Public Version PDF

THE EFFECTS OF BOND AND ANCHORAGE ON
THE BEHAVIOUR AND DESIGN OF

COMPOSITE SLABS
Mohammad Masud Rana
A thesis in fulfilment of the requirements for the degree of

Doctor of Philosophy
Centre for Infrastructure Engineering and Safety
School of Civil and Environmental Engineering, Faculty of Engineering
The University of New South Wales, Sydney, Australia
January 2016
Abstract
Composite steel-concrete slabs represent a popular form of flooring system in
high rise buildings and bridges. This thesis comprises of experimental studies and
numerical analyses into the behaviour of post-tensioned composite slabs and
composite slabs with end anchorage.
The shear bond between the profiled steel sheets and the concrete is very crucial
to the behaviour and strength of post-tensioned composite slabs. Therefore, the
possible influence of prestress on this bond behaviour is a fundamental issue for
these types of slabs. In this context, a series of push tests incorporating post-
tensioning at the steel bar to apply precompression to the concrete were conducted.
The test results indicated that the prestress has a detrimental effect on the bond
between the profiled steel sheets and concrete in post-tensioned composite slabs.
This research study developed a three dimensional nonlinear finite element model
of a post-tensioned push specimen using ABAQUS to simulate the bond stress-slip
relations accurately. The bond stress-slip relations obtained from push tests were
then assessed for post-tensioned profiled composite slabs numerically. A parametric
study was conducted using the verified finite element model to study the effects of
concrete strength, level of prestress, and thickness of the profiled steel sheets on the
overall performance of this type of slabs.
The end anchorages in combination with other shear transfer mechanism play an
important role into the strength, stiffness and ductility of composite slabs. A series of
tests were carried out on solid and profiled composite slabs under restrained and
i
unrestrained conditions. The test results showed that end anchorages have a positive
effect on the ultimate strength of both solid and composite slabs irrespective of the
long term loading conditions. A three-dimensional finite element model is developed
to simulate the behaviour of composite slabs with end anchorage. A parametric study
was conducted to investigate the influence of concrete strength, shear stud strength
and profiled sheeting thickness in profiled composite slabs.
In general, the findings of the experimental studies and numerical analyses
presented in this thesis herein are expected to be contributing a further knowledge to
the continuing improvement of composite steel-concrete structures.
ii
Preface
This thesis is submitted in fulfilment of the requirements for the degree of Doctor
of Philosophy in Structural Engineering at the University of New South Wales
(UNSW), Sydney, Australia. The candidate was supervised by Professor Brian Uy
during a period from August 2010 and August 2015. The work presented herein was
carried out by the candidate in the Centre for Infrastructure Engineering and Safety,
School of Civil and Environmental Engineering, UNSW. The papers submitted or
published in internationally renowned journals and conferences are listed in the
following.
Journal Papers
Rana, M., Uy, B. and Mirza, O. (2015). “Experimental and numerical study of the
bond stress-slip relationship for post-tensioned composite slabs”, Journal of
Constructional Steel Research, 114, 362-379.
Rana, M., Uy, B. and Mirza, O. (2015). “Experimental and numerical study of end
anchorage in composite slabs”, Journal of Constructional Steel Research, 115,
372-386.
iii
Rana, M., Uy, B. and Mirza, O. (2016). “Numerical study of post-tensioned
composite slabs”, Journal of Constructional Steel Research. (in preparation)
Rana, M., Uy, B. and Hulbert, B. (2016). “Ductility of profiled composite slabs - an
experimental and numerical study”, Journal of Constructional Steel Research. (in
preparation)
Conference Papers
Rana, M., Uy, B. and Mirza, O. (2013). “A finite element study on the behaviour of
post-tensioned composite steel-concrete slabs”, Proceedings of the 10th Pacific
Structural Steel Conference (PSSC 2013) - Advancements and Achievements in
Structural Steel, Research Publishing Services, Singapore, 1058 - 1063.
Rana, M., Uy, B. and Mirza, O. (2013). “A push test study on the behaviour of post-
tensioned composite steel-concrete slabs”, Composite Construction VII 2013
Conference, Queensland, Australia. (in press)
iv
Acknowledgements
I would like to express my sincere gratitude to my supervisor, Professor Brian Uy,
Director, Centre for Infrastructure Engineering and Safety, The University of New
South Wales, Australia for his boundless support, excellent advice and continuous
guidance throughout every stage of my PhD. Special thanks go to Dr. Olivia Mirza
for her co-supervision, suggestions and patience for reviewing of draft papers. I
would also like to acknowledge Prof. Mark Bradford and Prof. Zhong Tao for
serving on my supervisory panel.
The research work related to the behaviour of post-tensioned composite slabs was
supported under Australian Research Council’s Linkage projects (LP0990190) and
by Stramit Building Products and Arup as Partner Organisations and in collaboration
with University of Sydney. I would like to acknowledge the support and help from
Dr. Gianluca Ranzi, Associate Professor, The University of Sydney in this regard.
The research work related to the behaviour of composite slabs with end anchorage
was supported by Australian Research Council (LP0882929). I would also like to
acknowledge the support and help from all staff, friends and colleagues at The
University of New South Wales and University of Western Sydney.
Finally, I would like to thank my parents, Mohammad Abdul Matin and Halima
Khatun and other family members for their supports and inspirations. I am grateful to
my wife, Farhana Jesmin and my son, Rafael Rana for their constant love,
encouragement, inspirations and understanding.
v
vi
Table of Contents
Abstract ......................................................................................................................... i
Preface ......................................................................................................................... iii
Acknowledgements ...................................................................................................... v
Table of contents ........................................................................................................ vii
List of notation .......................................................................................................... xiii
List of abbreviations ................................................................................................. xvii
List of tables .............................................................................................................. xix
List of figures ............................................................................................................ xxi
CHAPTER 1
INTRODUCTION
1.1 Introduction ...................................................................................................... 1

1.2 Background and motivation for research ......................................................... 1
1.3 Objectives and scope of thesis ......................................................................... 6
1.4 Organisation of thesis ....................................................................................... 7
1.5 Summary of chapter ......................................................................................... 9
CHAPTER 2
LITERATURE REVIEW
2.1 Introduction .................................................................................................... 11

2.2 Post-tensioned composite slabs ...................................................................... 12
2.3 Bond stress-slip relationship .......................................................................... 16
2.4 Experimental studies of post-tensioned composite slabs ............................... 19
2.5 Push test study ................................................................................................ 21
2.6 Numerical modelling of post-tensioned composite slabs............................... 27
vii
2.7 Behaviour of composite slabs with end anchorage ........................................ 33
2.8 Numerical modelling of composite slabs with end anchorage ....................... 37
2.9 Summary of chapter ....................................................................................... 38
CHAPTER 3
PUSH TESTS
3.1 Introduction .................................................................................................... 47

3.2 Test specimens ............................................................................................... 47
3.3 Preliminary series ........................................................................................... 51
3.3.1 Material properties ................................................................................... 51
3.3.2 Specimen preparation and post-tensioning .............................................. 52
3.3.3 Instrumentation ........................................................................................ 53
3.3.4 Test setup ................................................................................................. 53
3.3.5 Summary of preliminary series ................................................................ 54
3.3.5.1 General.............................................................................................. 54
3.3.5.2 Push test specimens PT-500-0 and PT-500-3 ................................... 55
3.3.5.3 Push test specimens PT-1000-0 and PT-1000-3 ............................... 56
3.3.5.4 Push test specimens PT-2000-0 and PT-2000-3 ............................... 56
3.3.5.5 Discussion of preliminary test results ............................................... 57
3.4 Main series ..................................................................................................... 58
3.4.1 Material properties ................................................................................... 58
3.4.2 Specimen preparation, post-tensioning and instrumentation ................... 58
3.4.3 Test setup ................................................................................................. 59
3.4.4 Summary of main series........................................................................... 60
3.4.4.1 General.............................................................................................. 60
3.4.4.2 Push test specimens of 500 mm length ............................................. 60
3.4.4.3 Push test specimens of 1000 mm length ........................................... 62
3.4.5 Analysis of test results ............................................................................. 67
viii
3.4.5.1 Effects of prestress ........................................................................... 68
3.4.5.2 Effects of bond length ...................................................................... 69
3.5 Design guidance ............................................................................................. 69
3.6 Summary of chapter ....................................................................................... 71
CHAPTER 4
FINITE ELEMENT MODELLING OF PUSH TESTS
4.1 Introduction .................................................................................................. 109

4.2 Finite element model formulation ................................................................ 110
4.2.1 General ................................................................................................... 110
4.2.2 Model geometry and element types ....................................................... 111
4.2.3 Interactions............................................................................................. 111
4.2.4 Loading and boundary conditions.......................................................... 112
4.2.5 Solution method ..................................................................................... 113
4.2.6 Constitutive laws.................................................................................... 114
4.2.6.1 Concrete properties......................................................................... 114
4.2.6.2 Profiled steel sheeting properties ................................................... 116
4.2.6.3 Local bond-slip law ........................................................................ 117
4.3 Finite element analysis ................................................................................. 117
4.3.1 Mesh sensitivity analysis ....................................................................... 117
4.3.2 FE model validation ............................................................................... 118
4.3.3 Parametric study .................................................................................... 120
4.4 Summary of chapter ..................................................................................... 121
CHAPTER 5
FINITE ELEMENT MODELLING OF POST-TENSIONED COMPOSITE

SLABS
5.1 Introduction .................................................................................................. 141
ix
5.2 Description of experimental investigation ................................................... 142
5.3.1 General ................................................................................................... 143
5.3.3 Interactions ............................................................................................. 144
5.3.4 Loading and boundary conditions .......................................................... 145
5.3.5 Solution method ..................................................................................... 145
5.3.6 Constitutive laws .................................................................................... 146
5.3.6.1 Concrete properties ......................................................................... 146
5.3.6.2 Profiled steel sheeting properties .................................................... 146
5.3.6.3 Reinforcing steel and prestressing steel properties......................... 147
5.4.3 Parametric study..................................................................................... 151
5.4.3.1 Effects of concrete compressive strength ....................................... 152
5.4.3.2 Effects of profiled steel sheeting thickness .................................... 152
5.4.3.3 Effects of level of prestress ............................................................ 153
5.5 Ultimate capacity of post-tensioned composite slabs .................................. 154
CHAPTER 6
EXPERIMENTAL PROGRAMME OF COMPOSITE SLABS WITH END

ANCHORAGE
6.1 Introduction .................................................................................................. 175

6.2 Experimental programme ............................................................................. 175
6.2.1 Test specimens ....................................................................................... 175
6.2.2 Specimen preparation............................................................................. 177
6.2.3 Material properties ................................................................................. 178
x
6.2.4 Instrumentation ...................................................................................... 178
6.2.5 Test setup ............................................................................................... 179
6.3 Results and discussion ................................................................................. 179
6.3.1 General ................................................................................................... 179
6.3.2 Test observations ................................................................................... 180
6.3.2.1 Solid unrestrained slabs .................................................................. 180
6.3.2.2 Profiled unrestrained slabs ............................................................. 180
6.3.2.3 Solid restrained slabs ...................................................................... 181
6.3.2.4 Profiled restrained slabs ................................................................. 182
6.3.3 Effects of creep and shrinkage ............................................................... 183
6.3.4 Effects of end anchorage........................................................................ 184
6.3.5 Effects of profiled steel sheeting ........................................................... 185
CHAPTER 7
FINITE ELEMENT MODELLING OF COMPOSITE SLABS WITH END

ANCHORAGE
7.1 Introduction .................................................................................................. 205

7.2.1 General ................................................................................................... 205
7.2.3 Interactions............................................................................................. 206
7.2.4 Loading and boundary conditions.......................................................... 208
7.2.5 Solution method ..................................................................................... 208
7.2.6 Constitutive laws.................................................................................... 209
7.2.6.1 Concrete properties......................................................................... 209
7.2.6.2 Steel properties ............................................................................... 210
xi
7.3.3 Parametric study..................................................................................... 214
7.3.3.1 Effects of concrete compressive strength ....................................... 215
7.3.3.2 Effects of profiled steel sheeting thickness .................................... 215
7.3.3.3 Effects of headed shear stud strength ............................................. 216
CHAPTER 8
CONCLUSIONS
8.1 Concluding remarks ..................................................................................... 233

8.2 Recommendations for further research ........................................................ 236
REFERENCES ......................................................................................................... 239

APPENDIX A .......................................................................................................... 251
APPENDIX B .......................................................................................................... 254
APPENDIX C .......................................................................................................... 262
xii
List of Notation
Ac cross-sectional area of concrete
Ap cross-sectional area of prestressing steel
As cross-sectional area of profiled steel sheeting
b width of slab
C concrete compressive force
da maximum aggregate size
dn depth of neutral axis
dp depth of tendons
ds depth of profiled steel sheeting
e plastic flow potential eccentricity
Ec elastic modulus of concrete
eck eccentricity of post-tensioned composite slab
Ep elastic modulus of tendon
Es elastic modulus of steel
ess eccentricity of post-tensioned solid slab
fbo equibiaxial concrete strength
fbu maximum average bond stress in push test specimen
f'c concrete compressive strength
fct tensile strength of concrete
fpy yield strength of tendons
fu ultimate strength
fy yield stress of steel
xiii
Gf fracture energy
H height of concrete block
Hr mechanical resistance of profiled steel sheeting
I moment of inertia
Kc ratio of the second stress invariant on the tensile meridian to that on the
compressive meridian
l slab span of composite slab
L length of push test specimen
Mdeck flexural capacity of steel deck
Mfsc flexural capacity of composite section for full shear connection
Mosc flexural capacity of composite section for zero shear connection
Mp plastic moment capacity
Mpsc flexural capacity of composite section for partial shear connection
Mu ultimate moment capacity
P level of prestress
Pe effective prestress
t slab thickness
T resultant tensile force in steel
tp profiled steel sheeting thickness
w self-weight of slab
W width of push test specimen
γ neutral axis parameter
ƞ degree of shear connection
µ friction co-efficient
xiv
εc strain in concrete
ε'c strain corresponding to f'c
εce strain in the concrete due to effective prestress at the level of the
prestressing steel
εcu ultimate strain in concrete at extreme compression fibre
εpe initial strain in the tendons due to effective prestress
εpt ultimate strain in the concrete at the level of the prestressing steel
εpu ultimate strain in prestressing steel
εpy yield strain of prestressing tendon
εs strain of profiled steel sheeting
εy yield strain of profiled steel sheeting
σbu bearing stress at the base of profiled steel sheeting
σc compressive stress in concrete
σpu stress at tendons at ultimate
σsu stress at profiled steel sheeting at ultimate
τ average bond stress
xv
xvi
List of Abbreviations
ALLIE total internal energy
ALLKE total kinematic energy
AS Australian Standard
CK post-tensioned composite slab with Condeck profile
COV coefficients of variation
DEMEC demountable mechanical
FE finite element
FEA finite element analysis
FEM finite element modelling
FSC full shear connection
LSCT linear strain conversion transducers
MPC multi-point constraints
PSC partial shear connection
PT push test specimen
PTC post-tensioned composite
SS post-tensioned solid slab
xvii
xviii
List of Tables
Table 3.1 Details of push specimens of preliminary series ....................................... 73
Table 3.2 Details of concrete mix of preliminary series ........................................... 73
Table 3.3 Material properties of concrete, profiled steel sheets and prestressing steel
of preliminary series ................................................................................................... 74
Table 3.4 Results of concrete compressive strength test for preliminary series ....... 74
Table 3.5 Summary of bond stress of push specimens of preliminary series............ 75
Table 3.6 Details of push test specimens of main series ........................................... 75
Table 3.7 Details of concrete mix of main series ...................................................... 76
of main series ............................................................................................................. 76
Table 3.9 Results of concrete compressive strength test for main series .................. 77
Table 3.10 Test results of main series ....................................................................... 78
Table 4.1 Concrete properties used in FE model .................................................... 123
Table 4.2 Steel properties used in FE model ........................................................... 123
Table 4.3 Comparison of numerical results with experimental results ................... 124
Table 5.1 Concrete properties used in FE model .................................................... 160
Table 5.2 Material properties of concrete, profiled steel sheets, prestressing steel and
reinforcing steel ........................................................................................................ 160
Table 5.3 Steel properties used in FE model ........................................................... 161
Table 5.4 Comparison of ultimate moment capacities ............................................ 161
Table 6.1 Details of test specimens ......................................................................... 187
xix
Table 6.2 Material properties of concrete, profiled steel sheets, reinforcing steel and
headed shear studs .................................................................................................... 187
Table 6.3 Test results ............................................................................................... 188
Table 7.1 Properties used in FE model .................................................................... 219
Table 7.2 Comparison of numerical results with experimental results ................... 219
Table A.1 Material properties of profiled steel sheets of push tests........................ 251
Table A.2 Material properties of prestressing bar of push tests .............................. 251
Table C.1 Material properties of N12 Reinforcing steel bars ................................. 262
Table C.2 Material properties of headed stud shear connectors.............................. 262
Table C.3 Material properties of profiled steel decking - W-DEK ......................... 263
xx
List of Figures
Figure 1.1 Typical post-tensioned composite slab (Ranzi et al., 2013a) .................. 10
Figure 1.2 Typical composite steel-concrete slab (Ganesh et al., 2005)................... 10
Figure 2.1 Modes of failure of composite slab (Johnson, 1994) ............................... 40
Figure 2.2 Different types of profiled steel sheeting ................................................. 40
Figure 2.3 Push out test (Schuster, 1970) ................................................................. 40
Figure 2.4 Push out test (Stark, 1978) ....................................................................... 41
Figure 2.5 Push off test (Jolly and Zubair, 1987) ..................................................... 41
Figure 2.6 Pull out test (Daniels, 1988) .................................................................... 42
Figure 2.7 Slip block test (Patrick and Poh, 1990) ................................................... 42
Figure 2.8 Block bending test (An, 1993) ................................................................. 43
Figure 2.9 Push test and tension push test (Veljkovic, 1996) ................................... 43
Figure 2.10 Push test (Burnet, 1998) ........................................................................ 44
Figure 2.11 Push test (Tremblay, 2002) .................................................................... 44
Figure 2.12 Small scale test setup (Abdullah, 2004)................................................. 45
Figure 3.1 Geometry of the push specimen............................................................... 79
Figure 3.2 Details of end plate, channels and support block..................................... 80
Figure 3.3 Geometry of Stramit Condeck HP® ......................................................... 81
Figure 3.4 Formwork ready for pouring (a) without post-tensioning (b) with post-
tensioning. .................................................................................................................. 81
Figure 3.5 Post-tensioning ........................................................................................ 82
Figure 3.6 Load-strain curve for PT-500-3 ............................................................... 82
Figure 3.7 Instrumentation of push tests ................................................................... 83
xxi
Figure 3.8 Linear potentiometers for measurement of slip and separation ............... 83
Figure 3.9 Push test setup .......................................................................................... 84
Figure 3.10 Loading plate ......................................................................................... 84
Figure 3.11 Push test specimen ready for testing ...................................................... 85
Figure 3.12 Bond stress-slip curves of PT-500-0 & PT-500-3 ................................. 85
Figure 3.13 Initiation of slip from one side ............................................................... 86
Figure 3.14 Bond stress-slip curves of PT-1000-0 & PT-1000-3 ............................. 86
Figure 3.15 Local buckling of the profiled steel sheeting ......................................... 87
Figure 3.16 Bond stress-slip curves of PT-2000-0 & PT-2000-3 ............................. 87
Figure 3.17 Linear potentiometer dropped ................................................................ 88
Figure 3.18 Comparison of bond stress-slip behaviour of push specimens .............. 88
Figure 3.19 Bond stress-slip curves of PT-500-0 ...................................................... 89
Figure 3.20 Predelamination of PT-500-0-A ............................................................ 89
Figure 3.21 Failure of PT-500-0-A through uneven slip........................................... 90
Figure 3.22 Failure mode of PT-500-0-B .................................................................. 90
Figure 3.24 Bond stress-slip curves of PT-500-2.4 ................................................... 91
Figure 3.25 Bond stress-slip curves of both sides of PT-500-2.4-A ......................... 92
Figure 3.27 Bond stress-slip curves of PT-1000-0 .................................................... 93
Figure 3.28 Failure of PT-1000-0-B through buckling ............................................. 93
Figure 3.30 Failure pattern of PT-1000-2.4-A .......................................................... 94
Figure 3.31 Bond stress-slip curves of PT-1000-2.4 ................................................. 95
xxii
Figure 3.33 Corner distortion of the profiled steel sheeting ..................................... 96
Figure 3.35 Buckling of PT-1500-0-B ...................................................................... 97
Figure 3.37 Bond stress-slip curves of PT-1500-2.4................................................. 98
Figure 3.38 Buckling of PT-1500-2.4-B ................................................................... 98
Figure 3.40 Buckling of PT-1500-3-B ...................................................................... 99
Figure 3.41 Bond stress-slip curves of PT-2000-0 .................................................. 100
Figure 3.42 Buckling of PT-2000-0-A .................................................................... 100
Figure 3.44 Buckling and bond failure of PT-2000-1-A ......................................... 101
Figure 3.45 Buckling and bond failure of PT-2000-1-B ......................................... 102
Figure 3.46 Bond stress-slip curves of PT-2000-2.4............................................... 102
Figure 3.47 Buckling pattern observed after the test .............................................. 103
Figure 3.49 Buckling of PT-2000-3-A .................................................................... 104
Figure 3.50 Buckling of PT-2000-3-B .................................................................... 104
Figure 3.51 Variation in chemical bond stress due to prestress .............................. 105
Figure 3.52 Variation in mechanical bond stress due to prestress .......................... 105
Figure 3.53 Variation in bond stress at 5.0 mm due to prestress ............................ 106
Figure 3.54 Variation in chemical bond stress with bond length ............................ 106
Figure 3.55 Variation in mechanical bond stress with bond length ........................ 107
xxiii
Figure 3.56 Variation in bond stress at 5.0 mm with bond length .......................... 107
Figure 4.1 Overview of the finite element mesh ..................................................... 125
Figure 4.2 Radial thrust connector at steel-concrete interface ................................ 125
Figure 4.3 Energy level for analysis ........................................................................ 126
Figure 4.4 Stress-strain curve of concrete ............................................................... 126
Figure 4.5 Stress-strain curve of profiled steel sheeting ......................................... 127
Figure 4.6 Local bond-slip law................................................................................ 127
Figure 4.7 FE model for mesh sensitivity analysis ................................................. 128
Figure 4.8 Mesh sensitivity analysis ....................................................................... 128
Figure 4.9 Push specimen of height 500 mm without prestress .............................. 129
Figure 4.10 Push specimen of height 500 mm with prestress of 1 MPa ................. 129
Figure 4.11 Push specimen of height 500 mm with prestress of 2.4 MPa .............. 130
Figure 4.12 Push specimen of height 500 mm with prestress of 3 MPa ................. 130
Figure 4.13 Push specimen of height 1000 mm without prestress .......................... 131
Figure 4.14 Push specimen of height 1000 mm with prestress of 1 MPa ............... 131
Figure 4.15 Push specimen of height 1000 mm with prestress of 2.4 MPa ............ 132
xxiv
Figure 4.25 Re-entrant profile (a) Bond stress-slip curve (b) Stress distribution ... 137
Figure 4.26 Trapezoidal profile (a) Bond stress-slip curve (b) Stress distribution . 138
Figure 4.27 Bondek profile (a) Bond stress-slip curve (b) Stress distribution ........ 139
Figure 5.1 Ultimate strength test setup (Ranzi et al., 2013) .................................... 162
Figure 5.2 Overview of the finite element mesh ..................................................... 162
Figure 5.3 Radial thrust connector at steel-concrete interface ................................ 163
Figure 5.4 Energy level for analysis........................................................................ 163
Figure 5.6 Stress-strain curve of profiled steel sheets ............................................. 164
Figure 5.7 Stress-strain curve of reinforcing bar..................................................... 165
Figure 5.8 Stress-strain curve of prestressing bar ................................................... 165
Figure 5.9 Local bond stress-slip law ..................................................................... 166
Figure 5.10 Mesh sensitivity analysis ..................................................................... 166
Figure 5.11 Load-deflection curve of post-tensioned solid slab ............................. 167
Figure 5.12 Failure mode of post-tensioned solid slab ........................................... 167
Figure 5.13 Yielding of post-tensioning steel strand of post-tensioned solid slab . 168
Figure 5.14 Load-deflection curve of post-tensioned composite slab .................... 168
Figure 5.15 Shear bond failure of post-tensioned composite slab .......................... 169
Figure 5.16 Failure mode of post-tensioned composite slab ................................... 169
Figure 5.17 Yielding of post-tensioning steel strand of PTC slab .......................... 170
Figure 5.18 Effects of concrete compressive strength (𝑓𝑐 ′)for PTC slab .............. 170
Figure 5.19 Effects of profiled sheeting thickness ((tp) for PTC slab .................... 171
Figure 5.20 Effects of level of prestress (P) for PTC slab ...................................... 171
xxv
Figure 5.21 Strain and stress distribution at the ultimate limit state (Ranzi et al.,
2013a) ....................................................................................................................... 172
Figure 5.22 Forces exerted on a concrete beam by a parabolic tendon (Gilbert and
Mickleborough, 2004) .............................................................................................. 172
Figure 5.23 Strain distribution at three stages of loading (Gilbert and
Mickleborough, 2004) .............................................................................................. 173
Figure 5.24 Variation of flexural capacity with degree of shear connection .......... 173
Figure 6.1 Details of test specimens ........................................................................ 189
Figure 6.2 Geometry of W-DEK profiled steel sheeting......................................... 190
Figure 6.3 Instrumentation ...................................................................................... 190
Figure 6.4 Specimen ready for testing..................................................................... 190
Figure 6.5 Load-deflection curves of LTSU-SW & LTSU-SL ............................... 191
Figure 6.6 Flexural cracking of LTSU-SW ............................................................. 191
Figure 6.7 Cracking across the width of the concrete slab of LTSU-SL ................ 192
Figure 6.8 Flexural cracking at ultimate load of LTSU-SL .................................... 192
Figure 6.9 Load-deflection curves of LTPU-SW & LTPU-SL ............................... 193
Figure 6.10 Overall crack pattern at ultimate load of LTPU-SL............................. 193
Figure 6.11 Shear bond failure of LTPU-SL ........................................................... 194
Figure 6.12 Load-deflection curves of LTSR-SW & LTSR-SL ............................. 194
Figure 6.13 Cracking around studs of LTSR-SL .................................................... 195
Figure 6.14 Crack pattern at ultimate load of LTSR-SL ......................................... 195
Figure 6.15 Load-deflection curves of LTPR-SW & LTPR-SL ............................. 196
Figure 6.16 Shear bond failure of LTPR-SW.......................................................... 196
Figure 6.17 Weld failure of a corner stud of LTPR-SL .......................................... 197
xxvi
Figure 6.18 Crack pattern of LTPR-SL................................................................... 197
Figure 6.19 Comparison between behaviour of profiled and solid restrained slabs 198
Figure 6.20 Effects of creep and shrinkage (solid unrestrained slabs).................... 198
Figure 6.21 Effects of creep and shrinkage (solid restrained slabs)........................ 199
Figure 6.22 Effects of creep and shrinkage (profiled unrestrained slabs) ............... 199
Figure 6.23 Effects of creep and shrinkage (profiled restrained slabs) ................... 200
Figure 6.24 Effects of end anchorage (LTPR & LTPU) ......................................... 200
Figure 6.25 Effects of end anchorage (LTSR & LTSU) ......................................... 201
Figure 6.26: Deflection of unrestrained slabs ......................................................... 201
Figure 6.27: Deflection of restrained slabs ............................................................. 202
Figure 6.28 Effects of composite action (LTPU & LTSU) ..................................... 202
Figure 6.29 Effects of composite action (LTPR & LTSR) ..................................... 203
Figure 7.1 Overview of FE model ........................................................................... 220
Figure 7.2 Energy level of LTPR-SW ..................................................................... 220
Figure 7.4 Stress-strain curve of profiled steel sheeting ......................................... 221
Figure 7.5 Stress-strain curve of steel reinforcing bar ............................................ 222
Figure 7.6 Stress-strain curve of headed shear studs .............................................. 222
Figure 7.7 Local bond-slip law (Burnet, 1998) ....................................................... 223
Figure 7.8 Effects of concrete element size ............................................................ 223
Figure 7.9 Effects of headed shear stud element size ............................................. 224
Figure 7.10 Load-deflection curves of solid unrestrained slab ............................... 224
Figure 7.11 Load-deflection curves of solid restrained slab ................................... 225
Figure 7.12 Load-deflection curves of profiled unrestrained slab .......................... 225
xxvii
Figure 7.13 Load-deflection curves of profiled restrained slab .............................. 226
Figure 7.14 Reinforcement yielding of LTSR-FE (concrete slab removed) ........... 226
Figure 7.15 Local buckling of sheeting near the load point of LTPU-FE............... 227
Figure 7.16 Local buckling of sheeting near ultimate load of LTPU-SL................ 227
Figure 7.17 Tensile rupture near stud area of LTPR-FE ......................................... 228
Figure 7.18 Stud deformation in LTPR-FE ............................................................. 229
Figure 7.19 Effects of concrete compressive strength (𝑓𝑐 ′) ................................... 230
Figure 7.20 Effects of profiled steel sheeting thickness (tp).................................... 230
Figure 7.21 Effects of headed shear stud strength (fy)............................................. 231
Figure A.1 Tensile test results of profiled steel sheets ............................................ 252
Figure B.1 Geometry of post-tensioned composite slab (Ranzi et al., 2013a) ........ 255
Figure B.2 Stress-strain distribution of post-tensioned solid slab ........................... 255
Figure B.3 Determination of ultimate moment capacity ......................................... 257
Figure B.4 Stress-strain distribution of post-tensioned composite slab (Condeck) 257
Figure B.5 Mu vs. ƞ relationship (Condeck) ............................................................ 260
Figure C.1 Load-slip curves of LTPU-SW ............................................................. 263
Figure C.2 Load-slip curves of LTPU-SL............................................................... 264
Figure C.3 Load-slip curves of LTPR-SW ............................................................. 264
Figure C.4 Load-slip curves of LTPR-SL ............................................................... 265
xxviii
CHAPTER 1- Introduction
CHAPTER 1
INTRODUCTION
1.1 INTRODUCTION
This chapter presents an overview of the research work carried out in the present
study and outlines its significance and contributions. Background knowledge and
motivation to conduct experimental tests and numerical analyses to study the
behaviour of post-tensioned composite (PTC) steel-concrete slabs and composite
slabs with end anchorage are described herein. The objectives and scope of the thesis
are also defined in this chapter, followed by a brief description of each chapter in the
thesis.
1.2 BACKGROUND AND MOTIVATION FOR RESEARCH
A composite slab is defined by Eurocode 4 (British Standards Institution, 2004) as
“a slab in which profiled steel sheets are used initially as permanent shuttering and
subsequently combine structurally with the hardened concrete and act as tensile
reinforcement in the finished floor.”
1
The concept of composite steel-concrete slabs originated in the 1930s to substitute
the traditional reinforced concrete slabs (Crisinel and O’Leary, 1996). But the
application of the composite slabs was first known to have been started in the 1950s
when steel mesh reinforcement was welded to the profiled steel sheeting, which had
twice the yield stress of mild steel. In general, profiled steel sheeting utilises most or
all of its tensile strength, acting as main tensile reinforcement in composite slabs.
The first major research study on the behaviour and performance of composite slabs
took place in the late 1960s. Schuster (1970) and Porter et al. (1976) published a
number of articles and recommendations which recognised the major features of
composite slabs.
Composite slabs reduce dead loads and construction time, being both of major
interests in high rise steel framed floor construction. The introduction of fast-track
construction methods during the late 1980s has increased interest in steel design and
consequently in composite flooring. Nowadays, composite systems are being
integrated with reinforced concrete, prestressed concrete and timber structures and
used for a wide range of construction projects (Crisinel and O'Leary, 2004). The
combination of post-tensioning and composite action between the profiled steel
sheets and concrete in a single building floor system reduces the concrete cracking
significantly and thereby provides enhanced flexural strength and stiffness of PTC
slabs compared with post-tensioned concrete slabs and conventional composite slabs.
Therefore, a PTC slab represents a very popular form of building floor system at
present due to the economic and technical advantages of using profiled steel sheets as
an alternative to conventional timber formwork systems and reinforcement in post-
tensioned concrete structures. It consists of concrete slab, profiled steel sheets, post-
2
tensioning strands and non-prestressed reinforcement (Figure 1.1). In this system,
profiled steel sheets act as permanent formwork during construction and external
reinforcement during service.
The bond between the concrete and profiled steel sheets is usually assumed to
have a significant effect on the strength, stiffness and ductility of composite slabs. A
significant step toward understanding the bond behaviour is to generate a bond
stress-slip relationship. The horizontal shear bond interaction between the profiled
steel sheets and concrete is usually determined through push tests. Push tests provide
a less expensive and convenient method to investigate the complex interaction
between the profiled steel sheets and concrete. The bond stress-slip relationship
obtained from a push test may lead to a better understanding of the bond behaviour
by taking into account the effect of parameters ignored by existing design methods.
However, the shear bond stress versus slip relationship obtained from these tests
need to be interpreted with a deep understanding of the fundamental mechanisms.
The objective of many research studies on composite slabs was directed towards the
development of a simple, efficient and economical small scale tests for composite
slabs, which is capable to provide the parameters needed for all design and analysis
methods. In order to reduce the dependency on the full scale bending test which is
required by the current design specifications, many researchers (Schuster, 1970;
Stark, 1978; Jolly and Jubair, 1987; Daniels, 1988; Patrick and Poh, 1990; An, 1993;
Veljkovic, 1996; Burnet, 1998, Tremblay, 2002; Abdullah, 2004) developed small
scale tests. One of the main objectives of this research is to determine the potential
influence of prestress on the bond between the profiled steel sheets and concrete.
3
Due to complexity of horizontal shear transfer between the steel deck and the
concrete in composite slabs, it is very difficult to develop an analytical method that is
suitable for all slab conditions without using data from tests. Therefore, researchers
inclined towards the complicated finite element modelling (FEM) of composite slabs.
Again, the shear connection property obtained from the elemental tests was the main
hindrance to produce satisfactory results in the finite element analysis (FEA) and
modification was proposed by the researchers (Veljkovic, 2004; Abdullah, 2004).
Hence, any analytical method that is derived from finite element analysis using this
shear connection property should be used with caution.
Limited research studies have looked into the investigation of the behaviour of
PTC slabs (Schravendeel et al., 1990; Koukkari, 1999; Koukkari and Eligehausen,
2001; TEKES, 2001; Bailey et al., 2006; Miller and Bailey 2007; Patrick and Lloyde,
2008; Ranzi et al., 2013a & b). Experimental data for post-tensioned composite slabs
are limited that they are not enough to understand the behaviour of these types of
slabs and produce design guidelines. Since different factors affect the behaviour of
this form of slabs, analytical and numerical studies are also very limited.
Consequently, at present there is no code which presents a united method for analysis
of post-tensioned composite slabs. Therefore, the contribution of profiled steel sheet
is usually ignored in the design. To the knowledge of the author of this thesis, the
objective of all previous studies was limited into the evaluation of the flexural
strength, ultimate and long term behaviour of post-tensioned composite slabs. No
experimental test was published in the literature to investigate how the prestress
influences the shear bond interaction of post-tensioned composite slabs. In the
context of limited experimental studies, this thesis intends to contribute to this
4
deficiency by providing experimental data to study the shear bond behaviour of post-
tensioned composite slabs accurately. An accurate and efficient finite element model
is also presented in this thesis to study the strength and behaviour of post-tensioned
composite slabs.
The type of end anchorage has a significant effect on the strength, stiffness and
ductility of the composite slabs. The end anchorages by through-deck welded shear
studs are the most common and popular forms of effective shear connection in
composite slabs. Initiation of shear bond-slip at the steel-concrete interface leads to a
loss in composite action and thereby affects the structural response of these types of
slabs significantly. This thesis investigates the strength and behaviour of composite
slabs with end anchorage provided by through-deck welded studs.
Several studies have looked into the effect of end anchorage on the strength and
behaviour of composite slabs experimentally (Porter and Greimann, 1984; Easterling
and Young, 1990; Jolly and Lawson, 1992; Daniels and Crisinel, 1993a & b; Chen,
2003, Mirza and Uy, 2013). Some analytical studies were published to predict the
ultimate strength of composite slabs with studs used for end anchorage (Degtyarev,
2013a & b). Daniels and Crisinel (1993b) presented a nonlinear finite element model
of composite slabs with end anchorage. However, the capability of the model to
simulate the behaviour of profiled composite slabs is deemed to be limited and it is
possible to model the behaviour more efficiently and accurately due to availability of
sophisticated finite element software packages. In this context, this research presents
a series of ultimate strength tests aimed at evaluating the effect of end anchorages on
the load carrying capacities and failure modes of both solid and profiled composite
slabs. An accurate and efficient nonlinear finite element model of the composite
5
slabs is also presented to investigate the behaviour of profiled composite slabs with
end anchorage using a commercially available software ABAQUS. The model was
validated against experimental results for its accuracy and a parametric study was
conducted to investigate the influence of concrete strength, shear stud strength and
profiled sheeting thickness in profiled composite slabs.
1.3 OBJECTIVES AND SCOPE OF THESIS
The research work carried out in this thesis is mainly divided into experimental
and numerical studies. The main objectives of this research are outlined as follows:
 To investigate the effects of prestress on the bond stress-slip behaviour of post-
tensioned composite steel-concrete slabs.
 To develop an accurate nonlinear 3D finite element model of shear bond
behaviour of a push specimen considering possible influence of prestress on
the bond between the concrete and the profiled steel sheets.
 To assess the bond stress-slip relationship obtained from push test study on
post-tensioned composite slabs through a finite element model and to conduct a
parametric study to investigate important parameters affecting the behaviour of
post-tensioned composite slabs.
 To investigate the effect of end anchorage on the ultimate load carrying
capacities and failure modes of both solid and profiled slabs through an
experimental study.
6
 To develop an accurate nonlinear 3D finite element model of composite
profiled slab with end anchorage and to conduct a parametric study to
investigate important parameters affecting the behaviour of composite slabs
with end anchorage.
1.4 ORGANISATION OF THESIS
The layout of the thesis is comprised of eight chapters. An overview of the
contents of each chapter has been given with an introduction at the beginning of the
corresponding chapter. At the end of each chapter, the key findings and contributions
which are derived from this research are summarised.
In Chapter 1, the background of this research work is presented and then the
objectives and scope of this research are described. A state-of-the-art literature
review on the existing research work which is relevant to the areas of interest in this
thesis is presented in Chapter 2. This includes the results and important findings
published in the open literature based on various experimental, analytical and
numerical studies. The discussion covers primarily the issue of bond stress-slip
behaviour of composite slabs, the existing experimental and numerical studies on the
behaviour of post-tensioned composite slabs and the behaviour of composite slabs
with end anchorage.
Chapter 3 presents a push test study comprising of two series, namely the
preliminary series and main series. The preliminary series of six test specimens and
main series of thirty two specimens are reported in this chapter along with the details
7
of the test specimens, specimen preparation, material properties, instrumentation, test
setup and loading condition. Detailed results and observations for each of the
specimens during the experiments are presented in this chapter. In particular, the
effects of varying level of prestress and bond length are appraised.
Chapter 4 presents the numerical modelling of the push test specimens. The
formulation of a nonlinear three dimensional finite element model using commercial
software package ABAQUS is presented in this chapter. These include model
geometry and element types, interactions, loading and boundary conditions and
solution method. The constitutive laws of materials to be used in the models are
illustrated and the details of the material properties are also provided. A mesh
sensitivity analysis is presented for the accuracy of the finite element models
developed herein.
Chapter 5 illustrates the numerical modelling of the post-tensioned composite
slabs. The finite element model formulation, mesh sensitivity analysis and validation
are described in detail. A parametric study is presented to investigate the effect of
concrete strength, level of prestress and profiled sheeting thickness. A calculation
procedure of ultimate strength of post-tensioned composite slabs is also described in
this chapter.
Chapter 6 explains the experimental programme of composite slabs to investigate
the effects of end anchorage. A detailed description of all test specimens, specimen
preparation, material properties, instrumentation, test setup and loading condition are
provided. Main results and observations for each of the specimens during the
8
experiments are presented in this chapter. In particular, the effects of end anchorage,
composite action and long term loading effects are appraised.
The finite element modelling of the composite slabs with end anchorage are
described in Chapter 7. The finite element model formulation, mesh sensitivity
analysis and validation are described in detail. A parametric study is presented to
investigate the effect of concrete strength, shear stud strength and profiled sheeting
thickness.
Finally, conclusions drawn from the present experimental and numerical study are
summarised in Chapter 8. Recommendations and suggestions for possible future
research work on composite steel-concrete slabs are also provided.
1.5 SUMMARY OF CHAPTER
This chapter has provided an overview of the research work to be presented in this
thesis. The general layout and contents of each chapter of the thesis have also been
illustrated. The scope and objectives of this thesis are defined to understand the
limitations and novelty of this research work.
9
Figure 1.1 Typical post-tensioned composite slab (Ranzi et al., 2013a)
Figure 1.2 Typical composite steel-concrete slab (Ganesh et al., 2005)
10
CHAPTER 2- Literature Review
CHAPTER 2
LITERATURE REVIEW
2.1 INTRODUCTION
This chapter presents an overview of the existing research work published in open
literature which is relevant to the objectives and scope of this thesis. A summary of
the important findings and developments from previous studies is provided herein to
find out existing gaps in current knowledge which led to this study. The research
work on composite steel-concrete slabs and post-tensioned concrete structures is
reviewed to the extent that is necessary to understand the methodology of the current
study presented herein in case of post-tensioned composite slabs.
In this chapter, the main issues on the bond stress-slip relations of post-tensioned
composite slabs are discussed. Existing research work that has been undertaken to
investigate the bond stress-slip behaviour of composite steel-concrete slabs is
reviewed herein. The discussion also covers the overall behaviour of the PTC slabs
that is known from previous studies. As part of the experimental programme
presented in this thesis, a series of push tests were conducted and therefore the
available literature on push tests is reviewed. Since the composite slabs with end
anchorage encompass a significant part of this thesis, the existing research work on
11
the effects of various forms of end anchorages of composite slabs is also reviewed.
This thesis is aimed to develop an accurate and efficient numerical model of post-
tensioned composite slabs and composite slab with end anchorage. Therefore, the
available FE approaches are studied and taken into account to develop a three
dimensional nonlinear finite element model. The challenging issues of numerical
modelling are identified and discussed in this chapter.
2.2 POST-TENSIONED COMPOSITE SLABS
Post-tensioned composite steel-concrete slabs represent an attractive and novel
form of floor construction in the building industry. It is capable of combining the
advantages of the two most commonly used flooring systems, which consist of post-
tensioned slabs in the case of concrete structures and of composite slabs for steel
structures. The advantages of this novel system cannot be exploited as yet, as no
design guidelines are currently available. The advantages and behaviour of both
composite slabs and post-tensioned concrete slabs are described herein to understand
the benefits and features of PTC slabs.
Composite slabs have following advantages (Wright et al, 1987):
 The profiled steel sheeting provides a platform to support construction load and
a safe surface for the construction workers.
 The profiled sheeting acts as the permanent form for the in-situ cast concrete
slab and thus saves the time and cost of removable forms.
12
 The profiled sheeting acts as the external tensile reinforcement and thus saves
the time and cost for placing and fixing of reinforcing bars for concrete slab.
 The geometry of the profiled steel sheeting can be made effective in reducing
concrete volume and thereby reducing total dead loads.
 Services can be installed and distributed within the depth of the floor slab
thanks to the geometry of the profiled steel sheeting.
 The profiled steel sheeting is easy to stack, transport and handle.
 In general, profiled composite slabs reduce the construction time significantly
and thus increase the economy of the construction considerably.
Composite slabs under bending can exhibit three major modes of failure: flexural
failure at section 1-1, vertical shear failure at section 2-2 and longitudinal shear
failure at section 3-3, as shown in Figure 2.1 (Johnson, 1994).
The flexural failure occurs usually in long thin slabs and when complete
interaction at the interface between the concrete and the steel is achieved. However,
this is not a governing design criterion because the steel and concrete interaction is
normally incomplete and the slab length is invariably limited by the deflection limit
(Easterling and Young, 1992).
The vertical shear failure occurs usually in very short and thick slabs with high
concentrated load acting near the supports (Patrick and Bridge, 1992). This is not so
common in construction practice and the effect is normally ignored in design.
13
The horizontal shear failure or shear bond failure is the most common mode of
failure for composite slabs under vertical loads (Porter and Ekberg, 1978). The
development of an approximate diagonal crack under or near one of the concentrated
loads just before failure and consequently an end-slip found between the concrete
and the profiled steel sheeting within the shear span are the main features of this type
of failure.
Post-tensioning is a technique of prestressing the concrete in a manner which
eliminates or reduces the tensile stresses induced by the dead and live loads. Due to
its economic and technical advantages, the proportion of concrete floors being post-
tensioned is growing (Khan and Williams, 1995). Both post-tensioning systems,
bonded and unbonded, provide safe and efficient methods of prestressing building
structures with correct design and procedures. Post-tensioned concrete floors offer
the following advantages (Charalambous et al., 2002):
 Reduction in the floor thickness, significant savings in construction time and
lower on-going maintenance costs make the post-tensioning economic.
 Post-tensioning provides excellent control of concrete cracks and deflections.
There are three possible modes of failure in case of post-tensioned concrete slabs,
depending on the level of prestress according to (Khan and Williams, 1995).
Case 1: For lightly loaded slabs, the magnitude of the prestress may be very low so
that the tensile strength of the concrete may provide the most of the flexural strength.
When overloaded, the concrete will crack and failure occurs immediately without
warning.
14
Case 2: At a higher level of prestress of than that of case 1, as load increases, the
tendon reaches its limit of proportionality. Beyond this point, the stress-strain curve
for the tendons becomes nonlinear, the neutral axis rises and the depth of the
compression block reduces. The slab fails either when the tendon breaks or the
concrete crushes. This mode of failure is characterised by a large deflection.
Case 3: At a very high level of prestress, the concrete crushes before the tendon
reaches its yield point and failure is without warning.
Post-tensioning offers the provision of achieving slender, stiff and strong
composite slabs with long spans. The load balancing method is usually applied to
design bonded post-tensioned concrete floors. The design load to be balanced may
only be the dead load or it may also include the calculated permanent load with a
proportion or all of the live load. There is theoretically no vertical deflection or
flexural cracking and the slab is under uniform compressive stress under the
perfectly balanced condition. However, the slab must still be design for serviceability
and ultimate strength and so some conventional reinforcement or more prestressing
steel is needed usually in critical bending regions.
The main advantage of the PTC slabs relies on the ability of the profiled steel
sheeting to act as permanent formwork during construction and as external
reinforcement during service (Ranzi et al., 2013a). The steel decking in the soffit of a
PTC slab has the potential to serve as a major part of the conventional main bottom
tensile reinforcement (Patrick & Lloyde, 2008). However, the sheeting is only
partially embedded in the concrete floor, unlike the case of reinforcing bars in
reinforced concrete structures. Therefore, the structural response of these types of
15
slabs depends significantly on the shear bond interaction between the concrete and
the steel sheeting. It is worth mentioning herein that concrete cracking is
significantly reduced by the post-tensioning action. Both composite action and post-
tensioning offer a significant enhancement in the flexural stiffness and strength when
compared to conventional composite slabs and post-tensioned concrete slabs.
Therefore, the bond stress-slip relationship in post-tensioned composite slab could be
different from the corresponding composite slabs with similar profiled sheeting due
to potential influence of prestress. One of the novelties of this thesis is to explore this
effect, which is presented in chapter 3.
2.3 BOND STRESS-SLIP RELATIONSHIP
The strength, stiffness and ductility of composite profiled slabs depend mainly on
the transfer of longitudinal shear between the concrete slab and the profiled sheeting
through the rib shear connectors (Burnet and Oehlers, 2001). The three generic types
of profiled ribs used in practice consist of the re-entrant or dove-tailed rib, the
trapezoidal, and L-shaped or re-entrant L (Figure 2.2). Combinations of these generic
forms are now commonly used in practice. These ribs transfer most if not all of the
horizontal shear between the concrete slab and the steel sheet of composite slabs
(Oehlers and Bradford, 1999). The rib extends the full length of the structure so that
it does not bear directly onto the concrete, as do mechanical shear connectors.
Rib shear connectors can transfer shear either by chemical adhesion, interface
interlock or by active friction (Oehlers and Bradford, 1995).
16
 Chemical adhesion
Chemical bond results from chemical adherence of the cement paste to the steel
sheeting. It offers resistance to applied shear with no slip at the interface.
Initially the longitudinal shear is transferred across the concrete-steel sheeting
interface by chemical adhesion, which is lost as soon as interface slip occurs
and does not reform. The longitudinal shear is then transferred by the interface
interlock component of the mechanical bond.
 Interface interlock
Interface interlock or mechanical bond results from the physical interlocking of
the concrete and the steel sheeting. The protrusions and indentations on the rib
cause the rib to compress like a spring within the concrete void moulded by
the rib, inducing passive normal and shear forces across the interface that
transfers the longitudinal shear.
Longitudinal shear is transferred through the ribs in the interface interlock
mechanism and not through the steel sheet between the ribs that tend to detach.
The ability of transfer the longitudinal shear by bearing and passive friction
depends on the passive resistance of the rib.
 Active friction
Longitudinal shear can also be transferred by the mechanical action of the
active friction across the sheet-concrete interface in conjunction with interface
17
interlock shear where the interface is subjected to active pressure such as
weight of the concrete resting on the sheet or at the supports where the sheet is
trapped between the concrete and the support. This bond is directly
proportional to the normal force so that if the normal force is zero, then the
frictional force is zero.
The strength of horizontal shear bond depends on many factors, such as the shape
and geometry of the profiled steel sheeting, type and frequency of embossments, type
of end anchorage, length of shear span, slenderness of the slab and arrangement of
load. Therefore, it is difficult to provide representative design values that can be
applied to all slab conditions.
The behaviour of composite slab is directly affected by the slip of the shear
connection at the steel-concrete interface. The flexural forces cause this sliding
action across the interface. This slip is resisted by the interface longitudinal shear
force. Composite behaviour of concrete, steel and the shear connection between steel
and concrete components are usually described by the degree of shear connection
and degree of interaction. Degree of shear connection deals with equilibrium of
forces within a composite member and degree of interaction deals with compatibility
of deformations at the steel-concrete interface. Therefore, degree of shear connection
represents a strength criterion and degree of interaction represents a stiffness
criterion (Oehlers and Bradford, 1999).
The evaluation of shear-bond behaviour in composite slabs has been the focus of
extensive research. However, there are no previous studies focused on the evaluation
of bond-slip behaviour of post-tensioned composite slabs. For accurate numerical
18
simulation of post-tensioned composite slabs, a bond stress-slip model that accounts
for the effects of prestress is required. Moreover, knowledge of the effects of various
levels of prestress on the bond-slip behaviour is important to optimise the design of
2.4 EXPERIMENTAL STUDIES OF POST-TENSIONED COMPOSITE
SLABS
Despite the popularity of post-tensioned composite slabs, very limited research
studies are undertaken. Schravendeel et al. (1990) conducted a full scale
experimental study on PTC slabs and pointed out the potential advantages in terms of
strength and cost. A method to calculate the shear stress distribution at the steel-
concrete interface was developed using typical bond stress-slip behaviour available
in the literature.
Koukkari (1999) conducted a test series of columns formed by two pieces of deep
steel decking encasing a compressed concrete block. Although this test setup seemed
to be similar to the test setup of the current study, the research was to investigate the
influence of the location of post-tensioning anchors on the transfer length and
concrete failure mode. An ultimate strength test was also carried out as part of this
study on a two span continuous composite slab with unbonded tendons to investigate
the load carrying capacity.
Bailey et al. (2006) presented a technique of achieving a long span composite
floor using prestressed composite slabs. However, in this research, prestressing was
applied to the steel deck prior to delivery on site, while in the current study
19
prestressing was applied to the concrete after 7 days of pouring, which is typical of
Australian construction code of practice. The bond stress-slip characteristics in both
cases are considered to be different due to the two different methodologies of
prestressing.
Patrick and Lloyde (2008) carried out ultimate strength tests of a composite slab, a
post-tensioned solid slab and a post-tensioned composite slab and provided design
capacity tables for different post-tensioned slab arrangements. This research assured
the attainment of enhanced flexural strength of post-tensioned composite slabs over
post-tensioned solid slabs and profiled composite slabs.
Xiao et al. (2009) analysed the flexural performance of the profiled sheet and
prestressed concrete composite slabs. Load-deformation curves of prestressed
composite slabs are shown as three stages: elastic, elastic-plastic and plastic.
Analysing the influence of concrete strength, steel thickness and initial prestressing,
the authors found that concrete strength and steel thickness have less influence on the
load-deflection curves, while initial prestressing has stronger influence on the initial
and last stage of load-deflection curves. The author also commented that by applying
prestressing on composite slab, the crack resistance and load carrying capacity can be
improved and deformation can be reduced.
Ranzi et al. (2013a & b) presented results of a series of ultimate and long term
tests on PTC slabs and an analytical model for the determination of the ultimate load
carrying capacity of PTC slabs. Based on experimental study of post-tensioned solid
slabs and post-tensioned composite slabs, the authors found that rigid plastic
calculations produced good predictions for the moment capacity of post-tensioned
20
solid slabs, while the values calculated for the post-tensioned composite specimens
based on full shear connection model overestimated the test results. An analytical
model was also presented for the determination of the ultimate flexural capacity of
post-tensioned composite slabs based on partial connection theory. This model
requires the value of mechanical resistance Hr of the profiled steel sheet, which is
measured experimentally from slip-block or slab tests. However, the effect of
prestress on the bond stress-slip relations is not reflected in the value of Hr.
The novelty of the current research in relation to these previous research studies is
to explore the concrete-steel interaction behaviour of the PTC slabs accurately. The
effect of both prestress and bond length on bond stress-slip relationship is a
fundamental issue, which needs to be addressed in determination of ultimate capacity
of post-tensioned composite slabs. An appropriate push test setup was chosen and
post-tensioning was incorporated to find accurate bond stress-slip relations, as
presented in Chapter 3.
2.5 PUSH TEST STUDY
It has been recognised that the mechanism of longitudinal shear transfer between
the profiled steel sheets and the concrete in composite slabs is complex and difficult
to model mathematically (Abdullah and Easterling, 2004). Therefore, the shear
interaction property is usually determined using a push test. This behaviour is
characterised mainly by the chemical bond and the mechanical bond. Although the
chemical bond strength is usually assumed to be controlled mainly by the material
strength of chemical adhesion, a push test study conducted by Burnet and Oehlers
21
(2001) indicated that it is rather a geometric property in profiled composite slabs and
failure of the chemical bond is more akin to a peeling action.
The author of this thesis assumed that the bond-stress slip relation in post-
tensioned composite slab is influenced by the effect of level of prestress and bond
length. Therefore, a need for a small scale test setup with feasibility of incorporation
of the effect of prestressing is obvious. The literature review presented herein will
discuss and compare existing push test studies of composite slabs conducted by other
researchers to determine the shear bond characteristics. The objective of this review
is to find out an appropriate push test setup for determining bond stress-slip
relationship of post-tensioned composite slabs accurately. It is noteworthy herein that
existing push tests can be categorised into two classes: 1. Push tests setup requiring
external lateral restraining forces and 2. Push tests setup without external lateral
restraining forces.
Schuster (1970) conducted one of the earliest push out type small scale test to
investigate the shear bond characteristics of profiled steel sheets and establish a
relationship between the maximum push out force and moment capacity of slabs with
the same shear span length (Figure 2.3). The push out test results were not in good
agreement with the shear bond strength for composite slabs in flexure. Therefore, the
full scale tests were considered as the preferred method to evaluate the behaviour of
composite slabs and became the basis for the development of the m-k design
procedure.
Stark (1978) configured a push out test to determine the influence of concrete
quality on the maximum load per embossment and evaluate the results of full-scale
22
bending tests that were used by the author in the partial shear connection (PSC)
method. This was an earliest attempt to apply partial interaction theory for composite
slab design. The concrete block was placed between steel sheets and the steel sheets
were clamped against the concrete block but the clamping force was not measured
(Figure 2.4). The ultimate shear load per embossment from a push test was found to
differ by 15% from the bending test. The author concluded that the difference was
due to the effect of edge webs in the bending test, which were unrestrained and free
to curl, but were clamped in the push test. Moreover, the applied lateral force can
induce frictional resistance at the steel-concrete interface and prevent the separation
of the sheeting from the concrete.
Jolly and Zubair (1987) carried out a push off test to evaluate the effect of various
types of indentations of the steel sheeting on the shear bond strength, as shown in
Figure 2.5. It was found that the discontinuous embossments enable the steel sheet to
distort and ride over concrete more easily. Embossment faces orthogonal to the
direction of slip were found to be more effective than the inclined. No advantage in
more numerous and smaller indentations was observed. The experimental results
were also used in theoretical calculation for predicting the load carrying capacity of
the composite slabs. Obviously, the test results are significant for the modification of
embossment shape, size, depth and spacing. However, this test setup is not
considered further for the current study as it was not feasible to incorporate post-
tensioning in this configuration.
Daniels (1988) developed a pull out test, as shown in Figure 2.6, to determine the
load-slip behaviour and strength of the shear connection provided by embossments or
by the form of decking ribs at the steel-concrete interface. In this test setup, axial
23
load was applied in the direction of the rib and lateral force was applied
simultaneously to induce normal force representing weight of the concrete at the
sheet-concrete interface. The shear resistance of profiled sheeting found from the test
was overestimated, which may be due to the application of the lateral force
producing additional frictional resistance at the interface. The steel deck which was
held against each other in this test may be prevented separating from the concrete
block.
Patrick and Poh (1990) developed the slip block test as shown in Figure 2.7 in
order to investigate longitudinal slip resistance between the concrete and the profiled
sheeting. The average shear stress per unit horizontal area and co-efficient of friction
between the concrete and the sheeting were found from this test. The sheeting was
welded to the rigid base plate and lateral load was applied against the concrete block.
Therefore, vertical separation of the sheeting from the concrete was most likely
restricted, which could result in overestimation of the shear bond stress. A point
worth emphasizing is that the slip block test was only suitable for profile that exhibit
ductile shear connection. In 1994, Patrick and Bridge developed a partial shear
connection strength model using the results from slip block test. The author stated
that this model has been verified by the full scale tests and can account for the effects
of variation in shear connection performance, loading pattern and end support
conditions.
An (1993) developed a block bending test as shown in Figure 2.8 to include the
effect of bending curvature in small scale tests and to determine the shear stress
between the concrete and the profiled sheeting. The parameters studied were
concrete type and the shear span to depth ratio. The authors used two types of test
24
setup. In the first type of test setup shown in Figure 2.8 (a), the steel sheeting was in
contact directly with the support so that the effect of frictional force on the slip
resistance could be obtained. The second setup shown in Figure 2.8 (b) did not have
the steel in direct contact with the support to obtain the shear strength without
support friction. The shear stress was determined by calculating tensile force in the
sheeting analytically. The results were verified with the strain in the sheeting whose
values were measured during the test. The shear resistance of specimen with the
sheeting extended into the support was found 20-30% higher than that without
sheeting. This clearly demonstrated the presence of shear resistance at the support.
The inclusion of the effect of bending curvature and vertical separation was the
major improvement in this small scale test. However, due to shortness of the
specimen and the nature of the test setup, bending effect might not be simulated
accurately in this test also.
Veljkovic (1996) investigated the behaviour of composite slab using finite
element method. The parameters used in FE model were mechanical interlocking
resistance, a friction co-efficient at supports and a reduction function of the shear
bond strength, which were obtained from push test, slip block test and tension-push
test respectively as shown in Figure 2.9. From the tension-push test, it has been
recognised that tensile strain could flatten the embossments which consequently
would lower the shear bond resistance and increase the corresponding slip.
Therefore, a reduction function was used to make correction to the mechanical
interlocking resistance obtained from the push test. Veljkovic (2000) presented a
design method for composite slabs with shallow sheeting based on small scale tests
and a finite element analysis, namely three parameters partial connection strength
25
method. The envelope of the bending moment resistance can be determined
according to the mechanical model, where the influences of friction at the support
and slab overhang are considered.
Most of the small scale tests mentioned so far restricted the sheeting deformation.
In order to overcome this drawback, Burnet (1998) used push test setup as shown in
Figure 2.10 similar to Stark (1978) except that no lateral pressure was applied to the
specimen and the steel sheet was free to move laterally. The bond characteristics of
variable shape profiles were studied in this experimental study. It was observed that
the longitudinal slip was occurred along with lateral separation of the sheeting from
the concrete for profiles with embossments. For re-entrant ribs, the separation was
less than the trapezoidal ribs. Burnet (1998) also determined the limit of rib opening
beyond which the shear capacity was completely lost after breaking of chemical
bond.
Tremblay et al. (2002) developed a push out test as shown in Figure 2.11. A
lateral load of 6 kN was applied at the end of the specimen to simulate an end
condition typical of simply supported slab. The authors used the test to study the
effects of steel thickness, steel grade, surface coating, deck position and concrete
curing age. As expected, the shear strength was higher for stronger and thicker steel
and for longer concrete curing age.
In order to simulate the bending effect of the real slabs accurately and determine
composite slab strength under bending, Abdullah (2004) presented a small scale test
setup, as shown in Figure 2.12. The size of the specimen was kept at one-fourth of a
standard bending test specimen of composite slabs. It was observed that edge web
26
curling, end anchorage details and types of support can influence the behaviour and
strength of composite slabs significantly. The small scale test data were found very
close to the full-scale test results provided the same end details were maintained in
both elemental and full-scale specimens. However, the bond stress-slip relationship
was derived using an analytical method for the small scale specimens.
Several other researchers also carried out push out test in a different arrangement
(Plooksawasdi, 1977; Airumyan et al., 1990). However, those setups are not covered
herein. Finally, a similar push test setup without external lateral restraining forces
such as Stark (1978) and Burnet and Oehlers (2001) was adopted with appropriate
modification to optimise the ease of testing and accuracy of results, although the
challenge of handling the delamination problem and local buckling at the base of the
profiled steel sheets exists. Despite the similarity in the push test setup used by
previous researchers and the one presented in this thesis, the objective for this
research is quite unique when compared with all the previous studies related to PTC
slabs.
2.6 NUMERICAL MODELLING OF POST-TENSIONED COMPOSITE
SLABS
Numerical modelling and analysis of composite structures using the finite element
method have become a very popular area of research recently due to its advantages
of higher efficiency and lower computing costs.
An and Cederwall (1992) conducted a two dimensional finite element analysis of
composite slabs using ABAQUS and used the shear stress-slip behaviour from a
27
concrete block bending test. The interaction between the profiled steel sheets and the
concrete was modelled with spring elements. An (1993) investigated the behaviour
of composite slabs by means of 2D nonlinear finite element analysis using
ABAQUS/Standard. The steel sheeting and the concrete were modelled as 2 node
Timoshenko beam elements. The interaction between the steel sheeting and the
concrete slab was modelled with spring elements and a set of imposed equations
between degrees of freedom of concrete, spring and steel deck beam elements. The
spring property was found from a block bending test. Concrete cracking was
represented by a smeared crack model while a nonlinear stress-strain curve was used
for the steel deck and the concrete. The load-slip relationship simulating vertical
separation between the concrete and the steel deck was assumed as linear elastic and
modelled with spring elements. The FE results were very close to full scale test
results for long slabs but underestimated the capacity of short slabs. The author
attributed the underestimation to exclusion of the effect of the reaction force in the
model. Analysis conducted with different vertical separation properties did not
produce variable results.
Daniels and Crisinel (1993b) presented a numerical analysis incorporating pull out
test results to predict the behaviour and strength of composite slabs. The authors used
plane beam elements for analysing single and continuous span composite slabs. The
procedure considered nonlinear material properties, additional positive moment
reinforcement, the load-slip property for the shear studs (from push tests) and the
shear interaction property between the steel deck and the concrete slab (from pull out
test). A ductile load carrying mechanism was applied in the model and the brittle
portion of the push test data was ignored. The compressive stress-strain behaviour of
28
the deck was specified differently from the tensile behaviour, where the relationship
was established from the flexural tests on bare deck. The maximum load carrying
capacities for models with different span lengths were found to lie along the straight
line if plotted on the m-k axes but fell outside the same straight line if the steel
thickness is varied. The author attributed the results to the shear stress distribution
which depended less on span length than it did on slab thickness.
Veljkovic (1996) performed a three dimensional finite element analysis of
composite slabs using a nodal interface element available in DIANA to investigate
the mechanism of the longitudinal shear transfer between the concrete and the steel
sheet in composite slabs. The interaction property was obtained from a push test and
modified with a reduction function to correspond with the strain in the profiled
sheets. Different uniaxial stress-strain curves were used for the web and the flange of
the sheeting, based on tensile tests. A reduction function obtained from push-pull
tests was used to modify the interface property. The sheeting was under tensile strain
while the concrete was pushed over it because this property was found inversely
proportional with increasing strain in the sheeting. The steel strain reduced the
resistance of the embossment interlock representing more closely the behaviour of
real slabs.
Widjaja (1997) presented a finite element analysis of composite slabs with shear
stud and weld anchorage at the supports following a similar model used by An
(1993). The load-slip relationship for the interface element was obtained from the
small scale tests similar to that used by Daniels (1988) and the end anchorage was
obtained from a modified composite beam push off test. But the vertical load-slip
29
relationship was ignored. The results of the FE analysis was very close to the full
scale test results.
Abdullah (2004) developed a 3D nonlinear FE model and analysis of composite
slabs using ABAQUS/Explicit 6.3. The concrete slab was modelled with 8-node
linear brick, reduced integration elements (C3D8R), while the steel deck was
modelled with 4-node doubly curved shell, reduced integration finite membrane
strains elements (S4R). The interaction between the concrete and the steel deck was
modelled with radial thrust type connector elements (CONN3D2). The vertical
interaction and the frictional resistance at the support were not considered in the
model .The author reasoned that since the horizontal shear property was derived
directly from the bending test, these effects were implicitly included. The FE results
found in this study indicated that accurate results for a particular slab can only be
obtained when the shear bond-slip property from the same slab is used because the
shear bond property is unique to a particular slab dimension and is not
interchangeable to other slab dimensions. The authors stated that all FE analysis of
composite slabs performed by many researchers in the past used a single shear-bond
property obtained from conceptually the same push off type tests. Tenhovuori &
Leskela (1998) even conducted the FE analysis using an assumed non-variable shear
bond-slip relationship on various slab geometries. Although those analyses showed
success to some extent, sufficient information regarding the accuracy of the analysis
with respect to variable geometry model was not presented in those studies.
Veljkovic (1995) had to conduct three types of small scale tests to obtain a good
model for the shear interface property. To achieve a correct simulation, a function
was used to reduce their interfacial shear stress level during the analysis depending
30
upon the strain level in steel sheeting. This modification is clear evidence regarding
the need for variable shear interface properties when dealing with variable geometry
models. The major drawback of FE modelling can be attributing to the use of single
shear property obtained from push off type tests. These tests do not represent the true
behaviour of slab bending, where clamping, slenderness and curvature are not
present (Abdullah, 2004).
Some other researchers attempted to model the shear bond behaviour as a
nonlinear contact problem using contact elements available in ANSYS. Ferrer et al.
(2006) developed a FE model to simulate the longitudinal slip mechanism in pull-out
tests in order to investigate the effects of the friction co-efficient and geometrical
parameters of profiled steel sheets.
Tsalkatidis and Avdelas (2010) modelled the shear bond interaction in composite
slabs as a unilateral contact problem with friction. Chen and Shi (2011) performed
FE analysis on both pull-out and bending tests of composite slabs considering
adhesion and friction. Both Ferrer et al. (2006) and Chen and Shi (2011) used
Daniel’s (1993a) pull-out test setup which is a laterally restrained type of small scale
test. But the push test setup used in this study is a laterally unrestrained type of push
test such as in Stark (1978) and Burnet (1998) push tests. Accurate modelling of
laterally unrestrained push tests using the contact approach would require significant
effort for calibration of cohesion and friction in order to achieve good agreement of
the FE model results with the experimental results. Since the incorporation of the
effects of prestress in the FE model was the main concern of this thesis, the
connector element approach was thought reasonable in terms of accuracy and
simplicity.
31
Ellobody and Bailey (2008) used 3D solid elements to model the unbonded post-
tensioned concrete slabs using ABAQUS. A combination of C3D8 and C3D6 was
used to model the concrete, tendon and anchorage elements. The stress-strain curve
of the tendon from tensile test was used in modelling tendons (using the *PLASTIC
option). The contact between the concrete and the tendon was modelled by interface
elements (using the *CONTACT PAIR option). Only a quarter of the slab was
modelled due to symmetry. The ultimate loads, load-central deflection curves, failure
modes and strains in the tendons obtained from FE model were compared against the
test results. Good agreement was reported by the authors between numerical and
experimental results.
Malaska and Pajari (1996) focused on the finite element modelling of post-
tensioned composite slabs with unbonded tendons in the serviceability limit state. A
nonlinear finite element program entitled COMPCAL was used for the simulation of
the long-term behaviour of composite slabs, particularly the effects of creep and
shrinkage of the concrete and those of relaxation of the prestressing steel. The
nonlinear stress-strain relationship of the steel and the concrete, the cracking of the
concrete and the bond-slip at the interface of the concrete and the steel sheeting were
taken into account. But the model was not verified against any experimental data.
The finite element analysis, if conducted on post-tensioned composite slabs with
different levels of prestress utilising a single shear bond property without considering
the effect of prestress on the bond between the concrete and the profiled steel sheets,
may lead to an inaccurate estimation of the ultimate capacity of the post-tensioned
composite slabs. Therefore, this thesis aims to study the shear bond stress-slip
behaviour in post-tensioned push specimens using a FE approach incorporating the
32
effects of prestress on the local bond-slip behaviour. This FE model employed a
layer of interface elements between the concrete and the profiled steel sheets. The
effects of various levels of prestress on the overall bond stress-slip behaviour as
observed in the experimental study were incorporated in the local bond-slip
properties of the interface element. The objective of this study is achieved by
validating the experimental results with the FE simulation results.
2.7 BEHAVIOUR OF COMPOSITE SLABS WITH END ANCHORAGE
Mechanical interlock by indentations or embossments in the profile, frictional
interlock, end anchorage by through-deck welded studs and end anchorage by
deformation of the ribs at the end of the steel sheeting are very common types of
effective shear connection in composite slabs (British Standards Institution, 2004).
The type of end anchorage has a significant effect on the strength, stiffness and
ductility of the composite slabs. The welded shear studs are the most common and
popular forms of end anchorages in composite floor slabs. A significant amount of
research has been published in the open literature to investigate the behaviour of
composite slabs. However, limited studies have been published to investigate the end
anchorage effect on the strength and behaviour of composite slabs with end
anchorage.
One of the earliest attempts reported in literature to investigate the effect of end
anchorage is the experimental study conducted by Porter and Greimann (1984). The
authors carried out a series of experiment on composite slabs comprising of eight
composite slabs with end anchorage and seven composite slabs without end
33
anchorage. The results of this study indicated an 8% to 33% increase of the shear-
bond strength of composite slabs due to studs installed at the end of the span
depending on span length and sheeting thickness. Several studies after this research
also confirmed the increase of load carrying capacity of composite slabs with studs
used for end anchorage (Jolly and Lawson, 1992; Chen, 2003). The authors
recommended same empirical design approach for both composite slabs with and
without studs. Therefore, the research for an analytical method for predicting the
strength of composite slabs with studs was left under consideration.
Easterling and Young (1990) carried out an experimental study to investigate the
effect of end-slip restraints provided by hot-rolled angles, cold-formed angles, shear
studs and adjacent spans on strength of composite slabs. This study was followed by
development of an analytical method for determination of strength and initial
stiffness of composite slabs based on conventional reinforced concrete concepts and
elastic analysis principles. However, the magnitude of the force required for the
sufficient anchorage for yielding of the entire cross section at ultimate failure was not
precisely established in this study (Degtyarev, 2014a).
Jolly and Lawson (1992) proposed a method to quantify the contribution of the
end anchorage to the shear-bond resistance of the composite slabs. A series of six
composite slab tests were conducted in order to quantify the combined effect of end
anchorage and shear bond, five of which were provided with shear connectors.
Instead of adding the moment capacity due to shear bond to the moment capacity due
to end anchorage directly, this study proposed that a reduction factor should be
applied to the moment capacity due to end anchorage. The authors recommended that
50% of this anchorage force needs to be included with the longitudinal shear capacity
34
to predict the increase of the shear-bond capacity of composite slabs. Decades later,
an experimental study conducted by Chen (2003) found that maximum 30% of the
end anchorage strength was mobilised in studded composite slabs.
Bode and Minas (1997) reported results of an experimental study on composite
slabs with and without end anchorages utilising three different types of profiled steel
sheeting. All slabs showed brittle longitudinal shear failure modes. End anchorages
provided a very high load carrying capacity and good serviceability properties under
static and dynamic loading.
Chen (2003) investigated contribution of shear-bond resistance in the load
carrying capacity of composite slabs through an experimental study. Seven simply
supported one-span composite slabs and two continuous composite slabs were tested
as part of this research. The results indicated that composite slabs would not be
capable of developing full plastic moments in the spans due to the shear-bond failure
at the steel-concrete interface. The slabs with end anchorage achieved significantly
higher shear-bond resistance than the slabs with no anchors. The continuous
composite slabs attained higher load and less deflection when compared with simply
supported single span composite slabs. Author also concluded that the longitudinal
shear bond-slip governs the contribution of the end restraints to the shear-bond
resistance in composite slabs instead of the strength of the stud connectors.
Chen et al. (2011) carried out an experimental study on the shear-bond failure
mechanism of composite deck slabs. Thirteen simply supported composite slabs were
tested, where five slabs were provided with shear studs on each side. The authors
35
also developed a method to assess the moment capacity of composite slabs based on
partial shear connection and verified against the test results.
The main objective of the experimental study by Mirza and Uy (2013) was to
investigate the short and long term behaviour of composite slabs. The test program
consisted of four slabs for short term ultimate strength test and eight slabs for the
long term test. The results of the short term test indicated that the end restraints
increased the capacity of the slabs significantly for both solid and profiled slabs,
which corresponds to the observations made in the previous researches. The effects
of creep and shrinkage in the concrete caused greater deflection in the long term tests
over a period of 710 days. The ultimate strength test has been conducted on these
eight long term slabs as part of the current study and reported in this paper.
Degtyarev (2014a) developed an analytical model for predicting strength of
composite slabs with end anchorage. This model accounts for end anchorage strength
and flexibility, longitudinal shear strength between the concrete and the steel
decking, slab geometry and properties of the deck and the concrete. The effects of
the slip on the stress-strain state of the slab and on the end anchorage strength
mobilisation were considered. He also conducted a parametric study to quantify the
effects of design parameters on strength of composite slabs with end studs and to
determine the minimum number of shear studs required to achieve the full moment
capacity of the section (Degtyarev, 2014b).
One of the primary objectives of this thesis is to present a series of ultimate
strength tests aimed at evaluating the effects of end anchorage on the load carrying
capacities of both solid and profiled composite slabs. Eight composite slabs were
36
tested under different boundary conditions. Details of the experimental investigation
into composite slabs are provided in chapter 7.
2.8 NUMERICAL MODELLING OF COMPOSITE SLABS WITH END
ANCHORAGE
Several empirical design methods and few analytical methods are available for
determining the ultimate strength of composite slabs with end anchorage. However,
the numerical studies on composite slabs with end anchorage are not well
established.
Daniels and Crisinel (1993a) presented a calculation procedure combining shear-
bond test results with a numerical analysis to predict the behaviour and strength of
one-way single and continuous span composite slabs with ribbed decking as an
alternative to full scale testing. Longitudinal shear bond-slip and end anchorage
characteristics obtained from pull-out tests were used in the nonlinear finite element
model. The author stated that composite slab behaviour and maximum load carrying
capacity should be modelled ignoring brittle load carrying mechanism so that the
resulting prediction is found conservative and ductile in nature. However, the
capability of the model to simulate the behaviour of profiled composite slabs is
deemed to be limited and it is possible to model the behaviour more efficiently and
accurately due to availability of sophisticated finite element software packages.
Therefore, one of the objectives of the present study is to develop an efficient and
reliable nonlinear three dimensional finite element model using a commercially
available software ABAQUS to account for the shear connection behaviour of the
37
composite slabs and to predict the ultimate load carrying capacities and failure
modes.
The challenging issues in the modelling of composite slabs with end anchorage
are the simulation of shear bond behaviour and the concrete-shear stud interaction.
The modelling of shear bond behaviour is discussed in detail in Section 2.6. Several
studies have been dedicated to develop finite element models to investigate the
performance of headed stud shear connectors (Ellobody & Young, 2006). Modelling
the shear stud interaction using surface to surface interaction is a popular approach to
gain an insight of the behaviour of concrete-shear studs interaction. However, some
other researchers used nonlinear spring to model the concrete-shear stud interaction
(Vasdravellis et al., 2009). In this thesis, the shear bond behaviour of composite slabs
with end anchorage is modelled using connector elements available in ABAQUS and
concrete-shear stud interaction is modelled using surface to surface interaction.
This chapter has presented a detailed review of the published literature which is
relevant to the objectives of this thesis. Although the application of post-tensioned
composite slabs has been increasingly demanded for the possibility of achieving a
slender, stiff and long span slab in construction industry, the limited experimental
work and lack of research on the issue of bond stress-slip relations in post-tensioned
composite slabs is evident. The novelty of the experimental work reported in this
thesis is that it incorporated post-tensioning at the steel bar to apply precompression
to the concrete of the push test specimens in order to determine the potential
38
influence of prestress on the bond between the profiled steel sheets and the concrete.
An accurate local bond-slip model is of fundamental importance in the modelling of
Several experimental studies on composite slabs with end anchorage are available.
However, the numerical studies on composite slabs with end anchorage are not well
established. This thesis presents a nonlinear finite element model developed using
commercial package ABAQUS, which provides a detailed insight into the behaviour
of composite slabs with end anchorage.
39
Figure 2.1 Modes of failure of composite slab (Johnson, 1994)
(a) Dove-tailed (b) Trapezoidal (c) Re-entrant L
Figure 2.2 Different types of profiled steel sheeting
Figure 2.3 Push out test (Schuster, 1970)
40
Figure 2.4 Push out test (Stark, 1978)
Figure 2.5 Push off test (Jolly and Zubair, 1987)
41
Figure 2.6 Pull out test (Daniels, 1988)
Figure 2.7 Slip block test (Patrick and Poh, 1990)
42
(a)
(b)
Figure 2.8 Block bending test (An, 1993)
(a)
(b)
Figure 2.9 Push test and tension push test (Veljkovic, 1996)
43
Figure 2.10 Push test (Burnet, 1998)
(a)
(b)
Figure 2.11 Push test (Tremblay, 2002)
44
(a) Elevation
(b) Side view
Figure 2.12 Small scale test setup (Abdullah, 2004)
45
46
CHAPTER 3-Push Tests
CHAPTER 3
PUSH TESTS
3.1 INTRODUCTION
This chapter presents the details of an experimental programme that was carried
out in the form of push tests to investigate the bond stress-slip behaviour of post-
tensioned composite slabs. Push tests were conducted in two separate series which
are referred to as preliminary series and main series. This chapter presents results
from these two series of push tests followed by discussion of the results. The
parameters investigated in this study were (a) level of prestress and (b) bond length.
These were varied to determine the potential influence of prestress on the bond
between the profiled steel sheets and the concrete in post-tensioned composite slabs.
3.2 TEST SPECIMENS
Two series of tests were carried out in the Structural Research and Testing
Laboratory of University of Western Sydney, Australia. The preliminary series,
which consisted of six push test specimens, was conducted to optimise the push
specimen configuration and test setup. The optimistic results of the preliminary
47
series led to the main series of thirty two push test specimens. The experience gained
from the preliminary series was implemented in the form of appropriate modification
in case of formwork preparation, concrete casting, post-tensioning, boundary
conditions and loading in order to obtain best possible results from the push tests of
main series. The geometry of the push test specimen and details of its components
are shown in Figures 3.1 & 3.2.
Test specimens have been referred throughout in this thesis in the form of “PT-H-
P” where PT, H and P indicate push test specimen, height of the push specimen in
mm and level of prestress in MPa respectively. Therefore “PT-500-3” refers to a
push test specimen where the height of the push specimen is 500 mm and the level of
prestress is 3 MPa. All specimens were constructed using 1.00 mm thick Stramit
Condeck HP® profiled steel sheets produced in Australia by Stramit Pty Ltd. The
sheeting was grade 2 galvanised sheeting with total coated thickness of 1.03 mm and
a base metal thickness of 1.00 mm (Stramit, 2007). The geometry of the sheeting is
shown in Figure 3.3.
Each specimen consists of two identical sections of profiled steel sheets which
have a concrete block cast in between them and post-tensioning was applied to the
concrete block. The level of prestress was 1-3 MPa concrete stress due to prestress as
these values were thought to represent the practical limits of prestress in composite
slabs with profiled steel sheets in floor construction.
A critical literature review of research work incorporating push tests of composite
slab to determine the optimum test setup has been presented in Chapter 2. The factors
that can influence the results of push tests include not only the physical configuration
48
but also the boundary conditions and the loading. Particularly in this study, the
method of post-tensioning is very vital. Since the objective of this push specimen is
to determine the possible influence of the prestress on the bond-slip behaviour, no
attempt was made to reflect the effect of bending, slab slenderness and frictional
resistance on supports.
All specimens had identical cross-section with length L and width W of 300 mm
and 180 mm respectively. The height of the bonding area H of the push-specimens
was varied as 500 mm, 1000 mm, 1500 mm and 2000 mm to investigate the variation
of bond stress with slab span. The effective bonding area between the concrete and
the profiled steel sheets was determined by the height H and the length L of the
concrete block.
Two factors were taken into account in determining the minimum and maximum
dimension of H. Firstly, the interface area is sufficient to address the effect of the
prestress in relation to the variable height. Secondly, the maximum load developed in
the sheeting must not cause buckling in the sheeting at the supports at any time
throughout the test. The bearing stress at the base of the profiled steel sheets for a
given profile is given by Equation 3.1 (Burnet, 1998):
𝜎𝑏𝑢 = 𝑓𝑏𝑢 𝐻 ⁄𝑡 (3.1)
Where 𝜎𝑏𝑢 is bearing stress at the base of the profiled sheeting, 𝑓𝑏𝑢 represents
maximum average bond stress in push test specimen, H depicts height of push test
specimen and t represents thickness of the profiled sheeting.
49
It is obvious from the equation (3.1) that the higher the bond length, the higher the
possibility of buckling to occur. That is why most of the previous studies used a very
small height for the push tests. In this study, the height was varied up to 2000 mm to
achieve more accurate bond-stress characteristics. Two channels have been placed
and finger tight only at the bottom to maintain the rotational equilibrium of the push
specimen and prevent buckling in the sheeting at the base of the push-specimens.
The height of the channel has been specified as 200 mm for all the push specimens. It
is important to mention that there is no lateral restraining forces to constrain the
sheeting in this test setup.
The length L was chosen to allow one complete width of the sheeting to be tested.
Since the width of the Condeck profile is 300 mm, the length of the push specimen
has been specified as 300 mm to simulate the overall performance. No lateral
restraint was required to limit the sheeting distortion. The width W was chosen to be
180 mm to avoid premature splitting of concrete between opposite ribs and to
provide enough clearance between the sheets to apply the load uniformly to the
concrete block.
The profiled steel sheets overhang the concrete block by 40 mm at the top and
bottom of the push-specimens to accommodate a slip of 10-20 mm for the concrete
block along the sheeting and also to keep a future provision of any restraint if
required at the base of the sheeting to prevent buckling. However, maximum slips
observed at the ultimate capacity of composite slabs are usually in the order of 5 mm
(Crisinel, 1990; Burnet, 1998). The base on which the specimen was placed
incorporated steel plates of 20 mm in height and one base is placed on a roller so that
the profiled steel sheets are free to deform away from the concrete.
50
3.3 PRELIMINARY SERIES
Six push test specimens were prepared and tested as a preliminary series of the
experimental programme. In this series, three specimens were post-tensioned and the
bond length H of the push-specimens was varied as 500 mm, 1000 mm and 2000
mm. A summary of the push specimens is provided in Table 3.1.
3.3.1 Material properties
All push specimens were cast on the same day. The details of the constituents of
the concrete mix used in the experiments are reported in Table 3.2. The material
properties of the concrete were determined from standard cylinder tests as reported in
Table 3.3. A minimum of three cylinders were tested each day. A total of 15
cylinders were tested over a period of 34 days. The average concrete compressive
strengths are presented in Table 3.4. The compression tests were performed on 100
mm diameter cylinders, using the procedure outlined in AS 1012.9 (Standards
Australia, 1999). The tensile capacity of the concrete was determined using the
Splitting Test as outlined in AS 1012.10 (Standards Australia, 2000). The concrete
cylinders were tested at the same day that the push tests were carried out. The
properties of the profiled steel sheets and prestressing bar were obtained from
Standard tensile tests as presented in Table 3.3. The tensile test for profiled steel
sheets was performed in accordance with AS 1391(Standards Australia, 2007). A
total of three coupons of profiled sheeting were prepared and tested. The nominal
tensile strength, nominal breaking load and nominal yield load of the prestressing
bar, as provided by the supplier, was 1030 MPa, 728 kN and 590 kN respectively.
Tensile test for prestressing bar was performed in accordance with AS/NZS 4672.1
51
(Standards Australia, 2007). A total of three coupons of prestressing steel bar were
prepared and tested. Further details of the tensile tests of profiled sheeting and
prestressing bar are provided in Appendix A.
3.3.2 Specimen preparation and post-tensioning
The first step was to cut the profiled steel sheeting to a length of 580 mm, 1080
mm, 1580 mm and 2080 mm to allow bond length of 500 mm, 1000 mm, 1500 mm
and 2000 mm respectively and 40 mm overhang at the top and bottom of each
specimen. The sheeting was as received from the manufacturer and no measures
were taken to treat the inner surface condition. The formwork ready for pouring is
shown in Figure 3.4.
The formwork was assembled with appropriate dimensions and post-tensioning
cable and PVC duct was placed with zero eccentricity. A lifting lug was placed on
the sides of the specimen. The specimens were poured along with test cylinders and
kept moist for two days. After this period they were left to cure in the air until post-
tensioning.
The three specimens were post-tensioned at day 10 after casting with a concrete
strength measured at day 7 of 23.0 MPa. Post-tensioning was carried out by
elongating the 29 mm threaded bar from the live anchor through tightening by a
wrench (Figure 3.4). Precompression was applied to the concrete through end blocks.
A load cell was used to measure the applied load to the bars during the post-
tensioning while three strain gauges fixed on the bar were used to record the strains
within the bar. Push-specimens were clamped well in a strong floor during post-
tensioning to minimise any uplift or external disturbance. A single load cell was used
52
to perform the post-tensioning by calibrating each bar for corresponding strains in
that bar. The applied load versus strain recorded by the gauges fixed on the bar for
PT-500-3 is illustrated in Figure 3.5. It was found that the strains recorded by the
three gauges are different. The centre strain gauge reading was used for calibration.
The full applied design prestressing force was 162 kN and is equal to about 22 % of
the nominal breaking load.
3.3.3 Instrumentation
Details of instrumentation for the push specimens are shown in Figure 3.7. Two
linear potentiometers were attached to the sheeting to measure the slip at the
interface directly (Figure 3.8). Three strain gauges (FLA-5-11-3LT) were placed on
each outer surface of profiled steel sheeting to observe the strains developed during
the debonding process. Three concrete strain gauges (PML-60-3LT) were placed
near the centre of the specimen to observe the concrete strains. Concrete strain
readings were also taken using DEMEC strain gauges. A data acquisition system was
used to record the readings of the load cell, linear potentiometer and strain gauges.
3.3.4 Test setup
Before moving the specimens for testing, clamps are placed on the outer edges to
prevent any predelamination to the sheeting during the test setup. Predelamination
would reduce the actual bond strength between the concrete and profiled sheeting.
Therefore, the bond stress-slip relations obtained from each specimen needs to be
analysed and interpreted carefully. The specimens are lifted up by the lifting lugs and
then placed onto the base of the loading rig. Two channels were fixed and finger
tight only at the base of the sheeting. The base of one side of the specimen was
53
placed on a roller so that the sheeting is free to move away from the concrete. Some
bars were placed as a precaution at the periphery of the specimen during the testing
which do not contribute any lateral load to the specimen. Two end blocks were
placed on the top of the specimen to distribute the load evenly onto the specimen. A
thin layer of plaster is used to ensure a levelled surface for uniform load distribution.
A specially designed steel plate was integrated with the spherical seat to provide a
uniform dispersion of the load from the loading jack to the specimen and thus
maintain an even slip over the specimen (Figures 3.9 & 3.10). A push specimen
ready for testing is shown in Figure 3.11.
The side clamps were removed just before the load application started. The extra
bars were placed at a distance and channels are made loose at the start of loading.
The load was applied to the specimen very slowly at a rate of 0.005 mm/sec until the
chemical bond had been completely broken as indicated by the linear potentiometer
recording a slip displacement. After failure of the chemical bond, the load was
increased gradually. The load and slip were logged continuously. The test was
terminated when either the specimen was unable to take any load or the limit of slip
had been reached.
3.3.5 Summary of preliminary series
3.3.5.1 General
It was assumed in calculation of bond stress that the ribs of the profiled steel
sheets transfer the total shear in the concrete-sheet interface. Since maximum slips
recorded from composite slab tests are usually less than 10 mm, push tests were
conducted and reported up to a slip of 10 mm. Chemical bond stress is derived from
54
the maximum load at which the chemical bond failed by dividing the load by the
entire contact area between the sheets and the concrete. The residual stress is
considered the minimum value after failure of the chemical bond and before the
effect of the mechanical bond. The average mechanical bond stress was determined
by dividing the average load between the maximum load and load at 10 mm slip by
the entire contact area between the sheets and the concrete. The bond stress at 5 mm
was reported to quantify and compare the ductile plateau after the breakdown of
chemical bond. The contact area is determined from the perimeter of the profiled
steel sheets in contact with the concrete and the height of the concrete specimen. A
sample calculation for bond stress is provided in Appendix A.
3.3.5.2 Push test specimens PT-500-0 and PT-500-3
The load-slip curves for these two specimens are shown in Figure 3.12. The
failure of the chemical bond was evidenced by a cracking sound, followed by the
reduction in load carrying capacity and progression of slip until the limiting slip had
been reached to terminate the test. A sudden drop of bond stress is due to the loss of
chemical adhesion between the concrete and the profiled steel sheeting. The rise in
shear stress with increasing slip is due to either the experimental setup or the
geometry of the profile sheeting itself. If the rise in shear stress is due to the
experimental setup, the possible mechanism can be the increased restraint from the
support of the sheeting. This restraint was minimised by placing the support on the
roller to allow the movement of the sheeting relative to the concrete. It is more likely
that increase in shear capacity after breakdown of the chemical bond is due to
increase in roughness of the profile-concrete interface as the slip progresses (Burnet,
1998).
55
PT-500-0 specimen carried a maximum load of 60 kN, with one side measuring
over 2.5 mm more slip than the other. This resulted from the breakdown of the
chemical bond and the initiating slip on one side of the specimen shown in Figure
3.13, before the chemical bond had fully broken on the other side. PT-500-3
specimen recorded a peak load of 47 kN, which resulted in a 20% reduction in
maximum load.
The load-slip curves for these two specimens have been shown in Figure 3.14.
PT-1000-0 specimen suffered pre-delamination before the test at one corner to some
extent during handling of the specimen. This specimen recorded a maximum load of
115 kN before failure.
PT-1000-3 specimen recorded a maximum load of 137 kN which was not
expected compared with other push specimens. This rise in shear capacity resulted
from a buckling at the base of the profiled steel sheets on one side of the specimen
that occurred during loading before complete failure of the chemical bond. The
deformed shape of the profiled steel sheeting is shown in Figure 3.15.
These specimens were handled very carefully using some clamps and no
delamination was observed before the test. The load-slip curves for these two
specimens have been shown in Figure 3.16. PT-2000-0 specimen recorded a
maximum load of 240 kN before failure. One linear potentiometer dropped due to
non-uniform distribution of loading at top of the specimen and this is shown in
Figure 3.17. So slip on one side only is reported for analysis purposes.
56
PT-2000-3 specimen reached a maximum load of 175 kN before failure which is
lower than that of the nonprestressed specimen PT-2000-0. Some local buckling was
also observed at the lower portion of the specimen after 5 mm slip occurred. The
effect of this local buckling was not considered effective in reducing the chemical
bond characteristics as it occurred after a slip of 5 mm had reached.
3.3.5.5 Discussion of preliminary test results
The chemical bond strength, residual strength and average mechanical bond
strength have been reported in Table 3.5. In general, this type of push specimen is
considered to transfer longitudinal shear across the concrete-sheeting interface by
chemical adhesion prior to slip and by mechanical bond after the commencement of
slip. If the chemical bond stress and mechanical bond stress depended purely on the
chemical adhesion between the concrete and the profiled sheeting and interface
interlock respectively, then it would be expected that the chemical bond stress and
mechanical bond stress would be constant and independent of the level of prestress.
But this is not the case. The comparison of bond stress-slip curves for six different
push specimens tested under a static load is shown in Figure 3.18. It can be seen
from this figure that the chemical bond appears to vary significantly with the
prestress. The average chemical bond stress in post-tensioned push specimens is
found to be approximately 20% lower than that in push specimens without post-
tensioning. Also the residual stress and mechanical bond stress in post-tensioned
push specimens is found to be lower than that of push specimens without post-
tensioning up to a slip of 5 mm. The reason for the decrease in bond stress may be
attributed to the shortening of the concrete block due to prestress and redistribution
of interface stress.
57
3.4 MAIN SERIES
A total of thirty two push test specimens were prepared and tested as a main series
of the experimental programme. In this series, twenty four specimens were post-
tensioned and the bond length H of the push-specimens was varied as 500 mm, 1000
mm, 1500 mm and 2000 mm. A summary of the specimens for the main series are
provided in Table 3.6. For each configuration, there were two identical samples and
these samples were differentiated by adding A and B after the specimen reference
“PT-H-P”.
All push specimens were cast on the same day with a target concrete strength of
25 MPa. The details of the constituents of the concrete mix used in the main series
are reported in Table 3.7. A total of 27 cylinders were tested over a period of 63
days. The material properties of the concrete were determined from standard cylinder
tests as reported in Table 3.8. The concrete compressive strengths are presented in
Table 3.9. The properties of the profiled steel sheets and prestressing bar were
obtained from standard tensile tests and are presented in Table 3.8.
3.4.2 Specimen preparation, post-tensioning and instrumentation
The specimens were prepared in the same way as in the preliminary series except
some improvement. However, the specimen was poured in an inclined position so
that no air trapped can help predelamination of the profiled steel sheet from concrete.
Sixteen specimens were post-tensioned at day 7 after casting with a concrete strength
measured at day 7 of 16.0 MPa and the remaining eight specimens were post-
tensioned at day 8. Post-tensioning was carried out by elongating the 29 mm

58
threaded bar from the live anchor. A hydraulic wrench was used instead of a normal
wrench to tighten the live anchor and a load cell was used to measure the applied
load to the bars during the post-tensioning while three strain gauges fixed on the bar
were used to record the strains within the bar. The specimens were equipped with
steel strain gauges (FLA-5-11-3LT), concrete strain gauges (PML-60-3LT) and
linear potentiometers in the same way as in the preliminary series. However, two
extra linear potentiometers were also attached to the specimen in this series to
measure the vertical separation of the sheets. A data acquisition system was used to
record the readings of the load cell, linear potentiometer and strain gauges.
3.4.3 Test setup
The specimens for the main series were set up for testing in the same way as in
preliminary series except some minor modification to boundary conditions. At the
base of the profiled steel sheet a filling of 20 mm thick plaster and some steel
packing was provided to prevent non-uniform local buckling and minimise the
distortion of the sheeting at the end of the specimen. Most of the specimens were
successful in preventing buckling while some specimens failed. It was observed that
the sheeting corner has some distortions during the cutting procedure. Since one base
is on roller, this constraint will not prevent the sheeting to deform away from the
concrete.
The load was applied to the specimen very slowly at a rate of 0.005 mm/sec until
the chemical bond had completely broken as indicated by the linear potentiometer
recording a slip displacement. After failure of the chemical bond, the load was
increased gradually. The load and slip were logged continuously. The test was
59
terminated when either the specimen was unable to take any load or the limit of slip
had been reached.
3.4.4 Summary of main series
3.4.4.1 General
The aim of the testing programme was to investigate the qualitative effects of
varying the level of prestress for a given profile and to determine general trends in
the bond stress-slip characteristics of the profiled sheets. The overall bond stress-slip
behaviour of each sample was considered as reasonable only when it agrees with the
general patterns found in the existing literature.
3.4.4.2 Push test specimens of 500 mm length
3.4.4.2.1 PT-500-0
The bond stress-slip curves of these specimens of 500 mm are shown in Figure
3.19. PT-500-0-A suffered premature delamination to some extent although
sufficient measures were adopted (Figure 3.20). It is evident from its reduced load
carrying capacity and the uneven slip between two sides of the specimen (Figure
3.21). Premature delamination may have occurred in this form of push test for two
main reasons: 1) Geometric imperfection of the profiled steel sheets resisting
development of a uniform bond across the entire concrete-sheet interface, and 2)
Movement of the specimens during various phases of experimental work disturbing
the bond directly. PT-500-0-B achieved a maximum load of 92 kN with a very small
average slip of 0.07 mm. The measured slip on both sides of the specimen was within
acceptable ranges (Figure 3.22). Therefore, instead of taking the average of the two
60
samples, PT-500-0-B was considered more representative of the bond stress-slip
behaviour for further analysis.
3.4.4.2.2 PT-500-1
The bond stress-slip curves of these post-tensioned specimens of 500 mm length
and 1 MPa prestress are shown in Figure 3.23. PT-500-1-A showed similar
behaviour of PT-500-0-A with a double peak in the chemical bond stress. This
increase in the shear capacity after breakdown of the chemical bond may have been
due to the increase in roughness of the concrete-sheet interface as slip progressed. As
soon as chemical adhesion is lost, the interface interlock component of the
mechanical shear capacity of ribs of the profiled steel sheets transfers the
longitudinal shear by inducing passive normal and shear forces across the concrete-
sheet interface (Burnet, 1998). Due to similar reasoning as in PT-500-0-A, PT-500-1-
B was reported for further analysis.
3.4.4.2.3 PT-500-2.4
Figure 3.24 presents the overall bond stress-slip behaviour of the two samples of
PT-500-2.4. The average values of the chemical and mechanical bond stresses were
recorded for comparison. One of the reasons for the significant deviation of shear
stresses could be the difference of slip from one side to another, as shown in Figure
3.25. This resulted from the breakdown of the chemical bond and the initiation of
slip on one side of the specimen, before the chemical bond had fully broken on the
other side.
61
3.4.4.2.4 PT-500-3
The overall bond stress-slip behaviour of the two samples of PT-500-3 is shown
in Figure 3.26. There was no buckling or premature delamination observed in any of
these specimens. So the average values of the chemical and mechanical bond stresses
were recorded for further investigation. However, there was some unusual behaviour
of a sudden increase and decrease in shear stresses which was observed towards the
end of the test of specimen PT-500-3-B. The effect of this behaviour was not
considered significant in the overall bond characteristics as it occurred after a slip of
7-9 mm had been reached. PT-500-3-A and PT-500-3-B achieved a maximum load
of 66 kN and 74 kN with a very small average slip of 0.65 and 0.03 mm respectively.
These two samples with 3 MPa prestress showed a 24% reduction in load capacity
when compared with non-post-tensioned sample PT-500-0-B.
3.4.4.3.1 PT-1000-0
The possibility of buckling at the base of the profiled steel sheets was taken into
account for all the specimens of length above 500 mm. Plastering was provided at the
base of the profiled steel sheets of these specimens in a way so that it did not restrain
the outward movement of the profiled steel sheets. The overall bond stress-slip
behaviour of PT-1000-0-A was found to be in agreement with general pattern of the
bond stress-slip behaviour, as shown in Figure 3.27. But PT-1000-0-B exhibited
buckling at the base of the profiled steel sheets on one side of the specimen that
occurred before achieving the maximum load recorded ((Figure 3.28). Therefore, the
bond stress-slip curve of PT-1000-0-A was used for further investigation.
62
3.4.4.3.2 PT-1000-1
in (Figure 3.29). Although PT-1000-1-B showed significant variation in maximum
bond stress compared with identical sample PT-1000-1-A, there was no buckling
observed in any of these specimens. Furthermore, both specimens failed by slipping
simultaneously on both sides of the specimens. So the average values of the chemical
and mechanical bond stresses are recorded for comparison.
3.4.4.3.3 PT-1000-2.4
PT-1000-2.4-A specimen showed a reasonable bond stress-slip curve with a
complete breakdown of the chemical bond on both sides of the specimen and this
was followed by interface interlock shear transfer without any signs of buckling
(Figure 3.30). Due to data logger problems after 1.2 mm slip during the testing, a
complete bond stress-slip curve could not be reported for specimen PT-1000-2.4-B
(Figure 3.31). Therefore, only the chemical bond stress for PT-1000-2.4-B has been
considered. Finally, the average chemical bond stress of both specimens and the
mechanical bond stress of PT-1000-2.4-A were used for further analysis.
3.4.4.3.4 PT-1000-3
in Figure 3.32. The significant difference between the bond stresses of both samples
indicates that PT-1000-3-A might suffer premature delamination in some inner part
of the concrete-sheet interface which was difficult to observe. However, there was no
sign of buckling in the profiled steel sheets and failure mode of both specimens
corresponds to the desired amount of slip. For consistency in the analysis and
63
comparison of the test results, the average values of the chemical and mechanical
bond stresses are recorded for comparison.
3.4.4.4.1 PT-1500-0
Two samples of PT-1500-0 exhibited buckling at the base of the profiled sheets
on one side of the specimen. It appeared that the buckling was due to either the bond
capacity exceeding the bearing capacity of the sheets or pre-distortion at the corner
of the sheets which might occur during the cutting, transporting or handling phases
which thus facilitated premature buckling (Figure 3.33). The first signs of buckling
were evident when the shear stress reached approximately 0.122 MPa in specimen
PT-1500-0-B whereas PT-1500-0-A exhibited buckling earlier (Figure 3.34).
Premature buckling prevented the specimens from achieving their bond capacity
(Figure 3.35). Therefore, the bond stress-slip curve of PT-1500-0-B was used for
further investigation and qualitative assessment only.
3.4.4.4.2 PT-1500-1
The bond stress-slip behaviour of the two samples of PT-1500-1 obtained from
the tests is presented in Figure 3.36. The results of these two samples correspond
closely to each other in the region of the chemical bond and there was no buckling
observed in any of these specimens. The variation in the mechanical bond stress-slip
behaviour could be attributed to the variation in the passive interface forces induced
in the steel-concrete interface. So the average values of the chemical and mechanical
bond stresses were recorded for comparison.
64
3.4.4.4.3 PT-1500-2.4
Figure 3.37 compares the overall bond stress-slip characteristics of PT-1500-2.4-
A and PT-1500-2.4-B. Steel bar packing at the base of the profiled steel sheets was
placed instead of plastering to prevent the buckling. Despite the use of the steel bar
restraints with clamping, PT-1500-2.4-B exhibited buckling at the base of the
profiled steel sheets on one side of the specimen (Figure 3.38). But it occurred after
achieving the maximum load observed. Therefore, the average of the chemical bond
stresses of both specimens and the mechanical bond stress of PT-1500-2.4-A was
used for further investigation.
3.4.4.4.4 PT-1500-3
The bond stress-slip curves of PT-1500-3-A and PT-1500-3-B are presented in
Figure 3.39. But the filling of plaster at the base of the profiled steel sheets could not
prevent buckling on one side of the specimen PT-1500-3-B (Figure 3.40). This
occurred after achieving the maximum load which was very close to the load
carrying capacity of its identical specimen PT-1500-3-A. Therefore, the average
values of the chemical bond stresses of both specimens and the mechanical bond
stress of PT-1500-3-A were used for further investigation.
3.4.4.5.1 PT-2000-0
Figure 3.41 compares the overall bond stress-slip behaviour of the two samples of
PT-2000-0. The results of these two samples correspond closely to each other in both
chemical and mechanical bond behaviour However, buckling was observed only at
65
the end of the both tests despite using buckling prevention measurements (Figure
3.42). A small increase in the bond stress-slip behaviour after a slip of 7.5 mm as
observed in Figure 3.38 could be attributed to the combined effect of the buckling
and stress redistribution at concrete-steel interface. So the average values of the
chemical and mechanical bond stresses are recorded for comparison.
3.4.4.5.2 PT-2000-1
The bond stress-slip curves of these specimens of 2000 mm are shown in Figure
3.43. Two samples of specimen PT-2000-1 exhibited a mixed mode of bond failure
and buckling. Local buckling of the profiled steel sheets occurred usually on one side
and thus the applied load becomes eccentric and causes the bond failure on other
side. Therefore, the chemical bond stress found with these specimens can be
considered as the lower bound. Sample PT-2000-1-A exhibited buckling at a
significantly lower load from that of the specimen PT-2000-1-B (Figures 3.44 &
3.45). Therefore, bond stress-slip curve of PT-2000-1-B was considered more
accurate compared with other specimens.
3.4.4.5.3 PT-2000-2.4
Figure 3.46 presents the bond stress-slip curves of these post-tensioned specimens
of 2000 mm. Two samples of specimen PT-2000-2.4 first exhibited a sign of bond
failure and then continued with buckling at one side of the profiled steel sheets
before complete breakdown of the chemical bond (Figure 3.47). However, the results
of these specimens of maximum length in this test series were important to
understand the effect of the bond length. Therefore, the average bond stress was
considered for the lower bound of the chemical bond.
66
3.4.4.5.4 PT-2000-3
The bond stress-slip curves of these identical specimens of 2000 mm length and 3
MPa prestress are shown in Figure 3.45. PT-2000-3-A and PT-2000-3-B also
exhibited a mixed mode of bond failure and buckling on one side of the specimen
(Figures 3.49 & 3.50). The chemical bond stress-slip behaviour resembles each other
while the mechanical bond deviates which could be attributed to the combined effect
of buckling and possible stress redistribution at the steel-concrete interface.
Therefore, average bond stress was considered for the lower bound of chemical bond
stress for further analysis.
3.4.5 Analysis of test results
The chemical bond stress, mechanical bond stress and bond stress at 5.0 mm slip
are reported in Table 3.10. The effort has been made to present a qualitative
assessment and identify general trends of the test results. Since the specimens of
2000 mm height showed premature buckling in general prior to bond failure, the test
results were analysed very carefully and general trends obtained with other samples
were considered acceptable.
A step by step procedure of detecting the effect of prestress on the chemical and
mechanical bond stress is as follows:
1) The overall load-slip curve for each specimen was recorded from the push test
and then the chemical and mechanical bond stresses were calculated according to the
procedures described in Section 3.3.5.1. The values are tabulated for all specimens in
Table 3.10.
67
2) The overall bond stress-slip behaviour of each specimen was investigated and
presented in Section 3.4.4. For each bond length, the average and/or acceptable
chemical bond stress and mechanical bond stress were plotted against the level of
prestress on Figures 3.51, 3.52 and 3.53 to determine the effect of prestress.
3) A regression analysis was conducted and values of R2 are reported.
The role of buckling of the profiled sheeting can be a significant issue in
continuous post-tensioned composite slabs. But the push test presented in this study
represents simply supported post-tensioned composite slabs. The buckling of the
profiled sheeting in push tests occurred due to the test setup and corner distortions of
the profiled sheeting. Some push specimens exhibited failure by buckling or a mixed
mode of bond failure and buckling on one side of the specimen. The bond-slip
behaviour of those specimens was investigated and reasons for buckling were
discussed in Section 3.4.4.
3.4.5.1 Effects of prestress
To understand in detail the effects of prestress on the bond stress-slip relationship,
various components of the bond stress-slip curve were analysed. The variations of
the chemical bond stress for different heights of push specimens are plotted against
the level of prestress in Figure 3.51. It can be seen that the chemical bond stress
appears to decrease with increasing level of prestress. Since the chemical bond is
brittle in nature, as soon as failure starts at one location of the interface, the
breakdown of the bond propagates rapidly across the entire interface. The reason for
the decrease in bond stress may be attributed to the shortening of the concrete block
due to prestress. Consequently, a redistribution of the shear stress happens in the
68
concrete-steel interface in prestressed specimens. Due to significant scatter of the
results it is difficult to quantify accurately the actual reduction of the chemical bond
stress. However, the values of coefficients of determination R2 for three out of four
regression lines also indicate that the fitted regression lines provide reasonable
approximations to the general trend of reduction in the chemical bond stress for this
particular profile. The mechanical bond stress and bond stress at 5.0 mm slip also
exhibited a similar trend as the chemical bond stress, as plotted in Figure 3.52 and
Figure 3.53 respectively. It would therefore appear that prestress is one of the
controlling factors of the bond stress in post-tensioned push specimens.
3.4.5.2 Effects of bond length
The effects of bond length on the chemical bond stress, mechanical bond stress
and bond stress at 5.0 mm have been shown in Figures 3.54, 3.55 and 3.56
respectively. From these figures, it is clear that the bond stress-slip behaviour is also
sensitive to the bond length of the specimen. The values of R2 for most of the
regression lines for the various levels of prestress indicate a good correlation with the
test data. It would therefore appear that the bond stress behaviour of push specimens
is not constant with the bond length.
3.5 DESIGN GUIDANCE
Despite the extensive use of post-tensioned composite slabs for building
applications, no design guidelines are currently available. To date, designers have
ignored the effect of prestress on the bond stress-slip behaviour of post-tensioned
composite slabs and based their composite design on the bond stress-slip relations
69
suitable for conventional composite slabs. This thesis identified and quantified this
effect of composite action in this form of construction accurately and presented in a
form that can be used as a guideline in the design of post-tensioned composite slabs
and development of new forms of profile.
It is important to determine the longitudinal shear bond strength in the post-
tensioned composite slabs to calculate the ultimate strength accurately. The bond
stress-slip between the concrete slab and the profiled steel sheeting comprises of both
chemical and mechanical bond strengths. The bond strength at 5.0 mm slip can be
used as an indicator of ductility of these types of slabs. This study evaluated the
effect of prestress on each of these components of the bond stress-slip behaviour
between the profiled steel sheets and the concrete in post-tensioned composite slabs.
The design guidance presented herein applies to simply supported post-tensioned
composite slabs.
The average chemical bond strength, mechanical bond strength and bond strength
at 5.0 mm slip are given by Equations 3.2, 3.3 and 3.4 respectively.
  0.2  0.02P (3.2)
  0.1  0.01P (3.3)
  0.1  0.01P (3.4)
Where 𝜏 is the average bond stress and P is the level of prestress and 𝜏 and P are in
MPa.
The above mentioned equations are linear representation of the effects of prestress on
the chemical bond strength, mechanical bond strength and bond strength at 5.0 mm
slip, as shown in Figures 3.51, 3.52 and 3.53. However, these equations are valid
70
only for 1.00 mm thick Condeck profile and level of prestress of 1 to 3 MPa.
Moreover, it is suggested that these equations should only be used as rule of thumb
guides.
The bond which must be achieved between the profiled steel sheets and the
concrete is very crucial to develop composite action of post-tensioned composite
slabs and therefore, this must be assessed accurately to determine the ultimate
strength of these types of slabs. This chapter presented the results of a preliminary
and main series of push test study in order to investigate the effects of prestress on
the bond stress-slip behaviour. This experimental programme provides a valuable
insight into the mechanism of longitudinal shear transfer between the concrete and
the profiled steel sheets. The following conclusions can be drawn from this chapter:
 An experimental parametric study of various levels of prestress and bond
length has been completed using thirty two push specimens. It was found that
both the chemical and mechanical bond stresses appear to decrease with an
increase in the level of prestress. Similarly, the overall bond stress-slip
behaviour is also sensitive to the bond length. This study suggests
consideration of the effect of prestress and bond length on the bond stress-slip
behaviour for accurate determination of ultimate flexural capacity of post-
tensioned composite slabs.
 This study also offers an optimal design of the push test specimen. The
premature buckling of the specimen suggests that these test procedures should
71
not be used in the determination of the bond stress for specimens of height
above 1500 mm due to buckling of the base of the sheets for sheet thicknesses
of 1 mm.
72
Table 3.1 Details of push specimens of preliminary series

Specimen L W H Effective Cross- Concrete stress
ID (mm) (mm) (mm) prestress sectional due to prestress
(kN) area (MPa)
(mm2)
PT-500-0 300 180 500 0 54000 0.00
PT-1000-0 300 180 1000 0 54000 0.00
PT-2000-0 300 180 2000 0 54000 0.00
PT-500-3 300 180 500 162 54000 3.00
PT-1000-3 300 180 1000 162 54000 3.00
PT-2000-3 300 180 2000 162 54000 3.00
Table 3.2 Details of concrete mix of preliminary series

Material Quantity
(per m3)
Coarse sand 510 kg
Fine sand 350 kg
10 mm aggregate 343 kg
Cement 300 kg
Fly ash 86 kg
Water 105 L
Water reducer 900 ml
73
of preliminary series
Material Property Value
(MPa)
Concrete Compressive strength, 𝑓𝑐′ 40.8
Tensile strength, fct 5.1
Young’s modulus, Ec 35,000
Profiled Yield stress, fy 670

steel sheets
Ultimate strength, fu 680
Young’s modulus, Es 217,000
Prestressing Yield stress, fy 1050
steel
Table 3.4 Results of concrete compressive strength test for preliminary series
Age Compressive Compressive Compressive Average
(Days) strength strength strength compressive
(MPa) (MPa) (MPa)
strength
Sample 1 Sample 2 Sample 3
(MPa)
7 22.1 22.6 23.1 22.6
14 32.4 31.8 33.1 32.5
21 36.6 36.1 37.5 36.8
28 38.8 43.8 39.7 40.8
34 42.6 40.5 42.1 41.7
74
Table 3.5 Summary of bond stress of push specimens of preliminary series

Specimen Maximum Chemical Residual Average
ID load before bond stress mechanical
breakdown of stress (MPa) bond stress up
chemical bond (MPa) to a slip of 5
(kN) mm
(MPa)
PT-500-0 60.61 0.26 0.24 0.24
PT-1000-0 115.03 0.25 0.22 0.23
PT-2000-0 233.43 0.25 0.20 0.20
PT-500-3 47.36 0.21 0.19 0.19
PT-1000-3 137.74 0.30 0.27 0.27
PT-2000-3 182.18 0.20 0.18 0.19
Table 3.6 Details of push test specimens of main series

Specimen L W H Effective Cross- Concrete
ID (mm) (mm) (mm) prestress sectional stress due
(kN) area to prestress
(mm2) (MPa)
PT-500-2.4 300 180 500 129.6 54000 2.40
PT-1000-2.4 300 180 1000 129.6 54000 2.40
PT-1500-2.4 300 180 1500 129.6 54000 2.40
PT-2000-2.4 300 180 2000 129.6 54000 2.40
PT-500-1 300 180 500 54 54000 1.00
PT-1000-1 300 180 1000 54 54000 1.00
PT-1500-1 300 180 1500 54 54000 1.00
PT-2000-1 300 180 2000 54 54000 1.00
PT-500-3 300 180 500 162 54000 3.00
PT-1000-3 300 180 1000 162 54000 3.00
PT-1500-3 300 180 1500 162 54000 3.00
PT-2000-3 300 180 2000 162 54000 3.00
PT-500-0 300 180 500 0 54000 0
PT-1000-0 300 180 1000 0 54000 0
PT-1500-0 300 180 1500 0 54000 0
PT-2000-0 300 180 2000 0 54000 0
75
Table 3.7 Details of concrete mix of main series

Material Quantity
(per m3)
Coarse sand 546 kg
Fine sand 350 kg
Cement 203 kg
Fly ash 108 kg
Water 98 L
Water reducer 885 ml
of main series
Material Property Value
(MPa)

steel sheets
steel
76
Table 3.9 Results of concrete compressive strength test for main series
Age Compressive Compressive Compressive Average
(Days) strength strength strength compressive
(MPa) (MPa) (MPa)
strength
Sample 1 Sample 2 Sample 3
(MPa)
7 15.2 14.6 15.4 15.0
8 16.8 16.6 16.4 16.6
14 19.0 18.8 18.5 18.8
21 20.7 20.8 22.1 21.2
28 21.9 22.1 24.3 22.8
34 24.9 24.2 25.5 24.9
41 26.2 27.0 24.3 25.8
50 22.8 25.9 25.8 24.8
63 29.7 30.7 28.8 29.7
77
Table 3.10 Test results of main series

Specimen Level of Bond Chemical Mechanical Stress at
prestress length bond stress bond stress 5.0 mm
ID
(kN) (mm) (MPa) (MPa) slip
(MPa)
PT-500-0 0 500 0.199 0.129 0.115
PT-500-1 54 500 0.199 0.129 0.111
PT-500-2.4 129 500 0.162 0.111 0.102
PT-500-3 162 500 0.151 0.105 0.095
PT-1000-0 0 1000 0.157 0.129 0.118
PT-1000-1 54 1000 0.169 0.128 0.115
PT-1000-2.4 129 1000 0.168 0.107 0.098
PT-1000-3 162 1000 0.147 0.117 0.113
PT-1500-0 0 1500 0.147 0.103 0.101
PT-1500-1 54 1500 0.133 0.105 0.096
PT-1500-2.4 129 1500 0.137 0.100 0.091
PT-1500-3 162 1500 0.125 0.088 0.076
PT-2000-0 0 2000 0.138 0.111 0.104
PT-2000-1 54 2000 0.125 0.102 0.084
PT-2000-2.4 129 2000 0.109 0.080 0.080
PT-2000-3 162 2000 0.116 0.083 0.078
78
(a) Isometric view (b) Plan view
(c) End view (d) Side view
Figure 3.1 Geometry of the push specimen
79
Plan View Side view
(a) Details of end plate
End view Side view
(b) Details of parallel flange channel
Plan view Side view
(c) Details of support block
Figure 3.2 Details of end plate, channels and support block
80
Figure 3.3 Geometry of Stramit Condeck HP®
(a) (b)
Figure 3.4 Formwork ready for pouring (a) without post-tensioning (b) with post-
tensioning.
81
Figure 3.5 Post-tensioning
Figure 3.6 Load-strain curve for PT-500-3
82
Plan view
Cross-sectional view (AA) End view Side view
Figure 3.7 Instrumentation of push tests
Figure 3.8 Linear potentiometers for measurement of slip and separation
83
Top view
Side view
End view
(a) Schematic of push tests (b) Details of loading jig
Figure 3.9 Push test setup
(a) End view (b) Side view

Figure 3.10 Loading plate
84
Figure 3.11 Push test specimen ready for testing
Figure 3.12 Bond stress-slip curves of PT-500-0 & PT-500-3
85
Figure 3.13 Initiation of slip from one side
86
Figure 3.15 Local buckling of the profiled steel sheeting
87
Figure 3.17 Linear potentiometer dropped
Figure 3.18 Comparison of bond stress-slip behaviour of push specimens
88
Figure 3.19 Bond stress-slip curves of PT-500-0
Figure 3.20 Predelamination of PT-500-0-A
89
Figure 3.21 Failure of PT-500-0-A through uneven slip
Figure 3.22 Failure mode of PT-500-0-B
90
Figure 3.24 Bond stress-slip curves of PT-500-2.4
91
Figure 3.25 Bond stress-slip curves of both sides of PT-500-2.4-A
92
Figure 3.28 Failure of PT-1000-0-B through buckling
93
Figure 3.30 Failure pattern of PT-1000-2.4-A
94
95
Figure 3.33 Corner distortion of the profiled steel sheeting
96
Figure 3.35 Buckling of PT-1500-0-B
97
Figure 3.38 Buckling of PT-1500-2.4-B
98
99
Figure 3.42 Buckling of PT-2000-0-A
100
Figure 3.44 Buckling and bond failure of PT-2000-1-A
101
Figure 3.45 Buckling and bond failure of PT-2000-1-B
102
Figure 3.47 Buckling pattern observed after the test
103
Figure 3.49 Buckling of PT-2000-3-A
104
Figure 3.51 Variation in chemical bond stress due to prestress
Figure 3.52 Variation in mechanical bond stress due to prestress
105
Figure 3.53 Variation in bond stress at 5.0 mm due to prestress
Figure 3.54 Variation in chemical bond stress with bond length
106
Figure 3.55 Variation in mechanical bond stress with bond length
Figure 3.56 Variation in bond stress at 5.0 mm with bond length
107
108
CHAPTER 4 - Finite Element Modelling of Push Tests
CHAPTER 4
FINITE ELEMENT MODELLING OF PUSH TESTS
4.1 INTRODUCTION
Numerical modelling and analysis of composite structures using the finite element
method have become a very popular area of research. Due to the complex interaction
between the concrete and the profiled steel sheets, the behaviour of composite steel-
concrete slabs is very challenging to be modelled accurately. Moreover, accurate FE
modelling of composite slabs requires shear-bond slip relationship to be obtained
from push tests. In order to develop an accurate FE model of post-tensioned
composite slabs, the need for accurate shear stress-slip behaviour is obvious. This
chapter presents a three dimensional nonlinear finite element modelling to simulate
the bond stress-slip behaviour that exists between the concrete and the profiled steel
sheets in a post-tensioned push specimen. The primary objective is to demonstrate
that incorporating the effect of prestress on local bond-slip behaviour will produce
satisfactory results for simulation of the global bond stress-slip behaviour of post-
tensioned composite slabs. In the following sections, development of FE model of
post-tensioned push specimens is presented. Next, the use of shear interaction
109
property by linear interpolation in local bond stress-slip law is demonstrated. Finally,
the results of FE analyses and parametric studies are presented and discussed.
4.2 FINITE ELEMENT MODEL FORMULATION
4.2.1 General
The experimental programme described in the previous chapter provided the bond
stress-slip relationship for a particular profiled steel sheet and for certain levels of
prestress. In this study, the finite element program ABAQUS was used to generalise
the test results of the bond stress-slip behaviour of a post-tensioned push test
specimen. ABAQUS/Explicit solver was chosen in this numerical study because it
can solve static problems with complicated contact problems more readily and
efficiently than ABAQUS/Standard does. Although ABAQUS/Standard is suitable
for a nonlinear static problem, numerical convergence often occurs due to highly
nonlinear material failure behaviour (ABAQUS, 2012). In the dynamic explicit
analysis method, the global mass and stiffness matrices need not to be formed and
inverted so each increment is relatively inexpensive compared to implicit analysis.
The size of the increment is determined based on the mesh size and material
properties. Since ABAQUS/ Explicit is a dynamic analysis procedure, a quasi-static
solution was sought by applying load slowly and checking the energy balance of the
finite element model. Both geometric and material nonlinearity were considered in
the FE analysis. Appropriate material models for all components and suitable contact
interactions with real boundary conditions were specified to model all test specimens
accurately and efficiently.
110
4.2.2 Model geometry and element types
Due to the symmetry of geometry and loading, only a quarter of the push
specimen was modelled. 8-node linear brick, reduced integration elements (C3D8R)
and 4-node doubly curved shell, reduced integration elements (S4R) were used to
model the concrete and the profiled steel sheets respectively. C3D8R is used
extensively to model nonlinear simulation involving contact between deformable
bodies. A good mesh of hexahedral elements with reduced integration usually
provides the best results in three-dimensional analyses for the minimum cost. S4R is
considered as the most appropriate type of element in the numerical modelling of
thin walled steel structures as they are more accurate and efficient in contact
modelling than conventional shell element (ABAQUS, 2012). S4R employs two-
sided contact by taking into account changes in thickness. An overview of the finite
element mesh is shown in Figure 4.1.
4.2.3 Interactions
The profiled sheeting elements were located at a distance of half of the steel
sheeting thickness from the concrete elements. The interaction between the concrete
and the profiled steel sheets was modelled with radial-thrust type connector elements
(CONN3D2). The reason for using radial thrust type elements is the ability to specify
the radial and thrust displacement separately. In this case, the radial displacement
refers to the slip between the concrete and the profiled steel sheets and the thrust
displacement refers to the separation between the concrete and the profiled steel
sheets. Each connector element connected a pair of concrete and steel nodes that
were closest to each other (Figure 4.2). The bond stress-slip curves derived from the
111
push tests were converted into the form of load versus slip in order to assign to the
radial components of connectors. The lateral separation was not simulated in the FE
model assuming that its effect is implicitly present in the local bond-slip law. Since
the local bond-slip relationship was derived from push tests, this assumption is
reasonable. Therefore, a relatively stiff value was assigned to the thrust components
of the connectors in order to prevent the overlapping of the concrete and the steel
elements during loading.
The procedure for modelling connectors involves the following steps (ABAQUS,
2012):
1) Create reference points and datum coordinate systems.
2) Create assembly-level wire features.
3) Create connector sections to define the connection types, connector behaviour,
and section data.
4) Create a connector section assignment that associates the connector section
with the selected wires and that specifies the orientations for the first and second
points of the selected wires.
5) Prescribe connector loads and connector boundary conditions.
6) Create field and history output requests for connectors.
4.2.4 Loading and boundary conditions
Prestressing was defined by a predefined stress field. In order to create an initial
stress field for the concrete, stress components were defined by selecting direct
112
specification in ABAQUS. An initial stress field can only be created and modified in
the initial step. The following procedure was followed to create an initial stress field
for concrete part of the model. 1) A new stress field was created under the category
of “Mechanical” and the types for selected step of “stress”.
2) “Direct Specification” option was chosen to specify six component values of
stress. In this case, prestress was acting in z direction only. Therefore, values of
stress component in other directions were zero. ABAQUS uses arrows to represent
each component of the prescribed conditions in the viewport. Appropriate arrows
ensure the application of stress in the desired direction.
A uniformly distributed load was applied throughout the top surface of concrete
using the displacement control. A roller support was applied to the nodes of the
bottom end of steel sheets. XSYMM (surface 1) and YSYMM (surface 2) symmetry
boundary conditions were applied to simulate the continuity of the push test
specimen in both axes as shown in Figure 4.1.
4.2.5 Solution method
In accordance with ABAQUS (2012), some special considerations are required to
apply the explicit dynamic procedure to quasi-static problems. A static solution is a
long-time solution. Therefore, it is often computationally inefficient to simulate an
event in its natural time scale. Because it would require an excessive number of small
time increments. The event must be accelerated in some way to obtain an economical
solution. However, as the event is accelerated, the state of static equilibrium evolves
into a state of dynamic equilibrium in which inertial forces become more dominant.
113
ABAQUS (2012) suggests modelling the process in the shortest time period in which
inertial forces remain insignificant.
For each result, internal (ALLIE) and kinetic (ALLKE) energies of the whole
model were compared to ensure a quasi-static response. It was found that the kinetic
energy level was significantly below the internal energy level in all simulations, as
shown in Figure 4.3. However, oscillating response in loading is unavoidable in FE
models conducted using ABAQUS/Explicit. A smoothing function named
butterworthFilter (X, F) available in ABAQUS was used to eliminate oscillation
effect from the final results.
4.2.6 Constitutive laws
4.2.6.1 Concrete properties
In this study, the concrete damaged plasticity model available in ABAQUS was
used to model the concrete material. This model is suitable for both static and
dynamic analyses. Two main failure mechanisms in the form of tensile cracking and
compressive crushing of concrete element are incorporated in this model. The
evolution of yield surface is controlled by two hardening variables causing failure
under tensile and compressive loading. Post-failure behaviour under compression is
defined by a softening stress-strain response. The strain softening behaviour of
cracked concrete in tension is specified by tension stiffening. Tension stiffening can
be specified either by means of post-failure stress-strain behaviour in tension or by
using a fracture energy cracking criterion. The concrete properties used in the FE
model are reported in Table 4.1.
114
The elastic properties were specified according to ACI 318 (American Concrete
Institute, 2011). All plasticity parameters were specified very carefully. The dilation
angle is taken as 30º to define plastic flow potential. Default values in ABAQUS
were adopted for the plastic flow potential eccentricity (e), the ratio of the second
stress invariant on the tensile meridian to that on the compressive meridian (Kc) and
viscosity parameter. The ratio of equibiaxial concrete strength (𝑓𝑏𝑜 ) to concrete
compressive strength (𝑓𝑏𝑜 ⁄𝑓𝑐′ ) is calculated based on Equation 4.1 as proposed by
Papanikolaou and Kappos (2007).
𝑓𝑏𝑜
′
= 1.5 𝑓𝑐′ − 0.075 (4.1)
𝑓𝑐
Plain concrete in uniaxial compression was modelled as an elastic-plastic material
with strain softening, as shown in Figure 4.4. The stress-strain curve for concrete
proposed by Carreira and Chu (1985) was adopted as Equation 4.2.
𝑓𝑐′ 𝛾(𝜀𝑐 ⁄𝜀𝑐′ )

𝜎𝑐 = (4.2)
𝛾 − 1 + (𝜀𝑐 ⁄𝜀𝑐′ )𝛾
Where σc is the compressive stress in concrete, εc is the strain in concrete, 𝑓𝑐′ is the
cylinder compressive strength of concrete, 𝜀𝑐′ is the strain corresponding to 𝑓𝑐′ and γ
is given by Equation 4.3.
3
𝑓𝑐′
𝛾= | | + 1.55 (4.3)
32.4
115
The strain 𝜀𝑐′ is taken as 0.002. In this FE model, it was assumed that the uniaxial
stress-strain behaviour of concrete in compression is linear elastic up to 40% of
concrete compressive strength.
The tensile strength is specified as 0.1𝑓𝑐′ . The uniaxial tensile response was
assumed to be linear until the tensile strength of concrete was reached. Beyond that
the tensile softening response was defined by means of fracture energy. The fracture
energy is determined by Equation 4.4 as proposed by Bažant and Becq-Giraudon
(2002).
0.7
2 𝑓𝑐′
𝐺𝑓 = (0.0469𝑑𝑎 − 0.5𝑑𝑎 + 26) ( ) (4.4)
10
Where da is the maximum aggregate size and 𝑓𝑐′ is in MPa and Gf is in N/m.
4.2.6.2 Profiled steel sheeting properties
The actual stress–strain curves of the profiled steel sheeting were obtained from
uniaxial tension tests on standard coupon specimens, as shown in Figure A.1. An
idealised piecewise linear representation of this experimental stress–strain diagram
was adopted in the finite element analysis. The stress-strain relationship of profiled
steel sheeting was defined as an elastic-plastic material with strain hardening, as
shown in Figure 4.5. The mechanical behaviour for both tension and compression is
assumed to be similar. The properties used in the FE model for the profiled steel
sheets are tabulated in Table 4.2.
116
4.2.6.3 Local bond-slip law
The critical issue for numerical analysis of a post-tensioned push specimen is to
find an appropriate local bond stress-slip law considering the effect of prestress. In
this study, the bond stress-slip curve obtained from PT-500-0 has been used as a
benchmark local-bond slip law for the particular profile used in this study. For push
specimens without post-tensioning and with post-tensioning, local bond-slip law can
be found by linear interpolation considering the effect of both bond length and
prestress. FE analyses were carried out for all push specimens model using the
corresponding shear bond properties. The interpolated shear bond property curves for
the specimens are shown in Figure 4.6. The FE analyses indicate that the accurate
result for a particular specimen can only be obtained when the shear bond properties
are varied according to the level of prestress.
4.3 FINITE ELEMENT ANALYSIS
4.3.1 Mesh sensitivity analysis
Mesh sensitivity analysis is an indispensable part of FE modelling to check the
accuracy of the model. This study put considerable effort into the generation of good
quality mesh with acceptable aspect ratio. The challenge of preventing excessive
distortion came from the geometry of the profiled sheeting used in this study. Mesh
sensitivity analysis was conducted in this study by reducing element size without
changing the element order. The goal is to determine optimal mesh size that provides
relatively accurate solution with low computational time. Mesh configurations for
both concrete and profiled sheeting was considered for this analysis because the
117
behaviour of the concrete-sheeting interface may depend on the element sizes of both
concrete and profiled sheeting.
Three element sizes of 10 mm, 25 mm and 50 mm were chosen in the preliminary
development of the model and are designated as fine, medium and coarse mesh, as
shown in Figure 4.7. The connector elements for all the element sizes were placed at
the same locations and the same material properties were used in every model. All
three simulations produced almost identical results, as shown in Figure 4.8. It was
decided to use a medium mesh of 25 mm for the rest of the analysis considering the
efficiency and accuracy of numerical simulations.
4.3.2 FE model validation
The accuracy of the finite element model was verified by comparing the
numerical results with the experimental results. A comparison between the chemical
bond stress and the mechanical bond stress obtained from the experiments and finite
element models is shown in Table 4.3. The mean values of τFE / τEXPT ratios for
chemical bond stress and the mechanical bond stress are 0.88 and 1.02 respectively,
with the corresponding coefficients of variation (COV) of 0.10 and 0.12,
respectively. The bond stress versus slip curves from the analyses with
corresponding experimental curves for push specimens with various levels of
prestress and bond length are presented in Figures 4.9-4.24. The significant scatter of
test results and the simplifications in FE models posed a difficulty in making the
correlation between the results. However, the comparisons presented in this section
indicate that in general the finite element approach combined with empirical local-
118
bond slip law successfully simulated the bond stress-slip behaviour of post-tensioned
push specimens.
The push specimens of height 1500 mm and 2000 mm experienced buckling in
the tests which may be due to some imperfections in the edge of the profiled sheets
and the loading may be affected by eccentricity. This problem was not present in the
FE model specimens and therefore they all exhibited failure by achieving the desired
amount of slip.
The maximum bond stress in the FE model results and experimental results are in
good agreement, but the corresponding slip in the FE model varies from
experimental results. It is acknowledged that accurate modelling is very challenging
using the average bond-slip law. Since there is no general bond-slip law established
from the existing study, the average bond-slip law was used to determine acceptable
results.
For consistency in comparisons, total bonded area has been considered to
calculate the bond stress. In fact, total bonded area should be used only to determine
the bond stress till the chemical bond exists. After the breakdown of the chemical
bond, it is more reasonable to use the rib surface area only to calculate the
mechanical bond stress. But this procedure will introduce a discontinuity in the bond
stress-slip curves when the chemical bond has broken due to the change in
contributing area. To overcome this, calculation of the bond stress-slip curves was
based on total bonded area (Burnet, 1998). Similarly, the bond stress-slip curves
from the FE model are also based on total bonded area. Since actual contributing area
to resist shear was different in FE model and test specimens due to predelamination
119
and other problems that occurred in real experimental setup, a difference between
results obtained from the FE model and experiments exists in case of some push
specimens.
In general, the FE model follows the same trend as the experimental curves of
bond stress-slip behaviour. However, the chemical bond is slightly over-estimated in
the FE models. This is due to the fact that predelamination happened in some test
specimens and reduced the actual chemical bond. Nevertheless, considering the
complexity of the experimental setup it can be stated that the FE model can simulate
all the components of the bond with reasonably good accuracy.
4.3.3 Parametric study
A parametric study was conducted to demonstrate that this FE approach can be
applied to any type of profiles with corresponding local bond stress-slip law. Three
profiles of re-entrant, trapezoidal and Bondek have been chosen arbitrarily to
simulate the bond stress-slip behaviour. The numerical results have been compared
with the push test study conducted by Burnet and Oehlers (2001). The bond stress-
slip behaviour and von Mises stress contours for these push specimens are shown in
Figures 4.25-4.27. It can be seen that overall bond stress-slip curves obtained from
the FE models are in good agreement with the experimental curves for the push
specimens without post-tensioning.
The chemical bond stress in case of re-entrant profile was found significantly
lower than that obtained from experimental results. But the mechanical bond stress
was very close to the experimental results. The FE models were idealised in terms of
geometry of the profiled sheeting and boundary conditions of the support of the
120
profiled sheeting. Burnet (1998) reported that the chemical bond stress appears to
vary with the rib geometry. Therefore, any idealization in the profiled sheeting will
affect the bond stress. Also, it is difficult to calculate the exact contributing area of
the concrete-sheeting interface which resisted the shear at various stages of loading.
The accurate measurement of the interface bond is also affected by the application of
external lateral restraining force and boundary conditions of the test setup. Symmetry
boundary conditions of this FE model which were created for numerical efficiency
may also affect the bond stress.
The mode of failure for the push specimens was slip at the concrete-steel interface
as observed in the experiments (Burnet, 1998). Higher stress developed at the base of
the profile steel sheeting as shown in Figure 4.27 supports the reason for initiation of
possible buckling. However, further experimental studies are required to develop
local bond stress-slip law incorporating the effect of prestress for these profiles.
Since the interaction between the concrete and the profiled steel sheeting depends
on various parameters, correct modelling of this interaction is very challenging. This
chapter presents a three dimensional nonlinear finite element model of a post-
tensioned push specimen to investigate the longitudinal shear behaviour of post-
tensioned composite slabs accurately. The accuracy of the numerical model was
verified by comparing the model results with the experimental ones. The following
conclusions can be drawn from this study:
121
 The FE model developed using the ABAQUS/Explicit program could be used
to simulate the overall bond-stress slip behaviour between the concrete and the
profiled steel sheets comprising of both chemical and mechanical bond stress.
 The effect of the various levels of prestress and bond length could be
incorporated in the FE model to obtain accurate results for the load carrying
capacity of post-tensioned push specimens. This indicates that this FE
approach could be extended to study the behaviour of post-tensioned composite
slabs accurately and efficiently.
 The results of parametric study indicates that the shear bond behaviour of post-
tensioned composite slabs with different profiles can be simulated using this
approach with a reasonable accuracy.
122
Table 4.1 Concrete properties used in FE model

Properties Values
Density 2400 kg/m3
Poisson’s ratio 0.2
Dilation angle 30°
Flow potential eccentricity 0.1
Ratio of the compressive strength under biaxial loading 1.16
to uniaxial compressive strength
Ratio of the second stress invariant on the tensile 0.6667
meridian to that on the compressive meridian
Viscosity parameter 0
Fracture energy 0.066 N/mm
Table 4.2 Steel properties used in FE model

Properties Values
Density 7800 kg/m3
123
Table 4.3 Comparison of numerical results with experimental results

Specimen Chemical Chemical τFE/ Mechanical Mechanical τFE/
bond bond bond stress, bond stress,
ID τEXPT τEXPT
stress, stress, τFE τEXPT
τFE τEXPT
PT-500-0 0.17 0.20 0.83 0.13 0.11 1.17
PT-500-1 0.15 0.20 0.78 0.12 0.10 1.14
PT-500-2.4 0.14 0.16 0.88 0.11 0.11 0.96
PT-500-3 0.13 0.15 0.83 0.10 0.10 0.96
PT-1000-0 0.15 0.16 0.97 0.12 0.12 0.98
PT-1000-1 0.14 0.17 0.82 0.11 0.12 0.96
PT-1000-2.4 0.12 0.16 0.73 0.10 0.13 0.78
PT-1000-3 0.12 0.15 0.80 0.10 0.12 0.82
PT-1500-0 0.14 0.15 0.98 0.11 0.10 1.15
PT-1500-1 0.13 0.13 1.00 0.11 0.10 1.01
PT-1500-2.4 0.12 0.14 0.89 0.10 0.10 0.95
PT-1500-3 0.11 0.13 0.92 0.09 0.09 1.03
PT-2000-0 0.14 0.14 1.03 0.11 0.11 0.98
PT-2000-1 0.11 0.13 0.89 0.10 0.08 1.23
PT-2000-2.4 0.10 0.11 0.92 0.09 0.08 1.13
PT-2000-3 0.10 0.12 0.82 0.09 0.08 1.04
Mean 0.88 1.02
COV 0.10 0.12
124
Figure 4.1 Overview of the finite element mesh
Figure 4.2 Radial thrust connector at steel-concrete interface
125
Figure 4.3 Energy level for analysis
Figure 4.4 Stress-strain curve of concrete
126
Figure 4.5 Stress-strain curve of profiled steel sheeting
Figure 4.6 Local bond-slip law
127
(a) Fine elements (b) Medium elements (c) Coarse elements
Figure 4.7 FE model for mesh sensitivity analysis
Figure 4.8 Mesh sensitivity analysis
128
Figure 4.9 Push specimen of height 500 mm without prestress
Figure 4.10 Push specimen of height 500 mm with prestress of 1 MPa
129
Figure 4.11 Push specimen of height 500 mm with prestress of 2.4 MPa
130
131
132
133
134
135
136
(a)
(b)
Figure 4.25 Re-entrant profile (a) Bond stress-slip curve (b) Stress distribution
137
(a)
(b)
Figure 4.26 Trapezoidal profile (a) Bond stress-slip curve (b) Stress distribution
138
(a)
(b)
Figure 4.27 Bondek profile (a) Bond stress-slip curve (b) Stress distribution
139
140
CHAPTER 5 - Finite Element Modelling of Post-tensioned Composite Slabs
CHAPTER 5
FINITE ELEMENT MODELLING OF POST-TENSIONED
COMPOSITE SLABS
5.1 INTRODUCTION
This chapter presents a reasonably accurate three dimensional finite element
model to investigate the behaviour of post-tensioned composite (PTC) slabs. The
primary objective is to assess the bond-stress slip characteristics obtained from the
push test programme of this study. This is achieved by incorporating the effect of
prestress on local bond-slip laws to produce an accurate FE model of PTC slabs and
simulate the load-deflection behaviour and failure modes up to ultimate loads. The
ultimate strength tests of a post-tensioned solid slab and a post-tensioned composite
slab conducted by Ranzi et al. (2013a) are used to validate the finite element model.
The material nonlinearities of concrete, prestressing strands, profiled steel sheeting
and reinforcing bars were included in the finite element model. A parametric study
was conducted to investigate the effects of concrete strength, level of prestress and
profiled sheeting thickness on the strength and behaviour of post-tensioned
composite slabs. In the following sections, a summary of the experimental
investigation conducted by Ranzi et al. (2013a) is presented. All necessary steps and
components of the FE model of PTC slabs are described and the use of bond stress-
141
slip law by linear interpolation is demonstrated in subsequent sections. Finally, the
results of FE analyses and parametric studies are presented and discussed.
5.2 DESCRIPTION OF EXPERIMENTAL INVESTIGATION
The overall dimensions of the slabs used in the ultimate strength tests conducted
by Ranzi et al. (2013a) were identical with a full length of 7200 mm, an internal
simply supported span of 6000 mm, a width of 900 mm and a thickness of 180 mm.
Each slab had two overhang ends 600 mm long to minimise the effects of the
prestressing anchorage along the internal simply supported span as shown in Figure
5.1. Transverse N10 reinforcements were placed at 300 mm spacing at the top and
bottom of the solid post-tensioned slab and only at the top for the post-tensioned
composite slab. The profiled sheeting was terminated 100 mm before the roller
supports so that there is no frictional resistance between the deck and the concrete at
the supports.
Each slab had three longitudinal parabolic tendons with a nominal diameter of
12.7 mm and area of 98.6 mm2. The profile of the prestressing duct was kept straight
over the two overhangs and was varied over the internal simply supported span, with
a distance of 30 mm specified between the bottom side of the slab and the underside
of the duct at mid-span. One dead end and one live end were specified for the
prestressing operations. Post-tensioning was carried out by elongating the strand, to
85% of their capacity, from the live anchor using a hydraulic jack. The first 25% of
the jacking force was applied on the day after casting with a concrete strength of 9.3
MPa. The remaining 75% of the jacking force was applied at day 7 with a concrete
142
strength measured at day 6 of 23.0 MPa. The live end was locked using wedges with
the load transferred to the concrete. The non-jacked end of the strand was pre-locked
using wedges at the dead anchor. The slabs were tested in a static simply supported
configuration subjected to two transverse line loads applied near the mid-span as
shown in Figure 5.1. The tests results are presented and compared with the FE model
results in section 5.4.2.
5.3.1 General
In this study, the finite element program ABAQUS was used to simulate the
behaviour of a post-tensioned solid slab and a post-tensioned composite slab.
ABAQUS/Explicit solver was chosen again in this numerical study because of its
efficiency in solving discontinuous and contact problems. Both geometric and
material nonlinearity were considered in the FE analysis.
Due to the symmetry of geometry and loading, only a quarter of the slab was
modelled. Eight-node linear hexahedral solid elements with reduced integration
(C3D8R) were used to model the concrete. These elements are used extensively for
nonlinear analysis including contact, large deformation, plasticity and failure.
Therefore, C3D8R was also used to model post-tensioning strands. The profiled steel
sheets were modelled with 4-node reduced integration shell elements (S4R). S4R
was chosen as they are more accurate and efficient in contact modelling than
143
conventional shell element (ABAQUS, 2012). The reinforcing bars were modelled as
two-node three-dimensional linear truss elements (T3D2). An overview of the finite
element mesh is shown in Figure 5.2.
5.3.3 Interactions
Appropriate constraints to define the interaction between components are essential
in order to obtain accurate results from the finite element analysis. The steel sheet
elements were located at a distance of half of the steel sheet thickness from the
bottom of the concrete slab. Radial-thrust type connector elements (CONN3D2) were
chosen to simulate the interaction between the concrete and the profiled steel sheets.
The reason for using radial thrust type elements has been described in Section 4.2.3.
Each connector element connected a pair of concrete and steel nodes that were
closest to each other (Figure 5.3).
The reinforcement in the slab was modelled using the embedded technique. This
technique is used to specify an element or a group of elements that lie embedded in a
group of host elements whose response will be used to constrain the translational
degrees of freedom of the embedded nodes. In this study, the truss elements
representing the reinforcement are the embedded region while the concrete slab is the
host region. It is assumed that the perfect bond exists between the reinforcing bars
and the surrounding concrete slab. End anchorage of the post-tensioning strands was
modelled using beam multi-point constraints (MPC) which provide a rigid beam
between nodes that can be utilised to simulate the compatibility of deformations
between the tendon and the anchorage (Kang and Huang, 2012).
144
Prestressing was defined by a predefined stress field. In order to create an initial
stress field for the concrete, stress components were defined by selecting direct
specification according to global coordinate system in ABAQUS. It was assumed
that prestress was applied uniformly on concrete slab. In this analysis, a uniformly
distributed load was applied on a strip of 50 mm width on the top surface of concrete
using the displacement control. The applied displacement was increased linearly by
amplitude function. The slip was measured as the relative displacement between the
nodes on the profiled steel sheeting and on the concrete slab. The load was measured
as the total reaction acting on the loading surface. A roller support was applied to the
nodes of a strip of 50 mm width on the bottom side of the concrete slab. XSYMM
(surface 1) and YSYMM (surface 2) symmetry boundary conditions were applied to
the appropriate surfaces to simulate the continuity of the slab specimen in both axes,
as shown in Figure 5.2.
In this numerical study, a quasi-static solution is desired and therefore load was
applied slow enough to ensure the inertia effect was insignificant. For each result,
total internal (ALLIE) and kinetic (ALLKE) energies of the whole model computed
as history variables were compared to ensure a quasi-static response. As concrete
failure occurs by sudden drop of load carrying capacity, kinetic energy of the system
is increased enormously. It was found that the kinetic energy level was significantly
below the internal energy level throughout the analysis, as shown in Figure 5.4.
145
used to model the concrete material. The details of this model are described in
Section 4.2.6.1. Plain concrete in uniaxial compression was modelled as an elastic-
plastic material with strain softening, as shown in Figure 5.1. The elastic part of the
stress-strain curve was defined with a value of 27,300 MPa for the Young’s modulus
and 0.2 for the Poisson’s ratio. The concrete damage plasticity model requires three
types of material behaviour data: 1) Plasticity, 2) Compressive behaviour and 3)
Tensile behaviour. Plasticity was defined by five parameters: 1) Dilation angle, 2)
Flow potential eccentricity, 3) Ratio of the compressive strength under biaxial
loading to uniaxial compressive strength, 4) Ratio of the second stress invariant on
the tensile meridian to that on the compressive meridian and 5) Viscosity parameter.
The stress-strain curve for concrete compressive behaviour as proposed by Carreira
and Chu (1985) was used.
The uniaxial tensile response was assumed to be linear until the tensile strength of
concrete was reached. Beyond that, the tensile softening response was defined by
means of fracture energy. The concrete properties used in the FE model are reported
in Tables 5.1 and 5.2.
5.3.6.2 Profiled steel sheeting properties
The stress-strain relationship obtained from material tests was used for the
modelling of the profiled steel sheeting. An elastic-plastic material with isotropic
strain hardening law was used as the constitutive law, as shown in Figure 5.6. The
146
mechanical behaviour for both tension and compression is assumed to be similar.
The classical metal plasticity model available in ABAQUS was used to model the
material nonlinearity of the profiled steel sheeting. The elastic part of the stress-strain
curve was defined with the * ELASTIC option with a value of 200,000 MPa for the
Young’s modulus and 0.3 for the Poisson’s ratio. The plastic part of the stress-strain
curve was defined with the * PLASTIC option with a value of 600 MPa for yield
strength and 626 MPa for ultimate strength. The engineering stresses and strains
obtained from the standard tensile coupon tests were converted to true stresses and
strains. Plastic strain values, not total strain values, are used in defining the strain
hardening behaviour. Furthermore, the plastic strain value must be zero in the first
data pair to correspond with the onset of plasticity (ABAQUS, 2012). The properties
used in the FE model for the profiled steel sheets are tabulated in Tables 5.2 and 5.3.
5.3.6.3 Reinforcing steel and prestressing steel properties
The stress-strain relationship obtained from material tests was used for the
modelling of the reinforcing steel and the prestressing steel. The classical metal
plasticity model available in ABAQUS was used to model the material nonlinearity
of the reinforcing steel and the prestressing steel. The reinforcing bars were modelled
using an elastic-plastic model with isotropic strain hardening, as shown in Figure 5.7.
The elastic part of the stress-strain curve was defined with the * ELASTIC option
with a value of 200,000 MPa for the Young’s modulus and 0.3 for the Poisson’s
ratio. The plastic part of the stress-strain curve was defined with the * PLASTIC
option with a value of 500 MPa for the yield strength and 640 MPa for the ultimate
strength.
147
An elastic-perfectly plastic model was used as the constitutive law for the
prestressing bars, as shown in Figure 5.8. The mechanical behaviour for both tension
and compression is assumed to be similar. The elasticity of the prestressing bars was
defined with the * ELASTIC option with a value of 197,000 MPa for the Young’s
modulus and 0.3 for the Poisson’s ratio. The plastic part of the stress-strain curve
was defined with the * PLASTIC option with a value of 1740 MPa for the yield
strength and 1764 MPa for the ultimate strength. The properties used in the FE model
for reinforcing steel and prestressing steel are tabulated in Tables 5.2 and 5.3.
The critical issue for numerical analysis of a post-tensioned composite slab is to
model the complex interactions between the concrete and the profiled steel sheets.
One of the approaches available in ABAQUS is to model this interaction is using
interface elements. Although this technique is quite cumbersome both in terms of
geometry construction and computational time, accuracy of the model could not be
compromised. The appropriate local bond stress-slip law considering the effect of
prestress obtained from push test study, as described in Chapter 3 has been used. For
slab specimens without post-tensioning and with post-tensioning, local bond-slip law
can be found by linear interpolation considering the effect of both bond length and
prestress. The interpolated shear bond property curves for the specimens are shown
in Figure 5.9. The FE analyses indicate that the accurate result for a particular slab
can only be obtained when the shear bond properties are accurate.
148
A mesh sensitivity analysis was carried out to determine optimal mesh size that
provides relatively accurate solution with low computational time. Four element
sizes of 12.5 mm, 25 mm 50 mm and 100 mm were chosen in the preliminary
development of the model, as shown in Figure 5.10. All of these element sizes were
chosen for the concrete, profiled sheeting, reinforcing bar and prestressing bar. This
is due to the fact that the overall load-deflection behaviour of post-tensioned
composite slabs may depend on the behaviour of each component. The connector
elements for all the element sizes were placed at the same locations and the same
material properties were used in every model. Figure 5.10 shows that mesh size of
100 mm produces higher stiffness and ultimate load carrying capacity compared to
the experimental values. The stiffness and ultimate load carrying capacity of the FE
model using mesh size of 50 mm are found to be in good agreement with the
experimental curve. But there was no convergence in the load-deflection curve when
the result of the FE model with mesh size of 25 mm were plotted. It was found that
mesh convergence occurred for finer mesh sizes of 12.5 mm and 25 mm. It was
decided to use a mesh of 25 mm for the rest of the analysis considering the efficiency
and accuracy of numerical simulations.
numerical results with the experimental results available in literature (Ranzi et al.,
149
2013a). The comparisons presented in this section indicate that in general the finite
element approach incorporating empirical local-bond slip law successfully simulated
the bond stress-slip behaviour of post-tensioned composite slabs. The plastic strain
and stress distribution of the critical component of the finite element model were
observed and the modes of failure were determined.
In case of post-tensioned solid slab (SS), the stiffness and ultimate moment
capacity are found to be in good agreement with the experimental curve, as shown in
Figure 5.11. However, the initial numerical response for cracking strength of the slab
is stiffer than the experimental one. This overestimation of stiffness can be partly
attributed to the tension stiffening in the concrete damage plasticity model which did
not account the initial damage of the material. Another reason is that prestressing
was defined by predefined field which distributes prestress uniformly in the concrete
of the FE model whereas the distribution of prestress in test specimen may be non-
uniform. This slab specimen achieved a maximum load of 42.1 kN at a midspan
deflection of 170.3 mm. However, FE model reached an ultimate load of 37 kN,
which is 12% less than the experimental value of 42 kN. The numerical response is
ductile up to failure with an indication of concrete crushing at the midspan region, as
shown in Figure 5.12. The yielding of post-tensioning strand has been observed from
the FE model, as shown in Figure 5.13.
The results of post-tensioned composite slab with Condeck profile (CK) are
compared with the FE results in terms of load-midspan deflection curve in Figures
5.14. The peak load of this specimen was 84.2 kN occurred at a deflection of 165.3
mm. But the numerical response is slightly stiffer than the experimental one, while
the ultimate load carrying capacity is in good agreement. The FE model reached an
150
ultimate load of 77 kN, which is 8% less than the experimental value of 84 kN. The
greater numerical stiffness can be partly attributed to the bond stress-slip model
which has been obtained from the push tests. It is acknowledged that the constraints
in push test differ from that in composite slabs. Moreover, the tension stiffening in
the concrete model increases the stiffness partly. Also, the prestressing strand level
of this specimen was disturbed due to the possible breakage of the wires installed to
hold down the ducts during the concrete pour (Ranzi et al., 2013a). This variation
was reported based on the actual height of the strands measured after cutting the
samples at the end of the ultimate tests. The FE model of post-tensioned composite
slab developed in this study was based on the actual height of the strands initially
specified. So this variation in the height of strands could be one of the reasons for
deviation between experimental and FE model results. However, considering the
complexity of modelling of accurate bond-slip behaviour, it can be concluded that
the FE model developed in this study can simulate the nonlinear behaviour with
fairly good accuracy. The failure mode in FE model was shear bond failure followed
by an indication of concrete crushing at the midspan region, as shown in Figures 5.15
and 5.16. The yielding of post-tensioning strands has been observed from FE model,
as shown in Figure 5.17.
It is shown that the finite element model presented in the previous section was
accurate and reliable to simulate the nonlinear behaviour of post-tensioned composite
slabs up to ultimate loads. Therefore, a parametric study was conducted to
investigate the effects of concrete strength, profiled steel sheeting thickness and level
of prestress.
151
5.4.3.1 Effects of concrete compressive strength
Figure 5.18 shows the load versus deflection relationship for the post-tensioned
composite slab considering the variation in concrete compressive strength (𝑓𝑐′ ). It can
be seen that the load carrying capacity of the slab is increased slightly with the
increase of concrete compressive strength. However, no significant increase in load
carrying capacity was found for the concrete strength between 40 MPa and 50 MPa.
This could be attributed to the fact that concrete compressive strength has little
influence on the shear bond resistance of composite slab (Daniels and Crisinel,
1993b). In this study, the post-tensioned composite slab with concrete strength of 25
MPa, 32 MPa, 40 MPa and 50 MPa achieved a load carrying capacity of 73 kN, 76
kN, 80 kN and 81 kN respectively. Therefore, 25 MPa increase in concrete
compressive strength contributed about 11% increase in ultimate load carrying
capacity of post-tensioned composite slabs.
5.4.3.2 Effects of profiled steel sheeting thickness
composite slab of different thickness (tp) of profiled steel sheeting ranging from 0.75
mm to 2.00 mm. It can be seen that the stiffness of the slab is increased with the
increase of profiled sheeting thickness. The modes of failure in all cases are shear
bond failure. These results are also supported by the findings of other researchers
(Seleim and Schuster, 1985).
In this FE model, the connector section properties were based on the bond-slip
characteristics of 1.00 mm thick profiled sheeting. The bond stress-slip behaviour for
any other thickness of this profiled sheeting is not available in open literature. The
152
use of single shear bond property may produce good results for one particular slab
geometry only but it may not be accurate when the slab slenderness and geometry are
changed (Abdullah, 2004). Therefore, this parametric study was conducted on a 6 m
span slab only.
5.4.3.3 Effects of level of prestress
composite slab for four levels of prestress (P) ranging from 0 MPa to 3 MPa.
Obviously, the post-tensioned composite slab showed greater load carrying capacity
compared to composite slab without prestress. Composite slab with prestress of 1
MPa achieved a load carrying capacity of 78 kN whereas composite slab without
prestress achieved only 42 kN. There was no significant difference found in terms of
slabs’ ductility. It can be seen that the stiffness of the post-tensioned composite slabs
is increased with the increase of level of prestress. However, no significant increase
in the load carrying capacity of the slabs with prestress from 1 MPa to 3 MPa was
observed. This can be attributed to the finding of the push test study presented in
Chapter 3 that prestress has a detrimental effect on the bond stress-slip behaviour of
post-tensioned composite slabs. With the increase of level of prestress, the
longitudinal shear strength of the slab decreased and thus prevented the increase in
ultimate strength of the slabs. However, further experimental data is required to
validate these results.
Since the chemical bond is brittle in nature, as soon as failure starts at one location
of the interface, the breakdown of the bond propagates rapidly across the entire
interface and then mechanical bond transfers the shear in the concrete-sheeting
153
interface. Prestress shortens the concrete slab by compressive force and this creates a
tensile force in the profile sheeting for equilibrium. Consequently, a redistribution of
shear stress happens in the concrete-steel interface in prestressed slabs. Depending
on the level of prestress, the amount of tensile force in profiled sheeting varies and it
affects both chemical and mechanical bond stress-slip behaviour of the concrete-
sheeting interface. Therefore, overall bond stress-slip characteristics vary in post-
tensioned composite slab based on the level of prestress.
5.5 ULTIMATE CAPACITY OF POST-TENSIONED COMPOSITE
SLABS
The ultimate moment capacity of the post-tensioned composite slabs was
determined following the procedures available for the design of post-tensioned
concrete members (Gilbert & Mickleborough, 2004) and composite steel-concrete
slabs available in the literature (British Standards Institution, 2004).
The analytical approach is based on the following assumptions:
 Concrete carries no tensile stresses.
 The variation of strain on the cross-section is linear, i.e. plane section remains
plane.
 Concrete stresses are idealised based on the rectangular stress block for the
compressive concrete above the neutral axis.
154
The adopted strain diagram and stress distributions at ultimate are illustrated in
Figure 5.21.
At the ultimate moment, the extreme fibre of the concrete slab is assumed to
possess a compressive strain εcu of 0.003. The depth of the rectangular stress block is
γdn and the uniform stress intensity is α2𝑓𝑐′ .
The concrete compressive force is therefore given by
𝐶 = 𝛼2 𝑓𝑐′ 𝑏𝛾𝑑𝑛 (5.1)
Where b and dn represents the width of slab and location of neutral axis respectively.
The value of α2 and γ is given by
𝛼2 = 1.0 − 0.003𝑓𝑐′ 0.67 ≤ 𝛼2 ≤ 0.85
𝛾 = 1.0 − 0.003𝑓𝑐′ 0.67 ≤ 𝛾 ≤ 0.85
The resultant tensile force in post-tensioned composite cross-section is given by
𝑇 = 𝐴𝑝 𝜎𝑝𝑢 + 𝐴𝑠 𝜎𝑠𝑢 (5.2)
Where σpu and σsu are the stress in the bonded tendons and profiled steel sheet at
ultimate moment and are determined from considerations of strain compatibility,
axial equilibrium and stress-strain curves.
The assumption is tentatively made, and subsequently checked, that all steels are
at yield and the neutral axis depth dn is then determined enforcing axial equilibrium
at the cross-section:
𝐶=𝑇 (5.3)
155
With the maximum eccentricity of e known, the value of effective prestress Pe
required to cause an upward force w, which exactly balances a selected portion of the
external load is given by
𝑤𝐿2
𝑃𝑒 = (5.4)
8𝑒
Under this balanced load, the beam suffers no curvature and is subjected to only to
the longitudinal compressive force P as shown in Figure 5.22.
The force resisted by the prestressing strands at ultimate is determined based on
its strain pu which is the sum of three deformations related to the loading stages
illustrated in Figure 5.23:
𝜀𝑝𝑢 = 𝜀𝑝𝑒 + 𝜀𝑐𝑒 + 𝜀𝑝𝑡 (5.5)
Where pe is the strain produced by the initial prestressing force Pi, ce represents the
strain to be applied to the strands to reach the condition of decompression at stage 2,
and pt specifies the additional strain to reach the ultimate condition.
In particular, the initial deformation pe is calculated based on (Gilbert &
Mickleborough, 2004)
𝑃𝑒
𝜀𝑝𝑒 = 𝐴 (5.6)
𝑝 𝐸𝑝
Where Ap and Ep represent the area and elastic modulus of the prestressing tendons.
The consequent change in strain ce (between stages 1 and 2) is calculated as
follows:
156
1 𝑃 𝑃𝑒 𝑒 2
𝜀𝑐𝑒 = 𝐸 ( 𝐴𝑒 + ) (5.7)
𝑐 𝐼
In this calculation the value for pt (between stages 2 and 3) is obtained from
geometric considerations as:
𝑑𝑝 −𝑑𝑛
𝜀𝑝𝑡 = 𝜀𝑐𝑢 𝑑𝑛
(5.8)
When the correct value of dn and corresponding tensile force are found, the
ultimate moment may be expressed as
𝛾𝑑𝑛 𝛾𝑑𝑛
𝑀𝑢 = (𝐴𝑝 𝜎𝑝𝑢 (𝑑𝑝 − ) + 𝐴𝑝 𝜎𝑠𝑢 (𝑑𝑠 − )) (5.9)
2 2
This moment corresponds to full shear connection which is associated with one of
the two elements (concrete or steel) being fully stressed and hence achieving its full
axial yield strength.
When there is zero shear connection, the strength of the composite section is
simply the summation of flexural strength of the solid slab and the profiled sheet, i.e.
𝑀𝑜𝑠𝑐 = 𝑀𝑆𝑜𝑙𝑖𝑑 𝑠𝑙𝑎𝑏 + 𝑀𝑑𝑒𝑐𝑘 (5.10)
Between these two extreme values, at Mosc at ƞ=o and Mfsc at ƞ=1, the flexural
capacity of the composite section is given by Mpsc as shown in Figure 5.24.
From the Mu vs. ƞ linear relationship as shown in Figure 5.24, the degree of shear
connection ƞ may be determined.
157
Therefore, the ultimate moment for partial shear connection when the strength of
the shear connectors controls the strength of the composite section may be expressed
as
𝛾𝑑𝑛 𝛾𝑑𝑛
𝑀𝑢 = 𝜂 (𝐴𝑝 𝜎𝑝𝑢 (𝑑𝑝 − ) + 𝐴𝑝 𝜎𝑠𝑢 (𝑑𝑠 − )) (5.11)
2 2
The ultimate moment capacities calculated based on FSC and PSC are
summarised in Table 5.4 and compared with values obtained from experiments and
finite element model. It has been noticed that the rigid plastic calculations produced
satisfactory predictions for the ultimate flexural capacity of the post-tensioned solid
slab with ratio between experimental and calculated values of 0.88. But the ultimate
capacity of the post-tensioned composite specimens based on full shear connection
considerations overestimated the experimental flexural capacity in the order of 25%.
It indicates that the post-tensioned composite specimen did not achieve full shear
connection. The ultimate moment calculated based on partial shear connection was
87% of the experimental moment capacities. This difference may be attributed to the
difference in degree of interaction in experiment and analysis. Finite element analysis
produced satisfactory predictions for the ultimate moment capacity of both post-
tensioned solid slab and post-tensioned composite slab with ratio between FE output
and experimental values of 0.91 and 0.93 respectively.
A three dimensional nonlinear finite element model has been developed to
investigate the behaviour of post-tensioned composite slabs accurately. The accuracy
158
of the numerical model was verified by comparing the model results with the
experimental results available in literature. The following conclusions can be drawn
from this study:
 Accurate bond-slip law considering the effect of the prestress is crucial for
numerical simulation of post-tensioned composite slabs. Ultimate strength of
post-tensioned composite slabs was determined with higher accuracy using this
FE approach.
 The results of parametric studies indicate that increase in stiffness and strength
of post-tensioned composite slabs occurs with stronger concrete. Ultimate
strength of post-tensioned composite slab with the particular profiled steel
sheeting used in this study can be improved by increasing the thickness of
profiled sheeting. The effect of various level of prestress to the ultimate load
carrying capacity of post-tensioned composite slabs indicated that the prestress
has a detrimental effect on the shear bond-slip behaviour of these types of
slabs.
 Rigid plastic calculations based on partial shear connection and finite element
models produced satisfactory predictions for the ultimate flexural capacity of
the post-tensioned composite slabs. However, further experimental data are
required to properly validate the proposed FE model.
159
Table 5.1 Concrete properties used in FE model

Properties Values
Density 2400 kg/m3
Dilation angle 30°
Flow potential eccentricity 0.1
Fracture energy 0.066 N/mm
Table 5.2 Material properties of concrete, profiled steel sheets, prestressing steel and
reinforcing steel
Material Property Values
(MPa)

steel sheets
steel
Reinforcing Yield stress, fy 500
steel
160
Table 5.3 Steel properties used in FE model

Properties Values
Density 7800 kg/m3
Table 5.4 Comparison of ultimate moment capacities

FSC PSC FE
Specimen Mu,exp Mu,calc Mu,calc Mu,calc Mu,calc Mu,FE Mu,FE
(kNm) (kNm) /Mu,exp (kNm) /Mu,exp (kNm) /Mu,exp
SS 73.62 65.03 0.88 - - 66.71 0.91
CK 130.45 162.65 1.25 114.17 0.88 120.71 0.93
161
Figure 5.1 Ultimate strength test setup (Ranzi et al., 2013)
Figure 5.2 Overview of the finite element mesh
162
Figure 5.3 Radial thrust connector at steel-concrete interface
Figure 5.4 Energy level for analysis
163
Figure 5.6 Stress-strain curve of profiled steel sheets
164
Figure 5.7 Stress-strain curve of reinforcing bar
Figure 5.8 Stress-strain curve of prestressing bar
165
Figure 5.9 Local bond stress-slip law
Figure 5.10 Mesh sensitivity analysis
166
Figure 5.11 Load-deflection curve of post-tensioned solid slab
Figure 5.12 Failure mode of post-tensioned solid slab
167
Figure 5.13 Yielding of post-tensioning steel strand of post-tensioned solid slab
Figure 5.14 Load-deflection curve of post-tensioned composite slab
168
(a) Experiment (Ranzi et (b) FE Model

al., 2013)
Figure 5.15 Shear bond failure of post-tensioned composite slab
Figure 5.16 Failure mode of post-tensioned composite slab
169
Figure 5.17 Yielding of post-tensioning steel strand of PTC slab
Figure 5.18 Effects of concrete compressive strength (𝑓𝑐′ )for PTC slab
170
Figure 5.19 Effects of profiled sheeting thickness ((tp) for PTC slab
Figure 5.20 Effects of level of prestress (P) for PTC slab
171
(a) Typical PTC slab (b) Strain (c) Stress

cross-sections
Figure 5.21 Strain and stress distribution at the ultimate limit state (Ranzi et al.,
2013a)
Figure 5.22 Forces exerted on a concrete beam by a parabolic tendon (Gilbert and
172
(a) Typical PTC slab (b) Strain

cross-section
Stages: (1) initial prestress, (2) decompression of the concrete at the prestressing
steel level and (3) ultimate state.
Figure 5.23 Strain distribution at three stages of loading (Gilbert and

Figure 5.24 Variation of flexural capacity with degree of shear connection
173
174
CHAPTER 6 - Experimental Programme of Composite Slabs with End Anchorage
CHAPTER 6
EXPERIMENTAL PROGRAMME OF COMPOSITE
SLABS WITH END ANCHORAGE
6.1 INTRODUCTION
This chapter presents the details of an experimental programme to investigate the
strength and behaviour of composite slabs with end anchorage. The test program
consisted of eight slab specimens and the variables investigated in this experimental
study were (a) slab type, (b) end restraint condition and (c) long term loading
condition. In the following sections, details of all specimens and test procedure are
provided. Finally, the test results were analysed and discussed to provide an insight
into the overall behaviour of composite slabs with end anchorage.
6.2 EXPERIMENTAL PROGRAMME
6.2.1 Test specimens
The scope of the experimental programme considered in the present study
includes the testing of eight composite slab specimens in the Randwick Heavy
Structural Laboratory at The University of New South Wales, Australia. A total of
175
twelve composite slab specimens were prepared in the Structural Research and
Testing Laboratory of the University of Western Sydney, Australia. These specimens
have been referred to herein as STSR and STSU for short-term solid slab with
restrained and unrestrained end support, STPR and STPU for short-term profiled slab
with restrained and unrestrained end support, LTSR-SW and LTSR-SL for long-term
solid slab with end restraint under self-weight and service load, LTPR-SW and
LTPR-SL for long-term profiled slab with end restraint under self-weight and service
load, LTSU-SW and LTSU-SL for long-term solid slab without restraining ends
under self-weight and service load and LTPU-SW and LTPU-SL for long-term
profiled slab without restraining ends under self-weight and service load
respectively. AS2327.1 (Standards Australia, 1996) was followed for the design of
these specimens. For the long term tests over a period of 710 days, four slabs were
loaded under self-weight only assuming that this will cause the concrete to be
subjected to shrinkage and there is no creep developed in the concrete slabs. The
other four concrete slabs were loaded with a sustained external load of 8.5 kN/m at
28 days after concrete pouring and the concrete slab was assumed to be subjected to
creep effects. The sustained loading was applied as a uniformly distributed load by
placing three reinforced concrete blocks on the slab specimens. The loading blocks
were placed at a spacing of 40 mm to prevent any possibility of contact in between
them due to the long term deflections of the slabs. The end supports for each
concrete block were provided as a fixed and a roller to prevent any potential strut and
tie effect between the concrete blocks and the concrete slab. The results of short term
strength tests of first four specimens and long term tests of eight specimens have
been published by Mirza and Uy (2013). The results of ultimate strength tests of
176
eight specimens have been presented in this study. A summary of the test specimens
is given in Table 6.1.
All these specimens represent a typical 3 m composite slab span between
secondary beams with a slab thickness of 150 mm as shown in Figure 6.1. The width
of all the samples was 700 mm. Two 310UB46.2 structural steel beams were
provided as end supports representing typical secondary beams with a clear span of
2834 mm. The beam-slab restrained connection was achieved through three 19 mm
diameter headed stud shear connectors at 345 mm and 265 mm spacing welded in a
single line along the centre of the top flange of the steel beam and according to the
troughs of the W-DEK profiled steel sheeting. N12 longitudinal reinforcement at 200
mm spacing and N12 transverse reinforcement at 300 mm spacing were placed at the
top and bottom of the solid slabs and only at the top for the profiled slabs
maintaining a clear cover of 30 mm in all cases. All specimens were constructed
using 1.00 mm thick W-DEK profiled steel sheets with total galvanised coating
thickness of 1.03 mm (BlueScope Steel, 2007). The geometry of the profiled steel
sheets is shown in Figure 6.2.
6.2.2 Specimen preparation
The first step of the specimen preparation was to cut the profiled steel sheets to a
length of 3000 mm. Then the formwork was assembled with appropriate dimensions
using 12 mm thick fibreboards. The headed stud shear connectors were welded
using a stud welding machine directly onto the top flange of the steel beam in the
case of solid slabs and through deck welding at the troughs in the case of profiled
slabs. The reinforcing bars were then positioned using steel chairs and tie wire and
177
the strain gauges on profiled steel sheets and the reinforcing bars were attached. The
universal steel beams were stiffened using timber blocks to prevent any rotation in
the web and flanges. The specimens were poured along with the test cylinders and
then covered with polythene sheets and kept moist for two days. After this period
they were left to cure in the air until testing took place.
All slab specimens were cast on the same day with a target concrete compressive
strength of 25 MPa. The material properties of the concrete were determined from
standard cylinder tests as reported in Table 6.2. A minimum of 3 cylinders of 100
mm diameter and 200 mm height were tested. The properties of the profiled steel
sheets, reinforcing bars and 19 mm headed stud shear connectors obtained from
standard tensile tests are presented in Table 6.2 and Appendix C (Tables C.1, C.2 &
C.3).
6.2.4 Instrumentation
Two linear strain conversion transducers (LSCT) were attached to each end of
concrete slab to measure the slip at the interface directly, as shown in Figure 6.3.
Four strain gauges (FLA-5-11-3LT) were placed into the troughs of profiled steel
sheets in profiled slabs and on the bottom reinforcement in solid slabs. Four concrete
strain gauges (PL-60-11-3L) were placed on surface of the concrete slab at midspan.
The deflections were monitored using three laser displacement transducers placed at
the midspan and at the one-third points. An automatic data acquisition system was
employed to record the readings of the load cell, LSCT and strain gauges throughout
the test.
178
6.2.5 Test setup
The specimens were lifted up by the lifting lugs and then placed in the loading rig.
Two spreader beams were placed on the top of the specimen to apply two line loads
over the entire slab width. A spreader beam on the top of two transverse beams was
placed on the slab specimens to distribute the load as two line loads onto the
specimen, as shown in Figure 6.4. A spherical seat was used to provide a uniform
dispersion of the load from the loading jack to the specimen.
The load was applied to the specimen very slowly at a rate of 0.5 mm/min until
the maximum load capacity had reached. The loads and deflections were logged
continuously over a period of 4 hours. The test was terminated when either the load
dropped 20% from the maximum values or the midspan deflections were close to one
thirtieth of the span length.
6.3 RESULTS AND DISCUSSION
6.3.1 General
The aim of the testing programme was to investigate the behaviour and strength of
the composite slabs. Maximum loads and the corresponding midspan deflections for
all the slab specimens are tabulated in Table 6.3. The load-deflection behaviour and
failure modes of all slab specimens are discussed in the following section.
179
6.3.2 Test observations
6.3.2.1 Solid unrestrained slabs
The load-deflection responses of the solid unrestrained slabs LTSU-SW and
LTSU-SL at midspan were almost linear before initiation of shear-bond slip at the
slab ends (Figure 6.5). As the slabs were previously tested under self-weight and
service loads, some cracks already developed. Cracking continued to enlarge and
propagate across the full width of the specimens as load progresses. Ductile
behaviour was achieved with large deflections in the inelastic region. Moreover,
these specimens showed considerably higher ductility compared with all other
specimens of this test series. The maximum loads recorded at failure of LTSU-SW
and LTSU-SL were 60 and 63 kN respectively. Even though both specimens were
loaded for 710 days at different service loads, they have demonstrated almost
identical performance in terms of strength and ductility. However, a small difference
in the stiffness is observed which could be attributed to the creep and shrinkage
effect of concrete slab. The cracking pattern was more uniform and closely spaced in
case of LTSU-SL compared with LTSU-SW (Figures. 6.6, 6.7 & 6.8). This is due to
the effect of uniform service loading which was applied on LTSU-SL for previous
long term tests. The failure modes for both specimens included yielding of
reinforcing bars and concrete crushing at midspan, which was observed as the
deflection increased until the test was terminated.
6.3.2.2 Profiled unrestrained slabs
The load-deflection responses of profiled unrestrained slabs LTPU-SW and
LTPU-SL at midspan are presented in Figure 6.9. These curves were almost linear
180
before the initiation of shear-bond-slip at the slab ends. The shear bond-slip
developed with further load increments, leading to nonlinear load-deflection curves
accompanying flexural cracks in the concrete and ultimate failure of the composite
slabs (Figure 6.10). LTPU-SW and LTPU-SL achieved an ultimate load of 39 kN
and 42 kN respectively. Both specimens exhibited a less ductile response in the post-
elastic region compared with the solid unrestrained slabs. Although the service
loading for the long term tests was different, the load-deflection curves of these
specimens were very close to each other throughout the entire loading period. The
failure modes for these specimens included longitudinal shear bond failure and
yielding of reinforcing bars (Figure 6.11). Web crippling of the profiled sheeting
were found near the load point. This was a localized bearing failure, initiated at an
interior web followed by the outer webs as load increased. The crippling of the webs
also caused deformation on the flanges of the slab specimens. In case of LTPU-SW,
the average maximum end slips between the concrete and the profiled sheeting at
ultimate load were 6.3 mm and 2.1 mm at the left and right ends respectively. The
difference between end slips can be attributed to the application of loading and
boundary conditions. In case of LTPU-SL, the average maximum end slips between
the concrete and the profiled sheeting at ultimate load were 5.8 mm and 2.7 mm at
the left and right ends respectively. The load-end slip relationships are provided in
Appendix C (Figures C.1 and C.2).
6.3.2.3 Solid restrained slabs
The load-deflection curves for specimen LTSR-SW and LTSR-SL are shown in
Figure 6.12. Despite differences in the loading condition for long term, both
specimens have shown almost similar performance in terms of stiffness, strength and
181
ductility. LTSR-SW and LTSR-SL reached an ultimate load of 90 kN and 87 kN
respectively. LTSR-SW specimen behaved linearly up to a load of approximately 86
kN with a deflection of 28 mm, at which point the force deflection curve began to
plateau to a peak of around 90 kN (Figure 6.12). The failure mode included yielding
of reinforcing bars and concrete crushing at midspan. Similar failure mode was also
observed in case of LTSR-SW which achieved an ultimate load of 90 kN at a
deflection of 45 mm (Figure 6.12). At the maximum load, the cracks widened up
considerably, accompanied by drops in loads (Figures 6.13 & 6.14). There was no
evidence of failure of the shear connection after the completion of the tests and
interface slip between the concrete and the steel beam was very small, not exceeding
0.5 mm.
6.3.2.4 Profiled restrained slabs
Figure 6.15 shows the load-deflection curves for specimen LTPR-SW and LTPR-
SL. The specimen LTPR-SW achieved a load of 102 kN at a deflection 46 mm. The
slip between the concrete and the sheeting recorded at the right end LSCT was
highest, measuring 12 mm, followed by the left end which had a slip of 8 mm. The
failure mode included longitudinal shear bond-slip and tensile rupture of the bottom
flange of the deck around shear connectors (Figure 6.16). This is same as the failure
mode observed by Porter and Greimann (1984) and Jolly and Lawson (1992).
However, no concrete crushing at the top was observed even in post-ultimate region.
On the other hand, the specimen LTPR-SL showed a significantly reduced load of 79
kN at a deflection of 91 mm. The reason can be attributed to the weld failure of one
shear stud, which was observed at the end of the test by breaking some portion of the
concrete slab (Figure 6.17). The overall crack pattern of LTPR-SL is shown in Figure
182
6.18. Comparing the load-deflection behaviour with short term specimen STPR, as
shown in Figure 6.19, it was found that the LTPR-SW showed similar performance
in terms of stiffness, strength and ductility. But LTPR-SL showed similar initial
stiffness as STPU which justified the reason for the reduced load carrying capacity of
LTPR-SL. Therefore, the test result of LTPR-SL is not reported for further
comparison. The comparison between the behaviour of solid slabs and profiled slabs
from this work indicated that the failure mode for solid restrained slab becomes the
ductile failure with cracking of concrete across slab width and around stud area while
the failure mode for profiled slabs was less ductile with shear bond failure. In case of
LTPR-SW, the average maximum end slips between the concrete and the profiled
sheeting at ultimate load were 2.9 mm and 1.8 mm at the left and right ends
respectively. In case of LTPR-SL, the average maximum end slips between the
concrete and the profiled sheeting at ultimate load were 8.1 mm and 7.9 mm at the
left and right ends respectively. The load-end slip relationships are provided in
Appendix C (Figures C.3 and C.4).
6.3.3 Effects of creep and shrinkage
Figures 6.20, 6.21, 6.22 & 6.23 compare the results of the test specimens of this
study with that of the short term strength tests by Mirza and Uy (2013). In general,
degradation in ultimate strength due to creep and shrinkage was observed. The
decrease in ultimate load of the solid unrestrained and restrained slabs in long term is
about 14% and 10% compared with that of the short term solid slabs. This could be
explained in accordance with the findings of the long term tests conducted by Mirza
and Uy (2013). Both free and restrained shrinkage in case of solid unrestrained and
restrained slabs respectively reduce stiffness of the slabs which resulted in greater
183
deflection and reduced load carrying capacity. In case of profiled unrestrained slabs,
it is interesting to observe 24% strength degradation with lesser stiffness. But the
effect of creep and shrinkage is found to be negligible in the case of profiled
restrained slabs. The residual tension in the profiled sheeting from creep and
shrinkage of the concrete can affect the longitudinal shear bond-slip between the
concrete and the profiled sheeting in case of profiled unrestrained slabs.
Consequently, the load carrying capacities of the slabs are affected. But the end
anchorage in profiled restrained slabs can help maintaining the shear bond-slip.
6.3.4 Effects of end anchorage
The effects of end anchorage are plotted in Figures 6.24 & 6.25. As expected, end
anchorage increased the load carrying capacity and stiffness of both types of slabs.
The increase in stiffness of the restrained slabs in relation to the unrestrained slabs is
attributed to the rotational stiffness increased due to the end support condition.
Figures 6.26 & 6.27 illustrate the effect of the end support conditions on the
deflection of both restrained and unrestrained slab specimens.
𝛿𝑟
From the load-deflection curves, it was found that <1
𝛿𝑢
Where 𝛿𝑟 and 𝛿𝑢 are the deflection of restrained and unrestrained slabs.
Solid restrained slabs reached an ultimate load which is 40% greater than that
achieved in solid unrestrained slabs. Compared with profiled unrestrained slabs, it is
not surprising that a 150% increase in ultimate strength was achieved as the end
anchorage was provided in profiled restrained slabs. However, profiled slabs with
end anchorage exhibited less ductile behaviour. This is due to the fact that premature
184
tearing of the profiled steel sheeting near the end anchorage area affected the shear
bond stress-slip behaviour and governed the failure mode of the profiled restrained
slabs.
6.3.5 Effects of profiled steel sheeting
Figures 6.28 & 6.29 illustrate the effect of the composite action between the
concrete and the profiled steel sheeting. Profiled slabs without end anchorage
showed a decrease of 35% in strength when compared to solid slab in unrestrained
condition, as shown in Figure 6.25. But profiled slab with end anchorage showed an
increase of 14% in ultimate strength compared to solid restrained slab (Figure 6.26).
Thus, the influence of end anchorage in composite action is significant in this case.
However, end anchorage is usually more flexible than relatively brittle bond stress-
slip behaviour between the concrete and the profiled sheeting. Therefore, it is
possible that the end anchorage force may not be fully mobilised when the ultimate
shear-bond capacity is reached (Jolly and Lawson, 1992).
An experimental investigation of composite steel-concrete slabs with end
anchorage has been carried out in this study. The results presented provide a valuable
insight into the behaviour of composite slabs under ultimate load. The following
conclusions can be drawn from this study:
 Eight slabs with different end restraint conditions were investigated to evaluate
the influence of end anchorage. The end restraint increased the capacity of both
185
solid and profiled slabs significantly regardless of the long term service load
conditions.
 There was a significant reduction in the ultimate strength of profiled
unrestrained slabs due to creep and shrinkage. However, the effects of creep
and shrinkage on profiled composite slabs with end anchorage were negligible.
186
Table 6.1 Details of test specimens

Specimen Slab End Long term
ID type anchorage loading condition
LTSR-SW Solid Yes Self-Weight

LTSR-SL Solid Yes Service load
LTPR-SW Profiled Yes Self-Weight
LTPR-SL Profiled Yes Service load
LTSU-SW Solid No Self-Weight
LTSU-SL Solid No Service load
LTPU-SW Profiled No Self-Weight
LTPU-SL Profiled No Service load
Table 6.2 Material properties of concrete, profiled steel sheets, reinforcing steel and
headed shear studs
Material Property Values
(MPa)
Concrete Compressive strength, 𝑓𝑐′ 24
(At age of 1366 Tensile strength, 𝑓𝑐𝑡 2.7
days after casting)
Young’s modulus, 𝐸𝑐 28,900
Profiled steel Yield stress, 𝑓𝑦 691

sheets
Ultimate strength, 𝑓𝑢 710
Young’s modulus, 𝐸𝑠 248,000
Reinforcing steel Yield stress, 𝑓𝑦 513
Headed shear studs Yield stress, 𝑓𝑦 399
187
Table 6.3 Test results

Specimen Maximum Deflection at
ID load maximum load
(kN) (mm)
LTSR-SW 90 41
LTSR-SL 87 37
LTPR-SW 102 46
LTPR-SL 79 91
LTSU-SW 63 78
LTSU-SL 60 94
LTPU-SW 39 47
LTPU-SL 42 48
188
Plan view
Section A-A (Solid Slab)
Section A-A (Profiled Slab)
Section B-B (Solid Slab) Section B-B (Profiled Slab)
Figure 6.1 Details of test specimens
189
Figure 6.2 Geometry of W-DEK profiled steel sheeting
Figure 6.3 Instrumentation
Figure 6.4 Specimen ready for testing
190
Figure 6.5 Load-deflection curves of LTSU-SW & LTSU-SL
Figure 6.6 Flexural cracking of LTSU-SW
191
Figure 6.7 Cracking across the width of the concrete slab of LTSU-SL
Figure 6.8 Flexural cracking at ultimate load of LTSU-SL
192
Figure 6.9 Load-deflection curves of LTPU-SW & LTPU-SL
Figure 6.10 Overall crack pattern at ultimate load of LTPU-SL
193
Figure 6.11 Shear bond failure of LTPU-SL
Figure 6.12 Load-deflection curves of LTSR-SW & LTSR-SL
194
Figure 6.13 Cracking around studs of LTSR-SL
Figure 6.14 Crack pattern at ultimate load of LTSR-SL
195
Figure 6.15 Load-deflection curves of LTPR-SW & LTPR-SL
Figure 6.16 Shear bond failure of LTPR-SW
196
Figure 6.17 Weld failure of a corner stud of LTPR-SL
Figure 6.18 Crack pattern of LTPR-SL
197
Figure 6.19 Comparison between behaviour of profiled and solid restrained slabs
Figure 6.20 Effects of creep and shrinkage (solid unrestrained slabs)
198
Figure 6.21 Effects of creep and shrinkage (solid restrained slabs)
Figure 6.22 Effects of creep and shrinkage (profiled unrestrained slabs)
199
Figure 6.23 Effects of creep and shrinkage (profiled restrained slabs)
Figure 6.24 Effects of end anchorage (LTPR & LTPU)
200
Figure 6.25 Effects of end anchorage (LTSR & LTSU)
Figure 6.26: Deflection of unrestrained slabs
201
Figure 6.27: Deflection of restrained slabs
Figure 6.28 Effects of composite action (LTPU & LTSU)
202
Figure 6.29 Effects of composite action (LTPR & LTSR)
203
204
CHAPTER 7 - Finite Element modelling of Composite Slabs with End Anchorage
CHAPTER 7
FINITE ELEMENT MODELLING OF COMPOSITE
SLABS WITH END ANCHORAGE
7.1 INTRODUCTION
This chapter presents a three-dimensional finite element model developed using
ABAQUS to simulate the behaviour of composite slabs with end anchorage. The
model was validated against experimental results for its accuracy and a parametric
study was conducted to investigate the influence of concrete strength, shear stud
strength and profiled sheeting thickness in profiled composite slabs.
7.2.1 General
The experimental programme described in the previous chapter provided the load-
deflection relationship for a particular concrete strength, stud strength and profiled
steel sheets. In order to generalise the results, the finite element method has been
employed. In this study, a three dimensional finite element model of the slab
specimens was developed using ABAQUS. ABAQUS/Explicit solver was also
205
chosen in this numerical study. The reason for this preference is stated in Section
4.2.1. Both geometric and material nonlinearity were considered in the FE analysis.
Appropriate material models for all components and suitable contact interactions
with real boundary conditions were specified to model all test specimens accurately
and efficiently.
Due to the symmetry of geometry and loading, only half of the slab was modelled.
Eight-node linear hexahedral solid elements with reduced integration (C3D8R) were
used to model the concrete, the headed shear studs and the steel beam. The reduced
integration was adopted to eliminate shear locking and increase computational
efficiency. The profiled steel sheets were modelled as 4-node doubly curved thin
shell, reduced integration elements (S4R). The reinforcing bars were modelled by
two-node three-dimensional linear truss elements (T3D2). An overview of the finite
element mesh is depicted in Figure 7.1.
7.2.3 Interactions
The contact interaction between the profiled steel sheets and the concrete is the
most critical part of this modelling process. The steel sheet elements were located at
a distance of half of the steel sheet thickness from the concrete elements. Radial-
thrust type connector elements (CONN3D2) were chosen to simulate the interaction
between the concrete and the profiled steel sheets. The reason for using radial thrust
type elements is described in Section 4.2.3. Each connector element connected a pair
of concrete and steel nodes that were closest to each other.
206
The reinforcement was assembled and embedded within the concrete slabs using
the embedded constraints. The effect of the bond-slip behaviour of reinforcing bars
was neglected. In addition, a contact interaction in the form of surface to surface
contact pair algorithm was applied between the top of steel beam and the bottom of
concrete slab or profiled steel sheeting which allowed separation of the surfaces after
contact in order to simulate any interface slip. The node to surface contact with finite
sliding was used while the hard contact penalty frictional formulation with a
coefficient of friction of 0.25 was specified as the contact property. Finite sliding
allows arbitrary separation, sliding, and rotation of the surfaces. This formulation
usually includes nonlinear geometric effects because such simulations generally
involve large deformations and large rotations (ABAQUS, 2012). The contact
surface of the steel beam was chosen as first surface (master surface in this case) and
the contact surface of the profiled sheeting was chosen as slave surface. A balanced
master/slave formulation is used so that potential contact penetrations are decreased.
The nodes on the bottom surfaces of headed shear studs were tied to the surface of
steel I-beam using Tie constraints. Tie constraint is a surface-based constraint based
on a master-slave formulation. Since the constraint prevents slave nodes from
separating or sliding relative to the master surface, it is an ideal contact interaction
for welded shear studs on the top of steel beam. The surface to surface contact
interactions were also used to specify contact between the shear studs and the
surrounding elements of the concrete slab. The contact surfaces of headed shear studs
were chosen as first surface (master surface in this case) and the contact surfaces of
concrete slab were chosen as slave surface. The default normal behaviour was
specified for all interactions. To allow the separation of concrete behind the stud,
207
“separation after contact” was allowed in normal contact property options available
in ABAQUS in case of stud-concrete interaction. The tangential properties of those
interactions were defined with a friction co-efficient of 0.25 (Nguyen and Kim,
2009). Penalty contact method using finite sliding formulation was specified for this
interaction.
A uniformly distributed load was applied using a plate of 20 mm thickness on the
top surface of the concrete using the displacement control and a smooth amplitude
function. A surface to surface interaction was used between the plate and the
concrete surface to ensure uniform load transfer. A roller support was applied to the
nodes of the profiled steel sheets in case profiled composite slab and to the nodes of
concrete bottom surface in case of solid slabs. In both cases, a strip of 50 mm width
of concrete or profiled steel sheeting surface was used. A ZSYMM symmetry
boundary condition was applied to simulate the continuity of the slab specimen in the
Z direction (Figure 7.1).
Since the test setup is under static load but the explicit solution method is used,
the approach followed in this numerical study is to analyse the problem as a quasi-
static solution. Therefore, load was applied slow enough to keep the inertia effect
insignificant. Internal (ALLIE) and kinetic (ALLKE) energies of the whole model
were compared to ensure a quasi-static response in each analysis. It was found that
the kinetic energy level was significantly below the internal energy level in all
208
simulations, as shown in Figure 7.2. ABAQUS post-processing program was used to
eliminate oscillation effect from the final results.
used to model the concrete material. The details of this model are described in
Section 4.2.6.1. The elastic properties were specified according to ACI 318
(American Concrete Institute, 2011). The elastic part of the stress-strain curve was
defined with a value of 28,900 MPa for the Young’s modulus and 0.2 for the
Poisson’s ratio. All plasticity parameters were specified very carefully. The value of
dilation angle was specified as 30°. Default values of the flow potential eccentricity
and the ratio of the compressive strength under biaxial loading to uniaxial
compressive strength were adopted. The value of ratio of the second stress invariant
on the tensile meridian to that on the compressive meridian was taken as 0.6667. The
value of viscosity parameter was zero. Plain concrete in uniaxial compression was
modelled as an elastic-plastic material with strain softening, as shown in Figure 7.3.
The stress-strain curve for concrete as proposed by Carreira and Chu (1985) was
used. The compressive strength at the age of 1366 days of casting was 24 MPa. The
uniaxial tensile response was assumed to be linear until the tensile strength of
concrete was reached. Beyond that the tensile softening response was defined by
means of fracture energy. The concrete properties used in the FE model are reported
in Tables 6.2 and 7.1.
209
7.2.6.2 Steel properties
The stress-strain relationship obtained from material tests was used for modelling
of the profiled steel sheeting, reinforcing bars and headed shear studs. An elastic-
perfectly plastic model was used as the constitutive law for the profiled steel sheets
and the reinforcing bars, as shown in Figures 7.4 and 7.5 respectively. The
mechanical behaviour for both tension and compression is assumed to be similar.
The elasticity of the profiled steel sheeting was defined with the * ELASTIC option
ratio. The plastic part of the stress-strain curve was defined with the * PLASTIC
option with a value of 691 MPa for the yield strength and 710 MPa for the ultimate
strength. The elasticity of the reinforcing bar was defined with the * ELASTIC
option with a value of 201,000 MPa for the Young’s modulus and 0.3 for the
Poisson’s ratio. The plastic part of the stress-strain curve was defined with the *
PLASTIC option with a value of 513 MPa for the yield strength and 637 MPa for the
ultimate strength. The headed shear stud connectors were modelled using an elastic-
plastic model with isotropic strain hardening, as shown in Figure 7.6. The elasticity
of the headed shear stud connectors was defined with the * ELASTIC option with a
value of 207,000 MPa for the Young’s modulus and 0.3 for the Poisson’s ratio. The
plastic part of the stress-strain curve was defined with the * PLASTIC option with a
value of 399 MPa for the yield strength and 549 MPa for the ultimate strength. The
properties used in the FE model for the profiled steel sheeting, reinforcing bars and
headed shear studs are tabulated in Tables 6.2 and 7.1. The loading plate has been
modelled using elastic properties only. The elasticity of the loading plate was defined
210
ratio.
The complex interaction between the concrete and the profiled steel sheets was
modelled using the interface elements available in ABAQUS. The local bond stress-
slip law for the W-DEK profile steel sheeting has been taken from push test study
conducted by Burnet and Oehlers (2001). However, this bond-stress slip
characteristic does not consider the effect of end anchorage. Therefore, it would
affect the results of FE model. Due to lack of accurate bond-slip law, it was decided
to use this law in the proposed FE model (Figure 7.7). The FE analyses indicate that
the accurate result for a particular slab can only be obtained when the shear bond
properties are accurate.
A mesh sensitivity analysis was carried out to determine the optimal mesh size
that provides relatively accurate solution with low computational time. Three
element sizes of 12.5 mm, 25 mm and 50 mm for concrete were chosen in the
preliminary development of the model to check the effect of discretization. The
results obtained from this mesh sensitivity study show that the overall load versus
deflection behaviour remains the same (Figure 7.8). Similarly, different sizes of stud
elements also produced identical load-deflection behaviour (Figure 7.9). It is worth
mentioning that the connector elements for all the element sizes were placed at the
same locations and the same material properties were used in every model for
211
consistency of analysis. Based on the results, it was decided to use a mesh of 25 mm
global seed size for the concrete and the profiled steel sheeting and 10 mm global
seed size for the headed shear studs for the rest of the analysis considering the
efficiency and accuracy of numerical simulations.
numerical results with the experimental results. Using the ABAQUS post-processing
program the plastic strain and stress distribution of the critical component of the
finite element model were observed and the modes of failure were determined. A
comparison between the load-deflection curves obtained from the experiments and
nonlinear finite element simulation is shown in Table 7.2. The mean value of
PFE/PEXPT ratios for maximum load carrying capacity is 1.01, with the corresponding
coefficients of variation (COV) of 0.11. The load-deflection curves from the finite
element analyses with corresponding experimental curves for all slab specimens are
presented in Figures 7.10, 7.11, 7.12 & 7.13. It can be observed that the FE model
follows the same trend as the experimental results in terms of strength and ductility
and trace the failure modes up to ultimate loads. However, the initial numerical
response for cracking strength of the slab is considerably stiffer than the
experimental one. This overestimation of stiffness can be partly attributed to the
tension stiffening in the concrete damage plasticity model which did not account the
initial damage of the material. In case of profiled slabs, the bond stress-slip model
has been obtained from push tests and the constraint in the push test differs from that
in composite slabs. Therefore, the bond stress-slip model could be introducing
greater stiffness in the FE model of profiled slabs. Additionally, end anchorages are
212
usually more flexible than relatively brittle longitudinal shear bond between the
concrete and the profiled sheeting (Degtyarev, 2014a). Therefore, the end anchorage
strength may not be fully mobilised in the test specimens when the longitudinal shear
strength is reached.
Limited experimental data regarding the bond stress-slip behaviour of W-DEK are
available in open literature. The bond stress-slip characteristics from push test study
conducted by Burnet (1998) were adopted to model the connectors used in this FE
model. ABAQUS (2012) requires the bond stress-slip law to be converted into force-
slip law for connector section properties. In this conversion, the total bonded area
between the concrete and the profiled sheeting was used which might be different
from the exact contributing area of the concrete-sheeting interface to resist the shear
at various stages of loading. This is also one of the limitations of this modelling
approach to obtain a good agreement between experimental and FE model results.
The experimental observations indicated that the difference of the long term
loading condition has no significant effect on the load-deflection behaviour.
Therefore, the load-deflection behaviour of specimens under self-weight only is
presented for comparison. The ultimate load carrying capacity obtained from FE
model of LTSU-SW is found to be in reasonable agreement with the experimental
curve, as shown in Figure 7.10. The FE model reached an ultimate load of 65 kN,
which is only 3% greater than the experimental value of 63 kN. The maximum load
of solid restrained slab LTSR-SW from FE model is 85 kN, which is a 5%
underestimation of experimental load carrying capacity (Figure 7.12). The numerical
response in cases of both solid slabs is ductile up to failure with an indication of
yielding of reinforcement, as observed in the experiment (Figure 7.14).
213
The FE model of the profiled unrestrained slab LTPU-SW achieved a maximum
load of 45 kN, which overestimates the experimental value of 39 kN (Figure 7.12).
This could be attributed to the fact that the bond-slip characteristics are obtained
from push tests and there is some obvious difference between the behaviour of
composite slabs and push tests. However, the failure mode in both FE model and
experiments is shear bond failure and local buckling of sheeting near load point, as
presented in Figures 7.15 & 7.16. The FE model of profiled restrained composite
slab LTPR-SW reached an ultimate load of 93 kN, whereas the experimental value of
maximum load is 102 kN (Figure 7.13). Again, the local bond-slip law obtained from
push tests is a possible explanation for the difference in maximum load carrying
capacity. The failure mode in FE model of LTPR-SW is shear bond failure
accompanied by tensile rupture of the bottom flange of the sheeting around shear
connectors, which is same as the failure mode observed in experiments (Figure 7.17).
In addition, the deformation of studs derived from FE model is very close to the
deformation observed in this slab specimens which was partly crushed after the test,
as seen in Figure 7.18.
It was shown that the finite element models yield accurate results to simulate the
nonlinear behaviour of composite slabs up to the ultimate load. Therefore, a
parametric study has been carried out to investigate the effect of the ultimate load
carrying capacity with the variation in concrete strength, stud strength and profiled
sheeting thickness. The FE model of profiled composite slabs was used for
parametric study.
214
7.3.3.1 Effects of concrete compressive strength
The load versus deflection relationship for the profiled composite slab considering
the variation in concrete compressive strength (𝑓𝑐′ ) is shown in Figure 7.19. It can be
seen that the load carrying capacity of the slab is increased in general with the
increase in the concrete compressive strength. This is due to the fact that when
concrete strength increased, the end anchorage strength mobilisation also increased.
However, increase in load carrying capacity of the slab is not significant for the
concrete strength above 40 MPa. This can be attributed to the fact that concrete
compressive strength has little influence on the degree of composite action
(Degtyarev, 2014a).
In this study, the profiled restrained composite slab with concrete strength of 25
MPa, 30 MPa, 35 MPa, 40 MPa, 45 MPa and 50 MPa achieved a load carrying
capacity of 91 kN, 94 kN, 97 kN, 97 kN, 95 kN and 96 kN respectively. Therefore,
25 MPa increase in concrete compressive strength contributed about 5% increase in
ultimate load carrying capacity of profiled composite slabs with studs.
7.3.3.2 Effects of profiled steel sheeting thickness
Figure 7.20 shows the load versus deflection relationship profiled composite slab
of different thickness (tp) of profiled sheeting ranging from 0.6 mm to 2.00 mm
respectively. It can be seen that the stiffness and capacity of the slab are increased
with the increase of profiled sheeting thickness. The modes of failure in all cases are
shear-bond failure accompanied by tensile rupture of the bottom flange of the
sheeting around shear connectors. The load carrying capacities of profiled composite
slabs with studs were 96 kN, 107 kN, 151 kN and 171 kN for 0.6 mm, 0.75 mm, 1.5
215
mm and 2.00 mm thick profile sheeting respectively. It is about 40% increase in load
carrying capacity when the thickness of this particular profile was doubled. However,
the load carrying capacity of the slab increased only 13% when the thickness of the
sheeting increased from 1.5 mm to 2.00 mm. Therefore, the optimum thickness of
this profiled sheeting ranges between 0.75 mm to 1.5 mm. In this parametric study,
only the thickness of profiled sheeting has been varied. In reality, the bond stress-slip
behaviour of composite slab varies also with the thickness of profiled sheeting
(Burnet, 1998). Because the purpose of this parametric study was to investigate the
effect of the thickness of the profiled sheeting on the behaviour of profiled composite
slabs with end anchorage, the model has been developed with single bond stress-slip
law that was adopted for W-DEK profile of thickness 1.00 mm. However, it is
acknowledged that more accurate stiffness and strength of the composite slabs with
studs can be found if the variable bond stress-slip laws with corresponding thickness
of profiled sheeting are used.
7.3.3.3 Effects of headed shear stud strength
Figure 7.21 shows the load versus deflection relationship for different headed
shear stud strengths (fy) in profiled composite slab. It can be seen that the stiffness
and capacity of the slab are not sensitive to the increase of stud strength from 300
MPa to 700 MPa. This is due to fact that the composite action between the deck and
the concrete reached full moment capacity of the section when an increase in the end
anchorage strength would not increase the load carrying capacity of the slab. It also
supports the findings of Chen (2003) that it is the shear bond stress-slip rather than
the strength of studs that governs the contribution of the end anchorage to the shear
bond capacity of composite slabs.
216
Strength of the composite slabs increases as the longitudinal shear strength
increases. Degtyarev (2014a) also found that composite slabs with higher
longitudinal shear strength require less shear studs to achieve the full moment
capacity of the composite section. The number of shear studs per trough is also
another factor in the determination of composite slabs strength. Chen (2003) found
that only increase of the numbers of shear studs per trough would not enhance the
shear bond capacity. This is because the shear forces transmitted to the studs are very
small when the shear bond-slip occurred. The FE model developed in this thesis
considered only one shear stud per trough. As the end anchorage is usually more
flexible than the concrete-sheeting interface, increase in strength of shear studs may
not be fully mobilised to the extent it requires to increase the stiffness and ultimate
strength of profiled composite slabs with studs.
A three dimensional nonlinear finite element model has been developed to
simulate the nonlinear behaviour of composite slabs with end anchorage accurately.
The feasibility of the model has been verified against the experimental results of the
test specimens. Results of parametric studies on profiled composite slab with end
anchorage provided by welded shear studs are presented. End anchorage strength can
be more fully mobilised in the composite slabs with stronger concrete. Shear bond
stress-slip between the concrete and the profiled steel sheet is more responsible
rather than the strength of studs to the shear bond capacity of composite slabs. The
217
thickness of profiled sheeting increases the load carrying capacity of profiled
composite slabs significantly.
218
Table 7.1 Properties used in FE model

Material Properties Values
Concrete Density 2400 kg/m3
Dilation angle 30°
flow potential eccentricity 0.1
Steel Density 7800 kg/m3
Table 7.2 Comparison of numerical results with experimental results

Specimen Maximum Maximum load PFE/PEXPT
ID load at at FE model,
experiment, PEXPT PFE
(kN) (kN)
LTSR-SW 90 85 0.94
LTPR-SW 102 93 0.91
LTSU-SW 63 65 1.03
LTPU-SW 39 45 1.15
Mean 1.01
COV 0.11
219
Figure 7.1 Overview of FE model
Figure 7.2 Energy level of LTPR-SW
220
Figure 7.4 Stress-strain curve of profiled steel sheeting
221
Figure 7.5 Stress-strain curve of steel reinforcing bar
Figure 7.6 Stress-strain curve of headed shear studs
222
Figure 7.7 Local bond-slip law (Burnet, 1998)
Figure 7.8 Effects of concrete element size
223
Figure 7.9 Effects of headed shear stud element size
Figure 7.10 Load-deflection curves of solid unrestrained slab
224
Figure 7.11 Load-deflection curves of solid restrained slab
Figure 7.12 Load-deflection curves of profiled unrestrained slab
225
Figure 7.13 Load-deflection curves of profiled restrained slab
Figure 7.14 Reinforcement yielding of LTSR-FE (concrete slab removed)
226
Figure 7.15 Local buckling of sheeting near the load point of LTPU-FE
Figure 7.16 Local buckling of sheeting near ultimate load of LTPU-SL
227
(a) FE model
(b) Experiment
Figure 7.17 Tensile rupture near stud area of LTPR-FE
228
(a) FE model
(b) Experiment
Figure 7.18 Stud deformation in LTPR-FE
229
Figure 7.19 Effects of concrete compressive strength (𝑓𝑐′ )
Figure 7.20 Effects of profiled steel sheeting thickness (tp)
230
Figure 7.21 Effects of headed shear stud strength (fy)
231
232
CHAPTER 8 - Conclusions
CHAPTER 8
CONCLUSIONS
8.1 CONCLUDING REMARKS
This thesis has presented a detailed investigation and analysis of the strength and
behaviour of post-tensioned composite steel-concrete slabs and composite slabs with
end anchorage. Experimental programme and numerical modelling conducted in this
thesis have provided valuable insight into the mechanism of longitudinal shear
transfer between the concrete and the profiled steel sheets.
One of the objectives of this thesis was to investigate the effects of prestress on
the bond stress-slip behaviour in post-tensioned composite slabs. In Chapter 3, the
results from a push test study are reported. The parameters investigated in this study
were (a) level of prestress and (b) bond length. This study explored that the prestress
has a detrimental effect on the bond between the profiled steel sheets and concrete in
post-tensioned composite slabs. The average chemical and mechanical bond stresses
in post-tensioned push specimens were found to be approximately 20% lower than
that in push-specimens without post-tensioning. The results of this study are
presented in a form that can be used as guidelines in the design of post-tensioned
composite slabs and development of new forms of profiled steel sheeting.
233
The effects of prestress on the chemical bond strength, mechanical bond strength
and bond strength at 5.0 mm slip of post-tensioned composite slabs have been
expressed as a simple linear relationship. The effects of prestress on the bond stress-
slip behaviour between the concrete and profiled steel sheeting should be considered
in order to determine the ultimate strength of post-tensioned composite slabs.
A significant outcome of this study is the development of a reasonably accurate
nonlinear 3D finite element model of shear bond behaviour of post-tensioned
composite slabs to simulate the influence of prestress on bond between the concrete
and profiled steel sheets. In Chapter 4, FE model was verified against the
experimental results of the load carrying capacities and failure modes of post-
tensioned push specimens. The FE model developed using the ABAQUS/Explicit
program was shown to be capable of simulating the overall bond stress-slip
behaviour between the concrete and profiled steel sheets comprising of both
chemical and mechanical bond stress. This indicates that this FE approach could be
used to study the behaviour of post-tensioned composite slabs with different types of
profile accurately and efficiently.
In Chapter 5, the bond stress-slip law obtained from the push test study was
assessed for post-tensioned composite slabs through finite element modelling. The
FE models of both post-tensioned solid and post-tensioned composite slabs were
verified against available experimental results in the literature. It was shown that an
accurate bond-slip law is of fundamental importance for the numerical modelling of
these types of slabs. Finite element models predicted the ultimate load carrying
capacities and failure modes with a reasonable accuracy. The ultimate moment
capacity of post-tensioned composite slabs obtained from FE model was 93% of the
234
experimental results. The results of a parametric study indicated that the stiffness and
strength of a post-tensioned composite slab can be increased with higher concrete
strength. The thickness of profiled steel sheeting has a significant effect on both
stiffness and strength of these types of slabs. Also, an increase in the level of
prestress increases the stiffness of the post-tensioned composite slabs.
In Chapter 6, the effects of end anchorage on the ultimate load carrying capacities
and failure modes of both solid and profiled slabs investigated through an
experimental study are presented. Eight slabs with different end restraint conditions
were investigated to evaluate the influence of end anchorage. The results from the
experimental programme of composite slabs with and without end anchorage have
highlighted the significant benefits that may be achieved with the use of end
anchorage at the end of slab span. The end restraint increased the capacities of both
solid and profiled slabs significantly, regardless of the long term service load
conditions. The ultimate load carrying capacity of solid restrained slabs was found
to be 40% greater than that of solid unrestrained slabs. On the other hand, the
ultimate load carrying capacity of profiled restrained slabs was found to be 150%
greater than that of profiled unrestrained slabs. There was a reduction of 24% in the
ultimate strength of profiled unrestrained slabs due to creep and shrinkage. However,
the effect of creep and shrinkage on profiled composite slabs with end anchorage is
negligible. The reduction in ultimate load carrying capacity of the solid unrestrained
and restrained slabs in long term was about 14% and 10% compared with that of the
short term solid slabs.
In Chapter 7, an accurate nonlinear 3D finite element model of composite profiled
slabs with end anchorage has been proposed. Results of parametric studies indicated
235
that the effects of concrete strength and sheeting thickness on the load-deflection
behaviour are significant. However, the shear bond stress-slip rather than the strength
of studs governs the contribution of the end anchorage to the shear bond capacity of
composite slabs.
It can be concluded that this thesis has achieved its objectives. It is expected that
experimental results and findings from numerical models in this research will be
useful for practising engineers and researchers of composite steel-concrete structures
and for the benchmarking of analytical and numerical modelling.
8.2 RECOMMENDATIONS FOR FURTHER RESEARCH
The experience and knowledge gained from this research lead to some
recommendations for further research.
 In Chapter 3, push tests on a particular profile were conducted. Further testing
is recommended for other available profiles in order to develop a unified bond
stress-slip model for post-tensioned composite slabs. Further experimental data
are also required to properly validate the adequacy of the proposed method of
finite element analysis using interface elements available in ABAQUS. This
study also offers an optimal design of the push test specimen for future push
test studies. The overhang of the profiled steel sheets can be reduced so that the
possibility of buckling can be eliminated. The premature buckling of the
specimen suggests that these test procedures should not be used in the
determination of bond stress for specimens of height above 1500 mm due to
buckling of the base of the sheets for sheet thicknesses of 1 mm.
236
 The findings of the push test study conducted in this thesis are assessed
numerically against limited test results available in the literature. Tests on post-
tensioned composite slabs with various levels of prestress to provide
benchmarking for further development of analytical and numerical models are
recommended. Finite element modelling for long term behaviour of post-
tensioned composite slabs using the FE approach used in this thesis is also
recommended.
 In Chapter 7, the local bond-slip law for the connector elements was obtained
from the limited available literature. Push tests representing end anchorage in
profiled composite slabs are recommended for more accurate numerical
modelling of composite slabs with end anchorage.
237
238
References
REFERENCES
ABAQUS (2012). ABAQUS CAE user's manual, Version 6.12, Dassault Systèmes
Simulia Corp., Providence, RI, USA.
Abdullah, R. (2004). Experimental evaluation and analytical modeling of shear bond
in composite slabs, PhD Thesis, Virginia Polytechnic Institute and State
University, Virginia, USA.
Abdullah, R. and Easterling, W. S. (2009). “New evaluation and modeling procedure
for horizontal shear bond in composite slabs”, Journal of Constructional Steel
Research, 65(4), 891-899.
Airumyan, E., Belyaev, V. and Rumyancev, I. (1990). “Efficient embossment for
corrugated steel sheeting”, IABSE Symposium on Mixed Structures Including New
Materials, Brussels, Belgium, 137-142.
American Concrete Institute (2011). Building code requirements for structural
concrete (ACI 318-11) and commentary, Farmington Hills, MI, USA.
239
References
An, L. (1993). Load bearing capacity and behavior of composite slabs with profiled
steel sheet, PhD Thesis, Chalmers University of Technology, Gothenburg,
Sweden.
An, L. and Cederwall, K. (1992). “Composite slabs analysed by block bending test”,
Proceedings of the 11th International Specialty Conference on Cold-Formed Steel
Structures, University of Missouri-Rolla, 268-282.
Bailey, C.G., Currie, P.M. and Miller, F.R. (2006). “Development of a new long span
composite floor system”, The Structural Engineer, 84(21), 32-39.
Bažant, Z. P. and Becq-Giraudon, E. (2002). “Statistical prediction of fracture
parameters of concrete and implications for choice of testing standard”, Cement
and Concrete Research, 32(4), 529-556.
BlueScope Steel (2007). LYSAGHT W-DEK® structural deck User’s Guide,
Lysaght Structural Solutions, Australia.
Bode, H. and Minas, F. (1997). “Composite slabs with and without end anchorage
under static and dynamic loading”, Composite construction-conventional and
innovative International conference, Innsbruck, Austria, 265-270.
Bode, H., Minas, F. and Sauerborn, I. (1996). “Partial connection design of
composite slabs”, Structural Engineering International, 6(1), 53-56.
British Standards Institution (2004). Eurocode 4: Design of composite steel and
concrete structures-Part 1-1: General rules and rules for buildings, EN 1994-1-1,
European Committee for Standardisation, Brussels.
240
References
Burnet, M. (1998). Analysis of composite steel and concrete flexural members that
exhibit partial shear connection, PhD Thesis, University of Adelaide, Australia.
Burnet, M. J. and Oehlers, D. J. (2001). “Rib shear connectors in composite profiled
slabs”, Journal of Constructional Steel Research, 57(12), 1267-1287.
Carreira, D. J. and Chu, K.H. (1985). “Stress-strain relationship for plain concrete in
compression”, ACI Structural Journal, 82(11), 797-804.
Charalambous, A., Ashford, P. and Sandie, R. B. (2002). “Post-tensioned concrete
floor system in Australia”, Advances in Building Technology, 2, 1051-1055.
Chen, S. (2003). “Load carrying capacity of composite slabs with various end
constraints”, Journal of Constructional Steel Research, 59(3), 385-403.
Chen, S. and Shi, X. (2011). “Shear bond mechanism of composite slabs - A
universal FE approach”, Journal of Constructional Steel Research, 67(10), 1475-
1484.
Crisinel, M. (1990). “Partial-interaction analysis of composite beams with profiled
sheeting and non-welded shear connectors”, Journal of Constructional Steel
Research, 15(1), 65-98.
Crisinel, M. and Marimon, F. (2004). “A new simplified method for the design of
composite slabs”, Journal of Constructional Steel Research, 60(3-5), 481-491.
Crisinel, M. and O'Leary, D. (1996). “Composite floor slab design and construction”,
Structural Engineering International, 6(1), 41-46.
241
References
Daniels, B. J. (1988). “Shear bond pull-out tests for cold-formed-steel composite
slabs”, ICOM - Construction Metallique, Department de Genie Civil, Ecole
Polytechnique Federale de Lausanne.
Daniels, B. J. and Crisinel, M. (1993a). “Composite slab behavior and strength
analysis. Part I: Calculation procedure”, Journal of Structural Engineering,
119(1), 16-35.
Daniels, B. J. and Crisinel, M. (1993b). “Composite slab behavior and strength
analysis. Part II: Comparisons with test results and parametric analysis”, Journal
of Structural Engineering, 119(1), 36-49.
Degtyarev, V. (2014a). “Strength of composite slabs with end anchorages. Part I:
Analytical model”, Journal of Constructional Steel Research, 94, 150-162.
Degtyarev, V. (2014b). “Strength of composite slabs with end anchorages. Part II:
Parametric studies”, Journal of Constructional Steel Research, 94, 163-175.
Easterling, W. S. and Young, C. S. (1990). “Strength of composite slabs”, Journal of
Structural Engineering, ASCE, 118(9): 2370-2389.
Ellobody, E. and Bailey, C. G. (2008). “Behaviour of unbonded post-tensioned one-
way concrete slabs”, Advances in Structural Engineering, 11(1), 107-120.
Ellobody, E. and Young, B. (2006). “Performance of shear connection in composite
beams with profiled steel sheeting”, Journal of Constructional Steel Research,
62(7), 682-694.
242
References
Ferrer, M., Marimon, F. and Crisinel, M. (2006). “Designing cold-formed steel
sheets for composite slabs: An experimentally validated FEM approach to slip
failure mechanics”, Thin-Walled Structures, 44(12), 1261-1271.
Ganesh, G. M., Upadhyay, A. and Kaushik, S. K. (2005). “Simplified design of
composite slabs using slip block test”, Journal of Advanced Concrete Technology,
3(3), 403-412.
Gilbert, R. I. and Mickleborough, N. C. (1990). Design of prestressed concrete, CRC
Press, United Kingdom.
Johnson, R. P. (2004). Composite Structures of Steel and Concrete, Volume 1 -
Beams, Slabs, Columns, and Frames for Buildings, 3rd Edition, Blackwell
Scientific Publications, Oxford.
Jolly, C. and Lawson, R. (1992). “End anchorage in composite slabs: an increased
load carrying capacity”, Journal of Structural Engineering, ASCE, 70(11), 202-
205.
Jolly, C. K. and Zubair, A. K. (1987). “The efficiency of shear-bond interlock
between profiled steel sheeting and concrete”, Proceedings of the International
Conference on Steel and Aluminium Structures, Composite Steel Structures -
Advances, Design and Construction, Cardiff, 127-136.
Kang, T. H.-K. and Huang, Y. (2012). “Nonlinear finite element analyses of
unbonded post-tensioned slab-column connections ”, Journal of Post-Tensioning
Institute, 8(1), 4-19.
243
References
Khan, S. and Williams, M. (1995). Post-tensioned concrete floors, Taylor & Francis,
London.
Koukkari, H. (1999). “Steel sheet in a post-tensioned composite slab”, Proceedings
of Eurosteel '99, Prague, Czech Republic, 199-202.
Koukkari, H. and Eligehausen, R. (2001). “Anchorage zone in a steel-concrete
composite slab with unbonded tendons”, International Symposium on
Connections between Steel and Concrete, Stuttgart, Germany, 1223-1232.
Liu, X., Wang, L. and Wang, B. (2009). “Full-range analysis on flexural
performance of profiled sheet and prestressed concrete composite slabs”, Journal
of Shenyang Jianzhu University (Natural Science), 5, 25.
Malaska, M. and Pajari, M. (1996). Long-term behaviour of composite slabs
prestressed with unbonded tendons, Research Notes 1778, Technical Research
Centre of Finland.
Miller, F.R. and Bailey, C.G. (2007). “Design concept for a novel long span
composite floor system”, The Structural Engineer, 85(21),38-47.
Mirza, O. and Uy, B (2013). “Experimental study of short and long term behaviour
of composite steel-concrete slabs”, Concrete in Australia, 38(4), 41-49.
Nguyen, H. T. and Kim, S. E. (2009). “Finite element modeling of push-out tests for
large stud shear connectors”, Journal of Constructional Steel Research, 65(10),
1909-1920.
244
References
Oehlers, D. J. and Bradford, M. A. (1995). Composite steel and concrete structural
members: fundamental behaviour, Pergamon Press, Oxford.
Oehlers, D. J. and Bradford, M. A. (1999). Elementary behaviour of composite steel
and concrete structural members, Butterworth-Heinemann, Oxford.
Papanikolaou, V. K. and Kappos, A. J. (2007). “Confinement-sensitive plasticity
constitutive model for concrete in triaxial compression”, International Journal of
Solids and Structures, 44(21), 7021-7048.
Patrick, M. (1990). “A new partial shear connection strength model for composite
slabs”, Journal of the Australian Institute of Steel Construction, 24(3), 2-17.
Patrick, M. and Bridge, R. Q. (1992). “Design of composite slabs for vertical shear”,
Proceedings of an Engineering Foundation Conference on Composite
Construction in Steel and Concrete II, ASCE, 304-322.
Patrick, M. and Bridge, R. Q. (1994). “Partial shear connection design of composite
slabs”, Engineering Structures, 16(5), 348-362.
Patrick, M. and Lloyde, M. (2008). “Development of moment capacity tables for
post-tensioned composite slabs incorporating high-tensile steel decking, validated
by test data”, Proceedings of the Australasian Structural Engineering Conference
(ASEC), Australia, 803.
Patrick, M. and Poh, K. (1990). “Controlled test for composite slab design
parameters”, IABSE Symposium on Mixed Structures including New Materials,
Brussels, Belgium, 227-231.
245
References
Plooksawasdi, S. (1977). Evaluation and design formulations for composite steel
deck, Phd Thesis, West Virginia University, Morgantown, West Virginia.
Porter, M. L., Ekberg , C. E., Greimann, L. F. and Elleby, H. A. (1976). “Shear-bond
analysis of steel-deck-reinforced slabs”, Journal of the Structural Division, 102
(ST12), 2255-2268.
Porter, M. L. and Ekberg C.E. (1978). “Compendium of ISU research conducted on
cold-formed steel deck-reinforced slab systems”, Engineering Research Institute,
Iowa State University, Ames, Iowa.
Porter, M. L. and Greimann, L. F. (1984). “Shear-bond strength of studded steel deck
slabs”, Proceedings of the 7th International Specialty Conference on Cold-Formed
Steel Structures, St. Louis, USA.
Ranzi, G., Al-Deen, S., Ambrogi, L. and Uy, B. (2013a). “Long-term behaviour of
simply-supported post-tensioned composite slabs”, Journal of Constructional
Steel Research, 88, 172-180.
Ranzi, G., Al-Deen, S., Hollingum, G., Hone, T., Gowripalan, S. and Uy, B. (2013b).
“An experimental study on the ultimate behaviour of simply-supported post-
tensioned composite slabs”, Journal of Constructional Steel Research, 89, 293-
306.
Schravendeel, P., de Sitter, W. and Klop, G. (1990). “Prestressed post-tensioned
composite floor slabs”, IABSE Symposium on Mixed Structures, including New
Materials, Brussels, 173-178.
246
References
Schuster, R. M. (1970). Strength and behavior of cold-rolled steel-deck-reinforced
concrete floor slabs, PhD Thesis, Iowa State University, Ames, Iowa.
Seleim, S.S and Schuster, R.M. (1985). “Shear-bond resistance of composite deck-
slabs”, Canadian Journal of Civil Engineering, 12(2), 316-324.
Standards Australia (1996). Australian Standard AS 2327.1-1996, Composite
structures Part 1: Simply supported beams, Standards Australia International Ltd.
Standards Australia (1999). Australian Standard AS 1012.9-1999, Methods of testing
concrete - determination of the compressive strength of concrete specimens,
Standards Australia International Ltd.
Standards Australia (2000). Australian Standard AS 1012.10-2000, Methods of
testing concrete - determination of the indirect tensile strength of concrete
cylinders (Brazil or splitting test), Standards Australia International Ltd.
Standards Australia (2007). Australian Standard AS 1391-2007, Metallic materials -
Tensile testing at ambient temperature, Standards Australia International Ltd.
Standards Australia / Standards New Zealand (2007). Australian/New Zealand
Standard AS/NZS 4672.1:2007, Steel prestressing materials Part 1: General
requirements, Standards Australia International Ltd.
Stark, J.W. (1978). “Design of composite floors with profiled steel sheet”,
Proceedings of the 4th International Specialty Conference on Cold-Formed Steel
Structures, St. Louis Missouri-Rolla, 893-922.
247
References
Stramit (2007). Stramit Condeck HP: composite slab system: product technical
manual, Stramit Building Products, Australia.
TEKES (2001). Finnsteel Technology Programme 1995-2000: Final Report
Technology Programme Report 7/2001, TEKES (National Technology Agent of
Finland).
Tenhovuori, A. and Leskela, M.V. (1998). “Longitudinal shear resistance of
composite slabs”, Journal of Constructional Steel Research, 46(1), 228-228.
Tremblay, R., Rogers, C.A., Gignac, P. and Degrange, G. (2002). “Variables
effecting the shear-bond resistance of composite floor deck systems”, Proceedings
of the 16th International speciality Conference on Cold-Formed Steel Structures,
Orlando, Florida, 663-676.
Tsalkatidis, T. and Avdelas, A. (2010). “The unilateral contact problem in composite
slabs: Experimental study and numerical treatment”, Journal of Constructional
Steel Research, 66(3), 480-486.
Vasdravellis, G., Uy, B., Tan, E. and Kirkland, B. (2012). “Behaviour and design of
composite beams subjected to negative bending and compression”, Journal of
Constructional Steel Research, 79, 34-47.
Veljkovic, M. (1995). “Longitudinal shear capacity of composite slabs”, Nordic steel
Construction Conference ’95, Swedish Institute of Steel Construction and
Division of Steel structures, Malmö, Sweden, 547-554.
Veljkovic, M. (1996). Behaviour and Resistance of Composite Slabs, PhD Thesis,
Luleå University of Technology, Luleå, Sweden.
248
References
Veljkovic, M. (1998). “Influence of load arrangement on composite slab behaviour
and recommendations for design”, Journal of Constructional Steel Research,
45(2), 149-178.
Widjaja, B. R. (1997). Analysis and design of steel deck-concrete composite slabs,
PhD thesis, Virginia Polytechnic Institute and State University, Virginia, USA.
Wright, H.D., Evans, H.R. and Harding, P. (1987). “The use of profiled steel
sheeting in floor construction”, Journal of Constructional Steel Research, 7(4),
279-295.
249
References
250
Appendix
APPENDIX A
Table A.1 Material properties of profiled steel sheets of push tests

Specimen Maximum Young's Yield Ultimate % elongation
Load Modulus Strength Strength at fracture
(kN) (MPa) (MPa) (MPa) (over 50 mm)
1 8.4 235731 678.3 685.7 7.16

2 8.3 226761 669 676.8 7.81
3 8.4 188640 674 681.5 6.6
Average 8.4 217044 674 681 7.19
Table A.2 Material properties of prestressing bar of push tests

Specimen Maximum Young's Yield Ultimate % elongation
Load Modulus Strength Strength at fracture
(kN) (MPa) (MPa) (MPa) (over 50 mm)
1 88.3 200085 1054.1 1123.9 14.72
2 87.7 200970 1049.6 1116.5 13.17
3 87.6 201271 1050.4 1115.2 11.49
Average 87.9 200775 1051 1119 13.13
251
Appendix
Figure A.1 Tensile test results of profiled steel sheets
252
Appendix
Sample calculation for Bond Stress:
Specimen ID: PT-500-0
Maximum load before breakdown of

chemical bond Fc 60.61 kN
Height of bond area H 500 mm
Perimeter of profiled sheeting L 461.52 mm
Contact area A = H*L 230760 mm2
Chemical bond stress τ = Fc*1000/A 0.26 MPa
Minimum load after breakdown of

chemical bond Fr 55.38 kN
Residual bond stress τ = Fr*1000/A 0.24 MPa
Average load between maximum

load and load at 10 mm slip Fm 55.38 kN
Mechanical bond stress τ = Fm*1000/A 0.24 MPa
253
Appendix
APPENDIX B
Slab span l 6 m
Slab width b 0.9 m
Slab thickness t 0.18 m
Cross-sectional area of concrete Ac 162000 mm2
Cross-sectional area of tendon
(3 strands @12.5 mm diameter) Ap 295.8 mm2
Cross-sectional area of profiled steel sheets
(Condeck 1.00 mm) As 1458 mm2
Moment of inertia I 437400000 mm4
Eccentricity of tendon (SS) ess 45 mm
Eccentricity of tendon (CK) eck 20 mm
Concrete compressive strength f'c 33.7 MPa
Yield strength of tendon fpy 1760 MPa
Yield strength of Condeck fy 599 MPa
Elastic modulus of tendon Ep 195218.4 MPa
Elastic modulus of profiled steel sheets Es 195000 MPa
Elastic modulus of concrete Ec 30673.75 MPa
Neutral axis parameter γ 0.85
254
Appendix
𝛾 = 1.0 − 0.003𝑓𝑐′
Figure B.1 Geometry of post-tensioned composite slab (Ranzi et al., 2013a)
Self-weight of slab w 3.888 kN
Effective prestress under

𝑤𝐿2
balanced load condition 𝑃𝑒 = 388.8 kN
8𝑒
Post-tensioned solid slab
Cross-section(Not in scale) Strain Stress Force
Figure B.2 Stress-strain distribution of post-tensioned solid slab
255
Appendix
𝑃𝑒
𝜀𝑝 = 𝐴
Strain compatibility 𝑝 𝐸𝑝 0.0067330
𝜀𝑐𝑒 = 𝐸 ( 𝐴𝑒 + ) 0.0001369
𝑐 𝐼
𝜀𝑝𝑡 = 0.003 𝑑𝑛
𝜀𝑝𝑢 = 𝜀𝑝𝑒 + 𝜀𝑐𝑒 + 𝜀𝑝𝑡
Force equilibrium 𝐶 = 0.85𝑓′𝑐 𝑏𝛾𝑑𝑛 21913.43dn
Assuming tendon yields 𝑇 = 𝐴𝑝 𝑓𝑝𝑦 295.8*1760
𝐶=𝑇
𝑑𝑛 = 23.76 m
135−𝑑𝑛
𝜀𝑝𝑡 = 0.003 0.014047255
𝑑𝑛
𝜀𝑝𝑢 = 𝜀𝑝𝑒 + 𝜀𝑐𝑒 + 𝜀𝑝𝑡 0.0209172
𝑓𝑦
𝜀𝑝𝑦 = 𝐸
𝑝 0.009015544
𝜀𝑝𝑢 > 𝜀𝑝𝑦
So tendons yielded.
Ultimate moment capacity
𝛾𝑑𝑛
𝑀𝑢 = 𝐴𝑝 𝜎𝑝𝑢 (𝑑𝑝 − ) 65.03 kNm
2
256
Appendix
Figure B.3 Determination of ultimate moment capacity
𝑤𝐿 𝑤𝐿
𝑀𝑢 = (𝑃𝑢 + ) ∗ 3 − 𝑃𝑢 ∗ 0.3 − ∗ 1.8
2 2
𝑀𝑢 = 2.7𝑃𝑢 + 4.32𝑤
Ultimate load carrying capacity
𝑃𝑢 = (𝑀𝑢 − 4.32𝑤)/2.7 35.73 kN
Post-tensioned composite slab (Condeck)
Figure B.4 Stress-strain distribution of post-tensioned composite slab (Condeck)
257
Appendix
𝑃𝑒
𝜀𝑝 = 𝐴
Strain compatibility 𝑝 𝐸𝑝 0.0067330
𝜀𝑐𝑒 = 𝐸 ( 𝐴𝑒 + ) 0.0000898
𝑐 𝐼
𝜀𝑝𝑡 = 0.003 𝑑𝑛
𝜀𝑝𝑢 = 𝜀𝑝𝑒 + 𝜀𝑐𝑒 + 𝜀𝑝𝑡
Force equilibrium 𝐶 = 0.85𝑓′𝑐 𝑏𝛾𝑑𝑛 21913.43dn
Assuming tendon and profiled
steel sheets yields 𝑇 = 𝐴𝑝 𝑓𝑝𝑦 + 𝐴𝑠 𝑓𝑦 295.8*1760+1458*599
𝐶=𝑇
𝑑𝑛 = 63.61 m
110−𝑑𝑛
𝜀𝑝𝑡 = 0.003 0.002187726
𝑑𝑛
𝜀𝑝𝑢 = 𝜀𝑝𝑒 + 𝜀𝑐𝑒 + 𝜀𝑝𝑡 0.0090105
𝑓𝑦
𝜀𝑝𝑦 = 𝐸
𝑝 0.009015544
𝜀𝑝𝑢 ≈ 𝜀𝑝𝑦
So tendons yielded and assumption is valid.
𝑓𝑦
𝜀𝑦 = 𝐸 0.003071795
𝑠
𝜀𝑠 0.004762252
𝜀𝑠 > 𝜀𝑦
So profiled steel sheets yielded.
258
Appendix
𝛾𝑑𝑛 𝛾𝑑𝑛 163.32 kNm

𝑀𝑢 = 𝐴𝑝 𝜎𝑝𝑢 (𝑑𝑝 − ) + 𝐴𝑠 𝜎𝑠𝑢 (𝑑𝑠 − )
2 2
𝑃𝑢 = (𝑀𝑢 − 4.32𝑤)/2.7 108.54 kN
Partial shear connection
Degree of shear connection 𝜂=0
𝑀𝑑𝑒𝑐𝑘 6.59 kNm
𝑀𝑠𝑜𝑙𝑖𝑑 𝑠𝑙𝑎𝑏 65.03 kNm
𝑀𝑜𝑠𝑐 𝑀𝑠𝑜𝑙𝑖𝑑 𝑠𝑙𝑎𝑏 + 𝑀𝑑𝑒𝑐𝑘 71.61 kNm
𝑀𝑓𝑠𝑐 163.32 kNm
From experiment, 𝑀𝑝𝑠𝑐 130.45 kNm
𝑦 = 𝑚𝑥 + 𝑐
71.61 = 𝑚. 0 + 𝑐 c 71.61
162.65 = 𝑚. 1 + 71.61 m 91.71
𝑦 = 91.04𝑥 + 71.61
130.45 = 91.04𝜂 + 71.61 ŋ 0.64
259
Appendix
ƞ Mu
(kNm)
0 71.61
1 162.65
0.65 130.45
Figure B.5 Mu vs. ƞ relationship (Condeck)
Using partial shear connection, 𝜂 = 0.64
Assuming profiled steel sheets
and tendon yields 𝐶 = 0.85𝑓′𝑐 𝑏𝛾𝑑𝑛 21913.43dn
0.64*(295.8*176+1
𝑇 = 𝜂(𝐴𝑝 𝑓𝑝𝑦 + 𝐴𝑠 𝑓𝑦 ) 458*599)
𝐶=𝑇 40.81 m
260
Appendix
𝑑𝑛 =
135−𝑑𝑛
𝜀𝑝𝑡 = 0.003 0.006924104
𝑑𝑛
𝜀𝑝𝑢 = 𝜀𝑝𝑒 + 𝜀𝑐𝑒 + 𝜀𝑝𝑡 0.0137469
𝑓𝑦
𝜀𝑝𝑦 = 𝐸
𝑝 0.009
𝜀𝑝𝑢 > 𝜀𝑝𝑦
So tendons yielded.
𝑓𝑦
𝜀𝑦 = 𝐸 0.003071795
𝑠
𝜀𝑠 0.00909932
𝜀𝑠 > 𝜀𝑦
So profiled steel sheets yielded.
𝛾𝑑𝑛 𝛾𝑑𝑛 113.45 kNm

𝑀𝑢 = 𝜂 (𝐴𝑝 𝜎𝑝𝑢 (𝑑𝑝 − ) + 𝐴𝑠 𝜎𝑠𝑢 (𝑑𝑠 − ))
2 2
𝑃𝑢 = (𝑀𝑢 − 4.32𝑤)/2.7 84.03 kN
261
Appendix
APPENDIX C
Table C.1 Material properties of N12 Reinforcing steel bars

Specimen Young's Yield Ultimate % elongation
Modulus Strength Strength at fracture
(MPa) (MPa) (MPa) (over 50 mm)
1 199948 518.6 639.8 14.00
2 200522 520.0 641.8 26.00
3 203565 501.5 630.9 12.00
Average 201345 513.4 637.5 17.3
Table C.2 Material properties of headed stud shear connectors

1 201993 376.2 526.4 26.67
2 210647 422.1 574.7 23.33
3 209296 400.1 547.9 26.67
Average 207312 399.5 549.7 25.56
262
Appendix
Table C.3 Material properties of profiled steel decking - W-DEK

1 243086 690.2 711.5 7.04
2 254057 697.6 714.2 8.88
3 247953 693.0 712.2 4.82
4 247071 681.6 704.0 6.82
Average 248041 690.6 710.5 6.89
Figure C.1 Load-slip curves of LTPU-SW
263
Appendix
Figure C.2 Load-slip curves of LTPU-SL
Figure C.3 Load-slip curves of LTPR-SW
264
Appendix
Figure C.4 Load-slip curves of LTPR-SL
265

Public Version PDF

Enviado por

Dados do documento

Título original

Direitos autorais

Formatos disponíveis

Compartilhar este documento

Compartilhar ou incorporar documento

Opções de compartilhamento

Você considera este documento útil?

Este conteúdo é inapropriado?

Direitos autorais:

Formatos disponíveis

Public Version PDF

Enviado por

Direitos autorais:

Formatos disponíveis

THE EFFECTS OF BOND AND ANCHORAGE ON

THE BEHAVIOUR AND DESIGN OF

Mohammad Masud Rana

A thesis in fulfilment of the requirements for the degree of

Centre for Infrastructure Engineering and Safety

School of Civil and Environmental Engineering, Faculty of Engineering

The University of New South Wales, Sydney, Australia

Composite steel-concrete slabs represent a popular form of flooring system in

numerical analyses into the behaviour of post-tensioned composite slabs and

composite slabs with end anchorage.

to the behaviour and strength of post-tensioned composite slabs. Therefore, the

possible influence of prestress on this bond behaviour is a fundamental issue for

of a post-tensioned push specimen using ABAQUS to simulate the bond stress-slip

then assessed for post-tensioned profiled composite slabs numerically. A parametric

overall performance of this type of slabs.

long term loading conditions. A three-dimensional finite element model is developed

and profiled sheeting thickness in profiled composite slabs.

In general, the findings of the experimental studies and numerical analyses

presented in this thesis herein are expected to be contributing a further knowledge to

the continuing improvement of composite steel-concrete structures.

of Philosophy in Structural Engineering at the University of New South Wales

(UNSW), Sydney, Australia. The candidate was supervised by Professor Brian Uy

School of Civil and Environmental Engineering, UNSW. The papers submitted or

published in internationally renowned journals and conferences are listed in the

bond stress-slip relationship for post-tensioned composite slabs”, Journal of

Constructional Steel Research, 114, 362-379.

anchorage in composite slabs”, Journal of Constructional Steel Research, 115,

composite slabs”, Journal of Constructional Steel Research. (in preparation)

experimental and numerical study”, Journal of Constructional Steel Research. (in

post-tensioned composite steel-concrete slabs”, Proceedings of the 10th Pacific

Structural Steel Conference (PSSC 2013) - Advancements and Achievements in

Structural Steel, Research Publishing Services, Singapore, 1058 - 1063.

tensioned composite steel-concrete slabs”, Composite Construction VII 2013

Conference, Queensland, Australia. (in press)

I would like to express my sincere gratitude to my supervisor, Professor Brian Uy,

serving on my supervisory panel.

supported under Australian Research Council’s Linkage projects (LP0990190) and

by Stramit Building Products and Arup as Partner Organisations and in collaboration

was supported by Australian Research Council (LP0882929). I would also like to

University of New South Wales and University of Western Sydney.

encouragement, inspirations and understanding.

1.1 Introduction ...................................................................................................... 1

2.1 Introduction .................................................................................................... 11

3.1 Introduction .................................................................................................... 47

FINITE ELEMENT MODELLING OF PUSH TESTS

4.1 Introduction .................................................................................................. 109

FINITE ELEMENT MODELLING OF POST-TENSIONED COMPOSITE

5.1 Introduction .................................................................................................. 141

EXPERIMENTAL PROGRAMME OF COMPOSITE SLABS WITH END

6.1 Introduction .................................................................................................. 175

FINITE ELEMENT MODELLING OF COMPOSITE SLABS WITH END

7.1 Introduction .................................................................................................. 205

8.1 Concluding remarks ..................................................................................... 233

REFERENCES ......................................................................................................... 239

Ac cross-sectional area of concrete

Ap cross-sectional area of prestressing steel

As cross-sectional area of profiled steel sheeting

C concrete compressive force

da maximum aggregate size

dn depth of neutral axis

ds depth of profiled steel sheeting

e plastic flow potential eccentricity

Ec elastic modulus of concrete