝ ‫ݏ‬ (3.26)
஺

93
where, ∑ is the summation over the three sides of the triangle, ni is the unit normal to
the side i, and s is the length of the side. The two components of the specific
discharge capacity (q) are:

ଵ
‫ݍ‬ଵ ൌ ஺ ሾ݇ଵଵ σ ܲ݊ଵ ‫ ݏ‬൅ ݇ଵଶ σ ܲ݊ଶ ‫ݏ‬ሿ (3.27)

ଵ
‫ݍ‬ଶ ൌ ஺ ሾ݇ଶଵ σ ܲ݊ଵ ‫ ݏ‬൅ ݇ଶଶ σ ܲ݊ଶ ‫ݏ‬ሿ (3.28)

where, q1 is the specific discharge capacity vector for triangle side 1, q2 is the
specific discharge capacity vector for triangle side 2 (Figure 3.2).
Considering, for example, the contribution of side (ab) of the triangle (Figure
3.2) to the summations:

ͳ
“ ͳ ሺƒ„ሻ ൌ ʹ ൣǦ ͳͳ ൫ ሺ„ሻ ൅ ሺƒሻ ൯൫šʹ ሺ„ሻ Ǧšʹ ሺƒሻ ൯൅ ͳʹ ൫ ሺ„ሻ ൅ ሺƒሻ ൯൫šͳ ሺ„ሻ Ǧšͳ ሺƒሻ ൯൧ (3.29)

ͳ
“ ʹ ሺƒ„ሻ ൌ ʹ ൣǦ ʹͳ ൫ ሺ„ሻ ൅ ሺƒሻ ൯൫šʹ ሺ„ሻ Ǧšʹ ሺƒሻ ൯൅ ʹʹ ൫ ሺ„ሻ ൅ ሺƒሻ ൯൫šͳ ሺ„ሻ Ǧšͳ ሺƒሻ ൯൧ (3.30)

The two other sides, (bc) and (ca) (Figure 3.2), provide similar contributions to ‫ݍ‬௜ .
This specific discharge vector contribution was then converted to scalar volumetric
flow rates at the nodes by making dot products with the normals to the three sides of
the triangle. The general expression is:

ܳ ൌ ‫ݍ‬௜ ݊௜ ‫ݏ‬ (3.31)

For example the flow rate into node (a) is:

ܳ ሺ௔ሻ ൌ ൛െ‫ݍ‬ଵ ൫‫ݔ‬ଶ ሺ௕ሻ െ ‫ݔ‬ଶ ሺ௖ሻ ൯ ൅ ‫ݍ‬ଶ ൫‫ݔ‬ଵ ሺ௕ሻ െ ‫ݔ‬ଵ ሺ௖ሻ ൯ൟȀʹ (3.32)

The factor of 2 accounts for the fact that the node only captures half the flow crossing a
neighboring edge (since the other half goes to the other node of the edge). Similar
expressions apply to nodes (b) and (c). Nodal flow rates are added from the three
triangles meeting at the node and are divided by 2 since the flow sum comes from
two overlaid grids. The “stiffness” matrix ሾࣧሿ of the whole quadrilateral element is

94
defined in terms of the relationship between the pressures at the four nodes and the
four nodal flow rates, as derived above:

ሼܳሽ ൌ ሾࣧሿሼܲሽ (3.33)

For the special case of a square zone, aligned with the coordinate axes, the stiffness
matrix has the following form:

ʹ െͳ Ͳ െͳ
ሾࣧሿ ൌ െ ൦െͳ ʹ െͳ Ͳ
௞
൪ (3.34)
ଶ Ͳ െͳ ʹ െͳ
െͳ Ͳ െͳ ʹ

where, k is the coefficient of isotropic mobility, which is the coefficient of pore

pressure term in Darcy’s law that can be defined as the ratio of intrinsic permeability
to fluid dynamic viscosity. This matrix is identical to the one that would be
obtained in a classical finite difference method.
By including gravity, when the grid point pressures around a zone conform
ப୔
to the gradient ப୶౟
ൌ g୧ ɏ୵ , where ‰ ‹ is the vector of gravitational acceleration, then

the nodal flow rates {Q} should be zero. Hence, Equation (3.33) is modified as
follows:

ሼܳሽ ൌ ሾࣧሿ൛ܲ െ ሺ‫ݔ‬௜ െ ‫ݔ‬௜ ሺଵሻ ሻg௜ ߩ௪ ൟ (3.35)

ሺଵሻ
where, ‫ݔ‬௜ is the x-coordinate of one of the corners.

3.2.5.3 Continuity Equation

The flow imbalance (∑ Q) at a node causes a change in the pore water pressure at a
saturated node as follows:

డ௉ ெ್ డ௏
ൌെ ቀσ ܳ ൅ ߙ ቁ (3.36)
డ௧ ௏ డ௧

where, V is the total volume associated with the node. The term ∑ Q includes
contributions from the four surrounding zones and any sources that are specified by

95
the user (e.g. outflow from a well). Equation (3.36) has the following finite difference
form:

ெ್ ሺσ ொο௧ାఈο௏೘೐೎೓ ሻ
ܲ௜ାଵ ൌ ܲ௜ െ  ௏

(3.37)

where, ∆Vmech is the equivalent increase in the nodal volume arising from
mechanical deformation of the grid. The term V was computed as the sum of the
contributions from all triangular subzones connected to the node. Each triangle
contributes a third of its volume, and the resulting sum is divided by two, to account
for the double overlay scheme in FLAC 2D.

3.2.5.4 Numerical Stability: Fluid Time-step

There are two aspects of numerical stability associated with the pore-fluid scheme.
First, an explicit solution of the fluid flow equations requires that the time-step be less
than a critical value. Second, as the bulk modulus of the water increases the
mechanical stiffness of a saturated zone, the effect of increased mechanical stiffness is
incorporated into quasi-static analysis in the density-scaling scheme in FLAC 2D. The
apparent mechanical bulk modulus of a zone is modified by the presence of water as
follows:

‫ܭ‬௜ାଵ ൌ ‫ܭ‬௜ ൅ ߙ ଶ ‫ܯ‬௕ (3.38)

where, α is the Biot coefficient and Mb is the Biot modulus, which are explained in
previous sections.
The explicit fluid time-step can be derived by imagining that one node at the
centre of four zones is given a pressure of ଴. The resulting nodal flow is then given
by Equation (3.35) as  ൌ ଴ σ ୩୩ , where, σ ୩୩ is the sum over the four zones of
the diagonal terms corresponding to the selected node. The excess nodal flow gives
rise to an increment in ∆P, according to Equation (3.37).

ெ್ ொο௧
οܲ ൌ െ ௏
 (3.39)

The new pressure at the node P1 is then

96
 ெ್ σ ࣧೖೖ ο௧
ܲଵ ൌ ܲ଴ ൅ οܲ ൌ ܲ଴ ቀͳ െ ௏
ቁ (3.40)

where, [ࣧ] is the stiffness matrix relating t he pore water pressure to the flow rate,
as defined by Equation (3.33). Equation (3.40) is stable and monotonic if the
following equation is satisfied:

௏
ο‫ ݐ‬൏ (3.41)
ெ್ σ ࣧೖೖ

The value of ∆t used in FLAC 2D is that given by Equation (3.41) multiplied by a

safety factor of 0.8.

3.2.6 Optimisation of the Mechanical and Fluid Time steps

The fluid flow solution is controlled by the “SET” and “SOLVE” commands in
FLAC 2D. The “SET” command allows certain options to be switched on or off, and
certain solution parameters to be specified. The keywords “flow” and “mech” are
used to define the method of analysis. For example, the “flow” and “mech” keywords
should be switched on using the “SET” command to conduct coupled fluid and
mechanical analysis.
In practice the mechanical effects occur instantaneously compared to the
diffusion effects, where no time is associated with any of the mechanical sub-steps
taken in association with fluid-flow steps in order to satisfy quasi-static equilibrium.
Each fluid step corresponds to a real period of time. By default, FLAC 2D alternates
the mechanical steps and fluid steps in the basic fluid flow scheme, by one
mechanical step and then by one fluid step.
At the beginning of the solution of a coupled problem, even one fluid step
may put the system a long way out of equilibrium (i.e., there will be large
unbalanced forces). Hence, many mechanical steps should be taken for each fluid
step. As consolidation continues, the changes in fluid pressure will become small, so
the system will remain in equilibrium. At this stage, several fluid time steps may be
taken for each mechanical step (the reverse of the previous strategy). Two methods
were proposed by Itasca (2008) to obtain the optimum combination of fluid and
mechanical time steps that would result in an acceptable accuracy, and they are
explained below.

97
3.2.6.1 Manual Method

In this method the number of fluid and mechanical steps is controlled manually. The
“SET nmech” and “SET ngw” commands are used to set the number of mechanical
and fluid steps for each cycle, as denoted by the “STEP” command, respectively. The
unbalanced force ratio (or unbalanced force), which is generally assumed to be 0.001
or 0.01, should be monitored during this process. The unbalanced force ratio can be
defined using “SET sratio” command. It should be noted that adopting large fluid
and mechanical steps may increase the required time for numerical analyses
significantly.
The optimum combination of fluid and mechanical steps can be selected
when two criteria are satisfied; (i) the unbalanced force ratio (sratio) falls below the
defined value (0.001 or 0.01), and (ii) a numerical analyses is being conducted in a
minimum period of time. In this study a parametric study was carried out to
determine the optimum combination of fluid and mechanical steps which indicated
that adopting “nmech” and “ngw” in the range of 5 to 10 can result in accurate
outcomes and acceptable time for numerical calculations.

3.2.6.2 Automatic Method

In this approach the number of fluid and mechanical sub-cycles is adjusted

automatically to keep the maximum unbalanced force ratio (or unbalanced force)
below a present value. This algorithm is activated by the “auto on” keyword in the
“SOLVE” command; the total groundwater time (consolidation time) to be modelled
is specified by the “age” keyword in the “SOLVE” command (SOLVE auto on age
=consolidation time), keeping the computation until the time given by the “age”
parameter is reached. Also, parameters “step” and “clock” should also be set, to
prevent premature truncation of the run due to limits on step number and elapsed
time, respectively, from being reached.
According to the investigations carried out in this study, applying the
automatic approach results in acceptable numerical predictions, and therefore the
number of fluid and mechanical sub-cycles was adjusted automatically to conduct the
coupled fluid-mechanical analyses.

98
3.2.7 Modified Cam-Clay Model

In this study the modified Cam-Clay (MCC) model was adopted in the developed
numerical code to investigate the consolidation of soft soil. The MCC model was
introduced by Roscoe & Burland (1968) to model the behaviour of clay soil. The
details of this model and its numerical implementation can be found in a number of
references such as Wood (1990), and Gens and Potts (1988). The MCC model is an
incremental hardening/softening elasto-plastic model. Its features include a
particular form of non-linear elasticity (semi logarithmic), and a
hardening/softening behaviour governed by volumetric plastic strain. The failure
envelopes are similar in shape, and correspond to ellipsoids of rotation about the
mean stress axis in the principal stress space. The shear flow rule is associated and
no resistance to tensile mean stress is offered in this model. T h e MCC model has
been discussed extensively by Roscoe & Burland (1968).
The MCC model is expressed in terms of three variables; the mean effective
pressure (‫݌‬Ԣ); the deviator stress (q); and the specific volume (v). The generalised
stress components ‫݌‬Ԣ and q may be expressed in terms of principal stresses, as
follows:

ଵ
‫݌‬Ԣ ൌ െ ଷ ሺߪଵ ൅ ߪଶ ൅ ߪଷ ൅ሻ (3.42)

ଵ
‫ݍ‬ൌെ ඥሺߪଵ െ ߪଶ ሻଶ ൅ ሺߪଶ െ ߪଷ ሻଶ ൅ ሺߪଵ െ ߪଷ ሻଶ (3.43)
ξଶ

3.2.7.1 Virgin Consolidation Line and Swelling Lines

In the MCC model the assumption is that when a sample of soft soil is slowly
compressed under isotropic stress and under perfectly drained conditions, the
relationship between specific volume (v) and Ž ‫݌‬Ԣconsists of a straight virgin
consolidation line (also known as the normal compression line) and a set of straight
swelling lines (Figure 3.3). Swelling lines are also called unloading-reloading lines.

99
 normal
consolidation line

κ B
1
λ
1
swelling lines

Figure 3.3 Normal consolidation line swelling lines for an isotropic compression test

The normal consolidation line (loading line) and swelling line (reloading line) are
defined by the Equations (3.44) and (3.45), respectively.

௣ᇲ
‫ ݒ‬ൌ ‫ݒ‬ఒ െ ߣ݈݊ ௣ ᇲ (3.44)
భ

௣ᇲ
‫ ݒ‬ൌ ‫ݒ‬఑ െ ߢ݈݊ ௣ ᇲ (3.45)
భ

where, ߣǡ ߢǡ ܽ݊݀‫ݒ‬ఒ are the constant properties of the soil. The parameter ɉ
represents the slope of the normal compression (virgin consolidation) line, or
critical state line in ‫ ݒ‬െ ݈݊‫݌‬ᇱ space, Ɉ is the slope of the swelling line in ‫ ݒ‬െ ݈݊‫݌‬ᇱ
space, ’ଵ ᇱ is the reference pressure, and ‫ݒ‬஛ is the value of the specific volume at the
reference pressure. The value of ‫ݒ‬఑ for a particular line depends on the location of
the point on the normal consolidation line from which unloading is performed.

3.2.7.2 Yield and Potential Functions

Under increasing triaxial shear loading (q), MCC soil behaves elastically until a yield
value of q is attained (Roscoe and Borland, 1968; Wood, 1990). Coordinates at the
yielding point (‫݌‬ᇱ ǡ ‫ )ݍ‬are determined from the following function:

݂ ൌ ‫ ݍ‬ଶ ൅ ‫ܯ‬ଶ ‫݌‬ሺ‫ ݌‬െ ‫݌‬௖ ሻ (3.46)

100
where, ‫݌‬௖ ᇱ is preconsolidation pressure and M is the slope of the critical state line
(CSL) in ‫݌‬ᇱ െ ‫ ݍ‬space. The yield condition f=0 is represented by an ellipse with
horizontal axis, ‫݌‬௖ ᇱ , and vertical axis, ‫݌ܯ‬௖ ᇱ , in the ‫݌‬ᇱ െ ‫ ݍ‬space. Figure 3.4 indicates
the yield function for the MCC model.

plastic dilation

qcr=M(p'c/2) plastic compaction

p'cr=p'c/2 p'c p'

Figure 3.4 Yield surface of the Modified Cam-Clay model in pˊ-q plane (after Roscoe
and Borland, 1968)

3.2.7.3 Determination of the Input Parameters

The slope M of the CSL in ‫݌‬ᇱ െ ‫ ݍ‬space can be calculated from the friction angle
߮ ᇱ (Huang et al., 2006) of the Mohr-Coulomb yield criterion through the following
equation:

଺௦௜௡ఝ ᇲ
‫ ܯ‬ൌ ଷି௦௜௡ఝᇲ (3.47)

The slopes ߣ and ߢ of the normal compression and swelling lines in ‫ ݒ‬െ ݈݊‫݌‬ᇱ space
are related to the compression index Cc, and the swelling index Cs, and can be
determined through the following equations, respectively (Huang et al., 2006):

஼೎
ߣ ൌ ୪୬ ଵ଴ (3.48)

101
 ஼ ೞ
ߢ ൌ ୪୬ ଵ଴ (3.49)

Often ߢ is chosen to be in the range of 0.1λ to 0.3 λ (Graham, 2010; Budhu, 2010;
Liu and Evett, 2013).

3.3 NUMERICAL CODE DEVELOPMENT

3.3.1 General

In the present study, a comprehensive numerical code was developed using the FLAC
2D program to simulate the process of PVD assisted preloading. The developed
code can simulate the laboratory consolidation test and fully instrumented trial
embankment by using the axisymmetric and plane-strain models. A systematic back
calculation procedure was embedded in the numerical code to conduct the parametric
study and back calculate the properties of the smear zone for a given set of
consolidation results obtained from laboratory or field measurements.
The Modified Cam-Clay (MCC) model was used to simulate the behaviour
of soft clay during the process of PVD assisted preloading. The smear zone was
simulated by applying the mechanical properties of undisturbed soil with a reduced
coefficient of permeability. The sub-routines were written using the built-in
programming language FISH (FLACish) to tailor the analyses to suit the specific
needs of the parametric study and by giving the following unique advantages to the
developed codes for this study:

(i) Automatic mesh generation by entering the parameters needed to modify the grid
pattern inside and outside the smear zone,
(ii) Considering variations in permeability with the void ratio during consolidation,
(iii) Ability to change different parameters such as the dimensions of the model, the
properties of the vertical drain, the profile of the subsoil, characteristics of the
smear zone, and the preloading conditions,
(iv) Integrated permeability axisymmetric to equivalent plane-strain permeability
conversion,
(v) The option to define the exact location of the desired points to generate and plot
any future history graphs,
(vi) Automatic solving process based on the modified input data, and
102
(vii) The option to adjust the number of mechanical and flow steps to optimise the
analyses and minimise the numerical calculation issues.

The solid body is divided into a finite difference mesh composed of quadrilateral
elements as described in Figure 3.2. A sample of the generated numerical mesh
pattern using the developed code is illustrated in Figure 3.5.

Symmetric
axis Trial embankment

Crust

Layer 1

Layer 2
h0
Layer 3

Layer 4

(a)

Ground surface

PVD

h1

Smear zone

Intact zone

(b)

Figure 3.5 Generated FLAC mesh, (a) sample of discretised finite-difference mesh of
trial embankment, and (b) the pattern of meshes in the smear zone and undisturbed
region

3.3.2 Numerical Code Structure

The structure of the developed numerical code is explained in this section. It was
implemented into the back calculation procedure to predict reliable smear zone
properties. A schematic diagram of the code’s structure is illustrated in Figure 3.6.

103
 Entering the input data Undrained Analysis

Simulation of
Grid and Mesh Generation
Preconsolidation process

Layering and Assigning the Simulation of Vertical Drains

Material Properties and Smear Zone

Simulation of staged loading

Defining the History
& possible application of
Locations
vacuum

Consolidation analysis to
Boundary Conditions and
obtain deformations & pore
Initial Stresses
water pressures

Figure 3.6 Structure of the developed numerical code

3.3.2.1 Input data: Geometry and properties of the Materials

The first section of the code was developed to enter the parameters needed for
numerical modelling, such as the model geometry, PVD properties, and mesh
generation parameters. A FISH sub-routine was embedded in the developed code to
automatically convert the coefficient of permeability from axisymmetric to plane-
strain condition using the conversion equations proposed by Indraratna et al. (2005a).
FLAC code can adopt the strain-dependent permeability using the
“per_table” keyword (Itasca, 2008). A table can be created to provide permeability
values as a function of accumulated volume strain. It should be noted that sufficient
scope must be provided in the table so that all possible volume excursions are
anticipated. Since changes in the permeability affect the critical timestep and element
matrices, a recalculation of these quantities is done if the table is active, but only at
every tenth step. Recalculating the quantities of groundwater increases the solution

104
time. The parameters applied in the numerical code are illustrated in Figure 3.7. A
sample of the developed code for this part is indicated in Appendix A (Section A-1).

Symmetric
axis
a0
a1
a2

Layer 1 ly2

Layer 2
h1 Layer 3 ly9
h0 Layer 4
Layer 5

Layer 6

Layer 7

h2 Layer 8

sdrain

gn1
l1 l2 sdrain

Smear zone Layer 6

Intact zone

Layer 7
Smear zone

PVD

Figure 3.7 Typical input parameters for numerical simulation

It should be mentioned that a specific spreadsheet was designed to calculate the

reference pressure and corresponding specific volume by considering the pre-
consolidation pressures at each depth and the initial void ratios. The loading and
reloading curves in Figure 3.3 were referred to when conducting the calculations.

105
3.3.2.2 Grid and Mesh Generation

This section of the code was developed to conduct the automatic process of
generating the grid and mesh for the numerical model by considering the dimensions
of the model, length of the PVD, radius of the mandrel, the extent ratio (rs/rm), the
number of mesh zones inside the smear zone, the length ratio (l1/l2), the length of the
vertical zones, and the drain spacing. For this purpose, the following stages were
followed:

x Calculation of the number of zones needed in the horizontal direction

x Calculation of the number of zones needed in the vertical direction
x Grid and mesh generation of the soil profile
x Grid and mesh generation of the trial embankment

A sample of the generated mesh is illustrated in Figure 3.6 and the developed code
for this section is presented in Appendix A (Section A-2).

3.3.2.3 Layering and Assigning Material Properties

In this part the layers of soil and the stages of constructing a trial embankment were
categorised in different groups and input soil properties were assigned to each of
them (Figure 3.8). The trial embankment stages were deactivated to generate the
initial stress condition in the soils. The corresponding developed code for this
purpose is presented in Appendix A (Section A-3).

106
 Stage 7 (group 16)
Stage 6 (group 15)
Stage 5 (group 14)
Stage 4 (group 13)
Stage 3 (group 12) Stage 2 (group 11)
Stage 1 (group 10)
Layer 1 (group 1)

Layer 2 (group 2)

Layer 3 (group 3)
Layer 4 (group 4)
Layer 5 (group 5)
Layer 6 (group 6)

Layer 7 (group 7)

Layer 8 (group 8)

Layer 9 (group 9)

Figure 3.8 Input parameters in terms of different groups for numerical simulation

3.3.2.4 Defining the location of instrumentations and transducers

The developed code has an option to predefine any grid point or zone to record
changes in the required variables such as vertical displacement and pore water
pressure as time-stepping proceeds. A summary of the requested histories may be
plotted at any time. Variables can be plotted versus step number or versus other
histories. An example of the developed code is reported in Appendix A (Section A-
4), where some grid points were defined to record the changes in vertical
displacement and pore water pressure during consolidation.

3.3.2.5 Boundary Conditions, Initial Stresses, and Undrained Analysis

A sub-routine was developed to define the boundary conditions, generate the initial
stresses in the soil profile, and conduct the undrained analysis to obtain the initial
equilibrium stress condition. The boundary conditions are shown in Figure 3.9.

107
 Drainage boundary

Layer 1

Layer 2
Vertical drain
boundary Layer 3 Impermeable
Layer 4 boundary
Layer 5
Layer 6

Layer 7

Layer 8

Layer 9

Impermeable boundary

Figure 3.9 Boundary conditions of the numerical model

The displacement of soil at the bottom boundary was fixed in all directions and the
side boundaries are vertical rollers, which allowed for vertical displacement. The
drainage boundary is at the ground surface and the bottom and side boundaries are
impermeable. A specific FISH function (ininv.fis) was used to automatically
compute the initial condition for the horizontally layered soil profile by using the
depth of the ground water table and lateral pressure coefficients as the input data.
Finally, an undrained analysis was conducted to obtain the initial state of
equilibrium. The code developed for this section is reported in Appendix A (Section
A-5).

3.3.2.6 Simulation of Preconsolidation Stage

This section of the code was designed to conduct the preconsolidation process by
activating the mesh of the trial embankment platform (drainage blanket) and
assigning the Mohr-Coulomb model and corresponding soil properties. Figure 3.10
shows the schematic numerical model of this stage. The corresponding code
developed for this purpose is presented in Appendix A (Section A-6).

108
 Trial embankment platform
or drainage blanket

Figure 3.10 Schematic model for the preconsolidation stage

3.3.2.7 Vertical Drains and Smear Zone

Specific functions were implemented in the developed code to automatically model

the vertical drains and smear zones. To simulate the vertical drains, the pore water
pressure was fixed to hydrostatic pressure at every node along the grids of the
vertical drain (i.e. zero excess pore water pressure). The smear zones were modelled
by assigning reduced permeability values to them. Figure 3.11 shows a sample of the
numerically simulated vertical drain and adjacent disturbed area. An example of the
developed code is reported in Appendix A (Section A-7).

109
Symmetric axis

Crust

Layer 1

Layer 2
h0
Layer 3

Layer 4

(a)

u=0
Ground surface
(Zero pore pressure)

PVD
(Hydrostatic pressure)
Vertical drain h1
with fixed pore
Smear zone with
pressure
reduced permeability

Intact zone
u= h1 Jw

(b) (c)

Figure 3.11 Sample of the simulated vertical drain and adjacent disturbed area, (a)
Discretised finite-difference mesh of the trial embankment, (b) Variation of the
hydrostatic pressure along the vertical drain and (c) the pattern of meshes in the
smear zone and undisturbed region elements

3.3.2.8 Vacuum Pressure

In FLAC 2D, the vacuum pressure can be simulated by applying and fixing a
negative pore water pressure along the vertical drain. The simulation of vacuum
pressure is indicated in Figure 3.12. For this purpose a FISH function was developed,
which is presented in Appendix A (Section A-8).

110
 -70 kPa Vacuum pressure at top

Vacuum pressure distribution

Ground surface

PVD

Vertical drain
Smear zone

Intact zone

-70 kPa Vacuum pressure at bottom

(a) (b)

Figure 3.12 Simulation of vacuum pressure, (a) variation of vacuum pressure along
the vertical drain and (b) the pattern of meshes in the smear zone and undisturbed
region elements

3.3.2.9 Construction of Trial Embankment and Consolidation Process

The final section of the numerical code was developed to simulate the time history of
constructing a trial embankment and conducting a consolidation analysis. The
changes in the required variables can be plotted against the consolidation time at
predefined gridpoints and zones after completing the numerical analyses. The
developed code for this section is presented in Appendix A (Section A-9).

3.4 SYSTEMATIC BACK CALCULATION PROCEDURE

In this study, the construction of a trial embankment was introduced as a practical

and reliable solution to predict the properties of the smear zone using the back
calculation approach. For this purpose, a systematic back calculation procedure was
designed and integrated with the developed numerical code to define the properties
of the smear zone using the field measurements. Figure 3.13 presents a detailed
flowchart, including the back calculation procedure used to find the characteristics of
the smear zone that resulted in the best fit with the field measurements.

111
 Start

Input (Stage 1)
Vertical drain pattern
Field performance observation Geotechnical site investigation
& loading history
(monitoring results) (Soil properties)
(improvement plan)

∆s, (rs/rm )0 ∆R, (kh/ks )0

(rs/rm )min, (rs/rm )max (kh/ks )min, (kh/ks )max

i=1, j=1
(rs/rm )i = (rs/rm )0
(kh/ks )i = (kh/ks )0

j=i

(rs/rm )i = (rs/rm )i + ∆s

NO

Systematic numerical
(rs/rm )min ≤ (rs/rm )i ≤ (rs/rm )max Save corresponding

parametric study
rs/rm & kh/ks to (Es)min

(Stage 2)
YES
(Es)min =min Es(j ≤ s < i)

NO
(kh/ks )min ≤ (kh/ks )i ≤ (kh/ks )max

(kh/ks )i = (kh/ks )i-1 + ∆R

YES

i = i+ 1 Conduct numerical simulation adopting FLAC,

predicting ground behaviour (Pk)

Estimate the inaccuracies predictions

Save Ei, (rs/rm )i & (kh/ks )i
Ei=∑ │Fk-Pk │, (1≤ k ≤ n)
zone characteristics

Reporting the back

calculated smear

Print rs/rm & kh/ks

(Stage 3)

corresponding to the minimum errors

Stop

Figure 3.13 Back calculation flowchart for smear zone characteristics and the
minimum required monitoring time for trial embankment

112
 assumed initial extent permeability ratio increment depending on
(rs/rm)0 ∆R
ratio the required accuracy
assumed initial extent ratio increment depending on the
(kh/ks)0 ∆S
permeability ratio required accuracy
error between numerical predictions and
(rs/rm)min minimum extent ratio Ei
field measurements at step i
(rs/rm)max maximum extent ratio (Emin)s minimum error corresponding to step s
minimum permeability extent ratio corresponding to step s that
(kh/ks)min (rs/rm)s
ratio gives the minimum error
maximum permeability ratio corresponding to step s
(kh/ks)max (kh/ks)s
permeability ratio that gives the minimum error
(rs/rm)i extent ratio at step i Fk Field measurements
permeability ratio at
(kh/ks)i Pk Numerical predictions
step i

The proposed back calculation procedure had three stages as follows:

(i) Stage 1: Entering the input data, including the ground conditions, soil
properties, details of installing the PVD, and the upper and lower bound values
for the properties of the smear zone (rs/rm and kh/ks).

(ii) Stage 2: Conducting a numerical analysis varying rs/rm and kh/ks in the given
range to predict the ground behaviour, determine the corresponding settlement
curves, and calculate the error that accumulated between the numerical results
and field measurements at every degree of consolidation.

(iii) Stage 3: Analysing the outcomes of the second stage to define the optimum
combination of rs/rm and kh/ks that gave the best predicted settlement curve, and
which is in best agreement with the field measurements.

As illustrated in the flowchart of Figure 3.13, the approach began with collecting the
input data, including the soil properties, the PVD assisted preloading specifications
(PVD pattern and loading history), and field monitoring results (settlement and
excess pore water pressure variations). The first stage was completed by defining the
initial values, the minimum and maximum values, and the incremental rates for the
extent and permeability ratios. In the second stage, a systematic parametric study was
designed to back calculate and predict the properties of the smear zone as follows:

x Defining an external loop to vary the extent ratio (rs/rm),

113
x Defining an internal loop for each step of the external loop to change the
permeability ratio (kh/ks),

x Implementing the developed FLAC code in each step of the internal loop to
conduct a numerical simulation using the properties of the smear zone and input
data,

x Completing each step of the internal loop by calculating the computational error,
and comparing the numerical predictions and field monitoring results,

x Stopping the internal loop once the permeability ratio (kh/ks) was out of range, and
saving the extent and permeability of the smear zone that corresponded to the
minimum error, and

x Continuing the procedure in Stage 2 as long as the extent ratio (rs/rm) was in the
defined range.

Stage two stopped as soon as the extent ratio was out of the input range. The last
stage of the flowchart shown in Figure 3.13 was to report the back calculated pairs of
the extent and permeability of the smear zone that corresponded to the minimum
errors. It should be mentioned that in Stages 2 and 3, the FLAC output was linked to
a visual basic code to calculate the error and report the results.
The normalised error between the numerical prediction and field
measurement in every step of the consolidation process was calculated using the
following equation:

ሺௌ೟ ሻ೔ ିሺௌ೟೛ ሻ೔
ሺ‫ܧ‬௜ ሻ௡ ൌ (3.50)
ௌ೑

where, (Ei)n is the normalised error at time t and step number n, St is the field
settlement at time t, Stp is the predicted settlement at time t, and Sf is the final primary
consolidation settlement.
The predicted errors in stage 3 were used to calculate cumulative error for
each case at every step of the consolidation process using the following equation:

ሺௌ೟ ሻ೔ ିሺௌ೟೛ ሻ೔
ሺ‫ܧ‬௧ ሻ௡ ൌ σ௡௜ୀଵ ேൈௌ೑
(3.51)

114
where, (Et)n is the normalised cumulative error at time t and step number n, and N is
the total number of steps. The most accurate smear zone properties belong to the case
with the minimum amount of cumulative error.

3.5 SUMMARY

In this chapter, a systematic procedure was proposed to back calculate the properties
of the smear zone using the consolidation data measured in the laboratory and the
field. A numerical approach was employed to simulate layered soft clay improved by
PVD, using the proposed back calculation procedure. The explicit finite difference
program FLAC 2D was used to develop the numerical code. The explicit scheme
could incorporate the large deformations that soft soil may experience. The built-in
programming language FISH (FLACish) was used to write the sub-routines and tailor
the analyses to suit specific needs. The governing differential equations
corresponding to FLAC’s numerical implementation and the numerical formulation
of the coupled fluid-deformation mechanisms in FLAC 2D (which is based on the
Biot theory of consolidation) was also discussed.
The methods existing in FLAC 2D for optimising the mechanical and fluid
time steps was described in order to keep the system in an equilibrium condition.
Two criteria were used to determine the optimum combination of fluid and
mechanical steps, (i) the unbalanced force ratio (sratio) falls below the defined values
(0.001 or 0.01), and (ii) the numerical analyses was conducted in a minimum period
of time.
The Modified Cam-Clay (MCC) theory, developed based on an
incremental hardening/softening elasto-plastic model, was used as the soil
constitutive model in the numerical simulations. The MCC model relates the
specific volume (v) to Ž ‫݌‬Ԣ with a straight virgin consolidation line as the soft soil
sample was slowly compressed under the isotropic stress and perfectly drained
conditions.
The structure of the developed code was explained and included, (i) the data
entering process, (ii) generation of the grid and mesh, (iii) layering and assigning
material properties, (iv) applying the boundary conditions and initial stresses, (v)
simulation of the pre-consolidation process, (vi) simulation of the vertical drain,

115
smear zone, and vacuum pressure and (vii) construction of a trial embankment and
the process of consolidation.
Finally, a diagram of the systematic back calculation procedure was
described (Figure 3.13). It was combined with the FLAC code, developed to reliably
predict the properties of the smear zone. A parametric study scheme was
implemented into the procedure proposed in Figure 3.13, using different
combinations of smear zone properties to predict the total cumulative error between
the numerical results and field measurements for each case. The extent and
permeability of the smear zone for the case with the minimum total cumulative error
can be reported as the optimum combination, which could be used by practicing
engineers.

116
 CHAPTER FOUR

4 LABORATORY STUDY TO INVESTIGATE THE

SMEARING EFFECT ON THE PERFORMANCE
OF PVD ASSISTED PRELOADING

4.1 GENERAL

In recent years a number of experimental studies were carried out to investigate the
formation of a disturbed zone near the vertical drains resulting from installing a
mandrel (e.g. Bergado et al., 1991; Indraratna and Redana, 1998; Sharma and Xiao,
2000; Sathanantha and Indraratna, 2006; Shin et al., 2009; Ghandeharioon et al.,
2012). However, because the parameters of the smear zone vary due to mandrel
insertion, using the experimental results to evaluate the accuracy of the analytical and
numerical solutions is very difficult.
In this study an array of laboratory tests were conducted using a large, fully
instrumented Rowe cell apparatus (250mm diameter by 200mm high) to investigate
how the reduced permeability surrounding the drain affects consolidation, and verify
the numerical procedure developed to back calculate the properties of the smear
zone. The disturbed/smear soil near the vertical drain was simulated using clay with
a reduced permeability and a compacted sand pile covered with flexible porous
geotextile that was installed in the centre to act as a vertical drain. The dissipation of
pore pressure was captured during consolidation using transducers installed at
various distances away from the vertical drain. At the same time the surface
settlement was measured using a vertical settlement transducer connected to a data
logger. Finally, the laboratory test results were used to verify the numerical
procedure developed to back calculate the permeability of the smear zone, and the
formulations available for converting the axisymmetric condition to a plane-strain
condition were evaluated by comparing the numerical results with the experimental
measurements.

117
4.2 TESTING APPARATUS AND EXPERIMENTAL
PROCEDURE

4.2.1 Apparatus
The large scale Rowe cell used in this study consists of the body, the base, and the
top. The internal diameter is 250 mm and the height is 200 mm (Figure 4.1a). The
base and the cover are bolted to flanges on the body in nine locations. An ‘O’ ring
seals the base, and a rubber loading jack and another ‘O’ ring seal the top part.
A uniform load was applied to the sample by water pressure acting on a
convoluted rubber jack made from rubber. Vertical settlement was measured at the
centre of the sample where it connected to a spindle that is attached to the jack and
passes through the cover of the cell using an LVDT (Linear Variable Differential
Transformer). The spindle is sealed through the centre of the rubber diaphragm on
the jack by two washers to eliminate any error in the settlement readings due to the
rubber diaphragm compressing under pressure. Nine pore water pressure transducers
(PWPTs) were installed in various distances and depths to capture the changes in
pore water pressure during consolidation (Figure 4.1a and 4.1b). All the transducers
are connected to a PC based data logger to continuously measure and record the test
data.
The Rowe cell was designed to have a single drainage system from the top
through a perforated brass disc situated between the sample and the jacket. The
drainage outlet is via the centre of the settlement spindle and a short flexible tube
leading to a Klinger valve at the edge and on top of the cell (Figure 4.1a).

118
 Settlement
Settlement
dial gauge
gauge support

Cell pressure
supply line Settlement
Cell top rod

de-airing Drainage control valve

Watter screw

Back pressure
Initial drainage &
de-airing valve
Diaphragm

110 mm

135 mm
Initial drainage & Porous plate
de-airing valve PWPT A4 250 mm Cell body
160 mm
PWPT A3

182 mm
200 mm
PWPT A2
Porous de-airing
de-airing PWPT A1
stone system
system PWPT B5
O-ring seal

PWPT B1 PWPT B4
PWPT B3 PWPT B2
Cell base

(a)

Smear zone boundary

Vertical drain boundary

PWPT B5

102 mm PWPT B2
m
m
27
75 mm

52
mm
PWPT B1 PWPT B3

PWPT B4
(b)

Figure 4.1 Large scale Rowe cell apparatus (a) schematic diagram of the cell and (b)
locations of the pore pressure transducers at the base of the cell

Pressure is applied to the cell, and back pressure to the jacket from a series of
pressure lines connected to the enterprise level pressure/volume controllers (ELDPC)
that are filled with de-aired water. The de-aired water in the ELDPC cylinder is
pressurised and displaced by a piston moving in the cylinder. The piston is actuated

119
by a ball screw that rotates in a ball nut held by an electric motor and gearbox that
move rectilinearly on a ball slide. The ELDPC instrument and its operational
schematic are illustrated in Figures 4.2a and 4.2b. The applied ELDPC had a
volumetric capacity (nominal) of 200cc for a pressure range of 1MPa. The resolution
of the measurements for pressure and volume were 1kPa and 1 mm3, respectively.
The pressure and volume measurements were accurate to 0.25% of the full range and
0.4% of the measured value with +/- 500mm3 back lash, respectively.

(a)

Ball screw
Pressure Piston
Stepper motor Cylinder Pressure
and gearbox
outlet
± steps
Digital
control
circuit Air

Linear bearing Pressure

de-aired water trancducer

Analogue feedback
(b)

Figure 4.2 Pressure/volume controller device, (a) a photographic view of the GDS
controller and (b) an operational schematic diagram of the instrument
(http://gdsinstruments.com)

The pressure was measured by an integral solid state transducer. Control algorithms
were built into the on-board micro-controller to enable the controller to seek a target
pressure or step to a target volume change. The change in volume was measured by

120
counting the steps of the incremental motor. The instrument was controlled via PC
based software.
When a single pressure/volume controller is used, , it will need to manually
fill or empty the controller once the volumetric capacity of the barrel has reached either
100% full or 100% empty. To fix this problem in this study, 2 parallel pressure/volume
controllers (primary and secondary) connected to an infinite volume controller (IVC)
device (Figure 4.3) were used for each pressure line. The IVC was designed to remove
constraints from the volume such that in a test, fluid can flow continuously under
pressure control or volume control. The IVC system automatically switches between
the two controllers when one runs out of volume.

Figure 4.3 Infinite volume controller instrument (http://gdsinstruments.com)

One controller was called the primary controller and it was used to provide a
pressure source, while the other was called the ‘master’ controller and it acted as the
pressure source (is this correct?). At the same time the secondary ‘slave’ controller
can refill and centre itself and become ready for switching when the master
controller exceeds its volumetric limit. When this happens the secondary controller
takes over by providing the pressure and thus becomes the master. The primary
controller becomes the slave and hence can refill/empty and centralise itself.
When this is complete, the primary controller resumes pressure control and
becomes the master again and the secondary controller centralises itself.
The secondary pressure controller is only temporarily used as the master
(to provide pressure or a change in volume) when the primary controller exceeds

121
its volume capacity and thus needs to refill/empty and centralise itself. The IVC
valve panel comes complete with 4 channel IVC control boxes t ha t a r e connect
the IVC valve panel to the computer. GDS software controls the system and
collects the data from data loggers in every stage of loading. The de-aired water
needed, is provided through an elevated water reservoir. A schematic diagram of the
experimental set up is shown in Figure 4.4. Figure 4.5 shows the set up established
in the laboratory.

Water
Tank

DL DL

P/V C P/V C
IVC IVC

P/V C P/V C
LVDT
CP

BP

DL PWPT

PC

DL: Data logger LVDT: Linear Variable Differential Transformer

IVC: Definite Volume Controller CP: Cell Pressure
P/V C: Pressure/Volume Controller BP: Back Pressure
PWPT: Pore Pressure Transducer PC: Computer

Figure 4.4 Schematic diagram of Rowe cell set-up

122
 Infinite volume
controller (IVC)
Water tank

Rowe cell

Data logger (DL)

Pressure/volume
Pore water pressure
controller (PVC)
transducers (PWPT)

Figure 4.5 Established setup in the laboratory

4.2.2 Material Properties

4.2.2.1 Soil samples

To select soils for the Rowe cell test with a reduced horizontal permeability for the
smear zone and the undisturbed region, a number of reconstituted clay samples were
prepared by mixing the following materials:

x Q38 kaolinite

x ActiveBond23 bentonite

x Fine sand

Kaolinite and bentonite were selected because they are common artificial clays with
significantly different properties. The kaolinite samples had an average liquid limit
of 50% compared to 340% for the bentonite samples. The kaolinite and bentonite
samples indicated 9% and 35%, respectively as their shrinkage limits, when

123
measured according to AS 1289.3.4.1. Table 4.1 summarises the properties of the
clay samples used in this study.

Table 4.1 Properties of the adopted soil samples in this study

Liquid Limit Plastic Limit Linear Shrinkage
Soil Type USCS
(%) (%) (%)
Q38 kaolinite 49.5 27.5 9 CL
ActiveBond23
340 50 35 CH
Bentonite

According to Nelson and Miller’s (1997) sorting, the kaolinite and bentonite clays
have had medium and very high expansion/shrinkage potentials, respectively.
ActiveBond 23 is a pure form of bentonite which is plastic and impermeable, and has
a high absorbing and swelling capacity, as well as being highly viscous when
suspended in water. Common uses of bentonite include constructing the diaphragm
wall, piling, tunnelling, and sealing the dam and containing waste. The high swelling
properties of bentonite on exposure to water facilitate sealing porous soils and
leaking dams. Q38 kaolinite clay is a dry milled creamy white kaolin China clay.
Kaolinite is one of the most abundant minerals in soil, and as such is often
encountered in on-site conditions. Kaolinite is formed by the breakdown of feldspar,
which is induced by water and carbon dioxide, and is often formed by the alteration
of aluminium silicate minerals in a warm and humid environment (Craig, 2000;
Murray, 1999). Uniformly graded sand (SP) was used for two purposes: (i) to
achieve the target permeability of the soil mixture, and (ii) represent the drain
material. The grain size distribution curve of the applied sand is illustrated in Figure
4.6, and some important sizes are shown in Table 4.2.

124
 100%

90% CU=D60/D10=2.3
80% CC=D302/(D60D10)=1.2
Percent passing (%)

70%

60%

50%

40%

30%

20%

10%

0%
0.01 0.1 1 10
Grain size (mm)

Figure 4.6 Grain size distribution curve for vertical drain sand

Table 4.2 Important sizes

Symbol Grain size (mm)

D10 0.24

D30 0.4

D60 0.55

Note: D10 is the effective particle size (the grain diameter at 10% passing), D30 is the grain
diameter at 30% passing, and D60 is the grain diameter at 60% passing

Table 4.3 shows the mixed design of the samples. The Australian Standard (AS
1289.3.5.2) was used to determine the plastic limit (PL) and liquid limit (LL) of the
mixtures, which are illustrated in Table 4.4. The specimens were thoroughly mixed
with a water content that was 1.4-1.8 times the liquid limit (LL) and kept in a closed
container for couple of days to ensure full saturation and homogeneity. The
properties of the reconstituted clay samples are shown in Table 4.4.

125
 Table 4.3 Mix design for the reconstituted samples
ActiveBond 23
Soil Reference Number Q38 Kaolinite (%) Fine sand (%)
Bentonite (%)
S1 70 15 15
S2 68 17 15
S3 65 20 15

Table 4.4 Properties of the reconstituted samples

Soil Reference Number S1 S2 S3

Water Content, w (%) 120 120 120

Liquid Limit, LL (%) 67 70 87
Plastic Limit, PL (%) 27 29 34
Plasticity Index, PI 40 41 43

4.2.2.2 Consolidation test on reconstituted samples

Materials for the disturbed/smear zone and the intact area have been selected from
the reconstituted samples after a number of oedometer tests were conducted. Samples
were subjected to a pre-consolidation pressure of 20 kPa before the consolidation test
was carried out. For this purpose, three cylinders were filled with the reconstituted
samples (Table 4.3) and submerged in buckets of water to maintain their saturation
condition. The pre-consolidation process is shown in Figure 4.7. Once pre-
consolidation was completed, samples were taken from the soils using the oedometer
ring (Figure 4.8).

126
 (a) (b)

Figure 4.7 Pre-consolidation process prior to the oedometer test; (a) cylinder
contacting reconstituted sample and (b) samples under pre-consolidation
pressure

(a) (b)

(c) (d)

Figure 4.8 Preparing the samples for the oedometer test, (a) placing the oedometer
ring, (b) cutting the extra top part, (c) cutting the extra bottom part, and (d) the
final sample

A conventional oedometer test was conducted based on the Australian Standard (AS
1289.6.6.1) to determine the coefficient of permeability for each sample by applying
five stages of loading, including: 12.5 kPa, 25 kPa, 50 kPa, 100 kPa, and 200 kPa
(Figure 4.9). The soil samples were 20 mm thick by 50 mm in diameter, respectively.

127
The settlement and pore water pressure data were collected continuously for 24 hours
for each load increment using a data logger.

(a) (b)
Figure 4.9 Consolidation test, (a) placing the prepared sample and (b) oedometer
apparatus connected to the data logger

The data collected from the oedometer tests were analysed to obtain the
consolidation parameters of the reconstituted samples. According to Taylor (1948),
the variation of permeability (k) with void ratio (e) for the clays can be determined
using the following equation:

௘బ ି௘
Ž‘‰ ݇ ൌ Ž‘‰ ݇଴ െ (4.1)
௖ೖ

where, ck is the permeability change index, k0 and e0 are the initial coefficient of
permeability and the void ratio, respectively. The consolidation results indicated that
the permeability change indices (ck) for the reconstituted samples S1 and S3 were 0.3
and 0.72, respectively. The variations of permeability against the void ratio, for both
samples, are shown in Figures 4.10 and 4.11. Table 4.5 shows the permeability of the
samples under a surcharge of 20 kPa, which is simulating 1m of working platform
installed before ground improvement.

128
 1.8
Q38 Kaolinite = 70%
1.6 ActiveBond 23 Bentonite = 15%
Fine sand = 15%
Void ratio (e) 1.4

1.2 ck=0.34

0.8

0.6

0.4
1.00E-10 1.00E-09 1.00E-08 1.00E-07
log k (m/s 2)

Figure 4.10 Variation of permeability against void ratio (sample S1)

1.8
Q38 Kaolinite = 65%
1.6 ActiveBond 23 Bentonite = 15%
Fine sand = 20%
1.4
Void ratio (e)

1.2
ck=0.72
1

0.8

0.6

0.4
1.00E-10 1.00E-09 1.00E-08
log k (m/s2)

Figure 4.11 Variation of permeability against void ratio (sample S3)

Table 4.5 Permeability of mixtures (Surcharge = 20 kPa)

Soil reference number S1 S2 S3

Permeability
1.060E10-8 7.446E10-9 2.710E10-9
(m/s2)

In this research the permeability ratio (kh/ks) of 4 and an extent ratio (rs/rw) of 3 were
used to conduct the consolidation test assisted with a vertical drain. Samples S1 and
S3 were chosen as the soils for the intact and smear zones, respectively, to meet the
permeability ratio criteria. The variations of void ratio versus effective stress for
samples S1 and S3 are plotted in Figures 4.12 and 4.13, respectively.

129
 1.6
Q38 Kaolinite = 70%
ActiveBond 23 Bentonite = 15%
1.4 Fine sand = 15%
Void ratio (e)
1.2

0.8

0.6
10 100 1000
Log σ´
Figure 4.12 Variation of void ratio versus effective stress (sample S1)

1.6
Q38 Kaolinite = 65%
1.4 ActiveBond 23 Bentonite = 20%
Fine sand = 15%

1.2
Void ratio (e)

0.8

0.6

0.4
10 100 1000
Log σ´

Figure 4.13 Variation of void ratio versus effective stress (sample S3)

Table 4.6 indicates the properties (permeability and extent) of the undisturbed/intact
zone, the disturbed/smear zone, and the vertical drain. The soil properties used to
conduct the numerical analysis are summarised in Table 4.7.

Table 4.6 Properties of the intact zone, the smear zone, and drain

Area Intact zone Smear zone Vertical drain

2
Permeability(m/s )
1.060E10-8 2.710E10-9 4.50E10-3
(ߪ௭ᇱ =20 kPa)
Radius (mm) 125 33 11

Selected Soil S1 S3 Uniformly graded sand (SP)

130
 Table 4.7 Adopted properties for the numerical simulation
γdry k
Sample Soil type M λ κ μ e○ kh/ kv
(kN/m3) -9
(10 m/s)
S1 Intact zone 1.1 0.37 0.049 0.3 1.57 10.4 18 1
S3 Smear zone 1.15 0.34 0.054 0.3 1.55 10.6 4.5 1

Note: M is the slope of the critical state line; λ and κ are slopes of the specific volume versus
ln(p´) curves for compression and swelling, respectively, where p´ is the mean effective
stress; μ is the Poisson’s ratio; e○ is the initial void ratio; and k is the permeability in the
axisymmetric condition.

4.2.3 Preparation of Rowe cell and initial sample

The following steps were carried out to prepare the Rowe cell apparatus for the
consolidation test:

x Applying a vacuum pressure to the water tank for the de-airing process,

x Connecting the pore pressure transducers, pressure lines, drainage and de-airing
pipes, and data transfer cables according to the illustrated diagram in Figure 4.4,

x Running the software to check the functionality of pressure controllers, infinite

volume controllers, and data loggers,

x De-airing process of the pore pressure transducers, and

x Levelling the main body of the Rowe cell.

After the initial preparation of the Rowe cell, two thin PVC and brass pipes with
diameters of 66 mm and 22 mm (rs/rw=3) respectively, were placed at the centre of
the cell at the base to act as the smear zone and vertical drain boundaries,
respectively. The pipe for the vertical drain was covered with a filter paper to prevent
the smear zone material from mixing with the sand used in the vertical drain. The
pipes and a cross sectional view of the cell are shown in Figures 4.14a to 4.14c.

131
 Inner pipe
(Boundary of vertical drain)
Outer pipe
Filter (Boundary of smear zone)

(a) (b)

Cell circumference
Outer pipe Smear zone
Inner pipe
Intact zone
Vertical drain
22 mm

66 mm
125 mm

(c)
Figure 4.14 Placing PVC and brass pipes as the smear zone boundary and the
vertical drain border, (a) top view, (b) side view and (c) a typical cross section of
the Rowe cell

The area between the circumference of the cell and the boundary of the smear zone
was filled with slurry that was prepared previously, based on the mixture and design
of the sample S1 from Table 4.3. The area between the two pipes (the smear zone)
was filled with a mixture of 65% kaolinite, 20% bentonite and 15% sand, and a water
content of 120% (sample S3 from Table 4.3). The top surfaces of the both areas were
levelled and covered with geotextiles. Extra blocks were placed on top of the pipes to
stabilise them and minimise any chance of sliding, and then a steel ring was placed

132
on the intact zone and the surfaces of the smear zone. The sample then was left in
this condition for 24 hours to stabilise (Figure 4.15).

(a) (b)

Figure 4.15 Sample placement, (a) filling the undisturbed area (intact zone) with the
prepared soil and (b) the setup after placing PVC and Brass pipes as the smear zone
boundary and vertical drain border

4.2.4 The pre-consolidation process and preparation of the final

sample
An initial pre-consolidation pressure of 20 kPa was applied before installing the
drain to stabilise the sample in the intact and smear zones. A number of 20mm thick
steel rings were manufactured and then placed on top of each zone for pre-
consolidation. One (1.8 kPa) ring was added each day to obtain the required pre-
consolidation load. The first ring of each set was designed to allow water to drain
from the top surface during pre-consolidation. The level of water was checked
continuously to keep the sample saturated. The pre-consolidation process is shown in
Figure 4.16.

133
 (a) (b)

(c) (d) (e)

Figure 4.16 Rig set up, (a) geotextile filters, (b) pre-consolidation loading rings, (c)
the first two loading rings with drainage grooves and holes, (d) placing of the first
loading ring and (e) full loading condition

After completing the pre-consolidation process the following steps were carried out
to prepare the sample for the main consolidation test:
- All but the last loading ring was removed.
- Sand was poured into the inner pipe to act as a vertical drain and to compact the
sand (Figure 4.17a),
- The outer pipe was removed (Figure 4.17b),
- The inner pipe was removed and an extra part of the filter paper was cut (Figures
4.14a and 4.18b),
- The top surface was levelled (Figure 4.19a),
- A geotextile filter and a porous metal plate were placed on the top surface (Figure
4.19b),

134
- The cell was filled with water (Figure 4.20a),
- A porous stone was placed at the centre of the cell to allow water to drain through
the settlement rod,
- The top of the cell, and the diaphragm and settlement rod were placed in position
(Figure 4.20b), and

- The top part of the cell was bolted to the body.

(a) (b)

Figure 4.17 Testing procedures, (a) Pouring the vertical drain material and (b)
Pulling out the outer pipe

(a) (b)

Figure 4.18 Testing procedures, (a) pulling out the inner pipe and (b) cutting the
extra part of the filter paper

135
 (a) (b)

Figure 4.19 Testing procedures, (a) leveling the top surface and (b) placing the
geotextile on top surface

(a) (b)

Figure 4.20 Testing procedures, (a) filling the cell with water and (b) placing the cell
top

4.2.5 Initial drainage and de-airing of the Rowe cell system

The cell and back pressures were set at 110 kPa and 100 kPa, respectively, to fill the
geomembrane diaphragm with de-aired water and carry out a de-airing process based
on the diagram shown in Figure 4.21. Water was pumped into the geo-membrane
diaphragm through the valve V1 which increased the volume of the diaphragm by the
application of 110kPa of pressure. At the same time, the de-airing screw on top of
the cell was opened to allow the air trapped in the diaphragm to escape, after which
the bolts were tightened again.

136
 Water Tank

P/V C

P/V C
P/V C

P/V C
IVC V5 V4
IVC
De-airing valve 1.0 m
F A De-airing valve
Cell pressure

Back pressure
G Settlement
Cell top rod

V2
De-airing
V3

O E V1 screw
B

Diaphragm
C D
Drainage & Porous plate
Watter
de-airing valve Drainage &
Cell body de-airing valve

Figure 4.21 Schematic diagram of the de-airing process

The air trapped between the diaphragm, the body of the cell, and the porous plate
was drained out due to a sudden increase in the volume of the diaphragm. By closing
valve V2 and opening valves V1 and V3, the trapped air escaped through the outlet
point O, passing the path DEO (i.e. connecting points D, E, and O). By opening
valve V4, water in the tank was discharged through the connecting points A, B, C, D,
E, and O. In this way, the air trapped in the pipe between points B and C was drained
from the outlet point O. Valves V3 and V4 were closed after this stage was
completed.
The air trapped in the pressure lines and controller devices was drained
through the de-airing valves installed on the infinite volume controller (IVC). To do
this, valves V1 and V2 were closed and valves V4 and V5 were opened; this allowed
the water to flow from the tank to the path and push the trapped air through the de-
airing valves. This process was repeated several times to ensure that the system was
completely de-aired. It should be noted that if the air is not completely removed, the

137
resulting change in volume, and the settlement and pore water pressure
measurements may not be accurate.

4.2.6 Vertical drain assisted consolidation test procedure

Although it was presumed that the soil sample was fully saturated as slurry with a
moisture content well above the liquid limit, a pressure of 110 kPa and a back
pressure of 100 kPa were applied for 24 hours to ensure full saturation. The criterion
ο௎
for a fully saturated condition and obtaining a pore pressure coefficient (ߚ ൌ οఙ
) of

more than 95%, was reached.

In this study, four loading stages (20 kPa, 50 kPa, 100 kPa, and 200 kPa)
were applied to conduct the vertical drain assisted consolidation tests. The cell and
back pressures applied in each stage of loading are summarised in Table 4.8. The
large Rowe cell with all its connections is shown in Figure 4.5.

Table 4.8 Details of consolidation loading stages

Loading
Applied effective Cell pressure Back pressure
Loading stage duration
pressure (kPa) (kPa) (kPa)
(day)
1 20 120 100 5
2 50 250 200 14
3 100 300 200 28
4 200 400 200 28

The loads mentioned above were applied instantaneously and then maintained for the
duration mentioned in Table 4.8, before moving on to the next stage.

4.3 TEST RESULTS

As mentioned earlier, the Rowe cell was filled with the reconstituted soil samples
(Table 4.3) to evaluate the undisturbed region and the smear zone with a circular
sand drain at the centre. A consolidation test was conducted after the pre-
consolidation process (i.e. 20 kPa) was completed. A consolidation surcharge was
applied by adjusting the water pressure inside the membrane diaphragm that had
been placed on top of the pre-consolidated sample. The time dependent vertical
displacements of the sample were recorded using GDSLab software, with an LVDT
(Linear Variable Differential Transformer) transducer. Several pore pressure

138
transducers were used to monitor the dissipation of excess pore water pressure during
consolidation. The next stage of loading commenced when the excess pore water
pressure from the previous stage had been totally dissipated (i.e. it reached less than
1 kPa). Figure 4.22 illustrates the instrumentation plan for the 250mm Rowe cell.

LVDT 1

110 mm

135 mm
PWP A4
160 mm

182 mm PWP A3
PWP A2
PWP A1
PWP B5

PWP B1 PWP B4
PWP B3 PWP B2

(a)
Smear zone boundary

Vertical drain boundary

PWPT B5

102 mm PWPT B2
m
m
27
75 mm

52
mm
PWPT B1 PWPT B3

PWPT B4
(b)

Figure 4.22 Schematic diagram of the instrumentation plan, (a) plan view of the
body of Rowe cell and (b) the cross section of bottom of the Rowe cell

Figure 4.23 shows the settlement-time curve along with the loading history. It could
be seen that the primary settlement of the sample was about 48 mm after a load of

139
200 kPa was applied. The total settlement at the end of each loading stage is
presented in Table 4.9.

Time (day)
0 5 10 15 20 25 30 35 40 45 50 55 60
0 300
Measured soil surface settlement
Loading history
10 250

20 200

Surcharge (kPa)
Settlement (mm)

30 150

40 100

50 50

60 0

Figure 4.23 Settlement and corresponding surcharge versus consolidation time

Table 4.9 Surface settlement at the end of each loading stage

Applied stress (kPa) Total settlement (mm)

20 6
50 21
100 35.5
200 48

Note: Ss is the surface settlement; and h is the initial height of the sample

Readings of pore water pressure transducers were used to monitor the dissipation of
excess pore water pressure and evaluate the rate of consolidation. The location of the
selected pore water pressure transducers are shown in Figure 4.22. The plotted
graphs in Figures 4.24 and 4.25 indicate the variation of excess pore water pressure
versus consolidation time at different loading stages.

140
 PWPT B5 (r=102 mm)
200 PWPT B4 (r=75 mm) 200
PWPT B3 (r=52 mm)
PWPT B2 (r=27 mm)
Loading Histroy
Excess pore water pressure (kPa)

150 150

Surcharge (kPa)
100 100

50 50

0 0
0 10 20 30 40 50 60
Time (day)

Figure 4.24 Measured excess pore water pressure at transducers located on the
bottom of the cell (r is measured from the centre of the drain)

200 200
PWPT A4 (h=90 mm)
PWPT A3 (h=64 mm)
PWPT A2 (h=40 mm)
Excess pore pressure (kPa)

PWPT A1 (h=18 mm)

150 150
Loading histories
Surchrge (kPa)

100 100

50 50

0 0
0 10 20 30 40 50 60
Time (day)

Figure 4.25 Measured excess pore water pressures from transducers located on the
sides of the cell (h is measured from the bottom of the impervious boundary of the
cell)

141
It was observed that the pore water pressure transducer B1 (PWPT B1), located
inside the vertical drain area (at the centre of the Rowe cell), recorded insignificant
excess pore water pressure during consolidation.
According to Figures 4.24 and 4.25, the excess pore water pressure was
fully dissipated when all the loading stages had ended, which confirmed reaching the
end of primary consolidation. That was the criterion used to commence the next
stage of loading. It was noticed that the excess pore water pressure had increased
almost immediately after increasing the surcharge, but this rise in pore water pressure
was slightly less than the increase in the applied pressure, which could be related to
air trapped in the system. According to Figures 4.23-4.25, the settlement has been
slightly increased after complete dissipation of pore water pressure in each stage of
loading, which can be related to the creep phenomenon (Le et al., 2012; Fatahi et al.,
2013).
The variations of normalised excess pore water pressure with vertical and
radial distances at certain consolidation times are shown in Figures 4.26 and 4.27,
respectively. The readings from the pore water pressure transducers located on the
outer impervious wall and the base of the cell have been normalised dividing by
surcharge of 200 kPa.

Normalised excess pore water pressure (EPWP/Surcharge)

0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35
100
Vertical distance from the base of the cell (mm)

80
70 mm

60

40

20

Consolidation time=38 day Consolidation time=40 day

Consolidation time=42 day Consolidation time=44 day
0

Figure 4.26 Variation of excess pore water pressures with the vertical distance from
the base of the cell

142
 0.35
Consolidation time=38 day Consolidation time=40 day
Consolidation time=42 day Consolidation time=44 day
Normalised wxcess pore water pressure

0.30

0.25

0.20

0.15

0.10

0.05

0.00
0 20 40 60 80 100 120
Radial distance from drain (mm)

Figure 4.27 Variation of excess pore water pressures with the radial distance from
the centre of the drain

Figure 4.26 shows that the excess pore water pressure changed inversely with the
vertical distance from the impermeable base, i.e., the excess pore water pressure
increased when the distance from the impervious bottom boundary decreased. In the
other words the longer vertical distance from the top drainage boundary resulted in a
higher remaining excess pore water pressure. For example, the excess pore water
pressure at PWPT A1 (h=18 mm) was 62 kPa after 38 days of consolidation, but this
was reduced by 70% to 18 kPa at PWPT A2 (h= 40 mm). Furthermore, it was noted
that the changes in excess pore water pressure were insignificant in one third of the
soil sample (70 mm from the top pervious boundary), indicating that vertical
consolidation dominated in this zone and thus the excess pore water pressure
dissipated faster.
Figure 4.27 shows that the excess pore water pressure followed an
incremental trend when the radial distance from the drain increased, for example,
increasing the radial distance from 40 mm to 100 mm resulted in a 70% rise in the
excess pore water pressure and a change from 35 kPa to 60 kPa. Moreover, it can be

143
observed that the excess pore water pressure increased at a sharper rate inside the
smear zone in the initial loading stages. Table 4.10 compares the increased rate of
excess pore water pressure (kPa/m) with the radial distance, inside and outside the
smear zone at different consolidation times.

Table 4.10 The EPWP increase rate from the centre of the drain to the boundary of
smear zone for selected consolidation times

Consolidation time (day) 38 40 42 44

EPWP increase rate inside the smear zone (kPa/m) 1.00 0.45 0.22 0.14

EPWP increase rate outside the smear zone (kPa/m) 0.44 0.38 0.21 0.12

4.4 VERIFICATION OF THE NUMERICAL CODE

The laboratory measurements were used to validate the developed numerical code,
and an axisymmetric finite difference model was used to simulate the Rowe cell. The
hydrostatic water pressure was fixed along the centre line of the cell to model the
vertical drain boundary. The loading stages were simulated by applying a uniformly
distributed load on top of the cell. The radii of the smear zone and undisturbed area
were defined as 33 mm and 125 mm, respectively. It should be mentioned that the
initial height of the pre-consolidated sample was 121 mm. The modified Cam-Clay
(MCC) constitutive model was used to analyse the consolidation of the reconstituted
sample using the soil properties listed in Table 4.7. Figure 4.28 demonstrates the
Rowe cell mesh grid that was generated using the numerical code.

144
 Axisymmetric P (Surcharge: 20kPa 200kPa) Top of the Rowe cell
boundary

u=0

Periphery of the Rowe cell

(Impermeable boundary)

h=121 mm
Drain
boundary

rs=33 mm
Smear zone
Intact zone

R=125 mm

u= h Jw

rd=11 mm Bottom of the Rowe cell

((Impermeable
p boundary)
y)

Figure 4.28 Sample of the grid pattern for the simulated Rowe cell using the
developed code (rs= smear zone extent, and R= width of the cell), and the condition
of the pore water pressure distribution boundary along the vertical drain.

The discretised finite difference mesh consisted of quadrilateral elements, where only
half of the Rowe cell was used by exploiting symmetry. FLAC was used to sub-
divide each quadrilateral element into triangular elements, and the pore water
pressures were assumed to vary linearly in a triangular element.
The hydrostatic pore water pressure along the vertical drains was used to
model the PVD. Zero excess pore water pressure was assumed at the ground surface
to model the drainage boundary. The number of fluid and mechanical sub-cycles
needed to keep the maximum unbalanced force ratio (or unbalanced force) below a
predefined value using “SOLVE auto on age=consolidation time” command was
adjusted automatically. The developed numerical code and related sub-routines to
simulate the Rowe cell are presented in Appendix B.
A parametric study was carried out by applying the proposed back
calculation procedure integrated with the developed numerical code to accurately
determine the properties of the smear zone. A number of cases with a constant
extent ratio (rs/rw) and different permeability ratios (kh/ks) were used to conduct the

145
parametric study, and they are summarised in Table 4.11. The numerical settlement
predictions were compared with the laboratory measurements and are shown in
Figure 4.29.

Table 4.11 Applied combinations of smear zone permeability and extent in numerical
analyses

Case Permeability ratio (kh/ks) Extent ratio (rs/rw)

1 2 3
2 3 3
3 4 3
5 5 3

Time (day)
0 10 20 30 40 50 60
0
Laboratory measurments
Case 1
-10 Case 2
Case 3
Case 4
Settlement (mm)

-20

-30

-40

-50

-60

Figure 4.29 Comparison between predicted numerical settlements and laboratory

measurements

Figure 4.29 shows that the numerical settlement curves corresponding to Case 3
(rs/rw=3 & kh/ks=4) and Case 4 (rs/rw=3 & kh/ks=5) can predict the consolidation of
the sample more accurately than the other cases. It should be noted that selecting the
optimum case based on observation is a challenging task, so a back calculation
procedure was proposed to rectify this problem. Figure 4.30 illustrates the calculated

146
final cumulative errors (Equation 3.51) for the selected cases using the proposed
back calculation procedure (Figure 3.13).

0.250
0.224

0.200
Accumulative error

0.150 0.145

0.097
0.100 0.086

0.050

0.000
Case 1 Case 2 Case 3 Case 4

Figure 4.30 Calculated final cumulative error for the selected cases by using the
proposed back calculation procedure

The back calculation results indicate that in Case 1 where rs/rw=3 and kh/ks=2 had the
maximum cumulative error, while the minimum error belonged to Case 3 with smear
zone properties of rs/rw=3 and kh/ks=4. Therefore, Case 3 indicated the best
agreement with the laboratory measurements, which matched the smear zone
properties of the sample used in the test. The results of the parametric study
confirmed that the designed back calculation procedure integrated with the
developed numerical code can be used to predict the smear zone properties more
accurately and reliably.
The numerical predictions for the variations of excess pore water pressure at
transducers A1 and B3 were compared with the laboratory measurements that are
shown in Figures 4.31 and 4.32.

147
 240 240
Lab measurements at PWPT A1
Numerical predictions
Loading history
Excess water pore pressure (kPa) 200 200

160 160

Surcharge (kPa)
120 120

80 80

40 40

0 0
0 10 20 30 40 50 60
Time (day)

Figure 4.31 Variation of excess pore water pressures at PWPT A1 versus

consolidation time

240 240
Lab measurements at PWPT B3
Numerical predictions
Loading history
200 200
Excess water pore pressure (kPa)

160 160
Surcharge (kPa)

120 120

80 80

40 40

0 0
0 10 20 30 40 50 60
Time (day)

Figure 4.32 Variation of excess pore water pressures at PWPT B3 versus

consolidation time

148
Figures 4.31 and 4.32 show that the excess pore water pressure curves for the
numerical predictions and laboratory measurements followed a similar pattern, where
the pore water pressures increased at each loading stage and then dissipated with
time. Figures 4.31 and 4.32 show that the numerical predictions at PWPT A1 and
PWPT B3 agreed with the consolidation test data, although there are some
discrepancies.
The graphs in Figures 4.31 and 4.32 indicate that the increase in excess pore
water pressures at the commencement of each loading stage that were predicted
numerically, was equal to the change in the applied load, but the measured excess
pore water pressures were slightly less than the predictions. According to Robinson
(1999), the maximum value of the excess pore water pressure is usually less than the
pressure increment applied, an effect that may be related to the relationship between
the volumetric compliance of the pore water pressure measuring system and the
volume compressibility of the soil skeleton (Whitman et al., 1961; Gibson, 1963;
Perlof et al., 1965; Sonpal & Katti, 1973).
The overall trends in Figures 4.31 and 4.32 confirmed the validity of the
developed FLAC code in modelling the preloading process with prefabricated
vertical drains to back calculate the smear zone properties.

4.5 EVALUATION OF AXISYMMETRIC TO PLANE-STRAIN

CONVERSION METHODS

As explained earlier in Chapter 2, converting the coefficient of permeability from an

axisymmetric to a plane-strain condition is essential to conduct two dimensional (2D)
numerical analyses. According to the available literature, different conversion
methods have been proposed by researchers (e.g. Hird et al., 1995; Indraratna and
Redana, 1997; Lin et al, 2000; Chai et al, 2001; Tuan and Toshiyuki, 2008). In this
study, the accuracy of different conversion methods was evaluated by comparing the
2D numerical predictions with the results of the axisymmetric analysis. According to
the laboratory measurements, the smear zone permeability ratio (kh/ks) and an extent
ratio (rs/rw) of 4 and 3 should be adopted in the axisymmetric analysis. The adopted
conversion methods are summarised in Table 4.12.

149
Table 4.12 Adopted equations for converting permeability from axisymmetric to
plane-strain condition

Method Intact zone permeability Smear zone permeability Reference

ʹ݇௛
݇௘௣ ൌ Hird et al.
1 ܴ ݇ ͵
͵ ൤݈݊ ቀ ቁ ൅ ൬ ௛ ൰ ݈݊ሺ‫ݏ‬ሻ െ ൨ (1992)
‫ݎ‬௦ ݇௦ Ͷ

ߨ݇௛
݇௘௣ ൌ Lin et al.
2 ݊ ݇ ͵
͸ ൤݈݊ ቀ ቁ ൅ ൬ ௛ ൰ ݈݊ሺ‫ݏ‬ሻ െ ൨ (2000)
‫ݏ‬ ݇௦ Ͷ

݇௛௣ ߚ
݇௦௣ ൌ
݇௛௣ ݊ ݇௛
൤݈݊ ቀ ቁ ൅ ൬ ൰ ݈݊ሺ‫ݏ‬ሻ െ ͲǤ͹ͷ൨ െ ߙ Hird et al.
݇௛ ‫ݏ‬ ݇௦
(1992)
ʹ ሺ݊ െ ‫ݏ‬ሻଷ combined
3 ͲǤ͸͹݇௛ ߙൌ with
݇௛௣ ൌ ͵ ሺ݊ െ ͳሻ݊ଶ
ሾ݈݊ሺ݊ሻ െ ͲǤ͹ͷሿ Indraratna et
ʹ ሺ‫ ݏ‬െ ͳሻ al.
(ignoring the smear effect) ߚൌ ሾ͵݊ሺ݊ െ ‫ ݏ‬െ ͳሻ
͵ ሺ݊ െ ͳሻ݊ଶ (2005a)
൅ ሺ‫ ݏ‬ଶ ൅ ‫ ݏ‬൅ ͳሻሿ

݇௛௣ ߚ
݇௦௣ ൌ
݇௛௣ ݊ ݇௛
൤݈݊ ቀ ቁ ൅ ൬ ൰ ݈݊ሺ‫ݏ‬ሻ െ ͲǤ͹ͷ൨ െ ߙ Lin et al.
݇௛ ‫ݏ‬ ݇௦
(2000)
ߨ݇௛
݇௛௣ ൌ ʹ ሺ݊ െ ‫ݏ‬ሻଷ combined
4 ͸ሾ݈݊ሺ݊ሻ െ ͲǤ͹ͷሿ ߙൌ with
(ignoring the smear effect) ͵ ሺ݊ െ ͳሻ݊ଶ
Indraratna et
ʹ ሺ‫ ݏ‬െ ͳሻ al.
ߚൌ ሾ͵݊ሺ݊ െ ‫ ݏ‬െ ͳሻ
͵ ሺ݊ െ ͳሻ݊ଶ (2005a)
൅ ሺ‫ ݏ‬ଶ ൅ ‫ ݏ‬൅ ͳሻሿ

Note: where B is the half-width of unit cell, kh and khp are axisymmetric and plane-strain
horizontal permeability values of the intact zone respectively, ks and ksp are the axisymmetric
and plane-strain horizontal permeability values of the smear zone respectively, kep is the
plane-strain equivalent horizontal permeability of the drain’s surrounding soil, n is the
spacing ratio equal to R/rw=B/bw, s=rs/rw, α, and β are geometric coefficients.

The radial consolidation equation proposed by Hansbo (1981) was applied by Hird et
al. (1992) to develop the axisymmetric to plane-strain permeability conversion
equation (Method 1). This equation was then used to match the rate of consolidation
in a plane-strain and axisymmetric unit cell by incorporating the equality of the
average degree of consolidation at every time and every level in the cell. The
thickness of the drain and smear zone for matching purposes was omitted by Hird et
al. (1992). Lin et al. (2000) assumed an equivalent discharge rate under confined
flow to develop the axisymmetric to plane-strain conversion equation (Method 2).

150
Indraratna et al. (2005a) derived the average excess pore water pressure between the
smeared and intact zones by using Darcy’s law in the axisymmetric to plane-strain
matching process (Methods 3 and 4). Here, the vertical drain system was converted
to equivalent parallel drain walls and the total capacity of the drain was assumed to
be the same in both systems to determine the equivalent width of the drain and drain
spacing. The radial consolidation solution proposed by Hansbo (1981) was equated
with the derived radial consolidation equation based on the average excess pore
water pressure in order to develop an equation for converting the axisymmetric to
plane-strain permeability (Indraratna et al., 2005a).
The plane-strain permeability coefficients of the intact and smear zones
were determined using the equations proposed in Table 4.8 to conduct 2D analyses,
and which are summarised in Table 4.13. It should be mentioned that methods 1 and
2 used the conversion equations to calculate an equivalent coefficient of permeability
to replace the combined smear and intact zones.

Table 4.13 Adopted permeability coefficients for 2D analyses

Method khp (m/s) ksp (m/s)

1 1.47×10-9 (combined)

2 1.16×10-9 (combined)

3 4.38×10-9 6.81×10-9

4 3.42×10-9 7.16×10-9

The permeability coefficients listed in Table 4.10 were used to conduct the plane-
strain analyses. The corresponding numerical settlement predictions are compared to
the results of the axisymmetric analyses and laboratory measurements in Figure 4.33.

151
 Time (day)
0 10 20 30 40 50 60
0
Stage 1 Lab measurements
Axisymmetric analysis (rs/rw=3 & kh/ks=4)
Method 1
Method 2
-10
Method 3
Method 4
Stage 2
Settlement (mm)

-20

-30
Stage 3

-40

-50

Figure 4.33 Variation of settlement versus consolidation time for different equations
used to convert permeability from an axisymmetric to a plane-strain condition

Figure 4.34 illustrates cumulative errors between the plane-strain and axisymmetric
settlement values calculated for each stage of loading, in order to systematically
evaluate the efficiency of the applied conversion methods.

120 111.2 Method 1 Method 2

Errors between plane-strain and

Method 3 Method 4
100
axisymmetric results

80.4
80

60.4 63.3
59.7
60
44.3
38.6
40 32.6
27.5
20.4
20 11.0 10.6

0
Stage 1 Stage 2 Stage 3

Figure 4.34 Cumulative error between plane-strain and axisymmetric results in

each loading stage

152
Figure 4.34 shows that using conversion method 1 (Hird et al., 1992) resulted in the
best fit between the plane-strain and axisymmetric results in the early stages of
consolidation (Stage 1), when there were significant changes in the void ratio and
permeability. The plotted charts in Figure 4.34 indicate that the accuracy of Methods
3 and 4 increased by a gradual growth in the surcharge pressure and reduction of the
void ratio variation rate. It can be seen that in the last two loading stages (Stage 2 and
3) a minimum error occurred when the equation proposed by Lin et al. (2000) was
combined with Indraratna et al. (2005a) approach to conversion (Method 4) to
determine the permeability of the intact and smear zones for a plane-strain condition.
The variations of excess pore water pressure predicted at the location PWPT
A1 for the permeability conversion equations are plotted and compared with the
laboratory measurements in Figure 4.35.

250 250

Laboratory measurements at PWPT A1

Method 1
Method 2
200 Method 3
200
Excess pore water pressure (kPa)

Method 4
Loading Histroy

150 150

100 100

50 50

0 0
0 5 10 15 20 25 30 35 40 45 50
Time (day)

Figure 4.35 Variation of predicted excess pore water pressure versus consolidation
time for the adopted equations for permeability conversion from axisymmetric to
plane-strain condition

Figure 4.35 shows that the excess pore water pressure graphs obtained from
conversion Methods 3 and 4 agreed with the laboratory measurements in the first
loading stage, but these methods resulted in an accelerated dissipation of pore water
pressure in the initial period of loading Stage 2 compared to the laboratory
measurements. In the last stage of loading (Stage 3), the results obtained from
Method 1 (Hird et al., 1992) had the best agreement with the test measurements. The

153
graphs shown in Figure 4.35 confirmed the settlement-time curves shown in Figure
4.33.

4.6 SUMMARY

In this chapter, a fully instrumented large scale Rowe cell apparatus was used to
investigate the effect of the smear zone on consolidation and to verify the numerical
procedure developed to back calculate the smear zone properties. Staged uniform
loads were applied to the sample by water pressure acting on a convoluted rubber
jack. An LVDT (Linear Variable Differential Transformer) and pore pressure
transducers were used to monitor the variations of settlement and excess pore
pressure dissipation respectively, during consolidation.
A number of samples were prepared by mixing different percentages of Q38
kaolinite, ActiveBond23 bentonite, and uniformly graded sand (SP) to conduct an
oedometer test and select proper mixtures for the PVD assisted consolidation test
using the Rowe cell. Based on the results of the odometer tests, two samples were
selected as the intact and smear zones materials that meet the criteria of kh/ks=4.
After preparing the Rowe cell, two 66mm diameter and 22mm diameter pipes
(rs/rw=3) were placed in the centre of the cell at the base, as the smear zone and
vertical drain boundaries, respectively. The area between the circumference of the
cell and the boundary of the smear zone was filled with slurry that was prepared on
the selected mix-design. The reconstituted sample was then pre-consolidated under a
20 kPa load by placing a number of 20mm thick steel rings on top of each other.
A vertical drain was installed by pouring fine sand inside the central pipe
after pre-consolidation was complete, and then both pipes were pulled out. A
geotextile filter and porous metal plate were placed on the top surface to provide a
one way drainage condition. The whole system was de-aired as the final stage of the
sample preparation process. Four loading stages (20 kPa, 50 kPa, 100 kPa, and 200
kPa) were applied to conduct the PVD assisted consolidation tests.
The developed numerical code was used to simulate the consolidation test.
Laboratory measurements were used to evaluate the validity of the numerical
analyses and the proposed back calculation procedure. The results indicated that the
back calculation procedure integrated with the developed numerical simulation can
reliably predict the smear zone properties.

154
 In addition, further numerical analyses were conducted to evaluate the
accuracy of the equations proposed by Hird et al. (1992), Lin et al. (2000), and
Indraratna et al. (2005a) for converting the coefficient of permeability from an
axisymmetric to a plane-strain condition, and also comparing the results of a 2D
plane-strain and axisymmetric numerical analysis. According to these results, using
conversion method proposed by Hird et al. (1992) resulted in more accurate
predictions in the early stages of consolidation while better predictions were obtained
in later stages of consolidation by applying the conversion equations proposed by
Indraratna et al. (2005a), combined with Lin et al. (2000) and Hird et al. (1992)
methods.

155
 CHAPTER FIVE

5 CASE STUDIES AND FURTHER VALIDATION

EXERCISES

5.1 GENERAL

According to the Australian Bureau of Statistics (2008), the population of Australia

would experience a growth of 60%, from 22 million to 35 million by 2056, and the
population of Queensland and New South Wales would expand by about 110% and
50% by 2056, respectively. This rapid growth in population and urbanisation has
increased the need to build and expand infrastructure such as highways, railways and
ports. In Australia, the major cities lie along coastlines, consisting of deposits of deep
soft clay, especially in Northern Queensland and New South Wales. These deposits
can be categorised as problematic soils due to their characteristic low bearing
capacity and high compressibility.
Finding efficient ground improvement techniques to modify the properties
of soft soil, while considering the project time limitation and construction cost, has
been a continuous challenge for construction companies. Various ground
improvement methods have been proposed to improve the strength of the soft soil.
Over the pasts two decades prefabricated vertical drain (PVD) assisted preloading
has been used in Australia as a very efficient method of ground improvement for
sites with deposits of deep soft soil.

In this chapter, five case studies of PVD assisted preloading (mainly

focusing on projects in Australia that have already been completed) were numerically
simulated. For all case studies, the properties of the smear zone accurately estimated,
using the proposed back calculation procedure to validate the applied method. The
developed FLAC code was used to conduct the numerical simulations. The PVD
assisted preloading projects used in this study are summarised in Table 5.1.

156
 Table 5.1 Summary of simulated case studies
Embankment PVD Drain
Preloading Soil
Case Study Location Height Length Spacing
Method Profile
(m) (m) (m)
Cumbalum
AU PVD assisted 5.0 22 1.35 DSS
TE
Ballina Bypass PVD/Vacuum
AU 8.5 24 1.0 DSS
TE assisted
Sunshine
AU PVD assisted 2.8 11 2.0 SSS
TE
Chittagong
BA PVD assisted 3.0 9 1.0 SSS
Sea Port TE
Large-scale
UOW PVD assisted 0.95 0.95 0.225 SS
Consolidometer
Note: AU: Australia, UOW: University of Wollongong (Australia), BA: Bangladesh, TE:
Trial embankment, DSS: Deep soft soil, SSS: Shallow soft soil, SS: Soft soil

It can be noted that these case histories were selected to cover different possible
preloading conditions and evaluate the validity of the back calculation approach in
each case. For example, the application of vacuum pressure was modelled to
simulate the Ballina Bypass trial embankment. The back calculation procedure was
verified for those cases with a shallow soft soil profile by modelling the Sunshine
and Chittagong trial embankments, while the deep soft soil profile was considered in
the Cumbalum and Ballina Bypass trial embankments. Furthermore, different drain
spaces (between 0.5 m to 2.0 m) were applied in these numerical analyses. A large
scale consolidometer test was numerically modelled to validate the use of the back
calculation method for laboratory investigations.

5.2 CUMBALUM TRIAL EMBANKMENT

5.2.1 Introduction
Expanding populations in the coastal regions of Eastern Australia are gradually
outgrowing the existing transport infrastructure. The North-South link connecting
Brisbane and Sydney is one of the most important routes and the volume of traffic is
growing exponentially. A large proportion of this coastal highway is currently being
upgraded, with much of the new construction confined to marginal land corridors
that had previously been avoided. The eastern seaboard of Australia is dominated by
a succession of easterly flowing rivers that drain the adjacent ranges into the Pacific
Ocean. In the lower reaches of these river valleys are extensive estuarine deposits

157
with complex geological structures (Bishop 2004). The site of the Ballina Bypass
was constructed over the floodplain and underlying estuarine deposits near the mouth
of the Richmond river in Northern NSW, as shown in Figure 5.1.

Cumbalum
Emigrant Trail Embankment
Creek
North N
Creek
Proposed
Bypass
Ballina Bypass
Trail Embankment

Pacific
Highway
Ballina
Town
Richmond Pacific
River Ocean

0 5 10km

Figure 5.1 A map of the Ballina Bypass upgrade route and surrounding surface
features (modified after Bishop, 2004)

The Cumbalum trial embankment was constructed near Cumbalum on the Pacific
Highway, 6.3 km north of Ballina, to provide field data to be used in the design of
the Ballina Bypass section of the Pacific Highway upgrade in New South Wales
(Figure 5.2). The embankment was constructed in 1998 by the Roads and Traffic
Authority (RTA), Northern Road Services. This trial embankment also formed part
of the northern abutment of a bridge over Emigrant Creek.
The Ballina Bypass traverses a flood plain associated with the Richmond
River and its tributary creeks (Bishop, 2004). Soft soil deposits within the flood plain
can be up to 25m thick. An embankment constructed on these soils may undergo a
considerable amount of settlement during construction, and the settlement process

158
may take a long time and post construction settlements may be large. PVD assisted
preloading can be used as a ground improvement method to accelerate the settlement
process.

Study Area

N
Cumbalum

Pacific
Highway

Cumbalum
North
Ballina

Cumbalum
Trial Embankment

Figure 5.2 Location of the Cumbalum trial embankment (courtesy of Google Maps)

5.2.2 Geological Model and Subsoil Condition

Bishop (2004) developed a geological model for the Richmond River floodplain that
includes Quaternary marine and estuarine deposits overlying Tertiary volcanic and
Palaeozoic deformed and folded meta-sediments. Three distinct stages of Quaternary
deposition have been identified in the geological model proposed by Bishop (2004)
as follows:

159
x Stage 1: 1m to 2m thick deposits of fluvial sandy gravels overlayed by very stiff
oxidised clays. These deposits have been eroded and exist in isolated mounds from
depth to RL -15m AHD. A highly oxidised and indurated alteration horizon overlies
the Stage 1 deposits.

x Stage 2 and 3: deposits graded from gravels and sandy clays at lower levels to
dark grey shelly muds in the upper levels. Stage 3 deposits are Holocene in age, and
uncomfortably overlay the Stage 2 deposits that mainly comprise clays above RL -
10m AHD. The Stage 3 deposit is known as Pimlico Clay.

Within the Pimlico Clay deposit, Bishop (2004) infers that the presence of inter-
bedded estuarine sands and muds below RL -10m AHD indicates deposition under
high-energy conditions. The absence of sands above RL -10m AHD was interpreted
to correspond with the formation of a coastal barrier that created a low energy
estuarine deposition environment behind it. The interpreted section derived from
borehole logs under the Cumbalum study area (see Figure 5.1) is shown in Figure
5.3. MZ1 and MZ2 are highly oxidised and indurated horizons, which separate the
stages.

0m
Emigrant
Creek
Slope wash
10 Sands and gravels

Stage 3: Pimlico Clay

20
MZ1

30 MZ2
0 100m

Paleozoic Meta-sediments
40
Stage 2: Gundarimba Clay

Stage 2 channel
sands and gravels
Stage 3 channel gravels

Figure 5.3 Interpreted section for Cumbalum trial embankment study area (after
Bishop, 2004)

Site investigations for the Cumbalum trial embankment were conducted by Robert
Carr & Associates (RTA, 2000), and then developed further by Coffey International
Limited (2007).

160
The moisture content and Atterberg limits along the soil profile were determined by
conducting site investigations and laboratory testing (Figure 5.4). According to Kelly
(2008) the subsoil profile consists of lightly consolidated soft clay deposits from the
ground surface level to a depth of 10 m. A 2-m thick layer of soft clay is located at
the depth of 12 m and is surrounded by two silty sand lenses. The soil deposit
between the depth of 15 m and 22 m is categorised as firm clay.

Moisture content (%)

0 20 40 60 80 100 120 140 160
0

5
Depth (m)

10

15

20
Liquid limit Plastic limit Moisture content

Figure 5.4 Moisture content, liquid limit and plastic limit (after RTA, 2000)

5.2.3 Installation of Vertical Drains and Embankment

Construction

Once the drainage blanket (crusher and aggregate) had been completed at the
Cumbalum trial embankment, a series of vertical wick drains were installed. The
wick drains were designed to facilitate the flow of water from the soil to allow the
embankment to achieve its final settlement in a shorter period of time. Vertical drains
were installed at a 1.35-m triangular spacing over the entire area of the embankment
to a depth of approximately 22 m. The total number of drains installed was 2772,
with a total length of 61890 m.
The embankment was constructed between 1998 and 1999 to a nominal fill
thickness of 4.5 m. Subsequent site investigations through the trial embankment
showed that approximately 5 m of fill had actually been placed on site. The
embankment was then allowed to consolidate for about 4.5 years (Kelly, 2008). The

161
embankment material was compacted with an 18t pad foot roller and a 14t self-
propelled smooth drum roller. The material was spread out by a D3-Dozer (RTA,
2000). The construction history of the embankment is shown in Figure 5.5.

γ t = 21.5 kN/m3
Embankment Height (m)

φ = 35 °
4 c = 10 kPa
Stage 3
E = 7.5 MPa
PVD
3 Installation

2
Stage 2

Stage 1

0
0 250 500 750 1000 1250 1500 1750 2000

Time (days)

Figure 5.5 Construction history of Cumbalum trial embankment (after Kelly 2008)

The layout of the instrumentation in the embankment is shown in Figure 5.6.

Measurements at the settlement plate 9 (SP9) were used for the numerical
verification and parametric study. Figure 5.7 presents a cross section of the
embankment and subsoil profile at SP9.

SP6 I2
12 m

SP7
LEGEND
I3
Settlement plate
35 m

Piezometer
Bore hole settlement profile PT1 PC1
SP2 SP3 SP8 SP14
Inclinometer
Centre Line
PT2 PC2
SP4 SP9

SP1 SP5 SP10 SP15

PT3 PC3

I4
SP11 PC4

SP13 I5

Figure 5.6 Layout of the Cumbalum trial embankment (after RTA, 2000)

162
 Symmetric
axis
Settlement plate
+5.0 mRL
Surcharge fill
Berm Drainage blanket Crust
CL γ t=21.5 kN/m3
+1.77 mRL
0.0 mRL
SP9
Layer 1 (Soft Clay) 4.5 m

PC2

22m
Layer 2 (Soft Clay) 5.0 m
Symmetric flow and displacement

Layer 3 (Silty Sand) 1.0 m

boundary conditions

26 m
Layer 4 (Soft Clay) 2.0 m

Layer 5 (Silty Sand) 2.0 m

Layer 6 (Firm Clay) 4.0 m

Layer 7 (Firm Clay) 3.0 m

PVD (S=1.35 m)
4.0 m
Layer 8 (Firm Clay)

0m 35 m 100 m

Figure 5.7 Cross section of the Cumbalum trial embankment and subsoil profile

5.2.4 Numerical Simulation

The developed numerical code was used to simulate the Cumbalum trial
embankment. The finite difference mesh used for the 2D plane-strain numerical
simulation is shown in Figure 5.8. The vertical drains were modelled by fixing the
pore pressure to the hydrostatic pressure from the top to the bottom of the drain. A
constant reduced permeability was assigned to the area surrounding the vertical drain
in each layer to simulate the smear zone as shown in Figure 5.8. The modified Cam-
Clay model was used to simulate the behaviour of the soil in the intact region and the
smear zone, while the Mohr-Coulomb criteria elastic-perfectly plastic model was
used to simulate the embankment. The properties adopted for subsoil layers are
summarised in Table 5.2. The vertical drain system was converted into an equivalent
parallel drain wall by adopting Equations 2.30 to 2.32.

163
 Smear zone
k
permeability s1 Layer 1

Intact zone kh1

permeability kh2
Symmetric Smear zone
k Layer 2
axis permeability s2

5.0 m PVD

Layer 1

Layer 2

Layer 3
26.0 m Layer 4
Layer 5

Layer 6

Layer 7

Figure 5.8 Finite-difference meshes for the plane-strain analysis

Layer of
8 Cumbalum trial
embankment

Table 5.2 Adopted properties for subsoil layers for Cumbalum trial embankment near
SP9
γs kh kv
Depth: m λ κ e○ OCR
(kN/m3) -10
(10 m/s) (10 m/s)-10

0.0-0.5 0.7 0.042 2.2 17.5 2.04 1.02 3.0

0.5-5.0 0.84 0.118 2.87 15.0 1.71 0.85 2.0
5.0-10.0 0.95 0.134 3.4 14.0 1.71 0.85 1.5
10.0-11.0 0.125 0.031 2.61 18.0 1260 630 3.0
11.0-13.0 0.61 0.087 3.0 15.0 1.74 0.87 1.3
13.0-15.0 0.125 0.031 2.61 18.0 1740 870 3.0
15.0-19.0 0.47 0.067 2.08 17.0 2.94 1.5 1.3
19.0-22.0 0.335 0.047 2.08 20.0 2.94 1.5 3.5
22.0-26.0 0.335 0.047 2.08 20.0 2.94 1.5 2.5

Note: λ and κ are slopes of the specific volume versus ln(p’) curves for compression and
swelling, respectively, where p’ is the mean effective stress; e○ is the initial void ratio; γs is
the unit weight; kh is the horizontal permeability of intact zone; kv is the vertical permeability
of intact zone; and OCR is the over consolidated ratio

The FLAC code was applied to calculate the properties of the smear zone by
adopting the proposed back calculation procedure. The variation of settlement with
time is plotted in Figure 5.9. It should be noted that the settlement curve
corresponding to the case with smear zone characteristics of n=kh/ks=5 and

164
s=rs/rm=5 agrees with the field measurements, which shows the developed back
calculation procedure was validated in predicting the properties of the smear zone.

Consolidation Time (day)

0 250 500 750 1000 1250 1500 1750 2000
0.0
Field measurements (SP9)
installation

-0.5 Numerical predictions (n=5 & s=5)

PVD

Stage 2
End of
Settlement (m)

-1.0

-1.5
R2=0.99
-2.0
Stage 3
End of

-2.5

-3.0

Figure 5.9 Numerical parametric study results; Cumbalum trial embankment at SP9

The predicted and measured values of excess pore water pressure at PC2, at the depth
of 5.8 m, for the optimum combination of smear zone properties (n=kh/ks=5 and
s=rs/rm=5) are shown in Figure 5.10. The overall numerical and measured variations
in excess pore water pressure follow a similar trend in Figure 5.10, which agrees
with the settlement curves obtained in Figure 5.9, although some discrepancies can
be observed.

165
 120

Field measurements (depth 5.8m)

Excess pore pressure (kPa) 100
Numercal Predictions (n=5, s=5)
80

60

40

20

0
0 250 500 750 1000 1250 1500 1750 2000
Consolidation time (day)

Figure 5.10 Variation of excess pore pressure over time for PC2 at a depth 5.8m

According to Figure 5.10 the excess pore water pressure curves experienced a growth
during each stage of embankment construction, followed by a minor dissipation. The
field observations by Kabbaj et al. (1988) and Leroueil (1997) showed that the
maximum excess pore water pressure in the embankment occurred at the end of the
construction process, which is confirmed in Figure 5.9, for both field measurements
and numerical predictions. According to Figure 5.10 the maximum predicted excess
pore water pressure of 100 kPa occurred at the end of the construction process, while
this value for the field measurements was 80 kPa. This discrepancy can be attributed
to numerous factors such as the uncertainty of the soil properties, the effect of the
smear characteristics, inaccurate assumptions of soil behaviour, and an improper
conversion of axisymmetric conditions to plane-strain (2D) analysis of multiple
drains. The graphs in Figure 5.10 show that the predicted and measured excess pore
water pressures gradually dissipated after the trail embankment had been constructed.

166
5.3 CASE STUDY 2: BALLINA BYPASS TRIAL EMBANKMENT

5.3.1 Introduction

The Ballina Bypass is part of the Pacific Highway upgrade in New South Wales,
Australia. The NSW Minister for Planning approved the Ballina bypass project on 22
May 2003. The project provides 12 kilometres of dual carriageway, extending from
south of Ballina at the intersection of the Bruxner and Pacific highway to north of
Ballina at the intersection of Ross Lane at Tintenbar. According to Kelly and Wong
(2009) the soft soil deposits within the flood plain can be up to 25 m thick, which
may cause considerable settlement due to construction of the embankment. The
deepest soft soil deposit (25 m) belongs to the section located beneath the southern
abutment of a bridge crossing Emigrant Creek, near Cumbalum, as shown in Figure
5.11.

Trial N
Embankment

Ballina
Ballina
Emigrant
Creek

Pacific
Ocean

Figure 5.11 Location of the critical section and trial embankment of the Ballina
Bypass (courtesy of Google Maps)

167
Conventional staged construction using surcharge and wick drains was assessed to
take many years to complete due to the long consolidation periods required to
maintain stability between fill lifts, and this time was not predicted within the
timeframe for this project. The other ground improvement techniques such as soil
mixing, stone columns, and piled embankments were considered to be too costly for
such deep soft soil deposits. The membrane type of vacuum consolidation was
considered to be the best ground improvement technique for more rapid construction
than PVD assisted preloading, as well as being potentially more cost effective than
ground inclusions (Kelly and Wong, 2009). This was the first application of the
vacuum consolidation technique in Australia. A trial embankment north-west of
Ballina was constructed to investigate the performance of this approach. Figure 5.11
shows the location of the trial embankment.

5.3.2 Geological Model and Subsoil Condition

The geology at the site of the vacuum consolidation trial consists of a uniform
deposit of soft to firm silty clay overlying residual soil and bedrock. The silty clay is
thickest at 25m, at the northern end of the site, adjacent to Emigrant Creek. The
geological section for the Ballina Bypass trial embankment was interpreted by
Bishop (2004), and is shown in Figure 5.12. A highly oxidised and indurated
alteration horizon called MZ2 was identified, which affects the surface of the
basement rocks and the full thickness of the overlying Stage 1 channel gravels and
clays. Stage 3 overlies Stage 2, but it does not conform exactly, and it is separated by
a second oxidation horizon called MZ1.
Kelly et al. (2008) reported that the geology at the site of the vacuum
consolidation trial consists of a uniform deposit of soft to firm silty clay overlying
residual soil and bedrock.

168
 Duck Emigrant
Creek Creek
0m
0 100m
Stage 3: Pimlico Clay
10
MZ1

20

Stage 2: Gundarimba Clay Stage Dungarubba Clay

30

South Casino
MZ2 Gravel
40
Paleozoic Meta-sediments
Stage 1 channel gravels
50

Figure 5.12 Interpreted section for Ballina Bypass trial embankment study area (after
Bishop 2004)

5.3.3 Vacuum Consolidation and Embankment Construction

The vacuum consolidation system includes disconnected horizontal and vertical

drains where hydraulic connectivity occurs through the sand blanket. The system was
sealed by a 1mm thick membrane. Construction of the trial embankment began with
the placement of a 1.5 m thick layer of sand with a unit weight of 21.5 kN/m3 to act
as a working platform for plants and the drainage layer. Then 34mm diameter
circular vertical drains were installed, 1.0 m apart in a square pattern, and were then
connected to 50mm diameter horizontal drains. The vertical drains were installed by
a wick drain rig with a 110mm square base plate.
After instrumentation, a layer of fine grained sand was placed on top of the
granular material to protect the membrane. The membrane was installed and a layer
of fine grain sand was placed on top to protect it from punctures, and then general fill
was placed on top of the sand. The embankment was separated into two sections,
called Section A, the non-vacuum area, and Section B, the vacuum area. Figure 5.13
shows the location of the field instrumentation, including the surface settlement
plates and piezometers.

169
 M1 M2 M3
M4
SECTION A

SP1 SECTION B M5
P1
SP3
SP2 SP5 SP11
SP7
SP4 P2 SP9 P3
SP6
SP8 SP10
SP12
M6
Piezometer

Settlement plate
M9
M8 M7
Field measurements of SP12
applied to conduct parametric study

Figure 5.13 Layout of instrumentation for the trial embankment at Ballina Bypass
(modified after Kelly and Wong 2009)

The embankment was constructed in stages using a granular material (γs = 20 kN/m3)
up to a height of 8.5 m. A detailed cross section of the embankment and subsoil
profile (at the location SP12) is shown in Figure 5.14, and was used for numerical
modelling. A vacuum pressure of -70 kPa was applied for a period of 400 days and
then removed. The construction history of the trial embankment is shown in Figure
5.15.

Embankment
6.2 m CL
γt=21 kN/m3 Vacuum Membrane
1.5 m Sand Drainage Layer
0.8 m
SP12
Layer 2 3.5 m Soft Clay
Symmetric flow and displacement

P3
Layer 3 6.0 m Soft Clay
boundary conditions

24 m

Layer 4 5.0 m Firm Clay

Layer 5 Firm Clay

10.0 m

PVD (S=1.0 m)
0m 60 m 120 m
Figure 5.14 Cross section of the Ballina Bypass trial embankment and subsoil profile

170
 8 γt = 21 kN/m3
Embankment Height (m) φ= 35 °
c = 10 kPa Vacuum off
E = 8.0 MPa
6
Stage 3

Vacuum on
4

PVD Installation

2 Stage 2

Stage 1

0
0 50 100 150 200 250 300 350 400

Time (days)

Figure 5.15 Construction history of Ballina Bypass trial embankment

5.3.4 Numerical Simulation

The consolidation behaviour of soft clay beneath the Ballina Bypass trial
embankment was simulated using the developed FLAC code, incorporating the
modified Cam-Clay model. A fully saturated and coupled flow-deformation
simulation was carried out to model the dissipation of pore water pressures. The
discretised plane-strain finite difference mesh, composed of quadrilateral elements, is
shown in Figure 5.16(a), where only half of the trial embankment was considered by
exploiting its symmetry. FLAC sub-divided each quadrilateral element into
triangular elements, as shown in Figure 5.16(d). The pore pressures were assumed to
vary linearly in a triangular element.

171
 Symmetric
axis

6.2m

Layer 1

Layer 2
25m
Layer 3

Layer 4

Sd=1m (a)
-70 kPa Vacuum pressure at top

Ground surface
Vacuum pressure distribution

(Zero pore pressure)

PVD
(Hydrostatic pressure)
Vertical drain
a
Smear zone
Intact zone b

-70 kPa Vacuum pressure at bottom

(b) (c) (d)

Figure 5.16 (a) Discretised finite-difference mesh of Ballina Bypass trial

embankment, (b) Variation of vacuum pressure along the vertical drain, (c) The
pattern of meshes in the smear zone and undisturbed region, (d) FLAC zone
composed of overlaid triangular elements

To model the PVD, illustrated in Figure 15.16(c), the hydrostatic pore water pressure
was considered to be along the vertical drains (before applying the vacuum pressure),
and to model the drainage boundary the excess pore pressure at the ground surface
was assumed to be zero. According to the field measurements, uniform distribution
of vacuum pressure was considered and a constant negative pore pressure of -70 kPa
was applied along the vertical drain, as shown in Figure 15.16(b). The parameters
adopted for the layers of subsoil, based on the results of the site investigation and
laboratory tests used in the numerical analysis, are summarised in Table 5.3.

172
Table 5.3 adopted for the layers of subsoil for Ballina Bypass trial embankment near
SP12
γs kh kv
Depth: m λ κ e○ OCR
(kN/m3) (10 -10
m/s) (10 -10
m/s)
0.0 - 0.5 0.57 0.057 2.75 14.5 3.0 1.5 2.5
0.5 - 4.0 0.57 0.057 2.75 14.5 6.0 3.0 1.8
4.0 - 10.0 0.67 0.067 2.87 14.5 6.0 3.0 1.7
10.0 -15.0 0.47 0.047 2.61 15.0 15 7.5 1.3
15.0 - 25.0 0.40 0.040 2.09 15.0 15 7.5 1.2

Note: λ and κ are slopes of the specific volume versus ln(p’) curves for compression and
swelling, respectively, where p’ is the mean effective stress; e○ is the initial void ratio; γs is
the unit weight; kh is the horizontal permeability of intact zone; kv is the vertical permeability
of intact zone; and OCR is the over consolidated ratio

Field measurements at settlement plate No.12 (SP12) and piezometer P3 (Figure

5.14) in Section B, in conjunction with the soil profile data, were used with the
FLAC code to validate the developed back calculation procedure. The numerical
results and field measurements are compared in 5.17. Both the measured and
predicted vertical displacements indicated that the consolidation rate had accelerated
due to the installation of vertical drains and applying the vacuum pressure.

Consolidation Time (day)

0 50 100 150 200 250 300 350 400
0.0
-0.5 Field meaurements (SP12)
Numerical predictions (n=4 & s=4)
-1.0
Settlement (m)

-1.5
PVD Vacuum on
-2.0 installation
-2.5
Vacuum off
-3.0
-3.5
-4.0
-4.5
-5.0

Figure 5.17 Results of numerical parametric study; Ballina Bypass trial embankment
at SP12

According to the numerical results, the case with smear zone properties of s=rs/rw=4
and n=kh/ks=4 had the best fit with the field measurements, which verified the

173
validity of the back calculation procedure, when it was integrated with the developed
numerical code and then used for a reliable estimation of the properties of the smear
zone.
The comparison between the predicted and measured variations of excess
pore water pressure over time for the transducer P3 located 11.8 m deep, and 0. 5 m
away from the centreline of the embankment for the case with back calculated smear
zone properties of n=4 and s=4 are shown in Figure 5.18.

Consloidation time (day)

190 9
170 Field measurements (-11.8m)
8
Numerical predictions (n=4, s=4)
150 7
Vacuum Guage
130
Excess pore pressure (kPa)

Embankment height 6
110 5

Embankmnet height (m)

90 Vacuum on
4
70
3
50 Vacuum off
2
30
1
10
0
-10 0 50 100 150 200 250 300 350 400
-30 -1
PVD
-50 Installation -2

-70 -3
-90 -4
-110 -5

Figure 5.18 Variation of excess pore pressure over time for P3 at a depth 11.8m

Figure 5.18 shows that the calculared excess pore water pressure increases due to the
construction of each segment of the trial embankment and then gradually dissipates.
A sudden reduction of -70 kPa occurred in the numerical excess pore pressure curve
immediately after the vacuum pressure was applied, followed by an incremental
trend due to the embankment being constructed. In Figure 5.18, the measured excess
pore water pressures due to the staged construction of the trial embankment and the
application of negative vacuum pressure are plotted separately, while the numerical
predictions are plotted as a combined graph.
Some discrepancies can be observed between the measured and predicted
excess pore water pressures in Figure 5.18. It can be noted that the maximum
predicted excess pore water pressure occurred at the end of trial embankment
construction and were equal to the surcharge applied due to this construction, while

174
the value for the field measurements were slightly less than the applied pressure.
Robinson (1999) reported that the maximum value of excess pore water pressure is
usually less than the applied pressure increment, which can be related to the
relationship between volumetric compliance of the pore water pressure measuring
system and the volume compressibility of the soil skeleton.
Furthermore, field measurements showed that the excess pore pressure
values did not dissipate immediately at the end of loading or construction, but in fact
increased or stabilised for a period before decreasing. The abnormal behaviour of
excess pore water pressure has been reported in many field studies (Conlin and
Maddox, 1985; Kabbaj et al., 1988; Rowe and Li, 2002). Two main reasons have
recently been proposed to explain that anomalous behaviour; the Mandel-Cryer
effect and volumetric strain softening. Schiffman et al. (1969) reported that the
Mandel-Cryer effect is due to the increase in total stress caused by the volumetric
strain compatibility. The Mandel-Cryer effect is called after Mandel (1953) and
Cryer (1963), based on their observations related to the abnormal generation of
excess pore water pressure. Cryer (1963) analysed the process of consolidation by
applying pressure all around a saturated porous sphere. Since the surface of the
sphere is free to drain, the total stress at the centre of the sphere increased
temporarily under the applied pressure, because the dissipation of excess pore water
pressure at the centre was delayed. This results in an increasing excess pore water
pressure for some time before dissipation begins. In addition, the increase or delay in
the dissipation of excess pore water pressure may be the result of volumetric strain
softening due to unstable behaviour during consolidation when the stress paths depart
from the failure line. The constitutive model developed by Kimoto and Oka (2005)
can capture the increase in pore water pressure due to stagnation. Moreover, as
Asaoka et al. (2000) reported, since the decay of over-consolidation is much faster
than degradation of the structure in clay during consolidation, softening becomes
possible with volume compression even under a considerably low stress ratio.

175
5.4 CASE STUDY 3: SUNSHINE MOTORWAY TRAIL
EMBANKMENT

5.4.1 Introduction

The Sunshine Coast is one the fastest growing regions in Australia. The continued
economic and population growth has increased the pressure on the region’s main
traffic corridor, the Sunshine Motorway. According to the results of a site
investigation, the subsoil at the proposed development route consists of very soft and
saturated marine clays, which presented difficulties in developing the new alignment.
In order to discover how the foundation would respond under loading, and to
evaluate the effectiveness of various ground improvement techniques on these
marine clays, a fully instrumented trial embankment was constructed in 1992, located
in the Maroochy Shire, Queensland, Australia. This trial embankment was monitored
by the Queensland Department of Main Roads (QDMR), Brisbane, Australia. The
location of the study area is illustrated in Figure 5.19.

Study
Area

Brisbane

Figure 5.19 Location of the Sunshine Motorway study area, (courtesy of Google
Maps)

176
5.4.2 Subsoil Condition

The subsoil profile consists of sensitive silty clay with an approximate thickness of
11m, overlying a layer of sand with a thickness of about 6m. A thin soft clay layer
underlies the sand and extends to a depth of 18m, and another layer of sand is
encountered below 18m. The properties of the top sensitive silty clay with depth,
including the water content, Atterberg limits, bulk unit weight, undrained shear
strength and the compression index ratio are indicated in Figure 5.20.
According to Figure 5.20, the compression index ratio varies from 0.15 to
0.50. The recompression ratio was found to be about 10 times smaller than the
compression index ratio (QDMR Report, No.R1765, June 1991). The soft clay was
classified as lightly over consolidated soil because the over consolidation ratio
(OCR) varied from 1.0 to 1.6.

Water Content Grading % Bulk Undrained Shear Cc/(1+e0)

γt Particle Cv Organic
% Smaller LI PI % Strength (kPa) (m2/yr)
t/m3 Density (m2/yr) Content
than 2μ
%
100
120

(BPD) 0.1

0.5
0.3
20
40
60

10

25
20

10

30
80

30

15

35
40

20

40
5

1.68 1.7 19 0.9 5.3

2 1.52 2.63 2.5 27
1.36 2.60 2.3 28
2.6 35
3 1.40
1.34 2.4 40 0.4 10.7
4 55
Depth (m)

1.38 2.55 1.9 46

5
1.7 31
6
1.54 2.60 1.4 31
7 1.54 2.55 1.3 33
0.3 4.6
8 1.3 25

9 1.64 2.62 1.0 31

10
Old Alignment
- Liquid Limit - Vane Shea
- UU Triaxial – 50mm thin walled samples
New Alignment
- UU Triaxial – DH 19 100mm Piston samples
DH19 – Vane Shear
DH20 – Vane Shear

Figure 5.20 Profile of the Geotechnical characteristics (Sunshine Motorway Stage 2

Interim Report, 1992)

177
5.4.3 Installation of Vertical Drains and Embankment Construction

Figure 5.21 illustrates the base area of the trial embankment with dimensions of
approximately 90m by 40m, which was divided into 3 separate sections (Sections A,
B, and C). Vertical drains (Nylex Flodrain) were installed in a triangular pattern in
sections A and C with the drain spacing of 1m and 2m, respectively. Section B
represents the zone without PVDs.

35m 35m 20m

A B C 40m

Figure (triangular
PVD at 1m spacing 5.21 Planpattern)
view of Sunshine Motorway trial
PVD at 2m embankment
spacing (triangular pattern)

The subsoil conditions are relatively uniform throughout the site, and consist of
silty/sandy clay approximately 11 m thick, overlying a layer of dense sand
approximately 6 m thick. Of the sections available in this trial project, the section
with vertical drain spacing of 2 m was selected for numerical simulation.
Prefabricated vertical drains (Nylex Flodrain, 100×4 mm2) were installed in a
triangular pattern. A working platform 0.65 m thick (500 mm thick drainage layer
composed of 7 mm gravel, plus 150 mm of selected fill) was placed on the natural
ground surface for access by construction traffic (Sathananthan et al. 2008). PVDs
were installed from the working platform to a depth of 11 m. The embankment was
constructed in stages using a loosely compacted granular material (γt ≈ 19 kN/m3) up
to a height of 2.3 m (Sathananthan et al. 2008).
A schematic cross section of the trial embankment with the selected
instrumentation points is shown in Figure 5.22. Two 5m and 8m wide berms were
constructed on the sides of the embankment to rectify the problem of stability. The
location of the settlement gauges (SC) and piezometers (pneumatic-PP) are
illustrated in Figure 5.22.

178
 CL

17m Working Platform (0.65m)

5m 8m 1.7m

1.15m
SCC5/SCA1/SCB3
11m
PPA13/PPB31
Silty Clay
PPC43

Dense Sand

Figure 5.22 Typical cross-section of embankment with selected instrumentation

points (after Sathananthan et al. 2008)

5.4.4 Numerical Simulation

The Sunshine trial embankment was modelled numerically to verify the numerical
analysis using the developed FLAC code. The equivalent plane-strain model was
used to simulate the embankment, and the modified Cam-Clay model was adopted as
the soil constitutive model to simulate the behaviour of the soft clay. To model the
prefabricated vertical drains, a zero excess pore pressure boundary was considered
along the PVD. The smear zone was simulated by applying the mechanical properties
of undisturbed soil with a reduced coefficient of permeability. The instrumented side
of the trial embankment was simulated due to symmetry and to reduce the time
needed for numerical analysis. A detailed cross section of the simulated embankment
with the selected instrumentation point is shown in Figure 5.23. A sample of the
generated mesh pattern adopting the developed code is shown in Figure 5.24.

CL 8.5 m
PVD Working Platform
0.85 m γ t=19 kN/m 3
0.85 m
5m
0.5 m 20 m
0.65 m
P1
Symmetric flow and displacement

Silty clay (CL) Layer 1 2.5 m

boundary conditions

Soft silty clay (CL) Layer 2 2.5 m

2m
Silty clay (CL) Layer 3 6m

Dense Sand Layer 4 6m

Figure 5.23 Cross section of the Sunshine trial embankment and the subsoil profile

179
 l1
h4
h3 l2
h2 l3
h1
Smear
Layer 1 zone

Layer 2
hd ht Vertical
drain
Layer 3

Layer 4 rs
LI
sd

Figure 5.24 Sample of mesh grid pattern for an embankment considering the smear
zone and undisturbed region; (rs= smear zone extent; LI= intact zone extent; ht=
depth of the soil profile; and hd= length of the vertical drain, sd= drain spacing)

It should be noted that the equivalent plane-strain permeability to analyse the PVDs
was calculated using Equations 2.30 to 2.32, as proposed by Indraratna et al.
(2005b). Figure 5.25 illustrates the construction history of the embankment. The
properties of the subsurface ground profile are summarised in Tables 5.4 and 5.5.

3.5

3 Stage 3
Embankment Height (m)

2.5
Stage 2
2

1.5
Stage 1
1

0.5

0
0 10 20 30 40 50 60 70
Time (days)

Figure 5.25 Construction history of the trial embankment

180
Table 5.4 Adopted properties for the numerical simulation (after Sathananthan et al.
2008)
γs khp
Layer Soil type M λ κ ߤ e○ kh/ kv
(kN/m3) (10-9 m/s)
1 Silty clay 1.20 0.494 0.0494 0.3 1.6 16.4 9.72 2
2 Soft Silty clay 1.20 2.016 0.2016 0.3 2.2 13.7 0.34 2
3 Silty clay 1.18 0.532 0.0532 0.3 1.8 15.9 0.42 2
Note: M is the slope of the critical state line; λ and κ are slopes of the specific volume versus
ln(p´) curves for compression and swelling, respectively, where p´ is the mean effective
stress; ߤ is the Poisson’s ratio; e○ is the initial void ratio; and khp is the permeability of intact
zone in the plane-strain condition.

Table 5.5 Applied Properties for Sand Layer (after Sathananthan et al. 2008)

Layer Soil type c´(kPa) φ´ (deg) E (MPa) ߤ

4 Dense Sand 13.5 35 7.5 0.3

The settlement gauge P1 (located on the centreline at the interface of the

embankment and the ground) was used to verify the numerical predictions against
the field measurements, and the results are illustrated in Figure 2.26. It can be noted
that the FLAC results are in the best agreement with the field data assuming kh/ks=4
(the permeability ratio) and rs/rm=3 (the extent ratio), which confirms the validity of
the developed FLAC code in a numerical simulation of PVD assisted preloading
projects.

181
 Time (Day)
0 20 40 60 80 100
0

100 Field Data (Sathananthan et al. 2008)

FLAC results (kh/ks= 4 & rs/rm= 3)
200
Settlement (mm)

300 Commencement of stage 2

400

500
Commencement of stage 3
600

700

800

Figure 5.26 Comparison of numerical results with filed data (for settlement plate P1)

Figure 5.26 shows there are more disparities between the numerical predictions and
the field measurements in the initial stages of loading, which can be the result of
surcharge placement. For example, the settlement predicted at the commencement of
stage 2 was 95 mm, which is 60% greater than the field measurement of 60 mm,
whereas the difference was about 10% at the commencement of stage 3.

5.5 CHITTAGONG AIRPORT TRIAL EMBANKMENT

5.5.1 Introduction

Chittagong Port is the largest sea port in Bangladesh. A container yard was
constructed which covered an area of 60700 m2 and was designed to support a
container load producing a contact pressure of approximately 56 kPa (Dhar et al.
2011). The site is located on the banks of the Karnafully River beside the Bay of
Bengal in the Indian Ocean, as shown in Figure 5.27. According to the site
investigations, the soil profile contains soft to very soft clayey silt/silty clay layer at
depths of 0 to 3.5m below the ground surface with a thickness of 3.0m to 7.0m. A
ground improvement work was designed and carried out to pre-consolidate the soft
subsoil before the yard was constructed.

182
 Karnafully
N River

Indian
Ocean

Container Yard
22°17'27.34"N
91°47'27.77"E

Bay of
Bengal

Figure 5.27 Location of the container yard at Chittagong Port (courtesy of Google
Maps)

5.5.2 Subsurface Conditions

The site is a tidal plain on a narrow strip between the hilly Chittagong uplands
and the Bay of Bengal. The surface geology is mainly governed by shallow sea
water and flood plain activities of the Karnafully River and its tributaries. Fifteen
boreholes were drilled to collect samples of the subsoil. The distribution of these
boreholes over this area is shown in Figure 5.28.

183
 BH-12
Proposed Entry
(RL=17.94m)
BH-6 (RL=18.37m) BH-11
(RL=18.59m) BH-7 BH-8
(RL=19.65m) (RL=18.04m)
BH-13

(RL=18.27m)
BH-9
(RL=19.21m)
(RL=19.57m) BH-14
BH-5 BH-10 BH-15
(RL=19.72m)
(RL=18.58m)
(RL=19.12m)

BH-3
BH-4 BH-2
(RL=19.63m) BH-1
(RL=19.09m) (RL=19.83m)
(RL=19.32m)

Figure 5.28 Approximate locations of the boreholes, Chittagong Airport trial

embankmnet (after Dhar et al., 2011)

125 mm diameter boreholes were drilled to depths ranging from 14m to 24.5m
below the ground surface to collect disturbed and undisturbed samples of soil.
The results of this site investigation show that the subsoil profile includes very soft
to firm silty clay, or clayey silt and fine grained silty sand, with some decomposed
materials near the ground surface. The general subsurface condition is illustrated in
Figure 5.29.
The condition of the ground at this site varied from borehole to borehole.
At the ground surface the soil was generally fill materials consisting of light
brown clayey silt or brown silty sand/sandy silt. The clayey silt was firm to stiff
and the silty sand or sandy silt was medium dense. The fill materials extended
from the ground surface and continued down to depths of 0 to 3.5 m below the
existing grade.

184
 SPT N-values

Depth (m)

Figure 5.29 General ground profile along with SPT N-values (after Dhar et al., 2011)

5.2.3 Soil Properties

An extensive laboratory investigation was carried out by Dhar et al. (2001) to

identify the soil and determine the geotechnical design parameters. The soil samples
were classified conducting the Unified Soil Classification System (USCS).
Unconsolidated undrained and consolidated undrained triaxial tests were performed
to determine the strengths of the cohesive soil. An analysis of the grain size and
consolidated drained direct shear tests were conducted for the non-cohesive soil.
Table 5.6 indicates the Atterbeg limits, moisture content, and specific gravity of the
samples. The bulk unit weight of the soil varied from 20 kN/m3 to 23 kN/m3.

185
 Table 5.6 Properties of the cohesive soil samples (Dhar et al., 2011)

Natural Water
Borehole No. Depth (m) Gs LL PL PI
Content (%)
BH-1 3.1 – 3.55 33.7 2.75 45 26 19
BH-3 34.10 – 4.55 41.15 - 47 28 19
BH-4 6.80 – 7.25 37.55 2.71 32 23 9
BH-5 1.10 – 1.55 35.8 - 56 31 25
BH-5 4.10 – 4.55 51 2.74 43 27 16
BH-7 3.10 – 3.55 50.8 2.75 43 28 15
BH-8 3.10 – 3.55 39.15 - 48 29 19
BH-9 4.10 – 4.55 48.25 2.74 47 27 20
BH-12 2.10 – 2.55 50.8 2.77 51 30 21
BH-13 2.10 – 2.55 36.5 - 41 25 16
BH-14 4.10 – 4.55 45.25 2.76 44 27 17
BH-15 4.10 – 4.55 38.35 - 42 27 15

Nine Shelby tube samples collected from the silty clay or clayey silt layer were used
to determine the compressibility and permeability properties using one-dimensional
consolidation tests. According to Dhar et al. (2011), the values of Cc from twelve
tests were between 0.17 and 0.45 with the average value being equal to 0.3. The
recompression index, Cr, was calculated to be between 0.05 and 0.07. The initial void
ratios (e0) of these samples varied from 1.04 to 1.62, with an average of 1.28. Pre-
consolidation pressures calculated using the Cassagrande method were found to
range between 30 kPa and 50 kPa, which were less than the ground stresses expected
under a container load of 56 kPa. The coefficient of vertical permeability and the
coefficient of horizontal permeability were found to range from 0.032 m/year to
0.063 m/year and from 0.047 m/year to 0.095 m/year, respectively, within the range
of design stress (i.e. 60 to 100 kPa).

5.5.4 Installation of Vertical Drain and Embankment Construction

The results from the laboratory tests and site investigation were used to design soil
improvement works for the Chittagong Port, as indicated in Figure 5.27. The
thickness of the soft soil and the length of prefabricated vertical drains (PVDs) were
7 m and 9 m long respectively. PVDs with the dimensions of 100mm ൈ 4mm were

186
installed in a square pattern at a spacing of 1.0 m (centre to centre) to accelerate
consolidation. The properties of the applied PVDs are summarised in Table 5.7.

Table 5.7 Properties of the applied PVD (Dhar et al., 2011)

Properties Specified value
Drains
Weight per unit length 70 gm/m
Width 100 mm
Thickness 4 mm
Water discharge capacity 2840 m3/year
Core
Tensile strength 750 N
Filter Jacket
Apparent opening size (AOS) 90 μm
Grab tensile strength 400 N
Elongation at break 50 %
Puncture resistance 130 N
Burst strength 800 kPa
Permeability 6310 m/year

According to the designs, a surcharge pressure of 56 kPa was needed for the
preloading process, which is equivalent to a 3.0 high embankment and a unit weight
of 18 kN/m3 for the sand.
According to Dhar et al. (2011), the following sequences were conducted to
construct the embankment on top of the soft soil:
x Preparation of the existing ground
x Placement of local sand to raise the ground level where required
x Placement of a drainage blanket of coarse sand
x Installation of Prefabricated Vertical Drains (PVDs)
x Pre-loading

An approximately 150 mm thick layer of sand was placed on top of the levelled
ground. A drainage blanket of coarse sand was then placed over the sand to facilitate
the drainage of water through the vertical drains. The expected consolidation was
expected to be compensated for by the 450mm thick layer drainage blanket. The
drainage blanket was constructed in two stages, (i) the lower 250 mm thick layer was

187
placed before the PVDs were installed to provide a working platform for their
installation, and (ii) the remainder of the drainage layer was placed after the PVDs
were installed to allow the drains to discharge into the sand layer. The PVDs were
installed using a mandrel combined with an anchor having a maximum cross
sectional area of 7000 mm2. The PVDs were installed to a depth of 9 m below
ground level to cover the full depth of the soft soil. The remaining 200 mm thick
layer of coarse sand was placed on top of the final surface after the PVDs
installation, and then compacted and levelled to place the settlement measuring
gauges.
Thirty settlement measuring gauges were used to measure the rate of
vertical displacement. A schematic plan of the approximate locations of settlement
gauges is shown in Figure 5.30. A settlement gauge is illustrated in Figure 5.30(a),
and it includes a base plate and a stand pipe. The base plate for the gauge was placed
on top of the levelled granular layer, while the elevation of the top of the stand pipe
was monitored to obtain the ground settlement.

Proposed Entry A=60700 m2

160 m

Stand pipe

Base plate

460 m

(a) (b)

Figure 5.30 Settlement monitoring program, (a) Settlement gauge and (b) Points of
settlement measurement (schematic plan) (after Dhar et al. 2011)

The pre-loading process was conducted by constructing a 3.0m high embankment

over the drainage. The surcharge material was placed in two layers of approximately
equal thickness. The sand bags or brick stacks were used to contain the sides of the

188
surcharge load vertically along the boundary of the area. Details of the ground
improvement work are shown in Figure 5.31.

Settlement plate
CL Sand Blanket
Surcharge Material Sand bags
PVDs

3.0 m

9m
1m

Figure 5.31 Schematic details of the ground improvement works (after Dhar et al.,
2011).

5.5.4.1 Monitoring of Settlement

The settlement rates were monitored immediately after the embankment was
constructed and continued until consolidation was completed. Immediately after
placing the surcharge to its full height (in 10 - 12 days), the ground settlements were
measured using the settlement gauges. The first measured settlement for each gauge
was called the “initial settlement”.
The settlement rates at different locations are plotted in Figure 5.32. These
measurements began with an initial settlement of 80-300 mm due to the placement of
the 3.0 m high surcharge. The maximum settlements measured during the monitoring
period varied from 220 mm to 415 mm. Figure 5.32 reveals that the settlements were
almost completed after 30 to 52 days of preloading.

189
 450

400

350
Settlement (mm)

300

250

200

150 SP 10 SP 9 SP 8
SP 7 SP 6 SP 5
100 SP 3 SP 2 SP 1
SP 11 SP 12 SP 22
50

0
0 10 20 30 40 50 60

Time (days)

Figure 5.32 Variation of ground settlement with time

5.5.5 Numerical Simulation

A soft to very soft clayey silt/silty clay deposit with a thickness of approximately 7 m
was used to conduct the numerical simulation. Details of the simulated model are
shown in Figure 5.33.

CL Surcharge Material Sand bags

PVDs

γt=20 kN/m3
3.0 m
0m G1
clayey
silt/silty clay
(CL/ML) 4m

G2 9m
LL=45
PI=18
Gs=2.74
1m
7m
Silty sand

10 m

Figure 5.33 Cross section of constructed embankment at Chittagong Port site

190
Points G1 and G2 were selected to monitor the settlement and changes in pore
pressure respectively. Figure 5.34 shows the construction history of the embankment.

Embankment Height (m)

3

2
Commencement of recording
settlement plate readings

0
0 10 20 30 40 50 60 70 80 90
Time (days)

\ Figure 5.34 Construction history (Chittagong Port embankment)

FLAC 2D numerical code incorporating the modified Cam-Clay constitutive soil

model was used to simulate the Chittagong Port preloading process by applying
plane-strain conditions. The simulated discretised finite-difference mesh is illustrated
in Figure 5.35. To model the PVD and surface drainage, a zero excess pore water
pressure was assumed along the vertical drains and ground surface boundary,
respectively. Equations 2.30 to 2.32 were used to calculate the equivalent plane-
strain permeability needed in this analysis. The soil properties used in the numerical
analysis are summarised in Table 5.8.

60 kPa
Settlement 20 m
gauge (G1)

Vertical drain

10 m

Smear zone
1m

Figure 5.35 Sample of mesh grid pattern for Chittagong Port embankment
considering the smear

191
 Table 5.8 Adopted soil properties in FLAC simulation (after Dhar et al. 2011)
γs kh
Layer Soil type M λ κ ν e○ kh/ kv
(kN/m3) (10-9 m/s)
Clayey
Soft soil 0.94 0.13 0.026 0.3 1.28 14.0 2.31 1.5
Silt

The numerical results are compared with the field measurements in Figure 5.36. The
average field settlement was calculated by averaging the results of 16 variable
settlement plates (SP1-10 and SP16-20) close to the centre of the embankment on the
ground surface.

Time (Days)
0 10 20 30 40 50 60 70
0
(Kh/Ks=2 & rs/rm=3)
50 Field measurements (Dhar et al. 2011)
Settlement (mm)

Numerical predictions (Point G1)

100

150

200
Ū=90%
250

300

Figure 5.36 Comparison of numerical results with filed data at Chittagong Port site

According to Figure 5.36, the FLAC predictions are in a very good agreement with
the field measurements where kh/ks=2 and rs/rm=3. As illustrated in Figure 5.36, field
settlements were measured immediately after placing the surcharge to the full height
of 3 m (12 days). Primary consolidation settlement was predicted to be
approximately 258 mm, which was confirmed by the settlement measured in the
field. It can be noted that 90% degree of consolidation was obtained after 34 days.
Figure 5.37 illustrates the numerical variation of excess pore water pressure
for the case where the smear zone properties were kh/ks=2 and rs/rm=3. It can be
noted that the predicted dissipation of excess pore water pressure commenced after
reaching a peak value of 60 kPa at the end of embankment construction process.

192
According to Figure 5.37, the excess pore water pressure was completely dissipated
after about 70 days, which shows the end of the primary consolidation process and
confirmed the settlement graph shown in Figure 5.36.

70 4
(Kh/Ks=2 & rs/rm=3)
Numerical predictions (Point G1)
60
Excess pore pressure (kPa)

Embankmnet loading history

Embankmnet height (m)

3
50

40
2
30

20
1
Ū = 90%
10

0 0
0 10 20 30 40 50 60 70 80 90 100
Time (Days)

Figure 5.37 Numerical variation of excess pore water pressure against

consolidation time

The plotted graphs shown in Figures 5.36 and 5.37 indicate that the proposed back
calculation procedure in conjunction with the developed numerical code can be used
as a reliable tool to accurately predict the properties of the smear zone.

5.6 CASE STUDY 5: LARGE-SCALE CONSOLIDOMETER

5.6.1 Test Apparatus

Bamunawita (2004) used a large consolidometer designed and built at the University
of Wollongong, Australia by Indraratna and Redana (1998) to investigate the
effectiveness of vacuum preloading (Figure 5.38). For this purpose, a central vertical
drain was installed in the soil specimen placed in a large stainless steel cell (450 mm
in diameter by 950 mm high) using a specially designed mandrel, and then the
surcharge load was applied under two different conditions, (I) without vacuum
pressure, and (II) with vacuum pressure. In this study, the condition without a
vacuum pressure was used validates the developed numerical code.

193
 Consolidometer

PVD

Figure 5.38 Large scale consolidometer (after Bamunawita, 2004)

A 1.5 mm thick sheet of Teflon was wrapped around the internal circumference of
the cell, and at the bottom, to reduce friction on the boundary of the cell. The
surcharge loading system was applied with an air jack compressor system via a rigid
piston, while an LVDT was placed at the top to measure vertical displacement. Strain
gauge type pore pressure transducers were installed to measure the excess pore water
pressures at various locations. Figure 5.39 illustrates the large consolidometer
schematically.

194
 Settlement
Load transducer

Permeable layer

470 mm
Smear zone

Intact zone

l =950 mm
Pore water T3
pressure
transducer kh ks rs

480 mm
Vertical drain

ds
R

Impermeable
boundary D=450 mm

Figure 5.39 Cross section of the large consolidometer (modified after Indraratna et al.
2004a)

5.6.2 Test Sample

The size of each soil sample meant that the large scale consolidometer was
approximately 0.15m3. Reconstituted alluvial clay from Moruya, NSW, was used to
make the large samples because obtaining an undisturbed sample of this size was not
feasible. The properties of the reconstituted alluvial clay are listed in Table 5.9.
Particles of clay (<2μm) and particles smaller than silt size (<6μm) accounted for
about 40%-50% and 90%, of the sample, respectively. The reconstituted clay was
categorised as CH (high plasticity clay), based on the Casagranade Plasticity Chart.

195
Table 5.9 Soil properties of the reconstituted sample of Moruya clay (after
Bamunawita 2004)

Properties Specified values

Clay Content (%) 40-50
Silt Content (%) 45-60
Water content, w (%) 40
Liquid Limit, wL (%) 70
Plastic Limit, wP (%) 30
Plasticity Index, PI (%) 40
Unit Weight, (t/m3) 1.81
Specific Gravity, Gs 2.63

5.6.3 Test Procedure

The saturated clay was placed in the cell and compacted in layers, and then the final
surface of the compacted clay was covered by the filter material to facilitate water
drainage. After filling the chamber with the soft saturated clay (γt=18.1 kN/m3), an
initial consolidation pressure of 20 kPa was applied via a rigid piston. When the pre-
consolidation phase was completed, a PVD 100mm × 3mm was installed with a
rectangular mandrel. The end of the drain was attached to a shoe (anchor) to ensure
that the drain remained in position when the mandrel was withdrawn. Then a
surcharge load without vacuum pressure was applied in two stages, (i) 50 kPa for 17
days, and (ii) 100 kPa for 14 days. The loading increments were equivalent to an
earth fill (γt=18.1 kN/m3) with the heights of about 2.5m and 6m, respectively. The
loading history of the test is presented in Figure 5.40.

196
 120
Stage 2
100
Surcharge load (kPa)

80

60 Stage 1

40

20

0
0 5 10 15 20 25 30
Time (days)

Figure 5.40 Loading history for the large scale consolidometer

5.6.4 Verification of the Developed Numerical Code

The axisymmetric finite difference model was used to simulate the large
consolidometer. As Figure 5.41 shows, the zero excess pore water pressure with the
hydrostatic water pressure was assumed on the boundary of the drain, and then a
uniformly distributed load was assumed on top of the cell to simulate the surcharge
pressure. Figures 5.41(a) and 5.41(b) show the large-cell mesh grid, which was
generated using the numerical code, and the pore pressure boundary condition
(hydrostatic with zero excess pore water pressure) along the vertical drain,
respectively. The properties of the soil adopted in this study are summarised in Table
5.10.

197
 P (surcharge)

u=0

Intact zone

rs hc =950 mm drain
boundary
li

Smear zone
Axi-symmetric
u= hc Jw
boundary

Sd=225 mm

Figure 5.41 Sample of grid pattern for the large consolidometer applying developed
code (rs= smear zone extent; li= intact zone extent; sd= drain space, hc= length of the
cell, w= width of the cell), (a) boundary conditions applied to simulated cell and (b)
pore pressure distribution boundary condition along the vertical drain

Table 5.10 Soil properties for Modified Cam-Clay (after Indraratna et al. 2004a)
khp
Property M λ κ ߤ ecs -11 LL PI Gs
(10 m/s)
Moruya clay 1.1 0.15 0.05 0.25 1.55 9.1 70 40 2.63
Note: ecs is the void ratio at the critical state

The top vertical displacement gauge and pore water pressure transducer T3 (Figure
5.39) were selected to verify the numerical predictions with the test measurements.
The numerical results were compared with the laboratory measurements in Figures
5.42 and 5.43, and indicate that the developed numerical code can predict the
behaviour of PDV improved soft soil under a surcharge pressure quite well.
According to the back calculation analyses and the proposed procedure, the smear
zone characteristics of kh/ks=1.7 (permeability ratio) and rs/rm=3 (extent ratio)
obtained a settlement curve having minimum error with the test measurements.

198
 Time (day)
0 5 10 15 20 25 30
0
Lab measurements

10 Numerical predictions (n=1.7 & s=3)

Settlement (mm)

20

30

40

50

Figure 5.42 Results of large consolidometer cell: Settlement versus consolidation

time

As Figure 5.42 shows, the disparities between the numerical predictions and
experimental measurements increased overtime, which can be associated with the
viscous behaviour of the soil. The current numerical method uses constant
compressibility parameters (ɉǡɈ) and ignores the viscous deformation of the soil
during consolidation. According to Yin and Graham (1989) soil creep occurs on pore
water pressure during dissipation, which contributes to larger deformations.

80
(kh/ks=1.7 & rs/rm=3)
70
Excess pore water pressure (kPa)

Laboratory measurements
Numerical Predictions
60

50
Commencement of stage 1
Commencement of stage 2
40

30

20

10

0
0 5 10 15 20 25 30
Time (day)

Figure 5.43 Results of large consolidometer cell: Excess pore pressure versus
consolidation time

199
It can be noted that the overall trends for variations in the numerical and
experimental excess pore water pressure were similar, which shows the accuracy of
the back calculated smear zone properties, even though there were still some
disparities.
Figure 5.43 shows that the predicted excess pore water pressure curve
experienced sudden changes after each stage of loading that were equal to the extra
pressure applied, although this increase occurred over a period of time for the
laboratory measured values. There were two main reasons for this abnormal excess
pore water pressure behaviour, the Mandel-Cryer effect and volumetric strain
softening, both of which were explained previously. Both curves of the experimental
and variation in the numerical excess pore water pressure shown in Figure 5.43
indicate that primary consolidation was not obtained at the end of consolidation time,
which is in agreement with the plotted settlement graphs in Figure 5.42.

5.7 SUMMARY

In this chapter five PVD assisted preloading case studies, including four trial
embankments and a large-scale consolidometer, were explained in detail. The
selected case studies were simulated numerically using the developed numerical code
that incorporated the back calculation procedure for predicting the properties of the
smear zone.
The Cumbalum trial embankment constructed near Cumbalum (New South
Wales, Australia) in 1998 with a fill thickness of 5.0 m, was as the first case study to
be simulated. Vertical drains were installed at 1.35 m apart in a triangular pattern, to
a depth of approximately 22 m, and the embankment was allowed to consolidate for
about 4.5 years. The second simulated case study was the Ballina Bypass trial
embankment, which was constructed north west of Ballina (New South Wales,
Australia) to a height of 8.5 m on 25 m thick soft soil deposits, to investigate the
performance of the vacuum preloading ground improvement technique. 34-mm
diameter vertical drains were installed 1.0 m apart in a square pattern, and were then
connected to horizontal drains. A vacuum pressure of 70 kPa was applied for a
period of 400 days and then removed. The Sunshine trial embankment was the third
case history, and it was constructed in 1992 in Maroochy Shire, Queensland,
Australia. The the soft soil was 11 m thick and was overlying a layer of dense sand.

200
Vertical drains were installed in a triangular pattern 2 m apart. The trial embankment
had a final height of 2.85 m. The fourth case study was the Chittagong Port
embankment that was constructed on top of a 7m thick deposit of soft clay on the
banks of the Karnafully River beside the Bay of Bengal. The site was designed to
support a container load producing a contact pressure of approximately 56 kPa.
Vertical drains were installed in a square pattern to a depth 9 m to cover the full
depth of the soft clay. The final height of the embankment was 3.0. A fully
instrumented large consolidometer built at the University of Wollongong, Australia
was simulated as the last case study. The height of the cell was 950 mm and the
diameter was 450 mm. The saturated clay was placed in the cell and a central vertical
drain was installed with a mandrel, after the pre-consolidation process was
completed. A surcharge pressure of 100 kPa was applied in two stages.
In the numerical analyses, the modified Cam-Clay model was used to
simulate the behaviour of soft soil in the intact zone and the smear zones, while the
Mohr-Coulomb model was used to simulate the embankment. The hydrostatic
pressure was fixed along the drain boundary to simulate the PVDs and a constant
negative pore pressure was used to simulate the uniform distribution of vacuum
pressure. A zero excess pore pressure was assumed at the ground surface to model
the drainage boundaries. The smear zone was modelled by assigning a constant
reduced permeability to the area surrounding the vertical drain and simulated fully
saturated coupled flow-deformation was carried out to model the dissipation of pore
water pressure.
The numerical results indicated that the settlement curves obtained by
adopting the properties of the back calculated smear zone were in good agreement
with the field and laboratory measurements. These results confirmed the validity of
using the back calculation procedure combined with the numerical code to accurately
predict the smear zone properties, even though there were some disparities between
the variations in the numerical and field pore pressure. According to these variations
in pore pressure, the maximum predicted excess pore water pressure occurred
immediately after completion of the trial embankment and it was equal to the applied
surcharge. However, this value was slightly less than the applied pressure for the
field measurements; this can be related to the relationship between the volumetric
compliance of the pore water pressure measuring system and the compressibility of

201
the soil skeleton. Furthermore, the field measurements showed that the excess pore
pressures did not dissipate immediately at the end of loading or construction, but
either increased or stabilised for a period before decreasing; this can be explained by
the Mandel-Cryer effect and volumetric strain softening.

202
 CHAPTER SIX

6 DETERMINING THE MINIMUM PERIOD OF

MONITORING FOR PREDICTING SMEAR ZONE
PROPERTIES OF A TRIAL EMBANKMENT

6.1 GENERAL

The construction of a fully instrumented trial embankment has been extensively used
as a reliable method to determine the feasibility of preloading with vertical drains
and to estimate the properties of the smear zone by applying a back calculation
procedure (Kelly, 2008). The long time required to construct the trial embankment
has bee a major challenge in using this method to conduct the back calculation
analysis and estimate the extent and permeability of the smear zone. Hence, in many
cases it may cause a considerable delay in constructing the actual embankment and
impose a significant increase in the project cost. Estimating the extent and
permeability of the smear zone in the early stages of constructing a trial embankment
can convert this method into a very practical, accurate, and cost effective approach.
This chapter is allocated to establishing the minimum degree of
consolidation (i.e. the minimum waiting time after constructing a trial embankment)
that results in accurate prediction of smear zone properties. For this purpose, a
systematic procedure, illustrated in Figure 6.1 has been used in conjunction with the
developed numerical code. As Figure 6.1 shows, the approach starts with collecting
the input data, including the soil properties, PVD assisted preloading specifications
(PVD pattern and loading history) and the results of field monitoring (i.e. variations
in settlements and excess pore pressures). The first stage is then completed by
defining the initial values, the maximum and minimum values, and the incremental
rates for the ratios of extent and permeability. In the second stage, a well-designed
parametric study is conducted to predict the behaviour of the ground, determine the
corresponding settlement curves for different combinations of smear zone
characteristics, and calculate cumulative error between the numerical results and
field measurements at every degree of consolidation. Stage two stops as soon as the

203
extent ratio is out of the input range. The minimum required degree of consolidation
is determined in the last stage and results in appropriate smear zone characteristics.

Start

Input (stage 1)

Vertical drain pattern

Field performance observation Geotechnical site investigation
& loading history
(monitoring results) (Soil properties)
(improvement plane)

∆s, (rs /rm )0 ∆R, (kh/ks )0

Sf ( Final settlement)
(rs /rm )min, (rs /rm )max (kh/ks )min, (kh/ks )max

i=1, j=1
(rs /rm )i = (rs /rm )0 & (kh/ks )i = (kh/ks )0

(rs /rm )i = (rs /rm )i + ∆s j=i

Numerical Parametric Study (stage 2)

NO

YES YES NO
(kh/ks )min ≤ (kh/ks )i ≤ (kh/ks )max (rs /rm )min ≤ (rs /rm )i ≤ (rs /rm )max

Save (E U%)i, (Ef )i, U% n=5

(kh/ks )i = (kh/ks )i-1 + ∆R i=i+1
(rs /rm )i & (kh/ks )i n=1

Numerical simulation U% Calculation of (EU%)min=min (E U%)s

predicting ground behavior (SFEM ) (St / Sf)) (eU%)i, (EU%)i, (Ef )i (1 ≤ s < i)

Find corresponding (Ef )s ,

Draw (Ef )s -U% graph
(rs /rm )s & (kh/ks )s to (EU%)min
coresponding degree of consolidation
Prediction of smear zone properties &

f luctuation in YES n=n+1

m consecutive points
U% n=U% n+1
on (Ef )s - U% cruve
(stage 3)

NO

Print U% n-m , (rs /rm )s & (kh/ks )s

Stop

Figure 6.1 Flowchart of the systematic procedure to determine the minimum degree
of consolidation resulting in an accurate estimation of smear zone characteristics

A four-step solution has been suggested to apply the procedure proposed in Figure
6.1, which computes the minimum degree of consolidation (minimum preloading
time), using the simulated case studies, explained in Chapter 5. The four-step
solution includes: (i) estimating the primary consolidation settlement, (ii) conducting
the parametric study, (iii) determining the error, and (iv) establishing the minimum

204
waiting time required. A detailed explanation of the solution procedure is provided in
the following sections.

6.2 STEP I: ESTIMATING THE PRIMARY CONSOLIDATION

SETTLEMENT

The final settlement for each case study under the embankment surcharge was
predicted by applying the developed numerical code. This final settlement can then
be used to determine the degree of consolidation at any time during or after
construction of the trial embankment, using Equation (6.1).

௦೟
ܷΨ ൌ ൈ ͳͲͲ (6.1)
௦೑

where, U% is the degree of consolidation at time t, st is the field settlement at time t,

and sf denotes the predicted final primary consolidation settlement. Table 6.1 shows
the primary consolidation settlement and the corresponding degree of consolidation
for the simulated case studies.

Table 6.1 Primary consolidation settlements and the corresponding degree of

consolidations
Chittagong Ballina
Case Cumbalum Sunshine Large-scale
Sea Port Bypass at
Study at SP9 at P1 consolidometer
trial embankment SP12
Sf (m) 0.258 3.3 5.2 1.75 0.086

Spr (m) 0.258 2.45 4.4 0.75 0.045

tpr (day) 55 2000 400 90 30

Upr (%) 100% 75% 85% 40% 52%

Note: Sf is the predicted final primary consolidation settlement; Spr is the field settlement at
the end of preloading time; tpr is the preloading time; Upr is the degree of consolidation at the
end of preloading

205
6.3 STEP II: CONDUCTING PARAMETRIC STUDIES

A parametric study was conducted for each case study to obtain the time-settlement
curves for different combinations of the extent and permeability of the smear zone
used to determine the error between the field measurements and numerical
predictions. The numerical analyses were performed by adopting different
combinations of the extent and permeability of the smear zone, and in each case the
results were then compared with the field measurements. Furthermore, the effects of
the smear zone properties on the preloading process were also investigated.

6.3.1 Chittagong Sea Port Trial Embankment

A comprehensive parametric study was conducted for the Chittagong Sea Port trial
embankment to obtain the variations in settlement against the consolidation time, and
also to investigate the influence that different combinations of the smear zone
characteristics have on the preloading design. For this purpose kh/ks (permeability
ratio) and rs/rm (extent ratio) were changed from 2 to 5. It should be noted that the
developed numerical code, including the supplementary FISH sub-routines were
used, which allowed for an automatic parametric study in the required range of
parameters, while the characteristics of the smear zone, drain spacing, and adjusted
finite difference mesh and grids changed simultaneously, which is a unique feature
facilitating the parametric study. Details of the geotechnical model and numerical
verification exercise for this case study has been explained in Section 5.5 in Chapter
5. Figure 6.2 illustrates the results of the parametric study in the form of settlement-
time curves.

According to Figure 6.2, the settlement curves are converged to a unique

value of approximately 258 mm, which is the primary consolidation settlement. The
time required to obtain 90% of primary consolidation settlement (232 mm) was used
to investigate the effect of the properties of the smear zone in the consolidation
process. The 90% degree of consolidation has been adopted by researchers as an
acceptable value for near completion of consolidation resulting in accurate settlement
predictions (Basu et al. 2006, Indraratna 2009, Saowapakpiboon et al. 2011). It
should be noted that 90% degree of consolidation is also often used by practising
geotechnical engineers as the indication of the time for the removal of the surcharge.

206
 Time (Days)
0 20 40 60 80 100
0
rs/rm=2
Settlement (mm)

50

100
Kh/Ks=2

150 Kh/Ks=3
Kh/Ks=4
200 Kh/Ks=5
Ū = 90%
250
(a)
300

Time (Days)
0 20 40 60 80 100
0
rs/rm=3
50
Settlement (mm)

100
Kh/Ks=2
Kh/Ks=3
150
Kh/Ks=4

200 Kh/Ks=5
Ū = 90%
250
(b)
300

Time (Days)
0 20 40 60 80 100
0
rs/rm=4
50
Settlement (mm)

100
Kh/Ks=2
150 Kh/Ks=3
Kh/Ks=4
200 Kh/Ks=5
Ū = 90%
250
(c)
300

207
 Time (Days)
0 20 40 60 80 100
0
rs/rm=5
50
Settlement (mm)

100
Kh/Ks=2
Kh/Ks=3
150
Kh/Ks=4
Kh/Ks=5
200
Ū = 90%
250
(d)
300

Figure 6.2 Parametric study results for Chittagong port case history at point G1, (a)
rs/rm=2, (b) rs/rm =3, (c) rs/rm =4, and (d) rs/rm =5

Figure 6.2(a) shows that a minimum of 33 days was required to achieve 90% degree
of consolidation, while considering kh/ks=2 and rs/rm=2. When the properties of the
smear zone were kh/ks=5 and rs/rm=5, the time required was the highest and equal to
67 days, which is approximately two times longer than the minimum duration.
According to the plotted curves in Figure 6.2, the influence of the variations in the
permeability of the smear zone was more critical when the ratio of the extent of the
smear zone was larger. For instance, the time required to obtain 90% degree of
consolidation was increased by 60% (from 33 days to 53 days) by changing the
permeability ratio from 2 to 5 and making the extent ratio equal to 2, whereas this
increase was 80% (from 37 days to 67 days) for an extent ratio of 5.
The general trend in Figures 6.2(a)-6.2(d) shows that changing the
permeability ratio in a smaller range resulted in larger variations of the time required
to obtain 90% degree of consolidation with a constant extent ratio. According to
Figure 6.2(a), the consolidation time increased by 23% by varying the permeability
ratio from 2 to 3, while this change was 17% and 12% when the permeability ratio
was changed from 3 to 4 and 4 to 5, respectively.
Figure 6.3 illustrates the results of the numerical parametric study
investigating the influence that the smear zone properties had on the dissipation of
excess pore water pressure (EPWP). Graphs were plotted for point G2, located at a

208
depth of 4m (see Figure 5.34). Figure 6.3 confirms that increasing the permeability
and extent ratios prolongs the dissipation of excess pore water pressure (EPWP)
quite considerably. According to Figures 6.3(a)-6.3(d), the permeability ratio was a
more critical parameter than the extent ratio, although the influence that the
variations of the extent ratio had on consolidation time cannot be neglected. For
example, according to Figure 6.3(b), there was a 160% difference between the
predicted excess pore pressure after 34 days (90% field degree of consolidation) for
kh/ks=2 (EPWP=13 kPa) and kh/ks=5 (EPWP=34 kPa), while keeping rs/rm=3 (Figure
6.3(b)).

70
rs/rm=2
Excess pore pressure (kPa)

60
Full height of embankment reached
50 Kh/Ks=2
Kh/Ks=3
40 Kh/Ks=4
Kh/Ks=5
30

20

10
(a)
0
0 20 40 60 80 100
Time (Days)

80

70 rs/rm=3
Excess pore pressure (kPa)

60 Full height of embankment reached Kh/Ks=2

50 Kh/Ks=3
Kh/Ks=4
40
Series1
30

20
Ū = 90%
10
(b)
0
0 10 20 30 40 50 60 70 80 90 100
Time (Days)

209
 80

Excess pore pressure (kPa) 70 rs/rm=4

60 Full height of embankment reached Kh/Ks=2

50 Kh/Ks=3
Kh/Ks=4
40
Kh/Ks=5
30

20

10
(c)
0
0 20 40 60 80 100
Time (Days)

80

70 rs/rm=5
Excess pore pressure (kPa)

60 Full height of embankment reached Kh/Ks=2

50 Kh/Ks=3
Kh/Ks=4
40
Kh/Ks=5
30

20

10
(d)
0
0 20 40 60 80 100
Time (Days)

Figure 6.3 Results of parametric study using the FLAC code developed to
investigate the influence of the smear zone properties on the dissipation of excess
pore water pressure for Chittagong port case history at point G2, (a) rs/rm=2, (b)
rs/rm=3, (c) rs/rm=4, and (d) rs/rm=5

The time required to obtain 90% degree of consolidation for the different smear
zones is illustrated in Figure 6.4 using the results of the parametric study, which
interprets the effects the smear zone properties on consolidation time better.
According to Figure 6.4, the consolidation time depends on the extent and
permeability of the smear zone, for example, assuming that rs/rm=2 for the cases
where kh/ks=2 and kh/ks=5, the time needed to obtain 90% degree of consolidation is

210
approximately 33 days and 53 days, respectively, indicating a 60% difference. This
difference is more significant for larger values of rs/rm.

6.0
5.5
rs/rm=2 rs/rm=3 rs/rm=4 rs/rm=5
5.0
Permeability ratio (kh/ks)

4.5
4.0
3.5
3.0
2.5
2.0
1.5 S4 S3 S2 S1

1.0
32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70
Time (Days)

Figure 6.4 Predicted time to obtain 90% degree of consolidation (Chittagong trial
embankment)

Figure 6.4 clearly indicates that the extent ratio (rs/rm) of the smear zone is an
important parameter influencing the consolidation time and cannot be neglected. By
varying rs/rm from 2 to 5 and assuming kh/ks as a constant parameter can influence
the consolidation time needed by more than 25%. A combination of the variables in
the extent and permeability in the smear zone will result in large changes in the
consolidation time. The results presented in Figure 6.4 clearly indicate that the
influence of variables in rs/rm becomes more important when the permeability of the
smear zone decreases.
According to the numerical results presented in Figure 5.37, the predicted
settlement curve is in the best agreement with the field measurements when the
properties of the smear zone are kh/ks=2 and rs/rm=3. The time required to obtain 90%
degree of consolidation for this condition is equal to 34 days, which is highlighted as
point S2 in Figure 6.4. A vertical line is plotted from t90%= 34 days, which intersects
the set of lines at points S1, S2, S3, and S4. The properties of the smear zone at these
points are summarised in Table 6.2.

211
Table 6.2 Back calculated smear zone properties to achieve t90%= 34 days
(Chittagong trial embankment)

Point S1 S2 S3 S4

kh/ks 2.10 2.0 1.85 1.75

rs/rm 2 3 4 5

Different combinations of the extent and permeability of the smear zone (Table 6.2)
that result in the same t90%= 34 days were used to compare the variations in
settlement and excess pore water pressure with the consolidation time by applying
the developed FLAC code; the results are presented in Figure 6.5.

300
(a)
250
Settlement (mm)

200

150
Kh/Ks=2.1 & rs/rm=2 (S1)
100 Kh/Ks=2 & rs/rm=3 (S2)
Kh/Ks=1.85 & rs/rm=4 (S3)
50 Kh/Ks=1.75 & rs/rm=5 (S4)

0
0 20 40 60 80
Time (Days)

(b)
Excess pore water pressure (kPa)

60 Kh/Ks=2.10 & rs/rm=2 (S1)

Kh/Ks=2 & rs/rm=3 (S2)
50 Kh/Ks=1.85 & rs/rm=4 (S3)
Kh/Ks=1.75 & rs/rm=5 (S4)

40

30

20

10

0
0 20 40 60 80
Time (Days)

Figure 6.5 Results of FLAC analysis for points in Table 7, using the Chittagong
port case history, (a) Settlement variation, and (b) dissipation of excess pore water
pressure

212
Figure 6.5 shows that the curves for the variations in settlement and dissipation of
excess pore water pressure over time follow the same trend for points S1, S2, S3, and
S4. Therefore, the properties of the smear zone of any of these points can be used for
practical design purposes to reduce inaccuracies. In other words, the extent ratio
(rs/rm) can assumed to be a constant value and the permeability ratio (kh/ks) can be
changed to conduct the parametric study and determine the optimum combination.
The numerical results (settlement variation versus consolidation time) were
compared with the field measurements in Figure 6.6 by varying the permeability
ratio (n=kh/ks) from 2 to 5, while the extent ratio (s=rs/rm) was kept at a constant
value of 3.

Time (Days)
0 20 40 60 80 100
0 0%

Degree of consolidation (%)

50 20%
Field measurements
Settlement (mm)

100 n=5 & s=3 40%

n=4 & s=3
n=3 & s=3
150 n=2 & s=3
60%

200 80%

250 100%

300

Figure 6.6 Settlement variations against consolidation time for the Chittagong trial
embankment

The overall trend in Figure 6.6 indicates that the case with the permeability ratio of 2
and smear extent of 3 had the best fit with the field measurements, but this was
impossible to clarify in the early stages of the preloading process by observation.

6.3.2 Cumbalum Trial Embankment

The Cumbalum trial embankment was simulated using the developed FLAC code
and the systematic analysis procedure (Figure 6.1) was used to conduct a parametric

213
study by varying the kh/ks and rs/rm from 2 to 6. The geotechnical model developed
for the site and the numerical modelling procedure for this case study have been
explained in detail in Chapter 5, Section 5.2. The numerical predictions are plotted
in Figure 6.7 and compared with the field measurements by adopting the selected
combinations of kh/ks and rs/rm. It can be observed that the variation of the extent and
permeability of the smear zone had a substantial influence on the settlement rate.

Consolidation Time (day)

0 250 500 750 1000 1250 1500 1750 2000
0.0
Field measurements (SP9)
installation

-0.5 Numerical predictions (n=6 & s=6)

PVD

Numerical predictions (n=5 & s=5)

Stage 2
End of

Numerical predictions (n=4 & s=4)

Settlement (m)

-1.0 Numerical predictions (n=3 & s=3)

Numerical predictions (n=2 & s=2)
-1.5

-2.0
Stage 3
End of

-2.5

-3.0

Figure 6.7 Results of numerical parametric study: Cumbalum trial embankment at SP9

Figure 6.7 shows that the settlement has increased from 2.3 m to 3.0 m by varying
kh/ks and rs/rm from 6 to 2 after 1900 days of consolidation. According to the graphs
plotted in Figure 6.7, the settlement curve corresponding to the case with smear zone
characteristics of kh/ks=5 and rs/rm=5 agrees well with the field measurement.
However, it was not possible to verify this agreement by observation at the initial
stages of construction, thus a rigours analytical procedure was required.
Figure 6.8 illustrates the predicted and measured values of excess pore
water pressures at the location of PC2, at a depth of 5.8m (see also Figures 5.7 and
5.8 for more details). It can be observed that the excess pore water pressure curves
for the numerical predictions and the field measurements followed a similar pattern.
Pore water pressures increased during the fill placement at each stage of construction

214
and then gradually dissipated. The maximum excess pore water pressure for each
case occurred at the end of the construction process, which was similar to the field
observations reported by Kabbaj et al. (1988) and Leroueil (1997). Figure 6.8 shows
that the maximum excess pore water pressure of 120 kPa occurred at the end of the
construction process by adopting kh/ks=6 and rs/rm=6. It can be noted that lower kh/ks
and rs/rm values caused a faster dissipation of excess pore water pressure. For
example, 90% of the excess pore water pressure had been dissipated after 1900 days
by adopting kh/ks=2 and rs/rm=2, while this dissipation was 70% when kh/ks=6 and
rs/rm=6.

140
Field measurements (depth 5.8m)
120 Numercal Predictions (n=2, s=2)
Numercal Predictions (n=5, s=5)
Excess pore pressure (kPa)

100 Numercal Predictions (n=6, s=6)

80

60

40

20

0
0 250 500 750 1000 1250 1500 1750 2000
Consolidation time (day)

Figure 6.8 Variation of excess pore pressure with time for the Cumbalum trial
embankment at PC2 (depth 5.8m)

Figure 6.8 shows that there are some discrepancies between the measured and the
predicted excess pore water pressures. Compared to the field values, the predicted
excess pore pressures were larger at the end of the construction process, and in better
agreement when the degree of consolidation increased. This discrepancy can be
attributed to numerous factors such as uncertainties in the soil properties, the effect
of smear zone characteristics, and the simplified improper conversion of the
axisymmetric condition to the plane-strain (2D) condition to analyse the multiple
drains.

215
6.3.3 Ballina Bypass trial Embankment

The developed FLAC code was applied to the Ballina Bypass trial embankment
using the second stage of the proposed procedure shown in Figure 6.1 to conduct the
parametric studies. The geotechnical model adopted and the details of this case
study, as well as the results from the numerical modelling verified against the field
measurements, have been reported earlier in Chapter 5, Section 5.3. Different
combinations of the permeability ratio (n=kh/ks) and extent ratio (s=rs/rm) were used
to investigate how variations in the properties of the smear zone could affect the
consolidation settlement and to compare the numerical results with the field
measurements. The results of the selected numerical parametric study are compared
with the field measurements in Figure 6.9.

Consolidation Time (day)

0 50 100 150 200 250 300 350 400
0.0
-0.5 Field meaurements (SP12)
Numerical predictions (n=5 & s=5)
-1.0 Numerical predictions (n=4 & s=4)
Numerical predictions (n=3 & s=3)
Settlement (m)

-1.5 Numerical predictions (n=2 & s=2)

PVD Vacuum on
-2.0 installation
-2.5
Vacuum off
-3.0
-3.5
-4.0
-4.5
-5.0

Figure 6.9 Results of numerical parametric study: Ballina Bypass trial embankment
at SP12

According to the numerical results in Figure 6.9, using rs/rw=2 and kh/ks=2 causes a
settlement of 4.8 m at the end of the vacuum process while the settlement is 4.0 m
when rs/rw=5 and kh/ks=5 are used. This indicates that by varying rs/rw and kh/ks
between 2 to 5 a considerable reduction is caused in the degree of consolidation.

216
Figure 6.9 shows that the smearing effect on the consolidation process is much more
in low ranges of rs/rw and kh/ks.
A comparison of the predicted and measured variations of excess pore water
pressure with time for the transducer P3 located 11.8 m deep and 0. 5 m away from
the centreline of the embankment (refer to Figures 5.15 and 5.16) is shown in Figure
6.10. According to the numerical predictions, the rate of consolidation accelerated,
due to installation of vertical drains. It can be observed that the numerical excess
pore pressure curve has experienced a sudden change to -70 kPa when the vacuum
pressure was applied, followed by an incremental trend due to the embankment
construction.

Consloidation time (day)

190.0 9.0
Field measurements (-11.8m)
170.0 8.0
Numerical predictions (n=2, s=2)
150.0 Numerical predictions (n=4, s=4) 7.0
130.0 Numerical predictions (n=5, s=5)
Excess pore pressure (kPa)

6.0
Vacuum Guage
110.0
Surcharge thickness 5.0

Surcharge thickness (m)

90.0
4.0
Vacuum on
70.0
3.0
50.0 Vacuum off
2.0
30.0
1.0
10.0
0.0
-10.0 0 50 100 150 200 250 300 350 400
-1.0
-30.0 PVD
-50.0 Installation -2.0

-70.0 -3.0

-90.0 -4.0

-110.0 -5.0

Figure 6.10 Variation of excess pore pressure variation with time for Ballina Bypass
trial embankment at P3 (depth 11.8m)

The numerical results in Figure 6.10 clearly indicate that the variations in the
properties of the smear zone affected the dissipation of excess pore water pressure
quite considerably, and as expected, the higher smear zone permeability accelerated
this dissipation. Figure 6.10 also shows that the excess pore water pressures were
fully dissipated at the end of vacuum process, where rs/rw=2 and kh/ks=2, confirming
the numerically predicted settlement results in Figure 6.9. Using rs/rw=5 and kh/ks=5
prolonged the dissipation of excess pore water pressure with the minimum settlement
at the end of the vacuum process. It can be noted that the variation of rs/rw and kh/ks

217
in the low ranges (2 to 3) was more critical and influenced the dissipation of excess
pore water pressure quite considerably, which is in agreement with the settlement
predictions given in Figure 6.9. In Figure 6.10, the measured excess pore water
pressures due to the staged construction of trial embankment and the application of a
negative vacuum pressure were plotted separately, while the numerical predictions
were plotted as a combined graph.
Figure 6.10 shows the disparities in the predicted excess pore water
pressures and field measurements. The field measurements showed that the excess
pore pressure did not dissipate immediately after loading or construction was
completed, but actually increased or stabilised for a period before decreasing. The
abnormal behaviour of excess pore water pressure has been reported in many field
studies (Conlin and Maddox, 1985; Kabbaj et al., 1988; Rowe and Li, 2002), and two
reasons have already been proposed to explain this anomalous behaviour; the
Mandel-Cryer effect and volumetric strain softening. Schiffman et al. (1969)
reported that the Mandel-Cryer effect is due to the increase in total stress caused by
the volumetric strain compatibility. The Mandel-Cryer effect called after Mandel
(1953) and Cryer (1963), based on their observations related to the abnormal
generation of excess pore water pressure. Cryer (1963) analysed the process of
consolidation by applying pressure around a saturated porous sphere whose surface is
free to drain under the pressure applied. Here the total stress at the centre of the
sphere increased temporarily because the dissipation of excess pore water pressure at
the centre was delayed, which increases the excess pore water pressure for some time
before dissipation begins. In addition, this increase or delay in the dissipation of
excess pore water pressure may be the result of volumetric strain softening due to
unstable behaviour during consolidation when the stress paths depart from the failure
line. The constitutive model developed by Kimoto and Oka (2005) can capture the
increase in pore water pressure due to stagnation. Moreover, Asaoka et al. (2000)
reported that as the decay of over consolidation is much faster than the degradation
of the structure in clay during consolidation, softening becomes possible as the
volume is compressed, even under a considerably low stress ratio.

218
6.3.4 Sunshine Trial Embankment

The proposed back calculation scheme was used while adopting different
combinations of smear zone properties to conduct a systematic parametric study
simulating the Sunshine trial embankment, the results of which are reported in Figure
6.11. For this purpose the extent ratio (s=rs/rw) was a constant value of 3, while the
permeability ratio (n=kh/ks) varied from 2 to 6. Details of the geotechnical properties
of the site, construction sequence, and the numerical simulation using FLAC have
been explained in Chapter 5, Section 5.4.

Consolidation time (day)

0 10 20 30 40 50 60 70 80 90
0.0
-0.1
Stage 3
-0.2
Stage 1
-0.3
Settlement (m)

-0.4 Stage 2

-0.5
Field measurements (P1)
-0.6 Numerical predictions (n=2 & s=3)
-0.7 Numerical predictions (n=3 & s=3)
Numerical predictions (n=4 & s=3)
-0.8
Numerical predictions (n=5 & s=3)
-0.9 Numerical predictions (n=6 & s=3)

-1.0

Figure 6.11 Results of numerical parametric study: Sunshine trial embankment at P1

Figure 6.11 shows that there was 1.0 m of settlement after 90 days of consolidation,
where kh/ks=2 and rs/rm=3, while using kh/ks=6 and rs/rm=3 could reduce
consolidation by 0.3 m after the same elapsed time. This result shows that the
variation of smear zone properties can affect the rate of consolidation rate quite
considerably. Figure 6.11 indicates that kh/ks=4 and rs/rm=3 provides the best fit with
the field measurements. However, according to the plotted settlement graphs in
Figure 6.11, the observational approach cannot be used to predict the optimum smear
zone properties in the early stages of constructing the trial embankment and a
systematic procedure for this calculation is needed to determine the case having
minimum errors with the field measurements.
219
It can be noted that the permeability ratio of the smear zone is not a key factor in the
first stage of constructing the embankment, which lasted 30 days. However, when the
height was increased from 1.15 m to 2.85 m, the variations in the permeability ratio
(kh/ks) played a more significant role in the predicted settlement curve. It is clearly
observed that the smear zone uncertainties can significantly affect the consolidation
time, particularly in the higher degrees of consolidation.

6.3.5 Large Scale Consolidometer

The large scale consolidometer test was simulated using the FLAC code (see Chapter
5, Section 5.6 for details) and the back analysis procedure was applied to perform the
parametric study with a constant extent ratio (s=rs/rw) of 3 and a smear zone
permeability ratio (n=kh/ks) between 1.5 and 3.0. The results of the numerical
parametric study are shown in Figure 6.12.

Time (day)
0 5 10 15 20 25 30
0 0%
Large-scale

Degree of consolidation (%)

consolidometer 10%
10
Settlement (mm)

20%
20

30%
30
Lab measurements
n=1.5 & s=3 40%
40 n=1.7 & s=3
n=2 & s=3
n=3 & s=3
50%
(b)
50

Figure 6.12 Results of numerical parametric study: Large scale consolidometer test

According to the graphs illustrated in Figure 6.12, the predicted settlement at the end
of the consolidation test increased by 30% (from 35mm to 46mm) by varying the
permeability ratio (n=kh/ks) from 3 to 1.5. This confirmed the substantial effect that
the smear zone properties had on the PVD assisted preloading design. Figure 6.12
indicates that where kh/ks =1.5 and kh/ks =1.7 they were in good agreement with the
test measurements, but selecting the optimum ratios by observation is not feasible.

220
The predicted settlements for different combinations of the extent and permeability
of the smear zone were located in a narrow band in the early stages of applying the
surcharge, while the settlement-time curves diverged due to the increase in
consolidation time. The tight settlement graphs caused problems in determining the
best fitted curve to the test measurements and corresponding smear zone
characteristics by observation in the early stages of preloading, which reconfirmed
the importance of using the proposed procedure shown in Figure 6.1.

6.4 STEP III: DETERMINING THE ERROR

The conducted parametric studies revealed that the observational technique cannot be
used to predict the smear zone properties precisely in the early stages of constructing
the trial embankment, which resulted in the settlement curves having the best
agreement with the measurements. Therefore the existing error between the field
measurements and numerical predictions for the smear zone properties applied must
be determined at every time step (corresponding to the degree of consolidation) in
order to find the optimum combination of rs/rw and kh/ks. For this reason, the
normalised cumulative error at each time step during the consolidation process was
calculated using Equation (6.2).

ሺௌ೟ ሻ೔ ିሺௌ೟೛ ሻ೔
ሺ‫ܧ‬௧ ሻ௡ ൌ σ௡௜ୀଵ ேൈௌ೑
(6.2)

where, Et is the normalised cumulative error at time t, n is the observation point

number, N is the total number of observation points, St is the field settlement at time
t, Stp is the predicted settlement at time t, and Sf is the final primary consolidation
settlement.
The curves in Figure 6.13 show the changes of normalised cumulative error
with the degree of consolidation for all the simulated case studies where selected
combinations of smear zone properties were used. The degree of consolidation was
determined for each case by dividing the predicted settlement at every numerical step
with the final calculated settlement.

221
 Cumbalum
(a)
Trial Embankment

0.10
Ballina Bypass (b)
0.09 n=2 & s=2
Trial Embankment
n=3 & s=3
0.08
n=4 & s=4
Cumulative Error (E)

0.07
n=5 & s=4
0.06
0.05
0.04
0.03
0.02
0.01
0.00
5% 15% 25% 35% 45% 55% 65% 75% 85%
Degree of Consolidation (U%)
0.08
n=2 & s=3
Sunshine (c)
0.07 Trial Embankment
n=3 & s=3
0.06 n=4 & s=3
Cumulative error (E)

n=5 & s=3

0.05
n=6 & s=3
0.04

0.03

0.02

0.01

0.00
0% 5% 10% 15% 20% 25% 30% 35% 40% 45%
Degree of consolidation (U%)

222
 0.30
n=2 & s=3 Chittagong (d)
(c)
Trial Embankment
n=3 & s=3
Accumulative error (E) n=4 & s=3
n=5 & s=3
Cumulative Error (E)

0.20

0.10

0.00
20% 30% 40% 50% 60%
Degree of consolidation (%)

0.60
n=1.5 & s=3 Large-Scale (e)
(b)
Consolidometer Test
0.50 n=1.7 & s=3
n=2 & s=3
Cumulative Error (E)
Accumulative error (E)

0.40 n=5 & s=3

0.30

0.20

0.10

0.00
0% 10% 20% 30% 40% 50% 60%
Degree of consolidation (%)

Figure 6.13 Normalised cumulative error versus degree of consolidation for different
smear zone properties, (a) Cumbalum trial embankment at SP9, (b) Ballina Bypass
trial embankment at SP12, (c) Sunshine trial embankment at P1, (d) Chittagong Port
trial embankment, and (e) large-scale consolidometer

In each case study the best smear zone properties (s=rs/rw and n =kh/ks) predicted
belonged to the case with the minimum final cumulative error. The final cumulative
errors for different combinations of n and s are shown in Table 6.3. The highlighted
cells are the smear zone properties that resulted in the minimum final cumulative

223
error and where the corresponding curve had the best fit with the field measurements.
These can be reported as the best smear zone characteristics predicted.

Table 6.3 The final cumulative errors for different combinations of smear zone
properties
Cumbalum Trial Embankment
n (kh/ks) 2 3 4 5 6
s (rs/rm) 2 3 4 5 6
Ef 0.113 0.083 0.032 0.011 0.029
Ballina Bypass Trial Embankment
n (kh/ks) 2 3 4 5
s (rs/rm) 2 3 4 5
Ef 0.095 0.032 0.021 0.037
Sunshine Trial Embankment
n (kh/ks) 2 3 4 5 6
s (rs/rm) 3 3 3 3 3
Ef 0.068 0.034 0.010 0.018 0.027
Chittagong Sea Port Trial Embankment
n (kh/ks) 2 3 4 5
s (rs/rm) 3 3 3 3
Ef 0.22 0.34 0.67 1.03
Large Consolidometer Test
n (kh/ks) 1.5 1.7 2 3
s (rs/rm) 3 3 3 3
Ef 0.2 0.15 0.3 0.53

Note: Ef denotes the final cumulative error

6.5 STEP IV: DETERMINING THE MINIMUM REQUIRED

MONITORING TIME

Figures 6.13 indicates that estimating the smear zone characteristics at the early
stages after applying the surcharge is a challenging task and accurate values may not
readily be obtained. To determine the minimum degree of consolidation (minimum
waiting time after applying the surcharge) that results in predicting the most accurate
rs/rw and kh/ks values, the final cumulative error corresponding to the case with the
minimum cumulative error at every degree of consolidation was determined and the
results are plotted in Figures 6.14(a) to 6.14(e). The minimum degree of
consolidation required and the corresponding time belong to the first point with the
minimum final cumulative error. The predicted smear zone properties (rs/rm and
kh/ks) at that point can be reported as reliable values for practical design purposes.

224
 0.120
Cumbalum trial (a)
embankement
Final cumulative error 0.100
minimum accumulative error belongs to the
case with kh/ks=2 & rs/rm=2
0.080
minimum accumulative error belongs to the
case with kh/ks=3 & rs/rm=3
0.060

0.040 minimum accumulative error belongs to the

case with kh/ks=5 & rs/rm=5

0.020

0.000
0% 10% 20% 30% 40% 50% 60% 70% 80%
Degree of consolidation (U%)

0.045
Ballina Bypass trial (b)
0.040 vacuum embankement
minimum accumulative error belongs
0.035
Final cumulative error

to the case with kh/ks=2 & rs/rm=2

0.030
minimum accumulative error belongs to
the case with kh/ks=3 & rs/rm=3
0.025
minimum accumulative error belongs to
0.020 the case with kh/ks=4 & rs/rm=4

0.015

0.010

0.005

0.000
5% 15% 25% 35% 45% 55% 65% 75% 85%
Degree of Consolidation

0.020
Sunshine trial (c)
embankement

0.015 minimum accumulative error belongs to

Final cumulative error

the case with kh/ks=5 & rs/rm=3

0.010

0.005 minimum accumulative error belongs to

the case with kh/ks=4 & rs/rm=3

0.000
0% 5% 10% 15% 20% 25% 30% 35% 40% 45%
Degree of consolidation (U%)

225
 1.40
Chittagong Sea Port (c)
(d)
trial embankement
1.20
error (E)

1.00
cumulative error

minimum accumulative error belongs to

accumulative

0.80 the case with k h/ks=3 & rs/rm=3

0.60 minimum accumulative error belongs to

the case with kh/ks=2 & rs/rm=3
Final

0.40
Final

0.20

0.00
20% 25% 30% 35% 40% 45% 50% 55% 60%
Degree of consolidation (U%)

0.60
Large-scale (b)
(e)
consolidometer
minimum accumulative error belongs
0.50
to the case with kh/ks=3 & rs/rm=3
error
cumulative error

0.40 minimum accumulative error belongs

Final accumulative

to the case with kh/ks=1.5 & rs/rm=3

minimum accumulative error belongs

0.30
to the case with kh/ks=1.7 & rs/rm=3

0.20

0.10

0.00
0% 5% 10% 15% 20% 25% 30% 35% 40% 45%
Degree of consolidation (U%)

Figure 6.14 Total cumulative error (for the smear zone properties resulting in
minimum cumulative error) versus degree of consolidation, (a) Cumbalum trial
embankment at SP9, (b) Ballina Bypass trial embankment at SP12, (c) Sunshine trial
embankment at P1, (d) Chittagong Port trial embankment at G1, and (e) Large-scale
consolidometer test

According to the graph plotted for the Cumbalum trial embankment in Figure
6.14(a), 33% degree of consolidation was the minimum period that must be
considered to estimate the smear zone characteristics accurately. Figure 6.14(b)
shows there was no change in the value of the final cumulative error after 26%
degree of consolidation for the Ballina Bypass trial embankment that corresponds to

226
a case with smear zone properties of rs/rw=4 and kh/ks=4, and therefore reliable
predictions for the smear zone properties predictions can be obtained after 26%
degree of consolidation. Figure 6.14(c) shows that the smear zone properties for the
Sunshine trial embankment can be predicted quite well when at least 16% degree of
consolidation has been obtained. Referring to Figure 6.14(c), the final cumulative
error for Sunshine trial embankment has a constant and minimum value of 0.01 after
16% degree of consolidation, which belongs to the case with kh/ks=4 and rs/rm=3.
Figure 6.14(d) shows a minimum final cumulative error of 0.17 after the
Chittagong trial embankment had reached 25% degree of consolidation; this belongs
to the case with smear zone properties of rs/rw=2 and kh/ks=3. According to Figure
6.14(e), the permeability ratio is 1.5 (rs/rm=3) with a final cumulative error of 0.20
between 10% and 24% degree of consolidation for the large consolidometer test. The
minimum final cumulative error has a constant value of 0.15 after 24% degree of
consolidation, which corresponds to the case with smear zone properties of rs/rw=1.7
and kh/ks=3. This variation in the smear zone properties that resulted in a minimum
final cumulative error was insignificant after 10% degree of consolidation, which can
be reported as the minimum degree of consolidation needed for acquiring reliable
smear zone properties. The minimum required degree of consolidation and
corresponding time resulting in predictable smear zone properties for each case study
is summarised in Table 6.4.

Table 6.4 Minimum required degree of consolidation and corresponding time

resulting in predicting reliable smear zone properties
Cumbalum Ballina Bypass Sunshine Chittagong Large
Case Study
TR TR TR Sea Port Consolidometer
U%min 33% 26% 16% 25% 10%

tmin (day) 600 104 15 14 3

(kh/ks)opt 5 4 4 2 1.5-1.7

(rs/rm)opt 5 4 3 3 3

Note: TR is trial embankment; U%min is the minimum required degree of consolidation; tmin
is the corresponding consolidation time to the U% min; (kh/ks)opt is the predicted permeability
ratio at U%min; (rs/rm)opt is the predicted extent ratio at U%min

227
The results summarised in Table 6.4 indicate that the designed systematic procedure
(Figure 6.1) can be used to accurately estimate the smear zone properties in the early
stages of constructing a trial embankment. Therefore, constructing a trial
embankment with field measurements, in conjunction with the proposed systematic
procedure (Figure 6.1), is a very practical solution for accurately estimating the
smear zone properties and significantly reducing the time and cost of construction.
According to Table 6.4, the minimum required degree of consolidation needed to
predict reliable smear zone characteristics varied between 10% and 33% for the large
scale consolidometer test and the Cumbalum trial embankment, respectively. This
showed that a shorter waiting time was needed to accurately predict the smear zone
properties in the laboratory PVD assisted consolidation test compared to the time
needed to construct the trial embankment.
According to the back calculation results, 33% can be the minimum
required degree of consolidation needed to reliably estimate the extent and
permeability of the smear zone during the trial embankment construction process.

6.6 SUMMARY

In this chapter, a systematic procedure (explained in Chapter 3) was used to

determine the minimum degree of consolidation needed to predict reliable smear
zone properties. For this purpose a four-step solution combined with the developed
numerical code (Figure 6.1) was proposed, that includes: (i) an estimation of
primary consolidation settlement, (ii) a parametric study to estimate various extent
and permeability of the smear zone, (iii) calculating the settlement prediction error,
and (iv) determination of the minimum time needed for back calculating reliable
smear zone properties. The PVD assisted preloading case studies, including four trial
embankments (i.e. Chittagong Sea Port, Cumbalum, Ballina Bypass, and Sunshine)
and one large scale laboratory consolidometer test (described in Chapter 5) were
used in this investigation.
The first three steps of the developed back calculation procedure, in
conjunction with the FLAC code, were used to calculate the normalised cumulative
error at each stage of consolidation in order to determine the error existing between
the field measurements and the numerical predictions for the applied smear zone
properties at every degree of consolidation, and find the optimum combination of

228
rs/rw and kh/ks. The smear zone properties that resulted in the minimum final
cumulative error were presumed to be the best predicted smear zone characteristics.
The results of this stage indicated that estimating the smear zone characteristics at the
early stages of constructing the trial embankment was a challenging task and
accurate values may not be obtained straightaway.
The parametric studies indicated that although both the extent and
permeability of the smear zone affected the consolidation process quite substantially,
the permeability ratio was more critical. Moreover, the influence of the variability in
rs/rm became important when the permeability of the smear zone decreased.
According to the numerical results, several combinations of smear zone properties
can be used for practical design purposes.
In the last stage of the proposed solution (Figure 6.1), the minimum degree
of consolidation needed to predict the most accurate rs/rw and kh/ks values were
determined using the outcomes of the previous stages. For this reason, the final
cumulative error corresponding to the minimum error accumulated at every degree of
consolidation was determined, and the results for all case studies were plotted. The
minimum required degree of consolidation belongs to the first point with the
minimum final cumulative error. The smear zone properties (rs/rm and kh/ks)
predicted at that point can be reported as reliable values for practical design
purposes.
The minimum required degree of consolidation for the Cumbalum, Ballina,
Sunshine, and Chittagong trial embankments were determined as 33%, 26%, 16%,
and 25%, respectively. This minimum value was determined as low as 10% for a
large consolidometer test, indicating that reliable smear zone properties can be
predicted much faster in PVD assisted laboratory tests than field trial embankment.
According to the analyses presented in this chapter, the extent and permeability of
the smear zone can be predicted using the proposed calculation procedure during
construction of a trial embankment, when at least 33% of the predicted final
settlement has been reached (i.e. after achieving 33% degree of consolidation).

229
 CHAPTER SEVEN

7 CONCLUSIONS AND RECOMMENDATIONS

7.1 SUMMARY

A comprehensive introduction and literature review of prefabricated vertical drains

(PVDs) assisted preloading were presented in Chapters 1 and 2 with a focus on the
effects of smearing. Chapter 2 reviewed the factors affecting the consolidation of
clay deposits improved with PVDs, the concepts of the smear zone generation, the
advancements in determining the properties of the smear zone, the developments in
radial consolidation theory by considering the characteristics of the smear zone, and
the methods available for converting permeability from an axisymmetric to a plane-
strain condition. In Chapter 3 the explicit finite difference program FLAC 2D was
introduced and used to develop a numerical code to simulate ground improvement
using PVD combined with surcharge and vacuum preloading. Furthermore, a
systematic back calculation procedure was proposed to determine the properties of
the smear zone using the data consolidated from laboratory or field preloading
projects. Chapter 4 presented an experimental procedure where a large Rowe cell
was used to simulate the combined vertical and radial consolidation process with
vertical drain by considering artificially introduced smear zone properties. The
laboratory results were used to validate the proposed back calculation procedure to
obtain the properties of the smear zone by adopting the developed numerical code. In
Chapter 5 the properties of the smear zone for five case histories of PVD assisted
preloading were back calculated by using the proposed systematic back calculation
procedure to examine the validity of the method developed for different conditions.
Chapter 6 presented an expanded back analysis procedure to determine the minimum
required degree of consolidation and corresponding time resulting in predictable
smear zone properties.

230
7.2 CONCLUSIONS

Preloading with prefabricated vertical drains is highly recommended as an effective

ground improvement technique in deposits of soft soil. Installation of vertical drains
accelerates the rate of settlement quite significantly by reducing the drainage path.
Prefabricated vertical drains were installed using a mandrel, but the mandrel disturbs
the soil around the drain to a certain extent and reduces the permeability of the soil in
this region, which is called the smear zone, and affects the consolidation rate quite
considerably. Generally, two major parameters were proposed to characterise the
smear zone; the permeability (ks), and the extent (rs) of the smear zone.
Determination of the smear zone properties is a challenging task and as such, has
been the subject of intense discussion in the literature. Despite several investigations
that were conducted by researchers, no comprehensive approach for predicting the
properties of the smear zone precisely, has been recommended. Thus, developing a
reliable method for estimating the properties of the smear zone can assist practising
engineers increase the accuracy of practical designs and predictions.
According to the current literature, the extent of the smear zone (rs) may
vary from 1.6 to 7 times of the radius of the drain (rw), or 1 to 6 times of the
equivalent diameter of the mandrel (rm) where the proposed range for the
permeability ratio (kh/ks) is 1.03 to 10. It can be noted that wide ranges were
proposed for kh/ks and rs/rm. The assumptive properties for the characteristics of the
smear zone may result in inaccurate predictions of ground behaviour. This can lead
to an early removal of the surcharge during construction resulting in excessive post
construction settlement or excessive construction time that increases the project cost.
In this study, the FLAC program was used to investigate how the properties
of the smear zone affect the consolidation process. According to the results of the
numerical parametric study, the combined effects of the variability in the extent and
permeability of the smear zone may increase the time required to obtain 90% degree
of consolidation with a factor of two. Furthermore, the numerical analyses revealed
that different combinations of the extent and permeability of the smear zone can be
reported as acceptable values, which results in similar predictions of settlement and
excess pore water pressure. Thus, any of the reported combinations can be adopted
for practical design purposes to predict the settlement and excess pore water

231
pressure. This outcome can be used to conduct the parametric study and determine
the optimum combination by assuming the extent ratio (rs/rm) as a constant value and
by changing the permeability ratio (kh/ks).
A fully instrumented large Rowe cell (Figure 4.1a) was used in this study to
conduct the PVD assisted consolidation tests. The intact and smear zones were
formed by the clay with different permeability coefficients, and a vertical sand drain
was placed at the centre of the cell. Readings from the pore water pressure
transducers were used to investigate the influence of the radial and vertical distances
on the excess pore water pressure. According to the laboratory measurements, the
excess pore water pressure reduced as the vertical distance increased from the
impervious boundary (cell base). The excess pore water pressures were dissipated
much faster in the top half of the soil layer indicating that the vertical consolidation
was dominant in the top part of the sample. The laboratory measurements indicated
that the incremental variation of excess pore water with radial distance was sharper
within the smear zone, especially during the initial stages of loading.
In order to facilitate and accelerate the numerical simulation process, a 2D
model can be applied instead of a 3D model, by adopting the equivalent plane-strain
permeability for the intact region and the smear zone. In this study, the equations
available for converting permeability from an axisymmetric state to a plane-strain
condition were evaluated by comparing the plane-strain analysis with the
axisymmetric results and laboratory measurements. Numerical analyses indicated
that adopting the method of conversion proposed by Hird et al. (1992) may result in
better agreement between the plane-strain and axisymmetrically predicted settlement
curves, in the early stages of consolidation. It was noted that adopting the method for
converting axisymmetric to plane-strain proposed by Indraratna et al. (2005a)
combined with Lin et al. (2000) and Hird et al. (1992) resulted in better predictions
in the later stages of consolidation by increasing the surcharge and decreasing the
void ratio.
In this research, a systematic back calculation procedure combined with the
developed numerical code has been proposed to estimate the predicted the properties
of the smear zone precisely using laboratory or field measurements (Figure 3.13). A
numerical code was developed with various capabilities of simulating PVD assisted
preloading with or without vacuum pressures. The code developed was also used to

232
conduct parametric studies. The proposed back calculation procedure can record the
existing error between the numerical results and field measurements in every
computational step. The final output of the back calculation procedure is a prediction
of the extent and permeability of the smear zone that results in the minimum
cumulative error between the numerical results and field measurements. The Rowe
cell test measurements (settlement and pore water pressure) were compared with the
numerical predictions to validate the proposed back calculation procedure. The
results confirmed that the designed back calculation procedure integrated with the
developed numerical code can be used as a reliable tool for accurately predicting the
properties of the smear zone.
Five case studies of PVD assisted preloading, including four trial
embankments and a large scale consolidometer, were numerically simulated and then
parametric studies were conducted for each case to determine the properties of the
smear zone using the proposed back calculation procedure. In each case study, a
specific aspect of the PVD assisted preloading process was included and simulations
were conducted to verify how well the proposed back calculation procedure could
predict the properties of the smear zone under different conditions. The results
validated the capability of this back calculation procedure, in conjunction with the
developed numerical code, to reliably estimate the properties of the smear zone in
PVD assisted preloading combined with vacuum pressure and embankment loading.
The parametric studies conducted on simulated case studies indicated that
the higher smear zone permeability (kh/ks) accelerates the dissipation of excess pore
water pressure. According to the results, the permeability ratio was more a critical
parameter than the ratio of the extent of the smear zone, although the influence of
variation in the extent ratio on the consolidation time cannot be neglected, because it
can affect the consolidation time by more than 25%. Furthermore, the variation of the
extent ratio (rs/rm) and the permeability ratio (kh/ks) in the low ranges (i.e. 2 to 3) was
more critical and substantially influenced the consolidation time and dissipation of
excess pore water pressure. According to the numerical analyses, it can be concluded
that the variability in rs/rm becomes more important when the permeability of smear
zone decreases.
The long time required to construct a trial embankment was a major
challenge in using this method to conduct the back calculation analysis for estimating

233
the characteristics of the smear zone and in many cases may cause considerable delay
in constructing the actual embankment and a significant increase in the project cost.
Estimating the extent and permeability of the smear zone in the early stages of
constructing the trial embankment can convert this method into a very practical,
accurate, and cost effective approach. In this study, and to rectify this problem, an
expanded back calculation procedure was proposed to determine the minimum
required degree of consolidation that would result in predictable smear zone
properties. The expanded back calculation procedure was designed in four steps,
including (i) estimating the primary consolidation settlement, (ii) conducting the
parametric study, (iii) determining he normalised error, and (iv) determining the
minimum waiting time (Figure 6.1). The PVD assisted preloading case studies
including four trial embankments (Chittagong Sea Port, Cumbalum, Ballina Bypass,
and Sunshine) and one large-scale consolidometer test was used to conduct the
analyses. According to the results, the extent and permeability of the smear zone can
be predicted quite well by using the proposed systematic back calculation procedure,
when at least 33% of the final settlement predicted, has been obtained (i.e. after
achieving 33% degree of consolidation).
According to the results of the parametric study, the back calculated smear
zone permeability ratio for the laboratory study on PVD assisted preloading test case
was less than the value obtained for the real trial embankments. Furthermore, the
results of the back calculation showed that the minimum time needed to reliably
predict the properties of the smear zone for the laboratory PVD assisted preloading
tests can be determined after obtaining 10% degree of consolidation, and was
approximately 65% faster than the value determined for the real embankment.
It is recommended that engineers use the construction of a trial embankment
in combination with the proposed systematic back calculation procedure in order to
determine the properties of the smear zone with 33% degree of consolidation as the
minimum waiting time needed to obtain reliable values. The back calculation
procedure can be conducted by assuming a constant extent ratio (e.g. rs/rm=3 or
rs/rm=4) and change the permeability ratio to obtain the optimum combination that
results in the minimum amount of cumulative error between the numerical results
and field measurements.

234
7.3 RECOMMENDATIONS FOR FUTURE RESEARCH

This area of research can be further expanded by conducting the following studies:

x In this study, a two-zone hypothesis was used to describe the disturbed area
near the drain. The three-zone hypothesis, which divides the soil surrounding
the drain into the inner smear zone, the outer smear zone (the transition zone),
and the undisturbed zone, can be applied in future research and the results can
be compared to the outcomes of this study.

x The test procedure can be conducted using a larger cell combined with vacuum
pressure to determine the minimum degree of consolidation needed to obtain
reliable smear zone properties. Furthermore, the accuracy of axisymmetric to
plane-strain permeability conversion methods can be evaluated while a vacuum
is being applied.

x The proposed minimum waiting time was obtained through the simulation of
five case studies. Further verifications, using more data from other case studies,
would strengthen this conclusion.

x The proposed back calculation procedure can be combined with the 3D

numerical modelling to predict the properties of the smear zone, and then the
results can be compared to the 2D analyses. Furthermore, the 3D simulation
can be used to evaluate the efficiency of the available axisymmetric to plane-
strain permeability conversion equations.

x The variations in the over consolidation ratio (OCR) inside the smear zone can
be captured to investigate the influence of OCR on mandrel driven PVD
consolidation process, particularly when determining the rates of settlement
and distribution of excess pore water pressure.

x Predicting soft soil creep is generally a challenging research because creep can
play a significant role in the long-term deformation of soil stabilised with PVD
assisted preloading. A large amount of deformation may take place even after
the excess pore pressure has dissipated. The effect of creep on an elasto visco-
plastic constitutive model, which influences the dissipation of excess pore

235
water pressure and settlement rates, can be included in the numerical analysis
to obtain more reliable results.

236
 REFERENCES

AS: 1289.3.5.2, 2002, ‘Methods of testing soils for engineering purposes - Soil
classification tests - Determination of the soil particle density of combined soil
fractions - Vacuum pycnometer method’, Standards Association of Australia.
Aboshi, H. 1992, ‘Applied ground improvement techniques, workshop’, Southeast
Asian Geotechnical Society (SEAGS), Asian Institute of Technology, Bangkok,
Thailand, p 40.
Abuel-Naga, H. and Bouazza, A. 2009, ‘Equivalent diameter of a prefabricated
vertical drain’, Geotextiles and Geomembranes, 27(3), pp. 227-231.
Akagi, T. 1976, ‘Effect of displacement type sand drains on strength and
compressibility of soft clays’, Doctor of Philosophy. thesis, University of Tokyo,
Japan.
Almedia, M. S. S., Danzigner, F. A. B., F., Almedia M. C., L., Carvalho S. R. and M,
Martins I. S. 1993, ‘Performance of an embankment built on a soft disturbed clay’, In
Proceedings of the 3rd International Conf. Case Histories in Geotechnical
Engineering, Missouri, pp. 351-356.
Almeida, S. S., Marques, M. E. S. and Spotti, A. P. 2005, ‘Two case histories of
vertical drain in very soft clays’, Ground improvement case histories, Elsevier, 3,
145-157.
Arulrajah, A., Nikraz, H. and Bo, M. W. 2005, ‘Finite element modelling of marine
clay deformation under reclamation fills’, Ground Improvement, 9(3), pp. 105-118.
Asaoka, A., Nakano, M., Noda, T. and Kaneda, K. 2000, ‘Delayed
compression/consolidation of natural clay due to degradation of soil structure’, Soils
and Foundations, 40(3), pp. 75-85.
Atkinson, K. and Han, W. 2005, ‘Finite difference method’, Theoretical Numerical
Analysis, Springer New York, 39: 249-271.
Atkinson, M. S. and Eldred, P. J. L. 1981, ‘Consolidation of soil using vertical
drains’, Geotechnique, 31(1), pp. 33-43.
Australian Bureau of Statistics 2008. ‘Population projections, Australia, 2006 to
2101’, Available from http://www.abs.gov.au/Ausstats/abs@.nsf/mf/3222.0.
Balasubramaniam, A. S., Huang, M., Bolton, M., Oh, E. Y. N., Bergado, D. T. and
Phienwe, N. j. 2007, ‘Interpretation and analysis of test embankments in soft clays
with and without ground improvement’, Sixteenth southeast Asian geotechnical
conference, Southeast Asian geotechnical society.
Bamunawita, C. 2004, ‘Soft clay foundation improvement via prefabricated vertical
drains and vacuum preloading’, Doctor of Philosophy. thesis, School of Civil,
Mining and Environmental Engineering, Faculty of Engineering, The University of
Wollongong, Australia.

237
Barron, R. A. 1948, ‘Consolidation of fine-grained soils by drain wells’,
Transactions of the American Society of Civli Engineering, 113(1), pp. 718-742.
Basu, D., Basu, P. and Prezzi, M. 2006, ‘Analytical solutions for consolidation aided
by vertical drains’, Geomechanics and Geoengineering, 1(1), pp. 63-71.
Basu, D. and Prezzi, M. 2007, ‘Effect of the smear and transition zones around
prefabricated vertical drains installed in a triangular pattern on the rate of soil
consolidation’, International Journal of Geomechanics, 7(1), pp. 34-43.
Basu, D., Prezzi, M. and Madhav, M. R. 2010, ‘Effect of soil disturbance on
consolidation By prefabricated vertical drains installed in a rectangular pattern’,
Geotech. and Geolog. Eng., 28(1), pp. 61-77.
Bellezza, I. and Fentini, R. 2008, ‘Prefabricated vertical drains: a simplified design
procedure’, In Proceedings of the the ICE - Ground Improvement, 161(4), pp. 173-
178.
Bergado, D. T., Asakami, H., Alfaro, M. C. and Balasubramaniam, A. S. 1991,
‘Smear effects of vertical drains on soft bangkok clay’, Journal of Geotechnical
Engineering, 117(10), pp. 1509-1530.
Bergado, D. T., Manivannan, R. and Balasubramaniam, A. S. 1996, ‘Proposed
criteria for discharge capacity of prefabricated vertical drains’, Geotextiles and
Geomembranes, 14(9), pp. 481-505.
Bergado, D. T., Mukherjeea, K., Alfaroa, M. C. and Balasubramaniama, A.S. 1993,
‘Prediction of vertical-band-drain performance by the finite-element method’,
Geotextiles and Geomembranes, 12(6), pp. 567-586.
Bergadoa, D.T., Singha, N., Sima, S.H., Panichayatuma, B., Sampacoa, C.L. and
Balasubramaniama, A.S. 1990, ‘Improvement of soft Bangkok clay using vertical
geotextile band drains compared with granular piles’, Geotextiles and
Geomembranes, 9(3), pp. 203-231.
Bhosle, P., Vaishampayan, V.V. 2009, ‘Case study for ground improvement using
PVD with preloading for coal and iron ore stack yard’, Annual Conference of the
Indian Geotechnical Society (IGC), Guntur, India, pp. 504-510.
Biot, M. A. 1941, ‘General theory of three-dimensional consolidation’, Applied
Physics, 12, pp. 155-164.
Bishop, D. 2004, ‘A proposed geological model and geotechnical properties of a
NSW estuarine valley: A case study’, In Proceedings of the 9th ANZ Conference on
Geomechanics, Auckland NZ, pp. 261-267.
Bo, M. W. 2004, ‘Discharge capacity of prefabricated vertical drain and their field
measurements’, Geotextiles and Geomembranes, 22(12), pp. 37-48.
Bo, M. W., Arulrajah, A. and Nikraz, H. 2007, ‘Preloading and prefabricated vertical
drains design for foreshore land reclamation projects: a case study’, Ground
Improvement, 11(2), pp. 67-76.

238
Bo, M.W., Chu, J. and Choa, V. 2003, ‘Soil improvement: prefabricated vertical
drain techniques’, Thompson, Singapore.
Britto, A. M. and Gunn, M. J. 1987, ‘Critical state soil mechanics via finite
elements’, Chichester:: Ellis Horwood Limited, p 486.
Budhi, M. 2010, ‘Soil Mechanics and Foundations’, Wiley.
Burns, S. E. and Mayne, P. W. 1998, ‘Monotonic and dilatory pore pressure decay
during piezocone tests in clay’, Canadian Geotechnical Journal, 35(6), pp. 1063–
1073.
Cao, L. F., Chang, M. F., Teh, C. I. and Na, Y. M. 2001, ‘Back-calculation of
consolidation parameters from field measurements at a reclamation site’, Canadian
Geotechnical Journal, 38(4), pp. 755-769.
Carillo, N. 1942, ‘Simple two and three dimensional cases in the theory of
consolidation of soils’, J. Math. and Phys, 21(1), pp. 1-5.
Carroll, R.G., 1983, ‘Geotextile filter criteria’, Engineering Fabrics in
Transportation Construction, Transportation Research Record 916, Transportation
Research Board, Washington, DC, USA.
Casagrande, L. and Poulos, S. 1969, ‘On the effectiveness of sand drains’, Can.
Geotech. J., 6(3), pp. 287-326.
Chai, J. and Miura, N. 1999, ‘Investigation of factors affecting vertical drain
behavior’, Journal of Geotechnical and Geoenvironmental Engineering, 125(3), pp.
216-226.
Chai, J. C. and Bergado, D. T. 1993, ‘Performance of reinforced embankment on
Muar clay deposit’, Soils and Foundations, 33(4), pp. 1-17.
Chai, J. C., Bergado, D. T., Miura, N. and Sakajo, S. 1996, ‘Back calculated field
effect of vertical drain’, In Proceedings of the 2nd Int. Conf. Soft Soil Engrg., Hohai
University, Nanjing, China, pp. 270–275.
Chai, J. C., Miura, N., Sakajo, S. and Bergado, D. T. 1995, ‘Behavior of vertical
drain improved subsoil under embankment loading’, Soils and Foundations, 35(4),
pp. 49-61.
Chai, J., Carter, J. P. and Hayashi, S. 2005, ‘Ground deformation induced by vacuum
consolidation’, Journal of Geotechnical and Geoenvironmental Engineering,
131(12), pp. 1552-1561.
Chai, J., Carter, J. P. and Hayashi, S. 2006, ‘Vacuum consolidation and its
combination with embankment loading’, Canadian Geotechnical Journal, 43(10),
pp. 985-996.
Chai, J., Miura, N. and Bergado, D. T. 2008, ‘Preloading clayey deposit by vacuum
pressure with cap-drain: Analyses versus performance’, Geotextiles and
Geomembranes, 26(3), pp. 220-230.

239
Chai, J. and Shen, S. L. 2001, ‘Simple method of modeling PVD-improved subsoil’,
Journal of Geotechnical and Geoenvironmental Engineering, 27(11), pp. 965-972.
Christopher, B. R. and Holtz, R. D. 1985, ‘Geotextile engineering manual’, FHWA-
TS-86/203, U.S. Federal Highway Administration.
Chu, J., Yan, S. and Indraranata, B. 2008, ‘Vacuum preloading techniques - recent
developments and applications’, GeoCongress , Geosustainability and Geohazard
Mitigation GPS 178, New Orleans, pp. 586-595.
Chu, J., Yan, S. W. and Yang, H. 2000, ‘Soil improvement by the vacuum preloading
method for an oil storage station’, Géotechnique, 50(6), pp. 625-632.
Chung, S. G., Lee, N. K. and Kim, T. H. 2009, ‘Hyperbolic method for prediction of
prefabricated vertical drains performance’, Journal of Geotechnical and
Geoenvironmental Engineering, 135(10), pp. 1519-1528.
Collins, I. F. and Yu, H. S. 1996, ‘Undrained cavity expansion in critical state soils’,
International journal for numerical and analytical methods in geomechanics, 20(7),
pp. 489-516.
Conlin, B.H. and Maddox, W.P. 1985, ‘An assessment of the behaviour of
foundation clay at Tarsiut N-44 Caisson retained island’, In Proceedings of the 17th
Offshore Technology Conference, Houston, Texas, pp. 379-388.
Craig, H. 2000, ‘Soil Mechanics’, Chapman & Hall, London.
Cryer, C.W. 1963, ‘A comparison of the three dimensional consolidation theories of
Biot and Terzaghi’, Quartery Journal of Mechanics and Applied Mathematics, 16(4),
pp. 401-412.
Das, B. M. 2008, ‘Advanced Soil Mechanics’, Taylor & Francis.
Den Hoedt, G., 1981. ‘Laboratory testing of vertical drains’, In Proceedings of the
10th International Conference on Soil Mechanics and Foundation Engineering,
Stockholm, 1, pp. 627–630.
Dhar, A. S., Siddique, A. and Ameen, S. F. 2011, ‘Ground improvement using pre-
loading with prefabricated vertical drains’, International Journal of Geoengineering
Case Histories, 2(2), pp. 86-104.
Eriksson, U., Hansbo, S. and Torstensson, B. A. 2000, ‘Soil improvement at
Stockholm-Arlanda airport’, Ground Improvement, 4(2), pp. 73-80.
Fatahi, B., Khabbaz, H. and Indraratna, B. (2009), ‘Parametric studies on
bioengineering effects of tree root-based suction on ground behaviour’, Ecological
Engineering, 35(10), pp 1415-1426.
Fatahi, B., Khabbaz, H. and Indraratna, B. (2010), ‘Bioengineering ground
improvement considering root water uptake model’, Ecological Engineering, 36(2),
pp 222-229.

240
Fatahi, B., Le, T. M., Le, M. Q. and Khabbaz, H. (2012), ‘Soil creep effects on
ground lateral deformation and pore water pressure under embankments’,
Geomechanics and Geoengineering, 8(2). pp 107-124.
Fellenius, B.H. and Castonguay, N.G, 1985, ‘The Efficiency of band shaped drains –
a full scale laboratory study’, Report to the National Research Council of Canada and
the Industrial Research Assistance Programme.
Fox, P. J. and Berles, J. D. 1997, ‘CS2: A piecewise-linear model for large strain
radial consolidation’, International journal for numerical and analytical methods in
geomechanics, 21(7), pp. 453-475.
Gao, C. 2004, ‘Vacuum preloading method for improving soft soils of higher
permeability’, Ground Improvement, 8(3), pp. 101-107.
Geng, Xueyu, Indraratna, Buddhima and Rujikiatkamjorn, Cholachat 2012,
‘Analytical solutions for a single vertical drain with vacuum and time-dependent
surcharge preloading in membrane and membraneless systems’, International
Journal of Geomechanics, 12(1), pp. 27-42.
Gens, A. and Potts, D. M. 1988, ‘Critical state model in computational
geomechanics’, Engng Comput., 5, pp. 178–197.
Ghandeharioon, A., Indraratna, B. and Rujikiatkamjorn, C. 2010, ‘Analysis of soil
disturbance associated with mandrel-driven prefabricated vertical drains using an
elliptical cavity expansion theory’, International Journal Geomechanics, 10(2), pp.
53-64.
Ghandeharioon, Ali, Indraratna, Buddhima and Rujikiatkamjorn, Cholachat 2012,
‘Laboratory and finite-element investigation of soil disturbance associated with the
installation of mandrel-driven prefabricated vertical drains’, Journal of Geotechnical
and Geoenvironmental Engineering, 138(3), pp. 295-308.
Gibson, R. E. 1963, ‘An analysis of system flexibility and its effect on time-lag in
pore-water pressure measurements’, Géotechnique, 13(1), pp. 1-11.
Graham, B. 2010, ‘Soil mechanics, principles and practice’, Palgrave.
Hansbo, S 1997, ‘Practical aspects of vertical drain design’, In Proceedings of the
14th International Conference on Soil Mechanics and Foundation Engineering,
Hamburg, pp. 1749-1752.
Hansbo, S. 1960, ‘Consolidation of clay: With special reference to influence of
vertical sand drains’, In Proceedings of the Swedish Geotechnical Institute,
Stockholm, Sweden, Swedish Geotechnical Institute.
Hansbo, S. 1979, ‘Consolidation of clay by band-shaped prefabricated drains’,
Ground Engineering, 12(5).
Hansbo, S. 1981, ‘Consolidation of fine-grained soils by prefabricated drains’, In
Proceedings of the 10th International Conference on Soil Mechanics and Foundation
Engineering, Stockholm, Sweden, Balkema, Rotterdam, the Netherlands, pp. 677-
682.

241
Hansbo, S. 1987, ‘Design aspects of vertical drains and lime-column installations’, In
Proceedings of the 9th Southeast Asian Geotech. Conf., Bangkok, Thailand, pp. 8-12.
Hansbo, S. 1994, ‘Foundation Engineering’, Elsevier Sience B. V., Amsterdam.
Hansbo, S., Jamiolkowski, M. and Kok, L. 1981, ‘Consolidation by vertical drains’,
Geotechnique, 31, pp. 45-66.
Hawlader, B. C. and Muhunthan, B. 2002, ‘Numerical study of the factors affecting
the consolidation of clay with vertical drains’, Geotextiles and Geomembranes,
20(4), pp. 213-239.
Hird, C. C. and Moseley, V. J. 2000, ‘Model study of seepage in smear zones around
vertical drains in layered soil’, Géotechnique, 50(1), pp. 89-97.
Hird, C. C. and Sangtian, N. 2002, ‘Model study of seepage in smear zones around
vertical drains in layered soil: further results ’, Géotechnique, 52(5), pp. 375-378.
Hird, C.C., Pyrah, I.C. and Russell, D. 1992, ‘Finite element modelling of vertical
drains beneath embankments on soft ground’, Géotechnique, 42(3), pp. 499-511.
Holtz, R.D. 1987, ‘Preloading with prefabricated vertical strip drains’, Geotextiles
and Geomembranes, 6(1-3), pp. 109-131.
Holtz, R.D. and Holm, B.G. 1973, ‘Excavation and sampling around some sand
drains at Ska°-Edeby, Sweden’, In Proceedings of the Sixth Scandinavian
Geotechnical Meeting, Trondheim, Oslo, Sweden, Norwegian Geotechnical Institute,
pp. 79-85.
Holtz, R.D., Jamiolkowski, M.B., Lancellotta, R. and Pedroni, R. 1991,
‘Prefabricated vertical drains: design and performance’, CIRIA, Butterworth-
Heinemann, London.
Huang, W., Fityus, S., Bishop, D., Smith, D. and Sheng, D. 2006, ‘Finite-element
parametric study of the consolidation behavior of a trial embankment on soft clay’,
International Journal of Geomechanics, 6(5), pp. 328-341.
Indraratna, B. 2009, ‘Recent Advances in the Application of Vertical Drains and
Vacuum Preloading in Soft Soil Stabilisation’, E.H. Davis Memorial Lecture,
Australian Geomechanics Society.
Indraratna, B. 2002, ‘Book chapter: Soft ground improvement by vertical drains’, In-
situ Characetrisation of Soft Soil. O. I. P. Saxena, India and Swets & Zeitlinger
Publishers, The Netherlands.
Indraratna, B., Aljorany, A. and Rujikiatkamjorn, C. 2008, ‘Analytical and numerical
modeling of consolidation by vertical drain beneath a circular embankment’,
International Journal of Geomechanics, 8(3), pp. 199-206.
Indraratna, B. and Bamunawita, C. I. 2002, ‘Soft clay stabilisation by mandrel driven
geosynthetic vertical drains’, Australian Geomechanics Journal, 37(5), pp. 57-86.

242
Indraratna, B., Bamunawita, C. and Khabbaz, H. 2004a, ‘Numerical modeling of
vacuum preloading and field applications’, Canadian Geotechnical Journal, 41(6),
pp. 1098-1110.
Indraratna, B. and Redana, I. W. 1997, ‘Plane-strain modeling of smear effects
associated with vertical drains’, Journal of Geotechnical and Geoenvironmental
Engineering, 123(5), pp. 474-478.
Indraratna, B. and Redana, I. W. 1998, ‘Laboratory determination of smear zone due
to vertical drain installation’, Journal of Geotechnical and Geoenvironmental
Engineering, 124(2), pp. 180-184.
Indraratna, B. and Redana, I. W. 2000, ‘Numerical modeling of vertical drains with
smear and well resistance installed in soft clay’, Canadian Geotechnical Journal,
37(1), pp. 132-145.
Indraratna, B. and Rujikiatkamjorn, C. 2004b, ‘Mathematical modeling and field
evaluation of embankment stabilized with vertical drains incorporating vacuum
preloading’, The Fifth International Conference on Case Histories in Geotechnical
Engineering, New York, pp. 201-208.
Indraratna, B., Rujikiatkamjorn, C. and Chu, J. 2007a, ‘Soft clay stabilization with
geosynthetic vertical drains beneath road and railway embankments: A Critical
Review of Analytical Solutions and Numerical Analysis’, New Peaks in Geotechnics,
Denver, ASCE, pp. 1-20.
Indraratna, B., Rujikiatkamjorn, C., Kelly , R. and Buys, H. 2010, ‘Sustainable soil
improvement via vacuum preloading’, Proc. of the ICE - Ground Improvement,
163(1), pp. 31-42.
Indraratna, B., Rujikiatkamjorn, C. and Sathananthan, I. 2005a, ‘Analytical and
numerical solutions for a single vertical drain including the effects of vacuum
preloading’, Canadian Geotechnical Journal, 42(4), pp. 994-1014.
Indraratna, B., Rujikiatkamjorn, C., Sathananthan, I., Shahin, M. and Khabbaz, H.
2005c, ‘Analytical and numerical solutions for soft clay consolidation using
geosynthetic vertical drains with special reference to embankments’, The Fifth
International Geotechnical Engineering Conference, Cairo, Egypt, pp. 55-86.
Indraratna, B., Sathananthan, I. , Rujikiatkamjorn, C. and Balasubramaniam, A. S.
2005b, ‘Analytical and numerical modeling of soft soil stabilized by prefabricated
vertical drains incorporating vacuum preloading’, International Journal of
Geomechanics, 5(2), pp. 114-124.
Indraratna, B., Rujikiatkamjorn, C., Ameratunga, J. and Boyle, P. 2011,
‘Performance and prediction of vacuum combined surcharge consolidation at Port of
Brisbane’, Geotechnical and Geoenvironmental Engineering, 137(11), pp. 1009-
1018.
Indraratna, B., Rujikiatkamjorn, C., Balasubramaniam, C., Wijeyakulasuriya, V.
2005d, ‘Predictions and observations of soft clay foundations stabilized with

243
geosynthetic drains and vacuum surcharge’, Ground Improvement: Case Histories,
Elsevier, 3(Chapter 7), 199-230.
Itasca Consulting Group, Inc. 2008, ‘FLAC – Fast lagrangian analysis of continua’,
Version 6.0 User’s Manual, Itasca, Minneapolis.
Jacob, C.E. 1940, ‘On the flow of water in an elastic artesian aquifer’, Transactions
of the American Geophysical Union 21, Part II, pp. 574–586.
Jamiolkowski, M., Lancellotta, R. and Wolski, W. 1983, ‘Precompression and
speeding up consolidation’, In Proceedings of the 8th European Conference on Soil
Mechanics and Foundation Engineering, Helsinki, 23-26 May 1983, Balkema,
Rotterdam, the Netherlands, pp. 1201-1226.
Johnson, S. J. 1970, ‘Foundation precompression with vertical sand drains’, Journal
of Soil Mechanics & Foundations Division, 96(1), pp. 145–175.
Kabbaj, M., Tavenas, F. and Leroueil, S. 1988, ‘In situ and laboratory stress-strain
relationships’, Geotechnique, 38(1), pp. 83-100.
Kamon, M. and Bergado, D. T. 1992, ‘Ground improvement techniques’, 9th Asian
regional conf. on Soil Mech. and Found. Eng'g, Bangkok, pp. 526-546.
Kawaida, M., Kosaka, T., Yamada, K. and Nakakuma, K. 2012, ‘Vacuum
consolidation for construction of expressway embankment on extremely soft
ground’, Journal of the Society of Materials Science, Japan, 61(1), pp. 46-51.
Kelly, R. B. 2008, ‘Back analysis of the Cumbalum trial embankment’, Australian
Geomechanics, 43(1), pp. 47-54.
Kelly, R. B. and Wong, P. K. 2009, ‘An embankment constructed using vacuum
consolidation’, Australian Geomechanics, 44(2), pp. 55-64.
Kelly, R., Small, J. and Wong, P. 2008, ‘Construction of an embankment using
vacuum consolidation and surcharge fill’, In Proceedings of the GeoCongress 2008:
Geosustainability and Geohazard Mitigation (GSP 178), Orleans, Louisiana, 9–12
March 2008, American Society of Civil Engineers, New York.
Kianfar, K., Indraratna, B. and Rujikiatkamjorn, C. 2013, ‘Radial consolidation
model incorporating the effects of vacuum preloading and non-Darcian flow’,
Geotechnique, 63(12), pp. 1060-1073.
Kim, R., Choi, Y., Lee, J. and Lee, W. 2010, ‘Evaluation of the PVD smear zone
using micro penetrometer’, In Proceedings of the GeoFlorida 2010: Advances in
Analysis, Modeling & Design (GSP 199), 20-24 Feb 2010, American Society of Civil
Engineers, New York, pp. 998-1007.
Kim, T., Kim, N., Tumay, M. and Lee, W. 2007, ‘Spatial distribution of excess pore-
water pressure due to piezocone penetration in overconsolidated clay’, Journal of
Geotechnical and Geoenvironmental Engineering, 133(6), pp. 674–683.

244
Kimoto, S. and Oka, F. 2005, ‘An elasto-viscoplastic model for clay considering
destructuralisation and consolidation analysis of unstable behaviour’, Soils and
Foundations, 45(2), pp. 29-42.
Kjellman, W. 1952, ‘Consolidation of clayey soils by atmospheric pressure’, In
Proceedings of the Conference on Soil Stabilization, Massachusetts Institute of
Technology, Boston, pp. 258-263.
Koda, E., Szymanski, A. and Wolski, W. 1984, ‘Laboratory tests on geodrains
durability in organic soils’, Seminar on Laboratory Testing of Prefabricated Band-
Shaped Drains, Milan, Italy.
Koerner, R. M. 1987, ‘Soft soil stabilization using geosynthetics’, Elsevier, Applied
Science, London.
Koerner, R. M. and Ko, F. K. 1992, ‘Laboratory studies on long-term drainage
capability of geotextiles’, In Proceedings of the 2nd Int. Conf. on Geotextiles, Las
Vegas, pp. 91-95.
Kremer, R., De Japer, W., Maagdenberg, A., Megrogel, I. and Oostveen, J. 1982,
‘Quality standards for vertical drains’, In Proceedings of the Second International
Conference on Geotextiles, Las Vegas, Nevada, pp. 319–324.
Kremer, R. R. H. J., Oostven, J. P., Van Weele, A. F., Dejager, W. F. J. and
Meyvogel, I. J. 1983, ‘The quality of vertical drainage’, In Proceedings of the Eighth
European Conference on Soil Mechanics and Foundation Engineering, Helsinki, pp.
721–726.
Le, T. M., Fatahi, B. and Khabbaz, H. (2012), ‘Viscous Behaviour of Soft Clay and
Inducing Factors’, Geotechnical and Geological Engineering, 30(5), pp 1069-1083.
Leo, C. J. 2004, ‘Equal strain consolidation by vertical drains’, Journal of
Geotechnical and Geoenvironmental Engineering, 130(3), pp. 316-327.
Leroueil, S. 1997, ‘Closure to 'Compressibility of clays: Fundamental and practical
aspects’, J. Geotech. and Geoenvir. Engrg., 123(9), 895p.
Lin, D. G., Kim, H. K. and Balasubramaniam, A. S. 2000, ‘Numerical modelling of
prefabricated vertical drain’, Geotechnical Engineering Journal of the Southeast
Asian Geotechnical Society, 31(2), pp. 190-125.
Lin, D.G. and Chang, K.T. 2009, ‘Three-dimensional numerical modelling of soft
ground improved by prefabricated vertical drains’, Geosynthetics International,
16(5), pp. 339-353.
Liu, C. and Evett, J 2013, ‘Soils and foundations’, Pearson.
Liu, H. l. and Chu, J. 2009, ‘A new type of prefabricated vertical drain with
improved properties’, Geotextiles and Geomembranes, 27(2), pp. 152-155.
Liu, H. L., Chu, J. and Ren, Z. 2009, ‘New methods for measuring the installation
depth of prefabricated vertical drains’, Geotextiles and Geomembranes, 27(6), pp.
493-496.

245
Lo, D. O. K. 1991, ‘Soil improvement by vertical drains’, Doctor of Philosophy.
thesis, The University of Illinois at Urbana.
Logan, D. L. 2007, ‘A First course in the finite element method’, Thompson,
Tampa, USA.
Long, R. and Covo, A. 1994, ‘Equivalent diameter of vertical drains with an pblong
cross section’, Journal of Geotechnical Engineering, 120(9), pp. 1625-1630.
Madhav, M. R., Park, Y. M. and Miura, N. 1993, ‘Modelling and study of smear
zones around band shaped drins’, Soils and Foundations, 33(4), pp. 135-147.
Mandel, J 1953, ‘Consolidation des sols (Etude Mathematique)’, Geotechnique, 3(7),
pp. 287-299.
Masse, F., Yee, K. and Varaksin, S. 2002, ‘Successful application of Vacuum
Consolidation method to Nakdong River soft clay in Kimhae, South Korea’, 4the
Conference on Ground Ipmrovement Techniques, Kuala Lumpur, Malaysia.
Mesri, G., Kelly, W. E., Vallee, R. P. and Andersland, O. B. 1973, ‘Coefficient of
secondary compression (discussion)’, Journal of Soil Mechanics & Foundations,
Div 99, no. Proc Paper.
Mesri, G. & Lo, D.O.K. 1991, ‘Field Performance of Prefabricated Vertical Drains’,
In Proceedings of the International Conference on Geotechnical Engineering for
Coastal Development-Theory to Practice, Yokohama, Japan, 1, pp 231–236.
Mesri, G., Lo, D.O.K. and Feng, T-W. 1994, ‘Settlement of embankments on soft
clays’, In Vertical and Horizontal Deformations of Foundations and Embankments,
Proceedings of Settlement 94, GSP 40, ASCE, New York, pp. 8-56.
Mohamedelhassan, E. and Shang, J. Q. 2002, ‘Vacuum and surcharge combined one-
dimensional consolidation of clay soils’, Canadian Geotechnical Journal, 39(5), pp.
1126-1138.
Morris, P. H. 2005, ‘Analytical solutions of linear finite- and small-strain one-
dimensional consolidation’, International journal for numerical and analytical
methods in geomechanics, 29(2), pp. 127-140.
Murray, H. H. 1999, ‘Applied clay mineralogy today and tomorrow’, Clay Minerals
34(1), pp. 39–49.
Nagtegaal, J. C., Parks, D. M. and Rice, J. R. 1974, ‘On numerically accurate finite
element solutions in the fully plastic range’, Comp. Meth. in Appl. Mech. and Eng.,
4, pp. 153-177.
Nelson, J.D. and Miller, D.G. 1997, ‘Expansive soil: problems and practice in
foundation and pavement engineering’, John Wiley & Sons, New York.
Onoue, A. 1988, ‘Consolidation by vertical drains taking well resistance and smear
into consideration’, Soils and Foundations, 28(4), pp. 165-174.

246
Onoue, A., Ting, N. H., Germaine, J. T. and Whitman, R. V. 1991, ‘Permeability of
disturbed zone around vertical drains’, In Proceedings of the Geotechnical
Engineering Congress (GSP 27), American Society of Civil Engineers, New York,
pp. 879-890.
Oostveen, J. P. 1986, ‘Research acticities on vertical drainahe in the Netherland’,
Seminar on vertical drainage, Delf University of Technology, pp. 8-9.
Perlof, W. H., Nair, K. and Smith, J. G. 1965, ‘Effect of measuring system on pore
water pressures in the consolidation test’, In Proceedings of the 6th ICSMFE,
Montreal, pp. 338-341.
Pradhan, T. B. S., Imai, G., Murata, T., Kamon, M. and Suwa, S. 1993, ‘Experiment
study on the equivalent diameter of a prefabricated band-shaped drain’, In
Proceedings of the Proc. 11th Southeast Asian Geotech. Conf., 1.
Qian, J. H., Zhao, W. B., Cheung, Y. K. and Lee, P. K. K. 2003, ‘The theory and
practice of vacuum preloading’, Computers and Geotechnics, 13(2), pp. 103-118.
Queensland Department of Transport, Sunshine Motorway Stage 2, 1992,
‘Performance of the trial embankment area 2A (Ch 28490-28640)’, Report R1802,
Materials and Geotechnical Services Branch.
Rixner, J. J., Kraemer, S. R. and Smith, A. D., 1986, ‘Prefabrieated vertical drains’,
Engineering Guideline, FHWA/RD-86/168, Federal Highway Administration,
Virginia.
Robinson, R. G. 1999, ‘Consolidation analysis with pore water pressure
measurements’, Géotechnique, 49(1), pp. 127-132.
Roscoe, K. H. and Burland, J. B. 1968, ‘On the generalised stress-strain behaviour of
‘wet clay’, Engineering Plasticity’, Heyman and Leckie (ed), pp. 535-609.
Rowe, R. K. and Li, A. K. 2002, ‘Behaviour of reinforced embankments on soft rate
sensitive soils’, Geotechnique, 52(1), pp. 859-869.
RTA, 2000, ‘Trial embnakment construction report’, Roads and Traffic Authorithy of
NSW, Pacific highway office, Grafton.
Rujikiatkamjorn, C., Ghandeharioon, A. and Indraratna, B. 2009, ‘Finite element
simulation of mandrel penetration in a normally consolidated soil’, Geotechnics of
Soft Soils – Focus on Ground Improvement –, Leoni, K., Taylor & Francis Group,
London, 287-292.
Rujikiatkamjorn, C. and Indraratna, B. 2009, ‘Design procedure for vertical drains
considering a linear variation of lateral permeability within the smear zone’,
Canadian Geotechnical Journal, 46(3), pp. 270-280.
Rujikiatkamjorn, C. and Indraratna, B. 2010, ‘Radial consolidation modelling
incorporating the effect of a smear zone for a multilayer soil with downdrag caused
by mandrel action’, Canadian Geotechnical Journal, 47(9), pp. 1024-1035.

247
Rujikiatkamjorn, C., Indraratna, B. and Chu, J. 2008, ‘2D and 3D numerical
modeling of combined surcharge and vacuum preloading with vertical drains’,
International Journal of Geomechanics, 8(2), pp. 144-156.
Saowapakpiboon, J., Bergado, D. T., Voottipruex, P., Lam, L. G. and Nakakuma, K.
2011, ‘PVD improvement combined with surcharge and vacuum preloading
including simulations’, Geotextiles and Geomembranes, 29(1), pp. 74-82.
Saowapakpiboon, J., Bergado, D. T., Youwai, S. J., Chai, C., Wanthong, P. and
Voottipruexe, P. 2009, ‘Measured and predicted performance of prefabricated
vertical drains (PVDs) with and without vacuum preloading’, Geotextiles and
Geomembranes, 28(1), pp. 1-11.
Saowapakpiboon, J., Bergado, D. T., Youwai, S., Chai, J. C., Wanthong, P. and
Voottipruex, P. 2010, ‘Measured and predicted performance of prefabricated vertical
drains (PVDs) with and without vacuum preloading’, Geotextiles and
Geomembranes, 28(1), pp. 1-11.
Sathananthan, I. 2005, ‘Modelling of vertical drains with smear installed in soft
clay’, Doctor of Philosophy. thesis, School of Civil, Mining and Environmental
Engineering, The University of Wollongong, Wollongong, NSW.
Sathananthan, I. and Indraratna, B. 2006, ‘Laboratory evaluation of smear zone and
correlation between permeability and moisture content’, Journal of Geotechnical and
Geoenvironmental Engineering, 132(7), pp. 942-945.
Sathananthan, I., Indraratna, B. and Rujikiatkamjorn, C. 2008, ‘Evaluation of smear
zone extent surrounding mandrel driven vertical drains using the cavity expansion
theory’, International Journal of Geomechanics, 8(6), pp. 355-365.
Schiffman, R. L., Chen, A. T.-F. and Jordan, J. C. 1969, ‘An analysis of
consolidation theories’, Journal of the soil mechincs and fondation division,
95(SM1), pp. 285-312.
Scott, R. F. 1963, ‘Principles of soil mechanics’, Addison-Wesley, Massachusetts.
Seah, T. H. 2006, ‘Design and construction of ground improvement works at
Suvarnabhumi Airport’, Geot. Eng, J. of Southeast Asian Geot. Society, 37, pp. 171-
188.
Sharma, J. and Xiao, D. 2000, ‘Determination of smear zone around vertical Drains’,
Proceedings of the Soft Ground Technology Noordwijkerhout, the Netherlands, GSP,
pp. 270-279.
Sheng, D., Eigenbrod, K. D. and Wriggers, P. 2005, ‘Finite element analysis of pile
installation using large-slip frictional contact’, Computers and Geotechnics, 32(1),
pp. 17-26.
Shin, D. H., Lee, C., Lee, J. S. and Lee, W. 2009, ‘Detection of smear zone using
micro-cone and electrical resistance probe’, Canadian Geotechnical Journal, 46(6),
pp. 719-726.

248
Sivaram, B. and Swamee, P. 1977, ‘A computational method for consolidation
coefficient’, Soils and Foundations, 17(2), pp. 48-52.
Skempton, A. W. 1954, ‘The pore-pressure coefficients A and B’, Géotechnique,
4(4), pp. 143-147.
Small, J.C., Booker, J. R. and Davis, E. H. 1976, ‘Elasto-Platic consolidation of soil’,
International Journal of Solids and Strucrures, 12, pp. 431-448.
Songa, Y.S. and Kim, T. H. 2004, ‘Improvement of estuarine marine clays for
coastal reclamation using vacuum-applied consolidation method’, Ocean
Engineering, 31(16), pp. 1999-2010.
Sonpal, R. C. and Katti, R. K. 1973, ‘Consolidation analysis with pore pressure
measurements’, 8th Int. Conf. on Soil Mechanics and Foundation Eng., Moscow, pp.
385-388.
Spaulding and Porbaha, A. 2004, ‘Embankment construction on marshland using
vacuum consolidation technology’, Geotechnical Engineering for Transportation
Projects (GSP 126), ASCE, Los Angeles, CA, pp. 1943-1950.
Stapelfeldt, T., Vepsäläinen, P. and Yin, Z. Y. 2008, ‘Numerical modelling of a test
embankment on soft clay improved with vertical drains’, In Proceedings of the
Second International Workshop on Geotechnics of Soft Soils, Glasgow, Scotland, 3-5
September 2008, Taylor & Francis Group, London, pp. 173-179.
Tang, X. W. and Onitsuka, K. 2000, ‘Consolidation by vertical drains under time-
dependent loading’, International journal for numerical and analytical methods in
geomechanics, 24(9), pp. 739-751.
Tavenas, F., Tremblay, M., Larouche, G. and Leroueil, S. 1986, ‘In situ measurement
of permeability in soft clays’, ASCE Spec. Conf. on Use of in Situ Tests in Geotech.
Engrg., ASCE, New York, pp. 1034–1048.
Teparaksa, W. and Trung Ngo, D. 2012, ‘Performance of vacuum consolidation on
very soft clay at Nakorn Sri Thammarat Airport’, International Conference on
Ground Improvement, Wollongong, Australia.
Terzaghi, K. 1925, ‘Erdbaumechanik auf bodenphysikalischer Grundlage’, Deuticke,
Vienna, Austria.
Terzaghi, K. T. and Frohlich, O. K. 1936, ‘Theorie der Setzung von Tonschichten’,
Leipzig, pp. 166.
Tran-Nguyen, H.H. and Edil, T.B. 2011, ‘The characteristics of PVD smear zone’, In
Proceedings of the Geo-Frontiers 2011: Advances in Geotechnical Engineering, 13–
16 March 2011, ASCE, New York, pp. 748-757.
Tran, T. A. and Mitachi, T. 2008, ‘Equivalent plane strain modeling of vertical drains
in soft ground under embankment combined with vacuum preloading’, Computers
and Geotechnics, 35(5), pp. 655-672.

249
Vaziri, H. H. and Christian, H. A. 1994, ‘Application of Terzaghi's consolidation
theory to nearly saturated soils’, Canadian Geotechnical Journal, 31(2), pp. 311-317.
Walker, J. and Yu, H. S. 2006, ‘Adaptive finite element analysis of cone penetration
in clay’, Acta Geotechnica, 1, pp. 43-57.
Walker, R. and Indraratna, B. 2006, ‘Vertical drain consolidation with parabolic
distribution of permeability in smear zone’, Journal of Geotechnical and
Geoenvironmental Engineering, 132(7), pp. 937-941.
Walker, R. and Indraratna, B. 2007, ‘Vertical drain consolidation with overlapping
smear zones’, Géotechnique, 57(5), pp. 463-467.
Walker, R. and Indraratna, B. 2009a, ‘Consolidation analysis of a stratified soil with
vertical and horizontal drainage using the spectral method’, Géotechnique, 59(5), pp.
439-449.
Walker, R., Indraratna, B. and Sivakugan, N. 2009b, ‘Vertical and radial
consolidation analysis of multilayered soil using the spectral method’, Journal of
Geotechnical and Geoenvironmental Engineering, 135(5), pp. 657-663.
Wang, T. R. and Chen, W. H. 1996, ‘Development in application of prefabricated
drains in treatment of soft soils’, In Proceedings of the Third Symposium on Weak
Ground Improvement using PVD, China, pp. 13–40.
Wang, X. S. and Jiao, J. J. 2004, ‘Analysis of soil consolidation by vertical drains
with double porosity model’, Int. J. Numer. Anal. Meth. Geomech., 28, pp. 1385–
1400.
Welker, A. L., and K. M. Herdin. 2003, ‘Evaluation of Four Equivalent Diameter
Formulations for Prefabricated Vertical Drains Using Flow Rates’, Geosynthetics
International, 10(3), pp. 103-109.
Whitman, R. V., Richardson, A. M. and Healy, K. A. 1961, ‘Time-lags in pore
pressure measurements’, In Proceedings of the 5th ICSMFE.
Wood, D. M. 1990, ‘Soil behaviour and critical state soil mechanics’, Cambridge
University Press, Cambridge, UK.
Xie, K. H. 1987, ‘Consolidation theories and optimisation design for vertical drains’,
Doctor of Philosophy. thesis, The Zhejiang University, China.
Xie, K. H., Xie, X. Y. and Gao, X. 1999, ‘Theory of one dimensional consolidation
of two-layered soil with partially drained boundaries’, Computers and Geotechnics,
24(4), pp. 265-278.
Xie, K. H., Xie, X. Y. and Jiang, W. 2002, ‘A study on one-dimensional nonlinear
consolidation of double-layered soil’, Computers and Geotechnics, 29(2), pp. 151-
168.
Xie, K.H. and Leo, C.J. 2004, ‘Analytical solutions of one-dimensional large strain
consolidation of saturated and homogeneous clays’, Computers and Geotechnics,
31(4), pp. 301-314.

250
Yan, H.S. and Cao, D.Z. 2005, ‘Application of low-level vacuum preloading
technique in offshore projects’, Ocean and River Hydraulics, 3, pp. 41-43.
Yan, S.-W. and Chu, J. 2005, ‘Soil improvement for a storage yard using the
combined vacuum and fill preloading method’, Can. Geotec. J., 42(4), pp. 1094-
1104.
Yildiz, A. and Karstunen, M. 2009, ‘Numerical modeling of vertical drains with
advanced constitutive models’, Computers and Geotechnics, 36(6), pp. 1072-1083.
Yildiz, A., M., Karstunen and H., Krenn 2006, ‘Numerical modelling of vertical
drains with advanced constitutive models’, In Proceedings of the The sixth European
Conf. on Num. Meth. in Geotech. Engng., Graz, Austria.
Yildiz, Abdulazim 2009, ‘Numerical analyses of embankments on PVD improved
soft clays’, Advances in Engineering Software, 40(10), pp. 1047-1055.
Yin, J.-H. and Graham, J. 1989, ‘Viscous-elastic-plastic modelling of one-
dimensional time-dependent behaviour’, Canadian Geotechinical Journal, 26(2), pp.
199-209.
Yoshikuni, H. and Nakanodo, H. 1974, ‘Consolidation of fine-grained soils by drain
wells with finite permeability’, Soils and Foundations, 14(2), pp. 35-46.
Yu, H. 2000, ‘Cavity expansion methods in geomechanics’, Kluwer academic
Publishers.
Zeng, G. X. and Xie, K. h. 1989, ‘New development of the vertical drain theories’, In
Proceedings of the 12th International Conference on Soil Mechanics and Foundation
Engineering, Rio de Janeiro, Brazil, pp. 1435-1438.
Zhu, G. and Yin, J. H. 1998, ‘Consolidation of soil under depth-dependent ramp
load’, Canadian Geotechnical Journal, 35(2), pp. 344-350.
Zhu, G. and Yin, J. H. 2001, ‘Consolidation of soil with vertil and horizontal
drainage under ramp load’, Géotechnique, 51(4), pp. 361-367.
Zhu, X. R., Xie, K. H., Pan, Q.Y. 1993, ‘Soft clay ground improvement of Ningbo
international airport’, Third international conference on case histories in
geotechnical engineering, St. Louis, Missouri, pp. 983-987.
Zienkiewicz, O. C., Taylor, R. L. and Nithiarasu, P. 2005, ‘The Finite element
method for fluid dynamics’, Elsevier.

251
 Appendix A

DEVELOPED FLAC CODE TO SIMULATE

TRIAL EMBANKMENT

252
A-1 Geometry and Materials properties

;***Function to Enter Input Data***

;***Geometries, PVD properties, and required ratios***
def input1

;***Embankment & PVD dimension***

h0=25.0; depth of the soil profile
h1=24.0; length of the PVD
h2=h0-h1
a1=60; the total length of the trial embankment
a2=60.0; length of the ground at sides of the embankment
a0=a1+a2
rm=0.045; radius of mandrel or drain
smr=3.0; extend ratio (rs/rm)
sdrain=1.0; drain spacing
krat=3.0; permeability ratio (kh/ks)
khv=1.0 the ratio of the horizontal permeability over the vertical one (kh/kv)
gn1=1.0; the number of mesh zones inside the smear zone
glrat1=3; length of the mesh outside the smear zone over the inside one
glrat2=3; ratio for generating meshes and grids
gnv1=(h1/0.5); the length of the mesh zones in the vertical direction
glratv=1.0; ratio for generating meshes and grids

;***Depths of soil layer interfaces***

ly1=0.5; depth of layer 1 from the ground surface
ly2=4.0
ly3=10.0
ly4=15.0
ly5=24.0
ly6=25.0

;***Embank construction history***

stno=12; number of construction stages

;***SOIL PROPERTIES Layer 01***

kix01=1.0e-131.0e-13; horizontal mobility coefficient of the intact zone (kh)
kiy01=kix01/khv; vertical mobility coefficient of the intact zone (kv)
sden01=698; dray density of the soil (ρd)
spois01=0.27; Poisson’s ratio (μ)
sbulk01=5.0e7;5.0e6; soil bulk modulus (K)
smm01=1.2; frictional constant (M)
slanda01=0.569; slope of the normal consolidation line (λ)
skapa01=0.057; slope of the swelling line (к)
smpc01=5.51e3; tpreconsolidation pressure (pc)
smp101=1e2;1.76e4; reference pressure (p1)
smv101=5.93;2.55; specific volume at reference pressure (vλ)
sshear01=1.2e6; shear modulus

253
 spor01=0.73 porosity of soil (n)
;***SOIL PROPERTIES Layer 02***
kix02=6.0e-14
kiy02=kix02/khv
sden02=698
spois02=0.27
sbulk02=5.0e7;5.0e6
smm02=1.2
slanda02=0.569
skapa02=0.057
smpc02=3.33e4
smp102=1e2;1.76e4
smv102=6.97;2.55
;smv002=2.8
sshear02=1.2e6
spor02=0.73
;***SOIL PROPERTIES Layer 03***
kix03=3.0e-14
kiy03=kix02/khv
sden03=674
spois03=0.27
sbulk03=5.0e7;5.0e6
smm03=1.2
slanda03=0.672
skapa03=0.0672
smpc03=7.83e4
smp103=1e2;1.76e4
smv103=8.29;2.55
;smv003=2.8
sshear03=1.2e6
spor03=0.74
;***SOIL PROPERTIES Layer 04***
kix04=5.00e-15
kiy04=kix04/khv
sden04=720
spois04=0.3
sbulk04=5.0e7;5.0e6
smm04=1.18
slanda04=0.47
skapa04=0.047
smpc04=1.1e5
smp104=1e2;1.76e4
smv104=6.87;2.55
;smv004=2.8
sshear04=7.8e6
spor04=0.72
;***SOIL PROPERTIES Layer 05***
kix05=2.00e-15
kiy05=kix05/khv

254
 sden05=845
spois05=0.3
sbulk05=5.0e7;5.0e6
smm05=1.18
slanda05=0.40
skapa05=0.040
smpc05=1.59e5
smp105=1e2;1.63e5
smv105=6.04;2.55
;smv005=2.8
sshear05=7.8e6
spor05=0.68
;***SOIL PROPERTIES Layer 06***
kix06=1.00e-15
kiy06=kix06/khv
sden06=923
spois06=0.3
sbulk06=5.0e7;5.0e6
smm06=1.18
slanda06=0.368
skapa06=0.0368
smpc06=1.69e5
smp106=1e2;1.76e4
smv106=5.54;2.55
;smv006=2.8
sshear06=7.8e6
spor06=0.65
;***PERMEABILITY (SMEAR & INTACT ZONES)***
ks01=kix01/krat; smear zone permeability of the layer1 (ks1)
ks02=kix02/krat
ks03=kix03/krat
ks04=kix04/krat
ks05=kix05/krat
ks06=kix06/krat
;***WATER PROPERTIES***
wden=1000.0; density of water (ρw)
wbu=8.5e7;2.0e8; bulk modulus of water (Kw)

255
A-2 GRIDS AND MESH GENERATION

;***Horizontal Griding Calculations***

ls=rm*smr
li=(sdrain/2)-ls
ndrain0=a1/sdrain
delt=ndrain0-int(ndrain0)
if delt<0.5
ndrain=int(ndrain0)
else
ndrain=int(ndrain0)+1
end_if
a4=ndrain*sdrain
gl1=ls/gn1
gl20=glrat1*gl1
gn20=li/gl20
df1=gn20-int(gn20)
if df1<0.5
gn2=int(gn20)
else
gn2=int(gn20)+1
end_if
gl2=li/gn2
a5=sdrain/2
a6=a0-(ndrain+0.5)*sdrain
gl30=glrat2*gl2
gn30=a6/gl30
df2=gn30-int(gn30)
if df2<0.5
gn3=int(gn30)
else
gn3=int(gn30)+1
end_if
gl3=a6/gn3
gnt=(2*ndrain+1)*(gn1+gn2)
gnemb=gnt+1
znh=gn3+gnt;+gn4

;***Vertical Griding Calculations***

glv1=h1/gnv1
glv20=glratv*glv1
gnv20=h2/glv20
dfv1=gnv20-int(gnv20)
if dfv1<=0.5
gnv2=int(gnv20)
else
gnv2=1+int(gnv20)
end_if

256
 if gnv20=0
gnv2=0
glv2=0
else
if gnv20<1
gnv2=1
glv2=h2/gnv2
else
glv2=h2/gnv2
end_if
end_if
h00=glv1*gnv1+glv2*gnv2
znv=gnv1+gnv2

;*****embankment grids******
meshx1=znh+2
meshx2=znh+2*ndrain+2
meshy1=znv+2
meshy01=meshy1+1
meshy2=znv+stno+1
meshy02=meshy2-1
iembmax=meshx2+1
iemb01=imax+1
jmeshmax=meshy2+1
meshnullx1=znh+1
meshnullx2=meshx2-5
meshnullx3=meshx2-5-1
meshnully1=znv+1
meshnully2=meshy2-5
jemb01=meshy1+1
jemb02=jemb01+1
emb03=jemb02+1
jemb04=jemb03+1
jemb05=jemb04+1
jemb06=jemb05+1
jemb07=jemb06+1
jemb08=jemb07+1
jemb09=jemb08+1
jemb10=jemb09+1
jemb11=jemb10+1
jemb12=jemb11+1
jemb13=jemb12+1

;***Required Gird Lines***

i1=1.0
i2=i1+gn1+gn2
idrain=gnt+1
imax=znh+1
j1=1.0

257
 j2=gnv2+1
jmax=znv+1
end
;*** End of "input1" Function
input1
;***Mesh Generation Process***
config gw extra 3
grid meshx2 meshy2
mod cam i 1 znh j 1 znv
mod elas i meshx1 meshx2 j meshy1 meshy2
;mod null i meshnullx1
;mod null i meshnully1
;mod null i meshx1 meshx2 j 1 jmax
ini x=0.0 i=1
ini y=0.0 j=1
;***FUNCTION "hzone": Mesh Generating in Horizontal Direction***
def hzone
k1=2
xi=0.0
is1=1+gn1
loop while k1<=i2
loop while k1<=is1
xi=xi+gl1
command
ini x=xi i=k1
endcommand
k1=k1+1
endloop
xi=xi+gl2
command
ini x=xi i=k1
endcommand
k1=k1+1
endloop
k2=i2+gn2
step1=1
loop while step1<=ndrain
loop while k1<=k2
xi=xi+gl2
command
ini x=xi i=k1
endcommand
k1=k1+1
if xi<=hisxpp1
ipp1=k1
xpp1=xi+gl1
endif
if xi<=a1
iload=k1

258
 xload=xi+gl1
endif
endloop
k3=k2+gn1
loop while k1<=k3
xi=xi+gl1
command
ini x=xi i=k1
endcommand
k1=k1+1
if xi<=hisxpp1
ipp1=k1
xpp1=xi+gl1
endif
if xi<=a1
iload=k1
xload=xi+gl1
endif
endloop
k4=k3+gn1
loop while k1<=k4
xi=xi+gl1
command
ini x=xi i=k1
endcommand
k1=k1+1
if xi<=hisxpp1
ipp1=k1
xpp1=xi+gl1
endif
if xi<=a1
iload=k1
xload=xi+gl2
endif
endloop
k5=k4+gn2
loop while k1<=k5
xi=xi+gl2
command
ini x=xi i=k1
endcommand
k1=k1+1
if xi<=hisxpp1
ipp1=k1
xpp1=xi+gl1
endif
if xi<=a1
iload=k1
xload=xi+gl2

259
 endif
endloop
k2=k5+gn2
step1=step1+1
endloop
k6=imax
loop while k1<=k6
xi=xi+gl3
command
ini x=xi i=k1
endcommand
k1=k1+1
endloop
xmax=xi
end
;***End of "hzone" Function***
Hzone

;***FUNCTION "vzone": Mesh Generating in Vertical Direction***

def vzone
s1=2.0
loop while s1<=j2
y1=y1+glv2
command
ini y=y1 j=s1
endcommand
s1=s1+1
endloop
loop while s1<=jmax
y1=y1+glv1
command
ini y=y1 j=s1
endcommand
s1=s1+1
endloop
ymax=y1
end
;*** End of "vzone" Function***
vzone
;***Embankment Mesh Generating***
def embankmesh
iem2=imax+2
iemax=meshx2+1
lem1=sdrain/2
lem2=lem1-1
command
ini x add lem2 i iem2 iemax
endcommand
xmeshno=(26/(sdrain/2))

260
 ixnull=iemax-xmeshno
ixgroup=ixnull-1
iemb01=imax+1
iemb02=iemax-6;meshnullx2
iembmax=meshx2+1
jemb0=jemb01-1
meshnullx4=iemb02-1;meshx2-7
embankl=sdrain*(ndrain+0.5)
embank2=embankl-8.5
command
gen 0,25 0,33.5 embankl,33.5 embankl,25 i iemb01 iemax j jemb0 jemb12
gen same 0,27.3 55.4,27.3 same i iemb01 iemax j jemb0 jemb05
gen same 0,33.5 20,33.5 same i iemb01 ixnull j jemb05 jemb12
mod null i ixnull iemax j jemb05 jemb12
ini y add 0.1 j=jemb01
ini y add 0.1 j=jemb02
ini y add 0.2 j=jemb03
ini y add 0.2 j=jemb04
endcommand
end
;*** End of Function***
Embankmesh

261
A-3 LAYERING AND ASSIGNING SOIL PROPERTIES

; ***Required Grid Points for Layering Process***

def gridpoints
;******* group i & j*****
yg1=0
yg2=0
nn2=0
loop while yg1<ly1
yg1=glv1*nn2
nn2=nn2+1
endloop
yg2=glv1*(nn2-2)
delta01=yg1-ly1
delta02=ly1-yg2
if delta01>delta02
jg1=jmax-nn2+2
else
jg1=jmax-nn2+1
endif
yg1=0
yg2=0
nn2=0
loop while yg1<ly2
yg1=glv1*nn2
nn2=nn2+1
endloop
yg2=glv1*(nn2-2)
delta01=yg1-ly2
delta02=ly2-yg2
if delta01>delta02
jg2=jmax-nn2+2
else
jg2=jmax-nn2+1
endif
yg1=0
yg2=0
nn2=0
loop while yg1<ly3
yg1=glv1*nn2
nn2=nn2+1
endloop
yg2=glv1*(nn2-2)
delta01=yg1-ly3
delta02=ly3-yg2
if delta01>delta02
jg3=jmax-nn2+2
else

262
 jg3=jmax-nn2+1
endif
yg1=0
yg2=0
nn2=0
loop while yg1<ly4
yg1=glv1*nn2
nn2=nn2+1
endloop
yg2=glv1*(nn2-2)
delta01=yg1-ly4
delta02=ly4-yg2
if delta01>delta02
jg4=jmax-nn2+2
else
jg4=jmax-nn2+1
endif
yg1=0
yg2=0
nn2=0
loop while yg1<ly5
yg1=glv1*nn2
nn2=nn2+1
endloop
yg2=glv1*(nn2-2)
delta01=yg1-ly5
delta02=ly5-yg2
if delta01>delta02
jg5=jmax-nn2+2
else
jg5=jmax-nn2+1
endif
yg1=0
yg2=0
nn2=0
loop while yg1<ly6
yg1=glv1*nn2
nn2=nn2+1
endloop
yg2=glv1*(nn2-2)
delta01=yg1-ly6
delta02=ly6-yg2
if delta01>delta02
jg6=jmax-nn2+2
else
jg6=jmax-nn2+1
endif
ly1check=(jmax-jg1)*glv1
ly2check=(jmax-jg2)*glv1;h0-(gnv2*glv2)-(grj2-gnv2-1)*glv1

263
 ly3check=(jmax-jg3)*glv1;h0-(gnv2*glv2)-(grj3-gnv2-1)*glv1
gri1=1
gri2=imax-1
grj0=jmax-1
grj1=jg1-1
grj2=jg2-1
grj3=jg3-1
grj4=jg4-1
grj5=jg5-1
grj6=1
gri3=gnt+1
end
; ***End of "Grid point for history" Function***
gridpoints
; ***Layering Process***
group 'layer01' i 1 znh j grj1 grj0
group 'layer02' i 1 znh j grj2 grj1
group 'layer03' i 1 znh j grj3 grj2
group 'layer04' i 1 znh j grj4 grj3
group 'layer05' i 1 znh j grj5 grj4
group 'layer06' i 1 znh j grj6 grj5
group 'stage01' i meshx1 meshx2 j meshy1
group 'stage02' i meshx1 meshx2 j jemb01
group 'stage03' i meshx1 meshx2 j jemb02
group 'stage04' i meshx1 meshx2 j jemb03
group 'stage05' i meshx1 meshx2 j jemb04
group 'stage06' i meshx1 ixgroup j jemb05
group 'stage07' i meshx1 ixgroup j jemb06
group 'stage08' i meshx1 ixgroup j jemb07
group 'stage09' i meshx1 ixgroup j jemb08
group 'stage10' i meshx1 ixgroup j jemb09
group 'stage11' i meshx1 ixgroup j jemb10
group 'stage12' i meshx1 ixgroup j jemb11
attach aside from 1,jmax to idrain,jmax bside from iemb01,jemb0 to
iemax,jemb0
mod null group 'stage01'
mod null group 'stage02'
mod null group 'stage03'
mod null group 'stage04'
mod null group 'stage05'
mod null group 'stage06'
mod null group 'stage07'
mod null group 'stage08'
mod null group 'stage09'
mod null group 'stage10'
mod null group 'stage11'
mod null group 'stage12'
prop dens sden01 poiss spois01 bu sbulk01 mm smm01 lambda slanda01
kappa skapa01 group 'layer01'

264
prop shear_mod sshear01 mpc smpc01 mp1 smp101 mv_l smv101 group
'layer01'
prop k11 kix01 k22 kiy01 por spor01 group 'layer01'
prop dens sden02 poiss spois02 bu sbulk02 mm smm02 lambda slanda02
kappa skapa02 group 'layer02'
prop shear_mod sshear02 mpc smpc02 mp1 smp102 mv_l smv102 group
'layer02'
prop k11 kix02 k22 kiy02 por spor02 group 'layer02'
prop dens sden03 poiss spois03 bu sbulk03 mm smm03 lambda slanda03
kappa skapa03 group 'layer03'
prop shear_mod sshear03 mpc smpc03 mp1 smp103 mv_l smv103 group
'layer03'
prop k11 kix03 k22 kiy03 por spor03 group 'layer03'
prop dens sden04 poiss spois04 bu sbulk04 mm smm04 lambda slanda04
kappa skapa04 group 'layer04'
prop shear_mod sshear04 mpc smpc04 mp1 smp104 mv_l smv104 group
'layer04'
prop k11 kix04 k22 kiy04 por spor04 group 'layer04'
prop dens sden05 poiss spois05 bu sbulk05 mm smm05 lambda slanda05
kappa skapa05 group 'layer05'
prop shear_mod sshear05 mpc smpc05 mp1 smp105 mv_l smv105 group
'layer05'
prop k11 kix05 k22 kiy05 por spor05 group 'layer05'
prop dens sden06 poiss spois06 bu sbulk06 mm smm06 lambda slanda06
kappa skapa06 group 'layer06'
prop shear_mod sshear06 mpc smpc06 mp1 smp105 mv_l smv106 group
'layer06'
prop k11 kix06 k22 kiy06 por spor06 group 'layer06'
water bulk wbu tens wt den wden

265
A-4 Defining the location of instrumentations and transducers

history 1 gwtime
history 2 unbalanced
history 3 sratio
history 4 ydisp i=1, j=51; “ydisp” stands for vertical displacement
history 5 ydisp i=1, j=49
history 6 ydisp i=1, j=47
history 7 ydisp i=1, j=45
history 8 ydisp i=3, j=51
history 9 ydisp i=3, j=49
history 10 ydisp i=3, j=47
history 11 ydisp i=3, j=45
history 12 gpp i=3, j=50; “gpp” stands for pore pressure at gridpoint
history 13 gpp i=3, j=49
history 14 gpp i=3, j=48
history 15 gpp i=3, j=47
history 16 gpp i=3, j=46
history 17 gpp i=7, j=50
history 18 gpp i=7, j=49
history 19 gpp i=7, j=48
history 20 gpp i=7, j=47
history 21 gpp i=7, j=46

266
A-5 BOUNDARY CONDITIONS, INITIAL STRESSES, and
UNDRAINED ANALYSIS

;***Boundary Conditions***
fix x i=1
fix x i=imax
fix y j=1
initial pp 0 j jmax
initial sat 1 j jmax
initial pp 0 j jemb0
initial sat 1 j jemb0
fix pp j jmax
fix s j jmax
fix pp j jemb0
fix s j jemb0
;***Initial Conditions***
call ininv.fis
set grav 9.8; earth gravity
set wth=25 k0x=0.66 k0z=0.66; the water depth and the coefficients of lateral
pressure
ininv
;***Settings***
set flow off
;***undrained response***
solve sratio 1e-3

267
A-6 SIMULATION OF PRECONSOLIDATION STAGE

;***Simulation of Preconsolidation Process – without PVD***

;***Settings***
set sratio 1.0
set step 2000000
set nmech 1
;***Stage01***
model mohr group 'stage01'
prop sh 1.e8 bu 2.e8 coh 1.e3 fri 35 dens 2100 group 'stage01'
fix x i=meshx1
set flow off
solve
set flow on
set nmech 20
solve auto on age 4.32e5
;***Stage02***
model mohr group 'stage02'
prop sh 1.e8 bu 2.e8 coh 1.e3 fri 35 dens 2100 group 'stage02'
set flow off
set nmech 1
solve
set flow on
set nmech 20
solve auto on age 8.64e5
;***Stage03***
model mohr group 'stage03'
prop sh 1.e8 bu 2.e8 coh 1.e3 fri 35 dens 2100 group 'stage03'
set flow off
set nmech 1
solve
set flow on
set nmech 20
solve auto on age 3.71e6

268
A-7 SIMULATION OF PVDs SMEAR ZONE

;***FUNCTION "pdrain": Simulation of PVDs and Smear Zone***

def pdrain01
is01=1+gn1
ppd1=h1*9800
ppd2=(-1)*ppd1
command
initial pp ppd1 var 0 ppd2 i=1 j=j2,jmax
initial s=1 from 1,j2 to 1,jmax
fix pp i=1 j=j2,jmax
fix sat i=1 j=j2,jmax
prop perm ks01 i=1,is01 j=grj1,grj0;group 'layer01'
prop perm ks02 i=1,is01 j=grj2,grj1;group 'layer02'
prop perm ks03 i=1,is01 j=grj3,grj2;group 'layer03'
prop perm ks04 i=1,is01 j=grj4,grj3 group 'layer04'
prop perm ks05 i=1,is01 j=grj5,grj4 group 'layer05'
mark i 1 j j2,jmax
endcommand
end
def pdrain
idr=(gn1+gn2)*2+1
is1=idr-gn1
is2=idr+gn1
step2=1
loop while step2<=ndrain
command
initial pp ppd1 var 0 ppd2 i=idr j=j2,jmax
initial s=1 i=idr j=j2,jmax
fix pp i=idr j=j2,jmax
fix sat i=idr j=j2,jmax
prop perm ks01 i=is1,is2 j=grj1,grj0;group 'layer01'
prop perm ks02 i=is1,is2 j=grj2,grj1;group 'layer02'
prop perm ks03 i=is1,is2 j=grj3,grj2;group 'layer03'
prop perm ks04 i=is1,is2 j=grj4,grj3;group 'layer04'
prop perm ks05 i=is1,is2 j=grj5,grj4 group 'layer05'
mark i idr j j2,jmax
endcommand
step2=step2+1
idr=idr+2*(gn1+gn2)
is1=idr-gn1
is2=idr+gn1
endloop
end
;***End of "pdrain" Function***
pdrain01
pdrain

269
A-8 SIMULATION OF VACUUM PRESSURE

;***Function to Simulate Variation of Vacuum Pressure Alomg the Vertical

Drain***
def vacuum01
is01=1+gn1
command
initial pp -70000 i=1 j=j2,jmax
initial s=1 from 1,j2 to 1,jmax
fix pp i=1 j=j2,jmax
fix sat i=1 j=j2,jmax
mark i 1 j j2,jmax
endcommand
end
def vacuum02
idr=(gn1+gn2)*2+1
is1=idr-gn1
is2=idr+gn1
step2=1
loop while step2<=ndrain
command
initial pp -70000 i=idr j=j2,jmax
initial s=1 i=idr j=j2,jmax
fix pp i=idr j=j2,jmax
fix sat i=idr j=j2,jmax
mark i idr j j2,jmax
endcommand
step2=step2+1
idr=idr+2*(gn1+gn2)
is1=idr-gn1
is2=idr+gn1
endloop
end
vacuum01
vacuum02

270
A-9 CONSTRUTION OF TRIAL EMBANKMENT AND
CONSOLIDATION PROCESS

;***Stage04***
model mohr group 'stage04'
prop sh 1.e8 bu 2.e8 coh 1.e3 fri 35 dens 2100 group 'stage04'; tens 1e10
fix x i=meshx1
set flow off
set nmech 1
solve
set flow on
set nmech 20
solve auto on age 7.77e6
;***Stage05***
model mohr group 'stage05'
prop sh 1.e8 bu 2.e8 coh 1.e3 fri 35 dens 2100 group 'stage05'; tens 1e10
fix x i=meshx1
set flow off
set nmech 1
solve
set flow on
set nmech 20
solve auto on age 9.85e6
;***Stage05***
set sratio 1.9
model mohr group 'stage06'
prop sh 1.e8 bu 2.e8 coh 1.e3 fri 35 dens 2100 group 'stage06'; tens 1e10
set flow off
set nmech 1
solve
set flow on
set nmech 20
initial pp 0 j jemb06
initial sat 0 j jemb06
fix pp j jemb06
fix sat j jemb06
solve auto on age 1.28e7
model mohr group 'stage07'
prop sh 1.e8 bu 2.e8 coh 1.e3 fri 35 dens 2100 group 'stage07'; tens 1e10
set flow off
set nmech 1
solve
set flow on
set nmech 20
initial pp 0 j jemb07
initial sat 0 j jemb07
fix pp j jemb07
fix sat j jemb07

271
 solve auto on age 1.43e7
model mohr group 'stage08'
prop sh 1.e8 bu 2.e8 coh 1.e3 fri 35 dens 2100 group 'stage08'; tens 1e10
set flow off
set nmech 1
solve
set flow on
set nmech 20
initial pp 0 j jemb08
initial sat 0 j jemb08
fix pp j jemb08
fix sat j jemb08
solve auto on age 1.56e7
model mohr group 'stage09'
prop sh 1.e8 bu 2.e8 coh 1.e3 fri 35 dens 2100 group 'stage09'; tens 1e10
set flow off
set nmech 1
solve
set flow on
set nmech 20
initial pp 0 j jemb09
initial sat 0 j jemb09
fix pp j jemb09
fix sat j jemb09
solve auto on age 1.68e7;***195days***
model mohr group 'stage10'
initial pp 0 j jemb10
initial sat 0 j jemb10
fix pp j jemb10
fix sat j jemb10
prop sh 1.e8 bu 2.e8 coh 1.e3 fri 35 dens 2100 group 'stage10'; tens 1e10
set flow off
set nmech 1
solve
set flow on
set nmech 20
solve auto on age 1.81e7
model mohr group 'stage11'
initial pp 0 j jemb11
initial sat 0 j jemb11
fix pp j jemb11
fix sat j jemb11
prop sh 1.e8 bu 2.e8 coh 1.e3 fri 35 dens 2100 group 'stage11'; tens 1e10
set flow off
set nmech 1
solve
set flow on
set nmech 20
solve auto on age 1.94e7

272
 model mohr group 'stage12'
initial pp 0 j jemb12
initial sat 0 j jemb12
fix pp j jemb12
fix sat j jemb12
prop sh 1.e8 bu 2.e8 coh 1.e3 fri 35 dens 2100 group 'stage12'; tens 1e10
set flow off
set nmech 1
solve
set flow on
set nmech 20
solve auto on age 2.07e7
set nmech 20
solve auto on age 3.37e7

273
 Appendix B

DEVELOPED FLAC CODE TO SIMULATE

ROWE CELL TEST

274
;***Input Data Function for Cell & PVD Dimensions ***
def indata
hc=0.1213;h0
hd=0.1213;h1
h2=hc-hd
rc=0.125
rm=0.011
smr=3.0
krat=2.5
gns=8.0
lrat=1.5
gnv1=20.0
glratv=1.0
;***Permeability (Smear and & Intact Zones)***
ki=1.0e-12
ks=ki/krat
;***Soil Properties***
sden=935.0
spois=0.3
sbulk=2.0e7
smm=1.1
slanda=0.35
skapa=0.045
smpc=2.1e4
smp1=1.0e1
smv1=4.78
spor=0.57
;***Propertied of Water***
wden=1000.0
wbu=5.0e6;4.5e7
wt=1.0e10
;***Lateral Pressure Coefficients***
kx0=0.7
kz0=0.7
;***Calculations***
rs=rm*smr
li=rc-rs
gls=rs/gns
gli0=lrat*gls
gni0=li/gli0
df=gni0-int(gni0)
if df<0.5
gni=int(gni0)
else
gni=int(gni0)+1
end_if
gli=li/gni
gnh=gns+gni
rcc=gls*gns*gli*gni

275
;*** Calculations for Generating Vertical Grids ***
glv1=hd/gnv1
glv20=glratv*glv1
gnv20=h2/glv20
dfv1=gnv20-int(gnv20)
if gnv20=0
glv2=0
else
if dfv1<=0.5
gnv2=int(gnv20)
glv2=h2/gnv2
else
gnv2=1+int(gnv20)
glv2=h2/gnv2
end_if
end_if
znv=gnv1+gnv2
;***Required Grid Lines***
ismear=gns+1
imax=gnh+1
jmax=znv+1
jdrain=gnv2+1
i1=int((imax-ismear)/2)+ismear
jmid=int(jmax/2)
jtop=jmax-1
end
indata
;*** Mesh Generating ***
config ax gw
grid gnh znv
model cam
gen 0,0 0,hc rs,hc rs,0
gen rs,0 rs,hc rc,hc rc,0
ini x=0 i=1
ini y=0 j=1
;***FUNCTION "vzone": Mesh Generating in Vertical Direction***
def vzone
s1=2.0
loop while s1<=jdrain
y1=y1+glv2
command
ini y=y1 j=s1
endcommand
s1=s1+1
endloop
loop while s1<=jmax
y1=y1+glv1
command
ini y=y1 j=s1

276
 endcommand
s1=s1+1
endloop
end
vzone
;***End of " vzone" FUNCTION" ***
def szone
aa=2
i2=1+gns
loop while aa<=i2
x2=x2+gls
command
ini x=x2 i=aa
endcommand
aa=aa+1
endloop
end
def izone
i3=1+gnh
x3=rs
loop while aa<=i3
x3=x3+gli
command
ini x=x3 i=aa
endcommand
aa=aa+1
endloop
end
szone
izone
group 'initial01' notnull
prop dens sden poiss spois bu sbulk mm smm lambda slanda kappa skapa
group 'initial01'
prop mpc smpc mp1 smp1 mv_l smv1 group 'initial01'
prop perm ki por spor group 'initial01'
water bulk wbu tens wt den wden
;*** Boundary Conditions ***
fix x i=1
fix x i=imax
fix x y j=1
;*** Initial Conditions ***
call ininv.fis
set grav 9.8
set wth=hc k0x=kx0 k0z=kz0
ininv
;*** Recording the PWP and Settlement at the Location of PWPTs and LVDT***
history 1 gwtime
history 2 unbalanced
history 6 ydisp i=1, j=21

277
 history 10 ydisp i=3, j=21
history 11 ydisp i=14, j=21
history 14 gpp i=24, j=4
history 15 gpp i=24, j=8
history 18 gpp i=24, j=12
history 19 gpp i=24, j=16
history 20 gpp i=4, j=1
history 21 gpp i=8, j=1
history 22 gpp i=12, j=1
history 23 gpp i=16, j=1
history 24 gpp i=21, j=1
history 25 ydisp i=14, j=21
;*** Undrained Analysis***
set step 2000000
set flow off
solve sratio 1e-3
;*** Simulation of Vertical drain & Smear Zone ***
initial pp 1.96e3 var 0 -1.96e3 i=1 j=jdrain, jmax
initial s=1 from 1,jdrain to 1,jmax
fix pp i=1 j=jdrain, jmax
fix sat i=1 j=jdrain, jmax
def pp1
zh0=1
zh1=ismear-1
zh2=ismear
zh3=imax-1
zv0=1
zv1=jdrain-1
zv2=jdrain
zv3=jmax-1
if gnv20=0
command
group 'smear' i=zh0,zh1 j=zv0,zv3
prop perm ks por spor group 'smear'
endcommand
else
command
group 'smear' i=zh0,zh1 j=zv2,zv3
prop perm ks por spor group 'smear'
endcommand
end_if
end
pp1
;*** Staged Loading & Consolidation Process***
apply pres 2e4 hist ramp from 1,jmax to imax,jmax
set step 3000000
set flow on srat 1e-3
set nmech 1
solve auto on age 3.2e5

278
apply pres 5e4 hist ramp from 1,jmax to imax,jmax
solve auto on age 1.3e6
apply pres 1e5 hist ramp from 1,jmax to imax,jmax
solve auto on age 3.2e6
apply pres 2e5 hist ramp from 1,jmax to imax,jmax
solve auto on age 4.2e6

279