Blast Loading On Anchorage Zones in Cable-Stayed Bridges

BLAST LOADING ON ANCHORAGE ZONES IN
CABLE-STAYED BRIDGES
by
LUA WEI YEW
Submitted in total fulfilment of the requirements of the degree of

Doctor of Philosophy
2015
Department of Infrastructure Engineering

The University of Melbourne
ABSTRACT
Blast incidents, occurred under inadvertent and deliberate circumstances, will bring about
detrimental economic and social impacts, both domestically as well as internationally. It is an
irrefutable fact that blast-related issues are becoming overwhelmingly relevant to civil
infrastructure, including cable-stayed bridges, an essential element in the vibrant
transportation network.
The principal objectives of this ground-breaking research include: determination of the

fundamental blast response of cable anchorage zones, which leads to proposal of viable
retrofitting techniques for existing facilities and feasible design guidelines for future assets;
investigation of the disproportionate and progressive collapse mechanisms of cable-stayed
bridges, and subsequent delivery of sensible mitigation strategies.
Based on sophisticated finite element analysis, this study relies on: a global model, which
contains only the critical details ideal for the collapse analysis; a local model, which retains
the crucial features deemed indispensable for the blast simulation. Both person-borne-
improvised-explosive-devices (appeared below the girder) as well as vehicle-borne-
improvised-explosive-devices (positioned on top of the deck) are utilised.
Blast-induced material breakdown will substantially weaken the anchorage precinct housed in
the deck-and-girder system, and while residing in a deplorable state, if sufficient residual
capacity is not retained, brittle ultimate crushing failure is inevitable under the sustained
gravitational loading. The anchorage region managed to survive the event involving a
compact or standard automobile, but not an ordinary truck or a movable package/equivalent,
which gave rise to far more critical consequences. The anchorage point accommodated at the
tower will not be adversely affected, and likewise for the cable.
i
Abstract
With the portable charge, the anchorage precinct remained serviceable only when the inner
shield was installed along with the outer armour, but not independently; and likewise with the
vehicle bomb, with or without the overlaying shelter (incapable of acting alone). Steel,
instead of carbon-fibre-reinforced polymer, is competent to guarantee more remarkable
defence. The credibility of a fortification scheme, in terms of energy dissipation, stiffness and
confinement contribution, and restriction of matter disengagement, is not only dictated by
material mass and attributes, but also the composite behaviour developed.
Compatible with the traditional codes, the advanced design specifications are intended to be
implemented through the integrated logistics, by adhering to the performance-based design
principle. The innovative tactics identified include: alleviation of extreme blast action and
conventional load effects, by raising the standoff distance and expanding the bearing plate;
enhancement of blast endurance, by enlarging the anchorage region, augmenting the
reinforcement and improving the concrete quality. The relative performance of the anchorage
points is governed by both the blast resistance as well as the reaction force.
Withdrawal of multiple intermediate and/or counterbalancing stays is required to serve as an

initiating cause which triggers incremental load transfer and instigates uninterrupted loss of
consecutive supports (which releases the conveying spans and prompts the towers to topple).
The catastrophic collapse styles are a devastating advance and amplification of the most basic
failure modes. Occurrence hindrance is always better than resistance assurance; disastrous
destruction can be successfully prohibited through effective control of vulnerability, the
concept of redundancy and vigorous upgrade of robustness.
ii
DECLARATION
This is to certify that:
i. the thesis comprises only my original work,

ii. due acknowledgement has been made in the text to all other material used,
iii. the thesis is less than 100,000 words in length, exclusive of tables, maps,
bibliographies and appendices.
LUA Wei Yew
iii
PREFACE
This linkage project, jointly funded by the Roads and Maritime Services, in New South
Wales, together with the Australian Research Council, consists of two parallel studies carried
out by the University of Melbourne as well as the University of Technology, Sydney. The
former focusses predominantly on the local behaviour of cable anchorage zones subjected to
extreme impulsive blast loading, while the latter focusses primarily on the global
performance of cable-stayed bridges as a consequence of extreme loading effects.
Notwithstanding the close cooperation, the final products due to the distinct contributions are
published separately.
Independent of this project, the Department of Infrastructure Engineering in the University of

Melbourne has conducted numerous blast investigations involving concrete, steel, fibre-
reinforced polymer as well as polyurea, concerning not only the material behaviour but also
the structural response, with the aid of a series of experimental and theoretical programmes.
The relevant effort previously undertaken is further extended here, where the knowledge and
technology developed are appropriately transferred and adapted to suit the specific need of
this particular research.
iv
ACKNOWLEDGEMENTS
First and foremost, I would like to express my deepest gratitude to my family: my

grandparents; my father, Kun Eng; my mother, Len Yin; and my sister, Sin Yi. I am greatly
indebted to them for their perpetual love and care, which have granted me exceptional faith
and strength to achieve my aspirations. With great pleasure, I dedicate the success of this
doctoral research to them.
I am immensely grateful to my principal supervisor, Professor Priyan Mendis, for always

believing in me, and for his consistent and generous guidance and assistance, which have
given me absolute privilege and major inspiration to excel in this study. I also thank my
second supervisor, Dr. Tuan Ngo, for his constructive advice. The vibrant enthusiasm of my
third supervisor, Dr. Lihai Zhang, in offering critical ideas is also very much acknowledged.
I reserve my appreciation to the Roads and Maritime Services, in New South Wales, as well
as the Australian Research Council for the sponsorship of this linkage project which has
afforded me a precious opportunity to enhance my personal development, professional
experience, technical knowledge and academic skill to an extended horizon.
Last but not least, the continuous support afforded by my colleagues from the University of
Melbourne as well as all my friends, with special regard to Raymond Lumantarna,
Sudharshan Raman, Andrew McQuillen, Thomas Shaw, Chee Chang Gan, Pong Wee Yeo,
Dobrivoj Bojinovic, etc., is also sincerely valued.
v
TABLE OF CONTENTS
ABSTRACT ...................................................................................................................................I
DECLARATION ......................................................................................................................... III
PREFACE .................................................................................................................................. IV
ACKNOWLEDGEMENTS ..............................................................................................................V
TABLE OF CONTENTS .............................................................................................................. VI
LIST OF TABLES ..................................................................................................................... XIII
LIST OF FIGURES .................................................................................................................... XV
LIST OF ABBREVIATIONS AND ACRONYMS ......................................................................... XXIX
LIST OF NOTATIONS ............................................................................................................ XXXI
CHAPTER 1 INTRODUCTION.......................................................................................... 1
1.1 BACKGROUND AND CONTEXT ..................................................................................... 1
1.2 PROBLEM STATEMENT AND RESEARCH NEEDS ......................................................... 3
1.3 SCOPE, OBJECTIVES AND KEY CONTRIBUTIONS........................................................ 6
1.4 RESEARCH METHODOLOGY ....................................................................................... 9
1.5 THESIS ORGANISATION ............................................................................................. 11
CHAPTER 2 BLAST LOADING ON BRIDGES ............................................................ 13
2.1 OVERVIEW ................................................................................................................. 13
2.2 FUNDAMENTALS OF BLAST PHENOMENON .............................................................. 14
2.2.1 General Characteristics of Explosions ................................................................ 14
2.2.1.1 Explosions ................................................................................................... 14
2.2.1.2 Explosives.................................................................................................... 15
2.2.1.3 Explosion Energy ........................................................................................ 15
2.2.1.4 Formation of Blast Waves ........................................................................... 15
2.2.1.5 TNT Equivalency ........................................................................................ 16
2.2.1.6 Scaling Law ................................................................................................. 17
2.2.2 Blast Loading on Structures ............................................................................... 18
vi
Table of Contents
2.2.2.1 Blast Loading Categories ............................................................................ 18

2.2.2.2 Blast Wave Reflection ................................................................................. 19
2.2.2.3 Blast Wave Diffraction ................................................................................ 20
2.2.2.4 Blast Wave Parameters ................................................................................ 22
2.3 BLAST INVESTIGATION TECHNIQUES ....................................................................... 26
2.3.1 Coupled Methods ................................................................................................ 26
2.3.1.1 Numerical Simulation .................................................................................. 27
2.3.1.2 Experimental Procedure .............................................................................. 29
2.3.2 Uncoupled Methods ............................................................................................ 31
2.3.2.1 Static Analysis ............................................................................................. 32
2.3.2.2 Dynamic Analysis ....................................................................................... 33
2.4 BLAST RESPONSES..................................................................................................... 38
2.4.1 General Blast Effects .......................................................................................... 38
2.4.1.1 Local and Global Behaviours ...................................................................... 38
2.4.1.2 Strain Rate Effects ....................................................................................... 40
2.4.2 Blast Effects on Bridges ..................................................................................... 43
2.4.2.1 Decks ........................................................................................................... 43
2.4.2.2 Girders ......................................................................................................... 44
2.4.2.3 Towers and Piers ......................................................................................... 47
2.4.2.4 Cables and Anchorage Regions ................................................................... 51
2.5 BLAST COUNTERMEASURES...................................................................................... 52
2.5.1 Risk Assessment and Management .................................................................... 52
2.5.2 Security Measures............................................................................................... 53
2.5.3 Standoff Distance ............................................................................................... 54
2.5.4 Structural Geometry ........................................................................................... 55
2.5.5 Structural Dimension .......................................................................................... 55
2.5.6 Material Characteristics ...................................................................................... 56
2.5.7 Retrofitting Options ............................................................................................ 57
2.5.8 Detailing and Connections.................................................................................. 58
2.5.9 System Redundancy ........................................................................................... 62
2.5.10 Performance-Based Design Methodologies ....................................................... 63
vii
Table of Contents
2.6 SUMMARY .................................................................................................................. 64

CHAPTER 3 DISPROPORTIONATE AND PROGRESSIVE COLLAPSE
ANALYSIS ................................................................................................... 69
3.1 OVERVIEW ................................................................................................................. 69
3.2 INVESTIGATION METHODOLOGY ............................................................................. 70
3.3 FINITE ELEMENT MODELLING ................................................................................. 73
3.3.1 Geometries and Dimensions ............................................................................... 73
3.3.2 Elements and Meshes ......................................................................................... 76
3.3.2.1 Beam Elements ............................................................................................ 77
3.3.2.2 Shell Elements ............................................................................................. 78
3.3.3 Constitutive Material Formulations .................................................................... 79
3.3.3.1 Concrete ....................................................................................................... 79
3.3.3.2 Steel ............................................................................................................. 80
3.3.4 Connections ........................................................................................................ 81
3.3.5 Internal and External Loading ............................................................................ 81
3.3.5.1 Gravitational Load ....................................................................................... 81
3.3.5.2 Cable Forces ................................................................................................ 83
3.3.6 Boundary Concepts............................................................................................. 84
3.4 VULNERABILITY ASSESSMENT.................................................................................. 85
3.4.1 Elimination of Intermediate Supports................................................................. 85
3.4.1.1 Removal of Single Mid-Span Support ........................................................ 85
3.4.1.2 Removal of Multiple Main-Span Supports.................................................. 92
3.4.2 Elimination of Side Supports ............................................................................ 100
3.4.2.1 Removal of Single Tie-Down Support ...................................................... 100
3.4.2.2 Removal of Multiple Back-Span Supports ................................................ 107
3.5 SUMMARY ................................................................................................................ 115
CHAPTER 4 BLAST PERFORMANCE OF CABLE ANCHORAGE ZONES ........ 117
4.1 OVERVIEW ............................................................................................................... 117
4.2 NUMERICAL SIMULATION ....................................................................................... 118
4.2.1 Main Tension System, Anchorage Assembly, Ancillary Equipment ............... 118
4.2.1.1 Cables ........................................................................................................ 118
viii
Table of Contents
4.2.1.2 Anchorage Devices .................................................................................... 120

4.2.1.3 Auxiliary Components ............................................................................... 122
4.2.2 Geometries and Dimensions ............................................................................. 122
4.2.3 Elements and Meshes ....................................................................................... 126
4.2.3.1 Beam Elements .......................................................................................... 127
4.2.3.2 Solid Elements ........................................................................................... 127
4.2.4 Constitutive Material Formulations .................................................................. 129
4.2.4.1 Concrete ..................................................................................................... 129
4.2.4.2 Steel ........................................................................................................... 137
4.2.5 Connections ...................................................................................................... 140
4.2.5.1 Constraint Algorithm ................................................................................. 140
4.2.5.2 Node Merging ............................................................................................ 141
4.2.6 Conventional and Extreme Loading ................................................................. 142
4.2.6.1 Gravitational Load ..................................................................................... 142
4.2.6.2 Cable Force ................................................................................................ 142
4.2.6.3 Blast Action ............................................................................................... 143
4.2.7 Boundary Concepts........................................................................................... 146
4.3 CABLE ANCHORAGE ZONES ................................................................................... 147
4.3.1 General Definition ............................................................................................ 147
4.3.1.1 Local Zone ................................................................................................. 148
4.3.1.2 General Zone ............................................................................................. 148
4.3.2 Specific Definition ............................................................................................ 149
4.3.2.1 Stress Information ..................................................................................... 149
4.3.2.2 Force Paths ................................................................................................ 150
4.4 BLAST EFFECTS OF PERSON-BORNE-IMPROVISED-EXPLOSIVE-DEVICES
(PBIEDS) ................................................................................................................. 151
4.5 BLAST EFFECTS OF VEHICLE-BORNE-IMPROVISED-EXPLOSIVE-DEVICES
(VBIEDS) ................................................................................................................ 165
4.6 CRITICAL COMPARISON OF BLAST IMPLICATIONS................................................ 179
4.7 SUMMARY ................................................................................................................ 182
CHAPTER 5 BLAST-RESILIENT RETROFITTING TECHNIQUES...................... 187
ix
Table of Contents
5.1 OVERVIEW ............................................................................................................... 187

5.2 BASIC CREDIBILITY OF PROTECTIVE MATERIALS ................................................ 188
5.3 BLAST FORTIFICATION CONCEPTS ......................................................................... 190
5.4 COMPUTATIONAL MODELLING .............................................................................. 194
5.4.1 Elements and Meshes ....................................................................................... 194
5.4.2 Constitutive Material Formulations .................................................................. 194
5.4.2.1 Steel ........................................................................................................... 194
5.4.2.2 Carbon-Fibre-Reinforced Polymer ............................................................ 195
5.4.3 Contact and Connections .................................................................................. 197
5.5 HARDENING WITH STEEL ........................................................................................ 200
5.5.1 Feasibility of Defence Schemes ....................................................................... 200
5.5.2 Influence of Steel Class .................................................................................... 208
5.5.3 Significance of Plate Section ............................................................................ 211
5.5.4 Importance of Bonding Competency ................................................................ 214
5.5.5 Importance of Connection System ................................................................... 218
5.6 TOUGHENING WITH CARBON-FIBRE-REINFORCED POLYMER ............................. 221
5.6.1 Viability of Defence Plans ................................................................................ 221
5.6.2 Importance of Composite Quality .................................................................... 229
5.6.3 Significance of Layer Quantity......................................................................... 233
5.7 CRUCIAL COMPARISON OF STRENGTHENING ALTERNATIVES .............................. 236
5.8 BLAST RETROFITTING STRATEGIES ....................................................................... 237
5.8.1 General Comments ........................................................................................... 237
5.8.2 Roles and Functions.......................................................................................... 238
5.8.3 Configurations and Forms ................................................................................ 239
5.8.4 Material Selection ............................................................................................. 240
5.8.4.1 Steel Armour ............................................................................................. 241
5.8.4.2 Carbon-Fibre-Reinforced Polymer Layer .................................................. 241
5.8.5 Installation and Connection .............................................................................. 243
5.9 SUMMARY ................................................................................................................ 243
CHAPTER 6 PARAMETRIC STUDY AND BLAST-RESISTANT DESIGN
GUIDELINES ............................................................................................ 247
x
Table of Contents
6.1 OVERVIEW ............................................................................................................... 247

6.2 PARAMETRIC STUDY ............................................................................................... 248
6.2.1 Importance of Standoff Distance ...................................................................... 248
6.2.2 Influence of Concrete Quality .......................................................................... 256
6.2.3 Influence of Reinforcement Quantity ............................................................... 260
6.2.4 Significance of Bearing Plate Size ................................................................... 263
6.2.5 Significance of Anchorage Zone Size .............................................................. 268
6.2.6 Effect of Cable Force ........................................................................................ 274
6.2.7 Influence of Cable Inclination .......................................................................... 278
6.3 GENERAL DISCUSSION OF DESIGN VARIABLES ...................................................... 285
6.4 BLAST DESIGN SPECIFICATIONS ............................................................................. 286
6.4.1 General Concepts .............................................................................................. 286
6.4.1.1 Design Aim ................................................................................................ 286
6.4.1.2 Definition of Anchorage Zone ................................................................... 286
6.4.1.3 Responsibility Allocation .......................................................................... 286
6.4.1.4 Limit States ................................................................................................ 287
6.4.1.5 Adequacy of Conventional Design Codes ................................................. 287
6.4.2 Crucial Advice on Blast Analysis ..................................................................... 287
6.4.3 Advanced Design Strategies ............................................................................. 289
6.4.3.1 Blast Load .................................................................................................. 289
6.4.3.2 Anchorage Load ........................................................................................ 292
6.4.3.3 Material Properties .................................................................................... 294
6.4.3.4 Detailing .................................................................................................... 295
6.4.3.5 Dimensioning ............................................................................................ 296
6.4.3.6 Composite Enhancement ........................................................................... 297
6.4.4 Integrated Design Methodology ....................................................................... 298
6.5 SUMMARY ................................................................................................................ 301
CHAPTER 7 SUMMARY, CONCLUSIONS, RECOMMENDATIONS .................... 303
7.1 PREAMBLE ............................................................................................................... 303
7.2 INVESTIGATION PROGRAMME ................................................................................ 304
7.2.1 General Details ................................................................................................. 304
xi
Table of Contents
7.2.2 Local Blast Simulation ..................................................................................... 305

7.2.3 Global Collapse Analysis ................................................................................. 305
7.3 BASIC FINDINGS ...................................................................................................... 306
7.3.1 Local Blast Response........................................................................................ 306
7.3.2 Global Collapse Mechanism ............................................................................. 309
7.4 PROPOSED SOLUTIONS ............................................................................................ 310
7.4.1 Retrofitting Techniques .................................................................................... 310
7.4.2 Design Guidelines............................................................................................. 313
7.5 FUTURE RESEARCH ................................................................................................. 316
LIST OF REFERENCES ............................................................................................................ 319
APPENDICES
Appendix A Mitigation Strategies for Global Collapse ............................................A-1
Appendix B Analysis and Design of Anchorage Zones ............................................. B-1
xii
LIST OF TABLES
Table 2.1: TNT equivalent factors (UFC 3-340-01) ................................................................ 17

Table 2.2: Blast studies involving bridges ............................................................................... 66
Table 3.1: Element breakdown for the global model ............................................................... 77
Table 3.2: Mechanical properties of concrete and reinforcing steel in the global model ........ 79
Table 3.3: Reinforcement ratios in the global model ............................................................... 80
Table 3.4: Bridge response (due to the elimination of a main-span support) .......................... 92
Table 3.5: Collapse mode (upon the deactivation of intermediate anchorage points) ............. 93
Table 3.6: Bridge response (due to the elimination of a tie-down support) .......................... 107
Table 3.7: Collapse mode (upon the deactivation of side anchorage points) ........................ 108
Table 4.1: Element breakdown for the local model ............................................................... 126
Table 4.2: Self-generated concrete parameters for the local model....................................... 137
Table 4.3: Cowper-Symonds coefficients for steel ................................................................ 140
Table 4.4: Mechanical properties of the steel components in the local model ...................... 140
Table 4.5: Blast scenarios with PBIEDs and VBIEDs .......................................................... 145
Table 4.6: Blast implications due to PBIEDs and VBIEDs ................................................... 181
Table 5.1: Mechanical properties of the steel armour............................................................ 195
Table 5.2: Mechanical properties of CFRP (Tanapornraweekit, 2010) ................................. 196
Table 5.3: Mechanical properties of adhesive agent .............................................................. 198
Table 5.4: Mechanical properties of fasteners ....................................................................... 199
Table 5.5: Mechanical properties of steel .............................................................................. 208
Table 5.6: Armour sizes ......................................................................................................... 211
Table 5.7: Mechanical properties of epoxy resin ................................................................... 214
Table 5.8: Mechanical properties of CFRP of various qualities ............................................ 229
Table 5.9: Composite layers .................................................................................................. 233
Table 5.10: Relative merits of steel and CFRP ...................................................................... 236
Table 5.11: Qualitative appraisal of FRP (Christoffersen et al., 1999) ................................. 242
xiii
List of Tables
Table 6.1: Mechanical properties of concrete ........................................................................ 256

Table 6.2: Reinforcement efforts ........................................................................................... 260
Table 6.3: Basic influence of bearing plate size .................................................................... 264
Table 6.4: Bearing loads and capacities (as affected by the net contact area) ....................... 265
Table 6.5: Blast investigation of the effect of anchor pod size .............................................. 270
Table 6.6: Gravitational actions and cable loads ................................................................... 274
Table 6.7: Intricate effect of cable orientation ....................................................................... 280
Table A.1: Blast threat mitigation ..........................................................................................A-3
xiv
LIST OF FIGURES
Figure 1.1: Research methodology framework ........................................................................ 10

Figure 2.1: Blast wave pressure profile ................................................................................... 16
Figure 2.2: Blast loading due to an unconfined detonation ..................................................... 19
Figure 2.3: Blast wave reflection – (a) No reflection; (b) Normal reflection; (c) Oblique
reflection; (d) Mach reflection ......................................................................... 20
Figure 2.4: Blast wave diffraction effect (Kinney and Graham, 1985) ................................... 21
Figure 2.5: Blast wave parameters of unconfined spherical free air bursts (UFC 3-340-
02)..................................................................................................................... 23
Figure 2.6: Idealised blast wave pressure profile..................................................................... 26
Figure 2.7: Highway bridge model partially enclosed in the air (Yi, 2009) ............................ 28
Figure 2.8: Model boundary of the cable-stayed bridge (Son and Lee, 2011) ........................ 29
Figure 2.9: Euler domains for flow movement control (Son and Lee, 2011) .......................... 29
Figure 2.10: Connection details for the concrete-filled steel tube columns (Fujikura et
al., 2008) ........................................................................................................... 30
Figure 2.11: Blast test setup for the highway bridge piers (Williamson et al., 2011a)............ 31
Figure 2.12: Derivation of the blast load on a circular pier (Winget et al., 2005) ................... 34
Figure 2.13: Probabilistic assessment of blast vulnerability (Yan and Chang, 2009) ............. 36
Figure 2.14: Mesh distributions of the cable-stayed bridge model (Tang and Hao, 2010) ..... 37
Figure 2.15: Blast loading modes – (a) Contact burst; (b) Close-in detonation; (c) Far-
field explosion .................................................................................................. 38
Figure 2.16: Global collapse of a multi-storey building due to removal of columns .............. 39
Figure 2.17: Strain rates due to static and dynamic loading .................................................... 40
Figure 2.18: Strain rate effect on the compressive strength of concrete (Bischoff and
Perry, 1991) ...................................................................................................... 41
Figure 2.19: Strain rate effect on the tensile strength of concrete (Malvar and Ross,
1998)................................................................................................................. 42
xv
List of Figures
Figure 2.20: Strain rate effects on steel ................................................................................... 43

Figure 2.21: Local failure of a deck (Deng and Jin, 2009) ...................................................... 44
Figure 2.22: Transverse deformation of girders (Tokal-Ahmed, 2009) .................................. 45
Figure 2.23: Separation of girders from the piers (Tokal-Ahmed, 2009) ................................ 45
Figure 2.24: Confinement effect between girders (Winget et al., 2005) ................................. 46
Figure 2.25: Shear failure of a reinforced concrete column (Williamson et al., 2010) ........... 47
Figure 2.26: Flexural failure of a concrete-filled steel tube column (Fujikura et al.,
2008)................................................................................................................. 48
Figure 2.27: Local damage on reinforced concrete columns – (a) Breaching; (b)
Spalling (Williamson et al., 2010) ................................................................... 49
Figure 2.28: Lateral buckling of reinforced concrete columns (Tokal-Ahmed, 2009) ............ 50
Figure 2.29: Blast damage on a hollow tower (Tang and Hao, 2010) ..................................... 50
Figure 2.30: Blast impact on a hollow pier (Tang and Hao, 2010) ......................................... 51
Figure 2.31: Risk assessment and management (Williamson and Winget, 2005) ................... 52
Figure 2.32: Fuse System – (a) Enhancement I; (b) Enhancement II (Son, 2008) .................. 58
Figure 2.33: Detailing of discrete hoops according to scaled distance (m/kg1/3) – (a) Z >
1.2; (b) 1.2 ≥ Z > 0.6; (c) Z ≤ 0.6 (Williamson et al., 2010) ......................... 60
Figure 2.34: Fixed connection for simply-supported girders (Tokal-Ahmed, 2009) .............. 62
Figure 2.35: Wall configurations – (a) Intermediate; (b) Continuous (Tokal-Ahmed,
2009)................................................................................................................. 63
Figure 2.36: Ancillary cable system (Tokal-Ahmed, 2009) .................................................... 63
Figure 3.1: Collapse simulation procedure .............................................................................. 71
Figure 3.2: Labelling convention for the cable-stayed bridge ................................................. 72
Figure 3.3: Hypothetical cable-stayed bridge .......................................................................... 73
Figure 3.4: Cross section of the major and minor spans .......................................................... 73
Figure 3.5: A-shaped tower – (a) Cross section view; (b) Elevation view .............................. 75
Figure 3.6: Tower plan view – (a) Hollow section; (b) H-shaped section ............................... 76
Figure 3.7: Finite element constitution of the global model .................................................... 77
Figure 3.8: Beam elements in the global model ...................................................................... 78
Figure 3.9: Shell elements in the global model........................................................................ 78
Figure 3.10: Node sharing in the global model........................................................................ 81
xvi
List of Figures
Figure 3.11: M1600 traffic load (AS 5100.2-2004) ................................................................. 82

Figure 3.12: Traffic load distributions – (a) Case 1; (b) Case 2 .............................................. 83
Figure 3.13: Iterative procedure to determine the initial tensile forces for cables
(Walther et al., 1999)........................................................................................ 84
Figure 3.14: Boundary conditions for the global model .......................................................... 85
Figure 3.15: Targeted mid-span stay ....................................................................................... 85
Figure 3.16: Global response of the cable-stayed bridge in Case 1 (due to the
elimination of a mid-span support) .................................................................. 86
elimination of a mid-span support) .................................................................. 86
Figure 3.18: Main-span displacement (after the withdrawal of a mid-span stay) ................... 87
Figure 3.19: Back-span displacement (after the withdrawal of a mid-span stay) – (a)
Left span; (b) Right span .................................................................................. 88
Figure 3.20: Tower deflection (following the withdrawal of a mid-span cable) – (a)
West tower; (b) East tower ............................................................................... 89
Figure 3.21: Cable forces in Case 1 (after the removal of a mid-span stay) – (a) North
plane; (b) South plane ...................................................................................... 90
plane; (b) South plane ...................................................................................... 91
Figure 3.23: Targeted intermediate anchorage points.............................................................. 93
Figure 3.24: Progressive collapse of the cable-stayed bridge in Case 1 (upon the
deactivation of intermediate anchorage points) ............................................... 95
deactivation of intermediate anchorage points) ............................................... 96
Figure 3.26: Displacement of the supported spans in Case 1 (due to the destruction of
intermediate anchorage regions) ...................................................................... 97
intermediate anchorage regions) ...................................................................... 97
Figure 3.28: Deflection of the towers in Case 1 (upon the failure of centre anchorage
points) ............................................................................................................... 98
xvii
List of Figures
points) ............................................................................................................... 98
Figure 3.30: Cable force histories in Case 1 (following the withdrawal of main-span
supports) ........................................................................................................... 99
supports) ........................................................................................................... 99
Figure 3.32: Targeted tie-down stay ...................................................................................... 100
elimination of a tie-down support) ................................................................. 101
elimination of a tie-down support) ................................................................. 101
Figure 3.35: Main-span displacement (after the withdrawal of a tie-down stay) .................. 102
Figure 3.36: Back-span displacement (after the withdrawal of a tie-down stay) – (a)
Left span; (b) Right span ................................................................................ 103
Figure 3.37: Tower deflection (following the withdrawal of a tie-down cable) – (a)
West tower; (b) East tower ............................................................................. 104
Figure 3.38: Cable forces in Case 1 (after the removal of a tie-down stay) – (a) North
plane; (b) South plane .................................................................................... 105
plane; (b) South plane .................................................................................... 106
Figure 3.40: Targeted side anchorage points ......................................................................... 108
deactivation of side anchorage points) ........................................................... 110
deactivation of side anchorage points) ........................................................... 111
side anchorage regions) .................................................................................. 112
side anchorage regions) .................................................................................. 112
Figure 3.45: Deflection of the towers in Case 1 (upon the failure of end anchorage
points) ............................................................................................................. 113
xviii
List of Figures
points) ............................................................................................................. 113
Figure 3.47: Cable force histories in Case 1 (following the withdrawal of back-span
supports) ......................................................................................................... 114
supports) ......................................................................................................... 114
Figure 4.1: 7-wire strand (Gimsing, 1997) ............................................................................ 119
Figure 4.2: Idealisation of the cable ....................................................................................... 119
Figure 4.3: Conceptual depiction of the anchorage assembly ............................................... 120
Figure 4.4: Idealised anchorage assembly ............................................................................. 121
Figure 4.5: Local model – (a) Cross section view; (b) Elevation view ................................. 123
Figure 4.6: Edge beam – (a) Cross section view; (b) Elevation view.................................... 124
Figure 4.7: Anchor pod – (a) Cross section view; (b) Elevation view .................................. 125
Figure 4.8: Finite element constitution of the local model .................................................... 126
Figure 4.9: Beam elements in the local model ....................................................................... 127
Figure 4.10: Solid elements in the local model...................................................................... 128
Figure 4.11: Initial yield surface, maximum failure surface and residual strength surface
(Malvar et al., 1997) ....................................................................................... 131
Figure 4.12: Triangular shape with smooth corners (Malvar et al., 1997) ............................ 133
Figure 4.13: Pressure-volumetric strain relationship in Tabulated Compaction EOS ........... 136
Figure 4.14: An example of effective plastic strain contours from Mat_072R3 ................... 136
Figure 4.15: Bilinear stress-strain relationship of steel ......................................................... 138
Figure 4.16: Constraint mechanism in the local model ......................................................... 141
Figure 4.17: Node sharing in the local model ........................................................................ 142
Figure 4.18: Blast locations of VBIEDs and PBIEDs – (a) Cross section view; (b)
Elevation view ................................................................................................ 143
Figure 4.19: Safe standoff distances for PBIEDs and VBIEDs (FEMA, 2003) .................... 144
Figure 4.20: Boundary conditions for the local model – (a) Cross section view; (b)
Elevation view ................................................................................................ 147
Figure 4.21: Anchorage region: general zone and local zone ................................................ 148
Figure 4.22: Stress vector arrangement within a typical anchorage region ........................... 149
xix
List of Figures
Figure 4.23: Bursting stress plot for the model anchorage region ......................................... 150
Figure 4.24: Force paths in a common anchorage precinct ................................................... 151
Figure 4.25: Blast response of the cable anchorage zone under 30 kg TNT explosion ......... 152
Figure 4.26: Anchor pod deformation and hollow core distortion (30 kg TNT
explosion) ....................................................................................................... 153
Figure 4.27: Spalling damage on the anchorage region (30 kg TNT explosion) ................... 153
Figure 4.28: Damage distribution on the anchorage precinct (30 kg TNT explosion) .......... 154
Figure 4.29: General reinforcement stress data (30 kg TNT explosion) – (a) Front
section; (b) Rear section ................................................................................. 155
Figure 4.34: Blast effect on the edge beam due to PBIED detonations – (a) 10 kg TNT;
(b) 20 kg TNT; (c) 30 kg TNT ....................................................................... 160
Figure 4.35: Anchorage displacement records (after PBIED detonations)............................ 161
Figure 4.36: Transverse stress distribution (under 10 kg TNT explosion) ............................ 162
Figure 4.39: Mid-span displacement plots (after PBIED detonations) .................................. 164
Figure 4.40: Cable force histories (after PBIED detonations) ............................................... 164
Figure 4.41: Blast response of the cable anchorage zone under 1000 kg TNT explosion ..... 166
Figure 4.42: Damage distribution on the anchorage precinct (1000 kg TNT explosion) ...... 167
Figure 4.43: Spalling damage on the anchorage region (1000 kg TNT explosion) ............... 168
section; (b) Rear section ................................................................................. 169
Figure 4.45: Blast response of the cable anchorage zone under 500 kg TNT explosion ....... 170
Figure 4.46: Damage distribution on the anchorage precinct (500 kg TNT explosion) ........ 171
Figure 4.47: Blast response of the cable anchorage zone under 250 kg TNT explosion ....... 172
Figure 4.48: Damage distribution on the anchorage precinct (250 kg TNT explosion) ........ 173
xx
List of Figures
Figure 4.49: Blast effect on the edge beam due to VBIED detonations – (a) 250 kg
TNT; (b) 500 kg TNT; (c) 1000 kg TNT ....................................................... 174
Figure 4.50: Anchorage displacement records (after VBIED detonations) ........................... 175
Figure 4.51: Transverse stress distribution (under 250 kg TNT explosion) .......................... 176
Figure 4.52: Transverse stress distribution (under 500 kg TNT explosion) .......................... 176
Figure 4.53: Transverse stress distribution (under 1000 kg TNT explosion) ........................ 177
Figure 4.54: Mid-span displacement plots (after VBIED detonations) ................................. 178
Figure 4.55: Cable force histories (after VBIED detonations) .............................................. 178
Figure 5.1: Experimental setup for FRP-protected reinforced concrete slabs
(Tanapornraweekit, 2010) .............................................................................. 188
Figure 5.2: Experimental setup for polyurea-coated reinforced concrete slabs (Raman,
2011)............................................................................................................... 189
Figure 5.3: Experimental setup for polyurea-coated steel plates (Ackland et al., 2013) ....... 190
Figure 5.4: Retrofitting schemes for PBIED detonations – (a) External shield only; (b)
Internal armour only; (c) Integrated solution ................................................. 192
Figure 5.5: Retrofitting schemes for VBIED explosions – (a) Top cover only; (b)
External casing as well as internal lining only; (c) Combined solution ......... 193
Figure 5.6: Shell elements for the retrofitting system............................................................ 194
Figure 5.7: Stress-strain plot of CFRP ................................................................................... 195
Figure 5.8: Contact mechanism ............................................................................................. 197
Figure 5.9: Anchorage movement on account of PBIED detonations (with different
steel strengthening options) ............................................................................ 202
Figure 5.10: Anchorage movement as a result of VBIED explosions (with different
steel strengthening options) ............................................................................ 202
Figure 5.11: Cable force data under PBIED detonations (with different steel hardening
schemes) ......................................................................................................... 203
Figure 5.12: Cable force data following VBIED explosions (with different steel
hardening schemes) ........................................................................................ 203
Figure 5.13: Blast performance of the anchorage precinct confronted with PBIED
detonations (with unique steel retrofitting configurations) ............................ 204
xxi
List of Figures
Figure 5.14: Blast performance of the anchorage precinct subjected to VBIED

explosions (with unique steel retrofitting configurations) ............................. 205
Figure 5.15: General reinforcement stress information from PBIED detonations (with
the best steel protective solution) ................................................................... 206
Figure 5.16: General reinforcement stress information due to VBIED explosions (with
the best steel protective solution) ................................................................... 206
Figure 5.17: Effective stress records after PBIED detonations (with the best steel
fortification plan)............................................................................................ 207
Figure 5.18: Effective stress records produced by VBIED explosions (with the best
steel fortification plan) ................................................................................... 207
Figure 5.19: Anchorage displacement histories after PBIED detonations (with several
steel grades) .................................................................................................... 209
Figure 5.20: Anchorage displacement histories after VBIED explosions (with several
steel grades) .................................................................................................... 209
Figure 5.21: Effective stress information from PBIED detonations (with different steel
grades) ............................................................................................................ 210
Figure 5.22: Effective stress information from VBIED explosions (with different steel
grades) ............................................................................................................ 210
Figure 5.23: Anchorage displacement histories after PBIED detonations (with different
plate dimensions)............................................................................................ 212
Figure 5.24: Anchorage displacement histories after VBIED explosions (with different
Figure 5.25: Effective stress information from PBIED detonations (with several plate
dimensions) .................................................................................................... 213
Figure 5.26: Effective stress information from VBIED explosions (with several plate
dimensions) .................................................................................................... 213
Figure 5.27: Bond failure and contact disablement (under different bond capacities) –
(a) 30 MPa epoxy resin; (b) 60 MPa epoxy resin .......................................... 215
Figure 5.28: Anchorage displacement records following PBIED detonations (with
various bond capacities) ................................................................................. 216
xxii
List of Figures
Figure 5.29: Anchorage displacement records following VBIED explosions (with

various bond capacities) ................................................................................. 216
Figure 5.30: Effective stress data due to PBIED detonations (with distinct bond
capacities) ....................................................................................................... 217
Figure 5.31: Effective stress data due to VBIED explosions (with distinct bond
capacities) ....................................................................................................... 217
distinct connection arrangements) .................................................................. 219
distinct connection arrangements) .................................................................. 219
Figure 5.34: Effective stress data due to PBIED detonations (with various connection
arrangements) ................................................................................................. 220
Figure 5.35: Effective stress data due to VBIED explosions (with various connection
arrangements) ................................................................................................. 220
CFRP strengthening options) ......................................................................... 223
CFRP strengthening options) ......................................................................... 223
Figure 5.38: Cable force data under PBIED detonations (with different CFRP hardening
schemes) ......................................................................................................... 224
Figure 5.39: Cable force data following VBIED explosions (with different CFRP
hardening schemes) ........................................................................................ 224
detonations (with unique CFRP retrofitting configurations) .......................... 225
explosions (with unique CFRP retrofitting configurations) ........................... 226
the best CFRP protective solution)................................................................. 227
the best CFRP protective solution)................................................................. 227
xxiii
List of Figures
Figure 5.44: Maximum principal stress records after PBIED detonations (with the best
CFRP fortification plan) ................................................................................. 228
Figure 5.45: Maximum principal stress records produced by VBIED explosions (with
the best CFRP fortification plan) ................................................................... 228
Figure 5.46: Anchorage displacement traces on account of PBIED detonations (with
CFRP of distinct constitutive compositions) .................................................. 231
Figure 5.47: Anchorage displacement traces on account of VBIED explosions (with
CFRP of distinct constitutive compositions) .................................................. 231
Figure 5.48: Maximum principal stress plots after PBIED detonations (with CFRP of
different constitutive compositions) ............................................................... 232
Figure 5.49: Maximum principal stress plots after VBIED explosions (with CFRP of
different constitutive compositions) ............................................................... 232
several composite layers) ............................................................................... 234
several composite layers) ............................................................................... 234
Figure 5.52: Maximum principal stress plots after PBIED detonations (with various
composite layers)............................................................................................ 235
Figure 5.53: Maximum principal stress plots after VBIED explosions (with various
composite layers)............................................................................................ 235
Figure 5.54: Unique significance of material attributes......................................................... 240
Figure 6.1: Minimum safe standoff distances for PBIEDs and VBIEDs .............................. 248
Figure 6.2: Blast pressure histories (derived from the PBIED at several locations) ............. 249
Figure 6.3: Blast pressure histories (derived from the VBIED at several locations) ............. 249
Figure 6.4: Blast damage (produced from the desired blast threshold for PBIEDs) ............. 251
Figure 6.5: Blast damage (produced from the desired blast threshold for VBIEDs) ............. 251
different scaled distances) .............................................................................. 252
distinct charge centres) ................................................................................... 252
xxiv
List of Figures
Figure 6.8: Cable force histories on account of PBIED detonations (with various
standoff distances) .......................................................................................... 253
Figure 6.9: Cable force histories on account of VBIED explosions (with several
detonation points) ........................................................................................... 253
Figure 6.10: General reinforcement stress data (with the smallest standoff distance
required for PBIEDs) ..................................................................................... 254
Figure 6.11: General reinforcement stress data (with the nearest charge centre
acceptable for VBIEDs) ................................................................................. 254
Figure 6.12: Transverse stress distribution (with the lowest safe scaled distance for
PBIEDs) ......................................................................................................... 255
Figure 6.13: Transverse stress distribution (with the closest detonation point allowable
for VBIEDs) ................................................................................................... 255
Figure 6.14: Strain rate enhancement of compressive strength ............................................. 257
Figure 6.15: Strain rate enhancement of tensile strength ....................................................... 257
Figure 6.16: Anchorage displacement data under PBIED detonations (with different
concrete grades).............................................................................................. 258
Figure 6.17: Cable force histories after PBIED detonations (with different concrete
grades) ............................................................................................................ 258
Figure 6.18: Anchorage displacement data under VBIED explosions (with different
concrete grades).............................................................................................. 259
Figure 6.19: Cable force histories after VBIED explosions (with different concrete
grades) ............................................................................................................ 259
Figure 6.20: Anchorage displacement data under PBIED detonations (with distinct
reinforcement ratios) ...................................................................................... 261
Figure 6.21: Cable force histories after PBIED detonations (with distinct reinforcement
ratios) .............................................................................................................. 261
Figure 6.22: Anchorage displacement data under VBIED explosions (with distinct
reinforcement ratios) ...................................................................................... 262
Figure 6.23: Cable force histories after VBIED explosions (with distinct reinforcement
ratios) .............................................................................................................. 262
Figure 6.24: a/h ratio.............................................................................................................. 263
xxv
List of Figures
Figure 6.25: Effective, gross and net bearing surfaces .......................................................... 263
Figure 6.26: Bursting stress data (due to distinct a/h ratios) ................................................. 264
Figure 6.27: Anchorage displacement due to PBIED detonations (with various bearing
Figure 6.28: Cable force development following PBIED detonations (with various
bearing plate dimensions)............................................................................... 266
Figure 6.29: Anchorage displacement due to VBIED explosions (with various bearing
Figure 6.30: Cable force development following VBIED explosions (with various
bearing plate dimensions)............................................................................... 267
Figure 6.31: Expansion of the anchor pod ............................................................................. 268
Figure 6.32: Correlation of blast intensity with anchor pod dimension ................................ 269
Figure 6.33: Blast pressure histories (on distinct anchor pod sections) ................................. 269
Figure 6.34: Bursting stress data (for different anchor pod sizes) ......................................... 270
Figure 6.35: Anchorage displacement due to PBIED detonations (with several anchor
pod sections) ................................................................................................... 272
Figure 6.36: Cable force development following PBIED detonations (with several
anchor pod sections) ....................................................................................... 272
Figure 6.37: Anchorage displacement due to VBIED explosions (with several anchor
pod sections) ................................................................................................... 273
Figure 6.38: Cable force development following VBIED explosions (with several
anchor pod sections) ....................................................................................... 273
Figure 6.39: Bursting stress data (due to several anchorage loads) ....................................... 275
Figure 6.40: Anchorage displacement plots resulted from PBIED detonations (with
different cable loads) ...................................................................................... 276
Figure 6.41: Cable force records on account of PBIED detonations (with different cable
loads) .............................................................................................................. 276
Figure 6.42: Anchorage displacement plots resulted from VBIED explosions (with
different cable loads) ...................................................................................... 277
Figure 6.43: Cable force records on account of VBIED explosions (with different cable
loads) .............................................................................................................. 277
xxvi
List of Figures
Figure 6.44: Anchorage precincts configured in unique orientations – (a) 27.36°; (b)
47.60° ............................................................................................................. 278
Figure 6.45: Support inclinations and the corresponding cable loads ................................... 279
Figure 6.46: Bursting stress data (due to several support inclinations) ................................. 279
Figure 6.47: Blast response of the cable anchorage zone under PBIED detonations (in
various orientations) ....................................................................................... 281
Figure 6.48: Blast response of the cable anchorage zone under VBIED explosions (in
various orientations) ....................................................................................... 281
Figure 6.49: Damage distribution on the anchorage precinct due to PBIED detonations
(in various orientations) ................................................................................. 282
Figure 6.50: Damage distribution on the anchorage precinct due to VBIED explosions
(in various orientations) ................................................................................. 282
distinct support configurations) ...................................................................... 283
Figure 6.52: Cable force records on account of PBIED detonations (with distinct
support configurations) .................................................................................. 283
distinct support configurations) ...................................................................... 284
Figure 6.54: Cable force records on account of VBIED explosions (with distinct
support configurations) .................................................................................. 284
Figure 6.55: Architectural platform cantilevered from the longitudinal girder – (a)
Cross section view; (b) Plan view .................................................................. 290
Figure 6.56: Potential blast locations depending on cable orientations ................................. 290
Figure 6.57: Defensive barrier and sacrificial protection ...................................................... 291
Figure 6.58: Relief of anchorage load effects – (a) Normal design; (b) Blast design ........... 292
Figure 6.59: Correlation of resultant force with support orientation ..................................... 293
Figure 6.60: Stress-strain curves of concrete ......................................................................... 294
Figure 6.61: Anchorage zone enlargement – (a) Original unit; (b) Orthogonal extension;
(c) Width broadening; (d) Depth elongation .................................................. 297
Figure 6.62: Composite designs – (a) Typical design; (b) Internal lining; (c) External
casing; (d) Combined solution ....................................................................... 298
xxvii
List of Figures
Figure 6.63: Blast design methodology for cable anchorage zones....................................... 299
Figure A.1: Disproportionate and progressive collapse prevention.......................................A-1
Figure A.2: Support failure scenarios ....................................................................................A-4
Figure A.3: Stay positioning – (a) Mega cables; (b) Standard cables ...................................A-6
Figure A.4: Stay systems – (a) Single-plane; (b) Double-plane ............................................A-7
Figure B.1: Symmetrical prism .............................................................................................. B-2
Figure B.2: Strut-and-tie model ............................................................................................. B-2
Figure B.3: Stress vectors for an anchorage zone .................................................................. B-3
Figure B.4: Confinement reinforcement for the local zone – (a) Continuous spiral; (b)
Discrete ties (PTI, 2006) ................................................................................ B-6
xxviii
LIST OF ABBREVIATIONS AND ACRONYMS
AEM applied element method

AFRP aramid-fibre-reinforced polymer
ALE arbitrary Lagrangian-Eulerian
ANFO ammonium nitrate-fuel oil
AUSM advection upstream splitting method
CCTV closed-circuit television
CFD computational fluid dynamics
CFRP carbon-fibre-reinforced polymer
DEM discrete element method
DIF dynamic increase factor
ECC engineered cementitious composite
EOS equation of state
FEA finite element analysis
FEM finite element method
FOSM first-order second-moment
FRP fibre-reinforced polymer
FSI fluid-structure interaction
GFRP glass-fibre-reinforced polymer
HDPE high density polyethylene
MDOF multi-degree-of-freedom
MUSCL monotone upstream-centred schemes for conservation laws
PBIED person-borne-improvised-explosive-device
RL reduced level
SDOF single-degree-of-freedom
SIFCON slurry-infiltrated fibre concrete
SIMCON slurry-infiltrated mat concrete
xxix
List of Abbreviations and Acronyms
TNT trinitrotoluene
VBIED vehicle-borne-improvised-explosive-device
xxx
LIST OF NOTATIONS
A area/effective bearing area

Ab net bearing area
Acore effective restrained core area
Aduct duct area
Ag gross bearing area
As cross sectional area of a reinforcing steel bar
Ash total area of transverse reinforcing bars
a anchorage dimension in the direction of interest
ao speed of sound
a0, a1, a2 maximum failure surface coefficients
a0y initial yield surface cohesion
a1f, a2f residual strength surface coefficients
a1y, a2y initial yield surface coefficients
b waveform parameter
b1 compressive damage scaling exponent
b2 tensile damage scaling exponent
b3 input scalar multiplier
C compressive action
D diameter
DL dead load
dBurst centroid of bursting force
db nominal diameter of a reinforcing bar
E elastic modulus
EX longitudinal modulus
EY transverse modulus
e eccentricity of an anchorage assembly
xxxi
List of Notations
ec compressive fibre failure mode indicator

ed compressive matrix failure mode indicator
ef tensile fibre failure mode indicator
em tensile matrix failure mode indicator
F anchorage force/axial force/gravitational load
Fn normal force
Fs shear force
F0 initial tensile force
fb bonding competency
fc, f’c compressive strength
fci compressive strength of concrete during machine jacking
fcs static compressive strength
fd multiplication factor for incremental damage
fdu dynamic ultimate strength
fdy dynamic yield strength
flat lateral confinement pressure
fn design bearing resistance
fr reinforcement ratio
ft tensile strength
fts static tensile strength
fu ultimate strength
fy yield strength
G shear modulus
h depth or width of a structural cross section
hc core dimension in the direction of interest
hmin minimum thickness
h1 depth and length of a symmetrical prism
ir, ir+ positive reflected impulse
ir- negative reflected impulse
is, is+ positive incident impulse
is- negative incident impulse
xxxii
List of Notations
J2 second invariant of a deviatoric stress tensor

J3 third invariant of a deviatoric stress tensor
K stiffness
Kc concrete property
Km composite property
Kr reinforcement property
kd internal scalar multiplier
L distance/length
LL live load
Lw positive phase wave length
∆L change in length
NFLF normal failure force
NFLS normal failure stress
P, P(t) pressure
Pc pressure cutoff
Pcc concrete confinement contribution
PM(t) Mach stem pressure
Pn nominal load bearing capacity
Po ambient/atmospheric pressure
Pr(t) reflected pressure
Pr, Pr+ positive peak reflected pressure
Pr- negative peak reflected pressure
Prc reinforcement restraining contribution
Ps(t) incident pressure
Pso, Pso+ positive peak incident pressure
-
Pso negative peak incident pressure
Q, q DIF parameters
QTNT specific energy of TNT
QX specific energy of an explosive
qs peak dynamic pressure
R radius/standoff distance
xxxiii
List of Notations
R1 standoff distance from point A

R2 standoff distance from point B
rf strain rate enhancement factor
S shear strength
SFLF shear failure force
SFLS shear failure stress
s pitch of a continuous spiral bar/spacing of discrete hoops
T tensile action
TBurst bursting force
t time/thickness
ta arrival time
to, to+ positive phase duration
to - negative phase duration
tof + fictitious positive phase duration
-
tof fictitious negative phase duration
U, Us incident wave velocity
UM Mach stem velocity
Ur reflected wave velocity
W charge weight
WTNT equivalent mass of TNT
WX mass of an explosive
W1 charge weight at position A
W2 charge weight at position B
Xc longitudinal compressive strength
Xt longitudinal tensile strength
Yc transverse compressive strength
Yt transverse tensile strength
Z scaled distance
α angle of incidence
αfu DIF parameter for ultimate strength
αfy DIF parameter for yield strength
xxxiv
List of Notations
αs, γs DIF parameters for compressive strength

α0 angle of inclination of an anchorage force
βs, δs DIF parameters for tensile strength
ε strain/engineering strain
εf failure strain/fracture strain
εT true strain
εu ultimate elongation
εv volumetric strain
εv,yield volumetric strain at yielding
𝜀̇ strain rate
𝜀̇𝑠 static strain rate
̅̅̅
𝜀𝑝 effective plastic strain
𝑑𝜀̅̅̅𝑝 effective plastic strain increment
η parameter which indicates the relative location of a current surface
η0 calibration factor
θ support inclination
λ modified effective plastic strain
∆λ modified effective plastic strain increment
ν Poisson’s ratio
ρ density
ρs transverse reinforcement ratio
σ stress/engineering stress
σaa stress parallel to fibre direction
σab, σs shear stress
σbb stress normal to fibre direction
σn normal stress
σT true stress
σ0 applied stress
σ1, σ2, σ3 principal stresses
∆σ current surface
∆σ E enhanced value of a failure surface
xxxv
List of Notations
∆σ m maximum failure surface

∆σ r residual strength surface
∆σ y initial yield surface
φ resistance factor
ψ ratio of a tensile meridian to a compressive meridian
xxxvi
Chapter 1 INTRODUCTION
1.1 BACKGROUND AND CONTEXT
Blast-related issues have become extremely important for civil infrastructure, and no longer
confined to petrochemical and military facilities, as proven from the history of this modern
era. Blast incidents can happen under accidental or intentional circumstances, which are both
unpredictable since human behaviour is involved. These blast events could cause critical
injuries along with heavy casualties in addition to disastrous structural failure, thereby giving
rise to detrimental economic and social impacts, both domestically as well as internationally.
Bridges are an integral component of a vibrant transportation system. The intrinsic function
of a bridge is to afford convenient and efficient access to destinations separated by
geographical terrains and artificial obstacles. Apart from that, a high-profile bridge
constructed with cutting-edge technology as well as remarkable aesthetic appeal also serves
as a regional and national landmark (e.g. Sydney Harbour Bridge, Australia). Thus, bridges
are amongst the attractive targets for deliberate attacks with notorious intentions.
Additionally, bridges are also susceptible to accidental explosions, owing to the diverse
categories of users.
The cable-stayed bridge has been proven to be a reputable solution for a vast range of
applications: from light pedestrian weight (e.g. Sundial Bridge in the United States of
America) to heavy traffic loading (e.g. Stonecutters Bridge, China). With notable
aerodynamic stability and high stiffness, cable-stayed bridges are exceptionally attractive for
long-span crossing. The cantilever method (employed for span erection), together with the
1
Chapter 1
self-anchoring approach (which eliminates the necessity of enormous ground anchors), offer
considerable advantages in terms of constructability. Most fascinatingly, flexible
configuration of the inclined cables enables elegant and refreshing architectural profiles to be
achieved (e.g. Octavio Frias de Oliveira Bridge, Brazil).
In cable-stayed bridges, transmission of action to and from the cables happens within the
anchorage regions. An anchorage point failure will adversely affect the local load transfer,
and may lead to complete deactivation of the load-carrying function of the cable involved.
Consequently, the existing load-balancing mechanism will be unfavourably disrupted, and
catastrophic disproportionate and progressive collapse may be subsequently triggered under
severe circumstances. Thus, the well-being of the cable anchorage zones is of paramount
importance.
The anchorage zones positioned along the spans are naturally exposed to accidental blast
occasions on account of the potential close proximity of standard blast sources. On the
contrary, the anchorage points that reside in the towers are generally not accessible to
conventional explosives, but the blast action transmitted via the cables should not be
neglected. Furthermore, the anchorage regions are also commonly perceived to be susceptible
to intentional attacks, owing to their pivotal contribution to the ultimate survival of the entire
bridge. Thus, premium knowledge of the blast effects on the anchorage precincts as well as
the essential defence strategies is imperative.
Blast engineering regarding civil infrastructure has only received rapidly evolving interest in
recent time. More research is being conducted to advance the theoretical and experimental
investigation technology, as well as to enhance the level of understanding of the blast
implications on multistorey buildings, bridges, industrial structures and public facilities. Blast
solutions, which consist of retrofitting options, for existing provisions; design guidelines, for
future services; and deterrent measures, which aim to hinder blast occurrence and lower blast
severity, are under constant development. Nevertheless, many areas in this field, including
most of the aspects regarding the behaviour of the cable anchorage zones in cable-stayed
bridges, are still not well comprehended and demand intensive attention.
2
Introduction
This project is supported by a linkage grant from the Australian Research Council, and the
industry sponsor is the Roads and Maritime Services in New South Wales, which is one of
the largest infrastructure authorities in Australia that manages cable-stayed bridges.
1.2 PROBLEM STATEMENT AND RESEARCH NEEDS
Blast studies concerning the anchorage precincts in cable-stayed bridges appear to be

extremely limited, as no relevant literature has been discovered so far, notwithstanding the
provisional statements that suggested the probable vulnerability. Nevertheless, since blast
research is predominantly conducted for military applications and the associated findings are
usually restricted from the public, the exact situation might not be as indicated according to
the accessible sources, and recognising the actual progress of this engineering domain is
indubitably an incredible challenge.
Until now, a comprehensive understanding of the blast response of the cable anchorage zones
is still not reasonably well established. The effect of blast loading on the basic failure
mechanisms awaits in-depth investigation. The failure modes that arise exclusively due to
blast impact must be revealed along with their interaction with the standard limit states.
Performance criteria appropriate for both qualitative and quantitative blast appraisal should
also be defined.
Complete familiarisation with the decisive factors that govern the blast performance of the
anchorage regions has not been achieved, and a thorough evaluation of the corresponding
influence has not been subsequently completed. Without such information, little can be
known about the potential vulnerability of the individual anchorage points with distinct
integral features.
Also, critical blast locations that will bring about the most severe harm to the anchorage
precincts should also be confirmed, taking into account both accidental and intentional
circumstances.
3
Chapter 1
In-depth study of the broad influence exerted on the overall integrity of the whole bridge,
following the destruction of the anchorage regions, is strictly lacking. Moreover, the crucial
anchorage points that are competent to initiate disastrous collapse upon deactivation should
be identified.
Without extensive information on the blast consequences, viable strengthening concepts for
the anchorage regions in existing cable-stayed bridges are not able to be formulated. No
original innovations tailored specifically to the anchorage zones have been presented, and the
transferability of previously developed retrofitting measures has not been determined. No
retrofitting guidelines can be outlined at present to assist with the delivery of fortification
plans.
Furthermore, the feasibility of steel and fibre-reinforced polymer (FRP) as blast hardening
options for ordinary components (e.g. beams, columns, slabs, walls) has been widely
acknowledged, but remains unclear when anchorage zones are concerned. The underlying
uncertainties are not only attributable to the intrinsic characteristics of the protective
materials, but also lie within their practical forms of application. Additionally, without proper
appreciation of their relative merits, successful implementation of the most efficient and
effective strengthening schemes is often not assured.
Extreme blast loading is normally not considered in standard bridge designs. This is actually
motivated by the fact that blast incidents typically have relatively low probability of
occurrence during the service life of civil infrastructure, and therefore often not able to justify
the need for robust yet costly blast designs. Now, with hardening solutions remain
unavailable, the anchorage regions in existing cable-stayed bridges in operation are left with
no choice but to persist precariously in functioning under the severe risk of devastating blast
damage, especially when blast threat is becoming increasingly relevant in this modern era.
So far, no prescriptive and performance-based design recommendations have been proposed

for the conceivable incorporation of blast action into the future design of the cable anchorage
4
Introduction
zones. Apart from that, the adequacy of the various contemporary design specifications
dedicated to the conventional loading has not been carefully assessed, and, regrettably, no
suggestions can be made at this stage even if essential revisions are indeed indispensable and
urgent.
Besides enhancing the local resistance, general ideas that are capable of assuring a more
promising chance of either avoiding or surviving devastating global destruction should also
be put forward.
If blast provisions are not integrated within the anchorage precincts in new cable-stayed
bridges, both external fortification and internal upgrade in future will entail considerable
modification or reconstruction of the existing systems, thereby instigating undesirable
interruption to the routine operation of the affected facilities as well as the entire
transportation network.
The traditional theoretical approaches, which are incapable of handling awkward loading
patterns, generating substantial non-uniform damage which inflicts extremely chaotic
disruption to the original flow of action, managing highly non-linear behaviour, incorporating
complicated failure styles, as well as dealing with fluctuating anchorage loads, are indeed not
appropriate for impulsive blast impact.
Unfortunately, sophisticated computational models worthy of being employed for blast

investigation in this case are yet to be devised. The main challenges encountered include:
realistic estimation of the blast action, accurate representation of the anchorage zones (which
are heavily congested with steel reinforcement) together with the adopted strengthening
schemes, satisfactorily accounting for the complicated features and the key functions of the
cables and the anchorage devices, as well as successfully overcoming the shortcomings of the
current techniques.
5
Chapter 1
1.3 SCOPE, OBJECTIVES AND KEY CONTRIBUTIONS
This project involves a typical cable-stayed bridge intended for standard vehicular traffic at
normal-span crossing. A hypothetical bridge, instead of an actual one, is employed, owing to
the issues involved with the use of confidential and sensitive information and data in such
context due to security concerns. The anchorage zones located within the deck-and-girder
system are naturally vulnerable to direct blast impact; the anchorage points accommodated at
the elevated height of the towers are assumed to be susceptible only to the indirect blast
action transmitted via the cables.
Here, concrete (both reinforced and prestressed), which is widely utilised in civil
infrastructure, is taken as the construction material of interest. Within concrete units, load
transfer to and from the anchorage zones is most commonly achieved by means of direct
bearing support whereby special anchorage devices are brought into play, and such
mechanism (analogous to the concept practised in the post-tensioned system) is
commissioned in this case.
Blast incidents instigated either inadvertently or deliberately, involving vehicle-borne-

improvised-explosive-devices (VBIEDs) and person-borne-improvised-explosive-devices
(PBIEDs), containing trinitrotoluene (TNT), are considered. Blast locations that give rise to
the worst case scenarios will be indicated. The blast consequences to be anticipated from both
vehicular explosives and portable charges will be compared in order to derive a more holistic
understanding.
Only air-blast wave loading, which is the main concern in blast engineering, is taken into
account here, while the effect of heat radiation and high-velocity flying projectiles is
considered to be beyond the scope of this project.
The primary focus of this research is to investigate the fundamental blast implications on the
cable anchorage zones, under the concurrent presence of the conventional loading. Aside
from assessing the blast effects on the standard limit states, any additional failure modes
6
Introduction
noticed are to be made aware. Blast assessment criteria that are applicable in this context will
be highlighted.
Identifying the critical parameters that control the blast outcome generated on the anchorage
regions, along with their respective weight, is indeed imperative because this not only
provides valuable information about the blast resistance of the separate anchorage points
positioned across the whole bridge, but also enables the individual criteria to be manipulated
exactly to the desired functioning status.
Establishing general blast design specifications for the anchorage zones is also an important
goal of this study. The design solutions are intended to be incorporated conveniently and
pragmatically into the design schemes that adhere closely to the performance-based design
principle, in addition to the standard requirements of the traditional limit states. Furthermore,
the adequacy of the design provisions stipulated in the current design codes relevant to
anchorage zones will be verified, and any amendments found necessary are to be firmly
stated.
Developing advanced blast retrofitting techniques for the anchorage precincts is another
significant purpose of this investigation. The key variables of concern will be highlighted,
and their potential influence is determined. General retrofitting guidelines are formulated to
ensure satisfactory implementation of the hardening schemes.
Furthermore, this exercise also aims to expand the blast applications of steel and carbon-
fibre-reinforced polymer (CFRP), which are widely recognised in this engineering discipline,
each with distinct characteristics, and therefore not able to substitute one another and stand
out as the sole preference. The idea of attempting several protective materials here is to
ensure that possible options are available to satisfy the specific requirements emerged in
different situations. A solid confirmation of the relative strengths and weaknesses is also
sought after.
7
Chapter 1
Disproportionate collapse refers to a major effect instigated by relatively minor damage;

progressive collapse happens when successive failure followed the initial cause. It is
reasonable to claim that the ultimate objective of this study is none other than to prevent
disastrous disproportionate and progressive collapse of cable-stayed bridges. Thus, the
overall consequences initiated due to sudden deactivation of the anchorage points should be
determined. Highlighting the key anchorage locations that deserve extra attention is
exceptionally vital for resource allocation. These findings are very important for the
development of global-scale mitigation strategies.
A sophisticated numerical modelling approach, which adheres to the finite element method
(FEM), will be devised to fulfil the needs of this project. The main emphasis resides in
deciding on the suitable manner to derive the blast impact, together with discovering the
appropriate means that enable the structural response to be realistically captured. This
advanced simulation technique is essential to justify the existence as well as the significance
of this assignment.
Both existing and future cable-stayed bridges, which technically call for distinct defence
philosophies, are adequately catered for. Here, viable solutions, which provoke minimal
disturbance to the routine operation of current services, are offered, and feasible ideas for the
initial establishment of new facilities are put forward. Most importantly, this research wishes
not only to deliver original innovations, but also to verify the transferability of the current
concepts.
The far-reaching contributions afforded by this project extend broadly across multiple
disciplines, including blast and structural engineering, advanced material science, and
computational mechanics. Strictly speaking, it is acceptable to claim this study as a ground-
breaking blast research for the cable anchorage zones in cable-stayed bridges. Although
cable-stayed bridges enjoy the direct benefits, the overarching ideas presented in this
assignment can also be treated with a generic sense, and are, to a certain extent, transferable
to similar structures with cable systems, such as suspension bridges and cable-supported
roofs.
8
Introduction
1.4 RESEARCH METHODOLOGY
This project comprises three principal phases: the initiation stage, the delivery stage and the
closing stage. The logical sequence of the major activities constituting this work is explained
in Figure 1.1.
A comprehensive literature review will not only provide the fundamental knowledge required
and offer profound insights indispensable for the success of this research, but also facilitate
deeper appreciation of this study. Here, the literature survey will focus on the following
aspects: basic blast theories, blast implications particularly on bridges, collapse mechanisms
of cable-suspended bridges, available mitigation strategies as well as current investigation
techniques.
LS-DYNA, a commercially developed general-purpose finite element software capable of

dealing with sophisticated real-world problems, is employed for the finite element analysis
(FEA) in this study. LS-DYNA incorporates both advanced implicit and explicit solvers, and
relies on a virtual library well equipped with a vast collection of high-level modelling
features. This software has been extensively utilised for blast simulation and collapse
analysis, which technically entail extremely complicated non-linear dynamic computational
procedures.
The investigation process consists of two distinct parallel phases: global collapse analysis
(which is restricted only to demonstrate the criticality of anchorage point failures, as the
overall global behaviour under extreme action is actually the focus of the complementary
project completed in the University of Technology, Sydney), which is based on a full-size
hypothetical cable-stayed bridge; and local blast simulation, which relies solely on a full-
scale prototype extracted judiciously from the imaginary bridge. This particular approach is
devised with the purpose to ensure more notable computational efficiency whilst not
compromising the accuracy levels satisfactory to the specific research intentions.
9
Chapter 1
Figure 1.1: Research methodology framework
The preliminary analysis serves as an initial screening process which singles out the crucial
control factors in the local simulation and the global analysis for more in-depth examination,
aside from playing an important role during the conceptual development of the retrofitting
configurations. This preparation stage also assists in the determination of the conventional
10
Introduction
and blast loading setups, and affords an additional platform to aid the establishment of the
computational approach.
1.5 THESIS ORGANISATION
The contents of this dissertation are systematically presented in a total of seven chapters, as
follows:
Chapter 1 introduces the fundamental motivation that initiated this project, reveals the current
state of knowledge of this engineering topic, clarifies the overall focus and the anticipated
accomplishments of this research together with an explanation of the principal logistics
employed, highlights the significance of this investigation, as well as delivers an outline of
the contents contained within this thesis.
Chapter 2 covers the basic knowledge of blast engineering and contains a critical review of
the blast studies conducted specifically for bridges. A brief introduction of the general
characteristics and the loading mechanisms of explosions as well as the standard blast
consequences observed on civil infrastructure is accompanied by a broad discussion of the
investigation techniques applied to, the wide range of blast outcomes generated on, and the
whole list of mitigation strategies proposed for bridges.
Chapter 3 is reserved for the disproportionate and progressive collapse analysis. A full
description of the abstract cable-stayed bridge employed together with the conventional
loading considered is provided along with a detailed account of the finite element modelling
strategies commissioned to create the global model. An assessment of the overall condition of
the entire bridge, immediately after the elimination of the anchorage points with different
roles, is undertaken, and the collapse mechanisms observed are explained.
Chapter 4 gives extensive coverage of the local blast response of the cable anchorage zones
generated by both PBIEDs and VBIEDs, using both qualitative and quantitative performance
indicators. In order to clarify the initial conditions of the anchorage regions, a preliminary
11
Chapter 1
static analysis is also included. The vital information covered here will serve as a basis to
justify the defence strategies offered. The structural features that constitute the local model
are elaborated together with the finite element modelling approach employed, and the blast
scenarios involved are clarified.
Chapter 5 is dedicated purely to existing facilities. A thorough evaluation of the feasibility of

the strengthening concepts put forward is completed before the intensive sensitivity analysis
which enables the effectiveness of the protective measures to be subsequently highlighted.
The end product is a set of specifications which will benefit the delivery of the desired
fortification plans.
Chapter 6 is concerned primarily with future services. A rigorous parametric study is

conducted, followed by a critical comparison of the authority of the design variables
involved. Advanced design recommendations compatible with the design specifications
stipulated in the current codes are developed in accordance with the performance-based
design principle.
Chapter 7, the final chapter, concludes the project with general closing remarks, a concise
summary of all the research findings from the blast enquiry and the collapse analysis, a brief
recapitulation of the retrofitting techniques and the design guidelines, as well as a succinct
discussion of the potential research work that could be undertaken in future to further mature
this particular topic.
12
Chapter 2 BLAST LOADING ON BRIDGES
2.1 OVERVIEW
Blast study encompasses knowledge from a wide range of disciplines, including blast-related
physical and applied sciences, advanced material science and manufacturing engineering, as
well as structural and geotechnical engineering.
Within this chapter, the general attributes and the intrinsic loading characteristics of
explosions are introduced together with the conventional blast effects on civil infrastructure.
Condensed phase explosions derived from high explosives contained within PBIEDs and
VBIEDs are the primary focus, and the contents covered are mainly related to the air-blast
wave loading due to unconfined detonations.
A state-of-the-art review of the previous research projects concerning bridges is also

presented in this chapter. The simplistic and advanced theoretical and practical investigation
strategies employed are explained, and the distinct assumptions and justifications adopted are
highlighted. The blast consequences observed on the individual bridge elements considered
are described together with the subsequent impact on the overall integrity of the
corresponding bridge systems. The mitigation tactics proposed are discussed, with major
emphasis on the functions delivered as well as the inherent effectiveness.
Blast studies broadly differ in terms of objective, problem encountered, methodology,

expertise and technology required, as well as budgetary restriction and schedule constraint.
Consequently, the research findings derived under the discretion of separate investigations,
13
Chapter 2
which are either valid exclusively to the specific issues under consideration or justifiable
generically to much broader aspects, are not necessarily subject to direct comparison and
should always be interpreted independently and perceived as either mutually reinforcing or
disagreeing to a certain extent, while absolute generalisation should, however, never be
overly enforced.
2.2 FUNDAMENTALS OF BLAST PHENOMENON
2.2.1 General Characteristics of Explosions
2.2.1.1 Explosions
An explosion is a phenomenon where energy is released rapidly within an infinitesimal

period of time (Baker et al., 1983). According to Smith and Hetherington (1994), explosions
broadly fall into three main categories: physical events (e.g. volcanic eruptions), nuclear
reactions (i.e. fusion and fission) and chemical reactions (due to oxidation of explosives).
The ASCE Task Committee on Blast-Resistant Design (2010) identified four types of
chemical explosions: vapour cloud explosions (instigated by ignition of flammable vapour),
pressure vessel explosions (which involve pressurised gases and liquids), condensed phase
explosions (due to decomposition of solid and liquid explosives) and dust explosions (with
combustible dust).
Chemical explosions can take place through the process of detonation and deflagration
(Kinney and Graham, 1985). The former occurs at a supersonic speed when an explosive is
triggered by a shock incentive while the latter happens at a subsonic rate when an explosive is
decomposed by a thermal reaction.
14
Blast Loading on Bridges
2.2.1.2 Explosives
Baker et al. (1983) described four main classifications of explosives based on sensitivity to
initiation. Unstable substances (e.g. nitrogen trichloride) are able to instigate explosions as
long as enough quantity is present, and no detonators are needed. Primary explosives (e.g.
lead azide) are sensitive to stimuli such as spark and flame, but initiating secondary
explosives (e.g. TNT) is comparatively harder. Propellants, which are low explosives, are
commonly ignited by primers.
Alternatively, explosives may also be classified according to their physical state. Solid
explosives are mostly high explosives, while liquid and gaseous explosives cover a huge
range of substances utilised in chemicals, fuels and propellants (UFC 3-340-02).
2.2.1.3 Explosion Energy
The energy of an air-blast is released in the form of a blast wave, ground shock, fire and heat
radiation, as well as flying projectiles. Structural damage is primarily due to the action of
blast waves despite the destructive nature of fire and heat, while the impact of ground shock
instigated by shock waves travelling within the ground or transmitted via the air is
comparatively less significant. Flying fragments with high velocity are largely responsible for
inflicting heavy casualties aside from serious injuries (UFC 3-340-02).
2.2.1.4 Formation of Blast Waves
After an explosion is triggered, gases with elevated temperature and immense pressure are
produced. Owing to the violent expansion of the gases, the air in the immediate vicinity is
forced out of the occupied volume and travel outwards from the charge centre. This highly
compressed air layer is known as the blast wave. The pressure of the gases diminishes with
time and drops beyond the atmospheric pressure because of the residual momentum which
leads to over-expansion. This gives rise to a state of partial vacuum which subsequently
15
Chapter 2
provokes a reversal of flow (i.e. suction). Equilibrium will be attained eventually once the
ambient pressure is reinstated.
A typical blast pressure history profile is displayed in Figure 2.1. After the positive peak
incident pressure, Pso+, is observed at the arrival time, ta, the blast wave pressure decays until
the negative peak incident pressure, Pso-, is reached, and then gradually rises until the ambient
pressure, Po, is finally restored. The areas enclosed by the curve throughout the positive
phase duration, to+, and the negative phase duration, to-, denote the positive impulse, is+, and
the negative impulse, is-, respectively.
Figure 2.1: Blast wave pressure profile
2.2.1.5 TNT Equivalency
Charge weights are commonly expressed in terms of equivalent mass of TNT, by adhering to
the principle of conservation of energy:
𝑊𝑇𝑁𝑇 𝑄𝑇𝑁𝑇 = 𝑊𝑋 𝑄𝑋 (2.1)
16
where:
WTNT = equivalent mass of TNT
QTNT = specific energy of TNT
WX = mass of the explosive of interest
QX = specific energy of the explosive of interest
This standard conversion is a universal practice to enable a more convenient indication of the
relative intensities of the explosions derived from different explosives. Distinct TNT
equivalent factors have been proposed for blast pressure and impulse, as shown in Table 2.1.
Table 2.1: TNT equivalent factors (UFC 3-340-01)

Explosive Equivalent pressure factor Equivalent impulse factor
Composition B 1.11 0.98
Composition C4 1.37 1.19
HBX-1 1.17 1.16
HBX-3 1.14 0.97
Pentolite 1.42 1.00
PETN 1.27 -
Tetryl 1.07 -
Tritonal 1.07 0.96
However, specifying the TNT equivalence of certain explosives, such as ammonium nitrate-
fuel oil (ANFO), could be difficult because of the dissimilar formulations and the inconsistent
quality control procedures adopted by different manufacturers (Smith and Cormie, 2009).
2.2.1.6 Scaling Law
The scaled distance, Z, has been universally accepted as the standard means of blast
quantification:
17
Chapter 2
𝑅
𝑍= 1 (2.2)
𝑊3
where:
R = standoff distance
W = charge weight
For two charges of the same type and geometry, the relationship between the charge weight
and the standoff distance for identical blast wave parameters to be produced at a specific
location is established by this cube root scaling law (Hopkinson, 1915; Cranz, 1926):
1
𝑅1 𝑊1 3 (2.3)
=( )
𝑅2 𝑊2
where:
R1 = standoff distance from point A
R2 = standoff distance from point B
W1 = charge weight at position A
W2 = charge weight at position B
2.2.2 Blast Loading on Structures
2.2.2.1 Blast Loading Categories
The blast loading categories described according to confinement extent and loading style are
(UFC-3-340-02):
 unconfined explosions – free air bursts, air bursts (reflected waves from the
ground arrive before the incident waves) and surface bursts (ground reflections
amplify the initial waves)
18
 confined explosions – fully vented explosions, partially confined explosions and

fully confined explosions
The different forms of blast loading derived from an unconfined detonation are presented in
Figure 2.2.
Figure 2.2: Blast loading due to an unconfined detonation
2.2.2.2 Blast Wave Reflection
When an obstruction is encountered, a blast wave will undergo reflection if its angle of
incidence, 𝛼, happens to be less than 90°. The object will be subjected to a reflected pressure
as a consequence, even though the reflected wave travels away from it, and this is essentially
consistent with Newton’s laws of motion (Smith and Cormie, 2009).
The various forms of blast wave reflection illustrated in Figure 2.3 are:
a) no reflection at 𝛼 = 90° (only side-on pressure is exerted on the parallel surface)

b) normal reflection at 𝛼 = 0°
c) oblique reflection at 𝛼 < 40°
19
Chapter 2
d) Mach reflection at 𝛼 > 40° (the reflected wave converges with the incident wave,
thus creating a third wave called the Mach stem)
(a) (b)
(c) (d)
Figure 2.3: Blast wave reflection – (a) No reflection; (b) Normal reflection; (c) Oblique
reflection; (d) Mach reflection
2.2.2.3 Blast Wave Diffraction
A blast wave will also propagate or diffract around an obstacle. If the blast wave is
sufficiently large, the target will be fully engulfed in blast effect and could be shifted away by
the drag action; otherwise, sequential loading on its constitutive units could happen (Smith
and Cormie, 2009).
A reflected wave will seek relief to the free edges which experience lower pressure, by
forming rarefaction waves which travel from the low pressure regions to the high pressure
regions. The reflected pressure will decay until the stagnation pressure is reached. This entire
process is known as blast wave clearing.
20
Figure 2.4: Blast wave diffraction effect (Kinney and Graham, 1985)
21
Chapter 2
The effect of blast wave diffraction is shown in Figure 2.4. Even though the explosion was
initiated from the right, the destruction of the front face of the building is clearly visible.
Although all features were engulfed by the blast wave, the house and the tree still remain at
their original positions, but the roof tiles together with the windows (i.e. the weaker elements)
are completely shattered and displaced.
2.2.2.4 Blast Wave Parameters
The empirical equations developed by Kingery and Bulmash (1984) to determine blast wave
parameters are normally either presented in the form of charts (as shown in Figure 2.5 for
example), or incorporated into various programs, such as ConWep 1 (Hyde, 1993). The
formulae are applicable from 0.136 ft/lb1/3 (0.054 m/kg1/3), which corresponds to the surface
of an explosive, and only extend to 100 ft/lb1/3 (40 m/kg1/3), since beyond this limit blast
impact becomes negligible and blast waves are severely affected by the atmospheric
conditions (UFC 3-340-02). Other forms of empirical formulae are such as those given by
Wu and Hao (2005).
Analytical expressions have also been proposed. Theoretical simplifications are often adopted
to reduce the complexity of computational procedures, especially when preliminary or
conservative solutions are sought after.
The blast wave pressure profile can be derived from the Friedlander equation:
𝑡 𝑏𝑡
𝑃(𝑡) = 𝑃𝑠𝑜 (1 − ) exp (− ) (2.4)
𝑡𝑜 𝑡𝑜
where:
b = waveform parameter
1
Conventional Weapons Effects, a program intended for blast load estimation, based on the application of UFC
3-340-01
22
Figure 2.5: Blast wave parameters of unconfined spherical free air bursts (UFC 3-340-
02)
23
Chapter 2
Integrating the above equation at the positive and negative phases will yield the
corresponding impulses:
𝑡𝑎 +𝑡𝑜 +
𝑖𝑠 + = ∫ 𝑃(𝑡)𝑑𝑡 (2.5)
𝑡𝑎
𝑡𝑎 +𝑡𝑜 + +𝑡𝑜 −
𝑖𝑠 − = ∫ 𝑃(𝑡)𝑑𝑡 (2.6)
𝑡𝑎 +𝑡𝑜 +
Brode (1955) defined the peak pressure in the positive phase as:
6.7
𝑃𝑠𝑜 = +1 (Pso > 10 bar) (2.7)
𝑍3
0.975 1.455 5.85
𝑃𝑠𝑜 = + + 3 − 0.019 (0.1 bar < Pso < 10 bar) (2.8)
𝑍 𝑍2 𝑍
For negative phase:
0.35
𝑃𝑠𝑜 − = − bar (Z > 1.6) (2.9)
𝑍
1
𝑖𝑠 − ≈ 𝑖𝑠 (1 − ) (2.10)
2𝑍
Alternatively, Henrych (1979) proposed the following:
14.072 5.540 0.357 0.00625

𝑃𝑠𝑜 = + − + bar (0.05 ≤ Z < 0.3) (2.11)
𝑍 𝑍2 𝑍3 𝑍4
6.194 0.326 2.132
𝑃𝑠𝑜 = − + bar (0.3 ≤ Z < 1.0) (2.12)
𝑍 𝑍2 𝑍3
0.662 4.05 3.288
𝑃𝑠𝑜 = + 2 + bar (1 ≤ Z < 10) (2.13)
𝑍 𝑍 𝑍3
24
According to Bangash (2009), the blast wave velocity, the peak reflected pressure, together
with the peak dynamic pressure, qs, can be derived from the Rankine-Hugoniot relations:
6𝑃𝑠𝑜 + 7𝑃𝑜
𝑈𝑠 = 𝑎𝑜 √ (2.14)
7𝑃𝑜
7𝑃𝑜 + 4𝑃𝑠𝑜
𝑃𝑟 = 2𝑃𝑠𝑜 ( ) (2.15)
7𝑃𝑜 + 𝑃𝑠𝑜
5𝑃𝑠𝑜 2
𝑞𝑠 = (2.16)
2(𝑃𝑠𝑜 + 7𝑃𝑜 )
where:
ao = speed of sound
For the sake of simplicity, it is a common practice to assume a linear blast wave pressure
profile (refer to Figure 2.6) in analysis and design. The reflected impulses are preserved
through the introduction of fictitious durations:
1 + +
𝑖𝑟 + = 𝑃 𝑡 (2.17)
2 𝑟 𝑜𝑓
1 − −
𝑖𝑟 − = 𝑃 𝑡 (2.18)
2 𝑟 𝑜𝑓
where:
ir + = positive reflected impulse
+
Pr = peak positive reflected pressure
tof + = fictitious positive phase duration
ir - = negative reflected impulse
Pr - = peak negative reflected pressure
tof - = fictitious negative phase duration
25
Chapter 2
Figure 2.6: Idealised blast wave pressure profile
2.3 BLAST INVESTIGATION TECHNIQUES
Blast effects on bridges can be generated through the coupled methods, i.e. fluid-structure
interaction (FSI) simulation and the experimental approach, whereby the interaction between
the blast wave and the structure is taken into account, and the uncoupled methods, whereby
blast loads are estimated separately before being imposed either statically or dynamically on
the targets.
2.3.1 Coupled Methods
In the FSI procedure, the structure is modelled as a multi-degree-of-freedom (MDOF) system

with Lagrangian elements, while the air and the explosive are explicitly defined using
Eulerian elements. The recent development of the arbitrary Lagrangian-Eulerian (ALE)
formulation, which permits multiple substances within a single element, has resolved the
26
shortcomings of the Lagrangian approach2 and the Eulerian approach3 in computational fluid
dynamics (CFD) (Alia and Souli, 2006).
In blast trials, only the features of interest are usually presented (the prototypes are often
scaled down from the original sizes), and this is acceptable as long as sufficient details are
provided to fully capture the desired effects and attain realistic boundary conditions from the
reaction frames (Ray, 2006). The publication of sensitive data is usually restricted on account
of security concerns.
2.3.1.1 Numerical Simulation
Cimo (2007) created a two-span simply-supported composite girder bridge (a two-

dimensional model), using ANSYS AUTODYN 4 (ANSYS, 2010). A faster computation
speed was attained by remapping the blast pressures obtained from the one-dimensional
analysis onto the two-dimensional space. Unfortunately, the numerical outputs were only
recorded up to the instant when the computation was prematurely terminated because of an
unexpected interface-overlapping error.
Son (2008) modelled a steel orthotropic deck and a composite girder with MSC.Dytran 5
(MSC Software, 2002). Different levels of axial loads were assigned to replicate the distinct
conditions in the earth-anchored suspension bridge, the cable-stayed bridge and the self-
anchored suspension bridge. The nodal displacements recorded away from the blast impact
were taken to represent the global behaviour of the deck and the girder. The subsequent
heating effect was ignored in the study.
2
The matter is allowed to move together with the mesh, but loss of accuracy might occur during excessive
distortion, and the time step might decrease dramatically.
3
The matter is permitted to travel across the mesh which is fixed in position, but interaction and mixing of
different substances are unable to be handled.
4
ANSYS AUTODYN, a versatile explicit analysis tool for modelling the non-linear dynamics of solids, fluids
and gases
5
MSC.Dytran, an explicit FEA solution for analysing non-linear problems with extreme deformations
27
Chapter 2
With the aid of actual design drawings, Deng and Jin (2009) simulated a cable-stayed bridge
with ANSYS AUTODYN. The blast pressures derived from the one-dimensional analysis
were remapped into the final three-dimensional domain, and fine meshes were created only
for the air in the immediate vicinity of the structure, as an attempt to reduce the processing
duration.
A three-span simply-supported composite girder bridge model, as displayed in Figure 2.7,

was developed by Yi (2009), using LS-DYNA. Without modelling the explosive, the blast
pressures estimated from ConWep were applied directly onto the receiving faces of the
ambient elements. The desired mesh sizes were selected by reproducing the test data reported
by Magnusson and Hallgren (2004).
Figure 2.7: Highway bridge model partially enclosed in the air (Yi, 2009)
The study undertaken by Son and Lee (2011), which was related to an existing cable-stayed
bridge, involved a concrete-filled composite pylon and a hollow steel pylon. The FSI
simulations were implemented in MD Nastran SOL700 6 (MSC Software, 2009). Only the
tower and the steel orthotropic deck were extracted for the blast simulation (refer to Figure
2.8), while the absence of the cables was compensated by introducing axial compressive
forces and allowing only horizontal translation at the longitudinal ends. A total of four Euler
subdivisions were formed, as shown in Figure 2.9, in order to govern the flow movement
activated upon structural failure.
6
MD Nastran Explicit Nonlinear (SOL 700), a powerful explicit solution to simulate complicated FSI problems,
where MSC.Dytran is operated in conjunction with LS-DYNA
28
Figure 2.8: Model boundary of the cable-stayed bridge (Son and Lee, 2011)
Figure 2.9: Euler domains for flow movement control (Son and Lee, 2011)
2.3.1.2 Experimental Procedure
The investigation programme devised by Ray (2006) comprised a series of ½-scale blast
experiments and parallel numerical simulations involving steel towers, which focussed on the
following parameters: charge weight, span length of steel plate, compressibility of retrofitting
matter, axial load intensity, presence of clip angle, degree of fragmentation, steel quality and
type of fastener. Only individual steel components were included in Phase I; complex partial
models of the steel towers were incorporated in Phase II.
29
Chapter 2
A total of six ¼-scale circular concrete-filled steel tube columns, capable of withstanding
both blast and seismic action, were tested by Fujikura et al. (2008). The reaction frame setup
ensured a pinned condition at the top, whereas the base was secured with the steel plates and
the vertical channels embedded within the foundation (refer to Figure 2.10). The exact scaled
distances adopted to produce the elastic reactions, the maximum deformations and the
ultimate failures were, however, not revealed.
Figure 2.10: Connection details for the concrete-filled steel tube columns (Fujikura et
al., 2008)
Matthews (2008) reported two blast tests which involved full-scale prestressed concrete
girders. The reaction structures at both ends were intended to restrain uplift resulted from the
below-girder explosion, but treated as simple supports in the above-girder blast scenario.
Blast intensities just enough to induce ultimate responses were generated, and this ensured a
meaningful impact assessment.
Fujikura and Bruneau (2011) conducted blast experiments on two non-ductile reinforced
concrete piers surrounded with steel jackets and two ductile reinforced concrete piers, which
were designed to withstand seismic action. Adequate spacing was ensured between the piers
along the reaction frame so as to avoid blast wave interference. No instrumentation was
installed, however, because of the difficulty with data measurement in close-in detonations.
30
The blast trials reported by Holland (2008) as well as Williamson et al. (2011a) involved
highway bridge piers. Using ½-scale specimens, 10 small standoff tests and 6 local damage
tests were commissioned to investigate the blast consequences on the piers designed to
withstand gravity, seismic or blast action, by relying on the following variables: cross
sectional shape, length-to-depth ratio, type of transverse reinforcement, volumetric
reinforcement ratio, splice location on longitudinal reinforcement, standoff distance and
charge weight. The reaction frame was intended to provide a pinned condition at the top,
whereas the fixed connection at the bottom was achieved by grouting the large footing to the
reaction slab, as depicted in Figure 2.11.
Figure 2.11: Blast test setup for the highway bridge piers (Williamson et al., 2011a)
2.3.2 Uncoupled Methods
Blast load estimation is usually performed with the aid of empirical equations as well as
theoretical approaches. Alternatively, the first principles incorporated within hydrocodes, e.g.
31
Chapter 2
Air3d7 (Rose, 2001), can also be employed. The blast action is imposed either uniformly or in
a non-uniform manner on the targets.
In an equivalent static analysis where the inertial effect and dynamic material properties are
ignored, dynamic blast loads are converted into equivalent static loads; in a dynamic analysis,
an actual structural configuration can either be idealised as an equivalent single-degree-of-
freedom (SDOF) system where a lumped mass supported by a weightless spring is affected
by an axial force that brings about a displacement that corresponds to the deflection at the
critical location of the actual arrangement under the original loading (Biggs, 1964), or
represented as a MDOF system based on FEA or frame analysis.
2.3.2.1 Static Analysis
Mahoney (2007) applied the blast loads obtained from A.T.-BLAST8 (ARA, 2004) on a long-
span simply-supported prestressed concrete girder bridge, a three-span simply-supported
composite girder bridge and a three-span cantilever truss bridge, all simulated in SAP20009
(CSI, 2005). The arbitrary blast setups were decided with the aid of Monte Carlo simulation.
The consequence assessments were conducted by referring to the structural damage indicated
by the performance of the plastic hinges, along with the possible amount of downtime and
casualties.
A suspension bridge model was created by Suthar (2007) in SAP2000. The blast loads taken
from A.T.-BLAST were treated as equivalent static loads. The progressive collapse analysis
was conducted by relying on the formation of plastic hinges at the top and bottom chords of
the trusses.
7
Air3d, an Eulerian hydrocode which incorporates the advection upstream splitting method (AUSMDV)
scheme (Wada and Liou, 1997) and the monotone upstream-centred schemes for conservation laws (MUSCL)-
Hancock scheme (Toro, 1997)
8
A.T.-BLAST, a program which calculates the blast wave parameters of a hemispherical air-blast
9
SAP2000, a general-purpose structural analysis and design program
32
Islam and Yazdani (2008) modelled a two-span simply-supported prestressed concrete girder
bridge with STAAD.Pro10 (REI, 2004). 50% and 30% of the largest peak pressures given by
A.T.-BLAST were exerted as uniformly distributed static loads on the units closest to the
explosions and along the adjacent members respectively.
Zhou (2009) produced a two-span simply-supported composite girder bridge model, using
ANSYS11 (ANSYS, 2009). After the blast impact on a selected point was determined from
A.T.-BLAST, the pressure drop at another location at the instant that corresponds to the
arrival time recorded was computed, assuming a bilinear pressure profile, as consistent with
the adjustment proposed by McClendon (2007), to account for the actual quasi-exponential
decay. Then, the average pressure was uniformly distributed across the affected area.
2.3.2.2 Dynamic Analysis
Winget et al. (2005) investigated the blast response of a four-span simply-supported

prestressed concrete girder bridge, by relying on the SDOF analyses conducted using
SPAn3212 (USACE, 2002). The reduction in cross sectional area due to local damage was
calculated from the empirical equations developed by Marchand and Plenge (1998). The blast
action on a circular pier was derived from the procedure illustrated in Figure 2.12. The linear
wave profile, which preserved merely 80% of the original impulse on a square pier, was
adjusted to account for the clearing effect so as to obtain the impulse on the corresponding
circular pier, which was then used to scale the pressure history for the square pier, as
generated from BlastX13 (SAIC, 2001).
10
STAAD.Pro, a multi-purpose structural engineering software
11
ANSYS, structural mechanics and multiphysics solutions
12
Single-Degree-of-Freedom Plastic Analysis, a program intended for SDOF analysis
13
BlastX, a program written for blast load prediction
33
Chapter 2
Figure 2.12: Derivation of the blast load on a circular pier (Winget et al., 2005)
In the study conducted by Zheng (2007), the blast action determined from A.T.-BLAST was
imposed on the reinforced concrete piers simulated with ANSYS. For simplicity, the cross
sections of the piers were discretised and transformed into equivalent I-sections with
negligible web thicknesses.
Fujikura et al. (2008) devised a simplistic approach to predict the blast reaction of the
concrete-filled steel tube columns. The kinetic energy delivered by an impulse was assumed
to be fully converted into internal strain energy, ignoring spalling and breaching. A reduction
factor was introduced to incorporate close-in blast impact, clearing effect and the influence of
strain rate. The blast pressures were calculated from A.T.-BLAST, while the impulse
variations along the columns were derived from BEL14 (USACE, 2004).
14
Bridge Explosive Loading, a program devised for blast load generation
34
Matthews (2008) built a model of a simple-span simply-supported prestressed concrete girder

bridge with ABAQUS15 (ABAQUS, 2004). The segments near the supports were assumed to
behave in an elastic manner. The deck was also assigned a prestress action, owing to the
difficulty in integrating the slab after the introduction of the prestress forces in the girders.
The blast inputs were imported from BEL.
Tokal-Ahmed (2009) erected a two-span simply-supported prestressed concrete girder bridge

with ELS 16 (ASI, 2006), and claimed that, under extreme loading conditions, the applied
element method (AEM) is preferred over the discrete element method (DEM) as well as the
FEM. Nevertheless, the underestimation of the numerical results had been acknowledged,
since only incident pressures can be generated within ELS. A linear blast distribution (where
peak pressure was assumed to vary linearly with distance), together with a correction factor,
were proposed for the calculation of the uniformly distributed blast pressure in a SDOF
analysis.
Yan and Chang (2009) devised a probabilistic methodology, as presented in Figure 2.13, to
assess the blast vulnerability of cable-stayed bridges. Stage I predicts the reliability indices of
the individual bridge components, which are dependent on the given limit states, through
stochastic FEA and the first-order second-moment (FOSM) method; stage II estimates the
probability of progressive collapse under the damage inflicted at the component level, by
relying on the event tree approach. An imaginary cable-stayed bridge was created for
demonstration purpose.
15
ABAQUS, a complete solution for both routine and sophisticated engineering problems
16
Extreme Loading for Structures, an advanced non-linear structural analysis software based on the AEM
35
Chapter 2
Figure 2.13: Probabilistic assessment of blast vulnerability (Yan and Chang, 2009)
Tang and Hao (2010) constructed a cable-stayed bridge model with LS-DYNA. The main
span and the side spans were designed as composite hollow girders and prestressed concrete
box girders respectively. The blast pressures for scaled distances above and below 0.067
m/kg1/3 were generated using A.T.-BLAST and ANSYS AUTODYN respectively. Owing to
the massive scale of the bridge, smeared modelling of rebars, together with linear elastic
properties, were commissioned in the segments that were not anticipating blast destruction
and the mesh sizes were gradually increased from the blast vicinity onwards, as can be seen
in Figure 2.14.
36
Figure 2.14: Mesh distributions of the cable-stayed bridge model (Tang and Hao, 2010)
Fujikura and Bruneau (2011) described a moment-shear interaction model to estimate the
direct shear capacity which was expressed as the sum of the cohesion and frictional resistance
along the shear interface. Only the compressive region of the pier was of interest, while the
contribution from the reinforcing steel was ignored.
By relying on LS-DYNA, Williams and Williamson (2011) developed numerical models to

capture the spalling damage on the reinforced concrete piers tested by Williamson et al.
(2011a). The blast inputs were extracted from CFD simulation. The qualitative validation was
completed by matching the response shapes as well as the damage and crack patterns; the
quantitative verification was accomplished by comparing the peak residual displacements.
Interestingly, the strain rate behaviour of concrete was deliberately ignored, since strength
37
Chapter 2
enhancement was believed to be attributed to inertial restraint and non-homogeneous

deformation but not the intrinsic characteristic of the material itself, agreeing totally with
Schwer (2009).
2.4 BLAST RESPONSES
2.4.1 General Blast Effects
2.4.1.1 Local and Global Behaviours
Contact and close-in detonations produce non-uniform loading on the targets, giving rise to
mostly concentrated impact; far-field explosions impose essentially planar waves on the
objects, creating a more widespread effect, as illustrated in Figure 2.15.
(a) (b) (c)

Figure 2.15: Blast loading modes – (a) Contact burst; (b) Close-in detonation; (c) Far-
field explosion
At the component level, local damage arises because of material failure while global damage
is related to structural failure. However, if viewed at a broader perspective, local and global
effects may also refer to the general impacts generated on the individual members and the
overall systems respectively.
Global behaviours typically comprise flexural, membrane and shear reactions. The associated
P-Δ effect could be significant, since blast impact has the potential to provoke large
deflection. Axial crushing could happen on walls and slabs, but not on columns without
robust floor systems to effectively transmit the blast action (Crawford et al., 1999). During
38
the absence of alternative load paths, the demolition of critical units might trigger ultimate
collapse, as shown in Figure 2.16 for example, where a multi-storey building is experiencing
global progressive failure initiated by the removal of several of the exterior and interior
columns on the ground floor.
Figure 2.16: Global collapse of a multi-storey building due to removal of columns
Blast impact could induce compressive damage (cratering) on the immediate receiving
surfaces. Disengagement of matter due to excessive distortion is known as scabbing
(Smilowitz and Tennant, 2010), a term frequently used interchangeably with spalling which
is a tensile failure mode observed at the rear section where the stress wave is reflected off as a
tensile wave. Loss of matter throughout the depth of a cross section, known as breaching,
39
Chapter 2
could also occur in serious situations. Other forms of local behaviour include punching shear
response (controlled by diagonal shear resistance), direct shear reaction (due to both
geometric and load discontinuity), local hinge formation (when local moment capacity is
exceeded), as well as local buckling, yielding and fracture (for steel). However, local failures
are often not likely to pose serious threat to the ultimate survival of the structures (Crawford
et al., 1999).
2.4.1.2 Strain Rate Effects
Under high loading rates, materials will demonstrate dynamic behaviour which differs from
their static behaviour. The dynamic increase factor (DIF) denotes a ratio of the dynamic value
to the corresponding static value of a specific property.
The approximate strain rates produced from different loading regimes are explained in Figure
2.17. Blast-induced strain rates are significant and must not be neglected in analysis and
design, as overly conservative yet erroneous outcomes may ensue.
Figure 2.17: Strain rates due to static and dynamic loading
For concrete, greater compressive and tensile strength are detected with larger strain rates, as
noticeable in Figure 2.18 and Figure 2.19 respectively. The apparent scatter of the data is
attributable to the specimen sizes and shapes, the material properties, as well as the
experimental techniques adopted in different studies (Bischoff and Perry, 1991), but the
intriguing shift from a subtle to more drastic enhancement is evident nevertheless. The
formulations given in CEB-FIP (1993) suggested that the transition appears at approximately
30 s-1, but Malvar and Ross (1998) contended that a better agreement with the tensile test data
was reached with a threshold value of 1 s-1.
40
Before the transition, the moderate strength increase is attributable to the material or viscous
effects (Johansson, 2000). Cracks are forced to propagate through the stronger aggregates
instead of the weaker cement paste, since shorter paths are sought after, owing to insufficient
time. Fast compressive loading decreases the volume of the micropores, aside from causing
moisture movement which enhances the internal pressures; rapid tensile loading activates the
counteracting forces within the micropores wherein the free water fortifies the cement paste.
After the transition, the substantial strength improvement is accredited to the structural
effects (Johansson, 2000), i.e. the inertial factor, which prompts the stress distribution to
change, and the confinement factor, which invokes a plane strain state.
Figure 2.18: Strain rate effect on the compressive strength of concrete (Bischoff and
Perry, 1991)
41
Chapter 2
Figure 2.19: Strain rate effect on the tensile strength of concrete (Malvar and Ross,
1998)
For steel, the improvement in the yield strength and the ultimate strength during dynamic
loading is widely accepted as a consequence of insufficient time for deformation to occur
(Williamson et al., 2010). The influence of high strain rates, as observed in Figure 2.20, is as
follows:
 the elastic modulus, E, remains invariant

 the yield strength, fy, magnifies
 the ultimate strength increases, fu, but the percentage of increase is lower
 the ultimate elongation, εu, appears to be unaffected or slightly less, and therefore
the ductility reduces
42
Figure 2.20: Strain rate effects on steel
The strain rate behaviour of CFRP is strongly dependent on the constitutive fibres and
polymers. Carbon fibres alone are typically insensitive to strain rate, as proven by Zhou et al.
(2001) and Zhou et al. (2010), but the composite units are rate dependent. Kimura et al.
(2001) noticed that, at high strain rates, the carbon-fibre-reinforced epoxy resin exhibited a
non-linear stress-strain relationship where the elastic modulus and the tensile strength
improved but the failure strain remained unchanged, thereby agreeing with the outcome
reported by Welsh and Harding (1985).
2.4.2 Blast Effects on Bridges
2.4.2.1 Decks
Owing to the close proximity of the explosive, an above-deck explosion will normally inflict
local damage on the deck, as shown in Figure 2.21. A steel deck might fracture and suffer
plastic deformation (Son, 2008), while compressive crushing could happen on a concrete
deck (Tokal-Ahmed, 2009). During a below-deck blast event, the deck could separate from
the girders if the shear fasteners are inadequately designed, according to Winget et al. (2005)
and Yi (2009).
43
Chapter 2
Figure 2.21: Local failure of a deck (Deng and Jin, 2009)
Deck failure, albeit usually not devastating to the survival of the entire bridge (Matthews,
2008; Tokal-Ahmed, 2009), could be beneficial instead, since the dispersion of confined
waves will limit the harm imposed on the girders (Winget et al., 2005). Nevertheless,
compromising a deck system with a supportive role to the overall structural integrity (e.g. the
torsional strength of a curved trapezoidal steel bridge) is unacceptable (Roberts et al., 2003;
Williamson and Winget, 2005).
2.4.2.2 Girders
During a below-girder blast incident, the girders in a simply-supported bridge will mostly
experience flexural failure (Gannon, 2004; Tokal-Ahmed, 2009). Nevertheless, shear failure
is also possible (Islam and Yazdani, 2008). Upward loading is indeed harmful to the girders
which are normally intended to withstand predominantly downward loading, and the situation
could be worsened if prestressed members are involved (Williamson and Winget, 2005). The
girders might also endure transverse deflection (as suggested in Figure 2.22) aside from
vertical deformation, thereby suffering biaxial distortion (Tokal-Ahmed, 2009).
44
Figure 2.22: Transverse deformation of girders (Tokal-Ahmed, 2009)
The girders might undergo rigid body translation or uplift, as depicted in Figure 2.23, because
of the upward action. Apart from that, simultaneous movement of the abutments and the piers
due to the horizontal wave could also happen. Consequently, if inadequate seat width is
provided, the girders will collapse during the recovery phase after the upward pressure has
diminished (Winget et al., 2005; Tokal-Ahmed, 2009).
Figure 2.23: Separation of girders from the piers (Tokal-Ahmed, 2009)
A blast wave generated below the deck will travel towards the ground and compress the
surrounding air. The reflected wave assumes faster velocity in this heated medium and might
eventually merge with the incident wave approaching the girders, giving rise to a much
greater impact. However, if the explosive is situated higher above the ground, the girders
might encounter the incident wave before the arrival of the reflected wave, and therefore
sustain less damage (Winget et al., 2005). Wave reflection arising in the restricted spaces
45
Chapter 2
between the girders will prompt pressure amplification, as illustrated in Figure 2.24,
according to Winget et al. (2005) and Cimo (2007).
Figure 2.24: Confinement effect between girders (Winget et al., 2005)
The blast implications associated with a blast source placed on top of the deck are
comparatively less severe (Tokal-Ahmed, 2009). An above-girder explosion will most likely
impose merely local impact on the girders, in spite of the bending and shear failures reported
by Islam and Yazdani (2008) and Zhou (2009). The deck might also act as a defensive barrier
for the girders (Matthews, 2008).
It is unlikely for a truss girder with excessive redundancy to incur total collapse (Suthar,
2007), unless significant loss of units prevailed (Gannon, 2004). However, the functionality
of a truss element could be adversely affected by the lateral action as well as the reversal of
stress induced within (e.g. a tensile chord might buckle), as suggested by Williamson and
Winget (2005).
A blast wave will infiltrate into the interior of a hollow girder through a surface fracture (Son
and Lee, 2011). The stress waves transmitted from an affected area could also contribute to
the damage on the adjoining sections. Tang and Hao (2010) detected major destruction of the
base of the concrete box girder, but merely minor failure at the bottom of the hollow steel
girder. Box girders are also vulnerable to internal explosions (Williamson and Winget, 2005)
46
whereby confined pressures, which bring about higher impulses (Ray et al., 2003), are
generated.
The axial forces within a girder arrangement assume a significant role in governing the global
response of a cable-suspended bridge. Son (2008) claimed that a cable-supported bridge with
enormous axial forces along the girder system (e.g. a self-anchored suspension bridge) will
show an inferior blast performance. The blast deformation of a steel orthotropic deck could
be magnified by the destabilising P-Δ effect under the presence of the compressive action,
and this might lead to subsequent collapse of the entire span. With a composite girder, the
axial load initially received by the concrete slab will be transferred to the steel girders which
might buckle and give way eventually.
2.4.2.3 Towers and Piers
An explosive situated near the base of a column has a great potential to provoke shear failure
(as can be seen in Figure 2.25) which is the dominant failure mechanism for a wide range of
blast scenarios (Fujikura and Bruneau, 2011; Williamson et al., 2011a). However, flexural
failure (as shown in Figure 2.26) could also be a major concern (Winget et al., 2005; Yi,
2009).
Figure 2.25: Shear failure of a reinforced concrete column (Williamson et al., 2010)
47
Chapter 2
Figure 2.26: Flexural failure of a concrete-filled steel tube column (Fujikura et al., 2008)
The blast distribution on a column might vary with time because of the wave reflection
beneath the deck. Winget et al. (2005) noticed that the peak pressure was initially recorded at
the position corresponding to the charge height, but pressure build-up arose on top of the
column over time because of the restriction effect. Nonetheless, the first peak pressure, which
arrives earliest with the highest magnitude, will usually dictate the blast response
(Williamson et al., 2010).
Concrete spalling and breaching (as seen in Figure 2.27) due to a near-field explosion will
drastically reduce the capacity of a column (Williamson and Winget, 2005; Yi, 2009).
Interestingly, local damage could also be seen on the side covers of the column, in addition to
the front and rear faces, and this is noticeably distinct from the local effect normally seen on a
wall (Williamson et al., 2010).
48
(a) (b)
Figure 2.27: Local damage on reinforced concrete columns – (a) Breaching; (b) Spalling
(Williamson et al., 2010)
Detonations triggered simultaneously at the opposite sides of a column are expected to inflict
much greater harm in comparison with the equivalent explosion that only impacts on a single
face, and the relative effect could differ by “as much as a factor of three”, according to
Winget et al. (2005).
Gravitational load (up to the balance point on the axial load-moment interaction diagram)
will promote the flexural capacity of a column (Winget et al., 2005). A column is usually
competent to withstand the gravitational and blast loads transmitted via the spans
(Williamson and Winget, 2005; Winget et al., 2005), even though such extreme action may
prompt the column to buckle (Tokal-Ahmed, 2009), as suggested in Figure 2.28.
Upward blast loading will give rise to a net tensile action within a column that is rigidly
linked to the girders, and this may prompt the steel reinforcing bars to fracture, as observed
by Yi (2009).
49
Chapter 2
Figure 2.28: Lateral buckling of reinforced concrete columns (Tokal-Ahmed, 2009)
The blast outcome on a hollow structure strongly depends on the P-Δ effect, aside from the
presence of intermediate units. Son and Lee (2011) demonstrated the demolition of the steel
pylon under the influence of the P-Δ effect. Tang and Hao (2010) noticed that the protected
face of the concrete tower was not subjected to blast demolition (as can be seen in Figure
2.29), owing to the absence of interior units for stress wave transmission. However, the
destruction of the rear wall of the concrete pier (as obvious in Figure 2.30) was not only
evident, owing to the existing central wall, but also more violent, since concrete is weaker in
tension.
Figure 2.29: Blast damage on a hollow tower (Tang and Hao, 2010)
50
Figure 2.30: Blast impact on a hollow pier (Tang and Hao, 2010)
For simply-supported bridges, eliminating the piers will definitely bring the loaded spans to
fall, as observed by Tokal-Ahmed (2009) and Yi (2009). Bulson (1997) claimed that, in order
to induce collapse, it might be better to “shake down the piers” than to “shoot up the
superstructure”. Hao and Tang (2010) found that the cable-stayed bridge did not crumble
even though the pier and the tower suffered substantial blast damage, unless their cross
sections were completely destroyed.
2.4.2.4 Cables and Anchorage Regions
Williamson and Winget (2005) claimed that the cables in suspension bridges and cable-
stayed bridges are resilient to blast action. Prompt blast wave clearing around the flexible
cables with small exposed areas and rounded profiles enables equilibrium to be reached
faster. Moreover, multiple cable failures are usually required to collapse a span. The hangers
of suspension bridges might break under contact detonations, but not without the use of
unique charge shapes.
The anchorage precincts might be susceptible to blast impact (Williamson and Winget, 2005).
Tang and Hao (2010) noticed that the demolition of the concrete deck had resulted in the loss
of the anchorage points for several cables.
51
Chapter 2
2.5 BLAST COUNTERMEASURES
The two major objectives to be accomplished are threat and consequence mitigation. The
former aims to inhibit the occurrence and control the viciousness of blast instances while the
latter intends to minimise the detrimental implications on the structures.
2.5.1 Risk Assessment and Management
The risk assessment and management methodology, as presented in Figure 2.31, has been
devised by Williamson and Winget (2005) to address blast issues in a macroscopic
perspective.
Figure 2.31: Risk assessment and management (Williamson and Winget, 2005)
The risk assessment process consists of the following steps:
i. determination of importance level, depending on the inherent functions of the

bridge
52
ii. definition of threats, by taking into account internal problems (e.g. structural
deficiencies) as well as external factors which include human causes and natural
hazards
iii. evaluation of vulnerability, by allowing for both strategic and tactical issues in
the estimation of the probability of occurrence of the worst-case scenarios
iv. identification of consequences, in order to evaluate the severity of the detected
risks
The consequences that are deemed unacceptable must be mitigated in accordance with the
risk management procedure outlined below:
i. identification of potential countermeasures

ii. selection of feasible solutions, with the aid of a cost-benefit analysis, by factoring
in financial, environmental, social and technical constraints
iii. implementation of the preferred options, by utilising available resources
iv. evaluation of the effectiveness of the mitigation schemes
v. continuous monitoring of the performance (if the anticipated outcome is attained)
or reiteration of the entire procedure (otherwise)
2.5.2 Security Measures
The security measures that can be delivered to alleviate blast threat broadly fall into four
major categories (Roberts et al., 2003; Williamson and Winget, 2005):
a) strategic site layout, which involves practical landscaping (e.g. clearing of

overgrown vegetation provides better visual coverage to the surrounding
environment)
b) information and access deterrence, in order to frustrate attack plots (e.g.
prohibiting unauthorised admittance to the interior of a tower and removal of
sensitive data from accessible websites)
53
Chapter 2
c) regular monitoring and surveillance, to enable prompt discovery of suspicious

individuals and actions (e.g. installation of CCTV cameras at prominent spots and
good lighting setups)
d) efficient planning and coordination, to ensure prompt reactions during blast
incidents (e.g. emergency response plans)
Security plans may be customised flexibly to attain the desired security levels which are
dictated by the contemporary blast threat levels which, in turn, alter constantly with time
(Williamson and Winget, 2005). The security practices could be employed as short-term
relief while long-term solutions are under development (Roberts et al., 2003).
2.5.3 Standoff Distance
The best way of attaining the preferred level of blast protection is through the provision of
sufficient standoff distance (Tokal-Ahmed, 2009). Increasing the standoff distance even by
“as little as few inches” (Williamson et al., 2010) could immensely benefit the survival of a
structure. Interestingly however, Winget et al. (2005) realised that increasing the standoff
distance up to a particular range alleviated the blast impact to a considerable extent, but the
relative advantage derived beyond this point was not substantial.
Nevertheless, satisfactory standoff distances are often difficult to be attained (Winget et al.,
2005), since prohibiting full access to certain bridge components could be impractical or
impossible. Under permissible circumstances, vehicle barriers and security fences could be
erected (Williamson and Winget, 2005). Defensive armour could be placed at a slight
distance ahead of the target, and the intermediate space could be filled with energy-absorbing
substances (Tokal-Ahmed, 2009). Furthermore, offering sacrificial layers is also possible, but
the associated debris impact should not be neglected (Williamson et al., 2011b).
However, it must be emphasised that relocating a charge to a farther location might not
always end up with a positive outcome, as the loading mode involved might be altered. The
unification of incident and reflected waves could happen at a larger bridge clearance (Winget
54
et al., 2005), and greater harm could be resulted when local impacts are unfavourably
substituted with global effects (Hao and Tang, 2010).
2.5.4 Structural Geometry
A circular column with a curved surface will attract less reflected pressure and enable the
blast wave to clear more swiftly, in contrast to a rectangular column with a flat surface
(Williamson et al., 2011b). Under an identical explosion, the blast loads exerted on a round
and a square column could differ up to 34% according to Williamson et al. (2010), whereas
Fujikura et al. (2008) suggested a 55% decrease, as opposed to the 20% reduction adopted by
Winget et al. (2005).
Nonetheless, rectangular columns are not necessarily less favourable. A square column is
27% more massive per unit height in contrast to its counterpart of similar size (i.e. the width
equals to the diameter), and therefore incurs smaller shear demand under dynamic loading
(Williamson et al., 2011b). Besides, a round column of the same cross sectional area also
possesses lower blast endurance (Zheng, 2007). Further studies are needed, however, to draw
clearer conclusions regarding the relative merits of these two shapes (Williamson et al.,
2011b).
2.5.5 Structural Dimension
A structure with greater mass is likely to exhibit stronger blast resilience, since smaller
velocity will be resulted, thereby reducing the energy that needs to be dissipated through
strain (Dusenberry, 2010). Longer girders, which are usually bulkier and more flexible, could
afford greater deformation (Winget et al., 2005). Larger columns possess higher shear
capacities, and therefore able to sustain less damage albeit trialled under smaller scaled
distances (Yi, 2009; Williamson et al., 2010).
However, tall columns should be avoided if possible or braced at least (Zheng, 2007), as
stiffness and flexural resistance are inversely correlated to effective height (Gannon, 2004).
55
Chapter 2
Also, deck thickness should be kept minimal because blast transmission will cease following
deck failure, thereby limiting the harm confronting the girders which happen to be more
crucial (Gannon, 2004).
Williamson et al. (2010) recommended a minimum diameter of 30 in. (0.76 m) for circular
columns to counter close-in detonations. The current methods used to estimate the local
damage on walls are not suitable for columns (Davis et al., 2009). Mays and Feltham (2009)
proposed reasonable estimates of the minimum thickness, hmin, demanded by a reinforced
concrete wall or slab to effectively hinder spalling and breaching:
ℎ𝑚𝑖𝑛
1 > 0.07𝑍 −0.62 (no damage) (2.19)
𝑊3
ℎ𝑚𝑖𝑛
1 > 0.03𝑍 −0.62 (no breaching) (2.20)
𝑊3
2.5.6 Material Characteristics
The use of high strength steel will not bring any benefits, unless adequate energy-absorbing
capacity is assured (Williamson and Winget, 2005). Son (2008) observed that a thinner plate
fractured with larger displacement shortly after the explosion, although having greater
strength. This was accredited to the early yielding and the greater ductility of the lower grade
steel, aside from the lower flexibility of the thicker plate designed to have similar moment
capacity. However, if the member size is kept constant, higher strength steel is definitely
more advantageous (Gannon, 2004).
Judging from the failure modes, Yi (2009) noticed that the concrete piers with higher strength
were more superior, albeit observed to be more brittle because a constant reinforcement
quantity was adopted. Zheng (2007) also suggested that the blast endurance of a column is
proportional to concrete strength. Tokal-Ahmed (2009) recommended the use of slurry-
56
infiltrated fibre concrete (SIFCON) 17 , slurry-infiltrated mat concrete (SIMCON) 18 and

engineered cementitious composite (ECC)19 for enhanced ductility.
Concrete is more appropriate than steel when close-in detonations are concerned (Mays et al.,
2009). The rebound of a concrete member is less profound because cracking brings about
internal damping. Complex stress combinations could arise in steel members which are
comparatively more slender (hence more prone to local instability), and the ultimate
capacities are also difficult to be estimated.
2.5.7 Retrofitting Options
Malvar et al. (2007) as well as Buchan and Chen (2007) acknowledged the capability of FRP
in enhancing the blast resistance of timber, concrete, steel and masonry structures.
Nonetheless, the limitations of FRP have also been highlighted. FRP does not possess enough
strain capacity to meet the strain demand during a close-in detonation, and needs to be
shielded from flying fragments (Malvar et al., 2007). Silva and Lu (2007) and
Tanapornraweekit (2010) proved that strengthening solely the tensile side of the concrete
slabs was inadequate, but Wu et al. (2009) observed improvement even though only the
compressive face was protected. FRP promotes the shear capacity of a column, but might not
be credible in prohibiting concrete breaching (Winget et al., 2005). Furthermore, Buchan and
Chen (2007) also claimed that some fundamental aspects, such as the bonding characteristic
of FRP, the replacement criteria for FRP and the substrates, and the relative merits of
different composite types, are still not well understood.
A steel jacket shows better competence in preventing acute loss of cross section. It also
improves the flexural rigidity of a structural member. Although a minor gain in the shear
capacity is also achieved, it is often inadequate to satisfy the increasing shear demand, since
now less energy is consumed by means of flexural deformation (Winget et al., 2005). Such
17
Discrete steel fibres are concealed in slurry (Schneider and Mondragon, 1989)
18
Continuous steel fibre mats are covered with slurry (Malak, 1997)
19
Micromechanics is involved in the tailoring of the microstructures as well as the optimisation of the
substances (Li, 2003)
57
Chapter 2
claim is consistent with the test results reported by Fujikura and Bruneau (2011). Thus,
Winget et al. (2005) suggested that steel jackets could be used in conjunction with FRP wraps
in order to exploit the benefits of both materials.
Fujikura et al. (2008) found that the concrete-filled steel tube columns exhibited notable
ductility and minimal local damage. Son and Lee (2011) noticed that the concrete-filled steel
tube pylon suffered smaller lateral deformation and survived the blast incident, since the
associated P-Δ effect was less significant.
Son (2008) presented a strengthening concept known as the “Fuse System”, as depicted in
Figure 2.32, which aims to improve the blast performance of steel orthotropic decks. The
surrounding weak and strong elements are represented with steel plates thinner and thicker
than the deck respectively. The weak elements are expected to give way under blast action so
that the damaged section is isolated from the remainder of the deck, whereas the thick
elements contain the impact within the affected region.
(a) (b)
Figure 2.32: Fuse System – (a) Enhancement I; (b) Enhancement II (Son, 2008)
2.5.8 Detailing and Connections
In order to counter reverse loading, continuous top and bottom reinforcing bars should be
provided in the deck slabs and the girders, and undraped tendons are preferred over draped
versions in prestressed members (Williamson and Winget, 2005). Additionally, shear
58
fasteners should be carefully prescribed so that the required composite behaviour is attained
(Winget et al., 2005).
Williamson et al. (2010) recommended that the following critical variables ought to be dealt
with in such order of effectiveness: standoff distance, structural geometry, amount of
transverse reinforcement, type of ligature and anchorage, and splice location on longitudinal
reinforcement.
The ductility and the shear capacity of a column are governed by the transverse rebars
incorporated within. Williamson et al. (2011b) pointed out that the minimum amounts of
transverse reinforcement required along the total length of a circular and a rectangular
column, which are precisely 50% more than the seismic provisions stipulated in AASHTO
(2010), are given as:
𝑓𝑐
𝜌𝑠 ≥ 0.18 (for circular columns) (2.21)
𝑓𝑦
𝑓𝑐
𝐴𝑠ℎ ≥ 0.18ℎ𝑐 𝑠 (for rectangular columns) (2.22)
𝑓𝑦
where:
ρs = transverse reinforcement ratio
fc = compressive strength
Ash = total area of transverse reinforcing bars
s = spacing of discrete hoops
hc = core dimension in the direction of interest
The feasibility of other volumetric ratios apart from those considered above has not been
evaluated on account of the budgetary and schedule constraints of the experimental
programme (Williamson et al., 2011b).
59
Chapter 2
However, augmenting the transverse reinforcement ratio renders the placement of the
longitudinal rebars problematic (undesirable in terms of constructability) and invites the
formation of concrete voids. For this reason, mechanical couplers at splice locations, fibre-
reinforced concrete, concrete-filled steel tubes as well as retrofitting options may be
considered instead (Williamson et al., 2010).
Williamson et al. (2011a) showed that the blast resilience of the columns severely depreciated
following the loss of anchorage for the discrete hoops due to concrete spalling. Thus, the
spiral reinforcement, which ensures better concrete confinement, is more attractive.
Nonetheless, the blast performance could still be enhanced by augmenting the bend and the
length of the hooks to achieve sufficient anchorage. The detailing of discrete hoops, as
illustrated in Figure 2.33, are described in the blast-resistant design guidelines for highway
bridge piers published by Williamson et al. (2010), which covered three categories: Z > 1.2
m/kg1/3 (where no blast protection is required), 1.2 m/kg1/3 ≥ Z > 0.6 m/kg1/3 (where seismic
solutions are deemed adequate), and lastly Z ≤ 0.6 m/kg1/3 (which calls for stringent
fortification).
(a) (b) (c)

Figure 2.33: Detailing of discrete hoops according to scaled distance (m/kg1/3) – (a) Z >
1.2; (b) 1.2 ≥ Z > 0.6; (c) Z ≤ 0.6 (Williamson et al., 2010)
A 90° hook with an extension equals sixfold the nominal diameter of the rebar, db, is the
standard requirement in AASHTO (2010). The experimental work undertaken by Bae and
Bayrak (2008) demonstrated that an extension of 15db was more attractive than that of 8db for
a hook with a 135° bend. For more stringent protection against severe blast action, the length
60
of the 135° hook should not be less than the larger of 15db and 10 in. (≈ 0.25 m), as suggested
by Williamson et al. (2010).
Splices used to link longitudinal steel bars should be avoided whenever feasible, since
concrete breaching is possible if an explosion occurs near the splices. Williamson et al.
(2011b) advised that the lowest splice position should be at least 3.6 m above the ground or
the deck (by assuming the height of a typical truck bomb to be 1.8 m). Alternatively, splice
length could also be raised in order that the rebars remain anchored to the concrete. However,
no recommendations have been offered to define the acceptable splice lengths, owing to lack
of pertinent data.
Local buckling of steel members may be prevented with the aid of stiffeners (Roberts et al.,
2003; Williamson and Winget, 2005). Likewise, lateral bracings could also be attached to
girders and columns, as an attempt to improve their flexural capacities (Gannon, 2004;
Winget et al., 2005).
Enlarging the abutments and extending the expansion joints could avoid possible loss of
seating (Roberts et al., 2003). Hinge and cable restraints, which commonly serve in
earthquake-prone regions, may also be useful (Williamson and Winget, 2005). Tokal-Ahmed
(2009) described a fixed connection, as displayed in Figure 2.34, where the shear studs
embedded within the girder are welded to the plate which is held firm to the abutment or the
pier with the anchors.
61
Chapter 2
Figure 2.34: Fixed connection for simply-supported girders (Tokal-Ahmed, 2009)
2.5.9 System Redundancy
Both operational and structural redundancy ought to be taken into account (Winget et al.,
2005). The former (e.g. sound traffic management) ensures prompt reactions during the
aftermath of a blast event while the latter offers alternative load paths for the transfer of
action from the units that are dismissed.
Strong connections and continuous rebars are vital for the containment of blast effect
(Williamson and Winget, 2005). Additional girders and cables could also be delivered if
possible (Roberts et al., 2003). Aside from erecting extra columns, energy-absorbing walls
could also be installed either between the columns or as the primary support, as suggested in
Figure 2.35 (Tokal-Ahmed, 2009). However, the viability of the wall arrangement needs to
be affirmed, since larger reflected pressure is to be expected on a flat surface. Apart from
that, installing additional columns might not be beneficial, since less gravitational load will
be attracted to each individual column, and this will lead to a lower gain in the flexural
capacity.
62
(a) (b)
Figure 2.35: Wall configurations – (a) Intermediate; (b) Continuous (Tokal-Ahmed,
2009)
The auxiliary cable system developed by Tan and Astaneh-Asl (2003) to prohibit progressive
collapse of steel buildings could be adapted for use in bridges (refer to Figure 2.36), as
advised by Tokal-Ahmed (2009). The supplementary cables could be placed in the
intermediate diaphragms and the vertical stiffeners (along the longitudinal direction) of the
concrete and steel girders respectively. This secondary function is only activated if the
columns are removed, and the gravitational load will be transferred to the anchorage locations
behind the abutments or beneath the spans.
Figure 2.36: Ancillary cable system (Tokal-Ahmed, 2009)
2.5.10 Performance-Based Design Methodologies
Performance-based blast designs should be implemented in a multi-hazard perspective, taking

into account a diverse range of potential risks and consequences whilst factoring in available
63
Chapter 2
resources together with their possible allocation, when defining the performance objectives
which consist of design events and performance levels (Whittaker et al., 2003), making full
use of the prescriptive standards as a baseline (Thompson and Bank, 2007). The design
schemes must be capable of accommodating complicated modifications in a highly flexible
manner.
The design criteria corresponding to various protection levels, usually commence with
“superficial damage” (or equivalent) and terminate with “collapse prevention”, vary from
different sources depending on the focus (e.g. Winget et al., 2005; Mahoney, 2007; Son,
2008; Yi, 2009), and are normally qualitative for global integrity but quantitative for
individual components.
2.6 SUMMARY
Unconfined detonations derived from high explosives produce blast waves which result in
dynamic or impulsive loading (with high loading rates) which provokes local and global
reactions. Charge weights are normally converted into the equivalent sizes of TNT, and blast
intensity is indicated based on the concept of scaled distance. When an obstruction is
encountered, a blast wave will not only experience reflection whereby the incident pressure is
amplified, but also undergo diffraction whereby clearing and drag effect are observed.
A brief summary of the blast studies concerning bridges is given in Table 2.2.
Extensive blast investigations have been conducted on towers, piers and girders, whereas
fewer efforts have been dedicated to cables, anchorage points, decks and connections. The
simply-supported bridge, which is known unanimously to be the most common bridge type,
has received the most intensive attention, while restricted number of works related to cable-
suspended bridges have also been published.
Both coupled as well as uncoupled methods have been applied, not only in a deterministic
style, but also in a probabilistic manner. Although dynamic analysis (both SDOF and MDOF)
64
has expectedly exhibited strong popularity, the fact that static analysis has also been
attempted in numerous works in spite of its potential drawbacks is indeed surprising. FSI
simulation was preferred over the experimental approach.
Blast incidents on bridges may be broadly categorised as above-deck and below-deck

explosions. Deck failure usually does not pose devastating threat to the overall integrity of a
bridge, unless its strength and stability are compromised. A below-girder explosion is more
destructive, owing to its more widespread impact, its capability to provoke uplift and the
pressure amplification observed in the restricted spaces. A truss girder often enjoys excessive
redundancy, but could suffer from lateral action and stress reversal. For columns, shear
failure is more common than flexural failure; local damage could also be detected on the side
covers; and axial buckling and tensile pulling are possible as well. Blast wave infiltration,
stress wave transmission and significant P-Δ effect are also responsible for the demolition of
hollow structures. Cables are believed to be resilient to blast impact, as opposed to the
anchorage regions.
Threat and consequence mitigation can be achieved by means of: flexible security practices,
sufficient standoff distance, practical structural geometries and dimensions, desirable material
properties, effective retrofitting tactics, proper detailing and connections, as well as adequate
system redundancy. A risk assessment and management framework can be devised to deal
with blast issues in a systematic manner. Performance-based design principles are also
applicable to blast engineering.
Further enquiries should emphasise on innovative development of new techniques, along with
critical review of the current methods, for blast investigation and protection. Additionally,
robust expansion of the current state of knowledge regarding the blast response of bridges is
also vital. However, the focus of this study will be limited to the anchorage regions in cable-
stayed bridges.
65
Chapter 2
Table 2.2: Blast studies involving bridges

Investigation technique
Literature Coupled method Uncoupled method Bridge system Bridge component
Test FSI Static Dynamic
Winget et al. (2005) ● Simply-supported Concrete pier, prestressed girder
Ray (2006) ● Suspension Steel tower
Cimo (2007) ● Simply-supported Composite girder
Mahoney (2007) ● Simply-supported Prestressed & composite & truss girders
Suthar (2007) ● Suspension Steel truss girder
Zheng (2007) ● Simply-supported Concrete pier
Fujikura et al. (2008) ● ● Simply-supported Composite pier
Islam & Yazdani (2008) ● Simply-supported Concrete pier, prestressed girder
Matthews (2008) ● Simply-supported Prestressed girder
Son (2008) ● ● Cable-supported Steel orthotropic deck, composite girder
Deng & Jin (2009) ● Cable-stayed Steel truss girder
Tokal-Ahmed (2009) ● Simply-supported Prestressed girder, concrete pier
Yan & Chang (2009) ● Cable-stayed Steel girder & cable, composite tower
(Continued)
66
Table 2.2 (Continued)

Investigation technique
Literature Coupled method Uncoupled method Bridge system Bridge component
Test FSI Static Dynamic
Yi (2009) ● Simply-supported Composite girder, concrete pier
Zhou (2009) ● ● Simply-supported Concrete pier, composite girder
Tang & Hao (2010) ● Cable-stayed Hollow tower & pier, box girder
Fujikura & Bruneau (2011) ● ● Simply-supported Composite pier
Son & Lee (2011) ● Cable-stayed Hollow steel pylon, composite pylon
Williams & Williamson (2011) ● Simply-supported Concrete pier
Williamson et al. (2011a) ● Simply-supported Concrete pier
67
Chapter 2
68
Chapter 3 DISPROPORTIONATE AND PROGRESSIVE
COLLAPSE ANALYSIS
3.1 OVERVIEW
In this chapter, a complete description of the cable-stayed bridge employed as well as the
conventional loading details considered in the disproportionate and progressive collapse
analysis is included together with a clear explanation of the finite element modelling features
incorporated within the global model. The global performance of cable-stayed bridges, after
the destruction of the local anchorage regions, is determined by relying on the investigation
approach devised exclusively for the collapse analysis, and the critical blast locations are
subsequently highlighted.
Disproportionate collapse is a term used when the local failure of certain elements leads to a
substantial overall destruction (i.e. disproportionate in terms of cause and effect). On the
other hand, progressive collapse refers to the consecutive and uninterrupted demolition of the
adjoining elements (i.e. a domino effect or a chain reaction) precipitated due to incremental
load transfer after the initial damage. The latter is almost always the case of the former, but
not necessarily vice versa.
Cable-suspended bridges usually enjoy a certain degree of indeterminacy afforded by the stay
systems. Nonetheless, multiple anchorage point failures, especially when the anchorage
locations are clustered or appear in close proximity, are not impossible. Thus, the risk of
catastrophic collapse remains indisputably relevant.
69
Chapter 3
3.2 INVESTIGATION METHODOLOGY
A partial or total collapse can be provoked by: cable and anchorage zone failures which are
capable of leading to a zipper-type collapse20 (Starossek, 2007), when the deck-and-girder
system fails and during the destruction of the towers. These distinct activities can take place
either separately or simultaneously.
The non-linear dynamic analysis21, which gives the most credible outcome albeit entailing a
rather time-consuming processing (Marjanishvili, 2004; Marjanishvili and Agnew, 2006), is
implemented here, incorporating both geometric and material non-linearity, and without
being manipulated by artificial damping.
The simulation process of the collapse analysis is explained in Figure 3.1:
i. realisation of the initial state of equilibrium of the cable-stayed bridge

ii. definition of the assumed blast scenario
iii. identification of the anchorage points that are expected to become non-
operational as a consequence
iv. release of the link with the corresponding stays
v. if the bridge happens to either give way instantaneously or stabilise eventually,
the simulation is terminated at once; otherwise, the remaining stays that no longer
retain their load-carrying capability are dismissed, and this step is reiterated until
ultimate collapse is finally instigated
Being threat-dependent, since potential blast locations are factored in, the collapse analysis
assures a more convincing account of the actual sequel that might happen in reality, thereby
promising greater significance for blast mitigation.
20
Six categories of progressive collapse: pancake-type collapse and zipper-type collapse, domino-type collapse
and section-type collapse, instability-type collapse, as well as mixed-type collapse
21
Four classifications of collapse analysis: linear static analysis, non-linear static analysis, linear dynamic
analysis and non-linear dynamic analysis
70
Disproportionate and Progressive Collapse Analysis
Figure 3.1: Collapse simulation procedure
The gravitational load, F, is described as:
𝐹 = 𝐷𝐿 + 𝐿𝐿 (3.1)
where:
DL = dead load
LL = live load
Heavy traffic congestion during peak hour is taken into account, hence the factor of unity
assigned to the live load. The dynamic effect is naturally accounted for, and this excludes the
necessity of an overall amplification factor.
71
Chapter 3
The abrupt loss of supports is invoked rather conveniently with the aid of the built-in restart
function (LSTC, 2010), which allows the individual cables to be fully disabled at the correct
time during the simulation, instead of relying on equivalent forces (i.e. virtual step loads)
which are, nevertheless, commonly employed for this purpose (e.g. Feng and Chai, 2009;
Wolff and Starossek, 2009; Cai et al., 2012; Olmati and Giuliani, 2013).
The naming convention adopted, based on directional indication, is explained in Figure 3.2.
The x-axis is parallel to the north-south direction; the y-axis extends along the east-west (or
left-right) direction; the z-axis points towards the upward-downward direction.
Figure 3.2: Labelling convention for the cable-stayed bridge
72
3.3 FINITE ELEMENT MODELLING
3.3.1 Geometries and Dimensions
The cable-stayed bridge employed in this project is depicted in Figure 3.3. The bridge
represents a typical cable-stayed bridge in Australia. The 661-meter-long bridge comprised a
370 m concrete main span and two 145.5 m concrete back spans suspended by a total of 136
steel cables connected to two 124-meter-high concrete towers. The bridge is doubly
symmetrical along the longitudinal and lateral axes.
Figure 3.3: Hypothetical cable-stayed bridge
The centre span and the side spans have identical cross section, as shown in Figure 3.4. The
deck, which is 28.8 m wide, has a thickness of 0.25 m, with an additional 60 mm asphalt
layer on the surface. The depth and the width of the longitudinal girders are 1.9 m and 1.6 m
respectively. The lateral girders, with a cross sectional area of 1.5 m × 0.5 m, are installed
across the suspended spans at approximately 5 m intervals.
Figure 3.4: Cross section of the major and minor spans
73
Chapter 3
The A-shaped tower is shown in Figure 3.5. The tower legs, as constructed from the
ground/water level up to a distance of 89.7 m, are composed of solid blocks initially but
transform into hollow sections from a height of 5.65 m. For the cross section of the tower
stands (refer to Figure 3.6): the width maintains as 4.2 m throughout; the length varies from
7.15 m to 4.8 m; the wall thickness changes from 1.5 m (at RL 5.65) to 0.8 m (from RL 11 to
RL 30.96) and finally to 0.6 m (from RL 33.46 onwards). A 2.2 m cover is casted on top of
the two supports. The lateral bracing is designed as a box girder with a 4.3 m width, 3.3 m
depth and 0.6 m wall thickness. From 91.9 m to 123.6 m above the reference line, the tower
appears in a “H” shape, as seen from the plan layout (refer to Figure 3.6). The dimensions of
the side walls decrease from 4.8 m × 2.5 m to 4.8 m × 1.2 m, while the size of the lateral wall
reduces from 9.96 m × 2.05 m to 4.8 m × 2.05 m. A 0.4 m cover is placed on top of the H-
shaped section.
The cables are arranged in a fan configuration, with vertical inclinations ranging from 26.7°
to 79.7°, and individual lengths varying roughly between 62 m to 200 m. The live ends are
accommodated at the top section of the pylons, whereas the termination points are situated at
approximately 10.5 m spacing along the edge beams. The cables are sized accordingly, with
45% of the ultimate tensile strength as the upper limit of the allowable working load. The
largest stays contain 76 strands while the smallest stays possess 32 strands, where each strand
has a cross sectional area of 150 mm2 (prEN 10138-3).
74
(a) (b)
Figure 3.5: A-shaped tower – (a) Cross section view; (b) Elevation view
75
Chapter 3
(a)
(b)
Figure 3.6: Tower plan view – (a) Hollow section; (b) H-shaped section
3.3.2 Elements and Meshes
The finite element arrangement of the global model, as depicted in Figure 3.7, is constructed
with a sum of 21460 elements, and the corresponding breakdown is given in Table 3.1. For
the anticipated use of the global model, a modelling strategy with minimal discretisation,
which makes use of purely shell and beam elements, will suffice.
76
Figure 3.7: Finite element constitution of the global model
Table 3.1: Element breakdown for the global model

Element
Bridge component
Solid Beam Shell
Towers - - 7048
Suspended spans - 1952 9024
Stays - 3436 -
3.3.2.1 Beam Elements
The cables in the global model are produced as two-node discrete beam elements, as
displayed in Figure 3.8. This element type has six degrees of freedom, and the associated
mass is a product of material density and structural volume. The individual beams are
extended to a distance of approximately 5 m on account of the considerable length of the
stays.
77
Chapter 3
Figure 3.8: Beam elements in the global model
3.3.2.2 Shell Elements
Four-node shell elements are deployed for the towers and the spans, as displayed in Figure
3.9. The Belytschko-Tsay formulation, which is based on the co-rotational coordinate system
and the velocity-strain relationship, is selected for its high computational efficiency
(Hallquist, 2006). Owing to the enormous scale of the bridge, the length and the width of the
shell meshes are kept to be close to 100 cm whenever possible, but the upper limit is taken as
roughly 250 cm.
Figure 3.9: Shell elements in the global model
78
3.3.3 Constitutive Material Formulations
3.3.3.1 Concrete
Concrete in the global model is represented using Mat_Concrete_EC2 (Mat_172) which is

applicable to beam and shell elements only. The inherent capabilities of Mat_172, which
include concrete compressive crushing and tensile cracking as well as steel reinforcement
yielding, hardening and breaking, are governed by the formulations given in EN 1992-1-
2:2005.
The material properties of the concrete and the reinforcing steel in the global model are given
in Table 3.2.
Table 3.2: Mechanical properties of concrete and reinforcing steel in the global model
ρ ν E fc ft fy fu
Material
(kg/m3) (GPa) (MPa) (MPa) (MPa) (MPa)
Concrete 2400 0.2 33 50 4.1 – –
Steel 7850 0.3 200 – – 500 650
Smeared modelling of the steel reinforcement in the global model is performed through
Mat_172. By assuming uniform distribution of the rebars, the composite properties (e.g. bulk
modulus, shear modulus) can be determined through the following expression:
𝐾𝑚 = (1 − 𝑓𝑟 )𝐾𝑐 + 𝑓𝑟 𝐾𝑟 (3.2)
where:
Km = composite property
fr = reinforcement ratio
Kc = concrete property
Kr = reinforcement property
79
Chapter 3
The reinforcement ratios assigned to the bridge components, as listed in Table 3.3, are
believed to reflect the normal values adopted in a typical cable-stayed bridge.
Table 3.3: Reinforcement ratios in the global model

Bridge component Reinforcement percentage
Upper portion 4.0%
Towers Middle segment 2.5%
Lower section 2.1%
Deck 2.0%
Primary girders 2.0%
Secondary girders 3.6%
3.3.3.2 Steel
The steel cables in the global model are characterised using Mat_Cable_Discrete_Beam
(Mat_071) which is valid only for discrete beam elements. Since no compressive force will
develop within Mat_071, realistic modelling of the cables is possible through this material
model. The axial force, F, is non-zero only when the member is in pure tension (Hallquist,
2006).
𝐹 = max(𝐹0 + 𝐾∆𝐿, 0) (3.3)
where:
F0 = initial tensile force
K = stiffness
∆L = change in length
The prestress input serves as a convenient means to obtain the actual cable length intended
for the initial equilibrium.
80
The stress-strain curves for the cables are provided in terms of engineering stress vs.
engineering strain. The 0.1% proof force and the minimum breaking load for the strands are
240 kN and 279 kN respectively, as per prEN 10138-3.
3.3.4 Connections
In the global model, since the actual details of the anchorage regions are not afforded, the
cables are directly linked to the towers and the primary girders, as shown in Figure 3.10. This
approach has little impact on the anticipated axial forces in the cables, and is appropriate for
global modelling.
Figure 3.10: Node sharing in the global model
3.3.5 Internal and External Loading
3.3.5.1 Gravitational Load
The self-weight of the bridge is activated through the Load_Body function in which the
gravitational acceleration is specified as 9.81 m/s2.
The M1600 moving load (as per AS 5100.2-2004), which consists of a 720 kN truck load as
well as a 6 kN/m uniformly distributed component positioned along 15 m of a 3.2-metre-wide
design lane (refer to Figure 3.11), is considered here. A total of six standard traffic lanes are
81
Chapter 3
assumed to be operational on the bridge. By taking into account the accompanying lane
factors, a uniformly distributed load of 184 kN/m is applied as the traffic load in this study.
Figure 3.11: M1600 traffic load (AS 5100.2-2004)
Two traffic load scenarios, as shown in Figure 3.12, are taken into account in the collapse
analysis. The vehicular load is placed along the whole bridge in Case 1, but only across the
longer span (more critical than only on the shorter spans) in Case 2. Other load combinations
are merely the partial forms of these two, and the corresponding effects can be, thus, deduced
accordingly.
The gravitational load, which is sustained throughout the entire simulation, is initialised using
the dynamic relaxation and global damping features, whereby the system is critically damped
to attenuate the dynamic effect. Once the bridge structure is stabilised under the preloading,
the static (or quasi-static) process (Hallquist, 2006) is terminated, and the ordinary explicit
solver is subsequently invoked for the collapse analysis.
82
(a)
(b)
Figure 3.12: Traffic load distributions – (a) Case 1; (b) Case 2
3.3.5.2 Cable Forces
Cables are tensioned according to the desired forces during the construction stage in order to
achieve the intended structural equilibrium. In this study, the initial tensile forces required in
the individual cables, which are prescribed through Mat_071, fall within the range of 2600
kN and 6400 kN.
The initial tensile forces are calculated based on the iterative methodology developed by
Walther et al. (1999), as illustrated in Figure 3.13. During the absence of traffic load, the
spans deflect under the permanent loads and the deformation of the towers. Shortening the
main-span cables can balance the displacement of the major span due to the permanent loads
alone, while adjusting the back-span cables will counter both the deflection of the minor
spans as well as the movement of the towers.
83
Chapter 3
(i) Deformation of the spans and the towers without initial tensile forces
(ii) Fix the towers
(iii) Shorten the centre-span stays
(iv) Release the towers
(v) Adjust the side-span stays

Figure 3.13: Iterative procedure to determine the initial tensile forces for cables
(Walther et al., 1999)
3.3.6 Boundary Concepts
The foundations and the piers, which are believed to show a minimal role on the collapse
mechanisms of interest here, are not included in the global model. In order to simulate the
associated support conditions, a fixed boundary condition, which restrains movement in all
directions, is assigned to the nodes at the base of the towers, while the extreme ends of the
84
side spans, which correspond to the pier locations where the tie-down rods are found, are
given translational but not rotational constraints.
The boundary details imposed on the global model are shown in Figure 3.14.
Figure 3.14: Boundary conditions for the global model
3.4 VULNERABILITY ASSESSMENT
3.4.1 Elimination of Intermediate Supports
3.4.1.1 Removal of Single Mid-Span Support
In order to reveal the importance of the intermediate supports, a mid-span stay, which
appeared to attract the heaviest load, was deliberately removed (refer to Figure 3.15).
Figure 3.15: Targeted mid-span stay
85
Chapter 3
The resulted global effect in Case 1 and Case 2, as exhibited in Figure 3.16 and Figure 3.17
respectively, has been magnified considerably (i.e. by a factor of ten) for a better visual
effect.
Figure 3.16: Global response of the cable-stayed bridge in Case 1 (due to the elimination
of a mid-span support)
of a mid-span support)
The displacement histories of the primary and secondary spans (taken from the mid-points)
are displayed in Figure 3.18 and Figure 3.19 respectively, while the deflection of the towers
(measured from the top) is shown in Figure 3.20.
For Case 1, a large portion of the major span moved downwards, drawing the west tower
inwards, and this led to an uplift effect on the left span. Since the balancing action at the other
side had been weakened, the east tower leaned back slightly under the pulling action from the
tie-back stays, raising a small part of the longer span but lowering the right span.
As for Case 2, similar observations were made, except that the absence of traffic on both
sides (which serves as a counterbalance) allowed the pylons to bend at a much greater extent,
86
and as a consequence, the side spans experienced greater movement, whereas the middle span
fell farther downwards.
A partial collapse was unlikely, since a deflection of 1.09° and 1.15° was measured on the
affected segment of the longitudinal girder in Case 1 and Case 2 respectively, assuming 4° as
the maximum allowable rotation limit, by referring to UFC 4-023-03.
Fortunately, no cable failures were detected in both cases, as the final loads received by the
individual stays were well below the corresponding yield and ultimate strengths. The final
forces recorded in the critical cables are presented in Figure 3.21 and Figure 3.22.
The neighbouring stays at the south plane, together with those installed directly opposite (i.e.
at the north plane), felt a sharp rise in axial force, owing to the extra load released from the
targeted stay. As expected, a gain in tensile action was confirmed in the stays from the far
west, whereas the cables at the extreme east suffered a loss of internal load. Moving towards
the pylons, the load changes within the cables were less. It is also worth mentioning that the
tie-back stays were less strained in Case 1.
Figure 3.18: Main-span displacement (after the withdrawal of a mid-span stay)
87
Chapter 3
(a)
(b)
Figure 3.19: Back-span displacement (after the withdrawal of a mid-span stay) – (a)
Left span; (b) Right span
88
(a)
(b)
Figure 3.20: Tower deflection (following the withdrawal of a mid-span cable) – (a) West
tower; (b) East tower
89
Chapter 3
(a)
(b)
plane; (b) South plane
90
(a)
(b)
91
Chapter 3
The release of a mid-span stay did not give rise to a complete collapse. However, in Case 2,
the bridge ended up in a more critical state, owing to the extra deformation. A summary of
the outcome is provided in Table 3.4.
Table 3.4: Bridge response (due to the elimination of a main-span support)

Bridge element Response
West
Tower Inclined rightwards
East
Left
Major span Mostly displaced downwards
Right
Left Deflected upwards
Minor span
Right Deflected downwards
West
Centre stays Huge load increase near the damage location
East
West Axial force increase at the extreme end
Side stays
East Axial force decrease at the far end
3.4.1.2 Removal of Multiple Main-Span Supports
It was found that the cable-stayed bridge is indeed capable of withstanding the withdrawal of
no more than two sets of consecutive intermediate supports, without undergoing catastrophic
progressive collapse. For demonstration purpose, three separate couples of the strongest
anchorage points, as highlighted in Figure 3.23, were assumed to be fully demolished in a
blast incident.
Having multiple portable charges (which are relatively easier to be managed) targeting
different anchorage locations simultaneously is, in fact, not impossible, but a vehicle bomb is
also capable of severely damaging several anchorage points at once, depending on its size
and the spacing of the supports.
92
The collapse mechanisms in Case 1 and Case 2 are presented in Table 3.5, and displayed in
Figure 3.24 and Figure 3.25 respectively. Following the initial loss of supports, the remaining
stays were successively deactivated, and as a consequence, all the suspended spans, along
with the pylons, collapsed.
Figure 3.23: Targeted intermediate anchorage points
Table 3.5: Collapse mode (upon the deactivation of intermediate anchorage points)
Left Drawn inwards and then outwards
Pylon
Right Pulled towards the shorter span
West
Primary span Fell continuously
East
West Moved upwards and then downwards
Secondary span
East Lowered down
Left
Major stays Gradually failed due to cumulative loading
Right
Left Back to load rise after load drop
Minor stays
Right Load loss followed by load gain
93
Chapter 3
The displacement of the carried spans (at the middle) over time is traced in Figure 3.26 and
Figure 3.27, while the accompanying movement (at the top) of the towers is graphed in
Figure 3.28 and Figure 3.29.
The east-half of the centre span continued to sink; under the pulling effect from the side as
well as the ensuing P-Δ effect, the right tower was drawn towards the east span which
subsequently dropped.
The collapse of the west-half of the main span commenced later; it can be seen that the left
tower settled backwards after bending forward, and the west span experienced an uplift effect
before being lowered permanently.
Similar failure propagation was demonstrated in both cases: from the major span to the minor
spans; the right-half of the bridge crumbled earlier than the left-half. However, in Case 1, the
longer span showed a slightly smaller displacement, but a larger deformation was reached by
the pylons as well as the shorter spans within less duration, owing to the traffic load at both
ends.
The load histories of the selected cables were extracted to Figure 3.30 and Figure 3.31. The
continual outward bending of the pylons not only altered the inclination of the back stays but
also prompted them to gradually relax, since the side spans were falling at lesser rates. A total
loss of internal action is also clearly evident in the tie-back stays. As opposed to the
balancing stays from the east, a temporary huge load rise (which was greater in Case 2) was
also obvious at the west. A sharp rise in axial force, followed by abrupt failure, is indicative
of overloading.
Although both scenarios were confronted with total destruction, the aggressiveness of Case 1
slightly surpassed that of Case 2 because of the presence of traffic on the back spans.
94
At 0 s
At 1 s
At 2 s
At 3 s
At 4 s
deactivation of intermediate anchorage points)
95
Chapter 3
At 0 s
At 1 s
At 2 s
At 3 s
At 4 s
deactivation of intermediate anchorage points)
96
intermediate anchorage regions)
intermediate anchorage regions)
97
Chapter 3
points)
points)
98
supports)
supports)
99
Chapter 3
3.4.2 Elimination of Side Supports
3.4.2.1 Removal of Single Tie-Down Support
A strongest tie-back stay, which received the most profound tensile effect, was manually
excluded from the system (refer to Figure 3.32) so that the pivotal roles of the end supports
can be plainly highlighted.
Figure 3.32: Targeted tie-down stay
For the convenience of visual assessment, the global response observed in Case 1 as well as
Case 2, as demonstrated in Figure 3.33 and Figure 3.34 respectively, was exaggerated by
tenfold.
The deflection of the centre and side spans (measured from the mid-points) is plotted in
Figure 3.35 and Figure 3.36 respectively, while the displacement of the towers (as indicated
at the top) is recorded in Figure 3.37.
100
of a tie-down support)
of a tie-down support)
The right tower inclined towards the main span which was subsequently lowered, and as a
result, a pulling effect was induced on the left tower which, however, bent less in contrast to
its counterpart, and therefore the east span deflected more than the west span.
The overall response exhibited in Case 1 was much less dramatic, with smaller deformation
observed on the pylons and the deck-and-girder system. The initial equilibrium (i.e. before
cable loss was considered) attained in Case 2 was not as stable, since traffic was excluded on
the minor spans.
Furthermore, a segmental collapse of the primary girder was not to be expected, as a tie-back
stay assumes a more globalised role, instead of participating predominantly in local load-
carrying.
No cable failures were observed from either case. It is worth pointing out that, between the
two cases, wide discrepancies were found in the axial forces reported for the tie-back stays.
The final loads taken by the selected cables are given in Figure 3.38 and Figure 3.39.
101
Chapter 3
The additional loads assigned to the surrounding stays at the north and south planes,
following the removal of the targeted stay, were indeed obvious. As for the mid-span stays,
the internal action decreased at the east side, but increased at the west side (and so did the
tensile effect in the side stays). Approaching the pylons, the load transformation was noticed
to be not as profound.
The release of a tie-down stay was not enough to provoke total collapse. However, in Case 1,
the bridge was found to be in a less critical state, in comparison to Case 2, because the overall
deformation was less. The overall result obtained is summarised in Table 3.6.
Figure 3.35: Main-span displacement (after the withdrawal of a tie-down stay)
102
(a)
(b)
Figure 3.36: Back-span displacement (after the withdrawal of a tie-down stay) – (a) Left
span; (b) Right span
103
Chapter 3
(a)
(b)
Figure 3.37: Tower deflection (following the withdrawal of a tie-down cable) – (a) West
tower; (b) East tower
104
(a)
(b)
105
Chapter 3
(a)
(b)
106
Table 3.6: Bridge response (due to the elimination of a tie-down support)

West Inclined rightwards
Tower
East Bent leftwards
Left
Main span Moved downwards
Right
Left
Back span Raised upwards
Right
West Axial force increase
Centre stays
East Axial force decrease
West
Side stays Significant load rise at both ends
East
3.4.2.2 Removal of Multiple Back-Span Supports
For successive side supports, four pairs in Case 1, instead of only three pairs in Case 2, were
allowed to be taken away from the cable-stayed bridge, without instigating devastating
progressive collapse. However, for comparison purpose, the elimination of a total of four sets
of anchorage points located at the extreme end, as indicated in Figure 3.40, was considered.
Spaced within close distances, the anchorage points for the tie-back stays, which are
particularly susceptible to blast impact, are not only convenient targets for concurrent
detonations due to portable charges, but also within the coverage of the blast effect of a
typical vehicle bomb of a decent size.
The failure processes in Case 1 and Case 2 are reported in Table 3.7, and also presented in
Figure 3.41 and Figure 3.42 respectively. The whole deck-and-girder system totally gave
way, as a result of the consecutive dismissal of the stays after the initial damage. The pylons
were also expected to suffer complete failure.
107
Chapter 3
Figure 3.40: Targeted side anchorage points
Table 3.7: Collapse mode (upon the deactivation of side anchorage points)
Left Bent eastwards and then westwards
Pylon
Right Fell towards the longer span
West
Centre span Continuously sank
East
West
Side span Lifted initially, before being lowered
East
Left Sharp load rise succeeded by failure
Primary stays
Right Temporary load loss at the beginning
Left Back to load gain after load drop
Secondary stays
Right Incrementally loaded until elimination
The deflection histories of the carried spans (from the centre) are presented in Figure 3.43
and Figure 3.44, while the associated deformation of the pylons (top displacement) is shown
in Figure 3.45 and Figure 3.46.
As the initial restraint had been weakened, the east pylon inclined to the left, and a substantial
P-Δ effect was observed. The right span moved upwards before switching to a falling
108
position. The right-half of the middle span dropped accordingly, but at a smaller rate, and this
caused the tensile forces in the corresponding stays to decline.
The left-half of the longer span was pulled downwards at first. The west pylon was attracted
along, and this, in turn, brought the left span to a rising state, but the whole process was
quickly reversed following the uninterrupted loss of cables due to the additional load
transferred from the adjoining span which was confronted with a temporary loss of supports
during the relaxation stage.
The collapse sequence began from the east span and eventually encompassed the west span.
In Case 2, with no traffic load came into play at both ends, a larger displacement was
achieved by the primary span and the right tower within a shorter period, but the deformation
of the left tower and the secondary spans was smaller.
The load changes within the selected cables are presented in Figure 3.47 and Figure 3.48. It is
indeed interesting to notice that, owing to tensile relaxation and orientation modification, the
axial forces in the major stays on the east and the minor stays on the west actually dropped
continuously and could diminish eventually (slightly delayed in Case 1). However, the
internal forces could return to a rising state again until ultimate failure occurs, when
cumulative loading finally prevails. Also, the momentary load increase in the side stays on
the left due to the pulling action from the west pylon (larger in Case 2) was obvious.
The failure of Case 2 happened quicker than that of Case 1, which was also slightly more
difficult to be triggered.
109
Chapter 3
At 0 s
At 1 s
At 2 s
At 3 s
At 4 s
deactivation of side anchorage points)
110
At 0 s
At 1 s
At 2 s
At 3 s
At 4 s
deactivation of side anchorage points)
111
Chapter 3
side anchorage regions)
side anchorage regions)
112
points)
points)
113
Chapter 3
supports)
supports)
114
3.5 SUMMARY
The hypothetical cable-stayed bridge consisted of a medium-span concrete deck-and-girder

system carried by double-plane steel cables which were linked to two concrete towers of
moderate height.
The modelling effort undertaken for the global model, which was composed merely of beam
and shell elements, was less meticulous (i.e. with more implicit representations), as compared
to that for the local model (as described in the next chapter). The reinforcing steel bars were
smeared within the concrete, and no compressive action was allowed to develop in the cables.
Node merging was enforced for perfect and simplistic connections. Appropriate translational
and rotational constraints were prescribed at the reaction supports.
The load application process comprised two distinct stages: after the initial equilibrium was
introduced in a seemingly implicit manner, the collapse phase was implemented through the
normal explicit solver. The self-weight was determined internally, while the two critical
traffic load arrangements (i.e. on the entire crossing or only on the centre span) were
specified externally. Prestress actions were introduced to the individual cables, and the bridge
was initialised to its service state.
The iterative algorithm in the non-linear dynamic analysis was implemented with the aid of
the restart feature integrated within LS-DYNA; initial damage was introduced by disabling
the affected anchorage points (potential blast locations were taken into account), and the stays
that failed under cumulative burden were subsequently eliminated from the simulation.
The evaluation comprised two consecutive phases with growing complexity: the release of a
single mid-span or tie-down support as well as the elimination of multiple main-span or back-
span anchorage points.
115
Chapter 3
It was affirmed that the cable-stayed bridge was capable of withstanding the removal of any
one stay without undergoing partial failure, and the total destruction of multiple main-span or
back-span supports was required to give rise to a complete collapse.
Following the withdrawal of a single stay, the existing balancing system of the whole bridge
was disrupted, requiring the towers and the spans to participate in different extent to restore
equilibrium, apart from prompting varied degree of load changes within the stays. Owing to
direct load redistribution, the internal forces of the neighbouring stays around the immediate
locality of the affected area increased quickly, and this has the potential to trigger
catastrophic collapse.
The progressive collapse of a cable-stayed bridge is an extremely brittle failure style which is
significantly disproportionate to the triggering damage. The suspended spans totally gave
way under the uninterrupted loss of supports due to incremental load transfer and temporary
or permanent cable relaxation; the pylons toppled without surprise, owing to the destabilising
P-Δ effect. The collapse mechanisms were, typically, a further development and a substantial
magnification of the basic responses observed from the simplest failure settings. Technically,
the detrimental effect can either extend rapidly from the primary span or propagate violently
from one end to the other, depending on the initial damage.
Although the traffic load at both ends benefited the balancing arrangement when equilibrium
was reachable, it was, in fact, an utter disadvantage to the early survival of the secondary
spans during ultimate collapse.
Disproportionate and progressive collapse can be effectively prohibited by limiting the blast
vulnerability, enhancing the local structural robustness and improving the global collapse
resistance, by employing unique structural and non-structural tactics, through the direct and
indirect methods, as further discussed in Appendix A.
116
Chapter 4 BLAST PERFORMANCE OF CABLE ANCHORAGE
ZONES
4.1 OVERVIEW
Within this chapter, the finite element modelling techniques employed to present the local
model (which merely retains the pertinent features of the entire system that are deemed
essential for the blast analysis) in an elaborate manner, along with the principal assumptions
deployed to moderate the complexity of the cables and the anchorage components, are fully
covered.
The FEM is a numerical technique which spatially discretises conceivable subjects into
multiple divisions (commonly known as elements) defined by means of imaginary nodes with
distinct interrelationships collectively established by a stream of algebraic, integral as well as
differential equations, which are systematically solved through an implicit routine 22 or an
explicit procedure23 to reach approximate instead of exact solutions.
The local blast simulation in this research is conducted based on explicit time integration.
Here, the uncoupled method is employed: the blast impact is estimated solely by relying on
the empirical approach, while the dynamic structural response is predicted by characterising
the MDOF system purely under the Lagrangian domain.
22
The inverse of the stiffness matrix is applied to the force in order to yield the displacement. The process is
essentially static in nature. A relatively large memory is usually required, and the amount of iteration needed to
reach convergence is not predictable, and therefore end results are not always obtainable.
23
Dynamic equilibrium is satisfied by integrating the acceleration and the velocity in turn so as to attain the
displacement. Although the convergence issue is irrelevant here, the time step needed to ensure the stability of
the analysis is governed by the smallest element size, and this could lead to excessive computation time.
117
Chapter 4
An anchorage zone is usually subjected to a very large concentrated load, and the stress
distribution involved is truly complicated. In cable-stayed bridges, load transmission through
the anchorage regions can happen by means of saddles which are installed only on the towers,
transitional sections, or simply through direct bearing support which is the case in this
research.
Here, the common definition of a standard anchorage zone in the post-tensioned system is
introduced together with the associated failure mechanisms. Apart from that, the static stress
state and the force paths observed within a typical anchorage precinct in cable-stayed bridges
are also explored, and this will give a general idea of the critical areas.
The blast implications observed on the cable anchorage zone are presented in this chapter,
with the aid of a qualitative appraisal together with a quantitative assessment. A critical
comparison of the general blast consequences due to both PBIEDs and VBIEDs is also
included.
4.2 NUMERICAL SIMULATION
4.2.1 Main Tension System, Anchorage Assembly, Ancillary Equipment
4.2.1.1 Cables
The main tension system24 for the cable-stayed bridge considered in this project is in the form
of parallel strands, where each strand is composed of seven wires, i.e. six wires coiled around
a straight core wire in a helical manner at a constant pitch (refer to Figure 4.1), as consistent
with the typical configuration given in AS/NZS 4672.1:2007.
24
The four types of main tension elements for cable-suspended structures are: steel bars (spliced units are
connected using threaded couplers), locked-coil strands (round wires in the centre, followed by wedge-shaped
wires and finally Z-shaped wires) as well as parallel strands or wires (continuous strands or wires are arranged
parallel to each other)
118
Blast Performance of Cable Anchorage Zones
Figure 4.1: 7-wire strand (Gimsing, 1997)
Explicit modelling of the cables will inevitably lead to a prohibitively large amount of
elements. For instance, more than 10 million elements were created by Judge et al. (2011) for
merely a 250 mm segment of the 120 wires that were subjected to high velocity fragment
impact.
Thus, in this exercise, each individual cable is idealised as a circular bar with a specific
diameter defined by the total cross sectional area of the constitutive strands, as illustrated in
Figure 4.2, in order to control the total number of elements. The blast performance predicted
with this simplistic approach is believed to be acceptable for the purpose of this research.
Figure 4.2: Idealisation of the cable
119
Chapter 4
4.2.1.2 Anchorage Devices
A typical anchorage assembly 25 for parallel strands is exhibited in Figure 4.3. Individual
strands, which are normally equipped with high density polyethylene (HDPE) sheaths, are
firmly secured in an anchor block by means of conical wedges. Apart from frictional action,
radial forces also appear at the contact interfaces (Chacos, 1993). However, notches created
on the strands due to the sharp threads of the wedges are the chief concern for fatigue
strength (Tabatabai, 2005). The ring nut at the live end allows the initial tensile force in the
cable to be adjusted.
Figure 4.3: Conceptual depiction of the anchorage assembly
Filling substances are often injected for three reasons: as corrosion inhibitor, to supplement
load absorption, and to promote fatigue resistance by reducing the stress variation
experienced by the wedges. In recent time, metallic alloys 26 have been substituted by a cold-
casting compound27 (Podolny and Scalzi, 1986; Gimsing, 1997). Alternatively, the interior is
also allowed to remain empty for ease of cable inspection and replacement, and such
condition is considered in this exercise.
25
The three fundamental anchoring concepts are: individual anchoring by means of wedges or locking plates,
collective bonding in conical sockets, and the combination of both (Tabatabai, 2005)
26
Lead, copper and aluminium, for instance
27
Epoxy resin, zinc dust and steel balls
120
In the traditional FEA of the post-tensioned system, physical representation of the anchorage
components is usually avoided, and only a constant compressive action is exerted on the
loading surface instead. However, the presence of the anchorage units is indeed crucial in this
research, since the fluctuation of the anchorage forces due to blast loading as well as the
damage on the anchorage zones must be taken into account in order to generate realistic
outcomes.
Nevertheless, commercial anchorage devices are developed by the proprietary techniques

reserved by the manufacturers, and sensitive information, such as the interaction mechanisms
and the material properties, is usually confidential and concealed from the public. Moreover,
direct replication of such sophisticated features is beyond the scope here and should be
addressed in a separate study instead.
In this study, the anchorage system is simplified as a high strength steel cylinder with a
bearing plate, as illustrated in Figure 4.4. Actual modelling of the wedges is not attempted,
and the cable is connected directly to the anchor block. Load transmission is expected to
occur throughout the entire depth of the idealised unit with perfect gripping efficiency.
Figure 4.4: Idealised anchorage assembly
121
Chapter 4
The assumption of no slippage is, in fact, valid for both static and fatigue loading conditions,
where complete dislocation of any strands is absolutely forbidden, as firmly stated in PTI
(2007), while breakage of the cables, preferably far away from the anchorages, shall be the
sole failure mode observed, as clearly written in AS/NZS 1314:2003. However, no pertinent
blast specifications currently exist in this respect, and these failure criteria can be expected to
be applied for blast design.
At this stage, little is known about the blast implications on the anchorage fittings, as no
relevant studies and experimental data are available. Furthermore, genuine blast performance
of these complicated features is unlikely to be derived with the simplistic modelling approach
implemented here. In this project, the anchorage units are exempted from direct blast impact,
assuming excellent blast resilience, and the blast resistance of the individual anchorage
components is considered to be outside the scope of this research. This decision assures a
more conservative behaviour of the anchorage regions, since load transmission will not be
adversely affected owing to anchorage failure.
4.2.1.3 Auxiliary Components
In order to control the aerodynamic excitation induced by wind and rain, the HDPE tubes are
designed with low drag coefficients and furnished with helical ribs. Additional damping can
also be imposed with the installation of both internal and external dampers as well as cross
ties. Apart from that, better resilience to vibration is also assured with the presence of an
elastomeric bearing at the entrance of the delivery passage.
However, all these accessories are not considered in this study because their performance has
little impact on the blast response of the anchorage zones.
4.2.2 Geometries and Dimensions
The tributary zone for a cable extends from the median line of the deck towards the edge of
the supported primary girder and across the centres of the suspended spans (refer to Figure
122
4.5), covering a plan area of 16 m × 10.5 m. All the structural features encompassed within
this region, aside from the deck and the cross girders (which have minimal command on the
blast response of the anchorage region), are extracted to form the local model, as
demonstrated in Figure 4.5. This setup not only ensures that adequate matter is retained so
that genuine outcomes are attained, but also enables realistic boundary concepts to be
enforced.
(a)
(b)
Figure 4.5: Local model – (a) Cross section view; (b) Elevation view
The main girder is displayed in Figure 4.6. The longitudinal reinforcement comprised N36
bars (placed only at the top and bottom) and N20 bars (positioned at both sides) installed at a
spacing of roughly 160 mm, while the N20 transverse reinforcing bars are located 150 mm
apart throughout. The clear cover required is taken to be not less than 40 mm, as per AS
123
Chapter 4
5100.5-2004. The hollow space, which continues along the entire depth, serves as the
delivery passage.
(a)
(b)
Figure 4.6: Edge beam – (a) Cross section view; (b) Elevation view
A typical anchorage point, at the centre of the main span, is chosen for the blast investigation.
The stay, constituted with a total of 66 strands (hence a cross sectional area of 9900 mm2), is
installed at an inclination of 27.36°.
The corresponding anchor pod, which is essentially a protruding part casted along with the
edge beam, is elaborated in Figure 4.7. The square bearing surface is 0.9 m × 0.9 m. The local
zone reinforcement is fabricated as a continuous N16 spiral with a 0.65 m diameter and a
124
total of 9 turns at a constant pitch of 75 mm; the general zone reinforcement, which elongates
sufficiently well into the body of the major structure, consists of N20 open ties at varied
intervals. The central core is extruded to form a circular duct with a 0.2 m radius.
(a)
(b)
Figure 4.7: Anchor pod – (a) Cross section view; (b) Elevation view
125
Chapter 4
Owing to the high precision demanded in the blast simulation, solid meshes, aside from beam
and shell elements, are also created in the local model, as depicted in Figure 4.8. A total of
87173 elements are allocated accordingly, and the corresponding breakdown is disclosed in
Table 4.1.
Figure 4.8: Finite element constitution of the local model
Table 4.1: Element breakdown for the local model

Element
Bridge component
Solid Beam Shell
Edge beam and anchor pod 73052 - -
Steel reinforcement - 7326 -
Cable 5616 35 36
Bearing plate and anchorage unit 1108 - -
126
4.2.3.1 Beam Elements
Two-node discrete beam elements are assigned to the cable segment not anticipating direct
blast impact (i.e. from the upper end onwards, except the final 10 m stretch towards the lower
end).
All the steel reinforcing bars, as shown in Figure 4.9, are designated as Hughes-Liu beam
elements which are based on the degeneration of isoparametric solid elements (Hallquist,
2006). Each individual beam is defined using two end nodes and a unique reference node
indicating the orientation, each with six translational and rotational degrees of freedom. The
beam segments are mostly formed with a stretch of 15 cm.
Figure 4.9: Beam elements in the local model
4.2.3.2 Solid Elements
In the local model, the edge beam, the anchor pod, the bearing plate, the anchorage block as
well as the cable segment near the blast source (i.e. only the 10 m stretch from the bottom but
excluding the remainder towards the top) are discretised by means of constant stress solid
elements (refer to Figure 4.10).
127
Chapter 4
Figure 4.10: Solid elements in the local model
The eight-node hexahedral blocks are taken as the dominant units, while the six-node wedge
prisms are utilised as the transitional units, but the four-node tetrahedral mesh, which is
normally too rigid (and therefore reliable results are not always guaranteed) (Hallquist, 2006),
has been totally avoided here.
The one-point integration rule assigned to the hexahedrons ensures significant cost savings,
but zero energy modes, known as the hourglass modes, could arise. In this exercise, the
hourglass control is based on the Flanagan-Belytschko viscous form which is actually
recommended for problems where distortion develops at high velocities (LSTC, 2010). The
hourglass effect is irrelevant to pentahedrons where kinematic modes are non-existent
(Hallquist, 2006).
The edges of the solid blocks at close proximity to the explosion fall within the range of 5 cm
to 10 cm. It is difficult to maintain a regular mesh density, especially around the inclined
circular duct. Moving away from the anchor pod, the mesh for the edge beam is gradually
enlarged.
Mesh resolution is very important in this respect (Krauthammer and Otani, 1997). The mesh
sizes decided here are acceptable for the general comprehension of the blast response of the
128
anchorage zones and the roles of the various parameters, as well as the qualitative appraisal
of the effectiveness of the retrofitting techniques. Further refining the mesh is impractical
because the computation costs will become prohibitively large on account of the wide range
of scenarios considered.
4.2.4.1 Concrete
Concrete in the local model is replicated using Mat_Concrete_Damage_Rel3 (Mat_072R3),

an enhanced version of Mat_Pseudo_Tensor (Mat_016).
The competency of Mat_072R3 in blast simulation has been widely recognised. Magallanes
(2008) demonstrated that Mat_072R3 was more superior, as compared to
Mat_Brittle_Damage (Mat_096), Mat_Johnson_Holmquist_Concrete (Mat_111) and
Mat_Schwer_Murray_Cap_Model (Mat_145), which are all concrete models available in the
LS-DYNA library. Tanapornraweekit (2010) and Raman (2011) proved that Mat_072R3 was
more reliable, instead of Mat_Winfrith_Concrete (Mat_084) which has also been validated
against blast test data (Broadhouse, 1995).
Mat_072R3 is characterised by three independent surfaces (refer to Figure 4.11), i.e. the
initial yield surface, Δσy, the maximum failure surface, Δσm, and the residual strength surface,
Δσr, which are defined accordingly as:
𝑃
∆𝜎𝑦 = 𝑎0𝑦 + (4.1)
𝑎1𝑦 + 𝑎2𝑦 𝑃
𝑃
∆𝜎𝑚 = 𝑎0 + (4.2)
𝑎1 + 𝑎2 𝑃
𝑃
∆𝜎𝑟 = (4.3)
𝑎1𝑓 + 𝑎2𝑓 𝑃
129
Chapter 4
where:
P = pressure
a0y = initial yield surface cohesion
a1y, a2y = initial yield surface coefficients
a0, a1, a2 = maximum failure surface coefficients
a1f, a2f = residual strength surface coefficients
In Mat_016, the yield surface is conceived to be linearly related to the fixed maximum and
residual surfaces. However, this does not allow the brittle-ductile transition point to be
properly accounted for, and a disagreement with test data has been observed as a
consequence (Malvar et al., 1997).
The stress will remain purely elastic before the initial yield surface is exceeded. Beyond the
maximum failure surface, the response could either appear perfectly plastic or assumed to
soften towards the residual strength surface. The current surface, Δσ, is predicted by means of
linear interpolation between the corresponding surfaces, according to the following
equations:
∆𝜎 = 𝜂(∆𝜎𝑚 − ∆𝜎𝑦 ) + ∆𝜎𝑦 (4.4)
∆𝜎 = 𝜂(∆𝜎𝑚 − ∆𝜎𝑟 ) + ∆𝜎𝑟 (4.5)
where:
η = parameter which indicates the relative location of the current surface
130
Figure 4.11: Initial yield surface, maximum failure surface and residual strength
surface (Malvar et al., 1997)
131
Chapter 4
The parameter η, which indicates shear damage accumulation, is obtained as a function of the
modified effective plastic strain, λ. By incorporating the influence of strain rate and the
tensile damage scaling exponent into the original formulations in Mat_016, the modified
effective plastic strain is now written as:
̅̅
𝜀̅̅
𝑝
𝑑𝜀̅̅̅𝑝
∫ , 𝑃≥0
0 𝑃 𝑏1
𝑟𝑓 (1 + )
𝑟𝑓 𝑓𝑡
𝜆= ̅̅ (4.6)
𝜀̅̅
𝑝
𝑑𝜀̅̅̅𝑝
∫ , 𝑃<0
0 𝑃 𝑏2
𝑟𝑓 (1 + )
{ 𝑟𝑓 𝑓𝑡
where:
̅̅̅
𝜀𝑝 = effective plastic strain
𝑑𝜀̅̅̅𝑝 = effective plastic strain increment
rf = strain rate enhancement factor
b1 = compressive damage scaling exponent
b2 = tensile damage scaling exponent
Mat_072R3 is a three invariant model. The third invariant of the deviatoric stress tensor, J3, is
introduced so that the deviatoric section conforms to the shape displayed in Figure 4.12, as
adapted from Chen (1982), at low pressure and transforms into a circular shape at high
pressure. Thus, the tensile meridian is no longer obtained by rotating the compressive
meridian around the hydrostatic axis (whereby a circular deviatoric plane is resulted in
Mat_016).
Under triaxial tensile loading, the pressure reduction due to tensile failure is obtained by
incorporating the volumetric damage component into the deviatoric damage component only
when the stress path is near the negative hydrostatic axis based on the indication of the
|√3𝐽2 /𝑃| ratio. The effect of this ratio is implemented by introducing a multiplication factor,
132
fd, on the modified effective plastic strain increment, Δλ, which denotes the incremental
damage, as described below:
∆𝜆 = 𝑏3 𝑓𝑑 𝑘𝑑 (𝜀𝑣 − 𝜀𝑣,𝑦𝑖𝑒𝑙𝑑 ) (4.7)
|√3𝐽2 ⁄𝑃 |
1− , 0 ≤ |√3𝐽2 ⁄𝑃| < 0.1
𝑓𝑑 = { 0.1 (4.8)
0, |√3𝐽2 ⁄𝑃| ≥ 0.1
where:
b3 = input scalar multiplier
kd = internal scalar multiplier
εv = volumetric strain
εv,yield = volumetric strain at yielding
J2 = second invariant of the deviatoric stress tensor
Figure 4.12: Triangular shape with smooth corners (Malvar et al., 1997)
133
Chapter 4
The strength enhancement due to strain rate is assumed to occur along a radial stress path.
The enhanced value of the failure surface, ΔσE, is obtained by scaling the original strength
with the enhancement factor:
𝑃
∆𝜎𝐸 = 𝑟𝑓 ∆𝜎 ( ) (4.9)
𝑟𝑓
In this exercise, the influence of strain rate on the compressive strength is determined through
the formulations presented in CEB-FIP (1993):
𝜀̇ 1.026𝛼𝑠
( ) , 𝜀̇ ≤ 30 s−1
𝑓𝑐 𝜀𝑠̇
DIF = = 1 (4.10)
𝑓𝑐𝑠 𝜀̇ 3
𝛾𝑠 ( ) , 𝜀̇ > 30 s−1
{ 𝜀𝑠̇
where:
fcs = static compressive strength
𝜀̇ = strain rate (from 30 × 10-6 s-1 to 300 s-1)
𝜀̇𝑠 = static strain rate (30 × 10-6 s-1)
αs = 1 ⁄ (5 + 9𝑓𝑐𝑠 ⁄ 𝑓𝑐𝑜 )
log γs = 6.156𝛼𝑠 − 2
fco is a constant (i.e. 10 MPa)
As for the loading rate effect on the tensile strength, the modified version of the CEB-FIP
(1993) expressions, as proposed by Malvar and Ross (1998), is employed:
𝜀̇ 𝛿𝑠
( ) , 𝜀̇ ≤ 1 s−1
𝑓𝑡 𝜀𝑠̇
DIF = = 1 (4.11)
𝑓𝑡𝑠 𝜀̇ 3
𝛽𝑠 ( ) , 𝜀̇ > 1 s−1
{ 𝜀𝑠̇
134
where:
fts = static tensile strength
𝜀̇ = strain rate (from 10-6 s-1 to 160 s-1)
𝜀̇𝑠 = static strain rate (10-6 s-1)
δs = 1 ⁄ (1 + 8𝑓𝑐𝑠 ⁄ 𝑓𝑐𝑜 )
log βs = 6𝛿𝑠 − 2
fco is a constant (i.e. 10 MPa)
The Tabulated Compaction equation of state (EOS) defines the pressure in terms of
volumetric strain and internal energy. The pressure vs. volumetric strain relationship is shown
in Figure 4.13, where unloading happens along the bulk modulus from the current pressure to
the tensile cutoff, and reloading follows the identical path in the reverse direction. The
unloading bulk modulus in a compacted state is governed by the peak volumetric strain
(LSTC, 2010). In Mat_072R3, in order to avoid large negative pressure brought about by the
huge volumetric strain after cracking, the pressure cutoff, Pc, is assigned as the upper limit to
the negative pressure derived from this EOS:
𝑃𝑐 = −𝑓𝑡 (during hardening) (4.12)

𝑃𝑐 = −𝜂𝑓𝑡 (during softening) (4.13)
The damage distribution predicted by Mat_072R3 is displayed using the effective plastic
strain output with values between 0 and 2, as shown in the example given in Figure 4.14. A
value of 1 implies that the peak strength of the concrete has been reached. Softening of the
concrete beyond its ultimate strength is indicated with values between 1 and 2. At values
close to 2, cracking at tension or crushing at compression arises, and the concrete is
approaching failure.
135
Chapter 4
Figure 4.13: Pressure-volumetric strain relationship in Tabulated Compaction EOS
Figure 4.14: An example of effective plastic strain contours from Mat_072R3
The default automated parameter generation capability of Mat_072R3 is commissioned in

this study on account of the absence of essential information to comprehensively characterise
the concrete. The inputs disclosed in Table 4.2 are internally generated based solely on the
unconfined compressive strength which is 50 MPa in this case.
136
Table 4.2: Self-generated concrete parameters for the local model

a0y a1y a2y a0 a1 a2
1.116 × 107 6.25 × 10-1 5.15 × 10-9 1.478 × 107 4.463 × 10-1 1.616 × 10-9
a0f a1f a2f b1 b2 b3

0 4.417 × 10-1 2.366 × 10-9 1.6 1.35 1.15
The concrete elements that suffer excessive distortion can be eliminated from the analysis
with the aid of the erosion features incorporated within Mat_Add_Erosion. The failure
criteria decided for each study are usually selected either by trial-and-error or merely to
reproduce the experimental outcome. Both Tang and Hao (2010) and Wu et al. (2011)
assigned the maximum principal strain for concrete as 0.1, in contrast to the value of 0.03
adopted by Tanapornraweekit (2010) and Raman (2011), whereas two separate values were
suggested by Yi (2009), i.e. 0.002 and 0.005 for the concrete cover and core respectively.
These differing values do not appear to be mutually supportive by convincing justifications.
Thus, the implementation of the erosion algorithm is more commonly regarded as a
numerical manipulation.
Here, the concrete element is removed from the computation once the maximum principal
strain reaches 0.1. This failure strain value is, indeed, not only small enough to capture the
spalling damage and avoid computational instability, but at the same time also sufficiently
large to allow the elements to remain in the calculation for a longer period before being
eroded at a reasonable level of distortion, particularly when essential experimental data is
unavailable.
4.2.4.2 Steel
In the local model, Mat_Piecewise_Linear_Plasticity (Mat_024) is delegated to simulate the

elasto-plastic behaviour of the steel reinforcement, the bearing plate, the anchorage block and
the cable section that encounters blast pressures, while Mat_Cable_Discrete_Beam
(Mat_071) is selected for the cable segment far away from the blast source.
137
Chapter 4
The stress-strain relationship of steel is idealised as a bilinear function (refer to Figure 4.15),
with a definite yield point and linear hardening. When the fracture strain is reached, material
failure happens, and the element is subsequently removed from the calculation.
Figure 4.15: Bilinear stress-strain relationship of steel
Mat_024 allows arbitrary effective stress (i.e. true stress, σT) vs. effective plastic strain curves
to be specified for various strain rates. The true stress and the effective plastic strain can be
obtained from the following equations:
𝜎𝑇 = 𝜎(1 + 𝜀) (4.14)
𝜀𝑇 = ln(1 + 𝜀) (4.15)
𝜎𝑇
̅̅̅
𝜀 𝑝 = 𝜀𝑇 − (4.16)
𝐸
where:
σ = engineering stress
ε = engineering strain
εT = true strain
138
The strain rate behaviour of the steel reinforcement is described by the expressions
established by Malvar (1998), which are applicable only when the yield strength falls within
the range of 290 MPa to 710 MPa. As can be seen below, the static strain rate is taken as 10-4
s-1.
𝑓𝑑𝑦 𝜀̇ 𝛼𝑓𝑦
DIF = = ( −4 ) (for yield strength) (4.17)
𝑓𝑦 10
𝑓𝑑𝑢 𝜀̇ 𝛼𝑓𝑢
DIF = = ( −4 ) (for ultimate strength) (4.18)
𝑓𝑢 10
where:
αfy = 0.074 − 0.04(𝑓𝑦 ⁄414)
αfu = 0.019 − 0.009(𝑓𝑢 ⁄414)
As for the rest of the steel components, the strain rate effects are predicted through the
formulation developed by Cowper and Symonds (1957), which has been incorporated into
Mat_024 as an alternative means to account for loading rate dependency. The specific values
assigned to the coefficients in this equation are disclosed in Table 4.3.
1
𝜀̇ 𝑞 (4.19)
DIF = 1 + ( )
𝑄
where:
Q and q are coefficients
139
Chapter 4
Table 4.3: Cowper-Symonds coefficients for steel

Constant
Material
Q q
Normal strength steel* 40.4 5
High strength steel** 3200 5
* According to Cowper and Symonds (1957)
** From Paik and Thayamballi (2003)
The material properties of all the steel components are given in Table 4.4.
Table 4.4: Mechanical properties of the steel components in the local model
ρ ν E fy fu εf
Steel component
(kg/m3) (GPa) (MPa) (MPa)
Cable 195 1600 1860 0.05
Bearing plate 350 450 0.21
7850 0.3
Steel reinforcement 200 500 650 0.15
Anchorage unit 1000 1250 0.08
4.2.5 Connections
4.2.5.1 Constraint Algorithm
Physical modelling of steel reinforcement can be achieved through the discrete method or the
embedded method. Node sharing between the concrete and the reinforcing steel, as demanded
by the discrete approach, is rather impractical in the local model, especially within the
irregular elements around the circular duct and the anchorage zone with a spiral bar. Thus,
the embedded approach, which permits mesh independency, appears to be more appealing in
this case.
140
The Constrained_Lagrange_in_Solid command is invoked to constrain the velocity and the

acceleration of the steel reinforcement in the local model, while node coincidence with the
concrete is disregarded, as obvious in Figure 4.16. Perfect bond between the concrete and the
reinforcing steel is enforced, since bond slip is normally not critical under blast impact.
However, bond failure is assumed to have occurred when the coupling effect is deactivated
after the concrete elements (containing the rebars) are eroded on account of excessive
distortion.
Figure 4.16: Constraint mechanism in the local model
4.2.5.2 Node Merging
In the local model, the cable is secured to the anchorage block through node sharing (refer to
Figure 4.17), and this enables perfect gripping status to be attained.
Bearing plates are usually an integral part of the anchorage devices supplied by the
manufacturers. Thus, joining the coincident nodes between the bearing plate and the
anchorage unit, as seen in Figure 4.17, is acceptable.
141
Chapter 4
Figure 4.17: Node sharing in the local model
4.2.6 Conventional and Extreme Loading
4.2.6.1 Gravitational Load
The self-weight is initiated through the Load_Body command.
The traffic load considered in the global model is applied on the local model, and the
vehicular loading is assumed to act across the entire bridge.
Once the system is stabilised under the gravitational loading, which participates throughout
the whole simulation, the static (or quasi-static) routine is subsequently replaced by the
regular explicit procedure before the commencement of the blast phase.
4.2.6.2 Cable Force
The initial tensile force in the stay, as determined from the global model, is specified through
Mat_071.
142
4.2.6.3 Blast Action
The detonation points of a VBIED and a PBIED are shown in Figure 4.18. The VBIED is
situated at standard 1 m above the deck and minimum 2 m away from the cable; this
assumption practically defines the relative proximity of a vehicular explosion to the cable
anchorage zones. The PBIED, which is suspended from the deck, is stationed to coincide
with the centre of mass of the anchor pod at a standoff distance of 0.35 m; any attempt to
reduce such gap will demand positioning the charge farther beneath the edge beam, which is
technically difficult to be achieved in reality. The blast setups decided here intend to achieve
the worst case scenarios.
(a)
(b)
Figure 4.18: Blast locations of VBIEDs and PBIEDs – (a) Cross section view; (b)
Elevation view
143
Chapter 4
The size range for typical PBIEDs and VBIEDs is given in Figure 4.19. For PBIEDs, the
largest size available is given as 45 kg (≈ 100 lb) TNT, but the maximum explosive mass
considered here is restricted to 30 kg TNT, judging from the perspective of possibility based
on the nature of the current blast setting. For VBIEDs, the blast threat with the greatest
probability is linked to a charge weight of 227 kg (≈ 500 lb) TNT, according to Roberts et al.
(2003), but charge magnitudes up to 1000 kg TNT are also taken into account nevertheless,
owing to the enormous scale of the cable-stayed bridge in contrast to a typical highway
bridge.
Figure 4.19: Safe standoff distances for PBIEDs and VBIEDs (FEMA, 2003)
144
In this project, the blast implications of a vehicle bomb and a portable charge are dealt with
independently, since simultaneous existence of both types of explosives is not only rather
implausible during an unplanned occasion, but also commonly viewed as a redundant
decision in a deliberate offence. Nevertheless, while the concurrent presence of more than
one large-scale burst is indeed unlikely (except for extraordinary circumstances), the risk of
multiple occurrence of small-scale bursts is real and must not be taken lightly, as evident in
the past.
The blast scenarios considered in this research are detailed in Table 4.5. A total of three
distinct cases are established separately for the PBIED (which takes the form of a package,
with 10 kg rise in charge weight) and the VBIED (which appears as either a small or large
automobile or a truck, with the explosive size incrementally doubled from 250 kg onwards).
The blast magnitudes considered here are indeed realistic, as the values assigned are
generally consistent with the typical figures provided in actual cases. It is worth mentioning
that the scaled distances associated with both explosive types are not directly comparable
here, owing to the distinct loading nature; otherwise, a misleading perception of the blast
severity will prevail.
Table 4.5: Blast scenarios with PBIEDs and VBIEDs

Explosive Physical Charge weight Standoff distance Scaled distance
category appearance (kg) (m) (m/kg1/3)
10 0.162
Package or
PBIED 20 0.35* 0.129
equivalent
30 0.113
Compact sedan 250 0.159
VBIED Full-size sedan 500 1.00** 0.126
Van/Mini truck 1000 0.100
* From the anchor pod
** Above the deck
145
Chapter 4
In this research, the blast loads are generated internally within LS-DYNA, with the aid of the
Load_Blast_Enhanced command which is valid only for scaled distances ranged between
0.054 m/kg1/3 and 40 m/kg1/3 (LSTC, 2010). Aside from the enhancement in handling
reflected waves, moving warheads and multiple blast sources, this feature is similar to the
Load_Blast keyword where the governing functions originated from Kingery and Bulmash
(1984) are implemented based on the report published by Randers-Pehrson and Bannister
(1997).
Only the charge weight (in terms of equivalent mass of TNT) and the charge centre need to
be defined in order to derive the blast pressures, whereas physical modelling of the explosive
is not required. The impacted surfaces are identified using multiple segments, where each
segment will experience a unique blast pressure and impulse because of dissimilar arrival
times and angles of incidence. The negative blast phase is not disregarded so that genuine
outcomes are obtained.
4.2.7 Boundary Concepts
The axial loads induced by the neighbouring cables (due to the constitutive horizontal
components of the reaction forces) are applied externally along the longitudinal axis. The
interfaces that are in contact with the deck and the lateral girders in the transverse direction
are restrained against out-of-plane translations and rotations. The implementation of such
boundary solutions are capable of replicating the principal action exerted by the adjoining
members.
It is assumed that the anchorage points at the towers will not be adversely affected in this
study, and therefore fully restraining the upper end of the cable is acceptable.
The boundary conditions enforced in the local model are presented in Figure 4.20.
146
(a)
(b)
Figure 4.20: Boundary conditions for the local model – (a) Cross section view; (b)
Elevation view
4.3 CABLE ANCHORAGE ZONES
4.3.1 General Definition
An anchorage zone, as depicted in Figure 4.21, consists of a local zone (which is found
immediately ahead of the anchorage assembly) contained within or overlapping with a
general zone (which extends farther across the structural member), both defined based on the
inherent failure modes which are further explained in Appendix B.
147
Chapter 4
Figure 4.21: Anchorage region: general zone and local zone
4.3.1.1 Local Zone
Large compressive action with the potential to instigate crushing failure is the chief concern
in the local zone. The behaviour of the local zone is strongly dictated by the special
characteristics of the anchorage device (Burdet, 1990; Sanders, 1990).
4.3.1.2 General Zone
Within the general zone, the concentrated action will disperse laterally, and a more uniform
stress distribution is attained eventually. This leads to the development of lateral bursting
stresses which give rise to major cracking along the tendon path. Additionally, tensile stresses
will also arise near the loading and free edges, owing to the compatibility requirement for
differential deformation, especially when the anchorage units are eccentrically and widely
spaced.
148
4.3.2 Specific Definition
4.3.2.1 Stress Information
The principal stress vector plot (generated from a two-dimensional elastic FEA) produced in
Figure 4.22 shows the static stress flow induced within a typical anchorage precinct (using
the local model as an example) in cable-stayed bridges. The dispersal of the concentrated
action across the anchor pod and into the edge beam is obvious, and a uniform load
distribution is reached before the stress transmission is diverted horizontally. The unstressed
corner will incur tensile action which is capable of instigating undesirable cracking.
Figure 4.22: Stress vector arrangement within a typical anchorage region
The bursting stress distribution in this particular anchorage zone, which is subjected to a
constant anchorage force of 7300 kN, is shown in Figure 4.23. The bursting stress is
expressed as a percentage of the applied stress, while the distance from the loading edge is
given in terms of the anchor pod width. The tensile action extends from 0.23 m to 0.93 m
from the loading edge, with a maximum value reaching 0.87 MPa, and the resultant splitting
force is estimated to be approximately 333 kN.
149
Chapter 4
Figure 4.23: Bursting stress plot for the model anchorage region
4.3.2.2 Force Paths
A strut-and-tie model, as illustrated in Figure 4.24, is proposed based on the principal stress
vector plot. The flow of action is highlighted using compressive struts, C, while the resulted
tensile forces are indicated with intermediate ties, T.
The initial dispersal of the anchorage load is defined by C1 and C2; the ensuing stress
transmission is represented with C3 and C4; the eventual load transfer to the adjoining section
is outlined using C5 and C6. The lateral spread immediately ahead of the anchorage device is
shown using T1; the resultant bursting force due to the dispersal of the point load is identified
as T2; the tensile effect attributed to the horizontal diversion of the inclined action is
highlighted as T3.
150
Figure 4.24: Force paths in a common anchorage precinct
4.4 BLAST EFFECTS OF PERSON-BORNE-IMPROVISED-

EXPLOSIVE-DEVICES (PBIEDS)
The blast response of the cable anchorage zone, under the 30 kg TNT detonation, is
elaborated in Figure 4.25. Under the compressive blast action, the cratering effect on the
blast-receiving face of the anchor pod was indeed obvious. However, significant lateral
deformation of the anchor pod and severe distortion of the intermediate duct were not
observed, after all, as evident in Figure 4.26.
Matter disengagement was spotted not only at the external surfaces but also at the inner part
of the anchorage region (as shown in Figure 4.27), owing to stress wave transmission. The
intense spalling damage appeared at the bottom part of the anchor pod was attributable to the
transverse expansion instigated by both the anchorage force as well as the blast impact acting
on the other two local principal axes.
The damage distribution indicated through the effective plastic strain contours, as displayed
in Figure 4.28, revealed that the entire anchorage precinct was engulfed in blast effect.
However, the severity of the impact on the anchor pod (i.e. the blast recipient), which was
fully exposed to the surrounding environment, exceeded that on the volume contained within
the longitudinal girder.
151
Chapter 4
At 0 s At 0.001 s
At 0.010 s At 0.025 s
At 0.035 s At 0.045 s
Figure 4.25: Blast response of the cable anchorage zone under 30 kg TNT explosion
152
Figure 4.26: Anchor pod deformation and hollow core distortion (30 kg TNT explosion)
(Back view) (Front view)

Figure 4.27: Spalling damage on the anchorage region (30 kg TNT explosion)
The axial stress development in the general zone reinforcing steel, after the 30 kg TNT
detonation, is presented in Figure 4.29. The immediate sharp rise was due to the arrival of the
blast wave; it is interesting to note that, owing to strain rate enhancement, stress values
greater than the static ultimate strength had been achieved, without causing steel fracture. The
next build-up occurred during the crushing stage whereby reinforcement failures were
observed. In general, the front section was confronted with greater damage, in comparison to
the rear section which did not encounter the blast pressures.
153
Chapter 4
At 0 s At 0.01 s
At 0.02 s At 0.03 s
At 0.04 s At 0.05 s
Figure 4.28: Damage distribution on the anchorage precinct (30 kg TNT explosion)
154
(a)
(b)
section; (b) Rear section
155
Chapter 4
The loading capacity of the anchorage region was substantially undermined by the blast
damage, and crushing failure eventually occurred under the sustained loading (refer to Figure
4.25 and Figure 4.28). Concrete bulging ahead of the anchorage device, which was indicative
of compressive damage, was clearly visible. The crushing incident happened in a rather
brittle manner and represented the ultimate failure of the anchorage point.
When 10 kg TNT and 20 kg TNT were involved, the associated blast outcomes, as shown in
Figure 4.30 and Figure 4.32 respectively, were comparatively more moderate, as anticipated.
The cratering impact was much less noticeable, and the spalling damage alleviated
considerably. Nonetheless, compressive crushing had not been avoided regardless in both
instances. The damage contours generated from both cases, as displayed in Figure 4.31 and
Figure 4.33, revealed that the blast effect encompassed the whole anchorage zone.
At 0 s At 0.055 s
At 0.110 s At 0.130 s
156
At 0 s At 0.05 s
At 0.10 s At 0.15 s
It is apparent that the anchorage precinct was not large enough to avoid a “global blast
response” even under the close-in detonations due to the PBIEDs, although the blast impact
obtained here reflects predominantly the characteristics of the local effect understood in
normal sense. Alternatively, in order to avoid unnecessary confusion, it can also be claimed
that heavy “local blast damage” which extended across the entire anchorage region was
observed. The crushing failure of the anchorage zone should be known as the “secondary
global failure” instead in such context.
157
Chapter 4
At 0 s At 0.015 s
At 0.030 s At 0.056 s
158
At 0 s At 0.02 s
At 0.04 s At 0.06 s
On the other hand, the local effect produced on the longitudinal girder was indeed obvious.
Heavy damage concentrated around the middle part of the beam, owing to the unique blast
location selected; minimal damage was induced towards both sides of the beam, which
encountered lower blast pressures generated at larger angles of incidence. The blast impact on
the primary girder, which eased with a smaller charge weight, is highlighted accordingly in
Figure 4.34.
159
Chapter 4
(a)
(b)
(c)
Figure 4.34: Blast effect on the edge beam due to PBIED detonations – (a) 10 kg TNT;
(b) 20 kg TNT; (c) 30 kg TNT
160
The anchorage displacement histories are presented in Figure 4.35. The initial rising slope
was resulted from the vigorous thrust of the anchorage assembly as well as the beam
translation. The dramatic change of gradient indicates brittle crushing, and the brief
transitional period denotes the rapid initiation phase. With decreasing blast intensity, a
smaller axial stiffness reduction, together with a delay in the final destruction, suggested a
greater reserve of residual capacity.
Figure 4.35: Anchorage displacement records (after PBIED detonations)
Heavily non-uniform blast damage provoked a highly chaotic stress state in the concrete,
thereby giving rise to completely irregular stress dispersal across the anchorage region. The
transverse stresses at the mid-section of the anchorage zone are extracted for instance, as
shown accordingly in Figure 4.36, Figure 4.37 and Figure 4.38. The transverse stresses
became purely compressive, with an absolutely eccentric distribution, and converged swiftly
to zero on account of concrete failure.
161
Chapter 4
Figure 4.36: Transverse stress distribution (under 10 kg TNT explosion)
162
The mid-span displacement plotted in Figure 4.39 proved that the longitudinal girder
underwent a momentary translation in the opposite direction when the downward (negative)
gravitational load was overcome by the upward (positive) blast action, but began to fall soon
after the blast wave had diminished. With a larger bomb size, the uplift effect intensified, and
the edge beam sank lower within a shorter duration.
The axial force recorded in the cable is traced in Figure 4.40. The uplift effect experienced by
the primary girder, together with the sudden damage inflicted on the anchorage precinct,
contributed to the initial decline of the axial force below the static value. The reloading stage
(i.e. the ensuing climb) persisted until critical crushing prevailed; the function of the stay was
fully deactivated, and this led to a total loss of axial force. With increasing blast severity, a
greater initial drop in the cable force was noticed along with a shorter reloading process,
which again confirmed that less resistance was retained.
163
Chapter 4
Figure 4.39: Mid-span displacement plots (after PBIED detonations)
Figure 4.40: Cable force histories (after PBIED detonations)
In all three cases, the yield strength of the stay (i.e. 16 MN) was not exceeded. Apart from
that, no fractures had been spotted on the cable section at close proximity to the blast source,
since no direct blast transfer was possible across the central duct, aside from the existing
protection afforded by the surrounding concrete.
164
The corresponding anchorage point at the pylon was configured based on a design load of 11
MN, and therefore bursting cracking and bearing failure brought by the blast action
transmitted via the stay were not to be expected.
4.5 BLAST EFFECTS OF VEHICLE-BORNE-IMPROVISED-

EXPLOSIVE-DEVICES (VBIEDS)
The blast response of the cable anchorage zone, under the 1000 kg TNT detonation, is
exhibited in Figure 4.41. The initial resistance of the anchorage region was significantly
weakened by the blast damage, and compressive damage was evident when concrete bulging
arose ahead of the anchorage assembly. Crushing incident, an extremely brittle ultimate
failure mode, was inevitable, unfortunately, with the continual presence of the gravitational
loading.
The damage distribution indicated through the effective plastic strain contours, as displayed
in Figure 4.42, revealed that the whole anchorage precinct was completely engulfed in blast
effect, owing to the stress wave transmission occurred within as well as the anchorage force
amplified by the blast action.
Matter disengagement concentrated mostly at the internal part (around the intersection of the
longitudinal girder and the anchor pod, to be exact), but appeared to be very limited at the
outer faces of the anchorage zone, as can be seen in Figure 4.43. The anchor pod deformation
and the circular duct distortion due to the immediate blast impact were only superficial.
165
Chapter 4
At 0 s At 0.100 s
At 0.150 s At 0.200 s
At 0.205 s At 0.210 s
166
At 0 s At 0.10 s
At 0.15 s At 0.20 s
At 0.25 s At 0.30 s
167
Chapter 4
(Back view) (Front view)

Figure 4.43: Spalling damage on the anchorage region (1000 kg TNT explosion)
The axial stress recorded in the general zone reinforcing steel, following the 1000 kg TNT
detonation, is traced in Figure 4.44. The sudden stress change immediately after the blast
arrival, together with the major transformation (where a much greater value can be reached,
observed during the crushing process whereby reinforcement failures were resulted), are
indeed obvious. In general, the rear section, which was relatively closer to the blast source in
contrast to the front portion, incurred greater blast impact and much worse concrete damage,
which, in turn, weakened the stress transfer with the reinforcing steel.
168
(a)
(b)
section; (b) Rear section
169
Chapter 4
It is indeed interesting to notice that the anchorage region successfully survived the blast
incidents with 500 kg TNT and 250 kg TNT, as can be seen in Figure 4.45 and Figure 4.47
respectively, and continued to be functional albeit with relatively less resilience. The
associated blast consequences were relatively not as intense, as expected. The spalling
damage detected was very minor. The damage contours extracted from both cases, as
displayed in Figure 4.46 and Figure 4.48, suggested that the blast effect covered a large
portion of the anchor pod but merely a fairly small extent of the integrated region.
At 0 s At 0.2 s
At 0.4 s At 0.6 s
170
At 0 s At 0.2 s
At 0.4 s At 0.6 s
Here, the undesirable potential of heavy local effect in leading to a disastrous global response
was plainly demonstrated. However, it must be pointed out that, owing to the relative size of
the anchorage region, the local blast damage was inflicted on the global level, even though
the blast event associated with a VBIED can be arguably claimed as a close-in detonation in
this given context. Thus, careful use of the common terms is of paramount importance in
order to avoid false perception.
171
Chapter 4
At 0 s At 0.2 s
At 0.4 s At 0.6 s
172
At 0 s At 0.2 s
At 0.4 s At 0.6 s
Widespread damage was observed on the longitudinal girder, as evident in Figure 4.49, with
the small standoff distance, and the vehicle bombs were relatively large. The blast impact
extended throughout the depth of the beam, with 1000 kg TNT; but it focussed predominantly
at the top section which was closest to the blast source besides appearing in a tensile stress
state, when the blast magnitude was decidedly brought down to 500 kg TNT and 250 kg
TNT.
173
Chapter 4
(a)
(b)
(c)
Figure 4.49: Blast effect on the edge beam due to VBIED detonations – (a) 250 kg TNT;
(b) 500 kg TNT; (c) 1000 kg TNT
174
The anchorage displacement plots are given in Figure 4.50. With 1000 kg TNT, brittle
crushing was clearly indicated by the rapid movement of the anchorage device in the positive
direction (i.e. against and into the anchorage zone), while new equilibrium was achieved
when 250 kg TNT and 500 kg TNT were involved. The negative displacement (in the
direction away from the local structure) shown was due to the global translation of the overall
system. The diminishing rate of the resistance that remained was of course proportional to the
blast severity.
Figure 4.50: Anchorage displacement records (after VBIED detonations)
The non-homogeneous stress wave transmission instigated by the blast action gave rise to an
extremely disturbed stress condition in the concrete, which, in turn, led to totally eccentric
stress dispersal within the anchorage precinct. The transverse stresses extracted from the
centre of the anchorage region, as presented accordingly in Figure 4.51, Figure 4.52 and
Figure 4.53, are taken for example. Although the transverse stresses resolved into a steady
state (rather than converging to zero on account of concrete failure) when 250 kg TNT and
500 kg TNT came into play, the derivation of a general distribution rule seems to be
impractical.
175
Chapter 4
176
The mid-span displacement shown in Figure 4.54 suggested that the longitudinal girder
settled down steadily, except with 1000 kg TNT whereby it continued to fall. The blast-
induced translation, which reduced with a smaller bomb size, was controlled by not only the
gravitational and blast loading but also the blast performance of the edge beam and the
anchorage zone.
The axial force development in the cable is presented in Figure 4.55. The overall gain in axial
force, which increased with rising blast intensity, was contributed by both the immediate
blast action and the subsequent pulling of the primary girder. When 250 kg TNT and 500 kg
TNT were commissioned, the cable force gradually stabilised, indicating that the anchorage
region was still operational under the residual capacity retained. However, when 1000 kg
TNT was employed, the sustained loading persisted until compressive crushing arose, and the
presence of the stay became no longer meaningful.
177
Chapter 4
Figure 4.54: Mid-span displacement plots (after VBIED detonations)
Figure 4.55: Cable force histories (after VBIED detonations)
The yield strength of the stay was well beyond reach in all three cases, let alone the ultimate
strength (18 MN). No fractures were detected, even at the cable section extended beyond the
circular duct, which encountered direct blast impact, and this was attributable to the curved
profile and the small exposed surface of the flexible cable.
178
At the corresponding anchorage point at the tower, which was designed to withstand a
maximum force of 11 MN, splitting cracking and crushing failure under the blast action
transmitted via the stay were believed to be unlikely.
4.6 CRITICAL COMPARISON OF BLAST IMPLICATIONS
The blast action from a PBIED was directly applied to the anchorage zone, whereas indirect
transfer of the blast impact due to a VBIED occurred via the cable and the edge beam; this
contributed to completely unique loading conditions.
The criticality of the blast setting considered for PBIEDs was proven to surpass that for
VBIEDs. The anchorage region was fully demolished under all the blast magnitudes assigned
to the PBIED, but it gave way only when the most powerful VBIED came into play.
Compressive crushing was obtained as the common ultimate failure mode, after the
anchorage zone had been weakened considerably, but took longer to be initiated even with
the largest VBIED, and this, in turn, suggested that less resistance was preserved when a
PBIED was adopted.
In both situations, the blast effect (worsened under the direct blast impact) extended across
the entire anchorage precinct, but the exposed region, rather than the integrated content, was
usually confronted with more serious damage. The spalling damage focussed primarily at the
inner surrounding instead of the outer enclosure. The front section incurred heavier damage
with a detonation triggered from the right, and likewise for the rear section with an explosion
released from the left.
The blast-induced distortion of the anchorage zone, albeit not severe, was greater under the
lateral pressure produced by the small-scale bursts.
The anchorage region incurred significant non-uniform blast damage which brought about
totally random stress dispersal which failed to conform to a consistent distribution rule.
179
Chapter 4
However, under the indirect blast impact, the stress state induced was found to be much less
chaotic.
Even though both standoff distances of concern were sufficient to lead to close-in blast
situations, the vehicle bombs were large enough to inflict widespread effect on the
longitudinal girder, but the local impact associated with the portable charges, which were
relatively smaller in size, was generally more profound nevertheless. Overall, the top portion
of the primary girder, which resided in a tensile condition on account of the anchorage force,
suffered greater blast damage.
Without surprise, an explosive stationed above the deck instigated negative global translation,
but a prior temporary positive movement took place when the charge was positioned beneath
the girder.
The upward blast loading prompted a temporary axial load drop which was quickly
succeeded by the reloading stage; the downward blast loading caused the axial force to rise
steadily, and therefore a much higher value was resulted in comparison. However, the sharp
decline of cable load during compressive crushing was common; but when the anchorage
point happened to survive the catastrophe, the cable force gradually settled to a new
equilibrium.
In both circumstances, the blast action conveyed via the stay appeared to be incapable of
provoking bursting cracking and bearing crushing on the anchorage zone accommodated at
the pylon, and the original design load was still far beyond reach.
The adverse effect imposed on the stay was rather minimal, as no fractures were detected,
and the yield and ultimate strength were also not exceeded. However, it is worth noting that
only indirect blast application was involved in the blast events that occurred at the side of the
anchor pod, while direct blast contact with the cable also happened in the blast incidents that
took place on top of the edge beam.
180
The comparison of the blast influence of PBIEDs and VBIEDs in this specific context is also
presented in Table 4.6.
Table 4.6: Blast implications due to PBIEDs and VBIEDs

Bridge Blast implication
Main aspect
component PBIED VBIED
Loading manner Direct application Indirect transfer
Hostile Not as violent
Full survival except with
Complete destruction
the most powerful
Criticality under all circumstances
magnitude
Smaller reserve of Delayed failure
Anchorage residual capacity propagation
region (on Failure mode Brittle crushing as the common failure mechanism
the deck- Acute Relatively less profound
and-girder Spalling damage more dramatic at the interior
system) Blast damage Wide coverage of blast effect (much worse at the
anchor pod but not at the sheltered zone)
Graver at the front Heavier at the back
Lateral impact was
Blast distortion Predominantly axial
evident, but not drastic
Totally random Not as irregular
Stress dispersal
Non-conformity to a universal distribution rule
Loading Blast action transmitted via the cable could be treated
Anchorage
application essentially as an axial force
zone (at
Blast Negligible, compressive crushing and tensile cracking
the pylon)
consequence not to be anticipated
(Continued)
181
Chapter 4
Table 4.6 (Continued)

Bridge Blast implication
Main aspect
component PBIED VBIED
Loading direction Upward Downward
Loading nature Small standoff distances led to close-in detonations
Fell after being raised for
Edge Displacement Sank
a short duration
beam
Comparatively more
Concentrated in general
Damage extent widespread
Greater harm inflicted on the top part
Loading style Indirect Direct
Blast outcome Virtually superficial, no fractures were resulted
Abrupt load drop
followed by a momentary Substantial load gain
Axial force reloading
Cable
development Internal action totally diminished during ultimate
failure, but slowly stabilised when equilibrium was
able to be reinstated
Not beyond the yield capacity, let alone the ultimate
Load recorded
strength
4.7 SUMMARY
The cables were idealised as circular bars; the anchorage system was simplified as an
assembly of an anchor block and a bearing plate, assuming perfect gripping status and
superior blast resilience; all other auxiliary components were disregarded.
The stand-alone local model was extracted judiciously from the tributary zone for a given
stay. The modelling effort was much more rigorous (i.e. more explicit elaborations were
included), as compared to that dedicated to the global model (as described in the previous
182
chapter). Apart from the pros and cons of the specific modelling features, the research
purposes, the preferred accuracy levels, together with computational efficiency, were also
factored in while devising the modelling strategies, in choosing the element types and the
mesh distributions, selecting the constitutive material models, deciding on the connection
mechanisms, determining the loading details and the application modes, as well as
establishing the boundary concepts.
The local model contained beam, shell as well as solid elements. In the constitutive material
formulations selected for concrete and steel, strain rate behaviours were incorporated and
material failures were accounted for by invoking the erosion algorithms. When perfect or
simplistic connection was acceptable, node merging and the constraint approach were the two
options attempted. The two boundary tactics utilised were: translational and rotational
constraints, together with a longitudinal action introduced externally.
Once the initialisation stage (executed in a seemingly implicit manner) was finalised, the
blast simulation proceeded through the explicit routine. VBIEDs (250 kg TNT, 500 kg TNT
and 1000 kg TNT) were stationed on the deck while PBIEDs (10 kg TNT, 20 kg TNT and 30
kg TNT) were positioned beneath the primary girder. Large-extent bursts and small-extent
bursts were treated separately. A series of empirical formulae were employed to derive the
blast loads.
Based on the stress contour plot, the force paths indicating the stress dispersal within a typical
anchorage precinct was determined, and highlighted with the aid of multiple compressive
struts and tensile ties linked by imaginary nodes. Technically, the anchorage zone considered
in this research includes the anchor pod and extends into the edge beam until a uniform load
distribution is finally attained.
Under the blast action derived from all the charge weights allocated to the PBIED, the
anchorage region experienced devastating crushing failure in an extremely brittle manner,
owing to the sustained gravitational loading. On the other hand, the successful survival of the
anchorage point, where the system gradually stabilised, reaching new equilibrium, even under
183
Chapter 4
the presence of the VBIED, except with the most powerful explosive size adopted, was
accredited to the residual capacity retained.
The immediate blast-induced distortion was, overall, not profound, except when subjected to
the lateral blast pressure.
Disengagement of matter concentrated predominantly at the interior, but appeared not as

acute at the exterior. Material damage, which led to the deterioration of the initial resistance
of the anchorage precinct, should be viewed, in this particular context, as the predominant
failure mode that arises uniquely under blast loading.
Even though close-in detonations took place, the localised blast outcome, which exhibited
growing severity with more violent blast intensity for a given blast category, was actually
observed on the global extent, since the anchorage precinct was reasonably small, or in other
words, the blast sources were sufficiently large. The anchor pod generally sustained more
serious blast effect, in contrast to the sheltered volume.
The highly non-homogenous stress wave transmission caused a severely disturbed stress
condition, giving rise to an absolutely random stress distribution which was exceptionally
difficult to be explained using a universal rule. Meaningful comparison with the standard
stress distribution was, therefore, impractical.
The blast response of the longitudinal girder affected the blast performance of the anchorage
precinct by prompting a change in the axial load within the stay. However, anchorage point
failure provoked by the blast damage inflicted on the primary girder, which was obtained to
be relatively less dramatic, was not observed.
The cable was found to be resilient against the blast impact generated by the blast setups
employed here.
184
The blast loads delivered via the stay had virtually negligible effect on the performance of the
anchorage point fitted at the tower.
In this study, the criticality of the blast scenario with a PBIED, where direct blast impact was
generated from a closer location, generally surpassed that of the blast case with a VBIED,
where indirect blast action originated from a farther spot was involved. Nonetheless, the blast
implications due to the small and large explosives, at the same time, demonstrated both
similarities as well as differences, depending on the specific aspect of interest.
185
Chapter 4
186
Chapter 5 BLAST-RESILIENT RETROFITTING TECHNIQUES
5.1 OVERVIEW
Within this chapter, the strengthening proposals for the anchorage regions, which have been
formulated exclusively in response to the blast performance outcomes derived from PBIEDs
and VBIEDs, are described, and a discussion of the inherent capabilities of the different
options is presented. Additionally, the key variables of concern are also identified, and a
rigorous sensitivity analysis is undertaken. Finally, a set of essential guidelines is outlined to
facilitate the implementation of the fortification measures.
The Advanced Protective Technology for Engineering Structures Group in the University of
Melbourne conducted several projects to investigate the applicability of various materials in
blast protection. Those research findings are very useful for the development of the
retrofitting solutions in this study, as explained in this chapter.
Steel has been widely utilised throughout our history, and it continues to remain popular even
as of today; on the other hand, CFRP28 is invented only in more recent time but, nonetheless,
already gaining rising interest in its use in strengthening civil infrastructure. Here, a careful
evaluation of the accountability of these materials in blast hardening, particularly for the
anchorage zones in cable-stayed bridges, is conducted along with a judicious comparison of
their intrinsic strengths and weaknesses.
28
A composite of carbon fibres and a polymeric matrix
187
Chapter 5
5.2 BASIC CREDIBILITY OF PROTECTIVE MATERIALS
A total of nine 2000 mm × 1000 mm × 75 mm samples (four bare reinforced concrete slabs
and five strengthened specimens) were used by Tanapornraweekit (2010) in a research
project completed at the University of Melbourne. The glass-fibre-reinforced polymer
(GFRP), in a single-layer or double-layer arrangement, was attached either purely to the
bottom or to the top as well, with the aid of adhesive agents. The combination of GFRP and
CFRP was also attempted. The retrofitted panels were confronted with up to three
independent explosions. The TNT charge (0.45 kg or 0.90 kg or 1.36 kg) was raised to a
standoff distance of 0.5 m. The test setup is displayed in Figure 5.1.
Tanapornraweekit (2010) concluded that the desirability of the double-sided hardening

setups, which revealed nearly zero residual deflections even after the second detonations,
outshined that of the single-sided solution, but the provision of extra layers on either side
only led to marginal improvement. Both flexural and shear cracks were recorded. Only after
the third blast did the slab eventually give way, and concrete disintegration was observed
together with FRP delamination.
Figure 5.1: Experimental setup for FRP-protected reinforced concrete slabs

(Tanapornraweekit, 2010)
188
Blast-Resilient Retrofitting Techniques
In another study carried out at the University of Melbourne, Raman (2011) reported four
tests: one control panel and three polyurea-sprayed reinforced concrete slabs (all measured
1700 mm × 1000 mm × 60 mm). The coating, 4 mm or 10 mm, was sprayed directly onto the
designated surface (i.e. either on the shielded side alone or on the blast-receiving face as
well). The panels were secured to a rigid bunker, as shown in Figure 5.2, and the explosive, a
1 kg ammonite in circular shape, was suspended at a standoff distance of 1 m.
Raman (2011) confirmed that raising the thickness of the polyurea brought merely slight
benefit, but a huge advantage was evident with the sandwich arrangement. Only flexural
cracks were formed, while shear cracks were not detected. Based on careful visual
assessment, the coating was found to suffer only minor damage (which appeared in the form
of protuberances) while exhibiting a good bonding behaviour at the same time.
Figure 5.2: Experimental setup for polyurea-coated reinforced concrete slabs (Raman,
2011)
In a collaborative project between the Defence Science and Technology Organisation and the
University of Melbourne, Ackland et al. (2013) prepared nine samples: three 6 mm base steel
plates, and six polyurea-coated specimens. In order to give an identical areal density, the
spray, applied solely to the rear face of the 4 mm and 5 mm metal plates using primer, was
189
Chapter 5
15.7 mm and 7.7 mm thick (in average) respectively. As can be seen from the test
arrangement shown in Figure 5.3, the charge, a 0.5 kg spherical pentolite, was positioned
very close to the target.
De-bonding of the protective spray was observed, and it was much worse with the thicker
coating which exhibited very large hyperelastic extension (independent of the plate
deformation). Apart from that, Ackland et al. (2013) also noticed that, with a constant areal
density, it was much better to have a thinner layer of coating; in other words, the bare plates,
overall, showed lower residual displacements.
Figure 5.3: Experimental setup for polyurea-coated steel plates (Ackland et al., 2013)
5.3 BLAST FORTIFICATION CONCEPTS
Here, the elementary retrofitting alternatives attempted include applying the protective
material:
a) enclosing the exterior of the anchor pod

b) surrounding the internal face of the circular duct
c) covering the top surface of the edge beam
190
The basic defence options are either commissioned as stand-alone measures or implemented
as various integrated solutions: for a PBIED detonation, external encasing (Scheme A) or
internal lining (Scheme B) or both (Scheme C), as displayed in Figure 5.4, can be considered;
with a VBIED explosion, a top cover (Scheme A) or an external shield together with an
internal armour (Scheme B) or a combination of these three components (Scheme C), as
depicted in Figure 5.5, can be installed.
It is anticipated that multifaceted rewards will be obtained from the fortification strategies. It
has already been proven previously, by researchers such as Myers et al. (2004), Baylot et al.
(2005), Davidson et al. (2005), Hrynyk and Myers (2008) as well as Tan et al. (2012), that,
by sheltering the free surfaces, fragment ejection can be discouraged or debris will be
contained at least. Furthermore, utilisation of energy-absorbing materials helps to dissipate
the blast energy, and shielding from direct blast impact is believed to be crucial because this
will alleviate the subsequent stress transmission. The presence of retrofitting units improves
the overall stiffness, and Uy (2002) confirmed that attaching protective elements around a
column will actually restrict the axial deformation.
Also, it must be pointed out that hardening enclosures provides additional confinement which
substantially augments the overall capacity of the subject. Susantha et al. (2001) and Han et
al. (2008) demonstrated that a restraining effect was derived from the external steel jackets,
and likewise when substituted with the FRP wraps, according to Micelli and Modarelli
(2013). Apart from that, Han et al. (2010) advocated that internal tubes will lead to triaxial
restraint (where an extra restraining action will be activated in the radial direction) which is a
vital upgrade of the original biaxial mechanism which acts only along the circumferential
path.
191
Chapter 5
(a)
(b)
(c)
Figure 5.4: Retrofitting schemes for PBIED detonations – (a) External shield only; (b)
Internal armour only; (c) Integrated solution
192
(a)
(b)
(c)
Figure 5.5: Retrofitting schemes for VBIED explosions – (a) Top cover only; (b)
External casing as well as internal lining only; (c) Combined solution
193
Chapter 5
5.4 COMPUTATIONAL MODELLING
The strengthening members are generated using Belytschko-Tsay shell elements, as shown in
Figure 5.6. The membrane behaviour of these relatively thin components can be represented
reasonably well with shells, and therefore complicated solid elements are not required. The
dimensions of these shell units are selected to be compatible with the arrangement of the
solid mesh for ease of attachment.
Figure 5.6: Shell elements for the retrofitting system
5.4.2.1 Steel
Mat_Piecewise_Linear_Plasticity (Mat_024) is selected to describe the behaviour of steel.

The effect of loading rate is taken into account, and the bilinear stress-strain function is
defined by the material properties given in Table 5.1. Once the failure strain is exceeded, the
element is deleted from the calculation.
194
Table 5.1: Mechanical properties of the steel armour

ρ ν E fy fu εf
Steel component
Protective armour 7850 0.3 200 1000 1250 0.08
5.4.2.2 Carbon-Fibre-Reinforced Polymer
The material properties of CFRP employed in this study, which conform to the typical stress-
strain plot presented in Figure 5.7, are listed in Table 5.2.
Figure 5.7: Stress-strain plot of CFRP
CFRP is characterised using Mat_Enhanced_Composite_Damage (Mat_054) which takes into

account the anisotropic nature of the composite. This material model functions solely on shell
elements. The amount of unidirectional layer comprising the entire composite can be
assigned in Mat_054 by stating the desirable integration rule. Moreover, the stacking
sequence of the lamination can be prescribed by indicating the orientation angles.
195
Chapter 5
Table 5.2: Mechanical properties of CFRP (Tanapornraweekit, 2010)

Material attribute Input value
Poisson’s ratio, ν 0.27
Longitudinal modulus, EX 235.2 GPa
Transverse modulus, EY 78.4 GPa
Longitudinal tensile strength, Xt 2467 MPa
Transverse tensile strength, Yt 73 MPa
Longitudinal compressive strength, Xc 890 MPa
Transverse compressive strength, Yc 366 MPa
Shear strength, S12 115 MPa
Mat_054 considers the tensile and compressive failure of both the fibre and the matrix. The
softening of the tensile and compressive strength after matrix failure can be defined. The four
failure modes implemented in this study are described below:
a) For tensile fibre mode, if σaa > 0, then:
2
𝜎𝑎𝑎 2
𝑒𝑓 =( ) −1 (≥ 0 failed, < 0 elastic) (5.1)
𝑋𝑡
b) For compressive fibre mode, if σaa < 0, then:

𝜎𝑎𝑎 2
𝑒𝑐 2 = ( ) −1 (≥ 0 failed, < 0 elastic) (5.2)
𝑋𝑐
c) For tensile matrix mode, if σbb > 0, then:

𝜎𝑏𝑏 2 𝜎𝑎𝑏 2
𝑒𝑚 2 = ( ) + ( ) − 1 (≥ 0 failed, < 0 elastic) (5.3)
𝑌𝑡 𝑆
d) For compressive matrix mode, if σbb < 0, then:

𝜎𝑏𝑏 2 𝑌𝑐 2 𝜎𝑏𝑏 𝜎𝑎𝑏 2
𝑒𝑑 2
= ( ) + [( ) − 1] +( ) −1 (≥ 0 failed, < 0 elastic) (5.4)
2𝑆 2𝑆 𝑌𝑐 𝑆
196
where:
σaa = stress parallel to fibre direction
σbb = stress normal to fibre direction
σab = shear stress
Alas, Mat_054 is unable to incorporate the influence of strain rate. However, since CFRP is
relatively insensitive to loading rate effect (Welsh and Harding, 1985; Kimura et al., 2001),
only the static properties are adopted for the blast study.
CFRP is evidently much weaker across the fibre axis. In order to ensure adequate
strengthening in both ways, a bidirectional arrangement, whereby the transverse attributes are
overwritten by the longitudinal attributes, is assumed.
5.4.3 Contact and Connections
The Contact_Automatic_Surface_to_Surface_Tiebreak command is invoked to imitate the

adhesive effect of the epoxy resin used to bond the steel plates and the CFRP layers to the
concrete (refer to Figure 5.8).
Figure 5.8: Contact mechanism
The associated interfaces will remain attached until failure of the intermediate contact
surfaces happens when the condition stated in the equation below is fulfilled.
197
Chapter 5
2 2
|𝜎𝑛 | |𝜎𝑠 |
( ) +( ) ≥1 (5.5)
𝑁𝐹𝐿𝑆 𝑆𝐹𝐿𝑆
where:
σn = normal stress
σs = shear stress
NFLS = normal failure stress
SFLS = shear failure stress
The tensile and shear strength of the adhesive agent are given in Table 5.3.
Table 5.3: Mechanical properties of adhesive agent

Normal strength Shear strength
Substance
(MPa) (MPa)
Epoxy resin 100 100
The simulation tactics for connections, as described by Tabiei and Wu (2000), include:
a) node merging between the joined parts

b) node tying until failure occurs
c) explicit presentation of the fasteners
d) non-linear spring modelling
The first option will lead to an unreliable representation of the actual behaviour of the
fasteners, particularly during pull-out failure. Although the third option leads to the most
accurate outcome, it incurs huge simulation costs. The last technique, which demands
elaborate modelling in order to produce the force vs. displacement curves for the springs, is
not practical here when dynamic blast response and complicated loading directions are
considered. The second method is selected for this exercise; aside from being
computationally efficient, this approach also produces reasonably good results, and has been
198
verified by Mutalib and Hao (2011) based on the blast test data published by Muszynski and
Purcell (2003).
The bolts and the anchors used to improve the attachment of the steel plates as well as the
CFRP strips respectively are idealised as distinct nodes according to the
Contact_Tiebreak_Nodes_to_Surface option. Failure of the connections happens once the
condition prescribed in the following equation is satisfied, and the nodes are subsequently
released from the contact faces.
2 2
|𝐹𝑛 | |𝐹𝑠 |
( ) +( ) ≥1 (5.6)
𝑁𝐹𝐿𝐹 𝑆𝐹𝐿𝐹
where:
Fn = normal force
Fs = shear force
NFLF = normal failure force
SFLF = shear failure force
The normal and shear strength of the connections are given in Table 5.4. A tensile strength of
830 MPa (AS 4100-1998) is assumed for the 24 mm diameter steel bolts, while the capacity
of the CFRP anchors is taken from Hao and Tang (2010). The fasteners are installed
approximately 0.5 m apart along the edges of the retrofitting members.
Table 5.4: Mechanical properties of fasteners

Normal failure force Shear failure force
Connection unit
(kN) (kN)
Bolt 293 167
Anchor 261 100
199
Chapter 5
5.5 HARDENING WITH STEEL
5.5.1 Feasibility of Defence Schemes
The external shield, the internal tube as well as the top armour were installed using steel
plates which were 20 mm thick.
The anchorage movement and the cable force developed are presented in Figure 5.9, Figure
5.10 and Figure 5.11, Figure 5.12 respectively. For the case of below-girder blast impact, the
anchorage precinct managed to survive only when the integrated solution (Scheme C for
PBIEDs) was implemented, as it gradually stabilised, achieving new equilibrium eventually.
For the case of above-deck blast action, destruction of the anchorage point was not prohibited
when only the top cover was provided (Scheme A for VBIEDs), and the stay experienced
total loss of axial load.
Lining the inner surrounding (Scheme B for PBIEDs) was proven to be impressively
effective in limiting the blast damage, and debris ejection at the hollow core was completely
stopped. In comparison, covering the outer faces (Scheme A for PBIEDs) was noticed to be
much less effective, even in controlling the blast damage inflicted at the bottom part of the
target (as shown in Figure 5.13), and the reloading process persisted over a shorter period,
while brittle crushing took place at an earlier stage. Unfortunately, both strengthening options
managed to secure merely marginal improvement when implemented independently.
In Figure 5.14, it is apparent that, with the shelter, the detrimental impact on the centre
portion of the top section resolved to become more moderate, but the overall damage
produced on the edge beam was not necessarily less acute. Furthermore, it was confirmed that
the overlaying shield did not interfere directly with the integrity of the anchorage point, and
the overall benefit generated was only able to delay, but not prevent, the ultimate failure
process, as the constricted stress wave transmission was still able to provoke material
breakdown. Nevertheless, when it was installed in conjunction with the inner tube together
200
with the outer casing (Scheme C for VBIEDs), the best enhancement was resulted, in contrast
to the case which without (i.e. Scheme B for VBIEDs).
The axial stress experienced by the general zone reinforcing steel in both combined solutions
(i.e. the most robust setups) is traced in Figure 5.15 and Figure 5.16. Immediately after the
blast arrival, the sudden peaks appeared lower than the corresponding values reported from
the original configuration. Thus, it can be confirmed again that the anchorage region suffered
less severe blast damage. The overall stress development was also noticed to be less dramatic,
as anticipated. The steel bars gradually settled down to normal strain conditions, without
yielding or fracturing.
In Figure 5.17 and Figure 5.18, the effective stress histories extracted from the blast-receiving
shields in both integrated solutions are plotted. The reference nodes were taken at 0.5 m and 1
m intervals on the outer jacket and the overlaying shell respectively. At the locations nearest
to the blast sources (designated as the centre points), the highest peaks, shown almost
instantaneously after the arrival of the blast waves, were recorded. Fortunately, the armour
neither yielded nor fractured, but it must be mentioned that the stress development instigated
by the large-scale explosion was not as intense.
Bond removal occurred due to material disintegration, but complete detachment of the steel
plates was not detected, and this implies that fragments will be effectively contained behind
the armour, instead of being projected freely in a random manner.
201
Chapter 5
Figure 5.9: Anchorage movement on account of PBIED detonations (with different steel
strengthening options)
Figure 5.10: Anchorage movement as a result of VBIED explosions (with different steel
strengthening options)
202
Figure 5.11: Cable force data under PBIED detonations (with different steel hardening
schemes)
Figure 5.12: Cable force data following VBIED explosions (with different steel
hardening schemes)
203
Chapter 5
At 0.05 s
(Scheme A: External shield only)
At 0.08 s
(Scheme B: Internal armour only)
At 0.6 s
(Scheme C: Integrated solution)
detonations (with unique steel retrofitting configurations)
204
At 0.235 s
(Scheme A: Top cover only)
At 0.5 s
(Scheme B: External casing and internal lining only)
At 0.5 s
(Scheme C: Combined solution)
explosions (with unique steel retrofitting configurations)
205
Chapter 5
the best steel protective solution)
the best steel protective solution)
206
Figure 5.17: Effective stress records after PBIED detonations (with the best steel
fortification plan)
Figure 5.18: Effective stress records produced by VBIED explosions (with the best steel
fortification plan)
207
Chapter 5
5.5.2 Influence of Steel Class
Distinct steel grades from the low, normal and high strength bands, with details given in
Table 5.5, were employed for comparison purpose. With increasing strength, the fracture
strain drops, as the brittleness of the metal becomes greater.
Table 5.5: Mechanical properties of steel

ρ ν E fy fu εf
Steel category
Low strength 350 450 0.21
Normal strength 7850 0.3 200 500 650 0.15
High strength 1000 1250 0.08
Ultimate failure was prevented in all the cases considered, as evident in Figure 5.19 as well as
Figure 5.20. The weakest steel was already capable of preventing the anchorage zone from
experiencing total destruction, and the use of the strongest steel merely brought slightly better
outcome.
The effective stress data is plotted in Figure 5.21 and Figure 5.22. With the smaller charge,
the recorded peaks exceeded the smallest static yield point, but no fracture was detected. On
the other hand, the stress values resulted from the larger explosive, which was positioned
farther away, were very much lower, and not at all sufficient to induce any degree of damage
to the steel members.
Higher strength steel, albeit less adversely affected by blast pressure, may not be attractive in
terms of energy absorption which is usually dominated by ductility.
208
Figure 5.19: Anchorage displacement histories after PBIED detonations (with several
steel grades)
Figure 5.20: Anchorage displacement histories after VBIED explosions (with several
steel grades)
209
Chapter 5
Figure 5.21: Effective stress information from PBIED detonations (with different steel
grades)
Figure 5.22: Effective stress information from VBIED explosions (with different steel
grades)
210
5.5.3 Significance of Plate Section
The merits of increasing the thickness of the steel plates, from 5 mm to 10 mm and then
finally to 20 mm (i.e. progressively doubled), were assessed. The total tonnage of steel used
in each configuration is also given in Table 5.6.
Table 5.6: Armour sizes

Armour thickness Steel volume Estimated weight
Case
(mm) (m3) (ton)
1 5 0.131 1.03
2 10 0.261 2.05
3 20 0.518 4.07
According to Figure 5.23 and Figure 5.24, the anchorage region gave way only when the
thinnest option was adopted in the PBIED event, but, regardless of the plate section, it
survived the VBIED event which required less vigorous provision.
Reducing the armour size implied that less material was involved, and a lower degree of
fortification was achieved, thereby leading to less impressive improvement.
The effective stress data is presented in Figure 5.25 and Figure 5.26. It was observed that,
under an identical blast magnitude, the greatest peak was induced in the thinnest plate, as
expected (since the blast endurance of the steel members grows with their thickness). Under
the below-girder blast action, the stress rose beyond the static yield threshold and minimal
damage was inflicted on the blast-receiving face, but aside from that, no failure was detected
on the high strength armour in all the other setups.
211
Chapter 5
Figure 5.23: Anchorage displacement histories after PBIED detonations (with different
plate dimensions)
Figure 5.24: Anchorage displacement histories after VBIED explosions (with different
plate dimensions)
212
Figure 5.25: Effective stress information from PBIED detonations (with several plate
dimensions)
Figure 5.26: Effective stress information from VBIED explosions (with several plate
dimensions)
213
Chapter 5
5.5.4 Importance of Bonding Competency
Intermediate adhesive agents play an important role in affecting the blast performance of the
anchorage structure, and to prove this, three bond capacities which can be sourced from the
market (refer to Table 5.7), i.e. 30 MPa, 60 MPa and 100 MPa, were chosen.
Table 5.7: Mechanical properties of epoxy resin

Normal strength Shear strength
Case
(MPa) (MPa)
1 30 30
2 60 60
3 100 100
Following the small-scale burst: bond removal happened except with the strongest epoxy
resin, the bottom plate was isolated while the side shields were still held in place, but the
internal tube was untouched, as shown in Figure 5.27. Nonetheless, the anchorage zone was
totally demolished only when the weakest adhesion came into play (refer to Figure 5.28). On
the other hand, the anchorage region was still fully operational even after the large-extent
burst, as suggested in Figure 5.29, and the armour was found to remain firmly attached.
Figure 5.30 and Figure 5.31 show the effective stress plots. The metal neither fractured nor
yielded. It must be mentioned that, from the much higher peak recorded, the deactivation of
the weakest adhesion due to the below-girder blast action was confirmed.
Upgrading the quality of the bond only managed to secure marginal improvement when
dealing with the vehicle bomb, but the enhancement was more obvious when the portable
charge was concerned, as the material breakdown, which also contributed to the disablement
of surface contact (as can be seen in Figure 5.27), eased substantially.
214
Thus, the effectiveness of the strengthening approach is indeed dictated by the adhesive
substance adopted, since its presence is indispensable for the existence of the composite
behaviour between the subject and the fortification.
(a)
(b)
Figure 5.27: Bond failure and contact disablement (under different bond capacities) –
(a) 30 MPa epoxy resin; (b) 60 MPa epoxy resin
215
Chapter 5

various bond capacities)
Figure 5.29: Anchorage displacement records following VBIED explosions (with various
bond capacities)
216
Figure 5.30: Effective stress data due to PBIED detonations (with distinct bond
capacities)
Figure 5.31: Effective stress data due to VBIED explosions (with distinct bond
capacities)
217
Chapter 5
5.5.5 Importance of Connection System
Fasteners are often used in conjunction with adhesive agents to secure the defensive units.
Here, several ideas were attempted: besides distributing the bolts only along the perimeter or
throughout the whole designated surface, the consequence of omitting the connections was
also determined.
From Figure 5.32 and Figure 5.33, it is apparent that the anchorage point continued to remain
functional after both the PBIED and the VBIED events, even when the supplementary
fasteners were completely disregarded. This suggests that excellent adhesion was confidently
guaranteed by the bonding agent alone.
Detachment of the metal shields did not take place. Notwithstanding the stronger attachment
achieved, having additional bolts did not lead to an obvious advantage, rather, it may cause a
slight deterioration in the condition after blast, owing to stress concentration particularly on
the blast-receiving face, but the effect was insufficient to bring about a significant deviation
in the final response.
The effective stress plots are displayed in Figure 5.34 as well as Figure 5.35. The armour
maintained its perfect condition and did not suffer any form of damage from the extreme
loading.
218

distinct connection arrangements)
Figure 5.33: Anchorage displacement records following VBIED explosions (with distinct
connection arrangements)
219
Chapter 5
Figure 5.34: Effective stress data due to PBIED detonations (with various connection
arrangements)
Figure 5.35: Effective stress data due to VBIED explosions (with various connection
arrangements)
220
5.6 TOUGHENING WITH CARBON-FIBRE-REINFORCED POLYMER
5.6.1 Viability of Defence Plans
The outer shield, the inner lining as well as the overlaying shelter were constituted with 3 mm
thick CFRP sheets.
The anchorage displacement plots and the cable load histories are displayed in Figure 5.36,
Figure 5.37 and Figure 5.38, Figure 5.39 respectively. Upon encountering the below-girder
blast impact, the anchorage precinct was not destroyed, but settled down steadily (hence
remained operational), only when the integrated solution (Scheme C for PBIEDs) was
implemented. When confronted with the above-deck blast action, the anchorage zone
experienced disastrous failure when equipped only with the top cover (Scheme A for
VBIEDs), and total loss of axial force in the stay was recorded.
As can be seen in Figure 5.40, the external casing (Scheme A for PBIEDs) was not at all
effective, since heavy blast damage was still inflicted on the whole anchorage region, and
brittle crushing was not avoided. The advantage of employing the internal lining (Scheme B
for PBIEDs) was much greater, as material disintegration at the inner core was significantly
limited, but such benefit was only sufficient to prolong the reloading process, instead of
preventing total destruction. Thus, both hardening options were not useful when presented
alone.
Shielding the top of the edge beam brought merely marginally improvement, as the ultimate
failure of the anchorage point was not significantly delayed. While the adverse effect on the
middle section of the top portion eased slightly, the overall damage on the girder was not
necessarily less profound, as shown in Figure 5.41. The top cover had no direct contribution
to the integrity of the anchorage zone, and the constricted stress wave transmission was still
capable of instigating substantial material breakdown. This overlaying shell needs to be
installed along with the outer casing and the inner tube (Scheme C for VBIEDs) in order to
221
Chapter 5
obtain the most outstanding result, although the minimum expectation (i.e. eventual survival
of the anchorage point) was still able to be met even without its participation (Scheme B for
VBIEDs).
The axial stress development within the general zone reinforcing steel in both combined
solutions (i.e. the most impressive configurations) is traced in Figure 5.42 and Figure 5.43.
The peak values, appeared immediately after the blast arrival, were smaller, in comparison to
those reported from the unprotected setup, and this again indicated that the anchorage zone
resolved into a less destructive condition. The overall stress build-up was found to be less
intense, and the steel bars, which neither yielded nor fractured, gradually reaching new
equilibrium.
The maximum principal stress plots, obtained from the blast-receiving components in both
integrated solutions, are given in Figure 5.44 and Figure 5.45. The reference nodes were
taken at 0.5 m intervals on the external wrap but 1 m intervals on the overlaying sheet. Under
the blast action derived from the portable charge, the CFRP strip suffered acute rupture (as
obvious in Figure 5.40), as the stresses dropped to zero instantaneously after the ultimate
strength was exceeded. As for the case with the vehicle bomb, the stresses induced within the
CFRP layer were less dramatic (causing no rupture), and the highest peak was recorded at the
location closest to the blast source.
Following the elimination of the composite, it is believed that random ejection of fragments
will be inevitable, unfortunately. Bond removal due to material disintegration was not
avoided, but complete detachment of the CFRP sheets did not happen.
222

CFRP strengthening options)

CFRP strengthening options)
223
Chapter 5
Figure 5.38: Cable force data under PBIED detonations (with different CFRP
hardening schemes)
Figure 5.39: Cable force data following VBIED explosions (with different CFRP
hardening schemes)
224
At 0.03 s
(Scheme A: External shield only)
At 0.11 s
(Scheme B: Internal armour only)
At 0.6 s
(Scheme C: Integrated solution)
detonations (with unique CFRP retrofitting configurations)
225
Chapter 5
At 0.225 s
(Scheme A: Top cover only)
At 0.5 s
(Scheme B: External casing and internal lining only)
At 0.5 s
(Scheme C: Combined solution)
explosions (with unique CFRP retrofitting configurations)
226
the best CFRP protective solution)
the best CFRP protective solution)
227
Chapter 5
Figure 5.44: Maximum principal stress records after PBIED detonations (with the best
CFRP fortification plan)
Figure 5.45: Maximum principal stress records produced by VBIED explosions (with
the best CFRP fortification plan)
228
5.6.2 Importance of Composite Quality
CFRP is manufactured with a large range of constituents, and this warrants the necessity to
determine how the compositions govern the blast protection afforded. Here, three types of
composite units, composed of distinct constitutive fibres and polymers, were selected, and
their properties are listed accordingly in Table 5.8. It can be seen that, overall, Composite C
was, strictly speaking, merely slightly stronger than Composite B, whereas Composite A was
apparently the weakest.
Table 5.8: Mechanical properties of CFRP of various qualities

Material attribute Composite A* Composite B** Composite C***
ρ (kg/m3) 1512 1580 -
ν21 0.0127 0.021 -
ν31 - 0.021 -
ν32 - 0.49 -
G12 (GPa) 4.8 5.24 -
G23 (GPa) 4.8 2.24 -
EX (GPa) 118 138 235.2
EY (GPa) 5.5 9.65 78.4
Xt (MPa) 1095 2280 2467
Yt (MPa) 26.4 57 73
Xc (MPa) 712.9 1440 890
Yc (MPa) 84.4 228 366
S12 (MPa) 84.3 71 115
S23 (MPa) - 71 -
* Taken from Han et al. (2007)
** Reported by Chan et al. (2007)
*** Given by Tanapornraweekit (2010)
229
Chapter 5
The composites employed were competent enough to ensure that the anchorage precinct
continued to serve its intended use after the smaller charge and the larger explosive were
triggered, according to Figure 5.46 and Figure 5.47.
The maximum principal stress histories are plotted in Figure 5.48 and Figure 5.49. All the
composites attempted did not manage to satisfy the strain demand due to the direct blast
action from the PBIED burst, and thus fractured, predominantly at the front face, and it must
be noted that Compound B also suffered minor de-bonding. Conversely, no harm was done
by the VBIED burst, and it was realised that, in this situation, Compound A and Compound C
experienced the lowest and the highest stress development respectively, owing to their
distinct stiffness moduli.
Here, it was confirmed that the best protection was enabled by Composite C, followed by
Composite B, and lastly Composite A, as the material breakdown became noticeably not as
intense.
A compound with more superior strength is certainly less likely to be destroyed by extreme
blast loading, but may not reveal attractive energy-dissipating capability which is also shaped
by its flexibility together with its softening behaviour. Furthermore, the rigidity of the
compound is also a paramount criterion that affects the overall distortion and dictates the
restraining contribution.
230

CFRP of distinct constitutive compositions)

CFRP of distinct constitutive compositions)
231
Chapter 5
Figure 5.48: Maximum principal stress plots after PBIED detonations (with CFRP of
different constitutive compositions)
Figure 5.49: Maximum principal stress plots after VBIED explosions (with CFRP of
different constitutive compositions)
232
5.6.3 Significance of Layer Quantity
It is very common that several sheets of CFRP are attached together to achieve the required
layer thickness. Here, the individual strip was assumed to be 0.5 mm thick, and the composite
was decided to have a combined depth of 1, 2 or 3 mm, hence a steady augmentation of two
sheets (refer to Table 5.9).
Table 5.9: Composite layers

Strip size Layer quantity Composite thickness
Case
(mm) (mm)
1 2 1
2 0.5 4 2
3 6 3
From Figure 5.50 as well as Figure 5.51, it is obvious that, in both blast settings, the
anchorage region maintained in a reasonably good state, and no interruption was brought to
its service, regardless of the layer thickness of the composite.
A better defence effort, capable of inhibiting a heavier capacity loss, was made with the
provision of additional sheets of CFRP. Nonetheless, the amount of material increase
considered here was insufficient to give rise to a meaningful boost in the protection offered.
The maximum principal stress histories are displayed in Figure 5.52 and Figure 5.53. When
confronted with the small-extent burst, acute fracture of the CFRP concentrated
predominantly at the blast reception, but the strain demand due to the large-scale burst
permitted the composite to maintain in a perfect state. The fact that the thinnest setup
incurred the highest stress development, without surprise, indicated that, with extra layers of
constitutive strips, the composite will not break so easily.
233
Chapter 5

several composite layers)

several composite layers)
234
Figure 5.52: Maximum principal stress plots after PBIED detonations (with various
composite layers)
Figure 5.53: Maximum principal stress plots after VBIED explosions (with various
composite layers)
235
Chapter 5
5.7 CRUCIAL COMPARISON OF STRENGTHENING ALTERNATIVES
Steel has secured a very long history in the engineering industry, whereas CFRP consists of
advanced constituents innovated only at relatively recent times to satisfactorily fulfil the
increasingly stringent and updated requirements in multiple sectors.
A general comparison of the basic mechanical attributes of the retrofitting materials

attempted is presented in Table 5.10.
Table 5.10: Relative merits of steel and CFRP

Attribute* Steel CFRP
Density – +
Resistance 0 +
Rigidity + 0
Plasticity 0 –
* + = superior, 0 = moderate, – = inferior
Typically, CFRP enjoys exceptionally high strength and stiffness, but it is extremely brittle,
and for this reason, very unattractive when it comes to energy consumption. Overall, steel
demonstrates notable robustness in all the aspects of concern, while showing the largest mass
per unit volume.
Strictly speaking, both materials considered are suitable to be utilised as formidable blast
fortification for the anchorage zone, as it was able to remain fully operational, despite the
limited damage inflicted upon it, successfully avoiding catastrophic destruction which has the
potential to bring disastrous consequences. The anchorage point was not eliminated by the
small-scale burst when the inner tube was installed together with the outer armour, but was
deactivated when confronted with the large-extent burst if only the top cover was provided.
236
Nonetheless, the steel plates did not fracture, but removal of the CFRP sheets due to blast
failure, especially at the blast reception, was noticed, unfortunately. Complete detachment of
the retrofitting units did not happen.
Moreover, it was demonstrated that the fortification schemes involving steel, overall, offered
better defence, as under the protection of the strengthening setups utilising CFRP, the blast
damage suffered by the anchorage region was generally of greater severity.
5.8 BLAST RETROFITTING STRATEGIES
5.8.1 General Comments
The retrofitting solutions proposed concern chiefly with the application of steel (metal) as
well as CFRP (polymer) in blast defence. However, should other options be utilised instead,
the key concepts covered here, which can be conveniently adapted to suit the specific context,
can actually be taken for basic reference use.
The protective methods suggested here shall be treated only as general guidance, but as the
materials of interest are often composed of a diverse range of constituents with widely varied
attributes, rigorous investigations are, nevertheless, mandatory for the justification of the
decisions made in each unique situation.
Although dedicated primarily to the cable anchorage zones in existing bridges, the
instructions given here are equally valid for the implementation of composite design in new
cable-stayed bridges.
Apart from affording extra blast protection, the enhancement effort is also desirable for the
subject to better withstand the conventional loading encountered during the routine operation.
237
Chapter 5
Various protection categories can be commissioned by simply adjusting the strengthening

scheme, while ensuring the successful survival of the anchorage region shall be maintained as
the principal objective.
Modelling approaches that will enable genuine blast outcomes to be realistically captured
must be employed. Being reasonably slender, the hardening units should be replicated with
shell elements characterised by appropriate constitutive formulations. Adhesion due to the
bonding agent can be represented through proper attachment features, and the intermediate
fasteners are encouraged to be idealised as virtual nodal entities.
5.8.2 Roles and Functions
Vigorous blast hardening (which further upgrades the original resistance), along with robust
protection from potential blast damage, are indispensable for the anchorage precinct to
remain functional after experiencing extreme blast loading, and shall be offered through the
following means:
a) discouragement of matter disengagement (which can drastically undermine the

integrity of the subject), when the free surfaces at the interior and the exterior are
adequately sheltered
b) utilisation of substances well known for energy dissipation, as this guarantees
remarkable stress relief, and shielding the blast reception is, therefore, warranted
c) promoting the global axial and transverse as well as local rigidity, with the aid of
a proper strengthening mode, depending on the anticipated blast response, will
restrict the overall distortion
d) supplementary restraining effort from the armour covering the outer surrounding,
together with the inner lining at the hollow core, which is required to expand the
initial capacity
Of course, the most powerful defence will be afforded when all the merits outlined above are
achieved simultaneously within the entire fortification setting.
238
Aside from material enhancement, vital especially when dealing with near-field incidents,
structural improvement, although commonly emphasised for far-field events, should not be
neglected as well.
The exact roles as well as the distinct functions of the defence tactics adopted must be
cautiously distinguished, as this facilitates technology adaptation to and from other related
fields and sectors, besides enabling the most appropriate retrofitting strategy to be devised to
suit the unique situation encountered.
5.8.3 Configurations and Forms
The internal armour, arguably the most important piece of equipment in the entire retrofitting
setting, is indispensable, but the company of the external shield shall not be compromised, as
none is able to confidently guarantee alone that the anchorage point will remain operational
during the aftermath of a devastating blast catastrophe, even though both bring direct
advantages to the anchorage precinct.
All the exposed faces of the anchor pod must be protected. Strictly speaking, it is compulsory
to strengthen only the extent of the central core enclosed within the anchorage region, but
technical convenience should be factored in in actual practice; the guide tube, an accessory
incorporated within the cable-and-anchorage system, may be utilised to fulfil this specific
responsibility, but it must, nevertheless, never be perceived automatically as an existing
protection in this respect, owing to the fact that it is not an integral content essential for the
establishment of the anchorage zone.
If the blast action is generated on top of the deck, the top shelter (adequately catering for the
whole anchorage region) ought to be treated as an optional provision, although beneficial to
be included, owing to its indirect significance to the well-being of the anchorage point, but it
is completely irrelevant if the burst is released below the girder.
239
Chapter 5
A composite scheme that makes use of several substances should be strongly encouraged so
that their unique strengths can be fully exploited, while the shortcomings due to their inherent
weaknesses are appropriately countered.
5.8.4 Material Selection
In order to give basic advice on material selection (type and quality), the way their
characteristics serve in blast retrofitting is clarified with the aid of the example shown in
Figure 5.54.
Figure 5.54: Unique significance of material attributes
A substance with superior strength is less likely to suffer from blast damage, but its ability to
dissipate energy is dictated by the amount of work that can be done prior to ultimate failure.
The flexibility of the material is accountable for distortion control and confinement upgrade.
After encountering the blast action, if the hardening matter maintains in an excellent
condition, less work is done, but it will continue to serve its intended use; on the other hand,
if it is damaged or fractured, the remaining credibility of the fortification setup will be
unfavourably demoted, even though a larger amount of blast energy will be consumed.
240
It is absolutely pivotal to establish a clear idea of the desired behaviour to be anticipated

during an uninvited blast occasion so that the best option is assigned to undertake the unique
responsibilities.
For the sake of fidelity, strain rate behaviour, when applicable, must be considered in order to
arrive at realistic decisions.
5.8.4.1 Steel Armour
For steel, typically, resistance is inversely correlated to plasticity, and therefore both high and
low grade versions are indeed equally useful, depending on the exact service being asked for.
Technically speaking, steel has a rather consistent stiffness modulus. If the given applied
stress resides purely within the elastic range, a similar level of defence will be secured,
regardless of the quality of the metal, since a comparable response is to be expected; in other
words, damage or even failure of the hardening units is warranted for distinct degrees of
protection to be noticed.
The merits of a strengthening matter amplify with its member size. Thus, a metal shield,
which appears in the form of a flat plate or hollow tube, must have adequate thickness which
also makes itself less susceptible to the adverse effect of blast action.
Steel is suitable for casting the top shield, the outer jacket as well as the inner shell. Owing to
its overall robustness, steel is absolutely competent to promise that all the critical tasks
needed to be undertaken by the retrofitting effort are satisfactorily carried out. Nevertheless,
its heavyweight nature is often an unfavourable aspect.
5.8.4.2 Carbon-Fibre-Reinforced Polymer Layer
The pros and cons associated with the most common types of lightweight FRP are explained
in Table 5.11. In general, except for manufacturing and supply cost, the prestige of CFRP
241
Chapter 5
evidently surpassed that of both GFRP as well as aramid-fibre-reinforced polymer (AFRP),

judging from their intrinsic characteristics, hence the greater suitability to be chosen to
participate in harsh situations.
Table 5.11: Qualitative appraisal of FRP (Christoffersen et al., 1999)

Attribute* GFRP AFRP CFRP
Environmental resistance – + +
Tensile strength + + ++
Fatigue strength 0 – ++
Elastic modulus – – ++
Creep/relaxation – 0 ++
Stress fatigue – – ++
Density + ++ ++
Material price ++ – –
* ++ = very good, + = good, 0 = neutral, – = not good
CFRP is indeed capable of reaching ultra-high strength and exhibiting notable rigidity, but
usually lacks the generous strain capacity required to meet particularly the strict demand of
violent blast pressure which often prompts the compound to suffer critical rupture. With
brittleness standing as its major drawback, CFRP is less desirable for energy consumption,
especially at the blast reception.
The amount of matter involved, hence the number of layer assigned, is a crucial factor that
dictates not only the degree of protection offered but also the inherent endurance of the
composite.
Given the non-isotropic nature of CFRP, individual strips should be oriented in an orthogonal
manner so that the stronger attributes along the longitudinal axis are also valid across the
weaker transverse direction.
242
5.8.5 Installation and Connection
An adhesive agent not only dictates the quality but also stands as the dominant cause of the
existence of the composite behaviour between the anchorage region and the fortification.
Thus, the presence of the adhesive substance is mandatory, and assurance of excellent
bonding competency is absolutely crucial.
Fasteners, typically bolts and anchors, albeit useful to be included, shall never be treated as a
potential alternative to the bonding agent (which enables more effective stress transmission
and strain compatibility), but only assume an auxiliary role instead, and should be aligned
along the boundary rather than across the whole designated surface. If vigorous adhesion is
confidently guaranteed, reliance on this ancillary service may be disregarded.
It must be mentioned that, apart from bond removal, surface contact can also be disabled
through debris ejection. When complete detachment occurs, the fortification will no longer
fulfil its promises, and this totally defeats the purpose of its installation, while partial
deactivation of the composite effort will only undermine the defence, and the final
performance will not be able to meet the expectation.
5.9 SUMMARY
The elementary defence concepts formulated, which include the application of a shelter on
the top face of the edge beam, a set of armour surrounding the anchor pod as well as a tube
inserted into the hollow core, were either commissioned as independent options or selectively
combined to form integrated solutions.
The hardening units were created using shell elements which can be removed based on the
erosion features employed. For CFRP, the effect of loading rate was ignored, but this was not
the case for steel. Surface attachment was activated with the aid of a contact algorithm, while
discrete connection was enabled through node tying.
243
Chapter 5
Steel and CFRP have unreservedly demonstrated outstanding credibility in confidently

assuring the critical survival of the anchorage precinct after encountering the enormous blast
impact originated from both PBIEDs and VBIEDs, and are, thus, eligible to be designated as
viable blast retrofitting materials, each with unique strengths and weaknesses to be sensibly
exploited or appropriately managed, depending on the circumstances. Nevertheless, steel,
instead of CFRP, was able to offer better protection in this context. The steel plates did not
fracture, but failure of the CFRP sheets occurred.
With the smaller charge, the anchorage point remained serviceable only when the outer jacket
was installed in conjunction with the inner shell (the former was less reliable, as compared to
the latter), but not independently. With the larger bomb, the identical fortification setting was
also proven to be very effective, but extra benefit was obtained with the participation of the
top cover which was, however, found to be incompetent to act alone.
Based on the findings of the investigation programme and with the aid of the rigorous
sensitivity study, feasible retrofitting guidelines have been delivered. The defence
recommendations, which serve purely as general advice, are valid to not only current
facilities but also future assets. Although intended principally for blast hardening, the
advanced techniques established are indeed equally useful when dealing only with the
traditional loading. The ultimate objective of a fortification scheme shall be to prevent the
anchorage point from being destroyed by extreme blast action.
Apart from formidable fortification, generous enhancement of the intrinsic endurance must
also be strongly emphasised, and these two distinct contributions, which lead to both
structural and material advantages, shall be achieved by: shielding the boundary faces so as to
inhibit matter disengagement and offer additional restraint; employing substances renowned
for energy consumption, which will guarantee remarkable stress relief; and upgrading the
rigidity of the subject (especially for restricting distortion).
244
The resistance of a material controls the extent of the blast damage, but if the defensive unit
maintains in a perfect state, less energy is absorbed; the ability to consume energy is
equivalent to the amount of work that can be undertaken, but upon suffering a certain degree
of damage or experiencing acute fracture, the multifaceted services provided will be
adversely affected. A substance with higher stiffness, besides capable of showing a better
restraining effort, is also more attractive in limiting deformation.
The mass of the matter of concern, manipulated in terms of plate size or sheet layer, dictates
the basic endurance of the protective member as well as the effectiveness of the hardening
setup.
As a key criterion for the establishment of the composite behaviour between the subject and
the defence, it is mandatory to involve an adhesive agent of outstanding quality, since
supplementary fasteners only put in secondary effort. Material loss and bond failure both lead
to deactivation of the intermediate interaction, which partially or totally undermines the
fortification.
245
Chapter 5
246
Chapter 6 PARAMETRIC STUDY AND BLAST-RESISTANT
DESIGN GUIDELINES
6.1 OVERVIEW
Within this chapter, a comprehensive parametric investigation is conducted to not only

thoroughly evaluate but also carefully compare the influence of all the principal design
variables. Presented along is an in-depth discussion of the general blast design
recommendations proposed to address the conventional loading together with the extreme
action.
The key parameters identified here are not only based on the basic blast implications
observed on the cable anchorage zones, but also include the decisive criteria that traditionally
govern the ordinary behaviour of anchorage regions as well as the crucial factors that are
widely acknowledged to have absolute authority over the general blast response of civil
infrastructure.
While compatibility with the traditional codes is cautiously ensured, more importantly, the
integrated design standard, which consists of both innovative techniques together with
advanced logistics, is intended to be implemented sensibly in accordance with the
performance-based design principle which is fundamentally distinct from the prescriptive-
based design concept.
247
Chapter 6
6.2 PARAMETRIC STUDY
6.2.1 Importance of Standoff Distance
In order to determine the maximum allowable blast action whereby an ultimate failure will be
effectively avoided, various standoff distances were attempted for the largest PBIED and the
largest VBIED considered.
With 30 kg TNT, the smallest acceptable standoff distance was found to be approximately 1
m (i.e. 0.65 m farther away from the initial detonation point) (as indicated in Figure 6.1), with
a corresponding safe scaled distance of 0.322 m/kg1/3 (and therefore a 185% increase from
the original value of 0.113 m/kg1/3). The positive peak reflected pressure and impulse were
obtained accordingly as 67 MPa and 9.7 MPa.ms, hence a decrease of 78% and 83%
respectively (refer to Figure 6.2).
Figure 6.1: Minimum safe standoff distances for PBIEDs and VBIEDs
With 1000 kg TNT, the explosive must be at least 2.5 m from the cable (i.e. 0.5 m farther
from the original charge centre) (as shown in Figure 6.1). The standoff distance to the deck
248
Parametric Study and Blast-Resistant Design Guidelines
remained unchanged, and therefore the scaled distance maintained as 0.1 m/kg1/3. The
positive peak reflected pressure and impulse, at the centreline of the primary girder, were
measured accordingly as 24 MPa and 6.7 MPa.ms, hence a reduction of 41% and 46%
respectively (refer to Figure 6.3), owing to a rise in the angle of incidence.
Figure 6.2: Blast pressure histories (derived from the PBIED at several locations)
Figure 6.3: Blast pressure histories (derived from the VBIED at several locations)
249
Chapter 6
It was proven that, relocating the blast sources to slightly farther spots substantially relieved
the imposed burden. This benefitted the survivability of the target significantly, as aligned
with the expectation.
The anchorage zone, which sustained only minimal blast damage, with no sign of
compressive crushing aside from extremely limited concrete spalling, as evident in Figure 6.4
and Figure 6.5, remained functional albeit with less resistance preserved. The exposed region,
instead of the sheltered portion, continued to display blast outcomes of greater severity. The
blast impact observed on the edge beam also eased considerably, but surprisingly, widespread
overall damage was not obvious, as the safe standoff for the portable charge was still
insufficient to give rise to a far-field event which was, however, brought about by the vehicle
bomb.
The anchorage displacement settled down eventually, as indicated in Figure 6.6 and Figure
6.7. With the PBIED placed at the corresponding minimum standoff, the inclination of the
initial gradient was not at all comparable; as for the case with the VBIED, the overall
downward movement was obtained to be somewhat lower.
The cable force gradually stabilised, according to Figure 6.8 and Figure 6.9. The initial load
drop due to the PBIED was hardly noticeable; the overall load gain due to the VBIED was
reportedly smaller.
As shown in Figure 6.10 and Figure 6.11, the yield strength of the rebars was not exceeded
under the safe conditions. The initial peaks appeared lower, owing to the smaller blast
intensities, and emerged later, since the arrival times became longer.
As expected, the transverse stresses in the concrete, as given in Figure 6.12 and Figure 6.13
for example, still refused to reflect a consistent distribution pattern. With the direct blast
impact, the lateral stresses again resided purely in the compressive region; the deviation
shown under the indirect blast loading was undeniably not as profound.
250
Figure 6.4: Blast damage (produced from the desired blast threshold for PBIEDs)
Figure 6.5: Blast damage (produced from the desired blast threshold for VBIEDs)
251
Chapter 6

different scaled distances)
Figure 6.7: Anchorage displacement records following VBIED explosions (with distinct
charge centres)
252
Figure 6.8: Cable force histories on account of PBIED detonations (with various
standoff distances)
Figure 6.9: Cable force histories on account of VBIED explosions (with several
detonation points)
253
Chapter 6
Figure 6.10: General reinforcement stress data (with the smallest standoff distance
required for PBIEDs)
Figure 6.11: General reinforcement stress data (with the nearest charge centre
acceptable for VBIEDs)
254
Figure 6.12: Transverse stress distribution (with the lowest safe scaled distance for
PBIEDs)
Figure 6.13: Transverse stress distribution (with the closest detonation point allowable
for VBIEDs)
255
Chapter 6
6.2.2 Influence of Concrete Quality
It is a well-known fact that concrete grade governs the subject’s overall resilience. Apart
from the original 50 MPa, the uniaxial unconfined compressive strength attempted also
includes 70 MPa and 90 MPa, which can all be reasonably claimed to belong to the high
strength band29.
The inputs are listed in Table 6.1. However, it is worth highlighting that the amount of strain
rate enhancement will be compromised when the concrete becomes stronger (refer to Figure
6.14 and Figure 6.15).
Table 6.1: Mechanical properties of concrete

ρ ν E fc ft
Case
1 33 50 4.1
2 2400 0.2 40 70 5.1
3 45 90 6.0
Unfortunately, the anchorage region completely failed in most of the trial cases, as clearly
indicated in the cable force and anchorage displacement plots given in Figure 6.16 and Figure
6.17 respectively with the PBIED, and likewise with the VBIED except when the best grade
was adopted, as evident in Figure 6.18 and Figure 6.19.
When subjected to the small-scale burst, the blast endurance merely enhanced marginally
with a better concrete grade, since the blast damage turned vaguely less acute, according to
the initial decline of the axial stiffness of the target, apart from the delay of the crushing
failure.
29
i.e. > 40 MPa (ACI Committee 363, 1992)
256
In comparison, when the large-scale burst was concerned, with a more promising concrete
quality, the improvement, albeit self-explanatory, was, in fact, much more generous, and
eventually enabled the structure to survive.
Figure 6.14: Strain rate enhancement of compressive strength
Figure 6.15: Strain rate enhancement of tensile strength
257
Chapter 6
Figure 6.16: Anchorage displacement data under PBIED detonations (with different
concrete grades)
Figure 6.17: Cable force histories after PBIED detonations (with different concrete
grades)
258
Figure 6.18: Anchorage displacement data under VBIED explosions (with different
concrete grades)
Figure 6.19: Cable force histories after VBIED explosions (with different concrete
grades)
259
Chapter 6
6.2.3 Influence of Reinforcement Quantity
The various reinforcement plans implemented here are clarified in Table 6.2. Only the
general reinforcement was altered, as the spacing of the discrete ties was widened from 100
mm to 150 mm and finally 200 mm (and the number of bars inserted, therefore, shrunk
accordingly), while the diameter was maintained. The continuous spiral local reinforcement
remained untouched throughout. However, it is good be aware that the compact anchorage
precinct was highly crowded with steel bars, and room to accommodate additional changes
appeared to be strictly lacking.
Table 6.2: Reinforcement efforts

General reinforcement (N20) Local reinforcement (N16) Estimated
Case Spacing Number of bars Pitch Number of turns volume
(mm) (mm) (m3)
1 100 18 3.55 × 10-2
2 150 14 75 9 3.09 × 10-2
3 200 12 2.85 × 10-2
In all the trial settings, the anchorage region inevitably gave way, according to the blast
results extracted to Figure 6.20 and Figure 6.21 when the light charge was triggered, and
Figure 6.22 and Figure 6.23 when the heavy charge was activated. The harm inflicted on the
target intensified when the reinforcement effort, which not only provided basic hardening but
also offered extra confinement, was demoted, as suggested by the initiation phase of the
brittle crushing.
Even though the spacing was modified sequentially by merely 5 cm, the influence exerted on
the subject was, nonetheless, sharp and apparent, with greater sensitivity shown when the
vehicle bomb was used.
260
Interestingly however, it was noticed that downgrading the general reinforcement plan
undermined the local zone as well, especially when the portable charge was involved, owing
to both stress wave transmission and the manner of which the blast pressure was exerted.
Figure 6.20: Anchorage displacement data under PBIED detonations (with distinct
reinforcement ratios)
Figure 6.21: Cable force histories after PBIED detonations (with distinct reinforcement
ratios)
261
Chapter 6
Figure 6.22: Anchorage displacement data under VBIED explosions (with distinct
reinforcement ratios)
Figure 6.23: Cable force histories after VBIED explosions (with distinct reinforcement
ratios)
262
6.2.4 Significance of Bearing Plate Size
In order to establish the influence of bearing area as well as the anchorage dimension to
member depth or width ratio (commonly denoted as the a/h ratio) (refer to Figure 6.24 and
Figure 6.25), three sets of bearing plate dimensions, i.e. 0.7 m × 0.7 m, 0.98 m × 0.98 m and
1.12 m × 1.12 m, were selected for the anchor pod which measured 1.25 m × 1.25 m
(enlarged from 0.9 m × 0.9 m due to ease of demonstration).
Figure 6.24: a/h ratio
Figure 6.25: Effective, gross and net bearing surfaces
The various settings attempted here are explained in Table 6.3. The anchorage plate was first
expanded by 96% and then by another 59% (i.e. from 0.49 m2 to 0.96 m2 and eventually 1.25
m2), thereby giving rise to a total of three separate a/h figures and net contact areas.
263
Chapter 6
Table 6.3: Basic influence of bearing plate size

Bearing plate size Anchor pod size a/h ratio Net bearing area
Case
(m2) (m2) (m2)
1 0.7 × 0.7 0.56 0.36
2 0.98 × 0.98 1.25 × 1.25 0.78 0.83
3 1.12 × 1.12 0.90 1.13
It is well understood that the a/h ratio dictates the bursting action. Even though the anchorage
load was kept constant, with a smaller contact plate, more profound lateral stress dispersal
was evident, as plainly shown in Figure 6.26. The splitting force was found to ease
immensely, from 738 kN to 359 kN and then to 240 kN, hence a drop of approximately 67%
all together.
Figure 6.26: Bursting stress data (due to distinct a/h ratios)
The resulted bearing stress is inversely proportional to the net bearing area which is defined
as the gross bearing surface less the central opening. The expansion of the contact plate
264
prompted a much greater rise in the net bearing space but merely a relatively small drop in
the effective to gross bearing area ratio (i.e. the A/Ag ratio), and this led to a favourable gain
in the compressive resistance (around 96% in total), as clarified in Table 6.4.
Table 6.4: Bearing loads and capacities (as affected by the net contact area)
Bearing stress √(A/Ag) Bearing capacity*
Case
(MPa) (MN)
1 20 1.79 26
2 8.8 1.28 42
3 6.5 1.12 51
* Due to the concrete confinement contribution alone (refer to Appendix B)
When faced with the small-scale burst, although the target became non-operational in all the
trial setups, the blast performance indeed turned less disappointing when the contact surface
was enlarged, as shown in Figure 6.27 and Figure 6.28, and the most notable improvement
was observed when the reloading process managed to persist decisively after the consecutive
load drop.
On the other hand, when confronted with the large-scale burst, the influence of bearing area
was again indisputably confirmed. Except with the smallest plate, the subject continued to be
functional, according to Figure 6.29 and Figure 6.30. Also, it was noticed that, between the
two largest bearing areas, the outcome was merely slightly better with the use of the biggest
plate.
In short, the anchorage precinct was already in a deplorable state after the blast incident, but
the harm was further amplified by the sustained load effects which were dependent on the
bearing plate size.
265
Chapter 6
Figure 6.27: Anchorage displacement due to PBIED detonations (with various bearing
plate dimensions)
Figure 6.28: Cable force development following PBIED detonations (with various
bearing plate dimensions)
266
Figure 6.29: Anchorage displacement due to VBIED explosions (with various bearing
plate dimensions)
Figure 6.30: Cable force development following VBIED explosions (with various
bearing plate dimensions)
267
Chapter 6
6.2.5 Significance of Anchorage Zone Size
In order to study the influence of increasing mass, the anchor pod was enlarged by
transforming the cross section from 0.9 m × 0.9 m (0.81 m2) to 1.25 m × 1.25 m (1.56 m2), as
can be seen in Figure 6.31, thereby raising the volume by 1.15 m3 (i.e. 192%), from 0.6 m3 to
1.75 m3.
Figure 6.31: Expansion of the anchor pod
Unfortunately, it was difficult, if not impossible, to determine the contribution of the size
alone, by considering it as the sole variable, owing to the fact that the ordinary induced
actions will also vary along, as clarified later on. Thus, the contact plate was also adjusted
accordingly, from 0.7 m × 0.7 m to 0.98 m × 0.98 m (i.e. 0.49 m2 together with 0.96 m2),
albeit involuntarily.
Increasing the size of the anchor pod also implied moving closer to the light charge
suspended from above (by 0.175 m in this case), as demonstrated in Figure 6.32, and thus
facing a more extreme blast magnitude, with a rise of 272 MPa (about 90%) and 167 MPa.ms
(equivalent to 298%) for the positive peak reflected pressure and impulse respectively
(according to Figure 6.33), by halving the scaled distance. Such concern was irrelevant when
the heavy bomb positioned on the deck was of interest, as evident in Figure 6.32.
268
Figure 6.32: Correlation of blast intensity with anchor pod dimension
Figure 6.33: Blast pressure histories (on distinct anchor pod sections)
The blast cases considered here are described in Table 6.5. The anchor pod size together with
the steel plate area (selected in a way so as to give two separate a/h values and net bearing
areas, keeping one constant while altering the other), which stood out as the prime issues for
the VBIED blast setting, were also varied along with the standoff distance when the PBIED
was involved.
269
Chapter 6
Table 6.5: Blast investigation of the effect of anchor pod size

Anchor pod Bearing plate a/h ratio Net contact Standoff distance
Case (m2) (m2) (m2) (m)
PBIED VBIED
1 0.9 × 0.9 0.78 0.35
0.7 × 0.7 0.36
2 0.56 1.00
1.25 × 1.25 0.175*
3 0.98 × 0.98 0.78 0.83
* To be replaced also with 0.35 m for comparison purpose
As can be seen in Figure 6.34, for the enlarged subject, the transverse stress distribution
spread across a much longer distance, without surprise: the splitting force rose by roughly
122% (i.e. from 333 kN to 738 kN) when the net contact space (hence the bearing pressure)
was kept unchanged; on the other hand, with an identical a/h value, the bursting action was
found to be rather consistent, but the bearing stress was reportedly brought down by 56% (i.e.
from 20 MPa to 8.8 MPa).
Figure 6.34: Bursting stress data (for different anchor pod sizes)
270
Although unable to survive the direct blast impact, based on Figure 6.35 and Figure 6.36, the
expanded section exhibited blast endurance with greater superiority when accompanied with
a decrease in the bearing stress, even though trialled under a worse off condition (which gave
rise to a more dramatic load loss at the beginning); but proved to be ineffective when the
extra resistance was disastrously overwhelmed by an increase in the splitting action, even if
the scaled distance remained unaltered.
As anticipated, the significance of mass was much more apparent when the indirect blast
loading was concerned, as unreservedly suggested in Figure 6.37 as well as Figure 6.38. The
enlarged version not only managed to stage a better performance, albeit to a somewhat
marginal extent, even after the increase of the bursting force, but also effectively exhibited
excellent competence in avoiding tragic failure, as anticipated nonetheless, when the bearing
pressure was eased.
In fact, the rationale behind the enhanced robustness due to additional mass is none other than
the extra matter, which substantially supplemented the absorption of blast energy, and the
greater residual capacity preserved. The rigidity was also benefited, thereby limiting the
induced distortion.
271
Chapter 6
Figure 6.35: Anchorage displacement due to PBIED detonations (with several anchor
pod sections)
Figure 6.36: Cable force development following PBIED detonations (with several
anchor pod sections)
272
Figure 6.37: Anchorage displacement due to VBIED explosions (with several anchor
pod sections)
Figure 6.38: Cable force development following VBIED explosions (with several anchor
pod sections)
273
Chapter 6
6.2.6 Effect of Cable Force
The consequence of altering the cable action, hence the anchorage force, was assessed. While
it was unrealistic to vary the self-weight, only the existing traffic load was set to halve
progressively, i.e. from the initial 100% to 50% and finally to 0%.
The load information is given in Table 6.6.
Table 6.6: Gravitational actions and cable loads

Traffic load Self-weight Total cable force
Case
(kN/m) (kN/m) (kN)
1 184 7300
2 92 488 6300
3 0 5300
By varying the live load between 38% to 0% of the dead load, the internal action within the
stay eased incrementally by roughly 1000 kN, hence close to 14% decrease in the bearing
stress (equivalent to 2.8 MPa). Additionally, the bursting force also reduced accordingly,
from 333 kN to 244 kN, but the stress distribution remained consistent, as shown in Figure
6.39.
Here, a full cycle which fluctuated from rush hour to off peak, which is common in practice,
was assumed, and this also justified the decision to disallow the maximum burden to be
exceeded instead.
274
Figure 6.39: Bursting stress data (due to several anchorage loads)
In all the trial scenarios, the anchorage region was totally demolished, regrettably, as proven
in the blast results displayed in Figure 6.40 and Figure 6.41 when the portable charge was
triggered, and likewise in Figure 6.42 and Figure 6.43 when confronted with the vehicle
bomb.
Judging from the manner of the progression (based on the ascending/descending gradients
which reflect the relative hostility), the aggressiveness of the crushing mode was found to
heighten with the PBIED, but the effect of the coexistence of the anchorage action was
slightly more moderate with the VBIED.
Without surprise, the structure took longer to be disintegrated when the sustained loading was
smaller. The reason for the observation made is rather straightforward: the resistance that
remained, which was insufficient to even withstand the self-weight alone, was overwhelmed
much quicker when the extra burden grew bigger.
275
Chapter 6
different cable loads)
Figure 6.41: Cable force records on account of PBIED detonations (with different cable
loads)
276
different cable loads)
Figure 6.43: Cable force records on account of VBIED explosions (with different cable
loads)
277
Chapter 6
6.2.7 Influence of Cable Inclination
In cable-stayed bridges, the cables are installed in various orientations, when either a pure or
partial fan arrangement is implemented. For this reason, it is crucial to assertively affirm the
authority of support inclination as a control parameter. In order to serve this intention, two
separate configurations arranged in 27.36° and 47.60°, as displayed in Figure 6.44, were
selected.
(a)
(b)
Figure 6.44: Anchorage precincts configured in unique orientations – (a) 27.36°; (b)
47.60°
278
While the vertical component was maintained, lowering the angle instigated a rise in the
horizontal component, thereby raising the reaction force all together, by an amount of 2700
kN in this case with a drop of 20.24° (as clarified in Figure 6.45). A gain of 59% in the total
force not only prompted the bearing pressure to rise by 7.4 MPa, but also escalated the
bursting force from 209 kN to 333 kN (refer to Figure 6.46).
Figure 6.45: Support inclinations and the corresponding cable loads
Figure 6.46: Bursting stress data (due to several support inclinations)
The two settings involved here are presented in Table 6.7. The adoption of a constant bearing
surface of 0.9 m × 0.9 m meant that the length of the anchor pod was to be brought down
279
Chapter 6
from 1.74 m to 0.82 m when the inclination grew, and the volume, thus, shrunk accordingly
by approximately 53% (from 0.6 m3 to 0.28 m3). The reinforcement plans assigned also
differed: the inner spiral was shortened by 2 full turns and the pitch was widened by 25 mm,
whereas 7 sets of the discrete ties were deducted from the outer frame.
Table 6.7: Intricate effect of cable orientation

Case 1 2
Cable inclination (°) 27.36 47.60
Anchorage force (kN) 7300 4600
Anchor pod Area (m2) 0.9 × 0.9
dimensions Length (m) 1.74 0.82
Reinforcement General 14 bars 7 bars
details Local 9 turns @ 75 mm 7 turns @ 100 mm
It is absolutely vital to place a firm emphasis on the fact that the support angle was
responsible solely for the load variation shown within the stay, while the design adjustment,
albeit not mandatory, should be viewed merely as a sensible response (true in reality). Strictly
speaking, the support orientation was not at all a stand-alone variable in this respect; rather, it
was just a coming together of the imposed action as well as the inherent capacity of the
target.
The blast outcomes displayed in Figure 6.47 and Figure 6.48 are accompanied by the
response contours highlighted in Figure 6.49 and Figure 6.50. Regrettably, neither setup was
formidable enough to withstand the small-extent burst or the large-scale burst, as reported in
Figure 6.51 and Figure 6.52 as well as Figure 6.53 and Figure 6.54.
Notwithstanding the less acute blast damage incurred (as suggested by the degree of the
material breakdown), the anchorage zone, when constructed with a smaller inclination, was
noticed to fail in a more aggressive way, at a much earlier stage of the blast event.
280
The correct explanation for the observation is: for the steeper anchorage region, the blast
endurance was indubitably more inferior, owing to its smaller size and less reinforcement
provision; however, such disadvantage was substantially offset by the remarkable benefit
offered by the much lower reaction force.
At 0.045 s At 0.057 s
(θ = 27.36°) (θ = 47.60°)
Figure 6.47: Blast response of the cable anchorage zone under PBIED detonations (in
various orientations)
At 0.210 s At 0.425 s
(θ = 27.36°) (θ = 47.60°)
Figure 6.48: Blast response of the cable anchorage zone under VBIED explosions (in
various orientations)
281
Chapter 6
At 0.045 s At 0.057 s
(θ = 27.36°) (θ = 47.60°)
Figure 6.49: Damage distribution on the anchorage precinct due to PBIED detonations
(in various orientations)
At 0.210 s At 0.425 s
(θ = 27.36°) (θ = 47.60°)
Figure 6.50: Damage distribution on the anchorage precinct due to VBIED explosions
(in various orientations)
To hold a further enquiry, assuming similar resistance, a more superior performance will, no
doubt, be exhibited when handling a smaller cable load. On the other hand, supposing that the
same cable force is experienced, a much better outcome will be expected with a more
generous blast capacity.
282
Most importantly, the crucial understanding acquired here was actually intended none other
than to provide a profound insight regarding the blast vulnerability of all the anchorage points
on the entire bridge, or, in other words, how they might perform relative to each other.
distinct support configurations)
Figure 6.52: Cable force records on account of PBIED detonations (with distinct
support configurations)
283
Chapter 6
distinct support configurations)
Figure 6.54: Cable force records on account of VBIED explosions (with distinct support
configurations)
284
6.3 GENERAL DISCUSSION OF DESIGN VARIABLES
The key parameters selected for in-depth investigation, if viewed in terms of functionality,
may be collectively classified into two distinct categories: the decisive criteria that affect the
extreme and regular loading, as well as the dominant features that establish the intrinsic blast
and conventional resistance.
It was assertively affirmed that relief of blast impact serves the utmost significance, and thus
must be assigned foremost priority; but altering the standoff distance is more attractive than
controlling the charge weight, as suggested by the concept of cube-root scaling.
Adjusting the bearing space is preferred over manipulating the gravitational load, since a
huge change in the latter is equivalent to merely a slight difference in the former, but varying
the support orientation should only come into play last, despite the fact that all these options
virtually assume the identical role of governing the ordinary load effects exerted on the
anchorage precincts.
Although all unequivocally dictate the robustness of the anchorage regions, the following
activities shall be ranked in such order: size transformation, which affords the most superior
contribution, shall be designated the highest authority, while modification of reinforcement
arrangement should be given greater attention in comparison to variation of concrete
properties.
The evaluation of the relative authority is valid only in the perspective of effectiveness, and
merely serves to deliver a basic picture of the interrelationships between the principal
concepts. The common rule derived should be treated purely as a baseline which is to be
adapted flexibly to suit each specific situation bound up with unique practical and technical
constraints.
285
Chapter 6
6.4 BLAST DESIGN SPECIFICATIONS
6.4.1 General Concepts
6.4.1.1 Design Aim
In blast design, the chief objective is none other than to ensure that the anchorage precinct
possesses sufficient capacity to sustain adverse blast effects, while ultimately avoiding total
failure. A meticulous damage assessment must ensue immediately after a blast event so that
the most effective remedies can be identified promptly.
An anchorage point is said to have been completely dismissed if it is found to be no longer

operational (i.e. unable to participate in the intended load transmission), and the presence of
the stay is, therefore, rendered meaningless.
6.4.1.2 Definition of Anchorage Zone
An anchorage zone is a compact locality where a concentrated anchorage action disperses

characteristically (often in a very complicated manner) until a constant stress transmission is
eventually attained.
In this specific context, the anchorage region commences at the anchor pod and spreads well
into the primary beam until the load transfer is diverted laterally and carried over across the
deck-and-girder system.
6.4.1.3 Responsibility Allocation
It is commonly agreed that the anchorage manufacturer is held accountable for the integrity
of the local zone, whereas the general zone falls within the jurisdiction of the practising
engineer.
286
However, unreserved collaboration amongst the parties involved, together with the relevant
experts, is warranted if blast effects are to be factored in. Moreover, attempt to subdivide the
anchorage zone does not always hold when dealing with awkward blast situations, and a
more holistic thought is, therefore, very much preferred.
6.4.1.4 Limit States
Compressive bearing crushing, along with tensile bursting cracking and surface spalling, are
the conventional issues; the associated design principles are discussed in Appendix B30.
When blast impact is concerned, an additional failure style, i.e. blast-induced material
breakdown, must also be taken into account.
6.4.1.5 Adequacy of Conventional Design Codes
The traditional design procedures proved greater significance for explosions initiated above
the span, but strictly unreliable for detonations triggered at close proximity to the target.
Nevertheless, until essential revisions have been successfully finalised, the existing design
codes are not applicable to blast circumstances at all. Apart from addressing the possible
strategies to convincingly incorporate blast action, introduction of novel ideas, along with
proper modification of the current practices, are also imperative.
6.4.2 Crucial Advice on Blast Analysis
Judging from the impulsive nature of blast loading, an equivalent static approach is deemed
inappropriate, and only a dynamic procedure implemented satisfactorily in an uncoupled or a
coupled manner, depending on the specific issue encountered, is to be adhered to.
30
By referring to AS 5100.5-2004, AS 3600-2009, AASHTO (2010) and ACI 318-11
287
Chapter 6
Sophisticated FEA appears as a legitimate option to genuinely derive the blast implications
on an anchorage zone. The conventional analytical methods, i.e. the approximate expressions,
the strut-and-tie concept and the elastic FEM, are unequivocally known to be incompetent to
confidently incorporate blast pressure and generate credible blast consequences.
For blast action, aside from ensuring that accurate wave profiles are obtained from the unique
load distribution, the outcome due to wave propagation should not be disregarded all the
time, even though fully accounting for it is a truly challenging task.
The element choices and the mesh resolutions should be decided cautiously, catering for both
the desired precision whilst not entailing huge computational effort. The most ideal numerical
formulations for the materials of concern must be utilised (to be newly developed if not
already available), and excellent knowledge of the real behaviour under various loading
regimes is indispensable. Special tactics to establish the actual contact and connection
systems often call for various simplifications. Adequate provision of structural contents
should not be neglected, and the boundary concepts to be enforced, which are situational,
must not corrupt the result.
While explicit representation of the concrete is inevitable, smeared modelling of the

reinforcing steel is intolerable. The existence of the central opening shall not be omitted
purely for the sake of convenience, and the presence of the stay together with the anchorage
assembly is of paramount importance.
In order to paint the whole picture of the exact outcome, a sensible quantitative assessment,
which relies on anchorage displacement and cable force as the key performance indicators,
should be complemented by a judicious qualitative appraisal based on overall distortion,
damage contour and matter disengagement.
288
6.4.3 Advanced Design Strategies
6.4.3.1 Blast Load
Dedicated efforts to relieve the blast intensity, hence the unwanted yet unnecessary burden,
by raising the standoff distance and/or limiting the possible sizes of the blast source, shall be
firmly emphasised. Unfortunately, the exact charge weight that will take part in a blast
incident is unpredictable and beyond design control.
Installing outstretched barriers, which extend beyond the edge of the primary girder, fully
catering for the entire anchor pod, as demonstrated in Figure 6.55, is a direct approach in
response to the positioning of PBIEDs. The platforms can be perceived as an architectural
feature which contributes to the overall aesthetic, but must be made of tough matter
competent to withstand particularly huge impact which might take place in advance before
the actual event. The connection with the main structure should also be carefully configured
so as to defend the intended purpose. In order to counter vandalism, the panels can be
lowered, hence denying unauthorised access.
Furthermore, widening the edge beam can be attempted, but decreasing the width of the
anchor pod is not to be encouraged at all. Adapting a circular shape capable of attracting less
blast load might appear impractical for the anchor pod (i.e. the exposed target).
Under permissible circumstances, auxiliary features (e.g. fences and railings) that facilitate
the suspension of PBIEDs should always be avoided. Nevertheless, it is worth noting that the
charge could still be rather conveniently secured through the stay, but the potential proximity
to the anchorage point is dependent on the orientation (i.e. more critical with a larger
inclination), as clearly shown in Figure 6.56, and for this reason, the harp arrangement has an
obvious advantage over the fan configuration in this respect.
289
Chapter 6
(a)
(b)
Figure 6.55: Architectural platform cantilevered from the longitudinal girder – (a)
Cross section view; (b) Plan view
Figure 6.56: Potential blast locations depending on cable orientations
290
As for VBIEDs, blast-resistant walls can be erected (refer to Figure 6.57), in order to achieve
the intended inaccessible zone which is, however, restricted, in turn, by the service space
required. Additionally, sacrificial covers or protective layers can also be provided, as a means
to elevate the explosive (as shown in Figure 6.57), but it might not be appropriate to treat this
approach as the sole solution, depending on the material used, and the company of other
complementary measures are often called for.
Figure 6.57: Defensive barrier and sacrificial protection
Broadly speaking, VBIEDs are normally stationed above the deck, but the need to consider
the blast action generated beneath the girder must not be neglected, as bridge clearance, aside
from blast magnitude, must also be factored in. Although PBIEDs should be assumed to be
placed as close as possible to the anchorage precinct, the implications due to their presence
on the deck should not be taken lightly for all circumstances. Both small-scale detonations
and large-extent bursts shall be taken into account, and this is particularly crucial when
distinct blast consequences are anticipated.
Blast loading must be set to act concurrently along with all other conventional actions. Given
its extreme nature, a load factor of unity is commonly adopted (e.g. AS 5100.2-2004), but
proper adjustment, in the light of risk consideration, shall, nevertheless, not be ruled out
entirely.
291
Chapter 6
6.4.3.2 Anchorage Load
The bursting force exerted on the anchorage precinct should be as weak as possible, by
maximising the a/h ratio in a reasonable manner. This shall be achieved only by choosing a
larger bearing plate, hence a wider gross bearing surface (as shown in Figure 6.58), while the
cross section of the anchorage zone, which contributes to the effective bearing space, should
not be allowed to shrink.
In order to ensure that the bearing stress remains low, the net bearing space shall be enlarged
sufficiently. Aside from broadening the contact plate, the hollow core can also be formed
smaller if practical, as demonstrated in Figure 6.58.
(a) (b)
Figure 6.58: Relief of anchorage load effects – (a) Normal design; (b) Blast design
It is worth highlighting that the bearing plate, which is an integral feature of the anchorage
assembly, is usually provided by the manufacturer.
Under blast loading, crushing failure shall be the main concern, instead of splitting cracking
and edge spalling. The anchorage zones of interest (in cable-stayed bridges) are intended
principally for load support, while direct stress transfer between the structure and the entire
cable section passing through the circular duct is irrelevant here.
292
It should be clarified that the intent is not to influence the actual burden, although the security
level, as promised in the daily traffic management programme, must be heightened during
peak rush where the most disastrous scenarios are anticipated.
The aim shall rather be diverted to substantially alleviate the associated effects produced in
the orthogonal transverse directions. The relief is able to not only restrict the degree by which
the situation is worsened once the condition of the subject deteriorates, but also delay the
exhaustion of the resistance that remained.
Load changes likely to happen must be taken into account, especially for the anchorage point
accommodated at the pylon. Usually, the effect of portable charges on the anchorage force is
rather negligible, but the blast action exerted by vehicle bombs located on the deck will act
together with the gravitational load; downward pulling of the girder will also continuously
strain the stay.
Raising the inclination of the support eases the horizontal action, and hence the resultant
force. As can be seen in Figure 6.59, with a constant vertical component, bringing the stay to
the normal position from an angle of 30° potentially halved the reaction force. However,
albeit theoretically sound, any attempt to draw benefit from this concept in this particular
context is, nonetheless, not viable.
Figure 6.59: Correlation of resultant force with support orientation
293
Chapter 6
6.4.3.3 Material Properties
Under regular action, compressive strength is vital for bearing support, but tensile strength is
conservatively ignored when dealing with spalling and cracking; under extreme loading,
concrete strength plays a pivotal role against the blast-resulted material failure that seriously
weakens the subject.
Ductility, which is defined by the plastic region, governs the dissipation of blast energy, and
should not be overlooked. Apart from providing extra confinement, high performance
concrete may also be utilised for new structures. Since strength is typically inversely
correlated to ductility, as exhibited from the stress-strain relationships displayed in Figure
6.60, striking the most perfect balance should be strongly encouraged.
Figure 6.60: Stress-strain curves of concrete
In practice, the protruding anchor pod is casted together with the edge beam, thereby sharing
identical material attributes. In fact, the anchorage precincts cover only a fairly minor extent
294
of the suspended spans as well as the erected towers, and, in the light of economy, do not
have dominant command to dictate the selection of the preferable concrete grade. In this case,
reserving the heavy duty of blast hardening for other design variables might appear
reasonable.
Strength enhancement for both compressive and tensile properties due to rapid loading should
always be satisfactorily incorporated within.
It must be emphasised that, in this given situation, the use of stronger concrete shall not at all
be taken as a solid excuse to disadvantageously scale down the structure.
Following concrete failure, the whole responsibility will be automatically transferred to the
reinforcing steel. In an extreme situation, controlled yielding, which is absolutely intolerable
in normal practice, may be permitted (or perhaps warranted), but steel fracture is strictly
prohibited, with the effect of strain rate fully factored in.
6.4.3.4 Detailing
The local zone reinforcing steel, which contributes immensely to the bearing resistance,
should be carefully specified. Continuous spirals, which enjoy greater popularity, are
commonly adopted, but discrete ties could also be included (PTI, 2006). The special demand
of a unique anchorage device must be fulfilled.
In the general zone, the bursting reinforcement typically consists of individual ties which are
spread uniformly (but arranged closer near the resultant force) in accordance with the
characteristic flow of action. The spalling reinforcement shall be placed as close as
reasonable to the free edges whilst not violating the cover provision. Adequate elongation
into the adjoining body must be strictly enforced.
It is vital to realise that capacity reduction during a blast incident is by no means evitable, and
extra allowance is almost certainly indispensable. Furthermore, although the unique roles are
295
Chapter 6
outlined explicitly when traditional loading comes into play, both the general and local
reinforcement ought to be perceived as a whole in offering a joint effort against
unconventional blast impact.
Augmenting the reinforcement quantity, by inserting additional rebars and narrowing the
pitch and the spacing, leads to more superior robustness through basic hardening and
auxiliary restraint. However, further complicating the anchorage precinct, which is most
likely already heavily congested with reinforcing bars (hence the lack of room for
manoeuvre), is often highly unfavourable especially when viewed in the perspective of
constructability, and for this reason, adherence to the normal provision might be enforced,
while the help of other alternatives are sought.
6.4.3.5 Dimensioning
Mass is widely acknowledged as a vital factor that governs blast endurance, as it affects the
overall capacity, the ability to absorb energy as well as the structural rigidity.
It is favourable to configure an anchorage zone with a larger volume, by adopting an anchor

pod of a greater size (subject to the approval of the edge beam) and minimising the diameter
of the circular passage (whilst satisfying the opening space required). While lengthening the
perpendicular sides of the subject appears as an obvious option, increasing the width or the
depth (along with the length) alone, as illustrated in Figure 6.61, is also possible, but the
actual effectiveness of any preferred decision must be evaluated separately for each
individual case.
Extra caution is essential, however, if the proximity to the blast source is inevitably
improved, as intensifying the blast load will, in turn, offset a certain degree of the total
advantage brought by the additional mass. Thus, an optimum solution is called for to reach
the favoured balance.
296
Figure 6.61: Anchorage zone enlargement – (a) Original unit; (b) Orthogonal extension;
(c) Width broadening; (d) Depth elongation
Expansion of the anchor pod shall not lead to an intolerable a/h value which gives rise to a
huge increase in the internal action. This can be conveniently remedied by appropriately
adjusting the gross bearing space.
Nevertheless, it is worth highlighting that size is not explicitly embraced as a crucial

parameter in the normal design of anchorage regions, probably because an anchorage zone
(except for external anchorages) is usually fully incorporated within, hence completely
constricted by, the principal structure.
6.4.3.6 Composite Enhancement
Rather than coming into play only at a much later stage, fortification schemes might be
unreservedly exploited during the initial establishment, hence the early implementation of
composite design.
The far-reaching advantages that come along include supplementary restraining contribution,
effective control of matter disengagement and a bonus energy-dissipating mechanism.
297
Chapter 6
Most importantly, the composite solutions can be conveniently and flexibly treated as
feasible alternatives in controversial circumstances where other core variables are hard to be
manipulated with ease.
The strengthening matter may be equipped at the external surrounding and/or the internal free
surface, as illustrated in Figure 6.62.
Figure 6.62: Composite designs – (a) Typical design; (b) Internal lining; (c) External
casing; (d) Combined solution
6.4.4 Integrated Design Methodology
The advanced methodology illustrated in Figure 6.63 addresses the need of both current and
future cable-stayed bridges. The unique relationship between the separate logistics has been
established and incorporated within a single framework.
298
Figure 6.63: Blast design methodology for cable anchorage zones
Apart from the effectiveness as well as the prerequisite of the individual options, service
interruption (which is a social and economic issue) is also a vital concept that contributed
heavily to the establishment of the integrated logistics.
For a new structure, upon finalising the gravitational load together with the support
inclination, as well as the tentative blast magnitudes, the systematic design procedure to be
closely adhered to is explained below:
299
Chapter 6
i. determination of viable tactics to minimise the blast loads of the corresponding

worst case scenarios
ii. selection of a proper size for the anchor pod (hence also the volume contained
within the edge beam)
iii. identification of a desirable a/h ratio together with a favourable net contact area
(hence the adoption of a suitable bearing plate)
iv. specification of local and general reinforcement details (i.e. type, grade, diameter,
spacing and arrangement)
v. prescription of concrete attributes (usually consistent with the requirement of the
suspended span)
vi. consideration of composite configurations (entirely situational and not at all
indispensable)
After the construction stage, constant assurance of the adequacy of the design to continuously
fulfil the demand of the inconsistent blast threat during the operation phase must be strictly
enforced. Once additional fortification is found necessary, the following mitigation process,
which is also intended for the existing bridges erected without factoring in blast implications,
will prevail:
i. implementation of a flexible, complex and multifaceted security plan

ii. delivery of strategic measures to further improve the potential standoff
iii. deployment of feasible strengthening tactics
If modification work is preferred or deemed compulsory, the possibility of reconfiguration

can also be taken into consideration.
The performance levels that are of relevance here include: fully operational, minor damage,
major damage and total failure. Multiple iterations are entailed for the design to accomplish
the desired mission, commencing with the traditional practices which can be treated as a
baseline.
300
6.5 SUMMARY
The integrated guidelines developed in this study are compatible with the routine practices,
thereby satisfactorily addressing both the new issues as well as the old problems. It was
affirmed that, even though not appropriate for blast occasions, the conventional codes will
still demonstrate profound significance when indirect, rather than direct, blast loading is dealt
with, and this provides an intuitive perception of the survivability of the anchorage regions
configured without commissioning blast countermeasures.
Blast investigations are to be conducted through sophisticated FEA, replacing the common
channels. The basic instructions given focussed on two principal aspects: credible load
calculation and convincing response determination. The paramount indicators include but
certainly not restricted to: actual deformation and overall damage display, to assist with
qualitative judgement; as well as anchorage movement and cable load, for quantitative
valuation.
While distinct protection categories may be defined based on the extent of the blast outcome,
the final design shall, nonetheless, aspire to avoid complete destruction of the anchorage
precinct which comprises the anchor pod together with the assimilated volume, and it is good
to be aware that the idea of distinguishable zones is totally invalid under blast impact.
As blast threat must be regularly updated throughout the service life, the performance-based
design methodology is instituted to cater for the need of not only new facilities but also
existing assets.
By relying on the in-depth parametric investigation as the underlying foundation, innovative

tactics have been introduced in order that the abstract concepts conceived are able to be
implemented in a viable manner.
In order to improve the blast performance, the ordinary load effects must be substantially
eased (as this restricts the additional damage induced once the structure has been drastically
301
Chapter 6
weakened, and at the same time, defers the residual capacity from being overcome), but relief
of blast load (proven to have the most powerful authority) shall be viewed as the top priority.
Managing the standoff distance is preferable to dealing with the explosive size; and likewise
with the contact plate and the actual force which is jointly governed by both the gravitational
load as well as the cable orientation.
Although all lead to a gain in the blast endurance, enlargement of anchorage zone (which
expands the energy-absorbing capability, increases the overall resilience and gives extra
stiffness at the same time) deserves the greatest attention, while augmentation of steel
quantity (which extends the existing reinforcement and confinement input) is noticeably a
better decision as compared to upgrade of concrete quality (which can benefit the strength
and the ductility).
The critical contrast of the effectiveness of the tested variables is particularly useful to give a
priori information, but the order of importance should be re-established in each specific
occasion in the light of further constraints.
In a cable-stayed bridge styled in the old-fashioned way, the anchorage points associated with
steeper support inclinations and/or smaller imposed burden are typically furnished with less
formidable intrinsic contents, and therefore prone to suffer worse off blast damage.
Nonetheless, having less blast robustness does not deem more aggressive progression of the
final destruction absolutely compulsory, since the reaction forces of interest are also
correspondingly lower. Technically, this answered the question of how the anchorage zones
will behave relative to each other in an identical blast scenario.
302
Chapter 7 SUMMARY, CONCLUSIONS, RECOMMENDATIONS
7.1 PREAMBLE
All the objectives in this project have been satisfactorily accomplished. The major and minor
goals collectively fall into the following principal categories: implementation of a research
plan; determination of local blast consequences, confirmation of suitable fortification tactics
and delivery of sensible design solutions; examination of global collapse mechanisms as well
as formulation of relevant mitigation concepts.
Within this chapter, a concise summary of the research programme, together with its
execution, is delivered along with the key findings obtained. While the unique interpretations
offered to explain and justify the observations are outlined, the exclusive insights gained are
also revealed, and the product as a result, which consists of all the conclusions and
recommendations made, are succinctly recapitulated. Furthermore, a broad discussion of the
specific areas in which additional work is deemed compulsory in future is also included.
The information gathered, the knowledge established as well as the technology developed,
purely by theoretical and analytical means, are not only at the leading edge of this
engineering discipline, but also the most fundamental and essential to (hence the introductory
foundation for) this particular topic of interest.
303
Chapter 7
7.2 INVESTIGATION PROGRAMME
7.2.1 General Details
A typical cable-stayed bridge for normal-span crossing was created. The concrete deck-and-
girder system was carried by double-plane steel cables (installed in a fan arrangement)
terminated at the concrete pylons.
Supplied as parallel strands, the main tension elements were adapted as circular bars,
perfectly secured within the anchorage devices perceived as cylindrical blocks, each with an
integral bearing plate (and this implied that load transmission via the anchorage regions took
place in the form of direct support), while all other ancillary equipment was omitted from
consideration.
The self-weight together with the traffic load (only on the major span or including the minor
spans) were set to reach the serviceability limit state. Prestress forces were also assigned to
the individual cables.
In order to establish the worst case scenarios: PBIEDs, appeared as light packages or
equivalent, were suspended from above; on the other hand, VBIEDs, identified specifically as
either compact or standard automobiles or heavy trucks, were positioned on the deck. Large-
extent bursts and small-scale bursts were treated separately.
A sophisticated FEA was implemented based predominantly on explicit time integration. The
uncoupled method was involved, and the MDOF system resided purely in the Lagrangian
domain. Both precision as well as cost were carefully factored in, aiming to strike the most
desirable balance whilst adequately satisfying the whole range of research purposes. The
blast simulation and the collapse analysis were dealt with through two distinct but parallel
procedures.
304
Summary, Conclusions, Recommendations
7.2.2 Local Blast Simulation
The local model, extracted prudently from the tributary region for a specific stay, containing
all the constitutive features except the deck and the cross girders, was intended purely for the
blast investigation, and was strictly prepared with meticulous effort:
 while solid elements were indispensable, beams and shells were also incorporated
 the constitutive formulations adopted for concrete, steel and CFRP all included
the associated strain rate behaviours, and were equipped with erosion features
invoked to account for material failure
 aside from the constraint algorithm, node sharing was another way used to create
a perfect and simplistic connection; as with the contact approach and node tying,
failure criteria were prescribed
 nodal constraints were assigned to the reaction support and the interfaces where
the adjoining members met; the internal action induced along the major axis was
introduced using an external force
The gravitational load, initiated in a seemingly implicit manner, was sustained throughout the
blast phase (in which the impulsive blast loading was derived from a collection of empirical
expressions) managed with the aid of an explicit routine.
7.2.3 Global Collapse Analysis
The global model was created in a much less elaborate manner, affording only the optimum
level of details essential to the collapse analysis:
 only beam and shell elements were engaged

 the reinforcing steel was smeared within the concrete; only tensile action was
permitted in the cables
305
Chapter 7
 without properly presenting the anchorage zones, the stays were linked straight to
the primary girders and the pylons
 the foundation and pier system was substituted with translational and rotational
constraints
An iterative algorithm, programmed with the aid of the restart function, was instituted for the
non-linear dynamic analysis (preceded by the static stage), where the initial harm due to a
given blast incident was introduced by dismissing the anchorage points assumed to be no
longer operational, and the stays that gave way as a consequence of cumulative burden were
subsequently withdrawn from the simulation which was terminated either once equilibrium
was finally restored or when ultimate destruction was confirmed to be inevitable.
The collapse analysis was conducted merely to highlight the severity of anchorage point
removal, while a more detailed global investigation was actually covered by the parallel
project undertaken in the University of Technology, Sydney.
7.3 BASIC FINDINGS
7.3.1 Local Blast Response
The cable anchorage zone remained serviceable after the blast incidents involving a standard
and a compact automobile, amazingly, but failed to survive the blast events with light
packages and a heavy truck, regrettably.
The blast implications observed include:
 on the anchorage zone positioned on the suspended span –

o the blast impact, which was generally less acute under the large-scale
bursts, encompassed the entire region considered (can be explained based
306
on the idea of relative size), but the anchor pod suffered more, in contrast to
the assimilated volume
o the front section incurred heavier damage with a detonation released from
the right, and likewise for the rear section with an explosion activated from
the left
o matter disengagement at the interior was more intense than that faced by the
exterior
o the strictly non-uniform blast damage induced an extremely disturbed stress
state (the stress wave transmission was much more chaotic under the direct
blast impact), leading to a completely eccentric stress distribution which
refused to comply to a universal rule (practical comparison with the normal
stress distribution was, therefore, impossible)
o although more noticeable under the lateral blast pressure, the instantaneous
blast-induced distortion, was, after all, not dramatic at all
o after seriously undermined by the blast damage, brittle compressive
crushing was the ultimate failure style resulted, under the sustained
gravitational loading
o overall, greater reserve of residual capacity was permitted when subjected
to the indirect blast action
 on the anchorage zone configured in the pylon –
o the blast action conveyed via the stay appeared virtually in the form of axial
loading
o the consequence was essentially negligible, since bursting cracking and
bearing crushing were not to be anticipated
o the intended design capacity was far from being overcome
 on the longitudinal girder –
o the portable charges inflicted more focussed effect in contrast to the
vehicular explosives which led to more extensive damage, but both blast
locations considered were technically close-in in nature
o in contrast to the bottom part, the top portion (appeared in a tensile state)
was weaker against the blast impact
307
Chapter 7
o interestingly, severe demolition which led to total deactivation of the

anchorage point did not occur
o a bomb placed above caused downward displacement, whereas a charge
originated from below also led to a brief upward translation; this altered the
cable load (hence the anchorage force)
 on the cable –
o direct blast application was involved if the explosions were triggered on
top, but not when the detonations were initiated at the side, whereby only
indirect blast effect was relevant
o neither steel fracture nor yielding took place, thereby suggesting superior
blast resilience
o the presence of the stay was rendered meaningless when total loss of axial
load happened; if the anchorage region was not demolished, the cable force
settled down eventually
Aside from being dictated by its own characteristics, the blast performance of the anchorage
precinct was also affected, to varied degree, by the blast response of the primary beam and
the cable.
Here, material breakdown is the signature failure style associated with blast loading. The
localised blast effects, understood in normal sense, were inflicted on the global extent. For
this reason, the alternative usage of the common terms in this particular context must be
cautiously distinguished.
Direct blast impact, albeit due to PBIEDs, was actually found to be more detrimental, in
contrast to indirect blast loading imposed by VBIEDs. The blast consequences produced in
both situations exhibited similarities as well as differences, depending on the specific issue of
concern.
308
7.3.2 Global Collapse Mechanism
Owing to inherent indeterminacy, the destruction of multiple anchorage points was required
in order to bring down the entire bridge, whereas the elimination of a single support was not
even sufficient to instigate partial collapse.
The deactivation of the anchorage points unfavourably upset the initial stabilising
arrangement of the whole bridge; even though the initial damage was essentially localised,
the detrimental effect was carried across the global system:
 the suspended spans together with the pylons fully participated while attempting
to reinstate equilibrium –
o the towers were deflected when the balancing condition was disturbed,
inclining outwards following the release of the mid stays at the centre span,
but drawn inwards after the tie-back stays at the side spans were removed;
otherwise, they were attracted along by the falling major span
o the minor spans were being lifted up when the pylons bent towards the
longer span, but lowered down otherwise; however, the shorter spans
continued to sink during complete failure
o the primary span dropped when the centre supports were disabled; but
likewise with the secondary spans, its displacement was also strongly
dependent on the movement of the towers
 within the stays, the axial forces altered because of the following factors –
o change of inclination due to readjustment of the positions of the termination
points
o deformation of the pylons (which was worsened by the amplifying P-Δ
effect), accompanied by the displacement of the carried spans, which either
released (totally or partially) or strained the internal tensile effect
o load redistribution prompted by both permanent and provisional loss of
supports
309
Chapter 7
The withdrawal of intermediate and tie-back stays, which assume essentially distinct roles,
was capable of triggering progressive collapse, an exceptionally brittle failure manner which
is highly disproportionate to the initiating cause. In fact, the collapse modes were not only an
uninterrupted advance but also a devastating amplification of the simplest responses derived
from the most basic failure setups. Overall, the catastrophe can either spread swiftly from the
main span or propagate aggressively from one end to the other, depending on the triggering
setting. The conveying spans were brought down together with the pylons, and consecutive
removal of supports was observed.
It was affirmed that traffic distribution assumes a pivotal role in affecting the possible
outcome. The vehicular loads at the back spans served as an extra counterbalance when
equilibrium was restorable, but their presence was indeed disastrous to the early survival of
the end spans during tragic collapse.
7.4 PROPOSED SOLUTIONS
7.4.1 Retrofitting Techniques
The elementary strengthening concepts, developed exclusively based on the blast

implications observed on the anchorage precincts (due to both PBIEDs and VBIEDs), were:
 placement of an overlaying shelter on top of the edge beam

 provision of a circular tube around the hollow core
 installation of protective shields enclosing the anchor pod
The defence options were either implemented as independent measures or selectively

combined to form integrated solutions.
When confronted with the below-girder blast impact, the internal lining (which disallowed
matter disengagement at the interior) was more credible than the external casing (which was
310
not even competent to limit debris ejection at the exterior), although neither was capable of
preventing the anchorage point from experiencing total destruction which was successfully
prohibited only when both were installed simultaneously.
On the other hand, after encountering the above-deck blast action, the anchorage region
remained serviceable when equipped with the inner shell and the outer armour, but a better
improvement was evident when the top cover (which was not able to act alone, as it has no
direct contribution to the well-being of the anchorage zone) was also provided at the same
time.
Steel and CFRP are both reliable blast retrofitting materials in this context. Nevertheless,
CFRP is relatively more inferior. When defended by steel, the anchorage precinct resolved
into a much more satisfactory condition. Furthermore, the CFRP sheets, which did not
possess sufficient strain capacity to fulfil the stringent demand, fractured. However, it must
be mentioned that steel has a much higher density, and this makes it undesirable if
lightweight fortification is sought after.
A rigorous sensitivity analysis was conducted. The detrimental blast effect can be alleviated
by:
 employing steel and CFRP of appropriate quality (depending on the degree of

protection required)
 increasing the plate thickness and sheet layer (hence the amount of material
involved)
 upgrading the bonding competency (since this improves the composite behaviour)
 adopting a suitable connection arrangement (as the issue of stress concentration
must be addressed)
The advanced retrofitting guidelines established consist of three key aspects:
 general statements –
311
Chapter 7
o although steel and CFRP are the major concerns, the general advice is also
useful as a basic reference to other materials
o the proposed techniques, which are applicable to existing facilities and
future assets, are relevant to both conventional and extreme loading
o the principal objective of a fortification arrangement shall be to guarantee
the successful survival of the anchorage zone in a blast catastrophe
o hardening units, adhesive agents and supplementary fasteners must be
appropriately represented through advanced FEM
 fundamental principles of blast defence –
o external hardening must be emphasised along with internal upgrade in order
to obtain both material and structural enhancement
o the unique roles and functions of each strengthening option must be clearly
identified and distinguished
o shielding the boundary faces will discourage material disintegration which
can severely undermine the original resistance of the subject
o utilisation of substances excellent for energy absorption, for immense stress
relief
o the confinement effect due to the presence of the outer armour and the inner
tube will give extra capacity
o improving the global axial and transverse as well as local stiffness will limit
the overall distortion
 advanced fortification tactics –
o although the inner lining is the most crucial feature, the outer casing is also
essential, but the top shelter is purely optional
o aside from the inherent resistance, the stiffness and the ductility of the
material must also be taken into account
o adequate plate thickness or sheet layer must be ensured
o adhesive agents (which govern the performance as well as the existence of
the composite behaviour) of outstanding quality must be used, while
fasteners are considered to assume only a supplementary role; bond failure
and material loss will lead to disablement of the intermediate interaction
312
7.4.2 Design Guidelines
A rigorous parametric analysis has been conducted, drawing in the common criteria that
dictate the ordinary behaviour of anchorage regions as well as the prime factors that govern
the blast response both in general and specific to this particular topic. It was assertively
concluded that the adverse blast consequences on the anchorage precincts can be confidently
limited by:
 furthering the standoff distance (by doing so, the blast impact will be alleviated
immensely)
 reducing the explosive size (hence the blast magnitude)
 expanding the bearing plate (since this decreases the ordinary induced actions,
thereby restricting the extent by which the weakened subject further deteriorates)
 lowering the gravitational burden (which delays the exhaustion of the residual
capacity)
 raising the support inclination (as this decreases the reaction force)
 enlarging the anchorage zone (this extends the energy-dissipating capability,
enhances the overall capacity and provides additional stiffness)
 augmenting the reinforcement quantity (thus boosts up the strengthening effort
and restraining contribution)
 improving the concrete quality (lessens the blast damage)
The key parameters selected may be collectively classified into two distinct domains, if
viewed in the perspective of functionality, and their relative authority (only in terms of
effectiveness, while the ranking should be redefined by counting in the actual constraints in
reality) can also be highlighted, as evaluated below:
 fundamental criteria which control the extreme and regular loading –
313
Chapter 7
o blast relief, which brings the most profound implication, hence the most
pertinent measure, shall be viewed as the top priority; altering the
detonation point is preferable to manipulating the charge weight
o varying the actual load is less attractive than transforming the bearing space
 dominant features that constitute the inherent blast and conventional resistance –
o structural dimension is indeed a pivotal variable that deserves the utmost
attention
o reinforcement scheme exerts stronger influence in comparison to concrete
grade
The well-established design strategies broadly encompass four principal aspects:
 general considerations –
o the design objective is to afford confident assurance that the anchorage
precinct is sufficiently resilient to ultimately survive a blast catastrophe
o the anchorage zone not only consists of the anchor pod (i.e. the outer part),
but also extends deep into the inner part of the longitudinal girder
o technically subdividing the subject into distinguishable regions with unique
behaviour is not appropriate under blast impact
o being perfectly compatible with the routine principles, the blast-induced
issues are unreservedly addressed along with the traditional problems
 core advice on blast analysis –
o the usual approaches are unable to serve the purpose of blast investigation;
instead, sophisticated FEM with advanced competencies in both impulsive
load estimation and blast consequence prediction appears to be a reliable
channel
o a credible qualitative appraisal, which relies on the overall distortion, the
response contour and matter disengagement, should be accompanied by a
convincing quantitative assessment, whereby anchorage displacement and
cable action are utilised as the key performance indicators
 innovative design tactics –
314
o outstretched platforms, sacrificial covers and resistant barriers

o maximise the a/h ratio together with the net contact surface
o excellent strength and ductility must be guaranteed, but the concrete grade
will most likely be dictated by the dominant components nevertheless; the
steel bars may be permitted to yield but not fracture
o the integrated reinforcement plan must be carefully delivered, keeping in
mind that room available for manoeuvre might be strictly lacking
o proper sizing of the anchorage zone, with all the restrictions sensibly
factored in
o composite schemes at the exterior and/or the interior
 integrated design logistics –
o the interest of both future assets and existing facilities has been taken care
of under an identical framework
o the methodology is to be implemented in accordance with the performance-
based design concept; distinct protection categories can be identified based
on the severity of the blast outcome
Unfortunately, the existing codes are invalid for blast occasions. It must be mentioned that
the routine practices, nevertheless, typically exhibit more profound significance for blast
events that took place above the deck rather than below the girder, and this creates an
intuitive perception regarding the blast reliability of an anchorage zone that will be or already
configured without commissioning any blast provisions.
Furthermore, by adhering to the old-fashioned manner, an anchorage point that deals with a
stronger reaction force, which is jointly decided by the imposed burden together with the
support orientation, will definitely be equipped with more formidable features, thereby
enjoying more superior blast endurance. Nonetheless, this by no means implies that the
progression of the ultimate destruction will be less violent. This virtually addressed the issue
regarding the relative performance amongst all the anchorage points contained within the
whole bridge.
315
Chapter 7
7.5 FUTURE RESEARCH
For comparison purpose, rigorous coupled analysis should be attempted. Blast wave
reflection around the restricted vicinity beneath the girder, which was not included in this
exercise, can be taken into account. However, it is believed that the expected failure mode
will not be altered through FSI simulation.
Owing to restriction in funding and other resources, blast experiments could not be conducted
to verify the results. The test data collected will allow proper fine-tuning of the modelling
strategies commissioned, further confirming the inherent reliability and improving the
precision at the same time. Full-scale specimens are preferable but not mandatory, and
realistic boundary conditions have to be provided by the reaction frames.
An in-depth explanation of the detailed mechanics of the dynamic deformation of anchorage

zones subjected to extreme blast impact will be interesting and useful. Furthermore, while
absolute destruction is the primary concern in this study, for the anchorage regions that
manage to survive the blast catastrophe, surface and body crack formation and propagation,
which have not been addressed here, are indeed a pertinent issue that ought to be looked into.
A mega explosion derived from either the ground or the water is expected to generate an
enormous blast wave with a wide coverage that might span across multiple anchorage points.
After factoring in the blast magnitude and the possible merging of the initial and reflected
waves, depending on the height clearance, it is believed that there is no absolute guarantee
that a huge standoff distance will definitely bring about negligible blast consequences, and
this needs further investigation.
Apart from the only arrangement of anchorage zone considered, similar need of other
configurations, including bearing units within hollow structures and for saddles, and
transitional sections secured externally to the deck-and-girder system, must also be
adequately catered for. An all-embracing contribution which gives a thorough knowledge of
this engineering topic, thereby facilitating technology transfer, is clearly warranted.
316
The anchorage precincts might give way if serious blast damage is inflicted on the deck-and-
girder system, which, in this case, should be viewed as the main subject of interest, with the
desired prestress effect included. A good understanding of this topic could be useful for
identifying the possible combined solutions.
Although believed to enjoy superior blast endurance, the cables are by no means invincible to
all circumstances. Unfortunately, comprehensive blast studies for cables are still strictly
lacking. The key variables should be revealed, and protective measures, if needed, must be
suggested, as it is fair to claim that weak cables often do not justify the exhaustive effort
devoted to ensure strong anchorage regions.
Very little is actually known about the blast effects on a compact anchorage assembly.
Elimination of the wedges, through either blast destruction or gripping failure, will, in a
sense, defeat the purpose of the existing anchorage precincts. However, it is recognised that
relevant expertise from mechanical and manufacturing engineering as well as advanced
material science, especially in seeking for potential remedies, is required.
The fact that the anchorage points contained within a given bridge very often appear in a vast
range of configurations, rather than being universally identical, implicitly implies that, apart
from the common deterministic evaluation, it is also worth making available some statistical
options to confirm the blast vulnerability in the probabilistic domain.
The potential of ultra-high strength concrete, both as a construction material as well as a

strengthening substance, in this context, should be thoroughly explored for new structures. It
is indeed crucial to ensure that as many alternatives as possible are readily available because
all sorts of situations can arise in reality.
Some of the simple ideas formulated for collapse prevention should be rigorously expanded,
with the aid of proper investigations, focussing heavily on giving clear recommendations
useful for actual implementation, thus leading to comprehensive guidelines for new facilities.
317
Chapter 7
318
LIST OF REFERENCES
AASHTO. (2010). “AASHTO LRFD bridge design specifications.” American Association of

State Highway and Transportation Officials (AASHTO), Washington.
ABAQUS. (2004). ABAQUS, ABAQUS, Inc., Providence, Rhode Island.
ACI Committee 318. (2011). “Building code requirements for structural concrete and
commentary (ACI 318-11).” American Concrete Institute (ACI), Farmington Hills,
Michigan.
ACI Committee 363. (1992). “Report on high-strength concrete (ACI 363R-92).” American
Concrete Institute (ACI), Farmington Hills, Michigan.
Ackland, K., Anderson, C., and Ngo, T. D. (2013). “Deformation of polyurea-coated steel
plates under localised blast loading.” International Journal of Impact Engineering, 51,
13-22.
Alia, A., and Souli, M. (2006). “High explosive simulation using multi-material
formulations.” Applied Thermal Engineering, 26(10), 1032-1042.
ANSYS. (2009). ANSYS, ANSYS, Inc., Pittsburgh, Pennsylvania.
ANSYS. (2010). “ANSYS AUTODYN user’s manual.” ANSYS, Inc., Pittsburgh,

Pennsylvania.
ARA. (2004). A.T.-BLAST, Applied Research Associates (ARA), Inc., Albuquerque, New
Mexico.
319
List of References
ASCE. (2010). “Minimum design loads for buildings and other structures.” American Society
of Civil Engineers (ASCE), Reston, Virginia.
ASCE Task Committee on Blast-Resistant Design. (2010). Design of blast-resistant buildings

in petrochemical facilities, American Society of Civil Engineers (ASCE), Reston,
Virginia.
ASI. (2006). Extreme Loading for Structures (ELS), Applied Science International (ASI),
Raleigh, North Carolina.
Bae, S., and Bayrak, O. (2008). “Seismic performance of full-scale reinforced concrete
columns.” ACI Structural Journal, 105(2), 123-133.
Baker, W. E., Cox, P. A., Westine, P. S., Kulesz, J. J., and Strehlow, R. A. (1983). Explosion
hazards and evaluation, Elsevier, Amsterdam.
Bangash, M. Y. H. (2009). Shock, impact and explosion: Structural analysis and design,
Springer, Berlin.
Baylot, J. T., Bullock, B., Slawson, T. R., and Woodson, S. C. (2005). “Blast response of
lightly attached concrete masonry unit walls.” Journal of Structural Engineering,
131(8), 1186-1193.
Biggs, J. M. (1964). Introduction to structural dynamics, McGraw-Hill, New York.
Bischoff, P. H., and Perry, S. H. (1991). “Compressive behaviour of concrete at high strain
rates.” Materials and Structures, 24(144), 425-450.
320
List of References
Breen, J. E., Burdet, O., Roberts, C., Sanders, D., and Wollmann, G. (1994). “Anchorage
zone reinforcement for post-tensioned concrete girders.” Transportation Research
Board (TRB), Washington.
Broadhouse, B. J. (1995). “The Winfrith concrete model in LS-DYNA3D.” Winfrith

Technology Centre, Winfrith Newburgh, England.
Brode, H. L. (1955). “Numerical solutions of spherical blast waves.” Journal of Applied

Physics, 26(6), 766-775.
Buchan, P. A., and Chen, J. F. (2007). “Blast resistant of FRP composites and polymer
strengthened concrete and masonry structures – A state-of-the-art review.” Composites
Part B: Engineering, 38(5-6), 509-522.
Bulson, P. S. (1997). Explosive loading of engineering structures, E & FN Spon, London.
Burdet, O. L. (1990). “Analysis and design of anchorage zones in post-tensioned concrete

bridges.” The University of Texas, Austin, Texas.
Cai, J.-G., Xu, Y.-X., Zhuang, L.-P., Feng, J., and Zhang, J. (2012). “Comparison of various
procedures for progressive collapse analysis of cable-stayed bridges.” Journal of
Zhejiang University – Science A, 13(5), 323-334.
CEN. (2000). “Prestressing steels – Part 3: Strand (prEN 10138-3).” Comité Européen de
Normalisation (CEN)/European Committee for Standardization, Brussels.
CEN. (2004). “Eurocode 2: Design of concrete structures – Part 1-2: General rules –
Structural fire design (EN 1992-1-2:2005).” Comité Européen de Normalisation
(CEN)/European Committee for Standardization, Brussels.
321
List of References
Chacos, G. P. (1993). “Wedge forces on post-tensioning strand anchors.” Post-Tensioning

Institute (PTI), Farmington Hills, Michigan.
Chan, S., Fawaz, Z., Behdinan, K., and Amid, R. (2007). “Ballistic limit prediction using a
numerical model with progressive damage capability.” Composite Structures, 77(4),
466-474.
Chen, W.-F. (1982). Plasticity in reinforced concrete, McGraw-Hill, New York.
Christoffersen, J., Hauge, L., Bjerrum, J., and Jensen, H. E. (1999). “Design and construction
of a CFRP cable-stayed footbridge.” Proceedings of the 1999 Conference of
International Association for Bridge and Structural Engineering (IABSE), Malmö,
Sweden.
Cimo, R. (2007). “Analytical modeling to predict bridge performance under blast loading.”
University of Delaware, Newark, Delaware.
Comité Euro-International du Béton (CEB)/Euro-International Committee for Concrete and

Fédération Internationale de la Précontrainte (FIP)/International Federation for
Prestressing. (1993). CEB-FIP model code 1990: Design code, Thomas Telford,
London.
Committee BD-001. (1998). “Australian Standard: Steel structures (AS 4100-1998).”

Standards Australia, Sydney.
Committee BD-002. (2009). “Australian Standard: Concrete structures (AS 3600-2009).”

Standards Australia, Sydney.
Committee BD-084. (2003). “Australian/New Zealand Standard: Prestressing anchorages

(AS/NZS 1314:2003).” Standards Australia, Sydney, and Standards New Zealand,
Wellington.
322
List of References
Committee BD-084. (2007). “Australian/New Zealand Standard: Steel prestressing materials

– Part 1: General requirements (AS/NZS 4672.1:2007).” Standards Australia, Sydney,
and Standards New Zealand, Wellington.
Committee BD-090. (2004a). “Australian Standard: Bridge design – Part 2: Design loads (AS
5100.2-2004).” Standards Australia, Sydney.
Committee BD-090. (2004b). “Australian Standard: Bridge design – Part 5: Concrete (AS
5100.5-2004).” Standards Australia, Sydney.
Cowper, G. R., and Symonds, P. S. (1957). “Strain-hardening and strain-rate effects in the
impact loading of cantilever beams.” Brown University, Providence, Rhode Island.
Cranz, C. (1926). Lehrbuch der Ballistik, Springer, Berlin.
Crawford, J. E., Krauthammer, T., Karagozian, J., and Hinman, E. (1999). “Structural
components – Analysis and design examples.” Structural design for physical security:
State of the practice, Task Committee: Conrath, E. J., Krauthammer, T., Marchand, K.
A., and Mlakar, P. F., American Society of Civil Engineers (ASCE), Reston, Virginia.
CSI. (2005). SAP2000, Computers and Structures (CSI), Inc., Berkeley, California.
Davidson, J. S., Fisher, J. W., Hammons, M. I., Porter, J. R., and Dinan, R. J. (2005). “Failure
mechanisms of polymer-reinforced concrete masonry walls subjected to blast.” Journal
of Structural Engineering, 131(8), 1194-1205.
Davis, C. E., Williams, G. D., Williamson, E. B., Marchand, K. A., McKay, A. E., and
Bayrak, O. (2009). “Design and detailing guidelines for bridge columns subjected to
blast and other extreme loads.” Proceedings of the 2009 Structures Congress, Austin,
Texas.
323
List of References
Deng, R.-B., and Jin, X.-L. (2009). “Numerical simulation of bridge damage under blast
loads.” World Scientific and Engineering Academy and Society (WSEAS) Transactions
on Computers, 8(9), 1564-1574.
Dusenberry, D. O. (2010). “General considerations for blast-resistant design.” Handbook for

blast-resistant design of buildings, Dusenberry, D. O., eds., John Wiley & Sons,
Hoboken, New Jersey.
FEMA. (2003). “Reference manual to mitigate potential terrorist attacks against buildings
(FEMA 426).” Federal Emergency Management Agency (FEMA), Department of
Homeland Security, United States of America.
Feng, J., and Cai, J. (2009). “Progressive collapse of cable-stayed bridges.” Proceedings of
the 2009 International Conference on Transportation Engineering, Chengdu, Sichuan.
fib. (2005). “Acceptance of stay cable systems using prestressing steels.” fédération
internationale du béton (fib)/International Federation for Structural Concrete, Lausanne,
Switzerland.
Fujikura, S., and Bruneau, M. (2011). “Experimental investigation of seismically resistant

bridge piers under blast loading.” Journal of Bridge Engineering, 16(1), 63-71.
Fujikura, S., Bruneau, M., and Lopez-Garcia, D. (2008). “Experimental investigation of

multihazard resistant bridge piers having concrete-filled steel tube under blast loading.”
Journal of Bridge Engineering, 13(6), 586-594.
Gannon, J. C. (2004). “Design of bridges for security against terrorist attacks.” The
University of Texas, Austin, Texas.
324
List of References
Gimsing, N. J. (1997). Cable supported bridges: Concept and design, John Wiley & Sons,
Hoboken, New Jersey.
GSA. (2003). “Progressive collapse analysis and design guidelines for new federal office
buildings and major modernization projects.” General Services Administration (GSA),
Washington.
GSA. (2005). “Facilities standards for the public buildings services.” General Services
Administration (GSA), Washington.
Guyon, Y. (1953). Prestressed concrete, John Wiley & Sons, Hoboken, New Jersey.
Hallquist, J. O. (2006). “LS-DYNA theory manual.” Livermore Software Technology

Corporation, Livermore, California.
Han, H., Taheri, F., Pegg, N., and Lu, Y. (2007). “A numerical study on the axial crushing
response of hybrid pultruded and ±45° braided tubes.” Composite Structures, 80(2),
253-264.
Han, L.-H., Liu, W., and Yang, Y.-F. (2008). “Behavior of thin walled steel tube confined
concrete stub columns subjected to axial local compression.” Thin-Walled Structures,
46(2), 155-164.
Han, T. H., Yoon, K. Y., and Kang, Y. J. (2010). “Compressive strength of circular hollow
reinforced concrete confined by an internal steel tube.” Construction and Building
Materials, 24(9), 1690-1699.
Hao, H., and Tang, E. K. C. (2010). “Numerical simulation of a cable-stayed bridge response
to blast loads, Part II: Damage prediction and FRP strengthening.” Engineering
Structures, 32(10), 3193-3205.
325
List of References
Hengprathanee, S. (2004). “Linear and nonlinear finite element analyses of anchorage zones
in post-tensioned concrete structures.” Virginia Polytechnic Institute and State
University, Blacksburg, Virginia.
Henrych, J. (1979). The dynamics of explosion and its use, Elsevier, Amsterdam.
Holland, C. E. (2008). “Blast-resistant design of highway bridge columns.” The University of

Texas, Austin, Texas.
Hopkinson, B. (1915). “British Ordnance Board minutes 13565.” British Ordnance Board,
London.
Hrynyk, T. D., and Myers, J. J. (2008). “Out-of-plane behavior of URM arching walls with
modern blast retrofits: Experimental results and analytical model.” Journal of
Structural Engineering, 134(10), 1589-1597.
Hyde, D. W. (1993). “User’s guide for microcomputer program ConWep – Applications of

TM5-855-1: Fundamentals of protective design for conventional weapons.” Waterways
Experiment Station, United States Army Corps of Engineers (USACE), Vicksburg,
Mississippi.
Islam, A. K. M. A., and Yazdani, N. (2008). “Performance of AASHTO girder bridges under
blast loading.” Engineering Structures, 30(7), 1922-1937.
Jo, B.-W., Byun, Y.-J., and Tae, G.-H. (2002). “Structural behavior of cable anchorage zones
in prestressed concrete cable-stayed bridge.” Canadian Journal of Civil Engineering,
29(1), 171-180.
Johansson, M. (2000). “Structural behaviour in concrete frame corners of civil defence

shelters: Non-linear finite element analyses and experiments.” Chalmers University of
Technology, Gothenburg, Sweden.
326
List of References
Judge, R., Yang, Z., Jones, S. W., and Beattie, G. (2011). “Numerical simulation of spiral-
strand cables subjected to high velocity fragment impact.” Proceedings of the 8th
European LS-DYNA Users Conference, Strasbourg, France.
Kimura, H., Itabashi, M., and Kawata, K. (2001). “Mechanical characterization of

unidirectional CFRP thin strip and CFRP cables under quasi-static and dynamic
tension.” Advanced Composite Materials, 10(2-3), 177-187.
Kingery, C. N., and Bulmash, G. (1984). “Airblast parameters from TNT spherical air burst
and hemispherical surface burst.” Ballistic Research Laboratory, Aberdeen Proving
Ground, Maryland.
Kinney, G. F., and Graham, K. J. (1985). Explosive shocks in air, Springer, Berlin.
Krauthammer, T., and Otani, R. K. (1997). “Mesh, gravity and load effects on finite element
simulations of blast loaded reinforced concrete structures.” Computers and Structures,
63(6), 1113-1120.
Li, V. C. (2003). “On engineered cementitious composites (ECC): A review of the material
and its applications.” Journal of Advanced Concrete Technology, 1(3), 215-230.
Loo, Y.-C., and Chowdhury, S. H. (2010). Reinforced and prestressed concrete: Analysis and
design with emphasis on application of AS 3600-2009, Cambridge University Press,
Cambridge.
LSTC. (2010). “LS-DYNA keyword user's manual.” Livermore Software Technology

Corporation (LSTC), Livermore, California.
327
List of References
Magallanes, J. M. (2008). “Importance of concrete material characterization and modelling to

predicting the response of structures to shock and impact loading.” Proceedings of the
10th International Conference on Structures under Shock and Impact, The Algarve,
Portugal.
Magnusson, J., and Hallgren, M. (2004). “Reinforced high strength concrete beams subjected
to air blast loading.” Proceedings of the 8th International Conference on Structures
under Shock and Impact, Crete, Greece.
Mahoney, E. E. (2007). “Analyzing the effects of blast loads on bridges using probability,
structural analysis, and performance criteria.” University of Maryland, College Park,
Maryland.
Malak, S. A. J. (1997). “Tensile and compressive behavior of slurry infiltrated mat concrete
(SIMCON).” Northeastern University, Boston, Massachusetts.
Malvar, L. J. (1998). “Review of static and dynamic properties of steel reinforcing bars.” ACI
Materials Journal, 95(5), 609-616.
Malvar, L. J., Crawford, J. E., and Morrill, K. B. (2007). “Use of composites to resist blast.”
Journal of Composites for Construction, 11(6), 601-610.
Malvar, L. J., Crawford, J. E., Wesevich, J. W., and Simons, D. (1997). “A plasticity concrete
material model for DYNA3D.” International Journal of Impact Engineering, 19(9-10),
847-873.
Malvar, L. J., and Ross, C. A. (1998). “Review of strain rate effects for concrete in tension.”
ACI Materials Journal, 95(6), 735-739.
Marchand, K. A., and Plenge, B. T. (1998). “Concrete hard target spall and breach model.”
Air Force Research Laboratory, Eglin Air Force Base, Florida.
328
List of References
Marjanishvili, S. M. (2004). “Progressive analysis procedure for progressive collapse.”

Journal of Performance of Constructed Facilities, 18(2), 79-85.
Marjanishvili, S., and Agnew, E. (2006). “Comparison of various procedures for progressive
collapse analysis.” Journal of Performance of Constructed Facilities, 20(4), 365-374.
Matthews, D. S. (2008). “Blast effects on prestressed concrete bridges.” Washington State

University, Washington.
Mays, G., and Feltham, I. (2009). “Design of elements in reinforced concrete and masonry.”
Blast effects on buildings, Cormie, D., Mays, G., and Smith, P., eds., Thomas Telford,
London.
Mays, G., Feltham, I., and Banfi, M. (2009). “Design of elements in structural steel.” Blast
effects on buildings, Cormie, D., Mays, G., and Smith, P., eds., Thomas Telford,
London.
McClendon, M. A. (2007). “Blast resistant design for roof systems.” University of Missouri,
Columbia, Missouri.
Micelli, F., and Modarelli, R. (2013). “Experimental and analytical study on properties
affecting the behaviour of FRP-confined concrete.” Composites Part B: Engineering,
45(1), 1420-1431.
MSC Software. (2002). MSC.Dytran, MSC Software Corporation, Newport Beach,

California.
MSC Software. (2009). MD Nastran Explicit Nonlinear (SOL 700), MSC Software
Corporation, Newport Beach, California.
329
List of References
Muszynski, L. C., and Purcell, M. R. (2003). “Use of composite reinforcement to strengthen

concrete and air-entrained concrete masonry walls against air blast.” Journal of
Composites for Construction, 7(2), 98-108.
Mutalib, A. A., and Hao, H. (2011). “Numerical analysis of FRP-composite-strengthened RC

panels with anchorages against blast loads.” Journal of Performance of Constructed
Facilities, 25(5), 360-372.
Myers, J. J., Belarbi, A., and El-Domiaty, K. A. (2004). “Blast resistance of FRP retrofitted
un-reinforced masonry (URM) walls with and without arching action.” The Masonry
Society Journal, 22(1), 9-26.
Nair, R. S. (2006). “Preventing disproportionate collapse.” Journal of Performance of

Constructed Facilities, 20(4), 309-314.
Nawy, E. G. (2010). Prestressed concrete: A fundamental approach, Prentice Hall, New

Jersey.
Olmati, P., and Giuliani, L. (2013). “Progressive collapse susceptibility of a long span
suspension bridge.” Proceedings of the 2013 Structures Congress, Pittsburgh,
Pennsylvania.
Paik, J. K., and Thayamballi, A. K. (2003). Ultimate limit state design of steel-plated
structures, John Wiley & Sons, Hoboken, New Jersey.
Podolny, W., and Scalzi, J. B. (1986). Construction and design of cable-stayed bridges, John
Wiley & Sons, Hoboken, New Jersey.
PTI. (2006). “Post-tensioning manual.” Post-Tensioning Institute (PTI), Farmington Hills,

Michigan.
330
List of References
PTI. (2007). “Recommendations for stay cable design, testing and installation.” Post-
Tensioning Institute (PTI), Farmington Hills, Michigan.
Raman, S. N. (2011). “Polymeric coatings for enhanced protection of reinforced concrete

structures from the effects of blast.” The University of Melbourne, Melbourne.
Randers-Pehrson, G., and Bannister, K. A. (1997). “Airblast loading model for DYNA2D and
DYNA3D.” Army Research Laboratory, Maryland.
Ray, J. C. (2006). “Validation of numerical modeling and analysis of steel bridge towers
subjected to blast loadings.” Proceedings of the 2006 Structures Congress, St. Louis,
Missouri.
Ray, J. C., Armstrong, B. J., and Slawson, T. R. (2003). “Airblast environment beneath a
bridge overpass.” Transportation Research Record, 1827, 63-68.
REI. (2004). STAAD.Pro, Research Engineers International (REI), Yorba Linda, California.
Roberts, C. L. (1990). “Behavior and design of the local anchorage zone in post-tensioned
concrete.” The University of Texas, Austin, Texas.
Roberts, J. E., Kulicki, J. M., Beranek, D. A., Englot, J. M., Fisher, J. W., Hungerbeeler, H.,
Isenberg, J., Seible, F., Stinson, K., Tang, M. C., and Witt, K. (2003).
“Recommendations for bridge and tunnel security.” American Association of State
Highway and Transportation Officials (AASHTO) and Federal Highway
Administration (FHWA), Washington.
Rogowsky, D. M., and Marti, P. (1996). “Detailing for post-tensioned.” VSL International,
Bern, Switzerland.
331
List of References
Rose, T. A. (2001). “An approach to the evaluation of blast loads on finite and semi-infinite
structures.” Cranfield University, Cranfield, England.
SAIC. (2001). BlastX, Science Applications International Corporation (SAIC), McLean,

Virginia.
Sanders, D. H. (1990). “Design and behavior of anchorage zones in post-tensioned concrete

members.” The University of Texas, Austin, Texas.
Schneider, B., and Mondragon, R. (1989). “Design and construction techniques for
SIFCON.” The University of New Mexico, Albuquerque, New Mexico.
Schwer, L. E. (2009). “Strain rate induced strength enhancement in concrete: Much ado about
nothing?” Proceedings of the 7th European LS-DYNA Conference, Salzburg, Austria.
Silva, P. F., and Lu, B. (2007). “Improving the blast resistance capacity of RC slabs with
innovative composite materials.” Composites Part B: Engineering, 38(5-6), 523-534.
Smilowitz, R., and Tennant, D. (2010). “Structural systems design.” Handbook for blast-
resistant design of buildings, Dusenberry, D. O., eds., John Wiley & Sons, Hoboken,
New Jersey.
Smith, P., and Cormie, D. (2009). “Blast loading.” Blast effects on buildings, Cormie, D.,
Mays, G., and Smith, P., eds., Thomas Telford, London.
Smith, P. D., and Hetherington, J. G. (1994). Blast and ballistic loading of structures,
Butterworth-Heinemann, Oxford.
Son, J. (2008). “Performance of cable supported bridge decks subjected to blast loads.” The
University of California, Berkeley, California.
332
List of References
Son, J., and Lee, H.-J. (2011). “Performance of cable-stayed bridge pylons subjected to blast
loading.” Engineering Structures, 33(4), 1133-1148.
Starossek, U. (2007). “Typology of progressive collapse.” Engineering Structures, 29(9),

2302-2307.
Starossek, U. (2009). “Avoiding disproportionate collapse of major bridges.” Structural

Engineering International, 19(3), 289-297.
Starossek, U., and Haberland, M. (2010). “Disproportionate collapse: Terminology and

procedures.” Journal of Performance of Constructed Facilities, 24(6), 519-528.
Susantha, K. A. S., Ge, H., and Usami, T. (2001). “Uniaxial stress-strain relationship of
concrete confined by various shaped steel tubes.” Engineering Structures, 23(10),
1331-1347.
Suthar, K. N. (2007). “The effect of dead, live and blast loads on a suspension bridge.”
University of Maryland, College Park, Maryland.
Tabatabai, H. (2005). “Inspection and maintenance of bridge stay cable systems: A synthesis
of highway practice.” Transportation Research Board (TRB), Washington.
Tabiei, A., and Wu, J. (2000). “Roadmap for crashworthiness finite element simulation of
roadside safety structures.” Finite Elements in Analysis and Design, 34(2), 145-157.
Tan, S., and Astaneh-Asl, A. (2003). “Cable-based retrofit of steel building floors to prevent
progressive collapse.” The University of California, Berkeley, California.
Tan, S. H., Poon, J. K., Chan, R., and Chng, D. (2012). “Retrofitting of reinforced concrete
beam-column via steel jackets against close-in detonation.” Proceedings of the 12th
International LS-DYNA Users Conference, Detroit, Michigan.
333
List of References
Tanapornraweekit, G. (2010). “Behaviour of fibre reinforced polymer (FRP) strengthened RC

slabs subjected to blast loading.” The University of Melbourne, Melbourne.
Tang, E. K. C., and Hao, H. (2010). “Numerical simulation of a cable-stayed bridge response
to blast loads, Part I: Model development and response calculations.” Engineering
Structures, 32(10), 3180-3192.
Thompson, B. P., and Bank, L. C. (2007). “Risk perception in performance-based building

design and applications to terrorism-resistant design.” Journal of Performance of
Constructed Facilities, 21(1), 61-69.
Tokal-Ahmed, Y. M. (2009). “Response of bridge structures subjected to blast loads and

protection techniques to mitigate the effect of blast hazards on bridges.” Rutgers, The
State University of New Jersey, New Brunswick, New Jersey.
Toro, E. F. (1997). Riemann solvers and numerical methods for fluid dynamics: A practical
introduction, Springer, Berlin.
USACE. (2002). Single-Degree-of-Freedom Plastic Analysis (SPAn32), United States Army

Corps of Engineers (USACE), Washington.
USACE. (2004). Bridge Explosive Loading (BEL), United States Army Corps of Engineers
(USACE), Washington.
United States Army Corps of Engineers, Naval Facilities Engineering Command, Air Force
Civil Engineer Support Agency. (2002). “Design and analysis of hardened structures to
conventional weapons effects (UFC 3-340-01).” Department of Defence, United States
of America.
334
List of References
Civil Engineer Support Agency. (2005). “Design of buildings to resist progressive
collapse (UFC 4-023-03).” Department of Defence, United States of America.
Civil Engineer Support Agency. (2008). “Structures to resist the effects of accidental
explosions (UFC 3-340-02).” Department of Defence, United States of America.
Uy, B. (2002). “Strength of reinforced concrete columns bonded with external steel plates.”
Magazine of Concrete Research, 54(1), 61-76.
Wada, Y., and Liou, M.-S. (1997). “An accurate and robust flux splitting scheme for shock
and contact discontinuities.” Journal on Scientific Computing, 18(3), 633-657.
Walther, R., Houriet, B., Isler, W., Moïa, P., and Klein, J. F. (1999). Cable stayed bridges,
Thomas Telford, London.
Warner, R. F., Rangan, B. V., Hall, A. S., and Faulkes, K. A. (1998). Concrete structures,
Longman, Melbourne.
Welsh, L. M., and Harding, J. (1985). “Dynamic tensile response of unidirectionally-

reinforced carbon epoxy and glass epoxy composites.” Proceedings of the 5th
International Conference on Composite Materials, San Diego, California.
Whittaker, A., Hamburger, R., and Mahoney, M. (2003). “Performance-based engineering of

buildings for extreme events.” Proceedings of the Steel Building Symposium on Blast
and Progressive Collapse Resistance, New York.
Williams, G. D., and Williamson, E. B. (2011). “Response of reinforced concrete bridge

columns subjected to blast loads.” Journal of Structural Engineering, 137(9), 903-913.
335
List of References
Williamson, E. B., Bayrak, O., Davis, C., and Williams, G. D. (2011a). “Performance of
bridge columns subjected to blast loads. I: Experimental program.” Journal of Bridge
Engineering, 16(6), 693-702.
Williamson, E. B., Bayrak, O., Davis, C., and Williams, G. D. (2011b). “Performance of
bridge columns subjected to blast loads. II: Results and recommendations.” Journal of
Bridge Engineering, 16(6), 703-710.
Williamson, E. B., Bayrak, O., Williams, G. D., Davis, C. E., Marchand, K. A., McKay, A.
E., Kulicki, J., and Wassef, W. (2010). “Blast-resistant highway bridges: Design and
detailing guidelines.” Transportation Research Board (TRB), Washington.
Williamson, E. B., and Winget, D. G. (2005). “Risk management and design of critical
bridges for terrorist attacks.” Journal of Bridge Engineering, 10(1), 96-106.
Winget, D. G., Marchand, K. A., and Williamson, E. B. (2005). “Analysis and design of
critical bridges subjected to blast loads.” Journal of Structural Engineering, 131(8),
1243-1255.
Wolff, M., and Starossek, U. (2009). “Cable loss and progressive collapse in cable-stayed
bridges.” Bridge Structures, 5(1), 17-28.
Wollmann, G. P. (1992). “Anchorage zones in post-tensioned concrete structures.” The

University of Texas, Austin, Texas.
Wu, C., and Hao, H. (2005). “Modeling of simultaneous ground shock and airblast pressure
on nearby structures from surface explosions.” International Journal of Impact
Engineering, 31(6), 699-717.
336
List of References
Wu, C., Oehlers, D. J., Rebentrost, M., Leach, J., and Whittaker, A. S. (2009). “Blast testing
of ultra-high performance fibre and FRP-retrofitted concrete slabs.” Engineering
Structures, 31(9), 2060-2069.
Wu, K.-C., Li, B., and Tsai, K.-C. (2011). “Residual axial compression capacity of localized
blast-damaged RC columns.” International Journal of Impact Engineering, 37(1), 29-
40.
Yan, D., and Chang, C. C. (2009). “Vulnerability assessment of cable-stayed bridges in

probabilistic domain.” Journal of Bridge Engineering, 14(4), 270-278.
Yi, Z. (2009). “Blast load effects on highway bridges.” The City University of New York,
New York.
Zheng, R. (2007). “Performance of FRP-concrete bridges under blast loading.” Florida

International University, Miami, Florida.
Zhou, F. (2009). “Blast/Explosion resistant analysis of composite steel girder bridge system.”
Florida Atlantic University, Boca Raton, Florida.
Zhou, Y., Jiang, D., and Xia, Y. (2001). “Tensile mechanical behavior of T300 and M40J
fiber bundles at different strain rate.” Journal of Materials Science, 36(4), 919-922.
Zhou, Y., Wang, Y., Xia, Y., and Jeelani, S. (2010). “Tensile behavior of carbon fiber
bundles at different strain rates.” Materials Letters, 64(3), 246-248.
337
APPENDICES
Appendices
Appendix A MITIGATION STRATEGIES FOR GLOBAL COLLAPSE
A.1 General Outline
The three key concepts which contribute to collapse prevention, i.e. overall resilience
assurance, robustness requirement and vulnerability control, comprise structurally oriented as
well as non-structural-related tactics applicable on the global and local extent, as explained in
Figure A.1.
Figure A.1: Disproportionate and progressive collapse prevention
Robustness is a term adopted for the local ability of the anchorage regions to endure blast
impact, while collapse resistance is commonly known as the global capability of the bridge to
arrest tragic failure after initial damage took place. Vulnerability is defined as the exposure to
blast threat.
The two general approaches to inhibit disproportionate and progressive collapse, as outlined
in ASCE (2010), are the direct method and the indirect method 31 . The former aims to
31
Both endorsed by GSA (2003) and UFC 4-023-03, which are two well-accepted codes to be closely adhered
to in this respect.
A-1
Appendices
prescribe explicit solutions, by relying fully on detailed collapse analysis; as for the latter, the
effectiveness of the suggested options can only be recognised in an implicit manner.
A.2 Threat Control
Reducing the probability of occurrence of blast incidents and restricting the magnitude of
blast action, with the aid of the control measures described in Table A.1, are crucial, as
hindrance of tragedies is always more preferable to assurance of superior endurance.
Public awareness plays a key role in avoiding tragic accidental events, and deep cooperation
between the community and the relevant officials is particularly vital to effectively counter
intentional plots.
On-going surveillance, which overcomes the existing obstructions that adversely affect the
desired visual coverage, enables prompt actions to be carried out once suspicious
circumstances arise.
Proper information and site restriction and management can be viewed as a non-structural
means of limiting the severity of blast threat, an inherent function of adequate standoff
allowance and innovative designs, which are both structural tactics.
While efficient planning and coordination is crucial for life safety, it is only beneficial when
sufficient ductility is ensured, as a highly brittle failure will definitely not permit enough
reaction time.
Close cooperation between various parties is sought after, since threat reduction requires
multidisciplinary expertise. An integrated mitigation scheme, which can be frequently
reviewed and updated, is exceptionally vital for the systematic implementation of the
individual options.
A-2
Appendices
Table A.1: Blast threat mitigation

Category Strategy Description and example
Standoff Practical erection of impact-resistant barriers and
provision introduction of ancillary extensions
Structural Strategic positioning of the anchorage points (e.g.
Feasible design within a hollow structure), effective dimensioning of the
girders and the anchorage zones
Sensible education programmes which enable the
Promotion of
responsibility of the community to be openly made
public
aware (so as to encourage more active participation) and
consciousness
ensure a guided crisis response
Regular patrols conducted by national security
Surveillance and
authorities, installation of reliable detection and alert
monitoring
systems
Excellent line-of-sight to the surrounding environment,
Proper site
proper arrangement of auxiliary components (which do
Non- layout
not facilitate the placement of blast sources)
structural
Sensitive official data, especially the intrinsic features of
Information
the cables and the anchorage regions, must be treated as
censorship
highly confidential
Unauthorised access to the immediate vicinity of the
Access
anchorage precincts is strictly prohibited (clear
deterrence
definition of “no trespassing zones” is essential)
Prior establishment of efficient planning and
Emergency
coordination schemes, prompt evacuation during the
response
aftermath is crucial to limit the residual burden that
management
remains on the bridge
A-3
Appendices
A.3 Local Enhancement
Improving the robustness of the anchorage regions, with the aid of appropriate retrofitting
techniques and design solutions, will substantially limit the initial damage.
Sufficiently protecting all the anchorage zones, if possible, is of course the most ideal
resolution. Nevertheless, as a more viable alternative, heavily fortifying alternate anchorage
regions at appropriate (not necessarily constant) intervals is also acceptable. After all, the
underlying intent is none other than to prevent the failure of multiple neighbouring anchorage
points that will eventually lead to a devastating collapse.
The concept of tributary area is exceptionally pivotal here. The deactivation of consecutive
anchorage points appears as the most dangerous failure scenario. It was found that
progressive collapse was triggered following the destruction of three successive anchorage
regions (Combination A), but not when two pairs (Combination B) or three alternate
(Combination C) cables were withdrawn, as shown in Figure A.2. This can be explained by
considering the extra burden allocated to the remaining stays, which can be reasonably
estimated by adjusting the enclosing centrelines.
Figure A.2: Support failure scenarios
A-4
Appendices
Neither downplaying nor overemphasising the roles of the stays should be encouraged at all,
since the exclusion of intermediate and end supports are equally competent to give rise to a
total failure. Reliance on internal action, which depends on both the imposed load as well as
the cable orientation, as the indicator of criticality for individual stays is not always
acceptable. However, closely spaced supports are relatively more susceptible to a given
explosion. Also, the cables that are subjected to heavier loading and/or with less room for
extra burden deserve greater attention.
A.4 Global Improvement
A.4.1 Alternative Load Path
Both fib (2005) and PTI (2007) stipulated that cable-stayed bridges must be able to withstand
the loss of at least one stay. However, when dealing with extreme blast events, a much higher
degree of indeterminacy is called for in order to confidently assure remarkable collapse
resistance.
Positioning extra cables on the bridge, as opposed to relying solely on a few mega supports,
as shown in Figure A.3, appears as the most straightforward option in this case. This is
consistent with the advice offered by GSA (2005), which suggests that column spacing
should be minimised.
Furthermore, adopting a double-plane stay configuration where the individual cables assume
less responsibility in comparison to their counterparts in an equivalent single-plane stay
arrangement, as evident in Figure A.4, is also a better decision in this respect.
However, the noble intention of providing extra cables could lead to anchorage zones with
not as much inherent blast endurance (since the supports now attract smaller conventional
loads, and their corresponding designs are, therefore, downgraded), hence agreeing totally
with Nair (2006) with similar stance on column. Apart from that, the fact that closing the
cable spacing might, in turn, cause additional anchorage points to be subjected to direct blast
A-5
Appendices
impact must also be taken into careful consideration. Thus, a balanced solution is of
paramount importance.
(a)
(b)
Figure A.3: Stay positioning – (a) Mega cables; (b) Standard cables
A-6
Appendices
(a)
(b)
Figure A.4: Stay systems – (a) Single-plane; (b) Double-plane
A.4.2 Interconnection
Excellent continuity can be achieved by providing strong connections and strengthening the
end anchorages for rebars. This concept is, strictly speaking, not a stand-alone approach
independent of redundancy provision and local protection.
However, the risk of counter-productivity must not be overlooked, since failure, rather than
being arrested immediately, might now be carried over involuntarily to a larger part of the
system (Nair, 2006; Starossek, 2009).
A-7
Appendices
A.4.3 Compartmentalisation
Isolation by means of segmentation appears as an alternative if, and only if, an acceptable
failure extent can be reasonably specified. It is useful in certain circumstances that do not
favour the existence of continuity, particularly when sufficient strength to withstand
overloading is not guaranteed (Starossek, 2009).
Weak elements are meant to be eliminated or separated instantly or as soon as possible,

whereas strong elements are intended to inhibit damage propagation. Starossek and
Haberland (2010) proposed that segment borders can also be created by ensuring large
inherent energy-dissipating competency.
A-8
Appendices
Appendix B ANALYSIS AND DESIGN OF ANCHORAGE ZONES
B.1 Traditional Analytical Methods
B.1.1 Approximate Methods
The bursting force, TBurst, and the corresponding centroid, dBurst, given in ACI 318-11 are
written as:
𝑎
𝑇𝐵𝑢𝑟𝑠𝑡 = 0.25∑𝐹 (1 − ) + 0.5|∑(𝐹 sin 𝛼0 )| (B.1)
ℎ
𝑑𝐵𝑢𝑟𝑠𝑡 = 0.5(ℎ − 2𝑒) + 5𝑒 sin 𝛼0 (B.2)
where:
F = anchorage force
a = anchorage dimension in the direction of interest
h = depth or width of the structural cross section
α0 = angle of inclination of the anchorage force
e = eccentricity of the anchorage assembly
The concept of symmetrical prism introduced by Guyon (1953) is embraced by AS 3600-

2009. The depth and the length of a symmetrical prism is taken as twice the distance from the
centre of the anchorage assembly to the nearest edge of the structure in the direction
considered, as depicted in Figure B.1.
In order to derive the edge tensile force, an anchorage zone may be conceived of as subjected
to a combined flexural and axial load. The transverse moments can be divided by the
corresponding lever arms (AS 3600-2009) (as further demonstrated by Loo and Chowdhury,
2010). Nevertheless, for further simplicity, Warner et al. (1998) advised that the spalling
force could be simply taken as 4% of the anchorage load, while Rogowsky and Marti (1996)
B-1
Appendices
believed that an estimate of 3% will suffice, as opposed to the less conservative suggestion of
2% by PTI (2006).
Figure B.1: Symmetrical prism
B.1.2 Strut-and-Tie Model
A strut-and-tie model approximates the flow of action using compressive struts, tensile ties
and virtual nodes (as illustrated in Figure B.2), by adhering to the theory of plasticity. The
forces in the struts and the ties are established by satisfying the equilibrium conditions. The
selected layout ought to reflect closely the actual stress distribution which is often not readily
known especially under a complicated situation, hence the main challenge for this particular
approach.
Figure B.2: Strut-and-tie model
B-2
Appendices
B.1.3 Finite Element Method
The stress state within an anchorage zone can be examined with both linear and non-linear
FEA (e.g. Jo et al., 2002; Hengprathanee, 2004). The bursting force may be obtained by
integrating the bursting stresses along the tendon. The stress vectors produced using the FEM
(as displayed in Figure B.3) are often utilised to indicate the possible force paths for a strut-
and-tie model. The elastic approach is normally acceptable, although cracking will provoke
stress redistribution (Burdet, 1990) which could still be accounted for depending on the
capability of the specific software (Nawy, 2010).
Figure B.3: Stress vectors for an anchorage zone
B.2 Standard Design Practices
B.2.1 General Notes
Most design codes deal with the design of ordinary (rectangular) end blocks, while additional
information related to intermediate anchorages and external diaphragms is covered
extensively by Wollmann (1992) and Breen et al. (1994).
The allocation of responsibility in regard to the design of anchorage zones is addressed

explicitly in AASHTO (2010). The anchorage supplier is liable for the well-being of the local
B-3
Appendices
zone, whereas the arrangement of the general zone is entirely under the jurisdiction of the
practising engineer.
A three-dimensional problem could be conveniently handled by dealing with the two-

dimensional orthogonal principal planes separately. A complex structure may also be
abstractly disassembled into multiple entities that are viewed as virtually planar.
The tensile strength of concrete is completely disregarded, hence the full obligation of the
steel reinforcement in handling the design actions.
B.2.2 Design Techniques
B.2.2.1 Bursting Stress
The bursting reinforcement can appear in the form of mats consisting of parallel rods, welded
meshes, closed ties or continuous spiral bars (AS 5100.5-2004). The maximum allowable
stress is restricted to 200 MPa, since most of the cracks will be contained within the body of
the structural member.
AASHTO (2010) specified that the splitting reinforcement shall be placed throughout the
distance given by the lesser of 2.5dBurst and 1.5h, extending across the full width and depth of
the cross section, with the centroid coincides with the location of the resultant bursting force,
and the maximum spacing not exceeding the smaller of 12 in. (≈ 300 mm) and 24 bar
diameter.
B.2.2.2 Edge Tensile Stress
The spalling reinforcement shall be positioned as close as possible to the free edges whilst
fulfilling the cover requirements. For crack control purpose, the induced stress shall be no
more than 150 MPa (AS 5100.5-2004).
B-4
Appendices
B.2.2.3 Bearing Stress
Without special anchorage units, the design bearing resistance, fn, is given in AASHTO
(2010) as the lesser of:
𝐴
𝜑0.7𝑓𝑐𝑖 √ and (B.3)
𝐴𝑔
𝜑2.25𝑓𝑐𝑖 (B.4)
where:
fci = compressive strength of concrete during machine jacking (ksi)
A = effective bearing area (in.2)
Ag = gross bearing area (in.2)
With special anchorage devices, PTI (2006) suggested that, based on the findings reported by
Roberts (1990), the nominal load bearing capacity, Pn, can be expressed as a collective
contribution from the concrete and the local zone reinforcement:
𝑃𝑛 = 𝜂0 (𝑃𝑐𝑐 + 𝑃𝑟𝑐 ) ≤ 3𝑓𝑐𝑖 𝐴𝑏 (B.5)
𝐴
𝑃𝑐𝑐 = 0.8𝑓𝑐𝑖 𝐴𝑏 √ ≤ 2𝑓𝑐𝑖 𝐴𝑏 (B.6)
𝐴𝑔
𝑃𝑟𝑐 = 4.1𝑓𝑙𝑎𝑡 𝐴𝑐𝑜𝑟𝑒 (B.7)
where:
η0 = calibration factor
Pcc = concrete confinement contribution
Prc = reinforcement restraining contribution
Ab = net bearing area
flat = lateral confinement pressure
Acore = effective restrained core area
B-5
Appendices
The lateral confinement pressure and the effective restrained core area due to the local zone
reinforcement are described in Figure B.4.
(a)
(b)
Figure B.4: Confinement reinforcement for the local zone – (a) Continuous spiral; (b)
Discrete ties (PTI, 2006)
B-6

Blast Loading On Anchorage Zones in Cable-Stayed Bridges

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Blast Loading On Anchorage Zones in Cable-Stayed Bridges

Uploaded by

Copyright:

Available Formats

BLAST LOADING ON ANCHORAGE ZONES IN

LUA WEI YEW

Submitted in total fulfilment of the requirements of the degree of

Department of Infrastructure Engineering

The principal objectives of this ground-breaking research include: determination of the

Withdrawal of multiple intermediate and/or counterbalancing stays is required to serve as an

This is to certify that:

i. the thesis comprises only my original work,

LUA Wei Yew

Independent of this project, the Department of Infrastructure Engineering in the University of

First and foremost, I would like to express my deepest gratitude to my family: my

I am immensely grateful to my principal supervisor, Professor Priyan Mendis, for always

2.2.2.1 Blast Loading Categories ............................................................................ 18

2.6 SUMMARY .................................................................................................................. 64

4.2.1.2 Anchorage Devices .................................................................................... 120

5.1 OVERVIEW ............................................................................................................... 187

6.1 OVERVIEW ............................................................................................................... 247

7.2.2 Local Blast Simulation ..................................................................................... 305

Table 2.1: TNT equivalent factors (UFC 3-340-01) ................................................................ 17

Table 6.1: Mechanical properties of concrete ........................................................................ 256

Figure 1.1: Research methodology framework ........................................................................ 10

Figure 2.20: Strain rate effects on steel ................................................................................... 43

Figure 3.11: M1600 traffic load (AS 5100.2-2004) ................................................................. 82

Figure 5.14: Blast performance of the anchorage precinct subjected to VBIED

Figure 5.29: Anchorage displacement records following VBIED explosions (with

AEM applied element method

A area/effective bearing area

ec compressive fibre failure mode indicator

J2 second invariant of a deviatoric stress tensor

R1 standoff distance from point A

αs, γs DIF parameters for compressive strength

∆σ m maximum failure surface

1.1 BACKGROUND AND CONTEXT

1.2 PROBLEM STATEMENT AND RESEARCH NEEDS

Blast studies concerning the anchorage precincts in cable-stayed bridges appear to be

So far, no prescriptive and performance-based design recommendations have been proposed

Unfortunately, sophisticated computational models worthy of being employed for blast

1.3 SCOPE, OBJECTIVES AND KEY CONTRIBUTIONS

Blast incidents instigated either inadvertently or deliberately, involving vehicle-borne-

Disproportionate collapse refers to a major effect instigated by relatively minor damage;

1.4 RESEARCH METHODOLOGY

LS-DYNA, a commercially developed general-purpose finite element software capable of

Figure 1.1: Research methodology framework

1.5 THESIS ORGANISATION

Chapter 5 is dedicated purely to existing facilities. A thorough evaluation of the feasibility of

Chapter 6 is concerned primarily with future services. A rigorous parametric study is

A state-of-the-art review of the previous research projects concerning bridges is also

Blast studies broadly differ in terms of objective, problem encountered, methodology,

2.2 FUNDAMENTALS OF BLAST PHENOMENON

2.2.1 General Characteristics of Explosions

An explosion is a phenomenon where energy is released rapidly within an infinitesimal

2.2.1.3 Explosion Energy

2.2.1.4 Formation of Blast Waves

Figure 2.1: Blast wave pressure profile

2.2.1.5 TNT Equivalency

𝑊𝑇𝑁𝑇 𝑄𝑇𝑁𝑇 = 𝑊𝑋 𝑄𝑋 (2.1)

Table 2.1: TNT equivalent factors (UFC 3-340-01)

2.2.1.6 Scaling Law

2.2.2 Blast Loading on Structures

2.2.2.1 Blast Loading Categories

 confined explosions – fully vented explosions, partially confined explosions and

Figure 2.2: Blast loading due to an unconfined detonation

2.2.2.2 Blast Wave Reflection