ANALYSIS OF DYNAMIC LOADING BEHAVIOUR FOR

PAVEMENT ON SOFT SOIL

Dissertation

Slamet Widodo

Technische Universität Bergakademie Freiberg

Freiberg, 2013
 ANALYSIS OF DYNAMIC LOADING BEHAVIOUR FOR PAVEMENT
ON SOFT SOIL

To the Faculty of
Geowissenschaften, Geotechnik und Bergbau
of the Technische Universität Bergakademie Freiberg
approved

THESIS

to attain the academic degree of

DOKTOR-INGENIEUR
DR.-ING.

Submitted

By Magister Teknik in Pavement Engineering, Slamet Widodo

Born on the 23rd December 1967 in Surabaya, Indonesia

Reviewers: Prof. Dr.-Ing. Herbert Klapperich, Freiberg

Prof. Dr.rer.nat. Rafig Azzam, Aachen
Prof. Dr. Masyhur Irsyam, Bandung-Indonesia

Date of the award: 19-11-2013

i
 ACKNOWLEDGMENTS

The author wishes to express his sincere gratitude to Prof. Dr.-Ing. Herbert Klapperich, the
chairperson of his dissertation committee, for his very kind encouragement, guidance,
friendship and inspiration given throughout the graduate work. The author is also grateful to
the other dissertation committee members, Prof. Dr.rer.nat. Rafig Azzam and Prof. Dr.
Masyhur Irsyam for their very precious time for reading this dissertation.
The author is very grateful to Dr.-Ing. Habil. Nandor Tamaskovics for his kind
encouragement, competent guidance, care and always inspires him. Thanks to all of the
author’s colleagues and staff at Soil Mechanics laboratory, Institut für Geotechnik for their
help, co-operation and valuable discussions and especially to Dr.-Ing. Ernst-Dieter Hornig
and Frau Helga Vanselow and also Hong Shen, Abdelazim Ibrahim, Jinyang Fu, Phan Luu
Minh Phuang, Irina, Anastasia Alferova, Ahmed Mohamed, and Sebastian Althoff.
Special thanks is addressed to some generous people who permit him using their data in part
of this thesis namely Suzanne van Eekelen, Prof. Marcio S.S. Almeida, Abdullah C. Hassandi
and Cao Wei-ping.
The financial support of the Overseas Scholarship Program of Master/PhD for Academic
Staff of Tanjungpura University with contract No. 1725.35/D4.4/2009 during the author’s
studies at the Technische Universität Bergakademie Freiberg is gratefully acknowledged. The
support by the GraFA (Graduierten-und Forschungsakademie) is also gratefully
acknowledged. Likewise, join working with DFG-Research Project is really helpful. The
author honored to thank to Prof. Dr. Chairil Effendy, the former Rector of Tanjungpura
University and Prof. Dr. Thamrin Usman as the Rector of Tanjungpura University.
Thanks warmly to the author’s Indonesian friends in Freiberg who made his years enjoyable:
families of Arief Wijaya, Herry Permana, Toni and Eki, Nazaruddin, Heru, Annisa, Anesia,
Miranti, Annas, Anto, and Linda. Also, special thanks goes to Adang Budiman that improves
his English and Prof. Zhuang for correcting the final manuscript.
The author is very grateful to his parents, special thanks to the author’s parents, Suhadi and
Sutrinah and also mother-in-law, Nurjanah who always pray for the success, happiness and
moral support, and encouragement.
Finally, the author is overwhelmed in thanking his wife, Juliana and his children, Basith,
Retnoayu, and Satrio, for their understanding and patience throughout the period of his study.
Also, special thanks goes to Fajar Indaharti, from her he learn deep patience and sincerity.

ii
 ABSTRACT

The increasing need for regional development has led engineers to find safe ways to construct
the infrastructure of transportation on soft soils. Soft soil is not able to sustain external loads
without having large deformations. The geotechnical properties of soft soil which is known
for its low bearing capacity, high water content, high compressibility and long term
settlement as well.
In pavement engineering, either highway or runway as an infrastructure, a pavement
encompasses three important parts namely traffic load, pavement and subgrade. Traffic load
generated from tire pressure of vehicle and/or airplane wheels are usually around 550 kPa
even more on the surface of the pavement. Pavement generally comprises granular materials
with unbounded or bounded materials located between traffic load and subgrade, distributing
the load to surface of subgrade.
One of the promising soil improvement techniques is a piled embankment. When
geosynthetics layer is unrolled over piles, it is known as geosynthetics supported piled
embankment. Particularly in deep soft soil, when piles do not reach a hard stratum due to
large thickness of the soft soil, the construction is an embankment on floating piles.
Furthermore, because of different stiffness between piles and subsoil, soil arching effect
would be developed there.
By using Finite Element analysis, some findings resulted from experimental works and
several field tests around the world as field case studies are verified. Some important findings
are as follows: the stress concentration ratio is not a single value, but it would be changed
depending on the height of embankment, consolidation process of subsoil, surcharge of traffic
load, and tensile modulus of geosynthetics as well. Ratio height of embankment to clear piles
spacing (h/s) around 1.4 can be used as a critical value to distinguish between low
embankment and high embankment. When geosynthetics is applied to reinforce a
pavement/embankment, the vertical distance of geosynthetics layers and number of
geosynthetics layers depend on the quality of pavement material. The lower layer of
geosynthetics withstands a tensile stress higher than upper layer. Primary reinforcements for
geosynthetics in piled embankments are located at span between piles with maximum strains
at zones of adjacent piles. Traffic load that passes through on the surface of the pavement can
reduce the soil arching, but it can be restored during the off peak hours. Settlements of
embankments on floating piles can accurately be modelled using the consolidation
calculation type, whereas the end-bearing piles may be used the plastic calculation type.
Longer piles can be effectively applied to reduce a creep. By applying length of floating piles
more than 20% of soft soil depth, it would have a significant impact to reduce a creep on a
deep soft soil.

iii
 TABLE OF CONTENTS

ACKNOWLEDGMENTS ii
ABSTRACT iii
TABLE OF CONTENTS iv
LIST OF FIGURES ix
LIST OF TABLES xiii
LIST OF NOTATIONS xv

CHAPTER 1. INTRODUCTION
1.1. Reinforced Base Course 2
1.2. Piles Supported Embankment 2
1.3. Motivation 3
1.4. Geological Indonesia 5
1.5. Floating Foundation 7
1.5.1. Wooden Mattress Foundation 7
1.5.2. Wooden Piled Foundation 8
1.5.3. Cementitious Piled Foundation 8
1.6. Research Aims 9
1.7. Layout of Thesis 9

CHAPTER 2. LITERATURE REVIEW

2.1. Basic Definitions of Soft Soil 11
2.1.1. Classification System for Soil 11
2.1.2. Fine-grained Inorganic Soil 11
2.1.3. Organic Soil and Peat 11
2.1.4. Soft Soil 12
2.1.5. Parameters for Soft Soil 13
2.1.6. Factors Affecting Behaviour of Clay 13
2.1.6.1. Organic Content 13
2.1.6.2. Sedimentation Rate 13
2.1.6.3. Chemical Weathering 14
2.1.6.4. Freshwater Leaching 14
2.1.6.5. Clay in South East Asia 14
2.1.7. Factors Affecting Behaviour of Peat 14
2.1.7.1. Specific Gravity 14
2.1.7.2. Liquid Limit 15
2.1.7.3. Compressibility 15
2.1.7.4. Permeability 15
2.1.7.5. Properties of Peat Soil in Indonesia 15
2.2. Bearing Capacity of Soft Soil 16
2.2.1. Bearing Capacity at Ground Surface of Soft Soil 16
2.2.2. Contact Area and Tire Pressure 16
2.2.3. Determining of Fill Thickness 17
2.2.4. Settlement Analysis 18
2.3. Quasi-static and Dynamic Loading 18
2.3.1. Definition of Terms 18

iv
 2.3.2. Stress Distribution on Unbounded Material 19
2.3.3. Stress Distribution on Bounded Material 20
2.3.4. Magnitude of Loading 20
2.4. Geosynthetics Reinforcement 21
2.4.1. Type of Geosynthetics 21
2.4.2. Geosynthetics Reinforcement Mechanism 21
2.4.2.1. Restraint and Confinement 22
2.4.2.2. Membrane Mechanism 22
2.4.2.3. Local Reinforcement 22
2.4.3. Geosynthetics Reinforcement Method 22
2.4.3.1. Giroud-Noiray Method (1981) 23
2.4.3.2. Giroud-Han Method (2004) 23
2.4.3.3. US Army Corps of Engineer Method (2003) 24
2.4.3.4. DuPont Method (2008) 24
2.5. Piled Embankments 26
2.5.1. Conventional Piled Embankment 26
2.5.2. Geosynthetics-Reinforced Piled Embankment 27
2.5.3. Soil Arching Concept 28
2.5.3.1. Rectangular Prism: Terzaghi (1943) 28
2.5.3.2. Rectangular Pyramid: Guido (1987) 29
2.5.3.3. Semicircular Arch: Hewlett & Randolph (1988) 30
2.5.3.4. Positive Projecting Subsurface Conduits: BS8006-1 (2010) 31
2.5.3.5. Multi Vaulted-Dome: German Standard (EBGEO 2010) 32
2.5.3.6. Arching Evolution 34
2.5.4. Definition of Terms 35
2.5.4.1. Stress Concentration Ratio 35
2.5.4.2. Column Stress Ratio 36
2.5.4.3. Efficacy 36
2.5.4.4. Stress Reduction Ratio 36
2.5.5. Load Transfer Mechanism 37
2.5.5.1. BS 8006 Method 38
2.5.5.2. Adapted Guido Method 38
2.5.5.3. Adapted Nordic Method 39
2.5.5.4. Adapted Terzaghi Method 39
2.5.5.5. Hewlett and Randolph Method 40
2.5.5.6. German Method (EBGEO 2010) 41
2.5.6. Column Design 41
2.5.7. Tension in Geosynthetics Reinforcement due to Vertical Stress 41
2.5.7.1. BS 8006 Method 43
2.5.7.2. EBGEO (2010) Method 43
2.5.8. Soil Resistance 44
2.5.9. Tension in Geosynthetics Reinforcement due to Lateral Sliding 44
2.5.10. Settlement Analysis 45
2.5.10.1. Different Settlement on the Surface of Embankment 45
2.5.10.2. Settlement on the Bottom of Embankment 46
2.5.10.2.1. Interaction Factor Method 46
2.5.10.2.2. Equivalent Raft Method 48
2.5.10.2.3. Equivalent Pier Method 48
2.5.10.2.4. Piled Raft Method 49
2.5.10.2.5. Japanese Method 49
v
 2.5.10.2.6. Public Works Center Research Method 50
2.5.10.2.7. Scandinavian Method 51
2.5.10.3. Relative Settlement Reduction (RSR) 54

CHAPTER 3. CHARACTERIZATION OF MATERIALS AND LOADING

3.1. Characteristic of Soft Soil in Indonesia 55
3.1.1. Physical Properties 55
3.1.1.1. Physical Properties of Soft Soil in Java Island 55
3.1.1.2. Physical Properties of Soft Soil in Pontianak 56
3.1.2. Mechanical Properties 57
3.1.2.1. Direct Shear Strength 57
3.1.2.2. Compression 57
3.1.3. Bearing Capacity 58
3.2. Embankment Material 58
3.2.1. Material Properties 58
3.2.2. Strength of Material 59
3.2.3. Maximum Height of Embankment 61
3.2.4. Dynamic Properties 61
3.3. Geosynthetics 63
3.3.1. Material Properties 63
3.3.2. Position of Geosynthetics in Pavement Design Practice 65
3.3.3. Tensile Strength of Geosynthetics 66
3.4. Characteristic of Piles 68
3.4.1. Wooden Pile 68
3.4.2. Concrete Pile 69
3.4.3. Stone Column 70
3.4.4. Soil Cement Column 71
3.5. Characteristic of Dynamic Loading 74
3.5.1. Vehicular Traffic 74
3.5.2. Airplanes 74
3.5.3. Trains 75

CHAPTER 4. EXPERIMENTAL WORKS AND FIELD CASE STUDIES

4.1. General 76
4.2. Experimental Works at the Laboratory 76
4.2.1. Geosynthetics 76
4.2.1.1. Horizontal pressure experiment 76
4.2.1.2. Load transfer mechanism experiment 78
4.2.1.3. Geosynthetics-soil-interaction behaviour experiment 79
4.2.2. Piled Embankment 80
4.2.2.1. Pile-soil relative displacement experiment 80
4.2.2.2. Cyclic loading experiment 81
4.3. Case Studies in the Field 83
4.3.1. Geosynthetics 83
4.3.1.1. Weesenstein railway rehabilitation project 83
4.3.1.2. Tabing-Duku road widening project 84
4.3.1.3. Setoko-Nipah road embankment project 84
4.3.1.4. Yamanote Line railway rehabilitation project 86
4.3.2. Piled Embankment 87
vi
 4.3.2.1. Barra de Tijuca field test 87
4.3.2.2. Gebeng Bypass Highway field test 89
4.3.2.3. Kyoto road field test 90
4.3.2.4. Büchen-Hamburg railway section field test 91
4.3.2.5. Brogborough Lake embankment test 93
4.4. Summary 94

CHAPTER 5. MODELLING AND NUMERICAL ANALYSIS

5.1. Analytical Modelling 96
5.1.1. Two-layer System Elastic Theory 96
5.1.2. Analytical Method for Dynamic Response to Beam and Plate
on Winkler Type Elastic Foundation under Moving Loads 97
5.1.3. Analytical Method for Dynamic Response of Layered Half-space
under Moving Loads 99
5.1.4. Critical Velocity 100
5.1.5. Explicit Model for Cyclic Accumulation 100
5.1.5.1. Model Sawicki & Swidzinski 101
5.1.5.2. Model of Bouckovalast et al. 102
5.2. Constitutive Models in Numerical Analysis 102
5.2.1. Mohr-Coulomb Model 102
5.2.1.1. Elastic Perfectly Plastic Behaviour 102
5.2.1.2. Formulation of Mohr-Coulomb Model 103
5.2.2. Hardening Soil (HS) Model 105
5.2.2.1. Constitutive Equations for Standard Drained Traxial Test 105
5.2.2.2. Stiffness for Primay Loading 106
5.2.2.3. Stiffness for un-/reloading 106
5.2.2.4. Yield Surface, Failure Conditions, Hardening Law 106
5.2.2.5. Flow Rule, Plastic Potential Functions 107
5.2.2.6. Parameters of the HS Model 108
5.2.3. Hardening Soil Small (HS Small) Model 108
5.2.4. Soft Soil Creep Model 109
5.2.5. Hypoplastic Model 113
5.2.5.1. Hypoplastic Model for Granular Material 114
5.2.5.2. Intergranular Strain Concept (Small Strain Behaviour) 115
5.3. Geometrical Idealization 116
5.4. Summary 117

CHAPTER 6. DISCUSSION USING FINITE ELEMENT CALCULATIONS

6.1. Introduction 119
6.2. Optimal Vertical Distance of Geosynthetics 119
6.3. Stress Concentration Ratio 128
6.3.1. Influence of Embankment Height on Stress Concentration Ratio 129
6.3.2. Differential Settlement and Critical Height of Embankment 130
6.3.3. Comparison of Stress Concentration Ratio between Experimental
Works and Analytical Methods 131
6.4. Load Transfer Platform 133
6.4.1. Settlement of Embankment over Floating Piles 135
6.5. Low Embankment on the End-bearing Piles 138
6.6. Influence of Traffic Load on Arching Effect 140
vii
 6.6.1. Load Distribution 142
6.6.2. Influence of Traffic Load 145
6.7. Summary 146

CHAPTER 7. CASE STUDY USING FEM

7.1. Location 149
7.2. Runway Reconstruction Work of Supadio Airport 150
7.3. Apron Widening Project of Supadio Airport 154
7.3.1. Settlement at Surface of Pavement 156
7.3.2. Block Behaviour of Settlements 158
7.3.3. Effectiveness of Piles Length on Floating Piles 160

CHAPTER 8. CONCLUSIONS AND RECOMMENDATIONS

8.1. Conclusions 161
8.2. Recommendations for Further Research 163

REFERENCES

viii
 LIST OF FIGURES

1-1 Cross-section of flexible pavement system (after Muench, 2006) 1

1-2 Stress distribution with depth in a flexible pavement, (a) High stress area
directly under wheel load (b) Reduced load at subgrade level 1
1-3 Relative load magnitude at subgrade layer level (a) Unreinforced flexible
pavement (b) Geosynthetics-reinforced flexible pavement 2
1-4 Geosynthetics-reinforced piles supported embankment (after Satibi, 2009) 3
1-5 Location of soft soil in Indonesia (after Puslitbang, 2001) 4
1-6 Geological time scale showing the appearance of life forms and
the occurrence of major geological events (after Mac Kinnon, 1996) 5
1-7 Neotectonic profile of Indonesian archipelago (after Simandjuntak, 1993) 6
1-8 Wooden mattress foundation 7
1-9 Wooden piled foundation 8
1-10 Cementitious piled foundation (after Scottwilson, 2009) 9
2-1 Atterberg limit for organic and inorganic soils (after Puslitbang, 2001) 11
2-2 Contact area and tire pressure (after TENAX, 2001) 17
2-3 Sources of dynamic actions (after EBGEO, 2010) 19
2-4 Typical of load time history (after EBGEO, 2010) 19
2-5 Load Distribution on granular material (after Giroud et al., (1981) 20
2-6 Cross-section for rail track (after Ril 836) 20
2-7 Three stabilization mechanisms (after Perkins and Ismeik, 1997) 22
2-8 Correlation between Cone index, CBR, and Shear strength (after Archer, 2008) 24
2-9 Correlation chart for estimating the subgrade value (after Barenberg, 1975) 25
2-10 Design chart DuPont method (a) Compacted crushed stone thickness
for 1000 axle loads, (b) Factor to determine curve Pi 25
2-11 Typical piled embankments (a) Conventional piled embankment (b) Piled
embankment with basal reinforcement 26
2-12 Soil arching concept (after McNulty, 1965) 28
2-13 Stress state of a differential element (after Terzaghi, 1943) 28
2-14 Guido’s experimental set-up (after Guido, 1987) 30
2-15 Arching through a piled embankment (after Hewlett & Randolph, 1988) 30
2-16 Stress on an element of soil arching (after Hewlett & Randolph, 1988) 31
2-17 Stress distribution of sub-surface conduit (after Marston et al.,1913) 32
2-18 Theoretical arching model (after Zaeske et al., 2001) 33
2-19 Vertical stress redistribution on arching (after EBGEO, 2010) 33
2-20 Arching evolution (after Iglesia et al., 1999) 34
2-21 Generalized ground reaction curve (after Iglesia et al., 1999) 35
2-22 Relationship between n and SRR (after Smith, 2006) 37
2-23 Load transfer mechanism (after Collin, 2004) 37
2-24 Unit cell utilization (after Jones et al., 1990) 38
2-25 Triangular soil arching of Nordic method (after Nordic guideline, 2003) 39
2-26 Dome soil arching model (after Hewlett & Randolph, 1988) 40
2-27 Effective tributary area of column, (a) Triangular spacing (b) Square spacing
(c) Hexagonal spacing 41
2-28 Tension on geosynthetics reinforcement (after Satibi, 2009) 42
2-29 Plan view of foundation and geosynthetics (after Kempfert et al., 2004) 43
2-30 Chart to determine strain in geosynthetics (after Kempfert et al., 2004) 44
2-31 Horizontal outward-thrust resisted by geosynthetics (after EBGEO, 2010) 45
ix
2-32 Piled embankments showing arching soil, notations for geometry and
Settlement (after Zhuang, 2009) 45
2-33 Superposition via the interaction factor method (after Poulos&Davis, 1980) 47
2-34 Interaction factor, (a) Simplified case (b) Influence factor for non-homogeneity
for soil layer (c) Influence factor for bearing stratum (after Poulos, 1968) 48
2-35 Settlement on the end-bearing piles (after PWRC method, 2000) 50
2-36 Influence factor (), (after Han, 2003) 51
2-37 Stress distributions for columns of deep mixed foundation (after Smith, 2005) 51
2-38 Stress-strain relationship in dry mixed column (after Broms, 1999) 52
2-39 Stress distributions beneath the stabilized columns for both cases (after
Broms, 1999) 53
3-1 Oedometer test of Pontianak soft organic soil (after Priadi, 2008) 58
3-2 Geosynthetics subdivision (after PrEN ISO 10318) 64
3-3 Positions of geosynthetics in pavement design practice (after TenCate
Mirafi,2010) 66
3-4 Typical strain vs. Force behaviour of reinforcement (a) Exxon, 1989
(b) Carlson, 1987 66
3-5 Typical of columns arrangements: Triangular and Square grids
(after Priebe, 1995) 70
3-6 Priebe's basic improvement factor (after Priebe, 1995) 71
3-7 Undrained shear strength of lime/cement column (after Kivelo, 1997) 73
4-1 Setup of sample on Ruiken and Ziegler’s laboratory test (after Ruiken and
Siegler, 2008) 76
4-2 Radial strains on Ruiken and Ziegler’s laboratory test (after Ruiken and
Siegler, 2008) 77
4-3 Installation for measuring horizontal pressure (after Ruiken et al., 2010) 77
4-4 Horizontal outward-pressure (after Ruiken et al., 2010) 78
4-5 Test setup with movable front wall (after Bussert, 2006) 78
4-6 Earth stress reduction at movable front wall with different geosynthetics
and without reinforcement of soil body (after Bussert, 2006) 79
4-7 Geosynthetics-soil-interaction-testing device (ITD) (after Althoff, 2010) 79
4-8 Layout of test setup (a) Top view and (b) Side view (after Wei-ping et al., 2007) 81
4-9 Setup of test model under cyclic loading (after Heitz, 2006) 82
4-10 Location of devices in laboratory test (after Heitz, 2006) 82
4-11 Weesenstein railway embankment (after Klompmaker et al., 2008) 83
4-12 Tabing-Duku road widening (after Klompmaker et al., 2008) 84
4-13 Cross-section of road embankment in Setoko and Nipah islands
(after Djarwadi, 2006) 85
4-14 Settlements during work execution (after Djarwadi, 2006) 85
4-15 History of elevated railway and highway structure in Japan
(after Palmeira et al., 2008) 86
4-16 GRS RWs having a FHR Facing in Yamanote Line , a) Typical cross-section
b) Wall under construction and c) Completed wall (after Palmeira et al., 2008) 86
4-17 Staged Construction of GRS RW with a FHR facing (after Palmeira et al., 2008) 87
4-18 Embankment at Barra da Tijuca a) General scheme of piled embankment
b) Pile caps above the initial fill c) Pile caps inside the initial fill
(after Almeida et al., 2008) 88
4-19 Measured strains in geogrid in points at: a) face of the pile cap, b) half distance
between caps (after Almeida et al., 2008) 88
x
4-20 Layout of the major and the control sections of test embankment
(after Hassandi et al., 2005) 89
4-21 Scheme of the full-sclae test Kyoto road (after Eekelen, 2009) 91
4-22 Foundation system at section of Büchen-Hamburg (after Schwarz et al., 2005) 92
4-23 Installation of MIP-columns and Placing of geogrids (after Schwarz et al., 2005) 92
4-24 Settlements of field measurements (after Schwarz et al., 2005) 93
4-25 Distribution of undrained shear strength at Brogborough lake
(after Scottwilson, 2009) 93
4-26 Execution of Brogborough Lake embankments (after Scottwilson, 2009) 94
5-1 Geometry of two-layer system (after Burmister, 1943) 96
5-2 Vertical stress distribution at the surface of second layer for two-layer system
(after Fox 1948) 97
5-3 Cyclic loading (a) Procedure of an explicit calculation of accumulation
(b) Evolution of total strain in a cyclic triaxial test (after Wichtmann, 2005) 101
5-4 Basic principle of elastic perfectly plasticity (after Hill, 1958) 103
5-5 Mohr-Coulomb yield surface in main stress space (c=0) (after Smith, 1982) 104
5-6 Hyperbolic stress-strain relation in primary loading for a standard drained
triaxial test (after Schanz et al., 1999) 105
5-7 Successive yield loci for various values of the hardening parameter p and
failure surface (after Schanz et al., 1999) 107
5-8 Yield surface of HS model, a) Successive yield loci for shear hardening
and compression hardeing in p-q space, b) Total yield contour in principal
stress space (after Schanz et al., 1999) 107
5-9 HS-small model, a) Initial stiffness modulus E0 in a triaxial test
b) Small strain parameters E0 and  0.7 (after Benz, 2007) 109
5-10 Consolidation and creep behaviour in a standard oedometer test
(after Vermeer and Neher, 1999) 110
5-11 Standard oedometer test, a) Stepwise loading in e-log vs.  ’plot
b) Void ratio vs. Time (after Vermeer and Neher, 1999) 111
5-12 Diagram of peq-ellipse in a p-q plane (after Vermeer and Neher, 1999) 112
5-13 Influence of n (a) and hs (b) on oedometric curves (after Herle&Gudehus, 1999) 114
5-14 Dependency of the reference void ratio of ed0, ec0 and ei0 on the mean stress
(after Herle and Gudehus, 1999) 115
5-15 Geometrical idealizations of one cell piled embankment (a) Axisymmetrical
idealization (b) Idealization for plane strain after Bergado (c) Idealization plane
strain using equivalent stiffness (d) 3D geometry (after Satibi, 2009) 116
6-1 Schematic work procedure for FE analysis 119
6-2 Horizontal displacement vs. Number of geosynthetics layers 126
6-3 Maximum stress using MC-model at different elastic modulus of base course 127
6-4 Maximum stress using HS-model at different elastic modulus of base course 128
6-5 Influence of embankment height due to vertical stress during embankment
filling of Test 7 129
6-6 Variations of vertical stresses and stress concentration ratios, (a) Vertical stress
during water discharge test 7, (b) Influence h/s on stress concentration ratios 130
6-7 Differential settlements of Test 1 (above) and Test 7 (below) 130
6-8 Shapes of deformation for h/s =1.4 131
6-9 Comparison between test results, Analytical methods and Finite element
method for stress concentration ratio: (a) Test 1, (b) Test 2,(c) Test 3,
xi
 (d) Test 4,(e) Test 5,(f) Test 6 and (g) Test 7 133
6-10 Schematic sections of LTP (after Hassandi et al., 2007) 134
6-11 Total settlement of LTP sections (at centre of square pattern of geopiers) and
Control sections: (a) Section 1 (b) Section 2 (c) Section 3, 4 (d) Control sections 137
6-12 Layout of field test at Barra da Tijuca district in Brazil (after Almeida, 2008) 138
6-13 Settlements at the centre of 4 piles 140
6-14 Layout of the Kyoto road field test (after Eekelen et al., 2009) 141
6-15 Isochrones of geogrids (after Eekelen, 2009) 142
6-16 Load distribution in geosynthetics-reinforced piled embankment
(after Eekelen et al., 2008) 142
6-17 Load Distribution for A, B and C of vertical stresses observed throughout
2 ½ years, Prediction using Analytical method and FE-calculation
(a) Vertical stress (load A) directly on pile above geosynthetics, (b) Vertical
stress (load B) under geosynthetics, (c) Vertical stress (load C) on subsoil 144
6-18 Vertical stresses on piled embankment 145
6-19 Daily arching cycle on the Kyoto road, (a) Load A, (b) Load B
(after van Eekelen et al., 2008) 146
6-20 Passage of a truck of 397 kN with 2x2-axles (after Eekelen et al., 2008) 146
7-1 Aerial view of Supadio airport 149
7-2 cross-section of Supadio airport runway (reconstruction pavement in the middle
And existing pavement on left-right side) 150
7-3 Execution of Supadio airport runway reconstruction work (a) Driving piles
(b) Pouring base course (c) Compaction 151
7-4 Incremental displacement 153
7-5 Settlements at the weak-spot of Supadio airport runway (a) with piles
(b) without piles 154
7-6 Cross-section of Supadio airport apron 155
7-7 Total settlements on the surface of new apron 156
7-8 Predicted settlement at surface of of parking stand 1 157
7-9 Settlements on the surface of concrete apron at Supadio airport 157
7-10 Settlements at different length of piles (a) Pile length of 2.5 m (b) Pile length
of 5.0 m (c) Pile length of 10 m (d) Pile length of 15 m 159
7-11 Predicted total settlements of apron pavement without piles 159
7-12 Effectiveness of pile length, (a) No load of airplane (b) Maximum load of
airplane 160

xii
 LIST OF TABLES

1-1 Distribution of peat soils in Indonesia 4

1-2 Distribution of thickness for peat soils in Indonesia 5
2-1 Soil classification based on organic content 12
2-2 Consistency of clay (after Indonesian geotechnical guide-4, 2001) 12
2-3 Field indicator of undrained shear strength for soft clay 12
2-4 Range of realistic values for parameters of soft soils 13
2-5 Permeability values for peat soils 15
2-6 Properties of peat soils in Sumatera and Kalimantan islands 16
2-7 Comparison of bearing capacity in different design methods 18
2-8 Type of geosynthetics used in reinforced earth (after Suryolelono, 1997) 21
2-9 Aggregate efficiency 26
2-10 Piled embankments geometry used over the past 35 years 27
2-11 Critical height for different design methods 46
3-1 Characteristics of soft soils in Java island 55
3-2 Subsoil properties at several zones in Pontianak 56
3-3 Empirical bearing capacity factor 58
3-4 Various materials for embankments 59
3-5 Design parameters for embankment material 59
3-6 Correlation between Vs and N 62
3-7 TBR and BCR resulted from laboratory and field test 65
3-8 Tensile strength of geosynthetics (after EBGEO 2010) 67
3-9 Conversion factors of geosynthetics reinforcements (after Nordic guideline, 2003) 67
3-10 Conversion factors for long-term properties (after Nordic guideline, 2003) 67
3-11 Conversion factors  for damage during installation (after Nordic guideline, 2003) 68
3-12 Strength class for wooden material (after PKKI, 1961) 68
3-13 Quality code for wooden material (after SNI Kayu, 2002) 69
3-14 Typical Japanese mixing installation parameters (after Kaiqiu, 2000) 72
3-15 Specified values of qu on deep mixing projects in the U.S. 72
3-16 Relationship between E50 and qu 73
3-17 Aerodrome reference code (after ICAO, 1999) 75
4-1 Specification of interaction testing device 80
4-2 Main characteristics of two piled embankments 87
4-3 Total and Differential settlements at the base of the embankments 90
5-1 Material parameters for the Soft Soil Creep model (after Vermeer and
Neher, 1999) 113
6-1 Resilient moduli of base course and subbase course 120
6-2 Material properties used in finite element analysis 120
6-3 Horizontal displacements for MC-model, E=200 MPa and EA= 30 kN/m 121
6-4 Horizontal displacements for MC-model, E=600 MPa and EA= 30 kN/m 121
6-5 Horizontal displacements for MC-model, E=200 MPa and EA= 300 kN/m 122
6-6 Horizontal displacements for MC-model, E=600 MPa and EA= 300 kN/m 122
6-7 Horizontal displacements for HS-model, E=200 MPa and EA= 30 kN/m 123
6-8 Horizontal displacements for HS-model, E=400 MPa and EA= 30 kN/m 123
6-9 Horizontal displacements for HS-model, E=600 MPa and EA= 30 kN/m 124
6-10 Horizontal displacements for HS-model, E=200 MPa and EA= 300 kN/m 124
6-11 Horizontal displacements for HS-model, E=400 MPa and EA= 300 kN/m 125
6-12 Horizontal displacements for HS-model, E=600 MPa and EA= 300 kN/m 125
xiii
6-13 Material properties used in finite element analysis 129
6-14 Settlements at surface of embankment using FE-analysis 131
6-15 Equations in analytical method of two-dimension 132
6-16 Properties of geogrids used in the field test (after Hassandi et al., 2007) 135
6-17 Material properties used in finite element analysis at Gebeng’s highway
embankment test 135
6-18 Material properties used in finite element analysis at Barra da Tijuca district
in Brazil 139
6-19 Comparisons of settlements at base of embankment from field measurement
and FE-calculation 139
6-20 Comparison of tensile strains for geosynthetics from field measurement
due to FE-analysis 140
6-21 Properties of fill material used in Kyoto road embankment 141
6-22 Young’s moduli of Kyoto road’s subsoil 141
6-23 Tensile stiffnesses of geogrids reinforcement (after Eekelen, 2009) 142
6-24 Material properties used in finite element analysis of Kyoto road 143
7-1 Material properties used in FE-analysis on Supadio airport runway reconstruction
work 152
7-2 Settlements and types of loading for pavement reinforced by wooden piles 152
7-3 Settlements and types of loading for pavement unreinforced by wooden piles 153
7-4 Weight and ACN for airplane of Boeing 737-400 (after ICAO, 1999) 155
7-5 Material properties used in finite element analysis of Supadio airport apron
widening project 156

xiv
 LIST OF NOTATIONS

a width of the pile cap

A total cross-sectional area of the 'unit cell' attributed to each column
a radius of the circular footing
A,B,C,D fitting parameters
A/Ac reciprocal area replacement ratio
A1 reduction factor for creep
A2 reduction factor for damage caused by transport, installation, compression
A3 reduction factor for processing connection
A4 reduction factor for environmental influences (weather, chemical, micro-
organism)
Ac cross-sectional area of one column
AG plan area of pile group including the soil between the piles
as area replacement ratio or coverage ratio
B width of underground structure
2B width of strip
bErs width of column
C competency ratio
c viscous damping constant
C1 , C2 material constant
CBR California Bearing Ratio
CBRbc CBR value of the base course
CBRsg CBR value of the subgrade soil
Cc soil arching coefficient depending on H and a
Cc = 1.5 H / a – 0.07 for floating piles
Cc = 1.95 H / a – 0.18 for end-bearing piles
Cc coefficient of curvature
cc compression index
ch average coefficient of horizontal consolidation
cr recompression index
cs swelling index
CR compression ratio
CSR column stress ratio
Cu coefficient of uniformity
cu undrained cohesion
cv consolidation coefficient
c  swelling index
d diagonal pile spacing
D Euler's stretching tensor an
D damping ratio
d thickness of pavement (or concrete layer),
D column diameter
d depth
D flexural rigidity of the plate strip
D strain rate
D thickness of stabilized clay layer (or sub-layer)
xv
d* underlying stratum
D10 diameter of grains passing sieve at 10 percentage
D60 diameter of grains passing sieve at 60 percentage
Dacc given accumulation rate
de equivalent diameter
di reinforced depth
di stratum (or sub-layer) thickness within the reinforced depth
Dmax maximum diameter of grains
Dmin minimum damping ratio
E modulus of elasticity
E Young's modulus,
E efficacy
E elastic stiffness
e void ratio
E1, E2 elastic modulus of first and second layer
E50 secant modulus
E50ref reference stiffness modulus corresponding to the reference stress pref
Ec modulus deformation of the columns
Ec Young’s modulus of concrete
ed, ec and ei reference void ratio corresponding to the non-zero stress
Ek horizontal outward-thrust
emax maximum void ratio
emin minimum void ratio
eo initial void ratio
Eoed oedometer compression modulus
Eoedref reference tangent stiffness for primary oedometer loading
ESAL equivalent single axle load
Eurref reference un-/reloading modulus corresponding to the reference pressure ref
f yield function
Fb flexural strength
fB low frequency
fc partial factor for construction safe
fc' characteristic compressive strength 28 day
Fc compressive strength of perpendicular fiber
fc factor for bearing capacity of subgrade soil
Fc tensile strength for perpendicular fiber
fcf flexural-tensile strength 28 day
fd partial factor for mechanical damage
Fd permissible long term tensile strength
fenv partial factor for environment
ffs partial load factor for soil unit weight
Fk short term tensile strength
Fk vertical load
fm partial factor for extrapolation of tensile strength
fq partial load factor for applied external load
fs factor for rutting depth of 75 mm
Fs,G,k vertical load at static loading

xvi
Fs,G+Q,k vertical load at static and variable-dynamic loading
Ft tensile strength for parallel fiber
Fv shear strength
g plastic potential function
G shear modulus of material
G specific gravity
G specific gravity of wooden material
G0 initial shear modulus
G0ref elastic small strain shear modulus at reference pressure pref
Gmax maximum shear modulus of material
Go initial shear modulus of material
h height of embankment
H thickness of soil above the point
h required base course thickness
H thickness of soil above the point
h thickness of the first layer
H thickness of fill granular material
h thickness of first layer
H thickness of soil layer
h/s ratio of embankment height to clear spacing
Hc critical height of embankments without ground treatment
hg height of arching
hg = h for h < s / 2
hg = s / 2 for h > s / 2
h'o thickness of unreinforced aggregate base
Hr thickness of reinforced granular material
Hu thickness of unreinforced granular material
I moment inertia of beam
Io Bessel function of the first kind and order of zero
Ip placticity index
J geogrids aperture stability modulus
Jk tensile stiffness of geosynthetics
K earth pressure coefficient
k modulus of subgrade reaction
k foundation stiffness
k reaction modulus of subgrade
k stiffness of geosynthetics
k1, k2,k3 regression coefficients obtained from regression analysis
Ka active earth pressure coefficient (Ka = tan2 (45+/2)
Kagh active lateral earth pressure coefficient according to DIN 4085
Kb empirical bearing capacity factor
ki permeability
Kkrit critical ratio of major stress = tan2 [45o + 'k / 2]
KoNC Ko-value for normal consolidation
Kp coefficient of passive earth pressure
kv permeability
Kw coefficient instead of K
L length of span
xvii
L length of the column
L pile length
L column length
ℒ fourth order constitutive tensor
l radius of relative stiffness
L/d to optimum column aspect ratio
LL liquid limit
Lw clear spacing
m power for stress-level dependency of stiffness
m dimensionless parameter
M slope of critical state line
m mass per unit length of the beam
ME compressibility modulus
MV rigidity modulus
Mr resilient modulus
Msoil oedometer compression modulus of surrounding soil
mv coefficient of compressibility of untreated soft clay
M  fourth-order tangent stiffness tensor of the material
n number of piles
N second order constitutive tensor
N number of axle passes
N number of blows from Standard Penetration Test (SPT)
N number of cycles
N number of passes of axle load
n porosity
n stress concentration ratio
Nc factor of bearing capacity
OC organic content
OCR state of preconsolidation stress
P constant concentrated vertical with moving load
P wheel load
p total applied pressure of the embankment
p mean principle stress
p support pressure from roof above underground structure
p tire inflation pressure
p* normalized loading (p/po)
PA axle load
Pa reference pressure (atmospheric pressure)
Pav average load on a pile within group
Paxle axle load
pc vertical stress on the column
peq constant on the ellipse in the p-q plane
pG,k static loading
pG+Q,k static and variable loading
Pi actual axle load
PI plasticity index
Po equivalent standard axle load of 80 kN

xviii
po nominal overburden total stress at the elevation of underground
po pressure on the circular footing
pz vertical pressure on the circular footing
q distributed surcharge load
q surcharge acting at the surface of the soil
q ultimate vertical stress in a stone column
q vertical stress at the subgrade surface
Q design dynamic load
qa quantity
qc cone tip resistance
qf ultimate deviatoric stress
Qr required design vertical load
qu unconfined compressive strength
qu unconfined compressive strength
r column radius
r horizontal distance from centerline
R radius of equivalent tire contact area
R aspect ratio of the group
R radius of circular area
R,r,,mR,mT five addtional parameters for iIntergranular strain concept
Rf failure ratio
RSR relative settlement reduction
s diagonal pile spacing
s allowable rutting depth
s column spacing
s pile spacing
s rut depth
S settlement of embankment supported by piles
S stiffness coefficient
s/d ratio of pile spacing to pile cap diameter
S0 embankment settlement constructed on soft soil without the support of piles
S1 compression of the stabilized volume for case 1
S1 settlement of a single pile under unit load
S2 compression of stabilized volume for case 2
s-a width of strip
Sc settlement of the columns
So settlement of untreated soil subjected to the actual load of embankment
Sp primary settlement
Sr degree of saturation
SRR stress reduction ratio
Ss settlement of untreated soil subjected to reduced pressure
Ss secondary settlement
su undrained shear strength
t time
T corrected thickness of base course
t1 time in first step
t2 time in next step
Teff effective aggregate thickness
xix
To initial aggregate thickness
TRP tension in geosynthetics
Jaumann stress rate
V constant velocity
v Poisson's ratio
vcr critical velocity
vur un-/reloading Poisson’s ratio
Vs shear wave
w lateral deflection of plate strip
w water content
w deformation at interface between two layers
w(r) vertical displacement of the surface
wi settlement of a pile i within a group of n piles
Wn water content
WT equivalent vertical load
z thickness of the soil overlying the element
ϵacc v volumetric strain
 dstab consolidation settlement of stabilized ground
d consolidation settlement of unstabilized ground
P different stress
Sr differential settlement between the columns and the untreated soil
 state ’compaction’  =  n /n0
ij interaction factor for pile i due to any other pile j within the group
  influencing factor due to the presence of geosynthetics layer
  settlement reduction ratio
 maximum sag (vertical deflection) of geosynthetics
 settlement of roof
 intergranular strain tensor
* normalized displacement ( / B)
 acc(N) average accumulation curve
 ampl strain amplitude
elastic strain rate
accumulation of deviatoric strain
accumulation of volumetric strain
e elastic strain
p plastic strain
plastic strain rate
 strain
v vertical strain
k strain in geosynthetics
’ effective internal friction angle
c, hs, n, ed0, ec0, ei0,  ,  eight material parameters for hypoplastic model
max peak secant angle of shearing resistance
 internal friction angle of the soil
xx
 unit weight of the soil
 partial safety coefficient for the consideration of possibility
k unit weight of embankment
  unit weight of the soil
ampl shear strain amplitude
H nominal vertical stress at the base of embankment
p plastic shear strain
  unit weight
av average stress ratio
 creep factor
  installation damage factor
 biological and chemical degradation factor
* modified swelling index
 plastic multiplied factor
* modified compression index
 Poisson's ratio
* modified creep index
1, 2 Poisson’s ratio of first and second layer
 density of soil
 average stress of applied embankment plus surcharge at level of top piles
a allowable strength
3 confining stress
c stress on the columns
c vertical stress on the pile cap
d deviatoric stress
creep strength of column
h' effective horizontal pressure on the columns
h horizontal stress
i radial (vertical) stress acting immediately beneath the crown of the arch
pile vertical stress on piles
rz vertical stress at interface between two layers with radius = r
s reduced pressure on the untreated soil due to embankment
s stress on the subsoil
soil vertical stress on soil
u ultimate strength
v vertical stress
z vertical stress at interface between two layers with radius = 0
flexural flexural strength of wooden material
'ro free-field lateral effective stress
compressive compressive strength of wooden material
comp or tensile tensile strength of wooden material
θ bulk stress (σ1 + σ2 + σ3 )
 shear strength
oct octahedral shear stress
u,col undrained shear strength of the columns
xxi
 shear strength of wooden material
(x0 , z) fixed coordinates along the length of strip and in the vertical direction

xxii
 CHAPTER 1

Introduction

The increasing need for infrastructure development has led engineers to find an alternative on
soft soils. Engineers are continually faced with maintaining and developing pavement
infrastructure with limited financial resources. Traditional pavement design and construction
practices require high-quality material for fulfilment of construction standards. In many areas
of the world, quality materials are unavailable or in short supply. Due to these restrictions,
engineers seek alternative designs or innovative design practices.
Weak foundation soils have always been a challenge to geotechnical engineers when
designing infrastructure. Low bearing capacity, slope stability, lateral pressure, movements
and differential settlements are some major concerns. However, a variety of techniques are
available to address these issues, including preloading, deep mixing columns, stone columns,
use of lightweight fill and soil replacement. Flexible pavement layered system is shown in
Fig.1-1. Meanwhile, vertical load distribution induced by a moving wheel load is illustrated
in Fig. 1-2.

Fig. 1-1 Cross-section of flexible pavement system (after Muench, 2006)

(a) (b)

Fig. 1-2 Stress distribution with depth in a flexible pavement, (a) High stress area directly
under wheel load (b) Reduced load at subgrade level

1
1.1. Reinforced Base Course
Fig. 1-3 shows a typical section of reinforced base course, which consists of an aggregate
base layer, a subgrade and a reinforcement layer usually placed between the base course and
subgrade. The base course and geogrid transmit the traffic load to the top of the subgrade.
Under repeated load, the behaviour of the base-geogrid-subgrade system is complicated. The
overall behaviour depends on the properties of geosynthetics, soil characteristics, and the
interaction between the soil and the reinforcement.

Moving Wheel Load

(a) (b)

Fig. 1-3 Relative load magnitude at subgrade layer level (a) Unreinforced flexible pavement
(b) Geosynthetics-reinforced flexible pavement

1.2. Piles Supported Embankment

Soil alone is only able to carry compressive and shear forces. However, through the use of
geosynthetics as reinforcing elements, soil structures can be built to carry tensile forces. It is
real vision that the reinforcement of soil with geogrids will be common in the future as
reinforcement of concrete with steel mesh is nowadays. In roadway, runway and railway
applications the insertion of horizontal geogrid layers in granular base course provide an
increased modulus, hence it provides lateral confinement to the system.
For soft soils, one of the most promising solution to these problems is to use piled
embankments. In many cases, this method appears to be the most practical, efficient (low
long-term cost and short construction time) and environmentally-friendly solution for
construction on soft soil. Field applications are mainly highways and railways.
Piles are installed through the soft subsoil and transfer load to deeper subsoil. The majority of
the load from embankments and surcharge (pavements and live load) are carried by the piles
and thus there is relatively little load on the soft subsoil. Piles are typically arranged in square
or triangular patterns in the field. By using a piled embankment technique, the construction
can be undertaken in a single stage without having to wait for the soft soil to consolidate.

2
Settlements and differential settlements are also significantly reduced once the technique is
used successfully.
Geosynthetics-reinforced piles supported embankment is a promising technique when it is
constructed over weak soil as depicted in Fig. 1-4. It is also valuable in two ways; if there is
adjacent structure that is sensitive to differential settlement and if there is a need for
accelerated construction. This complex foundation system consists of strong piles in soft soil
and compressible soil. A bridging layer is placed above the columns consisting of coarse
grained soil and one or more layers of geosynthetics is situated on the bridging layer.

(a) (b)
Fig. 1-4 Geosynthetics-reinforced piles supported embankment (after Satibi, 2009)

1.3. Motivation
Stiff soils are preferable as a foundation for buildings, roadways and other infrastructures.
However, these types of foundation soil are not always available on site. In tropic regions like
Indonesia, soft soil deposits (peat, organic and inorganic soft soil) occur in lowland and
highland areas and generally termed as basin and valley of peat. In some places, the depth of
soft soil can reache 30 meters or even more.
Soft soil exhibits very low bearing capacity and it is not suitable for constructing
embankment, highway, runway, railway, building or any other load bearing engineering
structure. Organic soft soil in its natural state consists of water and decomposing plant
fragments with virtually no measurable bearing strength. Peat and organic soil are considered
as soft soil because they have high settlement value and even under moderately applied load.
The most extensive areas of peat soil are located in the northern hemisphere. It is estimated
about 1 billion acres of the world are peat or equal to 4.5% of total land area. In United States
peat is found in 42 states, with a total area of 30 million hectares. Canada has 170 million
hectares and the former USSR has 150 million hectares. In Japan, peat is widely distributed
throughout Hokaido with an area approximately 200 thousand hectares (Alwi, 2007).
The world tropical peat land is located in Asia, South Africa and Latin America covering
around 30 million hectares. Around two thirds (66%) are in Southeast Asia (Global
Environtment Centre). While in Indonesia peat soil covers about 20 million hectares of the
country land. Most extensive areas are located across eastern Sumatra, western and southern
Kalimantan and southern Papua. The soft areas are mostly dominated by peat overlaying soft
clay soils, whereas there are mainly soft clay soils spreading out in other places as depicted in
Fig. 1-5.
3
 Kalimantan

Sumatera

Maluku
Sulawesi Papua

Java

Fig. 1-5 Location of soft soil in Indonesia (after Puslitbang, 2001)

The total area of soft soils (peat and soft clay soil) in Indonesia is about 60 million hectares,
which contribute around 30 percent of the total land area (Satibi, 2009). Distribution of peat
soils in some regions or provinces in Indonesia is reported by Radjaguguk (1991) resulted
from a study conducted by Euroconsult in 1983 and the final report by BB Litbang (2008).
This is shown in Table 1-1. Meanwhile, Table 1-2 illustrates the depth distribution of peat
soils.
Table 1-1 Distribution of peat soils in Indonesia
Area of peat soils in Indonesia (hectares)
Islands / Provinces
After Radjaguguk, 1991 After BB Litbang, 2008
Sumatera Island
Nanngroe Aceh 270 -
North Sumatera 335,000 -
West Sumatera 31,000 -
Riau 1,704,000 4,043,600
Jambi 900,000 716,839
South Sumatera 990,000 1,483,662
Bengkulu 22,000 -
Lampung 24,000 -
Kalimantan Island
West Kalimantan 4,610,000 1,729,980
Central Kalimantan 2,162,000 3,010,640
South Kalimantan 1,484,000 331,629
East Kalimantan 1,053,000 -
Sulawesi Island
Central Sulawesi 15,000 -
South Sulawesi 1,000 -
Southeast Sulawesi 18,000 -
Maluku Archipelago 20,000 -
Papua Island 4,600,000 7,001,239
Java Island 25,000 -
17,994,270 18,317,589
Source: Radjaguguk, 1999; BB Litbang, 2008
4
 Table 1-2 Distribution of thickness for peat soils in Indonesia
Distribution based on thickness of peat (%)
Total area
Province Shallow Intermediate Deep
(hectares)
0-150 cm 100-200 cm >200 cm
Riau 8.6 10.7 80.7 486,339
Jambi 33.4 9.3 57.3 168,163
South Sumatera 63.0 11.5 25.5 317,784
West Kalimantan 39.5 34.6 25.9 100,754
South & Central Kalimantan 62.6 19.6 17.8 190,145
Total 1,263,185
Source: Radjaguguk, 1991

1.4. Geological Indonesia

The biogeography of the whole Indonesian archipelago and the distribution of its soil, plants
and animals have been determined by the geological and climatic history. Fig. 1-6 shows a
story of plate tectonics and continental drift, climatic events and changing sea levels
(MacKinnon et al., 1996).

Fig. 1-6 Geological time scale showing the appearance of life forms and the occurrence of
major geological events (after MacKinnon et al., 1996)
5
As recently as of 25 million years ago, very recently on the geological time scale as in Fig. 1-
6, the Indonesian archipelago as we know it today simply did not exist, but the story began
much earlier than that.
The continental land masses are by no means permanent and the earth is in a dynamic state.
The outer solid part of the earth, the crust, is quite thin, like the rind of an orange. The are
two kinds of crust: oceanic and continental. Oceanic crust is usually young (0-200 millions
years), thin (5-15 km) and composed mostly of dense volcanic rocks. Continental crust often
has a core of older rocks (200 to 3,500 million years), is thicker (20-50 km) and is less dense
than oceanic crust, composed of rocks such as sandstones and granites. Western Indonesia,
comprised of Kalimantan, Sumatra, and West and Central Java, is composed predominantly
of continental crust, as is much of the shallow sea floor between these islands. Below the
crust of the earth is a zone where the rock is hotter and more plastic. Continental and oceanic
plates float on a fluid, underlying material (Simandjuntak, 1993).

Fig. 1-7 Neotectonic profile of Indonesia archipelago (after Simandjuntak, 1993)

Plate movements are very slow, only a few centimeters per year, but over a 60-million year
period (the time since the dinosaurs became extinct) a plate drifting only 1 cm per year would
have moved 600 km. The Indonesian region is dominated by three major plates, namely
Southeast Asian plate, the Indo-Australian plate and the Pacific plate, as well as several
smaller platelets that have sheared off Irian Jaya. The process is still continuing today
(Gorsel, 2012).
Soil conditions are very important in affecting the distribution of vegetation. There are five
factors in soil formation: lithology, climate, topography, biological organisms and time. In
Kalimantan, the majority of its soils have developed on the rolling plains and dissected hills

6
on sedimentary and old igneous rock. These soils range from strongly weathered and acid
ultisols to young inceptisols. In the south, extensive alluvial plains and peat soils extend into
the Java Sea. Weathering is strong in the humid tropics, favoured by both warmth and
moisture. Because of the high rainfall, soils are constantly wet and soluble constituents are
removed. This process is called leaching. High levels of weathering, leaching and biological
activity (degradation of organic matter) are characteristics of many Bornean soils (van
Gorsel, 2012).

1.5. Floating Fondation

Choosing floating foundation is economic and effective in the presence of deep soft soils
more than 30 meters below the surface. Traditional floating foundation using wooden piles is
still commonly used today‘s construction on soft soil in Indonesia, particularly in
Kalimantan. Mini wooden piles ranging in size from about 10 to 17 cm in diameter and
approximately 400 to 1500 cm in length are widely used as a foundation to support
construction of building, e.g. houses and shopping centres.
For highway or road constructions, there are two ways to install the wooden piles over soft
soils. The first is similar to a mattress foundation where wooden piles are laid down
horizontally over the ground surface and transversely axis of the road. The second using
wooden piles are driven into the ground surface vertically and lay square board caps at the
top piles.

1.5.1. Wooden Mattress Foundation

In the past, wooden mattress foundations were usually used for road access or temporary
constructions and for low volume traffic. After completing the installation of piles, selected
material would be filled over this foundation. After that base course and asphaltic layers can
be performed over the embankment. Later on, in accordance with the increased traffic
volume and axle load, this kind of construction is rarely performed anymore besides some
drawbacks such as weathering of material because the piles are not submerged fully under
water. Figure 1-8 illustrates wooden mattress foundation technique located on soft soil in
West Kalimantan.

Fig. 1-8 Wooden mattress foundation

7
1.5.2. Wooden Piled Foundation
Nowadays, embankment of granular soils supported by piles are often used for road and
railway constructions. The required load carrying capacity of the pile is typically less than
around 30 tonnes. This implies that the piles used for soil improvement are micropiles, i.e. a
pile with diameter less than 30 cm. At Kyoto Road in Netherlands, wooden pile 18 cm in
diameter is connected with concrete cap 30 cm in diameter and 40 cm deep (Eekelen et al.,
2009).
In West Kalimantan, the mini pile with board cap is sometimes used in highway construction
even for runway construction. A mini pile with 10 to 17 cm in diameter and 400 to 1500 cm
long is driven into the ground surface with piles spacing ranging between 40 and 50 cm.
Later on, a square board cap around 20 cm in diameter is nailed to the top of the pile. Post
levelling the ground surface as high as the top surface of board caps, geosynthetic sheets can
be unrolled over this surface and followed by a selected material or granular material. Even in
some cases, construction of piled embankment without using pile cap can be done. This
technique is illustrated in Figure 1-9.

Fig. 1-9 Wooden piled foundation

1.5.3. Cementitious Piled Foundation

The piled embankment supported by piles with geogrids has been considered as an
advantageous alternative metod for solving problem some transport infrastructures over soft
soil deposits. Piles and/or pile caps can be cementitious material which gives more high
compressive strength to support higher vertical stress when being subjected to a load on
surface of pile caps. Some places in the world have applied this type of construction to
overcome some problems related to soft soils.
These types of piles are generally installed in the field by driving them downward using
machine. When a depth of soft layer is shallow, bottom tip of piles can be driven until hard
stratum of soil in order to reach maximum bearing capacity. Otherwise, floating piles become
alternative way if hard stratum of soil is deep enough. After installation is completed,
geosynthetics can be readily unrolled over them before filling up embankment material.
Figure 1-10 shows cementitious piled foundation techniques.
8
 Fig. 1-10 Cementitious piled foundation (after Scottwilson, 2009)

1.6. Research Aims

This research is focused on understanding behaviour of dynamic loading on pavement over
deep soft soil. Some topics are studied, including soil arching analysis, geosynthetics
reinforcement, piled embankments, settlements of construction and stress analysis on
pavement (improved subgrade or embankment) resulted from static and/or live load. Several
methods are used for approaches such as Giroud-Han, British Standar BS8006, Nordic
Standard, German Method EBGEO 2010 and Finite Element Method. Behaviour of loading
over pavement or embankment is analyzed comprising of static loading, moving load and
repeated loading.

1.7. Lay Out of Thesis

Thesis is arranged into 8 chapters as indicated below:
Chapter 2 reviews some terminologies which include arching concept, capping ratio, the
stress concentration ratio, stress reduction factor and efficacy. Thus, providing theoretical
approaches in geosynthetic reinforcement and piled embankments. Followed by describing
the method by calculating of settlement and differential settlement for piled embankment.
Chapter 3 reviews characteristics of soft soil such as physical and mechanical properties and
also describing properties of embankment material. Meanwhile, properties of geosynthetics
as reinforcing material is provided, particularly tensile strength and also presented
compression strength of pile materials such as wooden pile, concrete pile, stone column and
soil cement column and characteristic of dynamic loading on pavement as well.

9
Chapter 4 presents some experimental works in laboratories and field case studies for
various situations over the world in handling the soft soil. This chapter describes some
problems and simultaneously solves on the pertaining realm in soft soil such as excessive
settlement, horizontal thrust, vertical distance of geosynthetics layer, various kinds of piles,
static and cyclic loading.
Chapter 5 provides some analytical approaches of rutting on the surface of the pavement as
material response under static and moving load as well settlement on foundation. Moreover,
some constitutive models are introduced that can be applied in finite element analysis.
Moreover, geometrical idealization is needed to transform the real geometric shapes to finite
element analysis.
Chapter 6 discusses some interesting findings in Chapter 4 and then verified with numerical
analysis using Plaxis software package. Some important findings is presented such as optimal
vertical distance of geosynthetics layer in base course, influence of very soft soil against soil
arching and/or stress concentration ratio, determining critical height of embankment, various
types of load transfer platform, tensile strain of geosynthetics in piled embankment and creep
phenomenon in soft soil.
Chapter 7 elaborates excessive settlements of pavement construction supported by floating
piles over deep soft soil of two case studies in Supadio Airport Pontianak regarding with
runway reconstruction work and apron widening project.
Chapter 8 presents the main findings and concludes this study. Later on, recommendation
for further works is highlighted.

10
 CHAPTER 2

Literature Review

2.1 Basic Definitions of Soft Soil

2.1.1. Classification System for Soil
The word of soil, in civil engineering, will refer to material that is not petrified, not bedrock,
composed of mineral grains, organic matter, has a shape and size, water and gas. Commonly,
soil consists of peat, organic soil, clay, silt, sand and gravel. Soil itself has been classified by
the United States, known as the Unified Classification (Unified Soil Classification System or
USCS). This system is classified based on particle size, distribution and properties of fine
grain. The USCS divides the soil into three main categories: coarse grained soil, fine grained
soil and high-yield organic soil. Meanwhile, inorganic soil is divided into sub-classification
of gravel, sand, silt and clay.

2.1.2. Fine-Grained Inorganic Soil

Fine-grained inorganic soil is divided into sub-groups i.e. silt (M) and clay (C). Silt is a fine-
grained soil that has a liquid limit and plastic index if it is plotted into Fig. 2-1 the line-A,
whereas the clay will be above the line.

Fig. 2-1 Atterberg limit for organic and inorganic soils (after Puslitbang, 2001)

2.1.3. Organic Soil and Peat

There is also fine grained soil, in which Indonesian Geotechnical Guide-1, 2001 divides into
three groups based on their organic content, as shown in Table 2-1. Organic soil (O) is the
soil which has the organic content 25 to 75%. Organic soil is categorised into OL and OH
appropriate level of plasticity. Peat is a soil that has organic content more than 75%. Based on
its fiber content, peat soil is grouped into two: amorphous with fiber content less than 20%
and fibrous with the fiber content exceeds of 20%.
11
 Table 2-1 Soil classification based on organic content
Organic content Soil group

>75% Peat
25 - 75% Organic soil
< 25% Low organic soil
Source: Indonesian Geotechnical Guide-1, 2001)

Peat as used today includes a vast range of peat, peaty organic soils, organic soils and soils
with organic content (Landva et al., 1983). According to the ASTM standard D 2487-00,
organic clay/silt with sufficient organic content will influence the soil properties. By
classification, an organic soil is a soil that would be classified as a clay/silt except that its
liquid limit value after oven drying is less than 75% of its liquid value before.

2.1.4. Soft Soil

There are two types of soft soil: soft clay and peat. Soft clay has clay minerals and a high
moisture content causing low shear strength. The undrained shear strength (cu) for the clay
soil as indicated in the following Table 2-2.
Table 2-2 Consistency of clay
(after Indonesian geotechnical guide-4, 2001)
Consistency cu (kPa)

Very soft < 12.5

Soft 12.5 - 24
Medium 25 - 49
Stiff 50 - 99
Very stiff 100 - 200
Hard > 200

In geotechnical engineering, soft terms specify to clay with shear strength ranging from 12.5
to 25 kPa, whereas the very soft clay is below 12.5 kPa. Table 2-3 gives some clues of an
indication of the shear strength when it was identified in the field (Indonesian Geotechnical
Guide-4, 2001).

Table 2-3 Field indicator of undrained shear strength for soft clay
Consistency Field indication

Soft It can be formed easily with fingers

Very soft If squeezed it will be out between fingers
Source: Indonesian Geotechnical Guide-4, 2001)

Indonesian Public Works (1999) defines the soft soil as the soil that can be penetrated with
the thumb a minimum 25 mm or has the undrained shear strength less than 40 kPa on the
basis of field vane shear test. Soft soil may compose of inorganic and organic soil. Inorganic
soft soil generally consists of clay or silt and it has organic content between 0 and 25% or ash
content ranging from 75 to 100%. Whereas organic soft soil consists of clay or silt and
organic content between 25 and 75%, or ash content ranging from 25 to 75%.
12
Peat is a soil-forming consisting of the primary of the remaining plants and the organic
content exceeding 75%. Basically, all types of soil are Resen old in geological terms that are
less than 10000 years old. The geological period is also commonly known as the Holocene.

2.1.5. Parameters for Soft Soil

Realistic parameter values for soft soil are important to be well prepared before taking the
samples from field work and testing them into the laboratory. This would assist in
anticipating if result gained from laboratory or field test varies from site to others. Varied
results that are measured from some parameters like water content, unit weight, Atterberg
limits, soil strength, coefficient of compressibility, consolidation index and permeability are
common when taking testing in the laboratory or field works.
Table 2-4 Range of realistic values for parameters of soft soil
Soil parameter Unit Clay Organic clay Fibrous peat

Water content, w % 20 – 150 100 – 500 100 – 4000

Unit weight,  kN/m3 14 – 17 12 – 15 10 – 12
Organic content % < 25 25 – 75 > 75
Undrained cohesion, cu kPa 5 – 50 5 – 50 10 – 50
Liquid Limit % 60 – 120 - -
Plasticity Index % 40 -80 - -
Friction angle, ’ [o] 21 – 27 25 – 35 30 – 40
Compression index, cc [-] 1– 2.5 1–5 1 – 20
Consolidation coeff.,cv m2/year 1 – 10 5 – 50 10 – 100
Swelling Index, c cm/s (0.03 -0.05) cc (0.04 – 0.06) cc 1–4
Permeability, ki cm/s 10-8 – 10-9 100 – 10-12 100 – 10-12
Source: Indonesian Geotechnical Guide-1, 2001

Undrained shear strength is an important parameter. This parameter for soft clay around the
surface in Indonesia ranges from 10 to 20 kPa. Undrained shear strength of 10 kPa is only
able to support an embankment about 2 m high.

2.1.6. Factors Affecting Behaviour of Clay

2.1.6.1. Organic Content
Organic content of clay or peat is generally derived from crop residues remaining on the
surface of the earth. Clay with low organic content values, for example below 10%, can be
found in clay estuarine and shallow marine sediments. Paul and Barras (1999) identified the
existence of this estuarine clay in the area Bothkenaar Scotland where the organic content
ranges from 2 to 4%. The influences of organic content of soil are: increased levels of
saturated water, high compressibility and low permeability.
Hobbs (1987) emphasized that the organic content can be calculated based on the weight but
the effect on soil properties also depend on its volume. He concluded that if the organic
content of about 27% by weight or about 55% of the volume, the material will give a great
influence on the properties of clay.

2.1.6.2. Sedimentation Rate

On sedimentary clay, rapid addition of layer thickness caused by high sedimentation rate will
cause pore water pressure. Sedimentation rate for the clay on a delta turned out to be a little
13
clay underconsolidated (Cox, 1970). Although Barry and Rachlan (2001) concluded there is
no existence of such phenomena.
Skempton (1970) presented the variation curves for a large number of clay deposits from the
present until the Pliocene period. Clay said as normally consolidated if the soil never
experienced the pressure is greater than the effective stress on it. Otherwise, over
consolidated clay. The process of sedimentation of clay in salt water occurs at high rate. For
Indonesia, sedimentation rate is 120 to 300 cm per 1000 years (Cox, 1970).

2.1.6.3. Chemical Weathering

Weathering is defined as a process leading to structural disintegration and decomposition of
geological materials due to direct influence of hydrosphere and atmosphere (Kenney, 1976).
Bjerrum (1967) explained the effects of weathering on shear strength and compressibility of
which shear strength and compressibility decrease due to weathering.

2.1.6.4. Freshwater Leaching

Leaching can be defined as a process caused by hydraulic or by a diffusion gradient that
removes material in solution (e.g. salt) from a passage in the soil profile. Rosenqvist (1953)
considered the process of leaching also occurred in the area of sediment under water (sub-
aquatic) from ground water when mixing with salt water steadily.
On soft clay deposits located in a flat delta plain of southeast Asia, leaching process is largely
caused by rains and floods from overflowing rivers. Therefore, salt concentration on the
surface will be low and tended to increase with depth. Salinity will also be lower with
increasing distance to the coastline.

2.1.6.5. Clay in Southeast Asia

According to Rahadian (1992) some properties of soft organic mineral in Southeast Asia
encompass water content varying from 60 through 150 %, having high plasticity, content of
clay ranging from 35 % and 60 % with ilite, caolynite and monmorilonite. In Bangkok, clay
mineral is dominantly ilite, meanwhile in Singapore is Caolynite and in peninsular Malaysia
is monmorilonite. Organic content varies from 2 and 5%, and the 22.5% organic content
founded in Peninsular Malaysia. Value of pH is between 3.1 and 8. Sensitivity of this soil
ranges from 1.5 to 18 (or medium quick clay). Compression index varies from 0.02 through
1.5. The overconsolidated ratio is less than 1.6. Effective internal friction angle varies from
20 to 25 degrees and leads to be reduced with increasing of the plasticity index.
In Indonesia, engineering properties along the coastline is generally close to the soft soil in
Southeast Asia such as natural water content, specific gravity and unit weight. Soft soil in
Sumatera and Java Island is mainly silty clay. Atterberg limit for this kind of soil has a liquid
limit between 40% and 160%. Generally, plasticity index is above or nearby A-line of
Cassagrande or in USCS so called CH OH soil. Compressibility of Indonesian marine clay is
high enough. Compression index varies from 0.5 to 2.

2.1.7. Factors Affecting Behaviour of Peat

2.1.7.1. Specific Gravity
Mineral soil generally has a specific gravity of 2.7 and soil containing organic matter ranges
from 1.4. Therefore, gravity influenced by organic content (Skempton and Petley, 1970).
Similarly to soils in Indonesia, specific gravity varied 2.7 to 2.9 while the peat varied
between 1.4 and 1.7 (Rahadian et al. 2001).
14
Besides specific gravity, the physical property of peat is bulk density. According to Adhi W.
(1984) peat soil is characterized by bulk density of 0.6 to 0.1 g/cm3. Moreover, further
experiment was undertaken using a ring sample method ranging from 0.14 g/cm3 to 0.22
g/cm3 (Muslihat, 2003).

2.1.7.2. Liquid Limit

Liquid limit testing requires adequate soil crushing. Therefore this test has very limited value
as a guide, especially peat properties of fibrous peat exist in Indonesia. According to data on
the area Berengbengkel for organic clay, it was confirmed that high organic content showed a
high liquid limit (Farrell et al., 1994). In terms of weight loss on heating, it was assumed to
be equal to the organic content of the missing.

2.1.7.3. Compressibility
Farrell et al. (1994) showed that the compression index Ireland peat associated with liquid
limit according to equation:
cc = k (WL – 10) (2-1)
where: k = 0.007 to 0.009
For fibrous peat the equation above cannot be applied. Consolidation tests on fibrous peat in
Barengbengkel show the value of cc up to 20. The vertical compression index is almost twice
of horizontal compression index values. Keep in mind that the value of high compression
index (cc) cannot be applied to the conventional calculation in small strain. Peat compression
index will decrease with increasing stress.

2.1.7.4. Permeability
Barry et al. (1992) performed the pumping test to support permeability in the forests of Riau
and revealed that the permeability between 10-2 and 10-4 m/s. They also compared it with
other research, as indicated in Table 2-5.
Table 2-5 Permeability values for peat soils
Discription of peat Permeability (m/s) Sources
-1
At the surface > 10 Hobbs (1986)
On the bottom 3 x 10-5 -idem-
Fen Acrotelm in Russia
near surface 3 x 10-5 -idem-
near bottom 6 x 10-7 -idem-
Peat soils in Irlandia 3 x 10-8 to 10-7 -idem-
Sphagnum peats
H8 to H10 6 x 10-8 -idem-
H3 10-5 -idem-
Sedge peat, H3 to H5 10-5 -idem-
Brushwood, H3 to H6 10-5 -idem-
Fibrous acidic Malaysia peat 2 x 10 to 6 x 10-8
-5
Toh et al. (1990)
After Barry et al., 1992 and Hobbs, 1986

2.1.7.5. Properties of Peat Soil in Indonesia

In Indonesia, soft soil is widespread in the big islands, namely Sumatera, Kalimantan, and
Papua. Property of peat soil in Papua island is rarely found caused by lack of infrastructures

15
development in this region. Some properties of peat soil both islands of Sumatera and
Kalimantan as reported Soepandji (1996, 1998) are depicted in Table 2-6.
Table 2-6 Properties of peat soils in Sumatera and Kalimantan islands
Kalimantan Sumatera
Properties
Pontianak Banjarmasin Duri Desa Tampan Musi
Ash Content (%) 1.2 3.29 21.96 25.2 50.7
Water Content (%) 632 198 235 338 235.4
Specific Gravity 1.42 1.47 1.6 1.55 1.82
Liquid Limit (%) 260 182 440 236 274
Plastic Limit (%) 196 148 377 309 194
Shrinkage Limit (%) - 28 - 59 -
pH value 4.8 6.5 3.9 3.6 3.3
Bulk Density - - 1.084 0.95 1.12
Compression Index, Cc - - 2.5-3.2 2.11 1.57
Recompression Index, Cr - - 0.07-0.13 0.107 0.05
Classification ,ASTM low ash, low ash, organic organic soil organic
D4427-92 (1997) acidic acidic soil soil
After Soepandji et al., 1996, 1998 cited Eka Priadi, 2008

2.2. Bearing Capacity of Soft Soil

2.2.1. Bearing Capacity at Ground Surface of Soft Soil
Before a complete failure of soft soil subgrade occurs, local over-stressing in shear takes
place and results in punching shear failure or local shear failure in the soil (Rodin, 1965). The
bearing capacity of subgrade under such condition is low and can be quantified by the
equation:
qu =  cu (2-2)
where cu is undrained shear strength of subgrade. On pavement engineering, the bearing
capacity is well known as California Bearing Ratio (CBR). For soft soil an empirical relation
between CBR value and undrained shear strength is:
cu = 30 CBR (kPa) (2-3)
When a general shear failure can be reached using reinforcement or a localized shear failure
of the subgrade can be prevented, the bearing capacity of the subgrade can be increased
maximum almost twice from previous state to:
qu = 2 cu (2-4)

2.2.2. Contact Area and Tire Pressure

Wheel load of vehicular traffic over the surface of the pavement for rubber tired vehicle
(single, dual, or tandem wheel) is equal to half of the axle load as depicted in Fig 2-2.
Vertical stress on the pavement surface is distributed downward to subgrade surface. Vertical
stress on the surface of subgrade is smaller than vertical stress on the surface of the pavement
depending on thickness (H) and stiffness of the base course.
Therefore design dynamic load, Q, for this situation is:
Q = Paxle / 2 (2-5)

16
 Paxle

p = Contact Pressure

H Pavement or Embankment

Subgrade

Fig. 2-2 Contact area and tire pressure (after TENAX, 2001)

For single or dual wheel, contact pressure (p) is equal to their tire inflation pressure. The
contact pressure is assumed to be a circular area and its radius is calculated by:

(2-6)

2.2.3. Determination of Fill Thickness

Boussinesq equations can be used here to calculate fill thickness. The criterion for calculating
the fill thickness is that the thickness of the fill must be large enough to allow the stress
transferred to subgrade surface is within the bearing capacity of subgrade as Equation (2-2)
and Equation (2-4).
For rubber tired traffic loading, the contact area of the wheel is assumed to be a circular form
and in this condition, the vertical stress at the subgrade surface, q, transferred from uniform
loading (contact pressure) can be provided:

(2-7)

Re-writing equation (2-7) we find:

(2-8)

The required thickness for an unreinforced section is:

(2-9)

It is similarly that the required thickness for a section reinforced using geosynthetics is:

(2-10)

Magnitudes of bearing capacity of the subgrade soft soil both unreinforced and reinforced
are different between methods each other as shown in Table 2-7.

17
 Table 2-7 Comparison of bearing capacity in different design methods
Unreinforced Reinforced
Method Material
section section
Giroud-Noiray (1981) 3.14 cu 5.14 cu Geotextile
Giroud-Han (2004) 3.14 cu 5.14 cu Geotextile
5.71 cu Geogrid
USACE (2003) 2.8 cu 3.6 cu Geotextile
5.8 cu Geogrid
Barenberg (1975) 3.0 cu 6.0 cu Geotextile
DuPont Typar (2010) 3.14 cu 5.14 cu Geotextile
Philips (1987) 2.8 cu 5.0 cu Geotextile
Rodin (1965) 3.1 cu 6.2 cu Geogrid
Roadex III (2008) 4.0 cu - -

2.2.4. Settlement Analysis

Estimation of settlement must involve both primary and secondary. For soft clay and organic
clay, Terzaghi consolidation theory can be applied.
Primary settlement for normally consolidated clay:

(2-11)

Primary settlement for over consolidated clay:

(2-12)

(2-13)

Secondary settlement:
(2-14)

2.3. Quasi-Static and Dynamic Loading

2.3.1. Definition of Terms

A cyclic loading on a foundation soil may be caused by moving vehicles (e.g. aircrafts, trains,
cars), by wind (e.g. wind power plant), or by waves (e.g. coastal structures). If the cycles are
applied with a low loading frequency fB, the inertia forces are not considered or negligible
and it is spoken of a quasi-static cyclic loading. Whereas if loading frequency is large so
inertia forces are relevant and loading is dynamic. Border between quasi-static and dynamic
loading also depends on an amplitude of the cycles. This amplitude dependence is ignored
and the border is said to lay at fB≈5 Hz (Wichtmann, 2005). At a certain strain amplitude test
up to fB≈30 Hz in the literature and Wichtmann’s test showed no influence of loading
frequency fB on the secant stiffness (elastic portion of the strain).

18
Sources of dynamic actions are described in Fig. 2-3.

Dynamic Actions

Natural actions Artificial actions

- seismic - roadway,runway, railway live loads
- avalanche - machine
- wind - explosion, blasting
- water - impact, collision

Fig. 2-3 Sources of dynamic actions (after EBGEO, 2010)

The actions in terms of DIN 1054 can be differentiated into dynamic, cyclic, and shock-like
actions. Dynamic actions refer to high frequency. Inertia forces are not negligible and it can
critically influence system behavior. Cyclic actions refer to low frequency actions where the
inertia forces can generally be ignored (frequencies ≤1 to 2 Hz). The shock-like actions refer
to actions acting over a short period only. The time may be in the range of milliseconds up to
several second. Their upper bound is not fixed and inertia forces may also act.
Additional distinguishing criteria include load-time history characteristic, effective spatial
direction, source and frequency of occurrence. Load time history is shown in Fig. 2-4.

Fig. 2-4 Typical of load time history (after EBGEO, 2010)

2.3.2. Stress Distribution on Unbounded Material

Generally, load from vehicle wheels at the surface of granular pavement will be distributed
on top of the subgrade layer taking assumption the circular form for the spreading angle of 45
degrees. Range of vertical stress magnitude on the formation level (on top of the subgrade
layer) was just between 20 kPa to 120 kPa. Meanwhile, horizontal stress due to a pavement
system having around 0.25 to 0.75 m total thickness on the formation level was around 30
kPa (Sasongko, H., 1996). Fig. 2-5 presents load distributions on granular material.

19
 Fig. 2-5 Load distribution on granular material (after Giroud et al., 1981)

2.3.3. Stress Distribution on Bounded Material

It is different pattern with granular materials that induced by vehicle wheels at the surface of
bounded material of pavement, e.g. slab of Portland cement concrete or bituminous layer,
will be distributed on top of subgrade layer taking assumption the circular form with a radius
of relative stiffness, l, as indicated in Equation (2-15).

(2-15)

Where: E = Young's modulus, k = the reaction modulus of subgrade, d = the thickness of

pavement (or concrete layer),  = Poisson's ratio.

2.3.4. Magnitude of Loading

Transportation infrastructure such as roadway, railway and runway is always subjected to
moving load. For vehicles crossing on streets, all kinds of vehicles refer to Equivalent Single
Axle Load (ESAL) 18 kips (or 8.12 tonnes) with tire pressure 85 psi (550 kPa). Meanwhile,
airplane passing at runway surface which loading coming from an airplane depends on
weight, tire pressure, wheel configuration (single, dual, tandem) of the airplane. Tire
pressures for airplanes vary from 140 psi to 200 psi.
In Germany, according to Ril 836 for rail infrastructure, the subgrade or improved subgrade
has to be able to support the load above this surface layer at least around 52 kPa as described
in Fig. 2-6. High speed trains from 100 to 300 km/h need the additional layer around 1 to 2
of the superstructure thickness (Muncke et al., 1999; Kempfert et al., 1999).

Fig. 2-6 Cross-section for rail track (after Ril 836)

Standard axle load is different for each country. In Greece, Beskou et al. (2011) reports that
for a locomotive (or engine) is around 210 kN and 150 kN for carriage. In Indonesia, railway
infrastructure is subjected to a maximum axle load of 180 kN (Kepmenhub, 2000).
20
2.4. Geosynthetics Reinforcement
2.4.1. Type of Geosynthetics
Geosynthetics for earth reinforcement purpose is divided into two groups: permeable matter
(geotextile, geogrid, geocomposite and geolinier) and impermeable matter (geomembrane
and intermembrane). For ground improvement or earth reinforcement, it must have some
properties such as shear strength, tensile strength, and punch strength (Suryolelono, 1997).
Types of geosynthetics used in earth reinforcement are shown in Table 2-8.
Table 2-8 Type of geosynthetics used in reinforced earth (after Suryolelono, 1997)
Material Type Form
Geotextile Woven and Unwoven
Geogrid Uniaxial and Biaxial
Geosynthetics
Geolinier Bar
Geostrip Strip/tape
Geocell Assembled cell

When designing geosynthetics as earth reinforcement, some reduction factors have to be

considered. Design tensile strength of geosynthetics follows this equation:
(2-16)

where: Fd is permissible long term tensile strength

Fk is short term tensile strength
A1 is reduction factor for creep
A2 is reduction factor for damage caused by transport, installation, compression
A3 is reduction factor for processing connection
A4 is reduction factor for environmental influences (weather, chemical,
micro-organism)
 is partial safety coefficient for the consideration of possibility
2.4.2. Geosynthetics Reinforcement Mechanism
Some small and full-scale studies have been performed to better understand how
geosinthetics interact with fill material to contrast their performance with unreinforced
conditions in a variety of civil engineering applications. Koerner (1998) discussed this
interaction and providing basic definition through geogrid. A geogrid is defined as a
geosynthetics material consisting of connected parallel sets of tensile ribs with apertures of
sufficient size to allow strike-through of the sorrounding soil, stone or other geotechnical
material (Koerner, 2005). Commercial geogrid products marketed and sold include extruded
punched-and-drawn geogrids, woven and coated geogrids, welded geogrids, and geogrids
composites. Structural biaxial geogrids can be used to reinforce earth fill over soft soil
ground and provide a stable subgrade under flexible and rigid pavements, unpaved roads,
railroad track beds, parking yards, work platforms and building foundations.
Subgrade soil beneath a paved or unpaved surface can fail under load in two ways: localized
shear failure and deeper-seated bearing capacity failure (ultimate failure). The subgrade
beneath an unreinforced fill will fail in the localized shear failure at around a half of stress
level than the ultimate bearing capacity of the subgrade. Geogrid reinforcement of granular
fills over soft soil ground can prevent the localized shear failure of the subgrade.There are
21
three primary mechanisms as being relevant to the interaction of geogrid and pavement
materials: lateral restraint, improved bearing capacity and tensioned membrane effect
(Perkins and Ismeik, 1997a).

2.4.2.1. Restraint and Confinement

There are two types of restraint as shown in Fig. 2-7. The first is related to the reverse
curvature of the geosinthetics outside the wheel path where a downward pressure is created.
This has the effect of a surcharge load, which levels out the deformation and enforces the
compression of the soil. Secondly, when aggregate particles attempt to move away under the
load, this provides tensile reinforcement to aggregate layer.This confinement of aggregate
increases its strength and modulus, which in turn decrease the compressive stress on subgrade
by spreading the load better underneath the wheel load.

1. Restraint + Confinement
2. Membrane mechanism
3. Local reinforcement

Fig. 2-7 Three stabilization mechanisms (after Perkins and Ismeik, 1997)

2.4.2.2. Membrane Mechanism

Membrane mechanism is effective when a geosynthetic is laid on deformable soil and vertical
loads are applied. In plane, tensile stress develops in the geosynthetics, relieving the soil
which is not capable of absorbing it. This is plane force induces a component of stress
perpendicular to the plane of the geosynthetic sheet, in the direction of the force. Therefore,
this is of great significance in temporary road construction, where it can reduce rutting the
tremendously. The use of the higher modulus of geosynthetics will reduce the rutting on the
surface of the pavement.

2.4.2.3. Local Reinforcement

Loads on individual stones can cause spot failure in the subgrade. A high initial modulus of
geosynthetics allows to distribute the load and reduce the vertical stress as well as provide
resistance to displacement. A high elongation avoids local puncturing of geosynthetics to
stretch around a penetrating aggregate/stone.

2.4.3. Geosynthetics Reinforcement Methods

More attention has been given to the important practical application of geogrid reinforcement
incorporated at the base of a layer granular fill placed on a soft clay subgrade. This kind of
construction is commonly used for low-cost unpaved roads such as temporary site access
roads, low embankments, car parks and the working platform. The purpose of the fill is to
provide a suitable operating surface on which concentrated loads may be carried without the
subgrade failing or deforming excessively. It is now common practice to use a layer of
polymer geosynthetics at the base of the fill layer in order to separate the fill from the soft
soil beneath and to improve its load-carrying capacity.
The behaviour of such a system is complex and a number of procedures for the design have
been proposed, notably by Barenberg et al. (1975), Giroud and Noiray (1981), Sellmeijer et
al. (1982), Giroud et al. (1984). These procedures which are based on simplified deformation
22
mechanisms and some empiricism have provided the basis for satisfactory design. Some
methods work better than others for certain site conditions, material and traffic volume.
However, none is principles of plasticity and fully addresses the significance of shear forces
acting between the base and the subgrade. In 2004 Giroud and J. Han introduced a method
and improve the previous method. Also US Army Corps of Engineer in 2003 launched a
method particularly for low volume.

2.4.3.1. Giroud-Noiray Method (1981)

Giroud-Noiray method (1981) is often used as a reference. The equation for this method is:
(2-17)

Where: h'o is thickness of unreinforced aggregate base (m); N is number of passes of axle
load; PA is axle load (Newton); s is rut depth (m) and cu is undrained cohesion of subgrade
soil (Pascal).

2.4.3.2. Giroud-Han Method 2004

Recognizing a need to advance geosynthetic design for unpaved surface, J.P. Giroud and Jie
Han, published a design method in the August 2004 edition of the American Society Civil
Engineers (ASCE). Their approach combines bearing capacity theory with empirical data
from full-scale test sections and monitored unpaved roads.
Some distinctions of the methods relative to conventional geosynthetics road design practice
including the following :
- considering of the effects of variation in base course strength
- considering of number and size of load cycles and desired roadway performance
- considering load distribution angle within the base course changes with time
- to recognize that geotextiles and geogrids perform differently in roads
- to recognize that not all geogrids perform the same
- to calibrate and validate of theoretical results with laboratory and full-scale test
Equation for Giroud-Han method is:

(2-18)

where: h = required base course thickness (m)

J = geogrids aperture stability modulus (m-N/degree)
N = number of axle passes
P = wheel load (kN)
r = radius of equivalent tire contact area (m)
CBRsg = CBR value of the subgrade soil
CBRbc = CBR value of the base course
s = allowable rutting depth (mm)
fs = factor for rutting depth (75 mm)
fc = factor for bearing capacity of subgrade soil (30 kPa)
Nc = 3.14 and J=0 for unreinforced base course
Nc = 5.14 and J=0 for geotextile-reinforced base course

23
 Nc = 5.71 and J=0.32 m-N/degree for Tensar BX 1100 reinforced base course
Nc = 5.71 and J=0.65 m-N/degree for Tensar BX 1200 reinforced base course

2.4.3.3. US Army Corps of Engineers Method (2003)

In February 2003, Corps of Engineers published a design method consideration use of
geogrids and geotextiles for paved and unpaved roads. Its approach for unpaved surface is
based on methodology originally developed by U.S. Forest Service. This method
distinguishes the performance of geotextiles and geogrids as reinforcement component in
subgrade improvement applications.
When using Corps of Engineers method, engineers have to select an appropriate bearing
capacity factor, Nc, for the geosynthetics type being considered. The Corps of Engineers
method recommended the following Nc values:
Nc = 2.8 without a geosynthetics
Nc = 3.6 with a geotextile for conservative designs
Nc = 5.8 with a geogrid
The first step in designing an effective reinforced pavement system is to determine the
properties of the subgrade including the grain-size distribution, Atterberg limits and in situ
shear strength or bearing capacity. The in situ shear strength can be measured directly using
vane shear devices or indirectly using a correlation from California Bearing Ratio (CBR)
using Dynamic Cone Penetrometer device. The design subgrade strength is defined as the
75th percentile strength of the top 18 in (45 cm) of the subgrade. Besides CBR value, cone
index value can be converted to shear strength using a chart as depicted in Fig. 2-8.

Fig. 2-8 Correlation between Cone index, CBR, and Shear strength (after Archer, 2008)

Next step is to determine the subgrade bearing capacity using equation below:
Subgrade bearing capacity = C. Nc (psi) (2-19)
Once the subgrade bearing capacity has been determined, the engineer or designer can refer
to one of the three relevant design charts (single wheel, dual wheel and tandem wheel).

2.4.3.4. DuPont Method (2008)

Degradation of pavement is caused by some factors including contamination of aggregate
base by fine-grained subgrade under dynamic loading so called pumping effect. This causes a
substantial reduction of the shear resistance of the aggregate. The thickness of ’clean’
aggregate is reduced down to unacceptable levels. Others are lack of subsurface drainage and
unpredicted traffic increase. Use of geotextile using DuPont guide will prevent aggregate
contamination. This guide uses the CBR value as a measure of soil strength. The correlation
24
factors between CBR, cu (undrained shear strength), MV (rigidity modulus), ME
(compressibility modulus) are given by a chart as shown in Fig. 2-9.

Soil

CBR
2
cu [kN/m ]
cu [psi]
2
ME [MN/m ]
2
MV [MN/m ]

Fig. 2-9 Correlation chart for estimating the subgrade value (after Barenberg, 1975)

An unpaved road normally consists of unbound aggregate base. Inclusion geotextile between
subgrade and aggregate base allows for better aggregate compaction, subgrade consolidation,
reinforcement of the structure and to increase the ultimate bearing capacity of subgrade
around (2+) cu.
The first procedure is to determine initial aggregate thickness (TO) according to load and
subgrade conditions and then consider service life and aggregate efficiency. Fig. 2-10 below
describes this stage easily. The left side of the chart is the subgrade CBR and axle load Pi to
determine TO or alternatively using right one.

(a) (b)

Fig. 2-10 Design chart DuPont method (a) Compacted crushed stone thickness for 1000 axle
loads, (b) Factor to determine curve Pi

The second step is to make an adjustment of TO for service life or corrected thickness of base
course (T) using equation (2-20).
T = C . TO = [ 0.27 log ( Ni . ESAL) + 0.19 ] . TO (2-20)
ESAL = ( Pi / Po )3.95 (2-21)

The service life is expressed as the total number of 80 kN axle load application. The actual
axle load (Pi) is first converted to an equivalent standard axle load (Po=80 kN).
The last step is to adjust the use of different kind of material. This difference is accounted for
by using the aggregate efficiency . Angular crushed aggregate is the best because it

25
interlocks well and provides a high bearing capacity. Depending on availability, other
materials or blends can be used and Table 2-8 indicates typical thickness efficiency factors of
various surfacing and base materials.
Table 2-8 Aggregate efficiency
Material Efficiency 
Paving Stone 2
Hot Mix / Dense-Macadam 2
Dense Surface Course 2
Soil-cement > 5 MPa compression 1.5
Soil bitumen 1.5
Hard crushed stone aggregate-‘standard’ 1.0
Medium crushed stone aggregate (CBR > 80%) 0.8
Hard round stone aggregate (CBR > 80%) 0.8
Medium round stone aggregate 0.5
Sandy gravel (CBR = 20-30%) 0.5
Crushed limestone 0.5
Loose gravel, compactable sand 0.4
Source: DuPont method, 2008

The effective aggregate thickness (Teff) can be expressed:

Teff =  Ti / i (2-22)

2.5. Piled Embankments

2.5.1. Conventional Piled Embankments
The development of infrastructure is moving rapidly to a life cost analysis approach, in which
the whole initial capital and on-going maintenance cost are balanced to ensure the best
performance of the structure. Since, the first used in Europe in the 1970's, piled embankments
have provided an excellent method of supporting embankments. The initial capital cost
associated with piled embankments has been higher than other techniques. Typical piled
embankment is illustrated in Figure 2-11.

(a) (b)
Fig. 2-11 Typical piled embankments (a) Conventional piled embankment (b) Piled
embankment with basally reinforcement

26
The conventional pile-supported system or piled embankments without reinforcement
requires large pile caps and very close piles. This relies on soil arching in the embankment
fill and can lead to punching failure. It is essential thing to transfer the large embankment
loads to the piles and to avoid surface deformation due to large differential settlement
between caps. Consideration required to ensure the edges piles can take the lateral loads
imposed by lateral thrust, whereas piled embankment with reinforcement the reinforcement
take places the lateral embankment thrust loads so raking piles not required anymore.
A survey of various projects (Han, 1999) found that in conventional piled embankments the
coverage ratio (the ratio area of pile caps to total foundation area) is 60-70%, whereas in
geosynthetics reinforced piled supported embankments is reduced to about 10-20%.

2.5.2. Geosynthetics-Reinforced Piled Embankment

Piled embankment with large pile caps and no reinforcement materials was firstly used in the
late 1960-70’s in Finland. This was to overcome the problem at the interface of a rigid bridge
abutment and the adjacent embankment on soft soil. This construction method proved very
successful and in the 1970’s was modified to include high strength geosynthetics material.
This material, firstly, can reduce the size of the pile caps, as vertical load would be
transferred by tension in the geosynthetics spanning between pile caps. Secondly, it removes
the need for expensive raking piles as the geosynthetics would resist outward lateral
movement of the embankment side slopes.
A piled embankment is a complex soil-structure interaction problem consisting of piles,
generally on a square grid, driven through the soft soil to a firm-bearing stratum. Due to the
higher stiffness of piles in relation to the surrounding soft soil, the vertical stress from the
embankment and surcharge load are concentrated on piles. Then, soil arching develops as a
result of different settlement between the stiff pile heads and the soft ground. Meanwhile
several methods currently exist for estimating the magnitude of arching (Kempfert et al.
(2004); Russel et al. (2003); Jenner et al. (1998); Hewlett & Randolph (1988); Guido (1987);
and Terzaghi (1943).
During more than three decades in many places in the world over 30 years have applied this
kind of system as shown in Table 2-10. Piles are generally installed on a square grid, whereas
triangular arrangements greatly complicate the analysis of the arching mechanism.
Table 2-10 Piled embankments geometry used over the past 35 years
Year of Country Piles spacing Pile caps Height
Reference
construction (m) (m) (m)
Reid & Buchanan (1984) 1973 Scotland 3.0 – 4.5 1.1 – 1.5 6
Holmberg (1978) 1978 Thailand 1.5 0.8 3.0 – 6.0
O'Riordan (1996) 1985 Ireland 3 1 8.6
Jones et al. (1998) 1989 UK 2.75 1.4 4
Rogbeck et al. (1998) 1996 Sweden 2.4 1.2 1.7
Habib et al. (2002) 2000 Netherlands 2.5 0.7 1.55
Van Eekelen et al (2007) 2004 1.27 0.3 1.15
Marchi et al. (2006) 2003 Italy 2 0.5 5.5
Almeida et al.(2008) 2004 Brazil 2.5 0.8 1.2

27
2.5.3. Soil Arching Concept
Arching is defined by Mc Nulty (1965) as ’’the ability of a material to transfer loads from
one location to another in response to a relative displacement between the locations. A
system of shear stress is the mechanism by which the loads are transferred ’’. Consider soil
on a rigid base, there is no tendency for differential movement and hence no soil arching as in
Figure 2-12 (a) the stress acting at point of a is the overburden stress (H), where  is the unit
weight of the soil and H is the height of the soil prism. When one of the local supports at the
point a is removed, the point a is in tension and a roof tension arch is formed. The true arch
collapses as the soil is not in equilibrium as depicted in Figure 2-12(b) meanwhile in the next
stage as in Figure 2-12(c) the soil settles in an inverted arch, the adjacent soil develops the
required shear strength and the soil reaches equilibrium state. The transfer of pressure from
the yielding portion to the stationary portion is so-called ’arching’.

(a) (b) (c)

Fig. 2-12 Soil arching concept (after Mc Nulty, 1965)

Deformation-based design is widely used in foundation engineering. It is currently used in the

design of such applications as foots, piles and drilled shafts. In order to accurately use
deformation-based design, there must be reliable ways to calculate the differential settlements
in the ground. The equal strain assumption is sometimes also a contributing factor to the
inaccurate calculation of differential settlements. This method implies that the settlements in
the soft clay and at top columns are equal, which implies that there is no stress concentration
in the bridging layer. This is inconsistent with the stress concentrations in the columns.
Therefore, it violates the vertical stress equilibrium. If equilibrium is violating, the
settlements are not being calculated accurately.

2.5.3.1. Rectangular Prism: Terzaghi (1943)

Terzaghi (1943) considered the equilibrium of a differential element and then integrates this
through the depth (z) of the moving soil mass. See Fig. 2-13 below where a rectangular soil
element, having a thickness (dH) and weight (dW) is depicted.

Fig. 2-13 Stress state of a differential element (after Terzaghi, 1943)

28
The vertical stress applied to its upper surface is:
v =  H + q (2-23)
where: v = the vertical stress (kN/m2)
  = the unit weight of the soil (kN/m3)
H = the thickness of soil above the point (m)
q = the surcharge acting at the surface of the soil (kN/m2)

The corresponding normal stress (horizontal stress) on the vertical surface of sliding ( h) is
given by:
h = K. v (2-24)
where: h = the horizontal stress (kN/m2)
 K = the earth pressure coefficient (dimensionless parameter)

The shear strength of the soil (assuming the soil to be cohesionless) is determined by:
   = h . tan  (2-25)
where:  = the shear strength (kN/m2)
  = the internal friction angle of the soil (degree)
When the element is in equilibrium, the summation of the vertical forces must be equal to
zero. Therefore, the vertical equilibrium can be expressed as:
dv/dz =  – K. v tan  / B (2-26)
where: 2B = the width of strip (m)
 z = the thickness of the soil overlying the element (m)
Using the boundary condition that v = 0 for z = 0, the partial differential equation can be
solved as follows (Terzaghi, 1943 and later McKelvey, 1994).
(2-27)

Krynine (1945) derived the earth pressure coefficient, K, as depicted below.

(2-28)

Handy (1985) proposed that the shape of the arched soil is catenary and suggested the use of
the coefficient Kw instead of K, by considering an arch of minor principal stress.
Kw = 1.06 (cos2  + Ka sin2 ) (2-29)
where: = 45 + /2
Russell et al. (2003) proposed that K could be conservatively taken as 0.5 and Potts &
Zdravkovic (2008) proposed that K = 1.0 gave good correspondence with the results of plane
strain finite element.

2.5.3.2. Rectangular Pyramid: Guido Method (1987)

This method is quite different from other methods of analysis for soil arching. The so-called
'Guido' design method is based on empirical evidence from model test carried out with
geogrid reinforced granular soil beneath a footing continued in a rigid box (see Fig. 2-14).
29
The results suggest that multiple layers of geogrid reinforcement increase the bearing
capacity, which could be interpreted as an improved angle of friction for the composite soil-
geogrid material (Slocombe&Bell, 1998). The 'load spread' angle in the reinforced soil
beneath the footing was proposed to be 45o (Bell et al., 1994).

0.5(s-a) a

45o

Fig. 2-14 Guido's experimental set-up (after Guido, 1987)

The stress on the subsoil (s) resulted from self weight of the unsupported soil mass is equal
to the volume of triangle/pyramid multiplied by the soil unit weight (). Then, it is divided by
the area over which the soil prism acts. For the two dimensional situation, the stress acting on
the subsoil is:
s =  (s-a) / 4 (2-30)
Meanwhile, for the three dimensional situation, the equation is modified to:
  s =  (s-a) / 3 √2 (2-31)
where: s-a = the width of strip (m)
From the equations above that the height of the embankment has no effect on the pressure
acting on the subsoil. Moreover, the friction angle of the fill material is not considered in this
case. Love&Milligan (2003) suggest that the Guido method may experience difficulties when
dealing with situations where support of the existing subsoil is very low. The Guido method
concentrates more on reinforcement rather than arching process.

2.5.3.3. Semicircular Arch: Hewlett & Randolph (1988)

Hewlett&Randolph (1988) derived theoretical solutions based on observations from
experimental tests of arching in a granular soil. Their analysis attempts to consider actual
arches in the soil as shown in Fig. 2-15.

a s

Fig. 2-15 Arching through a piled embankment (after Hewlett&Randolph, 1988)

30
The 'arches of sand' transmit the majority of embankment load onto the pile caps, with the
subsoil carrying load predominantly from the 'infill' material below the arches. The arches are
assumed to be semi circular (in 2D) uniformly without overlapping.
The analysis considers equilibrium of an element at the 'crown' as in Fig. 2-16. Here the
tangential (horizontal) direction is the direction of major principal stress and the radial
(vertical) direction is the minor principal stress. Yielding is in the 'passive' condition since the
horizontal stress is the major principal stress.

a 0.5s

(a) An element of sand (b) An element of sand

at the crown of arch above the pile cap

Fig. 2-16 Stress on an element of soil arching (after Hewlett&Randolph, 1988)

By using the boundary condition, considering vertical equilibrium, that the stress at the top of
arching layer is equal to the weight of material above acting on the outer radius of arch will
give a solution for the radial (vertical) stress acting immediately beneath the crown of the
arch ( i ). The vertical stress acting on the subsoil is then obtained by adding the stress due to
the infilling material beneath the arch, based on the maximum height of infill (s-a)/2.
s = i +  (s-a) / 2 (2-32)
When considering the three dimensional solution the equation above is modified to:
  s = i +  (s-a) / √2 (2-33)
The vertical stress (s) is considered uniform here. Though, Low et al. (1994) introduced this
parameter to allow a possible non uniform vertical stress on the soft ground.
In case at the pile cap, the tangential (vertical) stress is the major principal stress and the
radial (horizontal) stress is the minor principal stress. In conjunction with overall vertical
equilibrium of the embankment, a value of s is obtained in the limit when the ratio of the
major and minor principal stress is Kp. In fact, yielding occurs in an ’active’ condition, since
the vertical stress is the major principal stress.

2.5.3.4. Positive Projecting Subsurface Conduits: BS 8006-1 (2010)

The method used in the British Standard (BS 8006) for strengthened/reinforced soils and
other fills to design geosynthetics over piles was initially developed by Jones et al. (1990). A
2 dimensional geometry was assumed, which implies 'walls' in the soil rather than piles. This
method uses a modified form of Marston's equation for positive projecting subsurface
conduits to obtain the ratio of the vertical stress acting on top of the piles caps to the average
vertical stress at the base of the embankment ( s =  H ).

31
The equation proposed by Marston was derived from field tests at the Engineering
Experiment Station at Iowa State College in 1913. The equation is normally used to calculate
the reduced loads on buried pipes (see Fig. 2-17).

Fig. 2-17 Stress distribution of sub-surface conduit (after Marston et al., 1913)

For the two dimensional situation the Marston's equation is:

(2-34)

Meanwhile, the three dimensional analysis is modified to:

(2-35)

where: c = the vertical stress on the pile cap (kN/m2)

  = the unit weight of the soil (kN/m3)
a = the width of the pile cap (m)
H = the thickness of soil above the point (m)
H = the nominal vertical stress at the base of embankment (kN/m2)
Cc = the soil arching coefficient depending on H and a
Cc = 1.95 H / a – 0.18 for end-bearing piles
Cc = 1.5 H / a – 0.07 for floating piles

2.5.3.5. Multi Vaulted-Dome: German Standard (EBGEO, 2010)

In EBGEO 2010 for a three dimensional arching model proposed by Kempfert et al. (1997), it
appears similar to the Hewlett&Randolph (1988) approach. However, the average vertical
pressure acting on the soft subsoil was obtained by considering the equilibrium of dome
shaped arches of varying size in the 'infill' material beneath a hemisphere (see Fig. 2-18).

32
 Fig. 2-18 Theoretical arching model (after Zaeske et al., 2001)

EBGEO (2010) recommends the use of geosynthetics reinforcement but the arching effect
and the membrane tension are dissociated. After arching on embankments, it will develop a
redistribution of vertical stress at the surface between piles (zo,k) and vertical stress at
surface of piles ( zs,k) as depicted in Fig. 2-19.

Soft soil

Fig. 2-19 Vertical stress redistribution on arching (after EBGEO, 2010)

Vertical stress at the surface between piles after arching on embankments is divided into two
groups, namely the vertical stress for static loading (zo,G,k) and vertical stress for static and
variable-dynamic loading ( zo,G+Q,k ).

1+ 2 . 2− (2-36)

1+ 2 . 2− (2-37)

where: k = unit weight of embankment ( kN/m3)

pG,k = static loading (kN/m2)
pG+Q,k = static and variable loading (kN/m2)
hg = height of arching (m)
hg = s / 2 for h > s / 2
hg = h for h < s / 2
33
 h = height of embankment (m)
s = diagonal pile spacing (m)
d = diagonal pile spacing (m)
Kkrit = critical ratio of major stress = tan2 [45o + 'k / 2]
  = d. ( Kkrit – 1) / (2 . s )
 1 = (s – d)2/ 8
 2 = (s2 + 2 d. s – d2) / 2 s2

Furthermore, the vertical stress at the surface of piles after arching on embankments resulted
from surcharge divided into two groups, namely the vertical stress for static loading (zs,G,k)
and vertical stress for static and variable-dynamic loading ( zs,G+Q,k ).
  zs,G,k = [ ( k . h + pG,k ) – zo,G,k ] AE/As + zo,G,k (2-38)
zs,G+Q,k = [ ( k . h + pG+Q,k ) – zo,G+Q,k ] AE/As + zo,G+Q,k (2-39)
Where AE is the area of one cell pile embankment and As is a support surface for point and/or
linear bearing elements.
Vertical loads or forces on bearing element as:
  Fs,G,k = zs,G,k . As (2-40)
Fs,G+Q,k = zs,G+Q,k . As (2-41)
Generally, the resultant force on bearing element conservatively is calculated using:
  Fs,G,k = ( k . h + pG,k ). AE (2-42)
Fs,G+Q,k = ( k . h + pG+Q,k ). AE (2-43)

2.5.3.6. Arching Evolution

A novel approach for determining the vertical loading on underground structure in granular
soils has been developed (Iglesia et al. 1999). As can be seen in Fig. 2-20, it is proposed that
as the trap door is gradually lowered, the arch evolves from an initially curved shape (1) to a
triangular one (2), before ultimately collapsing with the appearance of a prismatic sliding
mass with two vertical shear planes (3).
Compared to analysis of a piled embankment the structure is analogous to subsoil. The
curved arch is similar to Hewlett&Rundolph’s semi-circular arch. The triangular arch is
similar to Guido's triangular arch and prismatic sliding mass is similar to Terzaghi’s sliding
block.

Fig. 2-20 Arching evolution (after Iglesia et al., 1999)

34
A methodology has been proposed by Iglesia et al. (1999) not only for determining the
vertical loading on the structure, but also for relating to this movement of the roof of an
underground structure. This is referred to as a ’Ground Reaction Curve’ (GRC) for overlying
soil. This is used dimensionless plotting of normalized loading (p*) vs. normalized
displacement (*).
(2-44)

(2-45)

where : p = the support pressure from roof above underground structure (kN/m2)
po = the nominal overburden total stress at the elevation of underground (kN/m2)
 = the settlement of roof (kN/m2)
B = the width of underground structure (m)

As shown in Fig. 2-21 that the GRC is divided into four parts namely the initial arching
phase, maximum arching, loading of the recovery stage, and the ultimate state.

Fig. 2-21 Generalized ground reaction curve (after Iglesia et al., 1999)

2.5.4. Definition of Terms

There are four terms commonly used to relate the total applied embankment pressure, the
pressure acting on top of the pile, and the pressure acting on the soil surface between piles
spacing.
1. the stress concentration ratio, n
2. the column stress ratio, CSR, or competency ratio, C,
3. the efficacy, E
4. the stress reduction ratio, SRR

2.5.4.1. Stress Concentration Ratio

The stress concentration ratio (n) is a parameter that is used to quantify load transfer. It is
defined as the ratio of the stress on the pile (caps) to the soil between the piles (caps). Ooi et
al., (1987) cited in Han (1999) indicated that the value of stress concentration ratio for
conventional pile embankments ranges between 1.0 to 8,0. This ratio increased with the
increase in ratio of the embankment height to the net spacing between pile caps. For structure
using Geosynthetics Reinforced Pile Supported (GRPS) embankments on vibro-concrete and
concrete piles, this value ranged from 8 to 25 which is much higher than conventional piled
embankments (Reid et al., 1993; Maddisson et al., 1996). This increase in n is due to the
inclusion of the geosynthetics layer and depends on the stiffness or rigidity of foundation.
35
 (2-46)

2.5.4.2. Column Stress Ratio

The ratio between the stress on the top of the columns, pile, and the average applied
embankment plus surcharge stress at the level of the top of the piles, , is referred as the
column stress ratio, CSR.
(2-47)
Low et al. (1994) use term of ’’competency ratio’’ using the symbol of C to refer the same
meaning for column stress ratio.

2.5.4.3. Efficacy
Hewlett &Randolph (1988); Low et al. (1994) defined efficacy, E, as the proportion of the
embankment weight carried by the piles rather than the subsoil. Efficacy increase (tending
towards 1.0) as the effect of arching increases.
(2-48)

where: as = area replacement ratio or coverage ratio which is defined as:

(2-49)

2.5.4.4. Stress Reduction Ratio

Soil arching causes vertical stress to be transferred from the soft soil to the columns (Han and
Gabr, 2002). The stress reduction ratio (SRR) is defined as the ratio of stress applied to the
foundation soil between the columns, soil, to the average stress applied by the embankment
plus surcharge, . A SRR of 1.0 implies no arching and the SRR reduces ultimately tending
to zero as the effects of arching increase.
(2-50)
Mainly current methods such as Terzaghi, Hewlett &Randolph, Guido, and Carlsson have a
main assumption that the geosynthetics reinforcement takes the entire load between the
columns and that none of the load is carried by the foundation soil (Russel and Pierpoint
1997). As stated previously, there is a large discrepancy in results from each of these
methods.
All parameters above have a close relationship to each other as expressed in equations below.
(2-51)

(2-52)

(2-53)

(2-54)

Arching in the embankment plays an important role in the behaviour of column-supported

embankments. For the condition of no soil arching, there is no reduction of pressure on the

36
foundation soil between the columns and hence the values of n, SRR, and CSR are equal to
one. If complete soil arching hypothetically develops, the entire applied embankment load
would be carried by the columns and no load would be carried by the soil between columns.
Relationship of the parameters is shown in Fig 2-22.

Fig. 2-22 Relationship between n and SRR (after Smith, 2006)

2.5.5. Load Transfer Mechanism

The inclusion of geosynthetics layer (see Fig. 2-23) is expected to reduce the differential
settlement between two piles. A single layer of reinforcement may be used at or near the base
of the embankment. Generally it is not placed directly on the pile caps due to the risk of
damage. A single geosynthetics layer acts as a tension membrane or catenary, whereas a
multi-layer system can interlock better with the surrounding soil and acts as a stiffened beam
or plate.

(a) Catenary (b) Beam

Fig. 2-23 Load transfer mechanism (after Collin, 2004)

The design of the load transfer platform (Collin, 2004) based on the use of multiple layers is a
refinement of a method referred to Guido method. The primary assumptions for the beam
theory are:
1. a minimum of three layers of reinforced is applied to create the platform
2. spacing between layers is 200-450 mm
3. platform thickness is greater than or equal to one half the clear spacing piles
4. soil arching is fully developed within the depth of the platform

When one or more layers of geosynthetics reinforcement are placed in the fill above the
columns, the stress that would otherwise be applied to the foundation soil between columns is
37
assumed to be carried by geosynthetics. Thus, the SRR value can be used to represent the
portion of embankment load carried by geosynthetics reinforcement.

2.5.5.1. BS 8006 Method

This method has adopted the empirical method developed for design of geosynthetic
reinforced piled embankment that was developed by Jones et al. (1990). The plan area of the
unit cell is s2 and the area not supported by columns is s2-a2, as shown in Fig. 2-24.

Fig. 2-24 Unit cell utilization (after Jones et al., 1990)

BS8006 defines a critical embankment height equal to 1.4(s-a). If the embankment height is
below the critical height, arching is not fully developed. The embankment and surcharge load
is converted to an equivalent vertical load, WT, between the piles which triangle load with WT
as maximum value at the centre of span, as determined by using equation:

(2-55)

If the embankment heights is greater than the critical height assumed that all load above the
critical height are transferred directly to columns as a result of arching in the embankment fill
(Kempton et al.1998).

(2-56)

where: pc = vertical stress on the column

 = unit weight of the embankment fill
H = embankment height
q = surcharge load
The stress reduction ratio, SRR, may be determined using the following equation:

(2-57)

2.5.5.2. Adapted Guido Method

Guido et al., (1987) presented that the inclusion of biaxial geogrids within granular soil below
spread footing could improve the bearing capacity of foundation soils. Angle of load spread
through the geogrid reinforcedless soil can conservatively be taken as 45 degrees. Bell et al.
(1994) applied this finding to perform with two layers of geosynthetic reinforced supported
on vibro-concrete columns. Russel and Pierpoint (1997) also assumed that geosynthetics
caries a pyramid of soil that not supported by column with angle of 45 degrees from the
horizontal level.
38
The SRR value for the Adapted Guido Method is based on the interior area of the unit cell,
2 2 2
i.e., (s-a) rather than s -a .

(2-58)

2.5.5.3. Adapted Nordic Method

This method adopted a triangular soil arching model of Carlsson's work (1987). This method
is similar to the Adapted Guido method but a wedge with an internal angle at the apex of the
wedge is equal to 30 degrees as shown in Fig. 2-25 .The Nordic method adopts a critical
height of 1.87 (s-a) and additional overburden above the top wedge will be directly
transferred to the columns.

HCritic

Fig. 2-25 Triangular soil arching of Nordic method (after Nordic guideline,2003)

The stress reduction ratio for the two dimensional approach, SRR2D, is given as:
(2-59)
Rogberck et al. (1998) provide a correction factor used to compute the geosynthetics tension
for an embankment supported by square piles caps. The net effect of this correction factor for
the three-dimensional stress reduction is the same as the two dimensional value and when a
surcharge load considered, the equation above is:

(2-60)

Equation 2-59 is applicable for embankment height greater than or equal to the critical height
and for lower embankments. The upper part of the triangle must be truncated to calculate the
stress reduction ratio.

2.5.5.4. Adapted Terzaghi Method

Russell and Pierpoint (1997) adapted Terzaghi's arching theory to develop a stress reduction
ratio. The settling mass is assumed to be cruciform in plan. The three-dimensional stress
reduction ratio is expressed as:

(2-61)

where: K = coefficient of lateral earth pressure, and  = internal friction angle of embankment
fill. The K value is assumed to be equal to one and the equation above is referred as Adapted
Terzaghi's method 1.
39
Several years later Russell et al. (2003) presented a modified version of the previous method.
They assume that the portion of the embankment fill that settles as a cruciform has a height of
n.H and the embankment fill above the settling cruciform is treated as surcharge. The stress
reduction ratio from latter method (Adapted Terzaghi Method 2) is expressed as:

(2-62)

where: q = embankment surcharge, K= coefficient of lateral earth pressure which Russel et

al.(2003) assumes to be equal to 0.5 , n = 1.0 for ultimate limit state (ULS) conditions and 0.8
for serviceability limit state (SLS) conditions.

2.5.5.5. Hewlett and Randolph Method

Hewlett and Randolph (1988) stated that the dome regions will fail either at the crown of the
arch or at the top of the column, but not elsewhere (see Fig. 2-26). Hewlett and Randolph
(1988) evaluated the load transfer mechanisms in terms of efficacy.

Fig. 2-26 Dome soil arching model (after Hewlett &Randolph,1988)

Efficacy at the arch crown is:

(2-63)
Efficacy at the top of pile is:
(2-64)

(2-65)

where: Kp = coefficient of passive earth pressure

The stress reduction ratio may be determined efficacy and its value at the arch crown is:

(2-66)

40
The stress reduction ratio at the top of the pile is:
(2-67)

2.5.5.6. German Method (EBGEO 2010)

In the German method, it is adopted from Kempfert et al. (2004). The arching theory from
Zaeske's work (2004) has been developed based upon plasticity theory, laboratory pilot scale
tests and numerical analysis.
The vertical stress on soft soil in between columns is expressed as:

1+ 2. 2 (2-68)

Kempfert et al. (2004) recommend using the equation to determine the load on the top of
geosynthetics reinforcement. Soil support from the subgrade beneath geosynthetics layer also
can be taken into account. The stress reduction ratio may be determined as follows:

(2-69)

2.5.6. Column Design

The embankment and/or any surcharge load is typically assumed to be carried by the
columns. The selection of column type is most often based on workability, load capacity and
cost. For purposes of determining the design vertical load in the column, it is convenient to
associate the tributary area of soil surrounding each column as illustrated in Fig. 2-27.

Fig. 2-27 Effective tributary area of column,

(a) Triangular spacing (b) Square spacing (c) Hexagonal spacing

The required design vertical load, Qr, in the column is determined according to the following
equation:

(2-70)

41
2.5.7. Tension in Geosynthetics Reinforcement due to Vertical Stress
Geosynthetics reinforcement is commonly used in soil by placing at the base of the
embankment. The tension will provide support between the pile caps (see Fig.2-28).
Geosynthetics reinforcement helps to transfer the weight of the embankment directly on to
the columns or piles (Lawson, 1992). At the edges of embankment it also prevents lateral
spreading (Hewlett&Rundolph, 1988). However, these two functions are normally considered
independently.

Fig. 2-28 Tension on geosynthetics reinforcement (after Satibi, 2009)

The effect of additional capacity to carry a vertical load could be added based on purely
tensile response. They proposed that assuming the geosynthetics was subjected to a uniform
vertical load and deforms as a parabola.

(2-71)

where : TRP = the tension in geosynthetic (kN/m)

WT = the uniform stress acting on the geosynthetics (kN/m2)
L = (s-a) the length of span (m)
 = the maximum sag (vertical deflection) of geosynthetic (m)
The average strain based on the total extension in the geosynthetics, , can be expressed in
terms of the maximum sag as follows:

(2-72)

As can be seen in equation (2-72) that the strain, , increases as the square of . The tension
in the geosynthetics is assumed as linear response against strain as below:
(2-73)
where: k = the stiffness of geosynthetics and in another expression:

(2-74)
This can be re-arranged to express how the load which can be carried theoretically increase
with the sag:

(2-75)

Giroud, J.P. (1995) presented a correlation between deflection and strain in geosynthetics, it
is similar to Eq. 2-71, and re-writing it as follows:
42
 (2-76)

2.5.7.1. BS8006 Method

The BS8006 method assumes that the geosynthetics reinforcement is placed above columns
installed in a square grid pattern. It is assumed that the embankment fill load is distributed
along the horizontal span of geosynthetics between pile caps and the entire load between pile
caps is carried by the geosynthetics and there is no support from the foundation soil. The
tensile load TRP per meter 'run' generated in the reinforcement resulting from the distributed
load WT is given by:

(2-77)

The tension in the reinforcement is calculated taking into consideration the maximum
allowable strain in the reinforcement. Six percent of strain is considered the upper limit for
transferring the load to the piles. The upper limit should be reduced for shallow embankments
to prevent differential movement on the surface of the embankment. To avoid long term
localized deformations at the surface of the embankment, the long term strain should be kept
to a minimum and a maximum creep strain of 2% is permitted for permanent construction.

2.5.7.2. EBGEO 2010 Method

The EBGEO 2010 method adopted Kempfert's work for evaluating the tension and strain in
geosynthetics reinforcement. For practical application, the strain in geosynthetics
reinforcement is determined using dimensionless design charts and then the tension is
determined by multiplying the strain by the geosynthetics stiffness. The force on the
geosynthetics, Fk, is equal to the vertical stress over subsoil, soil , multiply by a tributary
area associated with the strip of geosynthetics spanning between adjacent pile caps. The
resistance of foundation soil is included by using a modulus of subgrade reaction, k. The
modulus of subgrade reaction is the ratio of the pressure applied to the soil over a loaded area
divided by the resulting displacement.
The maximum stress in the geosynthetics reinforcement occurs over a width equal to bErs .
For a circular columns have to be converted to bErs of 0.886 dc, where dc = diameter of
column. Plan view of foundation and geosynthetics is as shown in Fig. 2-29.

Fig. 2-29 Plan view of foundation and geosynthetics (after Kempfert et al., 2004)

43
The maximum strain in the geosynthetic reinforcement is dependent upon the tensile stiffness
of geosynthetic, Jk, the modulus of subgrade reaction, k, the vertical load, Fk, clear spacing,
Lw, and width of column, bErs . The value of strain in geosynthetics, k, can be determined
using the design chart in Fig. 2-30.

Fig. 2-30 Chart to determine strain in geosynthetics (after Kempfert et al., 2004)

The tensile force in the reinforcement can be calculated using the following equation:
TRP = k . Jk (2-78)
When applied two layers of geosynthetics, the calculated tensile force is divided with respect
to the ratio of their tensile moduli (Kempfert et al. 2004).

2.5.8. Soil Resistance

Most current design methods for piled embankment ignore any support from the subsoil
between the pile caps. This is conservative and over estimate the magnitude of tension in the
reinforcement (Russell et al., 2001). Strain and deformation compatibility is required between
each layer (embankment fill, geosynthetics and subsoil). In practice, there will be some
support provided by soil below. This will considerably reduce the tension in the geosynthetics
reinforcement. Reid and Buchman (1983) found from their study that the resistance from the
subsoil is around 0.18  H. John (1987) found the soil resistance to be 0.15  H.

2.5.9. Tension in Geosynthetics Reinforcement due to Lateral Sliding

The reinforcement should resist the horizontal force due to lateral sliding. This tensile should
be generated at strain which compatible with allowable lateral pile movements. The need for
raking of the piles is eliminated. The reinforcement tensile load needed to resist the outward
thrust on the embankment in accordance with BS 8006 is:
Tds = 0.5 Ka (ffs. . H + 2fq. q) H (2-79)
where : Ka = the active earth pressure coefficient (Ka = tan2 (45+/2)
q = the uniformly distributed surcharge load (kN/m2)
ffs = the partial load factor for soil unit weight
fq = the partial load factor for applied external load

44
 = the unit weight of embankment fill (kN/m3)
 the height of embankment fill (m)
In EBGEO 2010, it is similar to the equation above, but any little bite different for coefficient
of active earth pressure (see Fig.2-31).

Fig. 2-31 Horizontal outward-thrust resisted by geosynthetics (after EBGEO, 2010)

Ek = [ 0.5. embk . (h-z).G + pk .Q ] .(h-z). Kagh (2-80)

Where Kagh is the active lateral earth pressure coefficient according to DIN 4085, which is
defined as :
Kagh = [ cos ' / (1+As) ]2 (2-81)
As = [ (sin( ' + s,k). sin  ' ) / cos s,k) ] 1/2 (2-82)
where: G = the partial load factor for soil unit weight
Q = the partial load factor for applied external load
 ' = the internal friction angle (degree)

2.5.10. Settlements Analysis

2.5.10.1. Different Settlement on the Surface of Embankment
Because of difference in stiffness between piles and soft soil, it possible occur a differential
settlement at the surface of the embankment. BS 8006 states that a plane of equal settlement
exists at an embankment height of 1.4 (s-a) from the top of the piles caps in which s is
spacing of pile caps and a is the width of the pile caps. Terzaghi (1943) carried out laboratory
tests and found that the plane of equal settlements exists at 1.5 – 2.5 times the width of the
void. As can be seen in Fig. 2-32 that s is the settlement of subsoil at the midpoint between
piles. ec is the settlement at the surface of the embankment at the centerline above the pile
cap. em is the settlement at the surface of the embankment at the midpoint between the pile
caps.

Fig. 2-32 Piled embankments showing arching soil, notations for geometry and settlement
(after Zhuang, 2009)
45
When a height of embankment is lower than the critical height, the differential settlement
potentially develops there and vice versa. Table 2-11 below shows the critical height of some
design methods.
Table 2-11 Critical height for different design methods
Method Critical Height, Hc
Terzaghi (1943) 2.5 (s-a)
Carlsson (1987) 1.87 (s-a)
Hewlett&Randolph (1988) 1.4 (s-a)
BS8006 (1995) 1.4 (s-a)
Horgan & Sarby (2000) 1.54 to 1.92 (s-a)
Kempfert et al. (2004) sg / 2

2.5.10.2. Settlement on the Bottom of Embankment

Soft clay and other compressible soils have a tendency to settle under heavy loading. There
are some techniques of soil improvement used to prevent these settlements. Piles, stone
columns, vibro-concrete columns and deep mixed columns are some of the commonly used
techniques.
Poulos (2005) reviewed the evolution of settlement analysis for pile groups and the transition
from research to practice over the past 30 years. Many methods exist for estimating the
settlement of piled foundation, ranging from empirical methods, through simple hand
calculation methods, to sophisticated numerical finite element and finite difference analysis.
There are a number of approaches commonly adopted for the estimation of the settlement of
pile groups:

 Methods which employ the concept of interaction factors and principle of

superposition;
 Methods which involve the modification of a single pile load-settlement curve, to take
account of group interaction effects;
 The settlement ratio method, in which the settlement of a single pile at the average
load level is multiplied by a groups settlement ratio (Rs), which reflects the effects of
group interaction;
 The equivalent raft method, in which the pile groups is represented by an equivalent
raft acting at some characteristic depth along the piles;
 The equivalent pier method, in which the pile group is represented by a pier
containing the piles and the soil between them. The pier is treated as a single pile of
equivalent stiffness in order to compute the average settlement of the group;
 Numerical methods such as the finite element method (FEM) and finite difference
method (FDM).

2.5.10.2.1. Interaction Factor Method

The interaction factor, denoted by ij, is defined as the additional displacement at the top of
pile i due to a loaded adjacent pile j, divided by the settlement of pile j under its own load,
and its application based on the theory of elasticity. Poulos (1994) had already successfully
employed the interaction factor for predicting the response characteristics of pile groups or
piled raft foundation. In this method as shown in Fig. 2-33, the settlement wi of a pile i within
a group of n piles is given as follows:

46
 wi =  Pav . S1 . ij (2-83)
where: Pav = average load on a pile within group; S1 = settlement of a single pile under unit
load; ij = interaction factor for pile i due to any other pile j within the group, corresponding
to the spacing sij between pile i and j.

Fig. 2-33 Superposition via the interaction factor method (after Poulos and Davis, 1980)

Simplified or closed-form expressions for the interaction factors have been developed, thus
enabling a simpler computer analysis. Mandolini and Viggiani (1997) have developed the
following expressions for the interaction factor, in one of the following forms:
 = A ( s / d )B (2-84)
or  = C + D ln ( s / d ) (2-85)
where: A,B,C,D = fitting parameters
s/d = ratio of pile spacing to pile cap diameter
The value of A ranged between 0.57 and 0.98, while the range of B was -0.60 to -1.20. The
value of C is equal to 1.0 and D = -0.26.
The original interaction factors published by Poulos (1968) were based on the assumption
that the soil was a homogenous elastic medium, having a constant modulus with depth. This
was clearly a great simplification of reality, and in subsequent years, some significant
improvements and extensions have been made including non-uniform soil modulus, influence
of bearing stratum and interaction between two dissimilar piles.
Influence of non-homogeneity (see Fig. 2-34) compares relationship between interaction
factor and s/d for three cases, namely a homogeneous soil layer with a constant modulus Es
with depth, a soil where the surface modulus (Es0) is 3 times that at the base (EsL), and a non-
homogeneous soil layer whose modulus varies linearly with depth from zero at the surface (a
Gibson soil) but which has the same average modulus as the uniform layer. Influence of
stiffness of the bearing stratum against interaction factor which is effect of the stiffness of
bearing stratum Es2 as a multiple of the overlying soil Es1.

47
 (b)

(a)

(c)

Fig. 2-34 Interaction factor, (a) Simplified case (b) Influence factor for non-homogeneity for
soil layer (c) Influence factor for bearing stratum (after Poulos,1968)

2.5.10.2.2. Equivalent Raft Method

The equivalent raft method has been used extensively for estimating pile group settlements. It
relies on the replacement of the pile group by raft foundation of some equivalent dimensions,
acting at some representative depth below the surface. There are many variants of this
method, but the one suggested by Tomlinson (1986) appears to be a convenient and useful
approach. In this approach, the representative depth varies from 2L/3 to L, depending on the
assessed founding conditions; the former applies to floating pile groups, while the latter value
is for end-bearing groups. The load is spread at an angle which varies from 1 in 4 for friction
piles, to zero for end bearing groups. Once the equivalent raft has been established, the
settlement can be computed from normal shallow foundation analysis.
Poulos (1993) has examined the applicability of the equivalent raft method to friction and
end-bearing pile group and concluded that this method gives a reasonably accurate prediction
for pile group containing more than about 16 piles with spacing of 3 times of pile diameter.
This method should be limited to cases in which the pile cross-sections exceed about 10% of
the plan area of the group (van Impe, 1991).

2.5.10.2.3. Equivalent Pier Method

The pile group is replaced by a pier of similar length to the piles in the group, and with an
equivalent diameter, de, estimated as follows (Poulos, 1993):
de  (1.13 to 1.27) . AG0.5 (2-86)
where: AG = plan area of pile group including the soil between the piles

48
The lower value is more relevant to predominantly bearing piles, while the larger one is more
applicable to predominantly friction or floating piles. Randolph (1994) has related the
accuracy of equivalent pier method to the aspect ratio R of the group, where:
R = ( n s / L )0.5 (2-87)
where: n = number of piles, s = pile spacing, L = pile length. The equivalent pier method
tends to overestimate stiffness for values of R less than about 3, for values of R of 1 or more
provide that the pile spacing is not greater than 5 diameters.

2.5.10.2.4. Piled Raft Method

Piled raft foundations are often used to improve the bearing capacity or to reduce the
differential deflection in the foundation structure. For preliminary estimates of piled raft
behavior, a convenient method of estimating the load-settlement behavior has been developed
by combining the approaches described by Poulos and Davis (1980) and Randolph (1994).
The method is described as the Poulos-Davis-Randolph (PDR) method. There are two main
steps: firstly, it is the estimation of the ultimate load capacity and secondly, it is the
estimation of the load-settlement behavior via a simple tri-linear relationship.

2.5.10.2.5. Japanese Method

Deep mixing method is also applied in Japan to improve soft clay and organic soils
(Takenaka, 1995; Bergado et al., 1999; Porbaha, 2000; CDIT, 2002). Because the deep mix
elements are stiffer than the surrounding soft soil, the stress concentration ratio develops
there. It is different with the dry deep mixing method which the stress concentration ratio
may be on the order of 4 to 6 (Kaiqiu, 2000), whereas the wet deep mixing method for soil-
cement is typically to be about 10 to 20 (CDIT, 2002).
The consolidation settlement of soil stabilized by deep mixed columns is determined the
following equation (CDM, 1985):
   dstab =  . d (2-88)
where:  dstab = consolidation settlement of stabilized ground, d = consolidation settlement
of unstabilized ground, and  = settlement reduction ratio.
The settlement reduction ratio is identically as stress reduction ratio, SRR. The stress
concentration ratio, n, and area replacement ratio, as, are used to determine the settlement
reduction ratio as follows:
 = SRR = soil/  = 1 / [ 1 +(n-1) as ] (2-89)
The consolidation settlement of the unstabilized ground, d, is calculated as follows:
d = mv .  . D (2-90)
where: mv = coefficient of compressibility of untreated soft clay, D = thickness of stabilized
clay layer (or sub-layer)

For shallow soft soils we have a possibility to build the piled embankment using end-bearing
piles which bottom of columns is located on the hard stratum, but for deep soft soil it is not so
effective and economic if we construct this kind of structure. Therefore, the floating piles are
the best choice to overcome this problem. Though, settlement on soft soil is too high because
of high compressibility and creep is also the main problem.

49
2.5.10.2.6. Public Works Research Center Method
Public Works Research Center method (2000) as cited Han (2003) has come up with a design
method for reinforced embankments on deep mixed column.
Settlement of the columns is given as:
Sc = c . L / Ec (2-91)
where: Sc = settlement of the columns
c = stress on the columns
L = length of the columns
Ec = modulus deformation of the columns

Modulus deformation has a correlation with unconfined compression strength (qu) of the
column depending on a kind of column.
Ec = 100 qu (2-92)
The settlement of the untreated soil is given by:
Ss = So. s / p (2-93)
where: Ss = settlement of untreated soil subjected to reduced pressure, s
So = settlement of untreated soil subjected to the actual load of embankment, p
s = reduced pressure on the untreated soil due to embankment
p = total applied pressure of the embankment

The differential settlement illustrated in Fig. 2-35 between the soil and the columns in the
absence of geosynthetics reinforcement is given by:
S = Ss - Sc (2-94)

Fig. 2-35 Settlement on the end-bearing piles (after PWRC method, 2000)

When there is an inclusion of geosynthetics layer, the differential settlement can be given into
account an influence factor due to the inclusion of the reinforcement.
  Sr = Ss / [1 + 2  (Ss / p)] (2-95)
where: Sr = differential settlement between the columns and the untreated soil
 = influence factor due to the presence of geosynthetics layer
This influence factor is related to the tensile stiffness of geosynthetics reinforcement. The
relation between the two factors can be seen in Fig.2-36.
50
 b = pile spacing

Fig. 2-36 Influence factor (), (after Han, 2003)

2.5.10.2.7. Scandinavian Method

Lime/cement and cement columns installed by the dry method of deep mixing have been used
extensively to support road and railroad embankment in Scandinavia country. In Sweden
almost 90% of all columns are installed to increase the stability and reduce the settlement of
embankments constructed on soft soils (Holm 1999; Broms 2003). It assumed that the same
strain occurs in the columns and the soil at every level (Broms, 1999) is shown in Fig. 2-37
below.

Fig. 2-37 Stress distributions for columns of deep mixed foundation (after Smith,2005)

The stress concentration in the stiffer column must greater than stress in the surrounding soil,
and then re-writing Eq. 2-46 for the stress concentration ratio, n. and coverage ratio, as, from
Eq. 2-49.
(2-96)

where: as = area replacement ratio or coverage ratio which is defined as:

(2-97)

The load applied on embankment is carried by both the column and the soft soil. The average
stress applied by embankment, , may be expressed as:
 = col as + soil (1-as) (2-98)

51
The vertical stress is carried by column and the surrounding soil as illustrated in Fig. 2-37
above as follows:
(2-99)

(2-100)

Depending on stiffness of column material, for lime/cement and soil-cement columns in

Sweden and Finland the stress concentration ratio of five is often used in design purpose
(Carlsten and Ekstrom, 1997). Based on large-scale field test performed on soil-cement
columns installed by the dry method Kaiqiu (2000) reported that stress concentration ratio
ranged from 4.7 to 5.7 under embankment loading.
The load deformation behavior of a dry mixed column is assumed to take place as shown in
Fig. 2-38. The load-deformation curve is linear up to long-term strength or creep strength and
the slope of the curve is equal to the modulus of elasticity of the column, Epile. Once, the
creep strength of the column is reached, additional loads are carried by the soil. The creep
strength is less than the column ultimate (or failure) strength, namely 65% to 90% of the
ultimate strength (Broms, 1999).

Fig. 2-38 Stress-strain relationship in dry mixed column (after Broms, 1999)

The column ultimate strength can be empirically determined using equation:

(2-101)
Where: u,col = the undrained shear strength of the columns and h' = the effective horizontal
pressure on the columns. For the Scandinavian applications, the undrained shear strength is
limited to 22 psi or 150 kPa (EuroSoilStab, 2002).
The compression of a volume of stabilized soil column is evaluated by considering two load
cases. The first one is when the creep strength of the columns is not reached and the second
case is when the creep strength of columns is reached. In the first case, the compression, S1,
and corresponding vertical strain, v, within the stabilized area are calculated based on the
following equations:
(2-102)

(2-103)

52
Where: S1 = the compression of the stabilized volume for case 1, di = stratum (or sub-layer)
thickness within the reinforced depth, Msoil = oedometer compression modulus of
surrounding soil. From both equations above that settlement of the stabilized volume
decreases with increasing area replacement ratio and with increasing the column stiffness.
In the second case when the creep limit of the columns is reached. The columns cannot take
load anymore. Therefore, subsequent loads carried by the unstabilized soil between columns
govern the settlement. Compression of the stabilized volume is calculated using the following
equation:

(2-104)

where: S2 = the compression of stabilized volume for case 2, = the creep strength
Broms (2003) suggests that the observed settlement can often be larger than the calculated
settlement. The differences between estimated and observed settlements generally increase
with increasing column lengths.
For partly penetrating columns or floating piles, the compression of a stratum of thickness d*
below the reinforced depth can be estimated for both cases. For case 1, the applied load is
assumed to be transferred directly down through the reinforced depth, di, and then it is
distributed through the underlying layer, d*, with angle of 1H:2V as shown in Fig. 2-39. The
compression of the underlying stratum, d*, and the reinforced depth, di, are the total
settlement.
For case 2 when the column creep strength is reached, the load be carried by columns is
transferred directly down through the reinforced depth (di) and then it is distributed through
the underlying layer (d*) at angle of 1H:2V. Meanwhile, the applied load that exceeds the
creep strength of the columns is applied to the ground surface and distributed through the
underlying soil at angle of 1H: 2V.

di

(a) Case 1 (b) Case 2

Fig. 2-39 Stress distributions beneath the stabilized columns for both cases
(after Broms, 1999)

For all methods mentioned above, settlement is a procress of water dissipation from soil body
in which it consists of lot of fine grained soil. Some methods can be applied to accelerate this
process such as vertical drains, pre-loading, vacuum technique, and electro-osmosis. Zhuang

53
et al. (2006) conducted electro-osmotic consolidation test for small-scale model. They
concluded that soil properties like c,  , max exhibit increasing after treatment of electro-
kinetic geosynthetics (EKG).

2.5.10.3. Relative Settlement Reduction (RSR)

A piled embankment using end-bearing piles is known as an effective method for soft soil
improvement. However, the end-bearing piles can be applied only when the soft soil
thickness is relatively shallow, e.g. up to 15 m. But, in many regions such as in Scandinavia
and some places in Southeast Asia, the soft soil thickness can be very deep up to 30 m or
even more. In such case, it is not possible economically to use the end-bearing piles.
Effectiveness of embankments on floating piles can be determined using its relative
settlement reduction (RSR). This terminology is defined as:
(2-105)

Where S0 is embankment settlement constructed on soft soil without the support of piles and
S is the settlement of an embankment supported by piles.

54
 CHAPTER 3

Characterization of Materials and Loading

There are four kinds of material to be characterized regarding the topic research: subsoil,
embankment, geosynthetics, and piles. It is important to know well some properties of them.
Because stress and strain relationship in material is induced by loading either static or
dynamic loading, and the characteristic of loading, including the magnitude and movement of
loading on transportation infrastructure is necessary to be understood.

3.1. Characteristic of Soft Soil in Indonesia

3.1.1 Physical Properties
A soil comprises of three basic constituents i.e. solids, liquids and gasses. Solids may be
either mineral or organic matter or both with their spaces filled with water and/or air. The soil
is saturated when all of pore spaces are filled with water. The purpose of the physical
properties testing is to obtain adequate information related to soil behavior. Some parameters
to describe the soil are as follows: water content Wn, unit weight , Atterberg limits (LL, PL,
PI), sieve analysis, degree of saturation Sr, organic content OC.

3.1.1.1. Physical Properties of Soft Soil in Java Island

Characteristic of soft soils at some sites in several provinces in Java Island as reported in
Development of Guidance for Roadway Construction over Expansive Soil (2003). Evaluation
for this soil is aimed to know characteristics and classification of soil as summarized in Table
3-1.
Table 3-1 Characteristics of soft soils in Java island
Provinces and Links of the observed roads
Central Java D.I.Y East Java W. Java
Parameters
Semarang - Dempet - Demak- Wirosari- Yogya- Ngawi - Surabaya Gresik- Jakarta-
Purwodadi Godong Kudus Cepu Wates Caruban -Gresik Lamongan Cikampek

1. Unit weigth (gr/cm3) 1.63-1.76 1.68-1.75 1.67-1.74 1.75 – 1.86 1.68 – 1.73 1.63-1.89 1.61-1.79 1.73 1.59 – 1.71
2. Clay content (%) 22-43 25-40 24-34 30 - 52 30 -52 30-61 23-45 44 25 - 58
3. Liquid limit (%) 80-110 72-108 83-94 53 -107 53 - 107 72-130 62-90 81 82 -104
4. Plastic Index (%) 52-79 40-74 44-58 24 - 57 24 - 97 39-79 28-45 48 46 - 62
5. Linear shrinkage (%) 18-22 12-26 - - 18 15-27 - - -
6. Water content (%) 32-48 37-53 29-49 24 - 40 27 -32 40-55 38-53 34 34 - 52
7. Passing # 200 (%) 83-98 82-98 95-98 73 - 96 76-96 92-98 89-94 97 92 - 94
8. Classification of soil CH, clay CH,clay CH CH,Silty clay CH, clay CH, silty clay CH,clay CH,clay MH,Silt
9. Mineral of clay Montmori Montmori Montmori Montmori Montmori Montmori Montmori Montmori Montmori
lonite lonite lonite lonite lonite lonite, 60 % lonite 45 % lonite lonite, 10 %
10.Colour Browny grey Browny grey Blackish Greyey black Greyey black Blackish grey Blackish Blackish grey Blackish grey
grey grey

Location : - Depth (m) 2.0 -.3.0 1.0 -3.0 2.0 -5.0 1.0 - 4.0 2.0 – 5.0 1.0 -.4.0 1.0 -3.0 1.0 -2.0 1.0 - 3.0
- Site (KM) 36-43 13.8 38.8 37.5 – 58.25 23.0 -27.0 5..0 – 19.0 12.8 – 14.45 16.50 25.50 – 69.60

Source: Final Report of Guidance for Roadway Construction over Expansive Soil, 2003

Soft soils are widespread in a lot of locations in Indonesia. In Java island, the soft soils
mostly consist of clay and/or silt, whereas in Sumatra and Kalimantan they are not only soft
clay but some regions covered by peat soils. Java Island is most dense in population and the

55
island has many infrastructures particularly roadways, railways and runways constructed over
the soft soil.
Chen (1975) uses a single index based on plasticity index to identify expansive soil. For
Plasticity Index (PI) values ranging from 20 to 55 are high for swelling level and very high
for PI values more than 55. Whilst Seed et al. (1962) use equation Ac = PI / (CF-10) to
identify activity level of soil. For value of Ac that more than 1.25 the soil has high level of
activity.

3.1.1.2. Physical Properties of Soft Soil in Pontianak

Most of Pontianak soil samples were blackish, blackish grey to dark in colour and had an
acidic smell. Organic content (OC) of about 10% was found (Priadi, 2008). These are
classified as fine grain soils because most of the samples having an average 84% pass
through sieve no. 200. Plasticity Index (PI) and the liquid limit (LL) varied widely from 5 to
35% and 20 to 70% respectively. According to ASTM standard D2487-00 and USCS based
on visually observation, organic content and distribution of grain size, these soils are
classified as organic soil. Organic clay deposits seem dominantly near the ground surface.
Sandy soil layer is founded 15 to 30 m in depth. The water content (Wn) varies widely
ranging from 25 to 200%, but decrease with greater depth. Generally, water content of
Pontianak soft organic soil was higher than its liquid limit. Furthermore, a cohesive soil with
water content higher than the liquid limit is defined as super soft soil clay.
Based on soil investigation for the project of Supadio airport runway expansion (2009) will
be a complement for data soil in Pontianak. The existing runway 2.250 m long and 30 m wide
would be extended 2.600 m long and 45 m wide, respectively. There were 10 sites of taken
samples using cone penetrometer test and the standard penetration test. As a comparison and
supplement that the testing of physical properties for Pontianak soil taken at a 1.5 m depth
has also been done in Geotechnical Laboratory TU Bergakademie Freiberg. Furthermore,
physical and mechanical properties of soils as described in Table 3-2.
Table 3-2 Subsoil properties at several zones in Pontianak
Project data of Sample of
Parameters Unit Priadi’s data Values in range
Supadio's runway Pontianak soil
of all data
Range of value Range of value Range of value

1. Soil classification - Low organic Low organic Low organic Low organic
2. Water content, Wn % 58.22-169.98 35 – 91 81.6-109.4 81.6-91
3. Organic content, OC % 10 - 11.88 10-11.9
4. Specific gravity , Gs - 2.2-2.6 2.36 – 2.70 2.611 2.4-2.6
5. Unit weight,  (kg/m3) kN/m3 12.3-16.49 13.9 – 18.5 13.94-14.31 13.9-14.3
6. Liquid Limit, LL % 20-70 17.10 - 62.46 84.52-99.25 62.5
7. Plastic Limit, PL % 17-35 14.62 – 38.44 39.6-50.22 35-39
8. Plasticity Index, PI % 5-35 2.48 – 29.67 34.3-59.6 29.7-35
9. Void ratio, e % 1.02-3.30 0.88 – 2.65 - 1-2.6
10. Cohesion (UD), c kN/m2 6-16.5 7.2-19.8 0.38-8.08 7.2-8
11. Friction angle (UD),  o 1.0-20.3 3.03-13.66 13.61-27.26 13.6-13.7
12. Compression index, Cc - 0.16-0.34 0.278-1.663 - 0.28-0.34
13. Coeff. of compressibility - - 1.4E-3 to 7E-3 - 1.4E-3 to 7E-3
14. Permeability, kv m/day 4.4E-6 to 7.8E-4 - - 4.4E-6 to 7.8E-4
15. Unconf. comp. strength kN/m2 - 8.4-65.5 - 8.4-65.5
16. Oedometer modulus kN/m2 550 - - 550
17. Young’s modulus kN/m2 650-1166 - - 650-1166

Based on Chen (1975) approach, for soil in this region with PI value average 19.5 can be
classified as high for swelling level.

56
3.1.2. Mechanical Properties
Mechanical properties are a necessary thing when investigating strength and deformation of
soil during loading. Shear strength of cohesive soil can be determined by using shear tests
and/or triaxial tests in which parameters cohesion, c, and internal friction angle, , can be
measured. Whilst in predicting the rate of settlement for the soil can be carried out by using a
consolidation test or oedometer testing in the other to obtain consolidation index, cc.

3.1.2.1. Direct Shear Strength

Normally, specimens of around 20 mm height and 40 cm2 cross-sectional circle surface are
utilized in the direct shear test, but some problems emerge when doing the kind of test for
very soft material. A little adjustment is an important thing in this case based on previous
experiences which the height of specimen needs to be higher than in the normal situation. For
consolidated drained (CD) shear test, height of the specimen around 30 mm was mounted to
overcome the high compressibility of soil in other to avoid friction between upper and lower
ribs in the shear box. The consolidated drained shear tests were conducted with normal
stresses 50, 75, 100, 150, 200 kN/m2 respectively and consolidation time was set up for 2
days before running shearing tests.

3.1.2.2. Compression
When soil undergoes a loading, because of their relatively low permeability, their
compression is controlled by the rate at which water is squeezed out of the pores. The slope e
against log ’ plotted in normally consolidated soil is referred to as the compression index,
cc. The load increment ratio was uniform where the loading was from 25 to 800 kN/m2. The
compression index, cc, varies widely with the increasing depth, however, the depth does not
influence of cc.
Priadi (2008) characterized the Pontianak soft organic soil compressibility behavior that the
top layer (around 10 m deep) is highly compressible ranging from 0.5 to 1.38 with an average
value of about 0.8, whereas at below this layer ranging from 0.2 to 0.5 with an average value
is about 0.3. Meanwhile, the recompression index, cs, ranges widely from 0.03 to 0.25. Some
of the 1-D Oedometer test results are shown in Fig. 3-1.

Fig. 3-1 Oedometer test of Pontianak soft organic soil (after Priadi, 2008)

57
The over consolidated ratio, OCR, is defined as the ratio between the pre-consolidation stress
and the effective in-situ stress. OCR is a state parameter that indicates the amount of over-
consolidation of the soil (Brinkgreve, 2001). This value notably reduces with a depth.
Pontianak soft organic soils are heavily over consolidated from the ground surface to about 5
m depth due to the wetting and drying cycles during deposition. The over consolidation ratio
ranges from 2 to 11 at this layer, whereas OCR values range from 1.3 to 2 are found at 5 to 20
m depth.

3.1.3. Bearing Capacity

Bearing capacity for subsoil particularly undrained condition can be expressed by several
parameters such as shear strength (su), unconfined compressive strength (qu) and California
Bearing Ratio (CBR). The three parameters mentioned above have a correlation with each
other. In case soft soil, subsoil is a cohesive material, generally like this, and also undrained
condition. Shear strength is approximately value of cohesion because the internal friction
angle is very small. Even, it is equal to zero when soil is the unconsolidated undrained
condition.
Unconfined compressive strength is nearly twice of shear strength (Braja M. Das, 1995).
When a bearing capacity is expressed as CBR, the empirical correlation is that shear strength
is 30 times of CBR (Barenberg, 1975).
Bearing capacity of subsoil can be improved by inserting piles. It can be the bored piles and
also the driven piles. After inserting piles with a certain depth, bearing capacity of soil in an
area that is replaced by the pile can be represented by the pile. This value depends on the
cone tip resistance linearly.
When pile is inserted in subsoil to support load over top of pile, estimation from Bustamante
and Gianeseli (1982) can be proposed to determine the bearing capacity (qu) from the cone tip
resistance (qc). In which qc is the required force to penetrate the cone divided by base area of
the cone. The equation is as follows:
qu = Kb. qc (3-1)
where Kb is an empirical bearing capacity factor that varies from 0.15 to 0.60 depending on
the soil type and pile installation procedure as depicted in Table 3-3.
Table 3-3 Empirical bearing capacity factor
Soil Type Bored piles Driven piles
Clay-silt 0.375 0.60
Sand-gravel 0.15 0.375
Chalk 0.20 0.40
After Bustamante & Gianeseli (1982)

3.2. Embankment Materials

3.2.1. Material Properties
Materials for embankment should have some properties such as durability, fire resistant and
compacted ability. Various materials can be used as embankment fills as shown in Table 3-4.
However, these materials have to fulfill some requirements if a good result wants to be
achieved.

58
 Table 3-4 Various materials for embankments
Unit weight
No. Materials
(kN/m3)
1 Sand 18-22
2 Cohesive soil 16-19
3 Corduroy 7.0
4 Rubber slag 4.0-6.0
5 Pumice 10.9
6 Dreg saws < 10.0
7 Peat bales < 10.0
8 Expanded polystyrene (EPS) 0.2-0.4
Source: Indonesian Geotechnical Guidance-4 (2001)

The Indonesian Geotechnical Guidance-4 (2001) gives the design parameters when using
material as embankment fill as depicted in Table 3-5.
Table 3-5 Design parameters for embankment material
Geographical Zone
Parameters Unit
A B
Unit weight, 
3
kN/m 18 20
Undrained shear strength, Su kN/m2 100 100
Cohesion, c ' kN/m2 10 5
Internal friction angle,  ' [O] 35 30
Source: Indonesian Geotechnical Guidance-4(2001)
A Java island (vulcanic rocks)
B Sumatra, Kalimantan, Sulawesi. Papua island (sedimentary and metamorphic
rocks)

3.2.2 Strength of Material

As illustrated in Table 3-4 above, the higher values for unit weight, , internal friction angle,
 and shear strength, su, will give a good result regarding with soil arching on embankment.
Usually, embankment fill is a cohesionless material or very small cohesion.
In pavement engineering, particularly flexible pavement system, there are some important
layers namely surface course, base course and/or subbase course and subgrade. Rosyidi et al.
(2004) reported the elastic modulus of pavement materials in situ measurement. These values
are in the range 3,000 to 7,200 MPa, 1.000 to 2,400 MPa, 480 to 600 MPa for surface course,
base course and subbase course respectively.
The concept of resilient modulus has been used to explain the nonlinear stress-strain
characteristics of soils. Generally, regarding with bearing capacity, the strength of pavement
materials can be expressed in CBR (California Bearing Capacity) or Modulus reslient, Mr.
Heukelom and Klomp (1962) reported a correlation between CBR value using dynamic
compaction and the in situ resilient modulus of soil. It has been extensively used for fine
grained soils with a soaked CBR of 10% or less.
(3-2a)
(3-2b)

59
Webb et al. (1986) have reported a number test of cohesionless soils in repeated load triaxial
test following the AASHTO procedure as equations follows:
(3-3a)
(3-3b)
Sasongko (1996) also reported a correlation between CBR and Mr using repeated load triaxial
test in laboratory as shown equations follows:
(3-4a)
(3-4b)
(3-4c)
NCHRP (2004) provided a correlation between CBR and Mr for a number of cohesionless
soils in repeated load triaxial test following the AASHTO procedure. A typical equation for
medium clay sand is shown equation (3-5).
(3-5a)
(3-5b)
Several constitutive models have been proposed by many researchers for modeling resilient
moduli of soils and aggregates. Dunlap (1963) suggested the following relationship for
presenting resilient modulus:

(3-6)

where k1, k2 = regression coefficients obtained from regression analysis

Pa = reference pressure (atmospheric pressure)
σ3 = confining stress
Seed et al. (1967) suggested a relation where resilient modulus is a function of bulk stress (θ),
also known as the K–θ model. This model, generally adopted for granular soils, uses θ as the
main attribute in the model.

(3-7)

where θ = bulk stress (σ1 + σ2 + σ3)

The main drawback of this model is that it does not account for shear stresses and shear
strains developed during loading. Moossazadeh and Witczak (1981) proposed a relation
known as the deviatoric stress model recommended for cohesive soil, known as K-σd model.

(3-8)

where σd = deviator stress (σ1 - σ3)

May and Witczak (1981) proposed a model to describe the non-linear behavior revealed in
the repeated load triaxial test. This model considers the effects of shear stress, confining
stress and deviatoric stress with the model formulated in terms of bulk and deviatoric stress.

(3-9)

Uzan (1992) introduced the octahedral shear stress in place of deviator stress in equation (3-
9), which provided a better explanation for the stress state of the material, in which the
60
normal and shear stress change during loading. The proposed model is known as the k1- k3
model. The universality of this model stems from its ability to conceptually represent all
types of soils from pure cohesive soils to non-cohesive soils.

(3-10)

where
The coefficients k1, k2, and k3 are constants, depending on the state and quality of unbound
granular materials. Since coefficient k1 is proportional to Young’s modulus, it should always
be positive as Mr can never be negative. The coefficient k2 should be positive, because
increasing the volumetric stress produces stiffening or hardening of the material, yielding
higher modulus. The coefficient k3 should be negative because an increase in the shear stress
softens the material, thereby yielding lower modulus. If nonlinear property coefficients k2 and
k3 are set to zero, then the model can be simplified as linear elastic. If k3 is zero, the behavior
could be non-linear hardening and if k2 is zero, the behavior is non-linear softening.

3.2.3. Maximum Height of Embankment

Stability for maximum or critical height of embankment for road according Roadex III (2008)
without ground treatment can be calculated using equation (3-11).
(3-11)

where  is unit weight of embankment fill and cu is the undrained shear strength of subsoil
beneath embankments.

3.2.4. Dynamic Properties

Shear modulus, damping ratio and shear wave velocity profiles are an important input
parameter in site response analysis. Jafari, M.K. et al. (2002), based on field geoseismic
investigation data for fine grained soil in Tehran, present new correlation for Shear-wave (Vs)
and Number of blows (N) from Standard Penetration Test (SPT). Some researchers also give
some equations for the correlation between Vs and N.
Shear wave velocity can be obtained directly from field investigation or laboratory testing of
soil samples of the studied area, but it is not always economical solution. However, when the
direct measurement of shear wave velocity for soil layers is not available, the existing or
developed correlations between N values of SPT and the shear wave velocity could be used.
Some researchers have carried out some correlation between the shear wave velocity and
number of blows from Standard Penetration Test for some kinds of material for both cohesion
and cohesionless material as shown in Table 3-6.

The behaviour of soil under cyclic loading is non-linear and dependent on some factors
including soil type, confining pressure, number of loading cycles and amplitude of loading.
Non linear hysteretic soil behaviour is commonly characterized by a viscous damping and
equivalent shear modulus (Seed and Idriss, 1970; Hardin and Drenevich, 1972). Definition of
damping is a measure of energy dissipation. It increases with increasing magnitude of cyclic
shear strain, whereas shear modulus decrease with increasing magnitude of cyclic shear
strain. It is also known that dynamic properties of soil are influenced by the plasticity index,
void ratio, relative density and number of cycles (Cabalar and Cevik, 2008).

61
 Table 3-6 Correlation between Vs and N
Shear wave velocity, Vs
Author(s) Soil type
(meter/second)
Kanai, et al. (1966) All Vs = 19 N 0.6
Shibata (1970) Sand Vs = 32 N 0.5
Ohba & Toriuma (1970) Alluvial Vs = 85 N 0.31
Ohta, et al. (1972) Sand Vs = 87 N 0.36
Ohsaki & Iwasaki (1973) All Vs = 82 N 0.39
Cohesionless Vs = 59 N 0.47
Imai & Yoshimura (1975) All Vs = 92 N 0.329
Imai, et al. (1975) All Vs = 90 N 0.341
Imai (1977) All Vs = 91 N 0.337
Imai & Tonouchi (1982) All Vs = 97 N 0.314
Imai & Yoshimura (1990) All Vs = 76 N 0.33
Ohta & Goto (1978) All Vs = 85 N 0.348
Sands Vs = 88 N 0.34
Gravels Vs = 94 N 0.34
JRA (1980) Clays Vs = 100 N 0.333
Sand Vs = 80 N 0.333
Seed & Idriss (1981) All Vs = 61 N 0.5
Seed, et al. (1983) Sands Vs = 56 N 0.5
Sykora & Stokoe (1983) Granular Vs = 100 N 0.29
Okamota, et al. (1989) Dilluvial sands Vs = 125 N 0.3
Lee (1990) Sands Vs = 57 N 0.49
Clay Vs = 114 N 0.31
Silts Vs = 106 N 0.32
Yokota, et al. (1991) All Vs = 121 N 0.27
Jafary, et al. (1997) All Vs = 22 N 0.85
Clayey soils Vs = 27 N 0.73
Silty soils Vs = 22 N 0.77
Clayey & Silty Vs = 19 N 0.85
Source: Jafary et al., (2002)

Once shear wave velocity is determined, and then shear modulus of material, G, is obtained
using equation G= .Vs 2, which  is the density of soil.
Shear modulus of air dry clean sands at a certain level of shear strain, , can be represented
approximately by the following empirical equation irrespectively of kinds of sands.

( = 10-6) (3-12)

( = 10-5) (3-13)

( = 10-4) (3-14)

62
Where G is shear modulus in kg/cm2, p is mean principle stress in kg/cm2 and e is the void
ratio. Eq. 3-5 is identical to the empirical equation for round Ottawa-sand proposed by
Hardin, et al. (1972).
All tests demonstrated the well known dependence of Gmax on effective confining pressure 'o
and density expressed in terms of the void ratio e. In other to an analytical expression for the
Gmax = Gmax ('o) relationship, the general equation suggested by Hardin (1972) is adopted.
(3-15)
Where: S is a stiffness coefficient. The experimental results are closely approximated by
setting S= 420 and n= 0.6. It should be noticed that the value of the exponent n is higher than
widely used for cohesive and cohesiveless soils which ranges between 0.4 and 0.5. The
variation of shear modulus with shear strain amplitude as expressed in the following
equation.
(3-16)

Damping ratio D was found to be essentially independent on confining pressure and density.
An average value Dmin=2% was determined from all tests. The increase of damping with
shear strain amplitude  may be approximately expressed by:
(3-17)

with A = 6.2 and  = 6.5.10-4.

The soil starts to exhibit hysteretic and non-linear behaviour in the shear strain range between
10-6 and 10-3, where the secant stiffness decreases with the increasing of the strain level.
Various authors have suggested several equations to connect the damping ratio and the shear
modulus when both are functions of the shear strain. Hardin and Drnevich (1972) derived the
simple relationship equation.

(3-18)

where: G is the secant modulus and Go is the initial shear modulus.

Park and Stewart (2001) have proposed equations for sandy soils and clayey soil respectively.

For sandy soils (3-19)

For clayey soils (3-20)

3.3. Geosynthetics
3.3.1. Material Properties
The geosynthetics terminology may be based on the subdivision by PrEN ISO 10318.
According to this standard “Geosynthetics” is a generic term describing a product at least one
of whose components is made from a synthetic or natural polymer, in the form of a sheet, a
strip or a three dimensional structure, used in contact with soil and/or other materials in
geotechnical and civil engineering applications. As depicted in Fig. 3-2 that geosynthetics can
be differentiated into permeable and impermeable products.

63
 Geosynthetics

Permeable Essentially impermeable

Geotextiles Geotextile-related Geosynthetic clay Geomembranes

to products liners
-woven -polymeric
-nonwoven -geogrids -elastomeric
-knitted -geonets -thermoplastic
-geocells -bituminous
-geostrips
-geomats
-geospacers

Geocomposites

Fig. 3-2 Geosynthetics subdivision (after PrEN ISO 10318)

The use of a geosynthetics in pavement system reinforcement is to aid in support of the traffic
load. Traffic loads may be vehicular loads experienced over the life of the pavement. Base
(or subbase) reinforcement is a treatment using of a geosynthetics as a tensile element at the
bottom of base (or subbase) or within a base course and is designed to address the pavement
distress mode of pavement surface deformation or rutting and asphalt fatigue cracking.Whilst
subgrade restraint is the use of geosynthetics at the subgrade/subbase or subgrade/base
interface to increase the support of construction over a weak or low strength subgrade
(Barenberg, 1980; Steward et al., 1977; Giroud and Noiray, 1982; Holz et al., 1987).
The following benefits of using geosynthetics in roadways are identified (TenCate Mirafi,
2010):
1. Reducing the intensity of stress on subgrade (function: separation).
2. Preventing subgrade fines from pumping into the base (function: filtration).
3. Preventing contamination of the base materials allowing more open graded, free-
draining aggregates to be considered in the design (function : filtration).
4. Reducing the depth of excavation required for removal of unsuitable subgrade
materials (function: separation and reinforcement).
5. Reducing the thickness of aggregate required to stabilize the subgrade (function:
separation and reinforcement).
6. Minimizing disturbance of the subgrade during construction (function separation
and reinforcement).
7. Assisting the increase in subgrade strength over time (function: filtration).
8. Minimizing the differential settlement of roadway, which helps maintain
pavement integrity and uniformity (function: reinforcement).
9. Minimizing maintenance and extending the life of the pavement (function: all).

Others important findings from laboratory and/or field studies include the following:

64
 1. An optimum benefits when the geosynthetics was placed at the bottom of a 200-
300 mm thick base layer.
2. For thicker base sections, the most beneficial reinforcement location appeared to
be in the middle of the base, where geogrids were found to perform best.
3. For thin bases (less than 200 mm), lack of separation was noted as a potential
problem for geogrids. Geogrid-geotextile composites tend to perform better for
thin bases, especially where subgrade strengths were below a CBR of 3%.
4. Reinforcement benefits were observed with subgrade strengths up to a CBR of
8%.

Benefits using geosynthetics as reinforcement can be defined by TBR (traffic benefit ratio)
and BCR (Base course reduction) as shown in Table 3-7. The TBR is defined as the ratio of
number of cycles necessary to reach the same rut depth for a test section containing
reinforcement to unreinforced section with the same section thickness and subgrade
properties. Furthermore, BCR is expressed as a percentage savings of the unreinforced base
course thickness.
Table 3-7 TBR and BCR resulted from laboratory and field test
Materials TBR BCR
Geotextiles:
Range 1 – 220 22 – 33 %
Typical value 1.5 – 10
Geogrids:
Range 0.8 – 670 30 – 50 %
Typical value 1.5 – 70
After TenCate Mirafi, 2010

Besides the ratio coefficients (TBR, BCR), the following properties are considered to
influence performance : tensile strength at 1%, 2% and 5% strain, coefficients of pullout and
direct shear, aperture size (grids) and percent open area (geotextiles) and stiffness properties
including the flexural rigidity and aperture stability. For subgrade restraint applications, the
properties of tensile strength at 2% and 5% strain are primarily related to geosynthetics
performance.

3.3.2. Position of Geosynthetisc in Pavement Design Practice

There are three general applications for the use of geosynthetisc reinforcement in pavements.
Therefore, the appropriate application with the ultimate objective of maximizing performance
as the following guidance for these three distinct applications:

1. Weak Subgrade (CBR <3), For Thin (≤250 mm) Base Sections and Thick (>250 mm)
Base Sections.

When a weak subgrade exists, woven geotextile should be placed at the surface interface. In
addition, if the required base course is greater than 250 mm (10 in), a second layer of
reinforcement, biaxial geogrid, should be placed in the middle of the base course section.

2. Firm Subgrade (CBR >3), For Thin (≤250 mm) Base Sections and Thick (>250 mm)
Base Sections.
65
For a firm subgrade and relatively thin base course section is designed biaxial geogrid at the
subgrade interface. While geogrid can be placed in the middle of the base course when firm
subgrade and relatively thick base course section is designed.

(a) Weak Subgrade, Base course ≤250 mm (b) Weak Subgrade, Base course >250 mm

Surface course Surface course

Base course Base course

Geogrid
Geotextile Geotextile
/////// Subgrade ////////// /////// Subgrade //////////

(c) Firm Subgrade, Base course ≤250 mm (d) Firm Subgrade, Base course >250 mm

Surface course Surface course

Base course Base course

Geogrid
Geogrid
/////// Subgrade ////////// /////// Subgrade //////////

Fig. 3-3 Position of geosynthetics in pavement design practice (after Mirafi, 2010)

3.3.3. Tensile Strength of Geosynthetics

Tensile strength of geosynthetics depends on raw material of geosynthetic such as aramid or
polyamide (PA), polyethylene of high density (PE-HD), polyester (PET), polypropylene (PP)
and polyvinyl alcohol (PVA). Range of tensile strength from various raw materials of
geosynthetics is shown in Fig. 3-4.

(a) (b)
Fig. 3-4 Typical strain vs. Force behaviour of reinforcement (a) Exxon, 1989
(b) Carlson, 1987
66
Typical short term strength of geosynthetics is described in Table 3-8 for various raw
materials (EBGEO, 2010; Althoff, 2011).
Table 3-8 Tensile strengths of geosynthetics (after EBGEO, 2010)
Typical short term strengths Typical elongation
Raw material Product types [kN/m] at failure [%]
from to max. from to
AR Woven geogrids 40 1200 2200 2 4
(aramides) Woven geotextile 100 1400 300 2 4
PE Woven geogrids 20 150 300 15 20
(polyethylene) Extruded geogrids 40 150 200 10 15
Woven geotextile 30 200 400 15 20
PET Woven geogrids 20 800 1200 8 15
(polyester) Bonded geogrids 20 400 500 6 10
Woven geotextile 100 1000 1600 8 15
PP Woven geogrids 20 200 500 8 15
(polypropylene) Bonded geogrids 20 200 400 8 15
Extruded geogrids 20 50 - 8 20
Woven geotextile 20 200 600 8 20
PVA Woven geogrids 30 1000 1600 4 5
(polyvinyl alcohol) Woven geotextile 30 900 1800 4 5

When choosing geosynthetics for reinforcement, there are two factors that must be
considered well namely internal and external factor. Internal factors such as tensile strength,
creep properties, whereas external factors such as kind of embankment fill, endurance against
environment (ultra violet, acidic or alkaline matter, micro-organism. In other to cover all
conditions, strength of geosynthetics has to be adjusted using some partial factors. According
to Nordic guideline (2003) some conversion factors for design purpose are listed in Table 3-9
through Table 3-11.
Table 3-9 Conversion factors of geosynyhetic reinforcements
(after Nordic guideline, 2003)
Conversion parameters Conversion factor
Creep factor 
Installation damage 
Biological and chemical degradation 

Table 3-10 Conversion factors  for long-term properties

(after Nordic guideline, 2003)
Conversion parameters Conversion factor,  Material factor, fm
Steel 0.8 1.25
Polyester (PET) 0.4 2.5
Polypropylene (PP) 0.2 5
Polyamide (PA) 0.35 2.8
Polyethylene (PE) 0.2 5

67
 Table 3-11 Conversion factors  for damage during installation
(after Nordic guideline, 2003)
Conversion parameters Conversion factor,  Material factor, fd
Clay/silt 0.91 1.1
Sand 0.83 1.2
Gravel (Natural) 0.77 1.3
Gravel (Broken) 0.72 1.4
Chrused Rockfill 0.67 1.5

According to Swedish Road Administration publication 1992:10, the material factors for
biological and chemical degradation, fenv, may be assumed 1.1 as long as the pH value ranges
between 4 and 9, which gives a conversion factor of  = 0.91.
Allowable tensile strength of geosynthetics for reinforcement design is defined as ultimate
tensile strength divided by the reduction factor (or partial factor).
(3-21)

where: a = allowable strength

u = ultimate strength
fd = partial factor for mechanical damage
fenv= partial factor for environment
fm = partial factor for extrapolation of tensile strength
fc = partial factor for construction safe

3.4. Characteristic of Piles

There are three classifications of columnar foundation include (Han and Wayne, 2000)
namely:
 flexible column (such as stone columns and lime columns)
 semi-rigid columns (such as lime-cement and soil-cement columns)
 rigid piles (such as concrete pile, timber piles, and vibro-concrete piles)

3.4.1. Wooden Pile

Strength of wooden material can be grouped into four classes of strength and type of stress as
in Table 3-12 (PKKI, 1961). Wooden material with Class 1 and 2 is usually used in
construction demand.
Table 3-12 Strength class for wooden material (after PKKI, 1961)
No. Type of stress Strength class of wood (kg/cm2)
1 2 3 4
1 flexural 150 100 75 50
2 comp or tensile 130 85 60 45
3 compressive 40 25 15 10
4  20 12 8 5
2
5 Young’s modulus, E (kg/cm ) 125000 100000 80000 60000

68
According to the new code SNI 2002 (Indonesian National Standardization, 2002), quality of
wooden material or quality code use mixed Letter and Number to declare Elastic Modulus as
in Table 3-13.
Table 3-13 Quality code for wooden material (after SNI, 2002)
Code of Young’s Flexural Tensile strength Tenslie strength Shear Compr. strength
quality modulus strength parallel fiber perpendic. fiber strength perpendic.
Fb Ft Fc Fv Fc
[MPa] [MPa] [MPa] [MPa] [MPa] [MPa]
E26 25000 66 60 46 6.6 24
E25 24000 62 58 45 6.5 23
E24 23000 59 56 45 6.4 22
E23 22000 56 53 43 6.2 21
E22 21000 54 50 41 6.1 20
E21 20000 50 47 40 5.9 19
E20 19000 47 44 39 5.8 18
E19 18000 44 42 37 5.6 17
E18 17000 42 39 35 5.4 16
E17 16000 38 36 34 5.4 15
E16 15000 35 33 33 5.2 14
E15 14000 32 31 31 5.1 13
E14 13000 30 28 30 4.9 12
E13 12000 27 25 28 4.8 11
E12 11000 23 22 27 4.6 11
E11 10000 20 19 25 4.5 10
E10 9000 18 17 24 4.3 9

Elastic modulus or Young’s modulus of wooden material can be estimated using equation
below (SNI, 2002).
(3-22)
Where: G is specific gravity of wooden material at water content 15%.

3.4.2. Concrete Pile

It is similar to wooden material that for cementitious pile (or concrete pile) Young’s modulus
is an important parameter. Cemented material such as concrete column, it is able to be
subjected to high compressive stress. Young’s modulus can be estimated using compressive
strength of concrete.
(kg/cm2) (3-23)
Relationship between characteristic compressive strength and flexural strength as described
in Eq. 3-24.
(MPa) (3-24a)
(kg/cm2) (3-24b)
where: Ec = Young’s modulus of concrete
fc' = characteristic of 28-day-compressive strength
fcf = flexural-tensile strength of 28-day
K = 0.7 for gravel and 0.75 for crushed stone

69
3.4.3. Stone Column
Hughes and Withers (1974) performed pioneering laboratory studies of sand columns within
a cylindrical chamber containing clay and used radiography to track the deformations
occurring within and outside the columns. They found that CCET (cylindrical cavity
expansion theory) represented the measured column behaviour very well and proposed that
the ultimate vertical stress (q) in a stone column could be predicted by:

(3-25)


where:  ' is the friction angle of stone infill, 'ro is the free-field lateral effective stress and c
is the undrained shear strength.
The equation above is widely used in practice today. There are alternative approaches for
estimating the bearing capacity of a single column and column group, such as that recently
published by Etezad et al. (2006). The authors report an analytical treatment of bearing
capacity failure mechanisms. Failure mechanisms adopted are based upon the output from a
combination of finite element analysis and field trials.
Absolute and differential settlement restrictions usually govern the length and spacing of
columns, and the preferred method of estimating post-treatment in European practice was
developed by Priebe (1995). Although, this method is strictly applicable to infinite array of
columns and has some empiricism in its development.
Priebe's settlement improvement factor, n, defined as:
(3-26)
It is a function of the friction angle of stone ' , the soil's Poisson's ratio and an area
replacement ratio dictated by the column spacing. The area replacement ratio is defined as
Ac/A , where Ac = cross-sectional area of one column and A = total cross-sectional area of the
'unit cell' attributed to each column. Ac/A is related geometrically to the column radius, r, and
column spacing, s, according to:

(3-27)

Where: k is  and 2/√3 for square and triangular column grids respectively.

Fig. 3-5 Typical of columns arrangements: Triangular and Square grids

(after Priebe, 1995)

70
Priebe's 'basic improvement factor' may be derived from the chart as shown in Fig. 3-6. Need
to be noted that the reciprocal area replacement ratio A/Ac is used on the chart.

Fig. 3-6 Priebe's basic improvement factor (after Priebe, 1995)

A lower limit to the undrained strength of cu = 15 kPa is suggested for treatment with stone
column, although there have been situations where softer soils have been successfully
improved (Raju et al., 2004). In other hands, UK National House Building Council (NHBC,
1988) suggests that stone columns should not be used when Ip >40%. Wood et al., (2002)
conducted what is considered to be the most comprehensive laboratory model investigations
of large groups of columns. The results suggest that significant improvement in the bearing
capacity requires an area replacement ratio of 25 % or greater.
McKelvey et al. (2004) used a transparent medium with ’clay-like’ properties to allow visual
monitoring of the columns throughout the foundation loading. The main findings of this
research relate to optimum column aspect ratio L/d (L=column length, d=column diameter)
that in the case of ’short column’ (i.e. L/d=6), bulging took place over the entire length of
column. The ’long column’ (L/d=10) deformed significantly in the upper region whereas the
bottom portion remained undeformed. McKelvey et al. (2004) postulated a ’critical column
length’ of L/d=6, which is in keeping with earlier work (Hughes and Withers, 1974; Muir
Wood et al. 2004).

3.4.4. Soil Cement Column

The deep mixing method is a technology that mixes in-situ soils with cementitious materials
to form a vertical stiffness in the ground. The deep mixing method (DMM) utilizes
quicklime, slaked lime, cement, fly ash, and/or other agents. The agents, widely referred to as
’binders’, may be introduced in the form of either a dry powder or slurry.
In the late 1960’s, Japan and Sweden independently began research and development of deep
soil mixing techniques using granular quicklime. The Japanese were focusing on soil
improvement techniques suited to large marine and estuarine projects, while Sweden was
primary focusing on soil improvement of soft clay for road and railway projects. The method
in which dry powdered lime and cement are used as the stabilizing agents is generally known
as the ’Dry Method of Deep Mixing’, whereas the use of stabilizing agents in slurry form
referred to as the ’Wet Method of Deep Mixing’ by the mid 1970’s, in effort to improve the
uniformity of soil treated by deep mixing. Typical operating parameters of the Japanese
mixing machines are summarized in Table 3-14.

71
 Table 3-14. Typical Japanese mixing installation parameters (after Kaiqiu, 2000)
Description Single drive shaft Double drive shaft Multibarrel drive shaft
Depth of stabilization 49 ft (15 m) > 49 ft (15 m) 98-131 ft (30 - 40 m)
Penetration velocity 2 - 3.3 ft/m 0.7 - 3.3 ft/m 3.3 - 6.6 ft/m
0.6 - 1.0 m/min 0.2 - 1.0 m/min 1.0 - 2.0 m/min
Withdrawal velocity 2 - 3.3 ft/m 0.7 - 3.3 ft/m 3.3 - 6.0 ft/m
0.6 - 1.0 m/min 0.2 - 1.0 m/min 1.0 - 1.5 m/min
Rotating speed 50 rpm 46 rpm 20 - 30 rpm (penetration)
40 - 60 rpm (withdrawal)

Deep mixing methods in the U.S. have been used on several projects either dry or wet
method. In general, dry mixed stabilization is appropriate for sites with relatively deep
deposits of very soft soil, and sufficient groundwater to hydrate both the lime and cement
(Esrig and Mac Kenna, 1999). Cohesive soils with moisture contents between 60% and 200%
are best suited for dry mixing.
While several different types of laboratory tests are used to evaluate the shear strength and
stiffness of deep mixed columns, the most frequently used is the unconfined compression test,
mainly because of the simplicity of the test. Many factors affect the unconfined compressive
strength because of a wide variety of soil types and binder mixes. The 28-day unconfined-
compressive strengths for soil treated by the wet method may range from 140 to 27000 kPa
(Haley&Aldrich, 2000; Kaiqiu, 2000; Tatsuoka&Kobayashi, 1983) whereas using the dry
method range from 14 to 2700 kPa (Hebib&Farrell, 2002; Jacobson et al, 2002; Kaiqiu,
2000). Unconfined compressive strengths, qu, for three projects in the U.S. are presented in
Table 3-15.
Table 3-15 Specified values of qu on deep mixing projects in the U.S.
Soil types / binder
Projects Specified qu Reference(s)
amount
Oakland Airport Wet method; Loose At 28 days, Average qu > Yang et al,2001
Roadway, California sandy fill and soft 1035 kPa, Minimum qu >
soil; 160-240 kg/m3 690 kPa
cement
Central Artery Project, Wet method; Fill At 56 days, Maximum qu Lambrechts et
Boston and organic soft > 26900 kPa, Minimum al,1998;
clay; 220-300 kg/m3 qu > 2100 kPa Maswoswe,2001
cement
I-95 Route 1, Wet method; Soft At 28 days, Average qu > Shiells et al,2003;
Alexandria organic clay; 300 1100 kPa, Minimum qu > Lambrechts et
kg/m3 cement 690 kPa al,1998
Source, M. Smith, 2008

Stabilization of soft organic soils with cement columns using the mix-in-place technique
(MIP) for a railway embankment at section of Büchen-Hamburg was upgrade in 2003 by the
German Railway company (Deutsche Bahn) to allow a train speed of 230 km/h
(Schwarz&Raithel, 2005). The cement columns (diameter 0.63 m and 5-8 m length) were
installed in a square 1.5×1.5 m grid, containing 2.5 to 3% cement, which can be characterized
72
as a wet deep mixing technique, the composition of binder (water, cement and bentonite) and
the water binder ratio (approx. 1.0). Each of 500 m3-treated soil, 6 unconfined compression
tests, was carried out after 28 days. According to the test, unconfined compressive strength
after 28 days of all samples exceeds the design criteria of qu ≥2.2 Mpa.
For cemented columns installed by the wet method, Takenaka (1995) reported that undrained
shear strength is equal to one-half of the unconfined compressive strength for those values
below several hundred kPa and become less than one-half when they are greater than several
hundred kPa. As a rule of thumb, Takenaka (1995) recommended that undrained shear
strength be taken as one-third of the unconfined compressive strength. Kivelo (1997) found
that the undrained shear strength can be less than one-half the unconfined compressive
strength at low confining pressures. However, when the total confining pressure exceeds 150-
250 kPa, the undrained shear strength becomes almost constant at a value equal to one half of
the unconfined compressive strength.

Fig. 3-7 Undrained shear strength of lime/cement column (after Kivelo, 1997)

The peak strength is typically reached at strains 0f 1% to 2% and decrease in strength once
the peak strength is exceeded (Kivelo, 1998). The residual strength of soil-cement is 65% to
90% of the unconfined compressive strength (Tatsuoka&Kobayashi, 1983).
The undrained secant modulus of elasticity, E50, which evaluated at 50% of the peak strength,
is a measure of soil-cement compressibility. Some researchers correlate E50 to unconfined
compressive strength for columns installed the dray method (Braker, 2000; Broms, 2003;
Jacobson et al., 2003; Navin&Filz, 2005). Whilst for cement treated soils using the wet
method also have been performed and presents relatively higher values of secant modulus of
elasticity than those using the dry method (Kawasaki et al., 1981, Navin&Filz, 2005, Fang et
al., 2001). The relationship between E50 and qu is provided in Table 3-16.
Table 3-16 Relationship between E50 and qu
Binder types E50 Reference(s)
50 - 180 qu Baker, 2000; Broms, 2003
Dry lime/cement
75 qu Jacobson et al., 2003
65 - 250 qu Baker, 2000; Broms, 2003
Dry cement
300 qu Navin and Filz, 2005
350 - 1000 qu Kawasaki et al., 1981
30 - 300 qu Fang et al., 2001
Wet cement
150 qu Mc Ginn and O'Rouke, 2003
300 qu Navin and Filz, 2005
Source:M. Smith, 2008

73
When the modulus of elasticity is used in design analysis, the secant modulus E50 is typically
used as the design value of column Ecol. The modulus of elasticity on samples prepared in the
laboratory is typically higher than modulus determined from coring test obtained in situ
actual columns (Broms, 2003).
The oedometer compression modulus Eoed is related to modulus of elasticity Ecol and
Poisson's ratio  as follows:

(3-28)

Generally the Poisson's ratio of deep mixed treated soil is around 0.25 to 0.45 (Terashi,
2003). Therefore, Eoed is equal to 1.2 to 3 Ecol.
The total unit weight of treated soil using the dry method increases from 3% to 15% above
the untreated soil. Whilst tensile strength of soil improved by the wet method, it is 10% to
20% of unconfined compressive strength. Moreover, permeability of treated soil ranges from
10-7 to 10-8 m/s is routinely achievable.

3.5. Characteristic of Loading

Transport infrastructure such as roadway, railway and runway is mainly subjected by moving
load. Although at a certain situation they are static loading such as car parking at the parking
lot, airplane parking on a parking stand at the apron. When a vehicle passes through a
roadway, time loading at certain point on the surface of the roadway section depends on
velocity of the vehicle. Load repetition induced by wheels of the vehicles on the surface is
able to result in a rut depth during a service period of roadway.

3.5.1. Vehicular Traffic

When vehicles crossing on the roadway, all kinds of vehicle refer to Equivalent Single Axle
Load (ESAL) of 18 kips (or 8.12 ton) with inflation pressure 80 psi (560 kPa). Wheel
configuration may be single, dual wheels and tandem. Meanwhile, light vehicles (passenger
cars) have an inflation pressure around 32 psi (225 kPa).

3.5.2. Airplane
Compared to a wheel load generated by traffic on the roadway, the wheel load of airplane has
quite higher in magnitude than that of on the roadway. It depends on weight, tire pressure,
wheel configuration of airplane. Tire pressures of the airplane vary from 0.5 MPa to more
than 1.5 MPa.
Because weight of the airplane is quite heavy that will be transferred to the surface of the
pavement, wheel configuration plays an important role. For light aircraft, it uses a single
wheel configuration. Dual wheel configuration can be seen for a moderately weight of the
airplane. Furthermore, another configuration for heavy weight of the airplane is tandem and
also dual tandem.
In other to accommodate the various airplanes for operational movement, it needs to provide
aerodrome areal as shown in Table 3-17. The aerodrome reference code uses number and
letter codes to express class of airport (ICAO, 1999). Maximum allowable tire pressure
category consists of High (no pressure limit), Medium (pressure limited to 1.50 MPa), Low
(pressure limited to 1.00 Mpa) and Very low (pressure limited to 0.50 MPa. The higher tire
pressure indicates the heavier weight of the airplane.

74
 Table 3-17 Aerodrome reference code (after ICAO, 1999)
Code element 1 Code element 2
Code Aeroplane reference field Code Wingspan Outer main gear wheel
number length letter span *
[1] [2] [3] [4] [5]
1 Less than 800 m A Up to but not including Up to but not including
15 m 4.5 m
2 800 m up to but not B 15 m up to but not 15 m up to but not
including 1200 m including 24 m including 24 m
3 1200 m up to but not C 24 m up to but not 6 m up to but not
including 1800 m including 36 m including 9 m
4 1800 m and over D 36 m up to but not 9 m up to but not
including 52 m including 14 m
E 52 m up to but not 9 m up to but not
including 65 m including 14 m
F 65 m up to but not 14 m up to but not
including 80 m including 16 m
* Distance between the outside edges of the main gear wheels

3.5.3. Trains
In Germany, according to Ril 836 for rail infrastructure, the subgrade or improved subgrade
has to be able to support a load above this surface layer with bearing capacity at least around
52 kPa. For high speed trains from 100 to 300 km/h, it needs the additional layer around 1 to
2 times of the superstructure thickness (Muncke et al., 1999; Kempfert et al., 1999).
It is different for each country with respect to a standard axle load. For instance in Greece,
Beskou et al., (2011) reported that locomotive (or engine) is around 210 kN and 150 kN for
carriage. In Indonesia, railway infrastructure is subjected to axle load maximum 180 kN
(Indonesian Railway Code, 2003).

75
 CHAPTER 4

Experimental Works
and Field Case Studies

4.1. General
To understand well the pavement behaviour and/or embankment overlying soft soil, some
researchers have performed some experimental works either in laboratories using small scale
approach or monitoring full-scale directly in the field (Zaeske, 2001); Heitz, 2006; Wei-ping
et al., 2007; Hassandi et al., 2005; Schwarz et al., 2005; Djarwadi, 2006; Almeida et al.,
2008; Eekelen, 2009). Treatments for ground improvement in this study only focus on
horizontally inclusion of geosynthetics, vertically inclusion of piles and the combined
technique both of them. Meanwhile, some tests have been done to know an influence of
geosynthetics into granular soil (Bussert, 2006; Ruiken&Ziegler, 2008; Ruiken et al., 2010).
Furthermore, some intensive tests regarding with geosynthetics-soil interaction behaviour
have been carried out using interaction testing device (ITD) at Geotechnical Institute of
Technical University Bergakademie Freiberg (Aydogmus&Klapperich, 2004; Aydogmus,
2006; Tamaskovics&Klapperich, 2010; Althoff, 2010, 2011).

4.2. Experimental Work at the Laboratory

4.2.1. Geosynthetics
4.2.1.1. Horizontal pressure experiment
Ruiken and Ziegler (2008) performed a test to know the influence of inclusion of
geosynthetics in unbounded granular material. Tubular sample of diameter 500 mm and
height 1100 mm consisting of gravel d50=12 mm and sand d50=0.5 mm was compacted
around 95-100% standard proctor. Geogrids with tensile strength of 30 kN/m and aperture 32
cm were inserted horizontally from only a layer until 7 layers in tubular sample. There were 9
tape gauges installed to measure the radial deformation of the sample. Vertical stress with
constant speed of 1 mm/min ( 0.1% per minute) was applied. Fig. 4-1 is showing the setup
test.

Fig. 4-1 Setup of sample in laboratory test

(after Ruiken and Ziegler, 2008)
76
Ruiken and Ziegler (2008) reported their result test that the inclusion of geosynthetics can
reduce the radial deformation. The more number of layers will reduce the magnitude of radial
deformation as shown in Fig. 4-2. This result indicated that more than 3 layers did not give
significant impact on radial strains. It suggests a hint that distance of geogrid probably was
effectively applied in range 280-360 mm because the radial deformations resulted from 2 to 3
layers of geogrids are very close.

Fig. 4-2 Radial strains in laboratory test (after Ruiken and Ziegler, 2008)

Some studies to observe the horizontal deformation subjected to a vertical load has been
performed by researchers (Clayton et al., 1993; Soong & Koerner, 1997; Yang et al, 2009).
Ruiken and Ziegler (2010) also performed the deformation by inclusion geosynthetic as
shown in Figure 4-3. A sand sample in box sized (H×B×L) 1m×1m×0.45m compacted up to
100% standard proctor was used to do the experiment as shown in Fig. 4-3. A couple side
was made of steel and another couple side was glass with 106 mm thick. To reduce friction
between sand and glass, a thin latex membrane inserted in between and a uniform load up to
50 kN/m2 was applied at the surface of the sample box.

Fig. 4-3 Installation for measuring horizontal pressure (after Ruiken et al., 2010)

Inclusion of geogrid from one up to 5 layers has been applied and then the experiment result
was a curve correlating between horizontal deformation of vertical movable plane and
horizontal outward pressure on the vertical plane as depicted in Fig. 4-4. It was interesting

77
that for 2 and 3 inclusion of geogrids they almost coincided with each other. Moreover, using
of 5 layers did not give lower horizontal pressure than that of using 4 layers. Overall, more
and more numerous layers suggest that horizontal outward pressure was reduced.

Fig. 4-4 Horizontal outward-pressure (after Ruiken et al., 2010)

4.2.1.2. Load transfer mechanism experiment

Bussert (2006) performed a research for studying the load transfer mechanism by soil/geogrid
with ’insoil’ testing using setup with movable front wall and introducing forces from the soil
into the forcing geogrids. Fig. 4-5 is showing the test setup.

1. Side frame
2. Base plate
3. Load plate with reinforcement
4. Threaded rods
5. Plug gauge with fine thread
6. Force measurement
7. Movable front plate
8. HDPE coating with PE membrance
9. Displacement transducer
10. Sand/gravel
11. Geosynthetic layer

Fig. 4-5 Test setup with movable front wall (after Bussert, 2006)

The test results showed that interaction of soil and geogrid mainly depends on the
geosynthetic layer spacing, soil grain size, geosynthetic aperture size as well as strength of
shape and extensional stiffness of geogrids. Load plate was subjected to a load (in kPa) and
strain of geosynthetics was measured based movement of movable wall. Magnitude of load
78
in kPa was recorded as the stress reduction at the front of the moving wall in horizontal
direction caused by different geogrid products as reinforcement and also without
reinforcement is shown in Fig. 4-6.

Fig. 4-6 Earth stress reduction at movable front wall with different geosynthetics
and without reinforcement of soil body (after Bussert, 2006)

4.2.1.3. Geosynthetics-soil-interaction behaviour experiment

A lot of tests have been carried out in the laboratory of the Geotechnical Institute of
Technical University Bergakademie Freiberg to understand the interaction-behaviour
between geosynthetics and cohesive soil (Aydogmus&Klapperich, 2004; Aydogmus, 2006;
Tamaskovics&Klapperich, 2010; Althoff, 2010, 2011). The geosynthetics-soil-interaction-
testing device (ITD) was provided. The ITD is a large shear frame device for shear tests
(ISV). It is possible to use this device for friction tests (abbreviation IRV) and pull-out tests
(abbreviation IPV). Figure 4-7 gives an overview of the testing device.

(b) running test

(a) preparing sample in shear box

Fig.4-7 Geosynthetics-soil-interaction-testing device (ITD)

79
The shear box consists of a lower and upper shear frame. This box can be filled with more or
less 80-120 kg soil mass depending on type of soil, water content and degree of compaction.
Before installing material sample in the shear box, the material is homogenized using a
mixing device. The soil mass is built-in several layers and each layer is compacted with a
hand compaction device. During testing, the upper shear frame can move vertically, but will
not go under the pre-adjusted target value. Therefore, the shear gap can adjust itself during
the experiment. The parameters of the device are listed in Table 4-1.
Table 4-1 Specification of interaction testing device
Parameters Size and/or rate Unit
Size of the shear box L=500, W=500, H=200 mm
Normal stress range 0 - 600 kN/m2
Shear and pull-out force 125 kN
Maximum shear displacement 160 mm
Maxsimum pull-out displacement 400 mm
Shear velocity 0.000001 – 12.5 mm/minute

Tests of friction and shear behaviour are costly and time-consuming. Pull-out test in
particular is one of the most expensive performance tests (Koerner, 2005). Thus, in previous
research a more economical technique for using the ITD (for IRV, ISV, and IPV) was
investigated in the institute (Tamaskovics and Klapperich, 2010; Althoff, 2010; 2011;
Widodo, 2013). Multi-Stage Large Shear-Frame test has some advantages. During
installation of sample, soil in the shear box is highly compacted and then the tests are carried
out at low velocity. These test series contain multi-stage tests (ISV, IRV, and IPV) using a
cohesive soil (Canitz-silt) in combination with twelve different geogrids from six different
producers which have some properties such as short time tensile strength (25-180 kN/m),
elongation, aperture size, junction strength, surface and thickness. According to the grain size
distribution, Canitz-silt is strongly sandy and slightly clayey silt.

My research activity under DFG-research project FY 2012/2013 in this laboratory was still
ongoing. Multi-stage tests for ISV, IRV, and IPV have been undertaken not only for Canitz-
silt soil, but also Hohen Bockär glassy sand and Kaolin soil. In order to reach maximum
density for three kinds of soil, Proctor test has been done in which the measured water
contents are 11.3%, 2.5%, and 28.07% for Canitz-silt soil, Hohen Bockär glassy sand and
Kaolin soil, respectively. All types of geosynthetics were tested in machine direction, with
and without cross-element in order to understand behaviour of geosynthetics-soil-interaction.

4.2.2. Piled Embankment

4.2.2.1. Pile-soil relative displacement experiment
Wei-Ping et al. (2007) conducted the experimental test to investigate soil arching within
reinforced and unreinforced piled embankments in 2D-layout. A total of 15-model tests were
conducted to evaluate the effects of pile-soil relative displacement, embankment height, cap
beam width and clear spacing, and geosynthetics with different tensile strength on the stress
concentraion ratio and settlements in the embankments.
Model test consisted of a bricked base, two rubber water bags and tank. Toughened glass was
used for the four walls of the tank to allow observation. The system was 1500 mm long, 1000
mm wide and 1440 mm high as shown in Fig. 4-8. The base was 140 mm high, two water
80
bags which each water bag with a dimension of 1000 mm long, 600 mm wide and 140 mm
high, were placed inside the base and filled with water.
(a) (b)

Fig. 4-8 Layout of test setup (a) Top view and (b) Side view (after Wei-ping et al., 2007)

The sand used in the model tests was come from Qiantang River Beach, China. Specific
gravity G = 2.64, the coefficient of uniformity Cu = 2.5 and the coefficient of curvature Cc =
0.96. Sand grains were subangular and predominantly quartz with D10 = 0.1 mm, D60 = 0.25
mm, Dmax = 2 mm, emax = 0.89 and emin = 0,54. The unit weight of the fill in the tank was
15.35-15.83 kN/m3 and corresponded to a relative density of 557%. Peak secant angle of
shearing resistance max was 44o. Three types of geosynthetics were used in the model tests
with biaxial tensile strength 0.35, 1.40 and 22.5 kN/m at 8% axial strain.

Based on 15 model tests having a ratio of embankment height to clear spacing, h/s, from 0.7
to 2.0 they obtained some interesting results. In Test 1 through Test 4, when h/s < 1.4, the
surface of the embankment was non-uniform. It implies that the differential settlement
occurred on the top of an embankment. In Test 5 through Test 7, when h/s > 1.6, the
settlements at the base of the embankment were non-uniform, but the embankment surface
remained almost horizontal. Deformations in embankments of model Tests 1 through Test 7
also suggest that the height of equal settlement plane is about 1.4 - 1.6 times the cap beam
clear spacing, i.e., he = (1.4-1.6)s. Model tests from 1 to 7 were dedicated to unreinforcement
embankment whereas model tests from 8 to 13 were intended to reinforcement embankment
using geosynthetic. By using geosynthetics either low embankment h/s = 0.7 or hight
embankment h/s = 1.8 produces the higher stress concentration ratio, because the
geosynthetics transfers the vertical load over geosinthetics to beam cap. In addition, from the
experimental test that the higher embankment indicated the higher stress concentration ratio
when using the same of tensile strength of geosinthetics.

4.2.2.2. Cyclic loading experiment

Zaeske (2001) conducted the experimental test for embankment over soft soil reinforced by
geosynthetics. Heitz (2006) developed this experiment to observe the influence of cyclic
loading on the embankment over soft soil. Fig. 4-9 and Fig. 4-10 illustrate the setup of test
model of the embankment over soft soil and placement of devices respectively.
81
 Fig. 4-9 Setup of test model under cyclic loading (after Heitz, 2006)

Fig. 4-10 Location of devices in laboratory test (after Heitz, 2006)

In this experiment, four cross-sectional 16×16 cm concrete piles were installed to support an
embankment. The unit weight of piles  = 24 kN/m3, modulus of elasticity E = 23,982
MN/m2, strain emax = 0.9%, Poisson's ratio  = 0.2, Some properties of soft soil (peat) used in
the experiment consist of unit weight of piles  = 8 kN/m3, modulus of elasticity E = 0.85
MN/m2 at ' = 100 kPa, water content w = 300-350%, cohesion c = 8.5 kPa, internal friction
angle  ' = 24o, compression index cc = 2.48, organic content OC = 80.2%, permeability ki =
4.1E-7 m/s.
Geogrid used in the research were Polyester GW 60 PET and Polyvinyl GW 180 PVA.
Tensile strength for longitudinal and transversal direction are 60 kN/m and 180 kN/m
respectively. Elastic modulus of geogrids are 850 kN/m at a strain of 2% for GW 60 PET
and 3,800 kN/m at strain of 1.25%.

82
Some interesting findings were obtained from the study. Settlements at surface of
embankment induced by cyclic loading with the same magnitude of the load was deeper than
static loading. Numerous number of geosynthetics layers were able to reduce settlement at
the surface of the embankment.
Highest tensile force occurred at an interface between the pile and geosynthetics, and then the
lower tensile force was located on between piles in the orthogonal direction, the lowest
tensile force was at the middle of geosynthetics between piles in a diagonal direction.
Moreover, maximum strain max for low embankment (h = 0.35 m) is higher than high
embankment (h = 0.70 m). Meanwhile, settlement at the surface of the embankment for low
embankment was deeper than the high embankment over the increasing external load.

The influence of cyclic and dynamic loading could decrease the arching effect. Moreover,
vertical stress in the embankment fill will increase over the large number of repetitions.

4.3. Case Studies in the Field

4.3.1. Geosynthetics
4.3.1.1. Weesenstein railway rehabilitation project
Klompmaker et al. (2008) showed the Elbe-river flood disaster in 2002 that destroyed
approximately 80% of infrastructure in the valley of Mügliztal near Dresden. To operate the
rail traffic again in the shortest possible time, the combination of geosynthetic reinforcement
together with the fill material as a compound was applied to ensure the internal and external
stability of the structure. The steel grid element stabilizes the slope face and geotextile non-
woven separation. Filter layer was provided to prevent the erosion of fill material. The
approximately 5 m high geogrid reinforced part on the bottom of the slope is constructed with
60o inclination and is superposed by a 4 m-high embankment (see Fig. 4-11). Structural
analysis for the reinforced slope was carried out on the basis of the German recommendations
for geosynthetic reinforcements.

a) Collapsed railway embankment b) Geogrid reinforced railway embankment

Fig. 4-11 Weesenstein railway embankment (after Klompmaker et al., 2008)

83
The applied product Secugrid 120/40 R6 has a Federal Railway Agency's certification for
this application. A 0/45 mm crushed mix acted as covering material. The geogrid was
installed in 10 layers with a distance of 0.5 m. Within just a few weeks, whole structure had
been completed, so that the flow of regular traffic was quickly reinstated.

4.3.1.2. Tabing-Duku road widening project

Klompmaker et al. (2008) also reported the road widening project on a section of Tebing-
Duku near the town of Padang on the largest of the Indonesian island, Sumatera, required an
existing road from the airport to city center. In spite of extremely problematic ground
conditions on the site with low bearing capacity and a high ground-water level, the solution
involved the road embankment reinforced with uniaxial Secugrid R-geogrids. Secugrid
120/40 R1 geogrids 15,200 m² made of polypropylene (PP) were installed with anchorage
lengths between 6.0 m and 10.0 m. The original design envisaged geogrid reinforcement
layers with 60-80 kN/m tensile strength with a layer spacing of 0.5 m as shown in Fig. 4-12.

a) Subgrade condition b) Reinforced slope c) Finished project

Fig. 4-12 Tabing-Duku road widening (after Klompmaker et al., 2008)

The Secugrid solution with geogrid width of 4.75 m allowed faster and more cost-efficient
installation. The deformations were remedied as further layers of Secugrid were installed to
reinforce the embankment. Measurements were taken from the upper edge of the
embankment to determine the degree of deformation. Any deformations were hardly noted,
which confirmed that Secugrid had allowed an existing road to be successfully and safely be
widened on an extremely soft subgrade at favourable costs.

4.3.1.3. Setoko-Nipah road embankment project

Djarwadi (2006) reported the execution of road embankment using geotextile reinforcement
in Setoko and Nipah islands in Province of Kepulauan Riau, Indonesia. The soil investigation
shows that soft soil layer up to 15 m deep below the ground surface is present. Water table at
a certain situation can achieve about 1.5 m above the elevation of subsoil. Road embankment
height of 3.5 m consisting of unit weight = 18.5 kN/m3, cohesion c = 18 kN/m2, and internal
friction angle  = 19o was constructed over soft soil. Peat and/or very soft marine clay height
of 4.5 m was existing beneath the embankment fill consisting of unit weight = 14.5 kN/m3,
undrained shear strength su = 5 kN/m2, and internal friction angle  = 2o. Underneath this
layer, very fine sandy clay height of 1.5 m was present consisting of unit weight = 16.0
84
kN/m3, undrained shear strength su = 7.5 kN/m2, and internal friction angle  = 3o.
Furthermore, marine clay consisting of unit weight = 18.0 kN/m3, undrained shear strength
su = 10.0 kN/m2, and internal friction angle  = 2o was present beneath sandy clay. Typical
cross-section of the road embankment over soft soil in Setoko and Nipah islands is depicted
in Fig. 4-13.

Fig. 4-13 Cross-section of road embankment in Setoko and Nipah islands

(after Djarwadi, 2006)

According to Djarwadi (2006) for internal stability using the limit equilibrium method, the
construction needs a tensile strength of geotextile around 87 kN/m, whereas foundation
stability (external stability) needs a tensile strength of geotextile about 272,8 kN/m.
Compaction has to be done in each embankment thickness of 0.3 m using vibratory smooth
drum compactor 15 ton weight until compaction degree of proctor standard ≥ 97%, in which
are 8 passing with a constant speed 10 km/h. The compaction tests using the sand cone test
are for each 2500 m³ volume of embankment. The observation for settlements of 400 m long
road embankments is depicted in Fig. 4-14. We can see that settlements at the final elevation
can be more than 1.0 m during less than 3 weeks of execution.

Fig. 4-14 Settlements during work execution (after Djarwadi, 2006)

85
4.3.1.4. Yamanote Line railway project
Palmeira et al. (2008) reported that since early 1990's reconstruction of railway that collapsed
by flooding with embankments having geosynthetic-reinforced steep slope or Geosynthetic-
reinforced soil retaining walls (GRS RWs), having a stage-constructed FHR facing or their
combination, started based on successful experiences of high cost-effectiveness and high
performance of GRS RWs. This construction was employed also in other similar cases after
this event of flooding. The gentle slope of the embankment and conventional retaining walls
that collapsed by an earthquake (the 1995 Kobe earthquake) were reconstructed using GRS
RWs. The GRS RWs having a stage-constructed full-height rigid (FHR) facing is now the
standard retaining wall construction technology for railways in Japan (Tatsuoka et al. (1997a,
2007).

Fig. 4-15 History of elevated railway and highway structure in Japan

(after Palmeira et al., 2008)

Fig. 4-16 and 4-17 show a typical GRS RWs having a FHR facing constructed in center
Tokyo, in Yamanote Line near Shinjuku station. It had been built during 1995-2000. This
new type GRS RWs has been constructed in more than 600 sites in Japan and the total wall
length is recently more than 100 km as of March 2008 (after Palmeira, 2008).

Fig. 4-16 GRS-RWs having a FHR facing in Yamanote Line, a) Typical cross-section
b) Wall under construction c) Completed wall, (after Palmeira et al., 2008)
86
 Fig. 4-17 Staged construction of GRS RWs with a FHR facing
(after Palmeira et al., 2008)

4.3.2. Piled Embankment

4.3.2.1. Barra da Tijuca field test
Almeida et al. (2008) performed a field investigation for a case ’’low embankment’’ at Rio de
Janeiro. They investigated low embankment over a 10 m depth of soft clay at Barra da Tijuca
district. The soft soil parameters consisted of water content w = 100-500%, Plasticity Index Ip
= 100-250%, the average unit weight  = 12.5 kN/m3, the undrained shear strength su = 4-18
kPa, Compression ratio CR=cc/(1-eo) = 0.5, the average coefficient of horizontal
consolidation ch = 6.5E-8 m2/s.
There were two low embankments observed on the site. Dimensions and characteristics of
two embankments are provided in Table 4-2 and illustrated in Fig. 4-18.
Table 4-2 Main characteristics of two piled embankments

Characteristics Embankment 1 Embankment 2

Construction year 2004 2004-2005
Number of piles 1900 10000
Pile spacings, s (m) 2.5 2.8
Square pile cap, b (m) 0.8 1.0
Clear spacing, a = s – b (m) 1.7 1.8
Embankment height, h (m) 1.2 1.4
Ratio h/a 0.7 0.78
Geogrid characteristics Fortrac R, Polyester, Fortrac, PVA,
Biaxial Biaxial
Nominal geogrid strength (kN/m) 200 200 and 240
Geogrid modulus (kN/m) 3600 3600 and 4400
Fill height below the pile cap (m) 2.0 0.6 – 1.0
Soft soil deposit thickness (m) 8-10 9-11

87
 b)

a)

c)

Fig. 4-18 Embankment at Barra da Tijuca, a) General scheme of piled embankment, b) Pile
caps above the initial fill, c) Pile caps inside the initial fill (after Almeida et al., 2008)

Some findings are gained from the field test. The strains of geogrids at the face of the pile cap
are higher than those at half distance between caps. Meanwhile, strains of geosynthetics in
between pile caps at an orthogonal direction were higher than those at a diagonal direction.
Based on the damaged geogrids, a circular cap could be more effective than the square caps
in thoses cases. In addition, the result suggested that subgrade reaction should not be
considered for designing a geogrids-piled embankment.

Fig. 4-19 Measured strains in geogrid in points at: a) face of the pile cap, b) half distance
between caps (after Almeida et al., 2008)

88
4.3.2.2. Gebeng Bypass Highway field test
Hassandi and Edil (2005) reported a full scale test embankment to evaluate the performance
of different types of loading platform (LTP) supported on geopiers. Three types of LTP were
constructed, they include reinforced concrete LTP, geosynthetic-reinforced LTP with two
layers of geogrid (catenary beam LTP), and geosynthetic-reinforced LTP with three or more
layers of geogrids (beam LTP).
To evaluate the performance of the catenary and the beam LTP. A well instrumented full-
scale test embankment was constructed in Gebeng, Pahang State, Malaysia as shown in
Figure 4-20. The length of the embankment was approximately 90 m, 13.5 wide, and 3.5 m
high. The test embankment was divided into 4 major sections (section 1 through 4) and 2
controlled sections (C1 and C2) at the two ends of embankments.

Fig. 4-20 Layout of the major and the control sections of test embankment
(after Hassandi et al., 2005)

Site investigation carried out indicated a silty clay layer as deep as 15 m at some locations.
However, this soft layer generally ends at approximately 5-6 m below the original ground
level. The soils are composed of highly plastic clay and silt with natural water content w
between 35% and 61%. The field vane tests that the shear strength Su lies between 14 to 60
kPa with most of the values less than 25 kPa and sensitivity S varies from 3 to 11.
Geopiers were 75 cm in diameter with 5.5 m deep installed at all sections. Section 1 has the
beam LTP with four layers of geogrids and 3.25 m geopier spacing designed by Collin
Method. Section 2 has the beam LTP with three layers of geogrids with 2.5 m geopier
spacing designed by the Collin Method. Section 3 has the catenary LTP with two layers of
high strength with 2.5 m geopier spacing design by BS 8006. Section 4 has a continuous
reinforced concrete slab as LTP. Settlement plates were set directly above the geopier
elements and at the center between groups of geopier elements. Earth pressure cells were
positioned on the geopiers and in between them. Piezometers and extenometers were placed
at different depths. Vertical inclinometers were placed at the toe of the embankment in

89
between the geopiers, and horizontal inclinometers immediately under and across the
embankment in between the geopiers.
Geogrid biaxial Tensar SS 20 was utilized for reinforcement in Section 1 and 2. The distance
between each geogrid layer in this section was 300 mm. In Section 1, aggregate blanket was
laid between layers of geogrid to a thickness of 1.5 m. Meanwhile in Section 2 was 1.0 m.
Geogrid uniaxial Miragrid 24 XT was used for reinforcement in Section 3. The distance
between the geogrid layers was 75 mm and the selected aggregate blanket thickness was 0.3
m.
According to Nordic Hanbook (2003) it recommends that the differential settlement between
the subsoil and the columns shall not be more than 0.1 m and 0.2 m. Total and differential
settlements of embankment test from different section was indicated in Table 4-3. We can
look that the use of piled embankment would reduce total settlement. Section 3 as catenary
LTP suggests the highest differential settlement and the lowest total settlement compared to
other sections.
Table 4-3 Total and differential settlements at the base of the embankments

Description Section 1 Section 2 Section 3 Section 4 C1 and C2

Total settlement (mm) 145 103 102 136 490
Differential settlement (mm) 9 14 23 6 NA
Maximum strain in geogrid (%) 1.4 1.2 0.6 NA NA
Source: Hassandi et al., 2005

Another finding resulted from the vertical inclinometer readings indicates that the use of LTP
supported on columns could reduce the lateral displacement of the subsoil at the edge of the
embankment.

4.3.2.3. Kyoto road field test

For the construction of roads on very soft soils, several construction methods are available.
The piled embankment using geosynthetic reinforcement (GR) is one of these methods,
becoming more and more popular in the Netherlands (Eekelen et al., 2009). Until 2009, at
least 20 piled embankments have been constructed underneath highways and local roads.
Regarding with Dutch guideline for the design of piled embankment, a full-sclae field has
been conducted.
A full-scale field test has been carried out on ’’Kyoto Road” in Giessenburg, in the
Netherlands (see Fig. 4-21). Results of 2 years of measurements (2005-2007) in a full-scale
field test were validated. The Kyoto road was constructed on 13 m long wooden piles with
square grid pile spacing of 1.27 m, concrete piles cap with a height of 0.4 m and 0.3 m in
diameter. The geogrid reinforcement consisted of two layers of uniaxial geogrids,
perpendicular on road axis Fortrac 400/30-30 M and along the road axis Fortrac 350/50-30
M. The embankment fill 1.15 m high made of a “Hegemann sludge mixture” was overlaid
over the geogrid layers. The Hegemenn sludge mixture is a mixture of dredged material and
additives containing mainly clay and cement having average unit weight  = 18.6 kN/m3,
90
natural water content w = 18.1%, permeability kv = 2.1E-9 m/s, internal friction angle =
33.8o and a cohesion of 11.5 kPa. The modulus of subgrade reaction of soft soil (peat),
determined from compression tests on samples, is k = 477 kN/m3.

Fig. 4-21 Scheme of the full-scale test Kyoto road (after Eekelen et al., 2009)

Comparison of two and a half years of measurements on a full-scale test compared with the
calculation based on German EBGEO and the British BS 8006 showed that EBGEO gives
better predictions of the load distributed in the piled embankment than BS 8006. The
dynamic load influences the arching and tensile stress in the geosynthetics. However, the
arching can be restored during the rest period.

4.3.2.4. Büchen-Hamburg railway section field test

Schwarz and Reithel (2005) reported ground improvement of very soft organic subsoil at two
sections of Hamburg-Berlin railway lines by means of installing cement columns with the
mixed-in place (MIP) method, which can be characterized as a wet deep mixing technique,
and by reinforcing the embankment with geogrids. The section Büchen-Hamburg was
upgraded in 2003 by the German Rail company (Deutsche Bahn), to allow a train speed of
230 km/h. Due to very soft organic soil layers (peat and mud) and insufficient bearing
capacity of embankment, an improvement of the railway embankment was necessary in two
sections with a total length of 625 m near the railway station Büchen (see Fig. 4-22 and Fig.
4-23).

Peat soil has a water content of 80 to 330% and an organic content between 25 and 80%.
Underneath these soil layers, slightly silty sand layers with a thickness up to 8 m are present,
which are medium dense packed. At the base of the sand layers, boulder clay is present,
which has a soft to stiff consistency and a water content of 10 to 20%.
The cemented columns (diameter 0.63 m), totally 3,260 MIP columns of a length between 5
and 8 m, were installed in square 1.5 × 1.5 m grids using the MIP-technique. Using a single
auger, a cement slurry is injected continuously into the soil during penetration as well as
during retrieval of the auger. On the top of the MIP-columns two layers of Fortrac R PVA
91
geogrid type M 400/30-30 were unrolled. To obtained an uniform bearing platform for the
ballast bed, 2.5 to 3% cement was added to the filling material. A gap graded gravel-sand
mixture (soil group SI according to the German Standard DIN 18196) with a coefficient of
uniformity ≥6 was used. The filling material was placed in layers of maximum 30 cm
thickness in accordance to the Ril 836. Each layer was compacted to a degree of compaction
at least 98%.

Fig. 4-22 Foundation system at section of Büchen-Hamburg (after Schwarz et al., 2005)

For laboratory testing purpose, wet grab samples were extracted from 4.5% of column every
500 m3 of treated soil, 6 unconfined compressive tests were carried out after 28 days, to
determine the unconfined compressive strength qu. According to the tests, unconfined
compressive strength after 28 days of all samples exceeded the design criteria of qu ≥ 2.2
MN/m2.

Fig. 4-23 Installation of MIP-columns and placing of geogrids (after Schwarz et al., 2005)
92
The settlement behaviour of tracks was monitored by means of geodetic measurements of the
outer rail of both tracks (see Fig. 4-24). The measurements were conducted in 3 measurement
sections each 20 m in length, consisting of 5 measurement points with spacing 5 m. The
effectiveness of the executed improvement measures was proved by means of settlement
measurements. The measurements show that the track Hamburg-Berlin has settled up 7 mm
in a period of 6 months after reopening the track. This settlement can be considered as small
since usually a settlement of 10 mm to 15 mm will occur, due to compaction of ballast bed,
the protective layer and embankment, even if the soil condition is favourable.

Fig. 4-24 Settlements of field measurements (after Schwarz et al., 2005)

4.3.2.5. Brogborough Lake embankment test

As reported Scottwilson (2009) that Brogborough’s road embankment at A421 M1 Junction
in the UK can be considered for piled embankment case. In this site, subsoil consisting of soft
clay to depths up to 20 m is present. Distribution of undrained shear strength of Brogborough
Lake is shown in Fig. 4-25.

Fig. 4-25 Distribution of undrained shear strength at Brogborough Lake

(after Scottwilson, 2009)
93
Precast driven piles of 275 mm square with 900 mm diameter of the pile cap were installed
with piles spacing of 1.75 m, 2.0 m, and 2.5 m square grid pattern (see Fig. 4-26). These piles
are intended to support the embankment height between 1.5 m and 7 m. Meanwhile, Huesker
Fortrac R-MP with a single layer (in each direction) of high strength low strain Polyvinyl
Alcohol (strain at peak strength ca. 6% was used.

Fig. 4-26 Execution of Brogborough Lake embankments (after Scottwilson, 2009)

Final erection and driving of piles have been executed in mid November 2009. Rod and plat
settlement gauges above some of the pile caps with the aim of monitoring the settlement of
the pile group were installed.

4.4. Summary

Some important findings regarding with experimental works in laboratories and case studies
in the fields are as follows:
 Inclusion of geosynthetics into granular material can reduce horizontal outward
pressure. The more amount of geosynthetic layers increases, lateral pressure will be
significantly decreased. It depends on properties of geosynthetics such as tensile
modulus, tensile strength, and size of aperture. Inclusion of geosynthetics in cohesive
soil exhibits a reduced cohesion of compound material.
 In the field, in case of high embankment, geosynthetics can be applied using multi-
layer ranging from 30 cm to 50 cm between layers. Moreover, reinforcement in soft
soil using geosynthetics cannot overcome excessive settlement.
 Properties of soft soil can be easily recognized such as high water content, low
undrained shear strength, low friction angle, low elasticity modulus and low
compressibility.
 Piled embankment is a promising method to solve problem when constructing
infrastructure over soft soil. In laboratory, it is mostly modelled as end-bearing piles
that refers to shallow soft soil. Meanwhile, there is no a model has been derived for
floating piles.
 Critical height of embankment is an important parameter in which differential
settlement on surface of embankment is equal to zero or very few. This parameter can

94
 be applied to determine minimum thickness of embakment in order to avoid excessive
roughness of surface pavement.
 Differential settlement at surface of pavement can be reduced by utilizing a high
tensile strength of geosynthetics. Particularly, in case of low embakment in which
effective high of embakment is lower than critical height.
 There are various types of piles made of such as concrete pile, wooden pile, stone
column or pier aggregate, stabilized column which can be used in a construction with
or without pile cap. Because of these various types of piles, it will provide a little bit
different final result for a construction.
 In many cases of geosynthetic-reinforced piled embankment, layer of geosynthetics
functionate as a load transfer platform (LTP) which is directly laid down at base of
embankment to transfer a load to piles. Multi-layer of geosynthetics can also be
applied.
 Strains of geosynthetics at the face pile are higher than those at half distance between
piles. Meanwhile strains at half distance between piles are higher than those at center
of four piles.
 In cases of floating piles over soft soil, total settlements at surface of embankment are
high enough and they are still occurred because of creep. Otherwise, differential
settlement at base of embankment is small enough. Furthermore, influence of piles to
reduce total settlement is really obvious.

95
 CHAPTER 5

Modelling and Numerical Analysis

5.1. Analytical Modelling

In normal situation, the dynamic response of road pavements to the moving loads on their
surface can be modeled as a beam, a plate, or the top layer of a layered soil medium. The
foundation soil can be modeled as a system of elastic springs and dashpot or a homogeneous
or layered half-space. The behaviour material of the pavement can be elastic or viscoelastic,
whilst the foundation layers can be elastic, viscoelastic, even inelastic. The loads,
concentrated or distributed of finite extent, may vary with time and move with constant or
variable speed. Some methods can be done by analytical, analytical/numeric and purely
numerical methods, such as finite element and boundary element methods, under conditions
of plane strain or full three-dimensionality.

5.1.1. Two-layered System Elastic Theory

For flexible circular foundation under uniform load, the deflection of a two-layered soil
system had been investigated by several researchers (Burmister, 1943; Meyerhof, 1978;
Huang, 1969). In case axi-symmetrical problem, the basic equation to determine stress
distribution satisfies equilibrium and compatibility relationships. Fig. 5-1 depicts geometry of
two-layered system.

Fig. 5-1 Geometry of two-layered system (after Burmister, 1943)

For a surface load of mIo/mr, the vertical displacement of the surface is given as follow
(Milovic, 1992):

(5-1)

Where: n = E2 (1+1) / E1(1+2)

K = 1-n / [1 + n (3-41)]
L = (3-42) – n (3-41) / [(3-42) + n]
Io= Bessel function of the first kind and order of zero; m = dimensionless parameter; r =
horizontal distance from centerline; h = thickness of the first layer; E1, E2= elastic modulus of
first and second layer; 1, 2= Poisson’s ratio of first and second layer.
96
For the stress and deformation at interface between two layers, Burmister (1943) obtained the
following equations:

(5-2)

(5-3)

(5-4)

If the elastic properties (E and ) are equal in the two layers, the coefficient of K and L are
equal to zero and the above equations reduce to Boussinesq’s equations. The main
assumption in layered elastic theory that the two-layer system is linear elastic and there is no
relative displacement at the interface between two layers (perfectly rough interface).
Based on elastic analysis, Fox (1948) provided a solution to the vertical stress z on the top
of the second layer for a perfectly rough interface and perfectly smooth interface. Fig.5-2
provides the vertical stress on the axis for the case with a/h=1. Here a= radius of the circular
footing, h= thickness of the first layer, d= depth, pz= the vertical pressure on the circular
footing, po= the pressure on the circular footing. The vertical ratio of rough interface is
0.644, 0.292, 0.081 and vertical ratio of smooth interface is 0.722, 0.305, 0.082 for E1/E2= 1,
10, 100 respectively.

Fig. 5-2 Vertical stress distribution at the surface of second layer for two-layered system
(after Fox, 1948)

5.1.2. Analytical Method for Dynamic Response of Beam and Plate on Winkler Type
Elastic Foundation under Moving Loads
First of all, consider an infinitely extended elastic beam-like plate strip (modeling the
pavement) on an elastic foundation (Winkler springs and daspot) under a constant
97
concentrated vertical load P moving with a constant velocity V. The equation of the free
motion of this system is written as (Thompson, 1963):

(5-5)

where w = w(x0,t) is the lateral deflection of plate strip, k and c is the foundation stiffness and
damping, respectively. D=E·h3/12(1-v2) is the flexural rigidity of the plate strip with E and v
being the modulus of elasticity and Poisson's ratio, respectively. Mass density  and the
thickness of plate strip h, while x0 and z are the fixed coordinates along the length of the strip
and in the vertical direction, respectively, whilst t denotes time.
Now, it is introduced a new coordinate system x, z ,which moves with the load P. Thus, one
has
(5-6)
with the new coordinate system, the previous equation becomes

(5-7)
indicating that w in the moving coordinate system is independent of time after all the
transient vibrations have disappeared and the motion of the plate is said to be steady-state.
By introducing some symbols

 

into Eq. 5-7 above, we may rewrite the equation into the form

(5-8)
Therefore, the characteristic equation of the equation is
(5-9)
with a discriminant , which may be positive, negative or zero. For instance, at case  > 0,
after application of boundary conditions (w=dw/dx=0) at x = ± ∞ , compatible of deflection,
slope and bending moment of the plate (pavement) under loading and formulation of the
discontinuity in the shear load. Finally, we can obtain:

cos2 2+ 2±2 / 1/2 (5-10)

where in case of double signs the upper sign correspond to x > 0, while the lower thing to
x<0. In the above  = A1 / , with A1 is the real positive part of the first root.

Now, consider an elastic beam of infinite extent (modeling the pavement) on elastic
foundation (Winkler springs and dashpot) under a vertical distributed (over a finite length)
and time-dependent load p (x0,t) moving with a constant velocity V. The equation of motion
of this system in fixed Cartesian coordinates (x0,z) is (Kim and Roesset, 2003):

(5-11)

98
where v = v (x0,t) is the lateral beam deflection, I the cross-sectional moment of inertia, m the
mass per unit length of the beam, k the foundation stiffness per unit length and c is the
viscous damping constant. By using coordinate system (x,y), which moving load during a
time x = x0 – V·t, we may rewrite the Eq. 5-11 in another form:

(5-12)

When a distributed load p moves harmonically with time, i.e., p(x,t) = p0(x)eit, where i= √-1
and  the operational load frequency, one has that v(x,t) = v0(x) eit , the Eq. 5-12 becomes:

(5-13)
For the general case of time variation of the load p, one can apply onto the double Fourier
transform with respect to x and t and then we obtain:
(5-14)
where:
Finally, consider the three-dimensional extension for an elastic plate of infinite extent on an
elastic foundation (Winkler springs and dashpots) under a vertical distributed (over a finite
surface) and time-dependent load p(x0,y0,t) moving with a constant velocity V along the x-
direction. The equation of motion of this system in fixed Cartesian coordinates (x0,y0,z) is
(Kim and Roesset, 2003):

(5-15)

where D is the flexural rigidity of plate, w is deflection along the z-direction and x0,y is the
middle plane of the plate.

5.1.3. Analytical Method for Dynamic Response of Layered Half-space under Moving Loads
Firstly, consider an elastic layered three-dimensional half-space (whose top layer can model
the pavement) under a concentrated vertical time-dependent load moving on its surface with
constant velocity V (Grundmann et al., 1999). For a homogeneous elastic layer, the equations
of motion may be written:
(5-16)
where ui (i=1,2,3) is the displacement vector,  and  are the Lame elastic constants,  is the
mass density, commas and overdots denote space and time differentiation, respectively. Axes
x and y corresponding to i=1 and i=2 denote the two horizontal directions, while axis z for i=3
denotes the vertical direction. The Eq. 5-16 above can be expressed using the aid of
Helmholtz’s decomposition as:
(5-17)
where the scalar  and vector i functions satisfy the wave equations
(5-18)

with ijk being the alternating tensor and cp = √(l+2)/ and cs=√/ the dilatational and
shear wave velocities.

99
5.1.4. Critical Velocity
The velocity of high-speed train can approach or exceed the characteristic wave velocity of
the dynamic system comprising of the underlying soft ground, trackbed/embankment, and the
moving load. As the train’s velocity approaches some ’critical velocity’ large deformation
can occur. These motions could be dangerous for train and the integrity of the structure. The
velocity of Rayleigh surface in soft sandy soils may be vary low (90-130 m/s) even such
wave velocity in some soft soils in Netherlands can be as low as 29-47 m/s, which is
considerably lower than the railway's proposed design speed, the matter clearly required
investigation.
Kenney (1954) gave some insight into parameters that seem to be of importance to the critical
velocity, vcr, as the equation below:

(5-19)

where k = spring constant per unit length of beam

E = modulus of elasticity of beam
I = moment of inertia of beam
 = mass per unit length of beam
Kinney discusses a point load moving with constant velocity over an Euler-Bernoulli beam
on visco-elastic Winkler medium. In Kenney’s analytical solution, the embankment sub-soil
system has to be simplified as a single beam supported by linear springs.

5.1.5. Explicit Model for Cyclic Accumulation

In the implicit procedure, each cycle is calculated at a rate of  constitutive model. The
accumulation results as a by-product due to the not perfectly closed stress or strain loops.
Elastoplastic multi-surface models (Mroz et al., 1978; Chaboche, 1994), endochronic model
(Valanis&Lee, 1984) or hypoplastic model intergranular strain (Kolymbas, 1991; Gudehus,
1996; Wolffersdorf, 1996; Niemunis&Herle, 1997) can be used. The applicability of the
implicit method is restricted to a low number of cycles (N<50) because with each increment
an accumulation of systematic errors of constitutive model or the integration scheme takes
places (Niemunis, 2005). Even small errors accumulate significantly (e.g. with a factor 106, if
104 cycles are calculated each with 100 increments). Therefore, a constitutive model of an
unreachable perfection would be necessary. Also the large calculation effort sets boundaries
to the application of the implicit method. Wolffersdorff & Schwab (2001) had to restrict their
implicit FE calculation of Watergate ’Uelzen 1’ to less than 25 cycles.
For high-cyclic loading, in general, explicit models are the better choice. Treatment of the
process of accumulation under cyclic loading is similar to a process governed by viscosity.
The number of cycles N replaces the time t. First, two are calculated implicity with strain
increments (see Fig. 5-3) using a rate of  constitutive model. This implicit calculation can
be performed quasi-static or dynamic. During the second cycle in each integration point the
strain loop is recorded as a series of discrete strain points. The recording follows some
predefined criteria. The strain amplitude ampl is determined from this strain loop. The first
cycle is not suitable for the determination of ampl, since the deformation in the first cycle can
significantly differ from those in the subsequent cycles (the first quarter of the first cycle up
to the maximum load is a first loading). The amplitude of the second cycle is more
representative for the amplitudes in the following cycles.

100
 (a) (b)
Fig. 5-3 Cyclic loading (a) Procedure of an explicit calculation of accumulation
(b) Evolution of total strain in a cyclic triaxial test (after Wichtmann, 2005)

The accumulation starting from the second cycle is calculated directly by means of an
equation of the shape:
(5-20)
where: : Jaumann stress rate, D: strain rate, Dacc: given accumulation rate, E: elastic
stiffness, without following the strain path during the particular cycles. The equation leads to
an accumulation of stress (e.g. = -E: Dacc at D=0) and/or strain (D= Dacc at =0). When =0
the strain follows the average accumulation curve  acc(N).
The strain amplitude  ampl is assumed constant for the explicit calculation. The explicit
calculation may be interrupted after definite numbers of cycles and  ampl can be updated in
the an implicit so-called control cycle.

5.1.5.1. Model of Sawicki & Swidzinski

Sawicki and Swidzinski (1989) basing on experimental results from their simple shear
device, proposed for the cyclic accumulation a purely volumetric accumulation rule
(5-21)
in which the volumetric strain, ϵacc v, was described by a so-called ’universal densification
curve’
(5-22)
with the state “compaction”  =  n /n0 , with n = porosity, the number of cycles weighted by
the amplitude

(5-23)
and the material constant C1 and C2. The tensor ampl contains the amplitudes of the particular
strain component. The latter transformation in Eq. (5-23) is valid for the case of cyclic simple
shear tests with a constant shear strain amplitude ampl. The compaction rate  is
obtained from Eq. 5-24:

(5-24)

101
The model of Sawicki & Swidzinski was performed on tests with a relatively low number of
cycles (N < 103). A major drawback of the model is that only the volumetric accumulation,
whereas the deviatoric one not be considered.

5.1.5.2. Model of Bouckovalas et al.

The model of Bouckovalas et al. (1984) presents both the accumulation of volumetric and the
accumulation of deviatoric strains as depicted the equation below:
(5-25)
(5-26)
The value of average stress ratio av can be positive or negative, whilst A, B, a, b and c are
material constants. The dependence of the accumulation rates on the historiotropy is
expressed:
(5-27)
For cycles with a constant shear strain amplitude, the rates are proportional to N c(c+1). The
parameter f is stress-dependent. It takes the value 1 on the p-axis and is zero on the critical-
state line.

The model of Bouckovalas et al. (1984) predicts the cyclic flow rule correctly: vacc vanishes
on the critical state line, while qacc becomes zero at av = 0. A power law is used for the
dependence of the accumulation rates on the number of cycles. In the model, a constant a = 3
and c = -1.5 are chosen and factor f remains vague in the mathematical definition. The model
uses a state variable for the historiotropy which considers also the amplitude of the cycles.
The model gives a too strong amplitude-dependence with approximate value  acc ~ (ampl)3.

5.2. Consitutive Models in Numerical Analysis

5.2.1. Mohr Coulomb Model

Plasticity has a relationship which strain is irreversible. To evaluate the plasticity in a
calculation, a yield function f, can be used as a function between stress and strain. Generally,
a plastic perfectly model is a constitutive model with a certain yield surface, a yield surface
defined by model parameters and not influenced by plastic strain. For the stresses under yield
surface, they behave fully elastic and strains are reversible.

5.2.1.1. Elastic Perfectly Plastic Behaviour

A basic principle of elastic plasticity model is that strain and strain rate are devided by elastic
and plastic parts:
(5-28)
Hooke law is used to correlate the elastic stress rate and strain rate. Substituting the equation
above into Hooke‘s law gives:
(5-29)
According to classical plasticity theory (Hill, 1958), a plastic strain rate is proportional to a
derivative of stress yield function. It means that the plastic strain rate may be expressed as a
vector perpendicularly with the yield surface (see Fig. 5-4). The classical form of plasticity

102
theory is so-called as associated plasticity. But, for Mohr-Coulomb yield function, this theory
will give overestimate dilatancy prediction. Therefore, besides yield function f, it is used a
plastic potential function g. At case g ≠ f, it is so-called as non-associated plasticity.
Generally, strain rate can be written as:
(5-30)

where  is a plastic multiplied factor. For pure elastic,  is equal to zero and for plastic
behaviour,  is a positive value.

(5-31)

(5-32)

Fig. 5-4 Basic principle of elastic perfectly plasticity (after Hill, 1958)

The equations can be used to obtain a relationship between effective stress rate and strain rate
for elastoplastic model (Smith & Griffith, 1982; Vermeer, 1982):

(5-33)

with:

Parameter  can be used as a switch. If material behaviour is elastic,  value is equal to zero,
whilst  value is one when material behaviour is plastic. Plastic theory above is limited only
for a continuous and smooth yield surface, and not included into a multi-surface of yield likes
MC model. For multi-yield surface, Koiter (1960) took into consider to flow vertices
involving two or more plastic potential functions:
(5-34)

5.2.1.2. Formulation of Mohr-Coulomb Model

The yield condition of Mohr-Coulomb encompasses six yield functions when formulated in
context of main stress (Smith & Griffin, 1982):
(5-35a)
(5-35b)
(5-35c)

103
 (5-35d)
(5-35e)
(5-35f)
two parameters of plastic model emerge into yield function are the friction angle  and
cohesion c well known in geotechnical engineering. These yield functions together will form
a hexagonal cone in main stress space as shown in Fig. 5-5.

Fig. 5-5 Mohr-Coulomb yield surface in main stress space (c=0) (after Smith et al., 1982)

Besides yield functions, there are six plastic potential functions for MC model:
(5-36a)
(5-36b)
(5-36c)
(5-36d)
(5-36e)
(5-36f)
Plastic potential function has a third plastic parameter, namely dilation angle . This
parameter is needed to model an increment of plastic volumetric strain positively as it
actually occurs on stiff soil.
For c > 0, Mohr-Coulomb criteria allow to tensile stress, but in fact, soil only reminds this
stress very small, even nothing. Plaxis software can model by using constraint of tensile
stress. This constraint incorporates with three additional yield function:
(5-37a)
(5-37b)
(5-37c)
The allowable tensile stress t can be setup (default) as zero. For three yield functions, these
can be used with an associated flow rule. The Mohr-Coulomb model needs five parameters.
Generally, these parameters can be obtained from laboratory tests as follows:
E : Young‘s modulus [kN/m2]
104
 v : Poisson's ratio [-]
 : friction angle [o]
c : cohesion [kN/m2]
 : dilation angle [o]

5.2.2. The Hardening Soil (HS) Model

In this model the total strains are calculated using a stress-dependent stiffness, different for
both virgin loading and un-/reloading. The plastic strains are calculated by introducing a
multi-surface yield criterion. Hardening is assumed to be isotropic depending on both the
plastic shear and volumetric strain. For the frictional hardening a non-associated and for the
cap hardening an associated flow rule is assumed. First the model is written in its rate form.
Therefore the essential equations for the stiffness modules, the yield, failure and plastic
potential surfaces are given.
In contrast to an elastic perfectly-plastic model, the yield surface of the Hardening Soil model
is not fixed in the principal stress space, but it can expand due to plastic straining. Distinction
is made between two main types of hardening, namely shear hardening and compression
hardening. Shear hardening is used to model irreversible strains due to primary deviatoric
loading, whereas compression hardening is used to model irreversible plastic strains due to
primary compression in oedometer loading and isotropic loading.

5.2.2.1. Constitutive Equations for Standard Drained Triaxial Test

A basic idea for the formulation of the Hardening Soil model is the hyperbolic relationship
between the vertical strain 1 and the deviatoric stress q in primary triaxial loading (see Fig.
5-6). In case a drained triaxial test, the observed relationship between the axial strain and the
deviatoric stress can be well approximated by a hyperbola (Kondner&Zelasko, 1963).
Standard drained triaxial test tends to yield curves that can be described by:

(5-38)

The ultimate deviatoric stress qf and quantity qa in Eq.5-38 are defined as:

(5-39)

The above relationship for q is derived from the Mohr-Coulomb failure criteria, which
involves the strength parameters c and p. As soon as q = qf , the failure criterion is satisfied
and perfectly plastic yielding occurs. The ratio between qf and qa is given by the failure ratio
Rf , which should obviously be smaller than 1. Rf = 0,9 often is a suitable default setting.

Fig. 5-6 Hyperbolic stress-strain relation in primary loading

for a standard drained triaxial test (after Schanz et al.,1999)
105
5.2.2.2. Stiffness for Primary Loading
The stress strain behaviour of primary loading is highly nonlinear. The parameter E50 is the
confining stress dependent stiffness modulus for primary loading. E50 is used instead of the
initial modulus Ei for small strain which, as a tangent modulus, is more difficult to determine
experimentally. It is given by equation:

(5-40)

E50ref is a reference stiffness modulus corresponding to the reference stress pref. The actual
stiffness depends on the minor principal stress 3', which is the effective confining pressure
in a triaxial test. The amount of stress dependency is given by the power m. Von Soos (2001)
showed that m values range between 0.4 and 1.0. In order to simulate a logarithmic stress
dependency, as observed for soft clay, the power should be taken equal to 1.0 and for m value
sands is 0.5. As a secant modulus E50ref is determined from a triaxial stress-strain-curve for a
mobilization of 50% of the maximum shear strength qf.

5.2.2.3. Stiffness for un-/reloading

The stress paths for unloading and reloading, another stress-dependent stiffness modulus is
used:

(5-41)

where Eurref is the reference Young's modulus for unloading and reloading, corresponding to
the reference pressure ref. The un-/reloading path is modeled as purely (non-linear) elastic.
The elastic components of strain e are calculated according to a Hookean type of elastic
relation.
(5-42)

For drained triaxial test stress paths with 2 = 3 = constant, the elastic Young's modulus Eur
remains constant and the elastic strain are given by equations:
(5-43)

5.2.2.4. Yield Surface, Failure Conditions, Hardening Law

For the triaxial case, the two yield functions f12 and f13 are defined according to Eq. 5-38 and
5-39. Here the measure of the plastic shear strain p is used as the relevant parameter for the
frictional hardening.

(5-44)

(5-45)

with the definition:

(5-46)
In reality, plastic volumetric strains vp will never precisely equal to zero, but for hard soils
plastic volume changes tend to be small when compared with the axial strain, so that the
approximation in Eq. 5-46 will generally be accurate. For a given constant value of the
106
hardening parameter, the yield condition f12 = f13 = 0 can be visualized in p'-q-plane by means
of a yield locus. For m = 1.0 straight line is obtained, but for slightly curved yield loci
correspond to lower of the exponent. Fig. 5-7 shows the shape of successive yield loci for m
= 0.5, being typical for hard soils.

Fig. 5-7 Successive yield loci for various values of the hardening
parameter p and failure surface (after Schanz et al., 1999)

In contrast to elastic perfectly Mohr-Coulomb (MC) model, in the Hard Soil (HS) model,
plastic strains may occur before the limit MC-failure stress reached. The HS model
incorporates two other yield surfaces, which are not fixed in principal stress space, but they
may expand and soil hardening is simulated due to plastic straining. As illustrated in Fig. 5-
8, distinction is made between two types of hardening, which are shear hardening and
compression hardening. For the shear hardening law, a yield function fs is introduced, which
is a function of the triaxial loading stiffness E50 and for the compression hardening a yield
function fc is formulated.

Fig. 5-8 Yield surface of HS model, a) Successive yield loci for shear hardening and
compression hardeing in p-q space, b) Total yield contour in principal stress space
(after Schanz et al., 1999)

5.2.2.5. Flow Rule, Plastic Potential Functions

As for all plasticity model, the HS model involves a relationship between rates of plastic
strain, i.e. a relationship between vp and p . This flow rule has the linear form:
(5-47)
The expression of the mobilized dilatancy angle m is:
(5-48)

107
where  cv is the critical state friction angle, being a material constant independent of density
(Schanz & Vermeer, 1996), and m is the mobilized friction angle:
(5-49)

The above equations correspond to the well-known stress-dilatancy theory (Rowe 1962,
Rowe, 1971). The essential property of the stress-dilatancy theory is that the material
contracts for small stress ratio m < cv, whilst dilatancy occurs for high stress ratio m < cv.
At failure, when the mobilized friction angle is equal to the failure angle, p, it is found the
previous equation, that:

(5-50)

Hence, the critical state angle can be computed from failure angle p and p . The definition
of flow rule is equivalent to the definition of the plastic potential functions g12 and g13
according to:
(5-51a)

(5-51b)
By using the Koiter-rule (Koiter 1960) for yielding depending on two yield surface (Multi-
surface plasticity) one finds:

(5-52)

5.2.2.6. Parameters of the HS model

Some parameters of the present hardening model coincide with those of classical non-
hardening Mohr-Coulomb model. These are the failure parameters p, c and p. Additionally
we used the basic parameters for the soil stiffness:
E50ref, secant stiffness in standard drained triaxial test
Eoedref, tangent stiffness for primary oedometer loading
m, power for stress-level dependency of stiffness
This set of parameters is completed by the following advanced parameters:
Eurref, unloading/reloading stiffness
vur, Poisson's ratio for unloading/reloading
pref, reference stress for stiffness
KoNC, Ko-value for normal consolidation
Rf, failure ratio qf / qa
Experimental data on m, E50 and Eoed for granular soils is given in (Schanz & Vermeer,
1998).

5.2.3. The Hardening Soil Small (HS Small) Model

The Hardening Soil Small (HS-Small) model is an extension of the HS model to incorporate
the small strain stiffness behaviour of soils. The behaviour of soil at small strains has been
studied by some researchers (Seed&Idriss, 1970; Burland, 1989; Atkinson, 2000; Benz
108
2007). At small strain levels, most soil exhibit a higher stiffness than at strain levels. The soil
stiffness decays with increasing strain. Figure 5-9 shows a small unloading-reloading stress-
strain paths result in a considerably higher elastic modulus Eo.

Fig. 5-9 HS-small model, a) Initial stiffness modulus E0 in a triaxial test

b) Small strain parameters E0 and  0.7 (after Benz, 2007)

In fact, maximum soil stiffness is observed at very low strain levels, e.g. Strains smaller than
10-5 (Atkinson and Sällfors, 1991). Special devices is needed to identify stiffness at very
small strain. Biarez and Hicher (1994) gave a simple correlation for quarts sand as follows:

(5-53)

where e is the void ratio of the soil and p is the mean stress. Alphan (1970) also estimated
preliminary estimation of the E0. Hardin and Drnevich (1972) formulated the decay of
stiffness when strains increase.
(5-54)

where E is the actual secant modulus at the corresponding shear strain  , E0 is the initial
stiffness of soil and  0.7 is the shear strain at 70 percent from the initial stiffness E0 .
The initial shear modulus G0 is determined from the relationship between E0 and Poisson's
ratio v as in the following equation
(5-55)

Whilst, the shear strain is expressed using the strain invariant

(5-56)
The stiffness degradation due to plastic straining is modelled by involving material
hardening. Therefore, before reaching plastic material behaviour, the formulation of small
strain stiffness is cut off at the unloading-reloading stiffness Eur. There are two additional
input parameters are required for the HS-Small model: the elastic small-strain shear modulus
G0ref at reference pressure pref and the curve-decay value  0.7 in primary loading.

5.2.4. The Soft Soil Creep Model

Soft Soil Creep model distinguishes between primary loading and unloading/reloading
behaviour and in this respect the model is similar to the Hardening Soil model. In the HS
model, there is no time dependency in the model. The cap expandeds instaneously if an
increase in the load would cause the stress state to fall outside the current cap. In the Soft Soil
109
Creep (SSC) model, this shift of the cap needs time. If a higher load is applied, the cap will
not follow immediately, but it will take a day to adapt to the new stress state.
The SSC model is suitable for estimating the viscous effects, i.e. creep and stress relaxation.
In fact, all soils exhibit some creep and primary compression is more often than not followed
by a certain amount of secondary compression. Buisman (1936) proposed the following
equation to describe creep behaviour under constant effective stress.

(5-57)

where c is the strain up to the end of consolidation, t is the time measured from the
beginning of loading, tc is the time to end of primary consolidation and CB is a material
constant (see Fig. 5-10). Rewriting above the equation as:

(5-58)

where t ' = t - tc being the effective creep time.

Fig. 5-10 Consolidation and creep behaviour in a standard Oedometer test

(After Vermeer and Neher, 1999)

Garlanger (1972) proposed the creep equation as follows:

(5-59)

where: C = CB (1 + eo)
Butterfield (1979) alsodescribed a slightly different possibility for secondary compression
as:

(5-60)

where  is the logarithmic strain defined as:

(5-61)

Originally, Hencky (1928) used the subscript '0' denoting initial values whilst the superscript
'H' for denotating logaritmic strain. For small strains it is possible to show that:
(5-62)
110
1D Creep model
The SSC model described here is on the basis of the work done by Vermeer and Neher
(1999) on elastic visco-plastic creep model. Vermeer and Neher (1999) adopted Bjerrum's
idea to find an analytical expression for quantity of c. For the one dimensional creep, two
strain components need to be modelled as shown in Fig. 5-11.

Fig. 5-11 Standard Oedometer Test, a) Stepwise loading in e-log vs. ’ plot
b) Void ratio vs. Time (After Vermeer and Neher, 1999)

First of them is the more or less elastic deformation, as directly observed in unloading and
reloading condition. The other component of strain is irreversible and time dependent.
Volumetric strain implies a change of void ratio and it is convenient to formulate the
deformation in terms of void ratio e, and the change of void can be expressed by a equation
below:
(5-63)
where the superscripts e and c refer to the elastic and creep component respectively. The
elastic change of void ratio is formulated as follows:
(5-64)
where Cs is the swelling index, which can be the unloading-reloading index Cur. Whilst the
creep deformation is represented using power law

(5-65)

where  is a particular reference time, which can mostly be taken equal to one day. Cis the
secondary compression index that is also referred to as the creep index and Cc is the
compression index obtained from an oedometer test. From the equation above that the creep
rate depends on the OCR value. Some typical soil data give Cs= Cc/10 and C = Cc/30. This
will give  of about 27.
The preconsolidation stress p increases during creep according to the differential equation

(5-66)

By integrating the equation (5-66), we will get the preconsolidation stress:

(5-67)
111
where p0 is the initial preconsolidation stress for ec = e0c. Subscript ‘0’ denotes the initial
value. Finally, the creep rate formulation can be obtained as the equation below:

(5-68)

The effective stress ’ may be either larger or smaller than p0. For a special case of a
constant effective stress, the differential creep formulation can be integrated analytically to
obtain the logarithmic creep law

(5-69)
where t = 0 for e = e0, and then

(5-70)

3D Soft Soil Creep model

On extending the 1-D model to general states of stress and strain, the well-known stress
invariants p and q for mean and deviatoric stress are adopted. These invariants are used to
define a new stress measure named peq, namely:

(5-71)

with
Fig. 5-12 shows that the stress measure peq is constant on the ellipse in the p-q plane. In fact,
the ellipses are from the Modified Camclay Model as introduced by Roscoe and Burland
(1968).

Fig. 5-12 Diagram of peq-ellipse in a p-q plane (after Vermeer and Neher, 1999)

Soil parameter M represents the slope of the so-called ’critical state line’ and it is defined as:
(5-72)

where cv is the constant volume friction angle, also referred to as critical-state friction angle.
The preconsolidation pressure changes during creep according to the law

(5-73)

where * and * are a modified compression index and a modified swelling index
respectively. In case of small strain, it gives

112
 (5-74)

and

(5-75)

Volumetric creep strain rate define as

(5-76)

Whilst the elastic volumetric strain rate can be expressed as

(5-77)

where bulk modulus Kur = p'/*

The total volumetric strain rate in 3D Soft Soil Creep model can be written as

(5-78)

Parameter used in Soft Soil Creep model has a relationship with Camclay parameters model.
The modified compression index * is  / (1+e) , the modified swelling index * is  / (1+e) ,
and the modified creep index * is C / 2.3(1+e). The ratio of the unloading/primary loading
stress, * / *, cannot be smaller than 1 and should normally be between 2 and 10. Users
should be very wary of values outside this range' for most practical cases the value falls
within the range of 3 to 7. Secondly, there is the creep ratio, (*-*)/*, to consider. This
ratio can have a wide range of values, normally between 5 and 25, where high values
represent stiff soil with little creep and small values represent soft soils with a considerable
amount of creep. For most practical cases the ratio falls within the range of 10 to 20 and if the
creep ratio is over 25 one could reconsider the use of the creep model.
Table 5-1 Material parameters for the Soft Soil Creep model
(after Vermeer and Neher, 1999)
Symbols Name of parameters
* a modified compression index
* = * / 2-10 a modified swelling index
* = (*-*) / 10-20 a modified creep index
vur = 0.15 un-/reloading Poisson's ratio
c' effective cohesion
 ' effective friction angle
K0NC height of normal consolidation surface
OCR the state of preconsolidation stress

5.2.5. The Hypoplastic Model

Hypoplasticity is a particular class of incrementally non-linear constitutive models. The basic
structure of the hypoplastic models has been developed during 1990's at the University of
Karlsruhe. It is unlike in elasto-plasticity, in hypoplasticity, the strain rate is not decomposed
into elastic and plastic parts, and the models do not use explicitly the notions of the yield
surface and plastic potential surface. But the models are still capable in predicting an
113
important feature of soil behaviour. This is achieved by the hypoplastic equation being non-
linear in the strecthing tensor D. The basic equation may be expressed as:
ℒ (5-79)
where is the objective (Jaumann) stress rate, D is the Euler's stretching tensor and ℒ and N
are fourth- and second order constitutive tensors, respectively. The early hypoplastic model
was developed by trial and error (Kolymbas, 1991). Gudehus (1996) proposed a modification
to include the stress level (barotropy) and density (pyknotropy), as the modified equation
below:
ℒ (5-80)
with fs and fd are scalar factors expressing the influence of barotropy and pyknotropy.
Hereafter, Wolffersdorf (1996) refined it with incorporating Matsuoka-Nakai critical state
stress condition. Nowadays, this model is considered as a standard hypoplastic model for
granular materials.

5.2.5.1. Hypoplastic Model for Granular Materials

The hypoplastic model for granular material has eight material parameters, consisting of c,
hs, n, ed0, ec0, ei0,  and . The critical state friction angle c can be obtained directly by the
measurement of the angle of repose. Two parameters hs and n can be directly computed from
oedometric loading curves. The parameter n controls the curvature of oedometer curve, whilst
hs controls the slope of oedometric curve (see Fig. 5-13).

Fig. 5-13 Influence of n (a) and hs (b) on Oedometric curves (after Herle and Gudehus, 1999)

The parameter n can be written using an equation:

(5-81)

where mean stress ps1 and ps2 may be calculated from axial stresses using the Jaky formula K0
= 1 – sin c, and ep1 and ep2 are the void ratios corresponding to the stress ps1 and ps2. Tangent
compression indices corresponding to the limit values of interval ps1 and ps2 (Cc1 and Cc2) can
be approximated by secant moduli between loading steps preceding and following steps ps1
and ps2. The parameter hs can be expressed as:

(5-82)

where Cc is a secant compression index calculated from limiting values of the calibration
interval ps1 and ps2, in which ps and ep are averages of the limit values of p and e in this range.
Further model parameters are the reference void ratio ed0, ec0 and ei0, corresponding to the
densest, critical state and loosest particle packing at the zero mean stress. Bauer (1996)
114
formulated the reference void ratio ed, ec and ei corresponding to the non-zero stress
depending on the mean stress, as the Eq. 5-83 and then shown in Fig. 5-14.

(5-83)

Fig. 5-14 Dependency of the reference void ratio of ed0, ec0 and ei0 on the mean stress
(after Herle and Gudehus, 1999)

The initial void ratio emax can be considered equal to the critical state void ratio at zero
presure ec0. Void ratio ed0 and ei0 can approximately be obtained from empirical relations.
The physical meaning of ed0 is the void ratio at maximum density, void ratio ei0 represents
intercept of the isotropic normal compression line with p = 0 axis. The ratio ei0 /ec0 ≈ 1.2 was
derived from considering skeleton consisting of ideal spherical particles (Herle and Gudehus,
1999). Whereas the minimum void ratio ed0 should be obtained by densification of a granular
material by means of cyclic shearing with small amplitude under constant pressure. If there
are no data, it can be estimated using an empirical relation ed0 /ec0 ≈ 0.4.
The last two parameters  and  should be calibrated by means of single-element simulations
of the drained triaxial tests. Parameter  controls independently the shear stiffness and 
controls the peak friction angle.

5.2.5.2. Intergranular Strain Concept (Small Strain Behaviour)

The hypoplastic models can predict successfully the soil behaviour in the medium to large
strain range. However, in the small strain range and upon cyclic loading, they fail in
predicting the high quasi-elastic soil stiffness. To overcome this problem, Niemunis and
Herle (1997) proposed an extension of the hypoplastic equation considering additional state
variable “intergranular strain” determine the direction of the previous loading. This
modification, often denoted as the “intergranular strain concept”, can be used for both the
model for granular materials and the model for clay.
The rate formulation of the enhanced model is given by the equation:
(5-84)
where Mis the fourth-order tangent stiffness tensor of the material. Strain can be thought of
as the sum of a component related to the deformation of interface layers at intergranular
contacts, quantified by the intergranular strain tensor ; and a component related to the
rearrangement of the soil skeleton.
Intergranular strain concept requires five additional parameters, R controlling the size of the
elastic range, r and  controlling the rate of stiffness degradation, mR controlling the initial
115
shear stiffness for the initial and reverse loading conditions and mT controlling the stiffness
upon neutral loading conditions.

5.3. Geometrical Idealization

There are some geometrical idealizations used for the FE-calculations such as axisymmetric,
plane strain according to Bergado and Long (1994), plane strain with equivalent stiffness and
3D geometry (Fig. 5-15).

Fig. 5-15 Geometrical idealizations of one cell piled embankment (a) Axisymmetric
idealization (b) Idealization for plane strain after Bergado (c) Idealization plane strain using
equivalent stiffness (d) 3D geometry (after Satibi, 2009)

For Axisymmetric, the three dimensional one cell of the piled embankment is transformed in
to a circular cell using the area of the pile and the soil the same. Fig. 5-15(a) shows the
transformation of the squared cell to a circular cell which one radian of the circular piled
embankment cell is used.
Plane strain after Bergado: The three dimensional grid of piles can be transformed into
continuous walls with an equivalent thickness teq in plane strain model as indicated in Fig. 5-
15(b). By keeping an improved area ratio (Ac/AE) is constant, the thickness of the continuous

116
wall is calculated based on the consideration of this ratio. Due to symmetrical geometry, only
a half of the plane strain geometry is used for the FE calculations.
(5-85)

Plane strain with equivalent pile stiffness: An alternative method to transform the three
dimensional grid of piles into a continuous wall in plane strain condition is by assuming
equivalent wall stiffness. The equivalent stiffness of the wall Eeq is taken as the proportional
average of the pile and soil stiffness. Hence,

(5-86)

where Ec, Es and Aw are pile stiffness, soft soil stiffness and wall area as illustrated in Fig.5-
15(c). It is worthy to note that when using this approach, the improved area ratio becomes
larger.
3D geometry: Three dimensional dimension of the actual case can be best analyzed using 3D
geometry. Here, half of the piled embankment cell is considered the 3D FE analysis as
described in Fig. 5-15(d).
According to Irsyam et al. (2008) he modeled bamboo piles in the Bamboo piles-Mattress
system as soil reinforcement for embankment on soft clay, and results of FE-analysis was in a
good agreement with field measurement. To model pile he applies ‘node-to-node anchor’ in
which it was commonly used and available in Plaxis. By modelling pile using elasto-plastic
‘node-to-node anchor’, soil in between wooden piles is not confined and then actual vertical
pile capacity can be obtained. The stiffness of elasto-plastic ‘node-to-node anchor’ is taken
from the estimated displacement of wooden piles in the vertical direction.

5.4. Summary

To explain distribution of stress, strain due to a load on surface of pavement, it can be

grouped into two types. First point of view is static load and another is dynamic one.
Analitical approach can be used to solve simple problem. Then, a numerical analaysis
provides more advance models that offers a good solution for modelling behaviour of
material. Some drawbacks and advantages of these approaches as a consideration in
analyzing a problem can be explained as follows:
 For a simple case which layers behave elastic material, two layered system can be
used which has two homogeneous material consisting of pavement and subgrade
subjected to a circular load. Hence, elasticity modulus and Poisson’ratio are important
parameters.
 Dynamic response of pavement and subgrade under moving load is derived from an
equation of free motion which pavement modeled as a plate and subgrade as an elastic
foundation.
 Motion of a load will emerge large deformation on a structure integrity and could be
dangerous when approaching or higher than wave velocity of soil foundation.
 In implicit procedure, each cycle is calculated at a rate of  constitutive model.
This model is to estimate strain due to stress and is restricted to a low number of
cycles (N<50).
117
 For high cyclic loading, in general, explicit models are the better choice.
Accumulation process under cyclic loading is similar to a process that is governed by
viscosity. The number of cyclic N replaces the time t.
 Constitutive models in numerical analysis provide some models that are able to model
the behaviour of material in which it behaves not simply elastic material . Plasticity
has a  relationship which can be non-linear and strain is irreversible.
 When undertaking finite element analysis, idealization of geometry is needed. Some
geometrical idealization, such as axisymmetric, plane strain, and 3-D geometry, can
be applied in dealing with several cases of geometry in laboratory or in the field.
Particularly, it would be important part, when available software for analysis is
limited merely two dimensional package software.

118
 CHAPTER 6

Discussion Using Finite Element Calculations

6.1. Introduction
This chapter discusses some findings resulted from some experimental works and several
field case studies. Most of the results are directly obtained from observations through devices
installed both in the laboratories and in the field. By using FE-calculations, the results are
then verified with the main aim to do comparisons in order to be able to properly understand
the behavior of material.

6.2. Optimal Vertical Distance of Geosynthetics

Granular materials in pavement engineering such as sub-base course and/or base course play
an important role in bearing a load on it and then distribute downward to surface of subgrade.
When thickness of base course and/or sub-base course is deep enough, inserting
geosynthetics is possible to enhance the bearing capacity and simultaneously reduce the
horizontal thrust.
With reference to Ruiken’s work procedure in laboratory (Ruiken&Ziegler, 2008), now we
use axisymmetrical approach of FE-analysis. It will be presented the results for the purpose
of the effectiveness of vertical distance of geosynthetics to support a load. Fig. 6-1 explains
the work procedure used in FE-analysis.

Constant Speed of 1 mm/min

( ± 0.1 % per minute )
d
Geogrid n+1
d
H = 110 cm Geogrid n
d
Geogrid n-1
d

D = 50 cm
Fig. 6-1 Schematic work procedure for FE analysis

Inclusion of geosynthetics in granular material will enhance bearing capacity. By inserting

geosynthetics in tubular sample soil, it will increase the confining pressure of material and
results in smaller horizontal strain. In other words, it means that the soil stiffness will be
higher and sample soil has eventually higher bearing capacity.
To prove this hypothesis, FE-analysis using Plaxis software with axisymmetric configuration
is presented. Tubular sample soil with 110 cm high and 50 cm in diameter is subjected to a
vertical compressive load with a constant speed of 1 mm/minute. In this model, number of
geosynthetics layer used is up to 4 layers with two types of tensile strength 30 kN/m and 300
kN/m respectively. Tensile strength of 30 kN/m represents a low tensile strength, whilst
tensile strength of 300 kN/m represents a moderate tensile strength.

119
Strength of granular material for the base course or subbase course can be expressed by CBR
(California Bearing Ratio) or resilient modulus (Mr). Table 6-1 provides a correlation
between CBR and Mr using Eq. 3-2 through Eq. 3-5, as explained in Chapter 3.
Table 6-1 Resilient moduli of base course and subbase course
Resilient moduli (Mr)
Source(s) Equations
CBR 40 CBR100 CBR150
Heukelom&Klomp Mr (psi) = 1500 CBR 60 ksi 150 ksi 225 ksi
(1962) Mr (MPa) = 10.34 CBR 413 MPa 1034 MPa 1551MPa
Webb&Campbell Mr (psi) = 3116 CBR 0.4779707 18 ksi 28 ksi 34 ksi
(1986) Mr (MPa) = 21.485 CBR 0.4779707 125 MPa 194 MPa 235 MPa
Sasongko (1996) Mr (MPa) = 10 CBR 400 MPa 1000 MPa 1500 MPa
Mr (MPa) = 4 CBR 160 MPa 400 MPa 600 MPa
NCHRP (2004) Mr (psi) = 2555 CBR 0.64 27 ksi 48 ksi 63 ksi
Mr (MPa) = 17.6 CBR 0.64 186 MPa 335 MPa 435 MPa

According to Sasongko (1996), resilient modulus correlates to 4 times CBR value when
CBR>28%. This is quite close with NCHRP’s equation. Whilst Heukelom’s equation
extensively developed for fine grained soil with maximum CBR of 10% provides too high
estimation, whereas Webb’s equation gives too small estimation. Based on data in Table 6.1,
the appropriate values of subbase course with CBR of 40% correlates to the resilient modulus
between 160 MPa and 190 MPa. Meanwhile, base course with CBR of 100% correlates to the
resilient modulus between 335 MPa and 400 MPa and base course very dense with CBR of
150% correlates to the modulus resilient between 435 MPa and 600 MPa.
In this FE analysis, three resilient moduli are used for representing resilient moduli of
pavement materials, namely 200 MPa, 400 MPa and 600 MPa respectively. Mohr-Coulomb
(MC) and the Hardening Soil (HS) model are used for analyzing this work. Besides elastic
modulus represented by resilient modulus, other parameters are also introduced for FA-
analysis, namely friction angle () and dilation angle (ψ). Dilation angle is an important
parameter because pavement material is a compacted soil. Material properties are shown in
Table 6-2.
Table 6-2 Material properties used in finite element analysis
Element and model
Properties Unit Granular material Geogrid
MC HS Elastic
Unit weight kN/m3 18 18 -
Friction angle [o] 38, 45 38, 45 -
Dilation angle [o] 8, 15 8, 15 -
Cohesion, c kPa 0.5 0.5 -
Young’s modulus, E MPa 200, 600 - -
Secant modulus MPa - 50, 100 , 150 -
Oedometer comp. modulusoed MPa - 50, 100 , 150 -
Un-/reloading modulusur MPa - 200, 400, 600 -
Poisson’s ratio, vur /v [-] -/0.3 0.2/- -
Tensile strength kN/m - - 30, 300
Drainage Drained Drained Non-porous
Calculation type Stage construction, plastic calculation

Pictures in the Table 6-3 through Table 6-6 show the patterns of failure which indicate zones
having the maximum horizontal displacements (red colour) when being subjected to a load
120
using MC-model for the elastic modulus of 200 MPa and 600 MPa. In Table 6-3 and Table 6-
4, some findings are resulted from FE-analysis using MC model, which the elastic modulus
of granular soil of 200 MPa and 600 MPa with tensile strength of geosynthetics of 30 kN/m.
It is clearly shown that granular material with high friction angle of 45o and dilation angle of
15o is able to support higher stress (around 37 kPa) than lower friction angle and dilation
angle, 38o and 8o respectively, which can support the vertical stress around 34 kPa. Moreover,
by using material with a higher friction angle and dilation angle can reduce horizontal
displacement smaller than lower one. Another important finding is that the number of
inclusion of geosynthetics in granular material contributes to reduce horizontal displacement.

Table 6-3 Horizontal displacements for MC-model, E=200 MPa and EA= 30 kN/m
Elastic modulus MC-model for E = 200 MPa , EA= 30 kN/m
Friction angle and
dilatant angle  = 45 ,  = 15o
o
 = 38o,  = 8o
Number of
geosynthtetics layer
0 1 2 3 0 1 2 3

Patterns of horizontal
displacements

Remarks More More More More More More More More

load step load step load step load step load step load step load step load step
Horizontal displacement 0.0697 0.0541 0.0527 0.0557 0.1118 0.1071 0.1065 0.1061
(mm)
Vertical displacement 0.0697 0.0700 0.0686 0.0713 0.1497 0.142 0.1529 0.1482
(mm)
Force (kN) 0.819 0.82 0.819 0.825 0.77 0.777 0.775 0.775
Stress (kPa) 37.27 37.381 37.296 37.774 34.3 34.711 34.472 34.668

Table 6-4 Horizontal displacements for MC-model, E=600 MPa and EA= 30 kN/m
Elastic modulus MC-model for E = 600 MPa , EA= 30 kN/m
Friction angle and
dilatant angle  = 45 ,  = 15o
o
 = 38o,  = 8o
Number of
geosynthtetics layer
0 1 2 3 0 1 2 3

Patterns of horizontal
displacements

Remarks More More More More More More More More

load step load step load step load step load step load step load step load step
Horizontal displacement 0.0408 0.0405 0.0415 0.0390 0.0396 0.0418 0.0417 0.0431
(mm)
Vertical displacement 0.0498 0.0499 0.0525 0.0499 0.0543 0.0565 0.0563 0.0580
(mm)
Force (kN) 0.827 0.827 0.827 0.825 0.775 0.777 0.783 0.784
Stress (kPa) 37.91 37.889 37.951 37.794 33.628 33.882 34.304 34.564

121
Similar trend in previous findings, by inclusion of higher tensile strength of 300 kN/m in
granular material with elastic modulus of 200 kPa, it is able to support the vertical stress
enormously, particularly when granular material has the higher friction angle and dilation
angle, as indicated in Table 6-5. However, inclusion tensile strength of 300 kN/m in granular
material with an elastic modulus of 600 MPa only contributes slightly vertical stress, as
shown in Table 6-6.
Table 6-5 Horizontal displacements for MC-model, E=200 MPa and EA= 300 kN/m
Elastic modulus MC-model for E = 200 MPa , EA= 300 kN/m
Friction angle and
dilation angle  = 45 ,  = 15o
o
 = 38o,  = 8o
Number of
geosynthtetics layer
0 1 2 3 0 1 2 3

Patterns of horizontal
displacements

Remarks More More More More More More More More

load step load step load step load step load step load step load step load step
Horizontal displacement 0.160 0.167 0.168 0.140 0.133 0.137 0.135 0.100
(mm)
Vertical displacement 0.190 0.192 0.200 0.197 0.178 0.82 0.178 0.169
(mm)
Force (kN) 0.900 0.918 0.961 0.978 0.772 0.781 0.803 0.820
Stress (kPa) 45.526 47.042 51.058 52.419 34.724 35.529 37.386 38.701

Table 6-6 Horizontal displacements for MC-model, E=600 MPa and EA= 300 kN/m
Elastic modulus MC-model for E = 600 MPa , EA= 300 kN/m
Friction angle and
dilation angle  = 45 ,  = 15o
o
 = 38o,  = 8o
Number of
geosynthtetics layer
0 1 2 3 0 1 2 3

Patterns of horizontal
displacements

Remarks More More More More More More More More

load step load step load step load step load step load step load step load step
Horizontal displacement 0.0408 0.044 0.0397 0.0376 0.0396 0.0409 0.0437 0.0559
(mm)
Vertical displacement 0.0498 0.0541 0.0526 0.0517 0.0543 0.0539 0.0581 0.0558
(mm)
Force (kN) 0.827 0.826 0.847 0.849 0.775 0.791 0.795 0.804
Stress (kPa) 37.910 37.832 39.599 39.774 33.628 34.965 35.376 36.072

We may resume some findings as presented in Table 6-3 through Table 6-6 that smaller
horizontal displacement can be obtained when we apply higher elastic modulus, higher
friction angle and dilation angle, higher tensile strength and number of geosynthetics.

122
Table 6-7 through Table 6-12 show some important results using HS-model for granular
material with an elastic modulus of 200, 400 and 600 MPa respectively. We may notice
results as shown in Table 6-7 through Table 6-9 that the higher values of elastic moduli,
friction angle and dilation angle provide higher vertical stress. When we apply the elastic
modulus of 200 MPa with tensile strength of geosynthetics around 30 kN/m, at least 3 layers
of geosynthetics is needed for granular materials having friction angle of 45o and dilation
angle of 15o whereas using lower friction angle and dilation angle (38o and 8o respectively),
sample soil fails even applied for 4 layers of geosynthetics.
Table 6-7 Horizontal displacements for HS-model, E=200 MPa and EA= 30 kN/m
Elastic modulus HS-model for E = 200 MPa , EA= 30 kN/m
Friction angle and
dilation angle  = 45 ,  = 15o
o
 = 38o,  = 8o
Number of
geosynthtetics layer
0 1 2 3 0 1 2 3

Patterns of horizontal
displacements

Remarks Soil body Soil body More More Soil body Soil body Soil body Soil body
collapses collapses load step load step collapses collapses collapses collapses

Horizontal displacement 0.720 0.753 0.845 0.858 0.521 0.544 0.589 0.634
(mm)
Vertical displacement 1.228 1.303 1.375 1.422 1.111 1.121 1.268 1.250
(mm)
Force (kN) 0.633 0.634 0.637 0.64 0.634 0.634 0.633 0.636
Stress (kPa) 40.302 39.578 39.325 39.491 35.846 34.051 35.491 36.315

Table 6-8 Horizontal displacements for HS-model, E=400 MPa and EA= 30 kN/m
Elastic modulus HS-model for E = 400 MPa , EA= 30 kN/m
Friction angle and
dilation angle  = 45 ,  = 15o
o
 = 38o,  = 8o
Number of
geosynthtetics layer
0 1 2 3 0 1 2 3

Patterns of horizontal
displacements

Remarks More More More More Soil body More More More
load step load step load step load step collapses load step load step load step
Horizontal displacement 0.339 0.323 0.361 0.348 0.357 0.330 0.346 0.324
(mm)
Vertical displacement 0.693 0.698 0.682 0.699 0.709 0.643 0.685 0.637
(mm)
Force (kN) 0.726 0.735 0.743 0.748 0.062 0.650 0.649 0.659
Stress (kPa) 39.614 39.925 40.427 42.273 30.267 35.026 36.017 35.798

123
At least a layer of geosynthetics has to be used when the elastic modulus of sample soil is 400
MPa with a friction angle of 38o and a dilation angle of 8o as shown in Table 6-8. Another
important finding is that use of the higher friction angle and dilation angle can slightly reduce
the horizontal displacement.
Table 6-9 Horizontal displacements for HS-model, E=600 MPa and EA= 30 kN/m
Elastic modulus HS-model for E = 600 MPa , EA= 30 kN/m
Friction angle and
dilation angle  = 45 ,  = 15o
o
 = 38o,  = 8o
Number of
geosynthtetics layer
0 1 2 3 0 1 2 3

Patterns of horizontal
displacements

Remarks More More More More More More More More

load step load step load step load step load step load step load step load step
Horizontal displacement 0.367 0.390 0.358 0.316 0.206 0.227 0.211 0.197
(mm)
Vertical displacement 0.642 0.657 0.651 0.639 0.467 0.483 0.473 0.453
(mm)
Force (kN) 0.823 0.822 0.83 0.84 0.741 0.737 0.745 0.75
Stress (kPa) 42.342 41.861 43.118 43.586 35.664 34.661 35.647 35.91

Table 6-10 through Table 6-12 are the results using HS-model for granular materials with
elastic modulus of 200, 400 and 600 MPa respectively and inclusion of geosynthetics with
tensile strength of 300 kN/m. Similar result as previous findings that higher values of elastic
modulus, friction angle, dilation angle and number of layers are able to reduce the horizontal
displacement and simultaneously increase the vertical stress.
Table 6-10 Horizontal displacements for HS-model, E=200 MPa and EA= 300 kN/m
Elastic modulus HS-model for E = 200 MPa , EA= 300 kN/m
Friction angle and
dilation angle  = 45 ,  = 15o
o
 = 38o,  = 8o
Number of
geosynthtetics layer
0 1 2 3 0 1 2 3

Patterns of horizontal
displacements

Remarks Soil body Soil body More More Soil body Soil body More More
collapses collapses load step load step collapses collapses load step load step
Horizontal displacement 0.780 0.931 1.09 0.995 0.521 0.689 0.744 0.707
(mm)
Vertical displacement 1.549 1.756 1.882 1.914 1.111 1.267 1.459 1.451
(mm)
Force (kN) 0.630 0.654 0.722 0.817 0.634 0.635 0.678 0.725
Stress (kPa) 39.048 46.547 41.453 43.401 35.846 36.987 35.500 35.745

124
 Table 6-11 Horizontal displacements for HS-model, E=400 MPa and EA= 300 kN/m
Elastic modulus HS-model for E = 400 MPa , EA= 300 kN/m
Friction angle and
dilation angle  = 45 ,  = 15o
o
 = 38o,  = 8o
Number of
geosynthtetics layer
0 1 2 3 0 1 2 3

Patterns of horizontal
displacements

Remarks More More More More Soil body More More More
load step load step load step load step collapses load step load step load step
Horizontal displacement 0.337 0.364 0.355 0.340 0.357 0.326 0.308 0.278
(mm)
Vertical displacement 0.690 0.726 0.687 0.729 0.709 0.641 0.641 0.639
(mm)
Force (kN) 0.726 0.776 0.826 0.868 0.622 0.677 0.694 0.718
Stress (kPa) 39.551 44.254 47.231 50.274 30.267 37.434 33.885 35.517

Table 6-12 Horizontal displacements for HS-model, E=600 MPa and EA= 300 kN/m
Elastic modulus HS-model for E = 600 MPa , EA= 300 kN/m
Friction angle and
dilation angle  = 45 ,  = 15o
o
 = 38o,  = 8o
Number of
geosynthtetics layer
0 1 2 3 0 1 2 4

Patterns of horizontal
displacements

Remarks More More More More More More More More

load step load step load step load step load step load step load step load step
Horizontal displacement 0.313 0.389 0.269 0.299 0.206 0.227 0.168 0.192
(mm)
Vertical displacement 0.585 0.584 0.596 0.632 0.453 0.454 0.446 0.461
(mm)
Force (kN) 0.819 0.859 0.910 0.952 0.738 0.756 0.787 0.806
Stress (kPa) 37.287 41.241 46.197 52.024 36.297 37.173 40.901 41.554

If we compare patterns of horizontal displacements between MC-model and HS-model, we

may notice that the influence of geosynthetics inclusion in granular material using MC-model
is not quite obvious compared with HS-model, particularly when using a moderate tensile
strength.
By inserting geosynthetics layers on material pavement, the horizontal thrust will be reduced.
It is a correct assumption that number of layer inserted in pavement material will reduce the
horizontal thrust. In other words, it the reduced horizontal displacement. Figure 6-2 shows an
influence of the number of geosynthetics layers inserted on granular material.

125
 1.2

EA 30 kN/m Phi 38-Psi 8- E 200

1 EA 30 kN/m Phi 38-Psi 8- E 400

EA 30 kN/m Phi 38-Psi 8- E 600

0.8
Horizontal displacement (mm)

EA 30 kN/m Phi 45-Psi 15- E 200

EA 30 kN/m Phi 45-Psi 15- E 400

0.6
EA 30 kN/m Phi 45-Psi 15- E 600

EA 300 kN/m Phi 38-Psi 8- E 200

0.4
EA 300 kN/m Phi 38-Psi 8- E 400

EA 300 kN/m Phi 38-Psi 8- E 600

0.2

EA 300 kN/m Phi 45-Psi 15- E 200

EA 300 kN/m Phi 45-Psi 15- E 400

0
0 layer 1 layer 2 layers 3 layers
Number of geosynthetics layers EA 300 kN/m Phi 45-Psi 15- E 600

Fig. 6-2 Horizontal displacements vs. Number of geosynthetics layers

Fig. 6-2 suggests that at least two geosynthetics layer (vertical distance of 36.7 cm) of 300
kN/m tensile strengths on the elastic modulus of 200 MPa pavement material are needed, if
not soil sample will be collapsed. Whereas use of low tensile strength of 30 kN/m does
contribute stability of a soil sample. Meanwhile, at least a layer of geosynthetics (vertical
distance of geosynthetics around 55 cm) is needed when applying the elastic modulus of 400
MPa with friction angle of 38o and dilation angle of 8o. Another finding using material with
the elastic modulus of 600 MPa without the geosynthetics layer (soil sample height of 110
cm), soil sample does not fail. It is worthly to mention here that the best result is the use of
300 kN/m-tensile stiffness with the soil stiffness, phi, psi are 400 MPa, 38 degree, 8 degree
respectively. The reason is that the increased number of geosynthetics layers applied causes
horizontal displacements is smaller.

No doubt, the base course or subbase course material must be compacted on a construction.
Hence, this material has a dilation angle and also the material should have a friction angle at
least 30o. Fig. 6-3 presents some findings using MC-model. Generally, the higher values of
friction angle, dilation angle, tensile strength of geosynthetics and number of geosynthetics
layers will provide the higher vertical stress as shown in Fig. 6-3. In x-axis, the numbers for
126
the first, second and third are the friction angle, dilation angle, and number of geosynthetics
layers respectively.
60

Max. stress achieved (kPa)

50 MC-model
E = 200 MPa
40

(a) 30

20 EA=30 kN/m
EA=300 kN/m
10

0
38-8-0 38-8-1 38-8-2 38-8-3 45-15-0 45-15-1 45-15-2 45-15-3
Friction angle, Dilation angle, Number of layers

41
Max. stress achieved (kPa)

40
39
38 MC-model
37 E = 600 MPa
36
35
(b) 34
33 EA=30 kN/m
32
EA=300 kN/m
31
30
38-8-0 38-8-1 38-8-2 38-8-3 45-15-0 45-15-1 45-15-2 45-15-3
Friction angle, Dilation angle, Number of layers

Fig. 6-3 Maximum stress using MC-model at different elastic modulus of base course
(a) E=200 MPa (b) E= 600 MPa

Fig. 6-4(a) through Fig. 6-4(c) reveal some findings using HS-model as given in Table 6-7
through 6-12. When using MC-model, the model did not show stability of sample during
subjected by a load. Vertical stresses are influenced by the elastic modulus of granular
material, friction angle, dilation angle, tensile strength of geosynthetics and number of layer
inserted in granular soil as shown in Fig. 6-4. In x-axis, number of first, second and third is
the friction angle, dilation angle and number of geosynthetics layers respectively.

50
Max. stress achieved (kPa)

45
40
35
30
25
(a) 20
EA=30 kN/m
EA=300 kN/m
15 HS-model
10 E = 200 MPa Soil body collapses
5
0
38-8-0 38-8-1 38-8-2 38-8-3 45-15-0 45-15-1 45-15-2 45-15-3
Friction angle, Dilation angle, Number of layers

127
 45

Max. stress achieved (kPa)

40
35
30
25
(b)
20
EA=30 kN/m Soil body collapses
15
EA=300 kN/m HS-model
10
E = 400 MPa
5
0
38-8-0 38-8-1 38-8-2 38-8-3 45-15-0 45-15-1 45-15-2 45-15-3
Friction angle, Dilation angle, Number of layers

60
Max. stress achieved (kPa)

50

40

30
(c) EA=30 kN/m
20 HS-model
EA=300 kN/m E = 600 MPa
10

0
38-8-0 38-8-1 38-8-2 38-8-3 45-15-0 45-15-1 45-15-2 45-15-3
Friction angle, Dilation angle, Number of layers

Fig. 6-4 Maximum stress using HS-model at different elastic modulus of base course

Circles signs designated in Fig. 6-4 are the stresses achieved when the soil sample is
subjected to a vertical load. It can be seen that the higher friction angles can bear the higher
stresses as well as higher tensile strength of geosynthetics.Though it is not so obvious for the
elastic modulus of 400 MPa when geosynthetics with tensile strength of 300 kN/m is applied.
No doubt, the higher elastic modulus suggests higher stability when being subjected to a load.
Fig. 6-4 shows that the elastic modulus of 200 MPa needs at least 2 layers of geosynthetics
(36.7 cm-vertical distance of layers which the soil sample is 110 cm high) to withstand the
loading. Whilst, materials having the elastic moduli of 400 MPa with a friction angle of 38o
needs at least one layer of geosynthetics (55 cm vertical distance). Granular soils having the
elastic modulus of 600 MPa can withstand a load with vertical stress around of 35 kPa
without using a layer of geosynthetics.

6.3. Stress Concentration Ratio

The simplest starting point for analysis of arching is a two-dimensional plane strain model.
Some findings of Cao Wei-Ping’s experimental work and then the use of finite element as a
comparison are a good description to understand well the arching behavior. One of important
parameters when soil arching occurred is stress concentration ratio, n.
Cao Wei-Ping et al. (2007) conducted experimental work for 15 test models (as explained in
Chapter 4). The Plaxis finite element software using MC model based on soil material of
Qiantang River Beach will be compared with the experimental work. Furthermore, material
properties are listed in Table 6-13.
128
 Table 6-13 Material properties used in finite element analysis
Element and model
Properties Fill Water
Brick
material bag
Linear Linear
Symbol Unit MC
elastic elastic
Unit weight,  kN/m3 22 15.6 15
Friction angle,  [o] - 34/44 -
Dilation angle,  [o] 0 0 -
Cohesion, c kPa - 0.5 0.5
Young’s modulus, E MPa 16000 160 50
Poisson’s ratio, v [-] 0.2 0.33 0.33
Permeability, ki m/s - - 100

6.3.1. Influence of Embankment Height on Stress Concentration Ratio

During running test, the soft soil support was imitated by using the water bag. Test 7
describes the vertical stress both on above soft soil and top pile during embankment filling as
shown in Fig. 6-5. The stresses increase linearly with height of embankment. Moreover, the
stresses on top pile will be higher than those of subsoil.

35
ST2
ST3
30 ST4
ST2a
Vertical pressure (kPa)

25 ST3a
ST4a
20 ST1
MC top, phi 34 deg.
MC bot. phi 34 deg.
15 MC top, phi 44 deg.
MC bot. phi 44 deg.
10

0
0 200 400 600 800 1000 1200
Embankment height (mm)

(a) Vertical stress resulted from Test 7 and model (b) Vertical stress using FEM
Fig. 6-5 Influence of embankment height due to vertical stress
during embankment filling of Test 7

During the consolidation process, in this model, imitated by discharging water gradually from
the water bags, the surface of subsoil will go downward, and it means that the reduced
vertical support occurs over the surface of subsoil. In FE-analysis the process applies
prescribed displacement. It turns out that the stresses at the top pile increase until certain
maximum value and then gradually going down and more stable circumstance afterwards. It
means that the stress concentration ratio, n, would be changing during the consolidation
process. The stress concentration ratio for higher value of h/s will result in the stress
concentration ratio to be higher than those of smaller h/s (see Fig. 6-6). There is strongly
dependent on pile-subsoil relative displacement Sc, which the soil arching is developed at
maximum value when Sc=8-13 mm, whilst beyond this range the soil arching may be less.

129
 80 12
ST2 h/s=0.7

Stress concentration ratio

70 10 h/s=0.9
ST3
Vertical stress (kPa)

60 h/2=1.2
ST4
50 8 h/s=1.4
ST2a
40 ST3a 6 h/s=1.6
30 ST4a
h/s=1.8
20 4 h/s=2
ST1
10 MC h/s=0.7
MC top 2
0 MC h/s=1.4
MC bottom
0 20 0 MC h/s=2.0
0 20
Pile-subsoil relative displacement (mm)
Pile-subsoil relative displacement (mm)

(a) (b)
Fig. 6-6 Variations of vertical stresses and stress concentration ratios, (a) Vertical stress
during water discharge Test 7, (b) Influence h/s on stress concentration ratios

6.3.2. Differential Settlement and Critical Height of Embankment

Test 1 through Test 4, when h/s≤1.4, the embankment height was relatively ’low’ and not
completed soil arching. It shows that the surface of the embankment was non-uniform
implying differential settlement occurred on the top of the embankment. Furthermore, in Test
5 through Test 7, when h/s>1.6, the embankment height was relatively ’high’, the settlements
at the base of the embankment were non-uniform, but at the surface of the embankment
remained almost flat. Both situations when pile-soil relative displacement S =35 mm are
depicted in Fig. 6-7.

(a) Measured settlement (b) Deformation (c) Total displacement

(a) Measured settlement (b) Deformation (c) Total displacement

Fig. 6-7 Differential settlements of Test 1 (above) and Test 7 (below)

130
A comparison for settlement at the surface of the embankment between measurement and
finite element analysis for Test 1 (C) and Test 7 (h/s=2) is shown in Table 6-14.
Table 6-14 Settlements at surface of embankment using FE-analysis
Settlements at surface of embankment (mm)
Finite element method
Measurement
Test number  = 34 o  = 44 o
Above Between Above Between Above Between
of pile piles of pile piles of pile piles
Test 1
11 37 13.73 35 13.5 34.5
(h/s=0.7)
Test 7
40 39 29.63 29.63 29.75 29.77
(h/s=2.0)

Experimental work and FE-analysis suggest that the height of the equal settlement plane, hc,
is about 1.4-1.6 times of the cap beam clear spacing or hc=(1.4-1.6)s. Furthermore, to ensure
that no differential settlement occurs at the surface of embankment, the embankment height
of 1.6s is necessary. Fig. 6-8 presents some deformation of embankment including mesh
deformation, the total vertical displacement, total vertical and horizontal strain for h/s=1.4
using FE analysis when subsoil is moved downward 35 mm.

(a) Deformation (b) Vertical strain (c) Horizontal strain (d) Vertical displacement
Fig. 6-8 Shapes of deformation for h/s=1.4

6.3.3. Comparison of Stress Concentration Ratio between Experimental Work and Several
Analytical Methods
It is useful to understand the parameter of stress concentration ratio on arching phenomenon
between experimental results and some analytical methods as well as the finite element
method. Several analytical methods have already been existed such as Method of Low et al.
(1994), Method of Terzaghi (1943), British Standard BS 8006 (1995).
There are seven model tests with different ratios of embankment height to clear spacing (h/s).
Test 1 has the smallest ratio of 0.7, whereas Test 7 has the highest ratio of 2. Meanwhile,
some values in between are 0.9, 1.2, 1.4, 1.6, and 1.8 respectively. Results of Test 1 through
Test 7 are then compared with analytical methods and also the finite element method.
Parameters of material fill are assumed for c=0 kPa, =44o and =15.5 kN/m3. Resume of
equations for 2D-analytical methods are shown in Table 6-15.

131
 Table 6-15 Equations in analytical method of two-dimension
Method Stress on pile, p Stress on subsoil, s Remarks
Low et
al.(1994) =a/(s+a)
=0.8
a= width of pile cap
s = piles clear spacing
h=high of embankment

Terzaghi
(1943) K= 0.7
Valid for h/s < 2

BS8006
(1995) Cc=1.95 h/a- 0.18 for
end-bearing piles
Cc=1.55 h/a- 0.07 for
friction piles

hc=1.4s

Then, stress concentration ratio can be easily obtained which n=p /s. Herein, the
comparison is only intended for embankment without reinforcement namely from Test 1 with
h/s=0.7 to Test 7 with h/s=2.0 as depicted in Fig. 6-9.

4 5
Stress concentration ratio
Stress concentration ratio

3.5
4
3
2.5 3
2
Test 1, h/s=0.7 2 Test 2, h/s=0.9
1.5 Low et al. Low et al.
1 Terzaghi Terzaghi
1 BS 8006 (2D)
0.5 BS 8006 (2D)
MC model MC model
0 0
0 10 20 30 0 10 20 30
Pile-subsoil relative displacement (mm) Pile-subsoil relative displacement (mm)
(a) (b)

7 10
Stress concentration ratio
Stress concentration ratio

6
8
5
4 6
Test 3, h/s=1.2 Test 4, h/s=1.4
3 Low et al. 4 Low et al.
2 Terzaghi Terzaghi
BS 8006 (2D) 2 BS 8006 (2D)
1 MC model MC model
0 0
0 10 20 30 0 10 20 30
Pile-subsoil relative displacement (mm) Pile-subsoil relative displacement (mm)

(c) (d)
132
 9 10
Stress concentration ratio

Stress concentration ratio

8
7 8
6
6
5
(e) Test 5, h/s=1.6 (f) Test 6, h/s=1.8
4
Low et al. 4 Low et al.
3
Terzaghi Terzaghi
2 2
BS 8006 (2D) BS 8006 (2D)
1
MC model MC model
0 0
0 10 20 30 0 10 20 30
Pile-subsoil relative displacement (mm) Pile-subsoil relative displacement (mm)

12
Stress concentration ratio

10
8
Test 7, h/s=2.0
(g)
6 Low et al.

4 Terzaghi
BS 8006 (2D)
2 MC model
0
0 10 20 30
Pile-subsoil relative displacement (mm)
Fig. 6-9 Comparison between test results, Analytical methods and Finite element analysis for
stress concentration ratio: (a) Test 1, (b) Test 2, (c) Test 3, (d) Test 4,
(e) Test 5, (f) Test 6, (g) Test 7

The figures show that Terzaghi method is always over-prediction in the stress concentration
ratio whereas the BS8006 method suggests strongly under-estimate. Meanwhile, Low et al.
method gives slightly larger result for low embankment h/s≤1.4 but this method good agrees
with high embankment h/s>1.4. Iglesias et al. (1999) have introduced a terminology namely
Ground Reaction Curve by means of normalising subsoil movement downwards to pile clear
spacing. This curve describes a curve of vertical stress over subsoil. When stress on top of the
piles is at a maximum value, the vertical stress over subsoil is at a minimum value.
Meanwhile, the curves above show that the critical values are between 8-13 mm. By using
pile clear spacing of 60 cm, it means that the critical values are from 1.3% to 2.2%.

6.4. Load Transfer Platform

In Malaysia particularly nearby Gebeng Highway, Hassandi et al. (2007) carried out a field
test using the various load transfer platform (LTP) on top of aggregate piers (called
’geopiers’) on soft soil. It would have shown the performance of several load transfer
platform used in the field test. The rammed aggregate piers as columns used in the field test
are ‘floating piles’ over soft soil. Test embankment was approximately 90 m long, 14.5 m
wide, and 3.5 m high. The side slopes of the embankment were 1V:1.5 H.
There are three types of load transfer platform (LTP) performed in the field test, namely:
o Geosynthetics-reinforced LTP with two layers of geogrids (catenary LTP)
o Geosynthetics-reinforced LTP with three or more layers of geogrids (beam LTP)
o Reinforced concrete LTP

133
Two control sections were also provided at the edge of the field test. These sections are
embankment without reinforcement. Layout and dimension of different LTP are shown in
Fig. 6-10.
The geosynthetic-reinforced LTPs consist of an aggregate layer with geogrid reinforcement
within the aggregate. Section 1 has a 1.5 m thick beam LTP with four layers of biaxial
extruded polypropylene geogrid (Tensar SS20) reinforcement spaced at 0.3 m apart vertically
within the LTP, and supported on geopiers spaced in a 3.25 m center-to-center square pattern.
Section 2 has a 1.0 m thick beam LTP spaced at 0.3 m apart vertically, and supported on
geopiers in spaced a 2.5 m center-to-center square patterns. Section 3 has a catenary LTP
with two layers of uniaxial woven polyester high-strength geogrid coated with polyvinyl
chloride (Miragrid 24XT) spaced at 75 mm apart vertically placed in a different direction
(longitudinal and transverse) of embankment. This section was also supported on geopiers
spaced at 2.5 m. Last section (section 4) was supported on 0.3 m thick continuous steel-
reinforced concrete LTP over geopiers spaced of 2.5 m.

Fig. 6-10 Schematic sections of LTP (after Hassandi et al., 2007)

134
 Table 6-16 Properties of geogrids used in the field test (after Hassandi et al., 2007)
High-strength uniaxial
Biaxial extruded geogrid
geogrid
Properties
Machine Cross-machine Machine Cross-machine
direction direction direction direction
Tensile strength at ultimate (kN/m) 20 20 370.3 43.6
Tensile strength at 5% strain (kN/m) 14 14 93.3 17.5
Tensile modulus (kN/m) 280 280 1870 350
Grid aperture (mm) 39 39 101 17.8
Mass/unit area (kg/m2) 0.220 1.289

The selected aggregate used in the LTPs consisted of well-graded crushed granitic rock with
fine material that was less than 3%. This aggregate is normally used as a sub-base layer for
road pavements. Then, over this aggregate blanket, the embankment was constructed using
gravelly sandy clay.
The rammed aggregate pier, called geopier, is a relative the new intermediate-depth columnar
foundation introduced in the construction industry (Fox and Cowell, 1998). Typically, drilled
holes were extended between 2 and 8 m below subsoil surface. In this field test, the initial
drilled diameter of the geopiers was 0.75 m and the initial depth of the drilled hole for the
geopier was 5.5 m.
Site investigation provides a soft silty clay/clayey silt layer as deep as 15 m at some location.
This layer is composed of highly plastic clay with natural content between 35% and 61%.
Field vane tests indicated that the shear strength ranges from 14 to 60 kPa, with most of the
values less than 25 kPa. Moreover, its value of sensitivity varies from 3 to 11.

6.4.1. Settlement of the Embankment over Floating Piles

This embankment is high embankment (h/s>1.4). Therefore, differential settlement at the
surface of the embankment may be omitted because it is very small. Then, total settlement
and differential settlement at the bottom of the embankment are the important things to be
discussed. Table 6-17 provides material properties are used in FE-analysis.
Table 6-17 Material properties used in finite element analysis of embankment test
on Gebeng highway
Element and model
Properties Geopier Fill Aggregate Subsoil Geogrid Concrete
material blanket Slab
MC MC MC MC SSC Elasto-plastic Elastic
Unit weight kN/m3 19 18.5 21 14 14 - 24
Friction anle,  [o] 48 35 38 8 8 - -
Dilation angle,  [o] 0 0 4 0 0 - -
Cohesion, c kPa 1 1 1 15 15 - -
Young’s modulus,
MPa 100 200 300 0.75 0.75 280, 1870 19650
E
Modified comp.
[-] - - - - 0.1183 - -
index, 
Modified swelling
[-] - - - - 0.0229 - -
index, 
Modified creep
[-] - - - - 0.0058 - -
index, 
Poisson’s ratio,
[-] -/0.3 -/0.3 -/0.3 -/0.4 0.15/- - -/0.15
vur /v

135
The total settlement resulted from field test and calculation using finite element method is
compared. Treatment between compaction and no compaction using a load of 550 kPa
generated from wheels roller during the compaction process is presented. In addition, Mohr
Coulomb model for whole soil material (embankment and subsoil) and Mohr Coulomb model
for embankment combined with Soft Soil Creep (SSC) model for soft soil are also presented.
Figure 6-11(a) through Figure 6-11(d) show a comparison between field measurement and
finite element calculation for several types of LTP. Fig. 6-11(a) and Fig. 6-11(b) show that
the total settlement of geogrid from field measurement at the bottom layer would be
deformed deeper than at the upper layer of geogrid. Mohr Coulomb model using plastic
calculation with compaction load of 550 kPa can be used to predict the total settlement only
for the final step of compaction at the end of project execution. Although, it cannot follow the
creep phenomenon on soft soil after completion of project execution. Meanwhile, the SSC-
model for the soft soil is better and can follow the creep phenomenon when using the
consolidation calculation type and no compaction load.

4 200
Embankment height (m)
3.5 Compaction 180
160 Model MC Plastic fully compaction
3 load at layer 1 (bottom)
Embankment height (m)

140 Model MC Plastic no compaction load

Settlement (mm)
2.5 120 at layer 1 (bottom)
Model MC-SSC Plastic fully
2 100 compaction load at layer 1 (bottom)

1.5 80 Model MC-SSC Consolidation no

compaction load at layer 1 (bottom)
60
1 Geogrid Layer 1 (Bottom)
40
0.5 20 Geogrid Layer 2

0 0 Geogrid Layer 3

Geogrid Layer 4

(a) Section 1

4
140
Embankment height (m)
3.5 Compaction
120
Model MC Plastic fully compaction
Embankment height (m)

3 load at layer 1 (bottom)

100
Settlement (mm)

Model MC Plastic no compaction

2.5
load at layer 1 (bottom)
80
2 Model MC-SSC Plastic fully
compaction load at layer 1 (bottom)
60
1.5 Model MC-SSC Consolidation no
compaction load at layer 1 (bottom)
1 40
Geogrid Layer 1 (Bottom)

0.5 20
Geogrid Layer 2

0 0
Geogrid Layer 3
01-Apr
01-May

15-Jul
04-Feb
10-Feb
10-Feb
15-Feb
20-Feb
01-Mar
01-Jan

15 Agu
15-Jun
27-Jan
30-Jan

(b) Section 2

136
 4 350 Embankment height (m)

Compaction
3.5 300 Model MC Plastic fully compaction
load at layer 1 (bottom), Sect. 3
3
250
Model MC-SSC Plastic no compaction

Settlement (mm)
2.5 load at layer 1 (bottom), Sect. 3

200
Embankment height (m)

Model MC-SSC Consolidation no

2 compaction load at layer 1 (bottom),
150 Sect. 3
1.5 Model MC Plastic fully compaction
load at the bottom, Sect. 4
100
1
Model MC-SSC Plastic fully compaction
50 load at the bottom, Sect. 4
0.5
Geogrid Layer 1 (bottom), Sect. 3
0 0
15-Apr

15-Jul
10-Feb
13-Feb
16-Feb
20-Feb

01-May
05-Mar
25-Mar

7 Agu
15-Jan

15 Agu
15-Jun
01-Jan

19-Jan

Concrete Slab 30 cm (bottom), Sect. 4

(c) Section 3 and 4

4
Embankment height (m)
Compaction 700
3.5
Model MC Plastic fully compaction
600 load
3
Embenkment height (m)

500 Model MC-SSC Plastic fully

Settlement (mm)

2.5 compaction load

400 Model MC-SSC Consolidation fully

2
compaction load

300
1.5 Model MC Plastic adjusted
compaction load
1 200
Model MC-SSC Plastic no compaction
load
0.5 100
Control Section 1 (C1)
0 0
15-Apr
01-May

15-Jul
05-Mar
25-Mar
10-Feb
13-Feb
16-Feb
20-Feb

7 Agu
15 Agu
15-Jun
01-Jan
15-Jan
19-Jan

Control Section 2 (C2)

(d) Control sections

Fig. 6-11 Total settlement of LTP sections (at centre of square pattern of geopiers) and
control sections: (a) Section 1 (b) Section 2 (c) Section 3 and 4 (d) Control sections

Similarly, settlement for a catenary LTP over the floating piles will be better predicted with
consolidation calculation type without compaction process regarding with creep as shown in
Fig. 6-11(c). Although, plastic calculation type can be used without considering creep with
compaction at the final of embankment hight. Meanwhile, Fig. 6-11(d) suggests that
settlement without piles will be deeper than those of using piles.

137
6.5. Low Embankment on the End-bearing Piles
In case, the use of end-bearing pile and load transfer platform of catenary (1 or 2 layers of
geosynthetics) Almeida et al. (2008) performed a field test of a low embankment (h/s <1.4) at
Barra da Tijuca district in Brazil. Two situations had already been done by means of
excavated and non-excavated zones. Purpose of the excavated zone is to know the influence
soil support on pile embankment. Lay out for the field test is described in Fig. 6-12.

Fig. 6-12 Layout of field test at Barra da Tijuca district in Brazil (after Almeida et al., 2008)

Properties of materials are shown in Table 6-18 and FE-calculation uses 2-D axisymmetric
and plane strain model. Pavement materials are approached with MC and HS-model and
subsoil with SSC-model. While, linear elastic model is applied for pile and geosynthetics.

138
 Table 6-18 Material properties used in finite element analysis
at Barra da Tijuca district in Brazil
Element and model
Concrete Working Fill Subsoil Biaxial
Properties pile platform embankment Geogrid
Linear Linear
MC MC HS SSC
elastic elastic
Unit weight, 
3
kN/m 24 17.5 18 18 12.5 -
Friction angle,  [o] - 33.8 37 17 5 -
Dilation angle,  [o] - 0 0 0 -
Cohesion, c kPa - 11 11 11 11 -
Young’s modulus, E MPa 24000 300 400 - 0.75 -
Secant modulus, E50 MPa - - - 130 - -
Oedometer comp. modulus, Eoed MPa - - - 150 - -
Un-/reloading modulus, Eur MPa - - - 400 - -
Modified compression index,  [-] - - - - 0.226 -
Modified swelling index,  [-] - - - - 0.012 -
Modified creep index,  [-] - - - - 0.005 -
Poisson’s ratio, vur /v [-] -
-/0.15 -/0.3 0.2/- 0.15/- -
/0.3
Tensile strength, EA kN/m - - - - - 1400

Settlements resulted from field measurement and FE-calculation can be looked into in Table
6-19 both 2D and 3D layout for various positions. Meanwhile, Fig. 6-13 shows the settlement
using FE-calculation with the model using both the MC model and HS model for pavement
material and SSC model for the soft soil at the centre of 4 piles.
Table 6-19 Comparison of settlements between at base of embankment
from field measurement and FE-calculation

Embank Half Settlements till end of

Plates Layout Position ment distance h/b /b construction ,  (m)
height, h between pile Field FE-calculation
(m) caps, b (m) measure MC- HS-
ment SSC SSC
SP 01 2D Midpoint 1.10 0.85 1.30 0.37 0.032 0.027 0.030
SP 05 2D between 1.14 0.85 1.33 0.27 0.022 0.027 0.030
SP 02 2D adjacent 1.28 1.35 0.95 0.27 0.037 NC NC
SP 06 2D pile caps 1.25 1.35 0.93 0.30 0.040 NC NC
SP 03 3D Centre of 1.28 0.85 1.49 0.43 0.360 0.720 0.060
SP 04 3D 4 piles 1.08 0.85 1.27 0.12 0.025 0.015 0.020
SP 07 3D Midpoint 1.23 0.85 1.45 0.19 0.017 NC NC
SP 08 3D between 1.24 0.85 1.45 0.20 0.017 NC NC
pile caps
Remarks: NC = no calculated

Unfortunately, FE-analysis finds the difficulties when modeling the excavated sector which
suggests huge deformation. Whereas, on non-excavated sector the settlements could be
approached using the plastic calculation type with involving loading during the compaction
process or consolidation calculation type . Fig. 6-13 shows settlements at the centre of 4 piles
between field measurement and FA-analysis. It can be seen that several weeks after the end
of construction, the settlement in the field was developed highly huge because of traffic and
weak anchorage.

139
 140 60
Embankment height (m), Non-
120 50 excavated sector
Embankment height (cm)

100 Embankment height (m),

Settlements (cm)
40 Excavated sector
80
Model MC-SSC Plastic calculation,
30
Non-excavated sector
60
20 Model HS-SSC Plastic calculation,
40 End of construction
Non-excavated sector

20 10 Settlements (cm), Non-excavated

sector
0 0
Settlements (cm), Excavated
0 50 100 150 200
sector
Time: days
Fig. 6-13 Settlements at the centre of 4 piles

Strains of 3D layout resulted from field measurement particularly on the excavated zone and
compared with axisymmetric configuration are shown in Table 6-20.
Table 6-20 Comparison of tensile strains for geosynthetics between field measurement
and FE-analysis
Strain (%)
Deformation h/b Field FE-analysis
gauge Layout Position measure MC-SSC HS-SSC
ment
DG 01 3D Face of pile caps 1.5 2.05 1.14 1.52
DG 02 3D 1.5 1.73 1.14 1.52
DG 03 3D 1.5 1.51 1.14 1.52
DG 05 3D Half distance between pile 1.5 0.51 0.27 0.36
DG 09 3D caps, parallel to pile face 1.5 0.32 0.27 0.36
DG 06 3D 90o to DG 05 1.5 1.50 0.85 1.14
DG 10 3D 90o to DG 09 1.5 1.36 0.85 1.14
DG 07 3D Centre of four pile caps, 1.5 1.14 0.63 0.83
DG 08 3D align to pile array 1.5 0.97 0.63 0.83

It is clear that the maximum tensile strength of geosynthetics would be occurred at zone of
adjacent piles, and then the lower tensile strength would be found at half distance between
piles with parallel to pile face. This zone is so-called ’primary reinforcement’, whereas other
as ’secondary reinforcement’.

6.6. Influence of Traffic Load on Arching Effect

A full-scale test of a low embankment using the end-bearing wooden piles had been carried
out in Giessenburg (so called ’Kyoto Road’) in the Netherlands. Suzanne van Eekelen at al.
(2008) have observed the influence of loading induced by traffic load during 3 ½ years after
completion of construction by installing cell pressures at the surface of piles located above
and below geosynthetics to measure vertical stress on the top of piles.
The Kyoto road was constructed on 13 m long wooden piles and configured using piles
spacing of 1.27 m grid pattern, concrete pile caps with a height of 0.4 m and 0.3 m diameter.
The geogrid reinforcement consisted of two layers uni-axial grids, perpendicular on road axis
Fortrac 400/30-30 M and along the axis Fortrac 350/50-30 M (see Fig. 6-14). On top of that,
a 1.15 m high compacted embankment fill of a ’Hegemann’ (sandy) sludge mixture was
constructed.
140
 Fig. 6-14 Layout of the Kyoto road field test (after Eekelen, 2009)

The Hegemann sludge mixture is a mixture of dredged material and additives containing
mainly clay and cement with the following properties: average unit weight =18.6 kN/m3, a
friction angle of =33.8o and a cohesion of 11.5 kPa. The Kyoto road was built over a 9 m
deep of soft soil with a reaction modulus k= 477 kN/m3. Table 6-21 shows soil properties
used in Kyoto road embankment.

Table 6-21 Properties of fill material used in Kyoto road embankment

wet dry avg w kv  c
3 3 3 o
kN/m kN/m kN/m % m/s kPa
22.2 17.0 18.6 18.1 2.1E-9 33.8 11.5
Source: van Eekelen, 2009

Table 6-22 shows Young’s moduli of subsoil used for analysis. These values are found from
soil investigation on site and then tested in the laboratory.

Table 6-22 Young’s moduli of Kyoto road’s subsoil

Position Material Thickness (m) E (kN/m2)
Top layer peat d1= 1.45 1077
Lower layer clay d2= 1.50 2000
Source: van Eekelen, 2009

The modulus of subgrade reaction, k, can be calculated as follows: k = (E1.E2)/ (E1.d2+E2.d1)

= 477kN/m3.
Performance of gosynthetics is time-dependent, so time will influence the performance of
material in construction along its life service. Fig. 6-15 shows their isocrones: time dependent
tensile stiffness J can be evaluated for a strain 2%.

141
 Fig. 6-15 Isochrones of geogrids (after Eekelen, 2009)

By using Fig. 6-15, tensile stiffness of material used on the test of road embankment can be
estimated for long-term as indicated in Table 6-23.
Table 6-23 Tensile stiffnesses of geogrids (after Eekelen, 2009)
Direction on Time under Ultimate tensile Tensile stiffness J (kN/m2)
road load strength, UTS J=(% of UTS/strain) x UTS
construction (kN/m) Herein values at 2 % strain
Longitudinal 1 day 350 (25.0/2)x350=4375
Perpendicular 1 day 400 (25.0/2)x400=5000
Longitudinal 1 year 350 (22.8/2)x350=3990
Perpendicular 1 year 400 (22.8/2)x400=4560
Longitudinal 10 years 350 (22.1/2)x350=3868
Perpendicular 10 years 400 (22.1/2)x400=4420

6.6.1. Load Distribution

To compare the calculated and measured load distributions, load parts A, B, and C are defined
as (Fig. 6-16):
o Load part A is directly transferred to the pile caps through arching effect,
o Load part B is transferred through the reinforcement to the pile caps,
o Load part C is resting on the subsoil
Unity for load A, B and C are given in kN.

Fig. 6-16 Load distribution in geosynthetics-reinforced piled embankment

(after Eekelen et al., 2008)

142
Devices TPC t1, t2 and t3 were installed on top of reinforcement. These TPCs measured the
pressure imposed directly on the piles. The total pressure cells above reinforcement layer
(TPCt1, TPCt2, TPCt3) measure load A, whereas total pressure cell below reinforcement
layer (TPCb1) for measuring load (A+B). The vertical load B, which is the load carried by the
geosynthetics reinforcement (GR). By means of tensile forces in the geosynthetics, this load
is transferred to the piles. The curve presenting load B was determined by subtracting the
average measured pressure of TPC t1, TPC t2 and TPC t3 from TPC b1.
The transferred vertical load on each pile was calculated using the equation of
*H+p=18.6*1.15+p= 21.39+p kPa, where p is a surcharge. As soon as arching occurred, the
load was transferred laterally to the piles. Therefore, the pressure measured at TPCs would be
more than 21.39+p. The vertical distance between TPCs and the horizontal line at 21.39 kPa
was an indication of arching.
There are various kinds of material used in the model such as embankment fill, pavement,
subsoil, geosynthetics, pile cap, and pile. Therefore, the behaviour of the material also varies.
HS model is used to model granular and embankment fill. Linear elastic is applied for
modeling concrete pile, wooden pile, and hotmix asphalt. Elastoplastic is used to model
geogrid. Futhermore, SSC model is applied for subsoil. Properties of materials in
axisymmetric configuration of FE-analysis are shown in Table 6-24.
Table 6-24 Material properties used in finite element analysis of Kyoto road
Element and model
Concrete Wooden Hotmix Granular Fill Subsoil Biaxial
Properties Pile caps piles asphalt mixture embkmnt upper lower geogrid
Linear Linear Linear Elasto-
HS HS SSC SSC
elastic elastic elastic plastic
 kN/m3 24 9.75 23 20 18.6 10.3 13.4 -
 [o] - - - 44 33.8 20.3 17.43 -
 [o] - - - 8 0 0 0 -
c kPa - - - 1 11.5 2.05 4.53 -
E50 MPa - - - 160 6.215 1.07 2.36 -
Eoed MPa - - - 130 4.975 - - -
E/Eur MPa 24000/- 12500/- 4000/- -/400 -/14.92 - - -
 [-] - - - - - 0.176 0.138 -

 [-] - - - - - 0.032 0.010 -

 [-] - - - - - 0.011 0.006 -
vur /v [-] -/0.2 -/0.2 -/0.2 -/0.2 -/0.2 0.15/- -
EA kN/m - - - - - - 4100
kv m/s - - - 0.1 1.728E-4 6.86E-4 4.1E-4 -
Konc [-] - - - - - 0.653 0.700 -
Drainage Non Non Non Non
Drained Drained Undrained
Porous Porous Porous Porous

Distribution of load A, B, and C along 2 ½ years for both field measurement and finite
element analysis is shown in Fig. 6-17. Herein, there are three important parts of field
measurement. The first part is the fluctuation of vertical stress on the top pile over geogrid
along 2 ½ years. Secondly, it is the distribution of vertical stress on the top pile below
geogrid. This stress is additional stress due to load below the soil arching which geogrid bring
it to top pile. Last part is a fluctuation of vertical stress on the surface of subsoil below
geogrid. Each part is also compared with results using the analytical method and finite
element method.

143
 At center of piles
(a)

At edge of piles

(b)

At center of span

At adge of span

At upper subsoil

(c)
At lower subsoil

Fig.6-17 Load Distribution for A, B and C of vertical stresses observed throughout 2 ½ years,
Prediction using analytical method and FE-calculation (a) Vertical stress (load A) directly on
pile above geosynthetics, (b) Vertical stress (load B) under geosynthetics, (c) Vertical stress
(load C) on subsoil

144
Fig. 6-17 shows that it took several months to develop the arching fully. Perhaps, this
phenomenon was caused by cementation and settlement process of the fill material. From
July 2006 onward, the arching measurements were relatively constant. The fluctuations were
mostly due to variation in the weather, moisture content and the alternating periods of traffic.
The prediction of the load acting directly on the piles using BS8006 is better than that of
EBGEO, which is much higher than the measured values. However, this value is not so
important because for the design purpose for piles the total load is conservatively assumed as
the applied load on piles.
When taking into account the support of the subsoil, the EBGEO gives a better approach for
predicting the vertical load on geosynthetics compared to BS8006. It is an utmost important
thing that the imposed load on geosynthetics directly determines the tensile forces on the
geosynthetics reinforcement.
By using HS model for fill material and SSC model for soft soil, also using plastic calculation
type with 550 kPa as traffic load to predict distribution of vertical stresses for load A, B, and
C, it shows that the vertical stresses resulted from field measurement is in range of FE-
analysis. Furthermore, contour images of stresses from axisymmetrical approach of FE-
analysis with1651 elements are depicted in Fig.6-18. As shown in the figure, critical stress on
pile is at the edge of the piles which punch take place.

Fig. 6-18 Vertical stresses on piled embankment

6.6.2. Influence of Traffic Load

Traffic started immediately after the completion of construction. Arching needed several
months to develop completely (increasing of load A, decreasing of load B, C and pore
pressures).
A heavy dynamic load induced by vehicular traffic, however, can cause a sudden short-term
decrease of arching (load A decreases and load B increase suddenly). This construction only
serves traffic during working hours. Fig. 6-19 shows a daily arching reduction during the first

145
passages of the day. However, during periods without traffic or rest period so the
geosynthetics reinforcement has an opportunity to restore again.

(a)

(b)

Fig. 6-19 Daily arching cycle on the Kyoto road, (a) Load A, (b) Load B
(after Eekelen et al, 2008)

The following Fig. 6-20 shows the increasing of load B after a passage of a rather heavy
truck, but the arching generally recovers during the rest of the day or weekend, although
other passage is still occurred.

Fig. 6-20 Passage of a truck of 397 kN with 2x2-axles (after Eekelen et al, 2008)

It is clear that higher loading implies the alteration of soil arching. When load at the surface
of the embankment is increased, geosynthetics will bear the additional load and then to be
transferred to pile.

6.7. Summary

Finite element analysis has huge ability and some advantages to solve a lot of problem in
geotechnical engineering. In Chapter 6, some interesting topics regarding with experimental

146
works and some field case studies in Chapter 4 were readily solved using 2D Plaxis software.
Some important findings can be explained as follows:
 In case of high embakment, because of certain circumstance, utilizing multi-layer of
geosynthetics provides some benefits. Effectiveness of vertical distance of
geosynthetics depends on properties of geosynthetics mainly tensile strength and
tensile modulus. That is also influenced by elasticity modulus of granular soil, friction
angle and dilation angle. By inserting geosynthetics into granular soil, compound
material is able to bear a higher load and simultaneously reduces the horizontal
outward thrust. For a compacted soil, HS-model provides a good agreement with soil
behaviour. It can reveal the influence of number of layer and pattern of horizontal
displacement. By using this model, it is clearly described that maximum lateral
outward thrust will be occurred at the lower part of embankment for low elasticity
modulus of granular soil, and then at the middle and the upper part for medium and
very high elasticity modulus, respectively.
 Vertical stress at top pile and on surface of subsoil will be increased in line with high
of embankment.
 When subsoil is moving downward because of consolidation process or low bearing
capacity, vertical stress at top pile will be increased until reaching a maximum value
and then goes slow down and stable afterwards. Otherwise, on surface of subsoil it
will be decreased before getting a stable stage.
 Stress concentration ratio (SCR) is influenced by a height of embankment. The more
higher of embankment, it will be increased. Maximum stress concentration ratio or
this critical value is reached when ratio relative displacement between pile and
subsoil to clear spacing of piles ranges from 1.3 to 2.2%.
 Maximum SCR using Low et al.’s method suggests more close to the measured
findings in experimental works compared to others (Terzaghi and BS8006). Terzaghi
method provides upper limit, and otherwise BS8006 as lower limit.
 Differential settlement at surface of embankment can be minimized or equal to be
nearly zero, when ratio height of embankment to clear spacing of piles is at least 1.4.
 Shape of soil arching for a low relative density (55±7%) is isosceles triangle.
 Load transfer platform (LTP) can be provided using catenary and/or beam system.
Placement of inclusion these materials must be inside of soil arching.
 Using soft soil creep (SSC) model with consolidation calculation type agrees with
behaviour of soft soil compared to other model (MC) and other type of calculation
(plastic).
 Influence of loading on piled embankment over undrained soft soil during a
compaction stage is slightly small.
 In case of embankment over end-bearing piles, plastic calculation type can be applied
and behaviour of embankment material is better modelled using HS-model.
Meanwhile, for soft soil it can be applied the SSC-model.
 Maximum strain of geosynthetics occurs at the face of the piles, and then the lower
one is at the half distance between piles. Meanwhile the lowest is at the centre of four
piles.

147
 Finite element provides good results. For instance, in case study of Kyoto road, by
using the HS-model to model the embankment and the SSC-model to model the
subsoil which agree with the field results. Here, axisymmetric configuration with a
load of 550 kPa is applied and plastic calculation type is used for FE-analysis. In
comparison with another analytical method (BS8006), the EBGEO method suggests
better result which is close to the measured findings in the field.
 When surcharge load induced by vehicular traffic passes through on surface of
pavement, shape of soil arching will be changed. Not only an increasing vertical stress
or increasing load at the top piles will be increased, but it also happens on surface of
subsoil.
 For an embankment with stabilized material, for instance using cement, soil arching
effect fully starts after a period of time.

148
 CHAPTER 7

Case Studies Using FEM

This chapter presents analyses of two sites encountered in the field measurement on Supadio
airport projects, namely runway reconstruction work and Apron widening project. My works
was included in these sites which my role was as a pavement engineer during design stage
and intensively as supervision team when executing the project. The first one is a small part
of Runway Overlay Project in 2009 to level up PCN 50 of Supadio airport runway in order to
be able to serve heavier airplane. Latter is a project that was completed in 2006 to expand the
apron to anticipate high demand of air traffic in the future. Furthermore, finite element
method is applied and then compared with some results of field measurements. Main topic of
this chapter is to discuss settlement of pavement construction over soft soil.

7.1. Location
For purpose of case studies, Supadio airport is chosen regarding with settlement of pavement
construction over soft soil. There are two sites for these case studies. The first site is located
in the middle of the runway and the second one is situated on the new apron. Main problem
in the first site is the settlement of partial weak segment of runway, particularly on wheel
track of the airplane, because of the heavier kind of airplane. Whilst, discussion at the second
site is different settlement in adjacent segment between new construction and existing
construction as well as creep impact for pavement construction over deep soft soil. Aerial
view of location for case studies is presented in Fig. 7-1.

Fig. 7-1 Aerial view of Supadio airport

149
In the airport, there is a runway with 2500 m long and 30 m wide which direction of the
runway is 33-15. Whilst, the widening concrete apron with 80×100 m is located in northern
part of existing asphaltic apron. Everyday the airport serves air traffic started from 7 a.m to
10 p.m which the heaviest airplane is Boeing 737-400 with a maximum wheel load around 30
tonnes.

7.2. Runway Reconstruction Work of Supadio Airport

Since several months of commencing and the airplane Boeing 737-400 was operated, it
appeared a weak section on the runway with 35 m long and 5 m wide that was started at
station 1+150. To overcome this problem, wooden piles of 12 cm in diameter and 5 m long
using a square pattern of 40 cm pile spacing were applied to support the weak section. The
heave was around 4 cm even more that was occurred on this weak spot of runway. The
reconstruction work had been completed in April 2009.
Soft clay underlying pavement of runway has the undrained shear strength varying from 7 to
20 kPa and the values of the unconfined compressive strength ranging from 8 to 65 kPa
collected from 10 boreholes with depth between 3 m and 20 m. Natural water content of this
type of soil is between 35 and 90 percent. Plasticity index is around 20 percent and internal
friction is laid between 3% and 13.6% with mean value of 8%.
During undertaking this work, time is very important and limited. Starting from digging and
removing of pavement around 90 cm deep until the finishing work of the bituminous layer as
runway surface had to be done within a night because at 7 am the runway has to be able to
operate the air traffic as the normal situation. Fig. 7.2 is showing the cross-section of Supadio
airport runway.

Center line Wheel of airplane

100 cm 500 cm 900 cm

Asphalt Concrete 10 cm

Asphaltic Crushed Stone 15 cm 55 cm

concrete

Cobblestone Stabilized Base Course 65 cm 25 cm

Sand 10 cm

Subgrade Wooden pile 12 cm,#40 cm, L 5 m

Fig. 7-2 Cross-section of Supadio airport runway (reconstruction pavement in the middle and
existing pavement on left-right side)
150
Wooden piles were driven downward as soon as possible after excavation was completed.
Base course as an important part of this work has to be stabilized. Stabilized base course
using cement content of 6% and then each more or less 30 cm deep of the base course layer
was compacted in this site to get maximum density until the whole depth of base course.
Above the stabilized layer, crushed stone with 15 cm thick was poured and then compacted.
This layer serves to prevent the poor impact of water contacting directly with the bottom
asphaltic surface on the top surface of stabilized base course due to an oxidation process
during curring time. Top layer of pavement is 10 cm thick asphaltic surface course with
Marshall stability 1000 kg. Fig. 7-3 shows the execution steps of runway reconstruction
work.

(a) (b) (c)

Fig. 7-3. Execution of Supadio airport runway reconstruction work (a) Driving piles (b)
Pouring base course (c) Compaction

The aim of reconstruction work of Supadio airport runway in Indonesia is to provide a stable
surface for the operation of the airplane B737-400 which this kind of airplane is heavier than
B737-300 operated before. After about 4 months airplane B737-400 was operated, a
deflection appears around 2 cm that occured in this weak spot of runway. To overcome this
problem, wooden piles 12 cm in diameter and 4 m long using a square pattern of 40 cm piles
spacing was applied to support the weak section and also stabilize base course using 6%
cement.

Everyday the airport serves more or less 30 departures, and if during ramping of airplane to
apron the wheels of airplane touch an observed point around 0.5 second, hence during
airplanes passing through 4 months is equivalent to an hour of parking load and similarly for
36 months around a half day. It is relatively difficult to ensure a settlement caused by the
moving load using finite element method. Though, the settlement coming from an airplane is
very significant. No load means that there is no loading at the surface, whereas fully load is
similar to an airplane that parks at the parking stand all time. Hence, the moving load may
take a value in between.
There are various kinds of material used in the model such as embankment fill, pavement,
subsoil, and pile. Therefore, the behaviour of the material also varies. MC model is used to
model sand, cobble stone, and gravel. Linear elastic behaviour is applied to model wooden
pile, stabilized base course, and hotmix asphalt. Furthermore, SSC model is applied for
subsoil. Properties of material for FE-analysis is shown in Table 7-1. In the FE-analysis the
plane strain after Bergado (1994) was applied to transform from 3D to 2D.
151
 Table 7-1 Material properties used in FE-analysis on Supadio airport runway reconstruction
work
Properties Element and model
Wooden Asphaltic Gravel Stabilized Subsoil Coble Sand
piles concrete base course stone
Symbol Unit Linear Linear MC Linear SSC MC MC
elastic elastic elastic
Unit weight,  kN/m3 9.5 22.5 20 20 15.5 20 18
Friction angle,
[o] - - 35 44 13.7 33 30

Dilation angle,
[o] - - 0 - 0 0 0

Cohesion, c kPa - - 1 210 8 1 1
Young’s
MPa 12000 4000 400 660 - 300 120
modulus, E
Modified
compression [-] - - - - 0.1008 - -
index, 
Modified
swelling index, [-] - - - - 0.0202 - -

Modified creep
[-] - - - - 0.00350 - -
index, 
Poisson’s ratio,
[-] -/0.33 -/0.2 -/0.33 -/0.2 0.15/- -/0.33 -/0.33
vur /v
Drainage Non- Non-porous Drained Non-porous Undrained Drained Drained
porous
Calculation type Stage construction, Plastic calculation

Characteristic of soft soil is low bearing capacity and high compressibility when being
subjected by a load. Table 7-2 presents FE-analysis for settlement of the pavement that is
reinforced by wooden piles.

Table 7-2 Settlements and types of loading for pavement reinforced by wooden piles

Settlements and types of loading

Description
Parking load Moving load

Settlement estimated for 4

months

6.60 mm 6.16 mm

Settlement estimated for 3

years

17.70 mm 6.23 mm

152
Table 7-3 presents FE-analysis for settlement of pavement without wooden piles. Settlements
of pavement without wooden piles are deeper than those of with wooden piles. It means that
piles contribute to reduce the settlement of the pavement.

Table 7-3 Settlements and types of loading for pavement unreinforced by wooden piles
Settlement and types of loading
Description
Parking load Moving load

Settlement estimated for 4

months

14.87 mm 13.93 mm

Settlement estimated for 3

years

33.07 mm 14.06 mm

During the settlement process, soil body beneath a load will shove the outer of adjacent soil
to heave up. It can be seen some pictures in Table 7-2 and Table 7-3. Furthermore, Fig.7-4
describes clearly this phenomenon by looking into incremental displacement.

Fig. 7-4 Incremental displacement

153
Fig. 7-5 provides settlements on the weak spot of runway both reinforced by wooden piles
and without piles. These results are compared to each other, as well as a comparison with the
settlement of existing pavement construction.

35

No load of airplane No load of airplane

30
Load of moving airplanes Load of moving airplanes

Load of parking airplane Load of parking airplane

25
Existing pavement with parking airplane

Settlements (mm)
20 20
18
16
15
Settlements (mm)

14
12
10 10
8
6
5
4
2
0 0
0 10 20 30 40 0 10 20 30 40
Months Months

(a) with piles (b) without piles

Fig. 7-5 Settlements on weak-spot surface of Supadio airport runway (a) with piles (b)
without piles

It is clear that by using piles to support the pavement construction contributes smaller
settlements than those of no piles. Fig. 7-5 (b) indicates that a new type of construction for
the reconstruction work is not better than the type of the existing pavement construction when
the piles do not be applied to support it. For settlements induced by the moving airplanes, the
FE calculation suggests that after the reconstruction work is completed. The settlement is
more or less 6 mm and this value is stable until 36 months.

7.3. Apron Widening Project of Supadio Airport

This project is as an effort to fulfill the increasing demand at Supadio airport in Indonesia. A
30 cm-slab concrete has already been chosen for apron expansion 80×100 square meters at
western side and 40×80 square meters at eastern side, respectively. To support the pavement,
totally 120 cm high, wooden piles with diameter 12 cm and 10 meter long had been arranged
on a square pattern of 50×50 cm.
This project was completed in August 2006 and everyday this facility is able to serve airplane
B737-400. There are two parking stands operated in the new apron. The occupancy level of
parking stand that close to existing apron is higher than other because this place is quite easy
for maneuver of airplanes. The whole structure of the pavement on the widening apron as
illustrated in Fig.7-6.

154
 2
Concrete slab 30 cm, fc 400 kg/cm

2
Soil cement 10 cm, qu 50 kg/cm
120 cm Base course 25 cm, CBR 80%

Subbase course 35 cm, CBR 40 %

Coarse sand 20 cm, CBR 15%

Subgrade

Wooden pile 12 cm

# 50 cm
L 1000 cm

Fig.7-6 Cross-section of Supadio airport apron

In airport pavement engineering, there are two important classification values, the values for
airplane and pavement. Value for pavement is so-called PCN (Pavement Classification
Number). Meanwhile value for airplane is so-called ACN (Aircraft Classification Number).
The PCN value must be higher than the ACN value. It means that the pavement is able to
serve a certain airplane. Characteristic of airplane weight and Aircraft Characteristic Number
(ACN) for airplane of Boeing 737-400 is briefly listed in Table 7-4.

Table 7-4 Weight and ACN for airplane of Boeing 737-400 (after ICAO, 1999)
ACN for Rigid Pavement ACN for Flexible Pavement
Subgrade (MN/m3) Subgrade (CBR)
Aircraft All-up Load on Tire High Med Low Ultra High Med Low Ultra
Model mass/ mass one main pressure low low
empty gear leg (MPa) 150 80 40 20 15 10 6 3
(kg) (%)

737-400 68.266 46.91 1.27 42 44 47 48 37 39 44 48

33.643 18 19 20 21 16 17 18 21

From data above we know that the maximum weight of one main gear leg is 68,266×46.91%
(32,000 kg or 30 tonnes). For taking an assumption that the shape of the tire imprint is a
circle, the radius of the circle is around 28.32 cm.
Finite element analysis is used to estimate settlements on apron pavement. The consolidation
calculation type is chosen in estimating settlement for soft soil. The SSC model is applied to
model the soft soil. For soft soil which creep phenomenon is really obvious, the model is able
to imitate the creep behavior when being subjected by a load. The MC model is applied for
granular soil. Meanwhile, elastic linear model is applied for wooden piles and concrete slab.
Material properties of some elements for FE-analysis are shown in Table 7-5.
155
 Table 7-5 Material properties used in finite element analysis of Supadio airport apron
widening project
Element and model
Wooden Portland Soil Base Sub- Sand Subsoil
Properties pile cement cement course base
concrete course
Symbol Unit Linear Linear MC MC MC MC SSC
elastic elatic
Unit weight,  kN/m3 9.5 24 19 20 19.5 18 15.5
Friction angle,  [o] - - 44 40 38 35 13.7
Dilation angle,  [o] - - - 8 5 3 0
Cohesion, c kPa - - 210 1 1 1 8
Young’s modulus,
MPa 12000 30000 200 500 300 15 -
E
Modified
compression [-] - - - - - - 0.1008
index, 
Modified swelling
[-] - - - - - - 0.0202
index, 
Modified creep
[-] - - - - - - 0.0035
index, 
Poisson’s ratio,
[-] -/0.33 -/0.2 -/0.2 -/0.33 -/0.33 -/0.33 0.15/-
vur /v
Permeability, ki m/day - - - 1 1 1 0.0001
OCR [-] - - - - - - 1.5
Drainage Non- Non- Non- Drained Drained Drained Un-
[-] porous porous porous drained
Calculation type Stage construction, Consolidation calculation

7.3.1. Settlement at Surface of Pavement

After completion of the apron widening project was around 5 years ago, settlement resulted
from field measurement on the surface of concrete pavement at the first parking stand is
around 22.7 cm and the second parking stand is 16.7 cm. Meanwhile, settlement at the edge
of the apron that no load is 16 cm. Generally, operation time of apron is started at 7.00 a.m to
10.00 p.m, and then at the first parking stand there will be an airplane that overnight until
7.00 a.m the next day. If there are two airplanes at that night, the second parking stand is
provided for parking the other. The first parking stand bears a load of airplane around 12
hours. Fig. 7-7 depicts the total settlements on the surface of new apron.

New apron elevation in 2006

New apron elevation in 2011

Transfer slope Parking stand 1 Parking stand 2 +10.160

200
1 +10.010
+9.933 +9.993 +10.000
Existing apron
New apron
100 m

Fig. 7-7 Total settlements on the surface of new apron

156
Settlements of apron pavement at the 5th year using 2D Plane strain of FE analysis is
described in the Fig.7-8. We can see that the estimated total settlement is around 25.0 cm and
there is no differential settlement (too small) at the base of the embankment between top piles
and surface of subsoil.

Fig. 7-8 Predicted settlement at surface of parking stand 1

Comparison of settlements between the settlement at the first parking stand resulted from
field measurement and settlements from FE-analysis with different time and type of loading
can be seen in Fig. 7-9.

60

50

40
Settlements (cm)

30
No load of airplane
Load on parking stand (12 hours)
20
Load on parking stand (24 hours)
No load (Field measurement)
10
Actual load (Field measurement)

0
0 5 10 15 20 25 30 35
Years

Fig. 7-9 Settlements on the surface of concrete apron at Supadio airport

157
Fig. 7-9 presents total settlements at surface of parking stand 1 during the 5 years. The
measured values from field measurement are located between settlements with no load of
airplane and those of 50% occupancy level of parking stand. Of course, actual settlements
from field measurement are higher than those of when no airplane that parks at the surface of
pavement. It is over-estimated when predicting the total settlement with airplanes fully park
on the apron which actually airplanes are coming, parking and then leaving alternately.

7.3.2. Block Behaviour of Settlements

When improving the weak-ground using piles, Pile spacing-Diameter pile ratio (s/D) is an
important parameter. Commonly, to keep the structure is able to behave block behavior so the
ratio is less than 5. Here, with s=50 cm and D=12 cm or (s/D=4.17), this work is presented
using 2D Plane strain of FE-analysis. When settlements at top piles are equal or relatively
close to settlements at the surface of soft soil in the middle of between piles, the settlements
behave as a block.

Fig. 7-10(a) through 7-10(d) show settlements at different length of piles either no load or
maximum load of airplane that parks on the surface of concrete pavement. From these
figures, we may notice settlements both on top piles and middle between piles at the base of
the pavement for two cases, namely no airplane parks at the parking stand and airplane parks
whole day at the parking stand. Firstly, settlements due to airplane that parks on the surface
of the pavement are deeper than those of no airplane. Secondly, use of longer piles is able to
reduce the settlement for two cases. Thirdly, it is very important to note that the pavement
structure behaves as a block because differential settlement on top piles and in the middle
point between piles is very small for 2 cases. Lastly, settlements on the surface and bottom of
the pavement are very small, it means that the material pavement does not deform.

80 at surface, no load 70 at surface, no load

of airplane of airplane
70 60
at top of piles, no at top of piles, no
60 load of airplane load of airplane
50
Settlements (cm)
Settlements (cm)

50 at base of at base of
embankment, no 40 embankment, no
load of airplane load of airplane
40
at surface, with 30 at surface, with
max. load of max. load of
30 airplane
airplane
20
20 at top of piles, with at top of piles, with
max. load of max. load of
airplane 10 airplane
10
at base of at base of
embankment, with embankment, with
0 0 max. load of
max. load of
0 20 40 airplane
0 20 40 airplane

Years Years

(a) Pile length of 2.5 m (b) Pile length of 5 m

158
 60 at surface, no load 60 at surface, no
of airplane load of airplane

50 50 at top of piles, no
at top of piles, no
load of airplane load of airplane

40 40
Settlements (cm)

Settlements (cm)
at base of at base of
embankment, no embankment, no
load of airplane load of airplane
30 30
at surface, with max. at surface, with
load of airplane max. load of
airplane
20 20
at top of piles, with at top of piles,
max. load of with max. load of
airplane airplane
10 10
at base of at base of
embankment, with embankment,
0 max. load of 0 with max. load of
airplane airplane
0 20 40 0 20 40
Years Years

(c) Pile length of 10 m (d) Pile length of 15 m

Fig. 7-10 Predicted settlements at different length of piles and years, (a) Pile length of 2.5 m
(b) Pile length of 5.0 m (c) Pile length of 10 m (d) Pile length of 15 m

Fig.7-11 below presents the estimated settlements of apron pavement without using piles.
These settlements at 30th year for no loading and maximum loading are 50.4 cm and 76.6 cm
respectively. Whilst, for the 5th year, the settlements are 31.9 cm and 54.9 cm respectively. Of
course, settlements without using piles in construction give the deepest value compared to
those of using piles. All settlements in different length of piles show that settlements behave
’block’ where differential settlements at top piles and subsoil in the middle of between piles
are very small.

90
80 at surface, no load of
airplane
70
Settlements (cm)

60 at base of embankment, no
50 load of airplane
40
at surface, with max. load
30 of airplane
20
10 at base of embankment,
with max. load of airplane
0
0 10 20 30 40
Years

Fig. 7-11 Predicted total settlements of apron pavement without piles

159
7.3.3. Effectiveness of Piles Length on Floating Piles
No doubt that use of longer piles on a construction over soft soil contributes a smaller
settlement. For construction over soft soil which creep behavior is really obvious, it is worth
to note that the effectiveness of piles length that is penetrated in soft soil must be taken into
account. Relative Settlement Reduction (RSR) can be used to describe the effectiveness of
pile length when constructing the floating pile group on soft soil. The RSR is a ratio for a
deviation of settlements between no piles and piles to settlement no piles or (So-SL)/So. Where
So is settlement without using piles and SL is a settlement with a pile at certain length of piles.
Fig. 7-12 provides the effectiveness of pile length using RSR parameter for case no load of
airplane and parking load of airplane. It shows when a ratio of pile length to soft soil
thickness is less than 0.2, the length of the piles will not quite significant reduce settlement.
Moreover, time of period and magnitude of loading are an important aspect when observing
the effectiveness of pile length. Short-term monitoring is more effective than long-term
period to know the effectiveness. Whilst, higher load implies that longer pile is needed to
reduce the excessive settlement. In other words, the use of pile length more than 0.2 of soft
soil thickness is an important consideration to overcome the creep behaviour.

1 1

0.9 0.9
Relative settlement reduction
Relative settlement reduction

0.8 0.8

0.7 0.7
30 years 30 years
0.6 0.6
20 years 20 years
0.5 0.5
10 years 10 years
0.4 0.4
5 years 5 years
0.3 0.3
3 years 3 years
0.2 0.2
1 year 1 year
0.1 0.1

0 0
0 0.5 1 0 0.5 1
Pile length / Soft soil thickness Pile length / Soft soil thickness

(a) (b)
Fig. 7-12 Effectiveness of pile length, (a) No load of airplane (b) Parking load of airplane

160
 CHAPTER 8

Conclusions and Recommendations

In the framework of research on analysis of dynamic loading behaviour for pavement on soft
soil, several research topics have been presented. The research can be divided into six main
topics consisting of the effective distance of geosynthetics for reinforcement, analysis of
stress concentration ratio on high compressibility of soil, various types of load transfer
platforms on floating columns, performance of low embankment on end-bearing piles, the
influence of traffic load on arching effect, settlement on floating piles under traffic load. The
most important findings of the research are summarized in the following sections and then
followed by recommendations for further research.

8.1. Clonclusions
The research that has been presented is aimed to establish the behaviour of pavement material
under traffic load over soft soil that being supported by piles including both with
geosynthetics reinforcement and without geosynthetics reinforcement. The findings have
been discussed in the previous chapters, and they can be concluded in several topics below:

The effective distance of geosynthetics for reinforcement

 Vertical distance of geosynthetics depends on the stiffness of pavement material.
Pavement material with low elastic modulus around of 200 MPa (e.g. sandy gravel)
needs at maximum distance of 36.7 cm each other of geosynthetics layers. Whereas,
higher modulus elastic around of 400 MPa (e.g. crushed stone) can achieve a longer
distance around of 50 cm to support the same load.
 Hardening Soil (HS) model can predict better than Mohr Coulomb (MC) model, in
this case in which it can reveal the influence of number of layer and pattern of
horizontal displacement.
 The higher value of the elastic modulus, higher friction angle and higher value of the
dilation angle are able to reduce horizontal strain.

Analysis of stress concentration ratio on high compressibility of soil

 The higher embankment fill will increase the stresses at the top surface of piles for
both low embankment (h/s≤1.4) and high embankment (h/s>1.4).
 During consolidation process because of high compressibility of soft soil, stress at the
top surface of the pile will increase up to a certain level and then decrease slowly and
stable afterwards. It is in contrary with the vertical stress at the surface of subsoil. In
the test, peak situation is developed when pile-subsoil relative displacement (Sc/s) is
between 8 mm and 13 mm. By using the pile clear spacing of 60 cm, it means that the
critical values range from 1.3% to 2.3%.
 In the test in which the sample has low relative density, MC-model can be well
performed to predict stress concentration ratio. When estimating the stress
concentration ratio, Terzaghi’s model always suggests overestimate than others, and
otherwise BS8006, but Low’smodel provides quite close with the test model.

161
  From laboratory work of Ping et al. (2007) and FE-calculation, h/s value around of
1.4 can be used to distinguish between low embankment and high embankment based
on the equal plane strain at the surface of the embankment.

Various types of load transfer platforms on floating columns

 Load transfer platform (LTP) can be a catenary (max. 2 geosynthetics layers), a beam
(at least 3 geosynthetics layers) and the reinforced concrete. All of them will give
high total settlements when being applied on the floating piles, even only within 6
months of observation time, namely:
- Catenary LTP: 181 mm
- Beam LTP: 147 mm at Section 1 and 119 mm at Section 2
- Reinforced concrete: 132 mm
- Without Piles: 198 mm at Control 1 and 628 mm at Control 2
 Differential settlement at the base of the embankment (difference settlement between
at top surface of column and surface of subsoil) is quite small, namely:
- Catenary LTP: 29 mm
- Beam LTP: 4 mm at Section 1 and 11 mm at Section 2
- Reinforced concrete: 16 mm
- Without Piles: 0 mm
 To model embankment over floating piles using FE-calculation, the use of SSC-model
for soft soil and consolidation calculation type without load during compaction
process can be taken into account.
 For construction with load transfer platform using more than a layer, the lower
geosynthetics is deformed deeper than the upper layer.

The low embankment performance on end-bearing piles

 The differential settlement at the base of the embankment using the end-bearing piles
on the shallow soft soil will be relatively greater than on floating piles. By using MC-
model or HS-model for pavement material and SSC-model for soft soil, plastic
calculation type with involving compaction load agrees with result of the field test.
 Regarding with zones for geosynthetics reinforcement, maximum tensile strength of
geosynthetics is occurred at the zone of adjacent piles and at the half distance between
piles with parallel to pile face. It is primary reinforcement, whereas other as
secondary reinforcement.

The influence of traffic load on arching effect

 Traffic load at the surface of pavement is always changing by time. By observing the
change of the vertical load on top piles for both upper and lower part of geosynthetics
inserted at the base of the embankment of Kyoto road, it is possible to monitor the
change of vertical stress over time. The vertical stress at top pile and on surface of
subsoil would be increased during surcharge load. It means that shape of soil arching
was changing.
 By using HS model for granular material of the pavement and SSC model for soft soil
with plastic calculation type and traffic load around of 550 kPa at the surface of the
pavement, this model agrees with the field test. Furthermore, for calculating the
vertical load directly transferred on top of piles, BS8006 method is better than
162
 EBGEO method, but for vertical stress over subsoil. Moreover, EBGEO method gives
better results than BS 8006 method.
 Another important finding is that the soil arching is changing when surface of
pavement is subjected to a traffic load and pavement material is able to restore its self
to an initial phase during a rest period.

The settlements on floating piles under traffic load

 Total settlement on construction over soft soil using floating piles is too huge.
Though, there is no traffic load. Because of creep, settlement still happens over a long
time.
 SSC-model for soft soil with the consolidation calculation type can be accurately
applied for predicting the settlement of a pavement construction over soft soil
supported by the floating piles.
 In practice, always keeping value of s/d <5 to obtain block behaviour in pile group.
 Relative settlement ratio (RSR) can be used to describe the effectiveness of pile
length when constructing the floating pile group. In this case of the apron at Supadio
airport, the ratio of pile length to soft soil thickness is around 0.20.

Furthermore, for practical purpose it is important to note here that main conclusions below
are utmost important thing to keep in mind. Firstly, height of the embankment must be more
than 1.4 times of clear spacing between piles to avoid differential settlement at the surface of
the embankment. Secondly, because floating piles foundation works as a pile group, it is
important to remember that ratio pile spacing to pile diameter must be kept always less than
5. Thirdly, the use of small size for diameter of the pile is an useful way when being applied
to low embankment construction. Last thing, when we construct an embankment over soft
soil supported by floating piles using FEM and estimate settlement because of creep,
consolidation calculation type must be run without taking into consider a load compaction
during construction period. However, plastic calculation type can be used to predict
settlement with involving a static traffic load. This type of calculation is relatively faster than
consolidation calculation type.

8.2. Recommendations for Further Research

Further research related to behaviour of embankment material over soft soil under the traffic
load for both end-bearing piles and floating piles is required to investigate the following
topics:
 The embankment supported by basally piles reinforcement is widely used for shallow
depth of soft soil but floating piles is rear encountered in the literature study. It is the
promising realm to be developed in the future for treating it on deep soft soil.
 Resistance of soil arching under cyclic loading and long-term condition for cemented
material is necessary to be observed. Because in some circumstances the pavement
material is stabilized with cement.
 It is also a good idea to study soil interaction with inclusion geosynthetics on a
cemented material using pull-out test.

163
 The phenomenon of increasing bearing capacity of displacement piles along the
increasing time after installation is well known and these findings are based on some
empirical field experiments. However, its mechanical behaviour has not been fully
understood yet.

164
 REFERENCES

Adhi, W. (1984), ’’Masalah Tanaman di Tanah Gambut’’, Prosiding Pertemuan Teknis

Penelitian Pola Usahatani Menunjang Transmigrasi, Badan Penelitian dan
Pengembangan Pertanian, Jakarta.
Almeida, M.S.S., Ehrlich, M., Spotti, A.P. and Marques, M.E.S. (2007), ’’Embankment
supported on piles with biaxial geogrids”, Journal of Geotechnical Engineering,
Institution of Civil Engineers, ICE, UK, Vol. 160, No. 4, pp. 185-192.
Almeida, M.S.S., Marques, M.E.S., Almeida, M.C.F., and Mendonca, M.B. (2008),
’’Performance of Two ‘Low’ Piled Embankments with Geogrids at Rio de Janeiro’’,
The First Pan American Geosynthetics Conferences & Exhibition, Cancun, Mexico,
pp.1285-1295.
Alphan, I. (1970), ’’The geotechnical properties of soils‘‘, Earth-Science Review, Vol.6, pp.
5-49.
Althoff, S. (2010)‚ ’’Contribution to use Multi-stage Large Shear-Frame tests to Investigate
the Interaction-Behaviour between Geosynthetics and Cohesive Soil’’, Technische
Universität Bergakademie, Freiberg.
Althoff, S. et al. (2011), ’’Bestimmung der verbundeigenschaften-innovatives
versuchskonzept mit 12 geogittern’’, FS-KGEO 2011, München.
Althoff, S. (2011), ’’Forschung am Institut für Geotechnik, neue Erkenntnisse beim Verbund-
& Systemverhalten von Geokunststoffen in bindigen Böden’’, Geotechnisches
Seminar, January 2012, Technische Universität Bergakademie, Freiberg.
Alwi, A. (2007), ’’Ground Improvement on Malaysian Peat Soils Using Stabilised Peat-
Column Techniques’’, PhD thesis, University of Malaya, Kuala Lumpur, Malaysia.
Anonim (1978), ’’Peraturan Konstruksi Kayu Indonesia’’, NI-5 PKKI 1961, Penerbit
Departemen Pekerjaan Umum, Direktorat Jenderal Cipta Karya, Direktorat
Penyelidikan Masalah Bangunan, Bandung, Indonesia.
Archer, S. (2008), ’’Subgrade Improvement for Paved and Unpaved Surefaces using
Geogrids’’, Tensar International, Construction Products inc.
Atkinson, J.H. and Sällfors, G. (1991), ’’Experimental determination of soil properties’’,
Proceedings of 10th ECSMFE, Vol.3, pp. 915-956, Florence.
Atkinson, J.H. (2000), ’’Non-linear soil stiffness in routine design’’, Geotechnique, Vol. 50,
pp. 487-508.
Aydogmus, T. and Klapperich, H. (2004), ’’The next generation: Enhanced shear-pullout-
testing device’’, ECI int’l conference, Schloss Pillnitz, Dresden.
Aydogmus, T. (2006), ’’Beitrag zum Interaktionsverhalten von Geokunstoff und
Lockergestein’’, Dissertation, TU-Bergakademie Freiberg.
Baker, S. (2000),’’Deformation Behaviour of Lime/Cement Column Stabilized Clay’’,
Swedish Deep Stabilization Research Centre, Rapport 7, Chalmers University of
Technology.
Barenberg, E. J., Hales, J. and Dowland, J. (1975), ’’Evaluation of Soil-AggregateSystems
with MIRAFI Fabric”, University of Illinois Report No. UILU-ENG-75-2020,
prepared for Celanese Fibers Marketing Company.
Barenberg, E.J. (1992), ’’Subgrade stabilization, in A Design Primer’’, Geotextiles and
Related Materials, IFAI, 10-20.
Barry, A. J., Bambang, T., Symes ,T. & Younger, J. S. (1992), ’’Piling for roads over peat in
Sumatra’’, Geotropika 92 International Conference on Geotechnical Engineering,
Johor Bahru.
Barry, A. J. & Rachlan, A. (2001), ’’Embankments on soft soils in North Java’’, Proceedings
International Conference on In Situ Measurement of Soil Properties and Case
Histories, Bali, Parahyangan Catholic University, Indonesia.
Bauer, E.(1996), ’’Calibration of a comprehensive constitutive equation for granular
Materials’’, Soils and Foundations 36, pp. 13–26.
BB Litbang (2008), ’’Konsorsium Penelitian dan Pengembangan Perubahan Iklim pada
Sektor Pertanian’’, Balai Besar Penelitian dan Pengembangan Sumberdaya Lahan
Pertanian, Laporan Tahunan 2008, Bogor.
Bell, A. L., Jenner, C., Maddison, J.D. and Vignoles, J. (1994), ’’Embankment support using
geogrids with vibro concrete columns’’, Proceedings, 5th International Conference on
Geotextiles, Geomembranes, and Related Products, 335-338.
Bender, D.A. and Barenberg, E.J. (1980), ’’Design and Behavior of Soil-Fabric-Aggregate
Systems’’, Transportation Research Record 671, TRB, National Research Council,
Washington DC, USA, pp. 64-75.
Benz, T. (2007), ’’Small strain stiffness of soils and its numerical consequences’’, PhD
Thesis, Institute of Geotechnical Engineering, University of Stuttgart.
Bergado, D.T. and Long, P.V.(1994),‘‘Numerical analysis of embankment on subsiding
ground improved by vertical drains and granular piles‘‘, Proceeding of the XIII
International Conferences on Soil Mechanics and Foundation Engineering, Vol.4,
Morgantown, Balkema, pp.1361-1366.
Bergado, D. T., Ruenkrairergsa, T., Taesiri, Y., and Balasubramaniam, A.S. (1999), ’’Deep
soil mixing used to reduce embankment settlement’’, Ground Improvement 3:145-162.
Beskou, N.D and Theodorakopoulos, D.D.( 2011), ’’Dynamic effects of moving loads on road
Pavement’’: A review, Soil Dynamics and Earthquake Engineering 31(2011), Elsevier
Ltd, p.547-567.
Biarez, J. And Hicher, P.-Y. (1994), ’’Elementary Mechanics of Soil Behaviour’’, Balkema.
Bjerrum, L. (1967), ’’Engineering geology of Norwegian normally-consolidated marine clays
as related to settlements of buildings’’, Geotechnique, 17,2, p.81-118.
Bouckovalas, G.,Whitman, R.V. and Marr, W.A.(1984), ’’Permanent displacement of sand
with cyclic loading’’, Journal of Geotechnical Engineering, ASCE, 110(11): pp.1606-
1623.
Broms, B. B. (2003), ’’Deep soil stabilization: design and construction of lime and
lime/cement Columns’’, Royal Institute of Technology, Stockholm, Sweden.
Broms, B. B., and Boman, P. (1979), ’’Stabilization of soil with lime columns’’,
GroundEngineering 12, no. 4: 23-32.
BS8006 (1995), ’’Berehcmung und Dimensionierung von Erdkörpern mit Bewehrungens-
einlagen aus Geokunststoffen’’, Deutsche Gesellscahft für Geotechnik e.V.
Burland, J.B. (1989),’’Small is beautiful-the stiffness of soils at small strains’’, Canadian
Geotechnical Journal, 26, pp.499-516 .
Burmister, D. M. (1943), ’’The Theory of Stresses and Displacements in Layered Systems
and Applications to the Design of Airport Runways”, Proceedings, Highway Research
Board, Vol. 23, pp. 126-144.
Bussert, F. (2006), ’’Verformungsverhalten geokunststoffbewehrter Erdstützkörper
Einflussgrößen zur Er-mittlung der Gebrauchstauglichkeit’’, Schriftenreihe des
Instituts für Geotechnik und Markscheidewesen der TU Clausthal. Heft 13/2006.
Bussinesq (1936), ’’Stress distribution in elastic solids’’, Proceeding of 1. ICSMFE
Cambridge, Vol. 2, S. 157.
Bustamante, M.& Gianeselli, L. (1981), ’’Prévision de la capacité portante des pieux isolés
sous charge verticale. Règles pressiométriques et pénétrométriques (Prediction of the
bearing capacity of single piles under vertical load. Pressuremeter and penetrometer
rules)’’, Bull. Liaison Labo. P. et Ch., no.113, May-June, Réf. 2536, 83-108 (in
French).
Butterfield, R.(1979), ’’A natural compression law for soils’’, Geotechnique 29.
Cabalar, A.F. and Cevik, A. (2008), ’’Modelling damping ratio and shear modulus of sand-
mica mixtures using neural networks’’, Engineering Geology, doi: 10.1016/ j.enggeo.
2008. 08.005.
Cabalar, A.F. (2009), ’’Dynamic Properties of Various Plasticity Clay’’, EJGE Vol. 14,
Bund. B.
Cabalar, A.F. (2010), ’’Applications of the oedometer, triaxial and resonant column tests to
the study of micaceous sands’’, Engineering Geology 112 (21010), pp.21-28.
Carlsson, B. (1987), ’’Reinforced soil, principles for calculation’’, Terratema AB, Linköping
(in Schwedish).
Carlsten, P. and Ekstrom, J. (1997), ’’Lime and lime cement columns, guide for project
planning, construction and inspection’’, Swedish Geotechnical Society, SGF Report
4:95 E.
CDIT (Coastal Development Institute of Technology). (2002), ’’The Deep Mixing Method:
Principle, Design and Construction’’, A.A. Balkema: The Netherlands.
Chaboche, J.L.(1994), ’’Modelling of ratchetting: evaluation of various approaches’’,
European Journal of Mechanics, 13(4):501-781.
Chen, F.H. (1975), ’’Foundation on expansif soils’’, Elsevier Scientifik Publishing Company,
Amsterdam .
Clayton, C.R.I., Milititsky, J. & Woods, R.I.(1993), ’’Earth Pressure and Earth-Retaining
Structures”, Blackie Academic & Professional, Glasgow, UK, 2nd Edition
Collin, J.G., Watson, C.H., Han, J. (2004), ’’Column-Supported Embankments
DesignConsideration’’, Proceedings of Minnesota 52nd Annual Geotechnical
Engineering Conference, Minneapolis, Minnesota.
Collin, J.G.; Watson, C.H.; Han, J. (2005), ’’Column-Supported Embankments solves time
Constraint for new road construction’’ , Proceedings of the Geo-Frontiers 2005
Congress, Austin, Texas.
Cox, J. B. (1970), ’’The distribution and formation of recent sediments in South East Asia’’,
Proceedings of 2nd South East Asian Conference on Soils Engineering, p.30-47.
Das,B.M. (1985), ’’Principles of Geotechnical Engineering’’, The Univesity of Texas at El
Paso, USA.
Dawson, A., Kolisoja, P. And Vuorimies, N. (2008), ’’Understanding Low-Volume Pavement
Response to Heavy Traffic Loading’’, Roadex III Project: Northern Periphery, The
Swedish Road Administration.
DB Netz AG, (1999), ’’Ril 836 – Erdbauwerke planen, bauen und instand halten’’, Fassung
vom 20.12.1999.
Djarwadi, D. (2006), ’’Konstruksi Jalan di Atas Tanah Lunak dengan Perkuatan Geotextile
(Geotextile Reinforcement for Road Construction on Soft Soil)’’, International Civil
Engineering Conference ‘Towards Sustainable Civil Engineering Practice’, Surabaya,
Indonesia.
Dunlap, W.S. (1963), ’’A report on a mathematical model describing the deformation
characteristics of granular materials’’, Technical Report 1, Project 2-8-62-27, TTI,
Texas A & M University.
EBGEO (2010), ’’Berehcmung und Dimensionierung von Erdkörpern mit
Bewehrungenseinlagen aus Geokunststoffen’’, Deutsche Gesellscahft für Geotechnik
e.V.
Eekelen, S.J.M. van, Bezuijen, A. (2008a), ’’Design of piled embankments, considering the
basic starting points of the British Standard’’, Proceedings of EuroGeo4, Edinburgh.
Eekelen, S.J.M. van, Bezuijen, A. Alexiew, D. (2008b), ’’Piled Embankments in the
Netherlands: a full-scase test, comparing 2 years of measurements with design
calculations’’, Proceedings of EuroGeo4, Edinburgh.
Eekelen, S.J.M. van ,Bezuijen, A. , Duijnen, P. Van and Jansen, H.L. (2009), ’’Piled
embankments using geosynthetic reinforcement in the Netherlands:design, monitoring
& evaluation’’, Proceedings of the 17th International Conference on Soil Mechanics
and eotechnical Engineering M. Hamza et al. (Eds.) pp. 1690-1693.
Eekelen, S.J.M. van and Bezuijen, A. (2010), ’’The Kyoto road piled embankment: 3 ½ years
of measurements’’, 9th International Conferences on Geosynthetics, Brazil.
Eekelen, S.J.M. van, Bezuijen, A., and Tol, A.F. van (2011), ’’Analysis and modification of
the British Standard BS8006 for the design of piled embankments’’, Geotextiles and
Geomembranes 29 (2011) pp. 345-359.
Esrig, M. I., and MacKenna, P.E. (1999), ’’Lime cement column ground stabilization for I-15
in Salt Lake City’’, 34th Annual Symposium on Engineering Geology and
Geotechnical Engineering.
Etezad, M., Hanna, A.M. and Ayadat, T. (2006), ’’Bearing capacity of groups of stone
columns’’, Proceedings of the 6thEuropean Conference on Numerical Methods in
Geotechnical Engineering, Graz, pp. 781-786.
EuroSoilStab (2002), ’’Development of design and construction methods to stabilise soft
organic soil’’, Design Guide Soft Soil Stabilization, CT97-0351 Project No. BE 96-
3177.
Exxon (1989), ’’Designing for Soil Reinforcement’’, Exxon Chemical, GeoPolymers Ltd.
Fang, Y. S., Chung, Y. T., Yu, F. J., and Chen, T. J. (2001),’’Properties of soil-cement
stabilised with deep mixing method’’, Ground Improvement 5, no. 2: 69-74.
Farrell, E. R., O'Neill, C. & Morris, A. (1994), ’’Changes in the mechanical properties of
soils with variations in organic content’’, Advances in Understanding and Modelling
the Mechanical Behaviour of Peat, Balkema Rotterdam.
Fox, L. (1948), ’’Computation of Traffic Stresses in a Simple Road Structure”, Proc. 2nd Int.
Conf. Soil Mech. Found. Eng. Vol. 2, pp. 236-246.
Garlanger, J. (1972), ’’The consolidation of soils exhibiting creep under constant effective
Stress’’, Geotechnique, Vol.22, pp. 71-78.
GEC (2012), ’’Global Environment Centre’’, www.peat-portal.net
Giroud, J.P. and Noiray, L. (1981), ’’Design of geotextile reinforced unpaved road’’, Journal
of Geotechnical Engineering, ASCE, 107, 1233-1254.
Giroud, J. P., Ah-Line, C., and Bonaparte, R. (1984), ’’Design of Unpaved Roads and
Trafficked Areas with Geogrids”, Polymer Grid Reinforcement: A Conference
Sponsored by SERC and Netlon, Ltd., Thomas Telford, London, England, pp.116-
127.
Giroud, J.P., 1995, ’’Determination of Geosynthetic Strain due to Deflection”, Geosynthetics
International, Vol. 2, No. 3, pp. 635-641.
Giroud, J.P., and Han, J., 2004, ’’Design Method for Geosynthetic-Reinforced Unpaved
Roads:Part I–Development of Design Method”, Journal of Geotechnical and
Geoenvironmental Engineering, in press, ASCE.
Gorsel, J.T.v. (2012), ’’Bibliography of the Geology of Indonesia and Surrounding Areas’’,
Edition 4.1., www.vangorselslist.com.
Grundmann, H.(1999), ’’Dynamic interaction of structures with the subsoil’’, In:Frybal,
Naprstek J., editors. Structural Dynamics—EURODYN’99, 1.Rotterdam:AA
Balkema; 1999, p.31–41.
Gudehus, G. (1996), ’’A comprehensive constitutive equation for granular materials’’, Soils
and Foundations 36, pp. 1–12.
Guido, V. (1987), ’’Plate loading tests on geogrid-reinforced earth slabs’’, Proceedings of
Geosynthetics 1987 Conf., New Orleans.
Haley & Aldrich and Virginia Geotechnical Services, P.C. (2000), ’’Final Data Report on
Ground Improvement for Embankment Construction’’, I-95/Route 1 Interchange,
VDOT Project No. 0095-96A-106, PE-101.
Han, J. (1999), ’’Design and Construction of Embankments on Geosynthetic Reinforced
Platforms Supported by Piles’’, Invited Speaker, Proceedings, 1999 ASCE
Geotechnical Seminar, Hershey, PA.
Han, J., and Wayne, M.H,(2000), ’’Pile-Soil-Geosynthetic Interactions in Geosynthetic
Reinforced/Piled Embankments over Soft Soil’’, Presentation and CD-Rom Paper at
79th Annual TRB meeting.
Han, J., Gabr, M.A. (2002), ’’Numerical Analysis of Geosynthetic-Reinforced and Pile-
Supported Earth Platforms over Soft Soil’’, Journal of geotechnical and
geoenvironmental engineering, Vol. 128, No. 1, S. 44-53.
Han, J. (2003), ’’Development of Design Charts for Geosynthetically Reinforced
Embankments on Deep Mixed Columns’’, Interim Report, Literature Review,
Principal Investigator, funded by the FHWA National Deep Mixing Program, 2002-
2003, Widener University, Chester, PA.
Han, J., Collin, J.G. (2005), ’’Geosynthetic Support Systems over Pile Foundations’’,
Proceedings of the Geo-Frontiers 2005 Congress, Austin, Texas.
Handy, R.L. (1985), ’’The arch in soil arching’’, ASCE Journal of Geotechnical Engineering,
Vol. 111, No. 3, pp. 302-318.
Hardin, B.O. and Drnevich, V.P. (1972), ’’Shear Modulus and Damping in Soils:
Measurement and Parameter Effects”, Journal of Soil Mechanics and Foundations,
ASCE, 98, SM6, 603-624.
Hassandi, A.C. and Edil, T.B. (2005), ’’Full scale Geosynthetic-Reinforced Column-
Supported Test Embankment’’, University of Wisconsin-Madison.
Hassandi, A.C. and Edil, T.B. (2007), ’’Behaviour of Geogrid-reinforced load transfer
platforms for embankment on rammed aggregate piers’’, Geosynthetics International,
Vol. 14 No. 3. Pp. 141-153.
Heitz, C. (2006), ’’Bodengewölbe unter ruhender und nichtruhender Belastung bei
Berücksichtigung von Bewehrungseinlagen aus Geogittern’’, Schriftenreihe
Geotechnik, University of Kassel, Heft 19.
Hencky, H. (1928), ’’Über die form des elastizitätsgesetzes bei ideal elastischen stiffen’’,
Zeitschrift fur Technische Physik 9, pp. 214–247.
Herle, I. and Gudehus, G.(1999), ’’Determination of parameters of a hypoplastic constitutive
model from properties of grain assemblies’’, Mechanics of Cohesive-Frictional
Materials 4: pp. 461–486.
Heukelom, W. and A.J.G. Klomp (1962), ’’Dynamic Testing as a Means of Controlling
Pavements During and After Construction”, Proceedings of the First International
Conference on Structural Design of Asphalt Pavements, University of Michigan.
Hewlett, W.J., Randolph, M.F., Aust, M.I.E. (1988), ’’Analysis of piled embankments’’,
Ground Engineering Vol.21, pp. 12-17.
Hill, R. (1950), ’’A general theory of uniqueness and stability in elastic-plastic solids’’,
Journal of the Mechanics and Physics of Solids 6, pp. 236–249.
Hobbs, N.B. (1986), ’’Mire Morphology and the Properties and Behaviour of some British
and Foreign peats’’, Quarterly Journal of Engineering Geology, London, Vol. 19,
pp.7-80.
Hobbs, N. B. (1987), ’’A note on the classification of peat’’, Geotechnique, 37, 3, p. 405-407.
Holm, G. (1999), ’’Keynote lecture:Applications of dry mix methods for deep soil
stabilization’’, International Conference on Dry Mix Methods: Dry Mix Methods for
Deep Soil Stabilization, Balkema, Rotterdam: 3-13.
Holtz, R.D & Sivakugan, N. (1987), ’’Design charts for roads with geotextiles’’, Journal of
Geotextile and Geomembranes, 5, pp. 191-199.
Huang, Y. H. (1969), ’’Computation of Equivalent Single-wheel Loads Using Layered
Theory’’, Proceedings, Highway Research Board.
Hughes, J.M.O. and Withers, N.J. (1974), ’’Reinforcing of soft cohesive soils with stone
columns’’, Ground Engineering, Vol. 7, No. 3, pp 42-49.
ICAO (1999), ’’Aerodrome Design and Operations’’, Aerodrome Standards, Juli 1999,
COSCAP South Asia.
Iglesia, G.R., Einstein, H.H. & Whitman, R.V. (1999), ’’Determination of vertical loading on
underground structures based on an arching evolution concept’’, Proceedings 3rd
National Conference on Geo-Engineering for Underground Facilities, pp. 495-506.
Indonesian Public Works (1999), ’’Guidance of timber pile foundation over soft soils and
peat soils’’, (translated), Publisher Mediatama Saptakarya, Jakarta.
Irsyam, M. & Krisnanto, S. (2008), ’’Pengujian Skala Penuh dan Analysis Perkuatan
Cerucuk Matras Bambu untuk Timbunan Badan Jalan Diatas Tanah Lunak di Lokasi
Tambak Oso Surabaya’’, Forum Teknik Sipil, Universitas Gadjah Mada, Vol.
XVIII/1, Januari, 2008.
Jacobson, J. R., Filz, G. M. and Mitchell, J. K. (2003), ’’Factors Affecting Strength Gain in
Lime-Cement Columns and Development of a Laboratory Testing Procedure’’, Report
prepared for the Virginia Transportation Research Council, Virginia Polytechnic
Institute and State University.
Jafari,M.K., Shafiee, A., Razmkhah, A. (2002), ’’Dynamic Properties of Fine Grained Soils
in South of Tehran’’, JSEE: Spring 2002, Vol.4, No.1, pp. 25-35.
Jenner, C. G., Austin, R. A., and Buckland, D. (1998), ’’Embankment support over piles
using Geogrids’’, Proceedings of 6th International Conference on Geosynthetics Vol.
1: pp.763-766.
John, N. W. M. (1987), ’’Geotextiles’’, Glasgow and London: Blackie & Son Ltd.
Jones, C. J. F. P., Lawson, C. R., and Ayres, D. J. (1990), ’’Geotextile reinforced piled
Embankments’’, Proceedings, 5th International Conference on Geotextiles,
Geomembranes, and Related Products, 155-60.
Kaiqiu, L. (2000), ’’Behaviour of DCM Columns under Highway Embankment at Bridge
Approaches’’, Thesis for the Degree of Doctor of Philosophy, Nanyang Technical
University.
Kawasaki, T., Nina, A., Saitoh, S., Suzuki, Y., Honjyo Y. (1981), ’’Deep Mixing Method
Using Cement Hardening Agent’’, Proceedings of the 10th International Conference
on Soil Mechanics and Foundation Engineering, Stockholm: pp.721-24.
Kempfert, H.-G., Stadel, M., Zaeske, D. (1997), ’’Design of geosynthetic-reinforced bearing
layers over piles’’, Bautechnik 74 Heft 12: 818-25.
Kempfert, H.G., Hu, Y. (1999), ’’Measured dynamic loading of railway underground’’,
Proceedings of the XI. Panamerican Conference on Soil Mechanics and Geotechnical
Engineering, Foz do Iguacu, Brazil.
Kempfert, H.G., Hu, Y. (1999), ’’Numerical modeling of the deformation in railway
foundation– A case study’’, Proceedings of the International Symposium on
Numerical Models in Geomechanics, NUMOG VII, Graz, A.A. Balkema, S 669-674.
Kempfert, H.-G., Gobel, C., Alexiew, D., and Heitz, C. (2004), ’’German recommendations
for reinforced embankments on pile-similar elements’’, EuroGeo3 - Third European
Geosynthetics Conference, Geotechnical Engineering with Geosynthetics, 279-84.
Kempfert, H.G. and Raithel, M. (2005), ’’Soil Improvement and Foundation Systems with
Encased Columns and Reinforced Bearing Layers”, Elsevier Geo-Engineering Book
Series, Volume 3, pp. 923-946.
Kempton, G., and Naughton, P. (2002), ’’Piled embankments with basal reinforcement:
development of design methods and state of the art’’, Proceedings of the XV Italian
Conference on Geosynthetics.
Kenney, J.T. (1954), ’’Steady-state vibrations of beam on elastic foundation for moving
load’’, Journal Appl. Mech. ASCE 1954;21:359–364.
Kenney, T. C. (1976), ’’Formation and geotechnical characteristics of glacial-lake varved
soils’’, In Laurits Bjerrum Mem Vol, eds N Janbu, F Jorsted & B Kjaernsli, pp.15-39,
Norwegian Geotech Inst, Oslo.
Kepmenhub (2000), ’’Keputusan Menteri Perhubungan Tentang Jalur Kereta Api’’, KM 52
Tahun 2000, Departement Perhubungan republik Indonesia.
Kim, S.M. and Roesset, J.M. (2003), ’’Dynamic response of a beam on a frequency-
independent damped elastic foundation to moving load’’, Can. J. Civ. Eng.
2003;30:460–7.
Kivelo, M. (1997), ’’Undrained shear strength of lime/cement columns’’, Proceedings, 14th
International Conference on Soil Mechanics and Foundation Engineering, 1173-80.
Kivelo, M. (1998), ’’Stabilization of embankments on soft soil with lime/cement columns’’,
Doctoral Thesis, Royal Institute of Technology.
Klapperich, H. (2004), ’’The Use of Geosynthetics in Soil Reinforcement and Dynamics’’,
ECI International Conference, Schloss Pilnitz, Dresden.
Klompmaker, J. and Bishop, D. (2008), ’’Improving the bearing capacity of soils with
Geosynthetics’’, NAUE GmbH & Co. KG, Germany.
Koerner, R.M. (2005), ’’Designing with Geosynthetics’’, 5th Edition, Pearson Prentice Hall,
pp. 331-348.
Koiter, W.T. (1960), ’’General theorems for elastic-plastic solids’’, Sneddon&Hill, Eds,
Progress in Solid Mechanics, North Holland Publishing Co., Amsterdam, pp.165-221
Kolymbas, D. (1991), ’’An outline of hypoplasticity’’, Archive of Applied Mechanics 61, pp.
143–151.
Kondner, R.L. & Zelasko, J.S. (1963), ’’A hyperbolic stress-strain formulation for sands’’,
Proc. 2nd Pan-American Conference of Soil Mechanics and Foundation Engineering,
Vol.1, pp. 289-394.
Krynine, D.P. (1945), ’’Discussion of Stability and stiffness of cellular cofferdams’’,
Transactions, ASCE, Vol. 110, pp. 1175-1178.
Landva, A. O., et al (1983), ’’Geotechnical classification of peats and organic soils’’, in
Testings of Peats and Organic Soils, ASTM STP 820 ed. P Jarrett, 37-51.
Lawson, C. R. (1992), ’’Soil reinforcement with Geosynthetics, Applied Ground Improvement
Techniques’’, Southeast Asian Geotechnical Society (SEAGS): 55-74.
Low, K., Tang, S.K., Choa, V. (1994), ’’Arching in Piled Embankments’’, Journal of
geotechnical engineering, Vol. 120, No. 11, S. 1917-1938.
Maddison, J.D., Jones, D.B., Bell, A.L. & Jenner, C.G. (1996), ’’Design and performance of
an embankment supported using low strength geogrids and vibro concrete columns,
 Geosynthetics – Applications’’, Design and Construction, De Groot, Den Hoedt &
Termaat, pp. 325-332.
Mandolini, A. & Viggiani, C. (1997), ’’Settlement of piled foundations”, Géotechnique,
47(4): 791-816.
Marston, A.& Anderson, A.O. (1913), ’’The theory of loads on pipes in ditches and tests of
cement and clay drain tile and sewer pipe’’, Bulletin No. 31, Iowa Engineering
Experiment Station, Ames, Iowa.
Matsuoka, H. and Nakai, T. (1982), ’’A new failure for soils in three-dimensional stresses,
In: Deformation and Failure of Granular Materials’’, Proc. IUTAM Symp. In Delft,
pp. 253-263.
May, R. W. and Witczak, M.W. (1981), ’’Effective Granular Modulus to Model
PavementResponse”, Transportation Research Record 810, TRB, National Research
Council, Washington, D.C.
McKelvey, D., Sivakumar, V., Bell, A. and Graham, J. (2004), ’’Modelling vibrated stone
columns in soft clay’’, Proceedings of the Institute of Civil Engineers Geotechnical
Engineering, Vol. 157, Issue GE3, pp 137-149.
McKelvey, J.A. (1994), ’’The anatomy of soil arching’’, Geotextiles and Geomembranes 13,
pp. 317-329.
MacKinnon, K., Hatta, G., Halim, H. & Mangalik, A. (1996), ’’The Ecology of Indonesia
Series’’, Vol. III: The Ecology of Kalimantan; Periplus Editions (HK) Ltd.
McNulty, J.W. (1965), ’’An Experiment study of arching in sand’’, Technical Report No. I-
674, U.S. Army Engineer Waterways Experiment Station, Corps of Engineers,
Vicksburg, Mississippi, 170.
Meyerhof, G.G. and Hanna, A.M. (1978), ’’Ultimate bearing capacity of Foundation on
Layaered Soil under Inclined Load’’, Canadian Geotechnical Journal , Vol. 15 No. 4,
pp. 565-572.
Milovic, D. (1992), ’’Stress and Displacements for Shallow Foundations’’, Elsevier Science
Publisher, Amsterdam, The Netherlands.
Mirafi, T., (2010), ’’Geosynthetic Reinforcement of the Aggregate Base/Subbase courses of
Pavement Structures’’, Technical Note.
Moossazadeh, J. and Witczak, M.W. (1981), ’’Prediction of Subgrade Moduli for Soil that
Exhibits Nonlinear Behavior”, Transportation Research Record 810, TRB, National
Research Council, Washington, D.C., pp. 9-17.
Mroz, Z., Norris, V.A. and Zienkiewicz,O.C.(1978), ’’An anisotropic hardening model for
soils and its application to cyclic loading’’, International Journal For Numerical and
Analytical Methods in Geomechanics, 2: pp. 203-221.
Muench, S. (2006), ’’Pavement engineering’’, http://www.pavementinteractive.org
Muncke, M., Meinlschmidt, A., Hillig, J. (1999), ’’Neue Erkenntnisse und Lösungen im
Brücken-,Tunnel-, Erd- und Grundbau’’, Eisenbahntechnische Rundschau, Band 48,
Heft 10, S. Pp. 622-638.
Muslihat, L.(2003), ’’Teknik Pengukuran Bobot isi Tanah Gambut di Lapangan dan di
Laboratorium’’, Buletin Teknik Pertanian Vol. 8 Nomor 2, 2003.
Navin, M. P. and Filz, G. M. (2005), ’’Statistical Analysis of Strength Data from Ground
Improved with DMM Columns’’, Proceedings of the International Conference on
Deep Mixing, Stockholm.
Niemunis, A. & Herle, I. (1997), ’’Hypoplastic model for cohesionless soils with elastic
strain Range’’, Mechanics of Cohesive-Frictional Materials 2: pp. 279–299.
Niemunis, A., Wichtmann, T., Triantafyllidis, T. (2005), ’’A high-cycle accumulation model
for sand’’, Computers and Geotechnics, 32(4): pp. 245-263.
Nordic Guidelines (2003),‘‘Nordic Guidelines for Reinforced Soils and Fills‘‘, The Nordic
Geotechnical Societies Nordic Industrial Fund.
Ooi, T.A., Chan, S.F., Wong, S. N. (1987), ’’Design, Construction, and Performance of Pile
Supported Embankments’’, 9th Southeast Asian Geotechnical Conference: 2-1 to 2-12.
Palmeira, E.M., Tatsuoka, F., Bathurst, R.J., Stevenson, P.E. and Zornberg, J.G. (2008),
’’Advances in Geosynthetics Materials and Applications for Soil Reinforcement and
Environmental Protection Works’’, EJGE, Bouquet 08, pp. 1-37.
Park, D. & Stewart, H.E. (2001), ’’Suggestion of Empirical Equations for Damping Ratio of
Plastic and Nonplastic Soils based on the Previous Studies’’, Proc. Fourth Internat.
Conf. on Recent Advances in Geotech. Earthquake Eng. and Soil Dynamics, San
Diego, California, Paper No. 1.21.
Paul, M.A, & Barras, B.F. (1999), ’’Role of organic content in the plasticity of Bothkennar
Clay’’, Geotechnique (Technical Note), 49, 4, pp. 529-535.
Perkins, S.W. and Ismeik, M. (1997), ’’A Synthesis and Evaluation of Geosynthetic
Reinforced Base Layers in Flexible Pavements: Part I”, Geosynthetics International,
Vol. 4, No. 6, pp. 605-621.
Porbaha, A. (2000), ’’State of the art in deep mixing technology: Part IV’’, Design
Considerations, Ground Improvement 3: 111-125.
Potts, V.J. & Zdravkovic, L. (2008), ’’Assessment of BS8006: 1995 design method for
reinforced fill layers above voids’’, Proceddings of the 4th European Geosynthetics
Conference, Edinburgh, UK, Paper number 116.
Poulos, H.G. (1968), ’’Analysis of the settlement of pile groups”, Geotechnique,18:449-471.
Poulos, H. G., (1971), ’’Behaviour of Laterally Loaded Piles I-Single Piles”, Journal of
SoilMechanics and Foundation Division, ASCE, Vol. 97, No. 5, pp. 711-732.
Poulos, H.G. (1993), ’’Settlement of bored pile groups”, Proc. BAP II, Ghent, Balkema,
Rotterdam, pp. 103-117.
Poulos, H. GC and Davis, E. H. (1974), ’’Elastic Solutions for Soil and Rock Mechanics’’,
Wiley, New York.
Poulos, H.G. and Davis, E.H. (1980), ’’Pile foundation analysis and design’’, New York,
John Wiley.
Poulos, H.G. (1994a), ’’An approximate numerical analysis of pile-raft interaction”,
IJNAMG, 18, pp.73-92.
Poulos, H.G. (1994b), ’’Settlement prediction for driven piles and pile groups”, Spec.Tech.
Pub. 40, ASCE, 2, pp. 1629-1649.
Poulos, H.G. (2002), ’’Simplified Design Procedure for Piled Soil Raft Foundations”, Deep
Foundations 2002, ASCE, Reston, Virginia, 441-458.
Poulos, H.G. (2005), ’’Pile behaviour-Consequences of Geological and construction
imperfections”, 40th Terzaghi Lecture, Jnl. Geot. & Geoenv. Eng., ASCE, 131(5), pp.
538-563.
Priadi, E. (2008), ’’Behaviour of Tiang Tongkat Foundation over Pontianak Soft Organic
Soil using 3D-Finite Element Analysis’’, PhD Thesis, Geotechnical Institute of
Technische Universität Bergakademie Freiberg, Germany.
Priebe, H. J. (1993), ’’Design Criteria for Ground Improvement by Stone Columns’’,
Proceedings of the Fourth National Conference on Ground Improvement.
Priebe, H. J. (1995), ’’The design of vibro replacement’’, Ground Engineering, 28, No. 10:
pp. 31-37.
PrEN ISO 10318 (2001), ’’Geosynthetics-Geotextiles, geotextile-related products,
geomembranes and geosynthetic clay liners, Terms and their definitions’’, Trilingual
version, February 2001.
Puslitbang Prastrans (2001), ’’Timbunan Jalan pada Tanah Lunak’’, (Indonesian
Geotechnical Guidance-4 ) Panduan Geoteknik-4: Desain dan Konstruksi, WSP
International, Bandung, Indonesia.
Puslitbang Prastrans (2001), ’’Timbunan Jalan pada Tanah Lunak’’, (Indonesian
Geotechnical Guidance-1 ) Panduan Geoteknik-1: Proses Pembentukan dan Sifat-sifat
Dasar Tanah Lunak, WSP International, Bandung, Indonesia.
Puslitbang Prastrans (2002), ’’Timbunan Jalan pada Tanah Lunak’’, (Indonesian
Geotechnical Guidance-2 ) Panduan Geoteknik-2: Penyelidikan Tanah Lunak, Desain
& Pekerjaan Lapangan, WSP International, Departemen Permukiman dan Prasarana
Wilayah,Bandung, Indonesia.
Radjagukguk, B. (1991), ’’Utilization and management of peatlands in Indonesia for
Agriculture’’, Tropical Peat-Proceedings of the International Symposium on Tropical
Peatland, Kuching 6 - 10 May 1991, pp. 21-27.
Rahadian, H. (1992), ’’Comparison of Engineering Properties of Soft Marine Clays and Soft
Lacustrine Clays from Indonesia’’, M.Sc. Thesis, University of Strathclyde.
Rahadian, H. et al. (2001), ’’Field and laboratory data interpretation of peats at
Berengbenkel trial site’’, International Conference on In situ Measurement of Soil
Properties and Case Histories, Bali, Volume: Embankments on Soft Soils.
Raju, D.M. and Fannin, R.J. (1998), ’’Load-strain-displacement Response of geosynthetics in
Monotonic and Cyclic Pullout”, Canadian Geotechnical Journal, Vol. 35, pp. 946-
956.
Raju, V.R., Hari Krishna, R. and Wegner, R. (2004), ’’Ground improvement using Vibro
Replacement in Asia 1994 to 2004 – a 10 year review’’, Proceedings of 5th Int. Conf.
on Ground Improvement Techniques, Kuala Lumpur, Malaysia.
Raithel, M., Küster, V. & Lindmark, A., (2004), ’’Geotextile encased Column-a Foundation
System for Earth Structures”, Illustrated by a Dyke Project for a Works Extension
inHamburg, Nordic Geotechnical Meeting, Sweden.
Raithel, M., Kirchner, A., Schade, C. & Leusink, E., (2005), ’’Foundation of Constructions
On Very Soft Soils with Geotextile Encased Columns-State of the Art”, ASCE-
Conference Geo-Frontiers, Austin, USA.
Randolph, M.F. (1994), ’’Design methods for pile groups and piled rafts”, Proc. 13th Int.
Conf. S.M. & Found. Eng., 5, pp. 61-82.
Reid, W.M., Buchanan, N.W. (1983), ’’Bridge approach support piling; Proceedings of
Conference on Advances in Piling and Ground Improvement’’, London, S. 267-274.
Rodin, S. (1965), ’’Ability of a clay fill to support construction plant’’, Journal of
Terramchanics, 2, pp. 51-68.
Rogbeck, Y., Gustavsson, S., Södergren, I., Lindquist, D. (1998), ’’Reinforced Piled
Embankments in Sweden – Design Aspects’’, Proceedings of the 6th International
Conference on Geosynthetics, S. 755-762.
Roscoe, K.H. and Burland, J.B. (1968), ’’On the generalised stress-strain behaviour of wet
Clay’’, Engineering Plasticity, Vol.1, London, UK. Cambridge University Press, pp.
535-609.
Rosenqvist, I. Th. (1953), ’’Considerations on the sensitivity of Norwegian quickclays’’,
Geotechnique, 3, 5, pp. 195-200.
Rowe, P.W. (1962), ’’The stress-dilatancy relation for static equilibrium of an assembly of
particles in contact’’, Proc. Roy. Soc. A. 269, pp. 500-527.
Rowe, P.W. (1971), ’’Theoretical meaning and observed values of deformation parameters
for soil’’, Proc. of Roscoe memorial symposium, Foulis, Henley-on-Thames, pp. 143-
194.
Ruiken, A. & Ziegler, M. (2008), ’’Effect of Reinforcement on the Load Bearing Capacity of
Geosynthetic Reinforced Soil”, Proc., 4th European Geosynthetics Conf., igs,
Edinburgh, UK.
Ruiken, A., Ziegler, M., Vollmert, L. And Duzic, I. (2010), ’’Recent findings about the
confining effect of geogrids from large scale laboratory testing’’, 9th International
Conference on Geosynthetics, Brazil.
Russell, D., Naughton, P. J., and Kempton G. (2003), ’’A new design procedure for
piledEmbankments’’, Proceedings of the 56th Canadian Geotechnical Conference and
the NAGS Conference, 858-65.
Russel, D., Pierpoint, N., MacDonald, M. (1997), ’’An assessment of design methods for
piled Embankments’’, Ground Engineering, Nov. 1997, S. 39-44.
Sasongko, H. (1996), ’’Determination of Subgrade Moduli under Repeated Triaxial Loading
Tests’’, M.Sc Thesis, Bandung Institute of Technology, Indonesia.
Satibi, S. (2009), ’’Numerical Analysis and Design Criteria of Embankments on Floating
Piles’’, PhD Thesis, Mitteilung 62 des Institüts fur Geotechnik, Universität Stuttgart.
Sawicki, A. and Swidzinski, W. (1989), ’’Mechanics of a sandy subsoil subjected to cyclic
Loadings’’, International Journal for Numerical and Analytical Methods in
Geomechanics, 13: pp. 511-529.
Schanz, T., and Vermeer, P.A. (1996), ’’Angles of friction and dilatancy of sand’’,
Geotecnique 46, No.1, pp. 145-151.
Schanz, T., Vermeer, P.A., and Bonnier, P.G.(1999), ’’The hardening soil model:
Formulation and verification’’, Beyond 2000 in Computational Geotechnics-10 years
of Plaxis, Rotterdam, Balkema.
Schwarz, W. and Raithel, M. (2005), ’’Stabilization of soft organic soils with cement columns
using the Mixed-in-Place technique (MIP) for a railway embankment’’, Bauingenieur
Band 79.
Scottwillson (2009), ’’Bassaly Reinforced Platforms for Piled Embankments-Two Case
Studies’’, Joint UK-IGS EGGS Meeting London 08 Dec 2009,www.scottwilson.com.
Seed, H. B. and Chan, C.K. (1959), ’’Structure and Strength Characteristics of Compacted
Clays”, Journal of Soil Mechanics and Foundation Engineering Division, ASCE, Vol.
85, No. 5, pp. 87-128.
Seed, H.B., Mitry, F.G., Monismith, C.L., and Chan, C.K., (1967), ’’Prediction of pavement
deflection from laboratory repeated load tests’’, NCHRP Report 35, USA.
Seed, H.B. and Idriss, I. (1970), ’’Soil moduli and damping factors for dynamic response
analysis’’, Report, Eartquake Engineering Research Center, University of California,
Barkeley.
Sellmeijer, J. B. and Kenter, C. J. (1983), ’’Increase in Bearing Capacity due to the
Application of Geolon Fabric below Granular Base of Road’’, Delft Soil Mechanics
Laboratory, Delft.
Sellmeijer, J. B. (1990), ’’Design of Geotextile Reinforced Paved Roads and Parking Areas”,
Proceedings of the Fourth International Conference on Geotextiles, Geomembranes,
and Related Products, Balkema, Vol. 1, The Hague, Netherlands, pp. 177-182.
Simandjuntak, T. O. (1993), ’’Neogene tectonics and orogenis of Indonesian’’, Jurnal
Geologi dan Sumberdaya Mineral, Geological Research & Development Centre.
Skempton, A. W. (1970), ‘’The Consolidation of clays by gravitational compaction’’,
Quarterly Journal of Engineering Geology, p. 373-411.
Slocombe, B.C. & Bell, A.L. (1998), ’’Setting on a dispute: discussion on Russell and
Pierpoint paper’’, Ground Engineering,pp. 34-36.
Smith, I.M. and Griffith, D.V. (1982), ’’Programming the Finite Element Method”, Second
Edition. John Wiley & Sons, Chisester, U.K.
Smith, M.E. (2005), ’’Design of bridging layers in geosynthetic-reinforced column-supported
embankments”, PhD Thesis, Virginia Polythechnic Institute and State University.
SNI (2002), ’’Tata Cara Perencanaan Struktur Kayu Untuk Bangunan Gedung’’, Beta
Version, SK SNI 03-2000, Bandung, Indonesia.
Soong, T.-Y. & Koerner, R.M. (1997), ’’On the Required Connection Strength of
Geosynthetically Reinforced Walls”, Journal of Geotextiles and Geomembranes, Vol.
15, pp. 377-393.
Steward, J.E., Williamson, R., and Mohney, J. (1977), ’’Guidelines for Use of Fabrics in
Construction and Maintenance of Low-Volume Roads’’, USDA, Forest Service
Report PB-276 972, Portland, OR.
Suherman, M. et al. (2003), ’’Guidance for Roadway Construction over Expansive Soil’’,
Final Report, Puslibang Penelitian dan Pengembangan Prasarana Transportasi,
Bandung, Indonesia.
Suryolelono ,K.B. (1997), ’’Soil Stabilization and Reinforcement, Compaction, Accelerated
Consolidation’’, Penataran Soil Improvement and Reinforced Earth, University of
Lambung Mangkurat, Banjarmasin, Indonesia.
Tamáskovics, N. & Klapperich, H. (2010), ’’Verbundverhalten von Geokunststoffen in
bindigem Boden’’, 9.Sächsisches Bautextilien-Symposium BAUTEX 2010 Bauen mit
Geokunststoffen, Chemnitz, pp. 1-6.
Takenaka, D. and Takenaka, K. (1995), ’’Deep chemical mixing method-using cement
ashardening agent’’, Takenaka Corporation, Tokyo.
Tatsuoka, F. (1993), ’’Role of facing rigidity in soil reinforcing’’, Keynote Lecture, Proc. Int.
Sym. On Earth reinforcement Practice, IS Kyushu ’92. Balkemia, 2: 831-870.
Tatsuoka, F., Tateyama, M. Uchimura, T., Koseki, J. (1997), ’’Geosynthetic-Reinforced Soil
retaining Walls as Important permanent Structures’’, Geosynthetics International
1997, Volume 4, No. 2.
Tatsuoka, F., Tateyama, M., Mohri, Y. and Matsushima, K. (2007), ’’Remedial treatment of
soil structuress using geosynthetic-reinforcing technology’’, Geotextiles and
Geomembranes, Vol. 25, Nos. 4 & 5.
Tatsuoka, F. and Kobayashi, A. (1983), ’’Triaxial strength characteristics of cement-treated
soft clay’’, Proceedings of the 8th European Conference of SMFE, 421-26.
Tenax (2001), ’’Design of Soft Soil stabilization with Tenax geogrids’’, Technical Reference
GRID-DE-1.
Tenax (2001), ’’Design of unpaved roads with Tenax geogrids’’, Technical Reference
TDS007,GRID-DE-3.
Terashi, M. (2003), ’’The state of practice in deep mixing methods’’, Grouting and
GroundTreatment, Proceedings of the 3rd International Conference, ASCE
Geotechnical Special Publication No. 120, New Orleans: 25-49.
Terzaghi, K. (1943), ’’Theoretical Soil Mechanic’’, John Wiley & Sons, New York.
Thompson, W.E. (1963), ’’Analysis of dynamic behavior of roads subject to longitudinally
moving loads’’, Highway Res. Rec. 1963;No. 39:1–24.
Tomlinson, M.J. (1977), ’’Pile design and construction practice’’, London: Cement and
Concrete Association.
Uzan, J. (1985), ’’Characterization of Granular Material”, Transportation Research Record
345, Washington, D.C. 1022, TRB, National Research Council, Washington, D.C., pp
52-59.
Valanis, K.C and Lee, C.F. (1984), ’’Endochronic theory of cyclic plasticity with
applications’’, Journal of Applied Mechanics, 51:367-374.
Van Impe, W. F., Madhav, M. R., Vandercruyssen, J. P. (1997), ’’Considerations in stone
column design’’, Ground Improvement Geosystems, Densification and
Reinforcement, 190-196.
Vermeer, P. (1982), ’’A five constant model unifying well established concepts’’, In: Constit.
Relat. for Soils, Balkema, Holland, Proceedings of the International Workshop in
Grenoble, pp. 175-198.
Vermeer, P.A. and Neher, H.P.(1999)‚’’A soft soil model that accounts for creep’’,
Proceeding of International Symposium Beyond 2000 in Computational Geotechnics,
Balkema, pp. 249-261.
Von Soos, P. (2001), ’’Eigenscfahten von Boden und Fels-ihne Ermittung im Labor’’,
Gundbau-Taschenbuch, Vol.1. chapter 1.4, pp. 117-201, Ernst and Sohn, Berlin.
Webb, W.M., and Campbell, B.E. (1986), ’’Preliminary Investigation into Resilient Modulus
Testing for new AASHTO Pavement Design Guide”, Office of Materials and
Research, Georgia Department of Transportation.
Wei-ping, C.,Yun-min, C., Ren-peng, C. (2007), ’’An experimental investigation of soil
arching within basal reinforced and unreinforced piled embankments’’, Geotextiles
and Geomembranes 26 (2008), Science Direct, Elsevier, pp. 164-174.
Wichtmann, T.( 2005), ’’Explicit accumulation model for non-cohesive soils under cyclic
Loading’’, Doctoral Dissertation, Heft-38, Ruhr-Universität Bochum.
Widodo, S. et al. (2013), ’’Zwischenbericht des DFG-Forschungsprojektes:
Verbundverhalten von Geokunststoffen in bindigen Böden’’, TU-Bergakademie
Freiberg (unpublished).
Wilson and Huesker (2009), ’’Basally Reinforced Platforms for Piled Embankments-Two
Case Studies’’, Joint UK-IGS EGGS Meeting London 08 Dec 2009.
Wolffersdorff, P.A.v.(1996), ’’A hypoplastic relation for granular materials with a
predefined limit state surface’’, Mechanics of Cohesive-Frictional Materials 1: pp.
251–271.
Wood, D.M., Hu, W., Nash, D.F.T. (2000), ’’Group effects in stone column foundations-
model tests’’, Geotechnique, Vol. 50, No. 6, pp. 689-698.
Yang, G., Zhang, B. & Zhou, Q. (2009), ’’Behaviour of Geogrid Reinforced Soil Retaining
Wall with Concrete-Rigid Facing”, Journal of Geotextiles and Geomembranes
(2009), article in press, doi: 10.1016/j.geotexmem. 2009.03.001.
Zaeske, D. (2001), ’’Zur Wirkungsweise von unbewehrten und bewehrten mineralischen
Tragschichten über pfahlartigen Gründungselementen’’, Schriftenreihe Geotechnik,
Universität Kassel, Heft 10.
Ziegler, M., Heerten, G., Retzlaff, J. (2007), ’’A new dimensioning approach for junction stiff
Geogrids’’, XIV European Conference on Soil Mechanics and Geotechnical
Engineering, Madrid, Spain.
Zornberg, J.G and Grupta, R. (2010), ’’Geosynthetics in pavements: North America
Contributions’’, The 9th International Conference on Geosynthetics, Brazil.
Zhuang, Y.F. (2006), ’’Model test study on soft clay reinforced with electro-kinetic
geosynthetics’’, Geosynthetics, J. Kuwano & J. Koseki (eds), p.531-534, Rotterdam.
Zhuang, Y. (2009), ’’Numerical Modelling of Arching in Piled Embankments Including the
Effect of Reinforcement and Subsoil’’, PhD Thesis, University of Nottingham.