Mathew Gima V 2010

ANALYSIS AND INTERPRETATION OF
GROUND AND BUILDING MOVEMENTS

DUE TO EPB TUNNELLING
by
Gima V. Mathew
This thesis is presented for the degree of
Doctor of Philosophy
of
The University of Western Australia
School of Civil and Resources Engineering
October 2010
DEDICATED TO
ALMIGHTY GOD
ABSTRACT
As the number of tunnels increases to meet the demands of a world with a rapidly
expanding population, the importance of tunnel design and construction efficiencies
also grows; such new efficiencies must ensure minimal damage to nearby existing
structures and facilities. When tunnels are constructed in an urban area, it is important,
at the design stage, that the response of nearby structures can be predicted with
sufficient accuracy. These predictions will affect the choice of tunnelling method and
dictate the form and scope of potential ground improvement measures. While the
Gaussian empirical method has been shown to provide good estimates of the shape of
the settlement profile in greenfield conditions (i.e. where no structures are present),
this method cannot be used to assess structural movements associated with the
tunnelling, for which the profession now generally employs the finite element method.
This thesis investigates the surface (greenfield) and building movements induced by
two bored tunnels in the central business district (CBD) of Perth, Western Australia.
The tunnelling was carried out by an Earth Pressure Balanced (EPB) Tunnel Boring
Machine and the stratigraphy along the tunnel route comprised normally consolidated
dune sand overlying the interbedded layers of alluvial silts, clays and sands. Volume
losses during EPB tunnelling were generally less than 0.29% and, as a consequence,
settlements associated with tunnelling were small and no building damage was
observed.
The greenfield measurements (observed using electro-level beams and settlement pins
in a railway yard) indicated wider tunnelling-induced settlement troughs in soil profiles
comprising a higher proportion of clay layers. Time dependent or consolidation
settlements observed after tunnelling was completed were not significant. Contrary to
expectations, the volume losses associated with the second tunnel boring were less than
those induced during the initial boring. All greenfield settlement troughs were of a
Gaussian nature and this form could not be predicted with an acceptable level of
accuracy using both 2D (Plaxis version 9.02) and 3D (Plaxis 3D Tunnel, Version 2.0)
i
finite element (FE) analyses combined with relatively advanced soil models. Reasons
for the mismatch are discussed in this thesis and ways in which the Gaussian form may
be predicted are investigated.
Observed movement data as well as finite element back analyses for a multi-storey
building in Perth CBD indicated that the stiffness of the building altered the free field
(Gaussian) form of the settlement trough in its vicinity. 2D FE analyses are used to
illustrate how building stiffness and soil type and/stiffness influence the shape of the
predicted settlement trough. A comparison of measurements with predictions suggests
that the FE approach adopted is a reasonable method for assessing soil structure
interaction effects.
To assist the numerical predictions, the thesis also present results from a targeted
laboratory and in-situ test investigation on the upper horizons of the alluvial deposit
found beneath Perths dune sand; little information was available on the mechanical
properties of these horizons. The effectiveness of using compensation grouting in sand
to reduce the volume loss due to tunnel boring and the applicability of numerical
methods to model the grouting is also investigated.
ii
TABLE OF CONTENTS
ABSTRACT........................................................................................................................... I
TABLE OF CONTENTS .................................................................................................. III
TABLE OF FIGURES ...................................................................................................... IX
LIST OF TABLES....................................................................................................... XVIII
ACKNOWLEDGEMENT ............................................................................................. XXI
DECLARATION ......................................................................................................... XXIII
SYMBOLS AND ABBREVIATIONS..........................................................................XXV
CHAPTER 1 INTRODUCTION.........................................................................................1
1.1 BACKGROUND ...............................................................................................1
1.2 SCOPE OF THIS RESEARCH .........................................................................1
1.3 RESEARCH SITES ...........................................................................................4
1.4 THESIS STRUCTURE......................................................................................5
CHAPTER 2 LITERATURE REVIEW.............................................................................7
2.1 INTRODUCTION .............................................................................................7
2.2 GROUND DEFORMATION ............................................................................7
2.2.1 Failure Mode .....................................................................................9
2.2.2 Centrifuge Model Testing in Sand ..................................................10
2.3 SETTLEMENT PREDICTION METHODS...................................................12
2.3.1 Empirical Method ...........................................................................13
2.3.2 Numerical Prediction ......................................................................15
2.4 BUILDING MOVEMENT DUE TO TUNNEL BORING .............................20
2.4.1 Mode of Building Deformation.......................................................22
2.4.2 Numerical Studies of Interaction Problem......................................25
2.5 COMPENSATION GROUTING ....................................................................28
2.5.1 Numerical Modelling of Compensation Grouting ..........................29
2.6 SUMMARY .....................................................................................................30
CHAPTER 3 GEOLOGY, TUNNELLING AND INSTRUMENTATION ..................32
3.1 INTRODUCTION ...........................................................................................32
3.2 PROJECT DESCRIPTION..............................................................................32
iii
3.3 GEOLOGY AND HYDROLOGY OF THE PERTH REGION......................35
3.4 GEOLOGY AND HYDROLOGY OF THE TUNNEL ROUTE ....................38
3.4.1 Geology...........................................................................................38
3.4.2 Hydrogeology .................................................................................41
3.5 TUNNEL ALIGNMENT AND GROUND CONDITIONS............................44
3.6 SITE INVESTIGATIONS ...............................................................................45
3.7 EARTH PRESSURE BALANCE TUNNEL BORING MACHINE...............48
3.7.1 Ground Anchor Detector.................................................................49
3.7.2 Special Cutters ................................................................................51
3.7.3 Operation in Curved Alignments ....................................................51
3.7.4 Ground Conditioning System .........................................................52
3.7.5 Back Fill Grout System...................................................................52
3.7.6 Advanced TBM Operation Monitoring System..............................52
3.7.7 Muck Volume Measuring System ..................................................52
3.7.8 Internal Grout Ports.........................................................................53
3.7.9 Compressed Air Work Facility .......................................................53
3.8 EPB-TBM TUNNELLING PROCEDURE .....................................................53
3.9 GEOTECHNICAL INSTRUMENTATION AND MONITORING ...............54
3.10 SCHEDULING OF TUNNEL CONSTRUCTION .........................................56
CHAPTER 4 SOIL CHARACTERISATION .................................................................59
4.1 INTRODUCTION ...........................................................................................59
4.2 SUMMARY OF DATA ON SPEARWOOD SAND ......................................59
4.3 LABORATORY INVESTIGATION ON PERTH FORMATION .................63
4.4 SOIL CLASSIFICATION ...............................................................................65
4.4.1 Soil Classification from in-situ Tests..............................................65
4.4.2 Classification based on Laboratory Tests .......................................68
4.4.3 Classification based on Soil Composition ......................................70
4.5 COMPRESSIBILITY ......................................................................................74
4.6 EFFECTIVE STRESS STRENGTH ...............................................................76
4.6.1 Triaxial Tests ..................................................................................76
4.6.2 Simple Shear Tests..........................................................................79
iv
4.6.3 Undrained Shear Strength ...............................................................80
4.7 STIFFNESS .....................................................................................................83
4.8 CONCLUSIONS..............................................................................................86
4.8.1 Spearwood sand ..............................................................................86
4.8.2 Perth Formation...............................................................................86
CHAPTER 5 SURFACE SETTLEMENT - GREENFIELD SITE ...............................89
5.1 INTRODUCTION ...........................................................................................89
5.2 GROUND CONDITION .................................................................................89
5.2.1 Ground Condition South of Track 1 ...............................................91
5.2.2 Ground Condition North of Track 5 ...............................................94
5.3 TUNNELLING ................................................................................................97
5.4 INSTRUMENTATION ...................................................................................97
5.4.1 Rail Settlement Points.....................................................................97
5.4.2 Electro Level Beams (EL Beams)...................................................98
5.5 FIELD MEASUREMENT.............................................................................102
5.5.1 Rail Settlement Points (RSPs).......................................................103
5.5.2 EL Beams ......................................................................................108
5.5.3 Final Transverse Settlement Troughs ...........................................110
5.5.4 Longitudinal Settlement Profile ....................................................116
5.6 COMPARISON OF TRANSVERSE SETTLEMENT TROUGHS WITH
GAUSSIAN APPROACH .............................................................................117
5.7 DISCUSSION ................................................................................................120
5.7.1 Factors Contributing to Heave ......................................................121
5.7.2 Factors Contributing to Settlement ...............................................124
5.7.3 Nature of Settlement Trough.........................................................126
5.7.4 Relationship between i , Tunnel Depth and Tunnel Diameter......128
5.8 SUMMARY ...................................................................................................129
CHAPTER 6 GREENFIELD SETTLEMENT-NUMERICAL MODELLING .........131
6.1 NUMERICAL MODELLING .......................................................................131
6.2 SOIL MODELS .............................................................................................131
6.2.1 Hardening Soil Model (HS) ..........................................................131
6.2.2 Hardening Soil Model with Small Strain Stiffness (HSSmall).....133
6.2.3 Mohr-Coulomb Model (MC) ........................................................134
v
6.3 PARAMETER SELECTION.........................................................................135
6.3.1 Hardening Soil Model (HS Model)...............................................135
6.3.2 Hardening Soil Model with Small Strain Stiffness (HSSmall
Model) 139
6.3.3 Mohr Coulomb Model (MC) ........................................................139
6.4 TWO DIMENSIONAL FE ANALYSIS .......................................................140
6.4.1 Hardening Soil Model (HS Model)...............................................143
6.4.2 Hardening Soil Model with Small Strain Stiffness (HSSmall
Model) 146
6.4.3 Mohr-Coulomb Model ..................................................................146
6.5 COMPARISON OF THE PREDICTIONS USING DIFFERENT SOIL
MODELS .......................................................................................................148
6.6 THREE DIMENSIONAL ANALYSIS (WITH HS MODEL)......................151
Input Parameters ...........................................................................................154
6.6.1 Transverse Settlement Profile .......................................................155
6.6.2 Longitudinal Settlement profile ....................................................158
6.7 COMPARISON OF OBSERVED AND PREDICTED SETTLEMENT
TROUGHS.....................................................................................................161
6.7.1 Transverse Settlement Trough ......................................................161
6.7.2 Longitudinal Settlement Trough ...................................................165
6.8 DISCUSSION ................................................................................................166
6.8.1 Effect of K0 on the Predicted Settlement Trough..........................167
6.8.2 Effect of Lining Stiffness Analysis...............................................168
6.8.3 Effect of Stress Reduction around Tunnel ....................................169
6.8.4 Effect of Soil Stiffness on Settlement Trough ..............................172
6.8.5 Stress Changes above the Tunnel during Numerical Analyses ....174
6.8.6 Suggested Approach for Predicting Greenfield Movements
(using Plaxis) .................................................................................................176
6.8.7 Interaction Effect of Tunnels ........................................................177
6.9 SUMMARY...................................................................................................179
vi
CHAPTER 7 MALAYSIAN AIRLINES BUILDING- SURFACE AND BUILDING
MOVEMENT ....................................................................................................................183
7.1 INTRODUCTION .........................................................................................183
7.2 BUILDING DETAILS...................................................................................183
7.3 GROUND CONDITIONS .............................................................................187
7.4 INSTRUMENTATION .................................................................................191
7.4.1 Surface Settlement Pins and Building Settlement Points..............191
7.4.2 EL Beams ......................................................................................191
7.5 FIELD MEASUREMENT (SSP AND BSP) ...............................................192
7.5.1 Final Transverse Settlement Troughs ...........................................193
7.6 BUILDING MOVEMENT (EL BEAMS).....................................................194
7.7 EFFECT OF SOIL STRUCTURE INTERACTION .....................................196
7.8 NUMERICAL ANALYSIS ...........................................................................198
7.8.1 2D FE Analysis .............................................................................199
7.8.2 Transverse Surface Settlement Trough .........................................200
7.8.3 Building Settlement.......................................................................201
7.9 3D ANALYSIS ..............................................................................................202
7.9.1 Transverse Movement of the Building..........................................204
7.9.2 Longitudinal Surface Movement ..................................................205
7.10 DISCUSSION ................................................................................................206
7.10.1 Influence of Building Stiffness on Settlement Trough .................207
7.10.2 Influence of Soil Type or Stiffness on Settlement Trough ...........208
7.10.3 Settlement, Bending Moment and Horizontal Strain Induced on
Building due to Different Volume Losses .....................................................209
7.11 CONCLUSIONS............................................................................................212
CHAPTER 8 COMPENSATION GROUTING WALSH BUILDING .....................215
8.1 INTRODUCTION .........................................................................................215
8.2 BUILDING PROTECTION ..........................................................................215
8.2.1 Compensation Grouting ................................................................216
8.2.2 Building Details and Arrangement of Grouting............................217
8.3 GROUND CONDITION ...............................................................................223
8.4 BUILDING MOVEMENT AFTER CONDITION PHASE GROUTING....227
8.5 TRANSVERSE SETTLEMENT TROUGH (AFTER TUNNELLING).......230
vii
8.5.1 South Side of Building (non-grouted area) ...................................230
8.5.2 North Side of Building (grouted area) ..........................................233
8.5.3 Comparison of Surface Movement at Grouted and Non-Grouted
area 234
8.5.4 Building Settlement ......................................................................235
8.6 NUMERICAL MODELLING .......................................................................236
8.6.1 Building Movement after Condition Phase Grouting ...................238
8.6.2 Building Movement after Active Grouting (after tunnelling).......239
8.7 CONCLUSIONS............................................................................................240
CHAPTER 9 OBSERVATIONS, CONCLUSIONS AND RECOMMENDATIONS 241
9.1 INTRODUCTION .........................................................................................241
9.2 OBSERVATIONS AND CONCLUSIONS ..................................................242
9.2.1 Soil Characterisation.....................................................................242
9.2.2 Greenfield Ground Movement ......................................................243
9.2.3 Numerical Modelling of Greenfield Movement ...........................244
9.2.4 Soil Structure Interaction Effect ...................................................246
9.2.5 Numerical Modelling of Soil Structure Interaction ......................246
9.2.6 Compensation Grouting and Applicability of Numerical
Modelling247
9.3 CONTRIBUTIONS TO THE FIELD OF KNOWLEDGE ...........................247
9.4 RECOMMENDATIONS FOR FUTURE RESEARCH................................248
REFERENCES .................................................................................................................251
viii
TABLE OF FIGURES
Figure 1-1 T Bored Tunnels and Research Sites in the Perth CBD .................................. 2
Figure 2-1 Source of Ground Loss (Attewell, 1978) ......................................................... 9
Figure 2-2Observed Failure Mode based on Centrifuge Model Tests............................. 10
Figure 2-3 Volume Loss Calculated from Soil Displacements Compared to Tunnel
Volume Loss (Marshall, 2009).................................................................... 11
Figure 2-4 Profile of Mobilised Lateral Earth Pressure Coefficient on Tunnel

centre line at Different Volume Losses (Jacobsz, 2002)............................. 12
Figure 2-5 Geometry of the tunnel induced settlement trough (after Attewell et
al.,1986)....................................................................................................... 13
Figure 2-6 Transverse Settlement Trough........................................................................ 14
Figure 2-7 Longitudinal Settlement Profile (Franzius, 2003).......................................... 19
Figure 2-8 Predicted Greenfield settlement and Observed Building Settlement of

Mansion House, London (after Frischmann et al., 1994) ........................... 20
Figure 2-9 Building and Subsoil Movement of Buildings in Frankfurt Clay (after
Breth and Chambosse, 1974)....................................................................... 22
Figure 2-10 Building Deformation above Single Tunnel (Mair et al., 1996) .................. 23
Figure 2-11 Layout of Mansion House and Location of Tunnel (Frischmann et al.,
1994)............................................................................................................ 24
Figure 2-12 Settlement profiles for East and West Walls of Mansion House
(Frischmann et al., 1994) ............................................................................ 25
Figure 2-13 Comparison of Damage Estimation Criterion and Damage Levels

(Son and Cording, 2005) ............................................................................. 27
ix
Figure 2-14 Example of Interaction Diagram to Assess the Damage for Isotropic
Beam with L/H =1 (Burland, 1995) .............................................................28
Figure 3-1, Bored Tunnel Alignment................................................................................33
Figure 3-2 Eustatic Sea Level Curve for the Past 270,000 Years with Oxygen
Isotope Stages Indicated (Gozzard, 2007)....................................................35
Figure 3-3 Perth Region Generalized Geomorphology (after B Gozzard, 2007) .............37
Figure 3-4 Perth Region Generalized Geology.................................................................38
Figure 3-5 Schematic Geology of the Project Route (Hudson-Smith and Grinceri,
2007).............................................................................................................39
Figure 3-6 Schematic Section Showing Stratigraphic Relationships of Superficial

Formations in the Vicinity of Perth..............................................................40
Figure 3-7 Piezometric Profiles at South End of William Street Station (Leighton
et al., 2004)...................................................................................................42
Figure 3-8 Piezometric Profile along the Project Route (Johnson, I.D., et al.,2007) .......43
Figure 3-9 Old Lakes near Project Alignment ..................................................................45
Figure 3-10 Specifications of EPB-TBM..........................................................................50
Figure 3-11 EPB Tunnel Boring Machine- View on Cutter head.....................................51
Figure 3-12 Schedule of Tunnelling South of William Street Station..............................56
Figure 3-13 Schedule of Tunnelling North of William Street Station..............................57
Figure 4-1 SPT N vs. Depth..............................................................................................61
Figure 4-2 Typical Gradation Curve of the Dune Sand (after Andrews, 1971)................61
Figure 4-3 Relationship between G0 (from SCPTs) and qc Observed in Perth Sand ........62
Figure 4-4 Relationship between G0 (from SCPTs) and qc Observed for Perth
Formation Sand and Spearwood Sand .........................................................63
Figure 4-5 Typical CPT Data at Test Location.................................................................66
Figure 4-6 Ic and ID Indices at the Site..............................................................................68
Figure 4-7 Classification Data; Separate Symbols for each Bore Holes ..........................69
x
Figure 4-8 Activity Chart and Location of Tested Soil Samples ..................................... 70
Figure 4-9 Colour changes due to Iron on Soil Samples ................................................. 72
Figure 4-10 SEM Images ................................................................................................. 73
Figure 4-11 Oedometer Data for Three Samples ............................................................. 75
Figure 4-12 Stress Paths Measured in CIU Triaxial Compression Tests......................... 77
Figure 4-13 (a) Values of t and s at Failure for all Triaxial Tests, (b) Dependence
of Triaxial Friction Angle on Fines Content ............................................... 78
Figure 4-14 Stress Paths Measured in Undrained Simple Shear...................................... 80
Figure 4-15 Undrained Strength Ratios Plotted as a Function of (a) consolidation

stress level and (b) Fines Content................................................................ 81
Figure 4-16 Variation of A (undrained strength ratio at OCR=1) with Fines

Content ........................................................................................................ 82
Figure 4-17 Stiffness Data Measured in CIU Triaxial Compression Tests ..................... 84
Figure 5-1 Geotechnical Investigation Locations near the Railway Tracks .................... 90
Figure 5-2 Borehole Information and Stratigraphical Profile South of the Railway
Tracks .......................................................................................................... 92
Figure 5-3 CPT Profile South of Track 1......................................................................... 93
Figure 5-4 Idealised Soil Profile South of Railway Tracks ............................................. 93
Figure 5-5 Bore Hole Information and Stratigraphical Profiling North of the
Railway Tracks............................................................................................ 95
Figure 5-6 CPT Profile North of Track 5......................................................................... 96
Figure 5-7 Idealised Soil Profile North of Railway Tracks ............................................. 96
Figure 5-8 Variations in the SSP Data over a Period of Two Months for Track3
and Track 5.................................................................................................. 98
Figure 5-9 Typical View of a Horizontal EL Beam......................................................... 99
Figure 5-10 Schematic View of Horizontal EL Beam ..................................................... 99
Figure 5-11 Longitudinal and Transverse EL Beams on the Track ............................... 101
Figure 5-12 Closer View of EL Beams on the Track..................................................... 101
xi
Figure 5-13 Schematic View of EL Beam on Railway Tracks.......................................102
Figure 5-14 Transverse Settlement Trough due to Tunnel 1 Boring (RSP) ...................104
Figure 5-15 Transverse Settlement Trough due to Tunnel 2 Boring (RSP) ...................106
Figure 5-16 Relationship between Maximum Heave and Maximum Settlement

Deduced from RSPs ...................................................................................108
Figure 5-17 Transverse Settlement Trough due to Tunnel 1 Boring (EL Beam) ...........109
Figure 5-18 Transverse Settlement Trough due to Tunnel 2 Boring (ELBeam) ............110
Figure 5-19 Final Transverse Settlement Trough for Three Tracks due to Tunnel 1 .....111
Figure 5-20 Final Transverse Settlement Trough for Three Tracks due to Tunnel 2 .....112
Figure 5-21 Transverse Settlement Troughs due to Tunnel 1 and Tunnel 2...................113
Figure 5-22 Combined Transverse Settlement Troughs after Tunnel 2 Boring .............114
Figure 5-23 Typical Longitudinal Profile for Tunnel 1 Boring......................................116
Figure 5-24 Linear Regression Lines for Tunnel 1 Boring............................................118
Figure 5-25 Linear Regression Lines for Tunnel 2 Boring.............................................119
Figure 5-26 Estimated Face Pressure along Northern Bored Tunnel (after Sigl and
Yamazaki, 2007) ........................................................................................121
Figure 5-27 Variation of FP and Horizontal Effective Stress during Tunnel 1 and
Tunnel 2 Boring (Note Tracks 1, 3 & 5 are at approximate chainages
of 138m, 152m and 177m for Tunnel 1 and 146m, 164m and 192m
for Tunnel 2)...............................................................................................122
Figure 5-28 Variation of Grout Volume / Ring .............................................................123
Figure 5-29 Variation of GP and Horizontal Effective Stress during Tunnel 1 and
Tunnel 2 Boring .........................................................................................124
Figure 5-30 Normalised Field Settlement Troughs for Tunnel 1 ...................................127
Figure 5-31 Variation of i with z0 for Tunnels in Sand and Gravels (Mair and
Taylor, 1997) ..............................................................................................128
xii
Figure 5-32 Relation between i , Tunnel depth and Diameter for Different Ground
Conditions (Peck, 1969) ............................................................................ 129
Figure 6-1 Figure showing the Relationship between G0ref and Strain at Reference
Stress of 100 kPa ....................................................................................... 137
Figure 6-2 North Profile Showing Mesh and Tunnels for 2D FE Analysis................... 141
Figure 6-3 South Profile Showing Mesh and Tunnels for 2D FE Analysis................... 141
Figure 6-4 Transverse Settlement Trough for North and South profiles for Volume
Losses of 0.25% & 0.19%; HS model....................................................... 144
Figure 6-5 Variation of Excess Pore Pressure for North Profile (vl=0.25%) ................ 145
Figure 6-6 Variation of Excess Pore Pressure for South Profile (vl=0.25%) ................ 145
Figure 6-7 Transverse Settlement Trough for North and South profiles for Tunnel
1 (Volume Losses of 0.15% & 0.25%); MC with E=E50 .......................... 146
Figure 6-8 Final Transverse Settlement Trough for North and South Profiles; MC
with E=Eur ................................................................................................. 147
Figure 6-9 Transverse Settlement Trough for South Profile; MC with E=Eur ............... 148
Figure 6-10 Comparison of the Transverse Settlement Trough Predicted using

Different Soil Models................................................................................ 149
Figure 6-11 Deformed Mesh using MC (Eur) Model ..................................................... 150
Figure 6-12 Deformed Mesh of the Geometry using HS Model ................................... 150
Figure 6-13 Deformed Mesh of the Geometry using HSSmall Model .......................... 150
Figure 6-14 Geometric Model of the Tunnel Excavation .............................................. 152
Figure 6-15 3D Mesh for the Plaxis Analysis North of the Rail Lines......................... 152
Figure 6-16 3D Mesh for the Plaxis analysis South of the Rail Lines........................... 153
Figure 6-17 Schematic View of the Modelling of Tunnel Boring ................................. 154
Figure 6-18 Deformed Mesh after Excavating Tunnel 1 (North profile, vl=0.25%)..... 155
Figure 6-19 Deformed Mesh after Excavating Tunnel 1 and Tunnel 2 (North
profile, VL=0.25% &0.19%) ..................................................................... 156
xiii
Figure 6-20 Deformation Plane after the Excavation of Tunnel 1 (North profile,
VL VL =0.25%) ...........................................................................................156
Figure 6-21 Deformation Plane after the Excavation of Tunnel 1 and Tunnel 2
(North profile, VL =0.25% &0.19%)..........................................................156
Figure 6-22 Transverse Settlement Trough of North and South Profile for Volume
Losses of 0.25% & 0.19% ..........................................................................158
Figure 6-23 Typical Longitudinal Settlement Trough ....................................................159
Figure 6-24 Boundary Effect for Longitudinal Trough .................................................159
Figure 6-25 Settlement Trough with Different FP and GP.............................................160
Figure 6-26 Longitudinal Settlement Trough for Different GP and FP..........................161
Figure 6-27 Comparison between Settlement Troughs Predicted for North Profile.......162
Figure 6-28 Comparison between Settlement Troughs Predicted for South Profile.......163
Figure 6-29 Comparison of Normalised Settlements for Tunnel 1.................................165
Figure 6-30 Predicted and Observed Normalised Longitudinal Settlement Trough ......166
Figure 6-31 Settlement Trough with Different K0 Value for North and South Soil
Profile .........................................................................................................168
Figure 6-32 Variation of Surface Movements for Different Lining Stiffness ................169
Figure 6-33 Soil Profile and Location of Tunnel for Parametric Study..........................170
Figure 6-34 Settlement Trough with Different K0 around the Tunnel............................172
Figure 6-35 Settlement Profile with Different Stiffness .................................................173
Figure 6-36 Mobilised Horizontal Earth Pressure ..........................................................173
Figure 6-37 Mobilised Lateral Earth Pressure Coefficient along the Tunnel axis.........175
Figure 6-38 Mobilised Stress and Km near the Ground Surface at RL=10.3m
(section B-B in Figure 6-33) ......................................................................176
Figure 6-39 Comparison of Predicted and Measured Settlement Trough ......................177
Figure 6-40 Interaction Effect of Tunnels using South Profile.......................................178
xiv
Figure 6-41 Interaction Effect of Tunnels using North Profile...................................... 178
Figure 7-1 Aerial View of MAS Building ..................................................................... 184
Figure 7-2 Schematic View of the Building and Location of Tunnels .......................... 185
Figure 7-3 Front View of MAS Building....................................................................... 186
Figure 7-4 View Opposite MAS Building ..................................................................... 186
Figure 7-5 Ground and First Floor Plan for MAS Building .......................................... 187
Figure 7-6 Geotechnical Investigation Locations near MAS Building.......................... 188
Figure 7-7 Borehole Information and Stratigraphical Profile Close to the MAS
Building ..................................................................................................... 190
Figure 7-8 CPT Profile................................................................................................... 190
Figure 7-9 Soil Profile and Location of Tunnel near MAS Building ............................ 190
Figure 7-10 EL Beams on the Wall with Transmitter on Each Beam ........................... 192
Figure 7-11 Ground and Building Movement due to Tunnel 1 and Tunnel 2 ............... 193
Figure 7-12 Transverse Surface Settlement Trough for Tunnel 1 and Tunnel 2 ........... 194
Figure 7-13 Movement of Building at Various Times during Tunnelling..................... 195
Figure 7-14 Transverse Building Movement for Tunnel 1 and Tunnel 2 ...................... 196
Figure 7-15 Linear Regression Analysis for Tunnel 1 and Tunnel 2............................. 197
Figure 7-16 Normalised Settlement for Tunnel 1 and Tunnel 2 .................................... 197
Figure 7-17 2D Mesh Showing Tunnels and Building .................................................. 199
Figure 7-18 Observed and Predicted Settlement Trough............................................... 201
Figure 7-19 Measured vs Predicted Settlement of the Building .................................... 202
Figure 7-20 3D Mesh for the Geometry........................................................................ 203
Figure 7-21 Deformed Mesh after Tunnel Boring ......................................................... 203
Figure 7-22 Predicted Movement of the MAS Building as the TBM Passes ................ 204
Figure 7-23 Schematic View of Building and Direction of TBM ................................. 204
Figure 7-24 Transverse Settlement Observed and Predicted by FE Analysis ............... 205
Figure 7-25 Longitudinal Settlement Profile from 3D Analysis.................................... 206
xv
Figure 7-26 Schematic View of Geometry in z Direction ..............................................206
Figure 7-27 Surface Settlement Trough with Different Building Stiffness ....................208
Figure 7-28 Surface Settlement Trough with Different soil type ...................................209
Figure 7-29 Building Settlement due to Different Volume Losses ................................210
Figure 7-30 Predicted BM due to Tunnelling for Different Volume Losses ..................211
Figure 7-31 Horizontal Strain induced by Tunnel 1 Boring on MAS ............................212
Figure 8-1 Aerial View of the Protected Buildings ........................................................216
Figure 8-2 Front View of Walsh Building ......................................................................218
Figure 8-3 Schematic View of the Location of Building and Tunnel.............................219
Figure 8-4 Church Opposite to Walsh Building .............................................................220
Figure 8-5 Tunnel Alignment and Geotechnical Investigation Locations near

WALSH Building.......................................................................................221
Figure 8-6 Typical Cross-section of Grouting Arrangement ..........................................222
Figure 8-7 Site Setup (top) and Schematic Layout for Grouting (bottom).....................222
Figure 8-8 Schematic View of the Protected Buildings, Grout Ports and Location
of EL Beams on the Building.....................................................................223
Figure 8-9 Borehole Information and Stratigraphical Profile Close to WALSH

Building......................................................................................................225
Figure 8-10 CPT Profile Close to Walsh Building .........................................................226
Figure 8-11 Soil Profile and Location of Tunnel near WALSH Building......................227
Figure 8-12 Movement of Building due to Condition Phase Grouting ..........................228
Figure 8-13 Variation of Grout Volume along the Transverse Direction of the
Building......................................................................................................229
Figure 8-14 EL Beam Location and Zone of Influence ..................................................229
Figure 8-15 Settlement Troughs at Non-Grouted Area for Tunnel 1 and Tunnel 2 .......231
Figure 8-16 Normalised Settlement for Tunnel 1 and Tunnel 2 .....................................232
xvi
Figure 8-17 Location of Tunnels and the Buildings at Non-grouted Area .................... 233
Figure 8-18 Surface movement at Grouted Location (from SSP and BSP data) ........... 234
Figure 8-19 Surface Settlements Due to Tunnel 1 Boring at Grouted and Non-
Grouted Locations ..................................................................................... 235
Figure 8-20 Building Movement due to Grouting and Tunnel 1 Boring ....................... 236
Figure 8-21 Mesh Showing the Location of the Grouting ............................................. 237
Figure 8-22 Observed and Predicted Movement after Condition Phase Grouting. ....... 238
Figure 8-23 Observed and Measured Movement of Building due to Tunnel 1 ............. 239
xvii
LIST OF TABLES
Table 3-1 Construction Methods of the Different Sections of the New Metro Rail
City Project (Part 4B Geotechnical Interpretive Report, 2004) ...................34
Table 3-2 Phases of Site Investigation..............................................................................46
Table 3-3 Main Characteristics of the EPB-TBM ............................................................49
Table 3-4 Instrument Types and Quantities...................................................................55
Table 3-5 Schedule of Tunnel Construction .....................................................................56
Table 4-1 Relative Phase Composition (wt.%) Derived from XRD Data ........................71
Table 4-2 Properties of Soil Samples Tested in 1-D Compression..................................74
Table 4-3 Approximate Values of for Clays at the St. Georges Terrace Site ................85
Table 5-1 Variation of Heave and Settlement for each Track after the Tunnel
Boring.........................................................................................................107
Table 5-2 Settlement Details of Three Tracks ................................................................115
Table 5-3 Estimated i Values ..........................................................................................120
Table 5-4 Estimated i and K Values at the Three Track Locations ..............................127
Table 6-1 Hardening Soil Model Parameters for North Profile......................................138
Table 6-2 Hardening Soil Model Parameters for South Profile......................................138
Table 6-3 Mohr-Coulomb Model Parameters for North Profile .....................................139
Table 6-4 Mohr-Coulomb Model Parameters for South Profile .....................................140
Table 6-5 Tunnel Details.................................................................................................141
Table 6-6 Volume Losses Specified in Numerical Analyses..........................................142
Table 6-7 Main Features of the Mesh .............................................................................153
xviii
Table 6-8 Face Pressure, Grout Pressure and TBM Thrust for 3D Analyses ................ 155
Table 6-9 Input Volume Loss and the Out put Volume Losses..................................... 163
Table 6-10 Combinations of Stress History used for Predicting Settlement Profile...... 167
Table 6-11 Vertical Movement of Tunnel Crown.......................................................... 169
Table 6-12 Different Cases for the Analyses ................................................................. 170
Table 7-1 Surface Settlement Details at MAS Building ................................................ 194
Table 7-2 Input Parameters for the Building Components ............................................ 198
Table 7-3 Face Pressure, Grout Pressure and TBM Thrust used for 3D Analyses........ 203
Table 8-1 Summary of Settlement Troughs at Non-Grouted Area of Walsh

Building ..................................................................................................... 231
Table 8-2 Details of the Building Component ............................................................... 237
xix
xx
ACKNOWLEDGEMENT
It has been a pleasure and privilege to be part of Geotechnical Research Group at The
University of Western Australia. The research for this dissertation was carried out in the
School of Civil and Resources Engineering during 2005 -2010 and was supported by the
University Postgraduate Award (UPA) and Othman Frank Blakey Scholarship, which are
gratefully acknowledged.
There are many people to thank for their support and encouragement, without whom
this thesis would not have been possible.
First of all, I would like to express my sincere gratitude to my supervisor, W/Prof. Barry
Lehane for his valuable discussions, advices, support and consideration throughout this
research. I have been very fortunate to work with him. I sincerely thank you Barry for
your patience, attention to detail and always having time for discussion.
I would like to thank Binaya, Claire, Natalie and Christine in the Geotechnical
laboratory for helping and making my life easy with laboratory testing. I also thank all
the staff in the general workshop and electronic workshop for their untiring help and
enthusiasm in finishing the job on time.
The help and support provided by Jane, Simone, Monica, Eileen, Lisa and Sharon in the
Civil engineering and COFS office is also acknowledged. My special
acknowledgements are expressed to Wenge and Keith for their friendly and skilful IT
support throughout this research.
I say a great thank you to Natusha for helping me during my stay here. I aso thank
Mostafa for his time and patience for helping me with my laboratory study and lending
me the Plaxis dongle.
I am very grateful for the company of all the research associates, postgraduate students
and visitors in Civil Engineering as well as in COFS, without the support of everybody
xxi
the life would be difficult. I specially thank Lina, Shazzad, Booning, Tarrant, Xuelin,
Bao and Kevin for their valuable help, advice and encouragement whenever in need.
I am very much indebted to Selina for taking care of my children during my research.
She was there for me whenever I needed help. Without you I wouldnt have been able to
do my research so relaxingly. I also thank Siraj and Sajid for their help and company.
I specially thank Fiona Chow for providing me with journal papers, books and thesis
whenever in need. I also acknowledge the help and support provided by Susan
Governec, Doug Stewart, Eric Hudson-Smith, Raghu, Peter McGough, Michael
Grinceri, Logan, Babu, Suman and Matt Williams.
The help from Leighton Kumagai Joint Venture (LKJV) by providing instrumentation
monitoring data is also greatly acknowledged. I also acknowledge the funding providing
by an Australian Research Council (ARC) linkage grant in association with LKJV to
support this research.
I am also very grateful and sincerely acknowledge the time and effort of all the
examiners evaluating this thesis.
I sincerely thank my husband Joyis and children Joshia,

Denzel and Paulus for their untiring support and
consideration throughout my research. I wouldnt have
been able to finish this research without Joyiss help,
patience and support.
I would like to thank my parents, brothers and sister-in-laws for their constant support
and effort throughout this research. I also thank Sony Aunty for coming and helping me
during my delivery time and helping me with the kids. I specially thank my mother-in-
law Rose for her constant prayer for me and support. I also thank all my in- laws for
their love and affection for me and supporting me.
Last but not least I thank almighty God for bestowing me with knowledge and wisdom
to complete the study successfully. When ever I was in need He gave it to me through
somebody.
xxii
DECLARATION
I hereby declare that, except where specific reference is made to the works of others, the
contents of this dissertation are original and have never been submitted, in part or as a
whole, to any other University for any degree, diploma or other qualifications. This
dissertation is the result of my own work and includes nothing which is the outcome of
work done in collaboration. This dissertation contains no more than 70,000 words and
180 figures.
Gima V. Mathew
October 2010
xxiii
xxiv
SYMBOLS AND ABBREVIATIONS
Symbols
c' Cohesion
D Outer tunnel diameter
D1 Outer diameter of the lining
D2 Diameter of the cutter face
e In situ void ratio
emax Maximum void ratio
E Modulus of elasticity of concrete
Esec= E Drained secant Youngs moduli of soil
E0 Youngs moduli at very small strains
E50 ref Secant stiffness in standard drained triaxial test at reference stress
Eoedref Tangent stiffness for primary oedometer loading at reference

stress
Eurref Unloading/reloading stiffness
EA Axial stiffness
EI Bending stiffness
Fr Friction ratio
fs Friction sleeve measurement
G Secant shear modulus
xxv
G0ref Small strain shear modulus at reference stress
Ic Soil behaviour type index
ID DMT material index
i Transverse horizontal distance from centre-line of the tunnel to

the point of inflection of the settlement trough (also ix)
k_x, k_y & k_z Permeability of the soil in the x, y and z directions
K Trough width parameter (also K1 and K2)
K0nc K0 value for normal consolidation
K0 Coefficient of earth pressure at rest
Km Mobilised earth pressure coefficient
L - Length of each ring
LL Liquid limits
m Power for stress-level dependency of stiffness
Nk Cone factor
p Mean effective stress
pa Atmospheric pressure or reference stress
p0 Lift-off pressure
pref Reference confining pressure
p1 Pressure measured at a membrane expansion of 1.1 mm
p0 Initial mean effective stress
q Deviator stress
qc CPT end resistance
xxvi
qf Maximum deviator stress
Smax Maximum settlement
su Undrained shear strength
Sv Settlement at offset y from the tunnel centre line
S Settlement at a distance x from tunnel centre line
sutc Undrained shear strength in triaxial compression test
suss Undrained shear strength in simple shear test
u0 Ambient pore pressure.
Uy Vertical movement
Ux Horizontal movement
VL Volume loss
Vs Shear wave velocity
Vs Excavated volume
w Water content
y Horizontal distance from the tunnel centre line
Z0 Depth of the tunnel axis below the ground surface
' Friction angle
Peak dilation angle
cv Critical state friction angle
m Mobilised dilation angle
m Mobilised friction angle
xxvii
sat Saturated unit weight of soil
unsat Unsaturated unit weight of soil
Angle of dilatancy (o)
Poissons ratio
Mean density
Constant varies with the nature and age of the deposit,
xy Shear stress
( L) Deflection ratio
Poissons ratio
Shear strain
ur Poissons ratio for unloading reloading
3 Minor and intermediate principal effective stresses
v0 Total vertical stress
v Vertical effective stress.
vy Vertical yield stress
vc Vertical consolidation stress
Abbreviations
1-D One dimensional
AHD Australian Height Datum (Mean Sea Level)
xxviii
BSP Building settlement points
CF Coarse fraction
CPT Cone penetration test
DMT Dilatometer test
FC Fine content
POP Past overburden pressure
OCR Overconsolidation ratio
PI Plasticity index
RSP Rail settlement points
SSP Surface settlement pins
SBT Soil behaviour type
TAM Tubes-a-manchettes
xxix
xxx
CHAPTER 1 INTRODUCTION
1.1 BACKGROUND
The growth of many cities has resulted in the need for underground structures such as
tunnels to provide efficient transportation, water supply, sewage disposal and
communications. To avoid any damage to the overlying structures, efficient and
economic tunnel design and construction methods should be implemented. There are
empirical, numerical and analytical methods to predict the movement associated with
tunnelling. When tunnels are constructed in an urban area, it is important, at the design
stage, that the response of nearby structures can be predicted with sufficient accuracy.
These predictions will affect the choice of tunnelling method and dictate the form and
scope of potential ground improvement measures. The current deformation prediction
approaches are often conservative and can lead to unnecessary expenditure for the
protective measures. A more reliable design method would resolve issues such as
excessive cost and damage to surface or sub surface facilities. The most widely used
Gaussian empirical method has been shown to provide good estimates of the shape
and size of the settlement profile in greenfield conditions (i.e. where no structural
interaction effects are present), this method cannot be used to assess structural
movements associated with the tunnelling. The Finite Element (FE) method is currently
the most popular approach to predict the surface and building movement associated with
tunnelling.
1.2 SCOPE OF THIS RESEARCH
The Perth MetroRail project connecting Perth to Mandurah was a major infrastructural
investment by the Western Australian State Government. In the Perth Central Business
District (CBD), the project consisted of two bored (6.9m outer diameter) underground
rail tunnel, open dive and cut-and-cover tunnel sections and two below ground stations
(Esplanade and William Street Station); these elements are shown in Figure 1-1.
Leighton Kumagai Joint Venture (LKJV) carried out the tunnelling work in the CBD.
1
The stratigraphy of the tunnel route consists mainly of normally consolidated dune sand
(Spearwood Sand) overlying the overconsolidated layers of alluvial silts, clays and
sands (Perth Formation), which are underlain by shale/ siltstone (Kings Park
Formation). Earth Pressure Balanced Tunnel Boring Machine (EPB-TBM) was used to
bore the two tunnels.
Figure 1-1 T Bored Tunnels and Research Sites in the Perth CBD
Walsh MAS
Greenfield site
2
Chapter 1 Introduction
Although significant portions of the bored tunnels went through the overconsolidated
Perth Formation, no systematic laboratory investigation of the Perth Formation has been
published to date. The design and construction of this project was constrained by the
need to minimise the effect of tunnelling on the existing buildings. As the tunnel was
passing directly underneath some of the buildings along the tunnel route, the potential
for imposition of large ground distortions on these buildings was very high. These
buildings were on shallow footings and compensation grouting was implemented under
some of the buildings to mitigate the effect of tunnelling. Hence the main objectives of
this research are;
1. To study the properties of the Perth Formation in detail by conducting state

of the art laboratory tests on undisturbed soil samples collected from the
research site. The laboratory measurements are subsequently related and
calibrated with in-situ test data to assist in the assignment of material
properties for the numerical analyses.
2. To investigate the effect of EPB tunnelling in the Perth soils. This is attained
by interpreting extensive instrumentation data collected from three adjacent
locations at the greenfield site shown on Figure 1-1. The nature of the
settlement troughs is assessed in detail and compared with the observations
made in the similar ground conditions around the world.
3. To examine the suitability of FE methods to predict the movement

associated with the tunnelling especially in the mixed stratigrapghy present
in Perth CBD. This is achieved by carrying out both 2D and 3D finite
element analyses using the popular Plaxis program.
4. To study the effect of tunnel construction on nearby buildings as well as the

influence of buildings on the observed pattern of surface or building
movements. The vertical ground and building movement at two multistorey
building locations are analysed in detail to study the soil structure interaction
effect.
5. To investigate the applicability of both two dimensional (2D) and three

dimensional (3D) numerical modelling in replicating the soil-structure
interaction effect as observed in the field.
3
6. To study the effectiveness of compensation grouting in sand to reduce the
soil volume loss associated with tunnelling. This is achieved by comparing
the settlement troughs at grouted and non-grouted areas close to each other.
7. To examine the applicability of FE modelling to model the compensation

grouting.
1.3 RESEARCH SITES
The research mainly focused on five monitored sites along the tunnel route and they are,
i) greenfield sites (three adjacent locations in railway track area) ii) Malaysian Airlines
(MAS) building and iii) Walsh building; the locations of the five sites are indicated in
Figure 1-1. The instrumentation was installed and monitored by LKJV. However, at the
MAS building, the author installed EL beams in addition to the instrumentation installed
by LKJV to monitor the movement of the building; LKJV also recorded these EL beam
data. Analyses and interpretation of all the instrumentation data presented here were
carried out independently by the author. The tunnelling took place during the initial
period of the PhD study.
Details of each of the monitored sites are given below.
(i) Greenfield Site (Railway Tracks): Part of the tunnel went underneath five
existing railway tracks. As there were no buildings around this area, this was
considered as a good location to esimate the greenfield movement. The
tunnels were about 11m below ground level and 14.4m apart (centre to
centre) at this location. Out of five tracks, three tracks (Track 1, Track 3 and
Track5) are considered here for the analysis. The tunnels encountered both
the Perth Formation and Spearwood sand at these locations.
(ii) Malaysian Airlines (MAS) Building: This building (constructed in 1929) is

a framed structure having seven storeys and a basement founded on shallow
footing. Full structural details of the building are not available and
assumptions regarding its structural form are based on visual inspection.
This building was about 6m from the centre line of Tunnel 1. The near by
tunnels were at about 18 m below ground level and 10.2m apart. Most of the
4
Chapter 1 Introduction
tunnels cross-section is located in the sand and clay layers of the Perth
Formation at this area.
(iii) Walsh building. The Walsh building is heritage listed and has five storeys
and a basement. This building is a steel framed structure encased in concrete,
with secondary reinforced concrete beams. The building is on shallow
footings. The tunnels springlines are at about 17m below ground level and
are 10.6m apart (centre to centre) at south side of the building. The
corresponding measurements at north side of the building are 15.45m and
12m. The tunnels are located primarily in silty and clayey layers of the Perth
Formation at this area. Compensation grouting was implemented over the
northern section of the building to mitigate the effects of tunnelling induced
settlements.
1.4 THESIS STRUCTURE
This thesis consists of nine Chapters. Following the introduction, which is Chapter 1,
Chapter 2 presents a review of relevant literature for the thesis topic. This chapter
covers patterns of soil movement as well as volumetric straining and stress changes
above the tunnel crown, as observed in centrifuge model testing. Different approaches
to predict the tunnelling induced ground movements are also discussed. Soil structure
interaction effects are reviewed as are criteria for assessing movement thresholds for
building damage. Current approaches for numerical modelling of compensation
grouting are also evaluated.
Chapter 3 gives the over all description of the MetroRail project (in the Perth CBD
area), tunnelling method and geology & hydrogeology of the tunnel route. The
characteristics of the TBM used to bore the two tunnels and summary of
instrumentation used to monitor the ground as well as building movement are also
detailed.
All the laboratory tests carried out in the alluvial deposits of the upper horizons of Perth
Formation along with the in situ tests carried out as part of the MetroRail project are
reported in Chapter 4. The laboratory measurements are subsequently related and
calibrated with available in-situ test data to assist in the assignment of material
properties in numerical analyses.
5
The surface settlement (greenfield) troughs induced by boring of the two tunnels are
the main focus of Chapter 5. The extensive instrumentation data at three locations close
to each other are analysed thoroughly and compared with the observations in similar
ground conditions around the world. The applicability of the widely used Gaussian
Method to predict the greenfield troughs in the predominantly sandy soils is also
examined in this Chapter.
In Chapter 6, 2D as well as 3D numerical analyses carried out using Plaxis program are
presented that examine the suitability of this (widely used) program to predict the
vertical surface movement (greenfield) associated with tunnelling. The method of
parameter selection (from the laboratory as well as from the field test) for the numerical
modelling is also discussed in detail. The observations from the comparison of predicted
and measured vertical surface movement are also discussed. The (often predicted) wide
and shallow settlement trough is investigated through FE parametric studies in
combination with data from available centrifuge model tests.
The monitored vertical movement at a multi-storeyed building (MAS) location along

the tunnel route is analysed in detail and discussed in Chapter 7. The influence of
surface structures in modifying the surface and the structural movement and the soil
structure interaction effect is also discussed. Both 2D as well as 3D FE analyses were
carried out to study interaction effects. This Chapter also investigates the influence of
building stiffness and soil type and/stiffness on the settlement troughs through
numerical parametric studies.
Chapter 8 assesses the effectiveness of the compensation grouting (hydraulic fracture)

in sand to reduce the volume loss due to tunnelling. The suitability of Finite Element
(FE) program to model compensation grouting is also evaluated.
The main conclusions drawn from Chapter 4 to 8 are summarised in Chapter 9.

Recommendations for future research are also highlighted.
6
CHAPTER 2 LITERATURE REVIEW
2.1 INTRODUCTION
Bored tunnels are constructed either by (i) Shield Tunnelling or (ii) New Austrian
Tunnelling Method (NATM). In Shield Tunnelling, the Tunnel Boring Machine (TBM)
is used to excavate the tunnel and this technique is usually used in soft ground
conditions to prevent ground collapse. The tail of the TBM supports the lining to be
erected and final lining is placed in position as soon as sufficient length is excavated.
The gap between the soil and lining is grouted immediately to reduce the volume loss.
In NATM, the tunnel is excavated in different parts such as the crown, bench and invert.
After each excavation section, the tunnel contour is stabilised by introducing temporary
lining of sprayed concrete. A final lining is installed later if one is required for long
term. The long term lining is usually provided by cast in situ concrete.
Irrespective of the method of construction, the excavation of tunnels induces

movements to the ground and structures in the tunnel vicinity. If the estimated
movement is above the permissible limits (for any given tunnelling method), additional
measures such as grouting are often carried out to improve the strength and stiffness of
the soil near and ahead of the tunnel. This Chapter reviews aspects of direct relevance to
this thesis, namely (i) the mode of ground deformation due to tunnelling, (ii) Settlement
prediction methods, (iii) building movement associated with tunnelling and soil
structure interaction effect and (iv) compensation grouting and its numerical modelling.
2.2 GROUND DEFORMATION
This thesis examines tunnelling-induced ground movement and hence it is worthwhile

to consider the various sources of these movements. In the first instance, it should be
noted that, regardless of the tunnel construction method, excavation induces stress
changes causing displacements and strains around a tunnel. The amount of ground
deformation around a tunnel cross-section is directly related to the additional amount of
7
excavation required to maintain that cross-section. The magnitude of the displacements
caused by the tunnelling is quantified in terms of volume loss VL, which is defined as
the excavated volume of the soil (Vg) in excess of the tunnel volume (V) divided by the
tunnel volume. The magnitude of volume loss varies for different ground conditions and
different tunnelling methods. Figure 2-1 shows the source of ground loss associated
with tunnel boring and each of these is explained below (Cording, 1991; Attewell,
1978).
1. Face loss: For the case of an open face shield tunnelling, this is the movement of
the soil towards the unsupported tunnel face during the tunnel excavation. If an
Earth Pressure Balanced (EPB) or slurry shield machine is used, heave may
develop in front of the cutter face due to excessive face pressure.
2. Shield loss: This is movement due to the radial ground loss around the shield to
fill the gap between the shield and the unsupported soil. This will be more if the
TBM is steering through a curve because of over-excavation.
3. Tail void loss: The erection of lining within the TBM introduces a gap between
the lining and the unsupported soil. The ground lost due to the squeezing of soil
into the void between the lining and the unsupported soil is referred to as the tail
void loss. This can be minimised by grouting the void between the soil and the
lining.
4. Lining loss: This movement is a continued radial loss of soil due to the
deformation of the lining as the overburden pressure is gradually redistributed to
its new equilibrium. If the tunnel comprises thick precast concrete segments, this
movement will be minimal.
5. Consolidation: This is movement due to dissipation of excess pore pressure

developed due to excavation or grouting. These pore pressure changes lead to
effective stress changes and hence additional ground movements over a long
time.
8
Chapter 2 Literature Review
For NATM, only face loss, lining loss and consolidation are applicable. Immediate
or short term deformation is caused by face loss, shield loss, tail void loss and lining
loss and the long term ground deformation is caused by consolidation.
Figure 2-1 Source of Ground Loss (Attewell, 1978)
The final settlement is a combination of the short term (immediate) and long term
(consolidation) settlement. Post construction settlement can be significant, particularly
in the case of tunnels in soft compressible clays (OReilly et al., 1991).
2.2.1 Failure Mode
The mode of soil failure (short term settlement) induced by the tunnel construction in
sand and clay is significantly different. The failure mechanisms inferred from centrifuge
model tests are illustrated in Figure 2-2 (a) and Figure 2-2 (b) for clay (Mair, 1979) and
sand (Chambon and Cort, 1994) respectively. In clays, the failure surface propagates
upwards and outwards from the tunnel invert and becomes significantly wider at the
surface. However, in sand, the failure surface is like a narrow chimney propagating
almost vertically from the tunnel invert to the surface. This same behaviour of sand was
also observed by Potts (1976) in his laboratory model tests.
9
It may be inferred from these trends that the settlement trough that develops at the
ground surface is wider in clays than in sands - as failure approaches.
(a) Clay (Mair, 1979) (b) Sand (Chambon and Corte, 1994)
Figure 2-2Observed Failure Mode based on Centrifuge Model Tests
2.2.2 Centrifuge Model Testing in Sand
In clays, (i.e. in undrained condition), the volume loss calculated from the settlement
trough at any given level above the tunnel does not change with this level (Mair and
Taylor, 1997) and therefore an accurate estimate of volume loss can be obtained from
the surface settlement trough. However, in drained conditions (e.g. in dense sand) where
tunnelling induces volume changes in the soil, the volume loss calculated from the
surface settlement trough will be different at different levels above the tunnel (Cording
and Hansmire, 1975). This is in agreement with the observations of Jacobsz (2002) and
Marshall (2009). They carried out centrifuge tests that modelled tunnelling in sand and
found that under drained condition, estimation of ground loss from the soil displacement
data is complicated because of the soils contractive/dilative nature as shown in Figure
2-3. Figure 2-3 illustrates the difference in soil volume loss (ground loss calculated from
the settlement trough) and tunnel volume loss (calculated from the actual volume
extracted from the tunnel model) at different levels above the tunnel. Consequently,
inputting a volume loss (in numerical parametric study) calculated from the observed
10
surface settlement trough will over predict the greenfield settlements at low volume
losses and under predict the settlements at higher volume losses.
Figure 2-3 Volume Loss Calculated from Soil Displacements Compared to Tunnel
Volume Loss (Marshall, 2009)
Figure 2-4 illustrates the mobilised lateral earth pressure coefficient calculated from the
stress measurements in the centrifuge model testing of sand performed by Jacobsz
(2002). This figure shows higher earth pressure coefficients near the tunnel crown at
250mm depth which increase as the volume loss increased. Coefficients reduce
sharply with distance above the tunnel crown. However it appears that, for low volume
losses (less than 0.5%), the mobilised earth pressure coefficients are not significantly
varying along the depth of the tunnel axis. Unfortunately these measurements are only
from prototype depth equivalent of 7.5 m (75g test) from the surface.
11
Figure 2-4 Profile of Mobilised Lateral Earth Pressure Coefficient on Tunnel
centre line at Different Volume Losses (Jacobsz, 2002)
2.3 SETTLEMENT PREDICTION METHODS
Methods available for prediction of surface and subsurface settlement trough include the
empirical (Gaussian) method, numerical methods such as Finite element and Finite
difference methods and analytical methods (closed form solutions). Closed form
solutions can provide a rough estimate of ground behaviour although they cannot
accommodate complexities of tunnel construction methods and soil properties such as
anisotropy. Hence numerical methods are becoming increasingly common in
engineering practice. The empirical, 2D and 3D FE methods are considered in this
thesis and each of these is explained in the following sections. It is important to note,
however, that all of these methods require the tunnel volume loss as an input
parameter. This volume loss is a function of all the factors discussed in Section 2.2,
many of which depend critically on the construction methodology, TBM excavation
parameters etc. and therefore cannot be predicted numerically or analytically. The
methods therefore predict a distribution of ground movements for a prescribed volume
loss. Designers prescribe volume losses on the basis of case history data reported in the
12
literature using, for example similar tunnelling method in a given soil type. Soil
structure interaction can be modelled using the FE methods.
2.3.1 Empirical Method
The development of surface settlement trough ahead of a tunnel heading is illustrated in

Figure 2-5. Peck (1969) and subsequently many other authors have concluded that the
transverse settlement trough, which forms immediately after the tunnel has been
constructed, can be described by a Gaussian distribution curve. Typical view of
transverse settlement trough is shown in Figure 2-6. The vertical settlement along the
transverse direction can be expressed as,
y2
S S max exp (2.1)
2i 2
Where,
S settlement
Smax - maximum settlement above the tunnel centre line
y horizontal distance from the tunnel centre line
i transverse horizontal distance from the centre line to the point of inflection of the
settlement trough.
Figure 2-5 Geometry of the tunnel induced settlement trough (after Attewell et
al.,1986)
13
i
Volume of settlement
trough V s =V g + V g
Volume change Point of inflection

Smax
in ground = Vg
Ground loss at
Tunnel, V g
Volume loss V L =V g /V
Tunnel Volume, V
Figure 2-6 Transverse Settlement Trough
The volume of the transverse surface settlement trough per meter length of tunnel, Vs,
can be evaluated by integrating equation 2.1 and may be expressed as:
Vs Sdx 2 i S max (2.4)
Although, for the undrained (constant volume) case, the ground loss close to the tunnel
(Vg) will be same as the volume of the surface settlement trough (Vs), this will be
different (see section 2.2.2) for the drained case where shear induced volume changes
within the soil takes place. However the volume loss (VL) is usually approximated as,
Vs
VL (This is normally expressed as a percentage value) (2.5)
D2
4
Where D, is the outer tunnel diameter.
14
A number of empirical correlations have been proposed which relate iy to the tunnel
depth to diameter ratio and soil type. For practical purposes it is often assumed that the
parameter i varies proportionally with z0 (OReilly and New, 1982):
i Kz0 (2.6)
Where,
K a constant called the trough width parameter
z0 depth of the tunnel axis below the ground surface.
Taking into account of different layers of soil in a soil profile, New and OReilly (1991)
suggested the following relationship:
i K1 z1 K 2 z2 .... ,
Where,
K1 and K2 are trough width parameter in layer 1 and layer 2 of depth z1 and z2
respectively.
Although the Gaussian approach is simple, the use of this method is limited to
Single tunnels or multiple tunnels where there is no significant interaction effect.
Estimate short term ground movement (no consolidation movement)
Greenfield sites where there are no structural interaction effects.
2.3.2 Numerical Prediction
Because of the limitations of the empirical method (e.g. the effect of structures on
tunnelling cannot be predicted), numerical modelling has increasingly been used to
examine effects of tunnelling. The most widely used numerical method appears is the
Finite Element method (FEM); the Finite Difference Method (FDM) is used
15
occasionally (e.g. using FLAC). The advantages of FEM over the empirical method are
its ability to:
Deal with complex ground conditions;
Model realistic soil behaviour;
Simulate the construction sequence;
Accommodate the interaction between multiple tunnels;
Account for adjacent services and structures;
Simulate immediate and long term conditions;
Deal with ground treatment (e.g. compensation grouting); and
Handle complex hydraulic boundary conditions.
The two dimensional and three dimensional Finite Element (FE) method is used to
predict the tunnel induced settlement, as described in the following sections.
Two Dimensional Finite Element Analysis
Tunnel excavation is a three dimensional problem. However three dimensional analyses

require excessive computational resources such as storage space and time of analysis.
Because of these complexities, two dimensional analyses are preferred over three
dimensional analyses. Various methods proposed to take account of the stress and strain
changes ahead of the tunnel in plane strain analysis are the: (i) Gap method (Rowe et
al., 1983), (ii) Convergence-Confinement method (Panet and Guenot, 1982), (iii)
Progressive Softening method (Swoboda,1979), (iv) Volume Loss Control method
(Addenbrooke et al., 1997, Liu et al., 2000, Franzius, 2003)
Three Dimensional Finite Element Analysis
Three dimensional FE analyses would consider the movement of the soil associated
with all the sources described in section 2.2. Different researchers used different
16
methods to model 3D tunnel construction and the modelling techniques of different

authors were summarised by Wongsaroj (2005). The modelling of shield tunnelling is
complicated as it involves modelling different parameters such as (i) face pressure at the
cutter face of the TBM (to stabilise the soil ahead of the TBM), (ii) movement of the
TBM, (iii) installation of lining at the back of TBM and closing the tail void using
grouting.
The face pressure at the cutter face is often modelled by applying prescribed pressure at
the excavated tunnel face; for example, Komiya et al. (1999) and Mroueh and Shahrour
(2003). The shield losses were modelled in three different ways such as by (i) applying
a percentage stress reduction at the tunnel boundary depending on the length of unlined
tunnel (Mroueh and Shahrour, 2002) (ii) applying a prescribed displacement to the
tunnel boundary (Lee and Rowe 1990a) and (iii) contracting the tunnel boundary
(Koelewijn and Verruijt, 2001). The tail void closure using grouting was modelled by
two methods. The first of these involves applying prescribed displacements or
contraction to the lining (Koelewijn and Verruijt, 2001). In the second method, grouting
is modelled either by applying an outward pressure to the lining (Ezzeldine, 1999) or by
expansion of the lining (Koelewijn and Verruijt, 2001).
Effect of Initial Stress Condition on Predicted Settlement trough
The influence of initial stress history on the predicted settlement trough both in 2D and
3D analyses have been reported by many authors. The predicted settlement troughs
using a higher K0 was often wider and shallower compared to the observed settlement
troughs in the field. Wider settlement troughs were reported by Addenbrooke (1996),
Franzius (2003) and many others for their 2D plane strain analysis even with different
soil models.
Previous literature also shows that width of the settlement trough predicted using 3D
modelling is also wider than observed in the field. Although the tunnelling can be
modelled in a more realistic manner in 3D, various authors such as Guedes de Melo &
Santos Pereira (2000) and Franzius (2003) also predict a wider and shallower trough in
soils with higher K0 values.
17
Franzius (2003) carried out both 2D and 3D FE analysis to study the influence of K0 and
soil anisotropy on the predicted settlement trough. His analyses concluded that, neither
the 3D effect nor the soil anisotropy can account for the wide transverse settlement
trough predicted by FE analyses in a high K0 regime. He found that although the trough
shape can be improved with unrealistically high soil anisotropy, this leads to
significantly higher volume loss. This is because of the reduction in vertical stiffness
due to the increase in horizontal stiffness (to meet certain shear modulus criteria). He
also found that with low K0 the magnitude of settlement and thus the volume loss
predicted was higher in a 3D analysis; he did not, however, comment on the effect of
mobilised strength. Franzius (2003) argues that the K0 value at the tunnel axis has a
significant influence on the nature of surface settlement trough and that a higher
specified K0 value also results in a wider prediction for the longitudinal trough. His 3D
analyses also showed that stabilised longitudinal movements at the start of the geometry
could not be obtained even after excavating a long distance (100m) - both for isotropic
and anisotropic cases; this is shown in Figure 2-7.
Ng and Lee (2005) also demonstrated through their three dimensional analysis (open
face tunnelling) that the surface settlement trough is governed by the combined effect of
stiffness ratio (Eh/Ev) and the K0 value. They found that as Eh increases, the predicted
settlement trough will be deep and narrow due to the mobilisation of small plastic
extension shear zones at the crown and at the invert. They also found that, for a given
stiffness ratio, the computed settlement trough is narrower if K0 is lower.
18
(a) Isotropic 3D analysis
(b) Anisotropic 3D analysis
Figure 2-7 Longitudinal Settlement Profile (Franzius, 2003)
19
2.4 BUILDING MOVEMENT DUE TO TUNNEL BORING
The ground movements arising from tunnelling are described in section 2.2 and have
been the subject of most previous research. However the movements occurring in built
up areas are less well understood due to the site-specific nature of soil structure
interaction effects. Unfortunately, apart from relatively recent work done on the Jubilee
line extension in London, few published data are available that report measurements in
the buildings and the subsoil subjected to tunnel induced settlements. The mode of
building movement observed by past researchers are summarised below.
The interaction between soil and structure controls the shape and magnitude of the
building settlement. Available examples illustrating the mode of building movement are
shown in Figure 2-8 and Figure 2-9. Figure 2-8 shows the predicted greenfield
movement and the measured movement of the Mansion House in London after a 3.05m
diameter tunnel was constructed 15 m beneath it. The predicted greenfield and the
measured movements are significantly different. Figure 2-9 shows the movement of a
building and the adjacent soil above subway tunnels in Frankfurt clay and this also
indicates a modification of the buildings movement compared to the adjacent ground.
Both of these cases indicate that the surface or building settlement due to tunnel boring
is an interactive problem as discussed by Potts & Addenbrooke (1997) and many others.
Figure 2-8 Predicted Greenfield settlement and Observed Building Settlement of

Mansion House, London (after Frischmann et al., 1994)
20
21
Figure 2-9 Building and Subsoil Movement of Buildings in Frankfurt Clay (after
Breth and Chambosse, 1974)
2.4.1 Mode of Building Deformation
In the past, the researchers have studied the effect of tunnelling on nearby buildings and
recommended certain building damage criteria for given modes of building
deformation. The monitored movement of the building in the past showed either a
sagging & hogging mode of deformation or tilt. An example from Mair et al. (1996) of a
building undergoing a sagging and hogging mode of deformation above a single tunnel
is shown in Figure 2-10. Mair et al. (1996) also categorised building damage in terms of
the deflection ratio ( L), with higher ratios being permissible for sagging than for
hogging.
22
Figure 2-10 Building Deformation above Single Tunnel (Mair et al., 1996)
Burland and Wroth (1975) also discuss sagging and hogging modes suggesting that
unreinforced walls are much more likely to be damaged by hogging than by sagging. An
exception to this was shown by Breth & Chambosse (1974) and Frischmann et al.,
1994. After observing the movements of concrete framed structures and masonry
structures above the subway tunnel in Frankfurt, Breth & Chambosse (1974) concluded
that the outside walls of many buildings behaved like rigid plate as long as the length-
height ratio was less than one.
The monitored movement (after Waterloo and City Line Link tunnel) of Mansion House
in London shows both sagging & hogging mode of deformation and tilt behaviour in the
same building. This building has 4 storeys and a basement under the front two thirds of
the building. This building is more than 200 years old and the main walls are of load
bearing brick walls. Figure 2-11 shows the layout of the Mansion house and the location
of tunnel under the building. Figure 2-12 indicates sagging and hogging mode of failure
for the west wall and tilt behaviour for the east wall of the Mansion House. This
example also shows that even the masonry building can have tilt behaviour agreeing
with the observation of Breth & Chambosse (1974). Hence it is clear that the length-
23
height ratio (which effectively controls the building flexural rigidity) plays an important
role in determining the mode of building deformation.
The settlement profiles of the building along the Jubilee Line Extension Project (JLEP)
also showed different building behaviour. For example, Neptune house (approx. L/H=
4) and Murdoch house (approx. L/H= 4) showed sagging and hogging and Clegg house
(approx. L/H= 2) showed a tilt behaviour (Dimmock, 2003). These examples also
clearly indicate the influence of building stiffness on settlement behaviour.
Figure 2-11 Layout of Mansion House and Location of Tunnel (Frischmann et al.,
1994)
24
Figure 2-12 Settlement profiles for East and West Walls of Mansion House
(Frischmann et al., 1994)
2.4.2 Numerical Studies of Interaction Problem
Although as mentioned in section 2.4, the surface or building movement associated with
tunnelling is an interactive problem, it has been studied less frequently, perhaps because
of the difficulty and complication involved in modelling the soil structure interaction
problem. As mentioned in section 2.3.2, the wider settlement trough often produced by
the numerical modelling has not been solved yet and this is another factor which
influences modelling and interpretation.
25
Some researchers have carried out 2D and 3D numerical modelling of the soil structure
interaction problem. Both masonry as well as framed structures has been modelled in
the past and a review of those studies is discussed below.
Potts and Addenbrooke (1997) studied the soil structure interaction problem through 2D
non-linear finite element analysis. They modelled the building as an equivalent
weightless beam resting on the ground surface. Although this doesnt depict the true
nature of the problem, they did a parametric study varying the axial and bending
stiffness of the building (represented as an equivalent beam). The London clay was
modelled using non-linear elastic plastic parameters. They concluded that the presence
of surface structure usually reduces the surface movement except when the building has
a low bending stiffness. Although they considered extreme cases of bending and axial
stiffness for the analysis, the tilt or rotation was not predicted as observed by Breth and
Chambosse (1974), or by Son and Cording (2008).
The interaction of soil with a masonry structure was studied by Burd et al. (2000)
through 3D FE analysis. They also adopted the soil properties representative of London
clay. From the numerical study, they concluded that the performance of building is
highly dependent on the mode of deformation (sagging or hogging) induced by the
tunnelling. They concluded that the soil structure interaction effect can reduce the
settlement induced building damage if the building deforms in a sagging mode.
However the building is less effective in reducing differential settlement if the building
deforms in hogging mode; this point is in agreement with the recommendation of Mair
et al. (1996).
Liu et al. (2000) carried out 2D FE analysis to study the interaction effect for masonry
structures and concluded that with increasing stiffness, the settlement trough becomes
flatter. Hence the stiffness of building faade can reduce the differential settlement
because of the flatter settlement trough. He also noticed tilt of the building which is
located partially over the tunnel.
Son and Cording (2005) studied the damage to a masonry building due to excavation
through physical model tests and 2D numerical simulation. They proposed a damage
estimation criterion based on angular distortion (rotation of the line joining two
26
reference points, relative to the tilt) and lateral strain, which they deduced from field
observations, physical model tests and numerical parametric studies; this is shown in
Figure 2-13. They also observed that cracking can significantly reduce the stiffness of
the building causing it to approach the Greenfield movement. This is in agreement with
Potts and Addenbrookes (1997) observation for building having low bending stiffness.
They found that the surface settlement can be greater than greenfiled movement if the
structure has low bending stiffness (but with a realistic axial stiffness).
Figure 2-13 Comparison of Damage Estimation Criterion and Damage Levels (Son
and Cording, 2005)
The merits of using deflection ratio to define the building distortion has been explained
by Burland et al.(2004) from a theoretical point of view comparing with field monitored
data from case studies. They concluded that the interpretation of building settlement
data in terms of deflection ratio is more closely related to tensile strain in the building
than with relative rotation (angular distortion). They suggested to use angular distortion
( if the building, or part of it, is undergoing pure vertical shear distortion. Burland
(1995) recommended an interaction diagram relating the deflection ratio and horizontal
strain for various L/H ratios and a typical example is shown in Figure 2-14.
27
Figure 2-14 Example of Interaction Diagram to Assess the Damage for Isotropic
Beam with L/H =1 (Burland, 1995)
Son & Cording (2008) studied the distortion and damage to masonry as well as frame
structures due to excavation through numerical modelling and physical model tests.
They observed and predicted tilt of masonry buildings as for framed structures. Based
on their study, they recommended that tilt effect and thus the angular distortion may be
a more appropriate parameter for building damage assessment than the deflection ratio
and this is contradictory to Burland et al. (2004).
In summary, the building can have sagging & hogging mode of deformation and /or
tilt/rotation depending on the length-height ratio; this ratio reflects the stiffness of the
structure.
2.5 COMPENSATION GROUTING
Compensation grouting is a process of injecting cement-based grout to consolidate or

compact the zones of soil under consideration. This procedure is widely used around the
world to mitigate the effect of tunnelling induced settlements. Compensation grouting
can either be compaction grouting (for granular material) or hydraulic fracture (or hydro
28
fracture for clay). To be successful, grouting needs detailed monitoring of the surface
movement in conjunction with careful grout injection control. The first published
application of the grouting was in UK at Waterloo Station (Harris et al., 1994). The
technique was successfully used around the world since then. The examples of
successful use of compensation grouting (fractured grouting) used in stiff clays were
reported for the Viennese subway project (Pototschnik, 1992) and during tunnel
construction at Waterloo station in London (Harris et al., 1994). Usually compensation
grouting is used in soft or stiff clays. However, there are a few cases of successful
compensation grouting carried out in granular soil. For example, protection of about 40
masonry buildings in Baltimore USA (Baker et al., 1983), protection of arch culvert in
Minneapolis USA (Cording et al., 1989), London Dockland extension project and
(Sugiyama et al., 2000 and Essler et al., 2000) and Bologna project (Kummerer et.al,
2007).
2.5.1 Numerical Modelling of Compensation Grouting
Although compensation grouting has been used successfully around the world,
published data of numerical modelling of compensation grouting are minimal because
of the very complex practical procedure. As mentioned in section 2.5, compensation
grouting needs detailed monitoring of the surface movement in conjunction with careful
grout injection control. Although all the various practical procedures of compensation
grouting cannot be modelled, it is possible to make an engineering judgement to attempt
to model the grouting. Previous research in this area uses either the prescribed strain
approach or the prescribed pressure approach method to model the grouting. In the
prescribed strain approach, appropriate values of strain are imposed on the elements
representing the grouted soil, whereas in prescribed pressure approach, grouting is
modelled by applying a pressure to a row of elements which represents the line of tube
manchettes (TAM). The prescribed pressure approach has been applied by Kovacevic
et al. (1996), Addenbrooke et al. (2002), Wisser et al. (2005), whereas Nicholson et al.
(1994) and Nicolini & Nova (2000) have used the prescribed strain approach.
Addenbrooke et al. (2002) carried out finite element modelling of a field trial of
compensation grouting trial in Singapore marine clay. However they had to use two
different grout propagation models to match the field monitored heave data and pore
29
pressure data (which is unrealistic). A numerical model was developed by Wisser et al.
(2005) to model compensation grouting in overconsolidated clay in a greenfield site.
They found that a more complex control algorithm is required to model compensation
grouting when a building is present above the tunnel.
While there are a few examples of modelling of compensation grouting in clay reported
in the literature, unfortunately none have been reported thus far in granular material.
2.6 SUMMARY
The first part of this chapter reviews the mode of failure induced by the tunnelling in
different soil conditions and the volumetric changes and stress changes around the
tunnel in drained soil conditions. The settlement prediction methods and factors
influencing the predicted settlement troughs are also reviewed. This Chapter also
discusses the influence of building in modifying the settlement profile of the ground as
well as the building and the suitability of numerical modelling to replicate the observed
movement. The effect of compensation grouting in reducing the tunnelling induced
movements and the usefulness of numerical modelling of compensation grouting are
also discussed. Some of main conclusions from this review include:
1) The mode of failure and hence the size and shape of the surface settlement
trough due to tunnelling is influenced by the type of soil. The settlement
trough is narrower in sand than in clay. (Chambon and Corte, 1994)
2) The settlement troughs predicted by numerical methods are significantly

influenced by the initial stress history (K0) of the soil; with high K0 value, the
settlement troughs tend to be wide and shallow compared to the field
measurements.
3) The buildings can experience sagging and/or hogging modes of deformation

as well as tilting/rotation. The mode of building movement depends on the
length-height ratio (or stiffness) of the structure.
4) Soil structure interaction effects can reduce the settlement of the building as
well as the ground surface.
30
5) Compensation grouting method is successfully used around the world to

reduce the effect of tunnelling. However the numerical modelling of
compensation grouting is in a relatively primitive stage.
Each of these five observations is addressed in this thesis by (i) interpreting field
measurements obtained during construction of the Perth MetroRail and (ii) through 2D
and 3D back analyses of these measurements.
31
32
CHAPTER 3 GEOLOGY, TUNNELLING
AND INSTRUMENTATION
3.1 INTRODUCTION
The New MetroRail City Project forms part of the recently completed Southern Suburbs
Railway Project in Perth, Western Australia. The railway project included 72 km of
double track railway between Perth and Mandurah, nine new suburban stations, thirteen
bridges and some additional structures. The Metro part of the project included two
underground city stations, two bored tunnels, cut and cover and dive structures. The two
tunnels pass underneath busy roads, bridges and existing railway lines. Both new and
relatively old multistorey commercial buildings exist on either side of the two tunnel
alignment. The stakeholders were very much concerned about the possibility of damage
to the existing infrastructure. Tunnelling was therefore carried out using an Earth
Pressure Balanced Tunnel Boring Machine (EPB-TBM) to minimise the volume loss
and thus to reduce or control the ground settlement. In this Chapter, a detailed
description of the Geology and Hydrogeology of the area, TBM tunnelling and the
instrumentation to monitor the ground and building movement is provided.
3.2 PROJECT DESCRIPTION
The New Metro Rail City Project consisted mainly of i) two bored underground rail
tunnel (total length of about 1.5 km), ii) cut and cover tunnel sections (total length of
about 1km) and iii) two Underground stations namely the Esplanade Station and
William Street Station, in the Central Business District (CBD) of Perth. The layouts of
the main features are shown in Figure 3-1. Chainage 0 (Ch 0) is at the centre of William
Street Station with increasing chainage (expressed in meters) towards north and south
directions. South of William Street Station (southern bored tunnel section) is referred to
as Perth to Mandurah (eg. 200 PM) and north of William Street Station (northern bored
tunnel section)is referred to as Perth to Butler (eg. 100PB). The construction of the 6.9m
33
diameter twin tunnels started from the Esplanade station towards William Street Station.
North of William Street Station, the twin tunnel passes underneath the Horse Shoe
Bridge and the existing railway lines and meets the existing rail network near Roe
Street. Table 3-1 lists the construction methods adopted for the various sections of the
New Metro Rail City Project. The main features of the project in the Perth CBD are
summarised below.
The two underground stations have an approximate length of 138m. The rail
level at the Esplanade station is approximately 9.0 m below ground level and at
William Street Station it is approximately 17.0 m below the ground level. An
underground concourse connects William Street Station and the existing Perth
Train Station.
Figure 3-1 Bored Tunnel Alignment
Two bored tunnels of 6.16m internal diameter with a segmental lining of

thickness 275.0 mm. The total length of the bored tunnel is approximately 1.5
34
Chapter 3 Geology, Tunnelling and Instrumentation
km. The rails are located between 9.0 m and 21.0 m below the existing ground
level in the bored tunnel location.
Chainage Structure Construction Method

(Approx.)
900PBdn to At-grade rail Ballasted track

670PBdn
670PBdn to Dive Structure and Cut Cut and Cover using temporary steel
340PBdn and Cover (C&C) tunnel sheet pile wall with struts
Roe Street
340PBdn to Bored tunnel Perth Excavation with earth pressure balanced

70PBdn railway yard tunnel boring machine (EPB TBM)
Pedestrian Access and Cut and Cover using temporary steel

Underpass sheet pile wall with struts
70PBdn to William Street Station Permanent diaphragm walls and top-

70PMdn (underground) down excavation method
70PMdn to Bored tunnel Excavation with EPB TBM
540PMdn
540PMdn to Esplanade station Cut and Cover using temporary steel
680PMdn sheet pile wall with struts
680PMdn to Dive structure and C&C Cut and Cover using temporary steel
1250PMdn tunnel Foreshore sheet pile wall with struts
1250PMdn to Freeway underpass Secant bored pile wall with top-down
1310PMdn construction sequence
1310PMdn to Dive structure and C&C Cut and Cover using temporary steel
1480PMdn tunnel Foreshore sheet pile walls and struts, grading to
open cut
1480PMdn to At-grade rail Ballasted track
1660PMdn
- Foreshore Bus Transit Predominantly near-grade civil works
lanes and road works` but combined with dive structure in some
locations.
Table 3-1 Construction Methods of the Different Sections of the New Metro Rail
City Project (Part 4B Geotechnical Interpretive Report, 2004)
The cut and cover tunnel at Roe Street is approximately 300m long and the rail
level is about 9.0m below ground level at its deepest point.
35
A cut and cover tunnel, open box structure and dive structure at Foreshore
extend from the southern end of the Esplanade Station until the rail rises to
ground surface
3.3 GEOLOGY AND HYDROLOGY OF THE PERTH REGION
Changes in the sea level during the Quaternary period as a result of the advance and
retreat of continental ice sheets have been well established for the Perth region
(Playford, 1988; Gozzard, 2007). The sea level variations discussed by Chappel and
Shackleton (1986), Playford (1988) and Gordon (2003), and summarised by Gozzard
(2007) are illustrated on Figure 3-2. It is apparent that there have been three major
periods of high sea level and three major periods of low sea level. These correspond to
interglacial periods and glacial periods respectively.
Figure 3-2 Eustatic Sea Level Curve for the Past 270,000 Years with Oxygen
Isotope Stages Indicated (Gozzard, 2007)
The sediments of the Perth region were formed during the late Tertiary to Quaternary
Periods (Davidson, 1995). Figure 3-3 shows the generalized geomorphology of the
Perth Region (Gozzard, 2007). To the east of Perth lies the Darling Plateau which
consists mainly of Pre-Cambrian laterite underlain by residual soil called lateritic
profile. Bordering the plateau is a narrow Ridge Hill Shelf consisting of residual
laterite and yellow sand on the surface. The yellow sand is believed to have originated
36
from weathering of this laterite. An unconsolidated riverine material, forming the

Pinjarra Plain, lies to the west of this shelf, the deposition of which begun in the early
Pleistocene (McArthur and Bettenay, 1960). The Bassendean siliceous sand dunes were
formed along the coastline and these dunes are believed to be younger than the Pinjarra
Plain because the rivers flowing from the Darling Plateau continued to build an alluvial
plain between the dune and the Plateau. The formation of the Spearwood and Quindalup
dunes is attributed to the fall in sea level (Andrews, 1971). The Quindalup dune is
believed to be formed over the last 5000 years and unconsolidated.
37
Figure 3-3 Perth Region Generalized Geomorphology (after Gozzard, 2007)
Figure 3-4 shows the generalized geology of the Perth region. The extent of the
Guildford Formation in outcrop and sub crop is seen from this figure. Most of the
Guildford formation overlies Jurassic and Cretaceous bedrock or the Pliocene Ascot and
Yogunup Formation (Davidson, 1995). The surface outcrop of the Guildford Formation
38
is dominated by clay-rich material while the subsurface is characterised by different

lithologies, which is a collection of sand, clay, silt or a combination of any of the two or
three with sand and conglomerate especially near the base of the formation (Gozzard,
2007). The lakes just north of Perth formed due to the ponding that occurred between
the Bassendean and Spearwood systems.
Figure 3-4 Perth Region Generalized Geology
3.4 GEOLOGY AND HYDROLOGY OF THE TUNNEL ROUTE
3.4.1 Geology
The stratigraphy comprises normally consolidated Spearwood (aeolian) sand overlying

the interbedded layers of alluvial silts, clays and sands of Perth Formation (over
39
consolidated), which are underlain by Kings Park Formation ( shale/ siltstone). The
schematic geology of the project route for the southern bored tunnel section is shown in
Figure 3-5 and each of these formations is described below.
Sand Fill
Spearwood Sand
Perth Formation SRA
Kings Park Formation
Figure 3-5 Schematic Geology of the Project Route (Hudson-Smith and Grinceri,
2007)
Spearwood (aeolian) Sand
Andrews (1971) has given a clear description of different layers of soil in different areas
of the Perth CBD. The Spearwood dune sands cover the central Perth area and formed
between the existing Bassendean dunes and the coastal line as a result of the lowering of
the sea level.
Perth Formation
The alluvial deposits underlying the dune sand are very complex in nature and comprise
interbedded sands, silts and clays. These alluvial deposits, commonly referred to as
Guildford Formation, are believed to have been deposited by the Swan river system and
it has proven difficult to clearly identify or define consistent soil layers throughout the
area. The deposition of the alluvium was believed to have started in the early
Pleistocene and continued to the early stages of the last interglacial period (Gozzard,
2007). However, it is now believed that the alluvial deposits present along the middle
reaches of the Swan River covering the Perth CBD are older than the Guildford alluvial
deposits further east and were deposited as an alluvial fill in a deep inset valley cut into
the Guildford formation. Gozzard (2007) refers to this alluvium as the Perth Formation
(see Figure 3-6), which is currently thought to have been deposited during the last
interglacial period between 75,000-130,000 years before present in a deep valley cut by
40
the ancestral Swan River during the second last glacial period (130,000 190,000 years
before present). At some locations the subsoil above the bedrock comprises two to four
strata, each being predominantly a sandy, clayey or occasionally a silty soil.
Figure 3-6 Schematic Section Showing Stratigraphic Relationships of Superficial

Formations in the Vicinity of Perth
The deposits are understood to be overconsolidated but the source of this

overconsolidation (which may be due to ageing/chemical processes) is unknown. The
formation also contains zones of cemented material with calcium carbonate and iron.
Typical pictures are given in Section 4.4.3. The lower layer is a combination of grey to
dark grey, occasionally brownish grey, mottled red clay and silty clay with some lenses
of sand. This unit is mostly found between about -6.0 m AHD and -16.5m AHD and
overlies calcareous shale of the Kings Park Formation.
Kings Park Formation (KPF)
The city of Perth which lies on the Swan Coastal Plain is developed on the sedimentary
rocks of the Perth basin. The sedimentary rock believed to have been formed during the
early Tertiary Age, consists of calcareous sandstone, siltstone or shale (Fahey et al.,
2003). The elevation of the sedimentary rocks beneath the Perth CBD varies between an
RL of -16.0m to -21.0m. The Sswan coastal plain, which includes the metropolitan area,
41
is formed almost entirely of depositional material either from fluviatile or aeolian
activity, with occasional marine transgressions (Andrews, 1971).
Swan River alluvium (SRA)
As mentioned in Section 3.5, southern part of the Esplanade was reclaimed in 1950s
and 1960s from Swan River. This reclaimed area consists of black organic silty clay to
clayey silt with occasional thin bands of quartz sand and this is termed as swan river
alluvium. This is believed to have been formed during the late Pleistocene to Holocene
age within marine, estuarine and reverine environment.
3.4.2 Hydrogeology
The hydrogeology of the project area is controlled by the ground water level in the Perth
Rail Yard area to the North and the ground water level in the Foreshore area to the
south. The ground water level in the Perth Rail Yard Area is controlled by original
water level in the dunal lakes and the water level in the foreshore area is controlled by
the Swan River. Water flows from the Perth Rail Yard to the Foreshore through the
interbedded layers of Perth Formation (PF) below the Spearwood sand (SS) and,
consequently there is potential for multiple aquifers. Figure 3-7 shows the typical
piezometric profiles at south end of William Street station. Figure 3-8 shows the
extrapolated piezometric profile along the project route. The investigation for the tunnel
project showed the evidence of water table below mean sea level at some locations in
Perth CBD and concluded that the groundwater abstraction due to construction and
other basement pumping systems could be the reason (Golder Associates, 2003).
42
Ground water pressure, kPa Ground water head (mAHD)

0 50 100 150 200 250
-4 -2 0 2 4 6 8 10 12
15
15
10
10
5
5
0
0
-5
-5
-10
-15 -10
-20 -15
-25 -20
-30 -25
-30
Figure 3-7 Piezometric Profiles at South End of William Street Station (Leighton
Kumagai Joint Venture., 2004)
43
Figure 3-8 Piezometric Profile along the Project Route (Johnson et al., 2007)
44
3.5 TUNNEL ALIGNMENT AND GROUND CONDITIONS
The TBM was launched at the Launching Shaft, which was excavated just north of the
Esplanade Station (see Figure 3-1), and moved northwards underneath William Street
and a group of buildings to reach the William Street Station. After reaching the station,
the TBM was re-launched to complete the part of the tunnel north of William Street
Station. The TBM then passed underneath the Horseshoe Bridge and the existing
Railway lines and the platforms of Perth Railway Station before turning into Roe Street
in a very tight curve of only 135 m radius. The tunnel then passed at very close
proximity to Claisebrook Sewer and arrived at the receiving shaft in Roe Street. The
TBM was then brought back to the Esplanade Tunnel Launching Shaft and launched for
the second tunnel.
The tunnel route passes through varied geological zones. The geology and the
hydrogeology of the tunnel route are described in Section 3.3. The tunnel route can be
divided mainly into four parts and are explained below.
i) Foreshore Reclamation area
The Esplanade Station and Foreshore cut-and-cover section for the project are located in
an area reclaimed (in 1950s and 1960s) from the tidal estuary known as Perth Water.
Consequently, this area is covered by hydraulically filled sand and is underlain by the
estuarine mud (referred to as Swan River alluvium). The Tunnel Launching Shaft
located at the northern end of the Esplanade station is at the edge of the Foreshore
reclamation area.
ii) Bored Tunnel between the Esplanade Station and William Street Station
The Tunnel leaves the Esplanade station and goes underneath William Street to reach
William Street Station. Multi-storey commercial buildings exist on both sides of
William Street and some of these are heritage listed. The tunnel mainly passes through
the Perth Formation at this location.
45
iii) Bored Tunnel North of William Street Station
In the north, after leaving the William street station which is in Perth Formation (PF),
the TBM gradually entered the Spearwood sand. The ground water profile south of the
Perth Rail Yard was sub-hydrostatic but was hydrostatic north of this yard.
iv) Old Northbridge Lake System
The northern part of the rail line passes mainly through reclaimed area. Figure 3-9
shows the plan of the area in 1838, which indicates the rail line passing through an area
covered by the lakes namely Kingsford, Sutherland and Irwin. These lakes were formed
by ground water exposed in shallow interdunal swales. In the mid 1800s, these lakes
were filled to eliminate mosquito problems and to create more land for development.
The upper 1m-2m of the ground in this area comprises organic peaty material
Figure 3-9 Old Lakes near Project Alignment
3.6 SITE INVESTIGATIONS
A detailed soil investigation including field investigation and laboratory studies was
carried out by a variety of geotechnical firms during the feasibility stage and just prior
to and during tunnel construction. As shown in Table 3-2, six phases of geotechnical
site investigations were carried out for this project.
46
Phase Time Period Details
1 Prior to calling for expressions Scoped and commissioned by Public

of interest (January March Transport Authority (PTA), of Western
2003) Australia
2 During Expression of Interest Scope proposed by five EOI participants and
(EOI) stage, (April June commissioned by PTA
2003)
3 During the tender design stage Scope proposed by two Proponents, partial
(April August 2003) scope commissioned by PTA
4 During tender evaluation stage Scoped in consultation with Preferred
(November 2003 January Proponent. Partial scope commissioned by
2004) PTA
5 Immediately following the Scoped and commissioned by (Leighton
award of contract (January Kumagai Joint Venture, the contractor)
April 2004) LKJV
6 During construction Scoped and commissioned by LKJV
Table 3-2 Phases of Site Investigation
Investigations were also carried out to assess the presence of acid sulphate soil,
contaminated soil and ground water along the tunnel route.
The following field investigations were carried out during the course of the Metro Rail
project.
84 No. Electric Friction-Cone penetrometer Tests, including standard friction-

cone, Piezocone and Seismic Cone Testing;
15 No. profiles of Marchetti Dilatometer Tests;
94 No. geotechnical boreholes with Standard Penetration Testing, thin-walled

tube sampling and HQ diamond coring;
9 No. borehole profiles of Field Vane Shear Strength Testing;
Self-Boring Pressure Meter Testing within 5 No. boreholes;

47
Spectral Analysis of Surface Waves (SASW)testing at 3 No. locations;
20 No. Groundwater Monitoring and sampling bores;
Field Permeability Tests within 5 No. boreholes;
3 No. 48-hour Pump Tests of production bores;
13 No. Backhoe Test Pits and 12 No. hand auger boreholes; and
Gas Monitoring in 10 No. boreholes.
A detailed laboratory study was also carried out on both disturbed and nominally
disturbed samples collected from the field. The samples were either push samples or PQ
cored samples typically 63mm diameter. The following laboratory tests were carried
out:
Classification tests to determine particle size distributions, Atterberg limits,

water contents, organic contents, mineralogies, dispersivity and abrasivity X-ray
diffraction and SEM photomicrographs.
Strength determinations (drained & undrained) in triaxial and shear box tests.
Determination of stiffness and consolidation properties in triaxial and oedometer

tests.
Chemical tests of the soil and ground water to detect the presence of sulphates
and chlorides.
Unfortunately, the specified laboratory investigations were not conducted to research

standards and few, if any, state-of-the-art tests (including, for example, Rowe cell tests,
simple shear tests and triaxial tests with local strain instrumentation and bender
elements) were conducted. The author was involved in a more detailed and systematic
separate study of the properties of the clay-rich horizons of the Perth Formation. This
research is summarised by Lehane, Mathew and Stewart (2007) and is presented in
Chapter 4.
48
3.7 EARTH PRESSURE BALANCE TUNNEL BORING MACHINE
Sigl and Yamazaki, 2007 have given a detailed description about the tunnel route and
TBM selection and all the aspects related to tunnelling. Since the two tunnels were
constructed beneath multi-storey buildings, active railway lines and busy roads, it was
necessary to keep the volume loss and thus the settlement to a minimum to reduce or
avoid damages to existing infrastructure. Figure 3-1 shows the aerial view of the
alignment of bored tunnel. Also the tunnel alignment had other technical issues such as
abandoned ground anchors and well casings, which demanded a careful monitoring of
the tunnelling. All these factors led to the final decision to adopt an Earth Pressure
Balanced Tunnel Boring Machine (EPB-TBM). The main issues which required special
attention for the operation of the TBM are:
High rise buildings on either side of William Street, some of which are heritage
listed;
A group of buildings on shallow foundations just south of William Street Station

where the cover was low;
Abandoned ground anchors from previous deep basement excavations along

William Street1;
The heritage listed Horseshoe Bridge and the rail lines and the platforms of the
existing Perth Railway Station,
Tunnelling through loose sand at locations where the tunnel axis is at shallow
depth (near the eastern end of Roe Street cut and cover section).
Tunnelling just south of the existing Rail Yard, where water pressures are at full
hydrostatic values.
1
A TBM hitting such anchors can lead to high volume losses as well as cutter head damage
49
Hence to minimize the settlement during tunnelling, the two tunnels were excavated by
an EPB-TBM equipped with special features. The main Characteristics of the EPB-
TBM are given in Table 3-3.
Feature Details
Manufacturer Mitsubishi Heavy Industries, Kobe, Japan

Installed cutter head power 540kW
Excavation diameter 6,900 mm
Shield length 8,480mm
Number of cutters Typically 80 nos. drag bits, no discs
Thrust force Maximum 20 x 2,000 kN = 40,000 kN
Articulation jacks 8 x 3,500 kN
Maximum torque 5,485 kNm (100%): 6,582 kNm (120%)
Segmental Lining 6,160 mm ID., 275 mm thick, 5+ 1key
Segmental length 1,000 mm and 1,200 mm
Table 3-3 Main Characteristics of the EPB-TBM
The main features of the EPBTBM are explained in the following sections.
3.7.1 Ground Anchor Detector
The TBM had three anchor detectors installed to detect abandoned anchors or old wells
(cast iron or brick) in the ground. The details and specifications of the anchor detectors
are shown in Figure 3-10 and Figure 3-11. On detecting an obstacle, the hydraulic
cylindrical shaft attached with the cutter head would be pushed backwards increasing
the pressure in the hydraulic system. This unusual pressure in the hydraulic system
triggers an alert on the control display panel.
50
Figure 3-10 Specifications of EPB-TBM
51
Figure 3-11 EPB Tunnel Boring Machine- View on Cutter head
3.7.2 Special Cutters
To minimize the risk of damage to the cutter head, knife edge cutters were installed on
the cutter face of the TBM to cut through the obstacles such as abandoned ground
anchors and old well casings. These cutters are in addition to the ordinary drag bits used
for the tunnel boring.
3.7.3 Operation in Curved Alignments
In curved alignment, TBM steering is difficult or, in some cases, impossible. To avoid
such situations, two hydraulically operated copy cutters and a shield articulation were
installed on the TBM. Copy cutters were installed on the hydraulic jacks on the outside
of the rotating cutter head. They can extend and retract during rotation and can make an
oval shaped boring. Since the over cut in curved alignment is localised, the settlement
52
can be reduced. Shield articulation is installed between the front and the tail shield and
they comprise hydraulic jacks. The tail shield axis can be angled by extending and
retracting the jacks to better fit the alignment curve.
3.7.4 Ground Conditioning System
Settlement due to EPB-TBM is best controlled by properly monitoring the face pressure
and the muck discharge volume. Control of these parameters can be improved by using
suitable soil conditioning agent through a conditioning system. The conditioning agent
may be water, bentonite slurry, various polymer materials or foam. Because the TBM
moving through mixed soil conditions, a foam/polymer soil conditioning system was
used.
3.7.5 Back Fill Grout System
To avoid excessive settlement, the space between the extrados of the segmental lining
and the line of excavation has to be pressurised and backfilled as the TBM is moving
forward. To achieve this, the TBM was equipped with a tail void grouting system with
accelerated grout mixer.
3.7.6 Advanced TBM Operation Monitoring System
The TBM operation system ENZAN Arigataya monitored and recorded more than
150 items (such as grout volume, grout pressure, earth pressure, TBM thrust, foam
pressure etc.) four times per second. This operation system could display the data on the
operation panel in the TBM operators room as well as in the TBM monitoring room in
the tunnelling office. Two CCTV monitors were also part of the system to monitor and
record the condition of the discharged soil and general status of the work. Therefore a
detailed operation report was available for each ring.
3.7.7 Muck Volume Measuring System
To compare the extracted soil volume with the calculated volume, the following
independent methods were adopted during tunnelling.
53
Screw conveyor discharge;
Conveyor belt muck volume scanning;
Conveyor belt muck weight scanning;
Manual counting of muck skips; and
Weighing of muck skips;
The first three parameters were monitored by the TBM monitoring system. The last two
parameters were carried out manually to counter check the excavation volume in
sensitive ground.
3.7.8 Internal Grout Ports
There were number of grout ports on the crown of the TBM to treat the soil; these ports
were used, for example, when the TBM passed through sandy soil.
3.7.9 Compressed Air Work Facility
The TBM was equipped with a compressed air work facility, which could be used if an
emergency repair was required for the cutter head.
3.8 EPB-TBM TUNNELLING PROCEDURE
Most tunnel boring machines (TBMs) have a supporting shield which restrains the soil.
The shield also accommodates all equipment and accessories associated with the
excavation and it also facilitates erection of the segmental lining. Earth Pressure
Balance (EPB) machines will provide a pressurised support at the face by filling the
chamber at the head of the shield with soil. This is to give stability to the soil in front of
the cutter head and thus avoid excessive soil movement. The pressure in the chamber is
gradually reduced to atmospheric pressure over the length of the screw conveyor which
removes the muck. The construction sequences of the EPB tunnelling are:
1. Excavate a distance in front of the TBM equivalent to the length of the lining.
2. Advance the shield one ring length forward by jacks, thrusting against the previously
installed lining.
54
3. Retract the jacks and install the prefabricated concrete ring segments (lining) which
were bolted together within the TBM.
4. Fill the gap between the lining and the ground (tail void) by grouting to reduce the
settlement.
5. Boring is resumed until sufficient soil has been excavated to erect the next lining.
3.9 GEOTECHNICAL INSTRUMENTATION AND MONITORING
Since the project involved deep excavations and tunnels, a significant numbers (~5200
instruments) of manual as well as automated instrumentations were installed to monitor
(McGough and Williams, 2007);
the settlement of buildings, services, roads, rails and bridges;
the performance of deep excavations;
sheet piling, bored piling and diaphragm wall construction;
tunnelling;
ground improvement activities such as jet grouting, soil mixing and

compensation grouting; and
Consolidation movement from ground water drawdown.
The ground movement, structural response and ground water pressure at the Metro Rail
Project site was monitored by Leighton Kumagai Joint Venture using the following
instrumentation. Some of this instrumentation was installed specifically at the request of
this author (e.g. Electro levels beams on the Malaysian Airlines building).
Vertical ground movement (Settlement Pins, Settlement Points and

Extensometers)
Lateral ground movement (Inclinometers)
Building movements including tilt (Electro Level Beams, Tilt Meter and Retro
Targets)
Structural response (Crack meters and Strain Gauges)
55
Vibration (Vibration Sensor) and
Ground water monitoring (Vibrating Wire Piezometers and Open Hole

Piezometers).
This thesis focuses on measurements of surface and building movement due to

tunnelling.
Table 3-4 summarises the instrumentation types and quantities installed along the 3.0km
stretch of the project (McGough and Williams, 2007).
Instrument Type Designation Quantity Installed

Surface Settlement Pin SSP-1 1021
Surface Settlement Retro SSP- 2 451
Bored Settlement Point SSP- 3 559
Deep Settlement Point SSP- 4 19
Building Settlement Point BSPB 449
Building Settlement Retro BSPR 1403
Building Settlement Prism BSPP 285
Tilt Meter, Manual TILTM 54
Tilt Meter, Automatic TILTA 33
Crack Meters CM 82
Electro Level Beams ELB 150
Strain Gauges SG 174
Vibration Sensor VS 12
Inclinometers INCL 64
Extensometers, Magnetic EXTM 187
Extensometers, Rod EXTM 25
Vibrating Wire Piezometers VWPZ 91
Open Hole Piezometers OHPZ 146
Total 5205
Table 3-4 Instrument Types and Quantities
56
3.10 SCHEDULING OF TUNNEL CONSTRUCTION
As explained in section 3.2 the tunnel can be divided into two sections, namely north of
William Street Station (PB) and south of William Street Station (PM). The detailed
schedule of tunnelling is shown in Table 3-5. Figure 3-12 and Figure 3-13 shows the
schedule of tunnelling south and north of William Street Station respectively.
Tunnel Chainage Starting Date Completion Date
Tunnel 1 (PMdn) 70 PMdn 540 PMdn November 5th 2005 February, 9, 2006
Tunnel 1 (PBup) 340 PBup 70 PBup May 3, 2006 June 5, 2006
Tunnel 2 (PMup) 70 PMup 540 PMup July 18,2006 August31,2006
Tunnel 2 (PBdn) 340 PBdn 70 PBdn September 23, 2006 October27,2006
Table 3-5 Schedule of Tunnel Construction
Bored Tunnel -South of William Street Station
0
Tunnel 1
100 Tunnel 2
200
300
400
500
600
25-Oct-05 24-Dec-05 22-Feb-06 23-Apr-06 22-Jun-06 21-Aug-06
Date
Figure 3-12 Schedule of Tunnelling South of William Street Station
57
Bored Tunnel- North of William Street Station
400
350
300
250
200 Tunnel 1
150 Tunnel 2
100
50
0
24-Mar-06 13-May-06 02-Jul-06 21-Aug-06 10-Oct-06 29-Nov-06
Date
Figure 3-13 Schedule of Tunnelling North of William Street Station
58
CHAPTER 4 SOIL CHARACTERISATION
4.1 INTRODUCTION
The stratigraphy in the Perth CBD (Central Business District) comprises three
formations. The bedrock, or the Kings Park Formation (KPF), underlies the over
consolidated Perth Formation (comprising inter bedded layers of clay, silt and sand);
normally consolidated Spearwood (Aeolian) sand overlies the Perth Formation. The
MetroRail tunnels passed through the overconsolidated Perth Formation south of the
William Street Station and a significant portion of the northern bored tunnel (north of
William Street Station) passed through the Spearwood sand. While there have been a
number of studies in the properties of the Spearwood sand, no systematic laboratory
investigation of the properties of the Perth Formation has been published to date.
Therefore, as part of this research, a relatively comprehensive programme of laboratory
tests was performed on the upper (clayey) horizons of the Perth Formation. This
Chapter therefore summaries available information for the dune sand (Spearwood sand)
and then presents results from a laboratory investigation of the Perth Formation. The
soil parameters for the numerical analyses (Chapter 6, 7 and 8) were deduced from these
laboratory tests in combination with the in-situ test data.
4.2 SUMMARY OF DATA ON SPEARWOOD SAND
The Spearwood sand is of quartz Aeolian sand and is approximately 8 m thick in area of
interest in the Perth CBD. The soil investigation for the MetroRail project suggests that
the upper facies of this formation is about 3m thick and comprises yellow to dark
yellow sand with traces of silt; the lower facies comprises white to light brown quartz
sand (Hudson-Smith and Grinceri, 2007). Differences in engineering behaviour
between the facies have not been found.
A detailed description of the dune sand has been given by Andrews (1971), who states
that the colour changes of white through yellow to dark brown indicate the presence of
59
iron deposits within the sand profile and the occasionally seen dark grey or black on the
surface is due to the organic matter or rutile. However, no iron cemented sand was
noticed in any of the bore holes along the tunnel route (Golder Associates, 2003).
Andrewss database shows that the sand is in a loose condition near the surface and
becomes denser with depth. He pointed out that the SPT (Standard Penetrometer Test)
value is greater than 30 at about 4.5m to 6m from surface and this is in agreement with
the observations made during the MetroRail project. The SPT N value at Perth CBD are
presented in Figure 4-1 and also show that material is generally medium dense.
Typical particle size distribution ranges measured for the dune sand are shown in Figure
4-2 and indicate that the sand is uniformly graded; more than ~80% fall within the
medium size classification. Lehane et al. (2004) reported a relative density (based on
sand replacement tests) of 4510 % for the dune sand at a test location just outside
Perth CBD (although this sand had qc values 20 to 40% less than typically found in the
tunnel area). Lehane et al. (2004) report effective particle sizes D50, D60 and D10 in the
order of 0.42 0.02mm, 0.470.02mm and 0.210.01mm respectively. Anisotropic
drained triaxial tests (with local strain measurements) on soil samples reconstituted to
average in situ void ratio of 0.63 corresponding to a relative density of 50% showed a
friction angle of 350 (Lehane et al. 2008). These triaxial samples contracted during
triaxial shear loading.
60
Chapter 4 Soil Characterisation
SPT - N
0 10 20 30 40 50 60 70 80 90 100 110 120 130
0
4 Spearwood sand
8
12
16
20
24
28
32
36
40
Figure 4-1 SPT N vs. Depth
Figure 4-2 Typical Gradation Curve of the Dune Sand (after Andrews, 1971)
61
Fahey et al. (2003) have recommended a relationship between G0 and qc derived from
Seismic Cone Penetration tests (SCPTs) carried out in Perth CBD sand; this is shown in
Figure 4-3. The upper and lower bounds to the data presented in Figure 4-3 are
expressed as:
G0 4503 qc ' v patm (Upper bound)
G0 1103 qc ' v patm (Lower bound)
Fahey et al (2007) inferred unload reload moduli (Gur) of 40% to 50% of G0 from
SCPT shear wave velocity measurements.
Upper bound
Lower bound
Figure 4-3 Relationship between G0 (from SCPTs) and qc Observed in Perth Sand
Figure 4-4 compares the seismic cone Go values for the Spearwood sand and the sandy
horizons of the Perth Formation. It is apparent that the relationship with stress
62
normalised cones resistance (qc1, as defined in Figure 4-3 for both sands appears to be
the same. Ageing effects for both Deposits thus appear similar. This trend is
incorporated later in the thesis, where the same relationship for G0 for both sands is
employed in Plaxis analyses (Section 6.3).
100
Perth Formation sands
Spearwood sand
10
1
10 100 1000
qc1
Figure 4-4 Relationship between G0 (from SCPTs) and qc Observed for Perth
Formation Sand and Spearwood Sand
Schneider et al. (2008) carried out a detailed study of Spearwood sand (just outside of
Perth CBD) using in-situ tests such as Seismic Cone Penetration Tests (SCPTs), flat
plate dilatometer (DMTs) and self boring pressuremeter test (SBPTs). They estimated
peak friction angles of between 360 and 400 and a K0 value of 0.4 to 0.45. Fahey et al.
(2007) report a peak (plane strain) friction angle from SBPTs of as high as 500 for Perth
CBD dune sand.
4.3 LABORATORY INVESTIGATION ON PERTH FORMATION
The undisturbed soil samples (rotary cored) used for the laboratory tests were collected
from a site (St. Georges Terrace) closer to the tunnel site. The undisturbed soil (rotary
63
cored PQ-3) samples were collected from two boreholes about 25m apart at the St.
Georges Terrace site. The sample which was collected from a depth of between 8m
(6m RL) and 17 m (-3mRL) depth comprised of grey very stiff to hard silty clay, sandy
clay and clayey sand. Cores from this horizon were recovered from the coring
equipment and excess moisture was removed quickly before carefully scraping away a
zone (~5mm thick) of softened soil around the samples perimeter with a spatula; this
zone was removed to minimise the loss of sample suction due to the presence of the
drilling fluid. The cores were then wrapped in two layers of polythene and transported
to the university (UWA) where they were waxed prior to placement in the sample
storage facility. The laboratory investigation of the intact samples comprised:
Grading, water contents and Atterberg limits
X-ray diffraction analyses
Electron microscope imaging
Oedometer (1-D compression) tests
Filter paper suction tests
Consolidated undrained triaxial tests (some with local strain instrumentation)
Undrained simple shear tests
A consolidated undrained test was also performed on a reconstituted sample to provide

information on the structure of the intact material. All of the laboratory test data are
compared with the in-situ test results and also with laboratory data recovered during the
site investigation for the MetroRail project.
Measured groundwater levels in the vicinity of the site are characterised by two distinct
aquifers, which are separated by the 8 to 9 m thick clayey layer between about 8 m and
17 m depth. Piezometers indicate that the piezometric level in the upper dune sand
aquifer is at a depth of ~6 to 7 m. The piezometric level recorded in the lower aquifer
below 17 m is at ~15 to 16 m depth, which is slightly below sea level. Historical data
indicate that the piezometric level in this lower aquifer would normally be close to or
slightly above sea level; construction related dewatering in the vicinity is thought to
64
have depressed this level by 1 to 2 m. The piezometric levels in the zone of interest
are interpreted here to be between the levels in the overlying and underlying sand units.
4.4 SOIL CLASSIFICATION
The boreholes (from which the undisturbed samples recovered) were located adjacent to
a number of cone penetration tests (one of which included seismic shear wave velocity
measurements over the depth of interest and another which provided pore pressure
measurements). The boreholes were also close to flat plate dilatometer test (DMT)
soundings, which comprised DMT readings at 0.25 m depth intervals. The soil which
was under investigation was classified based on the field (field data obtained from
Golders Associates, Perth) as well as from different laboratory tests and each of those
are explained below.
4.4.1 Soil Classification from in-situ Tests
Figure 4-5 presents typical cone penetration test (CPT) data at this site and indicates a 9
m thick deposit of soil with a high friction ratio, Fr (of typically 3-5%) underlies dune
sand with Fr ~0.5%. This high Fr material is typical of the clay-rich horizons of the
Perth Formation and was investigated thoroughly in a range of laboratory and in-situ
tests discussed in the following.
65
qc (MPa) Friction ratio (%)
0 10 20 30 40 50 0 1 2 3 4 5 6 7
16 16
14 14
Fill
12 12
10 Dune Sand 10
8 8
6 6
4 4
2 2
Sample Location
0 0
-2 -2
-4 Alluvial -4
-6 Sand
-6
-8 -8
-10 -10
Figure 4-5 Typical CPT Data at Test Location
The CPT is the most commonly employed profiling tool and soils are classified using
the soil behaviour type (SBT) chart proposed by Robertson (1990). Rather than using
the chart, it is often convenient to plot the SBT index, Ic, defined by Robertson (2004)
as:
Ic= [ (3.47- log Q)2 + (log F+1.22)2]0.5 (1a)
where Q = [(qc v0)/pref ] [ pref / v0]n (1b)
F = fs/ (qc vo) 100 (%) (1c)
v0 is the vertical total stress
fs is the friction sleeve measurement
66
pref is a reference stress = atmospheric pressure in the same units as qc
The value of n in equation 1b varies as follows:
n = 0.5 if Ic < 1.64 (1d)
n = 1.0 if Ic > 3.3 (1e)
n = (Ic-1.64) 0.3 + 0.5 if 1.64 < Ic < 3.30 (1f)
The profile of Ic derived for the data plotted on Figure 4-5 is shown on Figure 4-6a and
provides a clear picture of the soil type variation with depth. The material between 8 m
and 17 m evidently classifies as clayey silt to silty clay with some sand lenses. Both the
dune sand above this horizon and the sand of the Perth Formation below 17m classify as
sands and silty sands.
The Marchetti dilatometer (DMT) is occasionally used in Perth as a profiling in-situ

device. Soil type using this device is inferred using the DMT material index, ID, defined
as:
ID = (p1-p0)/(p0-u0) (2)
where p0 is the lift-off pressure, p1 is the pressure measured at a membrane expansion of

1.1 mm and u0 is the ambient pore pressure. ID values derived from a series of DMTs at
the site are shown on Figure 4-6b and suggest a generally similar profile of soil types to
those inferred from the CPT Ic value. Variations between the elevation of the different
soil types from the CPT and DMT are likely to reflect the actual lateral variation across
the site.
67
Ic ID
1 1.5 2 2.5 3 3.5 0 0.6 1.2 1.8 2.4 3 3.6 4.2
16 16
Clean 14 Clay Silt Sand
14 Grave
Sand to Silty
lly Clayey Silty Clay 12
12 Silty Sand
Sand Silt to Clay
Sand to
10 to to 10
dense Sandy Silty
8 sand Silt Clay 8
6 6
4 4
2 2
0 0
-2 -2
-4 -4
-6 -6
-8 -8
-10 -10
(a) (b)
Figure 4-6 Ic and ID Indices at the Site
4.4.2 Classification based on Laboratory Tests
The water contents and Atterberg limits determined for the core samples obtained
between 6 m RL and -3m RL in the two boreholes are plotted on Figure 4-7. These
indicate a relatively wide variation of plasticity index (PI), varying from 19% to 66%.
The liquid limits (LL) and PI values plot above and parallel to the A-line of the soil
plasticity chart and the liquidity index are typically about zero.
68
LL,w,PL (%) Fine content (%) Clay fraction (%)

0 20 40 60 80 100 0 20 40 60 80 100 0 10 20 30 40
6 6 6
5 5 5
4 4 4
3 3 3
2 2 2
1 1 1
0 0 0
-1 -1 -1
-2 -2 -2
-3 -3 -3
Figure 4-7 Classification Data; Separate Symbols for each Bore Holes
The grading curves obtained for the samples were also variable and, as shown in Figure
4-7, the fines content (FC) 1 ranges from about 20% to 95% and the clay fraction (CF)1
varies from 3% to 37%. According to the Unified Soil Classification System (AS 1726
Geotechnical site investigations), materials with a fines content less than 50% are
classified as inorganic clayey sands (SC) while those with a fines content in excess of
50% are referred to as inorganic silty clays and sandy clays (generally CI & CH). The
gravel fraction of all samples recovered was zero and the sand fraction was composed
primarily of particles less than ~150 m.
1
Fines content = proportion of particles less than 75 m in size, Clay fraction = proportion of particles
less than 2 m in size, Activity = Plasticity index/Clay fraction
69
80
Sodium Calcium
montmorillonite montmorillonite
60
Illite
40
XRD Samples
Kaolinite
20
0
0 20 40 60 80
Clay-sized particles (% )
Figure 4-8 Activity Chart and Location of Tested Soil Samples
Sample activity values1 are relatively constant at 1.6 0.4. As shown in the activity
chart on Figure 4-8, the range of activities are consistent with the presence of smectites/
montmorillonites within the clay fraction and do not suggest a dominance of kaolinite.
Tests to assess the actual soil composition and mineralogy are discussed in the next
section.
4.4.3 Classification based on Soil Composition
The soil composition was identified through a variety of tests such as X-ray Diffraction
(XRD) analyses, XRF (X-ray Fluorescent spectroscopy) and SEM (Scanning Electron
Microscopy). An initial batch of X-ray Diffraction (XRD) analyses was conducted on
representative soil sample. XRD patterns were recorded by a Siemens D500 device
using Cu-radiation (40kV, 30mA) over a 2 angular range of 5o to 80o at 2s/0.02o. The
measured diffraction patterns were interpreted using in-house database search/match
software. Chemical element information provided by EDS (Energy Dispersive X-ray
Spectroscopy) in the scanning electron microscopy (SEM) study was used for
search/match analysis.
70
The XRD results obtained for three typical samples are summarized in Table 4-1, which
also lists classification data (e.g. FC, CF and PI) for each of these samples as well as the
elements detected using EDS. It is evident that, contrary to the indications of the activity
chart on Figure 4-8, kaolinite (and not montmorillonite) dominates the clay fraction
with typically one third of the each sample comprising kaolinite. Another third
(approximately) of each sample comprises quartz. These relatively large proportions of
quartz and kaolinite suggest that the Perth Formation is a weathering product of the
granite on the Darling Scarp (e.g. see Mitchell & Soga, 2005), which was subsequently
deposited downstream in a fluvial/riverine environment in Perth CBD. Microcline,
Anorthite and Protoenstatite occupy between 12 and 19% of samples 1, 2 and 3
respectively and their presence is consistent with the occurrence of potassium in sample
1, calcium in sample 2 and magnesium in sample 3.
Sample 1 Sample 2 Sample 3
Grading and index Depth=8.9m Depth=10.8m Depth=15.4m

properties CF =15 % CF =9 % CF= 19 %
FC=74 % FC=51 % FC = 77%
PI= 19 % PI= 19% PI = 24 %
Activity=1.27 Activity=2.1 Activity = 1.27
Mineral
Quartz 34.2 44.3 24.7

Microcline 18.5 - -
Kaolinite 29.5 37.0 35.5
Muscovite 8.7 - -
Cristobalite 1.2 - -
Mannardite 3.2 - -
Iscorite 4.7 - -
Rutile - 3.6 2.2
Anorthite - 15.1 -
Protoenstatite - - 12.7
Elements Identified C, O, Fe, Al, Si, K, C, O, Fe, Al, C, O, Fe, Al, Si, K,
in EDS/SEM Ti, Mn, Zr, S, Ba Si, K, Ti, Ca Ti, Zr, Ca, Na, Mg
Table 4-1 Relative Phase Composition (wt.%) Derived from XRD Data
71
Small quantities of iron were found in each sample and these lead to significant colour
changes when exposed to air. Oxidation leads to grey ferrous iron altering to red ferric
oxide and yellow limonite (hydrated ferric oxide). An example of such colour
alterations and variations is provided on Figure 4-9. This iron presence, which is also
present in the groundwater in Perth (e.g. Appleyard, 2003), may give rise to chemical
bonding.
Figure 4-9 Colour changes due to Iron on Soil Samples
Scanning electron microscope (SEM) images of different (but representative) samples

of the clay-rich deposits of the Perth Formation showed iron stained aggregates of
kaolinite crystals combined with silt size grains of quartz; some of these SEM images
are shown in Figure 4-10. The SEM study confirmed the findings of the XRD analyses,
72
which are consistent with a material composition derived from weathering of granite in
a warm humid climate (Mitchell & Soga 2005).
Figure 4-10 SEM Images
The higher than expected activity of the samples was investigated in an additional XRD
analysis on the particles smaller than 2 m (i.e. the clay fraction). This analysis revealed
that kaolin was still the dominant clay mineral and only relatively minor traces of high
activity minerals such as smectites existed. This finding suggests that the high activity
(and plasticity) of the soils was due to the presence of particularly fine kaolinite
crystals, such as those described by Drummond et al. (2001). In support of this view,
Whittaker (1938) states that the plasticity of kaolinite is directly proportional to the
particles surface area and inversely related to particle size.
73
4.5 COMPRESSIBILITY
Three 50 mm diameter samples, with the properties listed in Table 4-2, were subjected
to 1-D compression in an oedometer to a maximum stress of ~ 3 MPa (which is
significantly greater than that used in standard practice in Perth).
Sample Depth LL PI FC CF
No. (m) (%) (%) (%) (%) Activity
1 13.7 48 29 70 22 1.32
2 12.4 62 36 90 28 1.29
3 10.9 51 27 82 28 0.96
Table 4-2 Properties of Soil Samples Tested in 1-D Compression
Because of the relatively high stiffness of the samples, it was essential that the
compliance of the apparatus was accounted for in the derivation of void ratio (e) -
vertical effective stress ( v) relationship. A dummy steel sample was therefore tested in
the same manner as the soil samples and the steel compression measured at a given
stress increment was deducted from that observed for the soil sample. The e vs. v data
deduced in this manner are plotted on Figure 4-11a, where v is plotted to a logarithmic
scale. Figure 4-11a also plots the 1-D compression line for reconstituted material with a
liquid limit of 50% (referred to as the intrinsic compression line, ICL), which was
estimated using the correlations of Burland (1990). To assist interpretation of the
vertical yield stress ( vy), (following the recommendations of Butterfield 1979), Figure
4-11b presents the same oedometer data as ln(1+e) vs. v.
The following observations are made:
(i) The relatively large range in initial sample void ratios at v=100 kPa is in
keeping with the range of water contents apparent on Figure 4-7.
(ii) Both Figure 4-11a and Figure 4-11b indicate that yield occurs at a vertical
effective stress of about 700 kPa and that the compression lines approach the
74
slope of the ICL. The greater post-yield compressibility of sample 2 is

consistent with its higher liquid limit
(iii) The in-situ overconsolidation ratios corresponding to a vertical yield stress

( vy) of ~700 kPa vary between about 3 and 4 (given the in-situ vertical
effective stress for the samples tested is in the range 175 kPa to 250
kPa).The relationship between K0 and OCR proposed by Mayne & Kulhawy
(1982) suggests that the in-situ K0 values lie between 0.8 and 1.1 i.e. are
approximately unity1.
0.65 0.50
0.48
0.60
0.46
0.55 0.44
0.50 0.42
0.40
0.45
0.38
Sample 1
0.40 Sample 2 0.36
Sample 1
Sample 3 Sample 2
0.34
0.35 Sample 3
Intrinsic Compression line 0.32 'vy
0.30
0.30
10 100 1000 10000
10 100 1000 10000
'v (kPa) ' v (kPa)
(a) (b)
Figure 4-11 Oedometer Data for Three Samples
1
Three filter paper tests were conducted following ASTM D5298-94. These indicated matric suction values (pk) of
255 kPa, 141 kPa and 122 kPa on samples from 8.9 m, 12.5 m and 13.9 m depth respectively. If no change in sample
mean effective stress occurred during sampling and extrusion, pk is equal to the in-situ mean effective stress
implying K0 values in the range 0.3 to 2.1 in the tested region. Such a range is highly unlikely and indicates that
either the filter paper technique is not suitable for the type of sample and/or materials recovered, or that soil suctions
were not maintained during the period between sampling and testing.
75
4.6 EFFECTIVE STRESS STRENGTH
4.6.1 Triaxial Tests
Initial suction measurements made on 60 mm diameter specimens trimmed from the

PQ-3 rotary cored samples were typically less than 100 kPa. The existence of low
suctions indicates that the sampling process led to a significant loss in the specimens
effective stress and therefore all triaxial tests involved re-consolidation of specimens to
effective stresses typical (or sometimes greater) than the estimated in-situ mean
effective stresses.
These specimens were first saturated under a backpressure of ~500 kPa and then
consolidated isotropically to an initial mean effective stress (p0) of between about
150kPa and 500kPa. Consolidation was isotropic as the estimated in-situ K0 value was
approximately unity. Some specimens were equipped with local strain instrumentation
to facilitate measurement of stiffness (discussed later). All specimens were then
subjected to undrained triaxial compression.
The stress paths measured in these CIU tests, which are plotted in deviator stress (q)
versus mean effective stress (p) space, are shown on Figure 4-12; this figure also labels
each path with the fines content of the respective specimens. It is apparent that samples
at lower effective stresses (and hence at higher OCR) and with lower fines content (i.e.
sandier samples) showed a greater tendency to dilate. Peak undrained strengths
developed at lower mobilised friction angles for samples with higher fines contents.
Axial strains to mobilise peak strengths varied between 2% and 6% after which a shear
plane developed in the samples.
A consolidated undrained test was also performed on a reconstituted sample to provide

information on the structure of the intact material. All of the laboratory test data are
compared with the in-situ test results and also with laboratory data recovered during the
site investigation for the nearby New MetroRail project
76
1000 o
41
Numbers denote fines content (%)
32o
800
o
600 22.5
Reconstituted at OCR=2.5
FC=77%
39
400
95
74
57
200
80
51
0
0 200 400 600 800
Mean effective stress, p' (kPa)
Figure 4-12 Stress Paths Measured in CIU Triaxial Compression Tests
The values of t (= [ 1- 3]/2) and s(= [ 1+ 3]/2) recorded at peak deviator stress (=
tf, sf) are plotted on Figure 4-13a and are combined on this figure with corresponding
data obtained from the final stage of multi-stage CIU tests on 63 mm diameter tube
samples of the Perth Formation recovered during investigations for the New
MetroRail City Project. Both datasets are seen to be in general agreement with a typical
mean ' of 32o and a c' value of zero. Although sampling disturbance may have led to
some destructuration of the material, the inference of a zero or near-zero c' value from
Figure 4-13a suggests that cementation/bonding in the in-situ material is not a
significant component of the strength on average.
77
600 50
40
400
30
o
'=32 20
200 (average)
Data from New MetroRail City project
10
St. Georges Tce
Reconstituted at OCR=2.5
0 0
0 200 400 600 800 1000 0 20 40 60 80 100
s' f (kPa) Fines content (%)
(a) (b)
Figure 4-13 (a) Values of t and s at Failure for all Triaxial Tests, (b) Dependence
of Triaxial Friction Angle on Fines Content
The potential presence of cementation/bonding was examined further by conducting a

CIU test on a reconstituted sample with a fines content of 77% and plasticity index of
17%; reconstituted material does not, by definition, possess a c' component of strength.
Air dried soil was mixed at a water content in excess of its liquid limit and, after de-
airing, consolidated one dimensionally to a vertical effective stress of 100 kPa in a
thick-walled cylinder A specimen was extruded from this cylinder and first consolidated
isotropically to a mean effective stress of 500 kPa before being allowed to swell to a
mean effective stress of 200 kPa (inducing an isotropic over consolidation ratio of 2.5)
prior to undrained shearing. The stress path observed during this undrained shearing
stage is provided on Figure 4-12 and the comparisons shown on Figure 4-13a &Figure
4-13b between ' values mobilised at peak strengths (assuming c' = 0) suggest that the '
angle for the reconstituted specimen is comparable to that of the intact soil with the
same fines content and at the same stress state. Although a small non-zero c' component
of ~10 kPa could also be reasonably interpreted from the data for intact samples on
Figure 4-13a, the evident similarity of the friction angles for the intact and reconstituted
78
soil suggests that cementation/bonding effects (due, for example, to the presence of iron
staining on fissures) in the in-situ material should not be relied upon in design
Closer examination of the triaxial data revealed that ' (derived assuming c' = 0 when t
= tf) showed no clear correlation with plasticity index but did reduce with the fines
content. The observed variation of ' with FC is shown on Figure 4-13b which reveals
that despite significant scatter, ' decreases from about 40o at FC = 20% to between
about 22o and 30 at FC = 80% (placing more reliance on the St. Georges Tce. triaxial
test data).
4.6.2 Simple Shear Tests
Undrained simple shear (SS) tests were conducted on nine specimens after 1-D
consolidation. The variations of shear stress ( xy) with vertical effective stress ( v)
measured during undrained shearing in SS tests are shown on Figure 4-14. Similar
trends to those seen for undrained triaxial tests on Figure 4-12 are evident with
specimens at lower initial stress levels and fines contents showing a greater dilatant
tendency. Specimens consolidated to v less than 200 kPa tend to dilate strongly (i.e.
v increases) while v reduces (i.e. a contractant response) during shearing of samples
with an initial v greater than 300 kPa.
79
500 38o
Numbers denote fines content (%)
400 33o
29
300 40
200
40
100
23
96 37 69 82
80 52
0
0 200 400 600 800
Vertical effective stress, 'v (kPa)
Figure 4-14 Stress Paths Measured in Undrained Simple Shear
As indicated on Figure 4-15, the peak friction angles (defined using the tangent
definition, 'ss= tan-1( xy/ v)) are generally developed at peak shear stress and vary from
between ~29o and 40o. These relatively high angles of friction and those measured in
triaxial compression are consistent with materials that do not contain significant
quantities of smectites/montmorillonites within the clay fraction.
4.6.3 Undrained Shear Strength
It is evident from Figure 4-12 and Figure 4-14 that the undrained strength of samples
varies both with the consolidation stress (pi in CIU tests and vi is SS tests) and the
fines content. For a typical non-fissured clay, the undrained strength ratio in CIU tests
(sutc/p0) and SS tests (suss/ vi) varies with the overconsolidation ratio (OCR) as:
sutc/p0 ~ A OCR0.8 or sutc/ vi ~ A OCR0.8 (3)
80
where A is typically about 0.25 for both SS and CIU tests (a higher value of about 0.3 is
typically measured in anisotropically consolidated triaxial compression tests; see
Leroueil & Hight 2003).
An A value of 0.25 matches the sutc value of the reconstituted specimen, which has an
OCR of 2.5 and a relatively high fines content (FC = 77%). The undrained strength
ratios observed in CIU and SS tests are plotted against the consolidation pressure on
Figure 4-15a, which also plots equation (3) for A values of 0.25 and 0.6 assuming an
isotropic preconsolidation pressure of 700 kPa (estimated from the oedometer data). As
may be expected, the curve corresponding to an A value of 0.25 forms a lower bound to
all undrained strength data. The same undrained strength ratio data are plotted against
fines content on Figure 4-15b, which illustrates the additional strength that may be
mobilised by materials with lower fines contents (because of their more dilative nature).
It is apparent from Figure 4-15a that a value of A in equation (3) of 0.6 forms a near
upper bound to the strengths mobilised by these materials.
3 3
CIU tests CIU tests
2.5 2.5 Undrained SS
Undrained SS
Reconstituted at
2 2 OCR=2.5
Upperbound, A=0.6
1.5 1.5
1 1
0.5 0.5
Lowerbound, A=0.25
0 0
0 200 400 600 0 20 40 60 80 100
p'0 in CIU tests or 'vi in SS tests, kPa Fines content (%)
(a) (b)
Figure 4-15 Undrained Strength Ratios Plotted as a Function of (a) consolidation

stress level and (b) Fines Content
A reduction in the value of A ( undrained strength ratio of the material at OCR = 1)

with an increase in fines content (FC) is consistent with the observed dependence of the
friction angle on FC (Figure 4-13b) and with experimental data reported in the literature
81
for silts (e.g. Brandon et al. 2006, Lehane & Faulkner 1998). The values of A (as
defined in equation 3) indicated by all samples assuming a uniform preconsolidation
stress of 700 kPa are plotted against FC on Figure 4-16. As expected, A reduces (albeit
with appreciable scatter) as FC increases. At the average in-situ FC of 60%, A = 0.4
0.15, which implies that the mean undrained strength (in simple shear) of the in-situ
material between 8 m and 17 m is approximately 225 kPa. It is noteworthy that this
strength is significantly lower than the undrained strength of 400 150 kPa inferred
from self-boring pressuremeter tests (SBPTs) at the same site by Fahey (1988).
1.2
0.8 CIU tests
s u/p'0= su/ 'vi=A OCR
1 Undrained SS
0.8 Reconstituted at
OCR=2.5
A 0.6
0.4
Mean trend
0.2
0
0 20 40 60 80 100
Fines content (%)
Figure 4-16 Variation of A (undrained strength ratio at OCR=1) with Fines

Content
The mean CPT qc between 8m and 18m depth for the data plotted on Figure 4-5 is
4MPa, excluding qc values in excess of 5.5MPa for which it is presumed that partial or
full drainage during cone penetration has taken place. This mean qc value combined
with the average estimated undrained strength of 225kPa implies a cone factor, Nk, of
16.8. This Nk value factor falls within the range of 14 to 18 quoted by Lunne et al.
(1997) for a wide range of overconsolidated clays and indicates that the higher shear
strengths inferred from SBPTs require further investigation. For horizons with lower
82
fines content values, undrained strengths may be expected to be larger (e.g. see Figure
4-16) but the degree of drainage taking place during cone installation is also expected to
be larger; use of an Nk value to infer an undrained strength may not be appropriate for
these lower fines content materials.
4.7 STIFFNESS
The drained secant Youngs moduli (Esec) inferred from CIU triaxial compression data
of 6 specimens, including one reconstituted specimen are plotted in normalized form
against axial strain on Figure 4-17; these drained E values were converted from the
measured (undrained) Eu values assuming a mean Poissons ratio, = 0.3). As Youngs
moduli at low strains generally vary with the square root of the vertical stress in the
direction of the principal strain, the Esec values are normalized by the square root of the
vertical consolidation stress ( vc) divided by atmospheric pressure (pa). Measurements
plotted at strains less than 0.01% were obtained using local displacement transducers
(LDTs, manufactured at UWA) attached to the specimen membrane.
Youngs moduli at very small strains (E0) may be calculated from shear wave velocity
(Vs) data assuming a small strain Poissons ratio of 0.1 and mean density ( ) of 1900
kg/m3 (E0 = 2.2 Vs2). Vs values measured in Seismic Cone Penetration Tests (SCPTs) at
the site varied from ~300 m/s to 400 m/s between 8 m and 18 m depth, giving the range
shown on Figure 4-17 for E0/( v/pa)0.5 of between 300 MPa and 450 MPa. E0 values are
usually assumed to be generally applicable at strains less than 0.001% at which strain
the laboratory measurements suggest an E0/( v/pa)0.5 range of between 200 MPa to 300
MPa. These somewhat lower laboratory stiffness values may reflect disturbance to the
structure of the in-situ soil by the sampling process.
83
.
FC=57%
400
Range measured FC=51%
in SCPTs FC=95%
300 FC=39%
FC=20%
200
100
Reconstituted at OCR=2.5, FC=77%
0
0.0001 0.001 0.01 0.1 1
Axial strain (%)
Figure 4-17 Stiffness Data Measured in CIU Triaxial Compression Tests
Further points of note include:
1. Specimens with a fines content less than 40% are almost twice as stiff as those
with FC > 50%.
2. The stiffness degradation curves for the reconstituted soil and specimens with a
comparably (high) fines content effects are similar, indicating that the effects of
structure on stiffness are insignificant at strains greater than about 0.005%.
3. The upper bound of the normalized E0 values inferred from SCPTs are
comparable in magnitude to those measured (at shallower) depths in the
Spearwood dune sand.
4. The stiffness measured at an axial strain of 0.01% is only about 50% of the
estimated very small strain stiffness (E0)
5. A review of the literature indicates that the lower bound to the stiffness values
plotted on Figure 4-17 is similar to normalized stiffness values indicated by
84
(high plasticity) London Clay, while the upper bound to the plotted values are
comparable to the stiffness of sands such as the Spearwood dune sand.
Correlations proposed by Lehane & Fahey (2004) between the CPT end resistance (qc)
and stiffness data inferred from Vs data and DMT results in sand are of the following
form:
E = [ qc 0.25 v0.5 pa0.25] (4)
where v is the in-situ vertical effective stress, pa is a reference stress (=100 kPa) and
the constant varies with the nature and age of the deposit, in addition to the strain
level 1. Values of obtained assuming that this relationship is also applicable to the silts
and clays considered here are summarised in Table 4-3. These values were derived by
assuming that the upper bound to the laboratory stiffness data on Figure 4-17
correspond with siltier deposits at the St. Georges Terrace site (with a typical qc value of
5 MPa) and the lower bound to the laboratory data relate to the clay rich horizons (for
which qc ~3 MPa).
Axial strain Value of (Equation 4)

Very small strain 1100 to 1500
0.01% 430 to 530
0.1% 210 to 270
1% 55 to 70
Table 4-3 Approximate Values of for Clays at the St. Georges Terrace Site
The use of Equation (4) combined with the values in Table 4-3 provides an
approximate means of estimating the elemental E value for these deposits. Direct
1
It is noteworthy that Equation (4) predicts a weak dependence of stiffness on qc (contrary to common
practice which assumes E varies in direct proportion with qc)
85
measurement of soil stiffness is, however, clearly more desirable particularly given
the dearth of reliable stiffness data.
4.8 CONCLUSIONS
This chapter describes the available soil properties of Spearwood sand and results of a
comprehensive laboratory investigation, supplemented by in-situ test data, of the upper
horizons of the Perth Formation in Perth CBD. The main conclusions are as follows:
4.8.1 Spearwood sand
1 The Spearwood Sand is a uniformly graded, medium quartz dune sand.

Colour changes (varying from yellow to light brown) are associated with
variations in iron oxide or other iron variants. Observed mechanical
properties show no dependence on colour (and hence iron compounds).
2 The Spearwood sand is generally medium dense, with some loose deposit
close to the surface. The peak friction angle estimated from qc varies from
36o for the loose deposits to 40o for the medium dense material (Schneider et
al., 2008). Self boring pressuremeter tests suggests plane strain friction
angles of as high as 50o (Fahey et a., 2007).
3 The normalised small strain stiffness value of the Perth Formation sand and
the Spearwood dune sand are comparable. Field tests indicate an unload-
reload shear moduli of about 40 to 50% of small strain shear modulus (Fahey
et al., 2003, Fahey et al., 2007).
4.8.2 Perth Formation
1 Standard in-situ test correlations applied to the upper horizons of the Perth
Formation indicate that material varies between being a clayey silt and silty
clay, with occasional sand seams/lenses. The inference of these soil types is
consistent with grading analyses of rotary cored samples.
86
2 Kaolin and quartz are the dominant minerals present in the soil and their
presence supports the view that the deposit was formed by weathering of
granite on the Darling Scarp and subsequent deposition downstream in a
fluvial/riverine environment in Perth CBD.
3 The soils high activity (i.e. relatively high plasticity index for a given clay
fraction) appears to arise because of the presence of very small kaolinite
crystals and not because of the presence of smectites/montmorillonites.
4 Samples compressed one dimensionally to high stresses in an oedometer

indicate yield at a vertical effective stress (or preconsolidation pressure) of
~700 kPa. In-situ overconsolidation ratios (OCR) of between 3 and 4 are
inferred using this stress. The correlation of Mayne & Kulhawy (1982)
predicts that the in-situ K0 value is approximately unity for this OCR range.
5 The materials behaviour in undrained triaxial compression and simple shear

depends largely on the consolidation pressure (defining the
overconsolidation ratio) and the fines content (with siltier materials showing
more dilative behaviour). The estimated mean undrained shear strength (in
simple shear) for the in-situ material (with a mean FC value of 60%) of 225
kPa is consistent with a cone factor, Nk, of about 17.
6 Strength tests did not reveal the presence of significant

cementation/structure in rotary cored samples and best estimate effective
stress strength parameters of ' of 35o to 40o at FC = 20% and ' = 25o at FC
= 90%, with c` = 0 were inferred.
7 Triaxial samples equipped with local strain instrumentation confirm the

strongly non-linear nature of the materials stiffness showing, for example,
that the stiffness at an axial strain of 0.1% is only about 20% of the small
strain stiffness inferred from shear wave velocity measurements. A simple
(approximate) means of assessing the in-situ stiffness is provided, which
was derived by combining in-situ test data with trends indicated by the
laboratory element stiffness data.
87
88
CHAPTER 5 SURFACE SETTLEMENT -
GREENFIELD SITE
5.1 INTRODUCTION
Tunnelling induced ground (greenfield) movement is discussed in this Chapter. Part of

the tunnel went underneath five existing railway tracks north of William Street Station.
Monitoring of these railway tracks was undertaken to detect the onset of any
movements that may cause derailment during the two tunnel drives. As there were no
buildings around this area, this was a good location to estimate the greenfield
movement. All the five tracks were instrumented using Rail Settlement Pins (RSP)
while Track1, Track 3 and Track 5 were also instrumented with Electro Level beams
(EL beams). Measurements are compared with movements predicted using the standard
Empirical method (Gaussian method) to study the nature of the observed settlement.
The factors influencing the observed patterns of surface movement are also evaluated.
5.2 GROUND CONDITION
The detailed geology and the hydrogeology of the tunnel route is discussed in Section
3.4. In the northern bored tunnel section (north of William Street Station), the tunnels
passed through the Perth Formation near the William Street Station and then gradually
entered the Spearwood sand. Figure 5-1 shows the geotechnical investigation locations
relative to the railway tracks and the instrumentation in the vicinity of the railway
tracks. Soil investigations were carried out north and south of the railway tracks.
However, due to accessibility issues, no soil investigation was carried out close to the
tracks.
The ground profile at this site was estimated from the boreholes, cone penetration tests
(CPT) and dilatometer tests (DMT) carried out close to the north of Track 5 and the
south of Track 1; see Figure 5-1. The distance between the two soil investigations
locations was about 120 m and the distance between Track 1 and Track 5 was about
89
46m. Since there were variations in the soil profile north and south of the railway tracks,
two soil profiles were assessed based on the soil investigations carried out north and
south of the railway tracks. Each of these profiles is discussed below.
Figure 5-1 Geotechnical Investigation Locations near the Railway Tracks
90
Chapter 5 Surface Settlement - Greenfield Site
5.2.1 Ground Condition South of Track 1
Boreholes BH-2149, BH-03, BH-22, BH-19 and CPT-11 were located to the south of
Track 1. Their locations were about 30m and 40m south of Track 1 for Tunnel 1 and
Tunnel 2 respectively. Figure 5-2 shows the borehole information and stratigraphical
profiling based on the Ic index derived from the CPT data. (Details are given in Section
4.4.1) The CPT end resistance (qc) and friction ratio (Fr) for CPT-11 are plotted in
Figure 5-3 and these data combined with the boreholes were used to derive the profile
shown on Figure 5-4 for ground conditions south of Track 1 (no dilatometer tests were
carried out close to Track 1). The best estimate average qc value for each layer is also
shown in Figure 5-4. It is apparent that stratigraphy comprises 7m of dune sand
(Spearwood sand) overlying the interbedded layers of alluvial silts, clays and sands of
Perth Formation, which are underlain by Kings Park shale/ siltstone. Figure 5-4
suggests that the tunnels are located in predominantly sandy layers of the Perth
Formation. The water table is at about 9.2 m AHD (Australisn Height Datum or Mean
Sea Level), which is ~1.8m below the existing ground surface. The ground water
pressure is estimated to be about 92kPa at the centre of the tunnel.
91
RL(m)
Figure 5-2 Borehole Information and Stratigraphical Profile South of the Railway
Tracks
92
CPT-11, qc (MPa) CPT-11, Fr (%)
0 10 20 30 40 50 60 0 2 4 6 8 10 12
15 15
10 10
5 5
0 0
-5
-5
-10
-10
-15
-15
-20
-20
-25
-25
Figure 5-3 CPT Profile South of Track 1
+11.0
Tunnel 2 Tunnel 1 Spearwood
+9.2
14.4m Sand
qc = 10 MPa ~11m Sand
+4.0 qc = 2.5MPa Clay/Silty clay
+2.0
Sand/Silty
qc = 45 MPa 6.9 6.9m Sand
-5.0
qc = 4 MPa Clayey Silt to Perth
Silty Clay Formation
-13.0
qc = 1.5 MPa Silty Clay to
Clay
-20 0
Sand
-24.0
Figure 5-4 Idealised Soil Profile South of Railway Tracks
93
5.2.2 Ground Condition North of Track 5
Soil investigations north of Track 5 consisted of 2 boreholes (BH-2169 and BH-02), one
cone penetration test (CPT-10) and one dilatometer test (DMT 01). These were
performed about 40 10m north of Track 5. Figure 5-5 shows the borehole information
and stratigraphical profiling based on the Ic index derived from the CPT and the ID
index from the DMT. Figure 5-6 shows the CPT profile north of Track 5 while Figure
5-7 shows the idealised soil profile based on all available data; this figure also shows
the mean qc value for each soil layer. It is apparent from the soil profile shown on
Figure 5-7 that sand is the predominant soil type from ground level to about 25m below
the ground level and the tunnels are located in the sand. The water table is at about 9.2m
AHD, which is ~1.8m below the existing ground surface.
94
RL(m)
RL(m)
Figure 5-5 Bore Hole Information and Stratigraphical Profiling North of the
Railway Tracks
95
CPT-10, qc (MPa) CPT-10, Fr (%)
0 10 20 30 40 50 60 0 2 4 6 8 10 12
10 10
5 5
0 0
-5 -5
-10 -10
-15 -15
-20 -20
-25 -25
Figure 5-6 CPT Profile North of Track 5
Both North and South profiles were considered separately for numerical analyses
presented later.
+11.0
+9.2
14.4m
Tunnel2
~11m Tunnel 1
Spearwood
qc =15 MPa Sand
Sand
-2.0
qc = 9 MPa 6.9m 6.9m Silty Sand
-6.0
qc = 15 MPa Clean Sand to Silty Sand
-12.0
qc = 7 MPa Silty Sand to Sandy Silt Perth
-14.0
qc = 5 MPa Clayey Silt to Formation
-19.0 Silty Clay
Sand
-24.0
Figure 5-7 Idealised Soil Profile North of Railway Tracks
96
5.3 TUNNELLING
An earth pressure balanced tunnel boring machine (EPB-TBM) was used to bore the
two tunnels. Details of the TBM and boring technique and schedule of tunnelling are
described in Section 3.7. The railway tracks were between the chainages of about 130m
and 200m, north of William Street Station. The relevant section of Tunnel 1 was bored
between 17th May 2006 and 22nd May 2006 while that for Tunnel 2 was bored between
28th September 2006 and 2nd October 2006.
5.4 INSTRUMENTATION
Since the movements of the existing railway tracks due to tunnelling were of major
concern, all of the five tracks were instrumented and monitored for movement. Each of
the five tracks was instrumented using Rail Settlement Points (RSP) and three (Track1,
Track 3 and Track 5) were also instrumented with automated Electro Level beams (EL
beams). The installation procedures and the measuring techniques for each of these
instruments are described below.
5.4.1 Rail Settlement Points
Rail Settlement Points (RSPs) were survey points on the rails that enabled monitoring of
the movements of the rails during TBM tunnelling. Because of the accessibility
problem, these survey points took the form of retro reflective targets. Surveying was
carried out using Leica Total Stations. Repeatability using this method was in the range
of 1.5mm (McGough & Williams, 2007).
During processing of the instrumentation data at the railway track area, drift in the RSP
data was noticed over a period of time. Figure 5-8 shows the apparent erratic movement
of the RSPs over a period of 2 months at Track 3 and Track 5 prior to Tunnel 1 boring.
At Track 3, the data varied from -2mm to +4mm, while at Track 5, the corresponding
values were -2 and +7mm. Hence the base line reading for monitoring movements due
to Tunnel 1 and Tunnel 2 was selected just before the start of respective tunnel boring.
97
Track 3 Track 5
8 400 8 400
6 Tunnel 1 6 Tunnel 1
300 300
4 4
2 200 2 200
0 0
100 100
-2 -2
-4 0 -4 0
13/1/06 4/3/06 23/4/06 12/6/06 13/1/06 4/3/06 23/4/06 12/6/06
Date Date
Figure 5-8 Variations in the SSP Data over a Period of Two Months for Track3
and Track 5
5.4.2 Electro Level Beams (EL Beams)
Electro Level beams are used to monitor rotation and differential movement in
structures. While horizontal beams are used to monitor settlement and heave, vertical
beam sensors monitor lateral displacements. Figure 5-9 shows the typical view of an EL
beam. The beam is one to two meters long, and is mounted on anchor bolts that are set
into the structure. The horizontal beam consists of a built-in compartment and cover for
the tilt sensor. Each beam has end brackets which can be secured directly to anchors.
Movement of the structure is measured by the tilt of the beam.
By linking end to end, the beam sensors can monitor differential movement; total
movements are established if the movement of one of the sensors (typically one at the
end of the string) is known. Figure 5-10 illustrates how EL beams measures the
differential settlement. Even with train traffic vibrations and changing weather
conditions, the repeatability of EL beams was in the range of 1.0mm (McGough &
Williams, 2007).
98
Figure 5-9 Typical View of a Horizontal EL Beam
Figure 5-10 Schematic View of Horizontal EL Beam
The electrolytic tilt sensor, at the core of sensor, is a precision bubble-level that is
sensed electrically as a resistance bridge. The bridge circuit outputs a voltage
proportional to the tilt of the sensor, which is then converted to a gradient i.e. reading in
mm per meter.
Three of the five tracks were instrumented with automated EL beams. Figure 5-1 shows
the location of EL beams on the rail line. Both longitudinal as well as transverse EL
beams were used and they were installed on the sleepers between the two rails.
Transverse EL beams were installed to monitor any differential settlement between the
two rails, which was crucial for the smooth running of the trains. The sensors were then
99
cabled to a data logger (CRIDX) adjacent to EL beams and from there results were
radio transmitted to the office and uploaded to a computer.
After the first tunnel boring, to reduce costs, the first few EL beams were relocated to
the end of the beam to measure the movement due to Tunnel 2 boring. Figure 5-11 and
Figure 5-12 show the arrangement of longitudinal as well as transverse EL beams on the
track. The data were collected at 15 min intervals using an automatic logger. The data
recorded at 4.00 AM were selected in the analyses discussed later, as these did not
include any movement component associated with train traffic.
From Figure 5-1, it is clear that, the railway tracks were not orthogonal to the tunnels
and thus to the longitudinal EL beams. This is shown schematically in Figure 5-13.
Hence settlement troughs derived by plotting data measured along the railway tracks are
not truly transverse to the tunnel axis. Therefore to derive the transverse settlement
trough, the distance along the railway track was converted into an equivalent transverse
distance. The procedure adopted is illustrated on Figure 5-13. Consider points A, B and
C marked in the figure. EL beams are aligned along AC and the transverse distance
(orthogonal distance) is BC ( BC AC 2 AB 2 ). It may be assumed that there is little

difference between the soil properties at A and B and hence the final settlement at B
was taken equal to that at A.
100
Longitudinal
EL beams
Transverse
EL Beams
Figure 5-11 Longitudinal and Transverse EL Beams on the Track
Longitudinal
EL beams
Transverse
EL Beams
Figure 5-12 Closer View of EL Beams on the Track
101
Figure 5-13 Schematic View of EL Beam on Railway Tracks
5.5 FIELD MEASUREMENT
The transverse surface settlement troughs were plotted from RSP and EL beam data to
examine the response of the ground before, during and after the TBM had passed
through a particular location. Although all the five tracks were instrumented with RSPs,
the transverse settlement troughs were plotted only for three tracks (Track 1, Track 3
and Track 5) where EL beams were also installed. This was to enable comparison
between the two different types of instrumentation. As there was no long term
settlement expected (see section 5.5.1), the settlement of two tunnel borings were
estimated separately. There fore, the settlement trough discussed in this section is either
due to Tunnel 1 only or Tunnel 2 only. The combined movements (assuming no
significant settlements took place after Tunnel 1 finished and before Tunnel 2 started)
are discussed in Section 5.5.3.)
The settlement troughs plotted from RSP and EL beam data show the following details:
102
Centre line of Tunnel 1 and Tunnel 2;
Maximum heave observed at that location before, during or after the TBM has
passed;
Date and time of each settlement trough;
Either the distance of the cutter face from the corresponding track at a given date
and time if the tunnelling hadnt finished (Negative distance indicates that the
cutter face was behind the corresponding track and the positive sign indicates
that the cutter face was ahead of the corresponding track.) or time (in days)
passed after boring at that location if the tunnelling had finished; and
Date at which the tunnel bored at that location.
The observations from the transverse settlement troughs obtained for the tunnels both
from RSP and EL beam are summarised in the following sections.
5.5.1 Rail Settlement Points (RSPs)
The transverse settlement troughs were plotted from RSP data to study the behaviour of
the settlement pattern during and after the two tunnels boring. The transverse settlement
troughs for Tunnel 1 and Tunnel 2 were plotted separately and the observations are
discussed below.
Tunnel 1
The behaviour of transverse settlement troughs plotted from RSPs for three tracks are
given in Figure 5-14. Observations from Tunnel 1 RSP data include:
Although all the three tracks showed heave during tunnelling, the heave
decreased such that, in all cases, net settlements were observed within a day or
two.
The maximum heave observed was for Track 5 (9.5 mm) and minimum for
Track 1 (3mm). Track 3 had a heave of about 6 mm
103
The maximum settlement was observed for Track 3 (8.9mm) and minimum for
Track 1 (4.5mm). Track 5 had a settlement of 6.2mm
All the tracks showed maximum settlement at the centreline of the tunnel
The settlement stabilised within a day or two.
Track 1, Tunnel 1 Track 3, Tunnel 1

10
10 17/5/06 16:00 (2.1m)
Tunnel 1
18/5/06 14:00 (14.1m) 8
8 20/5/06 14:30 (41.1m)
18/5/06 10:20-
6 19/5/06 5:35
6 Tunnel 2 21/5/06 11:30 (188.6m)
4 21/6/06 13:00 (34 days) 4
2 2
0 0
-2 -2 Tunnel 2
-4 -4 18/5/06 21:00, (3.7m)
-6 Tunnel 1 -6 19/5/06 6:00, (7.7m)

20/5/06 12:00, (24.7m)
-8 17/5/06 8:40- -8 20/5/06 17:00, (27.7m)
21/6/06 13:00, (33days)
-10 18/5/06 3:25 -10 27/9/06 15:00, (100days)
-30 -20 -10 0 10 20 30 -30 -20 -10 0 10 20 30

Distance from centre of tunnels, m Distance from centre of tunnels, m
Track 5, Tunnel 1
10
8 18/05/06 14:00 (-25.4)
20/05/06 18:00 (2.6m)
6 22/5/06 2:00 (20.6m)
23/5/06 8:00 (32.6m) Tunnel 1
4 23/5/06 18:00 (37.6m)
21/6/06 13:00 (31days)
2
0
-2
-4
-6
Tunnel 2 20/5/06 6:40-
-8 21/5/06 4:05
-10
-30 -20 -10 0 10 20 30
Distance from centre of tunnels, m
Figure 5-14 Transverse Settlement Trough due to Tunnel 1 Boring (RSP)
104
Tunnel 2
Since the settlement stabilised after few days of Tunnel 1 boring, it is unlikely that
significant long term settlements occurred. Although no pore pressure measurements are
available, the predominance of sandy soils at this location is such that long term
consolidation settlements are likely to be small. Hence to avoid any instrumentation
error (See Section 5.4.1 to examine drift of RSP measurements over time), the vertical
displacements for Tunnel 2 was calculated separately taking a base line reading closer
to the start of the Tunnel 2 boring. Figure 5-15 shows the transverse settlement troughs
observed from RSPs for Tunnel 2 boring. Observations from Tunnel 2 RSP data
include:
Both heave and settlement values for all three tracks were less than during
Tunnel 1 boring.
The maximum heave observed was 1.5 mm, 1.8mm and 8.4 mm for Track 1,
Track 3 and Track 5 respectively and the respective maximum settlement values
were 2.2 mm, 6.0mm and 3.2 mm.
Maximum heave was observed for Track 5 while maximum settlement was
observed for Track 3.
Track 1 and Track 3 experienced maximum settlement along the centreline of

the tunnel and Track 5 showed maximum settlement away from the tunnel
centre. This could possibly be because, as mentioned in Section 3.7, that TBM
encountered a loosely filled abandoned well in this area.
105
Track 1, Tunnel 2 Track3, Tunnel 2
10 28/9/06 14:00 (-61m) 10 29/9/06 1:30 (-15.4m)
1/10/06 4:00 (16.10m)
8 Tunnel 2
2/10/06 8:00 (54.9m) 8 6/10/06 10:30 (102.60m)
5/10/06 10:00 (105.9) Tunnel 2
6 9/10/06 15:00 (120.9m) 6 10/10/06 9:00 (102.60m)
20/10/06 15:00 (108.6m)
4 23/10/06 12 PM (170.9m) 4 31/10/06 8:00 (31 days)
31/10/06 8:00 (32 days) 16/11/06 15:00 (47 days)
2 2
0 0
-2 -2
Tunnel 1 Tunnel 1
-4 -4
-6 -6 29/9/06 19:40-
28/9/06 20:11-
-8 -8 30/09/06 18:15
29/09/06 10:35
-10 -10
-30 -20 -10 0 10 20 30 -30 -20 -10 0 10 20 30
Track 5, Tunnel 2
10 1/10/06 10:00 (-8.4m)
2/10/06 8:00 (8.6m)
8 4/10/06 8:00 (40.6m)
6 1/10/06 15:35 -
5/10/06 0:00 (59.6m)
2/10/06 08:20 6/10/06 10:30 (74.6m)
4
10/10/06 9:00 (74.6m)
2
0
-2
-4
Tunnel 1
-6
-8 Tunnel 2
-10
-30 -20 -10 0 10 20 30
Figure 5-15 Transverse Settlement Trough due to Tunnel 2 Boring (RSP)
Comparison of RSP data for Tunnel 1 and Tunnel 2
A summary of maximum soil surface heave generated during tunnelling and of the
maximum settlement recorded after tunnelling was completed, as indicated by RSPs, is
provided in Table 5-1. One of the (perhaps surprising) features of the data was the
amount of short term heave generated during tunnelling. Such heave is presumed to
have arisen due to the application of relatively large pressures (typically 195 15kPa) at
106
the tunnel face or the grout pressure (18040kPa) at the tail of the TBM. Although all of
the three monitored tracks experienced heave during tunnelling, incremental movements
after the TBM cutter face had passed any given location were always downwards and
resulting final ground movements were also downwards (i.e. settlements), varying
between 2mm and 9mm. Even though the maximum heave and settlement observed was
different for all the tracks during Tunnel 1 and Tunnel 2 boring, some common trends
were observed, namely,
Track 5 showed maximum heave and Track 1 showed minimum heave.
Track 3 experienced maximum settlement while Track 1 experienced the least

settlement.
There appears to be a trend for greater settlement to take place when preceded
by greater heave (see Figure 5-16), suggesting that the benefits of settlement
reduction caused by application of high face pressures are only short term.
Maximum settlement was observed along the axis of the tunnel for all the tracks
except for Track 5 during boring of Tunnel 2.
Tunnel 1 Tunnel 2
Max. heave Max. settlement Max. heave Max. settlement

(mm) (mm) (mm) (mm)
Track 1 3 4.5 1.5 2.2
Track 3 6 8.9 1.8 6.0
Track 5 9.5 6.2 8.4 3.2
Table 5-1 Variation of Heave and Settlement for each Track after the Tunnel
Boring
107
10
9
8
7
6
5
4
3
2
Tunnel1
1 Tunnel2
0
0 2 4 6 8 10
Maximum heave, mm
Figure 5-16 Relationship between Maximum Heave and Maximum Settlement

Deduced from RSPs
5.5.2 EL Beams
Since one end of the string of EL beams was fixed, the plotted transverse settlement
troughs show the relative movement with respect to this fixed end. To obtain the
absolute values of the settlement, the settlement of the fixed end inferred from RSP
readings should be added to the relative movements. Since the RSP readings were not
available for the same dates and times of the EL beam readings, relative settlement
troughs are presented initially inFigure 5-17 to illustrate settlement patterns for each
track during Tunnel 1 and Tunnel 2 boring. Absolute final settlement troughs for each
track for Tunnel 1 and Tunnel 2 are plotted later in this Chapter and compared with
those indicated by the RSPs.
Figure 5-17 and Figure 5-18 show the (relative) transverse settlement trough at various
times during boring of Tunnel 1 and Tunnel 2 respectively. Observations from Tunnel 1
and Tunnel 2 EL beam data include:
The settlement appears to have stabilised in few days time suggesting that long
term consolidation/creep effects were insignificant.
Maximum settlement was observed above the tunnel axis for all the tracks
except for Track 5 during Tunnel 2 boring.
108
The settlement at about 18m offset from the centre line of Track 3 and Track 5
could be due to the instrumentation error as this was not detected by the RSP
data.
Track 5 data during tunnel 2 boring are not typical (e.g. maximum settlements
appeared to have occurred 18m from the tunnel centreline) and it is not clear if
the anomalies seen on Figure 5-18 arose due to instrumentation errors or due to a
localised zone of very loose sand. RSP data (which are only available to a
distance of 18m from the tunnel centreline) suggest that the presence of such a
loose zone is unlikely.
Track 1, Tunnel 1 17/5/06 (-3.9m) Track 3, Tunnel 1

8 19/5/06 (21.1m) 16/5/06 (-29.3m)
21/5/06 (47.1m) 8 19/5/06 (6.7m)
6 22/5/06 (62.1m) 6 Tunnel 2 23/5/06 (56.7m)
Tunnel 2 23/5/06 (71.1m) 24/5/06 (68.7m)
4 24/5/06 (83.1m) 4 25/5/06 (83.7m)
2 25/5/06 (98.1m) 2
0 0
-2 -2
-4 Tunnel 1 -4
-6
-6 17/5/06 8:40- 18/5/06 10:20-
-8 Tunnel 1
-8 18/5/06 3:25 19/5/06 5:35
-10
-10
-30 -20 -10 0 10 20 30
-30 -20 -10 0 10 20 30
Track 5, Tunnel 1
8
6 Tunnel 1
Tunnel 2
4
2
0
-2
20/5/06 (-4.4m)
-4 21/5/06 (7.6m) 20/5/06 6:40-
22/5/06 (22.6m)
-6 23/5/06 (31.6m) 21/5/06 4:05
24/5/06 (43.6m)
-8 25/5/06 (58.6m)
-10
-30 -10 10 30
Figure 5-17 Transverse Settlement Trough due to Tunnel 1 Boring (EL Beam)
109
8
Tunnel 1 8
6 Tunnel 2
6 Tunnel 2
4 Tunnel 1
4
2
2
0
0
-2 29/9/06 (4.9m) -2 29/9/06 (-13.4m)
-4 28/9/06 20:11- 2/10/06 (51.9m)
5/10/06 (101.9m) 1/10/06 (16.6m)
29/9/06 10:35 9/10/06 (120.9m) -4 5/10/06 (83.6m)
-6 10/10/06 (120.9m)
29/9/06 19:40- 6/10/06 (102.6m)
11/10/06 (120 9m) -6 30/9/06 18:15 10/10/06 (102.6m)
-8 11/10/06 (102.6m)
-8
-30 -10 10 30
-30 -10 10 30
Track 5, Tunnel 2
8 28/9/06(-59.4m)
1/10/06 (-11.4m)
6 1/10/06 15:35- 2/10/06 (5.6m)
4/10/06 (39.6m)
4 2/10/06 8:20 6/10/06 (74.6m)
10/10/06 (74.6m)
2 11/10/06 (74.6m)
0
-2
-4 Tunnel 2 Tunnel 1
-6
-8
-30 -10 10 30
Figure 5-18 Transverse Settlement Trough due to Tunnel 2 Boring (ELBeam)
5.5.3 Final Transverse Settlement Troughs
The transverse settlement troughs indicated by the RSP and EL beams are compared to
assess consistency between the two sets of measurements The EL beam data were
corrected from the values given above to absolute vertical displacements using RSP data
recorded close to the fixed end of the EL beam; these corrections were relative minor,
typically representing 0-15% of the maximum settlement. As seen on Figure 5-19 and
Figure 5-20, discussed below, the final movement (settlement/heave) indicated by the
110
RSPs and EL beams were in close agreement for both tunnel runs. Therefore, an
average of the two measurement sets was employed to derive final transverse settlement
troughs at each location.
The averaged final transverse surface settlement profiles for each track are given in
Figure 5-19 and Figure 5-20 for Tunnel 1 and 2 respectively. Only RSP data were
considered for Track 5 for Tunnel 2, as the corresponding EL beam data has been
shown to be erroneous. The maximum settlement, calculated volume losses and trough
widths for Track 1, Track 3 and Track 5 are given in Table 5-2. These characteristics of
the settlement troughs are discussed in the following paragraphs.
Settlement of Track 1, Tunnel 1 Settlement of Track 3, Tunnel 1

1 1
0 0
-1 -1
-2 -2
-3 -3
Tunnel 1
-4 RSP -4
-5 ELB -5
average -6
-6 Tunnel 2
RSP
-7 -7 ELB
Tunnel 1
-8 -8 Tunnel 2 Average
-9 -9
-20 -10 0 10 20 30 -10 0 10 20 30
Settlement ofTrack 5, Tunnel 1

1
0
-1
-2
RSP
-3
ELB
-4
Average
-5
Tunnel 1
-6 Tunnel 2
-7
-8
-9
-10 0 10 20 30
Figure 5-19 Final Transverse Settlement Trough for Three Tracks due to Tunnel 1
111
Settlement of Track 1 ,Tunnel 2 Settlement of Track 3, Tunnel 2
0 0
-1 -1
-2 -2
-3 -3
-4 Tunnel 2 -4
Tunnel 1 RSP
-5 -5 ELB
-6 -6 Average
RSP
-7 ELB -7
-8 Average Tunnel 1
-8 Tunnel 2
-9 -9
-35 -25 -15 -5 5 15 -35 -25 -15 -5 5 15
Distance from centre of Tunnels, m Distance from centre of tunnels, m
Settlement of Track 5, Tunnel 2

0
-1
-2
-3
-4 Tunnel 1
Tunnel 2
-5
-6
-7 RSP
-8 ELB
Average
-9
-35 -25 -15 -5 5 15 25 35 45 55
Figure 5-20 Final Transverse Settlement Trough for Three Tracks due to Tunnel 2
Settlement
Figure 5-21 shows separate final transverse surface settlement troughs observed in the
field for the three tracks due to boring of Tunnel 1 and Tunnel 2. The combined
settlement troughs are produced by adding the settlement troughs of Tunnel 1 and
Tunnel 2 (assuming no significant settlements took place after Tunnel 1 finished and
before Tunnel 2 started) and are shown in Figure 5-22.
112
These figures show that:
The settlement induced by boring of the second tunnel (Tunnel 2) was less than
that for the first tunnel (Tunnel 1) for all three tracks; this trend contrast with
trends reported by Perez Saiz et al. (1981), Phienwej et al. (2006) and Sugiyama
et al. (1999), but agrees with that observed by Nyren (1998) for St.James Park
and (after construction of Jubilee Line extension, London).
The settlement of Track 3 was largest and that of Track 1 was lowest for both
tunnel drives. It is possible that this situation arose because of localised
variations in soil stiffness and strength.
Although Track 1 and Track 5 had same volume losses (discussed below), Track
5 had a narrower and deeper settlement trough compared to track 1 for both
Tunnel 1 and Tunnel 2 boring.
-30 -20 -10 0 10 20 30

0
-1
-2
-3
-4
-5
-6
Track 1
-7 Track 3
Track 5
-8
-10
Distance from centre of Tunnels, m
Figure 5-21 Transverse Settlement Troughs due to Tunnel 1 and Tunnel 2
113
-30 -20 -10 0 10 20 30
0
-1
-2
-3
-4
-5
Tunnel 2
-6
-7 Track 1
-8 Track 3
Tunnel 1
Track 5
-9
Disctance from centre of tunnels, m
Figure 5-22 Combined Transverse Settlement Troughs after Tunnel 2 Boring
Trough Width
The width of the transverse settlement troughs (assuming zero movement prior to the
boring of the respective tunnels) at ground level on either side of the tunnel centre line
was assessed from the observed RSP and EL beam data, assuming the trough extended
to a distance where the settlement was less than about 0.1mm. These widths are listed in
Table 5-2 and indicate that:
The settlement troughs were symmetrical during Tunnel 1 boring.
The settlement troughs induced by Tunnel 2 were not symmetrical and indicated
a reduced trough width on east side of the tunnel axis and a larger trough width
on the west side. Given the overall lower settlements induced during Tunnel 2,
the trend on the east side is possibly due to an increase in ground stiffness due to
the presence of the stiff lining of Tunnel 1 and increases in the in-situ mean
effective stress. This will be discussed in Section 6.8.7
114
Volume Loss
The additional soil that has to be excavated over and above the volume associated with
the tunnel area is quantified in terms of volume loss (see Section 2.3.1) and is usually
expressed as a percentage fraction. The areas of the settlement troughs indicated on
Figure 5-21 are used to derive the volume losses (assuming negligible volume change in
the soil above the tunnel) summarised in Table 5-2. It is evident that:
All three tracks showed lower volume losses during Tunnel 2 boring than during
Tunnel 1 boring. On average the volume losses for Tunnel 1 and Tunnel 2 were
0.18% and 0.1% respectively.
Tracks 1 and 5 experienced comparable and relatively low volume losses,

whereas Track 3 experienced significantly larger volume losses during both
tunnel bores.
Although the maximum settlement at Track 1 and 5 differ, the volume loss
calculated from the respective settlement troughs remained the same. This is
because Track 5 had a narrower and deeper settlement trough compared to Track
1 for both Tunnel 1 and Tunnel 2 boring.
Max.Settlement Volume Trough Width (m)

(mm) loss (%)
East West Total
Track 1 4.5 0.15 13.7 12.3 26.0
Tunnel 1
Track 3 8.9 0.25 13.1 11.5 24.4
Track 5 6.2 0.14 10.1 10.0 20.1
Track 1 2.2 0.07 7.1 14.1 21.2
Tunnel 2
Track 3 6.0 0.19 11.9 17.8 29.7
Track 5 3.2 0.06 5.4 11.5 16.9
Table 5-2 Settlement Details of Three Tracks
115
5.5.4 Longitudinal Settlement Profile
Typical longitudinal settlement profiles observed at different times (marked as 1 and 2)

for Tunnel 1 boring are shown in Figure 5-23. This figure also shows the location of
cutter face (face pressure, FP) and the location of grouting (grout pressure, GP) for each
of those observed profiles. The longitudinal settlement profiles (marked as 2) indicate
that the heave was due to both face pressure as well as grout pressure. Both profiles
show the same final settlement after advancement of the TBM, indicating that the
settlement stabilised and is marked by monitoring point A. It can be seen from the
figure that the settlement trough (marked as 1, which is the first settlement profile)
stabilised around 13 m behind the cutter face and it shows that the settlement stabilised
as soon as the tunnelling activity finished at that location (note that the length of the
TBM was 8.4m) agreeing with the transverse settlement measurement (see section
5.5.1). It is also of interest to note that all the volume losses associated with tunnelling
happened behind the tunnel heading.
1.0
G FP
1 FP
GP 2
0.5
0.0
-0.5
-1.0
A
-1.5
0 10 20 30 40
Distance of TBM face from monitoring point, m
Figure 5-23 Typical Longitudinal Profile for Tunnel 1 Boring
116
5.6 COMPARISON OF TRANSVERSE SETTLEMENT TROUGHS WITH

GAUSSIAN APPROACH
Numerical methods to predict the observed transverse settlement troughs are dealt with
in Chapter 6. This section examines the applicability of the widely used Gaussian
method to predict the shapes of the greenfield troughs observed during two tunnel
boring in the predominantly sandy soils encountered north of William Street Station.
Best fit parameters for use of this method in conditions similar to those observed in
Perth are established.
The equation for the Gaussian profile is,
y 2 2i 2 )
Sv S max e ( Equation 1- 1
Where,
Smax Maximum settlement
Sv Settlement at offset y from the tunnel centre line
i The transverse horizontal distance from the centre-line of the tunnel to the point
of inflection of the settlement trough
Equation 1-1 can be re-arranged to allow derivation of i, if the maximum settlement and
the settlement at an offset distance, y , is known:
1 y2
i Equation 1- 2
2 ln ( S v S max )
If the settlement profile is of Gaussian nature, then ln (Sv/Smax) should be proportional

to y2 , allowing i to be estimated from the constant of proportionality (slope). The linear
regression lines between ln (Sv/Smax) and y2 based on the field measurement for Track 1,
Track 3 and Track 5 are given in Figure 5-24 and Figure 5-25 (plotted separately for
east and west of centre line of tunnel) for Tunnel 1 and 2 respectively. A good linear fit
for the data for both tunnel runs indicates that the settlement profile is of Gaussian
nature. The best estimate i values were derived from the constant of proportionality for
117
each track for both tunnel borings; these are listed in Table 5-3. The i values east and
west differs for Tunnel 2 because of the asymmetrical nature of the associated
settlement troughs discussed previously.

0 0
y = -0.0198x
2 East of C/L
-1 R = 0.9981 -1 West of C/L
-2 -2
y = -0.0332x
-3 East of C/L -3 2
R = 0.9937
West of C/L y = -0.028x
2
-4 y = -0.02x -4 R = 0.996
2
R = 0.9951
-5 -5
0 30 60 90 120 150 180 210 0 30 60 90 120 150 180 210
2 2 2 2
y (m ) y (m )
Track 5, Tunnel 1
0
-1 y = -0.0414x
2
R = 0.9999
-2
East of C/L
-3 West of
C/L
-4 y = -0.0446x
2
R = 0.9967
-5
0 20 40 60 80 100 120
22
y (m )
Figure 5-24 Linear Regression Lines for Tunnel 1 Boring
118

0 0
y = -0.0237x y = -0.0141x
2
-1 2
R = 0.9959 -1 R = 0.9929
-2 -2
East of C/L
West of C/L
-3 -3 East of C/L
y = -0.0317x West of C/L
-4 y = -0.0421x -4 2
R = 0.9956
2
R = 0.9984
-5 -5
0 30 60 90 120 150 180 210 0 30 60 90 120 150 180 210
2 2 2 2
y (m ) y (m )
Track 5, Tunnel 2
0
-1 y = -0.0405x
2
R = 0.9976
-2
-3
East of C/L
West of C/L y = -0.0625x
-4
2
R = 0.9967
-5
0 20 40 60 80 100
2 2
y (m )
Figure 5-25 Linear Regression Lines for Tunnel 2 Boring
119
Tunnel 1 (i) (m) Tunnel 2 (i) (m)
East of West of Average East of West of C/L Average

C/L C/L C/L
Track 1 5.01 5.03 5.02 2.31 4.22 3.26
Track 3 4.2 3.8 4.0 3.97 6.43 5.2
Track 5 3.4 3.5 3.5 2.33 5.31 3.82
Table 5-3 Estimated i Values
The linear nature of the data plotted on Figure 5-24 and Figure 5-25 confirms the
Gaussian nature of the transverse settlement troughs. It should be noted, however, that i
values recorded for Tunnel 2 are not, on average, comparable to those for Tunnel 1;
these i values are lower to the east of the centreline (presumably due to the presence of
Tunnel 1) and higher to the west of the centre line.
5.7 DISCUSSION
The average volume losses of 0.18% and 0.1% for Tunnels 1 and 2 respectively are
within the range of expectations of the earth pressure balancing technique employed for
boring of the two tunnels at Perth. Some notable features of the surface movement
include:
The significantly higher final settlements measured at the location of Track 3
The relatively large amounts of heave recorded
The asymmetrical nature of the settlement trough for Tunnel 2
The lower volume losses and settlements associated with Tunnel 2
This section discusses (i) factors influencing the observed patterns of heave and
settlement during boring of the two tunnels, (ii) the nature of transverse settlement
120
troughs observed at three locations close to each other, (iii) recommendations for the
trough width parameter (K) appropriate to the ground conditions in Perth CBD and (iv)
the relationship between the parameter i , tunnel depth and tunnel diameter.
5.7.1 Factors Contributing to Heave
Some of the particular characteristics of the Perth tunnelling operation are discussed in
this section to shed light on the observed pattern of vertical surface movement. One of
the, perhaps surprising, features of the data was the amount of short term heave
generated during tunnelling. Such heave can arise due to the application of large
pressures at the tunnel face and/or during grouting of the tail void. The face pressure
adopted for the tunnel boring was estimated using the face stability approach. The
calculation investigated the stability of the tunnel face considering limit equilibrium of a
sliding wedge at the face loaded by a prismatic body (Sigl and Yamazaki, 2007). The
required face pressure was also estimated using analytical methods considering a
circular hole in an elastic plastic continuum. Figure 5-26 shows the estimated face
pressures for the northern bored tunnel section (area under consideration). The grout
pressure was calculated based on a stress deformation analysis around a circular
opening in ground subjected to internal pressure.
Figure 5-26 Estimated Face Pressure along Northern Bored Tunnel (after Sigl and
Yamazaki, 2007)
121
Face Pressure
The face pressure (FP) was applied at the cutter head to stabilise the soil and to
minimize the volume loss and the settlement. Details concerning the applied face
pressure are given in Section 3.7.4. Figure 5-27 shows the variation of FP during Tunnel
1 and Tunnel 2 boring at the railway tracks location. The figure also shows the
estimated in situ total horizontal stress ( h = K0 'v +u) at the tunnel springline
(assuming a K0 value of 0.5 for the Spearwood sand and 1.0 for Perth formation). As
seen from Figure 5-27, the FP/ h ratio is about 1.25. A comparison of Figure 5-26 and
Figure 5-27 shows that the ratio of the applied and estimated face pressure is about 1.4.
This high face pressure must have therefore contributed to the high heave levels.
Although the FP/ h ratio was similar for Tunnel 1 and Tunnel 2 in the area under
consideration, the heave observed was less for Tunnel 2 boring for all the tracks. This
could be because of the higher overall stiffness of the soil due to the presence of the first
tunnel.
Tunnel 1 & Tunnel 2 Face Pressure

Tunnel 1FP(kPa)
250 Tunnel 2 FP(kPa)
Tunne 2 hori. effe. stress
Tunne 1 hori.effe. stress
200
150
100
Railway
Tracks
50
100 125 150 175 200 225 250
Chainage, m
Figure 5-27 Variation of FP and Horizontal Effective Stress during Tunnel 1 and
Tunnel 2 Boring (Note Tracks 1, 3 & 5 are at approximate chainages of 138m, 152m and
177m for Tunnel 1 and 146m, 164m and 192m for Tunnel 2)
122
Grout Pressure and Grout Volume
The space between the tunnel lining and the shield tail extrados were filled with grout
under pressure to minimise settlement. The theoretical volume of the tail void,
excluding the effect of any over cutting, is given by
2 2
Vtailvoid ( D1 D2 ) L (6.90 2 6.712 ) 1.0 2.03m 2 Where,
4 4
D1 - Outer diameter of the lining
D2 - Diameter of the cutter face
L - Length of each ring
Actual grout volumes employed per ring are plotted on Figure 5-28, where they are
compared with the theoretical volume. It is apparent that typical volumes pumped were
about 10% greater than this theoretical volume, although two volume spikes were
observed in the vicinity of Track 3 during Tunnel 1 boring. Largest heaves were
measured at the location of Track 5 (Chainage ~180m) and therefore the data on Figure
5-23 do not suggest that tail void grouting was a major factor contributing to heave.
Grout volume for Tunnel 1 and Tunnel 2

4
Railway Tracks
3
Tunnel 1
1 Tunnel 2
Theoretical Grout Volume
0
100 125 150 175 200 225 250
Chainage, m
Figure 5-28 Variation of Grout Volume / Ring
123
Tunnel 1 & Tunnel 2, Grout Pressure
Tunnel 1GP(kPa)
300
Tunnel 2 GP(kPa)
Railway Tunne 2 hori. effe. stress
250 Tracks Tunne 1 hori.effe. stress
200
150
100
50
100 125 150 175 200 225 250
Chainage, m
Figure 5-29 Variation of GP and Horizontal Effective Stress during Tunnel 1 and
Tunnel 2 Boring
Figure 5-29 shows the variation of grout pressure (GP) for Tunnel 1 and Tunnel 2
boring. Although grout pressures for the Tunnel 2 boring were higher than those used
for Tunnel 1, less heave was observed for Tunnel 2. Heave was therefore not directly
related to grout pressure.
5.7.2 Factors Contributing to Settlement
Factors considered to have an effect on the observed tunnelling induced settlement

include (i) ground water flow into the tunnel, (ii) curved tunnel alignment, (iii) effects
of first tunnel on the second tunnel and (iv) effects of inducing heave during tunnelling.
These factors are discussed below.
Ground Water
Settlements could have been affected by ground water flow into the tunnel, bearing in
mind that tunnelling in this area was conducted in soil with full hydrostatic water
pressures. The tunnel designers state that this did not occur for the following reasons:
124
Water inflow into the tunnel during the excavation was prevented or
significantly reduced by maintaining the face pressure above the hydrostatic
water pressures.
The pore water pressure difference between inside the cutter head and the point
of muck release onto the TBMs belt conveyor was maintained by keeping a soil
plug within the screw conveyor by using appropriate soil conditioning agents
such as foam or polymer.
The joints between the linings were fully water proofed to avoid any leak in to
the tunnel.
Curved Tunnel Alignment
The bored tunnel had curves with radii ranging from 135m to 510 m along the
horizontal alignment and 2800 m to 3000 m along the vertical alignment (Sigl &
YamaZaki, 2007). The volume loss due to over cutting at curvature for easy movement
of the TBM should be considered depending on the soil condition. More volume loss
and more settlement were expected for the second tunnel because of its higher curvature
(Radius of tunnel 1 was 143m and for Tunnel 2 it was 135 at this location). However,
Tunnel 1 showed greater volume losses and settlement for all the tracks (Table 5-2).
Over cutting due to copy cutters and forcing the TBM Shield into curves did not
therefore appear to have a major influence on the observed settlements.
Effects of First Tunnel on the Second Tunnel Run
It is often considered that boring of given tunnel will reduce the stiffness of the soil in
its vicinity and that greater settlement will be induced during boring of a second
adjacent tunnel (Perez Saiz et al. (1981) and Sugiyama et al. (199)). As indicated by the
measurements discussed above, this was not the case and, in fact, lower settlements
were induced by the second tunnel. The first tunnel also evidently restricted the extent
of the transverse settlement trough for the second tunnel, which consequently had an
asymmetrical transverse settlement trough. The reinforcing effect of the high stiffness
lining would have possibly reduced the settlement and heave as well. The influence of
125
adjacent structural stiffness on the width of settlement trough is again discussed in
Section 7.7. and in Section 8.5.1
Effects of Preceding Ground Heave
Heave of the soil was evidently induced by significant face pressures and subsequently,
as the tunnel face advanced, the same soil experienced unloading. The data show that
while heave may have reduced the settlements occurring during the initial unloading
phase, larger long term final settlements usually occurred in areas where larger heave
values were recorded. It would appear that the higher settlements (and heave) that
occurred close to track 3 arose because the lower soil stiffness in this area.
5.7.3 Nature of Settlement Trough
The surface settlements normalised by the respective maximum settlements for Tunnel
1 are shown in Figure 5-30. These indicate that, as the tunnel moves northwards (from
Track 1 to Track 5); there is a general trend for the width of the settlement trough to
reduce. A similar trend was also observed for Tunnel 2 (with the exception of Track 3).
This trend is consistent with the inference from Figure 5-4 and Figure 5-7 that the
tunnel moves progressively out of the Perth Formation and into the Spearwood dune
sand as it moves northwards. A narrower trough width in sand has been seen in
centrifuge model tests, which, with sufficient volume loss, ultimately leads to a failure
mechanism involving a chimney propagating vertically from the tunnel to the surface
(Potts, 1976 and Chambon & Corte, 1994). Although as the depth of the tunnel
decreases, the settlement trough tends to be narrow, in this case, the variations in the
depth of tunnel axes were very small (Table 6-5)
126
0.0
0.5
1.0
Track 1
Track 3
Track 5
1.5
-20 -10 0 10 20
Distance from centre of tunnel, m
Figure 5-30 Normalised Field Settlement Troughs for Tunnel 1
The trough width parameter K values corresponding to the i values shown in Table 5-3
are summarised in Table 5-4 and lines of constant K are plotted along with the i values
on Figure 5-31 (Only the west side is plotted for Tunnel 2 except for Track 3); this
figure also shows the range of i values recommended by Mair and Taylor (1997) in
sands and gravels. It is apparent that the observations are in keeping with these
recommendations, noting that the decreasing K values (or slopes on Figure 5-31) arise
in deposits with reducing quantities of clay.
Tunnel 1 Tunnel 2
z0 (m) K = i/z K (west)
Track 1 11 0.46 0.42
Track 3 10.5 0.41 0.57
Track 5 10 0.34 0.35
Table 5-4 Estimated i and K Values at the Three Track Locations
127
Offset to point of inflection (i), m
0 2 4 6 8 10 12 14
0
5
10
15
20
25 i=0.45z0
30 i=0.35z0
35 i=0.25z0
40
Figure 5-31 Variation of i with z0 for Tunnels in Sand and Gravels (Mair and
Taylor, 1997)
5.7.4 Relationship between i , Tunnel Depth and Tunnel Diameter
Peck (1969) suggested a relationship, shown in Figure 5-32, between the parameter i ,
tunnel diameter (D) and tunnel depth (z0) for different ground conditions. The observed
relationships between 2 i /D and z0/D are, for both tunnel drives, seen to fall under the
category of sands above water table (which is in the same category as rock and hard
clays) although the tunnels were below ground water table. This could have been either
because of the achievement of the effective control of ground water flow into the tunnel
as detailed in Section 5.7.2.or, as observed by Mair and Taylor (1997), there is no
difference between tunnels above or below ground water table.
128
2i/D
0 1 2 3 4 5
0
2 Sands below
groundwater level
4
Soft to stiff
clays
6
Rock, hard clays,
sands above
8 groundwater level
10 Tunnel 1
Tunnel 2
12
Figure 5-32 Relation between i , Tunnel depth and Diameter for Different Ground
Conditions (Peck, 1969)
5.8 SUMMARY
The observed patterns of greenfield movement due to the EPB boring of two adjacent
tunnels have been described in this Chapter. The main findings are as follows:
1. The average volume losses of 0.18% and 0.1% for Tunnels 1 and 2
respectively are within the range of expectations of the earth pressure
balancing technique employed.
2. No long term settlement occurred in the predominantly sandy conditions
present at the site.
3. Larger volume losses occurred consistently at one location for both
tunnel drives (at the location of Track 3) indicating the presence of
localised weaker soil and the need for thorough site investigations in
advance of tunnelling projects.
4. The width of the transverse settlement trough associated with a second
tunnel drive was lower than the greenfield value due to the influence of
the first tunnels stiff lining and possibly also due to increased soil
stiffness.
129
5. The trough width parameter K estimated from the observed settlement
troughs at the three locations is in agreement with past observations
made in similar ground conditions around the world.
6. The first tunnel boring induced more settlement and heave than second
tunnel at all three locations.
7. The settlement/ heave governing factors such as grout pressure, grout

volume, curved tunnel alignment, ground water, etc. did not seem to
have an effect on the settlement or heave.
130
CHAPTER 6 GREENFIELD SETTLEMENT-
NUMERICAL MODELLING
6.1 NUMERICAL MODELLING
Two dimensional (2D) and three dimensional (3D) finite element (FE) analyses were
carried out for the first and second tunnel drives. Plaxis 2D (version 9.02) was used for
2D FE analysis and Plaxis 3D Tunnel (version 2.0) was used for 3D FE analyses.
Separate analyses were carried out for North and South soil profiles (see Section 5.2),
assuming two representative volume losses.
FE analyses were carried out using (i) Mohr-Coulomb model (MC), (ii) Hardening soil
model (HS) and (iii) Hardening soil model with small strain stiffness (HSSmall). The
settlement predicted using different models and two soil profiles are then compared to
examine the effect of the soil model. Predicted movements are also compared with those
measured in the field and the main factors influencing the nature of the predicted
settlement trough are discussed.
6.2 SOIL MODELS
The Mohr-Coulomb, HS (Hardening Soil) and HSSmall (Hardening soil model with
small strain stiffness) models were used for the 2D analyses. Only the HS model was
employed for the 3D analyses as HSSmall was not incorporated in Plaxis 3D Tunnel.
Each of the soil models is described in the following sections.
6.2.1 Hardening Soil Model (HS)
The hardening soil (HS) model, which is described in detail by Schanz et al. (1999), is a
non-linear elastic plastic formulation which adopts multiple yield loci as a function of
plastic shear strain and a cap to allow volumetric hardening. Loading and unloading
within the current yield surface are assumed elastic (defined by a separate modulus, Eur)
131
and the Mohr Coulomb failure criterion is employed. Specification of the peak dilation
angle ( ) and friction angle ( ) defines the critical state friction angle ( cv) according to:
sin sin (1)

sin cv
1 sin sin
The ratio of the plastic volumetric strain rate to the plastic shear strain rate is taken
equal to sin m, where m is the mobilised dilation angle which is derived from equation
(1) by setting equal to the mobilised friction angle ( m). Plastic shear strains are
derived assuming plastic volume changes are zero (which is a reasonable approximation
for a hard soil). When modelling undrained conditions with direct input of undrained
strength (su), m is assumed to be zero (with =0). m is also set to zero for low
mobilised friction angles (sin m/sin <0.75) and when calculated m values are
negative. The Plaxis H-S model allows specification of a cut-off for dilation when the
void ratio reaches a nominated maximum void ratio (emax). For triaxial conditions
(where the minor and intermediate principal effective stresses equal 3), the model
predicts the following hyperbolic relationship for the drained secant Youngs modulus,
E:
m
q ref '3 (2)
E 2 E 50 1 R f where E 50 E 50
qf p ref
where q is the deviator stress and E50ref is the E value at the reference confining pressure
pref, (which can be determined when q is 50% of the maximum deviator stress, qf), Rf is
the failure ratio set to 0.9 and the exponent, m, controls the stress level dependency of
stiffness. If is set to zero (for an undrained analysis), E50 is independent of stress level
and equal to the specified value of E50ref.
Input soil parameters for the HS model are given below.
______________________________________________________________________
sat Saturated unit weight of soil,
unsat Unsaturated unit weight of soil,
k_x, k_y & k_z Permeability of the soil in the x, y and z directions
c` Cohesion
132
Chapter 6 Greenfield Settlement-Numerical Modelling
` Friction angle(o),
Angle of dilatancy (o),
E50 ref Secant stiffness in standard drained triaxial test at reference

stress,
Eoedref Tangent stiffness for primary oedometer loading at reference

stress,
m Power for stress-level dependency of stiffness,
Eurref Unloading/reloading stiffness,
ur Poissons ratio for unloading reloading,
pref Reference stress for stiffness,
K0 Coefficient of earth pressure at rest, and
K0nc K0 value for normal consolidation,
6.2.2 Hardening Soil Model with Small Strain Stiffness (HSSmall)
The HS-Small model is an extension of the original Hardening Soil model (HS model),
developed by Schanz et al. (1999) and implemented in Plaxis. The extension uses a
Small Strain Overlay (Benz 2006) to detect small strain amplitudes and applies
comparatively higher stiffness values for computations at these small strain levels. The
overlay is inactive for large strain amplitudes, when the soil response is determined by
the HS model. Changes in loading direction may re-activate the overlay.
The HS-Small model uses all input parameters used by the HS model (refer to Plaxis
material model manual), along with two additional parameters. These are the reference
small strain shear modulus (G0ref) and the parameter 0.7, which is the shear strain
required to reduce the shear modulus from G0 to 70% of G0. The value of G0 is
described as a function of the minor principal effective stress, '3, after Benz (2006):
m1
ref ' 3 c' cot '
G0 G0 (2)
p ref c' cot '
133
where pref is the reference pressure (default 100kPa), c' is cohesion, ' is the friction
angle and m1 is a constant. It is noteworthy that Kohata et al. (1997), and others, show
that G0 in triaxial compression tests and footing tests (Gvh0) is expected to vary more
strongly with the vertical effective stress than the minor/horizontal effective stress.
The small strain overlay uses a relationship between shear strain ( ) and secant shear
modulus (G) proposed by Santos & Correia (2001):
G 1
G0 3
1
7 0.7
(3)
This relationship is only used when G lies between G0 and Gur, where Gur is the
unload/reload shear modulus. The degradation in stiffness continues through a transition
curve until the soil behaviour is controlled by the hardening soil (HS) model, where the
stiffness in primary loading is determined by two other stiffness parameters, namely E50
and Eoed.
6.2.3 Mohr-Coulomb Model (MC)
The Mohr-Coulomb model employed here is an elastic-perfect plastic model. The yield
surface is fixed i.e. fully defined by the model parameters and hence not affected by
(plastic) straining. The behaviour is purely linear elastic and isotropic within the yield
surface. No stress level dependence of soil stiffness is considered and hence the
equivalent stiffness appropriate for the envisaged stress and strain level in each layer
needs to be input.
The MC model is defined by the following five input parameters:
E - Youngs modulus
Poissons ratio
c' Cohesion/cementation
' Friction angle(o),
134
Angle of dilatancy (o),
It is noted that, in this model, the pre-consolidation pressure cannot be modelled for the
initial stress generation i.e. the effects of stress history are not taken into account
explicitly.
6.3 PARAMETER SELECTION
The selection of appropriate input parameters is the key to the success of any numerical
analysis. Hence a thorough analysis of the available field and laboratory test data was
carried out to best define the soils under consideration. A detailed separate laboratory
study has also been carried out on undisturbed soil samples collected from the site and
this study is summarised in Chapter 4. The estimation of the input parameters selected
for each of the soil models are explained below.
6.3.1 Hardening Soil Model (HS Model)

ref ref
The main parameters for the HS model are stiffness parameters (E50 , Eoed and
Eurref.), effective stress strength parameters ( ', c and and stress history (POP and
K0). The basis for estimation of each of these parameters is given below.
Stiffness Parameters E50 ref, Eoedref and Eurref
The relationship, shown in Figure 6-1, between shear modulus (G), at reference stress
(100 kPa), and shear strain was deduced from field and laboratory tests. The small strain
shear modulus (G0) was obtained from a wide range of seismic cone penetration tests
(SCPTs) conducted in Perth CBD (see Chapter 4), which indicated Go values in the
following range.
0.5
G0 p'
(1300 300) in clays with p ref 100kPa
p ref pref
0.5
G0 p'
(2200 400) in sands with pref 100kPa
p ref p ref
135
Large strain shear moduli (G) were derived from the Youngs moduli measured in
triaxial compression tests discussed in Chapter 4; these moduli were converted to shear
moduli by assuming a large strain Poisons ratio of 0.4. As indicated on Figure 6-1, the
best fit G0ref vs axial strain relationship was fitted using the hyperbolic relationship of
the Plaxis HS model corresponding to an E50 ref of 114MPa and 68 MPa for upper bound
and lower bound trends respectively. The fines contents of the samples (indicated in the
figure) show that upper bound stiffnesses relate predominantly to low fines content
samples and hence were assumed relevant to the sands for the numerical analyses. The
lower bound values relate primarily to high fines content samples and were assumed
relevant to the clays. It is noteworthy that the normalised stiffness of the sand of the
Perth (Guildford) formation was similar to that of the Spearwood dune sands (See
Section 4.2).
The Eoedref and Eurref values were estimated using the respective E50 values. Based on
ref
research on a wide variety of sand, Schanz (1998) recommended Eoed E50ref . Since
there are no such data available for the clay material, the same relationship was used for
clay as well. Schanz (1998) also found that, for various soil types, the average value
of Eur 3 E50ref ; this relationship is in agreement with the triaxial test data performed on
the undisturbed Perth Formation clay samples.
136
250
Reference stress = 100 kPa
FC=57%
FC=51%
FC=95%
200 FC=39%
FC=20%
Range measured
in SCPTs
150
100 Ranges Measured from the Lab
50
0
0.0001 0.001 0.01 0.1 1
Axial strain (%)
Figure 6-1 Figure Showing the Relationship between G0ref and Strain at Reference
Stress of 100 kPa
Effective Stress Strength Parameters
Based on the laboratory tests, ' values of 35o and 28o were adopted for sand and clay
respectively with a c' value of 0. The angle of dilatancy ( was taken as zero for sand
as no dilation effect was observed in the triaxial testing as discussed in Section 4.2 .
Stress History
A K0 value of 0.5 was adopted for the normally consolidated Spearwood sand. Based on
laboratory tests data described in Chapter 4., a K0 value of 1.0 and preconsolidation
pressure (POP) of 700 kPa was used for sand and clay deposits of the Perth Formation.
137
The design soil parameters adopted for each layer for the North and South soil profiles
are shown in Table 6-1 and Table 6-2 respectively.
Layer Description Top of the Type E50ref ' ur pref Rinter 1

No. of soil Layer (m (MPa) (0) (kPa)
AHD)
1 Sand +11 Drained 114 35 0.15 100 0.8

2 silty sand -2 Drained 114 35 0.15 100 0.8
3 Sand -6 Drained 114 35 0.15 100 1
4 silty sand to -12 Drained 114 35 0.15 100 1
sandy silt
5 clayey silt to -14 Undrained 68 28 0.15 100 1
silty clay
6 Sand -19 2 Drained 114 35 0.15 100 1
Table 6-1 Hardening Soil Model Parameters for North Profile
Layer Descriptio Top of Type E50ref ' ur Pref Rinter

No. n of soil the Layer (MPa) 0
() (kPa
(m AHD)
1 Sand +11 Drained 114 35 0.15 100 1
2 Clay/Silty +4 Undrained 68 28 0.15 100 0.8
clay
3 Sand +2 Drained 114 35 0.15 100 0.8
4 Clayey silt/ -5 Undrained 68 28 0.15 100 1
Silty clay
5 Clay -13 Undrained 68 28 0.15 100 1
6 Sand -202 Drained 114 35 0.15 100 1
Table 6-2 Hardening Soil Model Parameters for South Profile
1
Interface strength
2
Bed Rock starts at -24 m
138
6.3.2 Hardening Soil Model with Small Strain Stiffness (HSSmall Model)
ref
The two parameters for HSSmall in additional to those used in the HS model are G0
(the small strain shear modulus at reference stress) and 0.7 (the shear strain level at
which the secant shear modulus G is reduced to 70% of G0). The lower bound (clay)
ref
and upper bound (sand) G0 values of 220 MPa and 130 MPa respectively correspond
ref
to respective 0.7 values of 0.005% and 0.002% and these G0 and 0 .7 were selected
for the analyses.
6.3.3 Mohr Coulomb Model (MC)

ref ref
The E value in the Mohr-Coulomb model is selected as the E50 value in the HS
model. Since the model does not accommodate the stress dependency of the stiffness,
for each layer of soil, a stiffness value appropriate to the effective stress in that layer
needed to be assessed. Table 6-3 and Table 6-4 show the parameters for the Mohr-
Coulomb model. The MC model does not account for stress history and hence does not
require specification of a POP value for the initial stress generation.
Layer Description of Type Eref cref ' ur m Rinter

No. soil (MPa) (kPa) (0) (-) (-)
1 Sand Drained 92 0 35 0.15 0.5 0.8
2 Silty sand Drained 178 0 35 0.15 0.5 0.8
3 Sand Drained 196 0 35 0.15 0.5 1
Silty sand to
4 sandy silt Drained 208 0 35 0.15 0.5 1
Clayey silt to
5 silty Clay Undrained 130 0 28 0.15 0.5 1
6 Sand Drained 234 0 35 0.15 0.5 1
Table 6-3 Mohr-Coulomb Model Parameters for North Profile
139
Layer Description of Type Eref cref ' ur m Rinter
0
No. soil (MPa) (kPa) () (-) (-)
1 Sand Drained 68 0 35 0.15 0.5 1
2 Clay/ Silty Clay Undrained 86 0 28 0.15 0.5 0.8
3 Sand Drained 169 0 35 0.15 0.5 0.8
Clayey Silt to
4 Silty Clay Undrained 116 0 28 0.15 0.5 1
5 Silty Clay to clay Undrained 130 0 28 0.15 0.5 1
6 Sand Drained 235 0 35 0.15 0.5 1
Table 6-4 Mohr-Coulomb Model Parameters for South Profile
6.4 TWO DIMENSIONAL FE ANALYSIS
2D FE analyses were carried out for the single tunnel drive as well as for the second
tunnel drive for North and South soil profiles representative of the strata underlying the
railway tracks north of William Street station (see Section 5.2. for both soil profile). The
clay layers in all analyses were assumed to be undrained. The geometry, input
parameters for the analyses and the tunnel excavation modelling are explained in the
following.
Geometry
The two tunnels were bored at different elevations and the centre to centre distances
between the tunnels differed at each of the overlying track locations. Table 6-5 shows
the tunnel invert and crown level positions as well as the centre to centre distance
between the tunnels at each track location. Track 3 was selected for analysis (for which
the tunnel invert was -3.35m AHD and where the tunnel centrelines were 14.5 m apart).
Figure 6-2 and Figure 6-3 show the Finite Element mesh and location of the tunnel in
the North and South profiles respectively. The ground water level was at 9.2m AHD and
initial water pressure regime was assumed to be hydrostatic. The soil was modelled
using 15 node triangular elements. This element provides fourth order interpolation for
displacements and numerical integration involves twelve Gauss points.
140
The outer diameter of the tunnel was 6.9m and it had a lining thickness of 0.275m. The
length of model adopted was 110m to avoid any boundary effects. Both vertical
boundaries were restrained against horizontal movement and rotation. The bottom
boundary was restrained against both vertical and horizontal movement, and a no flow
and no consolidation boundary was specified at the base to represent impermeable rock.
Chainage Invert level Crown level Centre to centre distance

(m) (m AHD) (m AHD) between Tunnels (m)
Track 1 142 -3.35 3.56 14.08
Track 3 158 -3.06 3.85 14.52
Track 5 185 -2.44 4.47 14.45
Table 6-5 Tunnel Details
Spearwood sand
Silty sand
Perth Sand
Formation Silty sand
Clayey silt or silty clay
Sand
Figure 6-2 North Profile Showing Mesh and Tunnels for 2D FE Analysis
Spearwood sand
Clay or silty clay
Sand
Perth
Formation Clay
Sand
Figure 6-3 South Profile Showing Mesh and Tunnels for 2D FE Analysis
141
Volume Loss
The volume loss due to tunnelling is modelled by applying a contraction to the lining
(e.g. Franzius, 2005). The transverse surface settlement profile derived from field
measurements (see Chapter 5) was employed to calculate the volume loss for each of
the three tracks for Tunnel 1 and Tunnel 2 boring separately. Since the volume losses
were approximately the same for Track 1 and Track 5 for both Tunnel 1 and Tunnel 2
drives, the volume losses considered in Table 6-6 were considered sufficient to allow
comparisons between measurements and predictions at the three track locations and for
the two tunnel drives.
Volume losses (%)
Tunnel 1 Tunnel 2
0.15 0.06
0.25 0.19
Table 6-6 Volume Losses Specified in Numerical Analyses
Modelling of Tunnel Excavation
The tunnel excavation was modelled using the volume loss method. In the first
calculation step, the in situ soil was consolidated to a minimum excess pore pressure of
1kPa under the existing overburden pressure to simulate the in-situ initial field
condition. The soil and water was removed from the tunnel and the lining put in place in
the second calculation step. The tunnel lining was modelled using elastic shell elements
with a Youngs modulus of 35GPa, Poissons ratio of 0.15 and thickness of 0.275m.
Reduced strength properties in the interaction zone between the soil and structures are
given by introducing Rinter parameter of 0.8. In the third calculation step, a contraction
was applied at the centre of the tunnel to model the volume loss due to tunnelling. The
fourth calculation step involved prediction of changes that occurred as excess pore
pressure generated as part of the tunnel excavation were allowed to dissipate. HS,
142
HSSmall, and MC models were used for 2D FE analyses. Transverse settlement troughs
predicted by each of these models for both soil profiles are discussed below.
6.4.1 Hardening Soil Model (HS Model)
Figure 6-4 shows the transverse settlement trough predicted by Plaxis 2D using the HS
model for volume losses (Table 6-6) incurred in the north and south areas. It is
important to note that, although the volume loss is an input parameter, the volume loss
estimated from the volume of predicted settlement troughs were not same as that
inputted volume loss. This is because the predicted settlement trough is dependent on
factors that will be discussed later (in Section 6.8). Figure 6-4 show that settlements
were predominantly immediate for both soil profiles; this trend is consistent with the
field measurements. Slightly lower settlements are predicted after the consolidation step
reflecting heave that took place as negative pore pressures dissipated in the over
consolidated clay. This predicted relaxation was not identified in the field possibly
because it was too small to measure given the accuracy of the instrumentations installed
at the site.
The predicted excess pore pressure fields are presented in Figure 6-5 and Figure 6-6 (for
a typical volume loss of 0.25%), which indicate generation of both positive and negative
excess pore pressures. For both North and South soil profiles, negative excess pore
pressure was generated in the clay layer beneath the tunnel invert due to the relaxation
caused by the tunnelling. In the South profile, positive and negative excess pore
pressures can be seen in the clay layer through which tunnel was passing. The negative
excess pore pressure for south profile (~50kPa) was almost twice that for north profile
(~25kPa).
New and Bowers (1994) also found pore pressure reductions close to tunnels in London
clay, and a review by Mair and Taylor (1997) found that excess pore pressures
generated are generally negative when the ground is unloaded during tunnelling. For the
predictions shown on Figure 6-5 and Figure 6-6, the heave induced by swelling is small
due to the relatively localised nature of the negative excess pore pressure fields.
143
North profile, 2D Plaxis, vl=0.25 &0.19%, HS
0
-1
-2
-3
-4 Tunnel -2
-5
Before consolidation Tunnel -1
-6 After consolidation
Before consolidation
-40 -30 -20 -10 0 10 20 30 40
South profile, 2D Plaxis, vl=0.25 &0.19%, HS
0
-1
-2
-3
Tunnel -2
Tunnel -1
-4
-5
-40 -30 -20 -10 0 10 20 30 40
Figure 6-4 Transverse Settlement Trough for North and South profiles for Volume
Losses of 0.25% & 0.19%; HS model
144
Figure 6-5 Variation of Excess Pore Pressure for North Profile (vl=0.25%)
Negative excess
Figure 6-6 Variation of Excess Pore Pressure for South Profile (vl=0.25%)
145
6.4.2 Hardening Soil Model with Small Strain Stiffness (HSSmall Model)
Despite the additional small strain component of HSSmall model, the settlement
patterns predicted by this model are almost identical to those obtained using the HS
model. Both models predicted negative excess pore pressure and a reduction in
settlement after consolidation. A comparison of predicted settlement trough shapes is
presented later in this chapter.
6.4.3 Mohr-Coulomb Model
Figure 6-7 shows the predicted settlement trough obtained using the MC soil model for
Tunnel 1 employing E50 (see Section 6.3.3) as the input stiffness value. It is apparent
that the predicted settlement troughs especially for the South profile show heave rather
than settlement which is unrealistic. The use of the MC model with E50 is therefore not
discussed further.
Tunnel 1, Plaxis 2D,South and North Profile, MC

1
0
-1
-2
-3
Tunnel 1
-4
-5 Tunnel 2
South profile vl=0.15%
-6 South profile vl=0.25%
North profile vl=0.15%
-7 North profile vl=0.25%
-60 -40 -20 0 20 40 60

Figure 6-7 Transverse Settlement Trough for North and South profiles for Tunnel
1 (Volume Losses of 0.15% & 0.25%); MC with E=E50
Realistic soil parameters should be used for the analysis depending on the type of
problem investigated. Since tunnelling is an unloading problem, the Plaxis material
146
model manual suggests using Eur instead of E50 (Plaxis material model manual). Hence
the MC analyses were also carried out using Eur instead of E50. Eur was selected as 3
times E50, as taken for HS and HSSmall models. The predicted settlement troughs for
the North and South profiles for a typical volume loss of 0.25% using Eur are shown in
Figure 6-8, where it may be seen that the predicted settlement profile and the maximum
settlement are comparable with those of HS and HSSmall models (note, however, that
the volume loss estimated from the predicted settlement trough is less than 0.25%). As
for predictions using the HS models, the settlement also reduced during
consolidation/swelling (Figure 6-9). The comparison of the prediction using E50 and Eur
shows that when the stiffness is increased, the predicted settlement trough was deeper.
This is further discussed in Section 6.8.4
Tunnel 1, Plaxis 2D,South and North Profile, MCEur
-1
-3
-5
Tunnel 2 Tunnel 1
-7
-9 South profile vl=0.25%

North profile vl=0.25%
-11
-30 -20 -10 0 10 20 30
Figure 6-8 Final Transverse Settlement Trough for North and South Profiles; MC
with E=Eur
147
South profile, 2D Plaxis, vl=0.25 &0.19%, MCEur
0
-1
Tunnel 1
-2
-3
-4
-5
-6 Tunnel 2 Before consolidation
After consolidation
-7 Before consolidation
-9
-40 -20 0 20 40
Figure 6-9 Transverse Settlement Trough for South Profile; MC with E=Eur
6.5 COMPARISON OF THE PREDICTIONS USING DIFFERENT SOIL

MODELS
The predictions obtained using different soil models discussed above are compared here
to add some insights into the capabilities of each of the models. Figure 6-10 compares
the transverse settlement troughs predicted for an imposed volume loss of 0.25% at the
Northern and Southern profiles. This comparison shows that the troughs predicted using
the simple Mohr-Coulomb model with Eur are deeper and narrower compared to those
obtained using the more sophisticated HS and HSSmall models.
Although the MC model predictions (using Eur rather than E50) of surface settlements
were comparable to the HS and HSSmall models, the overall predicted patterns of
displacement fields for the MC model may be seen to differ from those of the HS and
HSSmall soil models on Figure 6-11, Figure 6-12 and Figure 6-13. These differences
arise due to well known limitations of the linear elastic MC model (e.g. no effect of
stress history, soil stiffness non-linear dependence on strain and stress level). Hence
only HS and HSSmall models are discussed further in this thesis.
148
Tunnel 1, North Profile(0.25%)
0
-1
-2
Tunnel 2
-3
Tunnel 1
-4
-5
-6 2D HS
-7 2D HSSmall
-8 2D MC
-20 -10 0 10 20 30 40
Tunnel 1, South Profile(0.25%)
0
-1
-2
-3 Tunnel 2
-4
Tunnel 1
-5
2D HS
-6
2D HSSmall
-7 2D MC
-8
-20 -10 0 10 20 30 40
Figure 6-10 Comparison of the Transverse Settlement Trough Predicted using

Different Soil Models
149
Figure 6-11 Deformed Mesh using MC (Eur) Model
Figure 6-12 Deformed Mesh of the Geometry using HS Model
Figure 6-13 Deformed Mesh of the Geometry using HSSmall Model
150
6.6 THREE DIMENSIONAL ANALYSIS (WITH HS MODEL)
Three dimensional (3D) analyses were carried out to investigate differences with the 2D
analyses discussed above. Although previous numerical studies have concluded that
neither 2D nor 3D analyses with realistic soil properties could predict the Gaussian type
surface settlement trough seen in the field, it was decided to perform 3D analyses using
an advanced soil model (HS model) and a new version of Plaxis 3D Tunnel (version 2).
This program is a finite element (FE) package specifically for the deformation analysis
of tunnels. 3D analyses were carried out for North and South soil profiles (assuming
undrained behaviour in the clay) for the first tunnel drive as well as for the subsequent
second tunnel drive. The same soil parameters to those used for the 2D analyses were
employed. The geometry, additional input parameters for 3D analysis and the modelling
of the tunnel excavation are explained in the following paragraphs.
Geometry
The 2D cross-section model was initially created with subsoil and the structures. Both
vertical sides were restrained against horizontal movement and rotation and the bottom
boundary was restrained against both vertical and horizontal movement. A no flow
and no consolidation boundary were specified at the base to represent the impermeable
rock. The 2D model was 110 m wide and was then extended in the 3rd (z) direction to
obtain the 3D model (i.e. Plaxis 3D tunnel does not allow modelling of a truly 3D
geometry). The model was extended 72 m in the z-direction. Figure 6-14 shows the
geometry of the tunnel excavation. In the z- direction the model was divided into a
number of slices. The length of the first and last slice was 25 m to reduce the influence
of boundary. The length of each slice excluding the end slices was equal to length of
each concrete liner (1.0m at this location). The springline of the tunnel was at 0.4m
AHD which was about 11.0m below ground level. The ground water table level was at
9.2m AHD.
The finite element mesh and location of tunnels in the North and South profiles are
shown in Figure 6-15 and Figure 6-16 respectively. The main features of the FE mesh
for North and South profile are summarised in Table 6-7. The interaction between the
TBM and the surrounding soil was modelled using an interface, which can
accommodate reduced friction between the TBM and surrounding soil.
151
~11.0m
Tunnel 2 Tunnel 1
y
z 14.5m 72.0m
x
110m
22.0m
25m 25m
1.0m TBM 1.0m
Slice #1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Figure 6-14 Geometric Model of the Tunnel Excavation
Figure 6-15 3D Mesh for the Plaxis Analysis North of the Rail Lines
152
Figure 6-16 3D Mesh for the Plaxis analysis South of the Rail Lines
Description North profile South profile
Elements 15- Node Wedge 15- Node Wedge

No. of Elements 15708 20608
Nodes 45996 58725
Gauss points 94248 123648
Average element size 472.0mm 432.2mm
Table 6-7 Main Features of the Mesh
The construction procedure for EPB TBM tunnelling is explained in Section 3.8. The
tunnel boring involves (i) applying face pressure at the front of the TBM, (ii) grout
pressure at the back of the TBM, (iii) thrust from the hydraulic jack pushing the TBM
forward. The volume loss due to tunnel boring has to be modelled along with the above
factors to completely define the tunnel boring in 3D. Plaxis 3D Tunnel version 2 (which
is used for the analysis in this thesis) models the face pressure by applying a prescribed
pressure at the cutter face. The grout pressure is modelled by applying an outward
pressure at the corresponding location. The volume loss is modelled by applying a
contraction to the shield. Since it was found that the predicted settlement trough is
153
influenced by the stiffness of the lining (see Section 6.8.2), the TBM shield was
provided with properties that were the same as those of the lining which was used in the
2D analysis (see Section 6.4). The force from the hydraulic jacks advancing the TBM is
modelled by applying a pressure at the back of the TBM.
The shield tunnel construction was modelled in each calculation phase in a step-wise
process as in the field. The tunnel boring machine was 8.4 m long and 6.9m in diameter.
The construction processes were divided into stages controlled by the length of each
ring (slice as shown in Figure 6-14) and the calculation consisted of a number of Staged
construction phases. A total of 14 staged construction phases were involved for
excavating Tunnel 1 alone (requiring 8 hours computational time). The following
features were modelled in each of these phases; see Figure 6-17 for a schematic view:
i) Support pressure at the face (face pressure) to prevent yielding at the face
during the excavation.
ii) Volume loss and pore pressure inside the TBM
iii) Installation of the tunnel lining and grouting the gap between the soil and the
lining (grout pressure); and
iv) TBM thrust.
Figure 6-17 Schematic View of the Modelling of Tunnel Boring
Input Parameters
The model parameters and the volume losses used for the 3D analysis are same as those
used for the 2D FE analysis. The additional parameters used for the 3D analysis are face
154
pressure (FP), grout pressure (GP) and TBM thrust. The average value of each of these
parameters during the actual tunnelling operations is given in Table 6-8, which in turn
was used as the input parameter for the analyses. The input parameters for each of the
calculation phases are the same, the only difference being that the face of tunnel will
have advanced by a distance of 1m (length of the ring) from that of the previous
construction phase.
Tunnel 1 Tunnel 2
Face Pressure (kN/m2/m) 194 194
Grout Pressure (kN/m2/m) 160 208
TBM Thrust (kN/m2) 520 502
Table 6-8 Face Pressure, Grout Pressure and TBM Thrust for 3D Analyses
6.6.1 Transverse Settlement Profile
The deformed mesh after excavating the full geometry using a combination of
representative volume losses is shown in Figure 6-18 and Figure 6-19 for Tunnel 1 and
Tunnel 1& 2 (combined) respectively. The deformation plane corresponding to each of
these analyses is shown in Figure 6-20 and Figure 6-21.
Figure 6-18 Deformed Mesh after Excavating Tunnel 1 (North profile, vl=0.25%)
155
Figure 6-19 Deformed Mesh after Excavating Tunnel 1 and Tunnel 2 (North
profile, VL=0.25% &0.19%)
Figure 6-20 Deformation Plane after the Excavation of Tunnel 1 (North profile, VL
VL =0.25%)
Figure 6-21 Deformation Plane after the Excavation of Tunnel 1 and Tunnel 2
(North profile, VL =0.25% &0.19%)
156
The settlement trough predicted by Plaxis 3D for a typical combination of volume

losses for both soil profiles shows the same behaviour as observed in the 2D analyses.
For example, the analysis result in Figure 6-22 shows slight reduction in settlement after
consolidation indicating the formation of negative excess pore pressure. The
comparison of the settlement troughs predicted using 2D and 3D FE analyses for both
soil profiles will be discussed later in Section 6.7.
North profile, 3D Plaxis, vl=0.15 &0.06%, HS
0.0
-1.0
-2.0
Tunnel 1
-3.0
Tunnel 2
-4.0
-5.0 After consolidation
-40 -30 -20 -10 0 10 20 30 40
157
South profile, 3D Plaxis, vl=0.25 &0.19%, HS
0.0
-1.0
-2.0
-3.0
-4.0
Tunnel 2
-5.0 Tunnel 1
-6.0
-7.0 Before consolidation
After consolidation
-8.0 Before consolidation
-40 -30 -20 -10 0 10 20 30 40
Figure 6-22 Transverse Settlement Trough of North and South Profile for Volume
Losses of 0.25% & 0.19%
6.6.2 Longitudinal Settlement profile
Typical longitudinal profiles predicted using 3D analyses are shown in Figure 6-23. The
position of the cutter face is marked on the profile and is also shown in the legends. The
longitudinal profile indicates a stabilised settlement from the start of the geometry.
Although this is unrealistic, past research (Franzius, 2003) shows difficulty in achieving
a stabilised settlement even after excavating a considerable distance (about 20 tunnel
diameters). The stabilised settlement from the start of the geometry could be because of
the fact that, in the analysis, the full volume loss (contraction) is applied in one
calculation step. In reality, however, the unsupported length of the soil will gradually
induce the settlement. The stabilised settlement indicates that there was no mesh
dependant settlement or boundary effect. The effect of boundary was significant when
the excavation was approaching the boundary on the other side (end of the geometry)
and this can be seen from Figure 6-24. As Franzius (2003) had experienced, in this
analysis also the mesh dependant effect was observed in the longitudinal profile
initially. However, after spending considerable amount of time, this could be solved and
a stabilised settlement at the start of the geometry could be predicted. The comparison
158
of predicted and observed settlement patterns will be discussed later in this chapter (see
Section 6.7.2).
North profile (vl=0.25%)

0
-1
-2
-3
Position of Cutter Face (m)
-4
36
-5 39
-6
-7
0 10 20 30 40 50 60 70 80
Distance from the start of tunnel, m
Figure 6-23 Typical Longitudinal Settlement Trough

0
-1
-2
-3
-4 Position of Cutter Face (m)

36
-5 39
42
-6
45
-7
-8
0 10 20 30 40 50 60 70 80
Distance from the start of tunnel, m
Figure 6-24 Boundary Effect for Longitudinal Trough
159
The predicted longitudinal settlement profile did not indicate any heave, as observed in
the field (See Section 5.5) although realistic input parameters were used for face
pressure and grout pressure. Hence, a parametric study was carried out varying the face
pressure (FP) and the grout pressure (GP) to evaluate the effect of those parameters in
3D analysis. Two combinations of FP and GP were used in the analysis. The first set of
pressures were double those applied in the field (FP=388 kPa and GP=320kPa) and the
second set were about half of those applied in the field (FP=110 kPa and GP= 80 kPa).
The transverse settlement profile and the longitudinal settlement profile are shown in
Figure 6-25 and Figure 6-26 respectively for both cases. It is interesting to note that
both combinations did not predict any heave either ahead or behind the TBM. Similarly
there was no effect on the maximum settlement either. This was contrary to the
prediction made by Kasper and Meschke (2006) for clay and is probably a Plaxis
modelling related issue. The inputted value of the face pressure does, however,
influence predictions as stability problems were noted in the analyses when FP dropped
below 100 kPa..
North profile, vl=0.25%

0
-1
-2
-3
-4
-5
-6 FP=388, GP=320
-7 FP=110, GP=80
-8
-60 -40 -20 0 20 40 60
Figure 6-25 Settlement Trough with Different FP and GP
160

0
-1
-2
-3 TBM FACE
-4 Grout pressure
FP=110, GP=80
-5
FP=388,GP=320
-6
-7
0 10 20 30 40 50 60 70 80
Longitudinal distance, m
Figure 6-26 Longitudinal Settlement Trough for Different GP and FP
6.7 COMPARISON OF OBSERVED AND PREDICTED SETTLEMENT

TROUGHS
6.7.1 Transverse Settlement Trough
In this section, the transverse settlement troughs predicted using 2D and 3D FE analyses
are compared (for both depth and width) with those observed in the field.
Depth of Settlement Trough
The field settlement troughs are compared with those predicted using North and South
profiles in Figure 6-27 and Figure 6-28 respectively. These figures indicate that the
North profile predicted a deeper settlement trough than that in the South profile.
The field volume losses (volume loss calculated from field settlement trough, which
was set equal to the input value of contraction for the numerical analyses) of the
analyses and the corresponding output volume losses (volume loss calculated from the
predicted settlement trough) are shown in Table 6-9, where different input and output
161
volume losses are apparent although these losses were expected to be approximately the
same (for low volumetric strains in the soil). The output volume losses were much
lower for the south profile for both 2D and 3D analyses (with all the soil models). In the
north, the tunnel passed primarily through sand with a relatively low K0 (earth pressure
coefficient at rest) value (K0=0.5). In the South, however, the tunnel passed through
overconsolidated sand and clay in a higher K0 regime (K0=1). The influence of K0 for
such a prediction is discussed separately later (Section 6.8.1).
Tunnel 1, North Profile (0.15%) Tunnel 1, North Profile(0.25%)
0 0
-1 -2
-2
-4
-3
Tunnel 2
Tunnel 2
-4 -6
Track 1 (0.15%)
-5 Track 5 (0.14%) Track 3 (0.25%)
2D HS -8 2D HS
-6 Tunnel 1
3D HS Tunnel 1 3D HS
2D HSSmall 2D HSSmall
-7 -10
-20 -10 0 10 20 30 -20 -10 0 10 20 30
Tunnel 2, North Profile(0.06%) Tunnel 2, North Profile, (0.19%)
0
0
-1
-1 -2 Tunnel
-3 1
-2 Tunnel
-4
Track 1 (0.06%) Tunnel 2
Track 5 (0.06%) -5
-3 Tunnel 2 2D HS
Track 3 (0.19%)
2D HS
3D HS -6 3D HS
2D HSSmall
-4 2D HSSmall
-7
-40 -30 -20 -10 0 10 20 -40 -30 -20 -10 0 10 20
Figure 6-27 Comparison between Settlement Troughs Predicted for North Profile
162
Tunnel 1, South Profile (0.15%) Tunnel 1, South Profile(0.25%)
0 0
-1 -2
-2 Tunnel 2
-3 Tunnel 2 -4 t
-4 -6
-5 Track 1 (0.15%)
Track 5 (0.14%)
-8 Track 3 (0.25%) Tunnel 1
-6 2D HS
Tunnel 1 2D HS
3D HS
3D HS
-7 2D HSSmall -10 2D HSSmall
-20 -10 0 10 20 30 -20 -10 0 10 20 30
Tunnel 2, South Profile(0.06%) Tunnel 2, South Profile, (0.19%)

0
0 -1
-2
-1
-3
-4
-5 Tunnel 2
-3 Track 1 (0.06%)
Track 5 (0.06%)
2D HS -6 Track 3 (0.19%)
Tunnel 2 2D HS
3D HS
-4 2D HSSmall -7 3D HS
2D HSSmall
-40 -30 -20 -10 0 10 20 -40 -30 -20 -10 0 10 20
Figure 6-28 Comparison between Settlement Troughs Predicted for South Profile
Input
(Field) Output volume losses (%)
volume
losses 3D HS 2D HS 2D HSSmall 3D HS 2D HS 2D HSSmall
(%) South South South North North North
0.15 0.06 0.03 0.05 0.17 0.15 0.13
0.06 -0.03 -0.04 0.00 0.02 0.02 0.02
0.25 0.15 0.10 0.16 0.29 0.27 0.25
0.19 0.10 0.08 0.12 0.18 0.17 0.16
Table 6-9 Input Volume Loss and the Out put Volume Losses
163
It is of interest to note that both 2D and 3D analyses involving low input volume losses
(<0.1%) lead to heave predictions and also that as the input volume loss increases, the
ratio of the output to input volume losses also increases (considering only tunnel 1, i.e,
only volume losses, 0.15% and 0.25%). This reflects different volumetric response in
different soils, as discussed in Section 2.2.2 (see Figure 2-3). Although the lining
stiffness was the same for both profiles (elastic shell elements with a Youngs modulus
of 35GPa, Poissons ratio of 0.15 and thickness of 0.275m), the input lining stiffness also
affects the settlement trough predicted by Plaxis and this will be discussed in Section
6.8.2. It was also seen that the settlement of the second tunnel is influenced by the first
tunnel. The interaction effect of multiple tunnels is discussed in Section 6.8.7.
Width of Settlement Profile
The trough width of the field and predicted settlement troughs (using 2D and 3D) are
compared for Tunnel 1 by normalising the settlements against maximum settlement.
Each volume loss for North and South profiles are compared separately (note that,
Track 1 and Track 5 has almost same volume loss) and are shown in Figure 6-29. It is
apparent that, the predicted troughs are wider and the trough shape is different. It is also
of importance to note that the linear regression lines between ln (Sv/Smax) and y2 (details
given in Section 5.6) indicate that none of the predictions are of Gaussian nature. It is
also worthy of note that the predicted trough using the South profile is narrower.
164
North profile, 0.15% South profile, 0.15%

-0.5 -0.5
Tunne 1 0.0
0.0
0.5 0.5 Tunnel 1
Track 1
1.0 Track 1 1.0 Track 5
Track 5 2D, HS
2D, HS 3D, HS
3D, HS 2D, HSSmall
1.5 2D, HSSmall 1.5
-10 -5 0 5 10 15 20 25 -10 -5 0 5 10 15 20 25
Distance from center of Tunnels, m Distance from centre of tunnels, m
North profile, 0.25% South profile, 0.25%

-0.5 -0.5
Tunnel 1
0.0 0.0
0.5 0.5
Tunnel 1
1.0 1.0 Track 3

Track 3 2D, HS
2D, HS 3D, HS
3D, HS 2D, HSSmall
2D, HSSmall 1.5
1.5
-10 -5 0 5 10 15 20 25
-10 -5 0 5 10 15 20 25 Distance from centre of tunnels, m
Figure 6-29 Comparison of Normalised Settlements for Tunnel 1
6.7.2 Longitudinal Settlement Trough
The predicted and the observed longitudinal settlement troughs are compared in Figure
6-30, where they are normalised by the maximum settlement. It should be noted that,
unlike the predicted results, the observed trough is influenced by strata variations along
the route; (note that at Track 3 location (point marked as B), the settlement was more
than that at Track 1 location (point marked as A), see Section 5.5.3). It can be seen from
the figure that the field settlement trough (marked as 1) stabilised around 13 m behind
165
the cutter face. This distance is 30% less than the predicted stabilisation distance of
about 19m. Therefore, as for a tunnels transverse trough, predicted longitudinal
movements tend to have a wider extent than measured movements.

-1.0
GP FP
2
GP FP
-0.5
Field Data
1
0.0
A
Position of Cutter Face (m)
0.5
36
39
1.0
B
1.5
0 10 20 30 40 50 60 70 80
Longitudinal distance along the tunnel axis, m
Figure 6-30 Predicted and Observed Normalised Longitudinal Settlement Trough
6.8 DISCUSSION
The notable features of the numerical prediction using both soil profiles include,
The difference between the input and output volume losses.
The influence of the soil profile (North & South)
The relatively wide transverse settlement trough.
The different parameters affecting the predicted pattern of settlement trough are
discussed in detail in this section. These include (i) initial stress history, (ii) effect of
166
lining stiffness, (iii) effect of soil stiffness and (iv) the interaction effect of multiple
tunnels.
This section also discuses the comparative study of stress changes in numerical analysis
and past centrifuge model tests to examine the possible reasons for the often predicted
wide and shallow settlement trough for over consolidated soil. The influence of reduced
K0 value around the tunnel is also evaluated.
6.8.1 Effect of K0 on the Predicted Settlement Trough
A deeper settlement trough was predicted using the North soil profile (compare Figure
6-27 and Figure 6-28 ). As stated previously, in the north, the tunnel passed primarily
through sand with a relatively low K0 value (K0=0.5), while in the south, the tunnel
passed through over consolidated sand and clay (K0=1). The effect of initial stress
history on the predicted settlement trough is investigated here using a parametric study.
Two combinations of stress history (Table 6-10) were used to predict the settlement of
the North and South profile using both the HS and HSSmall models. A typical volume
loss of 0.25% was selected for the analyses and the comparison is shown in Figure 6-31.
Consistently, deeper settlement troughs were produced by both profiles for K0 =0.5 (Set
2), indicating a similar influence of K0 at the tunnel to that predicted by Addenbrooke
(1996) and other researchers. The high mean effective stress p due to high K0 around
the tunnel in the South profile would have reduced the settlement. The deeper settlement
profile for lower K0 was also observed by Franzius (2003) in 3D analyses.
Set no. K0 Preconsolidation

pressure, POP (kPa)
Set 1 1 700
Set 2 0.5 0
Table 6-10 Combinations of Stress History used for Predicting Settlement Profile
167
North profile, vl=0.25 South Profile, vl=0.25
0 0
-1 -1
-2 -2
-3 -3
-4 -4
-5 -5
-6 -6
-7 -7 K0=0.5
K0=0.5
K0=1 -8 K0=1
-8
-20 -10 0 10 20 30 -20 -10 0 10 20 30
Distance fron centre of tunnels, m Distance fron centre of tunnels, m
a) HS Model b) HSSmall Model
Figure 6-31 Settlement Trough with Different K0 Value for North and South Soil
Profile
6.8.2 Effect of Lining Stiffness Analysis
It was found that the choice of lining stiffness had a significant influence on the
settlement trough predicted using the volume loss method in Plaxis (note that Plaxis
requires input of the volumetric contraction at the centre point of the tunnel). Numerical
instability arises if the lining is provided with stiffness equal to that of the soil.
The predicted settlement troughs (for a typical inputted volume loss of 0.15%) with
different lining stiffness, shown in Figure 6-32, demonstrate that, as the stiffness of the
lining decreases, the settlement increases and the output volume loss increases
(although, in reality, the settlement at the tunnel crown is relatively independent of the
lining stiffness). The predicted movement of the lining at the tunnel crown (with
different lining stiffness) shown in Table 6-11 indicate heave for the stiffer lining. The
analyses showed that heave produced by the stiff lining contributes to higher mean
effective stresses above the tunnel crown and hence lower settlements. It appears that
the stress change (and associated soil stiffness changes) around the tunnel will affect the
size of the predicted settlement trough.
168
Hence care should be taken to check that the inputted volume loss in an analysis leads to
an output volume loss consistent with expectations/requirements. Addenbrooke (1996)
also observed different responses of linings in soils with different K0 values, finding
similar behaviour for a rigid lining with a high K0 condition to a flexible lining with a
low K0 condition.
Tunnel 1, South Profile (0.15%)
0
-1
-2
-3
-4 Tunnel 1
-5
Tunnel 2 Field measurement
-6
E=1.5GPa
-7 E=2GPa
-8 E=35GPa
-20 -10 0 10 20 30 40
Figure 6-32 Variation of Surface Movements for Different Lining Stiffness
Flexible Lining Rigid Lining

(E=1.5GPa) (E=35GPa)
7.1 mm settlement 3.7mm heave
Table 6-11 Vertical Movement of Tunnel Crown
6.8.3 Effect of Stress Reduction around Tunnel

The influence of high mean effective stress p (due to high K0) on the predicted
settlement trough has already been discussed in section 6.8.1, while effects of tunnel
169
lining stiffness in the Plaxis method for modelling tunnels is described in Section 6.8.2.
Hence a parametric study was carried out that varied the K0 value around the tunnel (half
of the tunnel diameter around the tunnel) in a high K0 soil profile. The soil profile
comprising sand had the same properties as used for the sand layers of the
Perth/Spearwood Formation. The diameter and location of the tunnel remained the same
and a volume loss of 1% was adopted for the analysis. Figure 6-33 shows the soil
profile and location of tunnel. The lining was assumed to have a stiffness equivalent to
that of concrete (35GPa). Three cases were considered for the analysis as indicated in
Table 6-12. A POP (past overburden pressure) of 2800kPa was selected corresponding
to K0 = 2.
Cases K0
Case 1 K0 = 2 throughout the soil profile
Case 2 K0 = 0.5 around Tunnel and K0 =2 other
places
Case 3 K0 = 0.125 around Tunnel and K0 =0.5
other places
Table 6-12 Different Cases for the Analyses
B
A B
+11
10.35m
11m
6.9m
6.9m 10.35m
Sand
-24
A
Figure 6-33 Soil Profile and Location of Tunnel for Parametric Study
The settlement predicted using case 1 is compared with case 2 and case 3 and is
demonstrated in Figure 6-34(a) where it is seen that the stress reduction around a tunnel
has a significant influence on the magnitude of settlement predicted. The normalised
170
settlement shown in Figure 6-34(b) for the three cases indicates that a lower stress
around the tunnel can predict a narrow settlement trough also. The dependence of K0 on
the predicted settlement trough indicates that the reduction in the stress around the
tunnel (especially in overconsolidated soil) due to the tunnel excavation is not properly
modelled by the soils models employed. Hence by manipulating the at rest condition
near the tunnel to a lower stress state can lead to settlement profile predictions that are
closer to the Gaussian form. However detailed study is required to find out the exact
area to be treated with a reduced stress and the value of K0 to be assigned for the
corresponding area. Comparing the settlement trough in Figure 6-34, it is evident that
the settlement trough is narrow up to about 5 m from the Tunnel axis and it becomes
wider after that. This could be because of the low mobilised horizontal stress as
described in section 6.8.5.
Addenbrooke (1996) also found through 2D analysis that by assuming an initial reduced
stress regime around the tunnel can predict a closer match to the field data. Since
majority of the volume loss is happening due to the shield loss (radial ground loss
around the shield to fill the gap between the shield and the unsupported soil) in the
closed face tunnelling, the assumption of stress reduction around the tunnel in the plain
strain condition is justified. The transverse and longitudinal settlement troughs plotted
in Chapter 5 shows only heave ahead of the tunnel heading and the same observation
was made by Sugiyama et al.(1999). Whittle (personal communication) has conducted
analytical studies showing that the three dimensional effects are limited within a region
equal to twice the depth of tunnel axis (i.e., 2z0 where z0 = vertical distance to the
tunnel axis) and that there is little variation in the normalised transverse settlement
trough behind the tunnel heading. Whittle therefore argues that a uniform ground loss
can be assumed along the tunnel axis.
171
Maximum Settlement Normalised Settlement
5 0
0
-5
-10 0.5
-15
-20
-25 1
-30 Ko=0.5and 0.125
-35 Ko=0.5and 0.125 Ko=2 and 0.5
Ko=2 and 0.5 Full K0=2
-40 1.5
Full K0=2
-20 -10 0 10 20 -20 -10 0 10 20
Distance from centre of tunnel, m Distance from centre of tunnel, m
a) Maximum Settlement (b) Normalised Settlement
Figure 6-34 Settlement Trough with Different K0 around the Tunnel
6.8.4 Effect of Soil Stiffness on Settlement Trough

Past research shows that the settlement trough shape becomes slightly closer to the
Gaussian form when soil anisotropy is included in the analysis (Franzius, 2003, Ng and
Lee, 2005). As the Plaxis HS models do not incorporate stiffness anisotropy, analyses
were performed to check the influence of the HS Eur parameter on the predicted
settlement trough. It is seen on Figure 6-35 that a 66% increase in Eur (and hence
stiffness) induced a deeper and relatively narrow settlement profile (input volume loss
1% and K0=2 throughout the soil profile) 1. The stress pattern (for both stiffnesses)
indicated in Figure 6-36 shows a reduced mobilised horizontal earth pressure ( h)
above the tunnel crown (at about 5m RL) for the high stiffness soil. This weak area
1
Although the settlement profile with high stiffness looks deeper and narrow compared to that with low
stiffness, the output volume loss calculated from the predicted settlement profile differed being 0.36% for
low stiffness and 0.41% for high stiffness.
172
must be contributing to the higher settlement. Hence it is clear that even a higher
vertical stiffness can induce a deeper and narrow settlement trough; the same effect was
observed by Franzius (2003) and Ng & Lee (2005) when they increased the horizontal
stiffness above that of the vertical stiffness.
1 0
-1
-3
-5 0.5
-7 Eur=3*E50
-9 Eur=5*E50 Eur=3*E50
1 Eur=5*E50
-11
-13
-15 1.5
-30 -20 -10 0 10 20 30 -30 -20 -10 0 10 20 30
Distance from centre of tunnel, m Distance from centre of tunnel, m
Figure 6-35 Settlement Profile with Different Stiffness
Reduced stress
a) Eur=5*E50 b) Eur=3*E50
Figure 6-36 Mobilised Horizontal Earth Pressure
173
6.8.5 Stress Changes above the Tunnel during Numerical Analyses
The stress pattern above the tunnel crown as measured by others in centrifuge tests in
sand as well as that predicted by numerical modelling is discussed in this section. The
following sections examine the reasons for the shallow and wide settlement troughs
often produced by numerical analyses in over consolidated soil.
a) Magnitude of settlement
The stress pattern along the axis of the tunnel for the case of uniform K0 of 2 (see
Section 6.8.3) is analysed in this section. Volume losses of 0.75% and 1% are
considered for the analysis. Predictions for both volume losses show that the mobilised
earth pressure coefficient Km (= 'h/ 'v) is highest at the tunnel crown (indicating arching
of the horizontal stresses) and then decreases above that before increasing towards the
ground surface (Figure 6-37). The high Km at the tunnel crown (which is about twice
K0) and the subsequent reduction was also observed by other researchers in centrifuge
modelling (see Figure 2-4) of tunnels in sand (Jacobsz, 2002 and Marshall, 2009).
Jacobsz (2002) noticed a Km value of as much as 9.8 near the tunnel crown for a volume
loss of 2% and it sharply reduced with increasing distance from the tunnel crown; the
Plaxis predictions also show the same trend for an increase in Km near the tunnel crown
(Figure 6-37). However the increase in Km value close to the surface is not recorded by
Jacobsz (2002) (for comparison) as mentioned in Section 2.2.2 and hence it is not
possible to comment on the predicted higher Km value near the surface as seen in Figure
6-37. In the numerical analysis, if the K0 value around the tunnel is higher than that in
the field, the Km value also tends to be higher and will contribute to a lower level of
settlement. The higher predicted Km values above the tunnel may there fore explain the
lower level of settlements predicted compared to those observed in the field. (note that
the higher K0 at the tunnel location predicted a lower settlement , see Section 6.8.1)
174
Initial (K 0 ) and mobilised lateral earth pressure coefficient (K m )

12
10
6
Tunnel crown
0.75%
2 1%
K0
0
0 1 2 3 4 5
Initial and mobilised lateral earth pressure coefficient
Figure 6-37 Mobilised Lateral Earth Pressure Coefficient along the Tunnel axis
b) Width of Settlement Trough
The surface settlement troughs predicted by numerical methods are often much wider
than those observed in the field. The mobilised stresses and Km values close to the
surface (Section B-B in Figure 6-33, 0.7m from ground level) plotted in Figure 6-38
show that, although the mobilised vertical effective stress, y(m), remains relatively
constant throughout the length of the geometry, there is a significant reduction in
mobilised horizontal effective stress h(m) (and hence Km) as the distance from the
tunnel axis reduces. These low h(m) and Km values may be contributing to a wider
trough in numerical analysis.
175
Mobilised verical and horizontal Mobilised Earth Pressure Coefficient
effective stress (RL=10.3m)
0 0
0.5
10
1
20 1.5
2
30
2.5
40 3
3.5
50 4
horizontal effe. Stress
60 4.5
vertical effe. Stress
-50 -30 -10 10 30 50 -50 -30 -10 10 30 50
Distance from centre of tunnel, m Distance from centre of tunnel , m
a) Vertical and Horizontal effective stress b) Km
Figure 6-38 Mobilised Stress and Km near the Ground Surface at RL=10.3m
(section B-B in Figure 6-33)
6.8.6 Suggested Approach for Predicting Greenfield Movements (using

Plaxis)
Based on the observations made in Sections 6.8.2and Section 6.8.3, it is suggested that
better predictions for greenfield movements using the Plaxis volume-loss method can be
obtained by modifying the initial effective stresses in the vicinity of the tunnel and
reducing the lining stiffness. After (i) imposing a stress reduction by reducing K0 to a
value 0.5 in the area around the tunnel extending to half of the tunnel diameter around
the tunnel and (ii) employing a more flexible lining (E= 10GPa), a 2D FE analysis was
carried out using an input volume loss of 0.15% for the South profile (where tunnel
passed trough the overconsolidated Perth Formation). The settlement trough predicted
using this approach for Track 1 (which is close to South profile) is compared with the
measured trough in Figure 6-39. The good agreement of the measured and predicted
maximum settlement and trough width indicates that the tunnel excavation in
overconsolidated soil can be modelled in 2D FE analysis using these simple
modifications.
176
0 0.0
-1
-2 0.5
-3
-4 1.0
Tunnel 1
Field Field Tunnel 1
-5
Plaxis Plaxis
-6
1.5
-30 -10 10 30
-25 -15 -5 5 15 25
Distance from centre of tunels, m Distance from centre of tunnels, m
Figure 6-39 Comparison of Predicted and Measured Settlement Trough
6.8.7 Interaction Effect of Tunnels
The asymmetric nature of the settlement trough for Tunnel 2, discussed in Section 5.5.3,
appeared to have arisen due to the effect that the lining of Tunnel 1 had on the Tunnel 2
settlement. This is verified through the numerical analyses summarised in Figure 6-40
and Figure 6-41, which show the interaction effect of tunnels using the South and North
soil profiles (see Section 5.2); the tunnel lining parameters were same for both analyses
and the an input volume loss of 0.25% was applied for both tunnels. These figures show
significant differences between the predictions for the settlement trough of the second
tunnel. For the South profile, a deeper and (slightly) wider settlement trough for Tunnel
2 is predicted while, for the North profile, the predicted trough for Tunnel 2 is
shallower; normalised transverse settlement troughs are a little wider for the South
profile compared to the North profile (due to the presence of clay in the former profile).
The different pattern of predicted settlement for the second tunnel boring shows the
influence of soil type or stiffness. Contrary to the observation of past researchers, the
second tunnel may induce a lower and narrower settlement profile in sandy soils such as
those in the North profile. The results presented in Chapter 5 also show that the
settlement trough associated with the second tunnel had a lower maximum settlement
and a reduced width (on the side of the first tunnel). However more parametric studies
177
are needed to draw firm conclusions regarding the combined effect of soil type, stiffness
and lining stiffness.
South Profile South Profile

-0.2
0.5
0
0
-0.5 0.2
-1 0.4
-1.5 0.6
-2
0.8 Tunnel 1
-2.5 Tunnel 2
-3
1
Tunnel 1
-3.5 1.2
Tunnel 2
-30 -20 -10 0 10 20 30 -30 -20 -10 0 10 20 30
Distance from centre of tunnels, m Distance from center of tunnel, m
Figure 6-40 Interaction Effect of Tunnels using South Profile
1 North Profile North Profile

-0.2
0
0
-1
0.2
-2 0.4
-3 0.6
-4 0.8
Tunnel 1
-5 Tunnel 1 1 Tunnel 2
Tunnel 2
-6 1.2
-30 -20 -10 0 10 20 30 -30 -20 -10 0 10 20 30
Distance from centre of tunnels
Figure 6-41 Interaction Effect of Tunnels using North Profile
178
6.9 SUMMARY
The numerical predictions (using Plaxis program) of the greenfield movements caused
by two bored tunnels are presented and discussed in this Chapter. The field measured
surface settlement troughs were compared with those predicted using 2D and 3D FE
analyses at three locations close to each other. Three soil models were employed in 2D
analyses while one model was used in 3D analysis. Although the soil properties were
varied at these locations, two typical soil profiles were chosen for the numerical
analyses. The main observations from the analyses and the comparison of the field
measurement with those predictions are;
1. Although there was a minimal consolidation settlement, predominantly

immediate settlement was predicted by all the FE analyses with different models
based on both soil profiles. In fact, heave occurred during consolidation due to
the over consolidated nature of the deposit. However measurements reported in
Chapter 5 shows only immediate settlement. This could have been because of
the mixed nature of the soil, the negative excess pore pressure formed would
have dissipated fast.
2. Although the prediction by 2D and 3D analyses and using different models

showed different widths and depths of the settlement troughs, the normalised
settlement showed only slight variations in the trough width for all the
predictions.
3. The transverse settlement trough predicted using the simple Mohr-Coulomb

model was similar to that obtained using more sophisticated constitutive soil
models. This indicated that the predicted surface movements were not very
sensitive to the soil models and points to the necessity of a detailed evaluation of
the influence of tunnel excavation modelling on the predicted pattern of
settlement trough.
4. All the predictions for the south profile (high K0 value) showed consistently
smaller settlement compared to the north profile (low K0 value) indicating the
importance of the K0 value at the tunnel location. Narrower and deeper
settlement troughs were predicted when the mean effective stress p at the tunnel
179
location was lower. The dependence of the K0 value (and thus p') on the
settlement trough suggests that the reduction in the stress around the tunnel
(especially in overconsolidated soil) due to the tunnel excavation is not properly
captured by the volume loss method and/or the soil models employed.
5. It was found that the stiffness of the tunnel lining had an impact on the
maximum settlement predicted when the standard Plaxis volume loss approach
to predict ground movements is employed. Lower maximum settlements were
predicted using a rigid lining, while higher settlements were predicted using a
flexible lining. This is because the deformed shape of a more flexible lining
subjected to contraction within the soil gave rise to higher settlement. However
more detailed parametric studies are needed to comment on the general
behaviour of lining in different soil conditions and depth of tunnels. (Note that
in different soil conditions and with different depth to diameter ratio of the
tunnel, the shape of the distorted lining may be different).
6. Since the contraction (volume loss) is applied at the centre of the tunnel after
lining the excavation in the Plaxis analysis, the tunnel lining distortion has a
significant influence in modifying the stress regime around the tunnel and hence
the predicted settlement trough. Hence care should be taken to assign the
stiffness to the lining (in Plaxis analysis) to deform it in a given form.
7. The influence of initial K0 value and lining stiffness on the predicted settlement
trough indicates that the effective stresses generated in the vicinity of a tunnel
have a significant influence on the surface settlement trough.
8. A comparison of the findings from Plaxis parametric studies with centrifuge

model tests showed the importance of the stress changes, which for a given
volume loss, control the soil stiffness and the nature of the surface settlement
trough. Hence a surface settlement trough of Gaussian form can be achieved (in
numerical back analysis) by manipulating the at rest condition near the tunnel to
a lower stress state (to compensate for reduced stresses in the vicinity of the
tunnel due to excavation) combined with a reduced lining stiffness. The reduced
stress around the tunnel and the lining stiffness should be selected in a way to
180
get the output volume loss closer to the input volume loss (when ground losses
are assumed to be approximately equal to volume losses).
9. The numerical modelling of the tunnel excavation using both 2D and 3D Plaxis
program in this study demonstrates that neither the 3D modelling nor the
sophisticated soil models (i.e. HS and HSSmall) significantly improve the
predicted greenfield settlement trough. Less time and resource consuming 2D
Plaxis analysis can predict a settlement trough closer to the Gaussian form with
modifications to the volume loss method in Plaxis.
10. The predicted longitudinal surface settlement profile using the Plaxis 3D
program was wide relative to the measured longitudinal trough. However with
considerable effort and time, a stabilised settlement trough could be produced
from the start of the geometry.
11. The Plaxis 3D (Tunnel) analyses did not predict the heave observed ahead of
the tunnels in the field. The Plaxis analyses also predicted no relationship
between the predicted maximum settlement and the grout pressure and/or the
face pressure. This effect needs to be investigated by the Plaxis developers.
12. The analyses of the interaction effect of multiple tunnels indicate lower
settlements for a second tunnel located adjacent to the first tunnel in normally
consolidated sandy conditions and higher settlement where overconsolidated
clay layers are present. Also the settlement trough was narrow (on the side of
the first tunnel) in sandy soil and wide in clayey soil. This indicates the
influence of soil type or stiffness at the tunnel location on the predicted
settlement trough. However more study needed to comment on the effect of
lining stiffness.
181
182
CHAPTER 7 MALAYSIAN AIRLINES
BUILDING- SURFACE AND BUILDING
MOVEMENT
7.1 INTRODUCTION
The movement of a multistorey framed structure due to the boring of two adjacent
tunnels is analysed and discussed in this chapter. The effects of tunnel construction on
the building and of building stiffness on the observed pattern of surface and building
movement are analysed and discussed. The measured building movement is then
compared with predictions made using 2D and 3D FE analyses. The effect of building
stiffness relative to soil stiffness and soil type on the settlement profile is also evaluated.
7.2 BUILDING DETAILS
The building under consideration is the Malaysian Airlines building (MAS), which was
constructed in 1929. This building is located on William Street (in the southern bored
tunnel section) and has seven storeys and one basement. Figure 7-1 presents an aerial
view of the building and Figure 7-2 shows its location relative to the tunnels. The south
side of the building is seen to be 6.2 m from the axis of Tunnel 1 and 16.4 m from the
centre line of Tunnel 2. The spring lines of the tunnels are at about 18 m below ground
level at this location (=R.L. of -3.6 m). A single storey light structure (Qantas travel)
with a basement and Central Park Garden are situated on the west side of William
Street opposite the MAS building; this building is ~ 6.6 m from the vertical axis of
Tunnel 2. Front views of the MAS building and the building opposite the MAS are
shown in Figure 7-3 and Figure 7-4 respectively. The ground and first floor plan of the
MAS building is given in Figure 7-5. The building is a steel framed structure with
concrete slab floors finished with timber parquetry. The height of the building is about
27m. Unfortunately, plans for the foundations of the building could not be found.
However it is known that the building was on either pad or strip footings.
183
N
Figure 7-1 Aerial View of MAS Building
184
Chapter 7 Malaysian Airlines Building- Surface and Building Movement
11.2 m
MAS
South North
~18.1m
TBM
Tunnel
a) Longitudinal Section
MAS
~6.6m 6.2m 18.5m

Quantas
+ 14.5m AHD
~18.1m
6.9m 6.9m -0.15
Tunnel 1
Tunnel 2 -7.10
10.2m
b) Transverse Section at south side of the Building
Figure 7-2 Schematic View of the Building and Location of Tunnels
185
Figure 7-3 Front View of MAS Building
Figure 7-4 View Opposite MAS Building
186
Figure 7-5 Ground and First Floor Plan for MAS Building
7.3 GROUND CONDITIONS
In the southern bored tunnel section (south of William Street Station), the tunnels
passed mainly through the mixed soil profile of over consolidated Perth Formation. The
soil profile at this location was arrived at based on bore holes (BH-2178 and BH-07)
CPTs (CPT -21) and DMTs (D07) close to the MAS building (Figure 7-6). The
borehole information and stratigraphical profiling close to MAS are shown in Figure
7-7. The stratigraphy generally comprises 9m of dune sand (Spearwood sand) overlying
the interbedded layers of alluvial silts, clays and sands of Perth Formation, which are
underlain by Kings Park shale/ siltstone.
The CPT end resistance (qc) and friction ratio (Fr) for CPT-21 are plotted in Figure 7-8.
The soil profile based on the above details is shown in Figure 7-9, which also indicates
the best estimate average qc value for each sub layer. Figure 7-9 indicates that both
tunnels are located in sandy and clayey layers of the Perth Formation. The (perched)
water table in the Spearwood dune sand was observed to be at +8.6 mAHD and another
187
water table was encountered at -2.3m AHD in the Perth Formation (Geotechnical and
Hydro geological report (2003) by Golder Associates).
Figure 7-6 Geotechnical Investigation Locations near MAS Building
188
CPT-21, Ic DMT-07, ID BH-07
BH-2178 0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.6 1.2 1.8 2.4 3 3.6 4.2
15 Clayey 15 14.9
14.4 Clean Silty 146 Fill
13.9 Silt Silty Silt Sand
14 Gravelly sand to Sand to Sand to 14 Clay
to Clay
13 Dense Sand Silty Sandy 13
Silty to
12 Sand Silt
Clay Clay 12
Sand 11 Sand
11
Clean Sand to Silty Sand
10 10
Sand
9 9
8 8
6.1 7 7
Silty Sand/
5.0 6 6
Clayey Silt 5.5
5 5 Clay
4.1
Silty Clay 4 Clayey Silt to Silty Clay 4
3 3 Silt with clay and
RL(m)
2
RL(m)
1.1 Clayey Sand / 2
Clayey sand
Silty Sand 1 1
-0.7 0 0
Silty Sand /
- Silty Clay -1 -1 Silt with sand
-2 -2
-3 Silty Sand to Sandy
Sand/Sandy Silt -3
Silt
Building
-4 -4 -4
-5 Silty Clay to Clay -5 Clay
-6.2 Clay
Sandy Silt/ -6 -6
Silty Sand with
Lenses of Clay -7 Clean Sand to Silty Sand -7
-8.1
-8 -8 Sand with -7.8
-9 Clean Sand to Silty
Sand -9 Sand
-10 -10
-11.1 -10.2 Clay
Clay/Silty clay -11.2
Interbedded
-12.4
Layers of
Sand &Clay
-13.2
Sand/Gravelly Sand
Sand/
- Gravelly Sand
Silt Stone
-18
Sand
-19.5
Silt Stone
189
Building
CPT-21, qc (MPa) Friction Ratio, Fr (%)

0 10 20 30 40 0 2 4 6 8
15 15
10 10
5 5
0 0
-5 -5
-10 -10
Figure 7-8 CPT Profile
+14
qc=8 MPa Sand
+10.0 Spearwood Sand
+9.0
qc= 15 MPa Sand
+5.5
Tunnel 1 Clayey Silt to Silty
qc= 5 MPa
Tunnel 2 Clay
10.2m
+1.
qc=10 MPa Sand to Silty Sand
-2.0
-2.3 qc=25 MPa sand to Silty sand
-4.0 -3.6m
qc =3.5 MPa Silty Clay toClay
-6.0 6.9m
qc = 20 Mpa 6.9 Silty Sand to
Perth Formation
Sandy
-11.0
-13.0 Clay/Silty Clay
Dense Sand or
Gravel
-20.0
Figure 7-9 Soil Profile and Location of Tunnel near MAS Building
190
7.4 INSTRUMENTATION
The Malaysian Airlines Building (MAS) was between the chainages of about 240m and
260m south of William Street Station. The relevant section of Tunnel 1 was bored
between 12th January 2006 and 14th January 2006 while that for Tunnel 2 was bored
between 15th August 2006 and 16th August 2006. At the request of the writer, LKJV
monitored movement of the MAS building using EL beams. LKJV also installed
Building Settlement Points and monitored the ground surface using Surface Settlement
Pins (SSP).
7.4.1 Surface Settlement Pins and Building Settlement Points
Surface Settlement Pins (SSP) and Building Settlement points (BSP) were survey points
located either on the ground (SSP) or on the building (BSP) to enable monitoring of the
movements during TBM tunnelling. The SSPs were standard survey nails while the
BSPs took the form of threaded bolts grouted into the building wall with a hexagonal
nut protruding approximately 20-25 mm out from the wall. Both SSPs and BSPs were
measured by digital levelling using Leica DNA-10 Digital Level and Barcode Staff.
Both SSP and BSP results showed the same degree of repeatability of 1.5mm,
(McGough and Williams, 2007).
7.4.2 EL Beams
EL beams were installed on the buildings south wall at about 18 m RL. Full details
concerning the type of EL beams employed and their measuring technique are explained
in Section 5.5.2. At this location, the data were radio transmitted to the project office on
William Street every 15 minutes and uploaded to a computer. A photo showing the EL
beams on the MAS wall with transmitter on each beam is provided in Figure 7-10.
191
Figure 7-10 EL Beams on the Wall with Transmitter on Each Beam
7.5 FIELD MEASUREMENT (SSP AND BSP)
The settlements measured using SSPs (RL ~ 14.5) and BSPs (located at RL values
between 15.5m and 18m) are shown in Figure 7-11, which also indicates the extent of
the MAS building and the front of the Qantas building (Note: no discernible effect of
RL was evident and therefore settlements at RL values are combined). This figure
shows the pattern of movement during and after the tunnelling activity at this location
for both tunnel drives. Since Tunnel 1 movement had stabilised shortly after the TBM
passed that location, the settlement for Tunnel 2 was calculated separately taking a base
line reading closer to the start of the Tunnel 2 boring. This was to avoid any
instrumentation error (see Section 5.4.1 which discussed drift of RSP measurements
over time). As seen in Figure 7-11, both tunnels induced heave at the centre of the
tunnel as well as on the east and west side of the tunnel axis. The heave on either side of
the tunnel axis was at the building locations. This could be because of the low
overburden pressure due to the buildings basement construction combined with the
application of high face pressure and /or the grout pressure.
192
Tunnel 1 Tunnel 2
4 Extent of MAS
4
3 Qantas building
2 Extent of MAS 2
Qantas
building
1 0
0 -2
-1 -4
-2 Y
-6
-3
Tunnel 2 -8
-4 Tunnel 1
-5 -10
-30 -20 -10 0 10 20 30 40 -30 -20 -10 0 10 20 30 40
Figure 7-11 Ground and Building Movement due to Tunnel 1 and Tunnel 2
7.5.1 Final Transverse Settlement Troughs
The best estimate final transverse settlement troughs are shown in Figure 7-12 and
volume losses deduced from the areas of these troughs are summarised in Table 7-1. It
is apparent that maximum settlement and volume loss experienced during boring of
Tunnel 1 are about half of those induced during boring of Tunnel 2. The volume loss
associated with Tunnel 1 is comparable to that seen at greenfield locations (see Chapter
5) despite the presence of 4.5m thick clay layer above the tunnel crown at the MAS
location. Hence it is apparent that the lower volume loss associated with Tunnel 1
boring at the MAS location is because of the presence of the building and the higher
volume loss associated with tunnel 2 is due to the presence of clay layer and the absence
of building stiffness. Similar observations were made at the Walsh building (Chapter 8).
Although the tunnels were located in the clay layers (3.5m thick) of Perth formation, the
volume loss associated with both tunnels (0.12% and 0.14%) at Walsh location are
comparable to that due to Tunnel 1 boring at MAS location and the greenfield location.
This is indicating that the structural stiffness can reduce the volume loss associated with
tunnel boring. Soil structure interaction effects are discussed in Section 7.7.
193
2 Qantas Extent of MAS building
0
-2
-4
-6
Tunnel 1
Tunnel 2
-8
-10
-30 -20 -10 0 10 20 30 40
Figure 7-12 Transverse Surface Settlement Trough for Tunnel 1 and Tunnel 2
Max. Settlement (mm) Volume loss (%)
Tunnel 1 -3.4 0.14

Tunnel 2 -6.7 0.29
Table 7-1 Surface Settlement Details at MAS Building
7.6 BUILDING MOVEMENT (EL BEAMS)
The transverse movements (relative to the fixed end of the EL beam) along the south
side of the MAS building measured at various times during boring of Tunnel 1 and
Tunnel 2 are shown in Figure 7-13. The figure shows data recorded at various distances
between the cutter face and the MAS building (negative distances are those before the
TBM face has reached the building) and also shows data recorded at various time
periods after the TBM has moved well past the building. Although the measurements
are relative, the settlement pattern shows that the settlement stabilised as soon as the
tunnelling activity was finished at that location for both tunnel drives.
194
Tunnel 1, MAS building Tunnel 2, MAS building

0.6 0.6 Extent of MAS
Tunnel 2
0.4 0.4 building
Extent of MAS building 0.2
0.2
0 0
-0.2 Tunnel 1 -0.2 Tunnel 1 -42.2 m
-0.4 -24.2 m
-0.4 7.0 m
-0.6 -49.5 m
-0.6 44.0 m
-9.9 m -0.8 Tunnel 2 60.0 m
-0.8 4.5 m 5 days
17.5 m -1 10 days
-1
32.5 m -1.2 15 days
-1.2 14 days 27 days
22 days -1.4
-1.4
32 days -10 -5 0 5 10 15 20 25 30 35
-10 -5 0 5 10 15 20 25 30 35
Figure 7-13 Movement of Building at Various Times during Tunnelling
The absolute settlement values were obtained by adding the settlement deduced from
the BSP at the anchor point of the string of EL beams. The final settlement profile of the
building at ~18m RL (~3.5m above ground level) is plotted in Figure 7-14. It is apparent
that during boring of both tunnels, heave was induced at far end of the building (but
only a fraction of a millimetre). The kink seen at the far end for Tunnel 2 is within the
accuracy limit of the EL beams and is probably an instrumentation error. The measured
movement will be compared with both 2D and 3D FE analyses and discussed later in
this Chapter.
The settlement pattern of the building in Figure 7-14 indicates that the building tilted
uniformly and hence behaved like a rigid plate; this is similar to the observations made
by Breth and Chambosse (1974), when they monitored movement of buildings in
Frankfurt clay. The length to height ratio of MAS building is about 0.7.
195
Movement of MAS building
0.6
0.4 MAS
0.2
0
-0.2
-0.4 Tunnel 1
-0.6
Tunnel 1
-0.8
Tunnel 2
-1
Tunnel 2
-1.2
-1.4
-10 -5 0 5 10 15 20 25 30 35
Figure 7-14 Transverse Building Movement for Tunnel 1 and Tunnel 2
7.7 EFFECT OF SOIL STRUCTURE INTERACTION
As presented at the greenfield site (railway tracks), the MAS data are plotted in Figure
7-15 in the form of Ln (Sv/Smax) vs y2. A straight line relationship of the data when
plotted in this format indicates that the settlement trough is of a Gaussian form which
is the form observed in greenfield conditions (see Chapter 5). Figure 7-15 indicates the
settlement troughs for both tunnels are only of a Gaussian nature to the west of the
tunnel axes. This clearly demonstrates the influence of the MAS building on the
settlement trough. However, the low rise Qantas building 6.6m west of the Tunnel 2
axis did not alter the Gaussian nature of the settlement trough. The high volume loss
observed for Tunnel 2 and the Gaussian nature on the west of the tunnel axis indicates
the absence of structural influence. This is likely to be because of the relatively low
stiffness of the Qantas building. The sharp change in slopes shown on Figure 7-15 for
the east side of the tunnels occurs due to the presence (i.e. stiffness) of the MAS
building.
196
Tunnel 1 Tunnel 2
0.0 0.0
-0.5 -0.5
y = -0.0097x y = -0.0138x
2 West -1.0
-1.0 R = 0.996 2
-1.5 R = 0.9924
-1.5 East
-2.0 West
-2.0
-2.5 East
-2.5 -3.0
-3.0 y = -0.005x
y = -0.0138x -3.5 2
-3.5 R = 0.947
2
R = 0.93 -4.0
-4.0 -4.5
-4.5 -5.0
0 100 200 300 400 0 200 400 600 800 1000
2 2
y y
Figure 7-15 Linear Regression Analysis for Tunnel 1 and Tunnel 2
Normalised settlements are plotted in Figure 7-16 and, as expected from the preceding
plots, these indicate marked changes in curvature of the settlement troughs in the
vicinity of the MAS. Also the reduced trough width on east side of the tunnel axis for
Tunnel 1 boring indicates the presence of structural influence as seen in Walsh location
(Section 8.5.1).
Normalised Settlement
-0.5
0.0
0.5
Tunnel 1
1.0 Tunnel 2
1.5
-30 -20 -10 0 10 20 30 40
Figure 7-16 Normalised Settlement for Tunnel 1 and Tunnel 2
197
7.8 NUMERICAL ANALYSIS
Although considerable efforts have been made by different researchers in the past to
estimate the greenfield movement, the response of nearby structures due to tunnel
excavation has been studied less frequently. This is because of the difficulty in
modelling the structure and the complexity involved in assessing the soil structure
interaction effect. Furthermore, the settlement troughs predicted by numerical methods
are often wide compared to the field measurements. However, an attempt at predicting
the surface as well as the building movement (MAS building) using both the 2D and 3D
Plaxis program has been carried out in this section. The input soil parameters to model
the soil profile shown in Figure 7-9 are described in Section 6.3. Since the exact
building details were unavailable, input parameters for the building were selected based
on assumed dimensions for the building components. Appropriate dimensions were
assumed (based on engineering judgement) for modelling the building components such
as beams, columns and footing and are given below.
Columns - 450 mm 450mm
Beams - 300 mm 450 mm
Foundation - 1.2m thick Raft
Plate elements were used to model the building components and the input parameters
such as axial stiffness (EA) and bending stiffness (EI) for beams, columns and footing
are shown in Table 7-2. The modulus of elasticity (E) of concrete was assumed to be
20GPa. The tunnel lining was modelled using elastic shell elements with a Youngs
modulus of 35GPa, Poissons ratio of 0.15 and thickness of 0.275m.
EI (kNm2/m) EA (kN/m)
Beam 47840 2835000
Column 71760 4252500
Footing 3024000 25200000
Tunnel lining 60600 9600000
Table 7-2 Input Parameters for the Building Components
198
A dead load (DL) of 5 kPa and live load (LL) of 3 kPa were assumed for load take
down calculations:
Total DL + LL for each floor = 5+3 = 8 kPa
DL from raft + LL from Basement = 1.224+3 = 32 kPa
7.8.1 2D FE Analysis
2D FE analyses were carried out for the single tunnel drive as well as for the subsequent
second tunnel drive; the more advanced HSSmall constitutive soil model was used for
the analyses. The geometry and the modelling of the tunnel excavation are explained
below.
Geometry
The two tunnels of 6.9 m diameter were bored at an elevation of ~3.6 m AHD (centre of
tunnel) at this location. The centre to centre distance between the tunnels was 10.2m.
The tunnel crown and invert were at -0.15 mAHD and -7.1 mAHD respectively (Figure
7-2). Figure 7-17 shows the finite element mesh with tunnels and building. The length
of the model adopted was 100m to minimise boundary effects. Both vertical boundaries
were restrained against horizontal movement and rotation. The bottom boundary was
restrained against vertical and horizontal movement. A no flow and no consolidation
boundary were specified at the base to represent impermeable rock.
Figure 7-17 2D Mesh Showing Tunnels and Building
199
As in the greenfield site, the tunnel excavation was modelled using the volume loss
method and this involved different calculation steps. In the first calculation step, the
building was activated and the in situ soil was consolidated to a minimum excess pore
pressure of 1kPa to simulate the initial field condition. After the activation of the
building the tunnel excavation was modelled as explained in Section 6.4.
7.8.2 Transverse Surface Settlement Trough
The observed and predicted settlement troughs for Tunnel 1 and Tunnel 2 boring shown
in Figure 7-18 indicate that the settlement troughs predicted using (what is believed to
be) a realistic soil model are wider and shallower than the observed settlement troughs
for both tunnels. The volume losses inputted to Plaxis were varied to ensure that the
volumes of the predicted greenfield settlement troughs were the same as the measured
values indicated inTable 7-1. The method used in Chapter 6 to produce a Gaussian form
(i.e. reduced initial stresses and reduced lining stiffness) was not as successful here.
Additional analyses indicated that the relatively high depth or depth to diameter ratio of
the tunnel at the MAS led to a different pattern of movements (e.g. predicted wider
settlement trough) compared to those seen around the shallower tunnels in the
greenfield (railway track) area. More study is therefore needed to generalise the
procedure recommended in Chapter 6. The effect of building stiffness on the settlement
trough will be discussed later in Section 7.10.1.
200
Plaxis 2D, Tunnel 1 Plaxis 2D, Tunnel 2

0 0
-2 -2
Extent of MAS Extent of MAS
building building
-4 -4
Tunnel 1
Tunnel 1
-6 -6 Tunnel 2
Tunnel 2 Plaxis
-8 -8 Plaxis
Field
Field
-10 -10
-40 -30 -20 -10 0 10 20 30 40 50 -40 -30 -20 -10 0 10 20 30 40 50
a) Tunnel 1 b) Tunnel 2
Figure 7-18 Observed and Predicted Settlement Trough
7.8.3 Building Settlement
The observed and predicted building movement (RL ~ 18 m) in Figure 7-19 shows that
the numerical analyses predicted greater settlement than observed in the field for both
tunnel drives. The numerical modelling did not reproduce the (apparent) uplift that was
observed on the far side of the building. The predicted tilt of the raft due to construction
of Tunnel 1 is similar to that measured. However a hogging mode of deformation was
predicted for Tunnel 2 and also the (apparent) curvature change measured at a distance
of 20m is not predicted.
It is also worth noting that the predicted transverse settlement at the footing level was
identical to that at a height of 8m above the footing.
201
1 1 Extent of
Extent of MAS MAS
0.5 0.5
0 0
-0.5 -0.5
-1 -1
Field Field
-1.5 -1.5 Plaxis
Plaxis
-2 -2
0 4 8 12 16 20 24 28 32 0 4 8 12 16 20 24 28 32
(a) Tunnel 1 (b) Tunnel 2
Figure 7-19 Measured vs Predicted Settlement of the Building
7.9 3D ANALYSIS
Three dimensional FE analyses were also carried out using Plaxis 3D Tunnel version
2.4 to study the effect of tunnelling on buildings in a more realistic three dimensional
manner. The details of the program and the modelling technique are given in detail in
Section 6.6. A total of 20 staged construction phases were involved for excavating
Tunnel 1 alone (requiring 15 hours computational time). As explained in 2D analyses,
here also the building components and footing was modelled using plate elements of
appropriate stiffness as mentioned in Table 7-2. Figure 7-20 shows the geometry of the
tunnels and the building. The length of the model selected along the longitudinal axis of
the tunnel was 82m (z-direction) and the width of the model was 75m. The approximate
size of the building was 11.2m along the longitudinal direction of the tunnel and 18.5 m
across the tunnel. The near side of the tunnel was 6.2 m from the centre of Tunnel 1 and
both tunnels were at 10.2 m apart centre to centre (Figure 7-2).
202
Figure 7-20 3D Mesh for the Geometry
Tunnel 1 Tunnel 2
Face Pressure (kN/m2/m) 186 184
Grout Pressure (kN/m2/m) 212 183
TBM Thrust (kN/m2) 566 598
Table 7-3 Face Pressure, Grout Pressure and TBM Thrust used for 3D Analyses
The input parameters such as grout pressure, face pressure and TBM thrust are shown in
Table 7-3 and the volume loss was selected from Table 7-1. The deformed mesh after
the tunnel drives are shown in Figure 7-21.
a) after Tunnel 1 a) after both Tunnel 1 and Tunnel 2
Figure 7-21 Deformed Mesh after Tunnel Boring
203
7.9.1 Transverse Movement of the Building
The predicted transverse movement of the building as the TBM passes is shown in
Figure 7-22 (before consolidation) for Tunnel 1. The distance of the TBM cutter face
with respect to the north faade of the building is also shown in the legends. The
negative distance indicates that the cutter face is under the building. This is shown
schematically in Figure 7-23. Figure 7-22 indicates that as soon as the cutter face has
passed the building, predicted settlements were virtually fully stabilised.
Building Movement due to Tunnel 1

0
-0.2
-0.4 Tunnel 1 Extent MAS
-0.6 building
-0.8
-1 Distance of cutter face
-1.2 from the building (m)
-1.4 -6
-1.6 6
12
-1.8 20.4
-2
0 10 20 30 40
Figure 7-22 Predicted Movement of the MAS Building as the TBM Passes
11.2 m
Building
South facade North facade
(-) (+)
TBM
Tunnel
Figure 7-23 Schematic View of Building and Direction of TBM
204
The measured transverse movements of the building are compared with those predicted
using 2D and 3D FE analyses in Figure 7-24. It is evident that, as for 2D analyses, 3D
analyses also predict greater settlements than observed (noting that the analyses ensured
that area of the measured and predicted surface settlement troughs were identical). This
is because of the wider surface settlement troughs predicted by the FE analyses which
actually predicted lower maximum settlements than measured (see Figure 7-19). There
is little difference between the 2D and 3D predictions for the Tunnel 1 case. However,
as seen on Figure 7-24, the 3D analysis for Tunnel 2 indicates a more rigid raft response
than the 2D analyses. This trend may reflect differences in the structural stiffness
computed by Plaxis for the 2D and 3D cases.
Building settlement for Tunnel 1 Building Settlement for Tunnel 2

0.5
1
Tunnel 1 Tunnel 2
0.5
0
0
Tunnel 1
-0.5 -0.5
-1
Field
-1 Field
Plaxis 2D -1.5 Plaxis 2D
Plaxis 3D Plaxis 3D
-2
-1.5
-1 -6 -2 2 6 10 14 18 22 26 30
0 10 20 30 40 0
Figure 7-24 Transverse Settlement Observed and Predicted by FE Analysis
7.9.2 Longitudinal Surface Movement
The predicted longitudinal surface settlements when TBM cutter face is at different
distances (z value) from the start of the geometry are given in Figure 7-25. The
longitudinal geometry (z-direction) is shown schematically in Figure 7-26. The
prediction of steady state movement from the beginning of the geometry is consistent
with that predicted in the greenfield site (Section 6.6.2).
205
Longitudinal Settlement for Tunnel 2
0
-0.5 Extent of MAS
-1 Position of TBM
-1.5 cutter face
-2 35.6m
39.2m
-2.5
47.6m
-3
51.2m
-3.5
-4
0 20 40 60 80 100
Distance from the start of geometry, m
Figure 7-25 Longitudinal Settlement Profile from 3D Analysis
11.2 m
Building
South facade North facade
TBM
Tunnel
y
z
z=30.8 m
z=82 m
Figure 7-26 Schematic View of Geometry in z Direction
7.10 DISCUSSION
The influence of building stiffness and or the soil type/stiffness on the observed pattern
of surface movement is discussed in this section. Use is made of additional parametric
206
studies undertaken to provide further insights. The effectiveness of the FE model to

predict this behaviour is also assessed.
7.10.1 Influence of Building Stiffness on Settlement Trough
The influence of building stiffness on the predicted settlement trough was investigated
using 2D FE analyses. Two analyses were carried out changing the stiffness of the
building: one with stiffness of the building components 10 times greater than those
indicated in Table 7-2 and another one with stiffness 0.1 times those indicated in Table
7-2. The surface settlement troughs (noting that at the building location, it is the
building movement) normalised by maximum settlement for these two additional
analyses along with the greenfield prediction (for Tunnel 1 only) are shown in Figure
7-27 and indicate a considerable difference for both cases. The building tilted uniformly
for the high stiffness case. This is in agreement with the observations of Breth and
Chambosse (1974), who observed rigid plate behaviour for the outside walls of many
buildings above the Frankfurt/Germany subway system; they concluded that the
buildings above tunnels behave like rigid plates as long as the length to height ratio is
less than one. The monitored buildings were either concrete framed structures or brick
building.
Son and Cording (2008) also showed the importance of considering the tilt effect for
building damage assessment of either masonry or framed structures through the
numerical study. Figure 7-27 shows a significant hogging mode for the less stiff case
and this is in agreement with Frischman et al. (1994) who state that the unreinforced
walls are much more likely to be damaged by hogging than by sagging. When the
stiffness is increased, the settlement trough is flatter and this is in agreement with Liu et
al. (2000). Hence the above mentioned cases of having different stiffness of the building
indicates that there is significant influence of building stiffness on the settlement mode
of the building.
207
Different stiffness for building
0
Extent of MAS building

0.2
0.4
0.6
0.8 Greenfield
Stiffness 0 .1 times
1 Stiffness 10 times
1.2
-10 0 10 20 30 40 50
Figure 7-27 Surface Settlement Trough with Different Building Stiffness
7.10.2 Influence of Soil Type or Stiffness on Settlement Trough
A parametric study was also carried out to study the influence of stiffness or soil
layering on the normalised settlement response; this involved changing the stiffness and
the soil while retaining the same initial building stiffness (Table 7-2). Two analyses
(2D) were carried out with a uniform soil layer: one of them with drained soil behaviour
and another with undrained soil behaviour. The E50, ref value adopted for the soil was
34MPa (half of that of the clay layers of the Perth Formation). The normalised surface
settlement troughs for the drained and undrained soils cases are plotted in Figure 7-28
for the greenfield case and for the case including the MAS building. It is evident that,
for the case examined, the assumption of drained or undrained (constant volume) soil
behaviour has no significant effect on the normalised settlement trough for the
greenfield condition. However, as seen by a comparison of the greenfield and building
case on this figure, the building stiffness has a relatively large effect on the normalised
trough shape for the drained soil effectively leading to a rigid body rotation of the
building towards the tunnel. The normalised movements for the building base in the
undrained soil falls between that of the building in the drained soil and that predicted for
the greenfield condition.
208
Different type of soil

-0.2 Extent of MAS building
0
0.2
0.4
0.6
0.8
Greenfield-Drained
1 Greenfield-Undrained
Building-Drained
1.2 Building-Undrained
-10 10 30 50
Figure 7-28 Surface Settlement Trough with Different Soil Type
7.10.3 Settlement, Bending Moment and Horizontal Strain Induced on

Building due to Different Volume Losses
A parametric study was carried out to shed light into the behaviour of buildings in Perth
due to different tunnelling induced volume losses. The variation of building settlement
due to different volume losses (0.14%, 0.5% and 1%) is provided in Figure 7-29, which,
as expected, shows greater settlements at higher volume losses. Building gradients do
not however increase in direct proportion to the volume loss. For example, the
maximum gradient for the 0.5% volume loss case is about 1/3200, but this increases
fourfold to 1/800 for a doubling of the volume loss to 1%. This trend highlights the
influence of soil stiffness non-linearity and the need for its incorporation in soil
structure interaction analyses. A predicted gradient of 1/800 is relatively small at a
volume loss of 1%, suggesting that volume losses in excess of at least 2% would have
been required to cause damage to this building.
209
Higher differential movements associated with higher volume losses will induce
damage to the existing structure. However, with closed face tunnelling using EPB-
TBM, a high degree of settlement control can be achieved and in sand it could be less
than 0.5%. In Perth, the maximum volume loss observed was 0.29% after examining 5
locations along the Tunnel route.
Building Settlement
Extent of MAS
Building
0
-5
-10
0.14%
-15 0.5%
1%
-20
10 14 18 22 26 30
Figure 7-29 Building Settlement due to Different Volume Losses
210
BM due to Tunnelling
300
250
200
150
Extent of MAS
100 building
50
-50
Volume Loss.14%
-100 Volume Loss 0.5%
Volume Loss1.0%
-150
0 10 20 30 40
Figure 7-30 Predicted BM due to Tunnelling for Different Volume Losses
The variation of bending moment corresponding to different volume losses shown in

Figure 7-30 indicates that, even with 1.0% volume loss, only minimum steel is required
to compensate the bending moment due to tunnelling. However the 1.0% volume loss
due to EPB TBM in the Perth soil is unlikely as mentioned earlier.
211
Horizontal Strain at Footing Level
0.06
0.04
0.02
-0.02
0.14%
-0.04 0.5%
1.0%
-0.06
0 5 10 15 20 25 30 35
Figure 7-31 Horizontal Strain induced by Tunnel 1 Boring on MAS
The variation of horizontal strain induced on the building with different volume losses
in Figure 7-31 shows that the maximum strain of 0.05% (The maximum value of strain
was predicted at column location) induced by 1% volume loss (Figure 7-31) is in the
category of negligible damage (after Boscardin and Cording, 1989). These predictions
are based on the stiffness of the MAS building. However with different stiffness and
dimensions of the building, the value will be different.
7.11 CONCLUSIONS
The main conclusions drawn from the measurements made in the vicinity of the
Malaysian Airlines building during the tunnelling are as follows:
As at the green-field locations, settlements stabilised within a few days of tunnel

boring.
Volume losses experienced during boring of Tunnel 1 are about half of those
induced during boring of Tunnel 2. The volume loss associated with Tunnel 1 is
comparable to that seen at green field locations although a 4.5m thick clay layer
212
was present above the tunnel at the MAS location. Hence the lower settlement
due to Tunnel 1 can be attributed to the stiffness of the nearby (MAS) building.
The stiffness of the MAS building altered the free field (Gaussian) form of the
settlement trough in its vicinity.
The building tilted due to the volume loss induced by tunnel boring. Also, even
after the tunnels had passed, one part of the MAS building had experienced
permanent heave.
A clear indication of the effects of soil structure interaction was shown by a

comparison of the settlements experienced by the (flexible) Qantas building and
(stiff) MAS building
The main conclusions drawn from the FE studies of the MAS building are;
Both 2D and 3D FE analyses, employing the non-linear hardening soil model,

failed to reproduce the observed surface settlement as well as the building
settlement. As seen by others with even more sophisticated soil constitutive
models, the predicted trough widths are substantially larger than observed.
The 3D FE analysis could, however, predict the pattern of building displacement

more accurately than the 2D analyses.
The building stiffness and soil type and/stiffness has a significant influence on
the shape of the predicted settlement trough.
Volume losses of 5 and even 10 times those incurred during the MetroRail
project are not likely to have caused significant structural damage to a building
like the MAS.
213
214
CHAPTER 8 COMPENSATION
GROUTING WALSH BUILDING
8.1 INTRODUCTION
The effect of compensation grouting in reducing the settlement induced by tunnel

boring is discussed in this Chapter. The tunnels in Perth CBD went directly underneath
some of the buildings just before entering the William Street Station and these buildings
were either on pad or strip footings. Some of these buildings were very old and even
heritage listed. Hence compensation grouting was implemented to mitigate the effects of
tunnelling induced settlements on these buildings. The ground and building movement
with and without grouting are compared and the influence of grouting on the observed
pattern of movement is discussed. The suitability of finite element (FE) program to
model compensation grouting is also evaluated. 2D FE analyses were carried out to
model the grouting and the predictions are compared with the monitored movement
after condition phase grouting as well as after active phase grouting.
8.2 BUILDING PROTECTION
Building protection was needed for some of the buildings along William Street, as these
buildings were more vulnerable to tunnelling due to their proximity to the tunnel and
also due to their shallow foundations. Figure 8-1 shows the aerial view of the buildings
and TBM path closer to the William Street Station. The Gold group buildings and the
Walsh building were more vulnerable to the tunnel construction. Hence compensation
grouting was implemented to reduce the volume loss and settlement at this location.
After studying the feasibility of different methods, hydro-fracture (hydraulic fracture)
compensation grouting was selected to protect the buildings during tunnel boring.
215
William Street
underground
station
Gold group
building
Walsh
Figure 8-1 Aerial View of the Protected Buildings
8.2.1 Compensation Grouting
Compensation grouting is achieved either by compaction grouting or hydraulic

fracture (or hydro fracture); the latter was adopted to protect the Gold group buildings
and part of the Walsh building (which is the focus of this Chapter). Hydro fracture
compensation grouting involves two phases of grouting; the first phase, referred to as
the condition phase grouting, is prior to any tunnelling and the second phase, called
active phase grouting, takes place during the tunnelling. Condition phase grouting
preconditions the soil to fill in the voids and tighten the soil to improve the efficiency of
subsequent active phase grouting. This will also reverse any settlement or loosening
caused by TAM installation. Three passes of grouting were carried out during condition
phase grouting. Active phase grouting was carried out during the tunnelling to
216
Chapter 8 Compensation Grouting Walsh Building
compensate the volume loss due to tunnelling and thus to control the building
movement. Since no significant movement was observed during Tunnel 1 boring, the
Contractor took the decision not to conduct active phase grouting (during tunnelling)
under Walsh building during Tunnel 2 boring. It is understood that this decision was not
reversed.
8.2.2 Building Details and Arrangement of Grouting
The Walsh building, which is the focus of this Chapter, is heritage listed and has five
elevated storeys and a basement. This building is a steel framed structure encased in
concrete, with secondary reinforced concrete beams. The floors are also reinforced
concrete (according to the property condition report prepared by Airey, Ryan & Hill
Consulting Engineers). The original construction date, floor plans and foundation details
of the building are unavailable. The building is about 35 m wide and 53 m in length
(parallel to the tunnel). Figure 8-2 shows the front view of the Walsh building and
Figure 8-3 indicates its location with respect to the tunnels at north and south side of the
building. The tunnel axis was about 17m below ground level at the south side and
15.7m deep on the north side. While Tunnel 1 gradually entered underneath the building
at south side of the building, Tunnel 2 was away from the building through out the
length of the building. The alignment of the tunnels near Walsh building is shown in
Figure 5-1. Figure 8-4 shows the Church opposite to south side of Walsh building.
217
Figure 8-2 Front View of Walsh Building
218
52 m
WALSH
South North
~15.7m
TBM
Tunnel
a) Longitudinal Section
WALSH
4.0m 35m
+ 14.5m AHD
~17.05m
6.9m 6.9m
+0.9
Tunnel 1
-6
Tunnel 2 10.6m
Soil
b) Transverse Section (South faade )
35m
-6m WALSH
+ 14.5m AHD
~15.45m
6.9m 6.9m
+2.3
Tunnel 1
-4.6
Tunnel 2 12.06m
Soil
b) Transverse Section (North faade )
Figure 8-3 Schematic View of the Location of Building and Tunnel

219
Figure 8-4 Church Opposite to Walsh Building
220
CPT-18
DMT-04
CPT-19
Tunnel 2
BH-17
Tunnel 1
Walsh Building
BH-07
Figure 8-5 Tunnel Alignment and Geotechnical Investigation Locations near

WALSH Building
The grout was injected through arrays of TAM pipes (tubes--manchettes) which were
installed at about 3.5m and 4m from tunnel crown and building foundations
respectively. The TAM pipes were located in the Spearwood sand and the schematic
view of this arrangement is shown in Figure 8-6. The TAM pipes were installed sub
horizontally to a maximum length of 50m from a temporary access shaft located in
221
William Street. Figure 8-7 shows the site set up and the schematic layout of the
grouting. The schematic view of the buildings, arrangements of TAM pipes, grouting
ports and locations of EL beams on the buildings are indicated in Figure 8-8. Figure 8-8
indicates that only about one quarter of the Walsh building was treated using
compensation grouting.
35m
-6m WALSH
TAM Pipes
+ 14.5m AHD
Access
Grouting
Shaft ~15.7m
~+8 6.9m 3.5m
6.9m +2.3
Tunnel 1
-4.6
Tunnel 2 12.06m
Soil
Figure 8-6 Typical Cross-section of Grouting Arrangement
Figure 8-7 Site Setup (top) and Schematic Layout for Grouting (bottom)
222
Friendlies Chemist
Gold Group Buildings
Australian Reflections
Hungry Jacks
Grout ports EL Beams
WALSH BUILDING
Figure 8-8 Schematic View of the Protected Buildings, Grout Ports and Location of
EL Beams on the Building
8.3 GROUND CONDITION
The ground profile at this site was estimated using boreholes BH-07 & BH-17, CPT 18
&19 and DMT D04; each of which was in close proximity to the Walsh building (see
Figure 8-5). Figure 8-9 shows the borehole information and stratigraphical profiling
based on the Ic index derived from the CPT data and ID index derived from dilatometer
(see Chapter 4 for more details). Figure 8-10 shows the CPT profile close to Walsh
223
building. The CPT data combined with the boreholes and dilatometer data were used to
derive the soil profile shown on Figure 8-11. The best estimate average qc value for
each layer is also shown in Figure 8-11. Figure 8-11 suggests that the tunnels are
located in silty and clayey layers of the Perth Formation. The (perched) water table in
the Spearwood dune sand was observed to be at +8.6 mAHD and another water table
was encountered at -2.3m AHD in the Perth Formation (Geotechnical and Hydro
geological report (2003) by Golder Associates).
224
RL(m)
RL(m)
Figure 8-9 Borehole Information and Stratigraphical Profile Close to WALSH

Building
225
CPT-18 CPT-18
Cone resistance, qc (MPa) Friction ratio, Fr ( %)
0 5 10 15 20 25 0 2 4 6 8
15 15
10 10
5 5
Tunnel
0 0
-5 -5
-10 -10
-15 -15
CPT-19 CPT-19
Cone resistance, qc (MPa) Friction ratio, Fr ( %)
0 5 10 15 20 25 0 2 4 6 8
15 15
10 10
Tunnel
5 5
0 0
-5 -5
-10 -10
-15 -15
Figure 8-10 CPT Profile Close to Walsh Building
226
+14.5
qc=10 MPa
Sand
+10.0 Spearwood Sand
+8 6 Sand
qc= 15 MPa Tunnel 2 Tunnel 1
+5.5
qc= 5 MPa 12.0m
Clayey Silt to Silty
Clay
-1.2m
-2.0
-2.3 qc=15 MPa Clean sand to Silty sand
-4.0
qc =5 MPa -4.6 m Silty Clay toClay Perth Formation
6.9m 6.9m
-8.0
Silty Sand to Sandy
-10.0 Sil
Clay/Silty Clay
-13.0
Dense Sand or
Gravel
-19.0
Figure 8-11 Soil Profile and Location of Tunnel near WALSH Building
8.4 BUILDING MOVEMENT AFTER CONDITION PHASE GROUTING
The EL beams were located at about RL 21m AHD (6.5m above the ground surface) on
the Northern faade of the Walsh building. As mentioned earlier, the condition phase
grouting was carried out prior to any tunnelling activity to tighten the ground by filling
all the voids. This, in turn, improves the efficiency of grouting during tunnelling.
For the condition phase grouting, three passes of grouting were applied with a short rest
period of about a day between passes. The real-time structural monitoring data
summarised by Williams and Nobes (2007) suggest that there was only negligible
relaxation (settlement) after the rest period of the first and second passes of grouting,
while no relaxation occurred after the 3rd pass (on completion of the grouting). By
comparison, in soft Singapore marine clay, Addenbrooke et al. (2002) report significant
loss of heave (i.e. settlement) over a three day period immediately after the grouting
finished due to the dissipation of excess pore pressure built up during grouting; such a
phenomenon was not reported for London clay (stiff).
227
The movements of the Walsh building, as inferred from the strings of EL beams and
monitored anchor point, after condition phase grouting are shown in Figure 8-12 and
indicate a maximum heave of about 2.2 mm. The irregularity in the ground movement
can be attributed to the non uniformity in the application of grout and this can be seen
from Figure 8-13. The volume of the heave profile is ~ 0.04m3 per metre width in the
longitudinal direction. The total volume of the grout applied in the zone of influence
was 0.56m3 (assuming a zone of influence of 0.5m on either side of the EL beam, see
Figure 8-14). However detailed experimentation is required to calibrate the zone of
influence for different types of soil.
Movement due to condition phase grouting

3.0
Tunnel 1
Extent of
2.0
Walsh Building
1.0
0.0
-1.0
Tunnel 2
-2.0
-10 0 10 20 30 40
Figure 8-12 Movement of Building due to Condition Phase Grouting
228
Grout volume- Condition phase

160
140
120
100
80
Grout volume
60
EL Beam
40
20
0
53840 53850 53860 53870 53880
Easting
Figure 8-13 Variation of Grout Volume along the Transverse Direction of the
Building
Hungry Jacks
Zone of influence
EL Beams
WALSH BUILDING
Figure 8-14 EL Beam Location and Zone of Influence
229
Therefore the approximate ratio of the volume of grout injected to the heave volume
was about 14. This high ratio reflects the need for the grout to fill the voids in the sand
(noting that the grouting was performed in the Spearwood dune sand). The grouting
efficiency defined by Au et al. (2003) as the volume of heave to the injected volume of
grout is about 7%. However further research is needed to explain this low efficiency.
8.5 TRANSVERSE SETTLEMENT TROUGH (AFTER TUNNELLING)
Figure 8-8 indicates that only about a quarter (north side) of the Walsh building under
went compensation grouting. Therefore a comparison between the north (grouted) area
and South (non-grouted area) provides an immediate indication of the effects of the
grouting programme. The total volume of grout applied during tunnelling in the
(assumed one metre wide) zone of influence from the TAM borehole was 0.092m3; this
is about one sixth of that applied during the condition phase grouting (volume of
condition phase grouting ~0.56m3). The instrumentation at the Walsh building, which is
shown in Figure 8-5, included Surface Settlement Pins (SSP), Building Settlement
Points (BSP) and EL beams (note that EL beams were only on the north side of the
building).
8.5.1 South Side of Building (non-grouted area)
The chainage of the buildings north faade was at about 152.5m (average for Tunnel 1
and Tunnel 2) and of the south faade was at about 205m. Compensation grouting was
carried out up to a chainage of ~165m (chainage decreasing northwards). The settlement
troughs were plotted separately using data from between chainages of 170m to 210m
(un-grouted area) and 152m to 170m (grouted area). The surface settlement troughs
obtained from SSP and BSP data for both tunnel drives are shown in Figure 8-15, which
also indicates the extent of Walsh building and the front of the Church. Although the
BSPs were at different elevations, no discernible effect of elevation was observed and
hence they are plotted together with SSPs. The settlement trough measurements are
shown in Figure 8-15 are summarised in Table 8-1. The volume losses calculated from
these settlement troughs are 0.12% and 0.14% for Tunnel 1 and Tunnel 2 respectively.
This is comparable with the observed volume loss of 0.14% for Tunnel 1 boring at
MAS building (Section 7.5.1)
230
The settlement troughs, normalised by the maximum settlement (excluding the uplift at
the building location), are shown in Figure 8-16. On the east side of Tunnel 1 the
settlement trough is seen to be steeper than that of Tunnel 2 while the opposite is
evident on the west side. These trends can be explained with reference to Figure 8-17,
which shows the distance of each tunnel from buildings on either side of their axes. It is
apparent that trough widths are larger when buildings are further away from the tunnel
centre line (or conversely the presence of a stiff building close to a tunnel caused the
apparent settlement trough width to reduce). This observation is consistent with those
made for the MAS building. (Section 7.7)
Tunnel 1, Walsh Building Tunnel 2, Walsh Building

2 2
Front of the Church Extent of Walsh Building
Front of the Church Extent of Walsh Building
1 1
0
0
-1
-1
-2
-2
Tunnel 2 -3
-3 Tunnel 1 Tunnel 2 Tunnel 1
-4
-4
-5
-5 -30 -20 -10 0 10 20 30 40 50
-30 -20 -10 0 10 20 30 40 50 Distance from centre of tunnels, m
Figure 8-15 Settlement Troughs at Non-Grouted Area for Tunnel 1 and Tunnel 2
Trough width Smax Volume

East West (mm) Loss
(%)
Tunnel 1 17.5 20.9 -2.42 0.12
Tunnel 2 22.5 13.5 -3.04 0.14
Table 8-1 Summary of Settlement Troughs at Non-Grouted Area of Walsh

Building
231
0.0
0.5
Tunnel 1
Tunnel 2
1.0
1.5
-30 -20 -10 0 10 20 30 40
Figure 8-16 Normalised Settlement for Tunnel 1 and Tunnel 2
232
Figure 8-17 Location of Tunnels and the Buildings at Non-grouted Area
8.5.2 North Side of Building (grouted area)
The surface settlement trough determined for Tunnel 1 at the northern (grouted) side of
the Walsh building is plotted (from SSP and BSP) in Figure 8-18 and indicates a
volume loss of only 0.03%. No comparable data are available for Tunnel 2 as there was
no grouting under the Walsh building during Tunnel 2 excavation.
233
Tunnel 1, Walsh Building
1.0 Extent of Walsh Building
0.5
0.0
-0.5
-1.0
Tunnel 1
-1.5
Tunnel 2
-2.0
-2.5
-3.0
-20 -10 0 10 20 30 40
Figure 8-18 Surface movement at Grouted Location (from SSP and BSP data)
8.5.3 Comparison of Surface Movement at Grouted and Non-Grouted

area
The total volume of grout applied (assumed a one metre wide zone of influence around
the TAMs) during condition phase grouting and active phase grouting was 0.56m3 and
0.092m3 respectively. The surface movements observed after completion of Tunnel 1 in
the grouted and non-grouted areas are compared in Figure 8-19. It is immediately
apparent that grouting was successful in reducing settlements: maximum settlements
were reduced by a factor of 2.5 and volume losses were reduced from 0.12% to 0.03%.
Heave on the eastern side of the Walsh building is observed for both cases although it
should be noted that, even in the grouted area, grout did not extend over the full width
of the building.
234
Tunnel 1, Surface Movement

1.0
0.0
-1.0
-2.0
Tunnel 1
-3.0 Tunnel 2
without grouting
-4.0
with grouting
-5.0
-20 -10 0 10 20 30 40
Figure 8-19 Surface Settlements Due to Tunnel 1 Boring at Grouted and Non-
Grouted Locations
8.5.4 Building Settlement
Active phase grouting was carried out during the tunnelling with real time movement
monitoring of the building. The building movement that took place during Tunnel 1
boring is shown in Figure 8-20 and seen to be negligible. However, the movement
pattern of Walsh building shown in Figure 8-20 differs from that observed for the MAS
building (Section 7.6). While the MAS building tilted and was relatively rigid, the
Walsh building showed more flexible behaviour and evidence of both sagging and
hogging. The higher length to height ratio of the Walsh Building, which was greater
than one (H~18m, L~35m, L/H~1.9) clearly contributed to its more flexible behaviour
and this is in agreement with the observations of Breth & Chambosse (1974), and many
others. Since there was no active phase grouting carried out during Tunnel 2 boring, the
movement of the building is not considered for this case.
235
2.0
Tunnel 1
1.5 Extent of
1.0 Walsh Building
0.5
0.0
-0.5
-1.0
-1.5 Tunnel 2
-2.0
-10 0 10 20 30 40
Figure 8-20 Building Movement due to Grouting and Tunnel 1 Boring
8.6 NUMERICAL MODELLING
There are many published reports detailing how compensation grouting (CG) has been
successful in reducing the building settlement, as mentioned in Section 2.5 However in
all of these cases, the grout injection was dictated and controlled by contemporaneous
monitoring of movement. Research has now moved on into trying to develop an
improved understanding of compensation grouting using analytical and numerical
modelling. Very few publications are available describing these modelling techniques
because of the complexity of modelling the hydro-fracture technique. Improved
understanding of grouting mechanisms can improve efficiency and reduce grout
quantities and costs. The compensation grouting performed under the Walsh building is
clearly three dimensional However, bcause of the complexity of modelling the TAMs,
the effects of grouting are modelled here using Plaxis 2D version 9.0. The aim of these
analyses was to assess the potential of predicting effects of compensation grouting
numerically and to add to the limited number of published numerical back-analyses.
The soil parameters selected for each of the soil layers in Figure 8-11 are given in
Chapter 6. Since the building details were not available as mentioned earlier (Section
8.2.2), appropriate section sizes were assumed for the buildings beams, columns and
(assumed) raft. The details of the building components assumed and axial and bending
236
stiffness of each of those components are shown in Table 8-2. The mesh showing the
building, tunnels and the location of compensation grouting (CG) are shown in Figure
8-21. Although the area occupied by the grout is not known, it was assumed that
grouting was effective in the sand present from the foundation level (at 11.5mAHD)
down to the silty clay layer (with top level at +5.5mAHD). The compensation grouting
was modelled by applying a volumetric strain in the grouted area as indicated in Figure
8-21. The same technique of prescribed strain for modelling CG was adopted by
Nicolini and Nova (2000), Nicholson et al. (1994) and Schweiger and Falk (1998).
Size Bending stiffness, EI Axial stiffness, EA

width, b (m) Depth, d (m) (kNm2) (kN)
Beam 0.35 0.65 168207.8 4777500

Column 0.5 0.5 109375 5250000
Footing 1 1.35 4305656 28350000
Table 8-2 Details of the Building Component
Grouting
Tunnel 1
Figure 8-21 Mesh Showing the Location of the Grouting
237
8.6.1 Building Movement after Condition Phase Grouting
As mentioned in section 8.2.1, the condition phase grouting was carried out to tighten
the soil and fill the voids to improve the efficiency during active phase grouting. Prior to
modelling the grouting, the building was activated in Plaxis and the ground consolidated
to a minimum pore pressure of 1kPa to initiate the in situ condition. The condition
phase grouting was then modelled by imposing a volumetric strain in the grouted area as
indicated in Figure 8-21. To simulate the cementation caused by grouting, a nominal
smeared c' (cementation) value of 10kPa was specified to the sand in the grouted area.
A volumetric strain of 0.09% was applied to predict the heave profile shown in Figure
8-22 (the measured movement is also shown in Figure 8-22), which is similar to that
observed in the field. This selected volumetric strain of 0.09% predicted a heave profile
of same volume as that of the measured one (0.04m3). The applied volumetric strain
(0.09%) equates to about one sixth of the actual volume of grout (0.56m3) injected.
Since the volumetric stain was applied uniformly, a rather more uniform profile was
predicted.
At EL Beam Level
2.5 Tunnel 1
Extent of
1.5 Walsh Building
0.5
-0.5 Tunnel 2 condition phase movement

(field)
Plaxis 0.09% volumetric
-1.5 strain
-10 0 10 20 30 40
Figure 8-22 Observed and Predicted Movement after Condition Phase Grouting
238
8.6.2 Building Movement after Active Grouting (after tunnelling)
As mentioned in Section 8.2.1, the active phase grouting was carried out during the
tunnelling to compensate for the volume loss. The modelling of the tunnel excavation
was carried out as explained in Section 6.4. A volumetric strain of 0.03% was applied to
the improved sand to represent the active phase grouting (0.092m3 of grout) along with a
contraction of the tunnel of 0.03%, which was the volume loss due to tunnelling in the
grouted area, (see Section 8.5.3); this volumetric strain is one third of that applied
during the condition phase grouting (0.09%). However the volume of grout applied was
about one sixth of that applied during the conditioning phase, indicating that the
efficiency of grouting in the active phase improved due to the prior condition phase
grouting. The comparison of the predicted and the measured vertical movement of the
building after Tunnel 1 boring are shown in Figure 8-23. As seen in Chapter 7, here also
the numerical analysis couldnt predict a heave at the far end of the building. However,
the analysis could predict a similar (hogging and sagging) settlement profile.
3.0
Tunnel 1
2.0
Extent of Walsh Building
1.0
0.0
-1.0
Tunnel 2 Field Measurement

-2.0
Predicted Movement
-3.0
-10 0 10 20 30 40
Figure 8-23 Observed and Measured Movement of Building due to Tunnel 1
239
8.7 CONCLUSIONS
Trough widths have been seen to be larger when buildings are further away
from the tunnel centre line e.g. the presence of a stiff building close to a
tunnel caused the apparent settlement trough width to reduce.
Compensation grouting (in sand) can have a very beneficial effect on

tunnelling induced movements. Volume losses after grouting were only 25%
of those observed in nearby untreated and very similar ground conditions
Large grout volumes were required during the grout conditioning phase due
to the relatively high porosity of the Spearwood dune sand.
Grouting can be modelled effectively in FE analysis by imposing a

volumetric strain to the mass of soil subjected to grouting. However further
research is needed to allow designers to assess the relationship between the
degree of volumetric strain imposed and the grout volume injected.
240
CHAPTER 9 OBSERVATIONS,
CONCLUSIONS AND
RECOMMENDATIONS
9.1 INTRODUCTION
This thesis presents the observations and conclusions from detailed analyses of the
extensive instrumentation data collected in association with two bored tunnels in the
Central Business District (CBD) of Perth, Western Australia. The tunnelling was carried
out by an Earth Pressure Balanced Tunnel Boring Machine (EPB-TBM) and the
stratigraphy along the tunnel route comprised normally consolidated dune sand
(Spearwood sand) overlying interbedded layers of alluvial silts, clays and sands (Perth
Formation). The objective of this research was to investigate the mechanism of ground
movement (greenfield) in varying soil conditions such as seen in Perth, and also the
effect of soil structure interaction effect in controlling the ground or structural
movement associated with the closed face tunnel boring. This thesis also investigates
the effect of compensation grouting in reducing the volume loss associated with
tunnelling in dense sand and the possibilities of numerical modelling of compensation
grouting.
The greenfield surface movement was studied at three locations close to each other for
both tunnel drives; the tunnels encountered both the Perth Formation and Spearwood
sand at these locations. The soil structure interaction study was carried out for
multistorey framed structures on shallow footings; the tunnels were located in the
overconsolidated sand and silty layers of the Perth Formation at this area. The effect of
compensation grouting in sand was evaluated at a building location where the tunnels
were primarily in clayey layers of the Perth Formation.
241
The numerical analyses were performed using 2D and 3D versions of the Plaxis finite
element program. Soils were generally modelled using either the Plaxis HS or HSSmall
constitutive models.
This Chapter summarises the conclusions drawn from Chapters 4 to 9. It includes

observations concerning (i) the stress history and strength characteristics of the upper
clayey horizons of Perth Formation deduced from variety of laboratory tests carried out
on undisturbed soil samples, (ii) patterns of surface movement at three greenfield
locations close to each other, evaluated from extensive ground monitoring data, (iii)
results from 2D and 3D numerical predictions of greenfield movement using a variety
of soil constitutive models, (iv) the effect of soil structure interaction in modifying the
ground or structural movement (interpreted from the monitored movement of surface
and buildings), (v) the applicability of 2D and 3D FE analyses to model the soil
structure interaction problem, (vi) the effect of compensation grouting (hydro-fracture)
in reducing volume loss due to bored tunnelling, evaluated from the instrumentation
data and associated numerical modelling and (viii) recommendations for future
research.
9.2 OBSERVATIONS AND CONCLUSIONS
9.2.1 Soil Characterisation
Although a significant proportion of the bored tunnels went through the over
consolidated alluvial deposits of Perth Formation underlying the Spearwood dune sand,
no systematic laboratory investigation of the Perth Formation has been published to
date. This thesis presents the first comprehensive study of the alluvial clay deposits of
the upper horizons of Perth Formation. The results of the comprehensive laboratory
investigation, supplemented by in-situ test data, of the alluvial deposits are reported in
Chapter 4. The laboratory measurements were subsequently related and calibrated with
in-situ test data to assist in the assignment of material properties for the numerical
analyses.
The soil composition found from X-ray Diffraction Analysis (XRD) and Scanning
Electron Microscopy (SEM) analysis indicates that major portions of the soil comprised
242
Chapter 9 Observations, Conclusions and Recommendations
kaolin and quartz, suggesting the origin of the soil is the granite found in the Darling
mountain range, east of Perth. The high activity of the soil was found to be due to the
presence of the very small kaolinite crystals. Samples compressed one dimensionally to
high stresses in an oedometer indicated yield at a vertical effective stress (or
preconsolidation pressure) of ~700 kPa and in-situ overconsolidation ratios (OCR) of
between 3 and 4 were inferred using this stress. Earth pressure coefficient at rest (K0)
was estimated to be approximately unity using the correlation of Mayne & Kulhawy
(1982) for this OCR range. Although colour changes in the soil sample was noticed due
to the presence of iron, strength tests did not reveal the presence of significant
cementation/structure. The triaxial data revealed relationships between fines content
(FC) of the soil and friction angle '. Best estimate effective stress strength parameters
of ' of 35o to 40o at FC = 20% and ' = 25o at FC = 90%, with c' = 0 were inferred from
the strength tests. Triaxial samples equipped with local strain instrumentation showed
the non-linear nature of the soils stiffness and these data combined with the in-situ
seismic cone penetration test data allowed assessment of appropriate parameters for the
constitutive models in the finite element analyses.
9.2.2 Greenfield Ground Movement
The vertical surface movement (greenfield) measurements (observed using Electro

Level beams and Settlement pins on railway tracks) indicated wider tunnelling-induced
settlement troughs in soil profiles comprising a higher proportion of clay layers and this
is in keeping with the observations made in the centrifuge model tests (Mair, 1976).
Time dependent or consolidation settlements observed after tunnelling was completed
were not significant because of the sandy nature of the soil. Contrary to expectations,
the volume losses associated with the second tunnel boring were less than those induced
due to the first tunnel boring. Also the transverse settlement trough width (on the side of
the first tunnel) associated with the second tunnel drive was lower than the greenfield
value and this observation was consistent with the prediction using Plaxis. The average
volume losses of 0.18% and 0.1% for Tunnels 1 and Tunnel 2 respectively are within
the range of expectations of the earth pressure balancing technique employed.
Consistently higher settlement and volume loss were observed at one location (Tack 3)
between the two adjacent monitored locations (Track 1 and Track 5) and for both tunnel
243
drives; this finding suggests the presence of localised weaker soil and highlights the
need for thorough site investigations in advance of tunnelling projects.
The field observed settlement troughs were of Gaussian nature and the trough width
parameter K estimated from the observed settlement troughs at the three greenfield
locations were in agreement with past observations made in similar ground conditions
around the world. The relationship (suggested by Peck (1969) between distance from
the tunnel centre to the point inflection ( i ), tunnel diameter and tunnel depth fell under
the category of sands above water table, although the tunnels were below ground water
table. This could have been either because of the achievement of the effective control of
ground water flow into the tunnel or, as observed by Mair and Taylor (1997), there is no
difference between tunnels above or below ground water table.
9.2.3 Numerical Modelling of Greenfield Movement
The numerical modelling of the tunnel excavation using both 2D (Plaxis version 9.02)
and 3D (Plaxis 3D Tunnel, Version 2.0) Finite Element (FE) analyses demonstrated that
neither the 3D modelling nor the sophisticated soil models (i.e. HS and HSSmall) could
predict the Gaussian form of the settlement trough; this trend is in agreement with past
research (Addenbrooke; 1996, Van der Berg; 1999, Guedes de Melo and Santos Pereira;
2000, Franzius; 2003). The similarity of the settlement troughs predicted using the
simple Mohr-Coulomb model and more sophisticated constitutive soil models indicated
that the predicted surface movements were not very sensitive to the soil models.
2D and 3D FE analyses using two representative soil profiles (North profile and South
profile) showed a strong relationship between the earth pressure coefficient at rest (K0)
and the predicted settlement trough; e.g. deeper settlement troughs were predicted for
low K0 values. It was also found that the predicted settlement trough using the volume
loss method adopted by Plaxis was dependent on the tunnel lining stiffness with higher
settlement and volume loss being predicted for flexible linings. The dependence of the
predicted settlement trough on K0 and lining stiffness indicate that the effective stresses
generated in the vicinity of a tunnel have a significant influence on the surface
settlement trough. A comparison of a numerical parametric study and other centrifuge
model tests verified the importance of these stress changes, which for a given volume
244
loss, control the soil stiffness and the nature of the surface settlement trough. It was
found that a surface settlement trough of the Gaussian form can be predicted by
manipulating the at rest condition near the tunnel to a lower stress state (to compensate
for reduced stresses in the vicinity of the tunnel due to excavation) and by employing a
reduced lining stiffness (leading to a more squat deformed lining shape). Further
studies are required to assess the applicability of this approach for tunnels of different
depth and diameter.
Although tunnelling is a 3D problem, virtually no difference was observed between

(rapidly achieved) Plaxis 2D predictions and (time consuming) Plaxis 3D predictions of
the transverse settlement trough shape. This trend suggests that the final stress and
deformation state around a tunnel, as predicted by Plaxis, are not significantly
dependent on the stress path followed during 3D analyses, and depend primarily on the
volume loss input at each liner location. Less time and resource consuming 2D Plaxis
analysis are therefore recommended and these can be modified, as outlined above, to
predict a settlement trough closer to the Gaussian form.
2D FE analyses could reproduce the interaction of the tunnels in the greenfield site. The
analyses using the two representative soil profiles (North profile and South profile)
predicted lower settlements for a second tunnel located adjacent to the first tunnel in
normally consolidated sandy conditions and higher settlement where overconsolidated
clay layers are present. In addition, the settlement trough was narrower (on the side of
the first tunnel) in sandy soil and wider in clayey soil. These predictions highlight the
influence of soil type and/or stiffness at the tunnel location on the predicted settlement
trough. However more study is needed to enable definitive conclusions to be drawn
regarding the influence of depth and distance between the tunnels and the effect of
lining stiffness.
Although, with considerable effort and time, stabilised settlement could be predicted for
the longitudinal settlement trough using 3D analysis, the settlement trough was wider
than the observed longitudinal trough. The Plaxis 3D analyses failed to predict the effect
of face pressure and grout pressure on the settlement trough and this requires further
attention by the Plaxis developers.
245
9.2.4 Soil Structure Interaction Effect
The detailed analyses of the instrumentation data of two multistorey framed buildings
(the MAS and Walsh building) on shallow footings indicated a strong relationship
between the structural stiffness and the ground or building movement due to tunnelling
(as detailed in Chapter 7 and 8). The analyses of the measured data indicated that the
stiffness of the building altered the free field (Gaussian) form of the settlement trough in
its vicinity. The MAS building displayed a uniform tilt in the direction of the tunnel and
this response is in keeping with its relatively high rigidity -as indicated by its low length
to height ratio (L/H ~ 0.7); such a response was also seen by Breth and Chambosse
(1974) and Frischmann et al. (1994) for stiff structures. The Walsh building underwent
sagging and hogging and this response is in keeping with trends exhibited by Mair et al.
(1996) and others for flexible buildings (noting L/H ~1.9 for the Walsh building).
Measurements at the MAS location indicated reduced volume losses (by 50% ) and a
narrow trough width (on the side of the building). This soil-structure interaction effect
for Tunnel 1 and the absence of this effect for Tunnel 2 is believed to be the primary
reason for the volume losses induced during driving of Tunnel 2 being greater than
those of Tunnel 1 (noting that the MAS foundations were outside of the zone of
influence of Tunnel 2). The volume losses for the two tunnel drives at the Walsh
location (0.12% for Tunnel 1 and 0.14% for Tunnel 2) are about half of those due to
Tunnel 2 boring at the MAS building and are believed to be low due to the proximity of
the Walsh building to Tunnel 1 and the Church to west of Tunnel 2.
9.2.5 Numerical Modelling of Soil Structure Interaction
Numerical analyses, employing non-linear soil models were carried out to study the soil
structure interaction effect of the MAS building. Both 2D and 3D FE analyses failed to
reproduce the observed surface settlements as well as the building settlement. Although
both 2D and 3D could predict the pattern of movement (but not necessarily the form),
the 3D FE analysis could predict the pattern of building displacement more accurately
than the 2D analyses. A comparison between measurements with predictions indicates
that the FE approach adopted is a reasonable method for assessing soil structure
interaction effects. The parametric study using 2D FE analyses indicated that the
246
building stiffness and soil type and/stiffness has a significant influence on the shape of
the predicted settlement trough.
Parametric studies also showed that volume losses of 5 and even 10 times those incurred
during the MetroRail project are not likely to have caused significant structural damage
to building like MAS.
9.2.6 Compensation Grouting and Applicability of Numerical Modelling
The effect of compensation grouting (hydraulic fracture) in Spearwood sand was

assessed from the monitored movement of the Walsh building, the northern portion of
which had undergone grouting. The tunnels were located primarily in silty and clayey
layers of the Perth Formation at this area. Compensation grouting was implemented for
one half of the Walsh building to mitigate the effects of tunnelling induced settlements.
Less volume loss after grouting (25% of those observed nearby in untreated and very
similar ground conditions) indicated that the compensation grouting (in sand) can have a
very beneficial effect on tunnelling induced movements. However, the grouting
efficiency (defined as the ratio of the volume of heave to the injected volume of grout)
was only about 7%. The large grout volumes required during the grout conditioning
phase are expected given the relatively high porosity of the Spearwood dune sand.
The grouting was modelled using Plaxis 2D to examine the possibility of numerically
modelling the effects of grouting. It was found that grouting can be modelled effectively
in FE analysis by imposing a volumetric strain to the mass of soil subjected to grouting.
However further research is needed to allow designers to assess the relationship
between the degree of volumetric strain imposed and the grout volume injected.
9.3 CONTRIBUTIONS TO THE FIELD OF KNOWLEDGE
Through field observation and the numerical analyses described in the thesis, the
contributions to the field of knowledge of this thesis include:
1. The stress history and the strength characteristics of the upper horizons of Perth
Formation were assessed for the first time by a comprehensive laboratory
investigation.
247
2. The database of monitored movements due to earth pressure balancing tunnel
boring in interbedded sands and clays has been extended considerably, for both
greenfield and built-up areas.
3. A comparison of parametric studies performed using FE analyses with available

centrifuge data allowed a thorough investigation of the wide and shallow
settlement trough often predicted for overconsolidated soil.
4. Detailed 2D and 3D soil structure interaction analyses involving multi storey

buildings revealed the influence of building stiffness on the surface settlement.
5. Analysis of the monitored data showed the efficiency of compensation grouting

in sand for reducing the tunnelling induced volume loss. The potential for FE
modelling of compensation grouting is also shown.
9.4 RECOMMENDATIONS FOR FUTURE RESEARCH
1. A strong relationship was found between the mobilised stress and the
numerically predicted size and shape of the transverse settlement trough. Hence
numerical modelling of control tests (e.g. centrifuge tests) and a comparison of
the stress changes around and above the tunnel will assist in the development of
simple approaches.
2. More study (involving experimental and numerical modelling) is required to

understand the influence of tunnel diameter and depth on ground movement
patterns.
3. The interaction effect of multiple tunnels requires further study to investigate the
relative influence of soil stiffness and lining stiffness on predictions for the
second tunnel.
4. More well controlled physical models are required to validate numerical

approaches for soil structure interaction. In particular, centrifuge studies
involving structures of varying rigidities and using soils with well know
properties would provide designers with case histories that enable improvement
of FE approaches for modelling soil-structure interaction.
248
5. More physical investigations are required to understand the relationship between

input grouting volumes and the associated response of structures.
In summary, while field observations such as those obtained in Perth CBD will continue
to provide much needed data for designers, it is thought that significant advances of
numerical approaches to predict tunnelling effects are only possible using well
controlled physical model tests in the centrifuge. These tests involve well defined
boundary conditions, soil characteristics and structural properties and, as such, do not
suffer from the significant variabilities and complexities encountered in the field.
249
250
REFERENCES
Addenbrooke, T.I. (1996). Numerical Analysis of tunnelling in stiff clay. PhD Thesis,
University of London.
Addenbrooke, T.I., Ong, J.C.W. and Potts, D.M. (2002). Finite-element analysis of a
compensation grouting field trial in soft clay. Proc. Institution of Civil Engineers,
Geotechnical Engineering, Vol. 155, issue 1, 47-58.
Addenbrooke, T.I., Potts, D.M. and Puzrin, A.M. (1997). The influence of pre-failure
soil stiffness on numerical analysis of tunnel construction. Geotechnique, Vol. 47,
No.3, 693-712.
Airey, Ryan & Hill (2004). Property condition report- Building No.24 Walsh
building.
Andrews, D.C. (1971). Soils of the Perth area-City Centre. Division of Applied
Geomechanics technical report no.13, Commonwealth Scientific and Industrial
Research Organisation, Australia.
Appleyard, Y. (2003). Ground water quality in the Perth region. Australian

Geomechanics, Vol.38, No.4, 103-112.
AS 1726 (1993). Australian standard for Geotechnical site investigations
Attewell, P.B. (1978). Ground movements caused by tunnelling in soil. Proc.

International Conference on Large Ground Movements and Structures, Pentech
Press, London, 812-948.
Attewell, P.B., Yeates, J. and Selby, A.R. (1986). Soil Movements Induced by
Tunnelling and their Effects on Pipelines and Structures, Blakie, Glasgow
Au, S.K.A., Soga, K., Jafari, M.R., Bolton, M.D.,Komiya, K. (2003). Factors affecting
long-term efficiency of compensation grouting in clays. Journal of Geotechnical
and Geoenvironmental engineering, 129, 3, 254-262.
251
Baker, W.H., Cording, E.J. and MacPherson, H.H. (1983). Compaction grouting to
control ground movements during tunnelling. Underground Space, Vol.7, 205-
212.
Benz, T., (2006). Small-Strain Stiffness of Soils and its Numerical Consequences. Ph.D.
Thesis, Stuttgart Universitt.
Boscardin, M.D. and Cording, E.J. (1989). Building response to excavation-induced

settlement. ASCE, Journal of Geotechnical Engineering, Vol.115, No.1, 1-21.
Brandon T.L., Rose A.T. and Duncan J.M. (2006). Drained and undrained strength
interpretation for low-plasticity silts. Journal of Geotechnical and
Geoenvironmental Engineering, ASCE, Vol. 132, No.2, 250-257.
Breth, H and Chambosse, G. (1974). Settlement behaviour of buildings above subway

tunnels in Frankfurt clay. Proc. Conf. Settlement of Structures, Cambridge,
Pentech Press (Published 1975, London), 329-336.
Burd, H.J., Houlsby, G.T. and Augarde, C.E. (2000). Modelling tunnelling-induced
settlement of masonry buildings. Proceedings of Institution of Civil engineers,
Geotechnical Enginering, 2000, 143, 17-29.
Burland J.B. (1990). On the compressibility and shear strength of natural clays, 30th
Rankine lecture. Geotechnique, Vol. 40, No.3, 327-378.
Burland J.B. (1995). Assessment of risk damage to buildings due to tunnelling and
excavations. Invited special lecture, Proc. 1st international Conference on
Earthquake Geotechnical Engineering, IS-Tokyo95, 1189-1201. (has to get from
library)
Burland J.B., and Wroth C.P. (1975). Settlement of building and associated damage.
Proc. Conf. Settlement of Structures, Cambridge, Pentech Press (Published 1975,
London), 611-654.
Burland, J.B., Mair, R.J. and Standing, J.R. (2004). Ground performance and building
response due to tunnelling. Advances in Geotechnical Engineering: The Skempton
Conference, 291-342.
Butterfield R. (1979). A natural compression law for soils (an advance on e-log p).
Geotechnique, Vol.29, No.4, 269-480.
252
References
Chambon, P and J.F. Cort (1994), Shallow tunnels in cohesion less soil: stability of
tunnel face. Journal of Geotechnical Engineering, ASCE, Vol.120, No.7, 1150-
1163.
Chappell, J. and Shackleton, N.J. (1986). Oxygen isotopes and sea level. Nature,
Vol.234, 137-140.
Cording, E.J. (1991). Control of ground movements around tunnels in soil. General
report, 9th Pan-American Conference on Soil mechanics and Foundation
Engineering, Chile.
Cording, E.J. and Hansmire, W.H. (1975). Displacements around soft ground tunnels.
general report 5th Pan American conference on Soil Mechanics and foundation
engineering ,Buenos Aires, Session IV, 571-632.
Cording, E.J., Brierley, G.S., Mahar, J.W. and Boscardin, M.D. (1989). Controlling
ground movements during tunnelling. Art and Science of Geotechnical
engineering at the Dawn of the 21st Century, Pentice-Hall, New Jersey, USA, 477-
505.
Davidson, W.A. (1995). Hydrogeology and groundwater resources of the Perth region,
Western Australia. Geological survey, Bulletin 142, Western Australia.
Dimmock, P. S. (2003). Tunnelling induced ground and building movement on the

Jubilee Line Extension. PhD Thesis, University of Cambridge.
Drummond A.F., Varago C., Gilkes R.J. and Hart R.D. (2001). The relationships
between kaolinite crystal properties and the origin of materials for a Brazillian
kaolin deposit. Clays and minerals, Vol.49, No.1, 44-59.
Essler, R.D., Drooff, E.R. and Falk, E. (2000). Compensation grouting, Concept,
Theory and Practise. Proc. sessions of Geo-Denver 2000, Geotechnical special
publication no. 104, Advances in grouting and ground modification, 1-15.
Ezzeldine, O.Y. (1999). Estimation of the surface displacement field due to construction
of Cairo Metro Line EI Khalafawy- St. Therese. Tunnelling and Underground
Space Technology, Vol.14, No.3, 267-279.
Fahey M. (1988). Interpretation of self-boring pressuremeter tests at St. Georges Tce.

Report for Golder Associates, School of Civil & Resource Engineering, UWA.
253
Fahey M., Lehane, B, M. and Schneider, J.A., D. (2007). Self boring pressuremeter
testing in Spearwood sand. Australian Geomechanics, Vol. 42, N0.4, 57-71.
Fahey M., Lehane, B. and Stewart, D. (2003). Soil stiffness for shallow foundation
design in the Perth CBD. Australian Geomechanics, Vol. 38, N0.3, 61-89.
Franzius, J.N. (2003). Behaviour of buildings due to tunnel induced subsidence. PhD
Thesis, Imperial college, London.
Frischmann, W.W., Hellings, J.E., Gittoes, G. and Snowden, C. (1994). Protection of

the Mansion House against damage caused by ground movements due to the
Docklands Light Railway Extension. Proc. Inst. Civil Engineers, Geotechnical
Engineering, Vol.107, No.2, 65-76.
Golder Associates (2003). Geotechnical and Hydrogeological investigation report.
Gordon, R. (2003). Sea level changes and paleochannels in the Perth area. Australian
Geoechanics, Vol.38, No.4, 85-89.
Gozzard B. (2007). The Guildford Formation re-evaluated. Australian Geomechanics,

Vol. 42, No.3, 59-79.
Guedes de Melo, P.F.M., Santos Pereira, C. (2000). The role of the soil K0 value in
numerical analysis of shallow tunnels. Proc. of the International Symposium on
Geotechnical Aspects of Underground Construction in Soft Ground, Balkema,
Rotterdam, 379384
Harris, D.I., Mair, R.J., Love, J.P. Taylor, R.N. and Henderson, T.O. (1994).
Observations of ground and structure movements for compensation grouting
during tunnel construction at Waterloo Station, Geotechnique, Vol. 44, No.4, 691-
713.
Hudson-Smith E. and Grinceri, M. (2007). Ground conditions and building protection

for the new MetroRail City project. Australian Geomechanics, Vol. 42, No.3, 33-
58.
Jacobsz, S. W. (2002). The effect of tunnelling on piled foundations. PhD Thesis,

University of Cambridge.
254
References
Johnson, I.D., Gibson, S.P. and Aikawa, F. (2007). Bored tunnel lining design. Proc.
Seminar on New MetroRail City Project, Tunnelling and Underground Structures,
69-80
Kasper, T. and Meschke, G. (2006). On the influence of face pressure, grouting pressure
and TBM design in soft ground tunnelling. Tunnelling and underground space
technology, Vol.21, 160-171.
Koelewijn, A.R. and Verruijt, A. (2001). Simplified three-dimensional numerical

modelling of shield tunnel advancement, Proc. 15th International Conference on
Soil Mechanics nad Foundation Engineering, Istanbul, Vol.2, 1463-1466.
Kohata, Y., Tatsuoka F., Wang L., Jiang, G.L., Hoque E. and kodaka T. (1997).
Modelling of non-linear deformation properties of stiff geomaterials.
Geotechnique, Vol. 47, No.3, 563-580.
Komiya, K., Soga, K., Akagi, H., Hagiwara, T. and Bolton, M.D (1999). Finite element
modelling of excavation and advancement process of a shield tunnelling machine.
Soils and Foundation, Vol. 39, No.3, 37-52.
Kovacevic, N., Edmonds, H.E., Mair, R.J., Higgins, K.G., Potts, D.M. (1996).
Numerical modelling of the NATM and compensation grouting trials at Red cross
way. Geotechnical aspects of underground construction in soft ground, London,
Balkema:Rotterdam, 553-559.
Kummerer, C., Thurner, R., Rigazio, A., and Zamagni, A. (2007). Proc. 14th European
Conference on Soil Mechanics and Geotechnical Engineering, Madrid, Vol.3,
1269-1274.
Lee, K.M. and Rowe, R.K. (1990a). Finite element modelling of the three-dimensional
ground deformation to tunnelling in soft cohesive soils: Part 1 method of analysis.
Computers and Geotechnics, Vol. 10, No. 2, 87-110.
Lee, S.W. (2002). The use of compensation grouting in tunnelling: a case study. Proc.
Institution of Civil Engineers, Geotechnical Engineering, Vol. 155, N0.2, 101-
109.
255
Lehane B.M. and Fahey M. (2004). Using SCPT and DMT data for settlement
prediction in sand. Proc. 2nd International Conf. on site characterisation, Porto,
Portugal, Vol. 2, 1673-1680.
Lehane B.M. and Faulkner A. (1998). Stiffness and strength characteristics of a hard
lodgement till. Proc. 2nd Int. Symp. on the Geotechnics of hard soils and soft rocks,
Naples, Vol.2, 637-646.
Lehane B.M., Ismail, M.A. and Fahey, M. (2004). Seasonal dependence of in situ test
parameters in sand above the water table. Geotechnique, Vol.54, N0.3, 215-218.
Lehane, B.M. ,Mathew G. and Stewart, D. (2007). A laboratory investigation of the upper
horizons of the Perth/Guildford Formation in Perth CBD. Australian
Geomechanics, Vol. 42, No.3, 87-100.
Lehane, B.M., Doherty, J.P. and Schneider, J.A. (2008). Settlement prediction for
footings on sand. Fourth International Symposium on Deformation
Characteristics of Geomaterials, Atlanta, USA, Vol. 1, 133-150
Leighton Kumagai Joint venture (2004). New MetroRail City Project, Geotechnical
Interpretative Report, Perth, Australia.
Leroueil S. and Hight D.W. (2003). Behaviour and properties of natural soils and soft
rocks. Proc. Characterisation and Engineering Properties of Natural Soils,
Singapore Vol.1, 29-254.
Liu, D.G., Houlsby, G.T. and Augarde, C.E. (2000). Two dimensional analysis of
settlement damage to masonry buildings caused by tunnelling. The structural
engineers, Vol.79, No.1, 19-25.
Lunne T., Robertson P.K. and Powell J.J.M. (1997). Cone Penetration Testing in
Geotechnical Practice. Spon Press.
Mair, R.J. (1979). Centrifuge modelling of tunnelling construction in soft clay. PhD
Thesis, University of Cambridge.
Mair, R.J. and Taylor, R.N. (1997) Theme lecture: Bored tunnelling in the urban
environment. 14th International conference on Soil Mechanics and Foundation
Engineering, Hamburg, 2353-2385.
256
References
Mair, R.J., Taylor, R.N. and Burland, J.B. (1996). Prediction of ground movements and
assessment of risk of building damage due to bored tunnelling. Proc. Int.
Sympossium on Geotechnical aspects of Underground Construction in Soft
Ground, London, Balkema, 713-718.
Marshall, A.M. (2009). Tunnelling in sand and its effect on pipelines and piles. PhD
Thesis, University of Cambridge.
Mayne P.W. and Kulhawy F.H. (1982), K0-OCR relationships in soils. Journal
Geotechnical Engineering Division, ASCE, Vol. 108(GT6), 851-872.
McArthur, W.M. and Bettenay, E. (1960). The development and distribution of the soils
of the Swan Coastal Plane, Western Australia. CSIRO Australian Soils,
Publication No.16
McGough, P.G. and Wlliams, M. (2007). Geotechnical instrumentation and Monitoring.

Proc. Seminar on New MetroRail City Project, Tunnelling and Structures, Perth,
81-136
Mitchell J.K. and Soga K. (2005). Fundamental soil behaviour. 3rd edition, Wiley.
Mroueh, H. and Shahrour, I. (2002). A full 3-D finite element analysis of adjacent
structures. Computers and Geotechnics, Vol.30, 245-253.
New, B.M. and Bowers, K.H. (1994). Ground movement model validation at the
heathrow Express trial tunnel. Tunnelling94, Proc. 7th International symposium
of Inst. Of Mining and Mettalurgy and British Tunnelling Society, London,
Chapman and Hall , 301-329
New, B.M. and OReilly, M.P.(1991). Tunnelling induced ground movements:

predicting their magnitude and effects. 4th International Conference on Ground
Movements and Structures, Cardiff, invited review paper, Pentech Press, 671-697.
Ng, C.W.W and Lee, G.T.K. (2005). Three - dimensional ground settlements and stress
transfer mechanisms due to open-face tunnelling: Canadian Geotechnical
Journal, Vol. 42, 1015-1029.
Nicholson, D.P., Gammage C., Chapman, T. (1994). The use of finite element methods
to model compensation grouting. Grouting in the ground, The institution of civil
engineers, Thomas Telford,:London, 297-312.
257
Nicolini, E., Nova, R. (2000). Modelling of a tunnel excavation in a non-cohesive soil
improved with cement injections. Computers and Geotechniques,Vol.27, 249-
272.
Nyren, R. (1998). Field measurement above twin tunnel in London Clay. PhD Thesis,
Imperial College, University of London.
OReilly, M.P, Mair. R.J. and Alderman, G.H. (1991). Long-term settlements over
tunnels: an eleven year study at Grimsby, Tunnelling 91, London, 173-181.
OReilly, M.P. and New, B.M.(1982). Settlements above tunnels in the United
Kingdom-their magnitude and prediction. Tunnelling 82, London, IMM, 173-181.
Panet, M. and Guenot, A. (1982). Analysis of convergence behind the face of a tunnel.
Tunnelling82, London, 197-204.
Peck, R.B. (1969). Deep excavations and tunnelling in soft ground. Proc. 7th
International conference on Soil mechanics and Foundation Engineering, Mexico
city, State of the Art volume, 1969, 225-290.
Perez Saiz, A., Garami, J., Arcones, A. and Soriano A. (1981). Experience gained
through tunnel instrumentation. Proc. 10th International conference on Soil
Mechanics and Foundation Engineering, Stockholm, Vol.1, 1981, 345-352.
Phienwej, N., Sirivachiraporn, A., Timpong, S. and Tavaranum, S. (2006).

Characteristics of ground movements from shield tunnelling of the first Bangkok
subway line, Proc. International Symposium on Underground Excavation and
Tunnelling, Bangkok, Thailand, 319-330.
Plaxis 3D Tunnel Reference manual- version2
Playford, P.E. (1988).Guidebook to the geology of Rottnest Island. Geological Society

of Australia, western Australia Division, Excursion Guidebook No.2.
Pototschnik, M.J. (1992). Settlement reduction by soil fracture grouting. ASCE

Speciality conference on Grouting, Soil Improvement and Geosynthetics, New
Orleans, 398-409.
Potts, D.M. (1976). Behaviour of lined and unlined tunnels in sand. PhD Thesis,
University of Cambridge.
258
References
Potts, D.M., Addenbrooke, T.I. (1997). A structures influence on tunnelling-induced

ground movements. Proc. Institution of Civil Engineers, Geotechnical
Engineering, Vol. 125, 109-125.
Robertson P.K. (1990). Soil classification using the cone penetration test. Canadian
Geotechnical Journal, Vol.27, No.1, 151-158.
Robertson, P.K. (2004). Evaluating soil liquefaction and post-earthquake deformations

using the CPT. Proc. 2nd International Conf. on site characterisation, Porto,
Portugal, Vol.1, 233-252.
Rowe, R.K., Lo, K.Y. and Kack, G.J. (1983). A method of estimating surface settlement
above tunnels constructed in soft ground. Canadian Geotechnical Journal, Vol.
20, 11-22
Santos J.A., Correia A.G. (2001). Reference threshold shear strain of soil. Its
application to obtain a unique strain-dependent shear modulus curve for soil.
Proceedings 15th International Conference on Soil Mechanics and Geotechnical
Engineering (Istanbul, Turkey), Vol.1: 267270.
Schanz, T., (1998). Zur Modellierung des Mechanischen Verhaltens von

Reibungsmaterialen, Habilitation, Stuttgart Universitt.
Schanz,T., Vermeer, P.A. and Bonnier, P.G. (1999). Formulation and verification of the
Hardening-Soil Model. Beyond 2000 in Computational Geotechnics,. Balkema,
R.B.J. Brinkgreve, Rotterdam, 281-290.
Schneider, J.A., Fahey, M. and Lehane, B.M. (2008). Characterisation of an unsaturated

sand deposit by in situ testing. Geotechnical and Geophysical Characterisation,
Taylor and Francis, Balkema, The Netherlands, 633-638.
Schweiger, H.F. and Falk, E. (1998). Reduction of settlements by compensation

grouting-numerical studies and experience from Lisbon underground. Tunnels and
Metropolises: Proceedings of the World Tunnel Congress98, Sao Paulo, Brazil,
Balkema:Rotterdam, 1047-1052.
Sigl, O. and Yamazaki, H. (2007). TBM tunnelling. Proc. Seminar on New MetroRail
City Project, Tunnelling and Underground Structures, 43-67
259
Son, M. and Cording, E.J. (2005). Estimation of building damage due to excavation
Induced Ground Movements. Journal of Geotechnical and Geoenvironmental
engineering, 162-177.
Son, M. and Cording, E.J. (2008). Numerical model tests of building response to
excavation-induced ground movements. Canadian Geotechnical Journal, 45,
1611-1621.
Standing, J.R., Farina, M., Potts, D.M. (1998). The prediction of tunnelling induced
building settlements-a case study, Tunnels and Metropolises: Proceedings of the
World Tunnel Congress98, Sao Paulo, Brazil, Balkema:Rotterdam, 1053-1058
Sugiyama, T., Hagiwara, T ,Nomoto, T., Nomoto, M., Ano, Y., Mair R.J., Bolton, M.D.
and Soga, K. (1999). Observations of ground movements during tunnel
construction by slurry shield method at the Docklands Light Railway Lewisham
Extension- East London, Soils and Foundations, Vol.39, No.3, 99-112.
Sugiyama, T., Nomoto, T., Ano, Y. and Hagiwara, T. (2000). Compensation grouting at
the Docklands Light Railway Lewisham Extension project. Geotechnical Aspects
of Underground Construction in Soft Ground, Kusakabe, Fugita & Miyazaki
(eds), Balkema, Rotterdam, 319-324.
Swoboda, G. (1979). Finite element analysis of the New Australian Tunnelling Method
(NATM). Proc. third international conference on Numerical method in
Geomechanic, Aachen, Vol. 2, 581-586.
Whittaker, H. (1938). Effect of particle size on plasticity of kaolinite. Proc. 40th Annual
meeting of ceramic society, New Orleans, USA.
Williams, M. and Nobes, C. (2007). Soilfrac compensation grouting for building

protection on the new MetroRail City Project. . Proc. Seminar on New MetroRail
City Project, Tunnelling and Structures, Perth, 27-42.
Wisser, C., Augarde, C.E., and Burd, H.J. (2005). Numerical modelling of
compensation grouting above shallow tunnels. International journal for numerical
and analytical methods in geomechanics, 29, 443-471.
260
References
Wongsaroj, J. (2005). Three-dimensional finite element analysis of short and long-term

ground response to open face tunnelling in stiff clay. PhD Thesis, University of
Cambridge, London.
261

Mathew Gima V 2010

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Mathew Gima V 2010

Hochgeladen von

Copyright:

Verfügbare Formate

ANALYSIS AND INTERPRETATION OF

GROUND AND BUILDING MOVEMENTS

This thesis is presented for the degree of

The University of Western Australia

School of Civil and Resources Engineering

Figure 2-1 Source of Ground Loss (Attewell, 1978) ......................................................... 9

Figure 2-2Observed Failure Mode based on Centrifuge Model Tests............................. 10

Figure 2-4 Profile of Mobilised Lateral Earth Pressure Coefficient on Tunnel

Figure 2-6 Transverse Settlement Trough........................................................................ 14

Figure 2-7 Longitudinal Settlement Profile (Franzius, 2003).......................................... 19

Figure 2-8 Predicted Greenfield settlement and Observed Building Settlement of

Figure 2-13 Comparison of Damage Estimation Criterion and Damage Levels

Figure 3-1, Bored Tunnel Alignment................................................................................33

Figure 3-4 Perth Region Generalized Geology.................................................................38

Figure 3-6 Schematic Section Showing Stratigraphic Relationships of Superficial

Figure 3-9 Old Lakes near Project Alignment ..................................................................45

Figure 3-10 Specifications of EPB-TBM..........................................................................50

Figure 3-11 EPB Tunnel Boring Machine- View on Cutter head.....................................51

Figure 3-12 Schedule of Tunnelling South of William Street Station..............................56

Figure 3-13 Schedule of Tunnelling North of William Street Station..............................57

Figure 4-1 SPT N vs. Depth..............................................................................................61

Figure 4-5 Typical CPT Data at Test Location.................................................................66

Figure 4-6 Ic and ID Indices at the Site..............................................................................68

Figure 4-9 Colour changes due to Iron on Soil Samples ................................................. 72

Figure 4-10 SEM Images ................................................................................................. 73

Figure 4-11 Oedometer Data for Three Samples ............................................................. 75

Figure 4-12 Stress Paths Measured in CIU Triaxial Compression Tests......................... 77

Figure 4-14 Stress Paths Measured in Undrained Simple Shear...................................... 80

Figure 4-15 Undrained Strength Ratios Plotted as a Function of (a) consolidation

Figure 4-16 Variation of A (undrained strength ratio at OCR=1) with Fines

Figure 5-3 CPT Profile South of Track 1......................................................................... 93

Figure 5-4 Idealised Soil Profile South of Railway Tracks ............................................. 93

Figure 5-6 CPT Profile North of Track 5......................................................................... 96

Figure 5-7 Idealised Soil Profile North of Railway Tracks ............................................. 96

Figure 5-9 Typical View of a Horizontal EL Beam......................................................... 99

Figure 5-10 Schematic View of Horizontal EL Beam ..................................................... 99

Figure 5-12 Closer View of EL Beams on the Track..................................................... 101

Figure 5-16 Relationship between Maximum Heave and Maximum Settlement

Figure 5-23 Typical Longitudinal Profile for Tunnel 1 Boring......................................116

Figure 5-24 Linear Regression Lines for Tunnel 1 Boring............................................118

Figure 5-25 Linear Regression Lines for Tunnel 2 Boring.............................................119

Figure 5-28 Variation of Grout Volume / Ring .............................................................123

Figure 5-30 Normalised Field Settlement Troughs for Tunnel 1 ...................................127

Figure 6-10 Comparison of the Transverse Settlement Trough Predicted using

Figure 6-11 Deformed Mesh using MC (Eur) Model ..................................................... 150

Figure 6-14 Geometric Model of the Tunnel Excavation .............................................. 152

Figure 6-23 Typical Longitudinal Settlement Trough ....................................................159

Figure 6-24 Boundary Effect for Longitudinal Trough .................................................159

Figure 6-25 Settlement Trough with Different FP and GP.............................................160

Figure 6-26 Longitudinal Settlement Trough for Different GP and FP..........................161

Figure 6-29 Comparison of Normalised Settlements for Tunnel 1.................................165

Figure 6-34 Settlement Trough with Different K0 around the Tunnel............................172

Figure 6-35 Settlement Profile with Different Stiffness .................................................173

Figure 6-36 Mobilised Horizontal Earth Pressure ..........................................................173

Figure 6-39 Comparison of Predicted and Measured Settlement Trough ......................177

Figure 6-40 Interaction Effect of Tunnels using South Profile.......................................178

Figure 7-1 Aerial View of MAS Building ..................................................................... 184

Figure 7-3 Front View of MAS Building....................................................................... 186

Figure 7-4 View Opposite MAS Building ..................................................................... 186

Figure 7-6 Geotechnical Investigation Locations near MAS Building.......................... 188

Figure 7-8 CPT Profile................................................................................................... 190

Figure 7-13 Movement of Building at Various Times during Tunnelling..................... 195

Figure 7-17 2D Mesh Showing Tunnels and Building .................................................. 199