Tese Soares

Influence of slab continuity on punching resistance
A thesis submitted to Imperial College London for the degree of

Doctor of Philosophy (PhD)
Luis Fernando Sampaio Soares

BEng, MSc
Department of Civil and Environmental Engineering

Imperial College of Science, Technology and Medicine,
London, SW7 2AZ, United Kingdom
It’s not about what you have.
It’s what you do with what you have.
This work is dedicated to my parents
Luis Alfredo Soares and Ivone Soares
Declaration
I hereby confirm that this thesis is the result of my own work carried out in the Structures Section of
the Department of Civil and Environmental Engineering at Imperial College London, and that I give
appropriate references and citations whenever I referred to, described, or quoted any work from
others, whether published or unpublished.
The copyright of this thesis rests with the author and is made available under a Creative Commons
Attribution Non-Commercial No Derivatives licence. Researchers are free to copy, distribute or
transmit the thesis on the condition that they attribute it, that they do not use it for commercial
purposes and that they do not alter, transform or build upon it. For any reuse or redistribution,
researchers must make clear to others the licence terms of this work.
Luis Fernando Soares

London, February 2017
Abstract
This research addresses the design of reinforced concrete flat slabs for punching shear. Flat slabs
are two way spanning slabs without downstand beams that are directly supported by columns. The
thickness of flat slabs is typically governed by considerations of either deflection control or
punching shear for which there is still no widely accepted theoretical treatment. The research seeks
to develop a better understanding of the effects of in-plane restraint and flexural continuity on the
punching resistance of flat slabs. The overriding aim is to determine whether flexural continuity is
sufficient to explain the satisfactory performance of existing flat slabs which appear punching shear
deficient when assessed against test data from isolated punching shear specimens. The influence of
slab continuity on punching resistance is assessed through analysis of test data, nonlinear finite
element analysis (NLFEA) and fib MC2010 Level of Approximation (LoA) IV which is based on
the Critical Shear Crack Theory (CSCT) of Muttoni.
The CSCT relates shear resistance to the width of the so-called ‘critical shear crack’ which depends
on slab rotation. MC2010 has four LoA of which I to III are intended for design. Punching
resistance is normally determined from tests on isolated slab-column specimens which are intended
to simulate behaviour within the elastic line of contraflexure for radial moments, although the
benefits of continuity on punching shear strength is well known. MC2010 LoA IV, which uses
nonlinear finite element analysis to determine the slab rotation, is used to investigate and develop a
better understanding of the influence of continuity on the punching shear resistance of flat slabs at
edge and internal columns. The FE model is calibrated against relevant test data from both isolated
and continuous specimens.
Parametric studies are presented which show the superseded UK design code BS8110 to require
significantly less punching shear reinforcement than EC2 (2004) and MC2010 LoA II which give
reasonable strength predictions for isolated internal column punching specimens. Despite this, slabs
designed to BS8110 have performed satisfactorily for many years. MC2010 LoA IV is used to show
that punching resistance at internal columns can be increased significantly by restraint from the
surrounding slab. The increase in punching resistance due to continuity is shown to be sufficient to
explain the satisfactory performance of flat slabs designed to BS8110 at internal columns. However,
the strength increase is variable and, in the case of uniformly loaded slabs, already largely included
in BS8110 and EC2.
The research also considers the influence of slab continuity on punching resistance at edge columns.
Parametric studies are developed to examine the influence of shear force eccentricity, flexural
9
reinforcement distribution between support and span, and modelling assumptions on the punching
resistance of flat slabs at edge columns. Relating the punching resistance at edge columns to the
elastic unbalanced moment as done in MC2010 LoA II and III is shown to be overly conservative
for continuous slabs. The ACI 318 and EC2 practice of making the design punching resistance
independent of the unbalanced moment is reviewed and shown to be reasonable particularly for
continuous slabs. MC2010 LoA IV is shown to be a powerful tool for assessment, however the
predictions for the method are very sensitive to the modelling assumptions and boundary
conditions.
10
Acknowledgements
This section should have been the biggest one in here. There is absolutely no way I could have
accomplished all this on my own. So, I would like to thank:
Luis Alfredo Soares for being the best P3 (Pai × Professor × PhoDa);
Ivone Soares, Luis Alfredo Filho, Michelle, Polianna, Maria Fernanda, Henrique, Tom, and the
ones to come;
My whole family, specially my beautiful godmother Ana Maria, Nina Mochel, Cláudio Luis
Guterres and André Ricardo Soares;
Robert Vollum, who should definitely be teaching "Supervising 101". Brilliant, hands-on, and
always available. Thanks, Boss;
João Celso Marques, Henrique Mariano, Pnina Ari-Gur, and Mounir El Debs;
Frederico Araújo, Marcela Martins, and Osvaldo Barbosa;
My MdA friends;
My favourite Imperial folks: Ananth 'brown sugar' Gopalakrishnan, Marcus 'Bro' Pastore, Jason 'Be
Goode' Bennett, Nívea 'Gabizeira' Albuquerque, Clément 'nice place' Favier, Nicolas ‘Shaffle’ or
'Shuffle' or ‘Schaeffer’, Robert Wright 'Blues Band', Christina 'Pedro' Bakopoulou, Simon 'no
touching' Parker, Karl 'on the violin' Smith, Срђан 'daddy likes it' Станковић, Magnus 'once a year'
Beyer, Liz 'Liu Long', George ‘recognition’ Koudis, Laura 'quoi?' Bellamy, Spyros 'not on
keyboard' Masouros, and Joseph 'OCD' Sherwood;
The Vollum family, which includes Marianna Micallef, Mohemmed Abdelsalan, Jean Paul Vella,
Andri Setiawan and Abobakr Elwakeel;
Sloth 'the first' and Sloth 'the second', my beloved PCs during my 4 PhD years;
Thiago Revil for the help with the Personal Statement;
My musical friends from The Favourites (cheers, Christopher!), Sam TM and The Gastric Band;
The folks from the Futsal List. Those matches saved my week so many times... and you lot still owe
me money;
André 'Russo' Ramos, Francisco 'meia boca' Quim, and Rodrigo 'Maricota' Barros from Saint
Charles;
Rodrigo Neves, Elys Regina, Antonio Parga, Fernando Lima, Virgínia Freire and Evandro Gomes
from IFMA.
11
12
Contents
Abstract .................................................................................................................................... 9
Acknowledgements ................................................................................................................. 11
Contents .................................................................................................................................. 13
List of Figures ......................................................................................................................... 17
List of Tables .......................................................................................................................... 27
Notations ................................................................................................................................. 29
1 Introduction ....................................................................................................................... 37
1.1 General Aspects ............................................................................................................................................. 37
1.2 Aims and Objectives ...................................................................................................................................... 39
1.3 Outline of this thesis ...................................................................................................................................... 39
2 Background and Literature Review .................................................................................. 41
2.1 General Aspects ............................................................................................................................................. 41
2.1.1 Shear transfer actions............................................................................................................................ 41
2.1.2 Shear Reinforcement ............................................................................................................................ 45
2.1.3 Shear Failure Modes ............................................................................................................................. 48
2.1.4 Size Effect ............................................................................................................................................ 49
2.1.5 Flexural reinforcement ......................................................................................................................... 50
2.2 Influence of continuity on Punching Shear Resistance ............................................................................... 51
2.2.1 General Aspects .................................................................................................................................... 51
2.2.2 Continuity ............................................................................................................................................. 51
2.3 Mechanical Models ........................................................................................................................................ 55
2.3.1 General approaches............................................................................................................................... 55
2.3.2 Kinnunen and Nylander (1960) ............................................................................................................ 58
2.3.3 Critical Shear Crack Theory ................................................................................................................. 60
2.3.4 Tangential Strain Theory ...................................................................................................................... 73
2.4 Codes of practice ............................................................................................................................................ 78
2.4.1 General Aspects .................................................................................................................................... 78
2.4.2 BS8110 ................................................................................................................................................. 78
2.4.3 Eurocode 2 ............................................................................................................................................ 81
2.4.4 ACI 318-14 ........................................................................................................................................... 86
2.4.5 MC2010 ................................................................................................................................................ 89
2.5 Conclusions..................................................................................................................................................... 94
3 Finite Element Modelling .................................................................................................. 97
3.1 Introduction.................................................................................................................................................... 97
3.2 Diana ............................................................................................................................................................... 97
13
Contents
3.2.1 Structural Elements .............................................................................................................................. 97

3.2.2 Crack modelling in NLFEA ............................................................................................................... 102
3.2.3 Compressive Behaviour ...................................................................................................................... 105
3.2.4 Tension Softening ............................................................................................................................... 112
3.2.5 Behaviour in Shear ............................................................................................................................. 113
3.2.6 Reinforcement modelling ................................................................................................................... 115
3.2.7 Iterative solution algorithms ............................................................................................................... 116
3.3 Atena ............................................................................................................................................................. 119
3.3.1 Structural Elements ............................................................................................................................ 119
3.3.2 Material Modelling Formulation ........................................................................................................ 120
3.3.3 Model for Concrete Cracking ............................................................................................................. 121
3.3.4 Concrete Crushing .............................................................................................................................. 122
3.3.5 Tension Stiffening .............................................................................................................................. 123
3.3.6 Embedded Reinforcements ................................................................................................................. 124
3.3.7 Solution Procedure ............................................................................................................................. 124
3.4 Conclusions................................................................................................................................................... 126
4 Punching Shear Failure at Internal Columns ................................................................. 127
4.1 Introduction.................................................................................................................................................. 127
4.2 Evaluation of BS8110, EC2 and MC2010 design methods for punching with test data ........................ 128
4.3 Modelling of restraint from surrounding slab with MC2010 Level IV ................................................... 132
4.3.1 Test specimens.................................................................................................................................... 133
4.3.2 Sensitivity study ................................................................................................................................. 135
4.4 Assessment of Chana and Desai (1992) punching tests with membrane action ..................................... 140
4.5 Parametric studies to compare shear reinforcement requirements of BS8110, EC2 and MC2010 ...... 145
4.6 Influence of restraint from surrounding slab ............................................................................................ 150
4.7 Influence of continuity on punching resistance of nine-panel flat slab ................................................... 158
4.8 Conclusions................................................................................................................................................... 161
5 Punching Shear Failure at Edge Columns ...................................................................... 163
5.1 Introduction.................................................................................................................................................. 163
5.2 Interaction between punching resistance and unbalanced moment ........................................................ 166
5.3 Numerical investigation ............................................................................................................................... 169
5.3.1 Material modelling ............................................................................................................................. 169
5.4 Isolated slabs of El-Salakawy et al. (1998) ................................................................................................. 170
5.5 Influence of eccentricity .............................................................................................................................. 175
5.6 Investigation of shear forces on ACI 318 critical section ......................................................................... 180
5.7 Continuous specimens of Regan (1993) ...................................................................................................... 183
5.8 Influence of Continuity ................................................................................................................................ 190
5.9 Influence of reinforcement arrangement ................................................................................................... 194
14
Contents
5.10 Regan (1993) slabs with Shell elements ...................................................................................................... 197

5.11 Influence of Uncracked Column on Specimens of Regan (1993) ............................................................. 200
5.12 Strength assessment of Regan Slabs ........................................................................................................... 205
5.12.1 Sensitivity Study ................................................................................................................................. 207
5.13 Assessment of Continuity ............................................................................................................................ 208
5.14 Assessment of Perpendicular/Transverse Flexural Reinforcement ......................................................... 211
5.15 Assessment of Uncracked Column on Specimens of Regan (1993).......................................................... 216
5.16 Conclusions................................................................................................................................................... 219
6 Design Considerations ..................................................................................................... 221
6.1 General Aspects ........................................................................................................................................... 221
6.2 Numerical Modelling ................................................................................................................................... 224
6.2.1 Continuous slabs of Sherif and Dilger (2000a)................................................................................... 226
6.2.2 Influence of uncracked column on specimens of Sherif and Dilger (2000a) ...................................... 231
6.2.3 Subassembly of designed slab ............................................................................................................ 233
6.3 Influence of Reinforcement arrangement .................................................................................................. 235
6.4 Code predictions .......................................................................................................................................... 239
6.4.1 Critical direction for rotations ............................................................................................................ 239
6.4.2 Comparisons between MC2010 Levels II, III, IV and EC2 for edge columns ................................... 241
6.5 Conclusions................................................................................................................................................... 246
7 Conclusions ...................................................................................................................... 247
7.1 Recapitulation .............................................................................................................................................. 247
7.2 Conclusions from the literature review ...................................................................................................... 248
7.3 Conclusions from the finite modelling ....................................................................................................... 249
7.4 Conclusions from the Punching shear failure at Internal columns study ............................................... 250
7.5 Conclusions from the Punching shear failure at Edge columns study .................................................... 251
7.6 Remarks on Design Considerations............................................................................................................ 252
7.7 Recommendations for future research ....................................................................................................... 252
References ............................................................................................................................. 255
15
Contents
16
List of Figures
Figure 1.1 - Isolated test specimen. ............................................................................................... 38

Figure 2.1 – Assumed kinematics of the critical crack in one-way elements (adapted from Einpaul,
2016). .................................................................................................................................... 42
Figure 2.2 – (a) Aggregate interlock mechanism for cracks in normal strength concretes and (b)
push-off test arrangement. ..................................................................................................... 43
Figure 2.3 – Shear and normal forces in the crack as a function of crack width and shear slip
(Walraven and Reinhardt,1981). ............................................................................................ 44
Figure 2.4 – Dowel action of the longitudinal reinforcement (Adapted from fib, 1999). ................ 44
Figure 2.5 – Cantilever action in one-way elements....................................................................... 45
Figure 2.6 – Shear reinforcement systems. (Adapted from fib, 2001; and Einpaul et al., 2016b) .... 46
Figure 2.7 – Different layouts for shear reinforcement: (a) radial arrangement, (b) cruciform
arrangement (ACI-type) and (c) uniform arrangement (UK-type). ......................................... 47
Figure 2.8 – Potential failure modes: (a) failure within shear-reinforced area; (b) failure outside
shear-reinforced area; (c) failure close to column by crushing of concrete struts; (d)
delamination of concrete core; (e) failure between transverse reinforcement; and (f) flexural
failure. (Based on Lips et al., 2012). ...................................................................................... 48
Figure 2.9 – Test setup adopted by Regan (1986). Dimensions in mm. .......................................... 53
Figure 2.10 – Specimen tested by Sherif and Dilger (2000a). ........................................................ 54
Figure 2.11 – Choi and Kim’s (2012) test setup. ............................................................................ 55
Figure 2.12 – The failure mechanism for the plasticity approach. .................................................. 57
Figure 2.13 – Sector element on Kinnunen-Nylander’s (1960) model. ........................................... 58
Figure 2.14 - Force body diagram on sector element at Kinnunen and Nylander’s (1960) model. .. 58
Figure 2.15 - Rotation ψ and centre of rotation (CR) of the slab segment at Kinnunen and
Nylander’s (1960) model. ...................................................................................................... 59
Figure 2.16 – Specimens 11 and 12 by Bollinger (1985) with and without ring reinforcement placed
over the critical region. .......................................................................................................... 61
Figure 2.17 – Critical shear crack and the rotation ψ of the slab (based on Muttoni, 2008). ........... 61
Figure 2.18 – Elbow-shaped strut with a horizontal tie close to the column (based on Muttoni,
2008). .................................................................................................................................... 62
Figure 2.19 – Internal forces and moments over a slab sector (based on Muttoni, 2008). ............... 63
Figure 2.20 – Moment-curvature relationship assumed by Muttoni (2008). ................................... 63
17
List of Figures
Figure 2.21 – Comparison of load-rotation curves for axisymmetric tests by Kinnunen and Nylander
(1960) and CSCT equations (extracted from Muttoni, 2008). ................................................. 65
Figure 2.22 – Contribution of concrete and shear reinforcement as function of the slab rotation
(Fernández Ruiz and Muttoni, 2009). .................................................................................... 66
Figure 2.23 – Influence of an in-plane force over the CSCT punching resistance (Clément et al.,
2013). .................................................................................................................................... 69
Figure 2.24 – Redistribution of radial moments in a continuous slab due to cracking and/or
reinforcement yielding (Einpaul et al., 2015). ........................................................................ 71
Figure 2.25 – Effect of the hogging moment region dilation being constrained by the rest of the slab
(Einpaul et al., 2015). ............................................................................................................ 71
Figure 2.26 – Radius of the line of contraflexure varying with load (Einpaul et al., 2015). ............ 72
Figure 2.27 – Influence of hogging and sagging reinforcement distribution over slabs with the same
plastic moment capacity (Einpaul et al., 2015). ...................................................................... 72
Figure 2.28 – Load transfer from flat slab to column. (Broms, 2016). ............................................ 76
Figure 2.29 – Comparison of punching shear strength versus ultimate rotation between TST and
CSCT. (Broms, 2016). ........................................................................................................... 77
Figure 2.30 – Division of panels in flat slabs according to BS8110. .............................................. 79
Figure 2.31 – Control perimeter of BS8110. .................................................................................. 79
Figure 2.32 - Control perimeter at which shear reinforcement is not required at BS8110. .............. 81
Figure 2.33 – Control perimeter according to EC2......................................................................... 82
Figure 2.34 - Effective width of a flat slab on edge column. .......................................................... 82
Figure 2.35 – Shear distribution due to an unbalanced moment at a slab-internal column connection.
(BSI, 2004a). ......................................................................................................................... 83
Figure 2.36 - Interaction between punching resistance and unbalanced moment. (Regan, 1981) .... 84
Figure 2.37 - Reduced control perimeter of EC2. .......................................................................... 85
Figure 2.38 – Control perimeter where shear reinforcement is no longer required. (k=1.5). ........... 86
Figure 2.39 – ACI 318 control perimeter. ...................................................................................... 87
Figure 2.40 - Critical perimeter of ACI 318 and corresponding shear stress distribution. ............... 88
Figure 2.41 – Basic control perimeter according to MC2010 (fib, 2013). ....................................... 89
Figure 2.42 - Resultant of shear forces: (a) position with respect to the centroid of the supported
area; and (b) approximated basic control perimeter for calculation of the position of its centroid
and eccentricity between the resultant of shear forces and the centroid of the basic control
perimeter. .............................................................................................................................. 90
Figure 2.43 – Shear reinforcement arrangement on section. ........................................................... 91
18
List of Figures
Figure 2.44 – MC2010 control perimeter where shear reinforcement is no longer required. ........... 92
Figure 2.45 – Support strip widths. ................................................................................................ 93
Figure 3.1 – Degrees of freedom on a Curved shell. ...................................................................... 98
Figure 3.2 – Cauchy stresses in Diana’s curved shell. .................................................................... 98
Figure 3.3 – Positive convention for generalized Moments and Forces in Diana’s curved shell. .... 98
Figure 3.4 - Six-node triangular curved shell CT30S. .................................................................... 99
Figure 3.5 - Four-node curved shell Q20SH. ................................................................................. 99
Figure 3.6 - Eight-node curved shell CQ40S. ................................................................................ 99
Figure 3.7 - Degrees of freedom on a Brick. ................................................................................ 100
Figure 3.8 - Cauchy stresses in Diana’s solid elements. ............................................................... 100
Figure 3.9 - Eight-node isoparametric solid brick HX24L ........................................................... 101
Figure 3.10 - Twenty-node isoparametric solid brick CHX60. ..................................................... 101
Figure 3.11 – Composed solid element in Diana. ......................................................................... 102
Figure 3.12 – Eight-node quadrilateral curved base element. ....................................................... 102
Figure 3.13 – Node splitting at Discrete cracking method. ........................................................... 103
Figure 3.14 – Discrete cracking and rough crack. ........................................................................ 103
Figure 3.15 – Characteristic uniaxial stress-strain curve. ............................................................. 106
Figure 3.16 – Biaxial test results for concrete (Kupfer and Gerstle, 1973). .................................. 107
Figure 3.17 – Triaxial stress-strain relation from test of (a) Richard et al. (1928) and (b) Balmer
(1949). ................................................................................................................................ 107
Figure 3.18 – Three-dimensional representation of triaxial failure surface for concrete (Chen, 1982).
............................................................................................................................................ 108
Figure 3.19 - Thorenfeldt compression curve. ............................................................................. 109
Figure 3.20 - Parabolic compression curve in Diana. ................................................................... 110
Figure 3.21 – Stress-strain relation of plain concrete and confined concrete (Binici, 2005). ......... 110
Figure 3.22 – Influence of lateral confinement on compressive stress-strain (TNO Diana, 2014). 111
Figure 3.23 – Reduction factor due to lateral cracking according to Vecchio and Collins (1993). 111
Figure 3.24 – Linear diagram for tension softening. .................................................................... 113
Figure 3.25 - Hordijk diagram for tension softening. ................................................................... 113
Figure 3.26 – Constant shear retention for Total Strain crack models. ......................................... 114
Figure 3.27 – Shear retention factor according to different models (Sagaseta, 2008).................... 114
Figure 3.28 – Grid reinforcement in Diana. ................................................................................. 115
Figure 3.29 – Bar reinforcement in Diana. ................................................................................... 115
Figure 3.30 – Bars in curved shell elements................................................................................. 116
19
List of Figures
Figure 3.31 – Bars in Solid elements. .......................................................................................... 116

Figure 3.32 – Newton-Raphson iteration procedure. .................................................................... 117
Figure 3.33 – Quasi-newton (Secant) iteration procedure. ........................................................... 118
Figure 3.34 – Load Control incremental process.......................................................................... 119
Figure 3.35 – Eight-node/Twenty-node isoparametric solid brick. Adapted from Červenka et al.
(2016). ................................................................................................................................ 120
Figure 3.36 - Four-node/Ten-node isoparametric solid wedge. Adapted from Červenka et al. (2016).
............................................................................................................................................ 120
Figure 3.37 – Tensile softening in Atena. (Červenka et al., 2016). ............................................... 121
Figure 3.38 – Hardening/softening curve for compressive behaviour in Atena. Adapted from
Červenka et al. (2016). ........................................................................................................ 123
Figure 3.39 – Tension stiffening in Atena. (Červenka et al., 2016). ............................................. 124
Figure 3.40 – Bilinear stress-strain law for reinforcement. (Červenka et al., 2016). ..................... 124
Figure 3.41 – Arc-length method in Atena. (Červenka et al., 2016). ............................................ 125
Figure 4.1 – Influence of span on shear reinforcement area within 1.5d of the column face (Vollum,
2013). .................................................................................................................................. 128
Figure 4.2 - Influence of the shear reinforcement ratio on Vtest/Vin for BS8110, EC2 and MC2010
Level 2 with: (a) γm=γc=γs= 1.0 (5% = μ – 1.64SD); (b) code recommended partial factors γ m,
γc and γs. .............................................................................................................................. 132
Figure 4.3 – Conventional isolated test specimen adopted by Guandalini et al. (2009) and Lips et al.
(2012). ................................................................................................................................ 133
Figure 4.4 – Comparison between nodal rotations and rotations calculated from deflected shape for
slab Msup = 0.063FL and qk = 2.5 kN/m2. ............................................................................. 134
Figure 4.5 - Concrete tensile stress-strain diagram....................................................................... 135
Figure 4.6 – Mesh size comparison of specimen PL1 (Lips et al., 2012). ..................................... 136
Figure 4.7 - Mesh used for specimen PV1 (Lips et al., 2012). ...................................................... 136
Figure 4.8 – Model with axial restraint applied at both column and slab. ..................................... 137
Figure 4.9 – Model with axial restraint applied only at column. ................................................... 137
Figure 4.10 – Model without column and axial restraint. ............................................................. 138
Figure 4.11 – Comparison of a NLFEA of a flat slab with different boundary conditions simulating
the quarter symmetry: (a) specimen PG10 (Guandalini et al., 2009), and (b) Msup=0.063FL and
qk=2.5 kN/m2 from section 4.6. ........................................................................................... 138
Figure 4.12 - Comparison of measured and calculated load versus rotation ψ in radians for slabs
PL1, PV1, PL4, PL5, PG10 and PG11. ................................................................................ 139
20
List of Figures
Figure 4.13 – (a) Deflected shape and (b) Rotations of slab PL1 (Lips et al., 2012) with different
shear retention factors.......................................................................................................... 140
Figure 4.14 – Schematic of Chana and Desai (1992a) test arrangement. (Dimensions in mm). .... 141
Figure 4.15 – Reinforcement details for Chana and Desai (1992a) test. ....................................... 142
Figure 4.16 – Isolated specimen of Chana and Desai (1992b) compared to full panel of Chana and
Desai (1992a). ..................................................................................................................... 142
Figure 4.17 – Comparison of deflections between a full panel and an isolated specimen (Chana and
Desai, 1992a). ..................................................................................................................... 142
Figure 4.18 – Mesh for the NLFEA of the (a) conventional Chana and Desai’s (1992b) specimen,
and (b) continuous Chana and Desai’s (1992a) specimen. .................................................... 143
Figure 4.19 - Analysis of Chana and Desai’s (1992a) slabs FPS1 (no stirrups) and FPS5 (with
stirrups): (a) comparison of measured and predicted deflections; (b) calculation of resistance
with MC2010. ..................................................................................................................... 144
Figure 4.20 – Span/thickness chart for flat slabs extracted from Goodchild et al. (2009).............. 146
Figure 4.21 - Influence of design imposed load on VEd/Vmax for design support moments of 0.063FL
and 0.083FL. ....................................................................................................................... 147
Figure 4.22 - Comparison of (a) areas of shear reinforcement required within 1.5d of column, and
(b) required minimum distance to outer shear reinforcement for design support moments of
0.063FL and 0.083FL. ......................................................................................................... 149
Figure 4.23 - Influence of surplus flexural reinforcement of a 265 mm thick slab with 450 mm
square column on: (a) maximum possible shear resistance; (b) area of shear reinforcement for
F=1039 kN. ......................................................................................................................... 150
Figure 4.24 - Boundary conditions for MC2010 Level IV analysis of the interior panels of a flat
slab...................................................................................................................................... 151
Figure 4.25 – Models for the continuous slabs in the parametric studies: (a) rotational fixity and (b)
rotational plus in-plane fixities. ........................................................................................... 152
Figure 4.26 - Variation in the nodal rotation along the slab centreline for M=0.063FL and qk=2.5
kN/m2. ................................................................................................................................. 153
Figure 4.27 - Influence of slab continuity on shear resistance: (a) Msup=0.063FL and qk=2.5 kN/m2;
(b) Msup=0.063FL and qk=5.0 kN/m2; (c) Msup=0.063FL and qk=7.5 kN/m2; (d) Msup=0.063FL
and qk=10 kN/m2 ................................................................................................................. 154
Figure 4.28 - Influence of slab continuity on shear resistance: (a) Msup=0.083FL and qk=2.5 kN/m2;
(b) Msup=0.083FL and qk=5.0 kN/m2; (c) Msup=0.083FL and qk=7.5 kN/m2; (d) Msup=0.083FL
and qk=10 kN/m2 ................................................................................................................. 155
21
List of Figures
Figure 4.29 – Layout of test structure by Guralnick and Fraugh (1963). Dimensions in mm. ....... 159
Figure 4.30 – Comparison between rotations extracted from shell elements and calculated with
Equation (4.1). .................................................................................................................... 161
Figure 4.31 - Calculation of MC2010 shear resistance of the slab reported by Guralnick and Fraugh
(1963). ................................................................................................................................ 161
Figure 5.1 – Geometry and loading arrangement of (a) isolated slabs of El-Salakawy et al. (1998)
and (b) continuous slabs of Regan (1993) (dimensions: mm). .............................................. 164
Figure 5.2 – Punching and flexure interaction graph for flat slabs on a) internal and edge
connections with unbalanced moment axis perpendicular to the slab edge, and b) edge column
with ‘normal’ moment (Stamenković and Chapman, 1974). ................................................ 165
Figure 5.3 - Interaction between punching resistance and unbalanced moment at column face for
specimens in Table 5.1 with Mflex calculated using effective width of c2 + 2y with Mflex≤ Mtmax.
............................................................................................................................................ 167
Figure 5.4 – Test arrangement of El-Salakawy et al. (1998) (Dimensions in mm). ....................... 170
Figure 5.5 – Reinforcement arrangement for El-Salakawy et al. (1998) specimens: (a) tension
reinforcement and (b) compression reinforcement. .............................................................. 171
Figure 5.6 - Finite-element mesh used for analysis of slabs of El-Salakawy et al. (1998) with solid
elements. ............................................................................................................................. 172
Figure 5.7 – Comparison between different crack models for XXX (El-Salakawy et al., 1998). ... 173
Figure 5.8 – Comparison between different compressive curves for XXX (El-Salakawy et al.,
1998). .................................................................................................................................. 173
Figure 5.9 – Deflected shape and column stub displacement of XXX (El-Salakawy et al., 1998) with
different shear retention factors. .......................................................................................... 174
Figure 5.10 – Comparison between iterative procedures for XXX (El-Salakawy et al., 1998). ..... 175
Figure 5.11 - Comparison of measured and predicted deflections for slab XXX (El-Salakawy et al.,
1998) for: (a) the longitudinal direction at failure load, (b) the transverse direction at failure
load, and (c) the column stub for varies loads. ..................................................................... 176
Figure 5.12 - Comparison of measured and predicted deflections for slab HXXX (El-Salakawy et
al., 1998) for: (a) the longitudinal direction at failure load, and (b) the column stub for varies
loads.................................................................................................................................... 176
Figure 5.13 – Comparison between NLFEA longitudinal and transverse slab rotations relative to the
column for El-Salakawy et al. (1998) specimens. ................................................................ 177
Figure 5.14 - Assessment of slabs of El-Salakawy et al. (1998) with MC2010 for eccentricities of:
(a) 0.2 m; (b) 0.3 m (XXX); (c) 0.4 m; (d) 0.5 m; (e) 0.6 m; and (f) 0.66 m (HXXX). .......... 179
22
List of Figures
Figure 5.15 - Influence of eccentricity on calculated punching resistance of slabs of El-Salakawy et

al. (1998). ............................................................................................................................ 180
Figure 5.16 - Analysis of shear force distribution on ACI 318 critical perimeter for slabs of El-
Salakawy et al. (1998): (a) shear force distribution for XXX; (b) shear force distribution for
HXXX. ................................................................................................................................ 181
Figure 5.17 - Influence of eccentricity on γv on ACI 318 critical perimeter for XXX slabs of El-
Salakawy et al. (1998). ........................................................................................................ 182
Figure 5.18 - Reinforcement details for Regan slabs (dimensions in mm) .................................... 185
Figure 5.19 – Plan view of the reinforcement distribution for Regan slabs (dimensions in mm). .. 185
Figure 5.20 - Finite-element mesh used for slabs of Regan (1993). .............................................. 187
Figure 5.21 – Comparison between different smeared crack approaches for Slab 1 End 1 (Regan,
1993). .................................................................................................................................. 188
Figure 5.22 – Comparison between different compressive curves for Slab 1 End 1 (Regan, 1993).
............................................................................................................................................ 189
Figure 5.23 - Deflected shape and rotations of Slab 1 End 1 (Regan, 1993) with different shear
retention factors................................................................................................................... 189
Figure 5.24 – Comparison between different iterative solution algorithms for Slab 1 End 1 (Regan,
1993). .................................................................................................................................. 190
Figure 5.25 – FE rotations calculated from the deflected shape for the specimens by Regan (1993).
............................................................................................................................................ 191
Figure 5.26 - Comparison between NLFEA longitudinal and transverse slab rotations relative to the
column for Regan (1993) specimens. ................................................................................... 192
Figure 5.27 - Assessment of slabs of Regan (1993) with MC2010 for: (a) slab 1 end 1; (b) slab 2
end 1; (c) slab 2 end 2; (d) slab 4 end 1; (e) slab 4 end 2 ...................................................... 193
Figure 5.28 - Influence of varying reinforcement arrangement on rotation of Regan (1993) Slab 1
End 1................................................................................................................................... 195
Figure 5.29 - Influence of varying reinforcement arrangement on punching resistance of Regan
(1993) Slab 1 End 1. ............................................................................................................ 195
Figure 5.30 - Influence of varying reinforcement arrangement on eccentricity (M/V) of (a) Regan
(1993) Slab 1 End 1 with span = 0.25%, (b) Regan (1993) Slab 1 End 1 with span = 0.5%, (c)
Regan (1993) Slab 1 End 1 with span = 1.0% and (d) Regan (1993) specimens (measured). 196
Figure 5.31 – Mesh used for the validation. ................................................................................. 197
Figure 5.32 - Deflected shape and rotations of Slab 1 End 1 (Regan, 1993) modelled with shell
elements and different shear retention factors. ..................................................................... 198
23
List of Figures
Figure 5.33 - Comparison of measured and NLFEA load versus rotation for slabs of Regan (1993).
............................................................................................................................................ 199
Figure 5.34 - Comparison of measured and FE column reaction versus rotation for slabs of Regan
(1993) with linear columns. ................................................................................................. 200
Figure 5.35 - Comparison of measured and FE deflected shape for slabs of Regan (1993) with linear
columns. .............................................................................................................................. 201
Figure 5.36 – Influence of modelling assumption in the eccentricity of Regan’s (1993) specimens.
............................................................................................................................................ 202
Figure 5.37 – Influence of column modelling assumption with Diana.......................................... 204
Figure 5.38 – Edge column reaction to total load ratio for Regan’s (1993) specimens. ................ 205
Figure 5.39 - Finite element mesh of Regan (1993) slabs used in the Atena analysis. .................. 206
1993). .................................................................................................................................. 207
Figure 5.41 - Comparison between different smeared crack approaches for Slab 1 End 1 (Regan,
1993). .................................................................................................................................. 208
Figure 5.42 – Load-rotation relationship for Regan slabs a) S1 end 1, b) S4 end 1 and c) S4 end 2.
............................................................................................................................................ 210
Figure 5.43 - Influence of varying longitudinal reinforcement arrangement on rotation of Regan
(1993) Slab 1 End 1. ............................................................................................................ 212
Figure 5.44 - Load x Longitudinal Rotation for specimens with transverse reinforcement ratio
varying between 0.5% and 1.0%. ......................................................................................... 214
Figure 5.45 - Influence of varying reinforcement arrangement on eccentricity a) M/V span = 0.25%,
b) M/V span = 0.5% and c) M/V span = 1.0%. ..................................................................... 215
Figure 5.46 – Influence of column modelling assumption on failure load. ................................... 217
Figure 5.47 – Influence of column modelling assumption on eccentricity M/V............................ 217
Figure 5.48 – Influence of column modelling assumption on punching resistance. ...................... 218
Figure 6.1 – Plan view of the considered flat slab (Dimensions in mm). ...................................... 222
Figure 6.2 – Flexural reinforcement for the analysed Flat Slab. ................................................... 224
Figure 6.3 – ¼ of the designed flat slab. ...................................................................................... 225
Figure 6.4 – Specimen tested by Sherif and Dilger (2000a). ........................................................ 225
Figure 6.5 – Flexural reinforcement for the specimens tested by Sherif and Dilger (2000a). ........ 226
Figure 6.6 – Shear reinforcement for slab S2 of Sherif and Dilger (2000a). ................................. 227
24
List of Figures
Figure 6.7 – Comparison between longitudinal deflected shape of slab ρsup=0.6% ρspan=1.2% at
design load with different convergence tolerances. .............................................................. 227
Figure 6.8 – Mesh used for the NLFEA of Sherif and Dilger’s (2000a) tests. .............................. 228
Figure 6.9 – Comparisons between NLFEA results and experimental data from Sherif’s (1996) slab
(a) S1 and (b) S2. ................................................................................................................ 229
Figure 6.10 –NLFEA reinforcements strains of Slab S1 at (a) face of the edge column (Hogging),
(b) face of the internal column (Hogging), and (c) midspan (Sagging). ................................ 230
Figure 6.11 – Column reaction-rotation relative to the column for the slabs of Sherif and Dilger
(2000a) and resistance according to the CSCT. .................................................................... 230
Figure 6.12 – Longitudinal ( = − ) and Transverse ( = , ℎ =
, ℎ ) rotations used in the CSCT analysis. ............................................................. 231
Figure 6.13 – Influence of uncracked column on deflections for slabs (a) S1 and (b) S2 of Sherif
and Dilger (2000a). ............................................................................................................. 232
Figure 6.14 – Influence of uncracked column on eccentricity at edge columns of slabs (a) S1 and (b)
S2 of Sherif and Dilger (2000a). .......................................................................................... 233
Figure 6.15 – Subassembly of the designed flat slab. ................................................................... 234
Figure 6.16 – Deflected shape comparison between the ¼ model and the Subassembly at the design
ULS load. ............................................................................................................................ 234
Figure 6.17 – Bending moment comparison between the ¼ model and the Subassembly at the
design ULS load. ................................................................................................................. 234
Figure 6.18 – Longitudinal reinforcement designed for the subassembly. .................................... 235
Figure 6.19 – Transverse reinforcement designed for the subassembly. ....................................... 236
Figure 6.20 – Longitudinal rotation of the subassembly’s edge column from the parametric study.
............................................................................................................................................ 237
Figure 6.21 – Influence of varying reinforcement arrangement in the eccentricity of the
subassembly’s edge column................................................................................................. 238
Figure 6.22- Comparison between NLFEA reactions and LFEA. ................................................. 238
Figure 6.23 – Comparison between longitudinal and transverse rotations with and without cladding
load for slab ρsup = 0.8% ρspan = 0.6%. ................................................................................. 239
Figure 6.24 – Comparison between longitudinal and transverse rotations for the edge column of the
subassembly parametric study. ............................................................................................ 240
Figure 6.25 – Transverse rotations of the subassembly’s edge column from the parametric study.241
Figure 6.26 – Critical mechanisms considered for the subassembly’s parametric studies. ............ 242
25
List of Figures
Figure 6.27 – Longitudinal and Transverse rotations of the subassembly slabs from the parametric
study. .................................................................................................................................. 243
Figure 6.28 – Shear reinforcement in 1.5d from the column required by MC2010 and EC2. ........ 245
26
List of Tables
Table 2.1 – Simplified apportionment of bending moment for a flat slab. (BSI, 2004a) ................. 81
Table 2.2 – Values of for rectangular loaded areas. (BSI, 2004a) ............................................... 83
Table 4.1 – Details of the slabs in Figure 4.2. .............................................................................. 130
Table 4.2 – Properties of the slabs used in the calibration of the NLFEA ..................................... 134
Table 4.3 - Properties of slabs used in the parametric studies....................................................... 146
Table 4.4 - Influence of continuity on shear strengths calculated to MC2010 with = = 1 and
= 0.9.............................................................................................................................. 157
Table 5.1 - Database of tests. Edge column tests without shear reinforcement. ............................ 168
Table 5.2 – Details of Regan Slabs .............................................................................................. 186
Table 5.3 – Failure loads from different column modelling assumptions and . (Loads in kN). . 204
Table 5.4 - Comparison of measured and predicted failure loads of Regan slabs.......................... 210
Table 5.5 - Comparison of calculated punching resistance for slabs in parametric study. ............. 213
Table 5.6 – Failure loads from different column modelling assumptions and . (Loads in kN). . 219
Table 6.1 – Details of specimens tested by Sherif and Dilger (2000a). ......................................... 227
Table 6.2 – Details on the parametric study varying the reinforcement arrangement. ................... 237
Table 6.3 – , calculated with = 0.7 with and without cladding load. ........................... 240
Table 6.4 - , and 1.5dvAsw from MC2010 LoA II to IV, EC2, and Vflex. Shear forces in kN.
............................................................................................................................................ 244
27
List of Tables
28
Notations
A area of flexural reinforcement provided
A area of flexural reinforcement required
A area of reinforcement
A area of tensile reinforcement
A area of shear reinforcement in each perimeter
b , control perimeter beyond the outer layer of shear reinforcement
b punching control perimeter
dimensions of critical section b0 measured perpendicular and parallel to slab

b and b
edge
b effective width of flexural reinforcement for normal moment at edge column
b width of support strip
diameter of a circle with the same surface area as enclosed by the control
b
perimeter
c and c column side perpendicular and parallel to the slab edge
d mean aggregate size
d shear reinforcement diameter
d maximum aggregate size
d reference size for aggregate
d effective depth of top reinforcement normal to the slab edge
d reduced effective depth of slab
eccentricity of the shear force with respect to the centroid of basic control
e
perimeter
29
Notations
E concrete modulus of elasticity
EI plate flexural stiffness before cracking
EI plate tangential flexural stiffness after cracking
E modulus of elasticity of reinforcement
f bond strength
f compressive cylinder strength of concrete
f maximum principal stress
f characteristic compressive cylinder strength of concrete at 28 days
f tensile strength of concrete
f characteristic cube strength of concrete
f peak compressive force
f tensile strength
f yield strength of reinforcement
f design yield strength of reinforcement
f characteristic yield strength of reinforcement
f yield strength of the shear reinforcement
f , effective design strength of the shear reinforcement
f design yield strength of the shear reinforcement
G compressive fracture energy
G tensile fracture energy
g out-of-balance force vector
J polar moment of inertia of critical section
k effectiveness coefficient dependent on maximum aggregate size
k effectiveness coefficient for eccentric shear
30
Notations
k efficiency factor for punching shear reinforcement system
k coefficient relating shear resistance to slab rotation
m tangential moment at the slab edge
M bending moment across panel width at inner column face
M bending moment about centroid of critical section
m average bending moment per unit width of the support strip
M design unbalanced moment
M bending moment portion of unbalanced moment
M moment of resistance of slab column connection at column face
m bending moment (i = x, y, z) and (j = x, y, z)
m radial moment per unit width
m nominal moment capacity per unit width
m design average flexural strength per unit width of the support strip
m average bending moment per unit width
M moment at the support
M torsion portion of unbalanced moment
M experimental ultimate bending moment about column centreline
M maximum unbalanced moment due to flexure
M ultimate resisting moment when the shear force is zero
n normal forces (i = x, y, z) and (j = x, y, z)
q shear forces (i = x, y, z) and (j = x, y, z)
q characteristic design imposed load
r radius of the zone in which cracking is stabilized
r radius of a circular column (2c/π for square columns)
r radius of cracked zone
31
Notations
r radius of the circular area inside the shell
r radius of the critical shear crack
r radius of load introduction
r radius of an isolated slab element
r radius for calculating the slab rotation
s radial spacing to the first perimeter of shear reinforcement from column face
s radial spacing of shear reinforcement
u basic control perimeter
u displacements
control perimeter where shear reinforcement is no longer needed for cross-type

u ,
arrangement in EC2
u length of first control perimeter at which shear reinforcement is not required
V punching resistance provided by concrete
V concrete punching shear resistance calculated according to EC2
V design shear force
V column load corresponding to flexural failure at inner column face
V ultimate shear force according to NLFEA
V , shear resistance of slab without shear reinforcement
V , combined shear resistance of concrete and shear reinforcement
V , maximum possible shear resistance
V , shear resistance outside shear reinforced zone
V , shear resistance provided by shear reinforcement
v , design punching shear stress provided by concrete
V punching resistance in absence of unbalanced moment
V shear resistance provided by shear reinforcement
32
Notations
V calculated punching resistance
V measured failure load
V Ultimate shear load when the moment is zero
V punching failure according to the TST
V punching failure according to the TST
W shear distribution as a function of the basic control perimeter
α peak compressive strain
αx height of equivalent rectangular stress block
γ partial factor for concrete
γ proportion of unbalanced moment transmitted by flexure
γ Cauchy strain (i = x, y, z) and (j = x, y, z)
γ BS8110 partial safety factor for shear
γ partial factor for reinforcement
γ proportion of unbalanced moment transmitted by uneven shear
δ iterative displacement increment
ε limiting tangential compression strain according to the TST
ε concrete elasto-plastic strain
ε Cauchy strains (i = x, y, z) and (j = x, y, z)
ε reinforcement strain
ε reinforcement yield strain
ε reinforcement yield strain
ρ hogging flexural reinforcement ratio
ρ sagging flexural reinforcement ratio
ρ longitudinal flexural reinforcement ratio in the span
33
Notations
ρ longitudinal flexural reinforcement ratio in the support
ρ and ρ flexural tension reinforcement ratio in x and y directions
σ Cauchy stresses (i = x, y, z) and (j = x, y, z)
σ reinforcement stress
σ shear reinforcement stress
τ average bond stress
χ decrease in curvature due to tension stiffening
average shear force per unit length over the control perimeter
punching shear stress provided by concrete
peak shear force per unit length over the control perimeter
column rotation corresponded to the longitudinal slab deflected shape
longitudinal slab rotation
longitudinal slab rotation relative to the column
column rotation corresponded to the transverse slab deflected shape
transverse slab rotation
, left-hand side transverse slab rotation
, right-hand side transverse slab rotation
transverse slab rotation relative to the column
slab rotation at d from the column face
column rotation
slab rotation
diameter of shear reinforcement
B column diameter
c side length of a square column
C diameter of the slab
34
Notations
d average effective depth of the slab
e support eccentricity with respect to the column axis
E Longitudinal modulus of elasticity
epar eccentricity parallel to the slab edge
F total load applied to each panel of a flat slab
G Transverse modulus of elasticity
h slab thickness
k size effect factor
Ki stiffness matrix
L span between column centrelines
Lx and Ly span over the x and y direction
M applied bending moment about column centreline
n,s,t crack coordinates
p squeezing pressure caused by the inclined strut according to the TST
r,s,t local coordinates at a Finite Element
s shear slip
S1 concrete stiffness within the column
S2 concrete stiffness outside the column
T transformation matrix
u length of basic control perimeter
u1 * length of reduced EC2 control perimeter
V shear force at the column
y perpendicular distance from slab edge to inner column face
y depth to the neutral axis
35
Notations
α MC2010 factor for rotations
β coefficient to account for uneven shear
ΔT bond force between concrete and reinforcement
Δu total displacement increment
Δε strain increment
ε longitudinal strain evaluated in the critical region
ζ, ξ, η local coordinates at a Finite Element
κ correction factor to account for dowel action.
ρ flexural tension reinforcement ratio
σ normal stress
τ shear stress
φ creep factor
design shear stress
crack width
constant shear retention factor
slab rotation for a continuous flat slab
rotation of slab outside critical shear crack
36
1 Introduction
1.1 GENERAL ASPECTS
This research addresses the design of reinforced concrete flat slabs for punching shear. Flat slabs
are two way spanning slabs without downstand beams that are directly supported by columns. They
were initially designed in the Europe and USA at the beginning of the 20th century. Currently, flat
slabs are the most common form of concrete frame construction in the UK as they have a great
number of advantages including ease of construction, saving in direct and indirect material costs of
the frame, reductions in floor-to-floor heights etc.
Prior to the widespread use of punching shear reinforcement, column cross sections were locally
increased with large capitals in order to reduce punching shear stresses to acceptable levels. Later,
with the advent of punching shear reinforcement, flat slabs without drop panels or column capital
became more usual. The thickness of flat slabs is usually governed by either the serviceability limit
state of deflection control or the ultimate limit state of punching shear.
Punching shear like deflection control is still a subject of active research mainly because there is
lack of agreement on the failure mechanism. There are also significant differences between the
available empirical treatments as discussed by Alexander (2005) amongst others.
According to fib (2001), punching is a shear failure within the discontinuity region of the highly
stressed slab at the column. The punching resistance of flat slabs is usually determined from tests on
isolated specimens like that shown in Figure 1.1, which are unrepresentative of flat slab floor plates
due to the absence of flexural continuity and compressive membrane action.
Punching shear failure is brittle and is characterised by the formation of a conical failure surface.
The angle of failure cone varies dependent on the amount of reinforcement present. In the case of
shear reinforced slabs, failure can occur either inside or outside the shear reinforcement zone. Shear
reinforcement significantly improves both the punching resistance and ductility of flat slabs. For a
given slab thickness, adding shear reinforcement will allow greater live loads and also increase
deflections developed prior to failure.
37
Introduction
Figure 1.1 - Isolated test specimen.

A variety of reinforcing systems are used to increase punching shear capacity including double
headed shear studs (deformed, smooth, prestressed), steel offcuts, stirrups and post-installed bonded
bars. Although all these options have shown good results, the most widely used shear
reinforcements are double headed studs and stirrups.
Recently, EC2 (BSI, 2004a) has replaced the previous UK code BS 8110:1997 (BSI, 1997) for the
design of concrete structures. While there are similarities in terms of general approach, there are
also considerable differences in assumptions and the resulting design equations, raising questions
about their reliability. Collins et al. (2008) illustrated the inconsistency of some beam shear design
provisions, including EC2, ACI 318-08 (ACI, 2008) and BS8110, by comparing their predictions
against experimental results of 4 slab-strip specimens. They showed the highest calculated shear
stress to be 2.56 times the lowest one, with some of the predictions unsafe.
The most recent European thinking on punching shear is reflected in the fib Model Code 2010
(MC2010) (fib, 2013) which adopts the critical shear crack theory (CSCT) of Muttoni (2008).
MC2010 provides four Levels of Approximation (LoA) of which LoA I to III are intended for
design. LoA IV is intended for assessment. MC2010 is likely to provide the basis for future
revisions to EC2.
Vollum (2013) has shown that the superseded UK code BS8110 requires less shear reinforcement
than both EC2 and MC2010 LoA I to III, but there is no evidence that slab-column connections
designed to BS8110 are unsafe. Furthermore, the MC2010 design recommendations (i.e. LoA I to
38
Introduction
III) are more onerous for punching than EC2 (BSI, 2004a). A key aim of the present research is to
investigate why the level of safety provided by the BS8110 punching rules has been sufficient in
practice. This is done by exploring the benefits of slab continuity on punching resistance.
1.2 AIMS AND OBJECTIVES
This research seeks to develop a better understanding of the effects of in-plane restraint and flexural
continuity on the punching resistance of flat slabs. The overriding aim is to determine whether
flexural continuity is sufficient to explain the satisfactory performance of existing flat slabs which
appear punching shear deficient when assessed against test data from isolated punching specimens
(see Figure 1.1). To this end, specific objectives in descending order of importance are to:
1. Use NLFEA, test data and MC2010 LoA IV to quantify the effect of rotational and in-plane
restraint on the punching resistance of internal and edge slab-column connections.
2. Carry out parametric studies to evaluate the influence of slab continuity on the punching
resistance of full scale structures. The studies will focus on the redistribution of flexural
reinforcement between the span and support.
3. Assess whether the effects of combined rotational and in-plane restraint are sufficient to explain
the satisfactory performance of internal slab column connections designed to BS8110 which appear
shear deficient when assessed with EC2 and MC2010 LoA II.
4. Assess whether punching resistance is related to the slab rotation at the column as assumed in the
Critical Shear Crack Theory (CSCT) of Muttoni (2008), which underpins the design method for
punching shear in MC2010 (fib, 2013).
5. Use nonlinear finite element analysis to simulate and evaluate the behaviour observed in
structural tests by others.
6. Carry out parametric studies to compare punching shear requirements in BS8110 (BSI, 1997),
EC2 (BSI, 2004a), and MC2010 (fib, 2013).
1.3 OUTLINE OF THIS THESIS
This thesis consists of seven chapters, which are structured as follows:

 Chapter 1 introduces the research by presenting a general overview of the problem and
describes the research objectives.
 Chapter 2 presents a literature review of punching shear in flat slabs, providing enough
background for the analyses that follow. General aspects that influence the punching shear
resistance such as shear transfer mechanisms, shear and flexural reinforcements, size effect,
39
Introduction
failure modes, and the effects of continuity are discussed. A few mechanical models
developed to describe the phenomenon are presented and reviewed, with emphasis on the
Critical Shear Crack Theory (CSCT) of Muttoni (2008), as it is applied throughout the
thesis. Design recommendations related to punching shear requirements by BS8110, EC2,
ACI 318-14 and MC2010 are presented and discussed.
 Chapter 3 describes and details the structural elements, material modelling assumptions and
solution procedure used for the Finite Element Analysis in this research, which included two
commercially available codes: Diana v9.6 (TNO Diana, 2014) and Atena v5.1.2 (Červenka
et al., 2016).
 Chapter 4 deals with punching shear failure of flat slabs supported on internal columns. The
chapter is based on the published paper Soares and Vollum (2015). Evaluation of test data
against punching shear requirements of BS8110, EC2 and MC2010; assessment of
continuous punching tests with NLFEA and the CSCT; parametric studies comparing shear
reinforcement requirements of BS8110, EC2 and MC2010; as well as NLFEA using
MC2010 Level IV to assess the influence of continuity are presented and discussed.
 Chapter 5 considers punching shear failure at edge columns of reinforced concrete flat slabs
without shear reinforcement. The chapter is based on the published papers Soares and
Vollum (2016a), and Soares and Vollum (2016b). The interaction between punching
resistance and unbalanced moment is assessed with test data and NLFEA with MC2010
Level IV. ACI 318 recommendations for the proportion of unbalanced moment resisted by
eccentric shear are evaluated. The influence of slab continuity and reinforcement
arrangement is investigated with calibrated FE models using both Diana and Atena.
 Chapter 6 is concerned with the assessment of a typical full scale flat slab floor plate. The
influence of column modelling on punching resistance according to MC2010 is investigated.
The effect on slab rotation, and hence punching resistance, of redistributing reinforcement
between the span and support is assessed.
 Chapter 7 summarizes the key findings of the research and demonstrates that the research
objectives are satisfied.
40
2 Background and Literature Review
2.1 GENERAL ASPECTS
Punching resistance remains a subject of active research mainly because of the lack of a widely
accepted theory to describe the phenomenon. The structural behaviour of reinforced concrete
members in shear is much more complex than in flexure. Design recommendations in codes of
practice are largely empirical and based on results from isolated experimental specimens like that
shown in Figure 1.1. The earliest rational theory for punching shear was developed by Kinnunen
and Nylander (1960) who proposed that the punching resistance is reached for a given critical
rotation ψ.
Other mechanical models of similar relevance are Moe (1961), Bræstrup/Nielsen et al. (1976),
Broms (1990), Hallgren (1996), and Broms (2016). Models of the Kinnunen and Nylander (1960)
type give good results but are overly complex for code implementation. More recently, Muttoni
(2008) proposed the Critical Shear Crack Theory (CSCT) as a new failure criterion for punching
shear, which is consistent with the model of Kinnunen and Nylander (1960). The CSCT forms the
basis of the design recommendations for punching in the most recent Model Code 2010 (MC2010)
(fib, 2013). Some of these mechanical approaches are discussed in more detail in section 2.3.
2.1.1 Shear transfer actions
Understanding the mechanisms of shear resistance is very important to an understanding of

punching failure since these so-called ‘shear transfer actions’ are directly related to the causes of
failure and possible failure criteria. fib (2001) describes them according to their contributions in the
tension and compression zone. In the cracked tension zone, they can be distinguished as:
i. Cantilever action of concrete teeth;
ii. Friction stresses along cracks;
iii. Dowel action of tension chord;
iv. Residual tensile stresses across cracks;
whereas for the compression zone, fib (2001) list them as:
v. Shear stresses in compression zone, and;
vi. Inclined compression chord.
In the case of beams, crack kinematics and the contributions to shear resistance of the various shear
transfer actions for reinforced concrete beams have been studied by many authors including
41
Background and Literature Review
Walraven and Reinhardt (1981), Gambarova and Karakoç (1983), Millard and Johnson (1985),
Zararis (1997), Sagaseta and Vollum (2011), Mihaylov et al. (2013), Campana et al. (2013) and
Fernández Ruiz et al. (2015). The individual contributions of the shear resisting mechanisms actions
and their relationship to the crack opening and width can be quantified through stress-displacement
laws. However, Campana et al. (2013) showed that different crack patterns and associated
kinematics at failure may develop in similar reinforced concrete members. Despite that, the sum of
the various shear transfer actions may still be similar. Muttoni and Fernández Ruiz (2008) observed
that shear is initially resisted by three shear-carrying mechanisms: aggregate interlock, dowel action
and cantilever action as shown in Figure 2.1. As the shear force increases the development of tensile
stresses leads to the formation of the critical shear crack. The following sections describe these
mechanisms in more detail.
Figure 2.1 – Assumed kinematics of the critical crack in one-way elements (adapted from Einpaul,
2016).
2.1.1.1 Aggregate Interlock
The component of shear resisted by the transmission of shear forces across cracks in concrete is
commonly known as Aggregate Interlock, which is illustrated in Figure 2.2a. Push-off tests such as
that presented in Figure 2.2b are commonly used to assess the amount of shear transferred through
this action. Several studies have been carried out to quantify this phenomenon. Empirically based
formulations were proposed by Hamadi and Regan (1980), Walraven and Reinhardt (1981), Bazant
and Gambarova (1980), and Gambarova and Karakoç (1983), whereas Walraven (1980) and Li et
al. (1989) amongst others developed analytical rational formulations.
42
(a) (b)
Figure 2.2 – (a) Aggregate interlock mechanism for cracks in normal strength concretes and (b)
push-off test arrangement.
One of the most widely used models for aggregate interlock is the linear aggregate interlock by
Walraven and Reinhardt (1981). Based on observations of experimental data and how parameters
such as crack width, shear and normal stresses, and shear displacement relate, Walraven and
Reinhardt (1981) proposed the curves described by Equation (2.1) on the basis of best fit.
. .
=− + 1.8 + (0.234 − 0.20) s ( 0)
30
(2.1)
. .
=− + 1.35 + (0.191 − 0.15) s (σ 0)
20
where is the shear stress, is the cube concrete compressive strength, is the crack width, and
s is the shear slip. The relation described in (2.1) is shown in Figure 2.3 for a concrete with =
25 and maximum aggregate size of 16 mm. The dotted lines depicts Equation (2.1) for
different crack widths, whereas the solid lines are calculated with the commonly used contact
density model of Walraven and Reinhardt (1981).
43
Figure 2.3 – Shear and normal forces in the crack as a function of crack width and shear slip
(Walraven and Reinhardt,1981).
2.1.1.2 Dowel Action

Dowel action is the component of shear resisted by the reinforcement bar at a crack, Figure 2.4. It
can consist of bending, shear and kinking, although the last two ones are usually neglected. fib
(1999) describes the relationship between dowel forces and shear displacements as nearly linear
until 40% of the failure load. Beyond that, the behaviour gets nonlinear and the inclination is
reduced as the concrete under the bar is progressively damaged. Unless some type of transverse
reinforcement is provided, the dowel action contribution in the shear transfer ceases when the
concrete splits along the longitudinal reinforcement.
Figure 2.4 – Dowel action of the longitudinal reinforcement (Adapted from fib, 1999).
44
2.1.1.3 Cantilever action

Kani (1964) observed that shear can be transferred by the concrete in between two flexural cracks
acting as cantilever beams (concrete “tooth”) subject to the bond forces between concrete and
reinforcement, Δ , as shown in Figure 2.5. Around the crack, the shear is carried by the inclined
compression chord. Fernández Ruiz et al. (2015) states that this action is limited by the formation of
a vertical flexural crack into a quasi-horizontal crack, which disables the capacity of the tension tie
of the tooth.
Figure 2.5 – Cantilever action in one-way elements.

Muttoni and Fernández Ruiz (2008) comment that the formation of the critical shear crack does not
necessarily imply the collapse of the structure as new shear-carrying mechanisms are developed
during its propagation such as an elbow-shaped strut. More details are presented and discussed in
section 2.3.3.
2.1.2 Shear Reinforcement
The punching strength of flat slabs can be enhanced by enlarging the column section, increasing the
effective depth of the slab, providing surplus flexural reinforcement, increasing the concrete
compressive strength, or using shear reinforcement. Among these options, providing effective shear
reinforcement is usually the most economic solution. Shear reinforcement can intercept shear cracks
and delay or even prevent their widening, as well as providing ductility to the flat slab. Good
anchorage is crucial for the performance of shear reinforcement, although in some cases it can lead
to a time consuming installation process. There are currently several different types of shear
reinforcements in common use. Figure 2.6 shows a few available and already tested options, which
includes stirrups, bent-up bars, hooks, shear ladders, stud-rails, double headed studs, I-beam section
cuts, and hoop reinforcement. It is worth mentioning that research is still been carried out into the
development of new types of shear reinforcement systems.
45
Figure 2.6 – Shear reinforcement systems. (Adapted from fib, 2001; and Einpaul et al., 2016b)
Numerous studies on flat slabs with punching shear reinforcement have proved already the benefits
of using transverse reinforcing systems (Graf, 1938; Elstner and Hognestad, 1956; Moe, 1961;
Andersson, 1963; Narasimham, 1971; Seible et al., 1980; Broms, 1990; Regan, 1990; Oliveira and
Melo, 1999; Fernández Ruiz and Muttoni, 2009; Guandalini et al., 2009; Lips et al., 2012; Einpaul
et al., 2016b, amongst others). The improvement on deformation capacity can be three times
greater, while the strength almost double as a direct result of transverse reinforcement in
46
comparison to specimens with no punching shear reinforcing system. fib (2001) provides a vast
detailed databank on punching of shear-reinforced flat slabs.
Typical examples of layout patterns are: radial, cruciform, and uniform. The radial arrangement is
the most common one in European practice, though EC2 allows for other arrangements.
Conversely, a cruciform layout is the most adopted one in the US practice (ACI 318), and the
uniform type is usually referred to as the ‘UK-type’. Figure 2.7 illustrates these different
arrangements. Einpaul et al. (2016b) observed that tests on shear-reinforced flat slabs with radial
and cruciform arrangements showed similar performances regarding the maximum punching
capacity, as oppose to the findings of Vollum et al. (2010), who tested the efficiency of stirrups with
ACI-type (cruciform) arrangement. Vollum et al. (2010) concluded that stirrups in the ACI-type
layout can increase the shear strength of flat slabs by a multiple of up to 1.5, compared with
multiples of 2.0 or more for well anchored radial shear reinforcement.
(a) (b) (c)

Figure 2.7 – Different layouts for shear reinforcement: (a) radial arrangement, (b) cruciform
arrangement (ACI-type) and (c) uniform arrangement (UK-type).
Lips et al. (2012) observed an increase in rotation capacity of 220% and 421% for flat slabs with
stirrups and studs, respectively, in comparison to a reference specimen without shear reinforcement.
As can be seen, the enhancement varies significantly between different systems. A few studies, such
as Regan and Samadian (2001), who tested 5 types of shear reinforcement, have already raised the
issue of different levels of contribution for different types of shear reinforcement. Einpaul et al.
(2016b) tested 11 different types and reported the disparity among shear reinforcement systems
based on the MC2010 parameter ksys (more details in section 2.4.5), which varied between 2.0 to
3.8, and up to 4.5 in special cases. EC2 does not make any specific recommendation regarding the
choice of system, whereas other codes of practice such as ACI 318-14 and MC2010 acknowledge
their differences and provide some distinction between them.
47
2.1.3 Shear Failure Modes
Punching failure in reinforced concrete flat slabs without shear reinforcement is a brittle failure
mode in which the flexural reinforcement does not necessarily yield. Failure is preceded by the
formation of inclined cracks around the column leading to a punching cone being separated from
the slab and an almost complete loss of the load-carrying capacity. It is considered one of the most
critical D-regions in concrete structures, developing very high bending moments and an extremely
complicated three-dimensional state of stress. As previously discussed, shear reinforcement is often
added to the region around the slab/column connection in order to provide some ductility and
increase shear resistance.
Failure modes of shear reinforced flat slabs are highly influenced by the type of transverse
reinforcement system, its anchorage, spacing, distribution, and distance to the supported area.
According to Lips et al. (2012), for flat slabs designed with typical detailing rules, three potential
shear failure modes govern: punching within the shear-reinforced area (Figure 2.8a); punching
outside the shear-reinforced zone (Figure 2.8b); and failure of the concrete close to the column
(Figure 2.8c). Furthermore, other failures may occur such as delamination (Figure 2.8d), and failure
between the transverse reinforcement (Figure 2.8e). In some cases, depending on the ratio of
flexural reinforcement and shear reinforcement, the flexural capacity may govern (Figure 2.8f).
(a) (b) (c)
(d) (e) (f)

Figure 2.8 – Potential failure modes: (a) failure within shear-reinforced area; (b) failure outside
shear-reinforced area; (c) failure close to column by crushing of concrete struts; (d) delamination of
concrete core; (e) failure between transverse reinforcement; and (f) flexural failure. (Based on Lips
et al., 2012).
Both failures between the transverse reinforcement and outside the shear reinforced area can be
avoided with a proper amount and detailing of shear reinforcement. The failures of the compressive
struts close to the column and within the shear reinforced area demands a better understanding on
the interaction between the concrete and the shear reinforcement. These last modes require further
investigation due to limited research in the current available literature.
48
2.1.4 Size Effect
Graf (1938) was amongst the first to observe that nominal shear resistance and slab effective depth
could vary in a non-proportional way, by testing slabs with mean effective depth varying from 271
mm to 474 mm. He found that the shear stress of a 500 mm thick slab at punching failure is
approximately the same as beams failing in shear, which has around half of the punching capacity
of a 150 mm two-way thick slab. Richart (1948), while testing footings, observed that shear stresses
at failure reduces significantly with increasing effective depth. According to Muttoni (2008), in the
early 1960s when the ACI formula was originally proposed, most of the experimental data available
included slabs with small effective depth, where the size effect is not apparent. The neglect of size
effect can lead to catastrophic consequences such as the shear failures of a few 900 mm deep beams
in two different Air Force warehouses (Anderson, 1957), and the Laval bridge in Quebec, which
suddenly failed in shear after 35 years causing the collapse of a major overpass (Collins et al.,
2008).
Regan (1986) tested six specimens where the overall dimension, effective depth, size of loading
discs, bar size and spacing were scaled linearly in the ratios 1:0.64:0.32. The greatest slab thickness
and aggregate size were 250 mm and 20 mm, respectively. The concrete compressive strength was
kept similar at around 43 MPa for all specimens. The tests clearly show an influence of effective
depth on shear stress at failure. Regan (1986) suggests that the observed size effect agrees
reasonably well with that of BS8110, which is proportional to 1/ (more details in section 2.4.2).
Interestingly, this variation is reduced when the aggregate is scaled approximately with the effective
depth.
Birkle and Dilger (2008) observed a significant effect of size on shear resistance by testing and
evaluating slabs with thickness ranging from 160 mm to 500 mm. They found that, though nominal
shear strength reduces with increasing effective depth for all slabs, the reduction is less pronounced
for specimens with shear reinforcement. This is to be expected, as the shear carried out by
transverse reinforcement is not subject to size effect.
The great majority of current codes of practices account for size effect through empirical
expressions. EC2, for example, adopts = 1 + 200/ ≤ 2.0. Interestingly, the ACI 318-14 still
does not allow for size effect in its shear design formulations, and the only difference in the
calculation of concrete punching shear factored strength between ACI 318-14 and the Canadian
standard CSA A23.3-14 (2014) is the presence of a size factor for slabs with effective depth greater
than 300 mm. MC2010 is based on the critical shear crack of Muttoni (2008), which provides a
49
physical explanation for the size effect. More details on design equations from some codes of
practice are presented in section 2.4.
2.1.5 Flexural reinforcement
There are currently more than a thousand experiments on conventional punching shear specimens
(refer to Figure 1.1) reported in the literature. The major issue when evaluating those results is the
significant number of variables influencing the punching capacity, such as concrete strength,
column size, slab thickness, aggregate size, etc. The impact of flexural reinforcement ratio is also
included in that group. Although many studies have been carried out using the top reinforcement
ratio as a main variable, the results can be masked by the influence of some of the previously listed
as well as some unknown variables (test setup, boundary conditions). Anyhow, it has been
established that punching resistance increases with increasing flexural capacity of the member.
Vanderbilt (1972) tested 15 flat slabs-interior column connection with a different set up consisting
of square plates surrounded by ring beams in the attempt of better simulate a continuous two-way
structure. The slabs were 51 mm thick. The variables included the size and shape of the column, and
the amount of reinforcement. He concluded that the enhancement in strength due to the increase in
reinforcement ratio is independent of the column shape and size, although doubling the
reinforcement (from ρ = 1.3% to ρ = 2.6%) only increased the strength by 15% ~ 25%.
Marzouk and Hussein (1991) tested 17 isolated specimens on internal columns varying the slab
thickness from 90 mm to 150 mm, and the reinforcement ratio from as low as 0.491% up to 2.37%.
The concrete compressive strength was around 70 MPa. The results once more confirmed the
increase in punching strength with increased reinforcement ratio, though the authors also reported a
loss of ductility as a consequence. Hallgren (1996) carried out a study with similar purposes but
with thicker slabs, and observed a less pronounced increase of punching resistance with
reinforcement ratio.
Mamede et al. (2013) verified the increase in punching strength with reinforcement using 3D
NLFEA. The authors observed a similar loss in ductility as reported by Marzouk and Hussein
(1991). Mamede et al. (2013) observed that the relationship between NLFEA predicted strength and
reinforcement ratio followed a cubic root function, which is in accordance to the assumption
presented in a few codes of practice, as pointed out by Dilger et al. (2005). Both the superseded
BS8110 and the current European design code EC2 account for the influence of hogging flexural
reinforcement on the punching capacity of slab-column connections by increasing the punching
shear strength in proportion to the cube root of the flexural reinforcement ratio. Other design codes,
50
such as the ones based on the American ACI 318 (AS 3600-1994 and CSA Standard A23.3-14) do
not explicitly account for the contribution of the flexural reinforcement ratio on the punching
resistance of flat slabs.
It is worth mentioning that almost the entirety of the studies and recommendations on the influence
of flexural reinforcement refers to the top flexural reinforcement over the slab-column connection.
However, this research will later show and discuss that in some cases, the bottom flexural
reinforcement over the span in continuous slabs can have an even greater impact on punching
resistance (more details in Chapter 5).
2.2 INFLUENCE OF CONTINUITY ON PUNCHING SHEAR RESISTANCE
2.2.1 General Aspects
Punching tests are usually made on isolated slab elements like that shown in Figure 1.1 in which the
line of contraflexure for radial moments is fixed. The simplified boundary conditions in such tests
does not well represent the boundary conditions in flat slabs where flexural continuity and in-plane
restraint are present. In continuous floor plates, the line of radial contraflexure moves as concrete
cracks and reinforcement yields. This movement and its effects are not generally considered in
design methods for punching shear, which are mostly empirical based, but can be readily
investigated using the CSCT, or MC2010 which is based on the CSCT. Chapters 4, 5, and 6 include
studies verifying the influence of continuity on flat slabs, and how the CSCT can account for it. The
following will present a literature review on some of the experimental and analytical studies
concerning the influence of the surrounding structure on punching shear resistance.
2.2.2 Continuity
The vertical displacement of a reinforced concrete slab under transverse load generates a lateral
outward expansion which is prevented in some degree by the supports (columns, beams, walls, etc.)
or the adjacent panels outside the failure zone acting as a stiff diaphragm. The in-plane restraining
forces mobilized increase the flexural and shear resistance of the structure. This is commonly
known as compressive membrane action. Although fairly hard to quantify due to the amount of
parameters involved, mainly related to the stiffness of the structure itself, the benefits of
compressive membrane on punching resistance is well known and has already been proved by many
researchers.
51
Additionally, continuous slabs allow some level of moment redistribution between hogging
moments over the support and sagging moments over the span. As a consequence, the line of
contraflexure of the slab, which is fixed in conventional punching shear specimens such as Figure
1.1, shifts its location and changes the shear slenderness of the slab, taken as / (distance between
the center of the slab and the line of contraflexure, , and slab effective depth, ).
Westergaard and Slater (1921) were among the first to notice an increase on resistance of tested full
scale slab panels in comparison to the strength of beams with equivalent reinforcement. Ockelston
(1955) monitored portions of a ten year-old reinforced concrete building during its demolition,
showing deflections and strains under working load to be generally close to calculated with the
assumption of uncracked concrete, flexural failure loads to be considerably underestimated for
bending in two directions, and punching shear strength of similar slabs to differ considerably. These
behaviours are largely related to the membrane effect present in real structures such as the
mentioned building.
Guralnick and Fraugh (1963) presented a report with the results of a three-quarter scale flat plate
specimen with an overall dimension of 13.72 m x 13.72 m, supported by 16 columns and
surrounded by beams. The authors observed a significant increase on strength of the slab and
attributed that to strain hardening of the reinforcement and compressive membrane action. The
Guralnick and Fraugh (1963) study is used in the analysis developed in section 4.7 of this thesis,
where it is described and discussed in more detail.
Vanderbilt (1972) tested 15 flat slabs on an interior column with a set up that was believed to
simulate more closely a real continuous two-way slab. The specimens were cast integrally with
heavily reinforced surrounding ring beams. The author observed that the new type of specimen
response did not differ much from equivalent footing type tests. Regan (1986) carried out a series of
punching tests using the test setup shown in Figure 2.9 which he depicted as “Group IV, boundary
restraint”. The test arrangement was designed to investigate the influence of rotational edge restraint
by adjusting the ratios of the upwards and downwards loads. The tests demonstrated that punching
strength increases with the restraining moments. Regan (1986) attributed the increased punching
resistance to the combined effects of both the moment and the lateral restraint provided by the
uncracked concrete at the slab edges. The author also observed that, because the factors involved in
the enhancement by different boundary conditions were not fully understood, it could not be
allowed for in Codes of Practice at the time.
Rankin and Long’s (1987) experimental programme included both conventional and large panel
specimens with in-plane restraint provided by extending the slab beyond the loaded perimeter. The
specimens were built on a ¼ scale to its original size with reinforcement ratios varying from
52
0.517% to 0.802%. They found that the punching resistance of large panel specimens was 30-50%
greater than that of comparable conventional punching shear specimens, and attributed that increase
to the compressive membrane action.
Chana and Desai (1992a) tested five 9 m x 9 m and 250 mm thick panels representing a region of a
slab surrounding an internal column of a continuous flat slab. The authors compared the results to
that of an equivalent isolated specimen tested by Chana and Desai (1992b), finding a significant
increase in punching shear strength due to the membrane action developed by the slab portion
outside the failure zone. The authors proposed a design equation which accounts for an increase of
40% in resistance for slabs without links as shear reinforcement. This study, like the one by
Guralnick and Fraugh (1963), is particularly important for this research and it will be described in
more details over section 4.4.
Figure 2.9 – Test setup adopted by Regan (1986). Dimensions in mm.
Kuang and Morley (1992) tested twelve reinforced concrete slabs to investigate the influence on
punching resistance of restraining all four slab edges with beams. The study focused on evaluating
the influence on punching resistance of span-depth ratio, degree of edge restraint, and percentage of
flexural reinforcement. Kuang and Morley (1992) found that longitudinal rebar had an important
effect on the punching strength of the lightly reinforced restrained slabs, but little effect on heavily
reinforced slabs.
53
Regan (1993) studied the influence of continuity on flat slabs supported on edge columns by testing
five 200 mm thick, 5784 mm long and 3000 mm wide specimens. The slabs extended beyond the
hogging moment region, allowing for the redistribution of moment between support and span. In
general, the slabs failed in punching with the measured final load fairly close to that calculated to
correspond to the formation of plastic hinges at the column face and in the span. The slabs of Regan
(1993) are described and analysed in detail in Chapter 5.
Sherif and Dilger (2000a) presented a new test setup to simulate continuity on flat slabs. The
specimens were 150 mm thick, 5000 mm wide, 7500 mm long, and aimed to represent the region
limited by the lines of zero shear, Figure 2.10. Three of the slab edges were free to move vertically,
but restrained against rotation through clamping jaws. The load was applied at 24 points.
Surprisingly, Sherif and Dilger (2000a) concluded that, for an internal flat slab-column connection,
the shear strength of a continuous slab seems to be the same as for an equivalent isolated specimens
such as the one presented in Figure 1.1. This issue can be explained by the works of Einpaul et al.
(2015), who observed that the effect of continuity is dependent on reinforcement ratio between span
and support, and slab geometry.
Figure 2.10 – Specimen tested by Sherif and Dilger (2000a).
Salim and Sebastian (2003) simulated the restraining conditions by reinforcing internal slab column
specimens with hoop reinforcement. The experimental results showed that there was a definite
increase in ultimate punching load of the slab panels as the degree of restraint increased. The
ultimate nondimensional punching strength increased 38% as the number of hoops went from 1 to
2, and flattened out with 3 hoops. The model presented by Salim and Sebastian (2003) allows the
enhancement in punching shear strength due to the development of compressive membrane action.
54
The compressive membrane forces were calculated with a membrane-modified flexure theory of
elasto-plasticity, and the predictions showed good agreement with a range of experimental results.
Choi and Kim’s (2012) study was intended to evaluate the restraining conditions through a series of
tests with a very complex setup that was only partially successful, as the edge restraint system failed
in providing a zero rotation condition as initially intended. The setup is illustrated in Figure 2.11.
They concluded that the moment redistribution limit is directly affected by the nominal punching
shear resistance to the nominal flexural strength ratio of flat slabs, and that the 30% moment
redistribution established by EC2 might be conservative.
Einpaul et al. (2015) studied the influence of moment redistribution and compressive membrane
action on punching resistance of flat slabs on internal columns. Their study concluded that punching
strength may be underestimated by current codes for cases with low amounts of flexural
reinforcement due to neglect of the benefits of continuity.
Figure 2.11 – Choi and Kim’s (2012) test setup.
Previous research has shown that the effects of compressive membrane on punching resistance are
greatest for internal columns, reduced by slab openings, and marginal for corner columns.
2.3 MECHANICAL MODELS
2.3.1 General approaches
Though researchers have been working on developing theories and formulations to describe the
punching phenomenon for more than 50 years, there is still no generally accepted theory. Regan and
Bræstrup (1985) did a comprehensive review of models available at the time. The authors attempted
55
to divide those numerous models into groups with similar approaches. It is worth mentioning that
such a classification is only meant to help identifying and understanding these approaches, since in
some cases they do not exactly fit, as pointed out by fib (2001).
2.3.1.1 Flexural punching approach

Flexural punching approaches derive the punching capacity of a reinforced concrete slab from its
flexural capacity. This approach is based on the substantial number of flat slab tests results reported
to have failed in punching with the ultimate load fairly close to the flexural capacity. Gesund and
Kaushik (1970) have carried out an investigation of the ratio between shear forces corresponding to
flexural failure (Vflex) and reported failure loads (Vtest) of 106 tests reported as punching failures.
The authors found an average of Vflex/Vtest = 1.015, with a standard deviation SD = 0.248, which led
to the development of a punching shear formulation based on yield line analysis proposed by
Gesund and Dikshit (1971). Models with similar approaches were developed by Moe (1961),
Yitzhaki (1966), Long (1975), Long and Rankin (1987). fib (2001) also includes in this
classification a model by Nölting (1984) based on the inclined compression strains. As the flexural
reinforcement ratio increases, the punching resistance becomes critical. As a consequence, punching
failure, which is brittle and without much warning, can govern.
2.3.1.2 Plasticity Approach

Plasticity based approaches for punching shear have been developed by Bræstrup et al. (1976),
Nielsen et al. (1978), Hess et al. (1978), and Bræstrup (1978) amongst others. The failure
mechanism is shown in Figure 2.12, where the deformations are concentrated in an axisymmetric
surface, initiating at the face of the loaded area and propagating towards the slab thickness. The
failure is not necessarily conical and the contribution from flexural reinforcement is not accounted
for. Analytical expressions for the shear force corresponding to the punching load are derived by
equating the rate of internal work dissipated by the failure with the external work done by the
applied load. Regan and Bræstrup (1985) state that, in order to improve the agreement between
plasticity approaches predictions and test results, the concrete compressive strength needs to be
reduced. The method introduced more recently by Hoang and Pop (2016) can also fit in this
classification, since it is based on a rigid-plastic upper-bound approach, called the crack sliding
model. Lower bond solutions based on the plasticity approach were developed by amongst others,
Marti (1980) and Pralong (1982).
56
Figure 2.12 – The failure mechanism for the plasticity approach.
2.3.1.3 Control Surface Approach

Analogously to the traditionally adopted procedure for reinforced concrete beams, punching shear
design for flat slabs can also be based on a nominal shear stress over a section of the member. This
stress is defined by the shear force V divided by a control surface around the loaded area (control
perimeter), which is normal to the slab plane. This control perimeter is usually assume to be either
at a fixed distance from the loaded area, or to be in conformity with the shape of the loaded area at a
defined minimum distance. This method was initially introduced by Talbot (1913) who found that
nominal shear strength of square footings centrally loaded by square columns were similar to the
ultimate shear strength of simple beams without shear reinforcement. The punching capacity is
limited by some strength parameter, usually associated with the tensile concrete strength.
Regan and Bræstrup (1985) comment that this control surface approach, although purely empirical,
can lead to realistic and consistent strength predictions. This is significant, as a great number of
codes of practices such as BS8110, EC2 and ACI 318 adopts this method for punching shear
recommendations. Their equations and assumptions will be discussed later on over section 2.4.
Regan and Bræstrup (1985) and fib (2001) list a couple of more classifications for models: Strut-
and-tie based approaches such as the ones presented by Pralong (1982) and Andrä (1982); and
Fracture mechanics based approaches such as Hallgren (1996), who proposed a failure criterion
based on a simple fracture mechanical model, which accounts for both the brittleness of concrete
and size effect.
The following will describe in more details models that relate directly with the Critical Shear Crack
Theory of Muttoni (2008), as this was the main approach used in this research.
57
2.3.2 Kinnunen and Nylander (1960)
Kinnunen and Nylander (1960) developed a model based on 61 punching shear tests on circular
slabs with circular stubs. Kinnunen and Nylander’s (1960) tests provided an insight on the shear
crack formation, and the model was derived from the equilibrium of forces acting at sectors outside
the inclined shear crack, shown in Figure 2.13, limited on their sides by the radial cracks and in the
front by the tangential shear crack, Figure 2.14. Each sector is assumed to act as a rigid body
supported by a conical shell, and rotates around a centre of rotation located in the root of the shear
crack, as presented in Figure 2.15.
Figure 2.13 – Sector element on Kinnunen-Nylander’s (1960) model.
Figure 2.14 - Force body diagram on sector element at Kinnunen and Nylander’s (1960) model.
(a)
58
(b) (c)
Figure 2.15 - Rotation ψ and centre of rotation (CR) of the slab segment at Kinnunen and
Nylander’s (1960) model.
The internal forces are defined as functions of the mechanical properties of both concrete and
reinforcement steel, as well as the slab rotation presented in Figure 2.15. The failure criterion is
defined when the stress in the conical shell, and both the compression strain and stress in the
tangential direction at a point located vertically under the shear crack root reach critical values
simultaneously. The maximum strain depends on the ratio between the size of the column, , and
the slab effective depth, , whereas the values for the stresses are based on the concrete mechanical
properties. Kinnunen and Nylander’s (1960) model was derived for different situations relative to
the amount of flexural reinforcement yielding over the slab. The radius where yielding is reached,
, is calculated according to Equation (2.2).
= ( − ) (2.2)
where is the reinforcement yield stress, and is the steel Young’s modulus. Equation (2.2)
allows the model to predict the ultimate load irrespective to the type of failure (punching or
flexure). At failure, the rotation is calculated according to Equation (2.3).
= 1+ (2.3)
2
where is the ultimate compressive strain at the slab soffit located vertically underneath the root
of the shear crack. Because the model was originally developed for axisymmetric cases with ring
reinforcement, a correction factor κ = 1.1 is used to account for dowel action in two-way
reinforcement layouts. fib (2001) details the equations from Kinnunen and Nylander’s (1960) model
for the case where the yield stress is reached inside the area of the conical shell, with the radius
(refer to Figure 2.15). The first equation, (2.4), is derived by setting Σ = 0, and is a function of the
ultimate concrete stress, . Equation (2.5) comes from Σ = 0, and is related to the flexural
reinforcement. The failure load of a flat slab without shear reinforcement is then calculated by
59
setting (2.4) and (2.5) to be equal, and solving them with iterations on the ratio of the concrete
compression zone .
2
1+
, = ( ) (2.4)
1+
1−
= 4 1+ 3 (2.5)
,
2 −
( )
where = , ( )= , = , is the slope of the conical shell, as shown in
Figure 2.15, is the reinforcement ratio, and is the radius of the circular area inside the shell.
More reviews, including design examples, of this model can be found in CEB bulletin (1966) and
Schaidt et al. (1970).
Kinnunen and Nylander’s (1960) model was, since then, reviewed and improved by many
researchers such as Reimann (1963), who replaced the compressed conical shell by a yield hinge
near the column, and use only the circumferential concrete stress instead of both stress and strain as
a failure parameter; Long and Bond (1967), who developed their failure criterion based on a failure
envelope of the concrete subject to a biaxial compression state; Broms (1990), who included the
size effect and applied known values for concrete properties instead of calibrating the model against
specific test results as suggested by Kinnunen and Nylander (1960); Hallgren (1996), who proposed
a failure criterion based on a fracture mechanical and showed that punching shear resistance is
increased by using high strength concrete, amongst others authors. Models of the Kinnunen and
Nylander (1960) type give good results but are known as overly complex for code implementation.
2.3.3 Critical Shear Crack Theory
The origins of the Critical Shear Crack Theory (CSCT) can be found in Muttoni and Schwartz
(1991), who investigated the cause of shear failure in beams and slabs without shear reinforcement,
and observed that the crack formation and pattern had a significant impact on the failure load, and
could be influenced by placing or omitting reinforcement in particular regions close to the column.
This issue had also been previously confirmed by Bollinger (1985), who tested slabs reinforced
with concentric rings placed at the edge of the slab only, and with additional ring in the region close
to the column, Figure 2.16. It was shown that the additional reinforcement (Figure 2.16b) over that
critical region initiated the formation of a crack and reduced the punching resistance by around
40%.
60
(a) (b)
Vtest = 76.1 kN Vtest = 44.0 kN
3 reinforcement rings db 10 3 reinf. rings db 10 + 1 reinf. rings db 12
(3 # 3) (3 # 3 + 1 # 4)
Figure 2.16 – Specimens 11 and 12 by Bollinger (1985) with and without ring reinforcement placed
over the critical region.
Muttoni and Schwartz (1991) concluded that punching shear strength is inversely proportional to
the width of that critical shear crack, which is a diagonal crack that propagates through the slab to
the column by an inclined compression strut (Figure 2.17). Furthermore, Muttoni and Schwartz
(1991) stated that, because crack width is related to the slab thickness, it can be assumed to be
proportional to the product ψ.d, where d is the effective depth of the slab and ψ is the rotation in the
region of the column (refer to Figure 2.17).
Figure 2.17 – Critical shear crack and the rotation ψ of the slab (based on Muttoni, 2008).
These conclusions and assumptions led to a semi-empirical failure criterion for punching shear
relating the width of the critical crack to the relation ψ.d presented in (2.6):
1
=
(2.6)
1+
4
Where VR is the punching shear, b0 is a control perimeter located at d/2 from the column and fc is
the average compressive strength of the concrete. Additionally, Guandalini and Muttoni (2004),
who tested symmetric flat slabs without transverse reinforcement, showed the radial compressive
61
strain in the soffit of the slab close to the column to reduce at loads near punching failure due to the
opening of the critical shear crack. That would be later related to the development of an elbow-
shaped strut with a horizontal tie, Figure 2.18. A similar phenomenon had already been observed by
Muttoni and Schwartz (1991) while studying the crack pattern of beams without shear
reinforcement.
Figure 2.18 – Elbow-shaped strut with a horizontal tie close to the column (based on Muttoni,
2008).
Muttoni (2003) improved the failure criterion showed in equation (2.6) by introducing some of the
aspects presented in Walraven’s (1981) probabilistic model for shear transfer through aggregate
interlock, which expresses the stresses acting on the crack faces as a function of the crack
kinematics. The new failure criterion is also consistent with the Vecchio and Collins (1986)
relationship between shear resistance provided by aggregate interlock, crack width, and normal
compressive stress on the crack, that was derived using Walraven’s (1981) data.
The improved failure criterion introduced by Muttoni (2003) and presented in (2.7) was originally
developed for the Swiss code SIA 262 (2003). It accounts for the roughness of the critical shear
crack, which determines its capacity to carry shear forces by dividing ψ.d by (dg0 + dg), where dg is
the maximum aggregate size, and dg0 is a reference size equal to 16 mm.
3/4
=
1 + 15 (2.7)
+
Muttoni (2008) shows that predictions of equation (2.7) agrees very well with a great number of
punching tests from the literature. Failure is assumed to occur when the slab’s load-rotation curve
intersects the failure criterion produced by (2.7). Consequently, it is paramount that, in order to
calculate shear resistance according to the CSCT, the relationship between the specimen’s rotation
and the applied load must be known.
The load-rotation relationship can be obtained through measurement in the case of experimental
studies, by a nonlinear numerical simulation of the slab’s deformed shape, or be calculated by
means of analytical expressions based on simplifications of the structural behaviour. Muttoni (2008)
62
derived a general analytical expression for the load-rotation response of an axisymmetric slab. The
equation is based on the assumption that the deflected shape of the isolated slab element is conical
outside the critical shear crack. Inside the critical shear crack, it is assumed that the curvatures in
both directions are constant and equal, which means that the deflected shape is assumed to be
spherical. The internal forces of a slab sector limited by radial and tangential cracks (Figure 2.19)
can be calculated according to a quadrilinear moment-cruvature relationship presented in Figure
2.20.
Figure 2.19 – Internal forces and moments over a slab sector (based on Muttoni, 2008).
Figure 2.20 – Moment-curvature relationship assumed by Muttoni (2008).
The contribution of reinforcement before cracking is neglected, and the steel and concrete are
assumed to have a linear-elastic behaviour after cracking. A parameter to account for the change in
layout of the reinforcement from the typical orthogonal placement to a polar symmetry (tangential
and radial directions) for the axisymmetric case was taken as β=0.6 on the basis of good agreement
63
with tests. A perfectly plastic behaviour of the reinforcement after yielding along with a rectangular
stress block for concrete in the compression zone is also considered.
The final expression is shown in equation (2.8). Its detailed derivation can be found in Appendix 1
of Muttoni (2008):
2.
= . − . + .〈 − 〉 + . . . 〈ln( ) − ln( )〉 + . . .〈 − 〉
−
(2.8)
+ .〈 − 〉 + . . . 〈ln( ) − ln( )〉
The operator 〈 〉 is for ≥ 0 and 0 for < 0. The parameters are described as follows: rq is the
radius of the load introduction at the perimeter, rc is the radius of a circular column, mr is the radial
moment per unit width, ro is the radius of the critical shear crack, mR is the nominal moment
capacity per unit width, ry is the radius of yielded zone, E.I1 is the tangential flexural stiffness after
cracking, r1 is the radius of the zone in which cracking is stabilized, χTS is the decrease in curvature
due to tension stiffening, mcr is the cracking moment per unit width, rcr is the radius of cracked
zone, E.I0 is the flexural stiffness before cracking, rs is the radius of circular isolated slab element.
Muttoni (2008) also proposes a simplified load-rotation based on a simpler bilinear moment-
curvature relationship showed as a dotted line in Figure 2.20, by neglecting the concrete tensile
strength. The equations are shown in (2.9) to (2.11):
2.
= . . . 1+ (elastic regime) (2.9)
−
2.
= . . . 1+ (elasto-plastic regime) (2.10)
−
= 2. . (plastic regime) (2.11)

−
A further simplification can be applied to the axisymmetric load-rotation expression in which a

parabola with a 3/2 exponent is assumed for the rotation as a function of V/Vflex, and that 75% of
the isolated slab has reached its flexural strength. The equation is also a function of the slenderness
of the slab, which is represented by the factor rs/d, and the reinforcement strain (fy/Es), and it is
shown in (2.12).
= 1.5 (2.12)
The CSCT assumes that the punching shear resistance is a function of the slab rotation, it accounts
for size effect, and it relates the punching strength to the opening of a critical shear crack in the slab
64
(refer to Figure 2.17). Shear is assumed to be transferred through the critical shear crack by
aggregate interlock. Muttoni (2008) has shown that the design equations of the CSCT for the shear
resistance provided by concrete is consistent with the strength predictions of numerous
experimental results. Figure 2.21 represents that good agreement by comparing the predictions of
the CSCT with axisymmetric tests by Kinnunen and Nylander (1960) for different flexural
reinforcement ratios.
Curve Equation
(2.7)
(2.8)
(2.10)
Experimental
Figure 2.21 – Comparison of load-rotation curves for axisymmetric tests by Kinnunen and Nylander
(1960) and CSCT equations (extracted from Muttoni, 2008).
Around the same time, Muttoni and Fernández Ruiz (2008) investigated the significance of the
critical shear crack on the shear strength of beams and one-way slabs without transverse
reinforcement. In a similar methodology, they proposed a failure criterion accounting for the
aggregate size, concrete compressive strength and width of the critical shear crack, shown in (2.13).
1 2
=
6 1 + 120 (2.13)
16 +
where is the longitudinal strain evaluated in the critical region which is located at a depth 0.6d
from the compression face. Concrete tensile strength is neglected and the assumption of a linear
elastic behaviour in compression as well as plane sections to remain plane are adopted. A simplified
design method adopted by the Swiss Code for structural concrete (SIA Code 262, 2003) is also
presented. The expressions are shown to agree well with experimental studies.
Muttoni and his co-workers have since then extrapolated the CSCT to various other situations.
Guandalini et al. (2009) studied the punching strength of slabs with low flexural reinforcement
65
ratios in which shear failure followed yielding of the flexural reinforcement. 11 isolated specimens
without transverse reinforcement were tested in a similar arrangement to that presented in Figure
1.1. The size of the slab and of the aggregate was also varied in order to investigate their impact on
punching resistance. The radial deflected shape of the specimens was essentially conical as
observed by many researchers and assumed by the CSCT. Through measuring the changes in slab
thickness during the tests, Guandalini et al. (2009) observed that the internal critical shear crack
forms after 50 to 70% of the failure load, which supports the hypothesis of the CSCT. Comparisons
with the tests showed the CSCT to be effective for flat slabs failing in punching with and without
plastic deformations i.e. low and large flexural reinforcement ratios.
Fernández Ruiz and Muttoni (2009) propose a new theoretical model based on the CSCT to
investigate the resistance and ductility of flat slabs with transverse reinforcement. A portion of the
applied shear is resisted by the concrete, which is limited by the opening of the critical shear crack
and its roughness, and the remaining is resisted by the shear reinforcement. Figure 2.22 depicts the
contribution of concrete and shear reinforcement to the punching shear resistance according to the
CSCT. The concrete contribution has already been discussed (Muttoni, 2008) and can be calculated
by equation (2.7). The shear reinforcement contribution is estimated from the main CSCT
hypothesis of proportionality between the opening of the crack and the product ψ.d by equation
(2.14).
Figure 2.22 – Contribution of concrete and shear reinforcement as function of the slab rotation
(Fernández Ruiz and Muttoni, 2009).
= ( ). . sin( ) (2.14)
where is the shear reinforcement cross-sectional, is the angle between the shear reinforcing
bar and the slab plane, and ( ) is the shear reinforcement stress as a function of the crack lips
parallel to the shear reinforcement which is a function of the rotation of the slab, . Fernández Ruiz
and Muttoni (2009) derived a number of expressions for ( ) for different reinforcement systems
66
and bond conditions. The punching resistance is then obtained by intersecting the load-rotation
response of the slab with the failure criterion of + (Equation (2.7) + Equation (2.14)).
Fernández Ruiz and Muttoni (2009) calculate the punching shear resistance outside the shear
reinforced zone by finding the intersection of the load-rotation curve and the failure criterion
defined by (Equation (2.7)) with replaced by , , which is the control perimeter at /2
beyond the outer layer of shear reinforcement. This approach is found to be slightly conservative
because it neglects the portion of rotation developed within the shear-reinforced zone. For cases
where crushing of the concrete strut near the column governs punching failure, Fernández Ruiz and
Muttoni (2009) recommend the failure criterion to be set as a proportion of , by adopting the
parameter , taken as 3 for well-anchored shear reinforcement and 2 for other cases. A code-like
formulation is also proposed by Fernández Ruiz and Muttoni (2009) for , presented in equation
(2.15) for smooth and deformed shear reinforcement.
.
= . ≤ . Smooth shear reinforcement
6
(2.15)
.
= + . ≤ . Deformed shear reinforcement
6
where is the amount of shear reinforcement within a perimeter at d from the edge of the
column, is the shear reinforcement design yield strength, is the average bond stress with a
recommended value of 5 MPa for usual cases, and is the diameter of the reinforcing bar. Both
proposed and simplified relations are shown to give satisfactory predictions against experimental
results with different shear reinforcing systems.
Fernández Ruiz et al. (2010), who studied the shear strengthening of slabs by adding a new system
consisting of bonded bars post-installed, based their proposed design concept on the CSCT. This
concept was validated through 12 tests carried out with specimens with different flexural
reinforcement ratios and shear strengthening arrangements, showing an increase in both
deformation capacity and punching strength. It allows to account for different failure modes such as
crushing of the concrete strut near the support, and punching within and outside the shear reinforced
zone.
Rodrigues et al. (2010) applied the principles of the CSCT to develop an analytical expression in
order to calculate the rotation capacity of plastic hinges in one-way members subject to shear. The
research investigated the relation between shear strength and rotation of one-way slabs without
shear reinforcement. The resulting design equation accounts for effective depth, shear span, flexural
67
reinforcement ratio, yield strength of flexural steel, concrete strength in compression, and aggregate
size.
The CSCT was initially developed for an axisymmetric flat slab case as previously discussed.
Sagaseta et al. (2011) carried out a series of flat slabs on internal column tests without transverse
reinforcement and different loading conditions and flexural reinforcement ratios in each orthogonal
direction, and evaluated the CSCT predictions for these non-axisymmetric conditions. A physical
justification for the common practice of calculating the punching shear resistance by averaging the
orthogonal reinforcement ratio over the column, as adopted by many codes of practices such as EC2
and BS8110, is presented based on the redistribution of shear stresses over the control perimeter.
The new approach based on the CSCT and considering a non-uniform nominal shear distribution
gives good predictions for the tested slabs. Still, the CSCT(ψmax) approach of calculating the
resistance along the direction of maximum rotations, although slightly more conservative, gives
reasonable estimations and it is practical for design purposes, according to Sagaseta et al. (2011).
Maya et al. (2012) proposed a CSCT based model to evaluate punching resistance of steel fibre
reinforced concrete slabs. The model is shown to agree well with a wide number of available test
data, and a simplified code-like version is also presented with good and reasonably conservative
estimations. Lips et al. (2012) tested 16 isolated flat slabs supported on interior columns with and
without shear reinforcement in order to examine the influence of the slab thickness, the column
size, and amount of transverse reinforcement. Studs and stirrups, which are the most widely used
shear reinforcement systems were also investigated. It proves that shear reinforced slabs allow for
the critical shear crack to develop larger widths than equivalent non shear-reinforced members as
suggested by the CSCT, according to Fernández Ruiz and Muttoni (2009). The CSCT was once
more proved to give consistent predictions.
Clément et al. (2013) studied the influence of prestressing on the punching shear resistance of flat
slabs without shear reinforcement using the CSCT and comparing the results with empirical based
predictions of codes of practice such as ACI 318-11 (2011) and EC2. It is shown that the load-
rotation stiffness response is increased due to the presence of compression normal stresses, and
consequently increases the punching shear strength according to the CSCT when compared to an
equivalent non-prestressed slab (points B and A, respectively, in Figure 2.23). The benefits over the
shear transfer capacity (which would affect the failure criterion curve) is neglected.
68
Figure 2.23 – Influence of an in-plane force over the CSCT punching resistance (Clément et al.,
2013).
Clément et al. (2013) discuss in their conclusions that most empirical design models can lead to
inaccurate predictions to test results mostly due to neglecting the influence of bending moment from
prestressing which can be accounted for with the CSCT by means of the prestressing eccentricity.
MC2010 is shown to give the best estimates with both Levels of approximation II and III (more
details in section 2.4.5), although the results get better on average as the analysis is refined (from II
to III).
Sagaseta et al. (2014) presents the results from an experimental and analytical study on punching
shear of flat slabs without transverse reinforcement supported by rectangular columns. The study
introduces two approaches to estimate the shear-resisting control perimeter in order to better
account for the effect of both column geometry and deflected shape (neglected in most codes) on
the shear forces distribution around the supported region. The first approach is based on the shear
fields of the slab which can be obtained from a simple FE linear-elastic analysis, and the other one
is based on the contact pressure in the column. The authors also show that the CSCT based model
for non-axis-symmetrical punching of flat slabs on square columns developed on Sagaseta et al.
(2011) can be applied for rectangular column cases.
Clément et al. (2014) tested 15 slabs under different load conditions in order to investigate,
separately, the influence of bending moments, in-plane forces and bonded tendons on punching
shear resistance of flat slabs. The CSCT equations are modified to account for the prestressing by
considering a self-equilibrated state of stresses in the slab, and modifying the moment-curvature law
assumed for non-prestressed cases (refer to Figure 2.20). Clément et al. (2014) also comment that
the CSCT failure criterion (Equation (2.7)) leads to conservative estimates of the punching
resistance of prestressed flat slabs. The authors justify that issue by stating the fact that in-plane
69
stresses reduces the depth of the cracked zone, and so the hypothesis of crack widths being
proportional to the product ψ.d may overestimate the actual crack width of those specific cases.
Clément et al. (2014) discuss that the failure criterion can be modified to account for the benefit of
in-plane forces over the reduction of the cracked zone by adopting a reduced effective depth.
Alternatively, a reduced rotation can be adopted, which is more convenient for design according to
Clément et al. (2014). The modified CSCT equations are shown to give consistent assessment of the
prestressing effects.
Ferreira et al. (2014) tested 12 flat slabs, 11 of which shear reinforced with double-headed studs,
and compared the predictions of ACI 318-08 (ACI, 2008) and EC2 with the CSCT estimates. The
study concludes that, despite the accuracy of results, which improves from ACI 318 to EC2 to
CSCT, the CSCT assumptions of a 45 degrees slope for the critical shear crack, and that half of the
slab rotation occurs in the critical shear crack are incorrect.
The CSCT principles can also be applied for shear design of arch-shaped members, as done by
Campana et al. (2014). Their study tested nine arch-shaped specimens, and showed that the CSCT
can be adapted to such cases yielding reasonable predictions for resistance. Cavagnis et al. (2015)
carried out tests on thirteen beams and applied photogrammetric techniques at high frequencies
during the tests in order to assess the actual critical shear crack development leading to shear
failure. The results demonstrates that different cracking patterns may be found and they develop in
distinct manners.
Einpaul et al. (2015) used the CSCT to account for the potential benefits of moment redistribution
between hogging and sagging moments and compressive membrane effect on punching shear
strength. An axisymmetric numerical model is developed and combined with the CSCT failure
criterion to assess the resistance of continuous punching tests from the literature. The authors raise
the issue that typical punching test specimens, such as the one shown in Figure 1.1, does not
account for the shift in location of the line of contraflexure due to cracking of concrete and
reinforcement yielding, which influences the shear slenderness rs/d (Figure 2.24), and does not
provide the lateral constrain commonly observed in continuous slab, which causes in-plane forces
and consequent increase in the slab stiffness and strength (Figure 2.25). This point was also
discussed by Vollum (2013).
70
Figure 2.24 – Redistribution of radial moments in a continuous slab due to cracking and/or
reinforcement yielding (Einpaul et al., 2015).
Figure 2.25 – Effect of the hogging moment region dilation being constrained by the rest of the slab
(Einpaul et al., 2015).
The assessment of the slenderness influence by Einpaul et al. (2015) can be illustrated with the
graph in Figure 2.26 showing the load x rs/L for a continuous slab (ρhog = 1.0% and ρsag = 0.5%) in
comparison to the same slab with a linear-elastic material response and an uncracked behaviour,
which clearly shows the line of contraflexure shift in location as load develops. The study also
shows the difference in response of slabs with the same plastic moment capacity, but different
reinforcement arrangement between span and support, Figure 2.27. Einpaul et al. (2015) conclude
that the demonstrated benefits of continuity is not accounted for in the punching shear provisions of
current design codes, as previously observed by others, including Chana and Desai (1992a) and
Vollum (2013).
71
Figure 2.26 – Radius of the line of contraflexure varying with load (Einpaul et al., 2015).
Figure 2.27 – Influence of hogging and sagging reinforcement distribution over slabs with the same
plastic moment capacity (Einpaul et al., 2015).
However, Soares and Vollum (2015), who compared punching shear requirements in codes of
practice, observed that, in case of uniformly loaded slabs, the benefits from the restraint of the
surrounding slab is already largely included in BS8110 and EC2 empirical based design
recommendations. Soares and Vollum (2015) evaluated the influence of continuity by extending the
conventional isolated slab (Figure 1.1) until midspan, and comparing punching resistance calculated
using NLFEA in conjunction with MC2010 LoA IV to the BS8110, EC2, and MC2010 LoA II and
III requirements. A detailed description of the approach adopted by Soares and Vollum (2015) and
their results can be found in Chapter 4 of this thesis.
72
Fernández Ruiz et al. (2015) presented a research in which a detailed study on the specific
contribution of various shear transfer actions (discussed in section 2.1.1) to the shear resistance of
reinforced concrete beams with rectangular cross-section was presented and related to the CSCT
failure criterion. The CSCT was found to be consistent with the observed shear transfer actions
developed, and to properly reproduce the influence of size and strain effects on the shear strength of
concrete members.
Einpaul et al. (2016a) tested 13 symmetric flat slabs on interior column aiming to investigate the
effect of varying the size of the supported area and the slenderness of the slab. The test results were
compared to the CSCT predictions as well as the ACI 318-14, EC2 and MC2010 recommendations.
The study shows that slenderness influences the slab stiffness, and raises the issue that this effect is
not accounted for in EC2 and ACI 318-14. The authors also discuss the effect previously observed
by Vanderbilt (1972) that increasing the size of the column reduces the nominal punching shear
strength per unit length of a control perimeter close to the column face, and state that it can be
physically explained by the CSCT, which relates the variation in size of the support to its impact on
the rotations and crack width. Einpaul et al. (2016a) concludes that the CSCT provides the best
mean and coefficient of variation for the tested slabs when compared to EC2 and ACI 318-14.
Einpaul et al. (2016b) investigated the performance of 11 different shear reinforcement systems
against punching of flat slabs supported on internal column. All the systems, which included
individual stirrups, continuous cages, double-headed studs, two-leg stirrups, bent-up bars and
bonded post-installed reinforcement, improved both strength and deformation capacity of the slabs,
but exhibited different levels of enhancement. The CSCT was applied in order to evaluate the
factor which is the MC2010’s parameter accounting for the type of shear reinforcement system. The
values varied between 2 and 3.8, with higher values reached for slabs with headed studs. The
authors pointed out, while concluding, that punching resistance obtained at load levels close to the
flexural limit of a conventional punching test specimen (refer to Figure 1.1) may not be directly
valid for a continuous slab, and that the CSCT can be used to estimate the capacity of actual slabs
based on results from isolated specimens and accounting for compressive membrane action and
moment redistribution.
2.3.4 Tangential Strain Theory
The origins of the Tangential Strain Theory (TST) were initially introduced in Broms (1990) while
improving the model of Kinnunen and Nylander (1960), which showed a few limitations. Kinnunen
and Nylander’s (1960) model used concrete critical values calibrated against specific test results
73
carried out by themselves and by Elstner and Hognestad (1956), which showed poor agreement
with normal values of concrete ultimate stress and strain, according to Broms (1990). Furthermore,
Regan and Bræstrup (1985) pointed out that Kinnunen and Nylander’s (1960) model does not
predict the punching failure load any better than other methods based on nominal shear strength.
Broms (1990) greatest contributions were accounting for size effect through a factor that considers
the height of the compression zone in both radial and tangential directions, considering known
values for concrete properties instead of calibrating the model against particular test results as
Kinnunen and Nylander (1960) suggested, and accounting for possible moment transfer between
column and slab. Similarly to Kinnunen and Nylander (1960), Broms (1990) failure criterion is
related to critical values of either a high circumferential strain or a high radial stress at the concrete
in compression close to the column. The shear force related to these conditions are denoted as
and , respectively. Reinforcement is assumed to behave as perfectly elasto-plastic. The concrete is
also assumed elasto-plastic up to the strain = 0.2%. Conditions are established to predict that
punching would occur before or after yielding of support reinforcement.
The limiting tangential compression strain is calculated with (2.16), neglecting the strength increase
of biaxial compression, and adopting the value of 0.08% reported by Wesche (1974) as the limit for
elastic behaviour.
.
150 25
= 0.0008 . (2.16)
where is the cylinder strength in MPa, and is the height of the equivalent rectangular stress
block with the stress , in mm. The punching failure load is calculated from on the basis of
classical bending theory. This mechanism can also be adapted for cases where only moment,
without shear force, is transferred between column and slab.
The failure mechanism regarding the radial compression stress is assumed to occur when the stress
in a conical shell with a constant thickness and an inclination of 15 degrees in relation to the slab
soffit reaches 1.1 ′ . is calculated by the simplified equation (2.17).
.
1 sin 15 150
= +2 1.1 sin 15
tan 30 sin 30 0.5
(2.17)
.
300
≈ 0.46 ( + 3.5 )
where is the column diameter, and is the depth to the neutral axis calculated assuming linear
elastic behaviour for the reinforcement and concrete, but with the reinforcement ratio modified by a
74
factor p to allow for inclined cracking. In the case of a moment instead of a shear force being
transferred between column and slab, the column should be considered clamped by the slab, and
failure is assumed when that clamping stress reaches 1.1 . The shear force corresponding to
punching failure is determined as the lesser value of and . The model can also be adapted for
unsymmetrical punching, where both shear forces and moments are transferred between slab and
column.
Broms (2016) introduced the Tangential Strain Theory (TST) for punching failure of flat slabs. The
theory assumes that shear forces are transferred to the column through a circumferential inclined
compression strut. The column would then be squeezed by the horizontal component of the strut up
to the yield level, after which, this squeezing component would quickly reduce, while a portion of it
is anchored back to the surrounding concrete by means of a radial tensile force. Punching failure is
assumed to be initiated by the crack developed when the resulting radial tensile strain at the column
edge exceeds the radial compression strain due to bending. This is assumed to occur when the
tangential compression strain due to bending reaches a critical value of 1% at the column.
At load levels below that corresponding to the first yielding of flexural reinforcement, the equations
from the theory of elasticity for plates, well described in Timoshenko and Woinowsky-Krieger
(1959), are assumed to be valid. Deflections resulting from both bending deformation and shear
deformation (usually neglected) are superimposed resulting on a truncated cone like shape with a
radial rotation denoted by , and calculated from elasticity as (2.18).
= = 1− (2.18)
2 4 2
where is the tangential moment at the slab edge, is the flexural stiffness of the slab, is the
shear force, is the diameter of the slab and is the diameter of the column. Broms (2016)
criticises the assumption by Kinnunen and Nylander (1960) that sector elements rotate in the form
of rigid bodies from the initial elastic stage. He states that it can lead to erroneous bending moment
distributions, since the ratio between tangential moments at the column edge and specimen edge for
the rigid body assumption can be as much as 3 times greater than calculated with the theory of
elasticity.
Punching failure is assumed to occur when the compression zone adjacent to the column, imagined
as the column capital depicted in Figure 2.28, collapses. As load increases, shear force is transferred
to the internal column capital through inclined compression struts. The formation of a shear crack
with an inclination of approximately 1:2 follows, and any shear force transferred across this crack
by aggregate interlock is neglected.
75
Figure 2.28 – Load transfer from flat slab to column. (Broms, 2016).
The squeezing pressure p caused by the inclined strut is calculated according to (2.19), by adopting
the 1:2 inclination of the strut, bending moment near to the column = 0.9 /(2 ), and the
absolute value of the concrete stress in the slab, due to bending moment m, = 2 /(0.9 ).
2×2 0.9 2×4 4
= ≈ = (2.19)
2 0.9
The surrounding compression zone restrains the volume decrease resulting from the pressure p. The
resulting stress transferred back to the surrounding area is calculated as /( + ), where is
the concrete stiffness within, and the concrete stiffness outside the column. Broms (2016) states
that “With a triangular distribution of the pressure p over the compression zone height x, the ratio
/ is approximately 5, because the peak stress is applied on the free edge of the slab but strikes
the column at a considerable distance from the free edge”.
Recorded radial strains close to the column from normal strength specimens tested by Tolf (1988),
and high-strength concrete tested by Halgren (1996), confirm that compression strains approach
zero before punching failure. This pattern was also observed by Muttoni and Schwartz (1991) while
studying beams and slabs without shear reinforcement, which is in accordance with the described
TST failure mechanism. Based on a stress-strain diagram of different concrete grades, Broms
(2016) has stipulated 0.1% as the tangential compression strain corresponding to the “yield level”.
The failure criterion expression is presented in (2.20), and accounts for both size effect and levels of
concrete brittleness.
.
/ 10
= 0.001 (2.20)
where is the critical tangential compression strain at the column edge due to the bending
moment; is the height of the compression zone at linear elastic stress conditions; and is the
reference compression zone height ( = 150 ). The effect of maximum aggregate size, not
included in equation (2.20), can be taken into account by adjusting the reference size , though
Broms (2016) states that there are currently not enough systematic tests to support the refinement
aforementioned. The TST uses Equation (2.21) to determine if flexural failure ( ) happens before
punching failure ( ).
76
−
=
/
(2.21)
=
= if( < ; ; )
Similarly, the expression for the slab rotation depends on the amount (if any) of reinforcement
yielding over the column. If part or all of the reinforcement yields, the rotation is calculated by
superimposing the elastic and rigid body portions of . Eccentric shear is accounted for within the
TST by checking the rotation capacity, as opposed to the common practice (EC2, for instance) of
checking the unbalanced moment.
Broms (2016) validates the TST by checking it against test results, the CSCT, and codes EC2 and
ACI 318-08 (ACI, 2008). The theory is found to be as accurate as the CSCT and EC2, although
EC2 did perform better for the considered database. A comparison between the ultimate rotation of
the TST and the CSCT is shown in Figure 2.29, where can be seen that both theories present similar
results.
Figure 2.29 – Comparison of punching shear strength versus ultimate rotation between TST and
CSCT. (Broms, 2016).
However, Broms (2016) raises a couple of issues with the CSCT approach:
i. The CSCT failure scenario in which the shear crack propagates down through the
compression zone is seen as inconsistent with experimental observations that
failure starts at the bottom of the slab at the column edge;
ii. The CSCT assumption that the critical shear width is proportional to the slab
rotation implicates that punching capacity decreases with time due to the increased
77
rotation resulting from concrete creep and drying shrinkage. However,

experimental tests under sustained loading such as those of Moe (1961) show a
slight enhancement of punching capacity (around 4%) even though both the width
of flexural cracks and deflection increased by 80%. The TST can account for that
minor increase by employing a creep factor (1 + φ) over the critical strain and
Young’s modulus of its model.
2.4 CODES OF PRACTICE
2.4.1 General Aspects
This research includes some comparisons of punching shear requirements, recommendations and
assumptions between four codes of practices: BS8110, EC2, ACI 318, and MC2010. All four
methods take the punching resistance of slabs with shear reinforcement as the least of VR,cs and
VR,out but not greater than VR,max, where:
i. , = , + , ≥ , is the combined shear resistance of the concrete , , where
≤ 1 and shear reinforcement , .
ii. , = is the strength of an otherwise similar slab without shear reinforcement in
which is the design concrete shear stress, is the basic control perimeter and is the
average slab effective depth.
iii. , = ≥ , is the shear resistance provided by concrete along a perimeter of
length just outside the shear reinforcement.
iv. , is the maximum possible punching resistance for a given column size, slab
effective depth and concrete strength.
All four codes take the design yield strength of reinforcement as fyk/γs where γs=1.15. The following
sections describe the key features of these codes.
2.4.2 BS8110
The British code of practice for the design and construction of reinforced and prestressed concrete
structures, BS 8110:1997 - Structural use of concrete (BSI, 1997), has been superseded since 2010,
after almost 30 years of practice. Its predecessor was the CP110 – The structural use of reinforced
concrete in buildings (BSI, 1972), which had a fairly short life: introduced in 1972 and withdrawn
in 1985. BS8110 is divided into two parts: part 1 - Code of practice for design and construction,
and part 2 – Code of practice for special circumstances.
78
BS8110 allows slabs to be designed by the equivalent frame method, finite element analysis, or by
appropriate elastic analyses. For slabs without drop (scope of this research), the panel divisions are
assumed as presented in Figure 2.30, and the distribution of design moments between column and
middle strip is defined in percentage as 75:25 for hogging, and 55:45 for sagging moments. The
design moment transferred between slab and column is limited by = 0.15 .
The maximum shear stress in the slab around the column is limited to 0.8 . The nominal shear
stress is calculated at the basic control perimeter, and BS8110 adopts rectangular perimeters for
located at 1.5d from the column face, as shown in Figure 2.31. Locating at 1.5d is convenient as
the same vc term for beam shear can be used for flat slabs, as pointed out by Chana (1991).
Figure 2.30 – Division of panels in flat slabs according to BS8110.
Figure 2.31 – Control perimeter of BS8110.
BS8110 multiplies the design shear force by to account for the effects of uneven shear due to the
support reaction being eccentric. These values can be taken as = 1.15 for internal columns in
braced structures with approximately equal spans, = 1.25 for corner columns and edge columns
79
subject to a moment about an axis parallel to the free edge, and = 1.4 for edge columns subject to
a moment about an axis perpendicular to the free edge for approximately equal spans.
The punching resistance without shear reinforcement is given by (2.22).
100 400 (2.22)

, ( 8110) = 0.27 /
where , is the design shear stress for punching failure without shear reinforcement, =
/ ≤ 0.03 is the flexural reinforcement ratio, is the characteristic compressive concrete
cube strength and is the effective depth. has a design value of 1.25. BS8110 limits to 40
MPa in its design equations for shear, but this limit was not applied in the present study as it is
unduly restrictive.
BS8110 requires shear reinforcement to be provided in rectangular perimeters centred on the
column, as showed in Figure 2.7c, and the use of any shear reinforcement other than links is not
covered by the code. The required area of shear reinforcement is calculated in terms of the design
shear stress according to (2.23) and (2.24).
( − ) 0.4
≥ ≥ for ≤ 1.6 (2.23)
0.87 0.87
.
5(0.7 − )
≥ for 1.6 ≤ ≤2 (2.24)
0.87
.
where = / and , = , / . The required shear reinforcement ∑ . is provided

over at least two perimeters, of which the first perimeter should not contain less than 0.4 ∑ . .
The first perimeter of shear reinforcement is provided at 0.5d from the column face, with successive
perimeters positioned at spacings of 0.75d. The spacing of shear reinforcement around a perimeter
must not exceed 1.5d. Equations (2.23) and (2.24) are subsequently used to design the shear
reinforcement on successive square perimeters spaced at multiples of 0.75d from the basic control
perimeter. The control perimeter where shear reinforcement is no longer required, , is located at
1.5d from the outer perimeter of shear reinforcement, Figure 2.32.
80
Figure 2.32 - Control perimeter at which shear reinforcement is not required at BS8110.
2.4.3 Eurocode 2
Eurocode 2 and EC2 are common abbreviations for BS EN 1992, Eurocode 2: Design of concrete
structures. It is the current code within the European Union, and it has also been adopted by a
number of countries outside the Europe. It has four parts:
i. BS EN 1992-1-1:2004 – Design of concrete structures. General rules and rules for
buildings;
ii. BS EN 1992-1-2:2004 – Design of concrete structures. General rules. Structural
fire design;
iii. BS EN 1992-2:2005 – Design of concrete structures. Concrete bridges. Design and
detailing rules;
iv. BS EN 1992-3:2006 – Design of concrete structures. Liquid retaining and
containing structures.
Each country is expected to issue a National Annex, which allows them to choose values to suit
their local conditions such as wind, snow, earthquake, flooding etc. Annex I of EC2 provides
general recommendations regarding the design of flat slabs. It allows flat slabs to be designed using
a proven method of analysis such as Finite Element, Yield Line, Grillage, or Equivalent Frame. The
analysis with Equivalent Frame divides the slab in both plane directions into frames consisting of
columns and sections of the slab. The panels are divided into column and middle strips as presented
in Figure 2.30 for BS8110. The bending moments should be apportioned as shown in Table 2.1,
where negative and positive moments must always add up to 100%.
Table 2.1 – Simplified apportionment of bending moment for a flat slab. (BSI, 2004a)
Negative moments Positive moments
Column Strip 60 – 80 % 50 – 70%
Middle Strip 40 – 20% 50 – 30%
81
The main background for the EC2 provisions is considered to be the CEB-FIP Model Code 1990
(CEB-FIP, 1993). As oppose to the BS8110, and following the rules of CEB-FIP Model Code 1990
(CEB-FIP, 1993), EC2 locates the control perimeter at a distance 2d from the column face, as
shown in Figure 2.33.
Figure 2.33 – Control perimeter according to EC2.

The maximum shear stress in the slab around the column is limited to 0.3(1 – /250) / in
the UK National Annex to EC2 (BSI, 2004b), where is the partial factor for concrete which EC2
takes as 1.5. The punching resistance without shear reinforcement is given by (2.25).
200
, = 0.18(100 ) (1 + ) / (2.25)
.
where = . ≤ 0.02, in which and are the flexural tension reinforcement ratios
/ within a slab width equal to the column plus 3 to each side. For edge columns, EC2
recommends the effective width + presented in Figure 2.34a, though it also allows for +2
(Figure 2.34b), as commonly adopted in UK practice (ISE, 2006). is the characteristic concrete
cylinder strength, is the average effective depth of the tension reinforcement. The term (1 +
200/ ) represents the size effect which EC2 limits to a maximum of 2.0, with d in mm.
(a) (b)
Figure 2.34 - Effective width of a flat slab on edge column.
82
EC2 multiplies the design shear force by to account for the effects of uneven shear due to the
support reaction being eccentric. The design shear stress is given by (2.26).
= (2.26)
where the multiple accounts for the effects of uneven shear and is the basic control perimeter,
which is located at 2 from the column face. is calculated with Equation (2.27)
= 1+ (2.27)
where is a function of the proportions of the unbalanced moment transmitted by uneven shear and
by bending and torsion. EC2 presents a table (replicated here in Table 2.2) where values for are
presented according to the ratio between the column dimensions and . corresponds to the
assumption of shear distribution according to Figure 2.35, and is a function of the basic control
perimeter, as shown in Equation (2.28).
Table 2.2 – Values of for rectangular loaded areas. (BSI, 2004a)
/ ≤ 0.5 1.0 2.0 ≥ 3.0
0.45 0.6 0.70 0.80
Figure 2.35 – Shear distribution due to an unbalanced moment at a slab-internal column connection.
(BSI, 2004a).
= | | (2.28)
where is a length increment of the perimeter, and is the distance of from the axis about
which the moment acts. EC2 provides a few equations derived from (2.27) and (2.28) for usual
cases such as rectangular and circular internal columns.
The design philosophy of EC2 for punching at edge columns with normal moments is based on the
works of Regan (1981, 1999) who proposed the idealised moment-shear (M-V) interaction diagram
shown in Figure 2.36 for punching at edge columns subject to normal unbalanced moments. The
unbalanced moments in Figure 2.36 are calculated relative to the column centreline. According to
83
Regan (1981), the total moment at the inner column face Mcf is made up of a “component Mf
resisted by steel passing through the column face and two components each Mt resisted by steel
distributed within a width r on either side of the column. The components Mt are eventually
transmitted to the column through torsion on its side faces”.
Figure 2.36 - Interaction between punching resistance and unbalanced moment. (Regan, 1981)
Regan (1981) showed that, for practical purposes, the width r can be taken as the perpendicular
distance y from the slab edge to the inner column face. The increment from points C to B in bending
moment in Figure 2.36 is due to eccentric shear. The maximum shear resistance occurs at point A,
which corresponds to uniform shear at the column faces in contact with the slab (Regan, 1999).
According to Regan (1999), design for punching in flat slabs can normally be based on point B in
Figure 2.36, as tests on statically indeterminate flat slabs specimens indicate complete or nearly
complete development of flexural capacity at punching failure.
Where the eccentricity perpendicular to the slab edge is toward the interior and there is no
eccentricity parallel to the edge, EC2 considers shear stress to be uniformly distributed along the
reduced control perimeter ∗ depicted in Figure 2.37. This is equivalent to taking as / ∗ in
equation (2.26) and is intended to limit the maximum design shear force to the resistance at point B
in Figure 2.36, which eliminates the need to consider interaction between punching and unbalanced
moment. Where there are eccentricities in both orthogonal directions, is calculated according to
Equation (2.29)
= + (2.29)
∗
where is the eccentricity parallel to the slab edge resulting from a moment about an axis
perpendicular to the slab edge. EC2 also allows fixed values for for structures where adjacent
spans do not differ in length by more than 25% and the lateral stability does not depend on frame
84
action between the columns and slabs. In such cases, = 1.15 for internal column, = 1.4 for
edge columns and = 1.5 for corner columns.
Figure 2.37 - Reduced control perimeter of EC2.
EC2 requires the bending moment at the column face Mcf to be resisted by reinforcement centred on
the column within a width c2+y, where y is the perpendicular distance from the inner column face to
the slab edge. However, it is common UK practice (ISE, 2006; The Concrete Society, 2007) to
provide flexural reinforcement within = +2 and to limit to = 0.255( +
) / (where is the effective depth of top reinforcement normal to the slab edge and =
1.5) as required by Annex I (informative) of EC2 to prevent over reinforcement.
The required area of shear reinforcement is calculated according to equation (2.30).
− 0.75
1.5 ≥ (2.30)
,
where is the area of shear reinforcement in each perimeter, is the radial spacing of the shear
reinforcement and , = (250 + 0.25 ) ≤ , where , and are respectively the
effective design strength and the design yield strength of the shear reinforcement.
In addition to limiting the maximum shear stress around the column, EC2 limits the maximum
possible design shear force to VRdc where  is a Nationally Determined Parameter, with a
recommended value of 1.5 that is increased to 2 in the UK National Annex (BSI, 2004b).
Figure 6.22 in EC2 (replicated here in Figure 2.38) shows shear reinforcement being provided in
radial- or cross-type configurations. The radial spacing of the first perimeter of shear reinforcement
from the column must lie between 0.3d and 0.5d. The maximum radial spacing of successive
perimeters of shear reinforcement is 0.75d. The circumferential spacing of vertical legs of shear
85
reinforcement should not exceed 1.5d within the first control perimeter and 2d outside where ‘that
part of the perimeter is assumed to contribute to the shear capacity’. The control perimeter where
shear reinforcement is no longer needed, , is located at 1.5d from the outer perimeter of the
shear reinforcement, as shown in Figure 2.38, which also presents , , to be used for cross-type
arrangement.
Figure 2.38 – Control perimeter where shear reinforcement is no longer required. (k=1.5).
A few theories have been in development with the aim of introducing physical based punching
shear design equations for the EC2 updating, planned for the year 2020. The critical shear crack
theory of Muttoni (2008) and Fernández Ruiz and Muttoni (2009) is a strong candidate as it has
already been code formulated and introduced in the MC2010 (fib, 2013) recommendations, which is
described later in section 2.4.5.
2.4.4 ACI 318-14
The current code of practice for reinforced concrete structures in the US is the ACI 318 Building
Code Requirements for Structural Concrete (ACI, 2014). The ACI 318 was first implemented in
1971 and had only 750 provisions. Since then, the American Concrete Institute has introduced
significant changes to the code, with the last update had taken place in 2014.
The ACI code allows slabs to be designed by any procedure satisfying conditions of geometric
compatibility and equilibrium, if shown that every section has a resistance greater than or equal to
the required strength, and that serviceability conditions are satisfied.
Unlike EC2, ACI 318 adopts a rectangular control perimeter of length b0 and locates it at a much
closer distance from the perimeter of the concentrated load, 0.5d, presented in Figure 2.39. The
maximum shear stress is calculated as the greater of (2.31).
86
Figure 2.39 – ACI 318 control perimeter.
( ) = +
(2.31)
( ) = −
where is a property of the critical section analogous to polar moment of inertia. ACI-318 only
presents an equation for for an interior column, Equation (2.32). MacGregor and Wight (2005)
develops it for edge-column cases subject to moments about an axis parallel to the edge, (2.33);
edge-column cases subject to moments about an axis perpendicular to the edge, (2.34); and corner
columns, (2.35).
=2 + + 2( ) (2.32)
12 12 2
=2 + +( ) − + (2.33)
12 12 2
= + + 2( ) (2.34)
12 12
= + +( ) − + (2.35)
12 12 2
where is the length of the sides of the shear perimeter perpendicular to the axis of bending, is
the length of the sides of the shear perimeter parallel to the axis of bending, and are the
column dimensions perpendicular and parallel to the axis of bending, respectively. The dimensions
and are depicted in Figure 2.40 for interior and edge columns, which also shows the stress
distribution on the critical section. A proportion of the unbalanced moment about the
centroid of the critical perimeter is assumed to be resisted by flexure, with the remainder
resisted by eccentric shear, where =1− is given by (2.36).
1
=1− (2.36)
2
1+ 3 /
87
where and are the dimensions of the critical section measured parallel and perpendicular to
the slab edge.
On the basis of research by Moehle (1988), ACI 318-14 allows γv to be taken as 0 for edge columns
with unbalanced moments about an axis parallel to the slab edge provided V ≤ 0.75VR0 (where VR0 is
the shear resistance provided by concrete in the absence of unbalanced moment) and sufficient
flexural reinforcement is provided within a width of c2+3h, centred on the column to resist Mcg.
Taking γv = 0 and limiting the design shear resistance to 0.75VR0 is analogous to taking = / ∗
in EC2. The safety of this practice is questioned by Ghali et al. (2015). Consideration of other
design methods such as EC2 and MC2010 shows that ACI 318 is unique in not relating shear
resistance to the area of flexural reinforcement provided at the column.
Figure 2.40 - Critical perimeter of ACI 318 and corresponding shear stress distribution.
88
2.4.5 MC2010
The fib (Fédération internationale du béton) is a not-for-profit association dedicated to advancing

the economical, technical, aesthetic and environmental performance of concrete structures
worldwide. It was created in 1998 from the merger of the FIP - Fédération internationale de la
précontrainte and CEB - Comité euro-international du béton, and represents the latest European
thinking in reinforced and prestressed concrete structures. The fib Model Code for Concrete
Structures 2010 (fib, 2013), usually shorten to MC2010, introduces new developments and ideas
concerning concrete structures and structural materials, and is intended to provide the basis for
future codes.
MC2010 punching shear recommendations are based on the CSCT (refer to section 2.3.3), and so it
relates punching resistance to the rotation in the so-called critical shear crack. The basic control
perimeter is taken at a distance 0.5 from the column face, shown in Figure 2.41, where is
the effective depth for shear considering support penetration. For the majority of the studies carried
out in this research, = . The few exceptions are stated when fitting.
Figure 2.41 – Basic control perimeter according to MC2010 (fib, 2013).
MC2010 reduces the design shear resistance by the multiple to account for any eccentricity over
the support. MC2010 allows to be estimated as 0.9 for inner columns, 0.7 for edge columns and
0.65 for corner columns, for braced frames where the adjacent spans do not differ in length by more
than 25%. can also be calculated as a function of the moment transferred to the column,
according to Equation (2.37).
′
= 1/(1 + ) (2.37)
where ′ is the eccentricity of the shear forces resultant with respect to the centroid of the basic
control perimeter, as shown in Figure 2.42, and is the diameter of a circle with the same surface
as the region inside the basic control perimeter.
89
= +
(a) (b)
Figure 2.42 - Resultant of shear forces: (a) position with respect to the centroid of the supported
area; and (b) approximated basic control perimeter for calculation of the position of its centroid and
eccentricity between the resultant of shear forces and the centroid of the basic control perimeter.
Additionally, ke can be determined from LFEA as the ratio of the average to peak shear stress
( / ) on the basic control perimeter. This is equivalent to assuming punching failure occurs
when the peak stress on the control perimeter reaches the shear resistance.
The punching shear resistance is calculated as = , + , where the design shear resistance
attributed to the concrete, , is calculated according to equation (2.38).
, = (2.38)
in which is in megapascals (MPa). The parameter , accounting for the opening of the shear
critical crack and its roughness, depends on the rotations of the slab around the support region, and
is calculated according to equations (2.39) and (2.40).
1
= ≤ 0.6 (2.39)
1.5 + 0.9
32
= ≥ 0.75 (2.40)
16 +
where is the size of the maximum aggregate particles. For high strength and lightweight
concrete, the rupture may occur in the aggregate particles, in which case = 0.
The shear resistance provided by transverse reinforcement is calculated as (2.41).
, = (2.41)
90
where ∑ is the cross-sectional area of all shear reinforcement within the zone bounded by
0.35dv and dv from the border of the support region, Figure 2.43. σsw is the stress that can be
mobilised in the shear reinforcement, and is taken as (2.42).
Figure 2.43 – Shear reinforcement arrangement on section.
= 1+ ≤ (2.42)
6
where is the design yield strength of the shear reinforcement, is the shear reinforcement
diameter and is the bond strength. MC2010 recommends the value of = 3 MPa for
corrugated bars, but also allows to be calculated according to its subclause 6.1.3.2 (fib, 2013).
For the parametric studies carried out in Chapter 4 of this thesis, was taken as 4.5 MPa. The
maximum punching resistance is limited by crushing of the concrete struts near the support region,
according to Equation (2.43).
, = ≤ (2.43)
The coefficient accounts for the performance of punching shear reinforcing systems, and is
taken as 2.4 for stirrups and 2.8 for studs, provided the radial spacing to the first perimeter of shear
reinforcement from the column face is ≤ 0.5 and the spacing of successive perimeters of shear
reinforcement is < 0.6 . In the absence of specific data concerning the shear reinforcement,
MC2010 recommends = 2.0. The spacing of vertical legs of shear reinforcement around a
perimeter should not exceed 3 , where that part of the perimeter is assumed to contribute to the
shear capacity. The perimeter where shear reinforcement is no longer required, , is located at a
distance 0.5 from the outer perimeter of the shear reinforcement, as illustrated by Figure 2.44.
91
Figure 2.44 – MC2010 control perimeter where shear reinforcement is no longer required.
MC2010 has four levels of approximation (LoA), of which Levels I to III are intended for design
and Level IV for assessment. LoA I can be adopted for cases where the flat slab has regular layout
and is designed according to elastic analysis with no significant redistribution of internal forces. In
such cases, the slab rotation at failure is assumed to depend only on the shear slenderness / and
the flexural reinforcement yielding strain, as shown in Equation (2.44).
= 1.5 (2.44)
where denotes the position where the radial bending moment is zero with respect to the column
axis and can be approximated as 0.22 or 0.22 if 0.5 ≤ / ≤ 2.0, is the design yield
strength of the flexural reinforcement, and is the modulus of elasticity of reinforcement. LoA II
calculates the rotations as (2.45).
.
= (2.45)
where = 1.5, is the design average flexural strength per unit width of the support strip, and
is the average bending moment per unit width in the support strip, which is assumed to be of
width = 1.5 , . , ≤ , though where close to slab edges, is limited to according to
Figure 2.45. can be approximated, for each flexural reinforcement direction, as (2.46) for
concentrically loaded inner columns; (2.47) for edge columns considering tension reinforcement
parallel to the edge; (2.48) for edge columns considering tension reinforcement perpendicular to the
edge; and (2.49) for corner columns (tension reinforcement in each direction).
92
Figure 2.45 – Support strip widths.
1
= (2.46)
8
1
= + ≥ (2.47)
8 2 4
1
= + (2.48)
8
1 (2.49)
= + ≥
8 2
In LoA III, from Equation (2.45) can be taken as 1.2, provided that both and are
calculated with linear elastic finite-element analysis (LFEA), though should not be taken as less
than 0.67 at edge and corner columns. Although not stated in MC2010, to account for twisting
moments the reinforcement should be designed for the Wood moments (Wood, 1968) or equivalent.
Level IV of approximation (LoA IV) allows the rotation to be calculated with a nonlinear finite
element analysis of the structure, accounting for tension stiffening, cracking, yielding of
reinforcement, and any other significant non-linear effect that would help improving the simulation
of the structure and its behaviour.
93
2.5 CONCLUSIONS
The punching shear resistance of flat slabs is influenced by many different factors, from concrete
and reinforcement parameters, to more complex issues like size effect, slenderness, and boundary
conditions (test set up). It is somewhat understandable that even after more than 100 years of
research, and 50 years since the first rational model, there is still no consensus regarding a general
theory.
The number of models available developed from different backgrounds (empirical control surface
approaches, plasticity approaches, Kinnunen and Nylander (1960)-like models, etc.), and dating as
early as the 1910s (Talbot, 1913), and as late as the 2010s (Broms, 2016; Hoang and Pop, 2016),
shows that it is a very active field of research, and attempts are still being made in this regard.
The complexity of the phenomenon is reflected in the conflicting opinions expressed by researchers.
For instance, depending on the test arrangement, conclusions can become conflicted regarding the
influence of continuity, with works such as Sherif and Dilger (2000a) observing no difference
between isolated and continuous specimens, whereas Chana and Desai (1992a) found an increase of
around 40% on punching resistance from extending the panel beyond the line of contraflexure.
Issues like this one led fib (2001) to remark that, for design purposes, it is still not recommended to
rely on compressive membrane action.
Another example is related to the shear reinforcement layout. Einpaul et al. (2016b) observed that
cruciform and radial arrangements showed comparable performances regarding the punching
strength as oppose to Vollum et al. (2010), who concluded that cruciform layout increases the shear
strength of flat slabs by a multiple up to 1.5, in comparison to multiples of 2.0 or more for well
anchored radial arrangements.
Some of the assumptions of the different mechanical models discussed in section 2.3 differ
significantly, while presenting reasonable predictions for punching shear capacity. Even models
based on similar backgrounds such as the CSCT by Muttoni (2008) and the TST by Broms (2016)
can present some discrepancy, like the impact of time dependent effects such as creep and drying
shrinkage, discussed in the end of section 2.3.4.
The CSCT by Muttoni (2008) and Fernández Ruiz and Muttoni (2009) was chosen as the physical
theory behind the latest MC2010 (fib, 2013). Since its introduction, the model has been tested and
extrapolated to be applied over different situations (non-axis-symmetric slabs, steel fibre reinforced
slabs, prestressing, post-installed shear reinforcement, continuity, etc.). However, the model is still
far from being widely accepted, and works such as Ferreira et al. (2014) and Broms (2016) have
already identified and discussed a few issues with the theory.
94
For instance, MC2010 LoA I to III do not consider the effect of moment redistribution on punching
resistance. Consequently, they can give very conservative estimates of the shear resistance of slabs
designed in accordance to UK practice using the equivalent frame method, where moments are
typically redistributed downwards by 20% over columns. Both tests and the CSCT suggest that this
inward movement of the line of radial contraflexure should lead to an increase in shear resistance.
This increase is not presently accounted for in levels I to III of MC2010 which consequently
underestimate the punching resistance of such slabs.
This research addresses a few of these theoretical issues while giving design considerations.
Chapter 4 presents a few analysis and parametric studies using test data, designed slabs with and
without moment redistribution, and NLFEA, assessing the potential benefit of the surrounding
structure on flat slabs on internal column. Chapter 5 applies a similar methodology focusing on flat
slabs on edge columns, where the influence of eccentricity M/V and reinforcement arrangement
between span and support are also investigated. Chapter 6 considers the influence of slab continuity
on the punching strength of realistically proportioned flat slab plates.
95
96
3 Finite Element Modelling
3.1 INTRODUCTION
This chapter is concerned with listing and detailing the structural elements, material modelling
assumptions and solution procedure used for the Finite Element Analysis. The research used the
commercially available finite element codes Diana v9.6 (TNO Diana, 2014) and Atena v5.1.2
(Červenka et al., 2016). Diana is a finite element software dedicated to a wide range of applications
in civil engineering including structural engineering, and Atena is a software for nonlinear analysis
of reinforced concrete. The main goal for the FE analysis was to provide an insight on the deformed
shape and failure load of the analysed flat slabs, which implies that the model must be calibrated
before any study is performed. These calibrations were carried out with experimental punching
shear tests available in the literature and described in detail within their corresponding chapters as
appropriate. Due to the numerous available concrete and steel models, solution procedures,
boundary conditions, mesh size, and loading type possibilities, these validations were usually time
consuming and fairly specific for each case. Chapters that involve FE analysis will present their
own, or refer to similar sensitivity studies and calibrations. Conclusions drawn from the chapters
concerning the NLFEA modelling should be considered only as a guideline for similar models, as
some of the FE assumptions are shown to be case dependent.
3.2 DIANA
The following section describes the elements and modelling procedure adopted for the FE analysis
with Diana with theory as described in TNO Diana (2014).
3.2.1 Structural Elements
All the structural elements used for the Diana analyses in this research are listed and detailed as
follows.
3.2.1.1 Curved Shells

Diana bases its curved shell elements on isoparametric degenerated-solid approach introducing the
shell hypotheses of: straight-normal, which includes the Mindlin-Reissner theory (Mindlin, 1951)
for shear deformation and assumes that normal remains straight, but not necessarily normal to the
97
Finite Element Modelling
reference surface; and zero-normal-stress, that forces the normal stress component in the normal
direction of a lamina basis to zero.
Each node of a curved shell element has five degrees of freedom: three translations and two
rotations, Figure 3.1. The global strains are calculated according to (3.1), from which the Cauchy
stresses are derived, Figure 3.2. Figure 3.3 presents the generalized moments and forces along with
the adopted sign convention.
Figure 3.1 – Degrees of freedom on a Curved shell.

= = =
(3.1)
= + = + = +
=0
=
=
=
=
Figure 3.2 – Cauchy stresses in Diana’s curved shell.
=
=
= =
Figure 3.3 – Positive convention for generalized Moments and Forces in Diana’s curved shell.
98
Three different types of curved shells are used in this thesis. They are presented here with their
respective denominations from TNO Diana (2014).
i. CT30S is a six-node triangular isoparametric curved shell based on a quadratic
interpolation and area integration. The element is depicted in Figure 3.4. In order to avoid
membrane and shear locking, an integration scheme of 3 over the ξη area should be used,
and the ζ direction allows values greater than or equal to 2.
Figure 3.4 - Six-node triangular curved shell CT30S.
ii. Q20SH is a four-node quadrilateral isoparametric curved shell based on a linear

interpolation and Gauss integration. The element is depicted in Figure 3.5. The only
possible integration scheme in the ξη plan is 2x2, whereas the ζ direction allows values
greater than or equal to 2.
Figure 3.5 - Four-node curved shell Q20SH.
iii. CQ40S is an eight-node quadrilateral isoparametric curved shell based on a quadratic

interpolation and Gauss integration over its plan sides. The element is depicted in Figure
3.6. Similarly to CT30S, an integration scheme of 3x3 is recommended for the ξη plan to
avoid membrane and shear locking, whereas the ζ direction allows values greater than or
equal to 2.
Figure 3.6 - Eight-node curved shell CQ40S.
99
3.2.1.2 Bricks
Bricks are three-dimensional general purpose elements. Due to their characteristic lack of
simplifications, they tend to generate large system of equations and consequently require
significantly greater computing effort than shell elements. As a consequence, bricks are usually
used where other elements would produce inaccurate results.
Three translational degrees of freedom are available in each node of a brick, as presented in Figure
3.7. The global strains are calculated as shown in (3.1), and the derived Cauchy stresses are
presented in Figure 3.8.
Figure 3.7 - Degrees of freedom on a Brick.
= =
=
=
Figure 3.8 - Cauchy stresses in Diana’s solid elements.
The following two types of solid elements are used with Diana in this thesis:
i. HX24L is an eight-node isoparametric solid brick element based on a linear interpolation
and Gauss integration. The element is depicted in Figure 3.9. It allows an integration
scheme as low as 1x1x1 and not greater than 2x2x2.
100
Figure 3.9 - Eight-node isoparametric solid brick HX24L
ii. CHX60 is a twenty-node isoparametric solid brick element based on a quadratic

interpolation and Gauss integration. The element is depicted in Figure 3.10. It allows an
integration scheme as low as 2x2x2 and not greater than 3x3x3.
Figure 3.10 - Twenty-node isoparametric solid brick CHX60.
3.2.1.3 Composed Solid Elements

Composed solid elements are used to calculate generalized moments and forces from the primary
Cauchy stresses of a regular solid element. They do not have mechanical properties of their own,
therefore, they do not interfere on the behaviour of the FE model.
Composed elements can be specified over a reference plane, and each one of their base elements
will have a defined thickness (composition). Diana calculates generalized moments and forces by
integration in the local z direction over the elements in the respective composition as depicted in
Figure 3.11, which shows a composed solid element.
101
Figure 3.11 – Composed solid element in Diana.
Only one composed solid element is used in this study:

i. CQ8CM is an eight-node quadrilateral curved base element, which must be combined
with a composition of CHX60 (refer to previous section) to form a composed solid
element. The element is shown in Figure 3.12. The integration scheme must follow that
of the solid element.
Figure 3.12 – Eight-node quadrilateral curved base element.
3.2.2 Crack modelling in NLFEA
Cracking in NLFEA of quasi-brittle materials such as concrete is usually modelled with either the
discrete cracking method or the smeared cracking method. The discrete cracking method introduces
a gap into the mesh. It is more realistic and complicated to implement than the smeared approach,
which smears the crack within the element, treated as a continuum. The smeared approach is
currently more often than not adopted for concrete analysis.
3.2.2.1 Discrete crack method

The discrete cracking approach was first introduced for the analysis of concrete structures by Ngo
and Scordelis (1967) and is based on fracture mechanics. It is more realistic and theoretically more
102
suitable to model the failure localisation of concrete structures, since the crack is treated as a
geometric entity, and it is allowed to propagate by splitting nodes as they exceed a particular tensile
strength criterion, Figure 3.13. This requires cracks to propagate along element boundaries, which is
unrealistic, and demands re-mesh techniques leading to computationally challenging analysis. The
issue of change in topology can be dealt with by introducing meshless methods, like the element-
free Galerking method (Belytschko et al., 1994), for instance, or the extended finite element method
(XFEM) based on the smeared crack approach of Hillerborg et al. (1976) and the concept of
partition unity by Belytschko and Black (1999), amongst others. However, discrete crack models
are still generally adopted for specific fracture mechanics problems in very localised areas.
Figure 3.13 – Node splitting at Discrete cracking method.
Geometric discontinuities, such as discrete cracks, can be modelled in Diana using its multipurpose
structural elements. The constitutive law is based on a total deformation theory, which expresses the
tractions as a function of the crack width (Δω) and slip (Δs), and the total relative displacements
(Figure 3.14) with a nonlinear relationship assumed between both normal traction and crack width
(σ/Δω), and shear traction and crack slip (τ/Δs).
Figure 3.14 – Discrete cracking and rough crack.
103
As can be seen in equation (3.2) for the discrete crack model in Diana, normal and shear stresses are
uncoupled ( = 0 for ≠ ). The crack formation is initially governed by the tension criteria.
Once the maximum tensile stress is reached, the effect of tension softening is applied for normal
stresses while a constant shear retention is assumed in shear.
=
=0
(3.2)
=0
=
3.2.2.2 Smeared crack method

The smeared crack method was introduced as a tool to simulate concrete fracture by Rashid (1968).
It is the most commonly adopted model for NLFEA of reinforced concrete structures. As oppose to
the discrete method, the material is assumed to remain a continuum after cracking. Besides a fixed
mesh topology during crack formation and propagation, the method has also the advantage of
having an unrestricted orientation of the crack plane. It is usually formulated on the basis of either a
strain-decomposition or a total strain concept. The strain-decomposition method divides the total
strain into two components: the strain at the crack, and the strain in the concrete between cracks.
This division allows the model to combine different constitutive models such as elastic, plastic or
visco-elastic.
The total strain method calculates the stresses in terms of total strains. The crack inclination can be
assumed to be fully fixed at crack opening, or to rotate according to the direction of the major
principal stress. It is also possible to adopt a hybrid crack approach like the multi-directional fixed
crack model introduced by Rots and Blaauwendraad (1989), in which a new crack is formed
whenever the angle of inclination between the existing crack and the current direction of principal
stress exceeds the value of a user-specified threshold angle. The next section will discuss the two
main models.
ROTATING CRACK MODEL
Cope et al. (1980) introduced the rotating crack model in which the stress-strain relationships are
evaluated in the principal directions of the strain vector. A coaxiality between crack axes and
principal axes is imposed in the model, hence no shear retention factor is required, unlike the fixed
crack model. According to Vechio (2000) and Maekawa et al. (2003), this assumption provides
reasonable predictions with the exception for cases where shear slip and shear transfer along crack
is predominant. Some numerical studies on flat slabs have shown good results with the rotating
smeared crack method such as Nogueira (2011) and Mamede et al. (2013), among others.
104
FIXED CRACK MODEL

The stress-strain relationship for smeared cracking (Equation (3.3) has been traditionally set up with
reference to fixed principal axes of orthotropy, depicted here as n, s, and t, where n is the direction
normal to the crack, and s, t the directions tangential to the crack.
0 0 0
0 0 0
0 0 0
= (3.3)
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
Rots and Blaauwendraad (1989) discuss that earlier versions of the fixed crack method such as the
ones presented by Rashid (1968), Červenka (1970), and Valliappan and Doolan (1971) would adopt
the crude assumption that crack normal stress ( ) and shear crack stresses ( and ) abruptly
drop to zero upon crack formation, by setting = = = = = 0. This is a
simplistic assumption as cracks are capable of transmitting stresses through a few mechanisms such
as aggregate interlock. Furthermore, Schnobrich (1972) stated that the sudden change from an
initial isotropic linearly-elastic relationship to an orthotropic law with zero stiffness moduli led to
numerical difficulties by creating strong discontinuities. As a consequence, models were developed
in which the initial isotropic moduli were maintained but with some reduction after initial cracking.
Suidan and Schnobrich (1973) introduced the concept of shear retention factor by reinserting
and as reduced values of the initial linear-elastic shear modulus G. This parameter will be
discussed in more detail in section 3.2.5.
NLFEA of flat slabs have been carried out with both rotating (Nogueira, 2011; Mamede et al., 2013,
etc) and fixed smeared crack methods (Eder et al., 2010, Eder et al. 2012 etc) presenting good
results in both cases, which highlights the applicability of both methods as well as the case
dependency of numerical analysis.
3.2.3 Compressive Behaviour
3.2.3.1 General remarks

The compressive behaviour of concrete depends in a complex manner on the multiaxial stress state.
Under uniaxial compression, no cracks initiate in the matrix up to 30-50% of the failure stress,
although by then shear-bond cracks have already developed in the region of coarse aggregate.
Between 50% and 75%, cracks starts to develop rapidly in the mortar among aggregate particles. At
around 75% of the failure stress the cracks tends to become unstable i.e. their length increases even
if the applied stress is constant, and when the internal energy exceeds the crack-release energy, the
105
whole concrete system becomes unstable with both mortar and bond cracks propagating at a
significant rate, leading to failure of the test specimen. After reaching the failure stress, the concrete
demonstrates a residual descending strength in the form of a compression softening as shown in
Figure 3.15, which depicts the uniaxial stress-strain curve of a concrete.
Figure 3.15 – Characteristic uniaxial stress-strain curve.
For biaxial compression behaviour, Kupfer et al. (1969) observe that due to past difficulties in
simulating a proper biaxial state of stress, studies reporting increases in strength of up to 350% in
comparison to an equivalent uniaxial test can be found. However, Kupfer et al. (1969) improved test
arrangement presents a more reasonable and more acceptable increase value of up to 27% in some
cases, and approximately 16% for equal compressive stresses in two principal direction. On the
other hand, biaxial tension does not seem to have any influence on concrete tensile strength. Kupfer
and Gerstle (1973) presents a graph (Figure 3.16) showing a nearly linear relation between biaxial
tension and compression as well as the observed increase of biaxial over uniaxial test.
Richard et al. (1928) and Balmer (1949) conducted tests with low to high levels of confinement,
observing a shift in concrete behaviour from quasi-brittle, to plastic-softening or even plastic-
hardening, dependent on the level of confining stress (Figure 3.17). A significant increase in axial
strength and ductility is observed with increasing confining pressure, with failure loads of around
20 times that of an equivalent uniaxial test for cases with high confining stresses (refer to =
= −170 in Figure 3.17b).
106
Figure 3.16 – Biaxial test results for concrete (Kupfer and Gerstle, 1973).
(a) (b)
Figure 3.17 – Triaxial stress-strain relation from test of (a) Richard et al. (1928) and (b) Balmer
(1949).
Figure 3.18 shows a representation of a three-dimensional principal-stress state of a concrete under

triaxial loading. Three stages represented as surfaces can be distinguished: the start of stable crack
propagation; the start of unstable crack propagation; and the failure limit.
107
Figure 3.18 – Three-dimensional representation of triaxial failure surface for concrete (Chen, 1982).
3.2.3.2 Compressive Behaviour models

All the Diana NLFEA undertaken in this research utilised the ‘Total Strain Fixed crack model’, on
the basis of studies by Sagaseta (2008), Eder (2011) and Fang (2014) showing it to give good
results for shear dominant reinforced concrete analysis, and sensitivity studies presented in sections
5.4.1.1 and 5.7.1.1. In the total strain based constitutive model, the stress is described as a function
of strain and this stress-strain relationship is evaluated in a coordinate system fixed at first cracking,
i.e. the strain transformation matrix T in (3.4) is independent of the principal directions of the
current strain vector εxyz in the element.
∆
. = . + .
∆
. = .
(3.4)
∆
. = ( . )
∆
. = .
where xyz is the element coordinate system, Δε is the strain increment, σ is the stress vector in
the element, and nst is the crack coordinate system. The model is defined by the basic properties
such as Young’s modulus, Poisson’s ratio, etc. and the input for concrete behaviour under
compression, tension and shear, which must be described with one stress-strain relationship.
Diana offers a few predefined functions to describe the behaviour of concrete under compression,
varying from a simple Elastic diagram to a more elaborate fracture energy based Parabolic curve.
Thorenfeldt et al. (1987) presented a hardening and softening curve to express the behaviour of
108
concrete in compression. The curve is shown in Figure 3.19 and its correspondent equation in
(3.5).
Figure 3.19 - Thorenfeldt compression curve.
=− (3.5)
− 1−
with,
1 < <0
= 0.80 + and = (3.6)
17 0.67 + ≤
17
is the peak compressive stress, is the compressive strength and is the peak compressive
strain.
The parabolic curve is fracture energy based and defined by the compressive fracture energy Gc,
assumed to be 50 times the tensile fracture energy, and the crack bandwidth h of the element, which
is assumed to be related to the square root of the element area. Figure 3.20 presents a general graph
to describe the compressive behaviour when a parabolic curve is adopted.
109
Figure 3.20 - Parabolic compression curve in Diana.
3.2.3.3 Compressive behaviour with Lateral Confinement

The strength and ductility are significantly influenced by the level of confinement as previously
discussed (refer to section 3.2.3.1) and illustrated in Figure 3.21. The benefit of lateral confinement
is modelled in Diana using the four parameter Hsieh-Ting-Chen (Chen, 1982) failure surface to
account for the increase in strength due to an increase in isotropic stress according to (3.7).
Figure 3.21 – Stress-strain relation of plain concrete and confined concrete (Binici, 2005).
= 2.0108 + 0.9714 + 9.1412 + 0.2312 −1=0
1 (3.7)
= − + − + −
6
= + +
where is the maximum principal stress, is the compressive strength, and is the stress in
the concrete. Equation (3.7) provides a gradual increase of peak strength in confined compression
based on the linear slope, varying from an unconfined situation to a full triaxial case where the
initial stress-strain relationship is kept and the failure surface is not reached, Figure 3.22.
110
Figure 3.22 – Influence of lateral confinement on compressive stress-strain (TNO Diana, 2014).
3.2.3.4 Compressive behaviour with lateral cracking

Under a biaxial compression-tension stress state the concrete compressive strength is reduced, when
compared to an equivalent uniaxial compression test. If the material is cracked in the lateral
direction, and from Figure 3.19 and Equation (3.5) are reduced by the factors βσcr and βεcr
respectively to account for that compression softening. This stiffness reduction due to lateral
cracking is modelled in accordance to Vecchio and Collins (1993) study, who proposed (3.8):
1
= ≤1 (3.8)
1+
where = 0.27 − − 0.37 , = −( )

, βεcr is taken as 1, is the average lateral
damage, and the relationship between and − is extracted from Figure 3.23.
Figure 3.23 – Reduction factor due to lateral cracking according to Vecchio and Collins (1993).
111
3.2.3.5 Lateral expansion effects due to Poisson’s Ratio

A specimen loaded with uniaxial compressive or tensile forces is subject to lateral displacements
due to the Poisson effect. When these displacements are restrained, a passive lateral confinement
acts on the specimen. Diana models this effect following the works of Selby and Vecchio (1993),
who defined a pre-strain concept in which the lateral expansion effects are considered with
additional external loading on the specimen. For NLFEA, the stress vector in the principal
coordinate system, presented in Equation (3.9), is evaluated in terms of the equivalent uniaxial
strain vector instead of the principal strain vector . is calculated from the Poisson’s
ratio and .
. . (3.9)
= ( )
3.2.4 Tension Softening
3.2.4.1 General aspects

Concrete is a strain softening material in both compression and tension. A softening zone following
the peak stress is formed around the crack tips, which significantly influence the stress
redistribution. Similarly to the behaviour in compression, Diana offers a number of diagrams to
describe the concrete tension softening.
3.2.4.2 Tension softening models

Two different options were used in the NLFEA carried out in this research. A Linear diagram
(Figure 3.24) was used for the cases in which the slab was modelled with shell elements, following
the recommendations of Vollum and Tay (2007), who studied tension stiffening of concrete
structures modelled with curved shell elements and shear deformation described by the Mindlin-
Reissner theory (Mindlin, 1951).
Vollum and Tay (2007) concluded that a linear tension softening stress-strain relationship for
cracked concrete gives similar instantaneous curvatures to EC2 provided that ft and εu are chosen
accordingly, with their recommendation being ft = 0.5fct and εu = 0.5 εsy.
112
Figure 3.24 – Linear diagram for tension softening.
The Hordijk diagram (Cornelissen et al., 1986; Hordijk, 1991), shown in Figure 3.25 and described
in equation (3.10), was used for analyses where the slab was modelled with brick elements.
Figure 3.25 - Hordijk diagram for tension softening.
Δu Δu
1+ c exp −c …
(Δu ) Δu , Δu ,
= Δu (3.10)
− (1 + c )exp(−c ) if 0 < Δu < Δu ,
Δu ,
0 if Δu , < Δu < ∞
where the parameters = 3 and = 6.93, the ultimate crack strain calculated as Δu , =
5.136× , and G is the fracture energy.
3.2.5 Behaviour in Shear
Diana adopts the concept of a shear retention factor in its Total Strain Fixed crack models to
account for the reduction in shear stiffness after cracking. Equation (3.11) presents β reducing the
shear modulus G.
113
= , with 0 ≤ ≤1 (3.11)
β can be calculated in varies way in Diana. This research used two methods chosen on the basis of
sensitivity studies to give good predictions of experimental results (sections 4.3.2.4, 5.4.1.3, and
5.7.1.3). The choice of depends on a number of factors, including concrete constitutive model and
reinforcement arrangement. For some cases, a constant shear retention factor shown in Figure 3.26
was used. Suidan and Schnobrich (1973) suggests that should lie around 0.1~0.2, which gives
good results for shear dominant failure cases as shown in Sagaseta (2008).
Figure 3.26 – Constant shear retention for Total Strain crack models.
However, experimental studies suggest that shear stiffness is not constant, and its reduction is
related to the increase of crack width. Figure 3.27 presents a graph developed by Sagaseta (2008)
showing a comparison between studies carried out by Figueiras (1983), Rots and Blaaunwendraad
(1989) and Červenka et al. (2002) where the shear retention factor is shown to reduce with
increasing normal strain at the crack. Thus, in some other cases β was calculated in relation to the
mean aggregate size dagg according to (3.12).
Figure 3.27 – Shear retention factor according to different models (Sagaseta, 2008)
114
2
= 1−( ) ℎ, with 0 ≤ ≤1 (3.12)
where is the crack normal strain and ℎ the crack bandwidth calculated as ℎ = √ for solid
elements and ℎ = √ for two-dimensional elements, where is the element volume and is the
total area of the element.
3.2.6 Reinforcement modelling
Diana offers two types of embedded reinforcement to add stiffness to the finite element model
depicted as: (a) grids, where the total area of the reinforcement is divided in several particles, which
contributes to the stiffness of the respective element they embed, Figure 3.28; and (b) bars, which
are line-shaped reinforcement (excluding here point-shaped bars embedded in plane strain and
axisymmetric elements) with their total length divided into several particles, Figure 3.29.
Figure 3.28 – Grid reinforcement in Diana.
Figure 3.29 – Bar reinforcement in Diana.
Reinforcements can be assumed to be fully embedded (perfect bond) in their concrete elements, in
which case displacements and strains are fully coupled, or alternatively, modelled with bond-slip
properties. They do not have their own degrees of freedom and are defined by:
i. location in the model;
115
ii. material properties, which can be chosen from: linear elasticity, Von Mises
plasticity with optionally hardening diagrams; Monti-Nuti model for cyclic loading
conditions; and user-supplied subroutines;
iii. physical properties such as cross-sectional area;
iv. integration point schemes.
This study used mostly discrete bars embedded in curved shell elements, Figure 3.30 and solid
elements, Figure 3.31. For reinforcement in curved shells, two conditions must be satisfied: each
bar section must intersect one or two element edges, but none of them more than once; and all
computed location points for each bar must be inside the defined thickness of the shell.
Figure 3.30 – Bars in curved shell elements.
Figure 3.31 – Bars in Solid elements.
3.2.7 Iterative solution algorithms
NLFEA usually requires incremental methods to achieve an equilibrium path. The analysis uses an
already established equilibrium state to determine the following one along the equilibrium path.
Diana offers several options for incremental-iterative procedures. For the internal column NLFEA
carried out in Chapter 4, a regular Newton-Raphson method for iterative procedure presented
reasonable results concerning the deflected shape (rotations) of the specimens. For the Newton-
Raphson method, the stiffness matrix relation presented in (3.13) is updated at every iteration, i.e.
116
(3.14) is based on the last known or predicted situation, although that might not be an equilibrium
state. Figure 3.32 shows a graph illustrating the iterative procedure.
g
= (3.13)
Δu
= g (3.14)
Figure 3.32 – Newton-Raphson iteration procedure.
where is the stiffness matrix, and represents a linearized form of the relation between the force
and displacement vectors, Δu is the total displacement increment, is the iterative displacement
increment and g is the out-of-balance force vector at the start of iteration i.
Better results for the NLFEA of Chapters 5 and 6 were found by adopting a Quasi-Newton
(normally referred to as ‘Secant’) iterative procedure method, based on the recommendations of
Vollum et al. (2010) and Eder et al. (2012), who carried out similar studies on punching shear using
Diana. As opposed to the Newton-Raphson, the Secant method does not update the stiffness matrix
on each iteration, setting it up at the start of each step, using the information of previous out-of-
balance forces vectors and solution vectors throughout the increments. Bathe and Cimento (1980)
describe it as a method that provides a compromise between the full update of the stiffness matrix
performed in the full Newton-Raphson method, and the use of a stiffness matrix from a previous
configurations as performed in the Modified Newton-Raphson method.
Figure 3.33 illustrates the Quasi-Newton iteration procedure, whereas (3.15) shows its matrix
relation.
117
Figure 3.33 – Quasi-newton (Secant) iteration procedure.
= (3.15)
Equation (3.16) is known as the BFGS for being attributed to Broyden, Fletcher, Goldfarb and
Shanno. The iterative displacements is calculated by substituting (3.16) in (3.14). It was adopted
for all the Quasi-Newton NLFEA in this research.
= I+ I− + (3.16)
Also, all the NLFEA used an energy based convergence criterion with the energy ratio being
calculated as (3.17).
(f , +f , )
Energy norm ratio = (3.17)
Δ (f , +f , )
The chosen value for the tolerance in the majority of the analysis was 10-3, meaning that the
iteration process stops and the following step starts when (3.17) becomes less than 10-3. The few
exceptions for tolerance values are stated when fitting.
A Load control incremental process illustrated in Figure 3.34 was applied for all the analysis carried
out in this thesis.
118
Figure 3.34 – Load Control incremental process.
3.3 ATENA
The following will give an insight on the assumptions adopted for the FE analysis with Atena as
described in Červenka et al. (2016).
3.3.1 Structural Elements
Atena offers three groups of elements: plane elements for 2D, 3D and axisymmetric analysis; solid
3D elements; and special elements, used for modelling springs, gaps, external cable etc. Only 3D
solid elements were used with Atena in this research. The elements are constructed using
isoparametric formulation with linear and/or quadratic interpolation functions, integrated by Gauss
integration scheme that ensures n(n-1) order accuracy, with n being the degree of the polynomial
used to estimate the integrated function. The incremental strains are calculated according to (3.18),
where , , ∈< 1 … 3 . The displacement derivatives are presented in (3.19).
. ()
1 . () . ()
1 ∆ ( ). () ∆ ( ). () 1 . () . ()
= , + , + , , + , , + , + ,
(3.18)
2 2 2
∆ () ∆ ( )
. ()
( . − . )
, =
.
(3.19)
∆ ( )
∆ ( ) .
, =
.
119
3.3.1.1 Bricks
Atena offers the option of choosing between an eight-node and a twenty-node Brick element,
depicted in the input as CCIsoBrick<xxxxxxxx>, where the string in < > states the present element
nodes. The element is illustrated in Figure 3.35.
CCIsoBrick<xxxxxxxx>
CCIsoBrick<xxxxxxx…xxx>
Figure 3.35 – Eight-node/Twenty-node isoparametric solid brick. Adapted from Červenka et al.
(2016).
3.3.1.2 Tetrahedral
Two options of tetrahedral elements are available in Atena: a four-node and a ten-node. Figure 3.36
depicts that element, which has an input of CCIsoTetra<xxxxxx>, where the string in < > defines
the number of element nodes.
CCIsoTetra<xxxx>
CCIsoTetra<xxxxxxxxxx>
Figure 3.36 - Four-node/Ten-node isoparametric solid wedge. Adapted from Červenka et al. (2016).
3.3.2 Material Modelling Formulation
Concrete was modelled using the fracture-plastic model CC3DNonLinCementitious2 which

combines constitutive models for tensile (fracturing) and compressive (plastic) behaviour. In
accordance to the work of De Borst (1986), the material formulation is based on the strain
decomposition into elastic, plastic and fracturing components, as shown in (3.20), and the new
stress is calculated by equation (3.21).
120
= + + (3.20)
= + (Δ −Δ −Δ ) (3.21)
All analyses used the Atena ‘Fixed crack model’ in which strains and stresses are converted into the
material directions given by the principal directions at the onset of cracking. However, studies such
as Mamede et al. (2013) obtained good predictions for punching shear using Atena’s rotated crack
model. Section 5.12.1.2 presents a sensitivity study regarding the choice of smeared crack method
with Atena.
3.3.3 Model for Concrete Cracking
The Rankine criterion, presented in (3.22), is adopted for concrete cracking in conjunction with an
exponential softening curve for tensile softening, shown in Figure 3.37. For fixed crack models, the
strains and stresses are converted into the principal directions at the onset of cracking.
= − ≤0 (3.22)
where identifies the trial stress and is the tensile strength in the material direction i.
Figure 3.37 – Tensile softening in Atena. (Červenka et al., 2016).
The crack band size Lt is a projection of the element size into the crack direction (refer to Figure
3.37). The trial stress is computed according to the elastic portion of (3.21), and checked against
(3.22). When does not satisfy (3.22), the final stress must satisfy (3.23).
= − = − Δ − =0 (3.23)
Atena’s current formulation assumes that there is no interaction between shear and normal
components. The crack tensor is calculated by (3.24).
= 0 for ≠ and ≠ (3.24)
121
Mode I crack stiffness, which refers to the crack propagation due to tensile stresses normal to the
plane of the crack, is calculated by (3.25), whereas modes II – sliding mode due to shear stresses
acting perpendicular to the crack front and parallel to the crack – and III – tearing mode due to
shear stresses acting parallel to the crack front and the plane of the crack – are calculated with
(3.26).
( )
= , (no summation of indices) (3.25)
= min( , ), (no summation of indices) (3.26)
where ≠ , and is a shear factor coefficient that defines the relationship between shear and
normal crack stiffness. The default value is = 20, which was adopted for all the analysis with
Atena. The shear strength of cracked concrete was determined in accordance with the ‘Modified
Compression Field Theory’ of Vecchio and Collins (1986), shown in (3.27).
0.18
≤ , ≠ (3.27)
24
0.31 +
+ 16
where is the maximum aggregate size, is the maximum crack width at the given location and
is the compressive strength.
3.3.4 Concrete Crushing
The plastic model in Atena calculates the new stress state using the predictor-corrector equation in
(3.28).
( ) ( )
. = . + (Δ −Δ )= − Δ = − (3.28)
is computed by the return mapping algorithm presented in equation (3.29).
− = −Δ =0 (3.29)
is the return direction defined as (3.30).

∂ ( ) ∂ ( )
= then Δ =Δ (3.30)
∂ ∂
where ( ) is the plastic potential function. The return direction is determined by evaluating the
derivative of ( ) at the predictor stress state .
The Menétrey-Willam failure surface, presented in (3.31) and (3.32), was used for concrete in
compression with a hardening/softening law based on the uniaxial compressive test. The ascending
122
branch of the hardening curve is elliptical and strain based, described by equation (3.33), but the
softening branch is linear and displacement based to introduce mesh objectivity into the finite
element solution. Figure 3.38 shows the hardening/softening graph.
= √1.5 + ( , )+ − =0 (3.31)
√6 √3
where,
−
=3
+1
(3.32)
4(1 − ) cos + (2 − 1 )
( , )=
2(1 − ) cos + (2 − 1) 4(1 − ) cos +5 −4
is the compressive and is the tensile strength, ( , , ) are the Heigh-Vestergaard coordinates
and is a parameter that defines the roundness of the failure surface and it varies from 0.5 to 1.0.
The failure surface position can shift along the hydrostatic axis, simulating the hardening and
softening stages.
−
= +( − ) 1− (3.33)
Figure 3.38 – Hardening/softening curve for compressive behaviour in Atena. Adapted from
Červenka et al. (2016).
Lc is the length scale parameter, which is the crushing equivalent to the Lt showed in Figure 3.37 for
cracking, and transforms the plastic strain into displacements on the descending curve (Figure 3.38).
It is calculated by the element size projection into the direction of minimal principal stresses.
3.3.5 Tension Stiffening
In heavily reinforced structures where cracks cannot fully develop, Atena introduces a tension
stiffening factor cts to account for the concrete contribution to steel stiffness. The value assumed in
123
this research was 0.4, recommended by the CEB-FIP Model Code 1990 (CEB-FIP, 1993). Figure
3.39 illustrates the tension stiffening curve adopted in Atena.
Figure 3.39 – Tension stiffening in Atena. (Červenka et al., 2016).
3.3.6 Embedded Reinforcements
Atena allows reinforcement to be modelled as discrete or smeared. This work only used discrete
reinforcement in form of reinforcement bars modelled with truss elements. The degrees of freedom
from the bar are always kinematically dependent on the degrees of freedom of the respective solid
element. The same behaviour in tension was adopted for compression, according to the bilinear,
elastic-perfectly plastic, curve presented in Figure 3.40. The first slope is calculated according to
the elastic modulus Es of the reinforcement, whereas the second line represents the level of
plasticity adopted in the analysis, being null when perfect plasticity is assumed (Esh = 0).
Figure 3.40 – Bilinear stress-strain law for reinforcement. (Červenka et al., 2016).
3.3.7 Solution Procedure
An arc length solution procedure was adopted in which the stiffness was updated at each iteration.
Equation (3.34) shows the nonlinear equations written in a suitable form for iterative solution.
124
( )∆ = − ( )= −
= +∆ = +
(3.34)
∆ =∆ +
= +∆
where ( ) is the stiffness matrix, vector is a reference loading, is a scalar used to accelerate
solutions in cases of well-behaved load-deformation relationships, is the new degree of freedom
associated with the loading level, ( ) is the vector of internal joint forces, is the deformations of
the structure before the load increment, and ∆ is the deformation increment due to loading
increment. Figure 3.41 shows the notations presented in (3.34).
Figure 3.41 – Arc-length method in Atena. (Červenka et al., 2016).
Error tolerances were adopted of energy (10-4), displacement (10-2), residual error (10-2), and
absolute residual error (10-2) and are shown in equations (3.35), (3.36), (3.37), and (3.38)
respectively.
Δ ( − ( ))
≤ . (3.35)
( )
Δ Δ
≤ . (3.36)
− ( ) − ( ) (3.37)
≤ .
( ) ( )
125
max − ( ) max − ( )
≤ (3.38)
.
max ( ( ) max ( ( )
3.4 CONCLUSIONS
Finite element analysis should be carried out with caution due to the well-known nonlinear
behaviour of reinforced concrete structures. The several number of approaches and assumptions
available to simulate the response of crushing and cracking under a particular loading situation
demand a certain level of knowledge from the researcher and an essential set of validations against
similar experimental studies. This research based most of its initial parameters on the
recommendations of Sagaseta (2008), Eder (2011), and Fang (2014), all of which studied shear
dominant failure of reinforced concrete members with similar FE codes. However, the exact
parameters for the analysis were very case-dependent at times, and sensitivity studies were
developed when suitable to get the best response (Sections 4.3.2, 5.4.1, 5.7.1, and 5.12.1).
In general, Diana gave the best results when the concrete was modelled with the Thorenfeldt
(Thorenfeldt et al., 1987) for compression, and tensile behaviour varied between a Hordijk
softening curve (Cornelissen et al., 1986; Hordijk, 1991) for cases where the slab was meshed with
solid elements, and a linear graph based on the works of Vollum and Tay (2007) for cases where the
slab was meshed with shell elements. For the shear retention factor, the results also depended on the
type of elements adopted for the slab: a constant value in the range of 0.1 to 0.2 as recommended by
Suidan and Schnobrich (1973) was ideal for solid elements. A variable aggregate based shear
retention factor worked best with shell elements. The Quasi-Newton iterative solution procedure
showed satisfactory performance in simulating both the deformed shape and failure load of the
analysed slabs.
Atena is a FE code specialized in the analysis of concrete and reinforced concrete, unlike Diana,
which is a multi-purpose FE software package dedicated, but not exclusive, to a wide range of
problems in Civil Engineering including structural, geotechnical, tunnelling, earthquake and oil &
gas engineering. Because of that, though both codes showed very good results conditioned to the
choice of parameters, Atena required less sensitivity studies than Diana. Typically, the choice of the
recommended CC3DNonLinCementitious2, which was previously described along with an Arc
Length as a solution procedure produced good results for both deformed shape and failure load.
126
4 Punching Shear Failure at Internal Columns
This chapter, based on a paper published in the Magazine of Concrete Research (Soares and
Vollum, 2015), considers design for punching shear at the internal columns of solid flat slabs.
4.1 INTRODUCTION
The relative economy of flat slabs depends on their thickness, which is governed by either
deflection limits or punching resistance. Thinner slabs not only save on direct material costs of the
frame and supporting foundations but also reduce cladding costs due to reduced floor-to-floor
heights. As previously discussed, there is no generally accepted theoretical treatment of punching,
and design methods are calibrated using data from test specimens like the one shown in Figure 1.1,
in which the line of radial contraflexure is fixed and compressive membrane effects are minimal.
Neither of these assumptions is realistic for flat slabs, in which punching resistance can be
significantly increased by restraint from the surrounding slab (Ockleston, 1955).
The benefit of compressive membrane action on punching resistance has been demonstrated
experimentally by researchers including Rankin and Long (1987), Chana and Desai (1992a) and
Salim and Sebastian (2003). Regan (1986) carried out tests on cross-shaped solid slab specimens
which showed that punching resistance is increased by rotational restraint at the slab edges. More
recently, Choi and Kim (2012) tested three internal slab–column connections with a complex set-up
intended to provide zero rotation at the slab edges. The punching resistance was found to be almost
independent of the percentage of moment redistribution used in the calculation of design moments,
contrary to the recommendations of BS 8110:1997 (BSI, 1997) and EC2 (BSI, 2004a), which relate
shear resistance to the flexural reinforcement ratio over the columns. EC2 is the current UK
standard for concrete structures, having replaced BS 8110:1997 (BSI, 1997) in 2010. Interestingly,
BS 8110:1985 (BSI, 1985) could halve the area of shear reinforcement required by the previous UK
code CP110 (BSI, 1972).
Unsurprisingly, this caused concern, which resulted in a test programme (Chana and Desai, 1992b),
funded jointly by the Department of Environment and the British Cement Association, that led to
BS8110 being revised in 1992. EC2 and fib Model Code 2010 (MC2010) (fib, 2013) typically
require significantly more punching shear reinforcement within 1.5d of columns than the 1992
amendment to BS8110.
127
Punching Shear Failure at Internal Columns
Therefore, it is striking that there is no evidence of punching failure in flat slabs designed to
BS8110. The most recent international design recommendations for punching are found in
MC2010, which may influence future revisions of EC2. The MC2010 recommendations are based
on the critical shear crack theory (CSCT) of Muttoni (2008) and Fernández Ruiz and Muttoni
(2009), which relates punching resistance to the slab rotation outside the so-called ‘critical shear
crack’. EC2 and BS8110 do not explicitly consider the potential increase in punching resistance due
to restraint from the surrounding slab but its effect can be modelled using the CSCT (Einpaul et al.
2015) and MC2010 Level IV (Muttoni and Fernández Ruiz, 2012).
This chapter aims to assess the additional level of safety provided by flexural continuity and in-
plane restraint in slabs designed to BS8110 and EC2 by means of the CSCT. The motivation comes
from the feedback from UK’s industry regarding the much greater areas of shear reinforcement
required from BS8110 to EC2, to MC2010, as discussed by Vollum (2013). Figure 4.1 extracted
from Vollum (2013) demonstrates how the shear reinforcement within 1.5d from the column varies
with span for a typical flat slab designed with moment redistribution, when these three codes of
practice are compared.
Figure 4.1 – Influence of span on shear reinforcement area within 1.5d of the column face (Vollum,
2013).
4.2 EVALUATION OF BS8110, EC2 AND MC2010 DESIGN METHODS FOR PUNCHING WITH TEST
DATA
It is well established that EC2 and the CSCT (Muttoni, 2008), on which MC2010 is based, give
similar and reasonable predictions of the punching resistance of test specimens like the one shown
128
in Figure 1.1. For example, Ferreira et al. (2014) compared the predictions of EC2 and the CSCT,
with partial factors of 1.0, for 45 tests with shear reinforcement. They found the mean (μ) and
covariance (COV) of Vtest/Vcalc to be μ=1.19 and COV=0.136 for EC2, and μ=1.16 and COV=0.121
for the CSCT. Closer inspection shows the similarity of these statistics to be misleading, as in 14
cases EC2 falsely predicts failure to occur outside the shear reinforcement, compared with only 5
cases for the CSCT. This is concerning because the introduction of the partial factors γc=1.5 and
γs=1.15 into EC2 makes failure outside the shear reinforcement unlikely, as it causes a 50%
increase in uout which is not matched by a corresponding increase in punching resistance within the
shear reinforcement VR,cs, as discussed by Vollum et al. (2010).
The BS8110 design provisions are not strictly applicable to the test specimens considered by
Ferreira et al. (2014), as the detailing of the shear reinforcement did not comply with the onerous
requirement of BS8110 that the transverse spacing of vertical legs should not exceed 1.5d.
Nevertheless, an analysis was carried out to determine the accuracy of BS8110, EC2 and MC2010
in predicting the punching resistance inside the shear reinforcement Vin, and to evaluate and
compare the factors of safety provided by these codes, since their punching shear recommendations
can differ significantly as presented in Figure 4.1. Each method was used to calculate Vtest/Vin for 40
specimens that failed within the shear reinforcement.
The analysis considered 25 specimens analysed by Ferreira et al. (2014), specimens PL6 to PL12 of
Lips et al. (2012) and specimens 2 to 9 of Chana and Desai (1992b), which formed the basis of the
1992 amendment to BS8110. All the specimens were reinforced with studs, except those of Chana
and Desai, which were reinforced with stirrups. Table 4.1 provides more details on the specimens,
along with the resistance calculated from each code without and with partial factors. In the case of
MC2010, slab rotations were calculated with equation (2.12) (Muttoni, 2008), in which Vflex is the
shear force associated with the flexural capacity of the slab and it was also calculated in accordance
with the recommendations of Muttoni (2008), equation (2.11). The coefficient ksys was taken as 2.8
in the calculation of VRd,max using equation (2.43).
The results are shown in Figures 4.2a and 4.2b which also show the mean, lower characteristic
(5%), and coefficient of variation (COV) values of Vtest/Vin. MC2010 gives conservative estimates
of Vin, with γc=γs=1.0, unlike BS8110 and EC2, of which BS8110 is least safe. The BS8110 5%
values of Vtest/Vin increase from 0.60 to 0.72 for γc=γs=1.0, and from 0.74 to 0.89 for γc=1.5 and
γs=1.15 when the limit on VRd,max is omitted due to reduction in scatter.
129
Table 4.1 – Details of the slabs in Figure 4.2.
Tension
Effective Column
Reinforcement Experimental
Author depth size fc fy fytc Shear VBS8110 VBS8110e VEC2 VEC2e VMC2010 VMC2010e
Slabs dd c
ratio
Reinforcement
V
(kN) (kN) (kN) (kN) (kN) (kN)
(MPa) (MPa) ρ (MPa) (kN)
(mm) (mm) %
C1 a 143 270 47.8 540 1.48 535 10 ϕ10.0 x 6 858 752 940 764 898 647 777
a
C2 140 360 46.9 540 1.52 535 10 ϕ10.0 x 6 956 971 1214 772 861 683 821
C3 a 142 450 48.9 540 1.49 535 10 ϕ10.0 x 6 1077 1120 1378 820 894 744 896
C4 a 140 360 47.9 540 1.52 535 12 ϕ10.0 x 6 1122 981 1226 665 665 744 888
Ferreira et a
C5 140 360 49.7 544 2.00 535 10 ϕ10.0 x 6 1118 1001 1252 808 963 697 839
al. (2014)
C6 a 143 360 48.6 540 1.48 535 10 ϕ10.0 x 6 1078 1013 1266 796 904 699 841
C7 a 144 360 49.0 540 1.47 535 10 ϕ10.0 x 7 1110 1020 1275 915 915 787 968
a
C8 144 360 48.1 540 1.47 535 12 ϕ10.0 x 6 1059 1009 1262 897 1032 757 905
S2 143 300 49.4 540 1.48 535 12 ϕ10.0 x 4 1128 1017 1244 900 1011 768 918
1 150 300 33.1 550 1.45 500 10 ϕ10.0 x 4 881 890 1093 726 866 616 741
Regan 2 150 300 29.7 550 1.76 500 12 ϕ10.0 x 6 1141 877 1096 921 1011 679 815
(2009) 3 150 300 25.6 550 1.76 500 10 ϕ12.0 x 5 1038 815 1018 827 847 681 842
a
5 160 240 61.6 550 1.65 500 12 ϕ12.0 x 5 1268 847 1059 1120 1436 885 1072
PL6 198 130 36.6 583 1.63 519 12 ϕ14.0 x 6 1363 616 770 643 965 848 1060
PL7 197 260 35.9 583 1.59 519 16 ϕ14.0 x 7 1773 1098 1372 1260 1890 1057 1307
PL8 200 520 36.0 583 1.57 519 24 ϕ14.0 x 7 2256 2232 2791 2564 2771 1483 1815
Lips et al. 531 ø20
PL9 266 340 30.5 1.59 516 16 ϕ18.0 x 6 3132 1787 2234 1938 2906 1681 2275
(2012) 580 ø26
PL10 343 440 33.0 580 1.55 563 16 ϕ22.0 x 5 5193 3102 3877 3458 4814 2211 2983
PL11 201 260 33.0 554 1.56 592 8 ϕ10.0 x 7 1176 1074 1343 833 1128 648 851
PL12 201 260 34.6 554 1.56 592 16 ϕ10.0 x 7 1633 1100 1375 1198 1548 804 1102
12 ϕ8.0 at 0.5d
2 200 300 35.52b 520 0.79 520 1094 1021 1260 740 983 676 866
12 ϕ8.0 at 1.25d
12 ϕ8.0 at 0.5d
Chana 3 200 300 32.88 b 520 0.79 520 1139 999 1233 730 969 659 852
12 ϕ8.0 at 1.25d
and Desai
24 ϕ8.0 at 0.5d
(1992b) 4 200 300 36.32 b 520 0.79 520 1302 1183 1448 1105 1404 904 1079
24 ϕ8.0 at 1.25d
12 ϕ10 at 0.5d
5 210 400 30.64 b 520 0.86 520 1382 1330 1630 1193 1524 870 1134
20 ϕ10 at 1.25d
130
20ϕ10 at 0.5d
6 210 400 34.72 b 520 0.86 520 1283 1373 1684 1212 1551 1161 1384
12 ϕ10 at 1.25d
7 210 400 32.32 b 520 0.86 520 32 ϕ10.0 at 0.5d 1492 1348 1653 1201 1535 1186 1450
12ϕ8.0 at 0.5d
8 210 400 31.76 b 520 0.86 520 1324 1225 1511 925 1217 783 1012
20ϕ8.0 at 1.25d
20ϕ8.0 at 0.5d
9 188 300 34 b 520 0.86 520 1135 1115 1363 925 1136 798 956
28ϕ8.0 at 1.25d
Regan and
Samadian A2 160 200 43 570 1.64 519 8 ϕ10.0 x 4 950 751 939 699 921 530 715
(2001)
2 153 200 34.5 680 1.32 430 8 ϕ6.0 x 2 693 595 727 455 609 440 542
Gomes
3 158 200 39.2 670 1.27 430 8 ϕ6.9 x 2 773 685 834 550 659 506 617
and Regan
10 154 200 35.4 670 1.31 430 8 ϕ6.0 x 5 800 602 735 461 625 444 546
(1999)
11 154 200 34.6 670 1.31 430 8 ϕ6.9 x 5 907 648 795 521 693 481 586
S2 124 250 29 488 1.53 393 8 ϕ9.5 x 6 574 597 736 480 480 416 509
S3 124 250 31.6 488 1.53 393 8 ϕ9.5 x 6 572 611 753 513 513 424 519
Birkle S8 190 300 35 531 1.29 460 8 ϕ9.5 x 5 1050 1072 1327 918 1070 807 979
(2004) S9 190 300 35.2 531 1.29 460 8 ϕ9.5 x 6 1091 1074 1330 758 1025 651 849
S11 260 350 30 524 1.10 409 8 ϕ12.7 x 5 1620 1601 2100 1615 1626 1281 1537
S12 260 350 33.5 524 1.10 409 8 ϕ12.7 x 6 1520 1654 2047 1321 1687 994 1288
a
Circular Column
b
fc = 0.8fcu
c
Shear reinforcement yield stress
d
Average effective depth
e
γc=γs= 1.0
131
1.6
1.2
Vtest/VR,cs
0.8
BS8110
0.4 BS8110: mean 0.95; 5%, 0.60; COV, 0.22
EC2: mean 1.01; 5%, 0.74; COV, 0.17 EC2
MC2010 LoA II: mean 1.25; 5%, 1.01; COV, 0.12 MC2010
0
0 0.2 0.4 0.6
∑Aswfy/(udfc0.5)
(a)
2.4
1.6
Vtest/VR,cs
1.2
0.8
BS8110
BS8110: mean 1.18; 5%, 0.74; COV, 0.23
0.4 EC2: mean 1.34; 5%, 0.92; COV, 0.19 EC2
MC2010 LoA II: mean 1.56; 5%, 1.18; COV, 0.15 MC2010
0
0 0.2 0.4 0.6
∑Aswfy/(udfc0.5)
(b)
Figure 4.2 - Influence of the shear reinforcement ratio on Vtest/Vin for BS8110, EC2 and MC2010
Level 2 with: (a) γm=γc=γs= 1.0 (5% = μ – 1.64SD); (b) code recommended partial factors γ m, γc and
γs.
Despite the lack of practical evidence that the more economic punching shear design of the
superseded BS8110 is unsafe, Figures 4.2a and 4.2b raise the question of whether internal slab
column connections designed to BS8110 have adequate reserve against punching failure.
Considering that all 40 slabs evaluated were based on conventional isolated punching shear
specimens, the benefits from the surrounded structure could provide an explanation for the poor
performance of BS8110 with respect to the factor of safety.
4.3 MODELLING OF RESTRAINT FROM SURROUNDING SLAB WITH MC2010 LEVEL IV
MC2010 Level IV was used to investigate whether restraint from the surrounding slab (refer to
section 2.2) is sufficient to explain the satisfactory performance of flat slabs designed to BS8110.
Consideration of equations (2.38) – (2.45) shows that the calculated shear resistance is independent
132
of the axial force in the slab, which implies that increases in strength from compressive membrane
action result from reductions in rotation. Justification for this assumption is provided by the analysis
of tests on prestressed slabs by Clément et al. (2013), who observed that in-plane forces stiffen the
load-rotation response of the slab by delaying the concrete cracking, when compared to a non-
prestressed equivalent slab with the same flexural reinforcement ratio, as shown in Figure 2.23. The
first step involved the development of a nonlinear finite-element analysis (NLFEA) procedure for
calculating slab rotations.
4.3.1 Test specimens
The procedure was calibrated using data from the internal slab–column punching tests reported by
Guandalini et al. (2009) and Lips et al. (2012). All the slabs measured 3 m square on plan with their
geometry following the conventional punching shear specimen presented in Figure 4.3. The flexural
reinforcement ratios ranged between 0.33% and 1.63%, as shown in Table 4.2, which gives details
of the test specimens, including geometry, material properties, failure loads and ultimate rotations.
The rotations were measured using inclinometers positioned at 1.38 m from the column centreline
at the positions depicted by the small triangles in Figure 4.3.
Figure 4.3 – Conventional isolated test specimen adopted by Guandalini et al. (2009) and Lips et al.
(2012).
133
Table 4.2 – Properties of the slabs used in the calibration of the NLFEA
h ca d fc fct ρ fy Vr,test ψ
Specimen
(mm) (mm) (mm) (MPa) (MPa) (%) (MPa) (kN) (‰)
PL1 250 130 x 130 193 36.2 2.272 1.63 583 682 6.0
PV1 250 260 x 260 210 34.0 2.16 1.50 709 974 7.6
531 ø20
PL4 320 340 x 340 267 30.5 2.02 1.58 1625 6.5
580 ø26
PL5 400 440 x 440 353 31.9 2.076 1.5 580 2491 4.7
PG10 250 260 x 260 210 28.5 1.94 0.33 577 540 22.3
PG11 250 260 x 260 210 31.5 2.06 0.75 570 763 10.0
a
Dimension of loaded area
The slabs were modelled with four-node quadrilateral isoparametric curved-shell elements
incorporating embedded reinforcement bars. Plates and columns were modelled with eight-node
isoparametric solid brick (refer to section 3.2.1). A 2×2×9 integration scheme was adopted for the
curved-shell elements, where 9 denotes the number of integration points through the slab thickness,
as recommended by Vollum and Tay (2007). The use of shell elements allows rotations to be
extracted directly from the nodes and compared with the values measured by the inclinometers in
the tests of Guandalini et al. (2009) and Lips et al. (2012) at any position along the perimeter of a
circle of radius rψ, which is commonly rq = rψ = 0.22L, as presented in Figure 4.3. Furthermore, due
to the characteristic straight line deflected shape of flat slabs, the rotations calculated from tangents
around the inclinometers’ position give results as good as the nodal values extracted from the shell
elements, as shown in Figure 4.4 for slab Msup = 0.063FL and qk = 2.5 kN/m2 from the parametric
study developed in Section 4.6.
Msup = 0.063FL and qk = 2.5 kN/m2

800
700
600
Load (kN)
500
Nodal Rotation -- N-S direction
400
Nodal Rotation -- E-W direction
300
Nodal Rotation -- SW - NE direction
200
Rotations from Deflected Shape
100
0
0 0.02 0.04
ψ
Figure 4.4 – Comparison between nodal rotations and rotations calculated from deflected shape for
slab Msup = 0.063FL and qk = 2.5 kN/m2.
134
The concrete was modelled using the ‘total strain fixed crack model’ in Diana, which evaluates
stress–strain relationships in the directions of the principal axes at first cracking. A linear tension-
softening stress–strain relationship was used for concrete. Following the recommendations of
Vollum and Tay (2007), which was concerned with modelling tension stiffening in curved shell
elements that support nonlinear concrete material behavior, the tensile stress was assumed to reduce
from a peak value of 0.5fct, where fct is the mean indirect tensile strength calculated in accordance
with EC2 (see Table 4.2), to zero at half the reinforcement yield strain εsy, as depicted in Figure 4.5.
Figure 4.5 - Concrete tensile stress-strain diagram.
The Thorenfeldt model (Thorenfeldt et al., 1987) was used to model concrete in compression as
presented in section 3.2.3.2. The increase in concrete compressive strength with increasing isotropic
stress was modelled with the four-parameter Hsieh-Ting-Chen failure surface. The reduction in
concrete compressive strength due to lateral cracking was modelled as recommended by Vecchio
and Collins (1993). More details about the Finite Element parameters and elements used in this
analysis can be found in Chapter 3.
4.3.2 Sensitivity study
This section presents some sensitivity studies developed during the calibration of this model in
order to get a reasonable representation of the experimental results. The parameters displayed here
were chosen as a consequence of some issues raised throughout the analysis, which aimed to
estimate the deflected shape as accurately as possible in order to calculate punching resistance with
the CSCT.
4.3.2.1 Mesh size

A mesh sensitivity study was carried out in order to define a reasonable size for the elements in this
analysis. Figure 4.6 shows a comparison between the deflected shapes of specimen PL1 (refer to
Table 4.2) modelled with shell elements varying from 25 mm to 100 mm.
135
It was concluded that there was not much difference in the deformation pattern for flat slabs
modelled with curved shell elements smaller than 50 mm, and slight changes for meshes around 100
mm, especially in the region closer to the column. The element size was chosen to be around 50 mm
square, with the exact dimensions depending on the column size. Figure 4.7 shows the mesh used to
simulate the slab of specimen PV1, using a quarter symmetry.
Mesh Size Comparison

0
-1
0.25Vu
-2 100mm
-3 50mm
0.50Vu
Deflection (mm)
-4 25mm
-5
0.75Vu
Inclinometer
-6
Column
-7
-8 Vu
-9
0 0.05 0.1 0.15 0.2 0.25
r/L
Figure 4.6 – Mesh size comparison of specimen PL1 (Lips et al., 2012).
Figure 4.7 - Mesh used for specimen PV1 (Lips et al., 2012).
4.3.2.2 Boundary Conditions

The slabs described in Table 4.2 are of the conventional type used in punching shear tests at internal
columns, which have an uniformly distributed flexural reinforcement and are loaded through 8 steel
plates, as depicted in Figure 4.3. The specimens are symmetric about both plan axes. Therefore,
136
only a quarter of the slabs was analysed. The initial boundary conditions applied to the model are
shown in Figure 4.8, where in addition to the vertical restraint over the column bottom surface,
axial fixities for symmetry were applied at the surfaces of the column and lines of the slab edges,
which also had rotational fixities. This setup gave good agreement for most slabs presented in Table
4.2. However, for slabs that reached their plastic plateau (low hogging reinforcement ratio), the
boundary condition arrangement shown in Figure 4.8 resulted in membrane forces which increased
the stiffness of the slab unrealistically. In-plane fixities were removed from the slab but kept at the
column as shown in Figure 4.9, in order to check the impact on the slab stiffness. As a result,
compressive forces were reduced, but were still greater than expected, as the response was still
stiffer than the measured test data. Finally, only rotational fixities over the slab edges were used to
apply the symmetry, and the column was removed with its corresponded nodes vertically restrained
over the slab as shown in Figure 4.10.
Figure 4.8 – Model with axial restraint applied at both column and slab.
Figure 4.9 – Model with axial restraint applied only at column.
137
Figure 4.10 – Model without column and axial restraint.
Figure 4.11 shows a comparison between rotations extracted at the same position as in specimen
PG10 (Guandalini et al., 2009), and one specimen from the parametric study developed in section
4.6. The curves are described as follows:
i. Complete – Boundary condition shown in Figure 4.8 with a steel column;
ii. Complete Elastic Column – Boundary condition shown in Figure 4.8 with an
elastic concrete column;
iii. No Axial Restraint – Boundary condition shown in Figure 4.9 with a steel column;
iv. No Column – Boundary condition shown in Figure 4.10.
800 1000
PG10 Msup = 0.063FL and qk = 2.5 kN/m2
700
800
600
500
Load (kN)
600
Load (kN)
400 Experimental
400 Failure Load
300
Complete
200 Complete Elastic Column
200
No Axial Restraint
100 No Column
0
0
0 0.02 0.04
0 0.01 0.02
ψ ψ
(a) (b)
Figure 4.11 – Comparison of a NLFEA of a flat slab with different boundary conditions simulating
the quarter symmetry: (a) specimen PG10 (Guandalini et al., 2009), and (b) Msup=0.063FL and
qk=2.5 kN/m2 from section 4.6.
138
The set up described in Figure 4.10 was adopted for all the NLFEA carried out in this chapter, as it
gave the best predictions for rotations when compared to the experimental results of Guandalini et
al. (2009) and Lips et al. (2012), as shown in Figures 4.11a and 4.12.
4.3.2.3 Elastic modulus

The measured and NLFEA rotations agreed well up to around 50% of the failure load when
calculated with the full short-term concrete elastic modulus, but the measured slopes were
significantly underestimated at failure. Figure 4.12 shows that much better estimates were obtained
of the ultimate rotations when the concrete elastic modulus was reduced to half its short-term value,
calculated in accordance with EC2.
0.6 0.6
PL1 PV1
0.5 0.5
V/bodfck0.5 (MPa0.5)
0.4 0.4
0.3 0.3
Experimental
0.2 0.2 Failure Load
0.1 FEA -- Ec
0.1
FEA -- 0.5Ec
0
0
0 0.002 0.004 0.006
0 0.002 0.004 0.006 0.008
ψ ψ
0.6 0.6
PL4 PL5
0.5 0.5
0.4 0.4
0.3 0.3
Experimental
0.2
0.2 Failure Load
0.1 FEA -- Ec
0.1 FEA -- 0.5Ec
0
0 0.002 0.004 0.006 0
ψ 0 0.002 ψ 0.004 0.006
0.6 0.6
PG10 PG11
0.5 0.5
0.4 0.4
0.3 Experimental
0.3
0.2 Failure Load
0.2
FEA -- Ec
0.1
0.1
FEA -- 0.5Ec
0
0 0 0.003 0.006 0.009 0.012
0 0.01 0.02 0.03
ψ ψ
Figure 4.12 - Comparison of measured and calculated load versus rotation ψ in radians for slabs
PL1, PV1, PL4, PL5, PG10 and PG11.
139
The parametric studies carried out in section 4.6, based on this calibration, also adopted half of the
assumed elastic modulus in order to get a better estimate of the ultimate load with MC2010 LoA IV.
4.3.2.4 Shear Retention factor

Figure 4.13a shows how the deflected shape at the measured failure load varies when the shear
retention factor is increased from β=0.05 to β=0.9 for the flat slab-internal column PL1 tested by
Lips et al. (2012) and modelled with shell elements. It indicates that the calculated rotations are
almost independent of the shear retention factor β. Figure 4.13b presents a comparison between
rotations extracted at the inclinometer position shown as small triangles in Figure 4.3 up until the
measured failure load. The value taken for the analysis in Chapter 4 was β=0.9, as it provided the
best agreement with experimental data.
PL1 PL1
0 700
-1
600
-2
Experimental
500
-3 β = 0.05
Deflection (mm)
Load (kN)
-4 400 β = 0.20
-5 β = 0.50
300
-6 β = 0.75
200 β = 0.90
-7
Inclinometer
-8 100
Column
-9 0
0 0.05 0.1 0.15 0.2 0 0.002 0.004 0.006
r/L ψ
(a) (b)
Figure 4.13 – (a) Deflected shape and (b) Rotations of slab PL1 (Lips et al., 2012) with different
shear retention factors.
4.4 ASSESSMENT OF CHANA AND DESAI (1992) PUNCHING TESTS WITH MEMBRANE ACTION
The tests reported by Chana and Desai (1992a) are particularly pertinent to this investigation. They
tested five 9 m x 9 m large panels with a thickness of 250 mm simply supported at their centre on a
400 mm square plate and by block walls along all four edges. Figure 4.14 shows a schematic of the
test arrangement, and Figure 4.15 shows the reinforcement details. This set up was believed to
impose a minimum lateral restraint which would give a lower bound value of enhancement from
compressive membrane on punching shear resistance of flat slabs supported by an internal column.
The simulated span between columns centrelines was 6 m, although the specimen only extended 4.5
m each way from the column centre to the simply supported edge. The tensile reinforcement ratio
was 0.86% over the column, and the cube strengths of all tested slabs were similar at around 40
140
MPa. Four slabs had shear reinforcement. The slabs were loaded at eight equally spaced points,
which were centred on the loading plate at a radius of 1.2 m. Comparisons were made with the
results from Chana and Desai (1992b), who tested nine conventional flat slab specimens as result of
an increasing concern that BS 8110:1985 (BSI, 1985) design methods rules were optimistic and
possibly unsafe. The test programme considered the concrete term, location of the control
perimeter, limitation on punching shear capacity, distribution and anchorage of shear reinforcement
and membrane action.
Figure 4.14 – Schematic of Chana and Desai (1992a) test arrangement. (Dimensions in mm).
The specimens were 3 m square with thickness varying from 228 mm to 250 mm, supported by a
square 300 mm x 300 mm column and simply supported at the elastic line of contraflexure, assumed
to be at 0.2L. The tensile reinforcement ratio varied from 0.79% to 0.86%, and eight of the
specimens had links as shear reinforcement.
Figure 4.16 shows the comparison between Chana and Desai (1992a) and Chana and Desai (1992b)
continuous and conventional specimens respectively, while Figure 4.17 presents a graph extracted
from Chana and Desai (1992a) comparing deflections of full panels and an equivalent isolated
specimen. Based on their test results, Chana and Desai (1992a) concluded that the shear capacity of
a slab without shear reinforcement can be increased by 40% due to compressive membrane action.
141
Figure 4.15 – Reinforcement details for Chana and Desai (1992a) test.
Figure 4.16 – Isolated specimen of Chana and Desai (1992b) compared to full panel of Chana and
Desai (1992a).
Figure 4.17 – Comparison of deflections between a full panel and an isolated specimen (Chana and
Desai, 1992a).
142
The punching resistances of Chana and Desai’s (1992a) slabs FPS1 (without shear reinforcement)
and FPS5 (with shear reinforcement) were evaluated using MC2010 Level IV. Comparisons were
also made with the shear resistance of Chana and Desai’s (1992b) 3 m square punching specimens.
Rotations were calculated with NLFEA, using the procedure described previously (Section 4.3).
The 3 m square panels were modelled using 50 mm square elements, Figure 4.18(a), and the 9 m
square panels with 100 mm square elements, Figure 4.18(b). For both cases, only one quarter of the
slab was modelled using symmetry at both plan axes. The boundary conditions followed the
arrangement discussed at section 4.3.2.2, where the isolated specimen was modelled with rotational
fixities at both lines of symmetry. The continuous specimen had the in-plane restraint in addition to
the rotational fixity, which provided the best agreement with the experimental results of Chana and
Desai (1992a).
(a)
(b)
Figure 4.18 – Mesh for the NLFEA of the (a) conventional Chana and Desai’s (1992b) specimen,
and (b) continuous Chana and Desai’s (1992a) specimen.
143
Figure 4.19(a) shows that the measured and calculated deflections agree well, which is significant
as MC2010 attributes the increase in punching resistance from restraint to the reduction in rotation,
and hence deflection.
The resulting load–rotation responses are shown in Figure 4.19(b) along with the MC2010
punching resistances for slabs FPS1 and FPS5, calculated with Equations (2.38) and (2.41).
Rotations are shown along the slab centreline at the loading radius, as measured in the tests reported
by Lips et al. (2012), and additionally in the 9 m square slab at 0.7 m from the column centreline,
where rotations were greatest. These FE rotations were extracted from the nodes positioned at the
aforementioned locations. For comparison, Figure 4.19(b) also shows rotations calculated using
Equation (2.12), which is applicable to the 3 m square panels. The calculated failure load is given
by the intersection of the rotation and resistance curves.
2000
1800
1600
1400
Load (kN)
1200
1000
800 Experimental
600
Continuous
400
200
Conventional
0
0 2 4 6 8 10 12 14 16 18 20
Deflection (mm)
(a)
0.9
0.8
Failure Load
0.7 FPS1
Failure Load
0.6
FPS5
0.5 3 m square
Resistance FPS5
V/(udfck0.5)
0.4
9 m square
0.3 (r=1200mm)
0.2 9 m square
Resistance FPS1 (r =700mm) -
MAX
0.1 Equation (2.12)
0
0 0.01 0.02 0.03
ψ
(b)
Figure 4.19 - Analysis of Chana and Desai’s (1992a) slabs FPS1 (no stirrups) and FPS5 (with
stirrups): (a) comparison of measured and predicted deflections; (b) calculation of resistance with
MC2010.
144
When the effect of continuity is included, the ratio Vtest/Vcalc for maximum rotations is 1.12 for
FPS1 and 1.41 for FPS5. A measure of the influence of continuity is the ratio of the shear
resistances given by the Level IV analysis with continuity and Equation (2.12). This ratio is 1.52 for
FPS1, which is close to the measured ratio of 1.4, and 1.07 for FPS5, which is significantly less
than the measured ratio of 1.4. The underestimate in strength of FPS5 is a consequence of MC2010
neglecting shear deformation in the calculation of the shear reinforcement stress.
4.5 PARAMETRIC STUDIES TO COMPARE SHEAR REINFORCEMENT REQUIREMENTS OF BS8110,

EC2 AND MC2010
A parametric study was undertaken to investigate how the required areas of shear reinforcement
vary at the internal columns of flat slabs according to BS8110, EC2 and MC2010 Level II. The span
L between the column centrelines was taken as 7.5 m and the internal columns as 450 mm square.
The superimposed dead load was taken as 1.5 kN/m2 and the superimposed live load was varied
between 2.5 and 10 kN/m2. Dead and imposed load factors of 1.35 and 1.5 were used with EC2 and
MC2010, and corresponding load factors of 1.4 and 1.6 with BS8110.
Characteristic material strengths of fck=30 MPa and fyk=500 MPa were adopted in conjunction with
code-recommended material partial factors. The slab thickness was related to the design imposed
loading in accordance with Goodchild et al.’s (2009) recommendations for economical frame
construction, as presented in Figure 4.20. The resulting slab thicknesses and mean effective depths
are given in Table 4.3.
BS8110 and the UK National Annex to EC2 allow design moments for slabs to be calculated using
a single load case in which all spans are fully loaded, provided the support moments are
redistributed downwards by 20% and the span moments are increased accordingly. The effect of
this moment redistribution on the amount of shear reinforcement required by each code was
investigated in the parametric study. Consequently, the hogging moment at the column centreline
Msup was taken as either its elastic value of 0.083FL or 0.063FL from Table 3.12 of BS8110, which
includes the 20% moment redistribution mentioned above (F is the total load on each panel).
145
Figure 4.20 – Span/thickness chart for flat slabs extracted from Goodchild et al. (2009).
Table 4.3 - Properties of slabs used in the parametric studies.

Design Hogging Moment 0.063FL 0.083FLa
Design imposed live load: kN/m2 2.5 5.0 7.5 10.0 2.5 5.0 7.5 10.0
Slab thickness: mm 239 265 324 363 239 265 324 363
F: kN 779 1039 1362 1647 779 1039 1362 1647
Average effective depth: mm 194 220 279 318 194 220 279 318
Asl centre column strip 1917 2259 2316 2451 3255 3845 3897 4111
2
Msup: mm Asl edge column strip 959 1130 1158 1226 1627 1922 1949 2056
Asl middle strip 959 1130 1158 1226 1627 1922 1949 2056
Asl column strip 2498 2943 3017 3193 3580 4229 4287 4523
Mspan: mm2
Asl middle strip 2043 2408 2469 2613 2929 3460 3508 3700
L = 7.5 m, columns 450 mm square, fck = 30 MPa, design imposed dead load 1.5 kN/m2
a
Asl provided/Asl required = 1.04 to allow for rationalisation of reinforcement arrangement
In each case, the design span moment Mspan was taken as 0.063FL as given in Table 3.12 of BS8110
for interior panels. However, the same areas of hogging and sagging reinforcement were provided
in the panels designed for Msup=0.083FL to simulate the common practice of adding surplus
flexural reinforcement in the span to control deflection (Vollum, 2009). To maximise the difference
between the two cases, the design hogging moment for flexural reinforcement was taken at the
centreline of the column for elastically designed slabs and at hc/3 from the column centreline (where
hc is the column depth) for slabs designed for Msup=0.063FL. The latter moments satisfy the
BS8110 requirement that the sum of the maximum span moment and average support moments
across the panel width should exceed F(L – 2hc/3)2/8.
146
In addition, the calculated areas of flexural reinforcement were increased by 4% in the slabs
designed for Msup=0.083FL to allow for rationalisation of the reinforcement arrangement. The
design hogging moment was proportioned between the column and middle strips (refer to Figure
2.30) in the ratio 75:25, with two-thirds of the column strip reinforcement placed in its central half
in accordance with the requirements of BS8110 and EC2 as shown in Table 4.3. The design shear
force was multiplied by β=1.15 in accordance with the recommendations of BS8110 and EC2. In
the case of MC2010, the control perimeter u was multiplied by ke=0.9, and the maximum possible
shear resistance was calculated with ksys=2.8 in equation (2.43), as recommended for studs.
The punching shear reinforcement was arranged radially in the EC2 and MC2010 designs but in
square perimeters for the BS8110 designs. In the BS8110 and EC2 designs, the spacing of
perimeters of shear reinforcement was taken as 0.5d, 1.25d, 2.0d etc. from the column face, in
accordance with UK practice. The perimeter spacing was reduced to 0.5d in the MC2010 designs as
the required area of shear reinforcement doubles for spacings of 0.5d, 1.25d, 2.0d etc., because only
one perimeter crosses the critical shear crack.
Figure 4.21 shows the variation in V/VR,max EC2 (where VR,max EC2 = 2VRd,c EC2) with the design
imposed load, and hence slab thickness, according to BS8110, EC2 and MC2010 for design
hogging moments of 0.063FL and 0.083FL from Table 3.12 of BS8110. The economic slab
thicknesses reported by Goodchild et al. (2009) are seen to comply with the BS8110 and UK
National Annex to EC2 restrictions on VR,max=2VRd,c but not the recommended code limit of
VR,max=1.5VRd,c, which is intended for stirrups. In the case of MC2010, VR,max is critical for all slabs
with Msup=0.063FL.
1.8
1.6
1.4
1.2
1
VEd/VR,max
0.8
EC2 M=0.063FL
0.6 EC2 M=0.083FL
0.4 BS8110 M=0.063FL
BS8110 M=0.083FL
0.2 MC2010 M=0.063FL
MC2010 M=0.083FL
0
0 2.5 5 7.5 10
Design imposed load (kN/m2)
Figure 4.21 - Influence of design imposed load on VEd/Vmax for design support moments of 0.063FL
and 0.083FL.
147
Figure 4.22a compares the total areas of shear reinforcement required by each code within 1.5d of
the column face, neglecting the limit on VR,max, which invalidates the MC2010 designs with
Msup=0.063FL. BS8110 requires the least area of shear reinforcement and MC2010 the most. The
difference between BS8110 and the other codes is greatest for slabs designed for Msup=0.083FL, as
the design shear force is less than 1.6VRd,c, making equation (2.23) from BS8110 applicable.
The installation time for shear reinforcement depends on the total number of shear studs, which is
governed by uout and spacing rules, rather than Asw, which determines the stud diameter for a given
shear reinforcement arrangement. Therefore, the required normal distances from the column face to
the outer row of shear reinforcement are compared for each method in Figure 4.22b, which shows
remarkable disparities between the extents of shear reinforcement required by each code,
particularly for Msup=0.063FL, where the extent of shear reinforcement required by MC2010 is
much greater than for BS8110 or EC2. The difference is, in part, due to MC2010 basing the shear
resistance at uout on the critical shear crack width around the column, which is particularly
unrealistic once the reinforcement yields over the column.
6000 EC2: M=0.063FL

∑Asw in 1.5d zone around the column (mm2)
EC2: M=0.083FL
5000 BS8110: M=0.063FL
BS8110: M=0.083FL
4000 MC2010: M=0.063FL
MC2010: M=0.083FL
3000
2000
1000
0
0 2.5 5 7.5 10
(a)
148
EC2 M=0.063FL
9
EC2 M=0.083FL
Distance to outer ring of shear reinforcement/d

8 BS8110 M=0.063FL
BS8110 M=0.083FL
7 MC2010 M=0.063FL
MC2010 M=0.083FL
6
0
0 2 4 6 8 10 12
(b)
Figure 4.22 - Comparison of (a) areas of shear reinforcement required within 1.5d of column, and
(b) required minimum distance to outer shear reinforcement for design support moments of 0.063FL
and 0.083FL.
The minimum possible slab thickness can be limited by VR,max in thin slabs, with edge and corner
columns being most critical. In this case, VR,max can be increased by providing surplus hogging
flexural reinforcement, as shown in Figure 4.23a for a 265 mm thick slab. The shear resistances in
Figure 4.23a are normalised by VRd,c EC2 calculated with Asl provided = Asl required for design support
moments of 0.063FL and 0.083FL. Figure 4.23a shows that MC2010 (with ksys=2.8) gives
significantly lower maximum possible shear resistances than BS8110 or the UK National Annex to
EC2, which limit VR,max to 2VRd,c. Figure 4.23b compares the areas of shear reinforcement required
by each code, within 1.5d of the column face, for a design imposed load of 5 kN/m2. EC2 and
MC2010 require greatly more shear reinforcement than BS8110, particularly in cases where
increasing Asl provided/Asl required makes equation (2.23) govern.
149
2.5
VR,max/Vc EC2 Asprov=Asreq

2
EC2: M=0.063FL
1.5
EC2: M=0.083FL
1 BS8110: M=0.063FL
BS8110: M=0.083FL
0.5 MC2010: M=0.063FL
MC2010: M=0.083FL
0
1.00 1.25 1.50 1.75 2.00
Asprov/Asreq
(a)
3500 EC2: M=0.063FL
EC2: M=0.083FL
∑Asw in 1.5d zone around column
BS8110: M=0.063FL
3000
BS8110: M=0.083FL
MC2010: M=0.063FL
2500
MC2010: M=0.083FL
2000
(mm2)
1500
1000
500
0
1.00 1.25 1.50 1.75 2.00
Asprov/Asreq
(b)
Figure 4.23 - Influence of surplus flexural reinforcement of a 265 mm thick slab with 450 mm
square column on: (a) maximum possible shear resistance; (b) area of shear reinforcement for
F=1039 kN.
4.6 INFLUENCE OF RESTRAINT FROM SURROUNDING SLAB
MC2010 Level IV was used to assess the influence of restraint from surrounding bays on the
punching resistance of the slabs designed in the previous section. Rotations were calculated with
NLFEA using the procedure described in section 4.3. The boundary conditions were varied as
shown in Figure 4.24, where (i) represents a conventional punching shear specimen of width 0.44L,
(ii) represents one-quarter of an internal panel of a flat slab of span L with rotational restraint at
midspan, and (iii) represents a slab with in-plane and rotational restraint at midspan. The isolated
slab, (i) in Figure 4.24b, was loaded at eight points around its perimeter to simulate a conventional
punching test, whereas the continuous slabs were loaded uniformly. The mesh used for the
150
conventional slab presented in Figure 4.24b(i) is similar to that adopted for the Chana and Desai
(1992b) analysis, shown in Figure 4.18a. The continuous slabs were modelled as shown in Figure
4.25a for (ii), and Figure 4.25b for (iii), with the only difference between them being the addition of
in-plane fixities.
(a)
(b)
Figure 4.24 - Boundary conditions for MC2010 Level IV analysis of the interior panels of a flat
slab.
151
(a)
(b)
Figure 4.25 – Models for the continuous slabs in the parametric studies: (a) rotational fixity and (b)
rotational plus in-plane fixities.
The rotations were extracted from the NLFEA along the slab centreline at nodes positioned around
0.2L from the column centre. This position was chosen because it was used in the calibration of the
NLFEA, since it was where Guandalini et al. (2009) and Lips et al. (2012) placed the inclinometers
152
in their tests, and is close to the position of maximum rotation, as shown in Figure 4.26a for
Msup=0.063FL and qk=2.5 kN/m2 where qk is the characteristic design imposed load (rotations are
shown normalised by the peak value for each boundary condition case). The CSCT recommends the
rotation to be taken at a distance d from the column face, shown in Figure 4.26b, as it assumes a
straight line deformation pattern. Figure 4.27 shows the calculated load against rotation responses for
slabs designed for Msup=0.063FL, and Figure 4.28 shows them for Msup=0.083FL with design
imposed loads of 2.5, 5.0, 7.5, and 10.0 kN/m2. In addition to NLFEA, rotations were calculated
using equations (2.45), (2.12), and Muttoni’s (2008) quadrilinear moment–curvature relationship
(2.8), which includes tension stiffening. The NLFEA rotations are denoted as follows in Figures
4.27 and 4.28.
 Conventional: NLFEA of isolated slab (see (i) in Figure 4.24b).
 Continuous: NLFEA of continuous slab with rotational restraint (see (ii) in Figure 4.24b).
 Continuous+axial: NLFEA of continuous slab with rotational and in-plane restraint (see
(iii) in Figure 4.24b).
Conventional
1
Continuous
Continuous + axial restraint
0.8
Column face
Normalised rotation
0.5c + d
0.6
0.22L
0.4
0.2
0
0 0.1 0.2 0.3 0.4 0.5
r/L
a)
(b)
Figure 4.26 - Variation in the nodal rotation along the slab centreline for M=0.063FL and q k=2.5
kN/m2.
153
0.6 0.6 Design load
PYL test specimen
MC2010 (LoA II) Conventional
0.5 MC2010 (LoA II) 0.5
Continuous
MC2010 (EC2) MC2010 (EC2) Continuous + axial
0.4 0.4 Eq 2.45
Eq 2.12
V/(udfck0.5)
V/(udfck0.5)
Eq 2.8
0.3 0.3
VRmax
0.2 0.2
MC2010 (BS8110)
0.1 MC2010 (BS8110) 0.1
0 0
0 0.01 0.02 0.03 0.04 0.05 0.06 0 0.01 0.02 0.03 0.04 0.05 0.06
Rotation ψ [Radians] Rotation ψ [Radians]
(a) (b)
0.6 Design load
0.6 PYL test specimen
Conventional
0.5 MC2010 (LoA II)
0.5 MC2010 (LoA II) Continuous
MC2010 (EC2) MC2010 (EC2) Continuous + axial
0.4 Eq 2.45
V/(udfck0.5)
0.4
V/(udfck0.5)
Eq 2.12
Eq 2.8
0.3
0.3 VRmax
0.2 0.2
MC2010 (BS8110)
MC2010 (BS8110)
0.1 0.1
0 0
0 0.01 0.02 0.03 0.04 0.05 0 0.01 0.02 0.03 0.04 0.05
(c) (d)
Figure 4.27 - Influence of slab continuity on shear resistance: (a) Msup=0.063FL and qk=2.5 kN/m2; (b) Msup=0.063FL and qk=5.0 kN/m2; (c) Msup=0.063FL and qk=7.5 kN/m2;
(d) Msup=0.063FL and qk=10 kN/m2
Note: MC2010 (BS8110), MC2010 (EC2) and MC2010 (LoA II) are the resistances calculated with MC2010 with ∑ from BS8110, EC2, and MC2010 LoA II,
respectively (see Table 4.4).
154
0.6 0.6 Design load
MC2010 (LoA II) PYL test specimen
0.5 MC2010 (LoA II) 0.5
Conventional
Continuous
0.4 0.4
Continuous + axial
V/(udfck0.5)
V/(udfck0.5)
Eq 2.45
0.3 0.3
Eq 2.12
MC2010 (EC2)
MC2010 (EC2) 0.2 Eq 2.8
0.2
MC2010 (BS8110) 0.1

MC2010 (BS8110)
0.1
0 0
0 0.01 0.02 0.03 0.04 0.05 0 0.01 0.02 0.03 0.04
(a) (b)
0.6 0.6 Design load
PYL test specimen
0.5
MC2010 (LoA II) 0.5 MC2010 (LoA II)
Conventional
Continuous
0.4 0.4
Continuous + axial
V/(udfck0.5)
Eq 2.45
V/(udfck0.5)
0.3 0.3
Eq 2.12
MC2010 (EC2) Eq 2.8
0.2 MC2010 (EC2) 0.2
0.1 MC2010 (BS8110) 0.1 MC2010 (BS8110)
0 0
0 0.01 0.02 0.03 0.04 0 0.01 0.02 0.03
(c) (d)
Figure 4.28 - Influence of slab continuity on shear resistance: (a) Msup=0.083FL and qk=2.5 kN/m2; (b) Msup=0.083FL and qk=5.0 kN/m2; (c) Msup=0.083FL and qk=7.5 kN/m2;
(d) Msup=0.083FL and qk=10 kN/m2
Note: MC2010 (BS8110), MC2010 (EC2) and MC2010 (LoA II) are the resistances calculated with MC2010 with ∑ from BS8110, EC2, and MC2010 LoA II,
respectively (see Table 4.4).
155
Figures 4.27 and 4.28 also show the punching resistances according to MC2010 for the areas of
shear reinforcement required by BS8110, EC2 and MC2010 (denoted by ‘MC2010 (BS8110),
‘MC2010 (EC2)’ and ‘MC2010 (LoA II)’, respectively), as well as VR,max=2.8VRd,c, where critical.
The punching resistances were calculated with γc=γs=1.0, ke=0.9 and db=12 mm. Failure loads are
given by the intersection of the resistance and rotation curves. The design ultimate shear force
VEd=F is shown for comparison, as is the flexural failure load of a comparable conventional isolated
test specimen, which is denoted ‘PYL test specimen’.
The rotations given by Muttoni’s (2008) quadrilinear moment curvature relationship (2.8) compare
well with those given by NLFEA of isolated slabs up to around 70% of the flexural capacity given
by equation (2.11). Figures 4.27 and 4.28 show that rotational restraint at the panel edges increases
punching resistance above that given by (2.8) or equation (2.12). Even greater resistances are
obtained with rotational and full in-plane restraint, but the latter is not generally available in flat
slabs. Equation (2.45) of MC2010 is seen to conservatively estimate the benefit of rotational
restraint from surrounding panels.
The calculated punching resistances are summarised in Table 4.4, which also shows the areas of
shear reinforcement included in the calculation of punching resistance with MC2010. Strikingly,
when rotational restraint is included, Vcalc/F is only around 20% greater for MC2010 than BS8110,
despite MC2010 requiring 2.25–4 times the area of shear reinforcement within d of the column
face.
The ratio of resistances calculated with rotational restraint (continuous) and equation (2.12) gives an
upper bound to the increase in strength due to rotational restraint. As can be seen from Table 4.4,
this ratio varies between 1.05 and 1.49, and is typically less than the 30–50% found by Chana and
Desai (1992a). Table 4.4 also shows the benefit of continuity to be more significant for slabs
designed with moment redistribution (Msup=0.063FL). This is a consequence of reduction in shear
slenderness due to redistribution between hogging moments around the column and sagging
moments at midspan, as previously discussed (refer to section 2.2.2), and observed by Einpaul et al.
(2015).
Despite this, rotational restraint appears sufficient to explain the satisfactory performance of
internal slab–column connections designed to BS8110, although according to MC2010, which tends
to underestimate strength, the factor of safety is close to 1.0. Einpaul et al. (2015) raised a similar
point, when discussing the assessment of existing structures where the effects of continuity should
be accounted for in order to avoid unnecessary strengthening. The EC2 designs appear optimum
from the view of economy and safety.
156
Table 4.4 - Influence of continuity on shear strengths calculated to MC2010 with = = 1 and = 0.9
Design Vcalc/F
method for
Boundary conditions
shear Msup = 0.063FL Msup = 0.083FL
reinforcement qk = 2.5 kN/m2 qk = 5.0 kN/m2 qk = 7.5 kN/m2 qk = 10 kN/m2 qk = 2.5 kN/m2 qk = 5.0 kN/m2 qk = 7.5 kN/m2 qk = 10 kN/m2
∑ [mm ] 2
433 b
852 b
1200 c
1613 c
368 d
486 b
534 c
733c
Equations (2.12) & (2.11) 0.74 0.78 0.79e 0.79e 0.90 0.85 0.90 0.83
Equation (2.45) 0.77 0.80 0.83 0.86 0.92 0.87 0.92 0.84
BS8110 Conventional a 0.82 0.84 0.85 0.87 1.02 0.94 1.01 0.93
Continuous a 0.95 0.95 0.99 1.02 1.08 0.99 1.07 1.00
Continuous/Equation (2.12)f 1.28 1.20 1.25 1.28 1.20 1.16 1.18 1.21
a
Continuous + axial 1.22 1.14 1.17 1.16 1.23 1.29 1.21 1.33
∑ [mm2] 878 1223 1538 1831 758 1080 1365 1635
Equations (2.12) & (2.11) 0.79 e 0.79 e 0.79 e 0.79 e 1.04 1.01 1.01 1.00
Equation (2.45) 0.92 0.91 0.90 0.90 1.05 1.03 1.03 1.02
EC2 Conventional a 0.90 0.89 0.89 0.89 1.15 1.12 1.08 1.04
Continuous a 1.06 1.03 1.05 1.04 1.21 1.15 1.11 1.08
Continuous + axiala 1.36 1.22 1.20 1.18 1.34 1.35 1.21 1.37
∑ [mm2] 1508 2090 2710 3290 1161 1683 2156 2696
Equations (2.12) & (2.11) 0.79 e 0.79 e 0.79 e 0.79 e 1.14 1.18 1.18 1.19
Equation (2.45) 1.04 0.98 0.99 0.99 1.16 1.20 1.20 1.21
MC2010 Conventional a 0.93 0.91 0.92 0.92 1.24 1.26 1.27 1.28
Continuous a 1.18 1.14 1.16 1.15 1.30 1.32 1.33 1.25
Continuous + axiala 1.68 1.52 1.38 1.33 1.49 1.42 1.31 1.41
a
Rotation calculated with NLFEA.
b
50:50 split of ∑ between first and second perimeters.
c
40:60 split of ∑ between first and second perimeters.
d
Asw BS8110 min.
e
Flexural Failure
f
Upper bound of enhancement in strength of isolated test specimen due to rotational restraint.
157
4.7 INFLUENCE OF CONTINUITY ON PUNCHING RESISTANCE OF NINE-PANEL FLAT SLAB
Experimental evidence for the effect of structural continuity on the punching resistance of
uniformly loaded flat slabs can be seen by analysing the work of Guralnick and Fraugh (1963), who
tested a three-quarter scale reinforced concrete flat plate based on a prototype designed by
consulting engineers Di Stasio and van Buren. The model had nine square 4.57 m x 4.57 m panels
arranged three-by-three resulting on a 13.72 m x 13.72 m, 133 mm thick floor system supported on
16 columns. The exterior columns were connected by 209.55 mm deep and 304.8 mm wide shallow
spandrel beams along the west and north edges while the east and south edges used 400.05 mm deep
and 152.4 mm wide deep beams, providing a diagonal symmetry to the structure. The internal
columns supported the slab without any beams, drop panels, or column capitals. Figure 4.29 shows
a plan view and a section of the test.
Although the prototype was designed as a typical floor in which columns extend both below and
above the slab, the column above had to be removed for the test set up. The columns below the slab
had their heights halved to compensate the loss of stiffness, and were fixed at their lower ends. #4
round bars were used for the slab reinforcement, with an 11.12 mm cover (3/4 of the 14.29 mm
prototype cover). A uniformly distributed load was simulated by applying 25 concentrated loads
through 203.2 mm x 203.2 mm steel plates to each panel.
The slab had a mean effective depth of 109.5 mm, and spanned 4.57 m between the column
centrelines. The internal columns were 457 mm square. No shear reinforcement was provided.
Deformed reinforcement bars with a mean yield strength of 276 MPa were used. The concrete
cylinder strength was 32.5 MPa. The uniformly distributed failure load was 1.05 times that given by
yield line analysis. Immediately before failure, the average recorded steel strain at the four faces of
the critical column was around seven times the yield strain. The corresponding maximum strains in
the span reinforcement were around three-quarters of the yield strain. The load deflection response
also indicated that the slab was close to flexural failure, even though it failed in punching. The
failure occurred on internal column 7 (refer to Figure 4.29) at a shear force of around 399 kN.
However, Guralnick and Fraugh (1963) concluded the study observing that, had the slab not failed
in punching, it would probably have failed at a much higher load due to the contributions of strain
hardening of reinforcement and compressive membrane action.
158
Figure 4.29 – Layout of test structure by Guralnick and Fraugh (1963). Dimensions in mm.
The shear resistance was calculated with BS8110, EC2 and MC2010 with γc=γs=γm=1.0. The effect
of moment transfer to the column was included in the BS8110 and EC2 strength assessments by
multiplying the applied shear force by β=1.15. In the case of MC2010, ke was taken as 0.9 as
159
recommended in the code. Slab rotations were calculated using equations (2.45) and (2.12) and
from the measured slab deflections as (4.1).
3
~ (4.1)
where w is the mean midspan deflection in the four panels surrounding the critical column, and L is
the span.
Equation (4.1) was derived from NLFEA of the continuous slabs with rotational restraint considered
in the previous section. The equation is remarkably accurate as shown in Figures 4.30a and 4.30b
for all the slabs from the parametric study carried out in section 4.6, and in 4.30c and 4.30d for the
same slabs with a 750 mm square column, which is one-tenth of the span, as in the Guralnick and
Fraugh test. Figure 4.31 shows the calculated rotations as well as the punching resistance according
to MC2010. Equation (2.12) gives the rotation of a conventional isolated punching specimen with
the same hogging reinforcement as the tested slab and rs=rq=0.22L, for which Vflex=299 kN. The
shear resistances corresponding to Equations (2.12), (2.45) and (4.1) are 299, 325 and 393 kN,
respectively. The latter agrees well with the measured strength of 399 kN, and illustrates the benefit
of flexural continuity, which is only partly included in equation (2.45) which is used in MC2010
Level II. The shear resistances given by BS8110 and EC2 are 413 and 377 kN, respectively, which
suggests that punching resistance was increased by rotational restraint but not compressive
membrane action, due to the high utilisation of flexural reinforcement in surrounding panels. If the
slab strength had been increased by in-plane forces, both BS8110 and EC2 predictions would have
been a lot more conservative.
2500 3500
M = 0.063FL M = 0.083FL
3000
2000
2500
qk=2.5kN
Load (kN)
Load (kN)
1500 2000 Equation 4.1: qk=2.5kN

qk=5.0kN
1500
1000 Equation 4.1: qk=5.0kN
1000 qk=7.5kN

500 qk=10kN
Equation 4.1: q=10kN
0 0
0 0.01 0.02 0.03 0.04 0.05 0.06 0 0.01 0.02 0.03 0.04 0.05 0.06
ψ ψ
(a) (b)
160
3000 4000
M = 0.063FL M = 0.083FL
Column = 750mm 3500 Column = 750mm
2500
3000
2000
2500
Load (kN)
Load (kN)
qk=2.5kN
Equation 4.1: qk=2.5kN
1500 2000
qk=5.0kN
1000
qk=7.5kN
1000
Equation 4.1: qk=7.5kN
500
500 qk=10kN
Equation 4.1: q=10kN
0 0
0 0.01 0.02 0.03 0.04 0.05 0.06 0 0.01 0.02 0.03 0.04 0.05 0.06
ψ ψ
(c) (d)
Figure 4.30 – Comparison between rotations extracted from shell elements and calculated with
Equation (4.1).
700
600 Resistance
500 Equation 2.45
Equation 2.12
400
Load (kN)
Equation 4.1
300
200
100
0
0 0.01 0.02 0.03 0.04 0.05 0.06
ψ
Figure 4.31 - Calculation of MC2010 shear resistance of the slab reported by Guralnick and Fraugh
(1963).
4.8 CONCLUSIONS
The relative safety of the design rules for punching shear in BS8110, EC2 and MC2010 were
compared for flat slabs on interior columns. With partial material factor of 1.0, BS8110 is shown to
significantly overestimate the punching of isolated test specimens within the shear reinforced zone.
EC2 is more conservative than BS8110, but its level of safety reduces as the shear reinforcement
ratio increases with strengths overestimated, within the shear reinforced zone, at higher
reinforcement ratios. MC2010 performs noticeably better in this respect, which is significant
161
because failure outside the shear reinforcement, which controls the calculated strength of many test
specimens, is unlikely in practice, because the introduction of partial factors tends to make failure
inside the shear reinforced zone critical.
Parametric studies (see Figure 4.21) show that limiting VR,max to ksysVRd,c in MC2010 prevents 20%
downwards moment redistribution over the columns of flat slabs, which has been allowed in the UK
for many years. Assessment with MC2010 Level IV shows that punching resistance is increased by
rotational continuity at midspan, and even more so by combined in-plane and rotational restraint.
The increase in calculated punching resistance due to rotational continuity is best seen by
comparing strengths calculated with rotations from equation (2.12) from the CSCT, which is
derived for isolated test specimens, and NLFEA of a complete panel with rotational restraint at
midspan. The increase in strength due to rotational continuity is partially included in MC2010 Level
II if rotations are calculated using equation (2.45). This has practical implications as the complete
panel with rotational restraint at midspan can provide a lower bound of the behaviour of continuous
flat slabs according to Einpaul et al. (2015), and MC2010 LoA II is easier to implement than LoA
IV.
It is necessary to invoke flexural continuity to explain the observed strength of the Guralnick and
Fraugh slab with MC2010, but not BS8110 or EC2, as it increases punching resistance by reducing
rotations below those in the comparable isolated punching specimens with which MC2010 is
calibrated. The wide variation in calculated failure loads of identical slabs evident in Figures 4.27
and 4.28 suggest that the adoption of a rotational-based failure criterion could lead to disagreements
between designers and checking engineers. Consequently, the more empirical design methods of
BS8110 and EC2 seem better suited for normal design, although MC2010 is useful for assessment.
Rotational restraint from surrounding panels, along with rounding up of calculated areas of
reinforcement, seem sufficient to explain the satisfactory performance of flat slabs designed to
BS8110.
The analysis carried out in this Chapter shows the potential benefit of continuity on punching
resistance of flat slabs on internal column. However, the numerical model is seen to be highly
dependent on the choice of boundary conditions, which are hard to model accurately. Furthermore,
the original CSCT failure criterion implemented in MC2010 neglects any increase in strength due to
axial compression. Clément et al. (2014) have addressed that issue by proposing a reduced effective
depth to account for the reduction of the cracked zone depth as a result of in-plane forces.
162
5 Punching Shear Failure at Edge Columns
This chapter, based on a paper published in the Magazine of Concrete Research (Soares and
Vollum, 2016a) and a paper published and presented at the Joint ACI-fib International Symposium
on Punching Shear of Structural Concrete Slabs (Soares and Vollum, 2016b), considers punching
failure at edge columns of reinforced concrete flat slabs without shear reinforcement and
unbalanced moments about an axis parallel to the slab edge.
5.1 INTRODUCTION
There is no generally accepted theoretical treatment of punching shear, and design methods are
calibrated largely with data from tests on isolated internal slab column specimens. Punching at edge
columns is much less researched than at internal columns, despite the fact that buildings typically
have more edge than internal columns. Significantly, in typical flat slab buildings, moment transfer
from the slab to column has a greater influence on the design punching resistance of edge and
corner than internal slab column connections, and practical experience shows that design for
punching shear is frequently more critical at edge than internal columns.
Flat slabs are normally designed using either elastic finite element analysis (FEA) or equivalent
frame analysis. In UK practice, slab column connections of equivalent frames are modelled as rigid
(The Concrete Society, 2007). Conversely, ACI 318-14 (ACI, 2014) introduces torsional members
between the slab and column to account for the flexibility of the slab column connection.
Equivalent frame analysis according to UK practice tends to overestimate column moments since
the flexibility of the slab column connection is not modelled. Consequently, UK practice allows
moments about an axis parallel to the slab edge (i.e. normal moments) to be redistributed by up to
50% at edge columns of flat slabs designed using the equivalent frame method. Only 30% moment
redistribution is allowed if the design moments are calculated with elastic finite element or grillage
analysis.
Significant investigations into the strength of slab edge column connections without shear
reinforcement have been carried out by, amongst others, Zaghlool (1971) (eight isolated
1830×965×152 mm slabs with various column sizes), Stamenković and Chapman (1974) (six
isolated 914.4×914.4×76.2 mm slabs with 127 mm square columns), Regan et al. (1979) (21 tests
on continuous slabs with thickness ranging between 80 and 125 mm), Regan (1993) (ten tests on
five continuous 200 mm thick slabs), Rangan (1990) (four continuous specimens with slab
163
Punching Shear Failure at Edge Columns
thickness of 80 mm and 100 mm), El-Salakawy et al. (1998) (two isolated 1540×1020×120 mm
slabs with 250 mm square columns) and Sherif et al. (2005) (five isolated 1000×1200×120 mm
slabs with various column sizes). As noted, these tests were carried out on a variety of isolated and
continuous specimens of which the test arrangements of El-Salakawy et al. (1998) and Regan
(1993) shown in Figure 5.1 are representative.
The loading eccentricity is typically fixed in tests on isolated specimens, but varies with loading in
continuous specimens which are most representative of flat slabs. In tests with constant eccentricity,
the ultimate column load is limited by the least of the punching and flexural resistances.
Conversely, in continuous specimens, like flat slabs, the punching shear force can increase
subsequent to the moment at the column face reaching its ultimate resistance due to support
moments being redistributed back into the span (Soares and Vollum, 2016a). It is notable that of the
56 tests listed above, 38 were on slabs with thickness of 120 mm or less, eight on 152 mm thick
slabs and only ten (Regan, 1993) on 200 mm thick slabs. This is significant due to the ‘size’ effect,
which causes the shear stress at failure of geometrically similar specimens to reduce with increasing
slab depth, as described in section 2.1.4. Only slabs with thickness of 120 mm or greater are
considered in the strength assessments of this research due to difficulties in assessing the
contribution of the size effect to the strength of thinner slabs.
(a) (b)
Figure 5.1 – Geometry and loading arrangement of (a) isolated slabs of El-Salakawy et al. (1998)
and (b) continuous slabs of Regan (1993) (dimensions: mm).
Stamenković and Chapman (1974) were amongst the first to systematically examine the interaction
of punching and flexure at internal, edge and corner connections. They found the interaction to be
164
almost linear at internal connections and edge connections when the axis of the unbalanced moment
is perpendicular to the slab edge, Figure 5.2a. Significantly, they found the interaction to be almost
square for ‘normal’ moments where the axis of the unbalanced moment is parallel to the slab edge,
Figure 5.2b.
(a) (b)
Figure 5.2 – Punching and flexure interaction graph for flat slabs on a) internal and edge
connections with unbalanced moment axis perpendicular to the slab edge, and b) edge column with
‘normal’ moment (Stamenković and Chapman, 1974).
In Figure 5.2, V is the shear force, Vu is the calculated ultimate shear load when external moment
about an axis perpendicular to the slab edge (Figure 5.2a) or parallel to the slab edge (Figure 5.2b)
M=0, M is the resisting moment and Mu is the calculated ultimate resisting moment when the shear
force V=0. Subsequently, Regan (1981, 1999) and Moehle (1988) determined conditions under
which the interaction between normal bending and punching can be neglected at edge columns.
Their recommendations form the basis of the design provisions for punching at edge columns in
EC2 (BSI, 2004a) and ACI 318-14 (2014), respectively. The critical design case for punching shear
in flat slabs dimensioned in accordance with UK practice is commonly the maximum possible
design punching resistance VR,max at edge columns. Both EC2 (BSI, 2004a) and fib Model Code
2010 (MC2010) (fib, 2013) limit the maximum possible punching resistance VR,max to a multiple of
the shear resistance provided by the concrete alone Vc.
According to EC2 and MC2010, Vc can be increased by increasing the provided area of flexural
reinforcement over that required for strength. MC2010 attributes the increase in Vc to a reduction in
the rotation of the critical shear crack which increases the shear resistance provided by aggregate
interlock. As well as enhancing Vc, providing surplus flexural reinforcement normal to the slab edge
165
potentially increases the moment transferred to the column at the design ultimate load which is
detrimental due to the consequent increase in maximum shear stress. Unlike EC2, MC2010 relates
punching resistance at edge columns with normal moments, and inwards eccentricity, to the
eccentricity of the shear force. ACI 318 also relates punching resistance at edge columns to the
eccentricity of the shear force. However, under certain circumstances it allows the interaction
between punching resistance at edge columns and normal moments to be neglected. This chapter
examines the case for neglecting the interaction between punching resistance and unbalanced
moments in design, the influence of flexural continuity on punching resistance, and the effect on
punching strength of providing surplus flexural reinforcement in continuous flat slab-edge column
subassemblies. The studies considered in this chapter are carried out for flat slabs without shear
reinforcement.
5.2 INTERACTION BETWEEN PUNCHING RESISTANCE AND UNBALANCED MOMENT
The interaction between punching resistance and unbalanced moment is assessed in Figure 5.3
using the data summarized in Table 5.1, which includes all the slabs considered by Moehle (1988)
with depths of at least 120 mm not subject to inelastic load reversals. Figure 5.3 shows the
interaction between / and / where is the ultimate bending moment across
the panel width at the inner column face, is the moment of resistance of the slab column
connection at the column face, is the measured punching strength and is the punching
∗
resistance calculated with EC2 using its reduced perimeter (refer to Figure 2.37 in section
2.4.3). is calculated assuming reinforcement to be effective if placed within an effective
width of be = c2+2y (Figure 2.34b) as commonly adopted in UK practice (ISE, 2006) and is limited
to a maximum of = 0.255( + ) / (with = 1) in accordance with Annex I of
EC2.
166
1.8
1.6
1.4
1.2
V/Vshear 1
0.8
0.6
Regan (1993) continuous
0.4
Continuous
0.2 Isolated
EC2
0
0 0.5 1 1.5 2
Mcf/Mflex c2+2y with Mtmax
Figure 5.3 - Interaction between punching resistance and unbalanced moment at column face for
specimens in Table 5.1 with Mflex calculated using effective width of c2 + 2y with Mflex≤ Mtmax.
was calculated from statics assuming the support reaction to be linearly distributed around the
simply supported edges of the isolated test specimens. Although approximate, FEA shows this
procedure to be reasonable. The development of at the column face does not lead to flexural
collapse of continuous slabs if the span reinforcement is still elastic. However, all the slabs in Table
5.1 are reported as failing in punching, typically subsequent to yielding of flexural reinforcement at
the column face. Consequently, all the points in Figure 5.3 correspond to punching failure.
Figure 5.3 shows that, although punching resistance reduces with eccentricity, the design punching
resistance can be safely assumed to be independent of eccentricity if limited by . Table 5.1
shows / and / for calculated with be = c2+2y omitting the EC2 Annex I
limit on without which EC2 overestimates the strength of Hawkins and Corley’s (1974)
specimen CN1. Cases where the punching resistance is calculated to have been limited by flexure at
the column face are highlighted in bold. There are no significant differences between the punching
resistances of isolated and continuous specimens evident in Figure 5.3, which suggests that it is
unnecessary to differentiate between the two types of specimen when evaluating test data.
167
Table 5.1 - Database of tests. Edge column tests without shear reinforcement.
Experimental EC2
c1 c h d fc fy d Vtest Mtest
Author Slabs
mm mm mm mm MPa MPa % kN kNm
c2+2y c2+2y
XXX 250 250 120 89 33.0 545 0.79 125 38 1.01 1.35 1.25
El-Salakawy et al.
(1998) Isolated HXXX 250 250 120 89 36.5 545 0.79 69 46 1.35 1.63 0.67
CN1 305 305 152 121 22.7 459 1.29 111 61 0.81 0.97 0.61
Hawkins and Corley
(1974) Isolated DN1 305 203 152 121 22.6 425 1.38 101 55 1.01 1.21 0.58
JS1 254 254 152 122 43.2 421 0.75 141 60 1.67 2.10 0.79
Mortin and Ghali
(1991) Isolated JS4 254 254 152 122 32.2 421 1.01 141 60 1.29 1.63 0.79
EX-S1 300 300 120 99 35.6 394 1.07 167 50 0.98 1.38 1.14
EX-S2 225 225 120 99 31.7 394 1.07 155 47 1.23 1.72 1.25
Sherif et al.
(2005) Isolated EX-S3 150 150 120 99 31.0 394 1.07 95 29 1.13 1.59 0.90
EX-S4 265 150 120 99 31.0 394 1.07 150 45 1.19 1.68 1.26
EX-S5 150 375 120 99 31.0 394 1.07 172 52 1.34 1.89 1.31
Z-IV(1) 178 178 152 121 27.3 476 1.30 122 45 0.94 1.15 0.77
Z-V(1) 267 267 152 121 34.3 474 1.30 215 85 0.98 1.27 1.08
Z-V(2) 267 267 152 121 40.5 474 1.70 247 94 0.90 1.18 1.07
Zaghlool Z-V(3) 267 267 152 117 38.7 475 2.08 268 104 0.82 1.07 1.15
(1971) Isolated Z-V(4) 267 267 152 121 35.0 437 1.30 0 81 1.31 1.31 0
Z-V(5) 267 267 152 121 35.2 476 1.30 279 0 0.19 0.00 1.39
Z-V(6) 267 267 152 121 31.3 476 1.30 117 88 1.10 1.34 0.61
Z-VI(1) 356 356 152 121 26.0 476 1.30 265 107 0.98 1.38 1.29
SE1 300 200 125 98 35.7 480 1.03 198 40 0.60 0.99 1.52
SE2 300 200 125 101 43.7 480 0.47 192 34 1.53 2.76 1.70
SE4 200 300 125 98 27.4 480 1.03 152 31 0.49 0.80 1.27
SE5 200 300 125 98 44.2 500 0.78 164 39 0.94 1.40 1.28
Regan et al. SE6 200 300 125 99 32.0 500 0.57 149 28 1.03 1.77 1.42
(1979) Continuous SE7 200 300 125 99 39.6 500 0.81 129 32 0.94 1.37 1.02
SE8 300 100 125 98 41.6 480 0.88 136 34 0.88 1.30 1.15
SE9 250 250 125 98 41.4 480 0.73 123 36 1.59 2.19 1.00
SE10 250 250 125 98 40.7 480 0.73 114 36 1.65 2.21 0.93
SE11 250 250 125 98 50.0 480 0.73 138 40 1.74 2.41 1.06
1(1) 300 300 200 168 35.4 507 0.56 282 118 0.81 1.18 1.04
1(2) 300 300 200 168 35.4 507 0.56 264 138 1.03 1.37 0.97
2(1) 300 300 200 168 35.4 507 0.56 256 124 0.91 1.24 0.94
2(2) 300 300 200 168 35.4 507 0.56 285 129 0.92 1.29 1.05
Regana 3(1) 300 300 200 168 41.0 507 0.56 416 73 0.40 0.72 1.46
(1993) Continuous 3(2) 300 300 200 168 41.0 507 0.56 233 148 1.16 1.46 0.82
4(1) 300 300 200 165 42.7 507 0.66 289 149 0.88 1.17 0.98
4(2) 300 300 200 168 42.7 507 0.45 281 111 1.12 1.66 1.05
5(1)b 300 300 200 168 38.4 507 0.61 327 84 0.36 0.86 1.25
5(2)c 300 300 200 168 38.4 507 0.60 234 86 1.10 1.86 1.00
Sherif and Dilger S1-2 250 250 150 114 28.0 444 1.34 185 44 0.38 0.80 1.11
(2000a,2000b)
Continuous EC1(T2) 250 250 150 114 84.1 532 1.61 245 103 0.76 1.08 0.96
Note: a (1) denotes end 1 and (2) end 2 of continuous slab

b
y = 150 mm
c
y = 0 mm
d
= , where and are the longitudinal (perpendicular to the slab edge) and transverse (parallel to the slab edge) flexural
reinforcement ratio, respectively.
168
5.3 NUMERICAL INVESTIGATION
Nonlinear finite element models were developed of two isolated slabs tested by El-Salakawy et al.
(1998) and five continuous slab tests of Regan (1993). These tests were chosen because they were
highly instrumented, with reinforcement strains and rotations being measured. Details of the
specimens are given in Figure 5.1 and Table 5.1.
The objective of the NLFEA was to gain insight into: (a) the shear stress distribution in the slab
around the ACI 318 critical section; (b) the proportion of unbalanced moment resisted by eccentric
shear; and (c) the influence of flexural continuity on punching resistance. The main difference
between isolated and continuous edge column punching tests is that the eccentricity M/V is typically
constant in tests on isolated specimens, but varies with load in continuous specimens owing to
redistribution of bending moment between the span and support. The influence of flexural
continuity on punching resistance is investigated in a parametric study of specimens geometrically
similar to those tested by Regan (1993). The studies investigate the influences of: (a) redistributing
reinforcement between the span and support; and (b) providing surplus flexural reinforcement.
Comparisons are made between the punching resistances given by NLFEA, EC2, and MC2010.
5.3.1 Material modelling
The NLFEA was carried out with Diana 9.6 (TN0, 2014) and Atena (Červenka et al., 2016). Firstly,
the modelling and results from the analysis carried out with Diana will be discussed. The Atena
NLFEA is shown and discussed later in this Chapter. The following sections will give specific
details of both type of specimens namely isolated, from El-Salakawy et al. (1998), and continuous,
from Regan (1993). That description will also include a few sensitivity studies carried out in order
to obtain a reasonable numerical representation of the structures. The ideal choices of parameters
were fairly similar for both type of specimens.
Concrete was modelled with Diana’s ‘total strain fixed crack model’, which evaluates the stress-
strain relationship in the directions of the principal axes at first cracking, which is governed by a
tension cut-off criterion. The fracture energy based Hordijk (1991) model was used to simulate
concrete tensile behaviour after cracking. The Thorenfeldt model (Thorenfeldt et al., 1987) was
used for concrete in compression in conjunction with the four-parameter Hsieh-Ting-Chen (Chen,
1982) failure surface, which models the increase in concrete compressive strength with increasing
isotropic stress. The reduction in concrete compressive strength due to lateral cracking was
modelled following Vecchio and Collins (1993). After cracking, the shear stiffness was reduced by
a constant shear retention factor of 0.1. The concrete elastic modulus Ec and tensile fracture energy
169
/
Gf were calculated in accordance with fib Model Code 1990 (CEB-FIP, 1993) as = 10000
.
and = ( /10) , where depends on the maximum aggregate size and fc is in MPa.
Measured concrete tensile strengths were used. The reinforcement bars were modelled with fully
bonded embedded reinforcement bars, which do not have degrees of freedom of their own. More
details on the FE assumptions and equations can be found in Chapter 3.
5.4 ISOLATED SLABS OF EL-SALAKAWY ET AL. (1998)
El-Salakawy et al. (1998) tested two isolated punching specimens, without shear reinforcement,
denoted as XXX and HXXX, with geometry and test arrangement as shown in Figure 5.1 and
Figure 5.4. The slabs were tested upside-down, simply supported on an arrangement of steel I-
beams along the three edges. The concrete cylinder strengths of XXX and HXXX were similar, at
33 MPa and 36.5 MPa, respectively. The tensile strengths were 3.38 and 3.36 MPa, respectively.
The height of the column above and below the slab was 700 mm. The average tension
reinforcement ratio was 0.0075 parallel and perpendicular to the long edge using 11.3 mm diameter
bars. The average compression reinforcement ratio equalled 0.0045 in both directions using 7.0 mm
diameter bars. Figure 5.5 shows the tension and compression reinforcement arrangement. The M/V
ratio about the column centreline was 0.3 m for XXX and 0.66 m for HXXX.
Figure 5.4 – Test arrangement of El-Salakawy et al. (1998) (Dimensions in mm).
170
(a)
(b)
Figure 5.5 – Reinforcement arrangement for El-Salakawy et al. (1998) specimens: (a) tension
reinforcement and (b) compression reinforcement.
The slab, column and steel plates were modelled with twenty-node isoparametric solid elements as
shown in Figure 5.6. The mesh size was chosen based on the sensitivity study developed in section
4.3.2.1 for interior columns. Eight rows of solid elements were provided through the slab thickness
171
with plan dimensions of 50×50 mm. A 3 × 3 × 3 integration scheme was adopted for solid elements.
The mid-surface of the slab was meshed with non-structural composed elements in order to obtain
the generalised moment and force distribution within the slab. The reinforcement was modelled
with embedded bars assuming perfect bond. LFEA with shell elements was carried out with eight
node quadrilateral isoparametric curved shell elements using the same mesh arrangement on plan as
for the NLFEA. Columns were modelled with solid elements as in the NLFEA. Isotropic elasticity
was assumed in the LFEA with Poisson’s ratio equal to 0.2.
Figure 5.6 - Finite-element mesh used for analysis of slabs of El-Salakawy et al. (1998) with solid
elements.
Similarly to section 4.3.2, the following presents a few comparisons between experimental and FE
results with differing assumptions in the smeared crack method, modelling of compression and
shear behaviours, and iterative solution procedures. The studies were developed when the FE model
was being built so as to get the best numerical representation of the structure.
5.4.1.1 Smeared crack methods

Figure 5.7 compares experimental and NLFEA results from the representative specimen XXX
tested by El-Salakawy et al. (1998) modelled with both fixed and rotating crack models. V depicts
the measured failure load. Details of the smeared crack methods used in Diana are given in section
3.2.2.2. As shown in Figure 5.7, the rotating crack model did not give good predictions for slab
XXX of El-Salakawy et al. (1998), with the analysis stopping at around 50% of the measured
172
failure load. A fixed crack model was adopted in the subsequent sensitivity studies for specimen
XXX.
XXX XXX
-16
140
-14
120
-12
Failure Load
Column Reaction (kN)

V 100 (Experimental)
Deflection (mm)
-10
Experimental
80
-8
0.56V 60 Fixed
-6
-4 40 Rotating
-2 20
Column
0 0
0 200 400 600 800 1000 0 5 10 15 20
Slab (mm) Column stub (mm)
Figure 5.7 – Comparison between different crack models for XXX (El-Salakawy et al., 1998).
5.4.1.2 Compressive behaviour

Figure 5.8 presents a comparison between experimental results and FE models of XXX by El-
Salakawy et al. (1998) with solid elements and either Thorenfeldt or Parabolic curves for
compressive behaviour (refer to section 3.2.3.2). The analysis with Parabolic stopped at around
87% of the measured failure load, whereas analysis with the Thorenfeldt model gave excellent
predictions for both the deformed shape and failure load of XXX. Consequently, the Thorenfeldt
model (Thorenfeldt et al., 1987) was adopted in all NLFEA based on this model.
XXX XXX
-16
140
-14
120
-12
V 100 Failure Load
Deflection (mm)
-10 (Experimental)
80 Experimental
-8
60
-6 Thorenfeldt
0.86V 40
-4 Parabolic
20
-2 Column
0
0 0 5 10 15 20
0 200 400 600 800 1000
Column stub (mm)
Slab (mm)
Figure 5.8 – Comparison between different compressive curves for XXX (El-Salakawy et al.,
1998).
173
5.4.1.3 Shear behaviour

Figure 5.9 shows the effect of varying the shear retention factor on the deflected shape (at measured
failure load) and column stub displacement for specimen XXX tested by El–Salakawy et al. (1998),
which failed in a brittle sudden punching mode. As can be seen, the impact on shear stiffness is very
pronounced. A value of β=0.1 was adopted for all the analysis with constant shear retention factor
where both slab and column were modelled with solid elements.
XXX XXX
-16 Failure Load
(Experimental)
-14 200 Experimental
-12 β=0.1

150
Deflection (mm)
-10 β=0.2
-8 β=0.5
100
-6
β=0.7
-4 50 β=0.9
-2
Column
0 0
0 200 400 600 800 1000 0 10 20 30 40 50
(a) (b)
Figure 5.9 – Deflected shape and column stub displacement of XXX (El-Salakawy et al., 1998) with
different shear retention factors.
5.4.1.4 Iterative procedure

Figure 5.10 presents a comparison between experimental and FE results for XXX from El-Salakawy
et al. (1998) where the only variable is the choice of iterative solution algorithms between a regular
Newton-Raphson and Secant (Quasi-Newton). Both analysis adopted an energy based convergence
criteria with a tolerance of 10-3 as presented in section 3.2.7, a maximum of a 100 number of
iterations, and an initial load step of 5% of the measured failure value with automatic adaptive
increments after a limit loading, based on the initial step, is reached.
Figure 5.10 shows that the choice of iterative solution procedure had a significant impact on the
ultimate load for the analysed case, since the analysis with a regular Newton-Raphson diverged at
around 45% of the measured ultimate load. As a consequence of the good results shown in Figure
5.10, a Quasi-Newton (Secant) method was adopted for all the analysis derived from this model.
174
-16
XXX 140 XXX
-14
120
-12

V 100
-10
Deflection (mm)
80
-8
60 Failure Load (Experimental)
-6 Experimental
-4 40 Secant
0.46V Newton-Raphson
-2 20
Column
0 0
0 200 400 600 800 1000 0 5 10 15 20
Figure 5.10 – Comparison between iterative procedures for XXX (El-Salakawy et al., 1998).
5.5 INFLUENCE OF ECCENTRICITY
The NLFEA gave very good estimates of the measured deflections from El–Salakawy et al. (1998)
and El–Salakawy et al. (1999) as shown in Figure 5.11 for XXX and Figure 5.12 for HXXX in
comparison to the available measurements. Figure 5.11 also shows how rotations relative to the
column were calculated at Longitudinal ( = + ) and Transverse ( = ,
since = 0 due to symmetry) directions from the deflected shape. Once the NLFEA was
validated, a parametric study was carried out to determine the effect of loading eccentricity on the
punching resistance of XXX. Eccentricities M/V of 0.2 m, 0.3 m (XXX), 0.4 m, 0.5 m, 0.6 m and
0.66 m (HXXX) were modelled. The concrete properties of XXX were used, except for M/V = 0.66
m, where the properties of HXXX were adopted. The results of the analyses were used to evaluate
the MC2010 punching shear provisions at edge columns. According to MC2010, punching
resistance is governed by the maximum rotation ψ of the slab relative to the column about the two
principal axes of the slab. The critical axis is parallel to the slab edge for the analysed specimens as
shown in Figure 5.13.
XXX XXX
Deflections (Longitudinal) Deflections (Transverse)
-15 Experimental 0
-13 NLFEA -2
Deflection (mm)
Deflection (mm)
-11 Column
-4
-9 Column -6
-7 Experimental
-8 NLFEA
-5
-10 Column
-3 T
-12
-1 TColumn =0
-14
0 250 500 750 1000 -770 -385 0 385 770
r (mm) r (mm)
(a) (b)
175
XXX
Column stub
160

140
120
100
80
40 Experimental
20 NLFEA
0
0 5 10 15 20 25
Deflection (mm)
(c)
Figure 5.11 - Comparison of measured and predicted deflections for slab XXX (El-Salakawy et al.,
1998) for: (a) the longitudinal direction at failure load, (b) the transverse direction at failure load,
and (c) the column stub for varies loads.
HXXX HXXX
Deflections (Longitudinal) Column Stub
80
-7
Experimental 70
-5 Column
60
NLFEA
Deflection (mm)
-3 50
40
-1 Failure Load (Experimental)
30
1 Experimental
20
NLFEA
3 10
5 0
0 250 500 750 1000 0 2.5 5 7.5 10
r (mm) Deflection (mm)
(a) (b)
Figure 5.12 - Comparison of measured and predicted deflections for slab HXXX (El-Salakawy et
al., 1998) for: (a) the longitudinal direction at failure load, and (b) the column stub for varies loads.
Figure 5.14a to f compare the following rotations for all eccentricities:
i. test data for XXX and HXXX;
ii. NLFEA used for MC2010 Level IV (LoA IV);
iii. MC2010 Level II (LoA II) with ms from equation (2.48);
iv. MC2010 Level III (LoA III) with ms and rs from LFEA with shell elements.
Figure 5.14 also shows shear resistances calculated with Equation (2.38) of MC2010, with =
89 , = 16 , = 1, and rotations from Equation (2.45) with = 1.5 (corresponding to
LoA II). The reduction factor for eccentric shear was given by:
v. = 0.7;
vi. Equation (2.37);
176
vii. = / with and from LFEA. Both and were

calculated from the shear force distribution over the control perimeter located at
0.5d from the column. The values were averaged between adjacent nodes on plan
25 mm apart from one to another, which corresponded to approximately one-
quarter of the average effective depth. No reduction in torsion was adopted, but the
averaging between nodes helped reducing the shear forces which are mesh sensitive
and overestimated by FEA, especially at slab edge.
XXX XXX
e = 0.20 m e = 0.30 m
200 200

150 150
100 100
Longitudinal
50 50 Longitudinal
Transverse
Transverse
0 0
0 0.01 0.02 0.03 0.04 0 0.02 0.04
Rotation Rotation
XXX XXX
200 200
e = 0.40 m e = 0.50 m
150 150
100 100
50 Longitudinal 50 Longitudinal
Transverse Transverse
0 0
0 0.02 0.04 0.06 0 0.02 0.04 0.06
Rotation Rotation
XXX HXXX
e = 0.60 m e = 0.66 m
100 100
80 80
60 60
40 40
Longitudinal Longitudinal
20 20
Transverse Transverse
0 0
0 0.02 0.04 0.06 0 0.02 0.04
Rotation Rotation
Figure 5.13 – Comparison between NLFEA longitudinal and transverse slab rotations relative to the
column for El-Salakawy et al. (1998) specimens.
177
Resistance (ke = 0.7) e = 0.30 m Resistance (ke = 0.7)
e = 0.20 m
Resistance (ke=1/[1+(e'/bu)]) Resistance (ke=1/[1+(e'/bu)])
200
Resistance (ke = Shell LFEA) Resistance (ke = Shell LFEA)
200 Rotation - MC2010 LII 180 Rotation - MC2010 LII
180 Rotation - MC2010 LIII 160 Rotation - MC2010 LIII
Rotation - Test
NLFEA

160 NLFEA 140
140 120 Vtest
120 EC2
EC2 100
100
80
80
60
60
40 40
20 20
0 0
0 0.02 0.04 0.06 0 0.01 0.02 0.03 0.04 0.05
Rotation (radians) Rotation (radians)
(a) (b)
e = 0.40 m Resistance (ke = 0.7) e = 0.50 m Resistance (ke = 0.7)
Resistance (ke=1/[1+(e'/bu)]) Resistance (ke=1/[1+(e'/bu)])
160
Resistance (ke = Shell LFEA) Resistance (ke = Shell LFEA)
200
Rotation - MC2010 LII 140 Rotation - MC2010 LII
180
Rotation - MC2010 LIII 120 Rotation - MC2010 LIII
160

NLFEA NLFEA
140 100
120
80
100
60 EC2
80 EC2
60 40
40 20
20
0
0 0 0.02 0.04 0.06 0.08
0 0.02 0.04 0.06 0.08
(c) (d)
178
e = 0.60 m Resistance (ke = 0.7)

Resistance (ke=1/[1+(e'/bu)])
160
Resistance (ke = Shell LFEA)
140
Rotation - MC2010 LII
120
Rotation - MC2010 LIII

100 NLFEA
80
60
EC2
40
20
0
0 0.02 0.04 0.06 0.08
Rotation (radians)
(e)
160 e = 0.66 m Resistance (ke=0.7)
140 Resistance (ke = Shell LFEA)
120 Rotation - MC2010 LII


100
NLFEA
80
Rotation - Test
Vtest
60
EC2
40
20
0
0 0.02 0.04 0.06 0.08
Rotation (radians)
(f)
Figure 5.14 - Assessment of slabs of El-Salakawy et al. (1998) with MC2010 for eccentricities of:
(a) 0.2 m; (b) 0.3 m (XXX); (c) 0.4 m; (d) 0.5 m; (e) 0.6 m; and (f) 0.66 m (HXXX).
Figure 5.14b and 5.14f show reasonable agreement between the measured and NLFEA rotations
and failure loads, particularly for XXX. The MC2010 failure load is given by the intersection of the
resistance and rotation curves. Figure 5.14 shows that the resistances calculated with ke from LFEA
above are very similar but slightly less than those with ke from equation (2.37). Equation (2.45)
gives reasonable estimates of rotation up to near failure if ms is calculated with LFEA (LIII), which
takes the absolute values of bending and torsion ( =| |+ ) in accordance to Wood
(1968) at the face of the column at a width equal to the flexural reinforcement effective width.
Rotations are significantly overestimated if ms is calculated with equation (2.48) (LII).
Figure 5.15 shows the influence of eccentricity on failure loads calculated with NLFEA, EC2 and
MC2010 Levels II to IV. The Level IV resistances were calculated with NLFEA rotations and ke
from (v) to (vii) used in Figure 5.14 and previously described. The Level IV resistances are closest
179
to the test and NLFEA results when calculated with ke = 0.7. The Level III and IV predictions with
ke from equation (2.37) are similar and reasonable, but Level II is overly conservative.
Consideration of Figure 5.14 shows that Level III is less conservative than Level IV if ke = 0.7.
∗
According to EC2, the failure load is the least of Vshear calculated using and Vflex, which is
calculated in Figure 5.15 assuming that Mcf equals Mflex calculated with be = c2+2y (refer to Figure
2.34b). EC2 is seen to give conservative estimates of the strengths of XXX and HXXX, when
limited by Mflex at the column face, and is considerably simpler to implement than MC2010.
EC2 Vshear
160 Vflex
Test
140
NLFEA
120
MC2010 LIV ke LFEA
100 MC2010 LIV ke=0.7
80 MC2010 LIV ke=1/[1+(e'/bu)]
MC2010 LIII ke=1/[1+(e'/bu)]

60
MC2010 LII ke=1/[1+(e'/bu)]
40
20
0
0 100 200 300 400 500 600 700
Eccentricity (mm)
Figure 5.15 - Influence of eccentricity on calculated punching resistance of slabs of El-Salakawy et
al. (1998).
5.6 INVESTIGATION OF SHEAR FORCES ON ACI 318 CRITICAL SECTION
Both MC2010 and ACI 318 allow punching resistance to be calculated on the basis that failure
occurs when the peak shear stress reaches the available shear resistance. MC2010 allows the peak
shear stress to be determined with LFEA, whereas ACI 318 calculates the peak shear stress using
equation (2.31). This section compares the shear stresses given on the ACI 318 critical perimeter by
NLFEA and ACI 318.
Figure 5.16 shows the shear force distribution along the ACI 318 critical section (refer to Figure
2.39) in N/mm for XXX and HXXX. The shear force were derived using linear elastic FEA with
both shell elements and brick elements, as well as NLFEA with brick elements. The FEA shear
forces in Figure 5.16 are averages between adjacent nodes on plan which are at 25 mm centres. As
previously discussed, averaging was carried out to smooth shear force distribution and reduce the
calculated shear force at the slab edge, which is mesh sensitive and overestimated by FEA.
180
Physically, the averaging width of 25 mm is approximately one-quarter of the average effective

depth d. The peak shear forces per unit length given by FEA are seen to be model dependent,
suggesting that peak shear stress is not a good failure criterion.
XXX
Shear Distribution
400
300
200
100
Shear (N/mm)
0 Brick - Linear
-100 Brick - Nonlinear
-200 Shell - Linear
-300 ACI 318 γv NLFEA

-400 ACI 318 γv ACI
-500 ACI resistance
-475 -237.5 0 237.5 475
Control Perimeter (mm)
(a)
HXXX
Shear Distribution
800
600
400
Shear (N/mm)
200
Brick - Linear
0
Brick - Nonlinear
-200 Shell - Linear
ACI 318 γv NLFEA
-400
ACI 318 γv ACI
-600 ACI resistance
-475 -237.5 0 237.5 475
Control Perimeter (mm)
(b)
Figure 5.16 - Analysis of shear force distribution on ACI 318 critical perimeter for slabs of El-
Salakawy et al. (1998): (a) shear force distribution for XXX; (b) shear force distribution for HXXX.
Figure 5.16 also shows shear forces per unit length calculated using Equation (2.31) with v from (a)
Equation (2.36) as recommended by ACI 318 and (b) NLFEA. Both give peak shear forces greater
than the resistance calculated in accordance with ACI 318. These observations are consistent with
the conclusions of Moehle (1988), who suggested that the peak shear stress, which depends
181
significantly on the method of calculation, is not a good measure of punching failure at edge
columns.
For each analysis, the proportion of unbalanced moment resisted by eccentric shear v was
calculated on the ACI 318 critical section (Figure 2.39), at the NLFEA failure loads, with LFEA
and NLFEA using both solid and shell elements. The addition in moment due to eccentric shear is
illustrated in Figure 2.36, extracted from Regan (1981), by the increment from points C to B. For
each FE model, v was the ratio between the moment calculated from the shear forces per unit
length shown in Figure 5.16 and the applied moment calculated by multiplying the shear force at
the column by its distance to the centroid of the control perimeter, the position of which is depicted
as line c-c in Figure 2.40.
Figure 5.17 shows the variation in v with eccentricity for the above and v from equation (2.36).
Significantly, equation (2.36) underestimates the proportion of unbalanced moment resisted by
eccentric shear, which is greatest for the NLFEA, as found by Gayed and Ghali (2008), who
concluded that ACI 318 is incorrect to allow v = 0. Conversely, the author consider taking v = 0
when V ≤ 0.75VR0 to be a computational device equivalent to the EC2 practice of taking β as u1/u1*
in equation (2.26), which is supported by Figure 5.3 providing sufficient flexural reinforcement is
placed within be = c2+2y to resist unbalanced moment at the column face and Mcf ≤
0.255(c2+y)fcdn2/c.
0.7
0.65
0.6
ACI
0.55
Brick NLFEA
0.5
v
Brick LFEA
0.45 Shell NLFEA
Shell LFEA
0.4
0.35
0.3
200 300 400 500 600
Eccentricity (mm)
Figure 5.17 - Influence of eccentricity on γv on ACI 318 critical perimeter for XXX slabs of El-
Salakawy et al. (1998).
182
5.7 CONTINUOUS SPECIMENS OF REGAN (1993)
Regan (1993) tested five full-scale continuous edge column punching shear specimens. The slabs
were 200 mm thick, 3000 mm wide and 5784 mm long with 300 mm square columns at the centres
of their short edges as shown in Figure 5.1. The maximum aggregate size was 20 mm. The bottom
longitudinal flexural reinforcement consisted of 23 number 12 mm diameter bars with an effective
depth of 174 mm which equates to a reinforcement ratio of 0.5%. The cover was 20 mm to the top
and bottom longitudinal reinforcement which was in the outer layer. The flexural tension
reinforcement ratio normal to the short slab edge, within a width be = 2c1 + c2 centred on the
column, sup was 0.8% in slabs S1 to S3, 1.0% in S4 end 1 and 0.5% in S4 end 2.
The reinforcement sup was provided in the form of U bars and distributed over a width of 500 mm
in slabs S1 to S3 and S4 end 2. In S4 end 1, sup was distributed over a width of 600 mm. The U bar
diameter was 12 mm with the exception of S4 end 1 where 16 mm diameter bars were used. No
other longitudinal hogging reinforcement was provided within 500 mm of the column centreline.
The top transverse reinforcement at the ends of the slabs consisted of 10 number 12 mm diameter
bars positioned over a width of 1350, with three bars passing through the column. There were five
different arrangements of bottom transverse reinforcement within the support strip as shown in
Figure 5.18. Top transverse reinforcement consisting of 14 number 8 mm diameter bars were
positioned over the central region of the slab.
Figure 5.19 shows a plan view of the reinforcement distribution. Further details of the slabs are
summarised in Table 5.2 including the measured failure loads and the ratio of the moment of
resistance at the column face McfR to the ultimate elastic moment Mcfel across the slab width at the
column face. The ultimate elastic moment Mcfel was derived from reactions calculated with FEA at
the measured failure load. The moment of resistance at the column face McfR was calculated in
terms of the flexural reinforcement provided within a width be = c2 + 2y centred on the column. The
elastic moment is typically greater than the moment of resistance indicating that moment
redistribution is implicit in the calculation of reinforcement design moments. Table 5.2 also gives
concrete cylinder strengths, calculated as 0.8 times the cube strength of specimens cured with the
slabs, and reinforcement yield strengths.
183
Slab 1 Ends 1 & 2 Slabs 2 & 3 Ends 1& 2
Slabs 1,3 & 4 Ends 1& 2 Slab 2 End 1
Slab 2 End 2
Slab 4 End 1 Slab 4 End 2
184

Figure 5.18 - Reinforcement details for Regan slabs (dimensions in mm)
Figure 5.19 – Plan view of the reinforcement distribution for Regan slabs (dimensions in mm).
The two ends of each slab were tested separately, with the column providing the support at one end
and the other end simply supported across the slab width just inside from the column face. With the
exception of slab S3, at each loading stage equal and opposite horizontal forces were applied to the
column at the end under test “so as to keep the column free of rotation” (Regan, 1993). In slabs S1
185
and S2 with sup = 0.8%, the column transfer moments had become constant before the bottom steel
yielded in the span and punching occurred. In slab S4 end 1, with sup = 1.0%, the span steel had not
visibly yielded when punching occurred. Midspan yield was well developed in slab S4 end 2, with
sup = 0.5%, at punching failure. In test S3 end 1, the slab was loaded to failure with 4 equal loads
positioned at the loading points closest to the supported column (points a to d in Figure 5.1b).
The effect of this was to reduce the eccentricity M/V perpendicular to the slab edge, as well as the
maximum span moment for a given total applied load relative to that for the standard loading
arrangement. In the case of slab S3 end 2, a relatively small vertical load was applied, distributed as
for slabs 1 and 2, with the column prevented from rotating. The vertical load was then held constant
and the column moment increased until a plastic hinge formed at the slab column connection. The
vertical load was then increased to failure whilst maintaining the column moment. Specimen S5
investigated the influence of the column being outboard of the slab. At end 1, the centreline of the
column was flush with the slab edge whereas at edge 2 the inner face of the column was flush with
the slab edge. End 1 of slab S5 was loaded following the general procedure until hinges formed at
the slab-column connection and in the span. Subsequently, the specimen was unloaded and reloaded
with all the actuators applying a constant load of 25 kN apart from those at the slab edge (points a
and b in Figure 5.1b) where the loads were increased until failure.
Table 5.2 – Details of Regan Slabs
Experimental Calculated
c c h d f f  (1) Vtest Mtest Mcf R(2)
Author Slabs 1 2 c y
( )
mm mm mm mm MPa MPa % kN kNm kNm
1(1) 300 300 200 168 35.4 507 0.56 282 118 100.4 0.63
1(2) 300 300 200 168 35.4 507 0.56 264 138 100.4 0.67
2(1) 300 300 200 168 35.4 507 0.56 256 124 100.4 0.69
2(2) 300 300 200 168 35.4 507 0.56 285 129 100.4 0.62
Regan 3(1) 300 300 200 168 41.0 507 0.56 416 73 101.7 -
3(2) 300 300 200 168 41.0 507 0.56 233 148 101.7 0.77
(1993) 4(1) 300 300 200 165 42.7 507 0.66 289 149 127.3 0.78
4(2) 300 300 200 168 42.7 507 0.45 281 111 66.7 0.42
5(1) 300 300 200 168 38.4 507 0.61 327 84 98.2 -
5(2) 300 300 200 168 38.4 507 0.60 234 86 46.3 -
(1) Calculated in accordance with EC2 punching shear requirements;
(2) McfR is moment of resistance at column face;
(3) Mcfel is elastic moment across slab width at column face
In all specimens, strains were measured in the top bars perpendicular to the slab edge close to the
inner column face. The strain measurements indicate that the bars passing into the column generally
yielded. The strains in the outer bars to either side of the column were typically around 2/3rds of
186
those in the bars going into the column. These bars did not generally yield although yielding
occurred in ends 2 of slabs S3 and S4. All the slabs are reported to have failed in punching.
However, flexural failure appears to have been imminent in both tests of slabs S1 and S2 as well as
end 2 of S4. Further details of the slabs were summarised in Table 5.1, including the punching
resistances Vshear calculated according to EC2.
The same FEA procedures were used as for section 5.4. The slab, columns and steel plates were
meshed with twenty-node isoparametric solid elements. The elements’ sizes were chosen based on a
sensitivity study previously developed in section 4.3.2.1 for internal column connections, the results
from the FEA carried out in section 5.4, and the work of Eder et al. (2010). Symmetry and a
gradated mesh were used to reduce the number of elements, which varied from 75x75 mm to
150x150 mm, with the region around the connection being most refined. Eight rows of solid
elements were provided through the slab thickness. A 3 × 3 × 3 integration scheme was adopted for
the solid elements at the refined region, and a 2 × 2 × 2 scheme was used elsewhere. The mid-
surface of the slab was once again meshed with non-structural composed elements to get the
generalised moment and force distribution within the slab. The final mesh is shown in Figure 5.20.
Figure 5.20 - Finite-element mesh used for slabs of Regan (1993).
This section provides some comparisons between experimental results extracted from Regan (1993)
and FEA of the representative specimen Slab 1 End 1 (refer to Table 5.2) using different modelling
187
assumptions, as carried out in sections 4.3.2 and 5.4.1. The goal was once more to adjust the
numerical model in order to obtain the best representation of the real structure.
5.7.1.1 Smeared crack methods

Figure 5.21 presents the comparison between Fixed and Rotated models of Slab 1 End 1 by Regan
(1993). As for the isolated specimen (refer to Figure 5.7), the rotating model analysis only reached
50% of the measured failure load. A fixed crack model was adopted for all the parametric studies
based on this model.

20 350
10 300

250
Deflection (mm)
0 Experimental
0 1000 2000 3000 4000 5000
200
-10 Failure Load
(Experimental)
150
Fixed
-20
0.50V 100
Rotating
-30 V 50
Column
-40 0
0 0.01 0.02 0.03
Slab (mm) Rotation
Figure 5.21 – Comparison between different smeared crack approaches for Slab 1 End 1 (Regan,
1993).
5.7.1.2 Compressive behaviour
Figure 5.22 shows a comparison between experimental results and FE models of Slab 1 End 1 with
Thorenfeldt (Thorenfeldt et al. 1987) and Parabolic (refer to section 3.2.3) stress strain curves for
concrete in compression. Rotations relative to the column, deflected shape, and failure load are
shown not to be much influenced by the choice between Thorenfeldt and Parabolic compressive
curves. The Thorenfeldt model was chosen for all the studies from this model.
188

20 350
10 300

250
Deflection (mm)
0
0 1000 2000 3000 4000 5000 Experimental
200
-10 Failure Load
150 (Experimental)
-20 Thorenfeldt
100
Parabolic
-30 50
Column
-40 0
0 0.01 0.02 0.03
Slab (mm) Rotation
Figure 5.22 – Comparison between different compressive curves for Slab 1 End 1 (Regan, 1993).
5.7.1.3 Shear behaviour

Figures 5.23 shows the influence of varying the shear retention factor over the slab rotation relative
to the column and deflected shape of Slab 1 End 1 tested by Regan (1993) modelled with solid
elements for both slab and column. It shows an increase in stiffness as β varies from β=0.1 to
β=0.9, with low values such as 0.1 and 0.2 giving good predictions for both deformed shape and
failure load, as suggested by Suidan and Schnobrich (1973). The value adopted for the following
analysis was β=0.1 as for the analyses of slabs tested by El-Salakawy et al. (1998).
Slab 1 End 1 350 Slab 1 End 1

20
300
10 Experimental
Column Reactiond (kN)
250
Failure Load
Deflection (mm)
0 (Experimental)
0 1000 2000 3000 4000 5000 200 β=0.1
-10 β=0.2
150
β=0.5
-20 100
β=0.7
-30 50 β=0.9
Column
-40 0
Slab (mm) 0 0.01 Rotation 0.02 0.03
(a) (b)
Figure 5.23 - Deflected shape and rotations of Slab 1 End 1 (Regan, 1993) with different shear
retention factors.

Figure 5.24 presents a comparison between test data from Slab 1 End 1 of Regan (1993) and FE
results where the only variable was the iterative solution algorithms, which varied from a regular
189
Newton-Raphson to a Secant (Quasi-Newton). Similarly to the NLFEA of specimens XXX and

HXXX of El-Salakawy et al. (1998) (section 5.4.1.4), this analysis adopted an energy based
convergence criteria with a 10-3 tolerance. The maximum number of iterations was 300 on the basis
of good agreement with experimental data. The initial load step was 5% of the test failure load with
automatic adaptive increments after a limit loading, based on the initial step, is reached.
350
20
300
10

250
Deflection (mm)
0 Experimental
200
0 1000 2000 3000 4000 5000
150 Failure Load
-10
(Experimental)
Secant
100
-20
Newton-Raphson
50
-30 Column
0
0 0.005 0.01 0.015 0.02 0.025 0.03
-40 Slab (mm) Rotation
1993).
The Newton-Raphson choice gave slightly greater failure load when compared to the model run
with the Quasi-Newton procedure. A Quasi-Newton method was adopted for all the analysis and
parametric studies carried out from the Regan (1993) study.
5.8 INFLUENCE OF CONTINUITY
Rotations were calculated from the deflected shape (nodal displacements) at positions close to the
inclinometers in the tests by Regan (1993) as shown in Figure 5.25 for the longitudinal and
transverse directions. Rotations calculated from the greatest deflected shape perpendicular and
parallel to the slab edge are depicted Longitudinal and Transverse respectively. For the case shown
in Figure 5.25, the longitudinal slab rotation relative to the column is = − in
which the sign of both rotations is positive. For Regan’s slabs, due to symmetry, the column
rotation for the transverse direction = 0. Consideration of Figures 5.25 and 4.6 show that
rotations should not be calculated too close to the column where the slab deformation can give
unrealistic results. One should look into the deflected shape beforehand in order to identify the
straight line pattern and position of maximum rotation, which was along the slab edge for the
transverse direction of both El-Salakawy (1998) and Regan’s (1993) specimens. Figure 5.26
compares the longitudinal and transverse slab rotations relative to the column in order to identify
the greatest one, as it governs the punching resistance according to MC2010.
190
Longitudinal Transverse
Inclinometers
Figure 5.25 – FE rotations calculated from the deflected shape for the specimens by Regan (1993).
191
350 350

300 300
250 250
200 200
150 150
100 100 Longitudinal
Longitudinal
50 Transverse 50 Transverse
0 0
0 0.01 0.02 0.03 0 0.01 0.02 0.03
Rotation Rotation
350
300
Column Reactiond (kN)

350
250 300
200 250
150 200
100 150
Longitudinal 100 Longitudinal
50 Transverse 50 Transverse
0
0 0.01 0.02 0.03 0
0 0.01 0.02 0.03
Rotation Rotation
350 Slab 4 End 2

300
250
200
150
100
Longitudinal
50 Transverse
0
0 0.01 0.02 0.03
Rotation
Figure 5.26 - Comparison between NLFEA longitudinal and transverse slab rotations relative to the
column for Regan (1993) specimens.
Figure 5.27 compares the measured and predicted slab rotations relative to the column for slab 1 (fct
= 3.36 MPa) end 1, slab 2 (fct = 3.14 MPa) ends 1 and 2 and slab 4 (fct = 3.5 MPa) ends 1 and 2. The
longitudinal slab rotations relative to the column are shown, as these were greatest (refer to Figure
5.26).
Resistance (ke=0.7)
350
Slab 1 End 1
300 Vtest (Test ULS e=M/V)
VEC2 Resistance (ke=1/[1+(e'/bu)])

250 (LFEA e=M/V)
Vflex Experimental
200
NLFEA
150
100
50
0
0 0.01 0.02 0.03
Rotation (radians)
(a)
192
Slab 2 End 1 Resistance (ke=0.7)
350 Slab 2 End 2
350 Resistance (ke=1/[1+(e'/bu)])
300 VEC2 (Test ULS e=M/V)
Vtest
Vtest

250 VEC2 (LFEA e=M/V)
Vflex 250
Experimental
200
Vflex
200
NLFEA
150 150
100 100
50 50
0 0
0 0.005 0.01 0.015 0.02 0.025 0.03 0 0.01 0.02 0.03 0.04

(b) (c)
Resistance (ke=0.7)
350 (Test ULS e=M/V)
VEC2 300
Vtest
300 VEC2 (LFEA e=M/V)
Vtest

250
Experimental
250 Vflex
Vflex 200
200 NLFEA
150
150 Rotation - MC2010 LII
100 100
50 50
0 0
0 0.01 0.02 0.03 0 0.01 0.02 0.03
(d) (e)
Figure 5.27 - Assessment of slabs of Regan (1993) with MC2010 for: (a) slab 1 end 1; (b) slab 2 end 1; (c) slab 2 end 2; (d) slab 4 end 1; (e) slab
4 end 2
193
The NLFEA gives good predictions of the measured rotations and failure loads of the slabs. Figure
5.27 also shows rotations calculated for each test with MC2010 levels II and III. The Level II
rotations were calculated with the measured ultimate eccentricity M/V, which is less than the elastic
eccentricity that would be used in design. The MC2010 resistances with ke from equation (2.37) are
evaluated with both the measured ultimate eccentricity and the elastic eccentricity. The latter are
almost identical to resistances calculated with ke derived from LFEA with shell elements. The
MC2010 failure loads are given by the intersection of the rotation and resistance curves in Figure
5.27. MC2010 Level IV, with rotations from the NLFEA, gives reasonable estimates of the
measured punching resistances, but no better than EC2, if ke = 0.7. However, resistances are
underestimated if ke is calculated with equation (2.37), even if the measured ultimate eccentricity is
used. MC2010 levels II and III significantly underestimate punching resistance particularly if ke is
calculated using the elastic eccentricity. The flexural failure loads in Figure 5.27 were calculated
neglecting strain hardening and the width of loading plates, assuming yielding in the span, and an
ultimate moment at the column face of Mflex calculated with be = c2 + 2y.
5.9 INFLUENCE OF REINFORCEMENT ARRANGEMENT
A series of parametric studies were carried out to investigate the influence of varying the
longitudinal flexural reinforcement in slabs with the same geometry as tested by Regan (1993).
Analyses were carried out with longitudinal hogging reinforcement ratios sup within be = 2c1 + c2
at the column support equal to 0.43%, 0.8%, 1.0%, 1.2% and 1.6%. For each of these ratios, the
longitudinal reinforcement ratio in the span span was taken as 0.25%, 0.5%, which corresponds to
Regan’s slabs, and 1.0%. The resulting load rotation diagrams are shown in Figure 5.28, which also
shows the punching resistance according to MC2010 with ke = 0.7. Despite giving good strength
predictions for the El-Salakawy et al. (1998) and Regan (1993) slabs, the NLFEA appears to
overestimate the punching resistance of the slabs with 1.0% span flexural reinforcement.
However, the NLFEA rotations are considered reasonable on the basis of good comparison with test
results in Figure 5.14 and Figure 5.27. These and other NLFEA validation studies (Soares and
Vollum, 2015) indicate that rotations can be predicted more reliably than punching resistances,
highlighting the benefit of the rotation based failure criteria of MC2010. Figure 5.28 shows that the
slab rotation relative to the column is largely governed by span and is almost independent of sup.
This is significant because EC2 and MC2010 Levels II and III relate punching resistance to sup and
not span.
194
500 ρsup = 0.43%

span = 1.0%
450 ρsup = 0.80%
400 ρsup = 1.0%
ρsup = 1.2%
350
300 span = 0.5% ρsup = 1.6% ρspan = 1.0%
250
Resistance (ke=0.7)
ρspan = 0.25% Test data

200
MC2010 LIII ρsup = 0.43%
150
MC2010 LIII ρsup=0.8%
100
50
0
0 0.01 0.02 0.03 0.04
Rotation (radians)
Figure 5.28 - Influence of varying reinforcement arrangement on rotation of Regan (1993) Slab 1
End 1.
Figure 5.29 illustrates the influence of sup on punching resistance according to MC2010 levels III
and IV. Test results, normalised by (35.4/fc)1/3 in accordance with equation (2.25), are also shown
for the Regan slabs considered in Figure 5.27, for which span = 0.5%. MC2010 Level IV compares
favourably with the test results, but Level III greatly overestimates the influence of sup on punching
resistance.
350
300
250
200 EC2 Vshear
150 MC2010 LIII
MC2010 LIV ρspan = 0.5%

100
MC2010 LIV ρspan = 1.0%
50
Test ρspan = 0.5%
0
0 0.5 1 1.5 2
Support reinforcement ratio sup (%)
Figure 5.29 - Influence of varying reinforcement arrangement on punching resistance of Regan

(1993) Slab 1 End 1.
195
This is largely explained by Figure 5.30, which shows the variation in eccentricity M/V with V for
the analysed slabs. Significantly, M/V is not constant, as assumed in MC2010 Level III, but reduces
with increasing V for all slabs (with the exception of span = 0.25%, which fails in flexure) due to
moment being redistributed into the span once torsional cracking and yielding of flexural
reinforcement occurs at the support. Furthermore, M/V is almost independent of sup for given span.
The underestimate of resistance is compounded if ke is calculated with equation (2.37) using the
elastic eccentricity.
0.8 0.7
0.7
0.6
0.6
0.5
0.5
M/V (m)
M/V (m)
0.4
0.4 ρsup = 0.43% ρspan = 0.25%
ρsup = 0.43% ρspan = 0.50%
ρsup = 0.80% ρspan = 0.25% 0.3
0.3 ρsup = 0.80% ρspan = 0.50%
ρsup = 1.0% ρspan = 0.25%
0.2 0.2 ρsup = 1.0% ρspan = 0.50%
ρsup = 1.2% ρspan = 0.25%
ρsup = 1.2% ρspan = 0.50%
0.1 ρsup = 1.6% ρspan = 0.25% 0.1
ρsup = 1.6% ρspan = 0.50%
0 0
0 50 100 150 200 250 0 100 200 300 400
Column Reaction (kN) Column Reaction (kN)
(a) (b)
ρsup = 0.43% ρspan = 1.0%

0.7 ρsup = 0.80% ρspan = 1.0% 0.8
ρsup = 1.0% ρspan = 1.0% 0.7
0.6
ρsup = 1.2% ρspan = 1.0%
0.6
0.5 ρsup = 1.6% ρspan = 1.0%
0.5
M/V (m)
M/V (m)
0.4
0.4
0.3 Slab 1 end 1 ρsup = 0.8%
0.3
Slab 2 end 1 ρsup = 0.8%
0.2 0.2 Slab 2 end 2 ρsup = 0.8%
0.1 0.1 Slab 4 end 1 ρsup = 1.0%
Slab 4 end 2 ρsup = 0.5%
0 0
0 200 400 600 0 100 200 300 400
(c) (d)
Figure 5.30 - Influence of varying reinforcement arrangement on eccentricity (M/V) of (a) Regan
(1993) Slab 1 End 1 with span = 0.25%, (b) Regan (1993) Slab 1 End 1 with span = 0.5%, (c)
Regan (1993) Slab 1 End 1 with span = 1.0% and (d) Regan (1993) specimens (measured).
196
5.10 REGAN (1993) SLABS WITH SHELL ELEMENTS
Since shell elements are more often adopted to model slabs in practice, a NLFEA modelling
procedure is carried out using punching shear specimens tested by Regan (1993), described in
section 5.7, Table 5.2, and Figures 5.1, 5.18, and 5.19. The slab was modelled with the eight-node
quadrilateral isoparametric curved shell element CQ40S rather than CHX60 brick elements (refer to
section 3.2.1), as done in Section 5.7. Reinforcement bars were modelled as embedded elements as
previously. The Von Mises yield criterion was adopted for reinforcement with no hardening. The
integration scheme was 3 x 3 in plan with 9 points through the slab thickness as recommended by
Vollum and Tay (2007). Concrete was modelled using the ‘total strain fixed crack model’ of Diana.
Following the suggestion of Vollum and Tay (2007), the concrete tensile was taken as 0.5 , where
is the mean indirect tensile strength. Linear tension softening was adopted in which the tensile
stress reduced to zero at a strain of 0.5 , where is the reinforcement yield strain. (see Figure
4.5).
For behaviour in compression, the Thorenfeldt model (Thorenfeldt et al., 1987) was adopted along
with the four-parameter Hsieh-Ting-Chen failure surface to simulate the increase in concrete
compressive strength with increasing isotropic stress. Concrete compressive strength reduction due
to lateral cracking was modelled in accordance with Vecchio and Collins (1993). A Quasi-Newton
iterative solution procedure was adopted in conjunction with an energy based convergence criteria
with 10-3 tolerance. Figure 5.31 shows the mesh of Regan’s (1993) slab adopted for the analysis.
The mesh refinement around the column is similar to that adopted for the analysis of this specimen
in section 5.7.
Figure 5.31 – Mesh used for the validation.
197
The shear stiffness was reduced after cracking following an aggregate based shear retention factor
and = 20 was the value for mean aggregate size used. The choice was based on the
sensitivity study shown in the following.
SENSITIVITY STUDY
Figure 5.32 shows the results of a sensitivity study carried out for Slab 1 End 1 tested by Regan
(1993). The column was modelled using solid elements and a constant shear retention factor of =
0.1 as previously described in section 5.7. The sensitivity analysis studied the effect of varying the
shear retention factor used for the slab. The effect of using a constant and variable, aggregate based,
shear retention factor was investigated. Analyses were carried out with constant shear retention
factors of = 0.1 and = 0.9 as well as the aggregate based variable shear retention factor in
Diana with maximum aggregate size varying from 10 to 20 mm.
Slab 1 End 1 350

Slab 1 End 1 Experimental
20 Failure Load
300
(Experimental)
10
dagg = 20
250
Deflection (mm)
0
dagg = 15
0 1000 2000 3000 4000 5000 200
-10
dagg = 10
150
-20
0.67V β=0.1
100
-30
β=0.9
V 50
-40
Column
0
0 0.01 0.02 0.03
Slab (mm) Rotation
(a) (b)
Figure 5.32 - Deflected shape and rotations of Slab 1 End 1 (Regan, 1993) modelled with shell
elements and different shear retention factors.
Analyses with = 0.1 did not reach the measured experimental failure load V, and stopped at
around 67% of the test’s final value. Not much change in deflected shape is seen for the other shear
retention factors, though = 10 and = 15 surprisingly reached greater ultimate
loads than = 20 . In subsequent analyses with shell elements, the variable shear retention
factor model is used with = 20 as in Regan’s tests (Regan 1993).
The comparison between measured and numerical rotations for all the analysed specimens from
Regan (1993) is presented in Figure 5.33. The good agreement of both rotations and failure loads
indicates that the model procedure can be adopted for similar studies. It should be noted that
analyses with shell elements overestimate the measured failure load since punching failure is not
simulated.
198

350 350

300 300
250 250
Experimental Experimental
200 200
150 Failure Load 150 Failure Load
100 (Experimental) 100 (Experimental)
NLFEA
50 50 NLFEA
0 0
0 0.01 0.02 0.03 0 0.01 0.02 0.03
Rotation Rotation

350 350

300 300
250 250
Experimental
Experimental
200 200
150 Failure Load 150 Failure Load
100 (Experimental) 100 (Experimental)
50 NLFEA 50 NLFEA
0 0
0 0.01 0.02 0.03 0 0.01 0.02 0.03
Rotation Rotation
Slab 4 End 2
350
300
250
Experimental
200
150 Failure Load
100 (Experimental)
50 NLFEA
0
0 0.01 0.02 0.03
Rotation
Figure 5.33 - Comparison of measured and NLFEA load versus rotation for slabs of Regan (1993).
The columns of Regan (1993) specimens cracked during loading of the slab. However, for flat slabs
extracted from lower levels of regular tall buildings, it is possible for the columns to remain
uncracked due to compressive forces provided by the upper floors. Consequently, an analysis was
carried out to evaluate the impact of adopting a linear behaviour for the columns of Regan’s (1993)
slabs.
199
5.11 INFLUENCE OF UNCRACKED COLUMN ON SPECIMENS OF REGAN (1993)
Slab 1 End 1, and Slabs 2 and 4 from Regan (1993) were analysed using Diana. The columns were
modelled elastically with twenty-node isoparametric solid elements with a Poisson’s ratio = 0.2, and
elastic modulus calculated from , according to EC2. The slabs were meshed with (i) solid
elements as described in sections 5.4 and 5.7; and (ii) eight-node quadrilateral isoparametric curved
shell elements as described in section 5.10.
Figure 5.34 compares the load rotation responses obtained with NLFEA and uncracked columns to
the experimental values reported by Regan (1993). The figure also shows the load rotation for the
model with nonlinear bricks for both slab and column, depicted as “Brick Slab”, and the punching
resistance calculated according to MC2010 with = 0.7.

350 350
Resistance (ke=0.7)
300 300
250 250 Experimental
200 200 Failure Load
150 150 Brick Slab
100 100 Brick Slab - Linear

Column
50 50 Shell Slab - Linear
Column
0 0
0 0.01 0.02 0.03 0 0.01 0.02 0.03
Rotation Rotation

350
350 Resistance (ke=0.7)
300
300
250 Experimental
250
200 200 Failure Load
150 150 Brick Slab
100 100 Brick Slab - Linear
50 50 Column
Shell Slab - Linear
0 0 Column
0 0.01 0.02 0.03 0 0.01 0.02 0.03
Rotation Rotation
350 Slab 4 End 2 Resistance (ke=0.7)

300
Experimental
250
Failure Load
200
Brick Slab
150
100 Brick Slab - Linear
Column
50 Shell Slab - Linear
Column
0
0 0.01 0.02 0.03
Rotation
Figure 5.34 - Comparison of measured and FE column reaction versus rotation for slabs of Regan
(1993) with linear columns.
200
Figure 5.34 shows that the assumption for the column behaviour has little impact on the rotation
and consequently on punching resistance according to the CSCT when the slab is meshed with solid
elements. Coincidently, despite poor agreement with the experimental load rotation curve, the shell
element model, depicted as Shell Slab – Linear Column, shows remarkable agreement with the
measured failure loads of the analysed slabs when the resistance is calculated with = 0.7.
Figure 5.35 shows a comparison between deflected shapes for the NLFEA with uncraked columns
along with the “Brick Slab” model (see Figure 5.20). The figure shows that the uncracked column
assumption reduced the slab deformation, as expected. It implies that the rotations relative to the
column for brick element based models, shown in Figure 5.34, are almost independent of column
modelling.
20
Slab 1 End 1
20 Slab 2 End 1
10 10
Column
0 0
Deflection (mm)
Deflection (mm)
-10 -10 Brick Slab

-20 -20
Brick Slab -
-30 -30 Linear Column
-40 -40 Shell Slab -
-50 Linear Column
-50
0 1000 2000 3000 4000 5000 6000 0 1000 2000 3000 4000 5000 6000
Slab (mm)
Slab (mm)
Slab 2 End 2 20 Slab 4 End 1 Column

20
10
10
Deflection (mm)
0 Brick Slab
Deflection (mm)
0
-10
-10
-20 -20 Brick Slab -
Linear Column
-30 -30
-40 -40 Shell Slab -
Linear Column
-50 -50
0 1000 2000 3000 4000 5000 6000 0 1000 2000 3000 4000 5000 6000
Slab (mm) Slab (mm)
20 Slab 4 End 2
Column
10
0 Brick Slab
Deflection (mm)
-10
-20 Brick Slab -
Linear Column
-30
-40 Shell Slab -
Linear Column
-50
0 1000 2000 3000 4000 5000 6000
Slab (mm)
Figure 5.35 - Comparison of measured and FE deflected shape for slabs of Regan (1993) with linear
columns.
201
Resistance can also be calculated with according to Equation (2.37) of MC2010, which depends
on loading eccentricity. The eccentricity M/V reduce below its elastic value in the tests of Regan
(1993), as shown in Figure 5.30. Figure 5.36 shows how the eccentricity varies with load when an
uncracked column is adopted. Calculated results are shown for NLFEA in which the slab is
modelled with both brick and shell elements. The column is modelled as both elastic and nonlinear,
making four analyses in total.

0.8
Experimental 1 Experimental
0.7
0.6 Brick Slab 0.8 Brick Slab
0.5
M/V (m)
M/V (m)
Shell Slab 0.6 Shell Slab
0.4
0.3 Brick Slab - 0.4 Brick Slab -
Linear Column Linear Column
0.2 Shell Slab -
Shell Slab - 0.2
0.1 Linear Column Linear Column
0 0
0 100 200 300 400 0 100 200 300 400

0.9 Experimental
1
0.8 Experimental
0.7 0.8 Brick Slab
0.6 Brick Slab
M/V (m)
Shell Slab
M/V (m)
0.5 0.6
Shell Slab
0.4 Brick Slab -
0.4
0.3 Brick Slab -
Linear Column
Linear Column
0.2 Shell Slab - 0.2 Shell Slab -
0.1 Linear Column Linear Column
0 0
0 100 200 300 400 0 100 200 300 400
Slab 4 End 2
0.8 Experimental
0.7
0.6 Brick Slab
M/V (m)
0.5
Shell Slab
0.4
0.3 Brick Slab -
Linear Column
0.2
Shell Slab -
0.1 Linear Column
0
0 100 200 300 400
Figure 5.36 – Influence of modelling assumption in the eccentricity of Regan’s (1993) specimens.
202
Figure 5.36 shows that modelling the column as uncracked causes a significant increase in the
eccentricity M/V. Furthermore, M/V is dependent on the modelling assumptions. This is significant
since MC2010 allows resistance to be calculated with from Equation (2.37) which depends on
the eccentricity. Both brick and shell models with nonlinear column give similar and consistent
predictions for rotations, eccentricities and failure loads. The punching resistance was calculated
with MC2010 with ke from Equation (2.37) using the (i) linear eccentricity; (ii) eccentricity from
Figure 5.36 “Brick Slab”, and (iii) eccentricity from Figure 5.36 “Brick Slab – Linear Column”.
The latter two eccentricities were determined by iteration at reactions equal to the punching
resistance given by MC2010. The resulting curves are presented in Figure 5.37, which also have the
resistance calculated with = 0.7. The failure loads are summarized in Table 5.3. The NLFEA
with brick elements predicts the punching resistance to increase when the column is modelled as
elastic whereas the CSCT predicts the resistance to reduce if is calculated with Equation (2.37).
Resistance (ke=0.7)
350 Resistance ke -- Linear e
350
300 Resistance ke -- iterated e

300
NonLinear Column
250 250 Resistance ke -- iterated e
Linear Column
200 200 Experimental
150 150
Failure Load
100 100
Brick Slab
50
50
0 Brick Slab - Linear Column
0 0 0.01 0.02 0.03
0 0.01 0.02 0.03
Rotation Rotation
Resistance (ke=0.7)
Resistance ke -- Linear e
350
300
Resistance ke -- iterated e
300
250 NonLinear Column
250 Resistance ke -- iterated e
200 Linear Column
200 Experimental
150
150 Failure Load
100
100 Brick Slab
50 50
Brick Slab - Linear
0 Column
0
0 0.01 0.02 0.03 0 0.01 0.02 0.03
Rotation Rotation
203
Resistance (ke=0.7)
350
Slab 4 End 2

300
200 Linear Column
Experimental
150
Failure Load
100
Brick Slab
50
Brick Slab - Linear Column
0
0 0.01 0.02 0.03
Rotation
Figure 5.37 – Influence of column modelling assumption with Diana.
Table 5.3 – Failure loads from different column modelling assumptions and . (Loads in kN).
ke -- Iterated ke -- Iterated
ke -- Linear
NLFEA ke = 0.7 Eccentricity Eccentricity
Eccentricity
Measured Nonlinear Column Linear Column
Failure Brick Brick Brick Brick Brick
load Brick Slab Brick Slab Brick Slab Brick Slab Brick Slab
Slab Linear Slab Linear Slab Linear Slab Linear Slab Linear
Column Column Column Column Column
Slab 1 End 1 282.5 296 326 252 252 206 210 220 215 197 202
Slab 2 End 1 256.1 292 316 251 251 206 207 212 213 196 198
Slab 2 End 2 284.5 300 328 253 255 206 210 214 216 197 200
Slab 4 End 1 288.7 300 326 270 263 220 219 232 230 213 213
Slab 4 End 2 281 294 322 265 265 217 223 229 230 215 220
Edge column reaction to total load ratio seems to be fairly constant as seen in Figure 5.38, with the
model with shell elements for the slab and linear column depicted as Shell Slab – Linear Column
giving a slight increase as a result of its stiffer behaviour previously discussed.

0.6
Edge Column Reaction / Total Load
0.6
0.5 0.5
0.4 Experimental 0.4 Experimental
0.3 Brick Slab 0.3 Brick Slab
Brick Slab - Linear Column 0.2 Brick Slab - Linear Column

0.2
Shell Slab Shell Slab
0.1 0.1
Shell Slab - Linear Column Shell Slab - Linear Column
0 0
0 200 400 600 0 200 400 600
Total Load (kN) Total Load (kN)
204

0.6

0.7

0.5 0.6
0.5
0.4
Experimental 0.4
0.3 Experimental
Brick Slab 0.3 Brick Slab
0.2 Brick Slab - Linear Column 0.2 Brick Slab - Linear Column
Shell Slab 0.1 Shell Slab
0.1
Shell Slab - Linear Column Shell Slab - Linear Column
0 0
0 200 400 600 0 200 400 600
Slab 4 End 2
0.6
0.5
0.4
0.3 Experimental
Brick Slab
0.2 Brick Slab - Linear Column
0.1 Shell Slab
Shell Slab - Linear Column
0
0 200 400 600
Total Load (kN)

Figure 5.38 – Edge column reaction to total load ratio for Regan’s (1993) specimens.
5.12 STRENGTH ASSESSMENT OF REGAN SLABS
Consideration of the design methods of ACI 318, EC2 and MC2010 (refer to section 2.4) shows
that ACI 318 is unique in not relating shear resistance to the area of flexural reinforcement provided
at the edge column. This assumption is well known to be overly simplistic for isolated slab column
connections but recent research (Einpaul et al., 2015; Soares and Vollum, 2015, 2016a, 2016b) into
the influence of flexural continuity on punching resistance shows it to be more reasonable for
continuous slabs.
According to MC2010 Level IV, this is because punching resistance depends on the rotation of the
critical shear crack which is influenced by the span as well as support flexural reinforcement.
Hence, the punching resistance at internal columns is less affected by redistribution of design
moments between the span and support than implied by EC2 or MC2010 Levels II and III. Slabs S1
end 1 and S4 ends 1 and 2 of Regan (1993), which were already fully detailed in Section 5.7, Table
5.2, and Figures 5.1, 5.18, and 5.19, were modelled here using the commercially available finite
element program Atena version 5.1.2 (Červenka et al., 2016). Section 5.12.1 presents a few
205
comparisons between modelling options available in Atena. The chosen parameters for this study
are summarized in the following.
Concrete was modelled using the fracture-plastic model CC3DNonLinCementitious2 which
combines constitutive models for tensile (fracturing) and compressive (plastic) behaviour. The total
strain fixed crack model was used in which strains and stresses are converted into the material
directions given by the principal directions at the onset of cracking. A Rankine tensile failure
criterion was adopted in conjunction with an exponential softening curve. The Menétrey-Willam
failure surface was used for concrete in compression with a hardening/softening law based on the
uniaxial compressive test.
Following cracking, the shear stiffness was assumed as 20 times the stiffness of the normal crack.
The shear strength of cracked concrete was determined in accordance with the ‘Modified
Compression Field Theory’ of Vechio and Collins (1986). An arc length solution procedure was
adopted in which the stiffness was updated at each iteration, and error tolerances were adopted of
displacement (10-2), energy (10-4), residual (10-2), and absolute residual (10-2). More details on the
assumptions for this NLFEA can be found in section 3.3 of this thesis and in Červenka et al. (2016).
In order to reduce computational time, which varied between 1 and 2 days depending on the
desktop computer used, only half of the slab was modelled with eight-node brick and four-node
tetrahedral elements using the mesh shown in Figure 5.39. The ends of the columns were laterally
restrained to keep the column vertical.
Figure 5.39 - Finite element mesh of Regan (1993) slabs used in the Atena analysis.
206
5.12.1 Sensitivity Study
This section presents a few sensitivity studies carried out with Atena applied to the slabs tested by
Regan (1993). Specimen Slab 1 End 1was chosen as representative, and the outcome was adopted
for the others as well as any parametric study developed afterwards.

Figure 5.40 presents a comparison with different choices of iterative solution procedure. All the
other parameters were kept constant during the analyses. The graph shows that a choice of Arc
Length with the tangent stiffness matrix updated at each iteration and a maximum of 40 iterations
for one analysis step gave the best prediction. The Modified Newton-Raphson analysis gave fairly
reasonable results, but the Full Newton-Raphson overestimated the deformation of the slab at
failure. For both Newton-Raphson options, a maximum of 200 iterations was adopted.
Slab 1 End 1 Experimental

ρsup = 0.8% ρspan = 0.5%
Failure Load (Experimental)
350
Arc Length
300
Modified Newton Raphson
250
200 Newton Raphson
100
50
0
0 0.01 0.02 0.03
Rotation
1993).
5.12.1.2 Smeared crack method

Considering that studies such as Mamede et al. (2013) found good results while analysing flat slabs
adopting the Rotated crack method with Atena, this research considered models of Slab 1 End 1
with both rotated and fixed crack method. The latter had already been successful in evaluating the
same specimens by Regan (1993) using Diana, as shown in section 5.7.1.1. Figure 5.41 shows the
outcome, with the fixed crack approach once again presenting better results. It is worth mentioning
that the rotated crack model did not converge in a few steps, possibly due to the choice of the Arc
Length iterative solution procedure, as previously described. Perhaps an analysis with one of the
207
Newton-Raphson options would have been better for the rotated model, but the author decided to
adopt the fixed crack method along with the Arc Length procedure as the result of the good
prediction shown in Figure 5.41.
Experimental
Slab 1 End 1
ρsup = 0.8% ρspan = 0.5% Failure Load
(Experimental)
350 Fixed
300 Rotated
200
150
100
50
0
0 0.01 0.02 0.03
Rotation
Figure 5.41 - Comparison between different smeared crack approaches for Slab 1 End 1 (Regan,
1993).
5.13 ASSESSMENT OF CONTINUITY
Figure 5.42 compares the measured and predicted slab rotations relative to the column for each
modelled slab. Rotations are shown about an axis parallel to the 3000 mm slab edge as these were
greatest and hence govern punching resistance according to MC2010, as showed for the Diana
analysis in Figure 5.26. The NLFEA is seen to give good predictions of the measured rotations and
failure loads of the slabs. Figure 5.42 also shows rotations calculated for each test with MC2010
Levels II and III. The Level II rotations were calculated with ms from equation (2.48) using the
measured ultimate eccentricity M/V which is significantly less than the elastic eccentricity which
would be used in design. Figure 5.42 also shows resistances calculated with MC2010 for ke = 0.7
and ke derived from elastic finite element analysis with shell elements as the ratio of the average to
peak shear stress. Levels II and III are seen to significantly overestimate the rotation particularly for
S4 end 2 with sup = 0.5%.
The MC2010 failure loads are given by the intersection of the rotation and resistance curves in
Figure 5.42. MC2010 Level IV, with rotations from the NLFEA, is seen to give reasonable
estimates of the measured punching resistances if ke = 0.7, but resistances are underestimated if ke is
derived from elastic FEA or calculated with equation (2.37) using the measured ultimate
208
eccentricity. Figure 5.42 also shows that MC2010 Levels II and III significantly underestimate
punching resistance particularly if ke is derived from elastic FEA. EC2 is shown to give good
estimates of punching resistance.
Table 5.4 shows the ratio of the measured Vtest to calculated flexural Vflex and punching resistances
Vshear according to MC2010 Level II, EC2 and ACI 318, with v = 0, as permitted by the code, and
v from equation (2.36), as suggested by Ghali et al. (2015). The flexural resistance Vflex is the
column load at flexural failure resulting from the development of plastic hinges in the span and at
the column face. It was calculated neglecting strain hardening and the widths of the loading plates.
The moment of resistance at the column face McfR was calculated assuming that reinforcement was
only effective if positioned within a width of be = c2+2y centred on the column which can be
conservative as shown by the third column of Table 5.4. This assumption follows UK practice (The
Concrete Society, 2007) which is based on research by Regan (1981, 1999) and is also consistent
with the recommendations of Joint ACI-ASCE Committee 352 (2011).
Slab 1 End 1
300 Vtest Resistance (ke = 1/[1+e'/bu)])
(Test ULS e=M/V)
250 EC2 Vshear
200 Experimental
150 NLFEA
100 Rotation - Level II (MC2010)
50 Rotation - Level III (MC2010)
0
0 0.01 0.02 0.03
Rotation (radians)
(a)
Resistance (ke=0.7)
Slab 4 End 1
400 Resistance (ke = 1/[1+e'/bu)])
(Test ULS e=M/V)
350
EC2 Vshear
300
Vtest Experimental
250
NLFEA
200
150 Rotation - Level II (MC2010)
100 Rotation - Level III (MC2010)
50
0
0 0.01 0.02 0.03
Rotation (radians)
(b)
209
Slab 4 End 2 Resistance (ke=0.7)

350
Resistance (ke = 1/[1+e'/bu)])
(Test ULS e=M/V)
300 Vtest
250 EC2 Vshear

Experimental
200
NLFEA
150
Rotation - Level II (MC2010)
100
Rotation - Level III (MC2010)
50
0
0 0.01 0.02 0.03
Rotation (radians)
(c)
Figure 5.42 – Load-rotation relationship for Regan slabs a) S1 end 1, b) S4 end 1 and c) S4 end 2.
Table 5.4 - Comparison of measured and predicted failure loads of Regan slabs.
Flexure EC2 ACI 318 MC2010 LII
Slabs
c2+2y
v = 0 v Eq (2.36) ke = 0.7 ke Eq (2.37)
1(1) 1.12 0.81 1.04 0.91 1.21 1.38 1.49
1(2) 1.05 1.03 0.97 0.85 1.30 1.44 1.65
2(1) 1.02 0.91 0.94 0.83 1.20 1.34 1.51
2(2) 1.13 0.92 1.05 0.92 1.28 1.45 1.60
3(1) 0.50 0.40 1.46 1.25 1.08 1.33 1.17
3(2) 0.92 1.16 0.82 0.70 1.22 1.34 1.63
4(1) 1.07 0.88 0.98 0.87 1.32 1.32 1.52
4(2) 1.21 1.12 1.05 0.83 1.06 1.57 1.67
5(1) 0.76 0.36 1.25 1.60 1.62 1.24 1.12
5(2) 1.05 1.10 1.00 1.41 1.39 1.64 1.65
Average 1.06 1.02 1.28 1.41 1.50
St deviation 0.18 0.30 0.15 0.12 0.20
5%(1) 0.76 0.53 1.03 1.20 1.18
COV(2) 0.17 0.29 0.12 0.09 0.13
(1)
5% lower characteristic value
(2)
COV = coefficient of variation
The ratio Mcftest/ McfR is more critical for ACI 318 than the ratio of the moment about the centroid of
the critical section to the moment of resistance provided by reinforcement within an effective width
of c2+3h centred on the column. All the slabs are reported as having failed in punching though
flexural failure appears to have been imminent in slabs where the ratio Vtest/Vflex is greater than 1.0.
The measured ultimate eccentricity was used in the assessments with MC2010 Level II and ACI
318 with v from equation (2.36). Table 5.4 shows that EC2 gives the best predictions of the failure
210
load of the Regan slabs S1 to S5. ACI 318 with v = 0 tends to give higher shear resistances than
EC2 and can significantly overestimate resistance if not limited by McfR.
It should, however, be noted that flexural failure (Vtest/Vflex) is calculated to be critical for slabs S1,
S2 and S4 than Vtest/Vcalc. Consequently, it is possible that the measured shear resistances were
reduced by flexural failure. ACI 318 gives reasonable but conservative predictions of shear
resistance when v is calculated with equation (2.36) but the predictions are overly conservative if
the initial elastic eccentricities are used to calculate the peak shear stress. The MC2010 Level II
resistance are overly conservative due the rotation being overestimated. Better estimates of shear
resistance are obtained with ke = 0.7 than for ke calculated with equation (2.37).
5.14 ASSESSMENT OF PERPENDICULAR/TRANSVERSE FLEXURAL REINFORCEMENT
A similar analysis to that conducted in Section 5.9 was carried out with Atena, also using the
specimens tested by Regan (1993). The study investigated the influence of the longitudinal
(perpendicular to the short slab edge in Figure 5.1) and transverse hogging reinforcement.
A series of parametric studies were carried out to investigate the influence on punching resistance
of varying the flexural reinforcement ratio in slabs with the same geometry and loading
arrangement as those tested by Regan (1993). The reinforcement arrangement was similar to Slab 1
End 1, as well as the concrete strength, which was taken as 35.4 MPa. The motivation is that EC2
and MC2010 Levels II and III relate the punching resistance provided by concrete to the area of
flexural tension reinforcement at the edge column. In the case of EC2, increasing the reinforcement
ratio increases the shear resistance vc according to equation (2.25). In the case of MC2010 Levels II
and III, increasing the area of flexural tension reinforcement over the column increases mR and
hence reduces rotation for a given ms which in turn increases Vc.
Analyses were carried out with longitudinal hogging reinforcement ratios sup within an effective
width be = 2c1 + c2 at the column support equal to 0.43%, 1.0% and 1.6%. The reinforcement ratio
sup = 1.6% gives a moment of resistance of approximately 0.17 , which is the maximum
design unbalanced moment permitted by Annex I of EC2. For each of these ratios, the longitudinal
reinforcement ratio in the span span was taken as 0.25%, 0.5%, which corresponds to Regan’s slabs,
and 1.0%.
The analyses were initially carried out with the area of transverse reinforcement provided in the
Regan slabs which corresponds to a reinforcement ratio of 0.5% within a width of 1350 mm from
the slab edge. Subsequently, the analyses with span equal to 0.5% and 1.0% were repeated with the
hogging transverse reinforcement ratio doubled to 1.0%. In all cases, the NLFEA rotations relative
211
to the support area about an axis parallel to the short slab edge were greatest and hence critical. The
resulting load rotation diagrams are presented in Figure 5.43 for 0.5% transverse reinforcement as
in the Regan slabs. Figure 5.43 also shows the punching resistance calculated with MC2010 for ke =
0.7, and it confirms the outcome from Figure 5.28 that slab rotation relative to the column is
essentially governed by the span reinforcement and is almost independent of sup. The MC2010
Level IV resistances are given by the intersection of the load rotation curves with the resistance
curve. Figure 5.43 shows that the Level IV punching resistances are slightly lower but broadly
consistent with the NLFEA predictions. In the absence of test data it is unclear which of the two are
most accurate. The slabs with span = 0.25% failed in flexure. Figure 5.43 also shows rotations
calculated with MC2010 Level III for sup = 0.43%, 1.0% and 1.6% which do not correlate with the
NLFEA rotations owing to the dependence of the Level III rotations on elastic moments.
500 Resistance (ke = 0.7)
450 ρsup = 1.6%
400 ρsup = 1.0%

ρspan = 1.0% ρsup = 0.43%
350
Test
300
ρspan = 0.50%
250
200 MC2010 LIII ρsup = 0.43%
150 ρspan = 0.25%
100
50
0
0 0.005 0.01 0.015 0.02 0.025 0.03
Rotation (radians)
Figure 5.43 - Influence of varying longitudinal reinforcement arrangement on rotation of Regan

(1993) Slab 1 End 1.
The relative independence of the calculated punching resistance on sup in Figure 5.43 is significant
because it is contrary to the predictions of EC2 and MC2010 Levels II and III. As previously
discussed, these methods relate punching resistance to the area of tension reinforcement provided
within the support strips adjacent to the column with punching resistance being independent of span
reinforcement. Both Diana and Atena FE analysis suggest that increasing the span reinforcement
ratio is more effective at reducing the maximum crack strain adjacent to the column than increasing
sup above that required for flexure. This is consistent with the prediction of the CSCT that the
width of the critical shear crack is related to the rotation , which Figure 5.43 shows to reduce with
increasing span.
212
Table 5.5 shows the punching resistances of the slabs considered in the parametric study, with span
equal to 0.5% and 1.0%, according to EC2, ACI 318 and MC2010 Level II, with ke = 0.7 and from
equation (2.37), and Level III with ke = 0.7. Additionally, Table 5.5 shows the ratio of the moment
of resistance at the column face McfR to the elastic moment at the NLFEA failure load across the
slab width at the column face Mcfel. The ratios of McfR/Mcfel corresponding to sup = 0.43%
correspond to a larger degree of moment redistribution than the maximum of 70% allowed by EC2
and UK practice for edge column moments calculated with elastic FEA (The Concrete Society,
2007). All the punching resistances in Table 5.5 were calculated using the elastic eccentricity M/V
as would be done in design. Consequently, the underestimate in resistance for MC2010 Level II and
ACI 318 with v from equation (2.36) is greater in Table 5.5 than Table 5.4 where the measured
ultimate eccentricity M/V was used in calculations of resistance.
Table 5.5 - Comparison of calculated punching resistance for slabs in parametric study.
EC2 ACI 318 MC2010
tran support span VNLFEA MC2010
VNLFEA/VLIII
shear
VNLFEA/Vshear
% % % kN Level II Level III
v = 0 v Eq (2.36) ke= 0.7 ke Eq (2.37) ke= 0.7
0.5 0.43 0.5 273 0.49 1.12 0.88 1.39 2.07 2.39 1.76
0.5 1 0.5 275 1.06 0.98 0.89 1.40 1.38 1.62 1.17
0.5 1.6 0.5 269.8 1.61 0.89 0.87 1.38 1.29 1.51 0.95
0.5 0.43 1.0 352.4 0.38 1.28 1.14 1.80 2.67 3.08 2.27
0.5 1 1.0 379.2 0.77 1.20 1.23 1.94 1.91 2.24 1.61
0.5 1.6 1.0 364.6 1.19 1.07 1.18 1.86 1.74 2.05 1.28
1 0.43 0.5 273.2 0.49 1.12 0.88 1.39 2.07 2.39 1.76
1 1 0.5 277.6 1.05 0.99 0.90 1.42 1.40 1.64 1.18
1 1.6 0.5 273.6 1.59 0.90 0.89 1.40 1.13 1.34 0.96
1 0.43 1.0 379.8 0.35 1.38 1.23 1.94 2.88 3.32 2.45
1 1 1.0 391.4 0.75 1.24 1.27 2.00 1.97 2.31 1.67
1 1.6 1.0 419.6 1.03 1.23 1.36 2.14 1.73 2.05 1.47
Average 1.11 1.06 1.67 1.85 2.16 1.54
St deviation 0.16 0.19 0.30 0.53 0.60 0.48
5% 0.86 0.75 1.18 0.98 1.17 0.76
COV 0.14 0.18 0.18 0.29 0.28 0.33
Table 5.5 shows that increasing the transverse reinforcement ratio to 1.0% at the edge column
marginally increased the punching resistances given by NLFEA with the increase greatest for span
= 1.0% and sup = 1.6%. Figure 5.44 shows Load x Longitudinal Rotation responses for specimens
with transverse reinforcement ratio (tran) of 0.5% and 1.0%. Increasing the transverse
reinforcement ratio to 1.0% is seen to have little to no influence on the longitudinal rotation, which
is critical for these slabs.
213
ρsup = 1.6% ρspan = 1.0% ρsup = 1.6% ρspan = 0.5%

500 500
450 450
400 400

350 350
300 300
250 250
200 ρtran = 0.5%
200
150 ρtran = 0.5%
ρtran = 1.0% 150
100 100 ρtran = 1.0%
50 50
0 0
0 0.005 0.01 0.015 0.02 0 0.005 0.01 0.015 0.02
Rotation Rotation
500 500
450 450
400 400
350
Column Reaction (kN) 350
300 300
250 250
200 ρtran = 0.5% 200 ρtran = 0.5%
150 ρtran = 1.0% 150 ρtran = 1.0%
100 100
50 50
0 0
0 0.005 0.01 0.015 0.02 0 0.01 0.02 0.03
Rotation Rotation
500
500
450
450
400
400
350 350
300 300
250 250
200 ρtran = 0.5% 200
150 ρtran = 1.0% 150 ρtran = 0.5%
100 100 ρtran = 1.0%
50 50
0 0
0 0.005 0.01 0.015 0.02 0 0.01 0.02 0.03
Rotation Rotation
Figure 5.44 - Load x Longitudinal Rotation for specimens with transverse reinforcement ratio
varying between 0.5% and 1.0%.
Table 5.5 also shows that EC2 gives the most accurate predictions of VNLFEA. MC2010 Level II and
ACI 318 with v from equation (2.36) are overly conservative partly because the eccentricity M/V
reduces with loading as shown in Figure 5.45. The reduction in M/V occurs as a result of moment
214
being redistributed into the span following torsional cracking and yielding of flexural reinforcement
at the support. The practical consequence is that calculating the punching resistance at edge
columns on the basis of the elastic eccentricity M/V is overly conservative because it overestimates
the design shear stress.
0.7
0.6
0.5
M/V (m)
0.4
ρsup = 1.6% ρspan = 0.25%
0.3
0.2 ρsup = 1.0% ρspan = 0.25%
0.1 ρsup = 0.43% ρspan = 0.25%

0
0 50 100 150 200
(a)
ρsup = 1.6% ρspan = 0.5%
0.7 ρsup = 1.0% ρspan = 0.5%
ρsup = 0.43% ρspan = 0.5%
0.6 Slab 1 End 1
0.5
Slab 4 End 1
Slab 4 End 2
M/V (m)
0.4
0.3
0.2
0.1
0
0 50 100 150 200 250 300
(b)
0.7
ρsup = 1.6% ρspan = 1.0%
0.6 ρsup = 1.0% ρspan = 1.0%

0.5
ρsup = 0.43% ρspan = 1.0%
M/V (m)
0.4
0.3
0.2
0.1
0
0 100 200 300 400
(c)
Figure 5.45 - Influence of varying reinforcement arrangement on eccentricity a) M/V span = 0.25%,
b) M/V span = 0.5% and c) M/V span = 1.0%.
215
It is suggested that when using MC2010, the effectiveness coefficient ke is taken as 0.7 at the edge
columns of braced frames rather than being calculated with equation (2.37). ACI 318 with v = 0
tends to overestimate the shear resistance of the slabs considered in this section partly due to its
neglect of the size effect and flexural reinforcement ratio in the calculation of vc. Neither ACI 318
nor EC2 capture the influence of the span reinforcement on punching resistance. It is however
notable in Table 5.5 that the EC2 predictions of the NLFEA column failure loads VNLFEA become
less conservative as sup increases whereas the accuracy of the ACI 318 predictions with v = 0 are
almost independent of sup for any given span reinforcement ratio. The MC2010 Level II
predictions are particularly poor for these slabs because the rotation  is significantly
overestimated.
5.15 ASSESSMENT OF UNCRACKED COLUMN ON SPECIMENS OF REGAN (1993)
The adoption of an uncracked column has influenced the punching resistance obtained with Diana
using solid elements, as shown in Figure 5.34 and Table 5.3. The study was repeated with Atena,
because previous analyses carried out in this thesis have showed it to give better predictions of
punching resistance than Diana, as can be seen in Figures 5.28 and 5.43.
The Atena model was identical to that previously described in Section 5.12 and shown in Figure
5.39. The analysed slabs had the same geometry and loading arrangement as those of Regan (1993).
The flexural reinforcement ratios were sup = 0.8% and span = 0.5% (Slab 1 End 1), and sup = 1.0%
and span = 1.0%. The span reinforcement ratio of span = 1.0% was chosen to eliminate the
possibility of flexural failure. Figure 5.46 shows that changing the modelling of the column from
elastic to nonlinear had little to no impact on the calculated failure load of the evaluated specimens.
Figure 5.47 shows how the eccentricity M/V varies with edge column reaction for both slabs and
both elastic and nonlinear columns.
It is seen that at service loads, the eccentricity is significantly greater for the model with an elastic
column. The punching resistance was calculated with MC2010 with ke from Equation (2.37) using
the (i) linear eccentricity; (ii) eccentricity from Figure 5.47 “Brick Slab”, and (iii) eccentricity from
Figure 5.47 “Brick Slab – Linear Column”. The latter two eccentricities were determined by
iteration at reactions equal to the punching resistance given by MC2010, as done for the Diana
analysis (refer to Figure 5.37). The resulting curves are presented in Figure 5.48, which also have
the resistance calculated with = 0.7. The failure loads are summarized in Table 5.6.
216
350
450
(Slab 1 End 1)
300 400
350

250
300
200 250
150 Experimental 200
150
Failure Load (Experimental)
100 Brick Slab
100
Brick Slab
50
0 0
0 0.005 0.01 0.015 0.02 0 0.005 0.01 0.015
Rotation Rotation
Figure 5.46 – Influence of column modelling assumption on failure load.

(Slab 1 End 1) 0.7
0.8
0.6
0.7
0.6 0.5
M/V (m)
M/V (m)
0.5 0.4
0.4
0.3
0.3
Experimental 0.2
0.2 Brick Slab
Brick Slab
0.1
0.1 Brick Slab - Linear Column Brick Slab - Linear Column
0 0
0 100 200 300 0 100 200 300 400
Figure 5.47 – Influence of column modelling assumption on eccentricity M/V.
217
Resistance (ke=0.7)
ρsup = 0.8% ρspan = 0.5%
350 (Slab 1 End 1) Resistance ke -- Linear e
Resistance ke -- iterated e Linear
Column
250
Experimental
Brick Slab
150
100
50
0
0 0.005 0.01 0.015 0.02 0.025 0.03
Rotation
(a)
Resistance (ke=0.7)
500 ρsup = 1.0% ρspan = 1.0%
450
350 Linear Column

300 Brick Slab
200
150
100
50
0
0 0.005 0.01 0.015 0.02
Rotation
(b)
Figure 5.48 – Influence of column modelling assumption on punching resistance.
There is a significant variation in punching resistance calculated with MC2010, due to sensitivity of
from Equation (2.37) on the modelling assumptions, which is not reflected in the NLFEA
resistances which are relatively insensitive to the modelling of the column as linear or nonlinear,
bringing into question the validity of Equation (2.37).
218
Table 5.6 – Failure loads from different column modelling assumptions and . (Loads in kN).
ke -- Iterated ke -- Iterated
ke -- Linear
NLFEA ke = 0.7 Eccentricity Eccentricity
Eccentricity
Nonlinear Column Linear Column
Brick Brick Brick Brick Brick

Brick Slab Brick Slab Brick Slab Brick Slab Brick Slab
Slab Linear Slab Linear Slab Linear Slab Linear Slab Linear
Column Column Column Column Column
ρsup = 0.8%
270 278 255 262 222 232 240 253 215 225
ρspan = 0.5%
ρsup = 1.0% 367 348 319 312 257 258 304 299 257 258
ρspan = 1.0%
5.16 CONCLUSIONS
This chapter investigates the influence of unbalanced moment and flexural continuity on punching
resistance at edge columns. It supports the rotational failure criterion of MC2010 and justifies the
practice of redistributing normal moments at edge columns within normal code limits. It is shown
that the current EC2 and ACI 318 practice of neglecting the interaction between unbalanced
moment and punching resistance, as suggested by Regan (1981) and Moehle (1988), is reasonable if
the unbalanced moment is limited to 0.255(c2+y)fcdn2/c (where c = 1.5) as required by Annex I
(informative) of EC2.
ACI 318 gives reasonable predictions of the shear resistance of Regan’s specimens if v is
calculated with equation (2.36), as recommended by Ghali et al. (2015), using the measured
eccentricity but can significantly underestimate resistance if the elastic eccentricity is used as shown
in Table 5.5. However, ACI 318 tends to overestimate capacity with v = 0 as shown in Table 5.4
and 5.5 in part due to its neglect of the size effect and reinforcement ratio in the calculation of vc.
The peak shear stress on the control perimeter is shown to depend on the method of calculation and
not to be a reliable indicator of punching failure.
MC2010 gives reasonable estimates of punching resistance if rotations are calculated with NLFEA
and ke = 0.7, but significantly underestimates the punching resistance at edge columns of continuous
slabs if rotations are calculated using Levels II or III, which are intended for design, with the
underestimate greatest when ke is calculated with equation (2.37) or LFEA. This is the case when
failure is assumed to depend on the maximum rotation as in MC2010. With an average rotation
failure criterion, taking ke = 0.7 would be unconservative. The conservatism of MC2010 Levels II
and III for Regan’s continuous slabs largely arises because M/V reduces below its elastic value as
the loading is increased to failure, owing to support moments being redistributed back to the span.
Significantly, MC2010 Level IV and NLFEA predict longitudinal span reinforcement to have a
219
much greater influence on punching resistance of continuous slab-edge column sub-assemblages

than hogging reinforcement normal to the slab edge at the column. This is contrary to the
predictions of ACI 318, EC2 and MC2010 Levels II and III which assume punching resistance to be
independent of the span reinforcement ratio, and calls into question the common practice of
increasing the area of hogging reinforcement at slab edges to increase punching resistance. In UK
practice, surplus reinforcement is often provided in the span for control of deflections at the
serviceability limit state but its potential benefit on punching resistance is neglected and merits
further study.
The moment transferred between slab-edge column connections is highly influenced by the
numerical modelling assumptions. The eccentricity is shown to reduce below its elastic value for
cases where the column has a nonlinear behaviour. On the other hand, the moment transferred and
consequently the eccentricity increases above its elastic value for cases where the column is
assumed uncracked, which can happen for floors at lower levels of a building. This assumption does
not significantly affect the critical rotation of the slab relative to the column, nor the failure load
calculated by NLFEA of well calibrated continuous models. It does, however, affect the punching
shear resistance calculated with Equation (2.37) of MC2010, as it relates the resistance to
eccentricity M/V.
The benefits of providing surplus span reinforcement on punching shear resistance are not
considered to be a substitute for the provision of shear reinforcement which is commonly required
in economically proportioned flat slabs. Also, these outcomes are valid for sub-assemblages like
that tested by Regan (1993). The following chapter will consider potential benefits of span
reinforcement on the punching resistance of realistically proportioned flat slabs.
Increasing Vc is beneficial because it reduces the required area of shear reinforcement as well as
increasing VR,max if considered to be a multiple of Vc as in EC2, MC2010 and ACI 318. For practical
purposes, the current EC2 design rules appear satisfactory and superior to MC2010 Level III for
continuous slabs. However, MC2010 Level IV is a powerful tool for assessment of existing
structures and gives valuable insights into structural behaviour.
220
6 Design Considerations
This chapter assesses the influence of slab continuity on the punching resistance of a realistically
proportioned flat slab floor plate. Flexural design is carried out in accordance with EC2. The edge
column punching resistance of a symmetric flat slab extending 3 bays in each direction is assessed
by means of NLFEA, MC2010 levels II, III, IV and EC2.
6.1 GENERAL ASPECTS
In order to assess the impact of continuity, a representative flat slab was designed using the
equivalent frame method. The floor plate is 250 mm thick and consists of 9 square bays spanning
7500 mm between column centrelines. The internal columns are 400 mm square in cross section.
Edge and corner columns are 250 mm x 400 mm in cross section. Figure 6.1 shows a plan view of
the slab which, although not shown, is considered to be braced with shear walls or steel bracing.
The slab was designed for a single load case of all spans fully loaded as permitted by the UK
National annex to EC2 (BSI, 2004b) (see section 2.4.3). The resulting bending moments were
redistributed downwards by 30% at the edge columns and 20% at internal columns as required by
UK National annex to EC2.
Following downwards redistribution of the support moment, the span moments were increased as
required for equilibrium. The characteristic concrete cylinder and reinforcement yield strength were
taken as = 30 and = 500 , respectively. The design loadings were slab self-
weight = 6.25 kN/m2, superimposed dead load = 1.5 kN/m2, imposed load = 2.5 kN/m2 and a
perimeter load of 10 kN/m for external cladding. The resulting uniformly distributed design
ultimate load, calculated with load factors of 1.35 for dead load and 1.5 for imposed load, is 14.21
kN/m2. The hogging flexural reinforcement was designed for the peak bending moments at the
centreline of the columns. The slab was divided into column and middle strips (as shown in Figure
2.30) and the design bending moments were proportioned between the strips in percentage as 75:25
for hogging, and 55:45 for sagging moments as recommended in BS8110 and allowed by the EC2
(refer to Table 2.1). The additional reinforcement required by the cladding load was distributed
across the entire edge column panel width in the same proportion as the UDL.
221
Design Considerations
Figure 6.1 – Plan view of the considered flat slab (Dimensions in mm).
Surplus span rebar was provided for deflection control in accordance to EC2 span-to-effective-
depth rules, which, for instance, increased the column strip sagging reinforcement ratio between
columns A2 and B2 (refer to Figure 6.1) from 0.51% to 0.73%, or its middle strip sagging from
0.41% to 0.45%. Two thirds of the hogging reinforcement within the column strip was placed
within a band over the columns of ¼ of the panel width. The flexural reinforcement is summarized
in Figure 6.2. The reinforcement effective depths were = 212 and = 196 depicted
in Figure 6.2 as layers T1/B1 and T2/B2 respectively, and calculated assuming 30 mm cover and 16
mm bars in both directions. Where no reinforcement is shown, a minimum area of 377 mm2/m is
provided calculated in accordance with EC2 as 0.0013. . , where is the width of the tension
zone. No shear reinforcement was designed at this stage, as a comparative study focusing on this
parameter is carried out in section 6.4.2.
222
223
Figure 6.2 – Flexural reinforcement for the analysed Flat Slab.
6.2 NUMERICAL MODELLING
The numerical analysis of the flat slab designed in section 6.1 used the same modelling assumptions
as described in section 5.10. The main aim of the NLFEA was to capture the deflected shape of the
slab, allowing the punching resistance to be calculated with MC2010 LoA IV. The slab was
modelled with eight-node quadrilateral curved shell elements with a size of around 50 mm square
near the slab/column connection, and 150 mm square elsewhere. Six-node triangular curved shell
elements were used to connect the refined mesh to the coarse mesh. The columns were modelled
with elastic twenty-node brick elements measuring 150 x 50 x 50 mm. The analysis modelled the
full column height of 3.75 m above and below the slab. Columns were fully fixed at each end.
Although unrealistic, this assumption has no significant influence on the NLFEA results since the
columns were modelled elastically. Initially, one quarter of the slab was modelled as shown in
Figure 6.3 using on a total of 14773 elements.
224
Figure 6.3 – ¼ of the designed flat slab.

The model took overly long to run, so instead parametric studies were carried out on a series of
subassemblies similar to that tested by Sherif and Dilger (2000a) (see section 2.2.2) and shown in
Figure 6.4. Alternatively, the element size could have been increased in the quarter panel model to
reduce computational effort.
Figure 6.4 – Specimen tested by Sherif and Dilger (2000a).
225
6.2.1 Continuous slabs of Sherif and Dilger (2000a)
Sherif and Dilger (2000a) tested two continuous flat slabs with a novel setup in which the panel
extended to the points of zero shear. Further details of the test set up are given in Sherif (1996). The
slabs measured 5000 mm x 7500 mm on plan and were 150 mm thick. Both slabs had the flexural
reinforcement depicted in Figure 6.5, with No. 15 bars (200 mm2) with fy = 444 MPa, and No. 10
bars (100 mm2) with fy = 523 MPa. Slab S2 was reinforced with ϕ = 9.53 shear studs at both
interior and edge columns in the arrangement shown in Figure 6.6. Slab S1 had no shear
reinforcement. The load was applied to the specimen at 16 points on the full panel and 8 points on
the half panel to simulate a UDL. The columns extended 1.5 m (around half of the story height)
above and below the slab and were heavily reinforced with six No. 25 (500 mm2) deformed bars.
Table 6.1 gives more details of the specimens.
(a)
(b)
Figure 6.5 – Flexural reinforcement for the specimens tested by Sherif and Dilger (2000a).
226
Figure 6.6 – Shear reinforcement for slab S2 of Sherif and Dilger (2000a).
Table 6.1 – Details of specimens tested by Sherif and Dilger (2000a).

Column size
h d Failure Failure
(edge and Failure Mode
Slabs internal)
Load
and Location
Shear Force Remarks
(mm) (mm) % (MPa) (kN/m2) (kN)
(mm)
Punching at No
S1 250x250 150 114 1.41 28 15.54 internal 399 Shear
column Studs
Studs at
edge
Punching at
S2 250x250 150 114 1.41 33 19.97 edge column
164 and
internal
column
The Sherif and Dilger (2000a) slabs were analysed with Diana using the modelling assumptions
described in section 5.10 as for the ¼ model shown in Figure 6.3. However, the current analysis
used a smaller tolerance of 10-4 for the energy based convergence criteria, compared with 10-3
previously used, in order to get a better indication of the slab failure load. The main objective of the
NLFEA was to capture the deflected shape, which is not affected by the choice of tolerance, as
shown in Figure 6.7 for the slab ρsup=0.6% ρspan=1.2% from the parametric study in section 6.3.
Deflected shape at design load
0
-2
-4
Deflection (mm)
-6
-8
-10
-12 Column
-14
Tolerance = 0.001
-16
-18 Tolerance = 0.0001
-20
0 5000 10000
Slab (mm)
Figure 6.7 – Comparison between longitudinal deflected shape of slab ρsup=0.6% ρspan=1.2% at
design load with different convergence tolerances.
227
Around the columns, the mesh was refined using 50 mm square eight-node quadrilateral curved
shell elements compared with 100 mm square elements elsewhere. The connection between the
refined and coarse mesh was made with six-node triangular curved shell elements as described
previously for the ¼ model, shown in Figure 6.3. The final mesh is presented in Figure 6.8. The
columns were modelled with the twenty-node brick and element size of 100 x 50 x 50 mm.
Figure 6.8 – Mesh used for the NLFEA of Sherif and Dilger’s (2000a) tests.
Figure 6.9 compares deflections from test data from Sherif (1996) and NLFEA. The deflected shape
was extracted along the slab centreline in the 7500 mm direction (Longitudinal), and the
displacement was extracted at midway between the edge and internal column.
Both slabs are reported to have failed in punching shear. The NLFEA gave satisfactory predictions
for the measured ultimate resistance despite the shell elements not modelling shear failure. The
analysis of S1 stopped fairly close to the flexural capacity as indicated by the reinforcement strain
distribution of the main flexural bars over the support and span (refer to Figure 6.5) shown in Figure
6.10. It can be seen that the reinforcement has yielded despite the lack of a plastic plateau in Figure
6.9a. The model of S2 stopped fairly close to the measured failure load.
228
Deflected Shape at failure load 35

0 2 4 6 8
0 30
5 25
10
Deflection (mm)
Load (kN/m2)
20
15
15
20 Experimental
25 10 Experimental Failure Load
30 5 NLFEA
35 Yield Line
0
40 0 50 100
Longitudinal direction (m)
Deflection (mm)
(a)
25
Deflected Shape at failure load
0 2 4 6
0 20
10
15
Deflection (mm)
Load (kN/m2)
20
10
30 Experimental
5 NLFEA
40
Experimental Failure Load
50 0
0 10 20 30 40 50 60
Longitudinal direction (m) Deflection (mm)
(b)
Figure 6.9 – Comparisons between NLFEA results and experimental data from Sherif’s (1996) slab
(a) S1 and (b) S2.
Top reinforcement at Edge Column Main top reinforcement at Internal column

20
8
18 Reinforcement
Reinforcement Strain (10-3)
7 Strain at each
16
bar
6 14 Yield Strain
5 12
10
4
8
3
6
2
4
1 2
0 0
-375 -225 -75 75 225 375 -1250 0 1250
c2 + 2y (mm) Transverse direction (mm)
(a) (b)
229
Bottom reinforcement at midspan

8
7

Reinforcement Strain at each bar
6
5 Yield Strain
4
3
2
1
0
-2500 -1250 0 1250 2500
Transverse direction (mm)
(c)
Figure 6.10 –NLFEA reinforcements strains of Slab S1 at (a) face of the edge column (Hogging),
(b) face of the internal column (Hogging), and (c) midspan (Sagging).
Sherif and Dilger (2000a) reports that specimens S1 and S2 failed in punching at the internal and
edge column, respectively. Figure 6.11 shows the prediction of MC2010 LoA IV, with resistances
calculated from Equations (2.38) (VR,c) and (2.41) (VR,s) using a = 0.9 and = 0.7 for internal
and edge columns, respectively. Rotations relative to the column for both perpendicular directions
on plan depicted as ‘Longitudinal’ ( = − ) when calculated from the longitudinal
deflected shape, and ‘Transverse’ ( = ,
= ,
, since = 0 due to
symmetry), when calculated from the deflected shape along the slab edge (see Figures 6.12 and
5.25), are also presented in Figure 6.11. Previous analysis have already shown that there is no
significant difference between extracting the rotations from the shell elements or calculating them
from the corresponded deflected shape (see section 4.3.1 and Figure 4.4). Figure 6.11 also shows
rotations calculated from Equation (4.1), which was derived from NLFEA parametric studies of
internal slab column connections (refer to section 4.7).
700 350
Slab S1 Slab S2
Internal Column Reaction (kN)
600 300
Edge Column Reaction (kN)
VR,c
500 250
VR,c+VR,s
400 200
300 Experimental Failure Load 150

Experimental Failure Load
200 NLFEA Transverse Rotation 100
NLFEA Transverse Rotation
100 NLFEA Longitudinal Rotation 50
NLFEA Longitudinal Rotation
Eq. (4.1)
0 0
0 0.01 0.02 0.03 0.04 0.05 0 0.01 0.02 0.03
Rotation Rotation
Figure 6.11 – Column reaction-rotation relative to the column for the slabs of Sherif and Dilger
(2000a) and resistance according to the CSCT.
230
The CSCT is seen to give reasonable predictions for punching failure for the slabs of Sherif and
Dilger (2000a). The longitudinal direction had the largest rotations, thus critical to calculate
punching resistance according to MC2010. Rotations from Equation (4.1) agree well with that from
NLFEA.
Rotations for the CSCT
Longitudinal Rotation Transverse Rotation

Figure 6.12 – Longitudinal ( = − ) and Transverse ( = ,
=
,
) rotations used in the CSCT analysis.
6.2.2 Influence of uncracked column on specimens of Sherif and Dilger (2000a)
Sections 5.11 and 5.15 discussed the impact of assuming elastic behaviour for the column on the
specimens tested by Regan (1993). Figure 5.36 shows that the assumption of an uncracked column
caused the shear force eccentricity M/V to increase with load rather than reduce as observed
231
experimentally. The influence of an uncracked column was also assessed for the test set up of Sherif
and Dilger (2000a), which is used in the following parametric study of practically proportioned flat
slabs.
Apart from the column, which was modelled elastically, the same modelling parameters were
adopted as in section 6.2.1. The column length was increased to 3 m above and below the slab
centreline and fixed at both ends as in the ¼ slab shown in Figure 6.3. The mesh size and
refinement were the same as presented in Figure 6.8. Figure 6.13 shows a comparison between the
experimental data, and analysis with both nonlinear and linear columns.
Deflected Shape at failure load 30
0 2 4 6 8
0 25
5
20
10
Deflection (mm)
Load (kN/m2)
15 15
Experimental
20 10 Experimental Failure Load
25 Yield Line
5 NLFEA
30
NLFEA - Linear Column
35 0
40 0 20 40 60 80 100
Longitudinal direction (m) Deflection (mm)
(a)
25
Deflected Shape at failure load
0 2 4 6 8
0 20
10
Deflection (mm)
15
Load (kN/m2)
20
10 Experimental Failure Load
30
Experimental
40 5 NLFEA
NLFEA - Linear Column
50 0
Longitudinal direction (m) 0 10 20 30 40 50 60
Deflection (mm)
(b)
Figure 6.13 – Influence of uncracked column on deflections for slabs (a) S1 and (b) S2 of Sherif
and Dilger (2000a).
The assumption of an uncracked column reduced the slab deflections, as found in the equivalent
analysis of Regan’s (1993) specimens. The model with linear column is representative of the lower
levels of a tall building in which the axial column load is sufficient to prevent cracking.
Figure 6.14 shows the calculated variation in eccentricity M/V with UDL for the edge column of
Slabs S1 and S2 of Sherif and Dilger (2000a) for different assumed column behaviours. The
232
experimental M/V for slab S1 was not available since, according to Sherif (1996), the deflection of
the testing frame prevented calculating of the moment transferred between slab and column.
0.8 Slab S1 0.8
Slab S2
0.7 0.7
0.6 0.6
0.5 0.5
M/V (m)
M/V (m)
0.4 0.4
0.3 0.3 Experimental Failure Load

Experimental
0.2 0.2
NLFEA
0.1 0.1 NLFEA - Uncracked Column
Linear
0 0
0 5 10 15 20 25 30 0 5 10 15 20
UDL (kN/m2) UDL (kN/m2)
(a) (b)
Figure 6.14 – Influence of uncracked column on eccentricity at edge columns of slabs (a) S1 and (b)
S2 of Sherif and Dilger (2000a).
Figure 6.14b shows that the model with linear column overestimates the observed eccentricity with
the calculated eccentricity M/V increasing beyond its elastic value. A similar outcome was observed
in sections 5.11 and 5.15 for the specimens by Regan (1993) where the increase in eccentricity for
the elastic column is explained by its increased stiffness compared to the column in Regan’s tests.
In MC2010, the basic control perimeter is reduced by a multiple to account for loading
eccentricity. If is calculated with Equation (2.37) of MC2010, the predicted punching resistance
depends on the column flexural stiffness which depends on the column axial load. It is unclear
whether this is the case in reality. In the limit, the maximum eccentricity is limited by the maximum
moment that can be transferred to the column.
6.2.3 Subassembly of designed slab
The subassembly shown in Figure 6.15 was extracted from the ¼ slab (refer to Figure 6.3), and
includes edge column 2A and internal column 2B (refer to Figure 6.1). The model is based on the
test setup by Sherif and Dilger (2000a) previously discussed in section 6.2.1, and will be used in the
following parametric study. The same element sizes were used as in the mesh shown in Figure 6.3.
Symmetry was simulated with only rotational fixities as done for the analysis in Chapter 4 and the
model in section 5.10, to avoid any additional in-plane forces. Figures 6.16 and 6.17 show that
deflections and bending moments, extracted at the same design ULS load and same position, are
very similar in the ¼ model of the full slab and the subassembly.
233
Figure 6.15 – Subassembly of the designed flat slab.

5
-5
-10
Deflection (mm)
-15
-20 Subassembly
-25
Full 1/4
-30
0 2000 4000 6000 8000 10000
Slab (mm)
Figure 6.16 – Deflected shape comparison between the ¼ model and the Subassembly at the design
ULS load.
-1000
-800
Subassembly
Moment (kNm/m)
-600
Full 1/4
-400
-200
200
0 2000 4000 6000 8000 10000
Slab (mm)
Figure 6.17 – Bending moment comparison between the ¼ model and the Subassembly at the
design ULS load.
234
6.3 INFLUENCE OF REINFORCEMENT ARRANGEMENT
A parametric study was carried out to investigate the influence of varying the reinforcement in the
slab of the subassembly shown in Figure 6.15. Columns were modelled as elastic and extended 3.75
m above and below the slab. The ends of the columns were fully fixed. Figure 6.15 gives further
details on the slab dimensions and adopted boundary conditions. The objective was to determine
whether the influence of varying the flexural reinforcement in the subassembly was similar to that
predicted for slabs with the geometry tested by Regan (1993). Figures 6.18 and 6.19 show the
adopted reinforcement arrangement which is the same as for the ¼ slab shown in Figure 6.2. A
minimum reinforcement area of 377 mm2/m was provided where no reinforcement is shown. Loads
were applied as shown in Figure 6.15. Both UDL and cladding loads were increased proportionately
until failure. This load case is critical for the internal column, but not the edge column where
pattern loading consisting of the full factored load on the end of span and factored dead load on the
internal span is more critical. The all spans fully loaded load case was adopted since the main
objective was to compare the punching resistances given by EC2 and MC2010 Levels II to IV.
Figure 6.18 – Longitudinal reinforcement designed for the subassembly.
235
Figure 6.19 – Transverse reinforcement designed for the subassembly.

In Figure 6.18, the support reinforcement ratio, normal to the free slab edge, is ρsup = 0.8% over a
width of + (where is the perpendicular distance from the slab edge to the inner column face
and is the parallel column dimension (see Figure 2.34a)). Analyses were also carried out with
ρsup doubled to ρsup = 1.0% and halved to ρsup = 0.5%. The span reinforcement ratio was also
doubled from ρsup = 0.6% in Figure 6.18 to ρspan = 1.2%, and halved to ρspan = 0.3%. All nine slabs
were loaded with a line load along the short external edge representing cladding. An additional
analysis was carried out with reinforcement ratio of ρsup = 0.8% and ρspan = 0.6% and no cladding
load, making a total of ten analyses. Table 6.2 gives more details of the slabs.
Figure 6.20 shows the resulting longitudinal rotations relative to the column, calculated from the
longitudinal deflected shape (see Figure 6.15), and the design shear force at the external column
extracted from LFEA. The rotations were calculated as previously described in Figure 6.12.
236
Table 6.2 – Details on the parametric study varying the reinforcement arrangement.
30 MPa
1.5
500 MPa
200 GPa
1.15
a 212 mm
b 196 mm
Self-weight 6.25 kN/m2
Superimposed dead load 1.5 kN/m2
Imposed load 2.5 kN/m2
Cladding load 10 kN/m
Edge Column (2A) 250 mm x 400 mm
Internal Column (2B) 400 mm x 400 mm
0.7
0.89
a
Top and Bottom bars in the transverse direction
b
Top and Bottom bars in the longitudinal direction
Longitudinal
600 ρspan = 1.2%
500
Edge Coulmn Reaction (kN)
ρspan = 0.6%
400
ρspan = 0.3%
300
Resistance (ke=0.7)
200 ρsup = 0.5%

ρsup = 0.8%
100 ρsup = 1.0%
Design Load (ULS)
0
0 0.002 0.004 0.006 0.008
Rotation
Figure 6.20 – Longitudinal rotation of the subassembly’s edge column from the parametric study.
Figure 6.20 shows that, as for Regan’s (1993) slabs (see section 5.16), longitudinal span
reinforcement is predicted to have significantly greater influence on the longitudinal rotation at
edge columns than longitudinal hogging reinforcement at the edge column. Figure 6.20 also shows
the shear resistance provided by the concrete calculated according to MC2010 with = 0.7,
chosen on the basis of good results for previous analysis of continuous flat slabs tested by Regan
(1993) (section 5.8 and Figure 5.27) and Sherif and Dilger (2000a) (section 6.2.1 and Figure 6.11).
The influence of reinforcement ratio on punching resistance is predicted to be relatively small.
Figure 6.21 shows the variation in edge column eccentricity M/V with edge column reaction. As
previously showed for the slabs of Sherif and Dilger (2000a) (refer to section 6.2.2), the adoption of
237
elastic columns and nonlinear shell elements results in the eccentricity (M/V) increasing above its
elastic value with increasing load. The reason for the increase appears to be that bending moment is
redistributed from the span back to the column after cracking in the span. It is notable that the
normal moment about the column centreline M is virtually independent of the provided support
reinforcement prior to yield. This was unexpected but appears to be due to the contribution of the
H12 Ubars normal to the slab edge which are not included in ρsup. Similar observations were made
for Regan’s (1993) slabs when modelled with solid elements using Atena, as shown in Figure 5.45.
0.6
ρspan = 0.3%
0.5 ρspan = 0.6%
0.4
Eccentricity (m)
ρspan = 1.2%
0.3
ρsup = 0.5%
0.2
ρsup = 0.8%
ρsup =1.0%
0.1 Linear Eccentricity
Design Load (ULS)
0
0 200 400 600
Figure 6.21 – Influence of varying reinforcement arrangement in the eccentricity of the

subassembly’s edge column.
The increase in M/V with load evident in Figure 6.21 has little impact on the reactions at edge and
internal columns, which were similar to those given by LFEA, as shown in Figure 6.22. The reason
for this is that the edge column moment is relatively small compared with that at the internal
support.
Edge Column Internal Column
1200 ρspan = 1.2%
ρspan = 1.2%
600
Internal Column Reaction (kN)
1000
500
ρspan = 0.6% ρspan = 0.3% ρspan = 0.6%
800
400 ρspan = 0.3%
ρsup = 0.5% 600
300 ρsup = 0.5%
ρsup = 0.8% ρsup = 0.8%
400
200 ρsup =1.0% ρsup =1.0%
LFEA 200 LFEA
100
Design Load Design Load
0 0
0 500 1000 1500 0 500 1000 1500
Figure 6.22- Comparison between NLFEA reactions and LFEA.
238
6.4 CODE PREDICTIONS
This section is used to compare predictions of levels II, III, IV of MC2010 between themselves and
EC2, for the slabs considered in the parametric study described in section 6.3.
6.4.1 Critical direction for rotations
According to MC2010, the punching resistance depends on the greater of the slab rotations relative
to the column in the longitudinal and transverse directions (refer to Figure 6.15). Chapter 5 shows
that longitudinal rotations (i.e. rotations calculated from longitudinal deflections) are critical for the
specimens of El-Salakawy (1998) and Regan (1993) in all cases. This finding is not true in general,
and is partly a consequence of the specimen geometry and loading arrangement. Longitudinal
rotations are also critical for the tests of Sherif and Dilger (2000a), but the difference between
longitudinal and transverse rotations is much less than for Regan’s (1993) slabs (see Figure 5.26).
The subassembly shown in Figure 6.15 realistically models the direction of span in both the
longitudinal and transverse directions as for the Sherif and Dilger’s (2000a) tests.
For practical slabs, the critical direction can vary and depends on parameters including the
reinforcement arrangement and magnitude of the cladding load. Figure 6.23 compares the
longitudinal and transverse rotations with and without cladding load for slab ρsup = 0.8% ρspan =
0.6%. The reinforcement is the same for both cases and is designed for the case with cladding.
Interestingly, the critical direction for rotations in Figure 6.23 changes from longitudinal to
transverse when the cladding load is included. The resistances in Figure 6.23 are calculated with
Equation (2.38) and = 0.7. Table 6.3 summarizes the values of , calculated at rotations
corresponding to the design ULS load at the edge column, which was calculated with LFEA.
600 600
500 500
Resistance (ke=0.7)
400 400
With Cladding Load
300 300 Without Cladding Load
200 Design Load (ULS)

200
100 100
0 0
0 0.005 0.01 0.015 0.02 0 0.005 0.01 0.015 0.02
Rotation Rotation
Figure 6.23 – Comparison between longitudinal and transverse rotations with and without cladding
load for slab ρsup = 0.8% ρspan = 0.6%.
239
Figure 6.24 compares longitudinal and right-hand side transverse rotations (greater than the left-
hand side transverse rotation for all cases due to asymmetric transverse reinforcement), relative to
the column, for all 9 slabs with cladding load. Also shown is the MC2010 punching resistance
provided by the concrete calculated with Equation (2.38) for = 0.7.
Table 6.3 – , calculated with = 0.7 with and without cladding load.
,
Slabs (kN)
With Cladding Load 290 257
Without Cladding Load 240 260
ρspan = 1.2% 600 ρspan = 0.6%

600
ρspan = 1.2%

500 ρspan = 0.6%
500
400
300 Longitudinal
300
Transverse
200 200
100 100
0 0
0 0.005 0.01 0.015 0.02 0 0.005 0.01 0.015
Rotation Rotation
Resistance (ke=0.7)
600
Longitudinal
500
Transverse
ρspan = 0.3%
400
ρspan = 0.3%
300
200
100
0
0 0.0025 0.005 0.0075 0.01
Rotation
Figure 6.24 – Comparison between longitudinal and transverse rotations for the edge column of the
subassembly parametric study.
It can be seen in Figure 6.24 that the transverse direction is critical for all cases. Neither the span
nor the support longitudinal reinforcement seems to have a significant impact on these transverse
rotations, as shown in Figure 6.25. However, the longitudinal reinforcement could still affect
punching resistance due to shear redistribution around the control perimeter of the type proposed by
240
Sagaseta et al. (2011) for slabs with non-axis-symmetrical reinforcement. The study suggests that
the boundary conditions adopted in punching shear tests could have a greater influence than
considered up to now.
Transverse
600 ρspan = 1.2%
500
ρspan = 0.6%

ρspan = 0.3%
ρsup = 0.5%
300
ρsup = 0.8%
200 ρsup = 1.0%
100
0
0 0.005 0.01 0.015 0.02
Rotation
Figure 6.25 – Transverse rotations of the subassembly’s edge column from the parametric study.
6.4.2 Comparisons between MC2010 Levels II, III, IV and EC2 for edge columns
The following section compares the areas of punching shear reinforcement required by the different
levels of MC2010 and EC2 for edge column A2 (refer to Figure 6.1). The effects of uneven shear
were accounted for using the simplified = 1.4 for EC2, and = 0.7 for MC2010, as allowed for
braced frames. The effective flexural width at the edge column was calculated according to Figure
2.34a for EC2, and Figure 2.45 for MC2010. Both codes adopt partial factors of = 1.5 and =
1.15 for concrete and reinforcement, respectively. was calculated with yield line analysis as
the least of the failure loads for the mechanisms shown in Figure 6.26.
241
Figure 6.26 – Critical mechanisms considered for the subassembly’s parametric studies.
For Level III, and were calculated in each direction with a LFEA using the same elements and
mesh shown in Figure 6.15. Furthermore, also included the twisting moments in accordance
with Wood (1968). Figure 6.27 shows rotations for longitudinal and transverse directions from
NLFEA, and calculated with LoA II and III of MC2010. It also shows , from Equation (2.25) of
EC2, and the resistance curve of Equation (2.38) from MC2010 using = 0.7. Both MC2010 and
EC2 need , to calculate the required amount of shear reinforcement. MC2010 and EC2 were
used to design shear reinforcement within a zone of 1.5d around the column for the design shear
force of 420 kN. MC2010 requires a minimum area of reinforcement to satisfy , ≥ 0.5 . This
was overlooked for all the LoA IV calculations as it is intended for assessment. Figure 6.27 also
shows the maximum punching resistance, limited by crushing of the concrete struts around the
column, , , from Equation (2.43) ( = 2.8). Table 6.4 shows , for all LoA (Levels of
Approximation) of MC2010 in both the longitudinal and transverse directions. The results in bold
are the smallest, thus critical according to MC2010. Also shown in Table 6.4 are , from
Equation (2.25) of EC2, from yield line analysis, and the shear reinforcement required by each
code and LoA for MC2010.
Interestingly, the rotations in the transverse direction calculated with LoA III are slightly greater
than LoA II. This is believed to be the result of the elastic shell elements overestimating torsion at
the side of the column, which led to a greater = + , and consequently greater
rotations. Shear reinforcement requirements for LoA III almost double as a result of the transverse
direction becoming critical. This issue merits further studies as reducing torsion stiffness in a LFEA
of a shell based model potentially reduces transverse rotations, but may affect rotations in the
longitudinal direction.
242
Longitudinal Transverse VRd,c (ke=0.7)

ρsup = 1.0% ρsup = 1.0% Level II (MC2010)
600 ρspan = 1.2% 600 Level III (MC2010)
ρspan = 1.2% VRd,max

500 ρspan = 0.6% 500

ρspan = 0.6%
400 Design Load (ULS) 400 Design Load (ULS)
ρspan = 0.3% ρspan = 0.3%
300 300
EC2VRd,c EC2VRd,c
200 200
100 100
0 0
0 0.005 0.01 0.015 0.02 0 0.005 0.01 0.015 0.02
Rotation Rotation

ρsup = 0.8% ρspan = 1.2% ρsup = 0.8% Level II (MC2010)
600 600 Level III (MC2010)
Edge Column Reaction (kN) VRd,max
500 500
ρspan = 0.6%
400 Design Load (ULS) 400 Design Load (ULS)
300 300
EC2VRd,c EC2VRd,c
200 200
100 100
0 0
0 0.005 0.01 0.015 0.02 0 0.005 0.01 0.015 0.02
Rotation Rotation

ρsup = 0.5% ρsup = 0.5% Level II (MC2010)
600 ρspan = 1.2% Level III (MC2010)
600
VRd,max
ρspan = 1.2%
500 ρspan = 0.6% 500

ρspan = 0.6%
400 Design Load (ULS) 400 ρspan = 0.3% Design Load (ULS)
ρspan = 0.3%
300 EC2VRd,c 300
200 200
EC2VRd,c
100 100
0 0
0 0.005 0.01 0.015 0.02 0 0.005 0.01 0.015 0.02
Rotation Rotation
Figure 6.27 – Longitudinal and Transverse rotations of the subassembly slabs from the parametric
study.
243
Table 6.4 - , and ∑ . A from MC2010 LoA II to IV, EC2, and V . Shear forces in kN.
MC2010
ke = 0.7 EC2
Longitudinal Transverse ∑ . [mm2] Vflex

∑ .
LoA IV LoA III LoA II LoA IV LoA III LoA II LoA IV LoA III LoA II , [mm2]
ρsup = 1.0% - 417.52
- -
ρspan = 0.3%
ρsup = 1.0%
285 260 788.18b 1358.52 1364.83 267.98 1107.72 626.87
ρspan = 0.6% 274.44 142.94c 144.22 c 155.48
ρsup = 1.0%
310 262 778.33b 711.36
ρspan = 1.2%
ρsup = 0.8%
- - - 410.14
ρspan = 0.3%
c
246.43 144.22 1358.52
ρsup = 0.8%
290 257 802.96b 620
ρspan = 0.6%
c
117.80 155.48 1488.67 260.09 1132.20
aρ = 0.8%
sup b
240 254.95 260 172.55 886.70 1218.97 620
ρspan = 0.6%
ρsup = 0.8%
315 246.43 262 144.22 c 778.33b 1358.52 711.36
ρspan = 1.2%
ρsup = 0.5%
- - - 399.43
ρspan = 0.3%
ρsup = 0.5%
275 192.74 79.59 c 255 144.22 c 155.48 812.81b 1358.52 1676.90 252.21 1153.62 610.10
ρspan = 0.6%
ρsup = 0.5%
309 262 778.33b 711.36
ρspan = 1.2%
a Without Cladding Load
b
Smaller than the minimum reinforcement by MC2010 corresponding to , ≥ 0.5 (1034.48 mm2)
c
Limited by the crush of the concrete strut.
244
As can be seen in Figure 6.27 and Table 6.4, the critical direction varies with the level of
approximation. Figure 6.23 and Table 6.3 shows that the omission of cladding load in the NLFEA
swapped the critical direction from transverse to longitudinal, which led to a slightly greater amount
of shear reinforcement required according to MC2010 LoA IV (refer to Table 6.4). The transverse
reinforcement was not reduced when the cladding load was removed. The extent to which this
influenced the rotation in the transverse direction merits further investigation as does the influence
of providing surplus transverse flexural reinforcement.
Most of the LoA II and III designs appear to be limited by the crushing of concrete near the support
region according to MC2010’s limit = 2.8 for double headed studs. For LoA II, the
assumption of = 0.22 significantly overestimated the rotations for the analysis. LFEA suggests
that the point of contraflexure is actually closer to = 0.15 in the longitudinal direction and =
0.18 in the transverse, which would result in smaller rotations. The overestimation of torsion from
shell elements based models is enough to explain the poor performance of LoA III, as previously
discussed. Analysis with LoA IV indicates that concrete crushing failure is not critical in reality for
these slabs and that shear failure could be avoided through the provision of shear reinforcement as
allowed by the UK National Annex to EC2 which limits Vmax to 2VRd,c.
Though slabs with ρspan = 0.3% would likely fail in flexure as suggested by Vflex from Yield Line,
and NLFEA, which stopped before the design load could be reached, they are included in Figure
6.28 for comparative purposes. Figure 6.28 compares the amount of shear reinforcement within a
zone of 1.5d around the column in relation to the longitudinal span reinforcement.
1700
ρsup = 1% (LoA IV)
ρsup = 1% (LoA III)
ρsup = 1% (LoA II)
∑Asv in 1.5d zone around the column (mm2)
ρsup = 1% (EC2)
1500
ρsup = 0.8% (LoA IV)
ρsup = 0.8% (LoA III)
ρsup = 0.8% (LoA II)
1300
ρsup = 0.8% (EC2)
ρsup = 0.5% (LoA IV)
ρsup = 0.5% (LoA III)
1100
ρsup = 0.5% (LoA II)
ρsup = 0.5% (EC2)
900
700
0.3 0.6 0.9 1.2
ρspan (%)
Figure 6.28 – Shear reinforcement in 1.5d from the column required by MC2010 and EC2.
245
According to MC2010 LoA II, III and EC2, the required area of punching shear reinforcement is
independent of the span reinforcement ratio. However, Figure 6.28 shows that even though LoA IV
captures the influence of span reinforcement, the impact on required shear reinforcement is
minimal, especially for slabs with ρsup = 1.0%. This is so since transverse rotations are critical for
the considered slabs, which is not always the case.
The impact of longitudinal support reinforcement is significantly greater for LoA II of MC2010
than LoA III and IV or EC2. Both LoA II and III, which are intended for design, required
significantly more punching shear reinforcement than EC2, and in some cases double the
reinforcement required by LoA IV, which is intended for assessment. LoA IV requirements are very
close to those of EC2 when the minimum reinforcement rule of MC2010 is implemented.
6.5 CONCLUSIONS
This chapter investigated the impact of slab continuity, modelling assumptions, and reinforcement
arrangement on the punching resistance of realistically proportioned flat slab floor plates. Elastic
columns were considered in the punching shear assessments of the full scale slab to simplify the
analysis. The effect of this was to increase the moment transferred to the edge column but also to
reduce longitudinal slab rotations calculated with NLFEA which were generally not critical for the
considered slabs. The effect of using an elastic column was investigated for two slabs tested by
Sherif and Dilger (2000a), and the effect was found to be similar to that found for Regan’s (1993)
slabs in section 5.11.
The direction of greatest rotation varies with reinforcement arrangement and between the levels of
approximation (LoA) in MC2010. For the modelled full scale flat slabs, longitudinal span
reinforcement has a significantly greater impact on longitudinal rotations than longitudinal support
reinforcement at the slab edge as for Regan’s (1993) specimens. However, the impact on punching
resistance was significantly less than observed for Regan’s specimens, partly because transverse
rotations are frequently critical.
The average moment to be adopted in MC2010 LoA III, which is calculated from LFEA with
shell elements, requires further study since it depends on the shear modulus adopted in the analysis.
For the considered full scale slab, the area of shear reinforcement required within 1.5d of the
column face by MC2010 LoA IV is around 35% less than that by EC2. In some cases, the area of
shear reinforcement required by MC2010 LoA IV is around half that required by LoA II and III,
which are over conservative due to slab rotations being overestimated.
246
7 Conclusions
7.1 RECAPITULATION
The main objective of this research is to develop a better understanding of the influence of slab
continuity on the punching shear resistance of flat slabs. To this end, a comprehensive literature
review is made into mechanisms of punching shear resistance and the parameters that influence it
such as flexural and shear reinforcement ratios, slab depth and continuity. Some of the most
relevant studies and tests on continuous flat slabs are presented and discussed. The historical
development of design methods for punching shear is reviewed from Talbot (1913) to Broms
(2016). Most emphasis is placed on the Critical Shear Crack Theory of Muttoni (2008) which forms
the basis of the current research. The review also covers the punching provisions of the withdrawn
UK code BS8110, the current European code EC2, the current American code ACI 318, and fib
MC2010 which is intended to provide the basis for future codes.
The findings of this thesis are largely based on the results of Finite Element analysis. Chapter 3
describes the basis of the FE modelling and the constitutive models adopted for concrete and
reinforcement, the types of structural elements, and the adopted nonlinear FE solution procedures.
Chapter 4 considers the influence of slab continuity on punching resistance at internal slab column
connections. The motivation is to explain the satisfactory performance of flat slabs designed with
the superseded UK code of practice BS8110 which requires significantly less punching shear
reinforcement than EC2 and fib MC2010 Levels of Approximation (LoA) I to III which are
intended for design. Comparisons are made between the strength predictions of BS8110, EC2 and
MC2010 for selected test data. BS8110 is shown to overestimate the measured punching resistance
of conventional isolated punching shear specimens unlike EC2 and MC2010 LoA II which give
much better strength predictions. MC2010 LoA IV is used to assess the influence of slab continuity
on punching resistance. In this approach, the punching resistance is related to the maximum slab
rotation relative to the supported area which was determined with NLFEA using shell elements. The
NLFEA modelling procedure was calibrated using data from isolated internal slab–column
punching tests reported by Guandalini et al. (2009) and Lips et al. (2012). MC2010 LoA IV in
conjunction with NLFEA is shown to give reasonable punching strength predictions of the
continuous slabs of Chana and Desai (1992a) and Guralnick and Fraugh (1963). Finally, a
parametric study is carried out to investigate the influence of slab continuity on punching resistance
according to MC2010 LoA IV. It is shown that the increase in strength due to restraint from the
247
Conclusions
surrounding slab is sufficient to explain the satisfactory performance of flat slabs designed to
BS8110 at internal columns.
Chapter 5 considers punching shear failure at edge columns without shear reinforcement and
unbalanced moments about an axis parallel to the slab edge. The interaction between punching
resistance and unbalanced moment is investigated through assessment of test data and FE
modelling. FE models of the isolated test specimens of El-Salakawy et al. (1998), and continuous
tests of Regan (1993) are developed and used to carry out parametric studies in order to check the
influence of eccentricity, continuity, and reinforcement arrangement on punching resistance at flat
slab-edge column connections. Comparisons are made between the strength predictions of NLFEA
with solid elements and MC2010 LoA IV.
In Chapter 6, the influence of slab continuity on punching resistance is assessed for a realistically
designed flat slab plate. The tests by Sherif and Dilger (2000a) are used to calibrate a FE model
with shell elements that represents a portion of a full scale flat slab including edge and internal
columns. The influence on punching resistance of modelling assumptions such as the use of an
elastic column and distribution of flexural reinforcement between span and support are evaluated.
Comparisons are made between the areas of punching shear reinforcement required for the full scale
slab by MC2010 LoA II, III, IV and EC2.
7.2 CONCLUSIONS FROM THE LITERATURE REVIEW
The literature review highlights that punching resistance is influenced by a great number of factors
and there is still no widely accepted theory. Attempts to physically describe or simply represent
punching failure have been going on for more than 100 years since the introduction of the control
surface approach by Talbot (1913). Considerable disagreements remain on issues such as the
influence of slab continuity which Sherif and Dilger (2000a) concluded to have no effect, which is a
consequence of the adopted flexural reinforcement ratio over the support and span in the tests, as
addressed by Einpaul et al. (2015). Conversely, Chana and Desai (1992a) found punching resistance
to increase by around 40% when the slab was extended beyond the line of radial contraflexure.
The difference in opinion extends to the mechanisms of shear resistance and influence of key
parameters. For example, the mechanically based models of Menétrey (1996) and Hallgren (1996)
make very different assumptions about the role of concrete tensile strength in punching resistance
yet give similar and reasonable strength predictions. These disagreements extend to the rotationally
based models of Muttoni (2008) and Broms (2016) which differ in their assessment of the influence
of creep on punching resistance.
248
Conclusions
Despite issues raised by Ferreira et al. (2014) and Broms (2016) among others, the CSCT of
Muttoni (2008) is very promising. The model is gaining considerable acceptance and was chosen as
the basis of the punching shear design provisions in MC2010 (fib, 2013). This theory is extensively
used in this research where LoA IV is shown to be consistent with test data and the strength
predictions of NLFEA.
7.3 CONCLUSIONS FROM THE FINITE MODELLING
Finite Element Analysis (FEA) is a powerful numerical tool for the analysis of reinforced concrete
structures. Its application to concrete structures is complex since there are many constitutive models
available to simulate material behaviour. Due to the nonlinear behaviour of reinforced concrete
structures, the best choice of constitutive model for a certain analysis is very case-dependent. Other
parameters like the type and size of the elements, or even the iterative solution procedure, can
significantly affect the results. As a consequence, validation against similar experimental tests is a
mandatory step when using Nonlinear Finite Element Analysis (NLFEA).
This thesis uses the commercially available FE codes Diana v9.6 (TNO Diana, 2014) and Atena
v5.1.2 (Červenka et al., 2016). Both codes have already been extensively used for the NLFEA of
flat slabs presenting good results, as seen in the works of Eder et al. (2010) and Mamede et al.
(2013) among others. The objective of the majority of the FE studies carried out in this research is
to capture the deflected shape since the CSCT determines punching resistance from rotations. For
that purpose, both codes give very good results. Atena required fewer sensitivity studies and gave
better punching strength predictions than Diana on average, though that could be a reflection of the
modelling procedures used.
In Atena, the recommended CC3DNonLinCementitious2 concrete model along with an arc length
solution procedure gives good results for both deflections and ultimate resistance. The Atena
analyses used solid elements to mesh slabs, columns and loading plates. On the other hand, the
analyses carried out with Diana included different types of elements, with slabs meshed with either
solid or shell elements. The best choice of modelling parameters depends on the element type and
FE code used. In Diana, a Hordijk softening curve (Cornelissen et al., 1986; Hordijk, 1991) for
tensile behaviour and a shear retention factor of 0.1 were found ideal for cases where the slab is
meshed with solid elements. For shell element based models, a linear stress-strain relationship
following the recommendations of Vollum and Tay (2007) along with a variable aggregate based
shear retention factor gave the best results. In Diana, the Thorenfeldt (Thorenfeldt et al., 1987)
249
Conclusions
compression model and Quasi-Newton iterative solution procedure gave good predictions with both
solid and shell elements.
7.4 CONCLUSIONS FROM THE PUNCHING SHEAR FAILURE AT INTERNAL COLUMNS STUDY
Chapter 4 addresses the benefits of slab continuity on punching shear resistance at internal columns
of flat slabs. Comparisons are made between the relative safety of the design rules for punching
shear in BS8110, EC2 and MC2010. The studies show that without the code recommended partial
factors, MC2010 gives better predictions of punching resistance within the shear reinforced zone of
internal slab column connections than EC2 and BS8110. Without partial factors, BS8110
considerably overestimates the punching resistance of tested isolated internal column punching
specimens. When code recommended partial factors are included, parametric studies show that
MC2010 LoA II can require significantly more punching shear reinforcement than EC2 and
BS8110. For instance, it is shown that limiting the maximum possible shear resistance , to
, in MC2010 prevents 20% moment redistribution between support and span of flat slabs,
which was common practice in the UK for many years. MC2010 LoA II is shown to require up to 4
times more shear reinforcement in a 1.5d wide zone around the column than BS8110. Analysis with
MC2010 LoA IV of continuous flat slabs shows punching resistance to be considerably enhanced
by rotational restraint at midspan, which allows moment redistribution from support to span, and
even more so by combined rotational and in-plane restraint. The increase in punching resistance due
to slab continuity is shown to be partially included in LoA II of MC2010 if rotations are calculated
with Equation (2.45).
The nine panel continuous flat slab test of Guralnick and Fraugh (1963), without punching shear
reinforcement, was analysed. BS8110 and EC2 give good predictions of punching resistance at
internal columns as does the CSCT if rotations are calculated using Equation (4.1) which accounts
for flexural continuity. It is suggested that the punching resistance was enhanced by flexural
continuity and not compressive membrane action which is thought to have been minimal due to the
high utilisation of flexural reinforcement in the surrounding bays. MC2010 LoA IV is shown to be a
powerful tool for assessment. However, the predictions of the method are sensitive to the assumed
boundary conditions and modelling assumptions which can significantly influence the calculated
slab rotation. Consequently, the adoption of a rotationally-based failure criterion is likely to lead to
disagreements between designers and checking engineers. In this respect, design methods which
provide a unique solution, like those of BS8110 and EC2, are preferable. The level of safety
provided by BS8110 is shown to be low for isolated test specimens but more justifiable for
250
Conclusions
continuous slabs where rotations are reduced, and hence punching resistance increased by the
rotational restraint provided by surrounding panels.
7.5 CONCLUSIONS FROM THE PUNCHING SHEAR FAILURE AT EDGE COLUMNS STUDY
Chapter 5 considers the influence of unbalanced moment, flexural continuity, and reinforcement
arrangement on punching resistance of flat slab-edge column connections without shear
reinforcement. The common practice of redistributing normal moments at edge columns from
support to span within code limits is justified. Neglecting the interaction between unbalanced
moments and punching resistance as done in the current EC2 and ACI 318 is shown to be
reasonable provided the unbalanced moment is limited to 0.255( + ) / (where = 1.5)
as required by Annex I (informative) of EC2.
The proportion of unbalanced normal moment resisted by bending and eccentric shear is evaluated
at the critical shear perimeter of ACI 318. The study shows that Equation (2.36) from ACI 318
underestimates the proportion of unbalanced moment resisted by eccentric shear compared to
NLFEA predictions. ACI 318 significantly underestimates the punching resistance of the
continuous specimens tested by Regan (1993) if calculated using the elastic eccentricity M/V. Better
predictions were found when using the measured eccentricity. However, ACI 318 overestimates the
punching resistance of Regan’s specimens if the interaction between unbalanced moment and
eccentric shear is neglected ( = 0) as allowed by the code (see Tables 5.4 and 5.5). This is
believed to be a consequence of ACI 318 neglecting the influence of flexural reinforcement ratio
and size effect in the calculation of punching strength. Furthermore, it is shown that the peak shear
stress on the ACI 318 control perimeter is not a good indicator of punching failure, because the
results are strongly dependent on the method of calculation.
If maximum rotation is considered to govern the punching resistance as assumed in MC2010, LoA
IV gives good estimations when = 0.7, but LoA II and III (Equation (2.45)), which are intended
for design, are overly conservative. This conservatism worsens when is calculated with Equation
(2.37) or LFEA. This is partially due to the fact that eccentricity M/V reduces below its elastic
values for continuous flat slabs like those tested by Regan (1993) owing to support moments being
redistributed back to span.
Both MC2010 LoA IV and NLFEA predict the span reinforcement normal to the slab edge to have
a much greater impact on punching resistance than the longitudinal support reinforcement for the
continuous tests of Regan (1993). EC2 and MC2010 LoA II and III consider punching resistance to
be independent of span reinforcement, and only relate punching strength to the hogging
251
Conclusions
reinforcement over the support. Additionally, the potential benefit on punching resistance as the
result of providing surplus reinforcement over the span for deflection control, commonly adopted in
UK practice, is currently neglected by these codes.
7.6 REMARKS ON DESIGN CONSIDERATIONS
Chapter 6 considers the impact of slab continuity, longitudinal reinforcement arrangement, and
modelling assumptions on the punching resistance of realistically proportioned flat slab panels.
Some of the conclusions from the FE analysis carried out for the specimens tested by Regan (1993)
were re-evaluated for the specimens tested by Sherif and Dilger (2000a). As for the analysis of
Regan’s (1993) specimens, parametric studies show longitudinal span reinforcement to have a
significantly greater impact on longitudinal rotations than longitudinal support reinforcement at the
edge column. However, the impact on punching resistance was considerably less for the slabs of
Sherif and Dilger (2000a) than observed for Regan’s partly because transverse rotations became
critical.
For practical flat slabs, the direction of greatest rotation, which governs punching shear resistance
according to MC2010, varies with longitudinal reinforcement arrangement and within the levels of
approximation of MC2010. The average moment per unit width calculated with LFEA of shell
based models to be adopted in Equation (2.45) of MC2010 for LoA III merits further investigation
as it is highly influenced by the choice of shear modulus in the analysis.
The current EC2 provisions for punching shear at edge columns require similar areas of shear to
MC2010 LoA IV which gives good strength predictions. MC2010 LoA II and III, which are
intended for design, can be very conservative for edge columns since rotations are overestimated.
The current EC2 design provisions for punching shear have the added attraction of being easy to
use.
7.7 RECOMMENDATIONS FOR FUTURE RESEARCH
This research raises a number of issues that merit further study. In particular, some of the
conclusions based on MC2010 LoA IV require experimental confirmation as suggested below.
As described in the thesis, MC2010 relates punching resistance to the maximum slab rotation rather
than reinforcement ratio as done in EC2 (BSI, 2004a). The impact on longitudinal rotations of
longitudinal reinforcement over the span is shown to be significantly greater than support
reinforcement normal to the slab edge. This effect, which is not recognised by EC2 (BSI, 2004a),
was initially observed in the NLFEA of test specimens by Regan (1993), and subsequently for the
252
Conclusions
test specimens of Sherif and Dilger (2000a) where the influence on punching resistance was less
due to the greater transverse rotations. The predicted influence of span reinforcement on punching
resistance is observed in NLFEA with Atena and for MC2010 LoA IV but is not verified
experimentally. The undertaking of such tests would be a useful assessment of the CSCT failure
criterion. The influence of transverse reinforcement, which was not investigated in depth in this
thesis, could also be included in the study.
The calculation of shear resistance with MC2010 in the presence of eccentric loading requires the
calculation of a reduced basic control perimeter. This is done by multiplying the basic control
perimeter at 0.5d from the column face by a reduction factor . Conditional on specified
geometrical restrictions, MC2010 allows to be taken as 0.7 at the edge columns of braced
frames. Alternatively, can be calculated with equation (2.37) or derived from FEA as the ratio of
average to maximum shear force per unit length on the control perimeter. The thesis shows that
calculating in accordance with MC2010 underestimates punching resistance at edge columns of
continuous slabs. However, strength predictions with calculated are more reasonable for isolated
test specimens as shown in Chapter 5. This suggests that punching resistance is increased by
redistribution of shear stress around the control perimeter in continuous slabs as suggested by
Sagaseta et al. (2011) for non-symmetrically reinforced internal slab column connections. Further
research is required to improve the calculation of for eccentric loading.
Finally, the implementation of MC2010 LoA III requires the calculation of , which is the average
design bending moment per unit length. In design, where is calculated with LFEA using shell
elements, is found to depend significantly on the shear modulus assumed in the analysis.
Consequently, the choice of shear modulus requires further investigation since it can significantly
influence loading eccentricity and the shear stress distribution on the basic control perimeter.
253
Conclusions
254
References
ACI (American Concrete Institute) (2008) ACI Committee 318: Building code requirements for
structural concrete and commentary. American Concrete Institute, Farmington Hills, MI, USA.
Alexander, S. D. B. (2005). A design perspective on punching shear. ACI Structural Journal, V.
232, Special publication, pp. 97-108.
Anderson, B.G. (1957). Rigid frame failures. Journal of the American Concrete Institute, V.28,
No.7, pp. 625-636.
Andersson, J.L. (1963). Punching of concrete slabs with shear reinforcement. Royal Institute of
Technology, Bull.212, Stockholm, 59 pp.
Andrä, H-P. (1982). Zum Tragverhalten des Auflagerbereichs von Flachdecken. Diss., Universität
Stuttgart.
AS 3600-1994 (1994). Australian Standard: Concrete Structures. Standards Association of
Australia, Homebush, NSW 2140.
Balmer, G. G. (1949). Shearing strength of concrete under high triaxial stress-computation of
Mohr's Envelope as a curve, Denver, Colo.
Bathe, K.J. and Cimento, A.P. (1980). Some practical procedures for the solution of nonlinear finite
element equations. Computer methods in applied Mechanics and Engineering, 22, 59-85, North-
Holland publishing company.
Bazant Z.P. and Gambarova P.G. (1980). Rough crack models in reinforced concrete. ASCE-
Journal of Structural Engineering, 106(4): pp. 819-842.
Belytschko, T. and Black, T. (1999). Elastic crack growth in finite elements with minimal
remeshing. Int J Numer Methods Eng; 45(5):601–20.
Belytschko, T., Lu, Y.Y., Gu, L. (1994). Element-free Galerkin methods. International Journal for
numerical methods in engineering, 37:229-256.
Binici, B. (2005). An analytical model for stress-strain behaviour of confined concrete. Engineering
structures, 27, 1040-1051.
255
References
Birkle, G. (2004). Punching of Flat Slabs: The Influence of Slab Thickness and Stud Layout, PhD
thesis, Department of Civil Engineering, University of Calgary, Calgary, AB, Canada, Mar.
2004, 152 pp.
Birkle, G. and Dilger, W.H. (2008). Influence of Slab Thickness on Punching Shear Strength. ACI
Structural Journal, V. 105, No. 2, March-April 2008.
Bollinger, K. (1985). Load-Carrying Behaviour and Reinforcement of Axisymmetrically Loaded
Reinforced Concrete Plates. Doctoral thesis, Abteilung Bauwesen der Universität Dortmund,
Dortmund, Germany, 262 pp. (in German)
Bræstrup M. W., Nielsen M.P., Jensen B.C., Bach F. (1976). Axisymmetric punching of plain and
reinforced concrete. Report R75, Structural Research Laboratory, Technical University of
Denmark, Copenhagen, 1976.
Bræstrup M.W. (1978). Punching shear in concrete slabs. IABSE Colloquium, Copenhagen,
Introductory Report Oct, pp 115-136.
Broms, C.E. (1990). Punching of flat plates – A question of concrete properties in biaxial
compression and size effect. ACI Structural Journal, V. 87, May-June, 292-304.
Broms, C.E. (2016). Punching of flat plates – Tangential Strain Theory for Punching Failure of Flat
Slabs. ACI Structural Journal, V. 113, January-February, 95-104.
BSI (British Standards Institution) (1972) CP 110: The structural use of reinforced concrete in
buildings. BSI, London, UK.
BSI (British Standards Institution) (1985) BS 8110:1985: Structural Use of Concrete, BSI, London,
UK.
BSI (British Standards Institution) (1997) BS 8110:1997: Structural Use of Concrete. BSI, London,
UK.
BSI (British Standards Institution) (2004a) BS EN 1992-1-1:2004: Eurocode 2, design of concrete
structures – part 1-1: general rules and rules for buildings. BSI, London, UK.
BSI (British Standards Institution) (2004b) NA to BS EN 1992-1-1:2004: UK National Annex to
Eurocode 2. Design of concrete structures. General rules and rules for buildings. BSI, London,
UK.
Campana S., Fernández Ruiz M., Anastasi A., Muttoni A. (2013). Analysis of shear-transfer actions
on one-way RC members based on measured cracking pattern and failure
kinematics. Magazine of Concrete Research, UK, 19 p.
Campana, S., Fernández Ruiz, M., Muttoni, A. (2014). Shear Strength of Arch-Shaped Members
without Transverse Reinforcement. ACI Structural Journal, V. 111, No. 1-6, USA.
256
References
Cavagnis, F., Fernández Ruiz, M., Muttoni, A.(2015). Shear failures in reinforced concrete
members without transverse reinforcement: An analysis of the critical shear crack development
on the basis of test results, Engineering structures, Vol. 103, UK, pp. 157-173.
CEB (1966). Dalles, structures planes – thème II: poinçonnement, Bulleting d’Information No 57,
Paris, September.
CEB-FIP (Comité Euro-International du Béton) (1993) CEB-FIP model code 1990. Comité Euro-
International du Béton, Thomas Telford, London, UK.
Červenka, V. (1970). Inelastic finite element analysis of reinforced concrete panels. Thesis
presented to the University of Colorado, at Boulder, Colo., in partial fulfilment of the
requirements for the degree of Doctor of Philosophy.
Červenka, V., Jendele, L., Červenka, J. (2002). ATENA Program Documentation Part 1- Theory,
Červenka Consulting: Prague.
Červenka, V., Jendele, L., Červenka, J. (2016). ATENA Program Documentation Part 1- Theory,
Prague.
Chana, P.S. (1991). Punching shear in concrete slabs. The Structural Engineer, Volume 69, No.15,
August.
Chana, P.S. and Desai, S.B. (1992a). Membrane action, and design against punching shear, The
Structural Engineer, 70 (19): 339-343.
Chana, P.S. and Desai. S.B. (1992b). Design of shear reinforcement against punching. The
Structural Engineer, 70(9): 159–164.
Chen, W.F. (1982). Plasticity in Reinforced Concrete. McGraw-Hill Co., New York, NY, USA.
Choi, J-W. and Kim, J-H.J. (2012). Experimental investigations on moment redistribution and
punching shear of flat plates. ACI Structural Journal, 109(3): 329–337.
Clément, T., Ramos, A.P., Fernández Ruiz, M., Muttoni, A. (2013). Design for punching of
prestressed concrete slabs. Structural Concrete, Vol. 14, No. 2, Lausanne, Switzerland, pp. 157-
167.
Clément, T., Ramos, A.P., Fernández Ruiz, M., Muttoni, A. (2014) Influence of prestressing on the
punching strength of post-tensioned slabs. Engineering Structures, 72(1): 56–69.
Collins M.P., Mitchell D., Bentz, E.C. (2008). Shear design of concrete structures. The Structural
Engineer, Institution of Structural Engineers, London, pp. 32-39.
Cope, R. J., Rao, P. V., Clark, L. A., Norris, P. (1980). Modelling of reinforced concrete behaviour
for finite element analysis of bridge slabs. Numerical Methods for Nonlinear problems 1.
Swansea.
257
References
Cornelissen, H. A. W., Hordijk, D. A., Reinhardt, H. W. (1986). Experimental determination of

crack softening characteristics of normal weight and lightweight concrete. Heron 31, 2.
CSA Standard A23.3-14, (2014). Design of Concrete Structures. Canadian Standards Association,
Mississauga, Ontario, Canada.
De Borst, R. (1986). Non-linear analysis of frictional materials, Ph.D. Thesis, Delft University of
Technology.
Dilger, W., Birkle G., Mitchell D. (2005). Effect of Flexural Reinforcement on Punching Shear
Resistance, ACI SP, V.232, 57-74.
Eder, M.A. (2011). Inelastic Behaviour of Hybrid Steel/Concrete Column-to-Flat Slab
Assemblages. PhD thesis, Imperial College London.
Eder, M.A., Vollum, R.L., Elghazouli, A.Y. (2012). Performance of ductile RC flat slab to steel
column connections under cyclic loading. Engineering Structures, 36 239–257.
Eder, M.A., Vollum, R.L., Elghazouli, A.Y., Abdel-Fattah, T. (2010). Modelling and experimental
assessment of punching shear in flat slabs with shearheads. Engineering Structures, 32 3911–
3924
Einpaul, J. (2016). Punching strength of continuous flat slabs, Thèse No 6928, École Polytechnique
Fédérale de Lausanne.
Einpaul, J., Brantschen, F., Fernández Ruiz, M., Muttoni, A. (2016b). Performance of punching
shear reinforcement under gravity loading: Influence of type and detailing, ACI Structural
Journal, Vol. 113, No 4, Farmington Hills, USA, 2016, pp. 827-838.
Einpaul, J., Bujnak, J., Fernández Ruiz, M., Muttoni, A. (2016a). Study on Influence of Column
Size and Slab Slenderness on Punching Strength. ACI Structural Journal, V. 113, Farmington
Hills, USA, 2016, pp. 135-145.
Einpaul, J., Fernández Ruiz, M., Muttoni, A. (2015). Influence of moment redistribution and
compressive membrane action on punching strength of flat slabs. Engineering Structures, 86:
43–57.
El-Salakawy, E.F., Polak, M.A., Soliman, M.H. (1998). Reinforced concrete slab-column edge
connections with shear studs. Canadian Journal of Civil Engineering, 27(2): 338–348.
El-Salakawy, E.F., Polak, M.A., Soliman, M.H. (1999). Reinforced concrete slab-column edge
connections with shear studs. ACI Structural Journal, V.96, No.1, January-February: 79–87.
Elstner R.C. and Hognestad, E. (1956). Shearing strength of reinforced concrete slabs. ACI Journal,
V.28, No.1, July, p.29-58.
Fang, L. (2014). Shear Enhancement in Reinforced Concrete Beams, PhD thesis, Imperial College
London.
258
References
Fernández Ruiz, M. and Muttoni, A. (2009) Applications of critical shear crack theory to punching
of reinforced concrete slabs with transverse reinforcement. ACI Structural Journal, 106(4): 485–
494.
Fernández Ruiz, M., Muttoni, A., Sagaseta, J. (2015). Shear strength of concrete members without
transverse reinforcement: A mechanical approach to consistently account for size and strain
effects, Engineering structures, UK, 2015, pp. 360-372.
Fernández Ruiz, M., Muttoni, A., Kunz, J.(2010). Strengthening of flat slabs against punching shear
using post-installed shear reinforcement, ACI Structural Journal, V. 107 N° 4, USA, pp. 434-
442.
Ferreira, M.P., Melo, G.S., Regan, P.E., Vollum, R.L. (2014). Punching of reinforced concrete flat
slabs with double-headed shear reinforcement. ACI Structural Journal, 111(2): 363–374.
fib (Fédération International du Béton) (1999), Structural Concrete: textbook on Behaviour, Design
and Performance – Volume 1: Introduction – Design process – Materials. Lausanne, Switzerland.
fib (Fédération International du Béton) (2001). Punching of structural concrete slabs. Lausanne,
Switzerland, fib bulletin 1.
fib (Fédération International du Béton) (2013). fib Model Code for concrete structures 2010.
Fédération International du Béton, Lausanne, Switzerland.
Figueiras, J.A. (1983). Ultimate Load Analysis of Anisotropic and reinforced Concrete Plates and
Shells. University of Swansea: Swansea.
Gambarova, P.G. and Karakoç, C. (1983). A new approach to the analysis of the confinement role
in regularly cracking concrete elements. Proceedings of 7th International Conference on
Structural Mechanics in Reactor Technology, Chicago. 251–261.
Gayed, R.B. and Ghali, A. (2008). Unbalanced moment resistance in slab-column joints: analytical
assessment. Journal of Structural Engineering, 134(5): 859–864.
Gesund, H. and Dikshit, O.P. (1971). Yield line analysis of the punching problem at slab/column
intersections, Publication SP30, ACI, pp 177-201.
Gesund, H. and Kaushik, Y.P. (1970). Yield line analysis of punching failures in slabs. IABSE
Proceedings, Vol 30-I, pp 41-60.
Ghali, A., Gayed, R.B., Dilger, W. (2015). Design of concrete slabs for punching shear:
controversial concepts. ACI Structural Journal, 112(4): 505–514.
Gomes, R. and Regan, P. E. (1999). Punching Strength of Slabs Reinforced for Shear with Offcuts
of Rolled Steel I-Section Beams, Magazine of Concrete Research, V. 51, No. 2, 1999, pp. 121-
129.
259
References
Goodchild, C.H., Webster, R.M., Elliot, K.S. (2009). Economic Concrete Frame Elements to
Eurocode 2. The Concrete Centre, London, UK.
Graf, O. (1938). Versuche über die Widerstandsfähigkeit von allseitig aufliegender dicken Eisen
Betonplatten unter Einzellasten, Deutscher Ausschuß fur Eisenbeton, Heft 88.
Guandalini, S. and Muttoni, A. (2004). Symmetric Punching Tests on Reinforced Concrete Slabs
without Shear Reinforcement. Test report, EPFL, Lausanne, Switzerland, 78 pp. (in French).
Guandalini, S., Burdet, O.L., Muttoni, A. (2009). Punching tests of slabs with low reinforcement
ratios. ACI Structural Journal, 106(1): 87–95.
Guralnick, S.A. and Fraugh, R.W. (1963). Laboratory study of a 45-foot square flat plate structure.
Journal of the American Concrete Institute, 60(9): 1107–1185.
Hallgren, M. (1996). Punching Shear Capacity of Reinforced High Strength Concrete Slabs. PhD
thesis, Bulletin 23, Department of Structural Engineering, Royal Institute of Technology,
Stockholm, Sweden, 206 pp.
Hamadi, Y.D. and Regan, P.E. (1980). Behaviour of normal and lightweight aggregate beams with
shear cracks. The Structural Engineer, 58(4): pp. 71-79.
Hawkins, N.M. and Corley, W.G. (1974). Moment Transfer to Columns in Slabs with Shearhead
Reinforcement. American Concrete Institute, Farmington Hills, MI, USA, ACI Special
Publication SP-42, Shear in Reinforced Concrete.
Hess, U., Jensen, B.C., Bræstrup, M.W., Nielsen, M.P. (1978). Gennemlokning af jernbetonplader,
Report R90, Structural Research Laboratory, Technical University of Denmark, Copenhagen.
Hillerborg, A., Modeer, M., Petersson, P-E. (1976). Analysis of crack formation and crack growth
in concrete by means of fracture mechanics and finite elements. Cem Concr Res; 6:773–82.
Hoang, L.C. and Pop, A. (2016). Punching shear capacity of reinforced concrete slabs with headed
shear studs. Magazine of Concrete Research, 68(3), 118–126.
Hordijk, D.A. (1991). Local Approach to Fatigue of Concrete. PhD thesis, Delft University of
Technology, Delft, The Netherlands.
ISE (The Institution of Structural Engineers) (2006). Standard Method of Detailing Structural
Concrete. The Institution of Structural Engineers, London, UK.
Joint ACI-ASCE Committee 352 (2011). Guide for Design of Slab Column Connections in
Monolithic Concrete Structures (ACI 352.1R-11). American Concrete Institute, Farmington
Hills, MI, 28 pp.
Kani, G.N.J. (1964). The riddle of shear failure and its solution. ACI J Proc, 61(4):441–68. USA.
Kinnunen, S., and Nylander, H. (1960). Punching of Concrete Slabs Without Shear Reinforcement,
Transactions of the Royal Institute of Technology, no. 158, Stockholm, Sweden, 112 pp.
260
References
Kuang, J.S. and Morley, C.T. (1992), Punching shear behavior of restrained reinforced concrete
slabs. ACI Structural Journal, V. 89, No. I, January-February, pp. 13-19.
Kupfer, H. and Gerstle, K.H. (1973). Behavior of concrete under biaxial stresses. J. Engng Mech.,
ASCE 99, 853-866.
Kupfer, H., Hilsdorf, K., Rusch, H. (1969). Behavior of concrete under biaxial stresses. ACI
Journal, V66, Issue 8, 656-666.
Li, N., Maekawa, L., Okamura, H. (1989). Contact density model for stress transfer across cracks in
concrete. Journal of the Faculty of Engineering, University of Tokyo, XL: pp. 9-52.
Lips, S., Fernández Ruiz, M., Muttoni, A. (2012). Experimental investigation on punching strength
and deformation capacity of shear-reinforced slabs. ACI Structural Journal, 109(6): 889–900.
Long, A.E. (1975). A two-phase approach to the prediction of the punching strength of slabs. ACI-
Journal, V.72 (1975), No.5, 37-45.
Long, A.E. and Bond, D. (1967). Punching failure of reinforced concrete slabs. Proceedings Vol 37,
Institution of Civil Engineers, London, pp 109-135.
Long, A.E. and Rankin, G.I.B. (1987). Prediction of the punching strength of conventional slab-
column specimens. Proc. Institution Civ. Engrs, 82, 327-345.
MacGregor, J.G. and Wight, J.K. (2005). Reinforced Concrete Mechanics and Design, 4th edn.
Prentice Hall, Upper Saddle River, NJ, USA.
Maekawa, K., Pimanmas, A., Okamura, H. (2003). Nonlinear Mechanics of Reinforced Concrete,
Spon Press.
Mamede, N.F.S., Ramos, A.P., Duarte, M.V.F. (2013). Experimental and parametric 3D nonlinear
finite element analysis on punching of flat slabs with orthogonal reinforcement. Engineering
Structures, 48, 442–457.
Marti, P. (1980). Zur plastischen Berechnung von Stahlbeton, ETH Zurich, Institut für Baustatik
und Konstruktion, Bericht Nr. 104, Oct.
Marzouk, H., and Hussein, A. (1991). Experimental Investigation on the Behavior of HighStrength
Concrete Slabs. ACI Structural Journal, V. 88, No. 6, Nov.-Dec., pp. 701-713.
Maya, L.F., Fernández Ruiz, M., Muttoni, A., Foster, S.J. (2012). Punching shear strength of steel
fibre reinforced concrete slabs. Engineering Structures, 40, July 83-94.
Mihaylov, B. I., Bentz, E. C. and Collins, M. P. (2013). Two-Parameter Kinematic Theory for
Shear Behavior of Deep Beams. ACI Structural Journal, 110-3, 447-456.
Millard, S. G. and Johnson, R. P. (1985). Shear transfer in cracked reinforced concrete. Magazine of
Concrete Research, 37(130), pp. 3-15.
261
References
Mindlin, R. D. (1951). Influence of rotatory inertia and shear on flexural motions of isotropic,
elastic plates, ASME Journal of Applied Mechanics, Vol. 18 pp. 31–38.
Moe, J. (1961). Shearing Strength of Reinforced Concrete Slabs and Footings under Concentrated
Loads. Development Department Bulletin D47, Portland Cement Association (PCA), Skokie, IL,
130 pp.
Moehle, J. (1988). Strength of slab-column edge connections. ACI Structural Journal, 85(1): 89–
98.
Mortin, J.D. and Ghali, A. (1991). Connection of flat plates to edge columns. ACI Structural
Journal, 88(2): 191–198.
Muttoni, A. (2003). Schubfestigkeit und Durchstanzen von Platten ohne
Querkraftbewehrung, Beton- und Stahlbetonbau, Vol. 98, No 2, Berlin, Germany, pp. 74-84.
Muttoni, A. (2008). Punching shear strength of reinforced concrete slabs without transverse
reinforcement. ACI Structural Journal 105(4): 440–450.
Muttoni, A. and Fernández Ruiz, M. (2008), Shear strength of members without transverse
reinforcement as function of critical shear crack width, ACI Structural Journal, V. 105, No 2,
Farmington Hills, USA, pp. 163-172.
Muttoni, A. and Fernández Ruiz, M.A. (2012). The levels-of-approximation approach in MC 2010:
application to punching shear provisions, Structural Concrete, 13(1): 32-41.
Muttoni, A. and Schwartz, J. (1991). Behaviour of Beams and Punching in Slabs without Shear
Reinforcement. IABSE Colloquium, V. 62, Zurich, Switzerland, pp. 703-708.
Narasimham, N. (1971). Shear reinforcement in reinforced concrete column heads. Doctor thesis,
Concrete Structures and Technology; Imperial College of Science and Technology; London.
Ngo, D. and Scordelis, A.C. (1967). Finite element analysis of reinforced concrete beams. Journal
of the American Concrete Institute, 64:152-163
Nielsen, M.P., Bræstrup, M.W., Jensen, B.C., Bach, F. (1978). Concrete plasticity – beam shear –
punching shear – shear in joints, Danish Society for Structural Science and Engineering,
Copenhagen, Oct.
Nogueira, J. (2011). Numerical analysis of punching shear behavior of flat slabs strengthened with
steel bolts. Master thesis. New University of Lisbon, Faculdade de Ciências e Tecnologia, Monte
da Caparica.
Nölting, D. (1984). Das Durchstanzen von Platten aus Stahlbeton: Tragverhalten, Berechnung,
Bemessung. Braunschweig. Diss., TU Braunschweig, Institut f. Baustoffe, Massivau und
Brandschutz, H.62.
262
References
Ockleston, A. J. (1955). Load tests on a three-storey reinforced concrete building in Johannesburg.

The Structural Engineer, 33: 304–322.
Oliveira, D. and Melo, G. (1999). Inclined stirrup as shear reinforcement in high performance
concrete flat slabs. In: Fifth International Symposium on the Utilization of High Strength/High
Performance Concrete. Proc., Sandefjord, Norway.
Pralong, J. (1982). Poinçonnement symétrique des plancher-dalles. IBK-Bericht Nr. 131, Institut für
Baustatik und Konstruktion ETH Zürich, Juni.
Rangan, B.V. (1990). Tests on slabs in the vicinity of edge columns. ACI Structural Journal, 87(6):
623–629.
Rankin, G.I.B. and Long, A.E. (1987), Predicting the enhanced punching strength of interior slab–
column connections. Proceedings of the ICE, Part 1 82: 1165–1186.
Rashid, Y. (1968). Ultimate strength analysis of prestressed concrete pressure vessels. Nuclear
Engineering and Design, 7, 334-344.
Regan, P. E. (2009). Unpublished tests for RFA-TECH at Cambridge University.
Regan, P. E. and Bræstrup, M. W. (1985). Punching Shear in Reinforced Concrete: A State-of-the-
Art Report. Bulletin d'lnformation No. 168, Comité Euro-International du Béton, Lausanne, 232
pp.
Regan, P.E. (1981). Behaviour of Reinforced Concrete Flat Slabs. Ciria, London, UK, Ciria Report
89.
Regan, P.E. (1986) Symmetric punching of reinforced concrete slabs. Magazine of Concrete
Research, 38(136): 115–128.
Regan, P.E. (1990). Punching test of a slab with welded shear reinforcement. Structures research
group, Polytechnic of Central London.
Regan, P.E. (1993). Tests of Connections between Flat Slabs and Edge Columns. School of
Architecture and Engineering, University of Westminster, London, UK.
Regan, P.E. (1999). Ultimate Limit State Principles in fib Bulletin 2 Structural Concrete Volume 2
Basis of Design. fib, Lausanne, Switzerland.
Regan, P.E. and Samadian, F. (2001). Shear reinforcement against punching in reinforced concrete
flat slabs. The Structural Engineer, Volume 79/No 10
Regan, P.E., Walker, P.R., Zakaria, K.A.A. (1979). Test of Reinforced Concrete Flat Slabs. School
of the Environment Polytechnic of Central London, London, UK, CIRIA Project RP 220.
Reimann, H. (1963). Zur Bemessung von dunnen Plattendecken auf Stützen ohne Kopf gegen
Durchstanzen, Thesis, Otto Graf Institut, Stuttgart.
263
References
Richart, F. E., Brandtzæg, A., Brown, R. L. (1928). A study of the failure of concrete under
combined compressive stresses. University of Illinois, Bulletin; v. 26, no. 12.
Richart, F.E. (1948). Reinforced concrete walls and column footings parts 1 and 2. ACI Journ., V
20, No.2, 97-127 and No.3, 237-260.
Rodrigues, R.V., Muttoni, A., Fernández Ruiz, M. (2010). Influence of shear on rotation capacity of
reinforced concrete members without shear reinforcement. ACI Structural Journal, V. 107, No.
5, September-October, pp. 516-525
Rots, J. and Blaauwendraad, J. (1989). Crack models for concrete, discrete or smeared? Fixed,
multi-directional or rotating? HERON, 34 (1), 1989, Delft University of Technology.
Sagaseta, J. (2008). The Influence of Aggregate Fracture on the Shear Strength of Reinforced
Concrete Beams. PhD thesis, Imperial College London.
Sagaseta, J. and Vollum, R.L. (2011). Influence of aggregate fracture on shear transfer through
cracks in reinforced concrete. Magazine of Concrete Research, 63, 119-137.
Sagaseta, J., Muttoni, A., Fernández Ruiz, M., Tassinari, L. (2011). Non-axis-symmetrical punching
shear around internal columns of RC slabs without transverse reinforcement. Magazine of
Concrete Research, 63(6): 441–457
Sagaseta, J., Tassinari, L., Fernández Ruiz, M., Muttoni, A. (2014). Punching of flat slabs supported
on rectangular columns. Engineering Structures, 77, pp. 17-33.
Salim, W. and Sebastian, W.M. (2003). Punching shear failure in reinforced concrete slabs with
compressive membrane action. ACI Structural Journal, 100(4): 471–479.
Schaidt, W., Ladner, M., Rosli, A. (1970). Berechnung von Flachdecken auf Durchstanzen,
Wildegg Technische Forschungs – und Beratungs stele der Schweizerischen Zementidustrie.
Schnobrich, W.C. (1972). Analysis of hipped roof hyperbolic paraboloid structures. J Struct Div
Proc ASCE, 98 (ST7): 1575-1583.
Seible, F., Ghali, A., Dilger, W. (1980). Preassembled shear reinforcing units for flat slabs, ACI
Journal, Jan-Feb, 28-35.
Selby, R. G. and Vecchio, F. J. (1993). Three-dimensional Constitutive Relations for Reinforced
Concrete. Tech. Rep. 93-02, Univ. Toronto, dept. Civil Eng., Toronto, Canada.
Sherif, A., Emara, M.B., Hassanein, A., Abul Magd, S. (2005). Effect of the Column Dimensions
on the Punching Shear Strength of Edge Column–Slab Connections. American Concrete
Institute, Farmington Hills, MI, USA, ACI Special Publication 232, Punching Shear in
Reinforced Concrete Slabs, pp. 175–192.
Sherif, A.G. (1996). Behaviour of reinforced concrete flat slabs. Department of Civil Engineering,
Calgary, Alberta. PhD Thesis.
264
References
Sherif, A.G. and Dilger, W.H. (2000a). Tests of full-scale continuous reinforced concrete flat slabs.
ACI Structural Journal, 97(3): 455–467.
Sherif, A.G. and Dilger, W.H. (2000b). Punching failure of full-scale high strength concrete flat
slabs. Proceedings of International Workshop on Punching Shear Capacity of RC Slabs,
Stockholm, Sweden. KTH, Stockholm, Sweden, TRITA-BKN, Bulletin 57, pp. 235–243.
SIA 262:2003 (2003). Code 262 for Concrete Structures, Swiss Society of Engineers and
Architects, Zürich, Switzerland, 94 p.
Soares, L.F.S. and Vollum R.L. (2015). Comparison of punching shear requirements in BS8110,
EC2 and MC2010. Magazine of Concrete Research 67(24): 1315–1328.
Soares, L.F.S. and Vollum R.L. (2016a). Influence of continuity on punching resistance at edge
columns. Magazine of Concrete Research 68(23): 1225–1239.
Soares, L.F.S. and Vollum R.L. (2016b). Influence of flexural continuity on punching resistance at
edge columns. ACI-fib International Symposium on Punching Shear in Structural Concrete
Slabs: Honoring Neil M. Hawkins.
Stamenković, A. and Chapman, J.C. (1974). Local strength at column heads at flat slabs subjected
to a combined vertical and horizontal loading. Proceedings of the Institution of Civil Engineers.
Part 2 Research and Theory 57(2): 205–232.
Suidan, M. and Schnobrich, W.C. (1973). Finite element analysis of reinforced concrete. Journal of
Structural Division ASCE, 99(10): pp. 2109-2122.
Talbot, A. N. (1913). Reinforced Concrete Wall Footings and Column Footings. Engineering
Experiment Station, University of Illinois, Bulletin 67.
The Concrete Society (2007). Guide to the Design and Construction of Reinforced Concrete Flat
Slabs. The Concrete Society, Camberley, UK, Technical Report 64.
Thorenfeldt, E., Tomaszewicz, A., Jensen, J.J. (1987). Mechanical properties of high-strength
concrete and applications in design. Proceedings of Symposium on the Utilization of High-
Strength Concrete, Stavanger, Norway. Tapir, Trondheim, Norway.
Timoshenko, S. and Woinowsky-Krieger, S. (1959). Theory of Plates and Shells, International
Student Edition, McGraw-Hill, New York, 580 pp.
TNO Diana (2014) Finite-element Analysis User’s Manual. Release 9.6. TNO Diana BV, Delft, The
Netherlands.
Tolf, P. (1988). Influence of the Slab Thickness on the Punching Strength of Concrete Slabs—Tests
with Circular Slabs, Bulletin No. 146, Department of Structural Mechanics and Engineering,
Royal Institute of Technology, Stockholm, Sweden, 64 pp. (in Swedish with a summary in
English).
265
References
Valliappan, S. and Doolan, T. F. (1971). Nonlinear Stress Analysis of Reinforced Concrete

Structures by Finite Element Method, UNICIV Report No. R-72, University of New South
Wales, Australia, September.
Vanderbilt, M. D. (1972). Shear Strength of Continuous Plates. Journal of the Structural Division,
ASCE, V. 98, pp. 961-973.
Vecchio, F. (2000). Disturbed stress field model for reinforced concrete: Formulation. Journal of
Structural Engineering, 126, 1070-1077.
Vecchio, F. J. and Collins, M. P. (1986). The Modified Compression-Field Theory for Reinforced
Concrete Elements Subjected to Shear. ACI Journal, Proceedings V. 83, No. 2, Mar.-Apr. pp.
219-231.
Vecchio, F.J. and Collins, M.P. (1993). Compression response of cracked reinforced concrete.
Journal of Structural Engineering, ASCE 119(12): 3590–3610.
Vollum, R. L. (2013). Comparison of punching shear requirements in BS8110, EC2 and MC2010,
Tel Aviv, fib Symposium Tel Aviv.
Vollum, R.L. (2009). Comparison of deflection calculations and span-to-depth ratios in BS8110 and
EC2. Magazine of Concrete Research, 61(6): 465–476.
Vollum, R.L. and Tay, U.L. (2007). Modelling tension stiffening in reinforced concrete with
NLFEA. Concrete, 41(1): 40–41.
Vollum, R.L., Abdel-Fattah, T., Eder M., Elghazouli, A. Y. (2010). Design of ACI-type punching
shear reinforcement to Eurocode 2, Magazine of Concrete Research, 62, No. 1
Walraven J.C. (1980). Aggregate Interlock: a Theoretical and Experimental Analysis. PhD Thesis,
1980, Delft University of Technology.
Walraven, J. C. (1981). Fundamental Analysis of Aggregate Interlock. Journal of Structural
Engineering, ASCE, V. 107, No. 11, pp. 2245-2270.
Walraven, J. C. and Reinhardt, H. W. (1981). Theory and Experiments on the Mechanical
Behaviour of Cracks in Plain and Reinforced Concrete Subjected to Shear Loading. Delft
University of Technology.
Wesche, K. (1974). Baustoffe für tragende Bauteile. Band 2. Nichtmetallisch anorganische Stoffe,
Beton, Mauerwerk, Bauverlag GmbH, Wiesbaden and Berlin.
Westergaard, H. M. and Slater, W. A. (1921). Moments and stresses in slabs. American Concrete
Institute Journal, 17: 415-538.
Wood, R.H. (1968). The reinforcement of slabs in accordance with a pre-determined field of
moments. Concrete 2(2): 69–76.
266
References
Yitzhaki, D. (1966). Punching strength of reinforced concrete slabs. ACI – Journal, V.63, No.5,
527-542.
Zaghlool, E.R.F. (1971). Strength and Behavior of Corner and Edge Column–Slab Connections in
Reinforced Concrete Flat Plates. PhD thesis, Department of Civil Engineering, University of
Calgary, Calgary, Canada.
Zararis, P. D. (1997). Aggregate interlock and steel shear forces in the analysis of RC membrane
elements. ACI Structural Journal, 94.
267

Tese Soares

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Tese Soares

Uploaded by

Copyright:

Available Formats

Influence of slab continuity on punching resistance

A thesis submitted to Imperial College London for the degree of

Luis Fernando Sampaio Soares

Department of Civil and Environmental Engineering

Luis Fernando Soares

3.2.1 Structural Elements .............................................................................................................................. 97

5.10 Regan (1993) slabs with Shell elements ...................................................................................................... 197

Figure 1.1 - Isolated test specimen. ............................................................................................... 38

Figure 3.31 – Bars in Solid elements. .......................................................................................... 116

Figure 5.15 - Influence of eccentricity on calculated punching resistance of slabs of El-Salakawy et

A area of flexural reinforcement provided

A area of flexural reinforcement required

A area of tensile reinforcement

A area of shear reinforcement in each perimeter

b , control perimeter beyond the outer layer of shear reinforcement

b punching control perimeter

dimensions of critical section b0 measured perpendicular and parallel to slab

b effective width of flexural reinforcement for normal moment at edge column

b width of support strip

c and c column side perpendicular and parallel to the slab edge

d mean aggregate size

d shear reinforcement diameter

d maximum aggregate size

d reference size for aggregate

d effective depth of top reinforcement normal to the slab edge

d reduced effective depth of slab

E concrete modulus of elasticity

EI plate flexural stiffness before cracking

EI plate tangential flexural stiffness after cracking

E modulus of elasticity of reinforcement

f compressive cylinder strength of concrete

f maximum principal stress

f characteristic compressive cylinder strength of concrete at 28 days

f tensile strength of concrete

f characteristic cube strength of concrete

f peak compressive force

f yield strength of reinforcement

f design yield strength of reinforcement

f characteristic yield strength of reinforcement

f yield strength of the shear reinforcement

f , effective design strength of the shear reinforcement

f design yield strength of the shear reinforcement

G compressive fracture energy

G tensile fracture energy

g out-of-balance force vector

J polar moment of inertia of critical section

k effectiveness coefficient dependent on maximum aggregate size

k effectiveness coefficient for eccentric shear

k efficiency factor for punching shear reinforcement system

k coefficient relating shear resistance to slab rotation

m tangential moment at the slab edge

M bending moment across panel width at inner column face

M bending moment about centroid of critical section

m average bending moment per unit width of the support strip

M design unbalanced moment

M bending moment portion of unbalanced moment

M moment of resistance of slab column connection at column face

m bending moment (i = x, y, z) and (j = x, y, z)

m radial moment per unit width

m nominal moment capacity per unit width

m average bending moment per unit width

M moment at the support

M torsion portion of unbalanced moment