Pham Thang Investigating Composite Behavior of Geosynthetic Reinforced Soil Mass Thesis

INVESTIGATING COMPOSITE BEHAVIOR OF GEOSYNTHETIC-
REINFORCED SOIL (GRS) MASS

by
Thang Quyet Pham
B.S., Hanoi University of Civil Engineering, 1993
M.S., Hanoi University of Civil Engineering, 2001
A thesis submitted to the

University of Colorado Denver
In partial fulfillment
of the requirements for the degree of
Doctoral of Philosophy
Civil Engineering
2009
Thang Quyet Pham (Ph.D., Civil Engineering)
“Investigating Composite Behavior of Geosynthetic-Reinforced Soil (GRS)
Mass”
Thesis directed by Professor Jonathan T. H. Wu
ABSTRACT
A study was undertaken to investigate the composite behavior of a Geosynthetic

Reinforced Soil (GRS) mass. Many studies have been conducted on the behavior
of GRS structures; however, the interactive behavior between the soil and
geosynthetic reinforcement in a GRS mass has not been fully elucidated. Current
design methods consider the reinforcement in a GRS structure as “tiebacks” and
adopt a design concept the reinforcement strength, Tf, and reinforcement spacing,
Sv, have the same effects on the performance of a GRS structure. This has
encouraged the designers to use stronger reinforcement at larger spacing, as the
use of larger spacing will generally reduce time and effort in construction.
A series of large-size Generic Soil-Geosynthetic Composite (GSGC) tests were

designed and conducted in the course of this study to examine the behavior of
GRS mass under well-controlled conditions. The tests clearly demonstrated that
reinforcement spacing has a much stronger effect on the performance of GRS
mass than reinforcement strength. An analytical model was established to
describe the relative contribution of reinforcement strength and reinforcement
spacing. Based on the analytical model, equations for calculating the apparent
cohesion of a GRS composite, the ultimate load carrying capacity of a reinforced
DEDICATION
This thesis is dedicated to my loving parents, Lam Van Pham and Khoi
Thi Pham, who have continuously given me unlimited support in achieving all my
life goals.
ACKNOWLEDGMENTS
I would like to express my most sincere gratitude to my thesis advisor,

Professor Jonathan T.H. Wu, for his dedicated support and guidance throughout
the course of this study. His clear insight of the subject has made my study both a
great learning experience and a joy. I also wish to thank members of my thesis
committee, Professors Hon-Yim Ko, John McCartney, Brian Brady, and Ronald
Rorrer for their helpful comments.
A special thank-you is extended to Michael Adams of the Federal
Highway Administration for his enthusiastic assistance and expert technical
support of the GSGC tests. My gratitude also goes to Jane Li and Thomas Stabile
for their help with the GSGC tests. Without their help, I could not have
conducted five successful tests during my three-month stay at the Turner-Fairbank
Highway Research Center in McLean, Virginia.
I truly appreciate the help of a dear brother and a loyal partner, Dr. Sang
Ho Lee, a visiting professor from Kyungpook National University, South Korea,
who helped me with all my experiments, including those I conducted at the
Turner-Fairbank Highway Research Center.
Last but not least, I would like to thank my wife Thuy Vu and our three
young daughters for standing by me and for encouraging me every step of the
way.
I feel blessed to have all these nice people around me in the course of this
study. Without them, this thesis would not be a reality.
CONTENTS
Figures.......................................................................................................................... xi
Tables ...................................................................................................................... xviii
Chapter
1. Introduction ........................................................................................................1
1.1 Problem Statement .............................................................................................1
1.2 Research Objectives ...........................................................................................4
1.3 Tasks of research................................................................................................5
2. Literature Review...............................................................................................9
2.1 Mechanics of Reinforced Soil ............................................................................9
2.2 Composite Behavior of GRS Mass ..................................................................15
2.3 Compaction-Induced Stresses in an Unreinforced Soil Mass ..........................27
2.3.1 Lateral Earth Pressure Estimation by Rowe (1954) .........................................27
2.3.2 Stress Path Theory by Broms (1971) and Extension of Broms’ Work by
Ingold (1979) ...................................................................................................31
2.3.3 Finite Element Analysis by Aggour and Brown (1974) ..................................36
2.3.4 Compaction-Induced Stress Models by Seed (1983) .......................................40
2.4 Compaction-Induced Stresses in a Reinforced Soil Mass ...............................62
2.4.1 Ehrlich and Mitchell (1994) .............................................................................62
2.4.2 Hatami and Bathurst (2006) .............................................................................65
2.4.3 Morrison et al. (2006) ......................................................................................66
2.5 Highlights on Compaction-Induced Stresses ...................................................68
3. Analytical Model for Calculating Lateral Displacement of a GRS Wall with
Modular Block Facing .....................................................................................71
3.1 Review of Existing Methods for Estimating Maximum Wall Movement .......73
3.1.1 The FHWA Method (Christopher et al., 1989) ................................................74
3.1.2 The Geoservices Method (Giroud, 1989) ........................................................76
vii
3.1.3 The CTI Method (Wu, 1994) ...........................................................................77
3.1.4 The Jewell-Milligan Method............................................................................78
3.2 Developing an Analytical Model for Calculating Lateral Movement and
Connection Forces of a GRS Wall ...................................................................84
3.2.1 Lateral Movement of GRS Walls with Negligible Facing Rigidity ................85
3.2.2 Connection Forces for GRS walls with Modular Block Facing ......................87
3.2.3 Lateral Movement of GRS Walls with Modular Block Facing .......................92
3.3 Verification of Analytical Model .....................................................................93
3.3.1 Comparisons with the Jewell-Milligan Method for Lateral Wall Movement..93
3.3.2 Comparisons of with Measured Data of Full-Scale Experiment by Hatami and
Bathurst (2005 and 2006).................................................................................99
3.4 Summary ......................................................................................................103
4. The Generic Soil-Geosynthetic Composite (GSGC) Tests ............................104
4.1 Dimension of the Plane Strain GSGC Test Specimen ...................................104
4.2 Apparatus for Plane Strain Test .....................................................................119
4.2.1 Lateral Deformation .......................................................................................119
4.2.2 Friction ...........................................................................................................119
4.3 Test Material ..................................................................................................121
4.3.1 Backfill...........................................................................................................121
4.3.2 Geosynthetics .................................................................................................128
4.3.3 Facing Block ..................................................................................................132
4.4 Test Program ..................................................................................................133
4.5 Test Conditions and Instrumentation .............................................................134
4.5.1 Vertical Loading System................................................................................134
4.5.2 Confining Pressure .........................................................................................134
4.5.3 Instrumentation ..............................................................................................134
4.5.4 Preparation of Test Specimen for GSGC Tests .............................................143
4.6 Test Results ....................................................................................................163
4.6.1 Test 1-Unreinforced Soil................................................................................163
4.6.2 Test 2-GSGC Test (T, Sv) ..............................................................................169
viii
4.6.3 Test 3-GSGC Test (2T, 2Sv) ..........................................................................187
4.6.4 Test 4-GSGC Test (T, 2Sv) ............................................................................199
4.6.5 Test 5-GSGC Test (unconfined with T, Sv) ...................................................211
4.7 Discussion of the Results ...............................................................................224
4.7.1 Effects of Geosynthetic Inclusion (Comparison between Tests 1 and 2) ......224
4.7.2 Relationship between Reinforcement Spacing and Reinforcement Strength
(Comparison between Tests 2 and 3) .............................................................226
4.7.3 Effects of Reinforcement Spacing (Comparison between Tests 2 and 4) .....228
4.7.4 Effects of Reinforcement Strength (Comparison between Tests 3 and 4) .....230
4.7.5 Effects of Confining Pressure (Comparison between Tests 2 and 5) ............231
4.7.6 Composite Strength Properties ......................................................................233
5. Analytical Models for Evaluating CIS, Composite Strength Properties of a
GRS Composite, and Required Reinforcement Strength...............................235
5.1 Evaluating CIS in a GRS Mass ......................................................................236
5.1.1 Conceptual Model for Simulation of Fill Compaction of a GRS Mass .........236
5.1.2 A Simplified Model to Simulate Fill Compaction of a GRS Mass................237
5.1.3 Model Parameters of the Proposed Compaction Simulation Model ..............239
5.1.4 Simulation of Fill Compaction Operation......................................................241
5.1.5 Estimation of K2,c ...........................................................................................246
5.2 Strength Properties of GRS Composite .........................................................250
5.2.1 Increase Confining Pressure ..........................................................................251
5.2.2 Apparent Cohesion and Ultimate Pressure Carrying Capacity of a GRS
Mass ...............................................................................................................257
5.3 Verification of the Analytical Model with Measured Data ...........................258
5.3.1 Comparison between the Analytical Model and GSGC Test Results............258
5.3.2 Comparison between the Analytical Model and Elton and Patawaran’s Test
Data ...............................................................................................................261
5.3.3 Comparison of the Results between the Analytical Model and Finite Element
Results ............................................................................................................266
5.4 Required Reinforcement Strength in Design .................................................268
ix
5.4.1 Proposed Model for Determining Reinforcement Force ...............................268
5.4.2 Comparison of Reinforcement Strength between the Analytical Model and
Current Design Equation................................................................................270
5.4.3 Verification of the Analytical Model for Determining Reinforcement
Strength ..........................................................................................................271
6 Finite Element Analyses ................................................................................275
6.1 Brief Description of Plaxis 8.2.......................................................................275
6.2 Compaction-Induced Stress in a GRS Mass ..................................................278
6.3 Finite Element Simulation of the GSGC Tests ..............................................280
6.3.1 Simulation of GSGC Test 1 ...........................................................................287
6.4 FE Analysis of GSGC Test 2 under Different Confining Pressures and
Dilation Angle of Soil-Geosynthetic Composites..........................................300
6.5 Verification of Compaction-Induced Stress Model .......................................302
7. Summary, Conclusions and Recommendations .............................................308
7.1 Summary ........................................................................................................308
7.2 Findings and Conclusions ..............................................................................309
Appendix A ...............................................................................................................311
Appendix B ...............................................................................................................345
References ..................................................................................................................349
x
LIST OF FIGURES
Figure
1.1 Typical Cross-Section of a GRS Wall with Modular Block Facing ..................8
2.1 Concept of Apparent Cohesion due to the Presence of Reinforcement
(Scholosser and Long, 1972) .........................................................................10
2.2 Concept of Apparent Confining Pressure due to the Presence of
Reinforcement (Yang, 1972) ...........................................................................11
2.3 Strength Envelopes for Sand and Reinforced Sand (Mitchell and Villet,
1987) ................................................................................................................14
2.4 Triaxial Compression Tests (Broms, 1977) .....................................................16
2.5 Reinforced Triaxial Test Specimen (Elton and Patawaran, 2005) ...................17
2.6 Stress-Strain Curves of Samples Reinforced at Spacing of 12 in. and 6 in.
in Large-Size Unconfined Compression Tests (Elton and Patawaran,
2005) ................................................................................................................18
2.7 Mini Pier Experiments (Adams, 1997) ............................................................19
2.8 Stress-Strain Curve (Adams, et al., 2007)........................................................20
2.9 Test Set-up of Large Triaxial Tests with 1,100 mm High and 500 mm in
Diameter (Ziegler, et al., 2008) ........................................................................21
2.10 Large-Size Triaxial Test Results (Ziegler, et al., 2008) ...................................22
2.11 Vertical Stress Distribution at 6-kN Vertical Load of the GRS Masses with
and without Reinforcement (Ketchart and Wu, 2001) .....................................24
2.12 Horizontal Stress Distribution at 6-kN Vertical Load of the GRS Masses
with and without Reinforcement (Ketchart and Wu, 2001) .............................25
2.13 Shear Stress Distribution at 6-kN Vertical Load the GRS Masses with and
without Reinforcement (Ketchart and Wu, 2001) ...........................................26
2.14 Schematic Illustration of Rowe’s Theory (Rowe, 1954) .................................29
2.15 Results of the Two-Directional Direct Shear Tests (Rowe, 1954) ..................30
2.16 Hypothetical Stress Path during Compaction (Broms, 1971) ..........................32
2.17 Residual Lateral Earth Pressure Distribution (Broms, 1971) ..........................34
xi
2.18 Hypothetical Stress Path of Shallow and Deep Soil Elements (Broms,
1971) ................................................................................................................35
2.19 A Sample Problem Analyzed by Aggour and Brown (1974) ..........................39
2.10 The First-Cycle K0-Reloading Model (Seed, 1983) ........................................41
2.21 Suggested Relationship between sinφ’ and α (Seed, 1983) .............................42
2.22 Typical K0-Reloading Stress Paths (Seed, 1983) .............................................43
2.23 K0-Unloading following Reloading (Seed, 1983) ............................................45
2.24 Unloading after Moderate Reloading (Seed, 1983) .........................................47
2.25 Basic Components of the Non-Linear K0-Loading/Unloading Model (Seed,
1983) ................................................................................................................49
2.26 Profile of Δσ h' ,vc , p against a Vertical Wall for a Single Drum Roller (Seed,
1983) ................................................................................................................50
2.27 Stress Path Associated with Placement and Compaction of a Typical Layer
of Fill (Seed, 1983) ..........................................................................................51
2.28 Bi-Linear Approximation of Non-Linear K0-Unloading Model (Seed,
1983) ................................................................................................................52
2.29 Relationship between K2 and F in the Bi-Linear Unloading Model (Seed,
1983) ................................................................................................................52
2.30 Relationship between K3 and β 3 in the Bi-Linear Model (Seed, 1983) .........53
2.31 Basic Components of the Bi-Linear Model (Seed, 1983) ................................55
2.32 Compaction Loading/Unloading Cycles in the Bi-Linear Model (Seed,
1983) ................................................................................................................57
2.33 An Example Problem for Hand Calculation of Peak Vertical Compaction
Profile (Seed, 1983) .........................................................................................60
2.34 Solution Results from the Bi-Linear Model and Non-Linear Model (Seed,
1983) ................................................................................................................61
2.35 Assumed Stress Path (Ehrlich and Mitchell, 1994) .........................................63
2.36 Compaction and Reinforcement Stiffness Typical Influence ..........................65
2.37 FE Model for FE Analysis (Morrison, et al., 2006) .........................................67
3.1 Basic Components of a GRS Wall with a Modular Block Facing ...................73
xii
3.2 Empirical Curve for Estimating Maximum Wall Movement during
Construction in the FHWA Method (Christopher, et al., 1989) ......................75
3.3 Assumed Strain Distribution in the Geoservices Method ................................77
3.4 Stress Characteristics and Velocity Characteristics behind a Smooth
Retaining Wall Rotating around the Toe (Jewell and Milligan, 1989) ............79
3.5 Major Zones of Reinforcement Forces in a GRS Wall and the Force
Distribution along reinforcement with Ideal Length (Jewell and Milligan,
1989) ................................................................................................................80
3.6 Charts for Estimating Lateral Displacement of GRS Walls with the Ideal
Layout (Jewell and Milligan, 1989) .................................................................83
3.7 Major Zones of Reinforcement Forces in a Reinforces Soil Wall (Jewell
and Milligan, 1989)..........................................................................................85
3.8 Forces acting on Two Facing Blocks at Depth zi.............................................88
3.9 Connection Forces in Reinforcement (q = 0)...................................................91
3.10 Connection Forces in Reinforcement (q = 50).................................................91
3.11 Comparison of Lateral Displacement Calculated by Jewell-Milligan
Method and the Analytical Model, γ b = 0 ......................................................95
Method and the Analytical Model, γ b = 10 ....................................................96
3.15 Configuration of a Full-Scale Experiment of a GRS Wall with Modular
Block Facing (Hatami and Bathurst, 2005 and 2006)....................................100
3.16 Comparisons of Measured Lateral Displacement with Jewell-Milligan
Method and the Analytical Model .................................................................102
4.1 Typical Geometric and Loading Conditions of a GRS Composite................107
4.2 Global Stress-Strain Curves for Soil-Geosynthetic Composites of Different
Dimensions under a Confining Pressure of 0 kPa .........................................109
4.3 Global Volume Change Curves for Soil-Geosynthetic Composites of
Different Dimensions under a Confining Pressure of 0 kPa ..........................110
xiii
4.4 Global Stress-Strain Curves for Soil-Geosynthetic Composites of Different
Dimensions under a Confining Pressure of 30 kPa .......................................111
4.5 Global Volume Change Curves for Soil-Geosynthetic Composites of
Different Dimensions under a Confining Pressure of 30 kPa ........................112
4.6 Global Stress-Strain Curves of the Unreinforced Soil under a Confining
Pressure of 30 kPa ..........................................................................................114
4.7 Global Volume Change Curves of the Unreinforced Soil under a Confining
Pressure of 30 kPa ..........................................................................................115
4.8 Specimen Dimensions for the GSGC Tests ...................................................116
4.9 Front View of the Test Setup .........................................................................117
4.10 Plan View of the Test Setup...........................................................................118
4.11 The Test Bin ...................................................................................................120
4.12 Grain Size Distribution of Backfill ................................................................124
4.13 Typical Triaxial Test Specimen before and after Test ...................................125
4.14 Triaxial Test Results ......................................................................................126
4.15 Mohr-Coulomb Failure Envelops of Backfill ................................................127
4.16 Uni-Axial Tension Test of Geotex 4x4 ..........................................................130
4.17 Load-Deformation Curves of the Geosynthetics ...........................................132
4.18 Locations of LVDTs and Digital Dial Indicator ............................................136
4.19 Strain Gauges on Geotext 4x4 Geotextile ......................................................138
4.20 Strain Gauges Mounted on Geotex 4x4 Geotextile .......................................139
4.21 Calibration Curve for Single-Sheet Geotex 4x4 ............................................141
4.22 Calibration Curve for Double-Sheet Geotex 4x4...........................................142
4.23 Applying Grease on Plexiglass Surfaces .......................................................145
4.24 Attaching Membrane .....................................................................................146
4.25 Placement of the First Course of Facing Block .............................................147
4.26 Compaction of the First Lift of Backfill ........................................................148
4.27 Placement of Backfill for the Second Lift .....................................................149
4.28 Placement of a Reinforcement Sheet .............................................................150
xiv
4.29 Completion of Compaction of the Composite Mass and Leveling the Top
Surface with 5 mm-thick Sand Layer ............................................................151
4.30 Completed Composite Mass with a Geotextile Sheet on the Top Surface ....152
4.31 Covering the Top Surface of the Composite Mass with a Sheet of
Membrane ......................................................................................................153
4.32 Removing Facing Blocks and Trimming off Excess Geosynthetic
Reinforcement ................................................................................................154
4.33 Insertion of the Strain Gauge Cables though Membrane Sheet .....................155
4.34 Vacuuming the Composite Mass with a Low Pressure .................................156
4.35 Sealing the Connection between Cable and Membrane with Epoxy to
Prevent Air Leaks ..........................................................................................157
4.36 Checking Air Leaks under Vacuuming..........................................................158
4.37 The LVDTs on an Open Side of Test Specimen............................................159
4.38 Location of Selected Points to Trace Internal Movements of Tests ..............160
4.39 Soil Dry Unit Weight Results during Specimen Preparation of Five GSGC
Tests ...............................................................................................................161
4.40 Soil Mass at Failure of Test 1 ........................................................................164
4.41 Results of Test 1-Unreinforced Soil Mass .....................................................165
4.42 Lateral Displacements on the Open Face of Test 1 .......................................166
4.43 Internal Displacements of Test 1 ...................................................................167
4.44 Composite Mass at Failure of Test 2 .............................................................170
4.45 Close-up of Shear Bands at Failure of Area A in Figure 4.44 .......................171
4.46 Failure Planes of the Composite Mass after Testing in Test 2 ......................172
4.47 Results of Test 2-Reinforced Soil Mass.........................................................173
4.50 Locations of Strain Gauges Geosynthetic Sheets in Test 2 ...........................178
4.51 Reinforcement Strain Distribution of the Composite Mass in Test 2 ............179
4.52 Aerial View of the Reinforcement Sheets Exhumed from the Composite
Mass after Test 2 ............................................................................................184
4.53 Location of Rupture Lines of Reinforcement in Test 2 .................................185
xv
4.54 Composite Mass after Testing of Test 3 ........................................................189
4.55 Global Stress-Strain Relationship of Test 3 ...................................................190
Mass after Test 3 ............................................................................................196
Mass after Test 4 ............................................................................................208
4.70 Composite Mass after Failure of Test 5 .........................................................213
4.73 Lateral Displacements on the Open Face of Test ..........................................216
Mass after Test 5 ............................................................................................221
5.1 Conceptual Stress Path for Compaction of a GRS Mass ...............................237
xvi
5.2 Stress Path of the Proposed Simplified Model for Compaction of a GRS
Mass ...............................................................................................................238
5.3 Locations of Compaction Loads and Stress Paths during Compaction at
Depth a along Section I-I as Compaction Loads Moving toward Section
I-I....................................................................................................................243
5.4 Locations of Compaction Loads and Stress Paths during Compaction at
Depth a along Section I-I as Compaction Loads Moving away from Section
I-I....................................................................................................................244
5.5 Stress Path at Depth a when Subject to Multiple Compaction Passes ...........245
5.6 Stress Path of the Proposed Model for Fill Compaction of a GRS Mass ......246
5.7 Concept of Apparent Confining Pressure and Apparent Cohesion of a GRS
Composite ......................................................................................................250
5.8 An Ideal Plane-Strain GRS Mass for the SPR Model....................................254
5.9 Equilibrium of Differential Soil and reinforcement Elements .......................254
5.10 Reinforced Soil Test Specimen before Testing (Elton and Patawanran,
2005) ..............................................................................................................262
5.11 Backfill Grain Size Distribution before and after Large-Size Triaxial Tests
(Elton and Patawanran, 2005) ........................................................................263
5.12 Large-Size Triaxial Test Results (Elton and Patawanran, 2005) ...................263
6.1 Distribution of Residual Lateral Stresses of a GRS mass with Depth due to
Fill Compaction .............................................................................................279
6.2 Comparison of Results for GSGC Test 1 .......................................................288
6.3 Comparison of Lateral Displacement at Open Face of GSGC Test 1 ...........289
6.4 Comparison of Global Stress-Strain Relationship of GSGC Test 2 ..............291
6.6 Comparison of Internal Displacements of GSGC Test 2 ...............................293
6.7 Comparison of Reinforcement Strains of GSGC Test 2 ................................294
6.8 Comparison of Global Stress-Strain Relationship of GSGC Test 3 ..............296
6.10 Comparison of Internal Displacements of GSGC Test 3 ...............................298
6.11 Comparison of Reinforcement Strains of GSGC Test 3 ................................299
6.12 FE analyses of Test 2 with Different Confining Pressures ............................301
xvii
6.13 FE Mesh to Simulate CIS of a Reinforced Soil Mass ....................................304
6.14 Lateral Stress Distribution of a GRS Mass from FE Analyses ......................305
6.15 Comparison of Residual Lateral Stresses of a GRS Mass due to Fill
Compaction between FE Analyses and Analytical Model ............................306
6.16 Comparison of Residual Lateral Stresses of a GRS Mass due to Fill
Compaction between FE Analyses with Coarse Mesh and Analytical
Model .............................................................................................................307
xviii
LIST OF TABLES
Table
2.1 Properties of material for the mini pier experiments (Adams, et al., 2007) ....19
2.2 None-Linear K0-loading/unloading model parameters (Seed, 1983) ..............48
2.3 Bi-Linear K0-loading/unloading model parameters (Seed, 1983) ...................54
4.1 Conditions and properties of material used in FE analyses ...........................108
4.2 Summary of some index properties of backfill ..............................................123
4.3 Summary of Geotex 4x4 properties ...............................................................128
4.4 Properties of Geotex 4x4 in fill-direction ......................................................131
4.5 Test program form the GSGC Tests ..............................................................133
4.6 Dimensions of the GSGC Test Specimens before Testing ............................162
4.7 Some Test Results for Test 1 .........................................................................168
4.12 Comparison between Test 1 and Test 2 .........................................................225
4.13 Comparison between Test 2 and Test 3 with the same Tf/Sv ratio.................227
4.17 Comparison of strength properties of five GSGC Tests ...............................234
5.1 Model parameters for the proposed compaction simulation model ...............240
5.2 Values of factor r under different applied pressure and reinforcement
lengths ............................................................................................................256
5.3 Comparison of the results between the analytical model and the GSGC
tests ................................................................................................................259
xix
5.4 Comparison of the results between Schlosser and Long’s method and GSGC
tests ................................................................................................................260
5.5 Comparison of the results between the analytical model and Elton and
Patawaran’s tests (2005) ................................................................................264
5.6 Comparison of the results between Schlosser and Long’s method and Elton
and Patawaran’s tests (2005) .........................................................................265
5.7 Comparison of the results between the analytical model and the FE results for
GSGC Test 2 ..................................................................................................267
5.8 Comparison of reinforcement forces between proposed model and current
design equation for a GRS wall .....................................................................272
5.9 Comparison of reinforcement forces between proposed model and the GSGC
tests ................................................................................................................273
5.10 Comparison of reinforcement forces between proposed model and test data
from Elton and Patawaran (2005) ..................................................................274
6.1 Parameters and properties of the GSGC Tests used in analyses ....................282
6.2 The steps of analysis for the GSGC Tests .....................................................284
xx
1. INTRODUCTION
1.1 Problem Statement
Over the past two decades, Geosynthetic-Reinforced Soil (GRS) structures, including
retaining walls, slopes, embankments, roadways, and load-bearing foundations, have
gained increasing popularity in the U.S. and abroad. In actual construction, GRS
structures have demonstrated a number of distinct advantages over their conventional
counterparts. GRS structures are generally more ductile, more flexible (hence more
tolerant to differential settlement and to seismic loading), more adaptable to low-
permeability backfill, easier to construct, require less over-excavation, and more
economical than conventional earth structures (Wu, 1994; Holtz, et al., 1997;
Bathurst, et al., 1997).
Among the various types of GRS structures, GRS walls have seen far more
applications than other types of reinforced soil structures. A GRS wall comprises two
major components: a facing element and a geosynthetic-reinforced soil mass. Figure
1.1 shows the schematic diagram of a typical GRS wall with modular block facing.
The facing of a GRS wall may take various shapes and sizes. It may also be made of
different materials. The other component of a GRS wall, a geosynthetic-reinforced
soil mass, however, is always a compacted soil mass reinforced by layers of
geosynthetic reinforcement.
It is a well-known fact that soil is weak in tension and relatively strong in

compression and shear. In a reinforced soil, the soil mass is reinforced by
1
incorporating an inclusion (or reinforcement) that is strong in tensile resistance.
Through soil-reinforcement interface bonding, the reinforcement restrains lateral
deformation of the surrounding soil, increases its confinement, reduces its tendency
for dilation, and consequently increases the stiffness and strength of the soil mass.
Many studies have been conducted on the behavior of GRS structures; however, the
interactive behavior between soil and reinforcement in a GRS mass has not been fully
elucidated. This has resulted in design methods that are fundamentally deficient in a
number of aspects (Wu, 2001). Perhaps the most serious deficiency with the current
design methods is that they ignore the composite nature of the reinforced soil mass,
and simply consider the reinforcement as “tiebacks” that are being added to the soil
mass. In current design methods, the reinforcement strength is determined by
requiring that the reinforcement be sufficiently strong to resist Rankine, Coulomb or
at-rest pressure that is assumed to be unaffected by the configuration of the
reinforcement. Specifically, the design strength of the reinforcement, Τrequired, has
been determined by multiplying an assumed lateral earth pressure at a given depth,
σh, by a prescribed value of reinforcement spacing, Sv, and a safety factor, Fs, i.e.,
Trequired = σ h ∗ S v ∗ Fs (1.1)
Equation 1.1 implies that, as along as the reinforcement strength is kept linearly
proportional to the reinforcement spacing, all walls with the same σh (i.e., walls of a
given height with the same backfill that is compacted to the same density) will behave
the same. In other words, a GRS wall with reinforcement strength of T at spacing Sv
will behave the same as one with reinforcement strength of 2*T at twice the spacing
2*Sv. Note that Equation 1.1 has very important practical significance. It has
encouraged designers to use stronger reinforcement at larger spacing, because the use
2
of larger spacing will generally reduce time and effort in construction.
A handful of engineers, however, have learned from actual construction that Equation
1.1 cannot be true. They realized that reinforcement spacing appears to play a much
greater role than reinforcement strength in the performance of a GRS wall.
Researchers at the Turner-Fairbank Highway Research Center have conducted a
series of full-scale experiments (Adams, 1997; Adams, et al., 2007) in which a weak
reinforcement at small spacing and a strong reinforcement (with several times the
strength of the weak reinforcement) at twice the spacing were load-tested. The
former was found to be much stronger than the latter. An in-depth study on the
relationship between reinforcement spacing and reinforcement stiffness/strength
regarding their effects on the behavior of a GRS mass is of critical importance to the
design of GRS structures and is urgently needed.
The effects of CIS in unreinforced soil masses and earth structures have been the
subject of study by many researchers, including Rowe (1954), Broms (1971), Aggour
and Brown (1974), Seed (1983), and Duncan, et al. (1984, 1986, 1991, and 1993).
These studies indicated that the CIS would increase significantly the lateral stresses in
soil (also known as the “locked-in” lateral stresses or “residual” lateral stresses),
provided that there was sufficient constraint to lateral movement of the soil during
compaction. The increase in lateral stresses will increase the stiffness and strength of
the compacted soil mass.
The effect of CIS is likely to be more significant in a soil mass reinforced with layers
of geosynthetics than in an unreinforced soil mass. This is because the interface
bonding between the soil and reinforcement will increase the degree of restraint to
lateral movement of the soil mass during fill compaction. With greater restraint to
lateral movement, the resulting locked-in lateral stresses are likely to become larger.
3
In numerical analysis of earth structures, the effects of CIS has either been overly
simplified (e.g., Katona, 1978; Hatami and Bathurst, 2005 and 2006; Morrison, et al.,
2006), or in most other studies, totally neglected. In the case of GRS walls, failure to
account for the CIS may be a critical culprit that has lead to the erroneous conclusion
by many numerical studies that Equation 1.1 is valid. Evaluation of compaction-
induced stresses in GRS structures is considered a very important issue in the study of
GRS structures.
In addition, GRS walls with modular block facing is rather “flexible”, hence the
design of these structures should consider not only the stresses in the GRS mass, but
also the deformation. Jewell-Milligan method (1989), recognized as the best
available method for estimating lateral movement of GRS walls applies only to walls
with little or no facing resistance. With increasing popularity of GRS walls with
modular block facing where facing rigidity should not be ignored, an improvement
over the Jewell-Milligan method for calculating lateral wall movement is needed.
1.2 Research Objectives
The objectives of this study were four-fold. The first objective was to investigate the
composite behavior of GRS masses with different reinforcing configurations. The
second objective was to examine the relationship between reinforcement strength and
reinforcement spacing regarding their effects on the behavior of a GRS mass. The
third objective was to develop an analytical model for evaluation of compaction-
induced stresses in a GRS mass. The fourth objective was to develop an analytical
model for predicting lateral movement of a GRS wall with modular block facing.
4
1.3 Tasks of Research
To achieve the research objectives outlined above, the following tasks were carried
out in this study:
Task 1: Reviewed previous studies on: (a) composite behavior of a GRS mass, (b)
compaction-induced stresses in a soil mass, and (c) reinforcing mechanism
of GRS structures.
Previous studies on composite behavior of a GRS mass were reviewed. The
review included theoretical analyses and experimental tests. Compaction-
induced stresses in an unreinforced soil mass that have been undertaken by
different researchers were also reviewed, including simulation models for
fill compaction. In addition, a literature study on reinforcing mechanisms of
GRS structures was conducted.
Task 2: Developed a hand-computation analytical model for simulation of

compaction-induced stresses in a GRS mass.
An analytical model for simulation of Compaction-Induced Stresses (CIS) in
a GRS mass was developed. The compaction model was developed by
modifying an existing fill compaction simulation model for unreinforced
soil. The model allows compaction-induced stress in a GRS mass to be
evaluated by hand computations. The CIS was implemented into a finite
element computer code for investigating performance of GRS structures.
Task 3: Developed an analytical model for the relationship between reinforcement

strength and reinforcement spacing, and derived an equation for calculating
composite strength properties.
An analytical model for the relationship between reinforcement strength and
reinforcement spacing was developed. Based on the model and the average
5
stress concept for GRS mass (Ketchart and Wu, 2001), an equation for
calculating the composite strength properties of a GRS mass was derived.
The model represents a major improvement over the existing model that has
been used in current design methods, and more correctly reflects the role of
reinforcement spacing versus reinforcement strength on the behavior of a
GRS mass. The equation allows the strength properties of a GRS mass to be
evaluated by a simple method.
Task 4: Designed and conducted laboratory experiments on a generic soil-

geosynthetic composite to investigate the performance of GRS masses with
different reinforcing conditions.
A generic soil-geosynthetic composite (GSGC) plane strain test was
designed by considering a number of factors learned from previous studies.
A series of finite element analyses were performed to determine the
dimensions of the test specimen that would yield stress-strain and volume
change behavior representative of a very large soil-geosynthetic composite
mass. Five GSGC tests with different reinforcement strength, reinforcement
spacing, and confining pressure were conducted. These tests allow direct
observation of the composite behavior of GRS mass in various reinforcing
conditions. They also provide measured data for verification of analytical
and numerical models, including the models developed in Tasks 2 and 3, for
investigating the behavior of a GRS mass.
Task 5: Performed finite element analyses to simulate the GSGC tests and analyze
the behavior of GRS mass.
Finite element analyses were performed to simulate the GSGC tests
conducted in Task 4. The analyses allowed stresses in the soil and forces in
the reinforcement to be determined. They also allowed the behavior of GRS
6
composites under conditions different from those employed in the GSGC
tests of Task 4 to be investigated.
Task 6: Verified the analytical models developed in Tasks 2 and 3 by using the
measured data from the GSGC tests and relevant test data available in the
literature.
The compaction model developed in Task 2 was employed to determine the
CIS for the GSGC tests; the results were then incorporated into a finite
element analysis to calculate the global stress-strain relationship and
compared to measured results. The measured data from the GSGC tests,
relevant test data available in the literature, and results from FE analyses
were also used to verify the analytical models developed in Task 3 for
calculating composite strength properties of a GRS mass and for calculating
required tensile strength of reinforcement based on the forces induced in the
reinforcement.
Task 7: Developed an analytical model for predicting lateral movement of GRS

walls with modular block facing.
An analytical model was developed for predicting the lateral movement of
GRS walls with modular block facing. The model was based on an existing
model for reinforced soil walls without facing (Jewell and Milligan, 1989).
The results obtained from the model were compared with measured data
from a full-scale experiment of a GRS wall with modular block facing.
7
Figure 1-1: Typical Cross-Section of a GRS Wall with Modular Block Facing
8
2. LITERATURE REVIEW
A GRS mass is a soil mass that is embedded with layers of geosynthetic

reinforcement. These layers are typically placed in the horizontal direction at vertical
spacing of 8 in. to 12 in. Under vertical loads, a GRS mass exhibits significantly
higher stiffness and strength than an unreinforced soil mass. This Chapter presents a
review of previous studies on the mechanics of reinforced soil, the composite
behavior of a Geosynthetic-Reinforced Soil (GRS) mass, and Compaction-Induced
Stresses (CIS) in a reinforced soil mass.
2.1 Mechanics of Reinforced Soil
In thee literature, three concepts have been proposed to explain the mechanical
behavior of a GRS mass: (1) the concept of enhanced confining pressure (Yang,
1972; Yang and Singh, 1974; Ingold, 1982; Athanasopoulos, 1994), (2) the concept of
enhanced material properties (Scholosser and Long, 1972; Hausmann, 1976; Ingold,
1982; Gray and Ohashi, 1983; Maher and Woods, 1990; Athanasopoulos, 1993; Elton
and Patawaran, 2004 and 2005), and (3) the concept of reduced normal strains (Basset
and Last, 1978).
The mechanics of a GRS mass has been explained by Schlosser and Long (1972) and
Yang (1972) by two concepts: (a) concept of apparent cohesion, and (b) concept of
apparent confining pressure.
9
a) Concept of apparent cohesion
In this concept, a reinforced soil is said to increase the major principle stress at failure
from σ1 to σ1R (with an apparent cohesion cR’) due to the presence of the
reinforcement, as shown by the Mohr stress diagram in Figure 2.1. If a series of
triaxial tests on unreinforced and reinforced soil elements are conducted, the failure
envelops of the unreinforced and reinforced soils shall allow the apparent cohesion
cR’ to be determined. Yang (1972) indicated that the φ value for unreinforced sand
and reinforced sand were about the same as long as slippage at the soil-reinforcement
interface did not occur.
Figure 2.1: Concept of Apparent Cohesion due to the Presence of Reinforcement

(Scholosser and Long, 1972)
10
b) Concept of increase of apparent confining pressure
In this concept, a reinforced soil is said to increase its axial strength from σ1 to σ1R
(with an increase of confining pressure Δσ3R), as shown in Figure 2.2, due to the
tensile inclusion. The value of Δσ3R can also be determined from a series of triaxial
tests, again by assuming that φ will remain the same.
Figure 2.2: Concept of Apparent Confining Pressure due to the Presence of

Reinforcement (Yang, 1972)
Note that the concept of apparent confining pressure allows the apparent cohesion to
be determined with only the strength data for the unreinforced soil as follows
(Schlosser and Long, 1972):
11
(1) Consider a reinforced soil mass with equally spaced reinforcement of strength
Tf (vertical spacing = Sv), it is assumed that the increase in confining pressure
due to the tensile inclusion Δσ3R is:
Tf
Δσ 3 R = (2.1)
Sv
(2) From Figure 2.1 and Figure 2.2 and using Rankine’s earth pressure theory to
equate the principal stress at failure σ1R,
Referring to Figure 2.1,
σ 1R = σ 3C K P + 2cR ' K P
(2.2)
Referring to Figure 2.2,
σ 1R = σ 3 R K P (2.3)
Knowing
σ 3 R = σ 3C + Δσ 3 R (2.4)
Equation 2.3 can be written as:
σ 1R = σ 3 R K P = (σ 3C + Δσ 3 R )K P (2.5)
Equating Equations 2.2 to Equation 2.5, we obtain
Δσ 3 R K P
c R '' = (2.6)
2
(3) Substituting equation 2.1 into equation 2.6,
Tf KP
c R '= (2.7)
2S v
Equation 2.7 may be very useful for evaluating the stability of a reinforced soil mass.
Given a granular soil with strength parameters c (c = 0) and φ, Equation 2.7 allows
the strength parameters of a reinforced soil mass (cR’ and φR) to be determined as a
function of T and Sv.
12
It should be noted that the validity of Equation 2.7 is rather questionable. There is a
Tf
key assumption involved in the derivation -- the assumption of Δσ 3 R = (this
Sv
expression implies that an increase in “Tf” has the same effect as a proportional
decrease in “Sv”.) Figure 2.3 shows the strength envelopes for sand and reinforced
sand based on the studies of Schlosser and Long (1972), Yang (1972), and Hausmann
(1976). Note that the increase in confining pressure was Δσ 3 = RT / S v (see Figure
2.2) based on the explanation given by Schlosser and Long (1972) and Yang (1972);
and it was Δσ 3 ≤ RT / S v in Hausmann’s study (1976), where RT ≡ T f .
13
Figure 2.3: Strength Envelopes for Sand and Reinforced Sand (Mitchell and
Villet, 1987)
14
2.2 Composite Behavior of GRS Mass
The behavior of soil-geosynthetic composites have been investigated through

different types of laboratory experiments, including: small-size triaxial compression
tests with the specimen diameter no greater than 6 in. (Broms, 1977; Gray and Al-
Refeai, 1986; Haeri et al., 2000; etc), large-size triaxial compression tests (Ziegler et
al., 2008), large-size unconfined compression tests (Elton and Patawaran, 2005),
unconfined compression tests with cubical specimens (Adams, 1997 and Adams et
al., 2007), and plane strain tests (Ketchart and Wu, 2001).
Figure 2.4 shows the effects of reinforcement layers on the stiffness and strength of
soil-geosynthetic composites conducted by Broms (1977). For unreinforced soil
specimen (number 1 in Figure 2.4) and the specimen with reinforcement at the top
and bottom (number 2 in Figure 2.4), the stress-strain curves are nearly the same.
This suggests that unless the reinforcement is placed at locations where lateral
deformation of the soil occurs, there will not be any reinforcing effect. For the
specimens with 3 and 4 layers, the stiffness and strength of the composites are
significantly higher as the reinforcement effectively restrains lateral deformation of
the soil.
15
Figure 2.4: Triaxial Compression Tests (Broms, 1977)
There are questions concerning the applicability of these small-size triaxial tests as
the reinforcement in these tests is very small compared with the typical field
installation, and factors such as gravity, soil arching, and compaction-induced stresses
are not simulated properly. For these reasons, a number of larger-size triaxial tests
and plane strain test specimens have been conducted. Elton and Patawaran (2005)
conducted seven unconfined compression tests on 2.5 ft diameter and 5 ft high
16
specimens with different types of reinforcement and spacing (see Figure 2.5). Six
types of reinforcement were used in the tests with spacing of 6 in. and 12 in. Figure
2.6 shows the stress-strain curves of the specimens reinforced by TG500 at spacing of
12 in. and 6 in. It can be seen that the strength of the soil-geosynthetic composite was
much higher at 6 in. spacing than at 12 in. spacing.
(a) (b)
Figure 2.5: Unconfined Test Specimen (a) before and (b) after Testing
(Elton and Patawaran, 2005)
17
Stress - Strain curves
300
TG500-12 (Spacing = 12 in.)
250
TG500 (Spacing = 6 in.)
200
Stress (kPa)
150
100
50
0
0 5 10 15 20 25 30
Vertical Strain εv (%)
Figure 2.6: Stress- Strain Curves of Specimens Reinforced at Spacing of 12 in.

and 6 in. in Large-Size Unconfined Compression Tests
Five unconfined “mini pier” experiments were conducted by Adams and his
associates (Adams, 1997 and Adams et al., 2007). The dimensions of the specimen
were 2.0 m high, 1.0 m wide and 1.0 m deep. The test results showed that the load-
carrying capacity is affected strongly by the spacing of the reinforcement, and not
significantly affected by the strength of the reinforcement. Figure 2.7 shows a photo
of the mini pier experiment and Table 2.1 shows the material properties and test
conditions for the tests. The stress-strain curves from the tests are presented in Figure
2.8. From the Figure, the effect of reinforcement spacing and reinforcement strength
on the behavior of the mini piers can be seen by comparing the difference between
curve B (at 0.4 m spacing) and curve D (at 0.2 m spacing) and the difference between
curve C (reinforcement strength = 21 kN/m) and curve D (reinforcement strength =
70 kN/m). The effect of reinforcement spacing is much more pronounced than the
effect of reinforcement strength.
18
Figure 2.7: Mini Pier Experiments (Adams, 1997)
Table 2.1: Properties of materials for the mini pier experiments (Adams et al., 2007)
19
Figure 2.8: Stress-Strain Curves of Mini-Pier Experiments (Adams et al., 2007)
Figure 2.9 shows the setup of the large-size triaxial tests conducted by Ziegler et al.
(2008). The specimens were 0.5 m in diameter and 1.1 m high. The results also
show that the behavior of the GRS specimens was strongly affected by reinforcement
spacing. Figure 2.10 shows the relationship between applied loads and vertical
strains of the test specimens. The strength of the specimen increases with increasing
number of reinforcement layers. The stiffness of the specimen also increases with
increasing number of reinforcement layers for strains more than about 1%. Below
1%, the stiffness is not affected by the reinforcement layers.
20
Figure 2.9: Test Setup of Large-Size Triaxial Tests with Specimens of 1100 mm
High and 500 mm in Diameter (Ziegler et al., 2008):
a) Schematic Diagram, and
b) A Photo of the Large-Size Triaxial Test
21
Figure 2.10: Large-Size Triaxial Test Results (Ziegler et al., 2008)
The behavior of soil-geosynthetic composites have also been investigated through

numerical analysis. Examples of these studies include Lee, 2000; Chen et al., 2000;
Holtz and Lee, 2002; Zhang et al., 2006; Ketchart and Wu, 2001; and Vulova and
Leshchinsky, 2003.
Vulova and Leshchinsky (2003) conducted a series of analyses using a two-

dimensional finite difference program FLAC Version 3.40 (1998). From the analysis,
it was concluded that reinforcement spacing was a major factor controlling the
behavior of GRS walls. The analysis of GRS walls with reinforcement spacing varied
from 0.2 m to 1.0 m showed that the critical wall height (defined as a general
characteristic of wall stability) always increased when reinforcement spacing
22
decreased. Reinforcement spacing also controls the mode of failure of GRS walls. In
these analyses, compaction induced stresses in soil were not included.
Comparisons of the stress distribution in a soil mass with and without reinforcement
were made by Ketchart and Wu (2001). The reinforcement was a medium strength
woven geotextile (with wide-width strength = 70 kN/m), the backfill was a
compacted road base material, and the reinforcement spacing was 0.3 m. Figures
2.11, 2.12 and 2.13 present the vertical, horizontal, and shear stress distributions,
respectively, at a vertical load of 6 kN. It is noted that the presence of the
reinforcement layers in the soil mass altered the horizontal and shear stress
distributions but not the vertical stress distribution. The horizontal and shear stresses
increased significantly near the reinforcement. The largest stresses occurred near the
reinforcement and reduced with the increasing distance from the reinforcement. The
extent of appreciable influence was only about 0.1~ 0.15 m from the reinforcement.
With the increased lateral stress, the stiffness and strength of the soil will become
larger. They emphasized the importance of keeping reinforcement spacing to be less
than 0.3 m for GRS walls.
23
Figure 2.11: Vertical Stress Distribution at 6-kN Vertical Load of the GRS Masses
(a) with and (b) without Reinforcement (Ketchart and Wu, 2001)
24
Figure 2.12: Horizontal Stress Distribution at 6-kN Vertical Load of the GRS
Masses (a) with and (b) without Reinforcement (Ketchart and Wu, 2001)
25
Figure 2.13: Shear Stress Distribution at 6-kN Vertical Load the GRS Masses (a)
with and (b) without Reinforcement (Ketchart and Wu, 2001)
26
2.3 Compaction-Induced Stresses in an Unreinforced Soil Mass
Many studies have been conducted to address the compaction-induced stresses (CIS)
in a soil mass. As early as 1934, Terzaghi noted that compaction significantly
affected lateral earth pressures. Rowe (1954) calculated lateral earth pressures for
conditions of wall deflection intermediate between the at-rest and fully active and
fully passive states. Rowe’s work did not directly address CIS, but it contributed
strongly to the later study by Broms (1971) on compaction-induced earth pressures.
Seed (1983), and Seed and Duncan (1984) had developed a simulation model called
the "bi-linear hysteretic loading/unloading model" to simulate the compaction effect
on vertical, non-deflecting structures. Duncan and Seed (1986) and Duncan et al.
(1991) also developed a procedure to determine lateral earth pressure due to
compaction. The studies were considered to have a strong impact on the
determination of CIS.
2.3.1 Lateral Earth Pressure Estimation by Rowe (1954)
This study was not directly related to CIS, but it addressed lateral earth pressures for
conditions of wall deflection intermediate between the at-rest and fully active and
fully passive states. Rowe’s stress-strain theory for calculations of lateral pressures
exerted on structures by cohesionless soils was based on the following hypotheses:
• The degrees of mobilization of the soil friction angle φ and the soil-wall
friction angle δ depend on the degrees of interlocking of the soil grains,
which in turn depend on the fractional movement of the shear planes or
slip strain (defined as the ratio of relative shear displacement to total slip
plane length). The friction angle developed increases from a relatively
low value to a higher limiting or ultimate value as slip strain increases.
27
• Earth pressures acting on a retaining wall or structure may be calculated
by conventional limiting equilibrium methods (i.e., gravity analyses of
sliding wedges) using the developed fractional φ and δ values.
The basic mechanics of Rowe’s theory are illustrated schematically in Figure 2.14.
The analysis is essentially a simple Coulomb analysis of sliding wedges. A sample
wedge adjacent to a wall or structures is considered as shown in Figure 2.14(a).
When the structure deflects, slip strain occurs along planes AB and AC as shown in
the figure. Assuming no soil compression, slip strain along each plane was calculated
as the ratio of shear displacement along the plane to the length of the plane. The
forces acting on the typical sliding wedge were shown in Figure 2.14(b).
28
Figure 2.14: Schematic Illustration of Rowe’s Theory (Rowe, 1954)
Rowe substantiated his theory by performing a series of direct shear tests on different
sands, recording the friction angle developed at various levels of slip strain, and using
these values to calculate lateral earth pressures for sample problems. Figure 2.15
shows some results of these tests. By considering “tamping” or compaction as
application and removal of a surcharge pressure γh0 (γ = unit weight of soil, h0 =
29
surcharge head), he postulated that slip strains would be induced by the load
application. Rowe suggested that in the compression and shear tests, unloading
results in relatively small strain reversals. Thus, after tamping a fill behind a wall, the
lateral pressure will be almost as great as the value, which acted under the
preconsolidation pressure. From that, the pressure coefficient (K0') could be
expressed as:
⎛ h ⎞
K 0 = K 0 ⎜1 + 0 ⎟
'
(2.8)
⎝ h⎠
where h0 is the surcharge head removed and h is the overburden head ( h = σ v / γ ). In
any case, K 0 ≤ K P (KP = coefficient of passive earth pressure at limiting condition).
Figure 2.15: Results of the Two-Directional Direct Shear Tests (Rowe, 1954)
30
It is interesting to note the similarity between Rowe’s early equation for residual,
compaction-induced lateral earth pressures and an equation proposed later by Schmidt
(1967) to explain residual lateral earth pressures resulting from overconsolidation of
soils under conditions of no lateral strain (i.e., the K0-condition). Schmidt's equation
which empirically allow for some degree of relaxation of lateral stresses following
surcharge removal, can be expressed as:
α
⎛ h ⎞
K 0 = K 0 ⎜1 + 0 ⎟
'
(2.9)
⎝ h⎠
where: α = 0.3 to 0.5 for most sands

α = 1.2 sinφ' for initially normally consolidated clays
2.3.2 Stress Path Theory by Broms (1971) and Extension of Broms’ Work by
Ingold (1979)
Broms (1971) developed a stress path theory to explain residual lateral earth presses
on rigid, vertical, non-yielding structures resulting either from compaction or
surcharge loading, which is subsequently removed. The theoretical basis for Broms'
theory is illustrated in Figure 2.16. An element of soil at some depth is considered to
exist at some initial stress state represented by point A with horizontal and vertical
effective stresses of σ ' ho and σ ' vo . Compaction of the soil is considered as a process
of loading followed by unloading. When the overburden pressure is increased (i.e.,

loading), there is a little change in lateral pressure until the ratio of lateral to vertical
effective stresses is equal to K0 (denoted by point B in Figure 2.16) where K0 is the
coefficient of earth pressure at rest. Thereafter, increased vertical stress is
accompanied by an increased lateral stress according to σ ' h = K 0 σ ' v , corresponding
to primary (or virgin) loading. When the overburden pressure is subsequently
decreased (say, from point C) the corresponding decrease in lateral pressure is small
31
until the ratio of lateral to vertical effective stresses is equal to some limiting constant
K1 (denoted by point D in Figure 2.16). Thereafter, continued decrease in vertical
pressure is accompanied by a decrease in lateral stress according to σ 'h = K1σ 'v . This
idealized stress path is in agreement with the earlier hypothesis of Rowe (1954) that
stress relaxation with unloading is negligible until some limiting condition, defined
by the K1-line, is reached. By following this type of stress path, an element of soil
can be brought to a final state represented by an effective coefficient Keffective varying
K 0 ≤ K effective ≤ K 1 . Having made this idealized assumption of the stress path, Broms
then postulated that the actual stress path followed by a real soil element might be as
represented by the dash line in Figure 2.17. Rowe (1954) as well as Ingold (1979)
suggested that K1 = KP (the coefficient of passive earth pressure) reasoning that the
limiting condition reached is essentially a form of passive failure.
σh =K0 σv
C
σv
Effective overbunden pressure,
B
σ'h0 /K0
σh =K1 σv
σ'v0 A D Assumed relationship
E Actual relationship
σ'h0
Effective lateral earth pressure, σh
Figure 2.16: Hypothetical Stress Path during Compaction (Broms, 1971)
32
By employing this theory to estimate the lateral pressure exerted on a vertical, rigid,
non-yielding structure with a compacted fill, Broms considered the compaction plant
to present a load applied to the fill surface inducing vertical stresses which may be
approximated as twice as those calculated by Boussinesq stress equation for an
infinite half space. Lateral earth pressures acting against the wall were then
calculated as σ ' h = K 0σ ' v . The resulting horizontal stress distribution calculated for a
10.2 metric ton smooth wheel roller is presented as line 1 in Figure 2.17(a).
33
Figure 2.17: Residual Lateral Earth Pressure Distributions (Broms, 1971)
34
Figure 2.18(a) shows the loading path for a soil element at a shallow depth z < zcr ,
where zcr is a critical depth that will be addressed later. The soil is loaded from points
A’ to C’, then latter stages of unloading following the vertical path to point D’ and
then following the stress path σ 'h = K1σ 'v to point E’, resulting in a final condition
K 0 ' = K1 .
Figure 2.18(b) shows the loading path for a soil element at deeper depth z > zcr. After
loading from A’’ to C’’, the unloading to E’’ is not sufficient to bring the soil element
to the limiting condition σ 'h = K1σ 'v at point D’’ because the vertical stress at point A
is greater than that at point D’’.
Figure 2.18: Hypothetical Stress Paths of Shallow and Deep Soil Elements
(Broms, 1971)
35
Following the process above and knowing values of K0 and K1, a residual lateral
pressure distribution, as shown by the shaded area in Figure 2.17(a), can be
determined. The line 23 in Figure 2.17(a) represents the residual stresses for
elements below zcr and line 02 represents the limiting condition σ 'h = K1σ 'v . Point 2,
where these two lines intersect, occurs at a depth zcr called the critical depth.
By considering the backfill process as the placement of a series of soil layers each
deposited and then compacted one after the other, the compaction-induced lateral
pressure for each new layer will be surpassed in magnitude by the at-rest earth
pressures due to the static overburden. A stress distribution as shown in Figure
2.17(b) can be calculated. This type of lateral pressure distribution may be
generalized as shown in Figure 2.17(c).
Ingold (1979) applied the extension of Broms’ theory in cases where wall deflection
during backfilling were sufficient to induce an active condition in the lower layers of
a backfill which is deposited and compacted in lifts by assuming the virgin loading
path to be σ 'h = K Aσ 'v instead of Broms’ σ 'h = K 0σ 'v . Ingold postulated that passive
failure controlled the other limiting condition, therefore K1 = KP.
2.3.3 Finite Element Analysis by Aggour and Brown (1974)
Aggour and Brown (1974) were the first to model compaction-induced lateral earth
pressure by two-dimensional finite element analysis. Aggour and Brown's analysis
involved the following steps for simulation of compaction operation:
1. A layer of soil elements adjacent to a wall was modeled with some initial
modulus E1.
36
2. Compaction was modeled as some increased vertical load acting uniformly
over the entire surface of the soil. Simultaneously the soil modulus is
increased to some new and stiffer value E2 to reflect the increase in density
during compaction.
3. The compaction load was then removed. The resulting strains, deflections,
and stress redistributions were modeled using a stiffer unloading modulus Eu2.
4. A new layer of fill with modulus E1 was added to the top of the preceding
layer. This increased the vertical stresses in the underlying layer. These
increased stresses in the underlying were modeled using the soil modulus E2.
5. A surface load to model compaction of the new layer was applied, increasing
the stresses in both these soil layers, and the modulus was increased to E2.
6. The compacting load was removed and the resulting wall deflections, strains
and stress redistributions were modeled using the modulus Eu2 for both soil
layers.
7. The entire process was then incrementally repeated for subsequent soil layers.
Figure 2.19 shows a sample problem analyzed by Aggour and Brown using the
procedure above. Figure 2.19(a) shows the method used to model the soil moduli.
The effects of increased numbers of compaction 'passes' were modeled by increasing
the soil modulus E2. The soil modulus was greater for unloading than for reloading.
The geometry and material properties used in the sample problem are shown in
Figure 2.19(b). Compaction loading was modeled as a uniform unit surface pressure
of unit width acting at all points along the full length of the fill, from the wall to the
right-hand boundary of the finite element mesh. The fill materials were placed in five
4-ft lifts. The interface between soil and wall was assumed to remain bonded at all
time.
37
The results of this analysis are shown in Figure 2.19(c). These results indicate:
• Increased wall deflections with increased compactive effort (increased number
of passes was modeled by increases E2 values)
• Increased residual lateral pressures near the top of the wall with increased
compactive effort.
The second sample problem analyzed by Aggour and Brown used the same soil and
wall geometry, but this time only the last soil lift placed was compacted. The results
of this analysis are shown in Figure 2.19(d). The effects of the compaction of the last
soil layer were: increased wall deflections, and increased residual lateral earth
pressures near the top of the wall. These results were found to be in agreement, on a
qualitative basis, with field and scale-model observations of the effects of compaction
on structural deflections and residual lateral earth pressures. The analyses suggest the
potential value of finite element analysis for determining compaction-induced earth
pressures on yielding structures.
38
Figure 2.19: A Sample Problem Analyzed by Aggour and Brown (1974)
39
2.3.4 Compaction-Induced Stress Models by Seed (1983)
Seed (1983) proposed a method to estimate the effects of CIS and associated
deflections. Seed's study will be summarized in detail as it represents the most in-
depth study on the subject of CIS. Seed’s study involved the following three areas:
• Compaction induced stresses due to different stress paths, including the first-
cycle K0–reloading stress path, typical K0–reloading stress path, multi-cycle
K0-unloading/reloading stress path, and K0-unloading following reloading
stress path.
• The nonlinear and bi-linear hysteretic loading/unloading compaction models
for simulation of fill compaction behind vertical, non-deflecting structures.
• Finite element analysis of fill compaction using the non-linear and bi-linear
models.
a) First-Cycle K0-Reloading Stress Path and Typical K0-unloading/Reloading

Stress Path
Seed (1983) proposed what is termed “first-cycle K0-reloading model,” as shown in

Figure 2.20. The model was derived from fitting available data. Upon loading, the
stress path is assumed to follow the K0-line to point A (see Figure 2.20), after which
the unloading stress path is followed by K ' 0 = K 0 (OCR ) to an arbitrary point B prior
α
to reloading. Reloading is then assumed to follow a linear path to point R, the

intersection of the reloading path with the K0-line, and again to follow the K0-line
thereafter. Point R, the intersecting point of the reloading path with the virgin loading
path is determined as:
σ *h ,r = σ 'h ,min + β ∗ Δ (2.10)
40
1 *
σ * v ,r = σ h ,r (2.11)
K0
In which Δ is the decrease in horizontal effective stress from the maximum loading at
point A to the minimum unloading at point B, β is assumed constant regardless of the
degree of unloading which precedes reloading (Figure 2.20), σ’h,min is minimum
horizontal stress, σ*v,r σ*h,r are vertical and horizontal residual stress after compaction.
The unloading curve from point A to point B in Figure 2.20, can be estimated by
K ' 0 = K 0 (OCR ) . The value α can be estimated by Equation 2.12 or by Figure 2.21:
α
α = 0.018 + 0.974 sin φ ' (2.12)
Figure 2.20: The First-Cycle K0-Reloading Model (Seed, 1983)
41
Figure 2.21: Suggested Relationship between sinφ' and α (Seed, 1983)
42
Some typical results of using the model are illustrated in Figure 2.22. In this figure, β
remains constant but Δ varies with magnitude of unloading.
Figure 2.22: Typical K0-Reloading Stress Paths (Seed, 1983)
In Figures 2.21 and 2.22, K1-line is at the limiting condition and defined as:
2c'
K 1 = K 1,φ + K 1,φ
σ '3 (2.13)
⎛ φ' ⎞
K 1, φ = tan 2 ⎜ 45 0 + ⎟ (2.14)
⎝ 2⎠
where K1, φ’ = coefficient of passive lateral earth pressure, and φ’ = effective internal
friction angle of soil.
43
b) Multi-Cycle K0-Unloading/Reloading Stress Path and K0-Unloading
Following Reloading Stress Path.
Seed (1983) proposed another model for multi-cycle K0-unloading/reloading

conditions. The values α and β are assumed to remain constant regardless of the
number of loading-reloading cycles. The model is illustrated in Figure 2.23.
Three situations of unloading were considered. The first situation, as shown in Figure
2.23(a), is for unloading after significant loading, and can be described as follows:
1. Loading by following the K0-line to point A;
2. Unloading by following the K0’-line, with K 0 ' = K 0 (OCR ) , to point B;
α
3. Reloading through point R then to point C, passing through previous

maximum loading point A;
4. Subsequent loading-unloading path is assumed to follow an α-type path CD.
44
Figure 2.23: K0-unloading following Reloading (Seed, 1983)
45
The second situation, unloading after intermediate reloading, is shown in Figure
2.23(b), and can be described as follows:
1. Loading by following the K0-line to point A;
2. Unloading by following the K0’-line, with K 0 ' = K 0 (OCR ) , to point B;
α
3. Reloading through point R to point C on the virgin K0-line, but at a stress less
than the maximum previous loading point A;
4. Subsequent loading-unloading path is assumed to follow an α*-type path CD
( α * ≥ α ).
The third situation, unloading after moderate reloading, is shown in Figure 2.24. In
this situation, the stress again begins with loading to point A and unloading to point
B, then reloading to point C. In this situation, the stress state at point C is not
sufficient to reach a stress state on the K0-line. Subsequent loading is then modeled
as follows:
1. Point C is projected vertically down to point C’ on the K0-line, and point B is
projected vertically down through the same distance to point B’;
2. From point C’ unloading to point B’ by following a α*-type ( α * ≥ α ) path;
3. The actual unloading path is assumed to be “parallel” to the imaginary path
above (dash lines).
Note that in this situation, point B is the “current” minimum stress point.
46
Figure 2.24: Unloading after Moderate Reloading (Seed, 1983)
c) Non-Linear, Multi-Cycle K0-Unloading/Reloading Compaction Model
The non-linear, multi-cycle K0-unloading/reloading model was developed based on

the stress paths described in Sections (a) and (b) above. The model requires five
material parameters: α, β, K0, K1,φ and c’. The name of each model parameter,
recommended limits of each parameter, and the correlations with φ’ are given in
Table 2.2.
47
Table 2.2: Non-linear K0-loading/unloading model parameters (Seed, 1983)
Parameter Name Recommended Method of Estimation

Limits Based on φ’
α Unloading coefficient 0 ≤α ≤1
β Reloading coefficient 0 ≤ β ≤1 β ≅ 0.6
K0 Coefficient of at-rest lateral 0 ≤ K0 ≤ 1 K 0 ≅ 1 − sin φ '

earth pressure for virgin
loading
K1,φ Frictional component of K 0 ≤ K1,φ ≤ K P ⎛ φ' ⎞

limiting coefficient of at-rest K 1,φ ≅ tan 2 ⎜ 45 + ⎟
⎝ 2⎠
lateral earth pressure for
unloading
c’ Effective stress strength

envelope cohesion intercept -- --
Note: K1 = Limiting coefficient of at-rest lateral pressure for unloading

2c'
σ 'h,lim = K1σ 'v and K1 = K1,φ ' + B K1,φ '
σ '3
KP = Coefficient of passive lateral earth pressure
Basic component of the non-linear K0-loading/unloading model is described in Figure

2.25. In the figure, the K1- line is the stress path at limiting condition; K0 – line is
virgin or initial stress path; MPLP is maximum loading point; RMUP is residual
minimum unloading point; RMLP is residual maximum loading point; point R is at
residual condition after compaction.
48
Figure 2.25: Basic Components of the Non-Linear K0-Loading/Unloading Model
(Seed, 1983)
The term Δσ'h,vc,p (same as “Δ” in Figure 2.25; the subscripts “h” denotes horizontal,
“vc” denotes virgin compression, and “p” denotes peak) is referred to as the peak
change in lateral stress induced by compaction (loading only). The value of Δσ'h,vc,p
resulted from surficial loading can be obtained from simple elastic solution, such as
Boussinesq’s solution (1885), by assuming the soil is previously uncompacted (i.e., a
“virgin” soil) and by varying the type and location of the compaction plant.
Boussinesq’s solution is for a semi-infinite mass. For any point along a vertical non-
deflecting wall, the change in lateral stress will be twice as much as the values
obtained from Boussinesq’s solution.
49
Figure 2.26 (a) shows the distribution of Δσ'h,vc,p for loads applied at different
distances under the condition that the soil is not underlain by a rigid base. The
distribution of Δσ'h,vc,p under the condition that the soil is underlain by a rigid base at
a depth of 6 ft is shown in Figure 2.26(b). In the latter case, attenuation of Δσ'h,vc,p
with depth beyond the maximum point is followed by an increase with depth when a
rigid base is approached. Figure 2.27 shows stress path associated with placement
and compaction of a typical layer of fill. For a typical layer of fill at point A, stress in
soil at this point is increased to point B (because of the soil weight of this layer), and
then to point D and C due to compaction loading. For the unloading path, stress at
point C then reduced to point D (the vertical load, Δσ’v,0 is remained because of the
soil weight of this layer). After compaction, Δσ’h,r is residual horizontal stress in soil
at this layer.
Figure 2.26: Profiles of Δσ'h,vc,p against a Vertical Wall for a Single Drum Roller
50
Figure 2.27: Stress Path associated with Placement and Compaction of a Typical
Layer of Fill (Seed, 1983)
d) Simplified Bi-Linear Approximation to the Non-Linear Model
In this model, the “α-type” non-linear unloading model for the first-cycle unloading
is approximated by a bi-linear unloading path, as shown by the dashed lines in Figure
2.28. The relationship between K2 and F (as defined in Figure 2.29) in the bi-linear
unloading model is shown in Figure 2.29. Figure 2.30 shows the relationship
between K3 (slope of reloading path) and β3 (as defined in Figure 2.30) of the model.
The bi-linear K0-loading/unloading model parameters are described in Table 2.3.
51
Figure 2.28: Bi-Linear Approximation of Non-linear K0-unloading Model
Figure 2.29: Relationship between K2 and F in the Bi-Linear Unloading Model

(Seed, 1983)
52
Figure 2.30: Relationship between K3 and β3 in the Bi-Linear Model (Seed, 1983)
53
Table 2.3: Bi-linear K0-loading/unloading model parameters (Seed,1983)
Parameter Name Recommended Method of Estimation Based on

Limits φ’
K0 Coefficient of at-rest Same as non-linear model

lateral earth pressure
for virgin loading 0 ≤ K0 ≤ 1 (Recommended K 0 ≅ 1 − sin φ ' )
K1,φ’,B Frictional component 2

of limiting coefficient K 1,φ ', B ≅ K1,ϕ '
of at-rest lateral earth K 0 ≤ K1 ≤ K P 3
pressure for ⎛ φ' ⎞
unloading K 1,φ ' ≅ tan 2 ⎜ 45 o + ⎟
⎝ 2⎠
cB’ Effective stress

strength envelope
Cohesion Intercept cB ' ≥ 0 c B ' ≅ 0.8c'
- Fraction of peak 0 ≤ F ≤1 - F (or K2) should be chosen

lateral compaction such that the bi-linear unloading
F
stress retained as stress path intersects the α-type
residual stress for non-linear unloading stress path
or virgin soil at a suitable OCR.
- Incremental Recommended:
coefficient of at-rest
F =1−
(OCR − OCR ) α
K2 K0 ≥ K2 ≥ 0
(OCR − 1)
K 2 = K 0 (1 − F )
for reloading
Where a suitable OCR for
“matching” the bi-linear and
non-linear unloading curves is
typical OCR ≅ 5 .
K3 Incremental 0 ≤ K3 ≤ K0 K 3 ≅ K 2 (= K 0 (1 − F ))
coefficient of at-rest
for reloading
2c' B 2
Note: σ 'h,lim = K1σ 'v , K1 = K1,φ ',B + K1,φ ',B , K 1,φ ', B ≅ K 1,φ ' , and K1,φ ' = K P .
σ '3 3
54
Figure 2.31 shows the basic components of the bi-linear model. Virgin loading is
assumed to follow the K0-line in the same manner as in the non-linear model.
Unloading initially follows a linear stress path according Δσ h ' = K 2 Δσ v ' until a K1-
type of limiting condition is reached, at which point further unloading follows a linear
stress path according to σ h ' = K1Δσ v ' . Reloading follows a linear stress path
according to Δσ h ' = K 3 Δσ v ' until the virgin K0-loading stress path is regained, after
that further reloading follows the virgin stress path.
Figure 2.31: Basic Components of the Bi-linear Model (Seed,1983)
From Figure 2.31, the parameters of bi-linear model can be calculated as:
Δσ h ,r ' = F .Δσ h, p ' (2.15)
F = 1−
K2
= 1−
(OCR − OCR α ) (2.16)
K0 (OCR − 1)
55
β3K2
K3 = (2.17)
K
1 − 2 (1 − β 3 )
K0
56
Figure 2.32: Compaction Loading/Unloading Cycles in the Bi-Linear Model
(Seed, 1983)
57
Seed (1983) also developed a simplified hand calculation procedure for computing
the CIS. The procedure can be described by the following steps:
1. Calculate the peak lateral compaction pressure profile (i.e., Δσ'h,vc,p vs. depth
relationship) by the method described in Section c, as shown in Figure
2.32(b).
2. Multiply the Δσ'h,vc,p values with the bi-linear model parameter F.
3. Calculate the lateral residual stress as σ h ,r ' = K 0σ v ' + FΔσ h' ,vc , p .
4. Reduce the near-the-surface portion of the σ’h,r distribution with

σ h , r ' ≤ K1,φ ', Bσ v ' at all depths.
5. Increase the residual effective stress distribution such that σ h ,r ' ≥ K 0σ v ' at all
depths.
Figure 2.33 shows an example problem given by Seed (1983) to show how to
determine the compaction-induced lateral pressure on a vertical non-deflecting
structure using bi-linear model and non-linear model. The material parameters of the
models based on the given angle of friction (φ = 35o) are shown in Table 2.4.
Table 2.4: Material parameters for non-linear and bi-linear models
Non-linear model parameters Bi-linear model parameters

K0 = 1 - sinφ' = 0.43 K0 = 1 - sinφ' = 0.43
⎛ φ'⎞ 2
K1,φ ' = tan 2 ⎜ 45o − ⎟ = 3.69 K1,φ ',B = K1,φ ' = 2.46
⎝ 2⎠ 3
α = 0.63 (from Figure 2.20) 5 − 5α
F = 1− = 0.44
β = 0.6 (assumed) 5 −1
K 3 = K 2 = 0.24
c' = 0
c'B = 0
58
Figure 2.34 shows the residual lateral stresses as a result of unloading due to fill
compaction and the at-rest earth pressure. The results of the bi-linear model (the solid
line) and the non-linear model (the bold dots) are approximately the same.
59
Figure 2.33: An Example Problem for Hand Calculation of Peak Vertical
Compaction Profile (Seed, 1983)
60
Figure 2.34: Solution Results from the Bi-Linear Model and the Nonlinear Model
(Seed, 1983)
The compaction models developed by Seed (1983) has been used to determine the
CIS for full-scale experiments and reported by Duncan and Seed (1986) and Duncan
et al. (1986). The papers showed that, the CIS could be calculated based on either the
simplified method (i.e., the bi-linear model) or the non-linear model with the aid of
finite element analysis. They have shown that the resulting lateral earth pressures
determined by the models are in good agreement with measured data.
61
2.4 Compaction-Induced Stresses in a Reinforced Soil Mass
Many researchers and practicing engineers have suggested that if a granular backfill
is well compacted, a GRS mass can usually carry a great deal of loads and experience
little movement. A handful of studies on performance of GRS structures have used
simplified, and somewhat arbitrary, procedures to simulate the effects of fill
compaction. In all other studies, the compaction-induced stresses (CIS) in the fills
have been ignored completely. The CIS in a GRS mass is likely to be more
pronounced than those induced in an unreinforced soil mass because soil-
reinforcement interface friction tends to restrain lateral deformation of the soil mass
and results in greater values of CIS. A review of previous studies on GRS masses
including the effects of CIS is presented below.
2.4.1 Ehrlich and Mitchell (1994)
Ehrlich and Mitchell (1994) presented a procedure to include CIS in the analysis of
reinforced soil walls and noted that CIS was a major factor affecting the
reinforcement tensions. The assumptions involved in the procedure are:
• The stress path was as shown in Figure 2.35.
• With the multi-cycle operations of soil placement and compaction during
construction, the soil surrounding the reinforcement maximum tension point
in each compaction lift was subjected to only one cycle of loading, as shown
in Figure 2.35.
Loading due to the weight of overlying soil layers plus some equivalent increase in
the stress state induced by the compaction operations is shown as paths 1 to 3 in
Figure 2.35. This is followed by unloading along paths 3 to 5 to the final residual
stress-state condition at the end of construction. Ehrlich and Mitchell (1994) noted
62
that by following this procedure, the stresses in each layer were calculated only once
and that each layer calculation was independent of the others.
The specific values of σ’z and σ’zc at point 3 in Figure 2.35 represent the maximum
stress applied to the soil at a given depth during the construction process. The
maximum past equivalent vertical stress, including compaction at the end of
construction (σ’zc), can be estimated using a new procedure, based on the method
given by Duncan and Seed (1986) for conventional retaining walls.
Figure 2.35: Assumed Stress Path (Ehrlich and Mitchell, 1994)
In Figure 2.35, the value of σ’zc can be estimated as the following:

σ xp '
σ zc ' = (2.18)
K0
where:
K 0 = 1 − sin φ ' (Jaky, 1944) (2.19)
63
0.5γ ' QN γ
σ xp ' = ν 0 (1 + K A ) (2.20)
L
and Q = maximum vertical operating force of the roller drum
L = length of the roller drum
γ = effective soil unit weight.
K0
υ0 = (Poisson’s ratio under K0-condition) (2.21)
1− K0
⎛ φ ' ⎞⎡ ⎛ φ' ⎞ ⎤
N γ = tan⎜ 45 o + ⎟ ⎢ tan 4 ⎜ 45 o + ⎟ − 1⎥ (bearing capacity factor) (2.22)
⎝ 2 ⎠⎣ ⎝ 2⎠ ⎦
Figure 2.36 shows the effects of CIS on compaction and reinforcement stiffness in
GRS walls. The conclusions drawn by Ehrlich and Mitchell from their study are:
- The soil shearing resistance parameters, the soil unit weight, the depth, the
relative soil-reinforcements stiffness index, Si, and compaction are the
major factors determining reinforcement tensions (typical Si for metallic
reinforcement: 0.500-3.200; plastic reinforcement: 0.030-0.120 and
geotextile reinforcement: 0.003-0.012);
- Increasing Si, usually means increased lateral earth pressure and
reinforcement tension, but at shallow depths the opposite effect can occur
depending on compaction conditions;
- The coefficient of horizontal earth pressure, K, can be greater than K0 at
the top of the wall and be greater than KA to depths of more than 6.1 m (20
ft) depending on the relative soil-reinforcements stiffness index, and the
compaction load; and
- K0 is the upper limit for the coefficient of horizontal earth pressure, K, if
there is no compaction of the backfill.
64
Figure 2.36: Compaction and Reinforcement Stiffness Typical Influence
(Ehrlich and Mitchell, 1994)
2.4.2 Hatami and Bathurst (2006)
Hatami and Bathurst (2006) noted that fill compaction has two effects on the soil: (1)
increase the lateral earth pressure, (2) reduce the effective Poisson’s ratio. They
suggested that the first effect can be modeled in a numerical analysis by applying a
uniform vertical stress (8 kPa, and 16 kPa depending on compaction load) to entire
surface of each newly placed soil layer before analysis and removed it afterwards.
This procedure was based on a recommendation by Gotteland et al. (1997).
Gotteland et al. simulated the compacting effect by loading and unloading of a
uniform surcharge of 50 kPa and 100 kPa on the top of the wall.
65
For the second effect of compaction on the reduction of Poisson’s ratio, Hatami and
Bathurst used the numerical simulation to find νmin from matching measured and
analysis data. The results (wall lateral movement and reinforcement forces) obtained
from the numerical analysis including compaction effect were in a very good
agreement with the measured data.
In their numerical analyses, the compaction effects were also account for by
increasing the elastic modulus number, Ke value from triaxial test results by a factor
of 2.25 for Walls 1 and 2. In other words, the elastic modulus was increased by the
factor of 2.25 for Walls 1 and 2.
2.4.3 Morrison et al. (2006)
Morrison et al. (2006) simulated the effects of fill compaction of shored mechanically
stability earth (SMSE) walls. A 50 kPa inward pressure was applied to the top,
bottom and exposed faces of each lift to simulate the effects of fill compaction. The
inward pressure was then reduced to 10 kPa on the top and bottom of a soil lift prior
to placement of the next lift to simulate vertical relaxation or unloading following
compaction. The inward pressure acting on the exposed face was maintained at 50
kPa as this produced the most reasonable model deformation behavior compared with
that observed in the field-scale test. They considered that the inward maintained
pressures are "locking-in" stresses in soil due to compaction.
The stiffness of soil was increased by the factor of ten (10) to consider the
compaction-induced stresses in the GRS mass. This factor in Hatami and Bathurst
(2006) was about 2.25.
66
Figure 2.37 shows the model for finite element analysis by Morrison et al. (2006).
The figure shows the simulation of fill compaction of lift 5 by the applying uniform
pressures.
Figure 2.37: Finite Element Model for Finite Element Analysis

(Morrison et al., 2006)
67
The results of the lateral movement and reinforcement forces showed a good
qualitative agreement with the measured data. However, general application of the
procedure may be questionable because neither the method of analysis nor the
magnitude of the applied inward pressure was properly justified.
2.5 Highlights on Compaction-Induced Earth Pressures in the Literature
A number of important highlights regarding compaction-induced lateral earth

pressures in the literature are summarized below:
1. Compaction of soil against a rigid, vertical, non-yielding structure appears to
result in the following residual lateral pressure distribution: (a) the lateral
pressures near the surface increase rapidly with depth, exceeding the at-rest
value, but limited a passive failure, (b) at intermediate depths, the lateral
pressures exceed the at-rest value, increase less rapidly with depth or remain
fairly constant with depth, and (c) at greater depths, the lateral pressures
appear to be the simple at-rest pressures, showing no affects of compaction
(Broms, 1971, Seed, 1983).
2. Compaction of soil against deflecting structures appears to increase near
surface.
3. The compaction-induced residual earth pressures are significant affected by
the compaction equipments. For compaction by small hand-operated rollers,
the increase in the lateral pressure occurs within a depth of about 3 to 4m, but
for very large rollers, the effect of compaction can be up to 15 to 25 m
(Duncan and Seed, 1986).
4. Structural deflections away from the soil, which occur during fill placement
and compaction, will reduce the residual lateral earth pressures. Reduction in
pressures appears to occur more rapidly in heavily compacted cohesionless
soil (Seed, 1983).
68
5. Compaction-induced residual lateral earth pressures in cohesive soils appear
to dissipate with time, even against non-deflecting structures, and eventually
approach at-rest values.
6. There is some evidence suggesting that the direction of rolling with the
compactor can have a significant effect in compaction-induced earth pressures
(Erhlich and Mitchell, 1993).
7. Field observations indicate that available overburden pressures are sufficient
that possible passive failure does not limit residual lateral earth pressure, a
high percentage of the peak lateral earth pressures induced during compaction
may be retained as residual pressures. In previously compacted soil, however,
additional compaction can result in only small increases in peak pressures, and
a negligible fraction of this (Aggour and Brown, 1974).
8. A number of simulation models have been proposed to explain and to evaluate
the residual lateral earth pressures induced by compaction. Common to all of
these theories is the idea that compaction represents a form of over-
consolidation wherein stresses resulting from a temporary or transient loading
condition are retained to some extent following removal of this peak load.
9. Many researchers, including Rowe (1954), Broms (1971), Gotteland et al.
(1997), and Hatami and Bathurst (2006), have simulated fill compaction by
application and removal of a surficial surcharge pressure.
10. Broms (1971) proposed a theory to calculate compaction-induced residual
lateral earth pressures against a rigid, vertical, frictionless, non-yielding wall.
The simulation results somewhat agree with available field data for walls
sustaining minimal deflections. Broms assumed that: (a) unloading results in
no decrease in lateral stress until a limiting passive failure-type condition is
reached, and (b) reloading results in no increase in lateral stress until the
virgin K0-loading stress path is regained. This type of model does not predict
well the peak lateral stresses induced by fill compaction, and is not suited for
69
computing lateral stresses induced by a surficial compaction plant of finite
lateral dimensions (not entire the surface). But Broms’ theory is very easy to
apply. Some researchers have adopted this theory for analysis of GRS
structures, including Gotteland et al. (1997), Hatami and Bathurst (2006), and
Morrison et al. (2006).
11. Seed (1983) developed two models for simulation of fill compaction: a non-
linear model and a bi-linear model. They are well suited for simulation of
compaction operation in GRS structures. The simulation results of the two
models are rather similar, and both agree well with measured data of
unreinforced earth retaining walls. The bi-linear model is easy to apply by
using hand calculation. Both models, bi-linear and non-linear models, are
based on the K0 condition. These are very useful to estimate CIS for soil only.
To use it for GRS structures, in-depth studies need to be carried out.
70
3. AN ANALYTICAL MODEL FOR CALCULATING LATERAL
DISPLACEMENT OF A GRS WALL WITH MODULAR BLOCK
FACING
Over the past two decades, Geosynthetic-Reinforced Soil (GRS) walls have gained
increasing popularity in the U.S. and abroad. In actual construction, GRS walls have
demonstrated a number of distinct advantages over the conventional cantilever and
gravity retaining walls. GRS walls are generally more ductile, more flexible (hence
more tolerant to differential settlement and to seismic loading), more adaptable to
low-permeability backfill, easier to construct, require less over-excavation, and
significantly more economical than conventional earth structures (Wu, 1994; Holtz
et al., 1997; Bathurst et al., 1997).
A GRS wall comprises two major components: a facing element and a GRS mass.
The facing element of a GRS wall have been constructed with different types of
material and in different forms, including wrapped geotextile facing, timber facing,
modular concrete block facing, precast concrete panel facing, and cast-in-place rigid
facing. Among the various facing types, modular concrete block facing has been
most popular in North America, mainly because of its ease of construction, ready
availability, and lower costs. The other component of a GRS wall, a GRS mass,
however, is always a compacted soil mass reinforced with layers of geosynthetic
reinforcement. Figure 3.1 shows the schematic diagram of a typical GRS wall with
modular block facing.
Current design methods for GRS walls consider only the stresses and forces in the
wall system. Even though a GRS wall with modular block facing is a fairly “flexible”
71
wall system, movement of the wall is not accounted for in current designs. A number
of empirical and analytical methods have been proposed for estimating lateral
movement of GRS walls. Most these methods, however, do not address the rigidity
of the facing although many full-scale experiments, numerical analysis, and field
experience have clearly indicated the importance of facing rigidity on wall movement
(e.g., Tatsuoka, et al., 1993; Rowe and Ho, 1993; Helwany et al., 1996; Bathurst et
al., 2006).
The prevailing methods for estimating the maximum lateral displacement of GRS
walls include: the FHWA method (Christopher, et al., 1989), the Geoservices method
(Giroud, 1989), the CTI method (Wu, 1994), and the Jewell-Milligan method (1989).
Among these methods, the Jewell-Milligan method has been found to give the closest
agreement with finite element analysis (Macklin, 1994). The Jewell-Milligan
method, however, ignores the effect of facing rigidity. Strictly speaking, the method
is only applicable to reinforced soil walls where there is little facing rigidity, such as a
wrapped-faced GRS wall.
A study aiming at developing an analytical model for calculating lateral movement of

a GRS wall with modular block facing was undertaken. The analytical model
modifies the Jewell-Milligan method (1989) to include the rigidity of facing element.
To verify the analytical model, the lateral wall displacements calculated by the
analytical model were compared with the results of the Jewell-Milligan method
(1989) for GRS walls with negligible facing rigidity. In addition, the lateral wall
displacements obtained from the analytical model were compared with the measured
data of a full-scale experiment of GRS wall with modular block facing (Hatami and
Bathurst, 2005 & 2006).
72
In addition to lateral displacement profiles, an equation for determining facing
connection forces (i.e., the forces in reinforcement immediately behind the facing) is
introduced.
Facing block
q
Reinforcement
Sv Soil
L.
Figure 3.1: Basic Components of a GRS Wall with a Modular Block Facing.
3.1 Review of Existing Methods for Estimating Maximum Wall Movement
The most prevalent methods for estimating the maximum lateral displacement of
GRS walls are the FHWA method (Christopher, et al., 1989), the Geoservices method
(Giroud, 1989), the CTI method (Wu, 1994), and the Jewell-Milligan method (1989).
A summary of each method is presented below.
73
3.1.1 The FHWA Method (Christopher, et al., 1989)
The FHWA method correlates L/H ratio (L = reinforcement length, H = wall height)
with the lateral displacement of a reinforced soil wall during construction. Figure 3.2
shows the relationship between L/H and δR, the empirically derived relative
displacement coefficient. Based on 6 m high walls, the δR value is to increase 25%
for every 20 kPa of surcharge. For the higher walls, the surcharge effect may be
greater. The curve in Figure 3.2 has been approximated by a fourth-order polynomial
as:
L
For 0.3 ≤ ≤ 1.175 ,
H
4 3 2
⎛L⎞ ⎛L⎞ ⎛L⎞ ⎛L⎞
δ R = 11.81⎜ ⎟ − 42.25⎜ ⎟ + 57.16⎜ ⎟ − 35.45⎜ ⎟ + 9.471 (3.1)
⎝H⎠ ⎝H⎠ ⎝H⎠ ⎝H⎠
For extensible reinforcement, the maximum lateral wall displacement, δmax, can be
calculated from by the following equation (δmax is in units of H):
⎛H⎞
δ max = δ R ⎜ ⎟ (3.2)
⎝ 75 ⎠
74
Figure 3.2: Empirical Curve for Estimating Maximum Wall Movement During
Construction in the FHWA Method (Christopher, et al., 1989)
75
The FHWA method was developed empirically by determining a displacement
“trend” from numerical analysis and adjusting the curve to fit with field measured
data. The method provides a quick estimate of the maximum lateral displacement.
Note that the maximum lateral displacement, δR, as obtained from Figure 3.2 has been
corrected for the wall with different height and surcharge.
3.1.2 The Geoservices Method (Giroud, 1989)
The Geoservices method relies on limit-equilibrium analyses to calculate the length

of the required reinforcement to satisfy a suggested factor of safety with regard to
three presumed external failure modes (e.g., bearing capacity failure, sliding and
overturning). The method provides a procedure for calculating the lateral wall
displacement.
The lateral displacement is calculated by first choosing a strain limit for the
reinforcement. This strain limit is usually less than 10 % and will depend on a
number of factors such as the type of wall facing, the displacement tolerances and the
type of geosynthetic to be used as reinforcement. Concrete facing panels, for
example, would not allow much lateral displacement without showing the signs of
distress. Therefore a low strain limit (1 to 3 %) should be selected.
Geosynthetics have a wide range of material properties depending on, among other
factors, the way they are manufactured. Non-woven geotextile exhibits low modulus
characteristics and if chosen as reinforcement for a wall, design would necessarily
imply that a large design strain is to be considered.
76
Once the strain limit has been selected, the method then assumes a distribution of
strain in the reinforcement, as shown in Figure 3.3 for calculating wall movement.
The horizontal displacement, δh, then becomes:
εd L
δh = (3.3)
2
where εd = strain limit (εmax), and L = reinforcement length.
εmax
Strain Distribution
Rankine Surface
Reinforcement
Φ/2
Figure 3.3: Assumed Strain Distribution in the Geoservices Method
3.1.3 The CTI Method (Wu, 1994)
Differing from all other design methods based on ultimate-strength of the

geosynthetic reinforcement, the CTI method is a service-load based design method.
77
The requirements of reinforcement are made in terms of stiffness at a design limit
strain as well as the ultimate strength.
In most cases, the designer will select a design limit strain of 1% to 3% for the
reinforcement. The maximum lateral displacement of a wall, δmax, can be estimated
by the following empirical equation:
⎛ H ⎞
δ max = ε d ⎜ ⎟ (3.4)
⎝ 1.25 ⎠
where εd = design limit strain (typically 1 % to 3 % for H ≤ 30 ft) and H =
wall height.
If the maximum wall displacement exceeds a prescribed tolerance for the wall, a
smaller design limit strain should be selected so that the maximum lateral
displacement of the wall will satisfy the performance requirement. Equation 3.4
applies only to walls with very small facing rigidity, such as wrapped-faced walls.
Walls with significant facing rigidity will have smaller maximum lateral
displacement. For example, a modular block GRS walls will have δmax about 15%
smaller than that calculated Equation 3.4.
3.1.4 The Jewell-Milligan Method
Jewell (1988) and Jewell and Milligan (1989) proposed a procedure for calculating
wall displacement based on analysis of stresses and displacements in a reinforced soil
mass. The method describes a link between soil stresses (stress fields) in a reinforced
soil mass in which a constant mobilized angle of friction is assumed with the resulting
displacements (velocity fields). There are two parameters for plane-strain plastic
deformation of soil: the plane strain angle of friction, φps, and the angle of dilation, ψ.
78
The planes on which the maximum shearing resistance φps is mobilized are called the
“stress characteristics” and are inclined at (45 0 + φ ps / 2 ) to the direction of major
principal stress, as shown in Figure 3.4(a). The directions along which there is no
linear extension strain in the soil are called the “velocity characteristics” and are
inclined at (45 0 + ψ / 2) to the direction of major principal stress, as shown in Figure
3.4(b).
Figure 3.4: (a) Stress Characteristics and (b) Velocity Characteristics behind a
Smooth Retaining Wall Rotating about the Toe (Jewell and Milligan, 1989).
Jewell and Milligan (1989) noted from limiting equilibrium analyses that there are
three important zones in a reinforced soil wall, as illustrated in Figure 3.5(a). The
boundary between zone 1 and 2 is at an angle (45 0 + ψ / 2) to the horizontal, and
between zone 2 and 3 at an angle φds. Large reinforcement forces are required in zone
1 to maintain stability across a series of critically inclined planes. In zone 2, the
required reinforcement forces reduce progressively.
79
The assumptions of the Jewell-Milligan method for "ideal length" of reinforcement
are:
• The reinforcement length at every layer extends to the back of zone 2, so
called "ideal length".
• The horizontal movement of the facing may be calculated by assuming the
horizontal deflections starting at the fixed boundary between zones 2 and 3
and working to the face of the wall.
• The stability on the stress characteristics and the velocity characteristics is
equally critical in soil and hence reinforcement must provide equilibrium for
both. The consequence is that behind the Rankine active zone in a reinforced
soil wall, the equilibrium is governed by φds mobilized on the velocity
characteristics.
Figure 3.5: Major Zones of Reinforcement Forces in a GRS Wall and the Force
Distribution along Reinforcement with Ideal Length (Jewell and Milligan, 1989).
80
In Figure 3.5, the maximum horizontal resultant force required for equilibrium, Prm, is
equal to the active force Pa:
⎛ γH 2 ⎞
Prm = Pa = K a ⎜⎜ + q s H ⎟⎟ (3.5)
⎝ 2 ⎠
in which, γ = unit weight of the soil; H = wall height; qs = uniform surcharge;

Ka = active earth pressure coefficient that can be expressed as:
⎛ ψ ⎞
tan⎜ 45 + − φds ⎟
(1 − sin φ ps ) ⎝ 2 ⎠ (3.6)
Ka = =
(1 + sin φ ps ) ⎛ ψ⎞
tan ⎜ 45 + ⎟
⎝ 2⎠
The required reinforcement force Pr in zone 2 at an angle θ, as shown in Figure

3.5(b), can be estimated from the maximum reinforcement force Prm as:
Pr tan(θ − φ ds )
= (3.7)
Prm K a tan θ
The results of the displacement analyses have been presented in the form of design
charts as shown in Figure 3.6. The charts can be used to determine the distribution of
lateral wall displacement along the wall face for different values of mobilized internal
friction φps and angles of dilation ψ.
To estimate the horizontal deflection at the GRS wall face for reinforcement with the
ideal length and uniform spacing, the charts as shown in Figure 3.6 can be used. The
horizontal deflection at the wall face depends on the wall height H, the mobilized soil
shearing resistance φds, the reinforcement force Pr, and the reinforcement stiffness K.
81
δhK
The charts in Figure 3.6 can be used to obtain a dimensionless factor, , and
HPbase
then the horizontal displacement δh can be calculated from this factor. The
reinforcement occurs at the base of the wall, Pbase, in the dimensionless factor above,
can be calculated by
Pbase = K a s v (γ H + q s ) (3.8)
82
Figure 3.6: Charts for Estimating Lateral Displacement of GRS Walls with the
Ideal Length Layout (Jewell and Milligan, 1989)
83
3.2 Developing an Analytical Model for Calculating Lateral Movement and
Connection Forces of a GRS Wall
Jewell and Milligan (1989) have presented design charts for estimating deformation
of reinforced soil walls where the rigidity of the facing can be ignored. Ho and Rowe
(1997) and Rowe and Ho (1993, 1998) pointed out that there is little variation in the
reinforcement forces and the lateral wall deformation when the reinforcement length
to wall height ratio, L/H, is equal to or greater than 0.7 (Note: L/H = 0.7 is commonly
used in practice and it is also suggested by the AASHTO). Based on a series of
numerical analyses of GRS walls, Rowe and Ho (1998) also showed that the
maximum lateral deformation obtained by the Jewell-Milligan method with an ideal
reinforcement length, as defined by Jewell and Milligan (1989), is generally in good
agreement with the numerical results for L/H = 0.7. For this reason, the Jewell-
Milligan method, with the ideal reinforcement length, can be used to estimate lateral
movement of a reinforced soil wall with L/H ≥ 0.7. The analytical model developed
in this study for GRS walls with modular block facing was base on Jewell-Milligan
method. The method is applicable to GRS walls with L/H ≥ 0.7.
The derivation of the analytical model is given below. It begins with the derivation of
the equations in the Jewell-Milligan method for predicting deformation of a
reinforced soil wall with negligible facing rigidity, followed by the derivation of the
equations for determining connection forces in the reinforcement for walls with
modular block facing, and ends with the equations for calculating lateral movement
of GRS walls with modular block facing.
84
3.2.1 Lateral Movement of GRS Walls with Negligible Facing Rigidity
Figure 3.7 shows the three major zones in a GRS wall and the force distribution in the
reinforcement at depth zi, used by Jewell and Milligan (1989) to develop an analytical
model for determination of wall deformation. Jewell and Milligan (1989) have
presented design charts based on the analytical model (without giving the derivation).
The following derivation is presented for completeness and for easier reference when
showing the derivation of the analytical model developed for this study.
x
45 + ψ/2 φds
Zone 1
z
zi Zone 2
Pr m
Zone 3
H
Lzone-1 Lzone-2
Reinforcement
(H - z i) Li
Figure 3.7: Major Zones of the Reinforcement Force in a Reinforced Soil Wall
(Jewell and Milligan, 1989)
The horizontal movement, Δh, of the wall face at depth zi can be evaluated as:
Δ h = Δ zone−1 + Δ zone−2 (3.9)
Lzone −1
Prm P
Δ zone −1 = ∫0
K re inf
dx = rm L zone −1
K re inf
(3.10)
85
Lzone −1 + Lzone − 2
Pr ⎛1⎞ P
Δ zone − 2 = ∫
Lzone −1
K re inf
dx ≈ ⎜ ⎟ rm L zone − 2
⎝ 2 ⎠ K re inf
(3.11)
where Kreinf stiffness of the reinforcement

Prm maximum reinforcement force at depth zi
Lzone-1 reinforcement length in Zone 1 at depth zi
Lzone-2 reinforcement length in Zone 2 at depth zi
Substituting Equation 2.10 and Equation 3.11 into Equation 3.9, we get
Prm ⎛ 1 ⎞
Δh = ⎜ Lzone −1 + L zone − 2 ⎟ (3.12)
K re inf ⎝ 2 ⎠
Since
⎛ ψ⎞
L zone −1 = (H − z i ) tan⎜ 45 o − ⎟ (3.13)
⎝ 2⎠
and
⎡ ψ ⎞⎤
( ⎛
)
L zone −2 = (H − z i ) ⎢ tan 90 o − φ ds − tan⎜ 45 o − ⎟⎥ (3.14)
⎣ ⎝ 2 ⎠⎦
Therefore, substituting Equation 3.13 and Equation 3.14 into Equation 3.12 leads to
Prm ⎧ ⎛ o ψ⎞ 1 ⎡ ⎛ o ψ ⎞⎤ ⎫
Δh = (
⎨(H − z i ) tan⎜ 45 − ⎟ + (H − z i ) ⎢ tan 90 − φ ds − tan⎜ 45 − ⎟⎥ ⎬
o
)
K re inf ⎩ ⎝ 2⎠ 2 ⎣ ⎝ 2 ⎠⎦ ⎭
(3.15)
Rearranging Equation 3.15, we have
⎧ ⎫
(H − z i ) ⎨ tan⎛⎜ 45 o − ψ ⎞⎟ + 1 ⎡⎢ tan 90 o − φ ds − tan⎛⎜ 45 o − ψ ⎞⎟⎤⎥ ⎬
( )
Prm
Δh = (3.16)
K re inf ⎩ ⎝ 2⎠ 2 ⎣ ⎝ 2 ⎠⎦ ⎭
or
⎧
Δh =
Prm
(H − z i ) ⎨ 1 ⎡⎢ tan⎛⎜ 45 o − ψ ⎞⎟ + tan 90 o − φ ds ( )⎤⎥ ⎫⎬ (3.17)
K re inf ⎩ 2 ⎣ ⎝ 2 ⎠ ⎦⎭
86
The value of Δh, lateral displacement of a GRS wall at depth zi, can be calculated
directly from the following equation:
⎛1⎞⎛ P ⎞
⎟⎟ (H − z i )
⎡ ⎛ 0 ψ⎞ ⎤
Δ h = ⎜ ⎟ ⎜⎜ rm ⎢ tan⎜ 45 − 2 ⎟ + tan(90 − φ ds )⎥
0
(3.18)
⎝ 2 ⎠ ⎝ K re inf ⎠ ⎣ ⎝ ⎠ ⎦
where Kreinf stiffness of the reinforcement
Prm maximum reinforcement force at depth zi
H wall height
φds effective direct shear friction angle of soil
ψ angle of dilation of soil
Prm reinforcement force in zone 1 or connection force (will be
discussed in Section 2.2)
3.2.2 Connection Forces for GRS walls with Modular Block Facing
The connection forces, Prm in Equation 3.18, are defined as the forces in the
reinforcement at the back face of the wall facing. The assumptions made for the
determination of the connection forces include:
• The wall face is vertical or nearly vertical.
• There is only friction connection between adjacent facing blocks (i.e., there is
no additional mechanical connection elements, such as lips, keys, or pins).
• A uniform surcharge is being applied over the entire horizontal crest of the
wall.
• Each facing block is a rigid body, i.e., movement is allowed, but not the
deformation. As a facing block moves, the frictional resistance, F, between
two adjacent blocks will reach Fmax.
87
Consider the reinforcement at depth zi, sandwiched between two adjacent facing
blocks, as shown in Figure 3.8. The frictional forces above and below these blocks
are Fi-1 and Fi+1. The horizontal resultant force of lateral earth pressure acting on the
two facing blocks is Pi. The tensile connection force in the reinforcement at depth zi
is Ti.
Top of the wall
Ni-1
Zi
± Fi-1
Ffi
reinforcement
Wi
Ti
Sv
Pi
± Fi+1
Ni+1
b
Figure 3.8: Forces Acting on Two Facing Blocks at Depth zi
The tensile connection force Ti in the reinforcement will be:

Ti = Pi ± ( Fi +1 − Fi −1 ) (3.19)
where Pi = K h (γ s z i + q )S v (3.20)
If Pi ± ( Fi +1 − Fi −1 ) < 0, Ti should be set equal to 0, as geosynthetic reinforcement can
resist only tensile forces, i.e., Ti is always ≥ 0.

From Figure 3.8,
Ti = Pi ± ( N i +1 − N i −1 ) tan δ ; Ti = 0 if Ti ≤ 0 (3.21)
where Ni+1 and Ni-1 are normal forces on the top and the bottom of the two adjacent
blocks.
Ti = Pi ± (Wi + F fi ) tan δ (3.22)
88
or,
Ti = Pi ± (γ b bS v + pS v tan β ) tan δ (3.23)
where
Ffi frictional resultant force between wall facing and soil
γb unit weight of facing block
b width of facing block
Sv reinforcement spacing
δ friction angle between modular block facing elements (δ can be the
friction angle between facing blocks if there is no reinforcement
between the blocks, or it can be the friction angle between facing
block and geosynthetic if there is reinforcement sandwiched between
blocks)
β friction angle between back face of wall and soil
p average net earth pressure acting on the facing, due to earth pressure
on the facing and the pressure caused by the reinforcement force. The
value of p can be estimated as:
F i +1− Fi −1
p= = γ bb tan δ (3.24)
Sv
Substituting Equation 3.24 into Equation 3.23, Ti , the connection force at depth zi,
can be determined as:
Ti = Pi ± (γ b bS v + γ b bS v tan δ tan β ) tan δ (3.25)
or,
Ti = K h (γ s z i + q )S v ± (γ b bS v + γ b bS v tan δ tan β ) tan δ (3.26)
Therefore, the tensile connection force in the reinforcement at depth zi can be
expressed as:
Ti = K h (γ s z i + q )S v ± γ b bS v tan δ (1 + tan δ tan β ) ; T = 0 if T ≤ 0 (3.27)
89
Note that if the friction between the back face of the wall facing and the soil behind
the wall is ignored, the connection force at depth zi becomes:
Ti = K h (γ s z i + q )S v ± (γ b bS v )(tan δ ) ; T = 0 if T ≤ 0 (3.28)
Calculating the reinforcement connection forces in a GRS wall with modular block
facing by a simple equation can reduce considerable amount of work for designing a
GRS wall. To design GRS walls with segmental facing, the connection forces are
concerned. A series of experimental tests to measure the connection forces were
conducted from many researchers. According to their results, if the blocks are heavy
and well connected, the GRS walls with block facing would perform very well
(Hatami and Bathurst, 2005 and 2006). A simple way to estimate the connection
forces of reinforcement in the GRS walls with block facing will be presented. The
values of the connection forces can be estimated by using the equation below for a
typical GRS wall with modular block facing:
Ti = K h (γ s z i + q ) S v − (γ b bS v )(tan δ ) (3.29)
The resistant connection force at depth zi at facing can be estimated by:
Fr = 2γ b bz i (tan δ ) (3.30)
The comparison of connection forces and maximum friction capacity at facing blocks
are shown in Figure 3.9 and 3.10.
90
Connection force
Connection force (kN)

0 5 10 15 20 25 30
0
0.5
Proposed
1 method
Depth (m)
1.5
2
Maximum
2.5 friction
capacity at
3 facing
3.5
4
Figure 3.9: Connection Forces in Reinforcement (q = 0)

(The Data of the Wall are from Section 3.3.1 with ψ = 150)
Connection force
Connection force (kN)

0 5 10 15 20 25 30 35 40
0
0.5
Proposed
1 method
Depth (m)
1.5
2
Maximum
2.5 friction
capacity at
3 facing
3.5
4
Figure 3.10: Connection Forces in Reinforcement (q = 50 kN/m)

(The Data of the Wall are from Section 3.3.1 with ψ = 150)
91
Figures 3.9 and 3.10 show the values of connection forces with the maximum
connection friction forces with the surcharge of zero and 50 kN/m. For most of the
cases, the lightweight blocks (without strong key connections between block and
block) can be used for GRS walls. With the heavy surcharge, some blocks at the top
of the walls may be unstable as shown in Figure 3.10.
3.2.3 Lateral Movement of GRS Walls with Modular Block Facing
From Equation 3.18 and Equation 3.29, the displacement of a GRS wall with modular
block facing at depth zi can be determined by:
⎛ K (γ z + q )S v − γ b bS v tan δ (1 + tan δ tan β ) ⎞ ⎡ ψ ⎤
Δ i = 0.5 ⎜⎜ h s i ⎟⎟ (H − z i ) ⎢ tan ⎛⎜ 45 0 − ⎞⎟ + tan(90 0 − φ ds )⎥
⎝ K re inf ⎠ ⎣ ⎝ 2 ⎠ ⎦
(3.31)
Equation 3.31 is referred to as the analytical model in this study. When there is the
frictional resistance between the back face of the wall facing and the soil can be
ignored, the displacement of the wall at depth zi will reduce to:
⎛ K (γ z + q )S v − (γ b bS v )(tan δ ) ⎞ ⎡ ψ ⎤
Δ i = 0.5 ⎜⎜ h s i ⎟⎟ (H − z i ) ⎢ tan ⎛⎜ 45 0 − ⎞⎟ + tan(90 0 − φ ds )⎥
⎝ K re inf ⎠ ⎣ ⎝ 2⎠ ⎦
(3.32)
Note that the lateral movement calculated by Equation 3.32 will be slightly larger
than those calculated by Equation 3.31.
92
3.3 Verification of the Analytical Method
To verify the analytical model developed in this study, the model calculation results
were first compared with the Jewell-Milligan method for GRS walls with negligible
facing rigidity. The model calculation results were then compared with measured
data from a full-scale experiment.
3.3.1 Comparisons with the Jewell-Milligan Method for Lateral Wall

Movement
The model calculation results of wall movement from the analytical model were first
compared with the results of the Jewell-Milligan method (1989) using an example.
The conditions of the wall in this example are as follows:
• Wall height H = 4.0 m.

• Geosynthetic reinforcement: vertical spacing Sv = 0.4 m, stiffness Kreinf = 200
kN/m.
• Backfill: free-draining granular soil, unit weight γs = 18 kN/m3, cohesion c =
0, friction angle φds = 35o, dilation angle ψ = 5o.
• Facing: modular concrete blocks, width b = 0.3 m, unit weight γb = 0, 20, and
30 kN/m3.
• Interface between adjacent blocks: cohesion, c = 0; friction angle, δ = 25o.
• Interface friction between back face of facing and soil, β = 0o.
The Results of the comparisons between the analytical model and the Jewell-Milligan
method are showed in Figures 3.11 to 3.14 for different weights of facing blocks (unit
weight of block γb = 0, 10, 20, and 30 kN/m3, respectively), each with different values
93
of surcharge pressure (q = 0, 10, and 50 kN/m; Note: q = 10 kN/m represents a typical
surcharge for highway design). These Figures indicate that, for all facing conditions
(i.e., for different values of γb), the effect of surcharge on wall movement is very
significant. The deformed shape of the wall face is rather different at q = 0 and q =
50 kN/m. At q = 0, the wall bulges at the mid-height, but as the surcharge increases,
the top of the wall face begins to move outward. At q = 50 kN/m, the largest wall
movement occurs near the top of the wall.
As can be seen from Figure 3.11, when the facing block is weightless, i.e., unit
weight of facing block γb = 0, the analytical model give nearly identical lateral wall
movement to the Jewell-Milligan method. This is to be expected because facing
rigidity is ignored in the Jewell-Milligan method. The Figures show that as the facing
blocks becomes heavier (i.e., as the unit weight of the blocks increases), the wall
movement becomes smaller. When a heavy facing block (γb = 30 kN/m3) is used, the
maximum lateral wall movement can be as much as 35% smaller than a wall with
negligible facing rigidity.
94
q=0
0 20 40 60
0 Lateral Displacem ent
0.4
Depth from the Top (m)

(m m )
0.8
1.2
1.6
2
2.4 Analytical model
2.8
3.2
Jew ell-Milligan
3.6
method
4
q = 10 kN/m
0 20 40 60
0.4
(m m )
0.8
1.2
1.6
2
2.8
3.2
Jew ell-Milligan
3.6
method
4
q = 50 kN/m
0 20 40 60
0.4
(m m )
0.8
1.2
1.6
2
2.8
3.2
Jew ell-Milligan
3.6
method
4
Figure 3.11: Comparison of Lateral Displacement Calculated by the Jewell-

Milligan Method and the Analytical Model, γb = 0 kN/m3
95
q=0
0 20 40 60
0.4

(m m )
0.8
1.2
1.6
2
2.8
3.2
Jew ell-Milligan
3.6
method
4
q = 10 kN/m
0 20 40 60
0.4
(m m )
0.8
1.2
1.6
2
2.8
3.2
Jew ell-Milligan
3.6
method
4
q = 50 kN/m
0 20 40 60
0.4
(m m )
0.8
1.2
1.6
2
2.8
3.2
Jew ell-Milligan
3.6
method
4

96
q=0
0 20 40 60
0.4

(m m )
0.8
1.2
1.6
2
2.8
3.2
Jew ell-Milligan
3.6
method
4
q = 10 kN/m
0 20 40 60
0.4
(m m )
0.8
1.2
1.6
2
2.8
3.2
Jew ell-Milligan
3.6
method
4
q = 50 kN/m
0 20 40 60
0.4
(m m )
0.8
1.2
1.6
2
2.8
3.2
Jew ell-Milligan
3.6
method
4

97
q=0
0 20 40 60
0.4

(m m )
0.8
1.2
1.6
2
2.8
3.2
Jew ell-Milligan
3.6
method
4
q = 10 kN/m
0 20 40 60
0.4
(m m )
0.8
1.2
1.6
2
2.8
3.2
Jew ell-Milligan
3.6
method
4
q = 50 kN/m
0 20 40 60
0.4
(m m )
0.8
1.2
1.6
2
2.8
3.2
Jew ell-Milligan
3.6
method
4

98
3.3.2 Comparisons with Measured Data of Full-Scale Experiment by Hatami
and Bathurst (2005 and 2006)
Figure 3.15 shows the configuration a full-scale experiment of GRS wall with
modular block facing. It is one of a series of laboratory experiments conducted by
Hatami and Bathurst (2005 and 2006), and was referred to as “wall 1” by the authors.
The parameters of the GRS wall are summarized below:
• Wall height H = 3.6 m high with a facing batter of 8° from the vertical and
seated on a rigid foundation.
• Soil: a clean uniform beach sand, γs = 16.8 kN/m3, φps = 44o, ψ = 11o, and c =
2 kPa.
• Geosynthetic reinforcement: a weak biaxial polypropylene (PP) geogrid,
vertical spacing = 0.6 m, reinforcement stiffness = 115 kN/m, ultimate
strength = 14 kN/m.
• Facing: solid masonry concrete blocks (300 mm wide by 150 mm high by 200
mm deep) with a shear key on the top surface of block, and γb = 20 kN/m3.
• Interface between facing blocks: δb-b = 57o and cb-b = 46 kPa.
Because the analytical model requires that the direct shear friction angle be used in
model calculations, and it also assumes a vertical wall face, the direct shear friction
angle of the soil and the facing batter factor were determined before using the
analytical model to evaluate the lateral movement of the facing.
(a) Direct shear friction angle:

sin φ ps cosψ sin 44 o cos11o
tan φ ds = = = 40 o (3.34)
1 − sin φ ps sinψ 1 − sin 44 o sin 11o
99
(b) Empirical facing batter factor, Φfb, from Allen and Bathurst (2001), with facing
batter of 8o:
d
⎛K ⎞
Φ fb = ⎜⎜ abh ⎟⎟ = 0.88 (3.35)
⎝ K avh ⎠
In the above equation, Kabh is the horizontal component of the active earth pressure
coefficient accounting for wall face batter, Kavh is the horizontal component of the
active earth pressure coefficient for a vertical wall, and d is a constant coefficient.
Allen and Bathurst (2001) found that the value of d = 0.5 would yield the best fit for
available Tmax data, and recommended using d = 0.5 for determination of Φfb.
Figure 3.15: Configuration of a Full-Scale Experiment of GRS Wall with Modular

Block Facing (Hatami and Bathurst, 2005 and 2006)
100
Figure 3.16 shows lateral movement of the GRS wall under surcharge pressures of 50
kN/m and 70 kN/m for: (a) mean value of measured displacement (Hatami and
Bathurst, 2006), (b) the Jewell-Milligan method, and (c) the analytical model (present
study). The Figure shows that the lateral movement calculated by the analytical
model is in very good agreement with the measured values. It is seen that that lateral
wall movement given by the analytical model agrees much better with the measured
values than that calculated by the Jewell-Milligan method for both surcharge
pressures. The lateral movement obtained by the Jewell-Milligan method was as
much as 3.5 times as large as the measured movement. Note that the analytical
model, not unlike the Jewell-Milligan method, produced a displacement profile that
had lower bending stiffness than the measured profile. However, considering that the
analytical model is a simplified model, the simulation is considered adequate.
101
q = 50 kN/m
0 30 60 90 Lateral
0 Displacem ent (m m )
0.4
0.8

1.2
1.6
Analytical
2 model
2.4 Jew ell-Milligan

method
2.8
3.2 Average
measured
3.6
q = 70 kN/m
0 30 60 90 Lateral
0 Displacem ent (m m )
0.4
0.8
1.2
1.6
Analytical
2 model
2.4 Jew ell-Milligan

method
2.8
3.2 Average
measured
3.6
Figure 3.16: Comparisons of Measured Lateral Displacements with Jewell-Milligan

Method and the Analytical Model
102
3.4 Summary
An analytical model has been developed to predict the profile of lateral movement of
a GRS wall with modular block facing. The connection forces in the reinforcement
can also be determined by a simple equation. The analytical model has been verified
through comparisons with Jewell-Milligan’s method. Jewell-Milligan’s method is
only a special case of the analytical model for GRS walls with negligible facing
rigidity. Comparisons were also made with a full-scale experiment of GRS wall with
modular block facing. It is shown that the analytical model offers a simple and
improved tool for predicting lateral movement of a GRS wall with modular block
facing.
103
4. THE GENERIC SOIL-GEOSYNTHETIC COMPOSITE (GSGC) TESTS
The understanding of soil-geosynthetic composite behavior in reinforced soil

structures has been lacking. As a result, current design methods have simply
considered the geosynthetics as added tensile elements, and have failed to account for
the interaction between soil and geosynthetics. A series of laboratory tests, referred
to as the “Generic Soil-Geosynthetic Composite” tests, or GSGC tests, were designed
and conducted to (a) examine the behavior of a generic soil-geosynthetic composite
with varying spacing and strength of reinforcement, (b) provide test data for verifying
the analytical model for calculating strength properties of a GRS composite as
described in Chapter 5, and (c) provide test data for calibration of Finite Element (FE)
model for a GRS mass. The GSGC tests were conducted at the Turner-Fairbank
Highway Research Center (TFHRC), Federal Highway Administration (FHWA) in
Mclean, Virginia.
4.1 Dimensions of the Plane Strain GSGC Test Specimen
A soil mass reinforced by layers of geosynthetic reinforcement is not a uniform mass.

To investigate the behavior of soil-geosynthetic composites by conducting laboratory
tests, it is necessary to determine the proper dimensions of the test specimen so that
the test will provide an adequate representation of soil-geosynthetic composite
behavior. Also, since most GRS structures resemble a plane strain condition, the test
specimen needs to be tested under a plane strain condition.
A number of factors were considered prior to determining the test specimen dimensions
of the GSGC test, including:
104
- Plane Strain Condition: As most GRS structures resemble a plane strain
condition, the test should be conducted in a plane strain condition.
- Backfill Particle Size: To alleviate the particle size effects on the test
specimen, the dimensions of a generic GRS mass should be at least 6 times as
large as the maximum particle size of the soil specimen, as suggested by the
U.S. Army Corps of Engineers, and 15 times larger than the average
particle size (D50) (Jewell, 1993). The recommended maximum particle
size for the backfill of GRS structures is 19 mm or ¾ in. (Elias and
Christopher, 1996). The specimen dimension, therefore, should be at least
120 mm.
- Reinforcement Spacing: The reinforcement spacing plays an important role
in the deformation behavior of GRS structures (Adams, 1999) and the load-
transfer mechanism of reinforced-soil masses (Abramento and Whittle,
1993). The height (H) of a generic GRS mass should be able to accommodate
the typical reinforcement spacing of 200 mm to 300 mm for GRS walls.
- Size of Reinforcement Sheet: The specimen dimensions in the plane strain
direction, referred to as the “width” (W) and in the longitudinal direction,
referred to as Length (L) should be sufficiently large to provide adequate
representation of the geosynthetic reinforcement. For polymer grids, enough
grid “cells” need to be included for a good representation of the polymer grid.
For nonwoven geotextiles, the aspect ratio of the reinforcement specimen
(i.e., the ratio of width to length) should be sufficiently large (say, greater
than 4) to alleviate significant “necking” effect. There will be little “necking”
effect for woven geotextiles, regardless of the aspect ratio.
A series of finite element analyses, using the computer code PLAXIS 8.2, were
conducted to examine the effect of specimen dimensions on the resulting global
stress-strain and volume change relationships of the composites. The objective of the
105
finite element analyses was to determine proper dimensions of a generic soil-
geosynthetic composite that will produce load-deformation behavior sufficiently close
to that of a large mass of soil-geosynthetic composite, referred to as the reference
composite.
Figure 4.1 shows the typical geometric and loading conditions of the GSGC tests.
The reference soil-geosynthetic composite is taken as a reinforced soil mass of
dimensions 7.0 m (23 ft) high and 4.9 m wide in a plane strain condition. Four
different dimensions of GRS composites were analyzed: specimen heights, H = 7.0
m, 2.0 m, 1.0 m and 0.5 m, while the width, W, of the test specimen was kept as
0.7*H. In these analyses, the soil was a dense sand. The sand was reinforced by a
medium-strength woven geotextile (Geotex 4x4) at 0.2 m vertical spacing. Table 4.1
lists the conditions and properties of the soil and reinforcement used in the analyses.
The global stress-strain curves obtained from the analyses are shown in Figures 4.2
and 4.4 for confining pressures, σc, of 0 and 30 kPa, respectively (note: a confining
pressure of 30 kPa is representative of the lateral stress at the mid-height of a 7.0 m
high wall). The corresponding global volume change curves are shown in Figures 4.3
and 4.5. The global vertical strain, εv, was calculated by the following equation:
⎛ ΔH ⎞
εv = ⎜ ⎟ 100% (4.1)
⎝ H ⎠
where ΔH = total vertical displacement of the specimen;

H = initial height of the specimen.
Figures 4.2 and 4.4 indicate that the composite with height H = 2.0 m, width W = 1.4
m, and under a confining pressure of 30 kPa yields stress-strain and volume change
relationships that are sufficiently close to those of the reference composite. Specimen
106
sizes with heights H = 1.0 m and 0.5 m appear too small for providing an adequate
representation of the reference composite.
σv
Reinforcement
Soil
σc σc
H
Sv = 0.2 m
W = 0.7 H
Figure 4.1: Typical Geometric and Loading Conditions of a GRS Composite
107
Table 4.1: Conditions and properties of the backfill and reinforcement used in
F.E. analyses
Description
A Dense sand: unit weight = 17 kN/m3; cohesion = 5 kPa;
Soil angle of internal friction, φ = 38o; angle of dilation, ψ = 8o; soil

modulus, E50 = 40,000 kPa; Poisson’s ratio = 0.3.
Reinforcement Geotex 4x4: axial stiffness, EA = 1,000 kN/m; ultimate

strength, Tult = 70 kN/m; reinforcement spacing = 0.2 m.
Confining Constant confining pressures of 0 and 30 kPa.

Pressure
108
2000
1500
Specimen Dimensio ns:

σ v -σ h (kPa)
1000
7.0 m x 4.9 m
2.0 m x 1.4 m
500
1.0 m x 0.7 m
0.5 m x 0.35 m
0
0 5 10 15 20 25
Global Vertical Strain εv (%)
Figure 4.2: Global Stress-Strain Curves for Soil-Geosynthetic Composites of

Different Dimensions under a Confining Pressure of 0 kPa
109
6
Specimen Dimensions:
Global Volumetric Strain Δ V/V (%)
4 7.0 m x 4.9 m
3
2.0 m x 1.4 m
2
1
1.0 m x 0.7 m
0
0 2 4 6 8 10 12 14 16
-1 0.5 m x 0.35 m
-2
Global Vertical Strain ε v (%)
Figure 4.3: Global Volume Change Curves for Soil-Geosynthetic Composites of

110
2000
1500
σv -σ h (kPa)
1000 7.0 m x 4.9 m
2.0 m x 1.4 m
1.0 m x 0.7 m
500
0.5 m x 0.35 m
0
0 5 10 15 20 25
Figure 4.4: Global Stress-Strain Curves for Soil-Geosynthetic Composites of

111
2.5
Global Volumetric Strain Δ V/V (%)
2
1.5 7.0 m x 4.9 m
1
2.0 m x 1.4 m
0.5
0
1.0 m x 0.7 m
0 2 4 6 8 10 12 14 16
-0.5
-1 0.5 m x 0.35 m
-1.5
Global Vertical Strain ε v (%)
Figure 4.5: Global Volume Change Curves for Soil-Geosynthetic Composites of

112
For comparison purposes, additional analyses were conducted on unreinforced soil.
Figures 4.6 and 4.7 show, respectively, the global stress-strain curves and global
volume change curves of the soil masses without any reinforcement for the different
specimen dimensions. The results indicate that the specimen height as small as 0.5 m
will yield nearly the same stress-strain and volume change relationships as the
reference soil mass of height = 7.0 m when reinforcement is not present.
113
140.0
120.0 Specimen Dimensions:
100.0
7.0 m x 4.9 m
σv-σh (kPa)
80.0
2.0 m x 1.4 m
60.0
40.0 1.0 m x 0.7 m
20.0
0.5 m x 0.35 m
0.0
0 1 2 3 4 5 6
Figure 4.6: Global Stress-Strain Curves of the Unreinforced Soil Under a

Confining Pressure of 30 kPa
114
1
Global Volumetric Strain Δ V /V (%)
0.8 Specimen Dimensions
0.6 7m x 4.9m
0.4
2m x 1.4m
0.2
1m x 0.7m
0
0 1 2 3 4 5 6
-0.2 0.5m x 0.35m
-0.4
Figure 4.7: Global Volume Change Curves of the Unreinforced Soil Under a
Confining Pressure of 30 kPa
115
Based on the results of the finite element analyses, a specimen height of 2.0 m, and
depth of 1.4 m, with 0.2 m reinforcement spacing, was selected as the specimen
dimensions for the GSGC tests. The actual specimen dimensions for the GSGC tests
are shown in Figures 4.8. Figures 4.9 and 4.10 show, respectively, the front and plan
views of the GSGC test setup.
Reinforcement
Sv = 0.2 m
H = 2.0 m
m
1 .2
L=
W = 1.4 m
Figure 4.8: Specimen Dimensions for the GSGC Tests
116
2.6 6.0
FRONT VIEW
(Note: All Dimensions in Inches)
A steel plate
(43''x34''x2'')
160.00
34.00
6.00
5.0 2.00
26.00 D=1.125''
6.00
30.75
2.625'' steel plate
Load cell LC1
steel plate
(1000 kip)
2.00
(26''x32.5''x2'')
2.63
12.25
5.62 16.93
steel plate
12.00
(43''x34''x2'')
25.19
51.98
jack
1.12 16.93 Load cell LC2
(300 kip)
2'' steel plate concrete pad
12.25 (49''x41 3/8''x12'')
1.94 4.00
1.50 9.63
49.00
12.00
134.63
concrete block
(7.625''x7.625''x15.625'')
81.65
76.25
70.00
30.00
33.00
7.63
3.00
53.375
2.87
11.38 70.00 11.38

74.00
74.25
36.00 108.00 36.00
1 A 2 3
Figure 4.9: Front View of the Test Setup
117
PLAN Anchor
(Note: All Dimensions in Inches)
Specimen Rigid frame

Open Face: Concrete blocks
(7.625''x7.625''x15.625'') would
47.00 be removed before testing
36.94
21.50
47.00
36.00
47.00
75.87
A
36.94
1.25
19.50 23.25 47.00
26.63 53.375 26.63

7.63
17.00 74.00 17.00
36.00 108.00 36.00
1 2 3
Figure 4.10: Plan View of the Test Setup
118
4.2 Apparatus for Plane Strain Test
To maintain a plane strain condition for the GSGC specimens throughout the tests,
two major factors were considered: (1) the test bin needs to be sufficiently rigid to
have negligible lateral deformation in the longitudinal direction (i.e. the length
direction, L), and (2) the friction between the backfill and the side panels of the test
bin needs to be minimized to nearly zero.
4.2.1 Lateral deformation
Five GSGC masses were tested inside a test bin. The test bin was designed to
experience little deformation for a surcharge pressure of up to 2,800 kPa. The test bin
is shown in Figure 4.1. The details of the test bin and the design calculations are
presented in Appendix B.
4.2.2 Friction
Two transparent plexiglass panels were attached inside steel tubing frame to form the
side surfaces of the test bin. To minimize the friction between the plexiglass and the
backfill in these surfaces, a lubrication layer was created inside surfaces of the
plexiglass panels. The lubrication layer consisted of a 0.5 mm-thick latex membrane
and an approximate 1 mm-thick lubrication agent (Dow Corning 4 Electrical
Insulating Compound NSF 6). This procedure has been successful in many plane
strain tests conducted by Professor Tatsuoka at University of Tokyo and Professor
Wu at University of Colorado Denver. The friction angle between the lubricant layer
and the plexiglass as determined by direct shear test was less than one (1) degree.
119
Figure 4.11: The Test Bin
120
4.3 Test Material
The backfill and geotextile reinforcement employed in the tests are described as
follows.
4.3.1 Backfill
The backfill was a crushed Diabase from a source near Washington D.C. Before
conducting the GSGC tests, a series of laboratory tests were performed to determine
the properties of the backfill, including:
• Gradation test
• Specific gravity and absorption test of the coarse aggregates
• Moisture-Density tests (Proctor compaction) with rock correction
• Large-size triaxial tests with specimen diameter of 152 mm (6 in.)
The details of these tests are presented in Appendix A. A summary of some index
properties is given in Table 4.2. The grain size distribution of the soil is shown in
Figure 4.12. Two gradation tests were performed. The results agree well with each
other.
Four triaxial tests were conducted at confining pressures of 5 psi, 15 psi, 30 psi, and
70 psi, and the results were compared with those performed by Ketchart et al. on the
same soil. The soil specimen was approximately 6 in. in diameter and 12 in. in
height. The shapes of a typical specimen before and after failure are shown in Figure
4.13. Figure 4.14 presents the stress-strain curves and volume change curves of the
tests. The stress-strain curves obtained by Ketchart et al. are also included for
comparisons and for furnishing a more complete set of data. The stress-strain
relationships agree well in trend with those by Ketchart et al. The Mohr-Coulomb
121
failure envelops of the backfill are shown in Figure 4.15. For confining pressures
between 0 and 30 psi, the strength parameters are: c = 10.3 psi, φ = 50o. For
confining pressures between 30 and 110 psi, the strength parameters are: c = 35.1 psi,
φ = 38o.
122
Table 4.2: Summary of some index properties of backfill
Classification Well graded gravel: A-1a, per AASHTO M-15; and

GW-GM, per ASTM D2487.
Maximum Dry Unit Weight 24.1 kN/m3 (153.7 lb/ft3)
Optimum Moisture Content 5.2 %
Specific Gravity of Soil 3.03

Solids
123
80
70
60
Percent finer (%)

50
40
30
20
10
0
100 10 1 0.1 0.01 Test 1
Grain size (m m ) Test 2
Figure 4.12: Grain Size Distribution of Backfill
124
(a) (b)
Figure 4.13: Typical Triaxial Test Specimen (a) before and (b) after Test
125
Deviatoric Stress versus Axial Strain Relationships
600
550
500
450 5 psi
Deviatoric Stress (psi)
400
15 psi
350
300 30 psi
250
200 70 psi
150
70 psi - Ketchart
100
50 110 psi - Ketchart
0
0 1 2 3 4 5 6 7 8 9 10 11
Axial Strain (%)
(a)
Volumetric Strain versus Axial Strain

0.6
Volumetric Strain (%)
0.4
0.2 5 psi
0
30 psi
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
-0.2
-0.4
Axial Strain (%)
(b)
Figure 4.14: Triaxial Test Results:

(a) Stress-Strain Curves of Backfill at 24.1 kN/m3 Dry Density and 5.2% Moisture
(b) Volume Change Curves of Backfill at Confining Pressures of 5 psi and 30 psi
126
σ
3 = 110 psi
Shear Streess (psi)
σ
3 = 70 psi
σ3 = 30 psi
σ
3 = 15 psi
C2 = 35.1 psi σ3 = 5 psi
C1 = 10.3 psi
0 30 98 172 265 433 616
Normal Stress (psi)
Figure 4.15: Mohr-Coulomb Failure Envelops of Backfill
127
4.3.2 Geosynthetics
The geosynthetic used in the experiments was Geotex 4x4 manufactured by Propex,
Inc. (formally known as Amoco 2044). This geosynthetic is a woven polypropylene
geotextile, with its strength properties provided by the manufacture shown in Table
4.3.
Table 4.3: Summary of Geotex 4x4 properties provided by the manufacture
Property Test Method Machine Direction Cross-Direction

(Wrap Direction) (Fill Direction)
Tensile Strength ASTM D-4632 2.67 kN 2.22 kN

(Grab)
(0.6 kips) (0.5 kips)
Wide-Width ASTM D-4595 70 kN/m 70 kN/m

Tensile Ultimate
(400 lb/in.) (400 lb/in.)
Strength
Wide-Width ASTM D-4595 21 kN/m 38 kN/m

Strength at 5%
(121 lb/in.) (217 lb/in.)
Strain
Wide-Width ASTM D-4595 10 % 10 %

Ultimate
Elongation
Puncture ASTM D-4833 0.8 kN (170 lb)
Trapezoid ASTM D-4533 1.11 kN (250 lb)

Tearing Strength
128
Two types of geosynthetics were used for the experiments: a single-sheet of Geotex
4x4, and a double-sheet Geotex 4x4 (by gluing two sheets of Geotex 4x4 together
using 3M Super 77 spray adhesive). The use of the double-sheet was to create a
geosynthetic that is approximately twice as stiff (and as strong), yet maintaining the
same interface condition as that of the single-sheet geosynthetic. Geotex 4x4
geotextile has been used in the construction of hundreds of production GRS walls and
in many full-scale experiments, including the FHWA GRS pier (Adams, 1997),
Havana Yard Test abutment and pier (Ketchart and Wu, 1997), Blackhawk preloaded
GRS bridge abutment (Ketchart and Wu, 1998), and NCHRP test abutments (Wu, et
al., 2006).
Uniaxial tension tests were performed on both types of geosynthetic to determine the
load-deformation behavior using specimen dimensions of Width = 305 mm (12 in.)
and Length = 152 mm (6 in.), see Figure 4.16. The stiffness and strength of the two
geosynthetics are shown in Table 4.4 and load-deformation curves are shown in
Figure 4.17. It is seen that the stiffness and the strength of the double-sheet Geotex
4x4 are approximately twice as much as those of the single-sheet Geotex 4x4, with
the breakage strain being almost the same.
129
Figure 4.16: Uni-Axial Tension Test of Geotex 4x4
130
Table 4.4: Properties of Geotex 4x4 in fill-direction
Geosynthetics Wide-Width Tensile Strength,
per ASTM 4595
Stiffness (kN/m) at Ultimate Strength (kN/m)

1% Strain (% at break)
Single-Sheet 1,000 70 (12 %)

Geotex 4x4
Double-Sheet 1,960 138 (12 %)

Geotex 4x4
131
160
140
120
Tensile Load (kN/m)
100
80
60
40
Single-Sheet
20
Double-Sheet
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Axial Strain (%)
Figure 4.17: Load-Deformation Curves of the Geosynthetics
4.3.3 Facing Block
Blocks used for the facing of the GSGC mass during specimen preparation were
hollow concrete blocks with the dimensions of 397 mm x 194 mm x 194 mm (15.625
in. x 7.625 in. x 7.625 in.) and with the average weight of 18.1 kg/block (40
lbs/block).
132
4.4 Test Program
The test program comprises five GSGC tests, with their test conditions shown in
Table 4.5. The plate compactor, MBW - GP1200, used for the tests has the following
specifications: weight = 120 lbs; plate dimensions = 12 in. x 21 in.; centrifugal force
= 1,500 lbf; rotation speed = 5000 vpm; and moving speed = 70 ft/min.
Table 4.5: Test program for the GSGC tests
Test Geosynthetic Confining Wide-Width Reinforcement

Designation Reinforcement Pressure Strength of Spacing, Sv
Reinforcement
Test 1 None 34 kPa None No

reinforcement
Test 2 Geotex 4x4 34 kPa T = 70 kN/m Sv = 0.2 m
Test 3 Double-Sheet 34 kPa T = 140 kN/m Sv = 0.4 m

Geotex 4x4
Test 4 Geotex 4x4 34 kPa T = 70 kN/m Sv = 0.4 m
Test 5 Geotex 4x4 0 T = 70 kN/m Sv = 0.2 m
(unconfined)
133
4.5 Test Conditions and Instrumentation
4.5.1 Vertical Loading System
The vertical loads were applied to the test specimens by using a 1,000,000-pound
capacity loading frame with a 1,000,000-pound hydraulic jack. Loads were measured
by load cells and by hydraulic jack pressure gauges. For Test 1, two load cells of
100,000 pounds and 300,000 pounds were used to measure the loads. For Tests 2 to
5, a 1,000,000 pound load cell was used to measure the loads. A 12 in.-thick concrete
pad was placed on top of the specimen before loading. Vertical loads were applied in
equal increments with ten-minute elapsed time between increments to allow time for
equilibrium. The elapsed time also allow manual recording of displacements of the
test specimen. The vertical loads were applied until a failure condition was reached
to determine the strength of the composite specimen. The applied pressures on the
composite specimens were determined from the applied vertical loads divided by the
surface areas of the composite specimens.
4.5.2 Confining Pressure
The confining pressure on the test specimens was applied by vacuuming. The entire
surface area of the test specimen was vacuum-sealed with a 0.5-mm thick latex
membrane. A prescribed confining pressure of 34 kPa (5 psi) for tests 1 to 4 was applied
by connecting the latex membrane to a suction device through two 6-mm diameter
flexible plastic tubes. Only Test 5 was conducted without confining pressure.
4.5.3 Instrumentation
The specimens were instrumented to monitor their performance during tests. The
instruments used include:
134
a. Vertical Movement
Three Linear Variable Displacement Transducers (LVDT) and two digital dial
indicators were installed on the top of the concrete pad to measure the vertical
movement of the specimen during loading. The vertical movement was measured
along the top surface of the concrete pad.
b. Lateral Movement
Ten LVDTs and two digital dial indicators were installed along the height of the
specimen (two open sides of the specimen) to measure the lateral movement of the
specimen. The location of the LVDTs and digital dial indicators are shown in Figure
4.18. Figure 4.37 shows the LVDTs on a test specimen to monitor the lateral
movement.
c. Internal Movement
The internal movement of the soil at selected points in the soil mass was traced by
marking the locations of pre-selected points on a 2 in. x 2 in. grid system drawn on
the membrane. The locations of the selected points for the tests are depicted in Figure
4.38.
135
Location of LVDTs and Dial Indicators
SIDE VIEW
FRONT VIEW
73.72 (1872.49)
72.25'' (1835.15) 47.37'' (1203.32)
Dial Indicator 1-1/4''
5.35''
LVDT LVDT LVDT

(59.49)
(59.49)
(59.49)
(264.16)(288.92)(288.92)(288.92)(288.92)(288.92)(288.92)
(264.16)(288.92)(288.92)(288.92)(288.92)(288.92)(288.92)
2.34
2.34
2.34
10.4 11.37 11.37 11.37 11.37 11.37 11.37
10.4 11.37 11.37 11.37 11.37 11.37 11.37

304.8 (332.1) (332.1) (297.4) (332.1)
304.8 (332.1) (332.1) (297.4) (332.1)
304.8 (332.1) (332.1) (297.4) (332.1)

13.0
13.0
13.0
Dial Indicator
81.65'' (2073.91)
81.65'' (2073.91)
Dial Indicator
11.0
11.0
11.0
76.3'' (1938)
76.3'' (1938)
Dial Indicator
13.0
13.0
13.0
2''
13.0
13.0
13.0
55.12'' (1400)
Center line
Cross Section A-A

72.25'' (1835.15)
TOP VIEW
55.12'' (1400) Footing on the top
5.35''
LVDT LVDT Dial Indicator on of GRS specimen

the top surface
(59.49)
(59.49)
2.34
2.34
LVDT on the top

surface
304.8 (332.1) (332.1) (297.4) (332.1)
304.8 (332.1) (332.1) (297.4) (332.1)
Dial Indicator LVDTs on the

13.0
13.0
Jack
Dial Indicator open side
81.65'' (2073.91)
6''
A
11.0
11.0
76.3'' (1938)
6''
GSGC MASS 4''
4''
13.0
13.0
Dial Indicator on
the open side
13.0
13.0
Dial Indicator on LVDT on the top surface (the

the top surface location can be adjusted
dependent on the size of the
loading system)
Figure 4.18: Locations of LVDTs and Digital Dial Indicators
136
d. Reinforcement Strain
To measure the strains in the geotextile, a number of high elongation strain gauge
(type EP-08-250BG-120), manufactured by Measurements Group, Inc., were used.
Each strain gauge was glued to the geotextile only at two ends to avoid inconsistent
local stiffening of geotextile due to the adhesive. The strain gauge attachment
technique was developed at the University of Colorado Denver. The gauge was first
mounted on a 25 mm by 76 mm patch of a lightweight nonwoven geotextile (see
Figure 4.19). A Microcystalline wax and a rubber coating (M-Coat B, Nitrile Rubber
coating) were used to protect the gauges from moisture. To check the effectiveness
of the moisture-protection technique, the geotextile specimens with the strain gauges
were tested after immersing in water for 24 hours. Before placing the reinforcement
sheet in the test specimen, an M-Coat FB-2, 6694, Butyl Rubber Tape was used to
cover the gauges to protect the gauges during compaction (see Figure 4.19). To
measure the strain distribution of the reinforcement, six strain gauges were mounted
on each Geotex 4x4 sheet (see Figure 4.20).
137
(a) (b)
Figure 4.19: Strain Gauge on Geotex 4x4 Geotextile:

(a) Before Applying Protection Tape
(b) After Applying Protection Tape (M-Coat FB-2, 6694, Butyl
Rubber Tape).
138
Figure 4.20: Strain Gauges Mounted on Geotex 4x4 Geotextile
139
Due to the presence of the light-weight geotextile patch, calibration of the strain
gauge is needed. The calibration tests were performed to relate the strain obtained
from the strain gauge to the actual strain of the reinforcement. Figures 4.21 and 4.22
show the calibration curves along the fill direction of Geotex 4x4 geotextile for the
single-sheet and the double-sheet specimens, respectively.
140
6
Strain from Instron Machine (%)
5
y = 1.172x
2
R = 0.9913
4
0
0 1 2 3 4 5
Strain from Strain Gage (%)
Figure 4.21: Calibration Curve for Single-Sheet Geotex 4x4
141
6
5
y = 1.078x
R2 = 0.9986
4
0
0 1 2 3 4 5
Strain from Strain Gauge (%)
Figure 4.22: Calibration Curve for Double-Sheet Geotex 4x4
142
4.5.4 Preparation of Test Specimen for GSGC Tests
The preparation procedure of a typical composite mass with the dimensions of 2.0 m
(H) x 1.4 m (D) x 1.2 m (L) is described as follows:
1. Mark the anticipated location of the GSGC mass on the plexiglass;
2. Apply approximately 1-mm thick lubricating agent (Dow Corning 4
Electrical Insulating Compound NSF 61) evenly on the inside surfaces
of the plexiglass (see Figure 4.23);
3. Attach a sheet of membrane (with 51 mm x 51 mm grid system pre-
drawn on membrane) over each plexiglass and at the bottom of the
specimen (see Figure 4.24);
4. Lay a course of the facing blocks on the open sides of the specimen (see
Figure 4.25);
5. Place the backfill in the test bin and compact in 0.2 m lifts (see Figures
4.26 and 4.27); check and adjust (if needed) the backfill moisture before
compaction to achieve the target moisture of 5.2%;
6. Check the water content and dry unit weight of each lift by using a
nuclear density gauge, Troxler 3440, by the direct transmission method
(note: the measured dry unit weights of five tests are shown in Figure
4.39);
7. Place the next layer of geosynthetic reinforcement (with strain gauges
already mounted) covering the entire top surface area of compacted fill
and the facing blocks (see Figure 4.28);
8. Repeat steps 4 to 8 until the full height of the composite mass is reached;
9. Sprinkle a 5 mm-thick fine sand layer over the top surface of the
completed composite mass to level the surface and protect the
membrane from being punctured by gravels in the backfill (see Figure
4.29);
143
10. Place a geosynthetic sheet on top of the composite mass (see Figure
4.30);
11. Glue a sheet of membrane to the top edge of the side membrane sheets
(see Figure 4.31);
12. Remove all facing blocks and trim off the excess geotextile (see Figure
4.32);
13. Insert strain gauge cables through the plastic openings that were already
attached on the membrane sheets at prescribed locations (see Figure
4.33);
14. Glue membrane sheets to enclose entire composite mass;
15. Apply vacuum to the composite mass at a low pressure of 14 kPa (see
Figure 4.34);
16. Seal the connection between cables and membrane with epoxy to
prevent air leaks (see Figure 4.35). The low vacuum pressure allows the
epoxy to seal the connection well;
17. Raise the vacuum pressure to 34 kPa and check air leaks under
vacuuming (see Figure 4.36) and measure the specimen dimensions (see
Table 4.6 for specimen dimensions of five tests).
144
Figure 4.23: Applying Grease on Plexiglass Surfaces
145
Figure 4.24: Attaching Membrane
146
Figure 4.25: Placement of the first Course of Facing Block
147
Figure 4.26: Compaction of the First Lift of Backfill
148
Figure 4.27: Placement of Backfill for the Second Lift
149
Figure 4.28: Placement of a Reinforcement Sheet
150
Figure 4.29: Completion of Compaction of the Composite Mass and Leveling the
Top Surface with 5 mm-thick Sand Layer
151
Figure 4.30: Completed Composite Mass with a Geotextile Sheet on the Top
Surface
152
Figure 4.31: Covering the Top Surface of the Composite Mass with a Sheet of
Membrane
153
Figure 4.32: Removing Facing Blocks and Trimming off Excess Geosynthetic
Reinforcement
154
Figure 4.33: Insertion of Strain Gauge Cables through the Membrane Sheet
155
Figure 4.34: Vacuuming the Composite Mass with a Low Pressure
156
Figure 4.35: Sealing the Connection between Cable and Membrane with Epoxy to
Prevent Air Leaks
157
Figure 4.36: Checking Air Leaks under Vacuuming
158
Figure 4.37: The LVDT’s on an Open Side of Test Specimen
159
(a): Tests 1 and 2 2x2'' Grid System
10.25
1 2 3
34.25
Heigth = 76.25 in. (Test 1), 76.35 in. (Test 2)
60.25
4 5 6
66.00
42.00
7 8 9
16.00
2.50
32.50 24.50
2.50
Width
(b): Tests 3,4 and 5 2x2'' Grid System

10.35
1 2 3
10 11
15.38
34.35
4 5 6
Height = 76.30 in.
54.00
42.00
7 8 9
16.00
3.38
31.38 3.38
Width
Figure 4.38: Locations of Selected Points to Trace Internal Movement of

(a) Tests 1 and 2 and (b) Tests 3, 4 and 5
160
2.00
1.80
1.60
1.40
Specimen Height (m)
Test 1
1.20
Test 2
1.00
Test 3
0.80
Test 4
0.60 Test 5
0.40
0.20
0.00
15 18 21 24 27 30
Dry Unit Weight (kN/m3)
Figure 4.39: Soil Dry Unit Weight Results during Specimen Preparation of the Five
Tests
161
Table 4.6: Dimensions of the GSGC specimens before loading
Height, Width, Length,

Test
m (in.) m (in.) m (in.)
Test 1 1.937 (76.25) 1.448 (57.00) 1.194 (47.00)
Test 2 1.939 (76.35) 1.372 (54.00) 1.187 (46.75)
Test 3 1.939 (76.35) 1.346 (53.00) 1.187 (46.75)
Test 4 1.938 (76.30) 1.492 (58.75) 1.187 (46.75)
Test 5 1.939 (76.35) 1.245 (49.00) 1.187 (46.75)
162
4.6 Test Results
4.6.1 Test 1 – Unreinforced Soil
This test is perhaps the largest plane strain test for soil with a confining pressure.
Test 1 was conducted as the base line for the other four GSGC tests.
The loading sequence of the soil mass was:

• Loading up to a vertical pressure of 250 kPa (nearly 1% vertical strain)
• Unloading to zero
• Reloading until a failure pressure of 770 kPa was reached.
The soil mass at failure is shown in Figure 4.40. Figure 4.41 shows the global
vertical stress-strain and volume change relationships of the soil mass. The average
lateral displacements, measured by LVDT’s, on the open faces of the soil mass under
different vertical stresses are presented in Figure 4.42. The internal displacements of
the soil at selected points under vertical applied pressures of 190 kPa, 310 kPa, 620
kPa, and 770 kPa are shown in Figure 4.43. The test results of Test 1 for
unreinforced soil are summarized in Table 4.7.
163
Figure 4.40: Soil Mass at Failure of Test 1
164
900
800
Applied Vertical Stress (kPa)
700
600
500
400
300
200
100
0
0 1 2 3 4 5 6
Global Vertical Strain (%)
(a)
1.2
0.8
0.6
0.4
0.2
0
0 0.5 1 1.5 2 2.5 3 3.5
-0.2
-0.4
(b)
Figure 4.41: Test 1-Unreinforced Soil Mass:

(a) Global Vertical Stress-Vertical Strain Relationship
(b) Global Volume Change Strain Relationship
165
2.00
1.75
1.50
Specimen Height (m)
1.25
Applied Pressure:
1.00
200 kPa
0.75 400 kPa
600 kPa
0.50
770 kPa
0.25
700 kPa
0.00
0 10 20 30 40 50
Lateral Displacement (mm)
Figure 4.42: Lateral Displacements on the Open Face of Test 1
166
Test 1
10.25
1 2 3
2'' x 2'' Grid System
34.25
4 5 6
76.25
Legend:
66.00
Before Loading
190 kPa (75 kips)
310 kPa (120 kips)
620 kPa (240 kips)
770 kPa (300 kips)
42.00
7 8 9
16.00
Note:
(All Displacements in inches, Drawn to Scale of the Soil Mass)
2.50
32.50 2.50
57.00
Figure 4.43: Internal Displacements of Test 1
167
Table 4.7: Some test results of Test 1
Test 1
Geosynthetic Reinforcement None

Test
Wide-Width Strength of Reinforcement None
Conditions
Reinforcement Spacing No Reinforcement
Confining Pressure 34 kPa
Applied Stress at Vertical Strain of 1% 335 kPa (48.6 psi)
Ultimate Applied Pressure 770 kPa (112 psi)
Test Results Vertical Strain at Failure 3%
Maximum Lateral Displacement of the Open

Face at Failure 47 mm
Stiffness at 1% vertical strain (Eat 1%) 33,500 kPa
Stiffness for Unloading-Reloading (Eur) 87,100 kPa
168
4.6.2 Test 2 – GSGC Test (T, Sv)
In this test, the GSGC mass was reinforced by nine sheet of single-sheet Geotex 4x4
with spacing of 0.2 m. The soil layer was compacted at 0.2 m-thick lifts. Each
reinforcement sheet was mounted with 54 strain gauges.
The failure load in this test was 1,000,000 pounds. All nine reinforcement sheets
were ruptured after testing. The composite mass after testing is shown in Figure 4.44.
The shear bands of the composite mass after testing are visible through the diagonal
lines of the mass (see Figures 4.44 and 4.45). Along the shear bands, the square grids
of 51 mm by 51 mm (2 in. by 2 in.) were severely distorted after testing (Figure 4.45).
These shear bands correspond exactly with the failure surfaces seen in Figure 4.46.
The location of rupture lines of all reinforcement sheets in the GSGC mass can be
seen in Figures 4.52 and 4.53.
The measured data of Test 2 are highlighted as bellows:

• Global stress-strain relationship: Figure 4.47 shows the global stress-strain
relationship of the composite up to and post failure. The maximum applied
vertical pressure was about 2700 kPa, where the corresponding vertical
displacement was 125 mm (6.5% vertical strain).
• Lateral displacement: The average lateral displacement profiles are on the
open faces of the composite under different vertical pressures shown in Figure
4.48. The lateral displacements were nearly uniform along the height of the
composite up to a pressure of about 600 kPa. At vertical pressures between
770 kPa and 1,500 kPa, the maximum lateral displacement occurred at about
3/8 H (H = the height of the composite mass) from the base. The locations of
the maximum displacements were about the same as those of Test 1
(unreinforced).
169
Figure 4.44: Composite Mass at Failure of Test 2
170
Figure 4.45: Close-Up of Shear Bands at Failure of Area A in Figure 4.44
171
Figure 4.46: Failure Planes of the Composite Mass after Testing in Test 2
172
Applied Vertical Stress (kPa) 3,000
2,500
2,000
1,500
1,000
500
0
0 1 2 3 4 5 6 7 8 9
Global Vertical Strain (% )
(a)
0
0 1 2 3 4 5 6 7
-0.2
-0.4
-0.6
-0.8
-1
-1.2
-1.4
-1.6
-1.8
(b)
Figure 4.47: Test 2-Reinforced Soil Mass:

(a) Global Vertical Stress-Vertical Strain Relationship
(b) Global Volume Change Strain Relationship
173
2
Applied Pressure:
1.75
200 kPa
1.5 400 kPa

Specimen Height (m)
600 kPa
1.25 770 kPa
1000 kPa
1
1250 kPa
1500 kPa
0.75
1750 kPa
0.5 2000 kPa
2250 kPa
0.25 2500 kPa
2700 kPa
0
0 10 20 30 40 50 60 70
174
The maximum lateral displacement at the mid-height of the composite under
the applied pressure of 2,700 kPa was 60 mm.
• Internal displacement: The internal displacements of the composite mass at
selected points under vertical applied pressures of 270 kPa to 2,700 kPa, at an
increment of 270 kPa are shown in Figure 4.49. At points 1 (and 3), 4 (and 6),
and 7 (and 9) near the open faces, the displacements moved downward and
outward with an angle, measured after testing, of about 67o, 47o, and 31o,
respectively, to the horizontal. The vertical displacements of the points at the
upper part of the soil mass were greater than those at the lower part. Along
the center line, there were only vertical displacements (Points 2, 5 and 8).
There was almost no displacement at Point 8 near the bottom and at the center
line. The maximum lateral displacement in the soil body was 60 mm at the
mid height on the open sides, and the maximum vertical displacement was
125 mm at the top of the specimen.
• Reinforcement strain: Figure 4.50 shows the locations of the strain gauges on
the geosynthetic sheets. The strain in the reinforcement of the GSGC mass is
shown in Figure 4.51. Most of the strain gauges performed well at strains less
than 4%. All reinforcement layers were found ruptured after the test
completed. The locations of the rupture lines can be seen from the aerial view
of the reinforcement sheets exhumed from the composite after testing (see
Figure 4.52). Based on the locations of the rupture lines, the rupture planes
can be constructed as shown in Figure 4.53. Note that this agrees perfectly
with the shear bands in Figure 4.46. The maximum strain in reinforcement at
ruptured was about 12 %, while the measured data from strain gauges were
only less than 4 %. From the strain distribution in Figure 4.51, the locations
of the maximum strain in reinforcement were different between layers. In
reinforcement layers near at the mid height of the GSGC mass (0.8 m and 1.0
175
m from the base), the maximum reinforcement strains were close to the
centerline, while the reinforcement layers at near the top and the base of the
GSGC mass (0,2 m and 1.8 m from the base), the maximum reinforcement
strains were at about 0.3 m from the edge of the composite mass. The
maximum strain locations in all reinforcement layers were at the ruptured
lines of reinforcement that can be seen in Figures 4.52 and 4.53.
• The test results of Test 2 are summarized in Table 4.8.
176
Test 2
10.35
1 2 3
38.35
4 5 6
76.35
Legend:
66.00
Before Loading
270 kPa
540 kPa
...
2,700 kPa
38.00
7 8 9
16.00
Note:
(All Displacements in inches, Drawn to Scale of the Composite Mass)
3.00
31.00 3.00
54.00
177
Reinforcement Strain Gages
7.63 7.63 7.63 7.63 7.63 7.63 7.63 7.63 7.63 7.63
Sheet No 9
Sheet No 8
Sheet No 7
Sheet No 6
Sheet No 5
Sheet No 4
Sheet No 3
Sheet No 2
Sheet No 1
11.00 8.00 8.00 8.00 8.00 11.00

54
(Note: All Dimensions in inches)
Figure 4.50: Locations of Strain Gauges on Geosynthetic sheets in Test 2
178
2.5 Applied Pressure:
2 200 kPa
400 kPa
Strain (%)
1.5 600 kPa

800 kPa
1
1000 kPa
0.5 1250 kPa
1500 kPa
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6
Distance from the Edge of the Composite Mass (m)
(a)
2.50 200 kPa

400 kPa
2.00
Strain (%)
600 kPa
1.50 800 kPa
1.00 1000 kPa
0.50 1250 kPa

1500 kPa
0.00
0.0 0.1 0.2 0.3 0.4 0.5 0.6
(b)
Figure 4.51: Reinforcement Strain Distribution of the Composite Mass in Test 2:

(a) Layer 1, at 0.2 m from the Base
(b) Layer 2, at 0.4 m from the Base
179
4 200 kPa
3.5 400 kPa
3
Strain (%)
600 kPa
2.5
2 800 kPa
1.5 1000 kPa
1 1250 kPa
0.5
1500 kPa
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6
(c)
3.5
Applied Pressure:
3
2.5 200 kPa
Strain (%)
2 400 kPa
1.5 600 kPa
1 800 kPa
0.5 1000 kPa
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6
(d)
Figure 4.51 (continued): Reinforcement Strain Distribution of the Composite

Mass in Test 2:
(c) Layer 3, at 0.6 m from the Base
(d) Layer 4, at 0.8 m from the Base
180
3
Applied Pressure:
2.5
200 kPa
2
Strain (%)
400 kPa
1.5
600 kPa
1
800 kPa
0.5
1000 kPa
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6
(e)
3
Applied Pressure:
2.5
200 kPa
2
Strain (%)
400 kPa
1.5
600 kPa
1
800 kPa
0.5
1000 kPa
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6
(f)

Mass in Test 2:
(e) Layer 5, at 1.0 m from the Base
(f) Layer 6, at 1.2 m from the Base
181
4 Applied Pressure:
3.5
3 200 kPa
Strain (%)
2.5 400 kPa

2 600 kPa
1.5 800 kPa
1 1000 kPa
0.5
1250 kPa
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6
(g)
3
Applied Pressure:
2.5 200 kPa
2 400 kPa
Strain (%)
1.5 600 kPa
1 800 kPa
1000 kPa
0.5
1250 kPa
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6
(h)

Mass in Test 2:
(g) Layer 7, at 1.4 m from the Base
(h) Layer 8, at 1.6 m from the Base
182
1.6
1.2
200 kPa
Strain (%)
1
0.8 400 kPa
0.6
600 kPa
0.4
0.2 800 kPa
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6
(i)

Mass in Test 2:
(i) Layer 9, at 1.8 m from the Base
183
Figure 4.52: Aerial View of the Reinforcement Sheets Exhumed from the
Composite Mass after Test 2
(The numbers next to each sheet indicate the sheet number in Figure 4.53)
184
Test 2
Reinforcement Rupture Lines
7.63 7.63 7.63 7.63 7.63 7.63 7.63 7.63 7.63 7.63
1
54
(Note: All Dimensions in inches, Drawn to Scale)
Figure 4.53: Locations of Rupture Lines of Reinforcement in Test 2;

Constructed based on Figure 4.52
185
Test 2
Geosynthetic Reinforcement Geotex 4x4

Test
Wide-Width Strength of Reinforcement 70 kN/m
Conditions
Reinforcement Spacing 0.2 m
Ultimate Applied Pressure 2700 kPa
Vertical Strain at Failure 6.5 %
Test Results Maximum Lateral Displacement of the Open

Maximum strain in reinforcement at ruptured 12 %
Maximum measured strain in reinforcement 4.0 %
186
4.6.3 Test 3 – GSGC Test (2T, 2Sv)
In this test, the GSGC mass was reinforced by four double-sheet Geotex 4x4 at 0.4 m
spacing. The strength and stiffness of the double-sheet reinforcement were nearly
doubled compared to those of the single-sheet reinforcement used in Test 2. The
GSGC mass after testing is shown in Figure 4.54.
The measured data of Test 3 are highlighted below:

• Global stress-strain relationship (see Figure 4.55): The maximum applied
vertical pressure was about 1,750 kPa. The vertical displacement at the failure
pressure was 118 mm (6.1% global vertical strain).
• Lateral displacement: The average lateral displacement profiles on the open
faces of the composite under different vertical pressures shown in Figure 4.56.
The maximum lateral displacement under the failure pressure 1750 kPa was
54 mm.
• Internal displacements at selected points are shown in Figure 4.57. The trend
of the internal movements in Test 3 was nearly the same as that in Test 2. The
points 1, 3, 4, 5, 6, 7 and 9 near the open faces, the displacements move
downward and outward with angles from 49o to 71o to the horizontal. The
points 2, 5 and 8 along the center line, the displacements were almost vertical.
The maximum vertical displacement at the top of the specimen was 118 mm.
shown in Figure 4.59. Three reinforcement layers near the top of the
composite were ruptured after testing. The locations of the rupture lines can
be seen from the aerial view of the reinforcement sheets exhumed from the
composite after testing (see Figure 4.60). Based on the locations of the
187
rupture line, the rupture planes can be constructed as shown in Figure 4.61.
This rupture line agrees perfectly with the failure line in Figure 4.54. The
maximum strain in reinforcement measured was 4 % and located at near the
ruptured line as shown in Figure 4.61.
188
Figure 4.54: Composite Mass after Testing of Test 3
189
2,000
1,800
1,600
1,400
1,200
1,000
800
600
400
200
0
0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00
Global Vertical Strain (% )
Figure 4.55: Global Stress-Strain Relationship of Test 3
190
2
1.75
Applied Pressure:
1.5
Specimen Height (m)
260 kPa
1.25
400 kPa
1 600 kPa
800 kPa
0.75
1000 kPa
0.5 1250 kPa
1500 kPa
0.25
1750 kPa
0
0.0 10.0 20.0 30.0 40.0 50.0 60.0
191
Test 3 2x2'' Grid System
10.35
1 2 3
34.35
14.5
10 11
60.36
4 5 6
2'' x 2'' Grid System Legend:
76.35
Before Loading
280 kPa
66.00
560 kPa
840 kPa
54.00
1120 kPa
1400 kPa
42.00
1680 kPa
1750 kPa
7 8 9
16.00
Note:
2.50
30.50 2.50
53.00
192
15.25
Sheet No 4
15.25
Sheet No 3
15.25
Sheet No 2
15.25
Sheet No 1
10.50 8.00 8.00 8.00 8.00 10.50 15.25

53.00
Figure 4.58: Location of Strain Gauges on Geosynthetic Sheets in Test 3
193
4
3.5 260 kPa
3 400 kPa
Strain (%)
2.5 600 kPa

2 `
1.5 800 kPa
1 1000 kPa
0.5 1250 kPa
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6
(a)
4.5
4 Applied Pressure:
3.5
260 kPa
3
Strain (%)
2.5 400 kPa

2 600 kPa
1.5
1 800 kPa
0.5 1000 kPa
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6
(b)

194
4.5
Applied Pressure:
4.0
3.5
260 kPa
3.0
Strain (%)
2.5 400 kPa

2.0
600 kPa
1.5
1.0 800 kPa
0.5 1000 kPa
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6
(c)
4.5
3.5
260 kPa
3.0
Strain (%)
2.5 400 kPa

2.0
600 kPa
1.5
1.0 800 kPa
0.5 1000 kPa
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6
(d)
Figure 4.59 (continued): Reinforcement Strain Distribution in Test 3:

195
Composite Mass after Test 3 (The numbers on each sheet indicate the sheet number in
Figure 4.61)
196
Test 3
15.25
1
15.25
2
15.25
3
15.25
4
15.25
53

Constructed based on Figure 4.60.
197
Test 3

Test
Conditions
Ultimate Applied Pressure 1,750 kPa

198
4.6.4 Test 4 – GSGC Test (T, 2 Sv)
The reinforcement used in this test was single-sheet of Geotex 4x4 at spacing of 0.4
m. The composite mass after testing is shown in Figure 4.62. The failure surfaces of
the composite mass after testing can be seen clearly in this Figure.
The measured data for of Test 4 are highlighted below:

• Global stress-strain relationship (see Figure 4.63): The maximum applied
pressure was 77 mm (4.0% global vertical strain).
• Lateral displacement: The average lateral displacements on the open faces of
the composite mass under different vertical pressures as shown in Figure 4.64.
The maximum lateral displacement under the failure pressure was 52 mm.
• Internal displacements at selected points are shown in Figure 4.65. The trends
of the internal movements in Tests 2, 3 and 4 were identical. The points 1, 3,
4, 5, 6, 7 and 9 near the open faces, the displacements move downward and
outward with angles from 30o to 63o to the horizontal. The points 2, 5 and 8
along the center line, the displacements were almost vertical. The maximum
vertical displacement at the top of the specimen was 77 mm.
shown in Figure 4.67. All reinforcement layers near the top of the composite
were ruptured after the test was completed. The locations of the rupture lines
can be seen from the aerial view of the reinforcement sheets exhumed from
the composite after testing (see Figure 4.68). Based on the locations of the
199
maximum strain in reinforcement at ruptured was about 12 % and located at
near the ruptured line as shown in Figure 4.69.
200
(a) (b)
Figure 4.62: Failure Planes of the Composite Mass after Testing in Test 4:
(a) Front View at the South
(b) Back View at the North.
201
1,400
1,200
1,000
800
600
400
200
0
0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00
202
2
1.75

Specimen Height (m)
1.25 200 kPa
400 kPa
1
600 kPa
0.75
800 kPa
0.5 1000 kPa
1250 kPa
0.25
1300 kPa
0
0.0 10.0 20.0 30.0 40.0 50.0 60.0
203
15.00
Test 4
10.35
1 2 3
10 11
34.35
4 5 6
76.30
Legend:
Before Loading
250 kPa
500 kPa
54.00
750 kPa
1000 kPa
42.00
1250 kPa
1300 kPa
7 8 9
Note: 16.00
3.00
31.00 3.00
58.00
204
15.25
Sheet No 4
15.25
Sheet No 3
15.25
Sheet No 2
15.25
Sheet No 1
15.25
13.35 8.00 8.00 8.00 8.00 13.35
58.75
205
1.8
Applied Pressure:
1.6
1.4 100 kPa
1.2 200 kPa
Strain (%)
1 300 kPa
0.8 400 kPa
0.6 500 kPa
0.4
600 kPa
0.2
800 kPa
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Distance from the Edge of the Composite M ass (m)
(a)
2 Applied Pressure:
1.8
1.6 100 kPa
1.4
Strain (%)
200 kPa
1.2
1 300 kPa
0.8
400 kPa
0.6
0.4 500 kPa
0.2 600 kPa
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
(b)

206
1
Applied Pressure:
0.8
Strain (%)
100 kPa
0.6
200 kPa
0.4
300 kPa
0.2
400 kPa
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
(c)
1.8
Applied Pressure:
1.6
1.4
100 kPa
1.2
Strain (%)
1 200 kPa
0.8 300 kPa
0.6
0.4 400 kPa
0.2 500 kPa
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
(d)
Figure 4.67 (continued): Reinforcement Strain Distribution of the Composite Mass in

Test 4:
207
Composite Mass after Test 4 (The numbers on each sheet indicate the sheet number in
Figure 4.69)
208
Test 4
15.25
1
15.25
2
15.25
3
15.25
4
15.25
58

209
Test 4

Test
Conditions

210
4.6.5 Test 5 – GSGC Test (unconfined with T, Sv)
The configuration of this test was the same as Test 2. The reinforcement was the
single-sheet Geotex 4x4 at spacing of 0.2 m. The confining pressure was not applied
for this test. Without applying confining pressure, the soil on the open faces fell of
continuously with increasing applied pressure. The composite mass and failure
surfaces after testing are shown in Figures 4.70 and 4.71.
The measured data of Test 5 highlighted below:

• The global stress-strain relationship (see Figure 4.72): The maximum applied
pressure was 111 mm (6.0 % global vertical strain).
• Lateral displacement: The average lateral displacements on the open faces of
the composite mass under different vertical pressures as shown in Figure 4.73.
The maximum lateral displacement at the open faces under the failure
pressure could not been measured because the soil at these faces dropped
during testing under the high applied pressures.
• Internal displacements at selected points are shown in Figure 4.74. The trend
of the internal movements in Test 5 was nearly the same as that in the other
Tests 2, 3 and 4. The points 1, 3, 4, 5, 6, 7 and 9 near the open faces, the
displacements move downward and outward with angles from 35o to 63o to the
horizontal. The points 2, 5 and 8 along the center line, the displacements were
almost vertical. The maximum vertical displacement at the top of the
specimen was 111 mm.
shown in Figure 4.76. Eight reinforcement layers near the top of the
211
composite were ruptured after testing. The locations of the rupture lines can
be seen from the aerial view of the reinforcement sheets exhumed from the
composite after testing (see Figure 4.77). Based on the locations of the
maximum strain in reinforcement measured was 3.2 % and located at near the
ruptured line as shown in Figure 4.78.
212
Figure 4.70: Composite Mass at Failure of Test 5
213
(a) (b)
Figure 4.71: Failure Planes of the Composite Mass after Testing in Test 5:
(a) Front View (South),
(b) Back View (North)
214
2,200
2,000
1,800
1,600
1,400
1,200
1,000
800
600
400
200
0
0 1 2 3 4 5 6 7 8
Global Vetical Strain (%)
215
2.000
1.750
1.500
Applied Pressure:
Specimen Height (m)
1.250
200 kPa
1.000
400 kPa
0.750
600 kPa
0.500 1500 kPa
0.250
0.000
0.0 10.0 20.0 30.0 40.0
216
12.50
Test 5
10.35
1 2 3
34.35
10 11
60.35
4 5 6
2'' x 2'' Grid System Legend:
76.35
Before Loading
300 kPa
600 kPa
900 kPa
54.00
1200 kPa
1500 kPa
42.00
1800 kPa
7 8 9
16.00
Note:
2.50
28.50 2.50
49.00
217
15.25
Sheet No 9
Sheet No 8
15.25
Sheet No 7
Sheet No 6
15.25
Sheet No 5
Sheet No 4
15.25
Sheet No 3
Sheet No 2
15.25
Sheet No 1
8.50 8.00 8.00 8.00 8.00 8.50

49.00
218
3 100 kPa
2.5 200 kPa
Strain (%)
2 300 kPa
1.5 400 kPa
1 500 kPa
600 kPa
0.5
750 kPa
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6
(a)
3 Applied Pressure:
2.5 100 kPa
2 200 kPa
Strain (%)
300 kPa
1.5
400 kPa
1 500 kPa
0.5 600 kPa
750 kPa
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6
(b)

219
1.2
Applied Pressure:
1
100 kPa
0.8
Strain (%)
200 kPa
0.6
300 kPa
0.4
400 kPa
0.2
525 kPa
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6
(c)
Figure 4.76 (continued): Reinforcement Strain Distribution of the Composite Mass in

Test 5:
220
Composite Mass after Test 5 (The number on each sheet indicate the sheet number in
Figure 4.78)
221
Test 5
7.63 7.63 7.63 7.63 7.63 7.63 7.63 7.63 7.63 7.63
1
49

222
Test 5

Test
Conditions
Confining Pressure 0

Face at Failure Not Measured
223
4.7 Discussion of the Results
The results of the GSGC Tests are discussed in term of the following:
• Effects of geosynthetic inclusion (comparison between Tests 1 and 2).
• Relationship between reinforcement spacing and reinforcement strength
(comparison between Tests 2 and 3).
• Effects of reinforcement spacing (comparison between Tests 2 and 4).
• Effects of reinforcement stiffness (comparison between Tests 3 and 4).
• Effects of confining pressure (comparison between Tests 2 and 5)
• Composite strength properties.
4.7.1 Effects of Geosynthetic Inclusion (Comparison between Tests 1 and 2)
Table 4.12 shows the result comparisons between an unreinforced soil mass (Test 1)
and a soil mass reinforced by Geotex 4x4 at 0.2 m spacing (Test 2). With the
presence of the reinforcement, the reinforced soil was much stronger than the
unreinforced soil. The ultimate applied pressure for the GSGC mass was about 3.5
times as large as the strength of the soil mass without reinforcement. The stiffness of
the unreinforced soil mass was 50 % of that for the reinforced soil mass. In addition,
the reinforced soil mass was much more ductile than the unreinforced soil mass. The
global vertical strain was 6.5 % at failure for Test 2; whereas it was only 3.0 % for
Test 1.
224
Table 4.12: Comparison between Test 1 and Test 2
Test 1 Test 2 (T, Sv)
Geosynthetic Reinforcement None Geotex 4x4
Wide-Width Strength of
Reinforcement None Tf = 70 kN/m
Reinforcement Spacing No Reinforcement Sv = 0.2 m
Confining Pressure 34 kPa 34 kPa
Ultimate Applied Pressure 770 kPa 2,700 kPa
Vertical Strain at Failure 3% 6.5 %
Maximum Lateral Displacement

of the Open Face at Failure 47 mm 60 mm
Stiffness at 1% vertical strain

(Eat 1%) 33,500 kPa 61,600 kPa
Stiffness for Unloading-Reloading

(Eur) 87,100 kPa Not Applied
225
4.7.2 Relationship between Reinforcement Spacing and Reinforcement
Strength (Comparison between Tests 2 and 3)
The relationship between reinforcement spacing and reinforcement strength can be

seen by comparing the results of Tests 2 and 3, as shown in Table 4.13. As noted in
Section 2.1, the current design methods are based on the concept that the
reinforcement spacing and reinforcement strength play an equal role on the
performance of a GRS mass. In other words, a GRS wall with reinforcement strength
of Tf at spacing Sv will behave the same as the one with reinforcement strength of
2*Tf at twice the spacing 2*Sv. The results of Tests 2 and 3 demonstrated that this
concept adopted in current design methods is not correct. With the same Tf / Sv ratio
(= 350 kN/m2) in Tests 2 and 3, the stiffness and strength of the Test 2 (with Tf and Sv
= 0.2 m) was much higher than that of the Test 3 (with 2Tf and Sv = 0.4 m). The
strength of the composite mass in Test 3 was only 65 % of the strength in Test 2 (see
Table 4.13). These results suggest that reinforcement spacing plays a much more
important role than strength of reinforcement in a reinforced soil mass.
226
Table 4.13: Comparison between Test 2 and Test 3 with the same Tf/Sv ratio
Test 2 (T, Sv) Test 3 (2 T, 2Sv)
Reinforcement Tf = 70 kN/m Tf = 140 kN/m
Reinforcement Spacing Sv = 0.2 m Sv = 0.4 m
Tf / Sv 350 kPa 350 kPa
Ultimate Applied Pressure 2,700 kPa 1,750 kPa

(Eat 1%) 61,600 kPa 48,900 kPa
Vertical Strain at Failure 6.5 % 6.1 %

227
4.7.3 Effects of Reinforcement Spacing (Comparison between Tests 2 and 4)
The effects of reinforcement spacing can be seen by comparing the results of Tests 2
and 4, as shown in Table 4.13. All test conditions in the two tests are the same,
except the reinforcement spacing in Test 4 was 0.4 m, while it was 0.2 m in Test 2.
The results demonstrate the importance of reinforcement spacing on the behavior of a
GRS mass. With reinforcement spacing of 0.2 m, the strength of the GRS mass was
about twice as high as the one with 0.4 m spacing. The corresponding increase in
stiffness at 1 % strain was about 30 %. The GRS mass at 0.2 m spacing also
exhibited significantly higher ductility than at 0.4 m spacing.
228
Test 2 (T, Sv) Test 4 (T, 2Sv)


(Eat 1%) 61,600 kPa 46,600 kPa
229
4.7.4 Effects of Reinforcement Strength (Comparison between Tests 3 and 4)
The effects of reinforcement strength can be seen by comparing the results of Tests 3
and 4, as shown in Table 4.15. All test conditions in the two tests are identical,
except that reinforcement strength in Test 3 was almost twice as high as that in Test
4. The results indicate that the increase in strength of the GRS mass due to doubling
the reinforcement strength was about 35%. The increase is much smaller than
doubling the reinforcement spacing (Section 4.7.3) where the increase in strength of
the GRS mass is over 100%. It is noted that the increase in stiffness at 1 % strain due
to doubling the reinforcement strength is only about 5%, compared to about 30%
increase due to doubling the reinforcement spacing.
Test 3 (2T, 2Sv) Test 4 (T, 2Sv)


(Eat 1%) 48,900 kPa 46,600 kPa
230
4.7.5 Effects of Confining Pressure (Comparison between Tests 2 and 5)
The effects of confining pressure can be seen by comparing the test results of Tests 2
and 5, as shown in Table 4.16. All test conditions in the two tests are identical,
except that the confining pressure in Test 2 was 34 kPa, while Test 4 was conducted
without confinement. The results indicate that the increase in strength due to the
confining pressure was about 40 %. The increase in stiffness at 1 % strain due to the
confining pressure is about 15 %.
231
Test 2 (T, Sv) Test 5 (T, Sv)
Confining Pressure 34 kPa 0

of the Open Face at Failure 60 mm Not Measured

(Eat 1%) 61,600 kPa 52,900 kPa
232
4.7.6 Composite Strength Properties
Table 4.17 shows a comparison of the composite strength properties of five GSGC
tests as obtained from the measured data and calculated from Schlosser and Long’s
method (1972) by assuming the friction angle remains the same as unreinforced soil
(i.e., same φ value as in Test 1). The apparent cohesion, CR, from Schlosser and
Tf
Long’s method (1972) is calculated as C R = K p + c , where c = the cohesion of
2 Sv
the backfill; Tf = the strength of reinforcement; Sv = reinforcement spacing; Kp =
coefficient of passive earth pressure. The values from Schlosser and Long’s method
were much higher than the measured by 20 % to 86 %.
233
Table 4.17: Comparison of strength properties of five GSGC Tests
Parameter Test 1 Test 2 Test 3 Test 4 Test 5

(un- (T, Sv) (2 T, 2Sv) (T, 2Sv) (T, Sv)
reinforced)
Wide-Width Strength of 70 140 70 70

Reinforcement, Tf (kN/m)
Reinforcement Spacing, 0.2 0.4 0.4 0.2

Sv (m)
Tf / Sv (kPa) 350 350 175 350
Confining Pressure (kPa) 34 34 34 34 0
Apparent Cohesion, CR,

(kPa) by Schlosser and 550 550 310 550
Long’s Method
Ultimate Applied
Pressure (kPa) from 770 2,700 1,750 1,300 1,900
Measured Data
Ultimate Applied
Pressure (kPa) by 3,250 3,250 1,930 3030
Schlosser and Long’s
Method
Difference in Ultimate
Pressure between
Measured Data and 20 % 86 % 48 % 59 %
Schlosser and Long’s
Method
234
5. ANALYTICAL MODELS FOR EVALUATING COMPACTION-
INDUCED STRESSES (CIS), COMPOSITE STRENGTH PROPERTIES
OF A GRS COMPOSITE, AND REQUIRED REINFORCEMENT
STRENGTH
This Chapter described three analytical models. The models are for evaluating (1)
Compaction-Induced Stresses (CIS) in a GRS mass, (2) strength properties of a GRS
mass, and (3) required tensile strength of reinforcement in the design of GRS
structures. The first analytical model is a simple compaction model that is capable of
estimating the compaction-induced stresses, or the increase of the horizontal stresses,
in a GRS mass due to fill compaction. The model was developed by combining a
compaction model developed by Seed (1983) and the companion hand-calculation
procedure by Duncan and Seed (1986) for an unreinforced soil mass, and (b) the
theory of GRS composite behavior proposed by Ketchart and Wu (2001).
The second analytical model is for determination of the strength properties of a GRS
composite. With the analytical model, a new relationship between reinforcement
strength and reinforcement spacing is introduced to reflect an observation made in
actual construction and in controlled experiments regarding the relative effects of
reinforcement spacing and reinforcement strength on the performance of GRS
structures.
The third analytical model is for determination of reinforcement strength in design.

This model was also developed based on the new relationship between reinforcement
strength and reinforcement spacing.
235
5.1 Evaluating Compaction-Induced Stress in a GRS Mass
5.1.1 Conceptual Model for Simulation of Fill Compaction of a GRS Mass
Based on previous studies regarding CIS for unreinforced soil masses, and very
limited study for reinforced soil masses, a conceptual stress path for loading-
unloading-reloading of a GRS mass is shown in Figure 5.1. An explanation of Figure
5.1 is given below:
- A GRS mass is loaded (due to application of compaction loads) from an initial
state (point A) following the Ki,c-line (with horizontal stress, σ ' h = K i , c σ ' v ,
where Ki,c = coefficient of lateral earth pressure of the GRS mass for initial
loading) up to point B. At point B, the GRS mass reaches a maximum stress
state with the vertical stress of σ 'v , max = σ 'v + Δσ 'v , c , max (σ'v = vertical stress at
the initial stress; Δσ'v,c,max = maximum increase in vertical stress due to

compaction loading).
- Upon unloading (i.e., upon removal of the compaction loads), the stresses in
the soil are reduced by following a non-linear path from point B to point C.
- In cases of “significant” unloading, i.e., during unloading, the unloading-path
reaches the limiting line (the “K1,c-line”) at point E, further unloading stress
path will follow line EF (with σ 'h = K1, c σ 'v ).
- Upon reloading due to the next cycle of compaction load application, the
stresses are to follow a K3,c-line (with σ ' h = K 3, c σ ' v ) from either point C or
pint F until it meets the initial loading path (line AB).

- The subsequent cycles of unloading and reloading shall not deviate much
from the K3,c-line, as suggested by Broms (1971), Seed (1983), and Erlich and
Mitchell (1993). Therefore, the same K3-line can be used for all subsequent
cycles of reloading and unloading.
236
K1,c-line
σh
K1,c
Ki,c-line
Unloading-curve
1
B
D
K3,c-line
C
G
E K3,c-line
Δσ'h,c,r
F σ'h = Ki,c σ'v
σv
σ'v σ'v + Δσ'v,c,max
Figure 5.1: Conceptual Stress Path for Compaction of a GRS Mass
5.1.2 A Simplified Model to Simulate Fill Compaction of a GRS Mass
A simplified compaction simulation model, as depicted in Figure 5.2, is proposed for

simulation of fill compaction of a GRS mass. By using the proposed model, the
increase of the horizontal stresses in a GRS mass due to compaction can be estimated.
These stresses, namely CIS, are represented by the horizontal residual stresses,
Δσ'h,c,r, in Figure 5.2. The proposed model is based on the bi-linear compaction
model suggested by Seed (1983) for an unreinforced soil mass, and the companion
hand-calculation procedure suggested by Duncan and Seed (1986). The proposed
model considers the presence of geosynthetic inclusions.
The stress path for fill compaction of a GRS mass in the proposed model can be
considered as a simplified form of the conceptual model described in Section 5.1.
The differences between the simplified model and the conceptual model are:
237
- Upon the removal of the compaction loads, the stresses in the soil are
reduced by following the K2,c-line from point B to point C, of which the
vertical stress, σ’v, equals to that of point A. The horizontal residual stress
due to the compaction loading is Δσ'h,c,r.
- In case of an unreinforced soil mass (or the reinforcement stiffness is
negligible), the stress in the soil in response to removal of compaction
loads will reduce to point S (instead of point C). The horizontal residual
stress due to compaction loading will be Δσ'h,s,r with Δσ'h,s,r ≤ Δσ'h,c,r.
- Upon application of the next cycle of compaction loads or placement of
new fill layers, the reloading path will follow the K3,c-line
(with σ ' h = K 3, c σ ' v , and K3, c ≤ K 2, c ). The stress path is to follow K3,c-line
from point C or point F until it meets the initial loading path, then follow
Ki,c-line to a new stress state.
K1,c-line K2,c-line
σh σ'h = K2,c σ'v K0-line
K1,c K3,c-line K0
1 σ'h = K3,c σ'v 1
B
σ'h,c C D
E σ'h = Ki,c σ'v Δσ'h,c,max

σ'h,s
Δσ'h,c,r S KA-line
Δσ'h,s,r F 1
KA
A
Unreinforced soil
Ki,c-line
σv
Figure 5.2: Stress Path of the Proposed Simplified Model for Fill Compaction of a
GRS Mass
238
5.1.3 Model Parameters of the Proposed Compaction Simulation Model
Four model parameters are needed for the proposed compaction simulation model,
including: Ki,c, K1,c, K2,c, and K3,c. These parameters can be estimated from soil and
reinforcement properties using correlations shown in Table 5.1, of which the
empirical coefficients (e.g., α and F) are to be calibrated by the measured data from
the Generic Soil-Geosynthetic Composite (GSGC) tests (to be described in Chapter
4). Recommended values of the empirical coefficients are to be given for routine
applications. Alternatively, the model parameters can be obtained directly from the
⎛ Er ⎞
results of GSGC tests. Note that the term ⎜⎜ 0.7 ⎟⎟ , for the estimation of
⎝ E s S v − 0.7 J r ⎠
K2, c, is to account for the presence of reinforcement. For hand calculations, the
maximum increase of "vertical" stress, Δσ 'v , c , max as shown in Figure 5.2, due to
compaction can be estimated simply by using the Westergaard's solution (1938).
239
Table 5.1 : Model parameters for the proposed compaction simulation model
Parameter Name Range of the Preliminary Estimation Based on Soil

Parameter and Reinforcement Properties
Values
Ki,c Coefficient of K A ≤ K i ,c ≤ K 0 K i ,c ≅ β K A
lateral earth
pressure of a 1.0 ≤ β ≤ 1.5
GRS mass for
⎛ φ' ⎞
initial loading K A ≅ tan 2 ⎜ 45 − ⎟
⎝ 2⎠
K 0 ≅ 1 − sin φ '
K1,c Limiting K 1,c ≅ K P ⎛ φ' ⎞

coefficient of K 1,c ≅ K P ≅ tan 2 ⎜ 45 + ⎟
⎝ 2⎠
lateral earth
pressure for
unloading
K2,c Coefficient of 0 ≤ K 2 ,c ≤ K 0 ⎡ ⎛ Er ⎞⎤
lateral earth K 2, c ≅ ⎢1 − F ⎜⎜1 + 0.7 ⎟⎟⎥ K i , c
⎣ ⎝ E s S v − 0.7 E r ⎠⎦
pressure for
unloading
where: F = 1 −
(OCR − OCR ) (Seed,
α
(OCR − 1)
1983); α ≅ sin φ '
K3,c Coefficient of 0 ≤ K 3, c ≤ K 0 K 3, c ≅ K 2 , c
lateral earth
pressure for
reloading
Note: K0 = coefficient of at-rest lateral earth pressure

KA = coefficient of active lateral earth pressure
KP = coefficient of passive lateral earth pressure
OCR = over-consolidation ratio
Es = soil stiffness (kPa)
Er = reinforcement stiffness (kN/m)
Sv = reinforcement spacing (m).
240
5.1.4 Simulation of Fill Compaction Operation
Fill compaction is a complex operation in term of change in stresses. This section

describes the maximum increase of vertical stress at depth z along a given section in a
soil mass due to typical compaction operation. Let us consider the change of vertical
stress at depth z in section I-I due to moving compaction loads, as shown in Figures
5.3 and 5.4. The directions of the moving compaction plant may be: (1) coming
toward section I-I, and (2) going away from section I-I. The compaction loads are
simulated by loading and unloading at different locations (locations 1, 2, and i), as
shown in Figures 5.3(a) and 5.4(a). Figure 5.3(b) and Figure 5.4(b) show the stress
path of the stresses at depth z along section I-I.
An explanation of Figure 5.3(b) is given below:
1. The initial stress condition at depth z along section I-I is denoted by
point A, with the initial vertical stress being σv;
2. With the compaction loads at location 1, the stresses are increased by
following the Ki,c-line to point B;
3. As the compaction loads are removed from location 1, the stresses will
reduce from point B to point C by following the K2,c-line;
4. When compaction loads move to a new location (location 2), the
stresses will increase from point C through point B to point D;
5. As the compaction loads are removed from location 2, the stresses will
reduce from point D to point E by following again the K2,c-line.
6. Steps 1 through 5 are repeated for all subsequent new locations as the
compaction plant moves toward section I-I. Note that as the
compaction plant moves closer to section I-I, the vertical stress at
depth z will become larger.
7. The maximum vertical stress condition will be reached when the
compaction plant is directly above section I-I. The corresponding
241
stress condition is represented by point F. Upon removal of the
compaction loads, the stress path will follow K2-line to point G.
In Figure 5.4, the compaction plant moves away from section I-I. Initially the
compaction is located directly above section I-I, which causes the stresses to increase
from point A to point F in Figure 5.4(b) due to the compaction loads. As the
compaction loads are removed from section I-I, the stresses are reduced from point F
to point G, following the K2,c-line. As the compaction plant moves away from
section I-I, the stress conditions will move to points D and B, along line FG. As the
compaction plant is finally removed, the stress condition will be at point G.
From Figures 5.3 and 5.4, it is noted that the residual stresses, as denoted by the
vertical distance AG, are the same for the two cases. It indicates that to determine the
compaction-induced stresses at a certain section due to a moving compaction plant,
one only need to determine the residual lateral stresses as the compaction loads are
directly above the section under consideration.
Figure 5.5 shows the conceptual stress path on the effect of the number of compaction
passes. For the first pass, the residual stress at point A is represented by point G.
With the subsequent pass of the compaction plant, the slope of the K2-line will
increase, and point G becomes point G'. As the number of compaction pass increases,
the final residual stresses will move from G’ to G’’, then to G'''.
242
compaction load
(a) I
location i location 2 location 1

Ζ
Δσ'v
changing K2,c-line due to

σh passes of compaction
(b)
F Ki,c-line
G
i
2 D
Δσ'h,c,r E
1
C K2,c and K3,c-line
B
(Assume: K2,c = K3,c)
A
σv
Figure 5.3: (a) Locations of Compaction Loads, and (b) Stress Paths during
Compaction at Depth z along Section I-I, as Compaction Loads Moving toward
Section I-I.
243
compaction load
(a) I
location 1 location 2 location i

Ζ
Δσ'v
P
changing K2,c-line due to

σh passes of compaction
(b)
1
F
G 2 D
B Ki,c-line
i
Δσ'h,c,r
K2,c and K3,c-line
A
σv
Figure 5.4: (a) Locations of Compaction Loads, and (b) Stress Paths during
Compaction at Depth z along Section I-I, as Compaction Loads Moving away from
Section I-I.
244
changing K2,c-line due to a
σh number of compaction passes
G''' F
G''
G'
G Ki,c-line
Δσ'h,c,r
K2,c and K3,c-line
A
σv
Figure 5.5: Stress Path at Depth z when Subject to Multiple Compaction Passes
245
5.1.5 Estimation of K2,c
The estimation of the K2,c in the proposed model for determining the compaction-
induced stresses in a GRS mass is presented in this section. Two stress reduction
factors “F” and “A” (as shown in Figure 5.6) are introduced. Factor F represents the
compaction-induced stresses for unreinforced soil in Seed’s model (1983), while
factor A is considered the presence of the reinforcement in the GRS mass.
σh
B
K2, s for soil
K2, c for composite
G*(Δσ) Δσ
S
F∗(Δσ) Ki, c - line
σv
σ'v σ'v + Δσ'v, c, max
Figure 5.6: Stress Path of the Proposed Model for Fill Compaction of a GRS Mass
(a) For Compacted Soil
For an unreinforced soil fill, the coefficient of lateral earth pressure for unloading in
Figure 5.2 can be estimated by using the expression suggested by Seed (1983):
K 2, s = (1 − F ) K i ,c (5.1)
246
where F = 1 −
(OCR − OCR ) ; α ≅ sin φ ' ; and OCR ≅ 5 for typical
α
OCR − 1
compacted sand (Seed, 1983).
(b) For GRS Composite
For a GRS mass, total residual strain in the soil can be determined as:
εs =
(F + G ) (Δσ ) (5.2)
Es
where F and G = stress reduction factors shown in figure 5.6; Δσ = increase
in horizontal stress due to compaction; Es = soil stiffness.
From Equation 5.2, the reinforcement force, T, due to residual strain in the soil can be
determined as:
⎛ (F + G ) Δσ ⎞
T = ⎜⎜ ⎟⎟ E r (5.3)
⎝ Es ⎠
where Er = reinforcement stiffness.
The average residual stress in the soil due to compaction, G * (Δσ ) , is:
⎛ (F + G ) Δσ ⎞ Er
G * Δσ = 0.7 ⎜⎜ ⎟⎟ (5.4)
⎝ Es ⎠ Sv
where Sv = reinforcement spacing.
or G E s S v = 0.7 F E r + 0.7 G E r (5.5)
Thus,
0.7 F E r
G= (5.6)
E s S v − 0.7 E r
Since
K 2 ,c = [ 1 − (F + G ) ] K i ,c (5.7)
Substituting (5.6) into (5.7), we have
247
⎡ ⎛ 0.7 E r ⎞⎤
K 2, c ≅ ⎢1 − F ⎜⎜1 + ⎟⎟⎥ K i , c (5.8)
⎣ ⎝ E s S v − 0.7 E r ⎠⎦
The Factor 0.7 in Equation 5.4 will be explained in the section 5.2.1. The increased
horizontal stress in a GRS mass due to compaction can be estimated as:
Δσ h , c = Δσ (G + F ) = Δσ ' v , c , max (K i , c − K 2, c ) (5.9)
Substituting (5.8) into (5.9), we have
⎧⎪ ⎡ ⎛ 0.7 E r ⎞⎤ ⎫⎪
Δσ h , c = Δσ ' v , c , max ⎨ K i , c ⎢ 1 − 1 + F ⎜⎜ 1 + ⎟⎟⎥ ⎬ (5.10)
⎪⎩ ⎣ ⎝ E s S v − 0.7 E r ⎠⎦ ⎪⎭
or
⎛ 0.7 E r ⎞
Δσ h , c = Δσ ' v , c , max K i , c F ⎜⎜ 1 + ⎟⎟ (5.11)
⎝ E s S v − 0 . 7 E r ⎠
Equation 5.11 shows calculation of the residual lateral stress in a GRS mass due to
compaction. The effect of compaction-induced stress in a GRS mass can be also seen
from Equation 5.11.
Using Equation 5.11, the increase of lateral stress in a GRS mass can be estimated
and the increase in soil stiffness can be evaluated. For example, the stiffness of a soil
can be evaluated as:
⎡ R f (1 − sin φ ) (σ 1 − (σ 3 + Δσ h ,c ))⎤
2
⎛ σ 3 + Δσ h ,c
n
⎞
E = ⎢1 − ⎥ K Pa ⎜⎜ ⎟⎟
2 c cos φ + 2 (σ 3 + Δσ h ,c ) sin φ ⎦⎥
(5.12)
t ⎢ ⎝ Pa ⎠
⎣
where
248
Et = tangent modulus
Rf = ratio of ultimate deviator stress to the failure deviator stress
c = cohesive strength
φ = angle of friction
Pa = the atmospheric pressure
K and n = material parameters.
249
5.2 Strength Properties of GRS Composite
Schlosser and Long (1972) proposed the concept of increase of apparent confining
pressure and concept of apparent cohesion of a GRS composite. The Mohr circles of
an unreinforced cohesive soil and a reinforced cohesive soil at failure are shown in
Figure 5.7.
φ
Shear Streess, τ
φ
Reinforced
Soil
Reinforced
Soil
CR
C Unreinforced Soil
σ3 σ3R σ1 σ1R
Δσ3R Normal Stress, σ
Figure 5.7: Concept of Apparent Confining Pressure and Apparent Cohesion of a

GRS Composite
The apparent cohesion of a GRS composite can be determined as:

Δσ 3 R K P
cR = +c (5.13)
2
where c R = apparent cohesion of a GRS composite
c = cohesion of soil
250
Kp = coefficient of passive earth pressure
Δσ 3 R = increase of confining pressure due to reinforcement
Schlosser and Long (1972) also proposed an equation to calculate increased confining
pressure as:
Tf
Δσ 3 = (5.14)
Sv
This expression implies that an increase in reinforcement strength, Tf, has the same
effect as a proportional decrease in reinforcement spacing, Sv. Many experimental
test results have shown that Equation 5.14 is not correct. Reinforcement spacing
plays a far more important role than reinforcement strength (Adams, 1997 and 2007;
Elton and Patawaran, 2004 and 2005; Ziegler et al., 2008). This point is supported by
the experiments conducted as a part of this study, as presented in Chapter 4.
5.2.1 Increased Confining Pressure
A new method to estimate the increased confining pressure in soil due to the presence
of reinforcement is presented. The proposed equation for the increased confining
pressure can be expressed as:
⎛ Tf ⎞
Δσ 3 = W ⎜⎜ ⎟⎟ (5.15)
⎝ Sv ⎠
where the factor, W, can be estimated as:
⎛ Sv ⎞
⎜ ⎟
⎜ S ref ⎟
W = r ⎝ ⎠
(5.16)
where Tf = extensile strength of reinforcement

r = a dimesionless factor (will be discussed later in this Section)
251
Sv = vertical spacing of reinforcement
Sref = the reference spacing (will be discussed further later)
To estimate the factor r in Equation 5.16, the concept of “average stresses” proposed
by Ketchart and Wu (2001) was employed. Instead of using average stresses,
however, average reinforcement forces were used.
a. Average Stress in GRS Mass by Ketchart and Wu (2001)
Ketchart and Wu (2001) developed a concept of “average stress” to determine the

behavior of a GRS composite based on a load-transfer analysis. From a simplified
preloading-reloading model for GRS mass, the equations to calculate stresses and
displacements of a GRS mass were developed using the idealized geometry of plain-
strain GRS mass and differential elements of the soil and reinforcement for
equilibrium equations (Hermann and Al-Yassin, 1978) as shown in Figures 5.8 and
5.9.
From the equilibrium equations and a number of assumptions, the stresses in soil and
the force in the reinforcement can be calculated as:
The force in the reinforcement:

⎡⎛ υ ⎞ ⎤ ⎛1 −υs2
⎟⎟ Pv − Ph ⎥ ⎜
⎞⎛
⎟ ⎜1 − β ⎞ ⎛ cosh (α x ) ⎞
Fx = Ar E r ⎢⎜⎜ s ⎟ ⎜⎜1 − ⎟⎟ (5.17)
⎣⎝ 1 − υ s ⎠
⎜
⎦ ⎝ Es
⎟⎝ α2
⎠ ⎠ ⎝ cosh (α L ) ⎠
The horizontal stress in the soil:
252
⎛ υ ⎞ ⎛⎛ υ ⎞ ⎞ ⎛ β cosh (α x ) ⎛ β ⎞⎞
σ x = − ⎜⎜ s ⎟⎟ Pv + ⎜ ⎜⎜ s ⎟⎟ Pv − Ph ⎟ ⎜⎜ 2 + ⎜ 1 − ⎟ ⎟⎟ (5.18)
⎜ 1−υ ⎟ α cosh (α L ) ⎝ α 2
⎝1 −υs ⎠ ⎝⎝ s ⎠ ⎠⎝ ⎠⎠
Based on a load-transfer analysis, the “average stresses” was determined.

The average vertical stress, σ v , is assumed to be equal to the boundary vertical
pressure, i.e.,
σ v = Pv (5.19)
The average horizontal stress, σ h , is:

L
∫σ x ( x)dx
σh = 0
(5.20)
L
or
⎛ υ ⎞ ⎛⎛ υ ⎞ ⎞⎛ β ⎛ tanh (α L ) ⎞ ⎞
σ h = ⎜⎜ s ⎟⎟ Pv − ⎜ ⎜⎜ s ⎟⎟ Pv − Ph ⎟ ⎜⎜1 + 2 ⎜ − 1⎟ ⎟⎟ (5.21)
⎜ 1−υ ⎟
⎝1 −υs ⎠ ⎝⎝ s ⎠ ⎠⎝ α ⎝ α .L ⎠⎠
253
Figure 5.8: An Idealized Plane-Strain GRS Mass for the SPR Model
Figure 5.9: Equilibrium of Differential Soil and Reinforcement Elements

(Reproduced from Hermann and Al-Yassin, 1978)
254
b. Average Reinforcement Forces in a GRS Mass
The equations of the forces in the reinforcement and the maximum force in the
reinforcement could be expresses as:
⎡⎛ υ ⎞ ⎤ ⎛1 −υs2
⎟⎟ Pv − Ph ⎥ ⎜
⎞⎛
⎟ ⎜1 − β ⎞ ⎛ cosh (α x ) ⎞
Fx = Ar E r ⎢⎜⎜ s ⎟ ⎜⎜1 − ⎟⎟ (5.22)
⎣⎝ 1 − υ s ⎠
⎜
⎦ ⎝ Es
⎟⎝ α2
⎠ ⎠ ⎝ cosh (α L ) ⎠
and
⎡⎛ υ ⎞ ⎤ ⎛1 −υs 2 ⎞⎛
⎟ ⎜1 − β ⎞⎛ 1 ⎞
Fmax = Ar E r ⎢⎜⎜ s ⎟⎟ Pv − Ph ⎥ ⎜ ⎟ ⎜⎜1 − ⎟⎟ (5.23)
⎣⎝ 1 − υ s ⎠
⎜
⎦ ⎝ Es
⎟⎝ α2
⎠ ⎠ ⎝ cosh (α L ) ⎠
The average force in the reinforcement may be calculated as:

L
∫ F dx
x
F= 0
(5.24)
L
Substituting Equation (5.22) into Equation (5.24), we have:
Ar E r L ⎡⎛ υ s ⎞ ⎤ ⎛1 −υs 2
⎟⎟ Pv − Ph ⎥ ⎜
⎞⎛
⎟ ⎜1 − β ⎞ ⎛ cosh (α x ) ⎞
L ∫0 ⎣⎜⎝ 1 − υ s
F = ⎢⎜ ⎟ ⎜⎜1 − ⎟⎟ dx (5.25)
⎠
⎜
⎦ ⎝ Es
⎟⎝ α2
⎠ ⎠ ⎝ cosh (α L ) ⎠
or
⎟ ⎜1 − β ⎞⎟ ⎜1 − cosh (α x ) ⎟ dx
A .E ⎡⎛ υ s ⎞ ⎤ ⎛1 −υs 2 ⎞⎛ L
⎛ ⎞
⎟⎟ Pv − Ph ⎥ ⎜
F = r r ⎢⎜⎜ ⎜ ⎟ ⎝ α 2 ⎠ ∫ ⎜ cosh (α L ) ⎟ (5.26)
L ⎣⎝ 1 − υ s ⎠ ⎦ ⎝ Es ⎠ 0⎝ ⎠
Thus,
A .E ⎡⎛ υ s ⎞ ⎤ ⎛1 −υs 2 ⎞⎛
⎟ ⎜1 − β ⎞⎛ sinh (α x ) ⎞
F = r r ⎢⎜⎜ ⎟⎟ Pv − Ph ⎥ ⎜ ⎟ ⎜⎜ L − ⎟ (5.27)
L ⎣⎝ 1 − υ s ⎠
⎜
⎦ ⎝ Es
⎟⎝ α2
⎠ ⎠⎝ α cosh (α L ) ⎟⎠
255
F
Introducing a factor r= , and from Equations (5.23) and (5.27), r can be
Fmax
determined as:
α L cosh (α L ) − sinh (α L )
r= (5.28)
α L (cosh (α L ) − 1)
Using the data from the calculation example in the SPR model (Ketchart and Wu,
2001), the values of factor r for different applied pressures and reinforcement lengths
are presented in Table 5.2.
Table 5.2: Values of factor r under different applied pressure and reinforcement
lengths
Increment of Reinforcement α r
Vertical Pressure Length, L (m)
F
(r = )
Fmax
Pv = 9.0 kPa 0.127 13.875 0.698
Pv = 18.0 kPa 0.127 14.616 0.701
Pv = 9.0 kPa 0.225 6.851 0.691
Pv = 18.0 kPa 0.225 6.966 0.692
It can be seen from Table 5.2, the average reinforcement forces are about 70% of the
maximum reinforcement force. The highest value of the maximum reinforcement
force, Fmax, can not exceed the tensile strength of reinforcement, Tf.
256
Fmax = T f (5.29)
and the average reinforcement force Tavg = 0.7 T f .
Equation 5.16 becomes:

⎛ Sv ⎞
⎜ ⎟
⎜ S ref ⎟
W = 0.7 ⎝ ⎠
(5.30)
5.2.2 Apparent Cohesion and Ultimate Pressure Carrying Capacity of a GRS

Mass
Substituting Equation 5.30 into Equation 5.15, the increased confining pressure in a
GRS mass becomes:
⎡ ⎛ Sv ⎞ ⎤
⎛ Tf ⎞ ⎢ ⎜⎜⎝ S ref ⎟⎟⎠ ⎥ ⎛ T f ⎞
Δσ 3 = W ⎜⎜ ⎟⎟ = 0.7 ⎜ ⎟⎟ (5.31)
⎝ Sv ⎠ ⎢⎢ ⎥ ⎜⎝ S v ⎠
⎣ ⎥⎦
Therefore, the apparent cohesion, CR, of a GRS composite can be evaluated as:
⎡ ⎛⎜ Sv ⎞⎟ ⎤ T
Δσ 3
K p + c = ⎢0.7 ⎝ ref ⎠ ⎥
⎜S ⎟
CR = Kp + c
f
(5.32)
2 ⎢ ⎥ 2 Sv
⎢⎣ ⎥⎦
where c = cohesion of soil

Kp = coefficient of passive earth pressure
Tf = extension strength of reinforcement
Sv = vertical spacing of reinforcement
Sref = the reference spacing and can be calculated by:
S ref = 6 d max (5.33)
(dmax = the maximum particle/grain size of soil)
Therefore, the ultimate pressure carrying capacity, σ 1R , of a soil-geosynthetic

composite mass is:
257
⎛ Tf ⎞
σ 1R = ⎜ σ 3 + W ⎟ Kp + 2 c Kp (5.34)
⎝ Sv ⎠
where σ 3 = confining pressure.
5.1 Verification of the Analytical Model with Measurement Data
Verification of the proposed analytical model for GRS composite strength properties
is made by comparing the model calculation results with the measured data from
GSGC tests (as presented in Chapter 4), and with the measured data by Elton and
Patawaran (2005).
5.1.1 Comparison between the Analytical Model and GSGC Test Results
The results of the GSGC tests have been reported in section 4.6. The dimensions of
the GSGC tests are 2 m high and 1.4 m wide in a plane strain condition. The soil
mass in the tests were reinforced with Geotext 4x4 geotextile at 0.2 m and 0.4 m
spacing. For the Diabase soil used in GSGC Tests, the maximum particle size was
about 1.3 in.; therefore, S ref = 6 d max = 7.8 in. (or 0.2 m). Comparisons of the results
between the analytical model and the GSGC tests are presented in Table 5.3. The
deviatoric stresses at failure calculated from the analytical model are in good
agreement with those of the GSGC tests. The differences between them are less than
10 %.
For reference purposes, comparisons of the results between Schlosser and Long’s
method and the GSGC tests are also presented (see Table 5.4). The deviatoric
stresses at failure calculated from the Schlosser and Long’s method are about 20 % to
86 % larger than the measured values.
258
Table 5.3: Comparison of the results between the analytical model and
the GSGC tests
Parameter Test 2 (T, S) Test 3 (2T, 2S) Test 4 (T, 2S)
Tf (kN/m) 70 140 70
Sv (m) 0.2 0.4 0.4
Δσ 3 (kN/m2) 245 172 86

by the Analytical Model
CR (kN/m2) 407 305 188

(σ 1R − σ 3 ) (kN/m2) 2,700 1,750 1,300

from Measured Data
(σ 1R − σ 3 ) (kN/m2) 2,460 1,900 1,250

Difference between -9% +8% -4%

the Analytical Model and
Measured Data
Note: Internal friction angle of soil, φ = 50o; cohesion of soil, c = 70 kPa.
259
Table 5.4: Comparison of the results between Schlosser and Long’s method and
the GSGC tests
Parameter Test 2 (T, S) Test 3 (2T, 2S) Test 4 (T, 2S)
Tf (kN/m) 70 140 70
Sv (m) 0.2 0.4 0.4
Δσ 3 (kN/m2) 350 350 175

by Schlosser & Long’s Method
CR (kN/m2) 550 550 310

(σ 1R − σ 3 ) (kN/m2) 2,700 1,750 1,300

from Measured Data
(σ 1R − σ 3 ) (kN/m2) 3,250 3,250 1,930

Difference between + 20 % + 86 % + 48 %
Schlosser & Long’s Method
and Measured Data
Note: Internal friction angle of soil, φ = 50o; cohesion of soil, c = 70 kPa.
260
5.3.2 Comparison between the Analytical Model and Elton and Patawaran’s
Test Results
Elton and Patawaran (2005) conducted seven large-size triaxial tests for reinforced
soil with the dimensions of 5 ft in height and 2.5 ft in diameter (see Figure 5.10). All
the tests were conducted under an unconfined condition. The properties of the tests
are summarized as follows:
• Backfill: The soil used in the test was a poorly graded sand with the gradation
test results shown in Figure 5.11; maximum dry unit weight γdry = 121 pcf;
optimum moisture content wopt = 9.3%; internal friction φ = 40o; and cohesion
c = 4 psi.
• Reinforcement: Six types of reinforcement (TG500, TG600, TG 700, TG800,
TG1000 and TG028) were used for the tests with the reinforcement spacing of
6 in. and 12 in. The strength, Tf, of the reinforcement and reinforcement
spacing, Sv, are shown in Table 5.5.
The maximum particle size (from gradation tests, Figure 5.11) of the backfill in the
large-size triaxial tests was d max = 0.5 in. (or 12.7 mm); therefore,
S ref = 6 d max = 3 in. (or 0.08 m). The measured results are shown in Figure 5.12.
The comparisons of Elton and Patawaran’s tests results with the analytical model and
with Schlosser and Long’s Method are presented in Tables 5.5 and 5.6, respectively.
The differences in the deviatoric stresses at failure calculated from the analytical
model and Elton and Patawaran’s measured data are less than 18 %. Whereas, the
results calculated from Schlosser and Long’s method are 69 % to 97 % larger than the
measured values.
261
Figure 5.10: Reinforced Soil Test Specimen before Testing
262
Figure 5.11: Backfill Grain Size Distribution before and after Large-Size Triaxial
Tests (Elton and Patawaran, 2005)
Figure 5.12: Large-Size Triaxial Test Results (Elton and Patawaran, 2005)
263
Elton and Patawaran’s tests (2005)
Reinforcement TG TG TG TG TG TG TG
Type 500 500 600 700 800 1000 028
Tf (kN/m) 9 9 14 15 19 20 25
Sv (m) 0.15 0.30 0.15 0.15 0.15 0.15 0.15
Δσ 3 (kN/m2)
by the Analytical 30 8 47 48 62 67 83
Model
CR (kN/m2)
60 36 78 79 94 99 116
by the Analytical
Model
(σ 1R − σ 3 ) (kN/m2)
from Measured 230 129 306 292 402 397 459
Data
(σ 1R − σ 3 ) (kN/m2)
256 153 333 341 402 426 498
by the Analytical
Model
Difference between
the Analytical 11 % 18 % 9% 17 % 0% 7% 8%
Model and
Measured Data
Note: Internal friction angle of soil, φ = 40o; cohesion of soil, c = 27.6 kPa.
264
Table 5.6: Comparison of the results between Schlosser and Long’s method and
Elton and Patawaran’s tests (2005)
Reinforcement TG TG TG TG TG TG TG
Type 500 500 600 700 800 1000 028
Tf (kN/m) 9 9 14 15 19 20 25
Sv (m) 0.15 0.30 0.15 0.15 0.15 0.15 0.15
Δσ 3 (kN/m2)
by Schlosser & 59 30 92 95 122 132 163
Long’s Method
CR (kN/m2)
91 59 126 130 158 169 202
by Schlosser &
Long’s Method
(σ 1R − σ 3 ) (kN/m2)
230 129 306 292 402 397 459
From Measured
Data
(σ 1R − σ 3 ) (kN/m2)
390 254 541 557 678 726 868
by Schlosser &
Long’s Method
Difference between
Schlosser & Long’s 70 % 97 % 77 % 91 % 69 % 83 % 89 %
Method and
Measured Data
Note: Internal friction angle of soil, φ = 40o; cohesion of soil, c = 27.6 kPa.
265
5.3.3 Comparison of the Results between the Analytical Model and Finite
Element Results
Finite element analyses were conducted to provide additional data for verifying the
analytical model. The test conditions and material properties used for the finite
element analyses were the same as those used in GSGC Test 2, but with confining
pressures of 34 kPa, 70 kPa, 100 kPa and 200 kPa. The confining pressure used in
GSGC Test 2 was 34 kPa. The comparison indicates that the results of the analytical
model are in good agreement with those obtained from the finite element analyses at
the different confining pressures. The largest difference in terms of the deviatoric
stress at failure is 9%.
266
results for GSGC Test 2 with different confining pressures from FE analyses
Parameter S3 = 34 S3 = 70 S3 = 100 S3 = 200

kPa kPa kPa kPa
Δσ 3 (kN/m2)
245 245 245 245
CR (kN/m2)
407 407 407 407
(σ 1R − σ 3 ) (kN/m2)
2,700 2,970 3,190 3,860
from FE Analysis
(σ 1R − σ 3 ) (kN/m2)
2,490 2,760 2,990 3,740
Difference between
-8% -7% -6% -3%
the Analytical Model and FE
analyses
Note: Internal friction angle of soil, φ = 50o; cohesion of soil, c = 70 kPa;

reinforcement strength, Tf = 70 kN/m; reinforcement spacing, Sv = 0.2 m.
267
5.4 Required Reinforcement Strength in Design
In this Section, an analytical model for determining required tensile strength in

reinforcement is developed, a comparison of the analytical model with current design
equation is made, and verification of the analytical model is presented.
5.4.1 Proposed Model for Determining Reinforcement Force
In current design methods the following equation is used to determine the required
reinforcement strength, Trequired, for the design of GRS structures:
Trequired = σ h ∗ S v ∗ Fs (5.35)
where
Trequired = required strength for reinforcement at depth z
σh = horizontal stress in a GRS mass at depth z
Fs = safety factor
Assuming Fs = 1, we have Trequired = T f (ultimate strength of reinforcement) and
Equation 5.35 becomes

Tf
=σh (5.36)
Sv
Tf
Note that when the horizontal stress, σ h , is a constant, the ratio becomes a
Sv
constant; i.e., Tf is linearly proportional to Sv. Using Equation 5.15 a new expression
for the increase of confining pressure due to tensile inclusion, a modified equation for
determination of required reinforcement strength can be obtained. The derivation of
268
the modified equation is described as follows. The horizontal stress, σ h , in a GRS
structure at depth z is:
σ h = σ 3 + Δσ 3 (5.37)
or Δσ 3 = σ h − σ 3 (5.38)
Since,
Tf
Δσ 3 = W (see Section 5.2.1) (5.39)
Sv
Substituting Equation 5.39 into Equation 5.36 leads to
Tf
σh − σ3 =W (5.40)
Sv
Tf (σ h − σ 3 )
or = (5.41)
Sv W
Since,
⎛ Sv ⎞
⎜ ⎟
⎜6d ⎟
W = 0.7 ⎝ max ⎠
(5.42)
Therefore,
Tf (σ h − σ 3 )
= ⎛ Sv ⎞
(5.43)
Sv ⎜
⎜6d
⎟
⎟
⎝ ⎠
0.7 max
where
Tf = ultimate strength of reinforcement at depth z
σ h = horizontal stress in a GRS mass at depth z
σ 3 = lateral constraint pressure at depth z, lateral earth pressure exerted by
external constraint. For a GRS wall with modular block facing, σ 3
can be estimated as:
σ 3 = γ b b tan δ (5.44)
269
where
γb = unit weight of facing block
b = width of facing block
δ= friction angle between modular block facing elements (δ can be
the friction angle between facing blocks if there is no
reinforcement between the blocks, or it can be the friction
angle between facing block and geosynthetic if there is
reinforcement sandwiched between blocks)
Sv = reinforcement spacing
dmax = maximum grain size of the backfill
The required tensile strength of the reinforcement in design can be expressed as:
⎡ ⎤
⎢ σh −σ3 ⎥
Trequired = ⎢ ⎛ S ⎞ ⎥ * S v * Fs (5.45)
⎜ v ⎟
⎢ 0.7 ⎜⎝ 6 d max ⎟⎠ ⎥
⎣ ⎦
Note that Trequired is always equal or greater than zero. For a GRS mass without
lateral constraint (e.g., a wrapped wall), σ 3 = 0, and Equation 5.45 becomes
⎡ ⎤
⎢ σh ⎥
Trequired = ⎢ ⎛ S ⎞ ⎥ * S v * Fs (5.46)
⎜ v ⎟
⎢ 0.7 ⎜⎝ 6 d max ⎟⎠ ⎥
⎣ ⎦
5.4.2 Comparison of Reinforcement Strength between the Analytical Model

and Current Design Equation
A comparison of reinforcement forces is made between the proposed model

(Equation 5.45) and the current design equation (Equation 5.35). The reinforcement
270
forces in a 6.0 m high GRS wall, as determined by the two equations, are shown in
Table 5.8. Note that the facing stiffness is ignored in the current design method;
while the facing rigidity, as denoted by the lateral constraint pressure, σ3, can be
accounted for in the proposed model.
5.4.3 Verification of the Analytical Model for Determining Reinforcement

Strength
Verification of the analytical model for determination of reinforcement strength is

made by comparing the results with the forces in reinforcement at failure from the
GSGC tests (see Chapter 4), with the measured data by Elton and Patawaran (2005),
and with a typical GRS wall.
Table 5.9 shows a comparison of the results from the proposed model and measured
data from the GSGC tests. The largest difference in reinforcement forces between the
two is 16 %, whereas there is 47 % between the current design methods and the test
data.
A comparison of the reinforcement forces between the proposed model and the
measured data from Elton and Patawaran’s tests (2005) is shown in Table 5.10. The
largest difference in reinforcement forces between the two is 13 %, whereas it was as
high as 74 % between the current design method (Equation 5.35) and the test results.
The proposed model clearly gives a much improved value for estimating
reinforcement strength compared to the current design methods.
271
Table 5.8: Comparison of reinforcement forces between proposed model and
current design equation for a GRS wall
Trequired (kN/m), for Fs = 1
Current Design Proposed Model

z (m) Equation
No Facing Modular Block Modular Block
Facing with Facing with
 bb = 35o  bb = 54o
0.4 0.3 1.1 0 0

0.8 0.6 1.5 0.1 0
1.2 1.0 2.0 0.5 0
1.6 1.3 2.4 1.0 0
2.0 1.6 2.9 1.4 0
2.4 1.9 3.3 1.9 0.1
2.8 2.3 3.7 2.3 0.6
3.2 2.6 4.2 2.8 1.0
3.6 2.9 4.6 3.2 1.5
4.0 3.2 5.1 3.6 1.9
4.4 3.6 5.5 4.1 2.4
4.8 3.9 6.0 4.5 2.8
5.2 4.2 6.4 5.0 3.2
5.6 4.5 6.8 5.4 3.7
6.0 4.9 7.3 5.8 4.1
Note: Internal friction angle of soil,  = 38o; cohesion of soil, c = 0; reinforcement

spacing, Sv = 0.2 m; maximum grain size of soil, d max  38 mm ; unit weight of soil,
 backfill  17 kN/m3; unit weight of facing block,  block  25 kN/m3;  b b = friction
angle between facing blocks; width of blocks, b = 0.3 m.
272
Table 5.9: Comparison of reinforcement forces between proposed model and
the GSGC tests
Parameter Test 2 Test 3 Test 4 Test 5

(T, S) (2T, 2S) (T, 2S) (T, S)
Reinforcement Force at
Failure Tf (kN/m) 70 140 70 70
Reinforcement Spacing
Sv (m) 0.2 0.4 0.4 0.2
Measured Failure Pressure

(kPa) 2,700 1,750 1,300 1,900
Lateral Constraint Pressure

σ 3 (kPa) 34 34 34 0
Maximum Reinforcement
Force from Current Design 62.4 74.4 50.5 41.2
Equation, Equation 5.35
(kN/m)
Difference between
- 11 % - 47 % - 28 % - 41 %
Current Design Equation
(Equation 5.35) and Tf
Maximum Reinforcement
Force from Proposed 79.4 124.1 75.4 58.8
Model, Equation 5.45
(kN/m)
Difference between
+ 13 % - 11 % +8% -16 %
Proposed Model (Equation
5.45) and Tf
Note: Internal friction angle of soil, φ = 50o; cohesion of soil, c = 70 kPa; unit weight
of soil, γ backfill = 24 kM/m3; maximum grain size of soil, d max = 33 mm .
273
Table 5.10: Comparison of reinforcement forces between proposed model and test
data from Elton and Patawaran (2005)
Parameter TG TG TG TG TG TG TG
500 500 600 700 800 1000 028
Reinforcement Force
at Failure Tf (kN/m) 9 9 14 15 19 20 25
Reinforcement 0.15 0.30 0.15 0.15 0.15 0.15 0.15

Spacing Sv (m)
Measured Failure
Pressure (kPa) 230 129 306 292 402 397 459
Maximum
Reinforcement Force 4.47 2.35 6.95 6.49 10.08 9.91 11.94
from Current Design
Equation, Equation
5.35 (kN/m)
Difference between - 50 - 74 - 50 - 57 - 47 - 50 - 52
Current Design % % % % % % %
Equation (Equation
5.35) and Tf
Maximum
Reinforcement Force 9.02 9.56 14.02 13.10 20.34 20.01 24.09
from Proposed Model,
Equation 5.45 (kN/m)
Difference between
0 +6 0 -13 +7 0 -4
Proposed Model
% % % % % % %
(Equation 5.45) and Tf
Note: Internal friction angle of soil, φ = 40o; cohesion of soil, c = 27.6 kPa; unit
weight of soil, γ backfill = 18.8 kM/m3; lateral constraint pressure, σ 3 = 0 kPa;
maximum grain size of soil, d max = 12.7 mm .
274
6. FINITE ELEMENT ANALYSES
In this chapter, the GSGC tests described in Chapter 4 were simulated using the Finite
Element (FE) method of analysis. The behavior of the reinforced soil mass in GSGC
Test 2 under different confining pressures, which was not part of the experimental
program, was investigated by using FE model. The angle of dilatation of a GRS
composite was investigated. In addition, the analytical model developed for
evaluation of compaction-induced stresses in a GRS mass was verified by comparing
the results with those obtained from the FE analysis.
Many finite element codes are readily available for the analysis of soil-structure
interaction problems, including Abacus, FLAC, LS-Dyna, Plaxis, Sage Crisp, Sigma
(Geoslope). In this study, Plaxis 8.2 code was selected for the analysis, due primarily
to the author’s familiarity with the code. This code has been used successfully for the
analysis of various earth structures, including GRS structures (e.g., Bueno, et al.,
2005), Christopher, et al., 2005, and Morison, et al., 2007).
6.1 Brief Description of Plaxis 8.2
Version 8.2 of Plaxis is a finite element code intended for two-dimensional analysis
of deformation and stability problems in geotechnical engineering. The details of
Plaxis 8.2 program can be found in the Plaxis manual (Plaxis, 2002). A brief
description of some key features of the program is presented bellow:
Graphical input of geometry models: The input of layers, structures, construction
stages, loads and boundary conditions is based on CAD drawing procedures, which
275
allows for a detailed modeling of the geometry. From the geometry model, a 2-
dimensional finite element mesh is easily generated.
Automatic mesh generation: Plaxis allows for automatic generation of 2-dimensional
finite element meshes with options for global and local mesh refinement.
High-order elements: Quadratic 6-node and 4th order 15-node triangle elements are
available to describe the stress-deformation behavior of the soil.
Plates/Beams: Special beam elements can be used to model the bending of retaining
walls, tunnel linings, shells, and other slender structures. The behavior of these
elements is defined using a flexural rigidity, a normal stiffness and an ultimate
bending moment. Plates with interfaces can also be used to model the behavior of
structures.
Interfaces: Joint elements are available to model soil-structure interface behavior.
For example, joint elements can be used to simulate the thin zone of intensely
shearing material at the contact between a retaining wall and the surrounding soil, a
tunnel lining and the soil, or between reinforcement and surrounding soil. Values of
the interface friction angle and adhesion are generally not the same as the friction
angle and cohesion of the surrounding soil.
Anchors: Elastoplastic spring elements can be used to model anchors and struts. The
behavior of these elements is defined using a normal stiffness and a maximum force.
A special option exists for the analyses of prestressed ground anchors and excavation
supports.
Geogrids/geotextiles: Geogrids or geotextiles in GRS structures can be simulated in
Plaxis by special tension elements. It is often convenient to combine these elements
with joint elements to model the interaction between geosynthetic reinforcement and
the surrounding soil.
Tunnels: The Plaxis program offers a convenient option to create circular and non-
circular tunnels using arcs and lines. Plates and interfaces may be used to model the
tunnel lining and the interaction with the surrounding soil. Fully isoparametric
276
elements can be used to model the curved boundaries within the mesh. Various
methods are available for the analysis of the deformation caused by various methods
of tunnel construction.
Mohr-Coulomb model: This non-linear model is based on soil parameters that are
well-known in geotechnical engineering practice. Not all non-linear features of soil
behavior are included in this model; however, the Mohr-Coulomb model can be used
to compute realistic support pressures for tunnel faces, ultimate loads for footings,
etc. It can also be used to calculate a safety factors using a “phi-c reduction’
approach.
Advanced soil models: In addition to the Mohr-Coulomb model, Plaxis offers a
variety of advanced soil model. A general second-order model, an elastoplastic type
of hyperbolic model, called the Hardening Soil model, is available. To model
accurately the time-dependent and logarithmic compression behavior of normally
consolidated soft soils, a creep model is also available, which is referred to as the Soft
Soil Creep model. In addition, a special model is available for the analysis of
anisotropic behavior of jointed rock.
User-defined soil models: A special feature in Plaxis 8.2 is that it has a user-defined
soil model option. This feature enables users to include self-programmed soil models
for the analysis.
Staged construction: This feature enables realistic simulation of construction of earth
structures by activating and deactivating compaction loads, and simulation of
excavation processes by activating and deactivating clusters of elements, application
of loads, changing of water tables, etc. The procedure allows for a realistic
assessment of stresses and displacements caused, for example, by compaction loads
or soil excavation during underground construction.
277
6.2 Compaction-Induced Stress in a GRS Mass
The residual lateral stresses in a GRS mass due to compaction can be evaluated by the
following equation (see Section 5.15 for details):
⎛ 0.7 E r ⎞
Δσ 3 = Δσ v , c , max K i , c F ⎜⎜ 1 + ⎟⎟ (6.1)
⎝ E s S v − 0.7 E r ⎠
where Δσ v ,c ,max is the maximum vertical stress due to compaction loading;
F =1−
(OCR − OCR ) ; sin φ
Es and Er are soil and reinforcement stiffness,
OCR − 1
respectively; and Sv is reinforcement spacing.
To illustrate how to compute the residual lateral stresses, or compaction-induced

stresses (CIS), in a GRS mass, a 6-m high GRS mass was chosen as an example. The
parameters used for the calculation of CIS are:
• Soil: A dense sand with unit weight γ = 17 kN/m3; angle of internal friction φ
= 45o; soil modulus, E s = 30,000 kPa; compaction lift S = 0.2 m.
• Geosynthetics: E r = 2,000 kPa; Sv = 0.2 m.

The vertical maximum pressures of 44 kPa, 100 kPa, 200 kPa, 300 kPa and 500 kPa
due to compaction were used for the calculation of the residual stresses. Note that the
maximum vertical compaction stress of 44 kPa was the contact pressure used in the
GSGC Tests described in Chapter 4.
Figure 6.1 shows the distribution of the lateral residual stresses with depth due to the
different maximum compaction pressures. The values of the lateral residual stresses
in this figure were calculated based on Equation 6.1 with the assumption that the
compaction lift of 0.2 m is small compared to the dimensions of the compaction plant.
As a result, the distribution of the lateral residual stress within the 0.2 m-thick lift due
278
to compaction is constant. Near the surface, the lateral residual stresses will follow
the limiting lateral earth pressure for unloading condition, i.e., the K1,c-line, as shown
in Figure 5.2 and Table 5.1.
Residual Lateral Stress (kPa)

0
0 25 50 75 100
-1
-2 Maximum Vertical
Pressure due to
Compaction:
Depth (m)
44 kPa
-3
100 kPa
-4 200 kPa
300 kPa
-5 500 kPa
-6
Figure 6.1: Distribution of Residual Lateral Stresses of a GRS mass with Depth
due to fill compaction
279
6.3 Finite Element Simulation of the GSGC Tests
The Finite Element (FE) program Plaxis 8.2 was used for the simulation of GSGC
Tests 1, 2 and 3, as described in Chapter 4. The backfill of all GSGC test specimens
was compacted at 0.2 m lifts using a plate compactor MBW-GP1200 with a contact
pressure of 44 kPa. The soil stiffness, E50ref , can be estimated from triaxial tests using
Janbu’s equation (1963). The Poisson’s ratio of the backfill can be estimated as:
ΔV σ
• = ε x + ε y + ε z = (1 − 2ν ) z (for uniaxial tests)
V E
ΔV (1 + ν ) (1 − 2ν ) σ z
• = εx +εy +εz = (for triaxial Tests)
V E (1 − ν )
For GSGC Test 1-unreinforced soil, the soil model and parameters used in the
analysis are as bellows:
• Soil model: Hardening model
• Dry unit weight γ = 24 kN/m3, wet unit weight γ = 25 kN/m3
• Angle of internal friction, φ = 50o; angle of dilation, ψ = 17.5o, cohesion c =
70 kPa (for confining pressure < 30 psi)
• Soil modulus: E50ref = 62,000 kPa, Eurref = 124,000 kPa, Poisson’s ratio ν =
0.37
• Power m = 0.5.
For the GSGC Tests, the backfill stiffness was increased due to CIS as described
below:
1. Place the first lift of facing blocks and soil. The compaction lift thickness was
0.2 m.
280
2. Apply a uniform vertical stress of 44 kPa over the entire surface of each newly
placed soil layer before analysis, and remove it afterwards (see Table 6.2).
This step was employed to simulate the compaction operation.
3. Place a sheet of reinforcement at pre-selected reinforcement spacing to cover
the entire backfill surface plus the facing blocks.
4. Place another lift of facing blocks and soil.
5. Repeat Steps 2 to 4 until the fill reached the total height of the specimen.
6. Remove facing blocks
7. Apply a prescribed confining pressure.
8. Apply vertical stresses on the top of the specimen at equal increments until
failure.
The properties of the backfill were modified due to compaction effects. The
Poisson’s ratio under plane strain condition was reduced to ν min , as suggested by
Hatami and Bathurst (2006). Soil modulus E50ref was calculated using the increase in
confining pressure due to compaction. The value of E50ref can be estimated by using
Equation 5.12 with σ '3 = σ '3 S + Δσ 3 , where Δσ 3 is the average residual lateral stress,
and can be estimated by Equation 6.1. Note that the elastic modulus was increased by
factor of 2.25 for walls 1 and 2 in Hatami and Bathurst’s numerical analyses (2006)
and by factor of 10 in the FE analyses by Morrison, et al. (2006).
The conditions and properties of the backfill and reinforcement of GSGC Tests 2 and
3 used in FE analyses are shown in Table 6.1.
281
Table 6.1: Parameters and properties of the GSGC Tests used in analyses
Description
Material: Diabase; Soil model: Hardening soil model; dry unit
Soil weight, γd = 24 kN/m3; wet unit weight, γw = 25 kN/m3;

cohesion, c = 70 kPa; angle of internal friction, φ = 50o; angle
of dilation, ψ = 17o; soil modulus, E50ref = 63,400 kPa,
Eurref = 126,800 kPa; Poisson’s ratio ν = 0.2; power, m = 0.5.
GSGC Single-sheet Geotex 4x4: axial stiffness, EA = 1,000

Test 2 kN/m; ultimate strength, Tult = 70 kN/m;
reinforcement spacing, Sv = 0.2 m;
Reinforcement GSGC Double-sheet Geotex 4x4: axial stiffness, EA = 2,000

Test 3 kN/m; ultimate strength, Tult = 140 kN/m;
reinforcement spacing, Sv = 0.4 m
FE Model: linear elastic model; modulus, E = 3*107 kPa; unit

Facing Block
weight, γ = 12.5 kN/m3 (hollow blocks); Poisson’s ratio, ν = 0.
Block-Block FE Model: Mohr-Coulomb model; modulus, E = 3*106 kPa;

Interface unit weight, γ = 0 kN/m3; cohesion, c = 2 kPa; angle of internal
friction, φ = 33o; Poisson’s ratio, ν = 0.45.
Confining Constant confining pressures of 34 kPa for the GSGC Tests 1 to

Pressure 3.
Note: Soil-reinforcement interface was assumed to be fully bonded in the analyses.
282
Table 6.2 shows the steps in the finite element analyses. The first 20 steps were used
for modeling the preparation of the specimen. The loading began from step 21.
283
Table 6.2: The steps of analysis for the GSGC Tests
Step Configuration
284
Table 6.2 (continued): The steps of analysis for the GSGC Tests
Step Configuration
20
Note: The full height of the GSGC mass was reached at the last step of the
specimen preparation
285
Table 6.2 (continued): The steps of analysis for the GSGC Tests
Step Configuration
21
Note: The facing blocks were removed, a confining pressure and vertical
pressures were applied (during loading)
286
6.3.1 Simulation of GSGC Test 1
The global stress-strain relationship and volume change relationship as obtained from
FE analysis and the GSGC tests are shown in Figure 6.2. The analysis results are in
good agreement with the measured data. The maximum differences of the results
between the FE analysis and the tests were about 5 %. Figure 6.3 shows the lateral
movements on the open faces of the specimen at 200 kPa, 400 kPa, 600 kPa and 770
kPa. The results from FE analyses and the tests are also in good agreement.
287
900
800
Applied Vetical Stress (kPa)
700
600
500
400
300
200 FE Analysis
100
Measured Data
0
0 1 2 3 4 5 6 7
(a)
1.4
1.2
1
0.8
0.6
0.4 Mearsued Data

0.2
FE Analysis
0
0 0.5 1 1.5 2 2.5 3 3.5
-0.2
-0.4
-0.6
(b)
Figure 6.2: Comparison of Results for GSGC Test 1:

(a) Global Vertical Stress-Strain Relationship, and
(b) Volume Change Relationship
288
2
1.75
Applied Pressure
1.5
200 kPa
(Measured)
Specimen Height (m)
1.25 400 kPa

(Measured)
600 kPa
1 (Measured)
770 kPa
(Measured)
0.75
200 kPa (FE)
0.5 400 kPa (FE)
600 kPa (FE)

0.25
770 kPa (FE)
0
0 10 20 30 40 50
Figure 6.3: Comparison of Lateral Displacements on Open Face of GSGC Test 1
289
6.3.2 Simulation of GSGC Test 2
The global stress-strain relationships, as obtained from FE analysis and the GSGC
tests, are shown in Figure 6.4. The results of FE analysis with and without
consideration of CIS are included in Figure 6.4. It is seen that the results with
consideration of CIS gives slightly better simulation of stress-strain curve. It should
be noted that the compaction energy used in the GSGC tests was very low. As a
result, the magnitude of CIS was very small, and the effect of CIS on the global
stress-strain relationship was not significant.
Figure 6.5 shows the lateral displacements on the open faces of the specimen at
applied pressures of 400 kPa, 1,000 kPa, 2,000 kPa and 2,500 kPa. The simulated
lateral displacements are no more than 5% greater than the measured values.
Figure 6.6 shows the simulated and measured displacements of the GSGC mass at
selected points in Figure 4.49. It is seen that the simulated displacements are in
agreement with the measured values.
A comparison of the distribution of strains in the reinforcement in GSGC Test 2

between the FE analyses and measured data are shown in Figure 6.7. It is seen that
the simulated strains are in good agreement with the measured values.
290
3500
3000
2500
Deviatoric Stress (kPa)
2000
1500
1000
Measured Data
500
FE Analysis-w ithout CIS
FE Analysis-w ith CIS

0
0 2 4 6 8 10
Figure 6.4: Comparison of Global Stress-Strain Relationship of GSGC Test 2
291
2
2500 kPa
1.5 (Measured)
2000 kPa
Specimen Height (m)
(Measured)
1.25
1000 kPa
(Measured)
1
400 kPa
(measured)
0.75 2500 kPa (FE)
2000 kPa (FE)

0.5
1000 kPa (FE)

0.25
400 kPa (FE)
0
0 20 40 60
Figure 6.5: Comparison of Lateral Displacement at Open Face of GSGC Test 2
292
0
-80 -60 -40 -20 0 20 40 60 80 Point 1 (Measured)
Point 2 (Measured)
-20
Point 3 (Measured)
-40 Point 4 (Measured)
Point 6 (Measured)
y (mm)
Point 8 (Measured)
-80
Point 9 (Measured)
Point 1 (FE)
-100
Point 4 (FE)
-120 Point 7 (FE)
Point 3 (FE)
-140 Point 6 (FE)
Point 9 (FE)
-160
Note: Locations of Points
x (mm) are show n in Figure 4.49
Figure 6.6: Comparison of Internal Displacements of GSGC Test 2
293
Applied Pressure
2.5
200 kPa
(Measured)
2 600 kPa
(Measured)
Strain (%)
1.5 800 kPa

(Measured)
1000 kPa
1 (Measured)
200 kPa (FE)
0.5
600 kPa (FE)
0
800 kPa (FE)
0.0 0.1 0.2 0.3 0.4 0.5 0.6
Distance from the Edge of the Composite Mass (m) 1000 kPa (FE)
(a)
Applied Pressure:
3.5
200 kPa
3 (Measured)
600 kPa
2.5 (Measured)
Strain (%)
800 kPa
2 (Measured)
1.5 1000 kPa
(Measured)
1 200 kPa (FE)
0.5 600 kPa (FE)

0
800 kPa (FE)
0.0 0.1 0.2 0.3 0.4 0.5 0.6
Distance from the Edge of the Composite Mass (m) 1000 kPa (FE)
(b)
Figure 6.7: Comparison of Reinforcement Strains of GSGC Test 2:

(a) At reinforcement Layer 1.6 m from the Base, and
(b) At reinforcement Layer 0.8 m from the Base
294
6.3.3 Simulations of GSGC Test 3
A comparison of the global stress-strain relationships of GSGC Test 3, as obtained

from FE analysis and measured data of GSGC Test 3, are shown in Figure 6.8. The
FE results are in good agreement with the measured data. The lateral displacements
on the open faces at applied pressures of 400 kPa, 600 kPa, 800 kPa, 1000 kPa and
1250 kPa are shown in Figure 6.9. Figure 6.10 shows a comparison of simulated and
measured internal displacements of the GSGC Test 3 specimen at selected points in
Figure 4.57. Once again, the results from FE analyses are in very good agreement
with the measured values.
The comparisons of the distribution of strains in the reinforcement in GSGC Test 3

between FE analyses and measured data are shown in Figure 6.11. It is seen that the
simulated strains are in agreement with the measured values.
295
2,000
1,800
1,600
1,400
1,200
1,000
800
600
400
Measured Data
200
FE Analyses
0
0 1 2 3 4 5 6 7 8
Figure 6.8: Comparison of Global Stress-Strain Relationship of GSGC Test 3
296
2
1.75
Applied Pressure:
400 kPa (Measured)

1.5
600 kPa (Measured)

Specimen Height (m)
1.25 800 kPa (Measured)
1,000 kPa (Measured)

1
1,250 kPa (Measured)
0.75 400 kPa (FE)
600 kPa (FE)

0.5
800 kPa (FE)
0.25 1,000 kPa (FE)
1,250 kPa (FE)

0
0.0 10.0 20.0 30.0
Lateral Movement (mm)
Figure 6.9: Comparison of Lateral Displacement at Open Face of GSGC Test 3
297
0
-80 -60 -40 -20 0 20 40 60 80 Point 1 (Measured)
-20
Point 2 (Measured)
Point 4 (Measured)
-60
y (mm)
Point 6 (Measured)
-80
Point 1 (FE)
-100 Point 3 (FE)
Point 4 (FE)
-120
Point 6 (FE)
-140
Note: Locations of Points
x (mm)
are show n in Figure 4.57
Figure 6.10: Comparison of Internal Displacements of GSGC Test 3
298
Applied Pressure
4.5
260 kPa
4.0 (Measured)
3.5 600 kPa
(Measured)
3.0
Strain (%)
1000 kPa
2.5 (Measured)
2.0 260 kPa (FE)
1.5 600 kPa (FE)

1.0
0.5 1000 kPa (FE)
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6
(a)
Applied Pressure
3.5
260 kPa
3 (Measured)
600 kPa
2.5 (Measured)
Strain (%)
1000 kPa
2 (Measured)
` 200 kPa (FE)
1.5
1 600 kPa (FE)
0.5 1000 kPa (FE)
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6
(b)
Figure 6.11: Comparison of Reinforcement Strains of GSGC Test 3:

(a) At reinforcement Layer 1.2 m from the Base, and
(b) At reinforcement Layer 0.4 m from the Base
299
6.4 FE Analysis of GSGC Test 2 under Different Confining Pressures and
Dilation Angle of Soil-Geosynthetic Composites
Figure 6.12 shows a comparison of the global stress-strain and volume change
relationships of GSGC Test 2 as obtained from FE analysis and the measured data.
GSGC Test 2 was conducted under a confining pressure of 34 kPa. It is seen that the
global stress-strain and volume change relationships under the confining pressure of
34 kPa as obtained from FE analysis are in good agreement with the measured data.
To provide additional data under different confining pressures, the FE model was
used to generate data under confining pressures of 100 kPa and 200 kPa, as shown in
Figure 6.12.
It is interesting to note that the reinforcing mechanism of GRS mass can be viewed in
terms of the angle of dilation. The angle of dilation of a geosynthetic-reinforced soil
mass is smaller than the angle of dilation of an unreinforced soil mass. Using the data
in Figure 6.12 as an example, the angles of dilation of the soil-geosynthetic
composites were approximately -8o, -11o and -12o (a negative dilation angle means no
dilation, the greater the absolute value the less likely the material will dilate) under
confining pressures of 34 kPa, 100 kPa and 200 kPa, respectively; whereas the angle
of dilation of the reinforced soil was +17o under a confining pressure up to 200 kPa.
This suggests that the presence of geosynthetic reinforcement has a tendency to
“suppress” dilation the surrounding soil. A soil having less tendency to dilate will
become stronger. The dilation behavior offers a new explanation of the reinforcing
mechanism, and the angle of dilation provides a quantitative measure of the degree of
reinforcing effect of a GRS mass.
300
4,500
4,000
3,500
3,000
2,500
Confining Pressure:
2,000
34 kPa (Measured)
1,500
34 kPa (FE)
1,000
100 kPa (FE)
500
200 kPa (FE)
0
0 1 2 3 4 5 6 7 8 9
(a)
Confining Pressure:
2
34 kPa (T2)
1
34 kPa (FE)
0
0 2 4 6 8 10
-1 100 kPa (FE)
-2
200 kPa (FE)
-3
34 kPa (Soil - T1)
-4
(b)
Figure 6.12: FE Analyses of GSGC Test 2 under Different Confining Pressures:

(a) Global Stress-Strain Relationship, and
(b) Volume Change Curves
301
6.5 Verification of Compaction-Induced Stress Model
The analytical model for evaluating compaction-induced stresses (CIS), as described

in Section 3.2, was verified using the finite element (FE) method of analysis. The FE
analysis was carried out by using Version 8.2 of Plaxis code (2002). To verify CIS
model, a 6 m-high GRS mass was chosen as an example. The parameters used for the
calculation of CIS in the analytical model and the FE analysis are:
• Soil: a dense sand with mass unit weight γ = 17 kN/m3; angle of internal
friction φ = 45o; loading modulus, E s = 30,000 kPa; unloading modulus,
Eur = 90,000 kPa; Poisson ratio, ν = 0.2; compaction lift S = 0.2 m (Note:
The behavior of soil was simulated in FE analysis by using the Hardening Soil
model in plane strain condition.)
• Geosynthetics: tensile modulus E r = 2,000 kPa; reinforcement spacing, Sv =
0.2 m.
• Interface: the interface between the soil and geosynthetic reinforcement is
fully bonded.
The very fine mesh of a FE analyses to simulate CIS in a GRS mass is shown in
Figure 6.13. Figure 6.14 shows the lateral stress distributions at the center line of the
GRS mass without considering CIS and with CIS under the maximum compaction
pressures of 200 kPa and 500 kPa. The compaction operation was simulated by
loading and unloading at different locations on the surface area of each lift. The
“residual lateral stresses” were the differences between the lateral stresses with
simulating CIS and those without CIS at the same location as shown in Figure 6.14.
Comparisons of residual lateral stresses distribution resulting from compaction

pressures of 200 kPa and 500 kPa between the CIS model and the FE analysis are
302
shown in Figures 6.15. It is seen that the compaction-induced stresses calculated
from the CIS hand-computation model are in very good agreement with the values
obtained from the FE analysis. Note that the residual lateral stresses in a GRS mass
are higher under a higher vertical compaction pressure. In actual construction, the
maximum vertical pressure of compaction is in the range of 200 kPa to 500 kPa. As
seen from Figure 6.15, the effects of CIS can be rather significant in actual
construction. The zigzag lines in FE analyses in Figure 6.15 were caused by the
thickness of compaction lift of 0.2 m. With the larger compaction lift, the amplitude
of the zigzag is larger and the effect of CIS is smaller.
The analyses form FE show more accurately with the finer meshes. But to get the
results for only one curve from the FE analysis as shown in Figure 6.14, it took more
than 20 hours for inputting data and running program with a strong PC configuration
e.g. Dual Core 1.86 GHz, 3 GB RAM, whereas, the analytical model is simple and
can use hand calculation for several minutes. To reduce the time consuming, a coarse
mesh can be used with somewhat tolerated error. Figure 6.16 shows the comparison
between FE results with the coarse mesh and the analytical model. The EF results in
Figure 6.16 obtained from simulating the compaction operation by applying the
compaction pressure over the entire surface area of the GRS mass at each compaction
lift. From Figures 6.15 and 6.16, it can be seen that the fine mesh should be used
when analyzing GRS structures.
303
Figure 6.13: FE Mesh to Simulate CIS in a Reinforced Soil Mass
304
Residual Lateral Stress (kPa) Residual Lateral Stress (kPa)
0 0
0 25 50 75 0 25 50 75 100 125
-1 -1
Lateral Stress Lateral Stress
Distribution Distribution
w ithout w ithout
Considering CIS Considering CIS
-2 -2
Lateral Stress Lateral Stress
Distribution w ith Distribution w ith
Considering CIS Considering CIS
Depth (m)
Depth (m)
(200 kPa) (500 kPa)
-3 -3
Residual Lateral Residual Lateral

Stress (due to Stress (due to
Maximum Vertical Maximum Vertical
-4 Compaction -4 Compaction
Pressure of 200 Pressure of 500
kPa) kPa)
-5 -5
-6 -6
(a) (b)
Figure 6.14: Lateral Stress Distribution of a GRS Mass from FE Analyses with:
(a) Maximum Vertical Compaction Pressures of 200 kPa
(b) Maximum Vertical Compaction Pressures of 500 kPa
305
0
0 25 50 75 100
-1
Maximum vertical
pressure due to
-2 compaction:
200 kPa (Model)

Depth (m)
-3
500 kPa (Model)
200 kPa (FE)

-4
500 kPa (FE)
-5
-6
Figure 6.15: Comparison of Residual Lateral Stresses of a GRS Mass due to Fill
Compaction between FE Analysis with Very Fine Mesh and the Analytical Model
306
0
0 25 50 75 100
-1
Maximum vertical
pressure due to
-2 compaction:
200 kPa (Model)

Depth (m)
300 kPa (Model)

-3
500 kPa (Model)
-4
200 kPa (FE)
300 kPa (FE)
-5
500 kPa (FE)
-6
Figure 6.16: Comparison of Residual Lateral Stresses of a GRS Mass due to Fill
Compaction between FE Analysis with Coarse Mesh and the Analytical Model
307
7. SUMMARY AND CONCLUSIONS
7.1 Summary
A study was taken to investigate the composite behavior of a geosynthetic-reinforced

soil (GRS) mass. The study focused on the strength of a GRS mass, the compaction-
induced stresses in a GRS mass, and the lateral deformation of a GRS wall with
modular block facing.
The following tasks were carried out:

1. Reviewed previous studies on: (1) composite behavior of a GRS mass and (2)
CIS in an unreinforced soil mass and in a GRS mass.
2. Designed a generic soil-geosynthetic composite (GSGC) test for investigating
the composite behavior of GRS mass, and conducted five GSGC tests with
well-controlled condition with extensive instrumentation to monitor the
behavior under different reinforcement spacing, reinforcement strength, and
confining pressure.
3. Developed an analytical model for the relationship between reinforcement
strength and reinforcement spacing, and derived an equation for calculating
composite strength properties.
4. Developed a hand-computation analytical model for simulation of
compaction-induced stresses in a GRS mass.
5. Performed finite element analyses to simulate the GSGC tests, generate
additional data (with different confining pressures) for verifying the analytical
models in this study, and investigate the behavior of GRS composites.
308
6. Verify the analytical models using measured data from the GSGC tests,
relevant test data available in the literature, and FE analyses.
7. Developed an analytical model for predicting lateral movement of GRS walls
with modular block facing.
7.2 Findings and Conclusions
The findings and conclusions of this study can be summarized as follows:
1. The results of the GSGC tests are consistent and appear very reliable. The
tests provide direct observation on the behavior of a GRS mass as related to
reinforcement strength and spacing. The tests also provide better
understanding of the composite behavior of GRS mass and can be used for
validation of analytical models in this study and other models of GRS
structures in the future.
2. An equation describing the relative effects of reinforcement spacing and
reinforcement strength was developed and verified. Based on the equation,
the required reinforcement strength in a GRS wall can be determined, and so
as the composite strength properties and ultimate pressure carrying capacity of
a GRS mass.
3. An analytical model for calculating lateral deformation of a GRS wall with
modular block facing was developed and verified.
4. An analytical model for simulating compaction operation of a GRS mass was
developed. The model allows the compaction-induced stresses in the fill to be
determined.
5. The presence of geosynthetic reinforcement has a tendency to suppress
dilation the surrounding soil, and reduce the angle of dilation of the soil mass.
The dilation behavior offers a new explanation of the reinforcing mechanism,
309
and the angle of dilation provides a quantitative measure of the degree of
reinforcing effect of a GRS mass.
310
APPENDIX A
MATERIAL TESTS
A.1 Backfill
A.1.1 Specific Gravity and Absorption of Coarse Aggregate (per AASHTO T85
and ASTM C127)
• Test Date: April 26 and 27, 2008.

The results of the tests are shown in Table A1 – 1.
311
Table A.1 Specific Gravity determination
A = mass of oven-dry test sample in air, g 184.7
B = mass of saturated-surface-dry test sample in air, g 185.5
C = mass of saturated test sample in water, g 123.4
1. Bulk Specific Gravity = A/(B-C) 2.974
2. Bulk Specific Gravity (Saturated-Surface-Dry) = B/(B-C) 2.987
3. Apparent Specific Gravity = A/(A-C) 3.013
Average Specific Gravity:
1
G = 3.03
P1 P2
+
100 G 1 100 G 2
G1 = specific gravity for size fraction passing sieve # 4: 3.038
G2 = apparent specific gravity for size fraction retained on sieve # 4: 3.013
P1 = mass percentage of size faction passing sieve # 4: 58.0
P2 = mass percentage of size faction retained on sieve # 4: 42.0
4. Absorption = [(B – A) / A] x 100, (%) 0.433
312
A.1.2 Moisture-Density (Compaction) Tests (per AASHTO T99 and ASTM
698, Method A)
• Test Date: from May 1st to May 3rd, 2008
A.1.2.1 Density
• Volume of mold = 944 cm3;

• Mass of mold = 4191.5 g.
Table A.2 Unit weight determination for size fraction passing sieve # 4
Compacted Soil number 1 2 3 4 5 6
Actual average water

content, % 6.06 7.63 8.57 9.63 11.24 11.92
Mass of compacted soil and

mold (g) 6396.9 6466.1 6543.9 6538.9 6521.6 6501.0
Wet mass of soil in mold

(g) 2205.4 2274.6 2352.4 2347.4 2330.1 2309.5
Wet density, (g/cm3) 2.34 2.41 2.49 2.49 2.47 2.45
Dry density, (g/cm3) 2.20 2.24 2.30 2.27 2.22 2.19
Dry density, (pcf) 151.05 153.52 157.39 155.54 152.16 149.9
313
2.40
2.35
Dry Density (g/cm )
3
3
γ d (max) = 2.3 (g/cm )
2.30
2.25
2.20
Wopt = 8.57 %
2.15
2 3 4 5 6 7 8 9 10 11 12 13
Water Content (%)
Figure A.1: Moisture-Density Curve for Size Fractions Passing Sieve # 4
314
A.1.2.2 Rock Correction Calculations
• Bulk Specific Gravity: 2.974

• Absorption: 0.433 %
• Optimum Moisture Content of - #4 Materials: 8.57 %
• Percent Retained on Sieve #4 by Weight: 42 %
The rock correction calculations for this material are follows:

• Corrected Maximum Dry Density = 153.7 pcf = 24.1 kN/m3; and
• Optimum Moisture Content = 5.2% (see Table A.3).
315
Table A.3 Rock correction calculations
Percent Percent
Retained Water Retained Water
on Sieve Content Dry on Sieve Content Dry
#4 (%) Density #4 (%) Density
0 8.57 144 24 6.6 149.5
1 8.5 144.2 25 6.5 149.8
2 8.4 144.5 26 6.5 150.0
3 8.3 144.7 27 6.4 150.2
4 8.2 144.9 28 6.3 150.4
5 8.2 145.2 29 6.2 150.7
6 8.1 145.4 30 6.1 150.9
7 8.0 145.6 31 6.0 151.1
8 7.9 145.8 32 6.0 151.4
9 7.8 146.1 33 5.9 151.6
10 7.8 146.3 34 5.8 151.8
11 7.7 146.5 35 5.7 152.1
12 7.6 146.8 36 5.6 152.3
13 7.5 147.0 37 5.6 152.5
14 7.4 147.2 38 5.5 152.7
15 7.3 147.5 39 5.4 153.0
16 7.3 147.7 40 5.3 153.2
17 7.2 147.9 41 5.2 153.4
18 7.1 148.1 42 5.2 153.7
19 7.0 148.4 43 5.1 153.9
20 6.9 148.6 44 5.0 154.1
21 6.9 148.8 45 4.9 154.4
22 6.8 149.1 46 4.8 154.6
23 6.7 149.3 47 4.7 154.8
316
A.1.3 Gradation (per ASTM D422)
• Test Date: April 20, 2008

• Weight of container: 171.3 g (Sample # 1); 171.2 g (Sample # 2)
• Weight container and dry soil: 920.0 g (Sample # 1); 853.8 g (Sample # 2)
• Weight of dry soil: 748.7 g (Sample # 1); 682.6 g (Sample # 2)
The results of these tests are presented in Table A.4 and Figure A.2.
317
Table A.4 Grain size analysis for Sample 1 and (Sample 2)
Mass of
Container +
Sieve Soil Percent Percent
Number Retained Retained Passing
Diameter Soil
(mm) (g) Retained (g) (%) (%)
389.2 217.9 29.1 70.9

3/8'' 9.5 (361.9) (190.7) (27.9) (72.1)
490.5 101.3 13.5 57.4

4 4.75 (451.6) (89.7) (13.1) (58.9)
598.4 107.9 14.4 43.0

10 2 (548.6) (97.0) (14.2) (44.7)
724.7 126.3 16.9 26.1

40 0.425 (668.1) (119.5) (17.5) (27.2)
756.8 32.1 4.3 21.8

60 0.25 (699.5) (31.4) (4.6) (22.6)
786.8 30.0 4.0 17.8

100 0.106 (725.9) (26.4) (3.9) (18.7)
810.6 23.8 3.2 14.6

200 0.075 (754.2) (28.3) (4.2) (14.6)
14.6 0
<200 (14.6) (0)
318
80
70
60
Percent finer (%)

50
40
30
20
10
0
100 10 1 0.1 0.01 Test 1
Grain size (m m ) Test 2
Figure A.2: Grain Size Distribution for Samples 1 and 2
319
A.1.4 Triaxial Compression Test
A.1.4.1 Test Material
A series of large-size triaxial tests were conducted on specimens of the backfill

material. The backfill was a crushed Diabase rock from a local source near
Washington D.C. The material was classified as a well graded gravel A-1a per
AASHTO M-15 or GW-GM per ASTM D2487. It has 14.6 % of fines passing sieve
#200. The triaxial tests were conducted in a consolidated drained condition. The
maximum dry density of the diabase was 24.1 kN/m3 (153.7 lb/ft3) and the specific
gravity was 3.03. The optimum moisture content was 5.2 %.
A.1.4.2 Test Procedure
Four large-size triaxial tests were conducted at different confining pressures, and the
results were compared with existing results by Ketchart et al. The soil specimen was
approximately 6 in. in diameter and 12 in. in height. The test procedure is described
as follows:
1. Use o-ring to attach a 0.35 – mm thick latex membrane to the base platen;
2. Place a filter paper and a copper porous stone at the base of the platen;
3. Measure the total height of 2 copper porous stone plates, 2 filter paper sheets
and a base plate (see Figure A.3);
4. Attach the fist latex membrane sheet to the base plate (see Figure A.4);
5. Place a metallic mold (a split-barrel type) around the latex membrane and fold
the top portion of the latex membrane down and over the mold;
6. Use vacuum pump to suck air between the latex membrane and the mold (see
Figure A.5);
7. Compact the soil inside the mold, in five layers at the prescribed density of
24.1 kN/m3, with 54 blows/layer as shown in Figure A.6;
320
8. Check the height of the specimen at the fifth (last) layer after 10 first blows,
and adjust (add/remove soil) to get approximately the specimen height of 12
in. if necessary (see Figure A.7);
9. Place a filter paper and a copper porous stone on the top of the specimen;
10. Place the top platen on the porous stone and roll the latex membrane over the
top platen;
11. Use vacuum pump to apply low confining pressure on the specimen to keep
the large-size specimen stable;
12. Remove the metallic mold and attach the second layer of latex membrane to
the specimen with o-rings on the top and base platens (see Figure A.8);
13. Obtain the average height in three different locations of the specimen by using
a stand ruler (see Figure A.9);
14. Obtain the average diameter at two ends and the middle of the specimen by
using a π-type;
15. Place the lucid cylinder on the cell base and fill water up in the cylinder
chamber;
16. Apply a predetermined confining pressure in the chamber using compressed
air, disconnect the specimen with the vacuum pump, and open the back
pressure valve connected to the base of the specimen (see Figure A.10);
17. After 24 hours of consolidation with the confining pressure, place the cell in a
loading frame and start to apply shear stress at a constant strain of 0.1 % per
minute (see Figure A.11);
18. Record the testing load versus strain at every increment and measure the
volume change of the specimen during testing;
19. After failure, remove the test chamber out of the loading machine, remove the
lucid cylinder covering the test specimen and clean up all devices.
Figures A.12 shows the specimen before and after failure.
321
Figure A.3: Measure the Total Height of a Top Plate, 2 Copper Porous Stone
Plates, 2 Filter Paper Sheets, and a Base Plate
322
Figure A.4: Attach the First Latex Membrane Sheet to the Base Plate
323
Figure A.5: Place a Metallic Mold and Apply Vacuum to Remove Air between the
Mold and the Membrane
324
Figure A.6: Compact the Soil in Five Layers
325
Figure A.7: Check the Height of the Specimen at the Last Layer
326
Figure A.8: Roll the Latex Membrane over the Top Plate and Attach the Second
Latex Membrane
327
Figure A.9: Measure the Height and the Diameter of a Specimen Using a Stand
Ruler and a π-Tape
328
Figure A.10: Apply Confining Pressure
329
Figure A.11: Test Cell in the Instron - 5569 Machine During Loading
330
` (a) (b)
Figure A.12: Specimen # 3 before (a) and after (b) Failure
331
A.1.4.3 Test results
The results of the test are shown in Table A.5. Figure A.14 shows the stress-strain
curves and the volume change curves. The stress-strain curves obtained by Ketchart
et al. are also included for comparison and for a more complete set of data. The Mohr
circles at failure are shown in Figure A.14.
Table A.5 Triaxial test results
Specimen Designation Test # 1 Test # 2 Test # 3 Test # 4
Date Test 05/08/08 05/06/08 05/09/08 05/13/08
Confining Pressure (psi) 5 15 30 70
Strain Rate (%/min) 0.1 0.1 0.1 0.1
Diameter (in.) 6.06 6.033 6.05 6.06
Height (in.) 12.08 11.95 12.01 12.11
Water Content (%) 5.2 5.2 5.2 5.2
Failure Condition
Peak Deviatoric Stress (psi) 93.4 157.1 235.5 363.8
Axial Strain at Failure (%) 2.3 2.9 3.7 5.0
Mohr-Coulomb Shear Confining Pressure C1 = 10.3 psi,

Strength Parameters from 0 to 30 psi
φ1 = 50o, ψ = 12.1o
Confining Pressure C2 = 35.1 psi, φ2 = 38o

from 30 psi to 70 psi
332
Deviatoric Stress versus Axial Strain Relationships
600
550
500
450 5 psi
Deviatoric Stress (psi)
400
15 psi
350
300 30 psi
250
200 70 psi
150
70 psi - Ketchart
100
50 110 psi - Ketchart
0
0 1 2 3 4 5 6 7 8 9 10 11
Axial Strain (%)
(a)
Volumetric Strain versus Axial Strain

0.6
0.4
0.2 5 psi
0
30 psi
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
-0.2
-0.4
Axial Strain (%)
(b)
Figure A.13: Triaxial Test Results:

(a) Stress-Strain Curves
(b) Volume Change Curves
333
σ
3 = 110 psi
Shear Streess (psi)
σ
3 = 70 psi
σ
3 = 30 psi
σ
3 = 15 psi
C2 = 35.1 psi σ
3 = 5 psi
C1 = 10.3 psi
0 30 98 172 265 433 616
Normal Stress (psi)
Figure A.14: Mohr Circles of the Stress at Failure
334
A.2 Geotextile
A.2.1 Uniaxial Load-Deformation Test (per ASTM D 4595)
• Material: In this study, geosynthetic used for the tests was Geotex 4x4, a
woven polypropylene geotextile (formally known as Amoco2044).
• Specimen dimensions: The dimensions of the geotextile specimen
accordance with ASTM D 4595 are: Width = 200 mm (8 in.) and Length =
100 mm (4 in.). For non-woven geotextiles, the aspect ration of the
reinforcement specimen (i.e., the ratio of width to length) should be
sufficiently large (say, greater than 4) to alleviate significant “necking” (i.e.
Poisson) effect. There is little “necking” effect for woven geotextiles,
regardless of the aspect ratio. Besides measuring stress-strain relationships,
the calibration curves (between actual geotextile strain and strain gages) were
also obtained from these tests. The length of the geotextile should be long
enough for a strain gage to be attached. The selected specimen dimensions
for these tests, therefore, were: Width = 305 mm (12 in.) and the Length =
152 mm (6 in.).
• Clamps: The grip portion at two ends of the Geotex 4x4 specimen was
treated with high-strength epoxy and two steel plates. The specimen then
was clamped tightly at two ends of the specimen by screwing total number
of 16 bolts in to stiff steel jaws (see Figure A.15) to prevent the specimen
slipping in the jaws.
• Strain Rate: The constant strain rate of 10 %/min was used.
A uni-axial load-deformation test of geotextile is shown in Figure A.15. A series of

tests were conducted using two types of geosynthetic: (1) single-sheet Geotex 4x4
and (2) double-sheet Geotex 4x4. The double-sheet was manufactured by gluing two
335
sheets of Geotex 4x4 geotextile together using spray glue “3M Super 77 Adhesive”.
Some index properties of the geotextile reinforcement are shown in Table A.6 and
stress-strain relationships are shown in Figure A.16.
336
Figure A.15: Uniaxial Load-Deformation Test of Geotex 4x4
337
Table A.6 Index properties of the reinforcement for fill direction
Wide-Width Tensile Strength
Geosynthetic ASTM D 4595
Type Stiffness at Ultimate Strain at Break

1% Strain Strength (%)
(kN/m) (kN/m)
Single-sheet Geotex 4x4 1,000 70 12 %
Double-sheet Geotex 4x4 1,960 138 12 %
338
160
140
120
Tensile Load (kN/m)
100
80
60 Single sheet
40 Double Sheets
20
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Axial Strain (%)
Figure A.16: Load-Strain Curves for Single-Sheet and Double-Sheet Geotex 4x4
339
A.2.2 Instrumentation for Measuring Strain Gage in Geotextile
To measure the strain in the geotextile, a high elongation strain gage (type EP-08-
250BG-120), manufactured by Measurements Group, Inc., was used. Each strain
gage was glued to the geotextile only at two ends to avoid inconsistent local stiffening
of geotextile due to the adhesive. The strain gage attachment technique was
developed at the University of Colorado Denver. The gage was first mounted on a 25
mm by 76 mm patch of a lightweight nonwoven geotextile (see Figure A.17). A
Microcystalline wax and a rubber coating (M-Coat B, Nitrile Rubber coating) were
used to cover strain gages to protect the gages from moisture. To check the effects of
the moisture-protection materials, the geotextile specimens with the strain gages were
tested after immersing in water for 24 hours. The rubber coating was selected to use
for the experiments. Before installing the reinforcement sheet in the composite mass,
the M-Coat FB-2, 6694, Butyl Rubber Tape was applied over the gages to protect
them during compaction as shown in Figure A.17.
340
(a) (b)
Figure A.17: Strain Gage on Geotex 4x4 Geotextile:
(a) Before Applying Protection Tape;
(b) After Applying Protection Tape (M-Coat FB-2, 6694, Butyl
Rubber Tape) in the Experiment.
341
A.2.3 Calibration of Strain Gages for Measuring Deformation Geotextile
Due to the presence of the light-weight geotextile patch, calibration is needed. The
calibration tests were performed to relate the strain obtained from the strain gage to
the actual strain of the reinforcement. Figures A.18 and A.19 show the calibration
curves along the fill direction of Geotex 4x4 geotextile for single-sheet and double-
sheet specimens, respectively.
342
6
5
y = 1.172x
2
R = 0.9913
4
0
0 1 2 3 4 5
Strain from Strain Gage (%)
Figure A.18: Calibration Curve for Single-Sheet Geotex 4x4
343
6
5
y = 1.078x
R2 = 0.9986
4
0
0 1 2 3 4 5
Strain from Strain Gauge (%)
Figure A.19: Calibration Curve for Double-Sheet Geotex 4x4
344
APPENDIX B
REVIEW DESIGN OF THE GSGC TEST FRAME
B.1 Pressure on Plexiglass (in the plane strain direction of GSGC Test)
The pressure on the plexiglass is calculated from the maximum vertical pressure on
the top of the specimen. The maximum lateral pressure on the plexiglass is:
q = 500 kPa (72psi).
Figure B.1: Dimensions of the Test Frame
345
B.2 Checking the Capacity of the Horizontal Tubings
B.2.1 Checking Bending Moment
The maximum required moment of horizontal tubing (like a beam with the simple
supports) is:
⎛ q l2 ⎞ ⎛ (500 kPa ) * (2.074 / 7 ) * (1.835 m )2 ⎞

M req = ⎜⎜ ⎟⎟ = ⎜ ⎟ = 62.4 kN.m
⎜ ⎟
⎝ 8 ⎠ ⎝ 8 ⎠
The tensile strength of the tubings is 400,000 kN/m2 (58,000 psi) and yield strength is
317,000 kN/m2 (46,000 psi). For the 2” x 12” x ¼” steel tubings used in the design,
the maximum allowable moment is:
M allow = M top and bottom + M 2 sides
in which,
( )( )( )[ ]
Mtop and bottom = A* h *[σ steel] = 6.35*10−3 * 50.8*10−3 * (304.8 − 6.35) *10−3 * 3.17*105 = 30.5
kN.m
M 2 sides = S x * [σ ] =
(2 * 6.35 *10 )* ((304.8 − 2 * 6.35) *10 ) * [3.17 *10 ] = 57.2
−3 −3 2
5
6
kN.m
Therefore,
Mallow = 30.5 + 57.2 = 87.7 kN.m. > Mreq = 62.4 kN.m. It is acceptable.
346
B.2.2 Checking Shear Stress
The maximum required shear stress of a horizontal tubing is:
⎛ (500 kPa) * (2.074/ 7) * (1.835 m) ⎞

⎛ql⎞
σ req = ⎜ ⎟ /( A) = ⎜ ( )
⎟ / (50.8 + 304.8) * 2 * 6.35*10 = 30,097
−6
⎝ ⎠
2 ⎝ 2 ⎠
kN.m = 4,334 psi.
Therefore, shear capacity of the tubings is acceptable.
B.2.3 Checking the Bolts
The maximum required shear force of one bolt is:
⎛ql⎞ (
⎛ (500 kPa ) * 11.4' '*25.4 * 10 −3 m
Qreq = ⎜ ⎟ / (2 bolts ) = ⎜⎜
) ⎞⎟ / 2 = 36.2 kN
⎟
⎝ 2 ⎠ ⎝ 2 ⎠
The required shear stress will be:

Q 36.2
σ req = = = 93,311 kPa = 13,436 psi = 13.4 kpsi
(
A π * 7 / 8' '*2.54 * 10 −3 )
2
/4
The shear capacity of 7/8 A325 bolt is: 0.4 * Fy = 0.4 * 36 = 14.4 kpsi.
Therefore, the bolt capacity is acceptable.
347
B.2.4 Checking the Capacity of the Plexiglass
The maximum required moment acting on Plexiglass is:
⎛ q l2 ⎞ ⎛ (500 kPa ) * (2.074 / 7m )2 ⎞

M req = ⎜⎜ ⎟⎟ = ⎜ ⎟ = 3.66 kN.m
⎜ ⎟
⎝ 12 ⎠ ⎝ 12 ⎠
The maximum required stress is:
⎛ ⎞
⎜ ⎟
⎛ M req ⎞ ⎜ 3.66 ⎟ = 21,784 kPa = 3,137 psi
σ req = ⎜⎜ ⎟⎟ =
⎠ ⎜⎜ (1m ) * (1.25' '*25.4 / 1000 ) ⎟
2
⎝ Sx ⎟
⎝ 6 ⎠
With the tensile strength of the Plexiglass acrylic sheet of 10,500 psi, the moment
capacity of the plexiglass is acceptable.
348
REFERENCES
Adams, M. T. (1997). “Performance of a Prestrained Geosynthetic Reinforced Soil

Bridge Pier.” Proc. of the International Symposium on Mechanically
Stabilized Backfill, J. T. H. Wu, ed., Balkema, Rotterdam, The Netherlands,
25-34.
Adams, M. T. (2007). “Mini Pier Experiments Geosynthetic Reinforcement Spacing

and Strength as Related to Performance.” Proc. of Geo-Denver 2007, ASCE,
Denver.
Aggour, M. S., and Brown, C. B. (1974). “The Prediction of Earth Pressure on

Retaining Walls Due to Compaction.” Geotechnique, 24(4), 489-502.
American Association of State Highway and Transportation Officials (AASHTO).

(2002). Standard Specifications for Highway Bridges, 17th Ed., Washington,
D.C., 829 pages.
Allen, T. M., and Bathurst, R. J. (2001). "Application of the K0-stiffness Method to

Reinforced Soil Wall Limit State Design." Final Research Report to
Washington State Department of Transportation, Seatle, Washington State.
Allen, T. M., and Bathurst, R. J. (2001). "Prediction of Soil Reinforcement Loads in

MSE Walls." Report WA-RD-522.1, Washington State Department of
Transportation and Federal Highway Administration (FHWA).
Athanasopoulos, G.A. (1994). “On the Enhanced Confining Pressure Approach to the
Mechanics of Reinforced Soil.” Geotechnical and Geological Engineering, 12,
122-132.
Athanasopoulos, G.A. (1993). “Effect of Particle Size on the Mechanical Behavior of

Sand-Geotextile Composite.” Geotextiles and Geomembranes, 12, 255-273.
349
ASTM D4595 (1986). “Standard Test Method for Tensile Properties of Geotextile by
the Wide-Width Strip Method.” American Society for Testing and Materials.
USA.
Bassett, A.K. and Last, N.C. (1978). “Reinforcing Earth below Footings and
Embankments.” Proc. Of the ASCE Spring Convention and Exhibit,
Pittsburgh, PA.
Bathurst, R. J. and Cai, Z. (1994). "In-isolation Cyclic Load-extension Behavior of

Two Geogrids." Geosynthetics International, 1(1), 1-19.
Bathurst R. J., and Simac, M. R. (1997). “Design and Performance of the Facing
Column for Geosynthetic Reinforced Segmental Walls.” Proc. of the
International Symposium on Mechanically Stabilized Backfill, J. T. H. Wu,
ed., Balkema, Rotterdam, The Netherlands, 193-208.
Bathurst R. J., Cai, Z., Alfaro, M., and Pelletier, M. (1997). “Seismic Design Issues
for Geosynthetic Reinforced Segmental Retaining Walls.” Proc. of the
International Symposium on Mechanically Stabilized Backfill, J. T. H. Wu,
ed., Balkema, Rotterdam, The Netherlands, 79-97.
Bathurst, R. J., Vlachopoulos, N., Walters, D. L., Burgess, P. G., and Allen, T. M.
(2006). “The Influence of Facing Stiffness on the Performance of Two
Geosynthetic Reinforced Soil Retaining Walls.” Canadian Geotechnical
Journal, 43, 1225-1237.
Bolton, M. D. (1986). "The Strength and Dilatancy of Sands." Geotechnique, 36(1),

65-78.
Broms, B. (1971). “Lateral Earth Pressures Due to Compaction of Cohesionless

Soils.” Geotechnical Engineering Report No. 83-1, Cornell University.
Broms, B. (1971). "Lateral Earth Pressure due to Compaction of Cohesionless Soils."

Proc., 4th Budapest Conference on soil mechanics and foundation
engineering, 373-384.
350
Broms, B. (1977). “Triaxial Tests with Fabric-Reinforced Soil.” Proc. of the
International Conference on the use of Fabric in Geotechnics, Paris, 3, 129-
134.
Broms, B., and Ingleson, I. (1971). "Earth Pressure against Abutment of Rigid Frame
Bridge." Geotechnique, 21(1), 15-28.
Bueno, B. S., Vinicius, C., Benjamim, S., and Zornberg, J. G. (2005). “Field
Performance of a Full-scale Retaining Wall Reinforced with Nonwonven
Geotextiles.” Geo-Frontiers, ASCE.
Carder, D. R., Pocock, R. G., and Murray, R. T. (1977). “Experimental Retaining

Wall Facility-Lateral Stress Measurements with Sand Backfill.” Report No.
LR 766, Transport and Road Research Laboratory (TRRL), England.
Chen, T.C, Chen, R.H, Lin, S.S. (2000). “A nonlinear Homogenized Model
Applicable to Reinforced Soil Analysis.” Geotextile and Geomembranes, 18,
349-366.
Chew, S.H., and Mitchell, J.K. (1994). “Deformation Evaluation Procedure for
Reinforced Walls.” Fifth International Conference on Geotextiles,
Geomembranes and related Products, Singapore, 5-9 September 1994.
Collin, J. G. (2002). "Design Manual for Segmental Retaining Walls", NCMA, 2nd
Ed., National Concrete Masonry Association, Herndon, Va.
Christopher, B. R., Gill, S. A., Giroud, J. P, Mitchell, J. K., Schlosser, F., and
Dunnicliff J. (1989). “Reinforced Soil Structures. Vol. 1: Design and
Construction Guidelines", FHWA-RD-89-043, Federal Highway
Administration, Washington, D.C.
Christopher, B.R., Leshchinsky, D., and Stulgis, R. (2005). “Geosynthetic-Reinforced

Soil Walls and Slopes: US Perspective.” GSP 141 International Perspectives
on Soil Reinforcement Applications, ASCE.
351
Day, R. (1993). Discussion of "Estimation Earth Pressures due to Compaction" by
Duncan, et al. (1991). Journal of Geotechnical Engineering, ASCE, 119(7),
1162.
Duncan, J. M., and Seed, R. M. (1986). "Compaction-Induced Earth Pressures Under

K0-Conditions." Journal of Geotechnical Engineering, ASCE, 112(1), 1-22.
Duncan, J. M., Williams, G. W., Sehn, A. L., and Seed R. M. (1991). "Estimation
Earth Pressures due to Compaction." Journal of Geotechnical Engineering,
ASCE, 117(12), 1833-1847.
Duncan, J. M., Williams, G. W., Sehn, A. L., and Seed R. M. (1993). Closure of
"Estimation Earth Pressures due to Compaction." Journal of Geotechnical
Engineering, ASCE, 119(7), 1172-1177.
Ehrlich, M., and Mitchell, J. K. (1994). “Working Stress Design Method for
Reinforced Soil Walls.” Journal of Geotechnical Engineering, ASCE, 120(4),
625-645.
Elias, V., and Christopher, B. R. (1996). "Mechanically Stabilized Earth Walls and
Reinforced Soil Slope Design and Construction Guidelines." Report FHWA-
SA-96-071, Federal Highway Administration, Washington, D.C.
Elias, V., Christopher, B. R., and Berg, R. R. (2001). "Mechanically Stabilized Earth
Walls and Reinforced Soil Slopes, Design and Construction Guidelines."
Report FHWA-NHI-00-043, Federal Highway Administration, Washington,
D.C.
Elton, D. J. and Patawaran, M. A. B. (2004). "Mechanically Stabilized Earth

Reinforcement Tensile Strength from Tests of Geotextile-Reinforced Soil."
Journal of the Transportation Research Board, No. 1868, TRB, National
Research Council, Washington, D.C., 81-88.
352
Elton, D. J. and Patawaran, M. A. B. (2005). "Mechanically Stabilized Earth (MSE)
Reinforcement Tensile Strength from Tests of Geotextile Reinforced Soil." A
Report to the Alabama Highway Research Center, Auburn University, 77
pages.
Giroud, J. P. (1989). “Geotextile Engineering Workshop-Design Examples.” Report

FHWA-HI-89-002, Federal Highway Administration, Washington, D.C.
Gray, D.H. and Ohashi, H. (1983). “Mechanics of fiber reinforcement in sand.”

ASCE, Journal of Geotechnical Engineering, 109, 335-353.
Hatami, K., and Bathurst, R. J. (2005). "Development and Verification of a

Numerical Model for the Analysis of Geosynthetic Reinforced-soil Segment
Walls." Canadian Geotechnical Journal, 42(4), 1066-1085.
Hatami K. and Bathurst R. J. (2006). "Numerical Model for Reinforced Soil

Segmental Walls under Surcharge Loading." Journal of Geotechnical and
Geoenvironental Engineering, ASCE, 132(6), 673-684.
Hausmann, M.R. (1976). “Strength of Reinforced Earth.” ARRB Proc., Vol. 8.
Helwany, M. B., Tatsuoka, F., Tateyama, M., and Kojima, K. (1996). “Effects of
Facing Rigidity on the Performance of Geosynthetic-Reinforced Soil
Retaining Walls.” Soil and Foundations, Japan, 36(1), 27-38.
Ho, S. K. and Rowe, R. K. (1997). “Effect of Wall Geometry on the Behavior of

Reinforced Soil Walls.” Geotextiles and Geomembranes, 14, 521–541.
Holtz, R. D., Christopher, B. R., and Berg, R. (1997). Geosynthetic Engineering,

BiTech Publishers, Vancouver, 452 pages.
Holtz, R. D. and Lee, W. F. (2002). "Internal Stability Analyses of Geosynthetic

Reinforced Retaining Walls." Report WA-RD 532.1, Washington State
Department of Transportation, Washington.
353
Ingold, T. S. (1979). “The Effect of Compaction on Retaining Walls.” Geotechnique,
29(3), 265-283.
Ingold, T.S. (1982). “Reinforced Earth.” Thomas Telford Ltd, London.
Jewell, R. A. (1985). “Limit Equilibrium Analysis of Reinforced Soil Walls.” Proc. of

11th International Conference of Soil Mechanics and Foundation Engineering,
San Francisco, 3, 1705-1708.
Jewell, R. A. and Wroth, C. P. (1987). “Direct Shear Tests on Reinforced Sand.”

Geotechnique, 37(1), 53-68.
Jewell, R. A. (1988). "Reinforced Soil Wall Analysis and Behavior." The Application
of Polimeric Reinforcement in Soil Retaining Structures, NATO Advanced
Research Workshop, 365-408.
Jewell, R. A., and Milligan, G. W. (1989). "Deformation Calculation for Reinforced

Soil Walls." International Conference on Soil Mechanics and Foundation
Engineering, 1259-1262.
Jewell, R. A. (1990). "Strength and Deformation in Reinforced Soil Design." The

fourth International Conference on Geotextiles, Geomembranes and Related
Products, The Hague, Netherlands.
Katona, M. G. (1978). "The Analysis of Long Span Culverts by Finite Element

Method." Paper submitted to Transportation Research Board, Washington
D.C.
Ketchart, K. (2000). "Behavior of Preloaded Geosynthetic-Reinforced Soil Mass."

PhD. Thesis, University of Colorado at Denver, 347 pages.
Ketchart, K., and Wu, J. T. H. (2001). "Performance Test for Geosynthetic

Reinforced Soil Including Effects of Preloading", Report FHWA-RD-01-018,
Federal Highway Administration, Washington, D.C.
354
Lee, W. F. (2000). "Internal Stability Analyses of Geosynthetic Reinforced Retaining
Walls." Ph.D. Thesis, Dept. of Civil and Environmental Engineering,
University of Washington, 379 pages.
Long N. T., Legeay, G., and Madani, C. (1983). “Soil-Reinforcement Friction in a

Triaxial Test.” Proc. of the 8th European Conference on Soil Mechanics and
Foundation Engineering, Hesinki, Finland, 1, 381-384.
Macklin, P. M. (1994). "A Comparison of Five Methods for Calculation Lateral

Deformation of Geosynthetically Reinforced Soil (GRS) walls." Master
Thesis, University of Colorado at Denver, Colorado, 70 pages.
Maher, M.H. and Woods, R.D. (1990). “Dynamic Response of Sand Reinforced with
Randomly Distributed Fibers.” ASCE, Journal of Geotechnical Engineering,
116, 1116-1131.
Mitchell, J.K., Villet, W.C.B. (1987). “Reinforcement of Earth Slopes and

Embankments.” National Cooperative Highway Research Program, Report
290, TRB, 323 pages.
Morrison, K. F., Harrison, F. E., Collin, J. G., Dodds, A., and Arndt, B. (2006).
"Shored Mechanically Stabilized Earth (SMSE) Wall Systems Design
Guidelines." Report FHWA-CFL/TD-06-001, Federal Highway
Administration, Lakewood, Colorado, 210 pages.
Morrison, K. F., Harrison, F. E., Collin, J. G., and Anderson, S. A. (2007). "Full-
Scale Testing of a Shored Mechanically-Stabilized Earth (SMSE) Wall
Employing Short Reinforcements." Proc. of Geo-Denver 2007, ASCE,
Denver.
Peck, R. B. (1993). Discussion of "Estimation Earth Pressures due to Compaction "

by Duncan et al. (1991), 117(12). Journal Geotechnical Engineering, ASCE,
119(7), 1171-1172.
Plaxis B.V. (2002). “Plaxis 2D - Version 8 Manual.” Balkema, Rotterdam, The

Netherlands.
355
Rowe, P.W. (1954). “A Stress-Strain Theory for Cohesionless Soil with Applications
to Earth Pressures at Rest and Moving Walls.” Geotechnique, 4(2), 70-88.
Rowe, R. K. and Ho S. K. (1993). “Some Insights into Reinforced Wall Behaviour on

Finite Element Analysis.” Earth Reinforcement Practice, Ochiai, Hayashi, and
Otani, eds., Balkema, Rotterdam, The Netherlands, 485-490.
Rowe, R. K. and Ho S. K. (1998). “Horizontal Deformation in Reinforced Soil

Walls.” Canadian Geotechnical Journal, 35, 312-327.
Schlosser, F., Long, N. (1972). “Comportement de al terre armée dans les ouvrages
de soutènement”. In: Proceedings of the European Conference on Soil
Mechanics and Foundation Engineering, vol. 1. Madrid, pp. 299–306.
Schlosser, F. and Long, N.T. (1974). ”Recent Results in French Research on

Reinforced Earth.” ASCE, Journal of Construction Division, 100, 223-237.
Seed, R. M. (1983). "Compaction-induced Stresses and Deflections on Earth

Structure." PhD. Thesis, University of California, Berkeley, CA, 447 pages.
Seed, R. M., Duncan, J. M. (1986), "FE Analyses: Compaction-Induced Stresses and

Deformations." Journal of Geotechnical Engineering, ASCE, 112(1), 23-43.
Sower, G. F., Robb, A. D., Mullis, C. H. and Glenn, A. J. (1957). “The Residual
Lateral Pressures Produced by Compacting Soils.” Proc. of 4th International
Conference on Soil Mechanics and Foundation Engineering, London.
Tatsuoka, F., Molenkamp, F., Torii, T., and Hino, T. (1997). "Preloaded and
Prestressed Reinforced Soil." Soil and Foundations, Japan, 37(3), 79-94.
Tatsuoka, F. (1993). “Roles of Facing Rigidity in Soil Reinforcing.” Earth

Reinforcement Practice, Ochiai, Hayashi, and Otani, eds., Balkema,
Rotterdam, The Netherlands, 831-867.
Terzaghi, K. (1943). “Large Retaining Wall Tests (I): Pressure of Dry Sand.”
Engineering News Record, 112, 136-140.
356
Vulova, C. and Leshchinsky, D. (2003). “Effect of Geosynthetic Reinforcement
Spacing on the Performance of Mechanically Stabilized Earth Walls.”
Publication No. FHWA-RD-03-048, Federal Highway Administration,
McLean, VA, USA.
Westergaard, C. M. (1938). “A Problem of Elasticity Suggested by a Problem in Soil

Mechanics: A Soft Material Reinforced by Numerous Strong Horizontal
Sheet,” in Contributions to the Mechanics of Solids, Stephen Timoshenko
60th Anniversary Volume, Macmillan, New York, 268-277.
Wu, J.T.H. (1989), “Behavior of Soil-Geotextile Composites and Its Application to

Finite Element Analysis,” Proc. of Geosynthetics’ 89, 365-372.
Wu, J.T.H. (1994). "Design and Construction of Simple, Easy and Low cost
Retaining Walls." Colorado Transportation Institute, Report CTI-UCD-1-94.
Wu, J. T. H. (2001). “Revising the AASHTO Guidelines for Design and Construction
of GRS Walls.” Report CDOT-DTD-R-2001-6, Colorado Department of
Transportation, University of Colorado at Denver, 148 pages.
Wu, J. T. H. "Design and Construction of Geosynthetic-Reinforced Soil Retaining

Walls." Draft of Textbook (unpublished).
Wu, J. T. H., Lee, K. Z. Z., Helwany, S. B., and Ketchart, K. (2006). "Design and
Construction Guidelines for GRS Bridge Abutment with a Flexible Facing."
NCHRP Report 556, National Cooperative Highway Research Program,
Washington, D.C.
Wu, J.T.H., Lee, K.Z.Z., and Pham, T. (2006). “Allowable Bearing Pressure of
Bridge Sills on GRS Abutments with Flexible Facing.” J. Geotechnical and
Geoenvironmental Engineering, ASCE, 132(7), 836-841.
Wu, J. T. H. (2007). "Lateral Earth Pressure against the Facing of Segmental GRS
Walls", Proc. of Geo-Denver 2007, ASCE, Denver.
357
Yang, Z. (1972). "Strength and Deformation Characteristics of Reinforced Sand."
PhD. Thesis, University of California at Los Angeles, CA, 236 pages.
Yang, Z. and Singh, A. (1974). “Strength and Deformation Characteristics of

Reinforced Sand.” International Meeting on Water Resources Engineering,
Los Angeles, CA.
Zhang, M.X., Javadi, A.A., Lai, Y.M., and Sun, J. (2006). “Analysis of Geosynthetic
Reinforced Soil Structures with Orthogonal Anisotropy.” Geotechnical and
Geological Engineering, 24, 903-917.
Ziegler, M., Heerten, G., and Ruiken. G. (2008). “Progress in the Understanding of
Geosynthetic/Soil Composite Material Behaviour in Geosynthetic Reinforced
Earth Structures.” The First Pan American Geosynthetics Conference &
Exhibition, 2-5 March 2008, Cancun, Mexico.
358

Pham Thang Investigating Composite Behavior of Geosynthetic Reinforced Soil Mass Thesis

Загружено:

Сведения о документе

Исходное описание:

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Pham Thang Investigating Composite Behavior of Geosynthetic Reinforced Soil Mass Thesis

Загружено:

Авторское право:

Доступные форматы

INVESTIGATING COMPOSITE BEHAVIOR OF GEOSYNTHETIC-

REINFORCED SOIL (GRS) MASS

A thesis submitted to the

A study was undertaken to investigate the composite behavior of a Geosynthetic

A series of large-size Generic Soil-Geosynthetic Composite (GSGC) tests were

I would like to express my most sincere gratitude to my thesis advisor,

1.1 Problem Statement

It is a well-known fact that soil is weak in tension and relatively strong in

1.2 Research Objectives

Task 2: Developed a hand-computation analytical model for simulation of

Task 3: Developed an analytical model for the relationship between reinforcement

Task 4: Designed and conducted laboratory experiments on a generic soil-

Task 7: Developed an analytical model for predicting lateral movement of GRS

A GRS mass is a soil mass that is embedded with layers of geosynthetic

2.1 Mechanics of Reinforced Soil

Figure 2.1: Concept of Apparent Cohesion due to the Presence of Reinforcement

Figure 2.2: Concept of Apparent Confining Pressure due to the Presence of

The behavior of soil-geosynthetic composites have been investigated through

Figure 2.6: Stress- Strain Curves of Specimens Reinforced at Spacing of 12 in.

The behavior of soil-geosynthetic composites have also been investigated through

Vulova and Leshchinsky (2003) conducted a series of analyses using a two-

2.3.1 Lateral Earth Pressure Estimation by Rowe (1954)

where h0 is the surcharge head removed and h is the overburden head ( h = σ v / γ ). In

any case, K 0 ≤ K P (KP = coefficient of passive earth pressure at limiting condition).

where: α = 0.3 to 0.5 for most sands

of loading followed by unloading. When the overburden pressure is increased (i.e.,

Figure 2.16: Hypothetical Stress Path during Compaction (Broms, 1971)

2.3.3 Finite Element Analysis by Aggour and Brown (1974)

a) First-Cycle K0-Reloading Stress Path and Typical K0-unloading/Reloading

Seed (1983) proposed what is termed “first-cycle K0-reloading model,” as shown in

to reloading. Reloading is then assumed to follow a linear path to point R, the

α = 0.018 + 0.974 sin φ ' (2.12)

Figure 2.20: The First-Cycle K0-Reloading Model (Seed, 1983)

Figure 2.22: Typical K0-Reloading Stress Paths (Seed, 1983)

Seed (1983) proposed another model for multi-cycle K0-unloading/reloading

3. Reloading through point R then to point C, passing through previous

c) Non-Linear, Multi-Cycle K0-Unloading/Reloading Compaction Model

The non-linear, multi-cycle K0-unloading/reloading model was developed based on

Parameter Name Recommended Method of Estimation

β Reloading coefficient 0 ≤ β ≤1 β ≅ 0.6

K0 Coefficient of at-rest lateral 0 ≤ K0 ≤ 1 K 0 ≅ 1 − sin φ '

K1,φ Frictional component of K 0 ≤ K1,φ ≤ K P ⎛ φ' ⎞

c’ Effective stress strength

Note: K1 = Limiting coefficient of at-rest lateral pressure for unloading

Basic component of the non-linear K0-loading/unloading model is described in Figure

d) Simplified Bi-Linear Approximation to the Non-Linear Model

Figure 2.29: Relationship between K2 and F in the Bi-Linear Unloading Model

Parameter Name Recommended Method of Estimation Based on

K0 Coefficient of at-rest Same as non-linear model

K1,φ’,B Frictional component 2

cB’ Effective stress

- Fraction of peak 0 ≤ F ≤1 - F (or K2) should be chosen

Figure 2.31: Basic Components of the Bi-linear Model (Seed,1983)

4. Reduce the near-the-surface portion of the σ’h,r distribution with

Table 2.4: Material parameters for non-linear and bi-linear models

Non-linear model parameters Bi-linear model parameters

2.4.1 Ehrlich and Mitchell (1994)

Figure 2.35: Assumed Stress Path (Ehrlich and Mitchell, 1994)

In Figure 2.35, the value of σ’zc can be estimated as the following:

2.4.2 Hatami and Bathurst (2006)

2.4.3 Morrison et al. (2006)

Figure 2.37: Finite Element Model for Finite Element Analysis

2.5 Highlights on Compaction-Induced Earth Pressures in the Literature